From mkesper at fsfe.org  Mon Sep  1 14:08:12 2008
From: mkesper at fsfe.org (Michael Kesper)
Date: Mon, 1 Sep 2008 14:08:12 +0200
Subject: [python-win32] Accessing properties of share on DFS
In-Reply-To: <48B55829.5020205@timgolden.me.uk>
References: <20080825143047.GD13124@localhost>
	<48B55829.5020205@timgolden.me.uk>
Message-ID: <20080901120812.GB4376@localhost>

Hi,

* Tim Golden <mail at timgolden.me.uk> [2008-08-27 14:35:37 +0100]:
?
> You might be able to get there with WMI. The following
> works on my own machine, but I'm not sure what machine
> you'd need to access to list the DFS shares. I'll see
> if I can persuade a friendly admin here to let me try
> things out:

Thanks for your answer. I try to try it. ;)
I can access the DFS share properties via mmc from one special server,
so I think I'll need to run my scripts there.
As you probably guessed, there's no python installed there.

> or you could come from the other end and enumerate
> the Win32_ServerConnection collection to start with.

Sorry, I do not exactly know how to do that.

Best wishes
Michael
-- 
Free Software Foundation Europe (FSFE) []         (http://fsfeurope.org)
Treten Sie der Fellowship bei!       [][][]       (http://fsfe.org/join)
Ihre Spende erm?glicht unsere Arbeit!  ||  (http://fsfeurope.org/donate)

From mail at timgolden.me.uk  Mon Sep  1 18:13:20 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Mon, 01 Sep 2008 17:13:20 +0100
Subject: [python-win32] Accessing properties of share on DFS
In-Reply-To: <20080901120812.GB4376@localhost>
References: <20080825143047.GD13124@localhost>	<48B55829.5020205@timgolden.me.uk>
	<20080901120812.GB4376@localhost>
Message-ID: <48BC14A0.3050202@timgolden.me.uk>

Michael Kesper wrote:
> Hi,
> 
> * Tim Golden <mail at timgolden.me.uk> [2008-08-27 14:35:37 +0100]:
>  
>> You might be able to get there with WMI. The following
>> works on my own machine, but I'm not sure what machine
>> you'd need to access to list the DFS shares. I'll see
>> if I can persuade a friendly admin here to let me try
>> things out:
> 
> Thanks for your answer. I try to try it. ;)
> I can access the DFS share properties via mmc from one special server,
> so I think I'll need to run my scripts there.
> As you probably guessed, there's no python installed there.

That's not problem. The great thing with WMI is
that it's installed *everywhere* and by the glory
of DCOM (!) using it is a breeze. Just take that code
snippet and change the line:

  c = wmi.WMI ()

to:

  c = wmi.WMI ("other-machine-with-dfs")

and Bob's your Uncle.

TJG

From mdriscoll at co.marshall.ia.us  Tue Sep  2 15:45:32 2008
From: mdriscoll at co.marshall.ia.us (Mike Driscoll)
Date: Tue, 02 Sep 2008 08:45:32 -0500
Subject: [python-win32] Pb IE8-b2 + Python like ActiveScripting language
In-Reply-To: <0CF37A7FCD3A4621BB844DF4008C17F9@MCI1330>
References: <0CF37A7FCD3A4621BB844DF4008C17F9@MCI1330>
Message-ID: <48BD437C.1020803@co.marshall.ia.us>

Michel Claveau wrote:
> <div class="moz-text-flowed" style="font-family: -moz-fixed">Hi!
>
> Impossible, for me, to use Python like ActiveScripting language, in a 
> HTA, since IE-8 beta-2 installed.
>
> Run OK with IE5, IE6, IE7. But IE8 does not recognize Python  (I 
> however recorded several times python like language of ActiveScripting)
>
> Any idea or solution?

I'll bee honest and say that I've never heard of an HTA. Upon doing a 
little research though, it looks like it's a self-contained web app of 
sorts. You might look at CherryPy. We use it for a web app on a couple 
of laptops and it works great. Of course, there's also IronPython + 
Silverlight. Both of these "solutions" are probably more complicated 
though.

Mike

From alecmunro at gmail.com  Tue Sep  2 18:44:02 2008
From: alecmunro at gmail.com (Alec Munro)
Date: Tue, 2 Sep 2008 13:44:02 -0300
Subject: [python-win32] Accessing other interfaces of a Dispatch created
	COM object
In-Reply-To: <9819d58b0808291047j3de5e9d8u7647b05e49a53f97@mail.gmail.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>
	<9819d58b0808280632x2b139645x278d5f0675dd46c7@mail.gmail.com>
	<9819d58b0808280633t13d89b77u433d55924cbb73a1@mail.gmail.com>
	<9819d58b0808280734o33af5721y741ca9b079d074af@mail.gmail.com>
	<00de01c9099c$b77731c0$26659540$@com.au>
	<9819d58b0808291047j3de5e9d8u7647b05e49a53f97@mail.gmail.com>
Message-ID: <9819d58b0809020944v488ab4a2n5127c71e7c69424b@mail.gmail.com>

I'm guessing, based on the lack of responses, and the consistency of
error messages, that my problem has more to do with getting the right
library registered, which isn't a python-win job.

So, can anyone recommend where I might go to try to solve this
problem? Also, is there a way to determine which library it is that is
not registered?

Thanks again,

Alec

On Fri, Aug 29, 2008 at 2:47 PM, Alec Munro <alecmunro at gmail.com> wrote:
> Thanks Mark.
>
> After a couple of misfires, I think I figured out how to get CastTo to
> work (the docstring is practically non-existent in the version I have,
> perhaps it's been updated?).
>
> However, that puts me back at the same place I was with some of the
> other attempts:
>
>>>> obj_i2 = CastTo(obj, "Interface_2")
>>>> obj_i2.secondaryMethod()
> Traceback (most recent call last):
> ...
>  File "C:\Python25\lib\site-packages\win32com\client\__init__.py",
> line 449, in _ApplyTypes_
>   dispid, 0, wFlags, retType, argTypes, *args),
> com_error: (-2147319779, 'Library not registered.', None, None)
>
> However, one interesting note is that the object generated by CastTo
> is of a different version of the Type Library than the object
> generated by Interface_2(obj). I don't know if that's useful
> information, however. They have the same CLSID, but different
> coclass_clsids.
>
> Thanks again,
>
> Alec
>
> On Fri, Aug 29, 2008 at 3:01 AM, Mark Hammond <mhammond at skippinet.com.au> wrote:
>> Look at the docstring for win32com.client.CastTo (and googling for that will
>> probably help too)
>>
>> Cheers,
>>
>> Mark
>>
>>> I just discovered a third possibility, and subsequent failure:
>>>
>>> >>> obj_i2 = Dispatch(obj, None, Interface_2.CLSID)
>>> >>> obj_i2.secondaryMethod()
>>> Traceback (most recent call last):
>>> ...
>>>   File "C:\Python25\lib\site-packages\win32com\client\__init__.py",
>>> line 449, in _ApplyTypes_
>>>     dispid, 0, wFlags, retType, argTypes, *args),
>>> com_error: (-2147352573, 'Member not found.', None, None)
>>>
>>> Alec
>>>
>>> On Thu, Aug 28, 2008 at 10:33 AM, Alec Munro <alecmunro at gmail.com>
>>> wrote:
>>> > Oh, and on re-reading, I do realize that the DLL itself doesn't
>>> > implement those interfaces. There's a class within the DLL that does
>>> > that. :)
>>> >
>>> > On Thu, Aug 28, 2008 at 10:32 AM, Alec Munro <alecmunro at gmail.com>
>>> wrote:
>>> >> Hi List,
>>> >>
>>> >> Hopefully I have my terminology right, I'm pretty new to COM and
>>> C++.
>>> >>
>>> >> My problem is that I have a COM DLL written in C++, that I would
>>> like
>>> >> to call from Python. This DLL implements 3 important interfaces,
>>> each
>>> >> of which is also an IDispatch implementation, like the following:
>>> >>
>>> >> IDispatchImpl<Interface_1, &__uuidof(Interface_1),
>>> &LIBID_Interface_1, 1, 0>
>>> >> IDispatchImpl<Interface_2, &__uuidof(Interface_2),
>>> &LIBID_Interface_2, 1, 0>
>>> >> IDispatchImpl<Interface_3, &__uuidof(Interface_3),
>>> &LIBID_Interface_3 1, 0>
>>> >>
>>> >>  However, when I do:
>>> >>>>> obj = win32com.client.Dispatch("a.b.c")
>>> >>>>> obj
>>> >> <COMObject a.b.c>
>>> >>
>>> >> The object I get only has methods available for the first of the
>>> >> interfaces. From my reading, this is simply the functionality of
>>> >> IDispatch, and if I wanted to get those additional methods through
>>> >> IDispatch, I would have to update the DLL, and restructure my
>>> >> interfaces. I'm not confident enough in my C++ to do this, or at
>>> least
>>> >> it's not my first choice.
>>> >>
>>> >> From what I've read, there are two ways that seem to be
>>> possibilities
>>> >> for being able to call the methods from the other interfaces.
>>> >>
>>> >> The simplest one I found was the suggestion to use MakePy on the
>>> type
>>> >> library containing Interface_2 and Interface_3. I did this, and
>>> copied
>>> >> the generated file to my working directory, then did:
>>> >>
>>> >>>>> from interfaces import Interface_2
>>> >>>>> obj_i2 = Interface_2(obj)
>>> >>>>> obj_i2
>>> >> <win32com.gen_py.typelibname 1.0 Type Library.Interface_2 instance
>>> at
>>> >> 0x17459602>
>>> >>
>>> >> Unfortunately when I try:
>>> >>
>>> >>>>> obj_i2.secondaryMethod()
>>> >> Traceback (most recent call last):
>>> >> ...
>>> >>  File "C:\Python25\lib\site-packages\win32com\client\__init__.py",
>>> >> line 449, in _ApplyTypes_
>>> >>   dispid, 0, wFlags, retType, argTypes, *args),
>>> >> com_error: (-2147319779, 'Library not registered.', None, None)
>>> >>
>>> >> The other suggestion I came across was to use QueryInterface. I
>>> wasn't
>>> >> sure what argument to pass to QueryInterface, so I tried using the
>>> >> CLSID of the interface I wanted.
>>> >>
>>> >>>>> obj._oleobj_.QueryInterface(Interface_2.CLSID)
>>> >> Traceback (most recent call last):
>>> >>  File "<interactive input>", line 1, in <module>
>>> >> TypeError: There is no interface object registered that supports
>>> this IID
>>> >>
>>> >> From these error messages, it seems like perhaps there's a problem
>>> >> with the registration of the type library that contains the
>>> >> interfaces. I'm not really sure how to fix that, or if that is even
>>> >> the problem.
>>> >>
>>> >> Any ideas?
>>> >>
>>> >> Thanks,
>>> >> Alec Munro
>>> >>
>>> >
>>> _______________________________________________
>>> python-win32 mailing list
>>> python-win32 at python.org
>>> http://mail.python.org/mailman/listinfo/python-win32
>>
>>
>

From codingjester at gmail.com  Tue Sep  2 19:25:55 2008
From: codingjester at gmail.com (John Bunting)
Date: Tue, 2 Sep 2008 13:25:55 -0400
Subject: [python-win32] A question about custom command options in
	win32serviceutil
Message-ID: <b146c4a40809021025o64746cd2ga893e5be1d907f16@mail.gmail.com>

Hey everyone,

I'm creating a program for my work that will let the users automatically
connect to a server that we have set up and read some data, etc.

I have been trying to figure out a way to use the win32serviceutil custom
options to pass some extra arguments for my users.

as far as I understand it, you have to use a custom function (I think) to
deal with all of your extra arguments that it doesn't recognize.

I would like to be able to apply to my windows class these extra arguments,
as they will just be used to pass data! is it correct in assuming

that i should probably set it up like this...


class service(win32service.ServiceFramework):
        def __init__(self):
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080902/ae190361/attachment.htm>

From codingjester at gmail.com  Tue Sep  2 19:27:10 2008
From: codingjester at gmail.com (John Bunting)
Date: Tue, 2 Sep 2008 13:27:10 -0400
Subject: [python-win32] A question about custom command options in
	win32serviceutil
In-Reply-To: <b146c4a40809021025o64746cd2ga893e5be1d907f16@mail.gmail.com>
References: <b146c4a40809021025o64746cd2ga893e5be1d907f16@mail.gmail.com>
Message-ID: <b146c4a40809021027y5e1cda44j5e17d0ec5c09feaf@mail.gmail.com>

sorry pressed enter too soon....

Am i just setting up a function to parse the arguments and set class
variables? or is it something else!

Sorry for the double email >.<

~John Bunting

On Tue, Sep 2, 2008 at 1:25 PM, John Bunting <codingjester at gmail.com> wrote:

> Hey everyone,
>
> I'm creating a program for my work that will let the users automatically
> connect to a server that we have set up and read some data, etc.
>
> I have been trying to figure out a way to use the win32serviceutil custom
> options to pass some extra arguments for my users.
>
> as far as I understand it, you have to use a custom function (I think) to
> deal with all of your extra arguments that it doesn't recognize.
>
> I would like to be able to apply to my windows class these extra arguments,
> as they will just be used to pass data! is it correct in assuming
>
> that i should probably set it up like this...
>
>
> class service(win32service.ServiceFramework):
>         def __init__(self):
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080902/c1f65d1a/attachment.htm>

From h.goebel at goebel-consult.de  Wed Sep  3 17:40:46 2008
From: h.goebel at goebel-consult.de (Hartmut Goebel)
Date: Wed, 03 Sep 2008 17:40:46 +0200
Subject: [python-win32] Can't import win32api from pywin32-egg in wine
Message-ID: <48BEAFFE.2090705@goebel-consult.de>

Hi,

when running Python in Wine, I can not import win32api which is
installed as pywin32-212-py2.5-win32.egg::

$ wine python
Python 2.5.2 (r252:60911, Feb 21 2008, 13:11:45) [MSC v.1310 32 bit
(Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import win32api
err:module:import_dll Library pywintypes25.dll (which is needed by
L"C:\\Python25\\lib\\site-packages\\pywin32-212-py2.5-win32.egg\\win32api.pyd")
not found
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: DLL load failed: Module not found
>>>

The dll exists, but

$ ls site-packages/pywin32*.egg/pywin32_system32/pywintypes25.*
site-packages/pywin32-212-py2.5-win32.egg/pywin32_system32/pywintypes25.dll

This tracks down to: pywintypes.py can not find pywintypes32.dll using
_win32sysloader.GetModuleFilename(filename) or
_win32sysloader.LoadModule(filename)

Any hints?

-- 
Sch?nen Gru? - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP

Goebel Consult
Spezialist f?r IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3195 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080903/e505cce6/attachment.bin>

From mhammond at skippinet.com.au  Thu Sep  4 09:18:22 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Thu, 4 Sep 2008 17:18:22 +1000
Subject: [python-win32] Can't import win32api from pywin32-egg in wine
In-Reply-To: <48BEAFFE.2090705@goebel-consult.de>
References: <48BEAFFE.2090705@goebel-consult.de>
Message-ID: <009d01c90e5e$6e459b60$4ad0d220$@com.au>

IIRC there is a bug in setuptools that prevented the post-install script
running in an egg - which would prevent those 2 DLLs being copied from
pywin32_system32 to either system32 or next to python.exe.  Try copying them
manually, or run 'pywin32_postinstall.py -install'

Mark

> -----Original Message-----
> From: python-win32-bounces+skippy.hammond=gmail.com at python.org
> [mailto:python-win32-bounces+skippy.hammond=gmail.com at python.org] On
> Behalf Of Hartmut Goebel
> Sent: Thursday, 4 September 2008 1:41 AM
> To: python-win32 at python.org
> Subject: [python-win32] Can't import win32api from pywin32-egg in wine
> 
> Hi,
> 
> when running Python in Wine, I can not import win32api which is
> installed as pywin32-212-py2.5-win32.egg::
> 
> $ wine python
> Python 2.5.2 (r252:60911, Feb 21 2008, 13:11:45) [MSC v.1310 32 bit
> (Intel)] on win32 Type "help", "copyright", "credits" or "license" for
> more information.
> >>> import win32api
> err:module:import_dll Library pywintypes25.dll (which is needed by
> L"C:\\Python25\\lib\\site-packages\\pywin32-212-py2.5-
> win32.egg\\win32api.pyd")
> not found
> Traceback (most recent call last):
>   File "<stdin>", line 1, in <module>
> ImportError: DLL load failed: Module not found
> >>>
> 
> The dll exists, but
> 
> $ ls site-packages/pywin32*.egg/pywin32_system32/pywintypes25.*
> site-packages/pywin32-212-py2.5-
> win32.egg/pywin32_system32/pywintypes25.dll
> 
> This tracks down to: pywintypes.py can not find pywintypes32.dll using
> _win32sysloader.GetModuleFilename(filename) or
> _win32sysloader.LoadModule(filename)
> 
> Any hints?
> 
> --
> Sch?nen Gru? - Regards
> Hartmut Goebel
> Dipl.-Informatiker (univ.), CISSP
> 
> Goebel Consult
> Spezialist f?r IT-Sicherheit in komplexen Umgebungen http://www.goebel-
> consult.de



From h.goebel at goebel-consult.de  Thu Sep  4 10:05:10 2008
From: h.goebel at goebel-consult.de (Hartmut Goebel)
Date: Thu, 04 Sep 2008 10:05:10 +0200
Subject: [python-win32] Can't import win32api from pywin32-egg in wine
In-Reply-To: <009d01c90e5e$6e459b60$4ad0d220$@com.au>
References: <48BEAFFE.2090705@goebel-consult.de>
	<009d01c90e5e$6e459b60$4ad0d220$@com.au>
Message-ID: <48BF96B6.5070209@goebel-consult.de>

Hallo Mark,

> IIRC there is a bug in setuptools that prevented the post-install script
> running in an egg - which would prevent those 2 DLLs being copied from
> pywin32_system32 to either system32 or next to python.exe.  Try copying them
> manually, or run 'pywin32_postinstall.py -install'

Neither of these seams to work.

1) Both Scripts/pywin32_postinstall.py and
Lib/site-packages/pywin32-212-py2.5-win32.egg/EGG-INFO/scripts/pywin32_postinstall.py
fail.

2) Manual copy to C:\Python25, C:\Python25\DLLs and C:\windows\system32 
does not help.

I figured out that LoadSystemModule wants to load
C:\Python25\Lib\site-packages\pywin32_system32\pywintypes25.dll
which doe not exist in an egg.

I prepared a patch which I will issue to the sourceforge project page.

-- 
Sch?nen Gru? - Regards
Hartmut Goebel
Dipl.-Informatiker (univ.), CISSP

Goebel Consult
Spezialist f?r IT-Sicherheit in komplexen Umgebungen
http://www.goebel-consult.de

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 4719 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080904/70543385/attachment.bin>

From mhammond at skippinet.com.au  Thu Sep  4 11:50:49 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Thu, 4 Sep 2008 19:50:49 +1000
Subject: [python-win32] Can't import win32api from pywin32-egg in wine
In-Reply-To: <48BF96B6.5070209@goebel-consult.de>
References: <48BEAFFE.2090705@goebel-consult.de>
	<009d01c90e5e$6e459b60$4ad0d220$@com.au>
	<48BF96B6.5070209@goebel-consult.de>
Message-ID: <010101c90e73$b94785f0$2bd691d0$@com.au>

See also http://bugs.python.org/setuptools/issue18 - it would be great if
you can work out what overlap exists...

Cheers,

Mark


> -----Original Message-----
> From: Hartmut Goebel [mailto:h.goebel at goebel-consult.de]
> Sent: Thursday, 4 September 2008 6:05 PM
> To: mhammond at skippinet.com.au
> Cc: python-win32 at python.org
> Subject: Re: [python-win32] Can't import win32api from pywin32-egg in
> wine
> 
> Hallo Mark,
> 
> > IIRC there is a bug in setuptools that prevented the post-install
> script
> > running in an egg - which would prevent those 2 DLLs being copied
> from
> > pywin32_system32 to either system32 or next to python.exe.  Try
> copying them
> > manually, or run 'pywin32_postinstall.py -install'
> 
> Neither of these seams to work.
> 
> 1) Both Scripts/pywin32_postinstall.py and
> Lib/site-packages/pywin32-212-py2.5-win32.egg/EGG-
> INFO/scripts/pywin32_postinstall.py
> fail.
> 
> 2) Manual copy to C:\Python25, C:\Python25\DLLs and C:\windows\system32
> does not help.
> 
> I figured out that LoadSystemModule wants to load
> C:\Python25\Lib\site-packages\pywin32_system32\pywintypes25.dll
> which doe not exist in an egg.
> 
> I prepared a patch which I will issue to the sourceforge project page.
> 
> --
> Sch?nen Gru? - Regards
> Hartmut Goebel
> Dipl.-Informatiker (univ.), CISSP
> 
> Goebel Consult
> Spezialist f?r IT-Sicherheit in komplexen Umgebungen
> http://www.goebel-consult.de



From usmelbourne at gmail.com  Fri Sep  5 06:37:43 2008
From: usmelbourne at gmail.com (Ujjaval Suthar)
Date: Fri, 5 Sep 2008 14:37:43 +1000
Subject: [python-win32] pythoncom question
Message-ID: <7cd54e080809042137w3fa74d1di668c6f2002d04953@mail.gmail.com>

Hi all,

I am working on a python application which is a multi-threaded application.
I just want to make sure that I am doing the right thing or if not, please
let me know.

I have a parent thread which initalize the COM object inside a python
object.

This parent thread instantiates 4 other child threads which are independent
of each other with the python object created above. As I understand this
will allow calling COM object functions inside all 4 child threads. The
program does not seem to complain. But my parent thread seems to go and call
join method and wait till all the threads stop executing and something is
not right.

So, just checking was this the right way? or I have to initialize COM object
separately inside all child threads?



Thanks heaps.

Ujjaval
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080905/a477b9ca/attachment.htm>

From gagsl-py2 at yahoo.com.ar  Fri Sep  5 09:52:54 2008
From: gagsl-py2 at yahoo.com.ar (Gabriel Genellina)
Date: Fri, 05 Sep 2008 04:52:54 -0300
Subject: [python-win32] Internet Explorer constants
Message-ID: <op.ugzz6gsxx6zn5v@gabriel2.softlabbsas.com.ar>

Hello

I want to control Internet Explorer. This is what I have so far:

import win32com.client
 from win32com.client.gencache import EnsureDispatch
url = "..."
IE = EnsureDispatch("InternetExplorer.Application")
IE.Navigate(url)
IE.Visible = 1

It works fine and shows the requested page. Now I want to use the "Save  
As" command. Looks like I should call the
IWebBrowser2.ExecWB method, passing IDM_SAVEAS as the command and  
MSOCMDEXECOPT_DONTPROMPTUSER in the options. But I don't know how to  
obtain the value for such constants... I could dig into the C header files  
looking for them, but is there some other way?

The ExecWB method is documented here  
http://msdn.microsoft.com/en-us/library/aa752117(VS.85).aspx

-- 
Gabriel Genellina


From usmelbourne at gmail.com  Fri Sep  5 10:19:27 2008
From: usmelbourne at gmail.com (Ujjaval Suthar)
Date: Fri, 5 Sep 2008 18:19:27 +1000
Subject: [python-win32] Handling COM events in a python thread
Message-ID: <7cd54e080809050119p699c7046me926c4e12d3ff343@mail.gmail.com>

Hi all,

I have a COM control which fires a status event called
ProcessingDataAvailable(). The COM control has a method call AddData() with
some data parameters.

Using python, in a simple program, I can instantiate COM object via Dispatch
and call AddData() method successfully.

Now, I want to have a thread which keeps running and listening for
ProcessingDataAvailable() event and do some processing in my program.

Can somebody please give me a simple example code that shows how it can be
done? I'm very new to this kind of programming.


Thanks.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080905/d90995fb/attachment.htm>

From mail at timgolden.me.uk  Fri Sep  5 10:23:01 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Fri, 05 Sep 2008 09:23:01 +0100
Subject: [python-win32] Internet Explorer constants
In-Reply-To: <op.ugzz6gsxx6zn5v@gabriel2.softlabbsas.com.ar>
References: <op.ugzz6gsxx6zn5v@gabriel2.softlabbsas.com.ar>
Message-ID: <48C0EC65.7000907@timgolden.me.uk>

Gabriel Genellina wrote:
> Hello
> 
> I want to control Internet Explorer. This is what I have so far:
> 
> import win32com.client
> from win32com.client.gencache import EnsureDispatch
> url = "..."
> IE = EnsureDispatch("InternetExplorer.Application")
> IE.Navigate(url)
> IE.Visible = 1
> 
> It works fine and shows the requested page. Now I want to use the "Save 
> As" command. Looks like I should call the
> IWebBrowser2.ExecWB method, passing IDM_SAVEAS as the command and 
> MSOCMDEXECOPT_DONTPROMPTUSER in the options. But I don't know how to 
> obtain the value for such constants... I could dig into the C header 
> files looking for them, but is there some other way?
> 
> The ExecWB method is documented here 
> http://msdn.microsoft.com/en-us/library/aa752117(VS.85).aspx

The usual way to get at constants after a Dispatch
is via the win32com.client.constants object. In this
case, the constant you want is OLECMDID_SAVEAS which...

<code>
import win32com.client

IE = "InternetExplorer.Application"
ie = win32com.client.gencache.EnsureDispatch (IE)

print win32com.client.constants.OLECMDID_SAVEAS 

</code>


TJG

From gagsl-py2 at yahoo.com.ar  Fri Sep  5 11:32:59 2008
From: gagsl-py2 at yahoo.com.ar (Gabriel Genellina)
Date: Fri, 05 Sep 2008 06:32:59 -0300
Subject: [python-win32] Internet Explorer constants
References: <op.ugzz6gsxx6zn5v@gabriel2.softlabbsas.com.ar>
	<48C0EC65.7000907@timgolden.me.uk>
Message-ID: <op.ugz4s9pex6zn5v@gabriel2.softlabbsas.com.ar>

En Fri, 05 Sep 2008 05:23:01 -0300, Tim Golden <mail at timgolden.me.uk>  
escribi?:

> Gabriel Genellina wrote:
>> Hello
>>  I want to control Internet Explorer. This is what I have so far:
>>  import win32com.client
>> from win32com.client.gencache import EnsureDispatch
>> url = "..."
>> IE = EnsureDispatch("InternetExplorer.Application")
>> IE.Navigate(url)
>> IE.Visible = 1
>>  It works fine and shows the requested page. Now I want to use the  
>> "Save As" command. Looks like I should call the
>> IWebBrowser2.ExecWB method, passing IDM_SAVEAS as the command and  
>> MSOCMDEXECOPT_DONTPROMPTUSER in the options. But I don't know how to  
>> obtain the value for such constants... I could dig into the C header  
>> files looking for them, but is there some other way?
>>  The ExecWB method is documented here  
>> http://msdn.microsoft.com/en-us/library/aa752117(VS.85).aspx
>
> The usual way to get at constants after a Dispatch
> is via the win32com.client.constants object. In this
> case, the constant you want is OLECMDID_SAVEAS which...
>
> <code>
> import win32com.client
>
> IE = "InternetExplorer.Application"
> ie = win32com.client.gencache.EnsureDispatch (IE)
>
> print win32com.client.constants.OLECMDID_SAVEAS </code>

Ok, thanks! Looks like these are the names to use instead:

IE.ExecWB(
   constants.OLECMDID_SAVEAS,
   constants.OLECMDEXECOPT_DONTPROMPTUSER,
   filename,
   None)

Altough the ...DONTPROMPTUSER option doesn't work - the Save As dialog  
always appears. Some kind of permission or safety measure, I presume.

My original intent was to save the file using the .mht format, but that  
format is not listed on the Save As dialog. I'll investigate using  
IMessage.CreateMHTMLBody but that's a different topic.

-- 
Gabriel Genellina


From mail at timgolden.me.uk  Fri Sep  5 17:13:59 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Fri, 05 Sep 2008 16:13:59 +0100
Subject: [python-win32] Handling COM events in a python thread
In-Reply-To: <7cd54e080809050119p699c7046me926c4e12d3ff343@mail.gmail.com>
References: <7cd54e080809050119p699c7046me926c4e12d3ff343@mail.gmail.com>
Message-ID: <48C14CB7.4060801@timgolden.me.uk>

Ujjaval Suthar wrote:
> Hi all,
> 
> I have a COM control which fires a status event called 
> ProcessingDataAvailable(). The COM control has a method call AddData() 
> with some data parameters.
> 
> Using python, in a simple program, I can instantiate COM object via 
> Dispatch and call AddData() method successfully.
> 
> Now, I want to have a thread which keeps running and listening for 
> ProcessingDataAvailable() event and do some processing in my program.
> 
> Can somebody please give me a simple example code that shows how it can 
> be done? I'm very new to this kind of programming.


Well here's a crude example to get you going. It uses the
MSAgent thingy which I imagine is installed (if little used)
on most Windows machines. Notice that it's launching a thread
to pick up events from a Queue -- the standard Python structure
for communicating between threads.

<code>
import win32com.client
import pythoncom
import Queue
import threading
import time

events = Queue.Queue ()

def do_things_with_events (events):
  while True:
    event = events.get ()
    open ("c:/temp/events.log", "a").write (str (event) + "\n")

class AgentEvents:
  def OnHide (self, character_id, cause):
    events.put (("hide", character_id, cause))

#
# Agent.Control.1 - Win2K characters
# Agent.Control.2 - Office2K characters
#
agent = win32com.client.DispatchWithEvents ("Agent.Control.2", AgentEvents)
agent.Connected = True

def start (name, filename=None, show_immediately=True):
  filename = filename or ("%s.acs" % name)
  agent.Characters.Load (name, filename)
  this_agent = agent.Characters (name)
  this_agent.Show ()

if __name__ == '__main__':
  import time
  threading.Thread (target=do_things_with_events, args=(events,)).start ()
  start ("merlin")
  while True:
    pythoncom.PumpWaitingMessages ()
    time.sleep (0.1)

</code>


From vernondcole at gmail.com  Fri Sep  5 21:54:09 2008
From: vernondcole at gmail.com (Vernon Cole)
Date: Fri, 5 Sep 2008 13:54:09 -0600
Subject: [python-win32] adodbapi 2.2.1 (bugfix) release.
Message-ID: <b7eb4a700809051254kfcc2097o2487094ebe64b647@mail.gmail.com>

I found a bug in adodbapi v2.2 already. When putting data into a numeric
database column, if the input was in a string (as for example the value of a
wxPython grid cell), then the data would be truncated to the binary length
of the field. So if I tried tried putting "12345" into an integer column,
the value 1234 would be stored, because integers are 4 bytes long. This took
me a while to find, since I kept looking at my wxwidgets code before finally
uncovering adodbapi as the culpret.

  The fixed version is in the pywin32 CVS and in
http://sourceforge.net/projects/adodbapi .

For some good news,  Mike Driscoll reports:

I used your adodbapi module in my wxPython / SqlAlchemy program today and it
> seemed to be returning the same data as the default module (which I think is
> pyobdc). I don't know if my application truly stresses adodbapi , but it
> does make lots of calls and accesses multiple tables in multiple databases.
>
>
If anyone can test SqlAlchemy -> adodbapi -> IronPython please let us know
if it works.
--
Vernon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080905/b1a995c3/attachment.htm>

From mdriscoll at co.marshall.ia.us  Fri Sep  5 22:38:35 2008
From: mdriscoll at co.marshall.ia.us (Mike Driscoll)
Date: Fri, 05 Sep 2008 15:38:35 -0500
Subject: [python-win32] Internet Explorer constants
In-Reply-To: <op.ugz4s9pex6zn5v@gabriel2.softlabbsas.com.ar>
References: <op.ugzz6gsxx6zn5v@gabriel2.softlabbsas.com.ar>	<48C0EC65.7000907@timgolden.me.uk>
	<op.ugz4s9pex6zn5v@gabriel2.softlabbsas.com.ar>
Message-ID: <48C198CB.7090501@co.marshall.ia.us>

Hi Gabriel,
> <div class="moz-text-flowed" style="font-family: -moz-fixed">En Fri, 
> 05 Sep 2008 05:23:01 -0300, Tim Golden <mail at timgolden.me.uk> escribi?:
>
>> Gabriel Genellina wrote:
>>> Hello
>>>  I want to control Internet Explorer. This is what I have so far:
>>>  import win32com.client
>>> from win32com.client.gencache import EnsureDispatch
>>> url = "..."
>>> IE = EnsureDispatch("InternetExplorer.Application")
>>> IE.Navigate(url)
>>> IE.Visible = 1
>>>  It works fine and shows the requested page. Now I want to use the 
>>> "Save As" command. Looks like I should call the
>>> IWebBrowser2.ExecWB method, passing IDM_SAVEAS as the command and 
>>> MSOCMDEXECOPT_DONTPROMPTUSER in the options. But I don't know how to 
>>> obtain the value for such constants... I could dig into the C header 
>>> files looking for them, but is there some other way?
>>>  The ExecWB method is documented here 
>>> http://msdn.microsoft.com/en-us/library/aa752117(VS.85).aspx
>>
>> The usual way to get at constants after a Dispatch
>> is via the win32com.client.constants object. In this
>> case, the constant you want is OLECMDID_SAVEAS which...
>>
>> <code>
>> import win32com.client
>>
>> IE = "InternetExplorer.Application"
>> ie = win32com.client.gencache.EnsureDispatch (IE)
>>
>> print win32com.client.constants.OLECMDID_SAVEAS </code>
>
> Ok, thanks! Looks like these are the names to use instead:
>
> IE.ExecWB(
>   constants.OLECMDID_SAVEAS,
>   constants.OLECMDEXECOPT_DONTPROMPTUSER,
>   filename,
>   None)
>
> Altough the ...DONTPROMPTUSER option doesn't work - the Save As dialog 
> always appears. Some kind of permission or safety measure, I presume.
>
> My original intent was to save the file using the .mht format, but 
> that format is not listed on the Save As dialog. I'll investigate 
> using IMessage.CreateMHTMLBody but that's a different topic.
>

You could also look at PAMIE: http://pamie.sourceforge.net/  or 
Mechanize: http://pypi.python.org/pypi/mechanize/0.1.7b

I don't think either uses the PyWin32 library, but I hear that they're 
both pretty good.

-------------------
Mike Driscoll

Blog:   http://blog.pythonlibrary.org
Python Extension Building Network:     http://www.pythonlibrary.org



From mkesper at fsfe.org  Fri Sep  5 23:47:20 2008
From: mkesper at fsfe.org (Michael Kesper)
Date: Fri, 5 Sep 2008 23:47:20 +0200
Subject: [python-win32] Accessing properties of share on DFS
In-Reply-To: <48BC14A0.3050202@timgolden.me.uk>
References: <20080825143047.GD13124@localhost>
	<48B55829.5020205@timgolden.me.uk>
	<20080901120812.GB4376@localhost>
	<48BC14A0.3050202@timgolden.me.uk>
Message-ID: <20080905214720.GC5062@localhost>

Hi,

* Tim Golden <mail at timgolden.me.uk> [2008-09-01 17:13:20 +0100]:
?
>> * Tim Golden <mail at timgolden.me.uk> [2008-08-27 14:35:37 +0100]:
>>  
>>> You might be able to get there with WMI. The following
>>> works on my own machine, but I'm not sure what machine
>>> you'd need to access to list the DFS shares. I'll see
>>> if I can persuade a friendly admin here to let me try
>>> things out:

It seems I need to access net32api.NetShareGetInfo or maybe
IADsFileShare, but I have to figure this out after my holiday.
Until now I used the trick to simply call os.listdir() on
the servers where I knew the homes are lying.
os.listdir is cool enough to work with UNC paths. :)

Best wishes
Michael 
-- 
Free Software Foundation Europe (FSFE) []         (http://fsfeurope.org)
Treten Sie der Fellowship bei!       [][][]       (http://fsfe.org/join)
Ihre Spende erm?glicht unsere Arbeit!  ||  (http://fsfeurope.org/donate)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 315 bytes
Desc: Digital signature
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080905/7b50337e/attachment.pgp>

From vernondcole at gmail.com  Sun Sep  7 16:20:10 2008
From: vernondcole at gmail.com (Vernon Cole)
Date: Sun, 7 Sep 2008 08:20:10 -0600
Subject: [python-win32] Fwd: adodbapi 2.2.1 (bugfix) release.
In-Reply-To: <b7eb4a700809051254kfcc2097o2487094ebe64b647@mail.gmail.com>
References: <b7eb4a700809051254kfcc2097o2487094ebe64b647@mail.gmail.com>
Message-ID: <b7eb4a700809070720ha858579saa9dcf3908d9d0ba@mail.gmail.com>

[I tried sending this to two lists at once, and it seems to have been
rejected by both, so this is a retry. -- VC]
I found a bug in adodbapi v2.2 already. When putting data into a numeric
database column, if the input was in a string (as for example the value of a
wxPython grid cell), then the data would be truncated to the binary length
of the field. So if I tried tried putting "12345" into an integer column,
the value 1234 would be stored, because integers are 4 bytes long. This took
me a while to find, since I kept looking at my wxwidgets code before finally
uncovering adodbapi as the culpret.

  The fixed version is in the pywin32 CVS and in
http://sourceforge.net/projects/adodbapi .

For some good news,  Mike Driscoll reports:

I used your adodbapi module in my wxPython / SqlAlchemy program today and it
> seemed to be returning the same data as the default module (which I think is
> pyobdc). I don't know if my application truly stresses adodbapi , but it
> does make lots of calls and accesses multiple tables in multiple databases.
>
>
If anyone can test SqlAlchemy -> adodbapi -> IronPython please let us know
if it works.
--
Vernon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080907/da9c0a2f/attachment.htm>

From patrick at raptr.com  Mon Sep  8 08:28:15 2008
From: patrick at raptr.com (Patrick Li)
Date: Sun, 7 Sep 2008 23:28:15 -0700
Subject: [python-win32] how to obtain list of connected IP/port pairs on
	user's computer?
Message-ID: <2b5a48740809072328h4cca2a56ve439212a4c399aa3@mail.gmail.com>

Hi,

I'm interested in writing a function that returns me the list of IP/port
that the user's machine is connected to.  The most naive way is to run
netstat -na on the user's computer and just parse the stdout.  Are there
libraries I can call to achieve this as opposed to running an external
command?

Thanks,
Patrick
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080907/e73ff4b1/attachment.htm>

From mail at timgolden.me.uk  Mon Sep  8 11:01:41 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Mon, 08 Sep 2008 10:01:41 +0100
Subject: [python-win32] how to obtain list of connected IP/port pairs
 on	user's computer?
In-Reply-To: <2b5a48740809072328h4cca2a56ve439212a4c399aa3@mail.gmail.com>
References: <2b5a48740809072328h4cca2a56ve439212a4c399aa3@mail.gmail.com>
Message-ID: <48C4E9F5.5000605@timgolden.me.uk>

Patrick Li wrote:
> I'm interested in writing a function that returns me the list of IP/port 
> that the user's machine is connected to.  The most naive way is to run 
> netstat -na on the user's computer and just parse the stdout.  Are there 
> libraries I can call to achieve this as opposed to running an external 
> command?

There was a thread about this a couple of months ago;
might help:

http://mail.python.org/pipermail/python-win32/2008-June/007786.html

TJG


From etiennepouliot at gmail.com  Tue Sep  9 17:13:27 2008
From: etiennepouliot at gmail.com (Etienne Pouliot)
Date: Tue, 9 Sep 2008 11:13:27 -0400
Subject: [python-win32] getting TerminalServicesProfilePath
Message-ID: <729c89ef0809090813u7b81a29by2276994e3c8d51a4@mail.gmail.com>

Hi,

  I'm trying to get TerminalServicesProfilePath property on user account in
my Active Directory.

adsi = win32com.client.Dispatch('ADsNameSpaces')
ldap = adsi.getobject("","LDAP:")
logon_ex='cn=Admin,dc=.....'
password='mypassword'
ex_path="LDAP://myserver/cn=test123,..."
obj = ldap.OpenDSObject(ex_path,logon_ex,password,0)
obj.Getinfo()

Then I tryed with :
print obj.TerminalServicesProfilePath
or
obj.get('TerminalServicesProfilePath')

None worked

If I use another properties instead, like DisplayName, it works.

Any idea what I am doing wrong ?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080909/2019a574/attachment.htm>

From Dominick.Lauzon at MSCsoftware.com  Wed Sep 10 14:02:35 2008
From: Dominick.Lauzon at MSCsoftware.com (Dominick Lauzon)
Date: Wed, 10 Sep 2008 08:02:35 -0400
Subject: [python-win32] Equivalent support on win64 ?
Message-ID: <589DC3F5067BDD479AAAE54CE398166D418303@NAAAEX01.na.mscsoftware.com>

Are we looking/expecting to obtain the same equivalent functionality and
support for all win32com on win64 platforms shortly ?

 

 

Dominick Lauzon

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080910/c45242a1/attachment.htm>

From Solomon.Zewdie.Altek at zf.com  Wed Sep 10 14:25:27 2008
From: Solomon.Zewdie.Altek at zf.com (Solomon.Zewdie.Altek at zf.com)
Date: Wed, 10 Sep 2008 14:25:27 +0200
Subject: [python-win32] ImportError: No system modul
Message-ID: <8A13418356E45E4784744371A8B4DD7971FFAF@frds00501.emea.zf-world.com>


 Hello, 
 I am tring to work on Excel File through python. 

 I am using Python 2.4.2 and work on xp with an administration authorization. 
  
 I implimented like that: 
  
 *********************************************************
 * 
 ** 
 ... 
 import win32com.client 
 ... 
 class Python_skript(Object): 
  
    def __init__(self, fileName=None): 
       self.xlapp = win32com.client.dynamic.Dispatch("Excel.Application") 
       if fileName: 
          self.xlbook = self.xlapp.Workbooks.Open(fileName) 
       else: 
          self.xlbook = self.xlapp.Workbooks.Add() ... 
 **********************************************************
 
  
 I got the error message: "ImportError: No system modul 'pythoncom' (pythoncom24.dll)"

 Although the file pythoncom24.dll does exist. 
  
 Can anyone help? 
  
 Many Thanx! 


From mhammond at skippinet.com.au  Wed Sep 10 14:48:17 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Wed, 10 Sep 2008 22:48:17 +1000
Subject: [python-win32] Equivalent support on win64 ?
In-Reply-To: <589DC3F5067BDD479AAAE54CE398166D418303@NAAAEX01.na.mscsoftware.com>
References: <589DC3F5067BDD479AAAE54CE398166D418303@NAAAEX01.na.mscsoftware.com>
Message-ID: <04ff01c91343$86cc1cb0$94645610$@com.au>

Regardless of what 'we' are expecting shortly, we can already find this
support if only we looked (for python 2.6 anyway - other versions will
require us to build it ourselves using vs2008)

Mark.

----

From: python-win32-bounces+skippy.hammond=gmail.com at python.org
[mailto:python-win32-bounces+skippy.hammond=gmail.com at python.org] On Behalf
Of Dominick Lauzon
Sent: Wednesday, 10 September 2008 10:03 PM
To: python-win32 at python.org
Subject: [python-win32] Equivalent support on win64 ?

Are we looking/expecting to obtain the same equivalent functionality and
support for all win32com on win64 platforms shortly ?


Dominick Lauzon


From timr at probo.com  Wed Sep 10 19:35:11 2008
From: timr at probo.com (Tim Roberts)
Date: Wed, 10 Sep 2008 10:35:11 -0700
Subject: [python-win32] ImportError: No system modul
In-Reply-To: <8A13418356E45E4784744371A8B4DD7971FFAF@frds00501.emea.zf-world.com>
References: <8A13418356E45E4784744371A8B4DD7971FFAF@frds00501.emea.zf-world.com>
Message-ID: <48C8054F.6040802@probo.com>

Solomon.Zewdie.Altek at zf.com wrote:
>  I am tring to work on Excel File through python. 
>
>  I am using Python 2.4.2 and work on xp with an administration authorization. 
>   
>  I implimented like that: 
>   
>  *********************************************************
>  * 
>  ** 
>  ... 
>  import win32com.client 
>  ... 
>  class Python_skript(Object): 
>   
>     def __init__(self, fileName=None): 
>        self.xlapp = win32com.client.dynamic.Dispatch("Excel.Application") 
>        if fileName: 
>           self.xlbook = self.xlapp.Workbooks.Open(fileName) 
>        else: 
>           self.xlbook = self.xlapp.Workbooks.Add() ... 
>  **********************************************************
>  
>   
>  I got the error message: "ImportError: No system modul 'pythoncom' (pythoncom24.dll)"
>
>  Although the file pythoncom24.dll does exist. 
>   

It must do more than "exist", of course.  It must be in the right spot. 
How did you install pywin32?  Where does it exist?  And what does this say?
    import sys
    print sys.path

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From alecmunro at gmail.com  Wed Sep 10 21:22:44 2008
From: alecmunro at gmail.com (Alec Munro)
Date: Wed, 10 Sep 2008 16:22:44 -0300
Subject: [python-win32] Accessing other interfaces of a Dispatch created
	COM object
In-Reply-To: <9819d58b0809020944v488ab4a2n5127c71e7c69424b@mail.gmail.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>
	<9819d58b0808280632x2b139645x278d5f0675dd46c7@mail.gmail.com>
	<9819d58b0808280633t13d89b77u433d55924cbb73a1@mail.gmail.com>
	<9819d58b0808280734o33af5721y741ca9b079d074af@mail.gmail.com>
	<00de01c9099c$b77731c0$26659540$@com.au>
	<9819d58b0808291047j3de5e9d8u7647b05e49a53f97@mail.gmail.com>
	<9819d58b0809020944v488ab4a2n5127c71e7c69424b@mail.gmail.com>
Message-ID: <9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>

Ok, perhaps this will be useful.

(Neither of these examples include the core arguments of the methods,
just the ones specific to InvokeTypes)

What I call a method from I1, it calls _oleobj_.InvokeTypes with the
following arguments:

18, 0x0, 1, (24, 0), ((8, 1),)

When I use the method from I2, it calls _ApplyTypes_(), which calls
_oleobj_.InvokeTypes with the following arguments:

1, 0x0, 1, (24, 0), ((3, 1), (12, 1), (16396, 2))

Now, from my reading (of old messages by Mark :) ), the last two
parameters specify method return type and argument type. In this case,
it seems like both methods return VOID, which seems fine. I'm a little
more concerned about the argument type. The first method is supposed
to take a single string, so that looks fine. The second method takes
an integer, a VT_TYPE, and a variant. I don't know if the
representation here accurately represents that.

The first parameter seems to be the ID of the method within the
interface (or something similar), and given that I2 only has a single
method, "1" seems like it could be correct.

So, I'm fairly certain this puts that problem at something that
happens in InvokeTypes, unrelated to the arguments I passed in. I get
the impression that InvokeTypes is something of Microsofts, and
there's isn't really any way to observe what it's doing?

Thanks, hopefully this helps someone.

Alec

On Tue, Sep 2, 2008 at 1:44 PM, Alec Munro <alecmunro at gmail.com> wrote:
> I'm guessing, based on the lack of responses, and the consistency of
> error messages, that my problem has more to do with getting the right
> library registered, which isn't a python-win job.
>
> So, can anyone recommend where I might go to try to solve this
> problem? Also, is there a way to determine which library it is that is
> not registered?
>
> Thanks again,
>
> Alec
>
> On Fri, Aug 29, 2008 at 2:47 PM, Alec Munro <alecmunro at gmail.com> wrote:
>> Thanks Mark.
>>
>> After a couple of misfires, I think I figured out how to get CastTo to
>> work (the docstring is practically non-existent in the version I have,
>> perhaps it's been updated?).
>>
>> However, that puts me back at the same place I was with some of the
>> other attempts:
>>
>>>>> obj_i2 = CastTo(obj, "Interface_2")
>>>>> obj_i2.secondaryMethod()
>> Traceback (most recent call last):
>> ...
>>  File "C:\Python25\lib\site-packages\win32com\client\__init__.py",
>> line 449, in _ApplyTypes_
>>   dispid, 0, wFlags, retType, argTypes, *args),
>> com_error: (-2147319779, 'Library not registered.', None, None)
>>
>> However, one interesting note is that the object generated by CastTo
>> is of a different version of the Type Library than the object
>> generated by Interface_2(obj). I don't know if that's useful
>> information, however. They have the same CLSID, but different
>> coclass_clsids.
>>
>> Thanks again,
>>
>> Alec
>>
>> On Fri, Aug 29, 2008 at 3:01 AM, Mark Hammond <mhammond at skippinet.com.au> wrote:
>>> Look at the docstring for win32com.client.CastTo (and googling for that will
>>> probably help too)
>>>
>>> Cheers,
>>>
>>> Mark
>>>
>>>> I just discovered a third possibility, and subsequent failure:
>>>>
>>>> >>> obj_i2 = Dispatch(obj, None, Interface_2.CLSID)
>>>> >>> obj_i2.secondaryMethod()
>>>> Traceback (most recent call last):
>>>> ...
>>>>   File "C:\Python25\lib\site-packages\win32com\client\__init__.py",
>>>> line 449, in _ApplyTypes_
>>>>     dispid, 0, wFlags, retType, argTypes, *args),
>>>> com_error: (-2147352573, 'Member not found.', None, None)
>>>>
>>>> Alec
>>>>
>>>> On Thu, Aug 28, 2008 at 10:33 AM, Alec Munro <alecmunro at gmail.com>
>>>> wrote:
>>>> > Oh, and on re-reading, I do realize that the DLL itself doesn't
>>>> > implement those interfaces. There's a class within the DLL that does
>>>> > that. :)
>>>> >
>>>> > On Thu, Aug 28, 2008 at 10:32 AM, Alec Munro <alecmunro at gmail.com>
>>>> wrote:
>>>> >> Hi List,
>>>> >>
>>>> >> Hopefully I have my terminology right, I'm pretty new to COM and
>>>> C++.
>>>> >>
>>>> >> My problem is that I have a COM DLL written in C++, that I would
>>>> like
>>>> >> to call from Python. This DLL implements 3 important interfaces,
>>>> each
>>>> >> of which is also an IDispatch implementation, like the following:
>>>> >>
>>>> >> IDispatchImpl<Interface_1, &__uuidof(Interface_1),
>>>> &LIBID_Interface_1, 1, 0>
>>>> >> IDispatchImpl<Interface_2, &__uuidof(Interface_2),
>>>> &LIBID_Interface_2, 1, 0>
>>>> >> IDispatchImpl<Interface_3, &__uuidof(Interface_3),
>>>> &LIBID_Interface_3 1, 0>
>>>> >>
>>>> >>  However, when I do:
>>>> >>>>> obj = win32com.client.Dispatch("a.b.c")
>>>> >>>>> obj
>>>> >> <COMObject a.b.c>
>>>> >>
>>>> >> The object I get only has methods available for the first of the
>>>> >> interfaces. From my reading, this is simply the functionality of
>>>> >> IDispatch, and if I wanted to get those additional methods through
>>>> >> IDispatch, I would have to update the DLL, and restructure my
>>>> >> interfaces. I'm not confident enough in my C++ to do this, or at
>>>> least
>>>> >> it's not my first choice.
>>>> >>
>>>> >> From what I've read, there are two ways that seem to be
>>>> possibilities
>>>> >> for being able to call the methods from the other interfaces.
>>>> >>
>>>> >> The simplest one I found was the suggestion to use MakePy on the
>>>> type
>>>> >> library containing Interface_2 and Interface_3. I did this, and
>>>> copied
>>>> >> the generated file to my working directory, then did:
>>>> >>
>>>> >>>>> from interfaces import Interface_2
>>>> >>>>> obj_i2 = Interface_2(obj)
>>>> >>>>> obj_i2
>>>> >> <win32com.gen_py.typelibname 1.0 Type Library.Interface_2 instance
>>>> at
>>>> >> 0x17459602>
>>>> >>
>>>> >> Unfortunately when I try:
>>>> >>
>>>> >>>>> obj_i2.secondaryMethod()
>>>> >> Traceback (most recent call last):
>>>> >> ...
>>>> >>  File "C:\Python25\lib\site-packages\win32com\client\__init__.py",
>>>> >> line 449, in _ApplyTypes_
>>>> >>   dispid, 0, wFlags, retType, argTypes, *args),
>>>> >> com_error: (-2147319779, 'Library not registered.', None, None)
>>>> >>
>>>> >> The other suggestion I came across was to use QueryInterface. I
>>>> wasn't
>>>> >> sure what argument to pass to QueryInterface, so I tried using the
>>>> >> CLSID of the interface I wanted.
>>>> >>
>>>> >>>>> obj._oleobj_.QueryInterface(Interface_2.CLSID)
>>>> >> Traceback (most recent call last):
>>>> >>  File "<interactive input>", line 1, in <module>
>>>> >> TypeError: There is no interface object registered that supports
>>>> this IID
>>>> >>
>>>> >> From these error messages, it seems like perhaps there's a problem
>>>> >> with the registration of the type library that contains the
>>>> >> interfaces. I'm not really sure how to fix that, or if that is even
>>>> >> the problem.
>>>> >>
>>>> >> Any ideas?
>>>> >>
>>>> >> Thanks,
>>>> >> Alec Munro
>>>> >>
>>>> >
>>>> _______________________________________________
>>>> python-win32 mailing list
>>>> python-win32 at python.org
>>>> http://mail.python.org/mailman/listinfo/python-win32
>>>
>>>
>>
>

From rblazecka at radical.ca  Wed Sep 10 21:22:49 2008
From: rblazecka at radical.ca (Ryan Blazecka)
Date: Wed, 10 Sep 2008 12:22:49 -0700
Subject: [python-win32] suggested change to pywintypes.py for
	non-admininstall
In-Reply-To: <45F19959.20906@activestate.com>
References: <014301c761fe$8ad76d70$070a0a0a@enfoldsystems.local>
	<45F19959.20906@activestate.com>
Message-ID: <ga96qs$109$1@ger.gmane.org>

Trent Mick wrote:
> Given that my suggestion is no panacea, perhaps it is better to just leave it 
> at status quo and tell the user about the requirement to update their PATH.
> 
> Thoughts?
> 
> Trent
> 

I have come across this issue, and am equally unsatisfied with the 
solutions I've come up with so far. all of them require the user to know 
too much about what's going on, and does not make the import "just work".

what if we added a python wrapper module (.py) around the native (.pyd) 
modules? it would contain something like the following:

import pywintypes
import pythoncom
from _win32ui import *

this would allow us to ensure pywintypes is loaded before the native 
code runs, and the python module would just pass all of the imported 
stuff from the binary module through to the caller. (some extra work to 
set __all__ properly is probably in order here too).


From timr at probo.com  Wed Sep 10 21:59:36 2008
From: timr at probo.com (Tim Roberts)
Date: Wed, 10 Sep 2008 12:59:36 -0700
Subject: [python-win32] Accessing other interfaces of a Dispatch created
 COM object
In-Reply-To: <9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>	<9819d58b0808280632x2b139645x278d5f0675dd46c7@mail.gmail.com>	<9819d58b0808280633t13d89b77u433d55924cbb73a1@mail.gmail.com>	<9819d58b0808280734o33af5721y741ca9b079d074af@mail.gmail.com>	<00de01c9099c$b77731c0$26659540$@com.au>	<9819d58b0808291047j3de5e9d8u7647b05e49a53f97@mail.gmail.com>	<9819d58b0809020944v488ab4a2n5127c71e7c69424b@mail.gmail.com>
	<9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>
Message-ID: <48C82728.8060206@probo.com>

Alec Munro wrote:
> Ok, perhaps this will be useful.
>
> (Neither of these examples include the core arguments of the methods,
> just the ones specific to InvokeTypes)
>
> What I call a method from I1, it calls _oleobj_.InvokeTypes with the
> following arguments:
>
> 18, 0x0, 1, (24, 0), ((8, 1),)
>
> When I use the method from I2, it calls _ApplyTypes_(), which calls
> _oleobj_.InvokeTypes with the following arguments:
>
> 1, 0x0, 1, (24, 0), ((3, 1), (12, 1), (16396, 2))
>
> Now, from my reading (of old messages by Mark :) ), the last two
> parameters specify method return type and argument type.

Well, it's the 4th argument that gives the return type.  24 is VT_VOID. 
That's unusual for a COM type.

> In this case,
> it seems like both methods return VOID, which seems fine. I'm a little
> more concerned about the argument type. The first method is supposed
> to take a single string, so that looks fine.

Right; 8 is VT_BSTR, which is the typical COM Unicope string type.

> The second method takes
> an integer, a VT_TYPE, and a variant. I don't know if the
> representation here accurately represents that.
>   

3 is VT_I4 (long), 12 is VT_VARIANT, and 16396 is a VT_VARIANT passed by
reference, meaning an output parameter.

> So, I'm fairly certain this puts that problem at something that
> happens in InvokeTypes, unrelated to the arguments I passed in. I get
> the impression that InvokeTypes is something of Microsofts, and
> there's isn't really any way to observe what it's doing?
>   

I've lost track.  What's the root problem?  Is it still  the "Library
not registered" issue when you call a method in Interface_2?  Was this
all registered through a type library, or through a regsvr32 call, or what?

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From Jim.Vickroy at noaa.gov  Wed Sep 10 21:35:15 2008
From: Jim.Vickroy at noaa.gov (Jim Vickroy)
Date: Wed, 10 Sep 2008 13:35:15 -0600
Subject: [python-win32] Microsoft Visual Sourcesafe automation --
Message-ID: <48C82173.6010205@noaa.gov>

Hello all,

I am attempting to automate Sourcesafe access.

Here is a sample session:

 >>> import sys
 >>> sys.version
'2.5.2 (r252:60911, Feb 21 2008, 13:11:45) [MSC v.1310 32 bit (Intel)]'
 >>> sys.subversion
('CPython', 'tags/r252', '60911')
 >>> sys.platform
'win32'
 >>> sys.getwindowsversion()
(5, 1, 2600, 2, 'Service Pack 2')
 >>> import win32com.client
 >>> vcs = win32com.client.Dispatch('SourceSafe')
 >>> config   = r'\\Tsg-admin-01\VSS\srcsafe.ini'
 >>> account  = 'jim.vickroy'
 >>> password = raw_input('password: ')
 >>> vcs.Open(SrcSafeIni=config, Username=account, Password=password)
Traceback (most recent call last):
  File "<interactive input>", line 1, in <module>
  File 
"C:\Python25\lib\site-packages\win32com\gen_py\783CD4E0-9D54-11CF-B8EE-00608CC9A71Fx0x5x1.py", 
line 179, in Open
    , Username, Password)
  File "C:\Python25\Lib\site-packages\win32com\client\__init__.py", line 
448, in _ApplyTypes_
    dispid, 0, wFlags, retType, argTypes, *args),
com_error: (-2147352567, 'Exception occurred.', (0, u'SourceSafe', 
u'Error loading resource string', u'ssusexp.hlp', -11250, -2147210254), 
None)
 >>>

I'm using pywin32 build 212 and Visual Sourcesafe 6.0c build 9447.

The \\Tsg-admin-01\VSS\ folder contains a SSUSEXP.CHM file but no 
SSUSEXP.HLP file.

Any suggestions on how to proceed would be much appreciated.

Thanks,
-- jv

From alecmunro at gmail.com  Thu Sep 11 00:23:26 2008
From: alecmunro at gmail.com (Alec Munro)
Date: Wed, 10 Sep 2008 19:23:26 -0300
Subject: [python-win32] Accessing other interfaces of a Dispatch created
	COM object
In-Reply-To: <48C82728.8060206@probo.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>
	<9819d58b0808280632x2b139645x278d5f0675dd46c7@mail.gmail.com>
	<9819d58b0808280633t13d89b77u433d55924cbb73a1@mail.gmail.com>
	<9819d58b0808280734o33af5721y741ca9b079d074af@mail.gmail.com>
	<00de01c9099c$b77731c0$26659540$@com.au>
	<9819d58b0808291047j3de5e9d8u7647b05e49a53f97@mail.gmail.com>
	<9819d58b0809020944v488ab4a2n5127c71e7c69424b@mail.gmail.com>
	<9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>
	<48C82728.8060206@probo.com>
Message-ID: <9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>

Thanks for the clarification Tim.

My root problem is still the "Library not registered" issue. As far as
how it was registered:

 * I maintain a dll that defines a class that implements I1 and I2. I
use regsvr to register this dll.
 * I1 and I2 come from a third party, and in order for my dll to work
properly, it has to implement them (actually I2 is optional).
 * I don't know how I1 and I2 are registered. I do know that if I use
the COM browser, I can go to Registered Type Libraries, find one with
the typelibname that is associated with my CastTo created objects,
which has "Type Library" underneath it, and underneath that are
entries for I1, I2, several related interfaces, and a variety of
coclasses. Like this:
 - TypeLibName 1.0 TypeLibrary
 -- IID: {00..00}
 -- Type Library
 --- I1 - Dispatch
 --- I2 - Dispatch
 --- IX - Dispatch
 --- CX - CoClass

I've just spent some time hunting through the registry to determine
the differences between I1 and I2.

Looking the gen_py files, they have the same package CLSID, and the
same class coclass_clsid. But their class CLSIDs differ.

So I searched through the registry for those CLSIDs. They show up in
the same places, once each under HKEY_CLASSES_ROOT and
HKEY_LOCAL_MACHINE. Under each of them there is a "TypeLib" folder,
and within that, a "default" entry. The data associated with this
entry differs. In the case of I2, it's the same as the IID of the Type
Library I found in the COM browser. In the case of I1, it's another
GUID, but I don't know what it corresponds to.

Hopefully that sheds some light on the situation (or at worst, doesn't
confuse things further).

Thanks again,

Alec




On Wed, Sep 10, 2008 at 4:59 PM, Tim Roberts <timr at probo.com> wrote:
> Alec Munro wrote:
>> Ok, perhaps this will be useful.
>>
>> (Neither of these examples include the core arguments of the methods,
>> just the ones specific to InvokeTypes)
>>
>> What I call a method from I1, it calls _oleobj_.InvokeTypes with the
>> following arguments:
>>
>> 18, 0x0, 1, (24, 0), ((8, 1),)
>>
>> When I use the method from I2, it calls _ApplyTypes_(), which calls
>> _oleobj_.InvokeTypes with the following arguments:
>>
>> 1, 0x0, 1, (24, 0), ((3, 1), (12, 1), (16396, 2))
>>
>> Now, from my reading (of old messages by Mark :) ), the last two
>> parameters specify method return type and argument type.
>
> Well, it's the 4th argument that gives the return type.  24 is VT_VOID.
> That's unusual for a COM type.
>
>> In this case,
>> it seems like both methods return VOID, which seems fine. I'm a little
>> more concerned about the argument type. The first method is supposed
>> to take a single string, so that looks fine.
>
> Right; 8 is VT_BSTR, which is the typical COM Unicope string type.
>
>> The second method takes
>> an integer, a VT_TYPE, and a variant. I don't know if the
>> representation here accurately represents that.
>>
>
> 3 is VT_I4 (long), 12 is VT_VARIANT, and 16396 is a VT_VARIANT passed by
> reference, meaning an output parameter.
>
>> So, I'm fairly certain this puts that problem at something that
>> happens in InvokeTypes, unrelated to the arguments I passed in. I get
>> the impression that InvokeTypes is something of Microsofts, and
>> there's isn't really any way to observe what it's doing?
>>
>
> I've lost track.  What's the root problem?  Is it still  the "Library
> not registered" issue when you call a method in Interface_2?  Was this
> all registered through a type library, or through a regsvr32 call, or what?
>
> --
> Tim Roberts, timr at probo.com
> Providenza & Boekelheide, Inc.
>
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32
>

From mhammond at skippinet.com.au  Thu Sep 11 01:29:02 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Thu, 11 Sep 2008 09:29:02 +1000
Subject: [python-win32] suggested change to pywintypes.py for
	non-admininstall
In-Reply-To: <ga96qs$109$1@ger.gmane.org>
References: <014301c761fe$8ad76d70$070a0a0a@enfoldsystems.local>
	<45F19959.20906@activestate.com> <ga96qs$109$1@ger.gmane.org>
Message-ID: <057001c9139d$075b0a90$16111fb0$@com.au>

I'm afraid I've missed most of the context here:

> I have come across this issue, and am equally unsatisfied with the
> solutions I've come up with so far. all of them require the user to
> know
> too much about what's going on, and does not make the import "just
> work".
> 
> what if we added a python wrapper module (.py) around the native (.pyd)
> modules? it would contain something like the following:
> 
> import pywintypes
> import pythoncom
> from _win32ui import *

In what cases is it necessary to explicitly import pywintypes before
win32ui?

Mark


From timr at probo.com  Thu Sep 11 01:33:50 2008
From: timr at probo.com (Tim Roberts)
Date: Wed, 10 Sep 2008 16:33:50 -0700
Subject: [python-win32] Accessing other interfaces of a Dispatch created
 COM object
In-Reply-To: <9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>	
	<9819d58b0808280632x2b139645x278d5f0675dd46c7@mail.gmail.com>	
	<9819d58b0808280633t13d89b77u433d55924cbb73a1@mail.gmail.com>	
	<9819d58b0808280734o33af5721y741ca9b079d074af@mail.gmail.com>	
	<00de01c9099c$b77731c0$26659540$@com.au>	
	<9819d58b0808291047j3de5e9d8u7647b05e49a53f97@mail.gmail.com>	
	<9819d58b0809020944v488ab4a2n5127c71e7c69424b@mail.gmail.com>	
	<9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>	
	<48C82728.8060206@probo.com>
	<9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>
Message-ID: <48C8595E.5040906@probo.com>

Alec Munro wrote:
> Thanks for the clarification Tim.
>
> My root problem is still the "Library not registered" issue. As far as
> how it was registered:
>
>  * I maintain a dll that defines a class that implements I1 and I2. I
> use regsvr to register this dll.
>  * I1 and I2 come from a third party, and in order for my dll to work
> properly, it has to implement them (actually I2 is optional).
>  * I don't know how I1 and I2 are registered. I do know that if I use
> the COM browser, I can go to Registered Type Libraries, find one with
> the typelibname that is associated with my CastTo created objects,
> which has "Type Library" underneath it, and underneath that are
> entries for I1, I2, several related interfaces, and a variety of
> coclasses. Like this:
>  - TypeLibName 1.0 TypeLibrary
>  -- IID: {00..00}
>  -- Type Library
>  --- I1 - Dispatch
>  --- I2 - Dispatch
>  --- IX - Dispatch
>  --- CX - CoClass
>
> I've just spent some time hunting through the registry to determine
> the differences between I1 and I2.
>
> Looking the gen_py files, they have the same package CLSID, and the
> same class coclass_clsid. But their class CLSIDs differ.
>   

Well, we may be getting confused about identifiers here.  The IID listed
below the typename is the type library UUID.  It identifies the type
library, which is just a collection of definitions.  Within the type
library, you have interfaces, and you have coclasses.  An interface also
has an IID, and that's the important IID to use when you query for an
interface.  A coclass is a concrete object that implements one or more
of the interfaces -- the "default" object for those interfaces, if you
like.  A coclass is identified by a CLSID.

In gen_py, the "CLSID = IID(...)" line at the top of the file should be
the type library UUID.  The "CLSID = IID(...)" line at the top of each
class is the IID for the interface being exposed by this coclass.  The
coclass_clsid is the model coclass that implements the interface.  If
you try to create an object of this type, PythonCom will essentially do
a CoCreateObject on the coclass_clsid, using the IID of the interface.

So, your two interface should have different CLSID lines at the top of
the class, and you should find those IIDs listed in
HKEY_CLASSES_ROOT\Interface.

> So I searched through the registry for those CLSIDs. They show up in
> the same places, once each under HKEY_CLASSES_ROOT and
> HKEY_LOCAL_MACHINE.

HKEY_CLASSES_ROOT is just a "symbolic link" shortcut to
HKEY_LOCAL_MACHINE\Software\Classes.  Everything in one will be in the
other.

No solution yet, but maybe this provides more background information.

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From timr at probo.com  Thu Sep 11 01:36:47 2008
From: timr at probo.com (Tim Roberts)
Date: Wed, 10 Sep 2008 16:36:47 -0700
Subject: [python-win32] Accessing other interfaces of a Dispatch created
 COM object
In-Reply-To: <9819d58b0809101553i62f21843g9394acabc31d6716@mail.gmail.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>	
	<9819d58b0808280632x2b139645x278d5f0675dd46c7@mail.gmail.com>	
	<9819d58b0808280633t13d89b77u433d55924cbb73a1@mail.gmail.com>	
	<9819d58b0808280734o33af5721y741ca9b079d074af@mail.gmail.com>	
	<00de01c9099c$b77731c0$26659540$@com.au>	
	<9819d58b0808291047j3de5e9d8u7647b05e49a53f97@mail.gmail.com>	
	<9819d58b0809020944v488ab4a2n5127c71e7c69424b@mail.gmail.com>	
	<9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>	
	<48C82728.8060206@probo.com>	
	<9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>
	<9819d58b0809101553i62f21843g9394acabc31d6716@mail.gmail.com>
Message-ID: <48C85A0F.6030706@probo.com>

Alec Munro wrote:
> I've searched through the registry for the GUIDs I found associated
> with the CLSIDs of the classes. They seem to be IIDs of type
> libraries.
>
> The first one, associated with I1, seems to refer to a _basicUHI.tlb
> file that was distributed with SDK the third-party provides for
> creating the DLL we use.
>
> The second one, associated with I2, refers to a .dll file distributed
> with the third-parties application (the one that makes use of our
> DLL).
>
> In Visual Studio, if I look at the definition of I2, it opens up a
> _basicUHI.tlh file in a temp directory. Am I correct in assuming the
> .tlh file is automatically generated by VS to display a representation
> of the .tlb file?
>
> I found regtlibv12, and ran it against _basicUHI.tlb. Now the COM
> Browser has a "TypeLibName 2.0 Type Library" entry, but if I try to
> access it, or run makepy against it, I get "_basicUHI.tlb can not be
> loaded" or "Error loading type library/DLL.", respectively.
>
> This feels like it must be the source of the problem, but I'm not
> really sure what to do next.
>   

Take the IID of the type library for _basicUHI.tlb.  Now, look that up
in HKEY_CLASSES_ROOT\TypeLib.  Somewhere in there, there should be a
path to the TLB.  Make sure the path is correct.

You might also try running regsvr32 on the DLL file associated with I2.

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From rblazecka at radical.ca  Thu Sep 11 01:55:42 2008
From: rblazecka at radical.ca (Ryan Blazecka)
Date: Wed, 10 Sep 2008 16:55:42 -0700
Subject: [python-win32] suggested change to pywintypes.py for
	non-admininstall
In-Reply-To: <057001c9139d$075b0a90$16111fb0$@com.au>
References: <014301c761fe$8ad76d70$070a0a0a@enfoldsystems.local>	<45F19959.20906@activestate.com>
	<ga96qs$109$1@ger.gmane.org>
	<057001c9139d$075b0a90$16111fb0$@com.au>
Message-ID: <48C85E7E.60401@radical.ca>

oops, sorry, I suppose I shouldn't be deleting so much of the message 
history. ;)

the specific circumstance that I've come across is with a non-admin 
install of the win32 module. if I run python from the python directory 
(so that the current directory is the directory where pywintypes25.dll 
lives), then everything works fine.

if I run python from some other directory (ie my application directory), 
imports of the win32 modules (win32gui, win32ui, etc) fail unless I 
first import pywintypes.

Mark Hammond wrote:
> I'm afraid I've missed most of the context here:
> 
>> I have come across this issue, and am equally unsatisfied with the
>> solutions I've come up with so far. all of them require the user to
>> know
>> too much about what's going on, and does not make the import "just
>> work".
>>
>> what if we added a python wrapper module (.py) around the native (.pyd)
>> modules? it would contain something like the following:
>>
>> import pywintypes
>> import pythoncom
>> from _win32ui import *
> 
> In what cases is it necessary to explicitly import pywintypes before
> win32ui?
> 
> Mark


From mhammond at skippinet.com.au  Thu Sep 11 01:59:35 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Thu, 11 Sep 2008 09:59:35 +1000
Subject: [python-win32] Accessing other interfaces of a Dispatch created
	COM object
In-Reply-To: <9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>
	<9819d58b0808280632x2b139645x278d5f0675dd46c7@mail.gmail.com>
	<9819d58b0808280633t13d89b77u433d55924cbb73a1@mail.gmail.com>
	<9819d58b0808280734o33af5721y741ca9b079d074af@mail.gmail.com>
	<00de01c9099c$b77731c0$26659540$@com.au>
	<9819d58b0808291047j3de5e9d8u7647b05e49a53f97@mail.gmail.com>
	<9819d58b0809020944v488ab4a2n5127c71e7c69424b@mail.gmail.com>
	<9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>
	<48C82728.8060206@probo.com>
	<9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>
Message-ID: <057d01c913a1$4e80dc20$eb829460$@com.au>

I'm still a little lost too, but:

>  * I maintain a dll that defines a class that implements I1 and I2. I
> use regsvr to register this dll.
>  * I1 and I2 come from a third party, and in order for my dll to work
> properly, it has to implement them (actually I2 is optional).
>  * I don't know how I1 and I2 are registered.

I don't understand the above.  If you maintain the class that implements I1
and I2 how can:

* they come from a 3rd party - either you maintain them, or they do?
* Surely you know how they are registered?


> I do know that if I use
> the COM browser, I can go to Registered Type Libraries, find one with
> the typelibname that is associated with my CastTo created objects,
> which has "Type Library" underneath it, and underneath that are
> entries for I1, I2, several related interfaces, and a variety of
> coclasses. Like this:
>  - TypeLibName 1.0 TypeLibrary
>  -- IID: {00..00}
>  -- Type Library
>  --- I1 - Dispatch
>  --- I2 - Dispatch
>  --- IX - Dispatch
>  --- CX - CoClass
> 
> I've just spent some time hunting through the registry to determine
> the differences between I1 and I2.

What I'm guessing is that the GetTypeInfo() and GetTypeLib() calls for the
other dispatch object are failing.  Below is how we expect things to work,
demonstratrated using for Internet Explorer:

>>> from win32com.client import Dispatch
>>> d=Dispatch("InternetExplorer.Application")
>>> d=d._oleobj_
>>> d.GetTypeInfoCount()
1
>>> d.GetTypeInfo(0)
<PyITypeInfo at 0x00D1B038 with obj at 0x0025285C>
>>> i=d.GetTypeInfo(0)
>>> tlb, index=i.GetContainingTypeLib()
>>> tlb.GetLibAttr()
(IID('{EAB22AC0-30C1-11CF-A7EB-0000C05BAE0B}'), 0, 1, 1, 1, 8)
>>>

Which is the GUID, version, locale etc information (a TLIBATTR structure)
for the typelib, and you will find this in the registry.  It sounds like
this works for one of your objects and fails for the other (well - "fails"
might mean that the call works, but returns a GID and version info for a
typelib that isn't registered - ie, attempting to load that typelib would
fail.)  ie, you would also need the following to work:

>>> guid, lcid, syskind, majver, minver, flags = tlb.GetLibAttr()
>>> import pythoncom
>>> pythoncom.LoadRegTypeLib(guid, majver, minver, lcid)
<PyITypeLib at 0x00D1B038 with obj at 0x00254530>
>>> _.GetDocumentation(1)
(u'DWebBrowserEvents', u'Web Browser Control Events (old)', 0, None)
>>> 

Hope this helps,

Mark


From alecmunro at gmail.com  Fri Sep 12 19:28:57 2008
From: alecmunro at gmail.com (Alec Munro)
Date: Fri, 12 Sep 2008 14:28:57 -0300
Subject: [python-win32] Accessing other interfaces of a Dispatch created
	COM object
In-Reply-To: <9819d58b0809120949v217de44bn8c71ac792dec70fa@mail.gmail.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>
	<00de01c9099c$b77731c0$26659540$@com.au>
	<9819d58b0808291047j3de5e9d8u7647b05e49a53f97@mail.gmail.com>
	<9819d58b0809020944v488ab4a2n5127c71e7c69424b@mail.gmail.com>
	<9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>
	<48C82728.8060206@probo.com>
	<9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>
	<057d01c913a1$4e80dc20$eb829460$@com.au>
	<9819d58b0809101835h1a581fpb67305217b39a40f@mail.gmail.com>
	<9819d58b0809120949v217de44bn8c71ac792dec70fa@mail.gmail.com>
Message-ID: <9819d58b0809121028w46cf0267l3c517a33289ea595@mail.gmail.com>

(I just noticed my last two responses only went to Mark. They were
intended for the Mailing List and Tim as well. Sorry for the
oversight.)

So I just ran some tests, fruitlessly. Here is what I did:

 - Removed all gen_py stuff
 - Set registry entries for both interfaces to point to the BASICUHI typelib.
 - Restarted
 - Use Make_py on BASICUHI typelib

 X: Same error

 - Use Make_py for other typelib

 X: Same error

 - Removed all gen_py stuff
 - Set registry entries for both interfaces to point to other typelib.
 - Restarted
 - Use Make_py for other typelib

 X: Same error

 - Use Make_py for BASICUHI typelib

 X: Same error

What I'm wondering at this point is how I determine what Library it is
that isn't registered. That might yield some clues, but right now I'm
pretty much out of ideas, and toying seriously with the idea of
rewriting the dll to allow it to export the interfaces I need through
IDispatch.

Thanks again for your time,

Alec


On Fri, Sep 12, 2008 at 1:49 PM, Alec Munro <alecmunro at gmail.com> wrote:
> Hi Mark,
>
> Ok, with my first object, everything works as you demonstrated. With
> my second object (the first one CastTo(,I2)), I get the "Library not
> Registered" message when I call GetTypeInfo(), which I guess is what
> you were expecting.
>
> As far as Tim's question goes, I looked up the ID for the BasicUHI
> typelib in the registry, and the path there does indeed point to the
> BasicUHI file.
>
> I've got some ideas for things I can test, for the moment, but if you
> have any ideas or additional information, I would love to hear it.
>
> Thanks again!
>
> Alec
>
>
> On Wed, Sep 10, 2008 at 10:35 PM, Alec Munro <alecmunro at gmail.com> wrote:
>> Thanks Mark, Tim!
>>
>> Lots of excellent information there for someone as COM-confused as I am. :)
>>
>> I'll try out your suggestions tomorrow, to see what I find, but for
>> now, here's a little more information on our situation.
>>
>> We use a third-party solution for testing our products. This solution
>> requires us to create a COM dll with a class implementing specific
>> interfaces (at least I1, any others are optional). The third party
>> defines the interfaces, we just provide implementations for them. The
>> third party provides us with their testing solution itself, as well as
>> an SDK for developing the dll.
>>
>> In order to use the dll, we do register it manually using regsvr32.
>> Any other registrations I would assume are performed by the installers
>> from the third party.
>>
>> We've been using this for several years now (I only took over
>> maintenance on the project 6 months ago), but this is the first time
>> we are trying to access the dll ourselves, so it's fairly new
>> territory.
>>
>> Hopefully that makes sense, thanks again,
>>
>> Alec
>>
>> On Wed, Sep 10, 2008 at 8:59 PM, Mark Hammond <mhammond at skippinet.com.au> wrote:
>>> I'm still a little lost too, but:
>>>
>>>>  * I maintain a dll that defines a class that implements I1 and I2. I
>>>> use regsvr to register this dll.
>>>>  * I1 and I2 come from a third party, and in order for my dll to work
>>>> properly, it has to implement them (actually I2 is optional).
>>>>  * I don't know how I1 and I2 are registered.
>>>
>>> I don't understand the above.  If you maintain the class that implements I1
>>> and I2 how can:
>>>
>>> * they come from a 3rd party - either you maintain them, or they do?
>>> * Surely you know how they are registered?
>>>
>>>
>>>> I do know that if I use
>>>> the COM browser, I can go to Registered Type Libraries, find one with
>>>> the typelibname that is associated with my CastTo created objects,
>>>> which has "Type Library" underneath it, and underneath that are
>>>> entries for I1, I2, several related interfaces, and a variety of
>>>> coclasses. Like this:
>>>>  - TypeLibName 1.0 TypeLibrary
>>>>  -- IID: {00..00}
>>>>  -- Type Library
>>>>  --- I1 - Dispatch
>>>>  --- I2 - Dispatch
>>>>  --- IX - Dispatch
>>>>  --- CX - CoClass
>>>>
>>>> I've just spent some time hunting through the registry to determine
>>>> the differences between I1 and I2.
>>>
>>> What I'm guessing is that the GetTypeInfo() and GetTypeLib() calls for the
>>> other dispatch object are failing.  Below is how we expect things to work,
>>> demonstratrated using for Internet Explorer:
>>>
>>>>>> from win32com.client import Dispatch
>>>>>> d=Dispatch("InternetExplorer.Application")
>>>>>> d=d._oleobj_
>>>>>> d.GetTypeInfoCount()
>>> 1
>>>>>> d.GetTypeInfo(0)
>>> <PyITypeInfo at 0x00D1B038 with obj at 0x0025285C>
>>>>>> i=d.GetTypeInfo(0)
>>>>>> tlb, index=i.GetContainingTypeLib()
>>>>>> tlb.GetLibAttr()
>>> (IID('{EAB22AC0-30C1-11CF-A7EB-0000C05BAE0B}'), 0, 1, 1, 1, 8)
>>>>>>
>>>
>>> Which is the GUID, version, locale etc information (a TLIBATTR structure)
>>> for the typelib, and you will find this in the registry.  It sounds like
>>> this works for one of your objects and fails for the other (well - "fails"
>>> might mean that the call works, but returns a GID and version info for a
>>> typelib that isn't registered - ie, attempting to load that typelib would
>>> fail.)  ie, you would also need the following to work:
>>>
>>>>>> guid, lcid, syskind, majver, minver, flags = tlb.GetLibAttr()
>>>>>> import pythoncom
>>>>>> pythoncom.LoadRegTypeLib(guid, majver, minver, lcid)
>>> <PyITypeLib at 0x00D1B038 with obj at 0x00254530>
>>>>>> _.GetDocumentation(1)
>>> (u'DWebBrowserEvents', u'Web Browser Control Events (old)', 0, None)
>>>>>>
>>>
>>> Hope this helps,
>>>
>>> Mark
>>>
>>>
>>
>

From siegfried at heintze.com  Sun Sep 14 09:09:23 2008
From: siegfried at heintze.com (siegfried)
Date: Sun, 14 Sep 2008 00:09:23 -0700
Subject: [python-win32] Windows GUI programming with Python?
Message-ID: <9D6B1ED566694FD7859249DE6088DF3D@kingmark>

I just finished trying out the sample program scribble in "Python
Programming in Win32". While I

carefully cut and pasted from my soft copy of the book, I could not

create a new document. When I finally found the source code download

URL hidden in the chm file, the zip file was corrupted!

 

This program is more what I am looking for:
http://aspn.activestate.com/ASPN/Mail/Message/ActivePython/1775844

 

Unfortunately it does not work. Apparently I don't have to use
ActiveStatePython (according to a poster in the nntp group). Where can I
find similar source code that works 

that displays a simple window and lets me call other exotic window

functions like LoadKeyboardLayout, MapVirtualKeyEx, ToUnicodeEx. That

latter function produces Unicode strings and since I don't know how to
display

Cyrillic in a console mode app, I'm thinking I'll create a window to

display it in. I'll need to know how to load a font in python too ---

anyone got an example of that too?

 

 

 

thanks!

Siegfried

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080914/f9433f2c/attachment.htm>

From Jim.Vickroy at noaa.gov  Sun Sep 14 14:31:23 2008
From: Jim.Vickroy at noaa.gov (Jim Vickroy)
Date: Sun, 14 Sep 2008 06:31:23 -0600
Subject: [python-win32] Windows GUI programming with Python?
In-Reply-To: <9D6B1ED566694FD7859249DE6088DF3D@kingmark>
References: <9D6B1ED566694FD7859249DE6088DF3D@kingmark>
Message-ID: <48CD041B.7080108@noaa.gov>

siegfried wrote:
>
> I just finished trying out the sample program scribble in "Python 
> Programming in Win32". While I
>
> carefully cut and pasted from my soft copy of the book, I could not
>
> create a new document. When I finally found the source code download
>
> URL hidden in the chm file, the zip file was corrupted!
>
I've attached the 3 "scribble" files.

I'm not a GUI programmer so others will have to answer your questions.
>
>  
>
> This program is more what I am looking for: 
> http://aspn.activestate.com/ASPN/Mail/Message/ActivePython/1775844
>
>  
>
> Unfortunately it does not work.
>
Could you provide the code and the error message?
>
> Apparently I don't have to use ActiveStatePython (according to a 
> poster in the nntp group).
>
That is true.
>
> Where can I find similar source code that works
>
> that displays a simple window and lets me call other exotic window
>
> functions like LoadKeyboardLayout, MapVirtualKeyEx, ToUnicodeEx. That
>
> latter function produces Unicode strings and since I don't know how to 
> display
>
> Cyrillic in a console mode app, I'm thinking I'll create a window to
>
> display it in. I'll need to know how to load a font in python too ---
>
> anyone got an example of that too?
>
>  
>
>  
>
>  
>
> thanks!
>
> Siegfried
>
>  
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32
>   

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080914/17374dbc/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: scribbleApp.py
Type: application/x-python
Size: 353 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080914/17374dbc/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: scribble1.py
Type: application/x-python
Size: 1335 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080914/17374dbc/attachment-0001.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: scribble2.py
Type: application/x-python
Size: 4216 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080914/17374dbc/attachment-0002.bin>

From Jim.Vickroy at noaa.gov  Sun Sep 14 14:40:33 2008
From: Jim.Vickroy at noaa.gov (Jim Vickroy)
Date: Sun, 14 Sep 2008 06:40:33 -0600
Subject: [python-win32] Microsoft Visual Sourcesafe automation --
In-Reply-To: <48C82173.6010205@noaa.gov>
References: <48C82173.6010205@noaa.gov>
Message-ID: <48CD0641.3000303@noaa.gov>

Jim Vickroy wrote:
> Hello all,
>
> I am attempting to automate Sourcesafe access.
>
> Here is a sample session:
>
> >>> import sys
> >>> sys.version
> '2.5.2 (r252:60911, Feb 21 2008, 13:11:45) [MSC v.1310 32 bit (Intel)]'
> >>> sys.subversion
> ('CPython', 'tags/r252', '60911')
> >>> sys.platform
> 'win32'
> >>> sys.getwindowsversion()
> (5, 1, 2600, 2, 'Service Pack 2')
> >>> import win32com.client
> >>> vcs = win32com.client.Dispatch('SourceSafe')
> >>> config   = r'\\Tsg-admin-01\VSS\srcsafe.ini'
> >>> account  = 'jim.vickroy'
> >>> password = raw_input('password: ')
> >>> vcs.Open(SrcSafeIni=config, Username=account, Password=password)
> Traceback (most recent call last):
>  File "<interactive input>", line 1, in <module>
>  File 
> "C:\Python25\lib\site-packages\win32com\gen_py\783CD4E0-9D54-11CF-B8EE-00608CC9A71Fx0x5x1.py", 
> line 179, in Open
>    , Username, Password)
>  File "C:\Python25\Lib\site-packages\win32com\client\__init__.py", 
> line 448, in _ApplyTypes_
>    dispid, 0, wFlags, retType, argTypes, *args),
> com_error: (-2147352567, 'Exception occurred.', (0, u'SourceSafe', 
> u'Error loading resource string', u'ssusexp.hlp', -11250, 
> -2147210254), None)
> >>>
>
> I'm using pywin32 build 212 and Visual Sourcesafe 6.0c build 9447.
>
> The \\Tsg-admin-01\VSS\ folder contains a SSUSEXP.CHM file but no 
> SSUSEXP.HLP file.
>
> Any suggestions on how to proceed would be much appreciated.
>
> Thanks,
> -- jv
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32
I finally got this to work by deleting *vcs* which apparently had 
"somehow" become unusable.  So the following does work:

 >>> import win32com.client
 >>> vcs = win32com.client.Dispatch('SourceSafe')
 >>> config   = r'\\Tsg-admin-01\VSS\srcsafe.ini'
 >>> account  = 'jim.vickroy'
 >>> vcs.Open(SrcSafeIni=config, Username=account)
 >>> vcs.Username
u'Jim.vickroy'
 >>> vcs.CurrentProject
u'$/RTGM/data archival'

From werner.bruhin at free.fr  Mon Sep 15 09:45:36 2008
From: werner.bruhin at free.fr (Werner F. Bruhin)
Date: Mon, 15 Sep 2008 09:45:36 +0200
Subject: [python-win32] Windows GUI programming with Python?
In-Reply-To: <9D6B1ED566694FD7859249DE6088DF3D@kingmark>
References: <9D6B1ED566694FD7859249DE6088DF3D@kingmark>
Message-ID: <gal3r6$pgr$1@ger.gmane.org>

Siegfried,

siegfried wrote:
> I just finished trying out the sample program scribble in "Python 
> Programming in Win32". While I
> 
> carefully cut and pasted from my soft copy of the book, I could not
> 
> create a new document. When I finally found the source code download
> 
> URL hidden in the chm file, the zip file was corrupted!
> 
>  
> 
> This program is more what I am looking for: 
> http://aspn.activestate.com/ASPN/Mail/Message/ActivePython/1775844
> 
>  
> 
> Unfortunately it does not work. Apparently I don't have to use 
> ActiveStatePython (according to a poster in the nntp group). Where can I 
> find similar source code that works
> 
> that displays a simple window and lets me call other exotic window
> 
> functions like LoadKeyboardLayout, MapVirtualKeyEx, ToUnicodeEx. That
> 
> latter function produces Unicode strings and since I don't know how to 
> display
> 
> Cyrillic in a console mode app, I'm thinking I'll create a window to
> 
> display it in. I'll need to know how to load a font in python too ---
> 
> anyone got an example of that too?
You might want to look into wxPython.  It is a GUI library wrapping 
wxWidgets.

Make sure to download the demo and use the Unicode version.  In the demo 
look for the styledtextctrl_1 and/or the RichTextCtrl it should be close 
to what you would like to do.

Werner
> 
>  
> 
>  
> 
>  
> 
> thanks!
> 
> Siegfried
> 
>  
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32


From keston.obendorf at intel.com  Tue Sep 16 00:19:01 2008
From: keston.obendorf at intel.com (Obendorf, Keston)
Date: Mon, 15 Sep 2008 15:19:01 -0700
Subject: [python-win32] COM: Parameterless functions seen as properties
Message-ID: <12A5C15467D5B94F8E0FF265D9498ADD03AC5794@orsmsx419.amr.corp.intel.com>

I'm doing COM development in both Python and C#. Things are working
well, but I'm running into an issue where trying to get parameterless
function returns the return value of that function, which Python then
tries to call, causing an error.

 

 

The C# class is an implementation of an IDL-defined interface.

 

//IDL 

interface IService : IDispatch

{

      [id(101)] HRESULT foo([out, retval] BSTR * pstr);

}

 

//C#

[ComImport]

[Guid("/*GUID of IService*/")]

interface IService{}

 

 

[ComVisible(true)]

[ProgId("bar")]

[Guid("Valid GUID")]

public class bar : IService

{

    public bar(){}

 

    public string foo()

    {

        return "foo"; 

    }

}

 

Now, getting a Dispatch of "bar" in Python is where the issue lies.

 

from win32com.client import Dispatch

bar = Dispatch("bar")

str = bar.foo()

 

>>TypeError: 'unicode' object is not callable

 

This is only a problem with the C# class. If the interface is
implemented in Pythong, the above code will work.

 

Now, this issue can be fixed by adding in _FlagAsMethod before calling
bar.foo()

 

bar._FlagAsMethod("foo")

 

I really don't want to have to do that every time I try to call a
parameterless method of a C# COM-object from Python. That's a huge
headache, especially because I am trying to keep the language any one
component is written in a non-issue. 

 

I've started to dive into the source code of dynamic.py and other Client
modules to try and figure out how I can make this distinction at
Dispatch-time, but it's rough going understanding everything that's
going on in there. Does anyone have any pointers or a straight up
solution to this problem? 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080915/62f9931d/attachment.htm>

From mhammond at skippinet.com.au  Tue Sep 16 01:06:01 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Tue, 16 Sep 2008 09:06:01 +1000
Subject: [python-win32] COM: Parameterless functions seen as properties
In-Reply-To: <12A5C15467D5B94F8E0FF265D9498ADD03AC5794@orsmsx419.amr.corp.intel.com>
References: <12A5C15467D5B94F8E0FF265D9498ADD03AC5794@orsmsx419.amr.corp.intel.com>
Message-ID: <009601c91787$a25c94d0$e715be70$@com.au>

It looks like 2 things are going on:

 

* Your object isn't returning typelib info to Python when it asks.  If your
IDispatch returned something from GetTypeInfo() we could use, this wouldn't
happen.

 

* Without a typelib, Python can't tell if a reference to 'a.foo' is going to
end up being a method call or not (ie, from Python's POV, a.foo() is just a
property fetch for 'foo' and then calling the result.)  When making a
property reference, win32com checks for the ERRORS_BAD_CONTEXT values
defined in win32com\client\dynamic.py - if it sees one of those errors, it
assumes it really hit a method and handles it accordgingly.  But - in your
case, your object is succeeding!  The simplest thing to do here is to have
your object refuse to work as a property using one of the standard error
codes, and Python will then try it as a method.

 

Mark

 

From: python-win32-bounces+skippy.hammond=gmail.com at python.org
[mailto:python-win32-bounces+skippy.hammond=gmail.com at python.org] On Behalf
Of Obendorf, Keston
Sent: Tuesday, 16 September 2008 8:19 AM
To: python-win32 at python.org
Subject: [python-win32] COM: Parameterless functions seen as properties

 

I'm doing COM development in both Python and C#. Things are working well,
but I'm running into an issue where trying to get parameterless function
returns the return value of that function, which Python then tries to call,
causing an error.

 

 

The C# class is an implementation of an IDL-defined interface.

 

//IDL 

interface IService : IDispatch

{

      [id(101)] HRESULT foo([out, retval] BSTR * pstr);

}

 

//C#

[ComImport]

[Guid("/*GUID of IService*/")]

interface IService{}

 

 

[ComVisible(true)]

[ProgId("bar")]

[Guid("Valid GUID")]

public class bar : IService

{

    public bar(){}

 

    public string foo()

    {

        return "foo"; 

    }

}

 

Now, getting a Dispatch of "bar" in Python is where the issue lies.

 

from win32com.client import Dispatch

bar = Dispatch("bar")

str = bar.foo()

 

>>TypeError: 'unicode' object is not callable

 

This is only a problem with the C# class. If the interface is implemented in
Pythong, the above code will work.

 

Now, this issue can be fixed by adding in _FlagAsMethod before calling
bar.foo()

 

bar._FlagAsMethod("foo")

 

I really don't want to have to do that every time I try to call a
parameterless method of a C# COM-object from Python. That's a huge headache,
especially because I am trying to keep the language any one component is
written in a non-issue. 

 

I've started to dive into the source code of dynamic.py and other Client
modules to try and figure out how I can make this distinction at
Dispatch-time, but it's rough going understanding everything that's going on
in there. Does anyone have any pointers or a straight up solution to this
problem? 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080916/eaf323be/attachment-0001.htm>

From martin.prochnow at team6-kg.de  Tue Sep 16 15:10:09 2008
From: martin.prochnow at team6-kg.de (martin.prochnow at team6-kg.de)
Date: Tue, 16 Sep 2008 15:10:09 +0200
Subject: [python-win32] VARIANT as byref parameter
Message-ID: <CHILKAT-MID-54fc71c2-cf92-1476-5e1c-07b7949f223e@Entwicklung12.TM>

Hi,

I'm trying to call a COM-function which first parameter is a reference to a VARIANT. The COM object that contains this function supports early-bound automation over MakePy.

The corresponding C-code  is:

LONG lValue;
lValue = plQueryNote->GetVarSize();
LPBYTE buff = new BYTE[lValue];
VARIANT varData;
varData.vt = VT_BYREF;
varData.byref = buff;
plQueryNote->ReadVarBlock(&varData,0,lValue);

I have no idea how to define the VARIANT and how to pass it as byref-style-parameter in python with win32com. I found no documentation for a problem like this in the web. It would be great if someone could help me.

Regards,
Martin

From mkesper at fsfe.org  Tue Sep 16 16:23:30 2008
From: mkesper at fsfe.org (Michael Kesper)
Date: Tue, 16 Sep 2008 16:23:30 +0200
Subject: [python-win32] getting TerminalServicesProfilePath
In-Reply-To: <729c89ef0809090813u7b81a29by2276994e3c8d51a4@mail.gmail.com>
References: <729c89ef0809090813u7b81a29by2276994e3c8d51a4@mail.gmail.com>
Message-ID: <20080916142330.GA4535@localhost>

Hi,

* Etienne Pouliot <etiennepouliot at gmail.com> [2008-09-09 11:13:27 -0400]:
?
> Hi,
> 
>   I'm trying to get TerminalServicesProfilePath property on user account in
> my Active Directory.
...
> If I use another properties instead, like DisplayName, it works.
> 
> Any idea what I am doing wrong ?

I always found it easiest to use Tim Golden's active_directory module [1]
for accessing ADS. There you have the possibility to search for your
object and do a dump() on it to see all attributes it provides.

Best wishes
Michael

[1] http://tgolden.sc.sabren.com/python/active_directory.html
-- 
Free Software Foundation Europe (FSFE) []         (http://fsfeurope.org)
Join the Fellowship!                 [][][]       (http://fsfe.org/join)
Your donation powers our work!         ||  (http://fsfeurope.org/donate)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 315 bytes
Desc: Digital signature
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080916/505c5379/attachment.pgp>

From codingjester at gmail.com  Tue Sep 16 16:27:00 2008
From: codingjester at gmail.com (John Bunting)
Date: Tue, 16 Sep 2008 10:27:00 -0400
Subject: [python-win32] Adding an exe to startup...
Message-ID: <b146c4a40809160727l22fc182cs53fc9b9bcad8b48b@mail.gmail.com>

Hey everyone,

I've been having some trouble finding information on how to add an exe that
I have created to the startup of my windows box.  I would like to know how I
should go about adding the exe to the startup menu, without using the gui
interface.  I want to be able to automate this because I'm also going to be
giving out some of this code to different users and I don't want to have to
have them manually add the exe.

Thank alot everyone,


~John
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080916/d84c7597/attachment.htm>

From mail at timgolden.me.uk  Tue Sep 16 16:51:58 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Tue, 16 Sep 2008 15:51:58 +0100
Subject: [python-win32] Adding an exe to startup...
In-Reply-To: <b146c4a40809160727l22fc182cs53fc9b9bcad8b48b@mail.gmail.com>
References: <b146c4a40809160727l22fc182cs53fc9b9bcad8b48b@mail.gmail.com>
Message-ID: <48CFC80E.2070109@timgolden.me.uk>

John Bunting wrote:
> I've been having some trouble finding information on how to add an exe 
> that I have created to the startup of my windows box.  I would like to 
> know how I should go about adding the exe to the startup menu, without 
> using the gui interface.  I want to be able to automate this because I'm 
> also going to be giving out some of this code to different users and I 
> don't want to have to have them manually add the exe.


The startup menu is basically a folder which you can put
shortcuts into in the normal way. For simplicity's sake,
I recommend using (my) winshell module:

http://timgolden.me.uk/python/winshell.html

but it's only a very simple wrapper around the pywin32
extensions so you might prefer to roll your own. At
least you can look at the source. The following code
will do what you want (assuming you want Python to
start up automatically!)

<code>
import os, sys
import winshell

startup = winshell.startup () # use common=1 for all users
print startup

winshell.CreateShortcut (
  Path=os.path.join (winshell.startup (), "Python.lnk"),
  Target=sys.executable,
  Icon=(sys.executable, 0),
  Description="Python"
)

</code>

TJG

From mkesper at fsfe.org  Tue Sep 16 16:23:30 2008
From: mkesper at fsfe.org (Michael Kesper)
Date: Tue, 16 Sep 2008 16:23:30 +0200
Subject: [python-win32] getting TerminalServicesProfilePath
In-Reply-To: <729c89ef0809090813u7b81a29by2276994e3c8d51a4@mail.gmail.com>
References: <729c89ef0809090813u7b81a29by2276994e3c8d51a4@mail.gmail.com>
Message-ID: <20080916142330.GA4535@localhost>

Hi,

* Etienne Pouliot <etiennepouliot at gmail.com> [2008-09-09 11:13:27 -0400]:
?
> Hi,
> 
>   I'm trying to get TerminalServicesProfilePath property on user account in
> my Active Directory.
...
> If I use another properties instead, like DisplayName, it works.
> 
> Any idea what I am doing wrong ?

I always found it easiest to use Tim Golden's active_directory module [1]
for accessing ADS. There you have the possibility to search for your
object and do a dump() on it to see all attributes it provides.

Best wishes
Michael

[1] http://tgolden.sc.sabren.com/python/active_directory.html
-- 
Free Software Foundation Europe (FSFE) []         (http://fsfeurope.org)
Join the Fellowship!                 [][][]       (http://fsfe.org/join)
Your donation powers our work!         ||  (http://fsfeurope.org/donate)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 315 bytes
Desc: Digital signature
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080916/505c5379/attachment-0001.pgp>

From kevin.horn at gmail.com  Tue Sep 16 20:40:58 2008
From: kevin.horn at gmail.com (Kevin Horn)
Date: Tue, 16 Sep 2008 13:40:58 -0500
Subject: [python-win32] Adding an exe to startup...
In-Reply-To: <48CFC80E.2070109@timgolden.me.uk>
References: <b146c4a40809160727l22fc182cs53fc9b9bcad8b48b@mail.gmail.com>
	<48CFC80E.2070109@timgolden.me.uk>
Message-ID: <562bcc10809161140g1a29fa97kbe2c600c927fd085@mail.gmail.com>

There are a number of ways to do this:

1) add a shortcut to the "Startup start menu folder
2) add registry entries to various places like *
[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows\CurrentVersion\Run]
*    see here for some other locations:
http://www.doshelp.com/HowToView/Registry_Keys.htm
3) make it a service and set it to "Automatic" (probably more work than you
want to do)
4) There are other more esoteric ways, but I think the above probably covers
anything you are likely to want to do.

Kevin Horn

On Tue, Sep 16, 2008 at 9:51 AM, Tim Golden <mail at timgolden.me.uk> wrote:

> John Bunting wrote:
>
>> I've been having some trouble finding information on how to add an exe
>> that I have created to the startup of my windows box.  I would like to know
>> how I should go about adding the exe to the startup menu, without using the
>> gui interface.  I want to be able to automate this because I'm also going to
>> be giving out some of this code to different users and I don't want to have
>> to have them manually add the exe.
>>
>
>
> The startup menu is basically a folder which you can put
> shortcuts into in the normal way. For simplicity's sake,
> I recommend using (my) winshell module:
>
> http://timgolden.me.uk/python/winshell.html
>
> but it's only a very simple wrapper around the pywin32
> extensions so you might prefer to roll your own. At
> least you can look at the source. The following code
> will do what you want (assuming you want Python to
> start up automatically!)
>
> <code>
> import os, sys
> import winshell
>
> startup = winshell.startup () # use common=1 for all users
> print startup
>
> winshell.CreateShortcut (
>  Path=os.path.join (winshell.startup (), "Python.lnk"),
>  Target=sys.executable,
>  Icon=(sys.executable, 0),
>  Description="Python"
> )
>
> </code>
>
> TJG
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080916/360137d7/attachment.htm>

From timr at probo.com  Wed Sep 17 00:50:05 2008
From: timr at probo.com (Tim Roberts)
Date: Tue, 16 Sep 2008 15:50:05 -0700
Subject: [python-win32] VARIANT as byref parameter
In-Reply-To: <CHILKAT-MID-54fc71c2-cf92-1476-5e1c-07b7949f223e@Entwicklung12.TM>
References: <CHILKAT-MID-54fc71c2-cf92-1476-5e1c-07b7949f223e@Entwicklung12.TM>
Message-ID: <48D0381D.5010907@probo.com>

martin.prochnow at team6-kg.de wrote:
> I'm trying to call a COM-function which first parameter is a reference to a VARIANT. The COM object that contains this function supports early-bound automation over MakePy.
>
> The corresponding C-code  is:
>
> LONG lValue;
> lValue = plQueryNote->GetVarSize();
> LPBYTE buff = new BYTE[lValue];
> VARIANT varData;
> varData.vt = VT_BYREF;
> varData.byref = buff;
> plQueryNote->ReadVarBlock(&varData,0,lValue);
>
> I have no idea how to define the VARIANT and how to pass it as byref-style-parameter in python with win32com. I found no documentation for a problem like this in the web. It would be great if someone could help me.

Is the parameter declared as [out,retval] in the type library?  If so,
then Python should create the variant and return it to you:
    varData = QueryNote.ReadVarBlock( 0, lValue )

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From larry.bates at websafe.com  Wed Sep 17 02:31:39 2008
From: larry.bates at websafe.com (Larry Bates)
Date: Tue, 16 Sep 2008 19:31:39 -0500
Subject: [python-win32] Adding an exe to startup...
In-Reply-To: <b146c4a40809160727l22fc182cs53fc9b9bcad8b48b@mail.gmail.com>
References: <b146c4a40809160727l22fc182cs53fc9b9bcad8b48b@mail.gmail.com>
Message-ID: <gapj67$bn4$1@ger.gmane.org>

John Bunting wrote:
> Hey everyone,
> 
> I've been having some trouble finding information on how to add an exe 
> that I have created to the startup of my windows box.  I would like to 
> know how I should go about adding the exe to the startup menu, without 
> using the gui interface.  I want to be able to automate this because I'm 
> also going to be giving out some of this code to different users and I 
> don't want to have to have them manually add the exe.
> 
> Thank alot everyone,
> 
> 
> ~John
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32

Get Inno Setup and create installer that will do this for you.

-Larry


From nytrokiss at gmail.com  Wed Sep 17 05:48:56 2008
From: nytrokiss at gmail.com (James Matthews)
Date: Tue, 16 Sep 2008 20:48:56 -0700
Subject: [python-win32] Fwd: Modifying the system menu
In-Reply-To: <bfa45e21-fb5e-454d-a4c2-e7cc00919f38@p31g2000prf.googlegroups.com>
References: <bfa45e21-fb5e-454d-a4c2-e7cc00919f38@p31g2000prf.googlegroups.com>
Message-ID: <8a6b8e350809162048n32ccf943i281f80b79b185402@mail.gmail.com>

---------- Forwarded message ----------
From: <raj.indian.08 at gmail.com>
Date: Tue, Sep 16, 2008 at 8:31 PM
Subject: Modifying the system menu
To: python-list at python.org


Hi all,
   I am trying to modify the system menu of all the applications in
my machine.
   For example say - I am creating multiple desktops for windows  -
and I want to give every application the capability to be moved across
different desktops. So I wanted to modify the basic system menu list
to include the new options.

   The methods I could think of are -
1. Modify the basic windows set of system menu values (best option)
      or
2. Poll every few milliseconds and modify the system menu values of
the active window

For (1) I couldnt think of a way to do it. Is it even possible using
python?

For (2) I wrote the following code for basic testing:
   hwnd = win32gui.GetForegroundWindow()
   hw = win32gui.GetSystemMenu(hwnd, False)
   if hw != None:
       win32gui.AppendMenu(hw,win32con.MF_SEPARATOR,0,'-');

and it shows the following error: pywintypes.error: (1401,
'AppendMenu', 'Invalid menu handle.')

If anyone could help me out in both the options (1) and/or (2), I
would be very thankful.

Thank you very much in advance,
A python newbie
--
http://mail.python.org/mailman/listinfo/python-list



-- 
http://www.goldwatches.com/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080916/cd251a13/attachment.htm>

From pramod.vaidyanathan at gmail.com  Wed Sep 17 13:04:03 2008
From: pramod.vaidyanathan at gmail.com (Pramod Vaidyanathan)
Date: Wed, 17 Sep 2008 07:04:03 -0400
Subject: [python-win32] Win32 Python Extensions Outlook.Application Question
Message-ID: <54ee9f9e0809170404m456d2d64m6d66fbf0612339e1@mail.gmail.com>

I've been struggling with a problem for the past couple of days.  Can anyone
provide me some help or insight?

The problem comes down to this.  I have an email that I have received in
Microsoft Outlook that contains characters outside of the ascii set.  I was
able to use your library to traverse through my outlook folders and select
the appropriate emails etc.  There are characters in the email that are
ascii, extended ascii, and other.  Really, the characters outside of ascii
are all "fraction" characters.  I am trying to replace these characters with
their respective floating point numbers (so I can parse the email and
picking out the numbers I need etc.).  The problem I am having is when I
read the "item.Body" into body, the characters "\xe2\x85\x9b",
"\xe2\x85\x9c", "\xe2\x85\x9d", "\xe2\x85\x9d" become unknown characters
(question marks).  These characters have unicode equivalents of u215b,
u215c, u215d, and u215e.  They are the fractions 1/8, 3/8, 5/8, 7/8
respectively.

I have the same problem when I open the outlook email and "SaveAs" a text
file or use the win32 interface to do that.

Interestingly enough, when I copy and paste the email text into a text file
and save it, I can parse the text file just fine using Python and none of
the characters are lost.  I get the feeling that the problem is occurring
when I read the email's body with "item.Body", but I am not sure how I can
read it using a different encoding scheme.  I have included my code below.
I would be extremely grateful if someone could provide some insight as to
how I can get around this.

Thanks!

Pramod




import codecs, win32com.client,string,re


def parseEmailsWithFractions():
    # Create instance of Outlook
    o = win32com.client.Dispatch("Outlook.Application")
    mapi = o.GetNamespace("MAPI")

    folder=mapi.Folders["Folder_NAME"]

    numItems=folder.Items.Count+1
    SubjectNotFound=1


    for i in range(1,501):

        item = folder.Items[numItems-i]
        for attribute in attributes:
            subject=item.Subject
            body=item.Body

            disclaimerStart=body.find("------")

            if subject.find("SUBJECT_TO_FIND") >=0 and SubjectNotFound:
                print subject

                EncodedBody=body.encode("utf-8")

                print repr(EncodedBody)
#Replacing all fractional characters with decimals
                EncodedBody=EncodedBody.replace("\xc2\xbe",".75")
                EncodedBody=EncodedBody.replace("\xc2\xbc",".25")
                EncodedBody=EncodedBody.replace("\xc2\xbd",".5")
                EncodedBody=EncodedBody.replace("\xe2\x85\x9b",".125")
                EncodedBody=EncodedBody.replace("\xe2\x85\x9c",".375")
                EncodedBody=EncodedBody.replace("\xe2\x85\x9d",".625")
                EncodedBody=EncodedBody.replace("\xe2\x85\x9d",".875")
                print EncodedBody
                SubjectNotFound=0

parseEmailsWithFractions()
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080917/04875e41/attachment.htm>

From mail at timgolden.me.uk  Wed Sep 17 13:38:40 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Wed, 17 Sep 2008 12:38:40 +0100
Subject: [python-win32] Win32 Python Extensions Outlook.Application
	Question
In-Reply-To: <54ee9f9e0809170404m456d2d64m6d66fbf0612339e1@mail.gmail.com>
References: <54ee9f9e0809170404m456d2d64m6d66fbf0612339e1@mail.gmail.com>
Message-ID: <48D0EC40.8010905@timgolden.me.uk>

Pramod Vaidyanathan wrote:
> The problem comes down to this.  I have an email that I have received in 
> Microsoft Outlook that contains characters outside of the ascii set.  I 
> was able to use your library to traverse through my outlook folders and 
> select the appropriate emails etc.  There are characters in the email 
> that are ascii, extended ascii, and other.  The problem I 
> am having is when I read the "item.Body" into body, the characters 
> "\xe2\x85\x9b", "\xe2\x85\x9c", "\xe2\x85\x9d", "\xe2\x85\x9d" become 
> unknown characters (question marks).  These characters have unicode 
> equivalents of u215b, u215c, u215d, and u215e.  They are the fractions 
> 1/8, 3/8, 5/8, 7/8 respectively. 

I'm not sure whether you're confused or I am.
Just to clarify: the Body attribute of an Outlook
MailItem is returned to Python as a unicode object.

In my case (having sent myself an email containing the
characters you mention) it looks like this:

u'\u215b\u215c\u215d\u215e'

Exactly how these chars will be output will depend on your
console, locale settings etc. If you want to replace
those as decimals, you can simply do this, eg:

body = message.Body.replace (u"\u215b", u"0.125")

You don't need to encode it to anything unless
you have some other reason to do that.

TJG

From pramod.vaidyanathan at gmail.com  Wed Sep 17 13:41:59 2008
From: pramod.vaidyanathan at gmail.com (Pramod Vaidyanathan)
Date: Wed, 17 Sep 2008 07:41:59 -0400
Subject: [python-win32] Win32 Python Extensions Outlook.Application
	Question
In-Reply-To: <48D0EC40.8010905@timgolden.me.uk>
References: <54ee9f9e0809170404m456d2d64m6d66fbf0612339e1@mail.gmail.com>
	<48D0EC40.8010905@timgolden.me.uk>
Message-ID: <54ee9f9e0809170441m2cc988cai11359cad83b7c9f9@mail.gmail.com>

Thanks for your quick reply.  I'm probably confused.  I'll give that a try.


On Wed, Sep 17, 2008 at 7:38 AM, Tim Golden <mail at timgolden.me.uk> wrote:

> Pramod Vaidyanathan wrote:
>
>> The problem comes down to this.  I have an email that I have received in
>> Microsoft Outlook that contains characters outside of the ascii set.  I was
>> able to use your library to traverse through my outlook folders and select
>> the appropriate emails etc.  There are characters in the email that are
>> ascii, extended ascii, and other.  The problem I am having is when I read
>> the "item.Body" into body, the characters "\xe2\x85\x9b", "\xe2\x85\x9c",
>> "\xe2\x85\x9d", "\xe2\x85\x9d" become unknown characters (question marks).
>>  These characters have unicode equivalents of u215b, u215c, u215d, and
>> u215e.  They are the fractions 1/8, 3/8, 5/8, 7/8 respectively.
>>
>
> I'm not sure whether you're confused or I am.
> Just to clarify: the Body attribute of an Outlook
> MailItem is returned to Python as a unicode object.
>
> In my case (having sent myself an email containing the
> characters you mention) it looks like this:
>
> u'\u215b\u215c\u215d\u215e'
>
> Exactly how these chars will be output will depend on your
> console, locale settings etc. If you want to replace
> those as decimals, you can simply do this, eg:
>
> body = message.Body.replace (u"\u215b", u"0.125")
>
> You don't need to encode it to anything unless
> you have some other reason to do that.
>
> TJG
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080917/50170402/attachment.htm>

From mail at timgolden.me.uk  Wed Sep 17 15:58:36 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Wed, 17 Sep 2008 14:58:36 +0100
Subject: [python-win32] Win32 Python Extensions
	Outlook.ApplicationQuestion
In-Reply-To: <1FD8FD0C19027145B941FAB2BC40C7B30259362C@extxmb17.nam.nsroot.net>
References: <54ee9f9e0809170404m456d2d64m6d66fbf0612339e1@mail.gmail.com>	<48D0EC40.8010905@timgolden.me.uk>	<1FD8FD0C19027145B941FAB2BC40C7B302593624@extxmb17.nam.nsroot.net>	<48D10511.4060503@timgolden.me.uk>
	<1FD8FD0C19027145B941FAB2BC40C7B30259362C@extxmb17.nam.nsroot.net>
Message-ID: <48D10D0C.3020109@timgolden.me.uk>

Vaidyanathan, Pramod wrote:
> Thanks Tim,
> 
> Unfortunately my resutls are still as follows.  I still have question
> mark characters.  Is it possibly a microsoft outlook versioning issue?
> 
> u' \xbe \xbc \xbd \xbd \xbd \xbd\r\n\xbe \xbc \xbd \xbd \xbd \xbd
> \xbd\r\n\xbe \xbc\r\n\xbe \xbc
> \r\n
> \r\n? \r\n? \xbd ? \xbe\r\n'


Must admit I'm mystified, if that's the result of running the
code I gave. You're clearly getting a unicode object, so the
only reason there should be question mark characters in there
would be if the text contained question mark characters. You
haven't done any kind of encoding for unknown-replacement to
occur.

To confirm: that output above is the result of doing:

print repr (message.Body)

Anyone else got any ideas?

TJG

From mail at timgolden.me.uk  Wed Sep 17 16:22:58 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Wed, 17 Sep 2008 15:22:58 +0100
Subject: [python-win32] Win32 Python
	ExtensionsOutlook.ApplicationQuestion
In-Reply-To: <1FD8FD0C19027145B941FAB2BC40C7B302593632@extxmb17.nam.nsroot.net>
References: <54ee9f9e0809170404m456d2d64m6d66fbf0612339e1@mail.gmail.com>	<48D0EC40.8010905@timgolden.me.uk>	<1FD8FD0C19027145B941FAB2BC40C7B302593624@extxmb17.nam.nsroot.net>	<48D10511.4060503@timgolden.me.uk><1FD8FD0C19027145B941FAB2BC40C7B30259362C@extxmb17.nam.nsroot.net>	<48D10D0C.3020109@timgolden.me.uk>
	<1FD8FD0C19027145B941FAB2BC40C7B302593632@extxmb17.nam.nsroot.net>
Message-ID: <48D112C2.5020208@timgolden.me.uk>

Vaidyanathan, Pramod wrote:
> Yes,
> 
> Print repr (message.Body)
> 
> I am using Microsoft Outlook (2000).  Are you using outlook 2003 by
> chance?  

Yep.

> I just discovered that when I do the exact same thing in Outlook 2003,
> the problem disappears. I believe it's some sort of settings issue in my
> version of Outlook 2000.  You wouldn't know what that might be would
> you?  If not, I'm sure I can dig it up by comparing.

I don't know of anything, I'm afraid.

> Thanks again for all of your help - turned out to be something stupid as
> it often is.   

Glad to be of help. And thanks for keeping the list updated;
it often helps when trawling the archives looking for
similar problems to one's own.

TJG

From timr at probo.com  Wed Sep 17 20:07:34 2008
From: timr at probo.com (Tim Roberts)
Date: Wed, 17 Sep 2008 11:07:34 -0700
Subject: [python-win32] Fwd: Modifying the system menu
In-Reply-To: <8a6b8e350809162048n32ccf943i281f80b79b185402@mail.gmail.com>
References: <bfa45e21-fb5e-454d-a4c2-e7cc00919f38@p31g2000prf.googlegroups.com>
	<8a6b8e350809162048n32ccf943i281f80b79b185402@mail.gmail.com>
Message-ID: <48D14766.6050203@probo.com>

James Matthews wrote:
>
> Hi all,
>    I am trying to modify the system menu of all the applications in
> my machine.
>    For example say - I am creating multiple desktops for windows  -
> and I want to give every application the capability to be moved across
> different desktops. So I wanted to modify the basic system menu list
> to include the new options.
>
>    The methods I could think of are -
> 1. Modify the basic windows set of system menu values (best option)

There is no central repository where this is stored, so there's no place
to change.


> 2. Poll every few milliseconds and modify the system menu values of
> the active window

Ick  Windows don't come and go that often, and you only need to do this
once for each application.

The correct option is (3) install a Windows hook to watch for new
application to activate, and modify the menu in the hook.  You will have
to use a hook in order to catch the menu item being chosen anyway. 
There are Python modules to handle some kinds of hooks, but overall this
would be much easier in C.

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From timr at probo.com  Wed Sep 17 20:43:36 2008
From: timr at probo.com (Tim Roberts)
Date: Wed, 17 Sep 2008 11:43:36 -0700
Subject: [python-win32] VARIANT as byref parameter
In-Reply-To: <CHILKAT-MID-ab8ba63a-9749-63d8-73bb-9dbad913fc79@Entwicklung12.TM>
References: <CHILKAT-MID-ab8ba63a-9749-63d8-73bb-9dbad913fc79@Entwicklung12.TM>
Message-ID: <48D14FD8.8050207@probo.com>

martin.prochnow at team6-kg.de wrote:
>
> > Is the parameter declared as [out,retval] in the type library?
>
> This is how the function is declared in the file that MakePy has
> generated:
>
>  def ReadVarBlock(self, pbBuff=defaultNamedNotOptArg,
> dwOffset=defaultNamedNotOptArg, dwLen=defaultNamedNotOptArg):
>   return self._oleobj_.InvokeTypes(58, LCID, 1, (3, 0), ((16396, 0),
> (3, 0), (3, 0)),pbBuff
>    , dwOffset, dwLen)
>
> > If so, then Python should create the variant and return it to you:
> >    varData = QueryNote.ReadVarBlock( 0, lValue )
>
> This line produces a type mismatch error.
>

Well, that implies that the buffer is an in/out variant, not an output. 
I'm not sure how to construct a buffer Pythoncom will correctly
translate in this instance.  Mark, maybe?

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From alecmunro at gmail.com  Wed Sep 17 21:10:51 2008
From: alecmunro at gmail.com (Alec Munro)
Date: Wed, 17 Sep 2008 16:10:51 -0300
Subject: [python-win32] Accessing other interfaces of a Dispatch created
	COM object
In-Reply-To: <9819d58b0809121028w46cf0267l3c517a33289ea595@mail.gmail.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>
	<9819d58b0808291047j3de5e9d8u7647b05e49a53f97@mail.gmail.com>
	<9819d58b0809020944v488ab4a2n5127c71e7c69424b@mail.gmail.com>
	<9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>
	<48C82728.8060206@probo.com>
	<9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>
	<057d01c913a1$4e80dc20$eb829460$@com.au>
	<9819d58b0809101835h1a581fpb67305217b39a40f@mail.gmail.com>
	<9819d58b0809120949v217de44bn8c71ac792dec70fa@mail.gmail.com>
	<9819d58b0809121028w46cf0267l3c517a33289ea595@mail.gmail.com>
Message-ID: <9819d58b0809171210h159b3402t654bd3e011e4bb@mail.gmail.com>

I'm back.

I've taken a crack at rewriting the DLL to export the methods I need
through IDispatch. Unfortunately, what I discovered (and I'm probably
describing this in the wrong way), is that the class is declared
ATL_NO_VTABLE, and as far as I can tell, the third-party app relies on
the interfaces being exported through IDispatch. This means that the
methods suggested here won't work:

http://www.sellsbrothers.com/tools/multidisp/

Or at least the first two.

So I'm back to trying to figure out what might be going wrong in my
attempts to access the current implementation through Python.

I wonder, is there any way to log IDispatch operations? If so, that
might tell me what the third-party operation was doing differently.

I would be happy to hear any other suggestions.

Thanks,

Alec

On Fri, Sep 12, 2008 at 2:28 PM, Alec Munro <alecmunro at gmail.com> wrote:
> (I just noticed my last two responses only went to Mark. They were
> intended for the Mailing List and Tim as well. Sorry for the
> oversight.)
>
> So I just ran some tests, fruitlessly. Here is what I did:
>
>  - Removed all gen_py stuff
>  - Set registry entries for both interfaces to point to the BASICUHI typelib.
>  - Restarted
>  - Use Make_py on BASICUHI typelib
>
>  X: Same error
>
>  - Use Make_py for other typelib
>
>  X: Same error
>
>  - Removed all gen_py stuff
>  - Set registry entries for both interfaces to point to other typelib.
>  - Restarted
>  - Use Make_py for other typelib
>
>  X: Same error
>
>  - Use Make_py for BASICUHI typelib
>
>  X: Same error
>
> What I'm wondering at this point is how I determine what Library it is
> that isn't registered. That might yield some clues, but right now I'm
> pretty much out of ideas, and toying seriously with the idea of
> rewriting the dll to allow it to export the interfaces I need through
> IDispatch.
>
> Thanks again for your time,
>
> Alec
>
>
> On Fri, Sep 12, 2008 at 1:49 PM, Alec Munro <alecmunro at gmail.com> wrote:
>> Hi Mark,
>>
>> Ok, with my first object, everything works as you demonstrated. With
>> my second object (the first one CastTo(,I2)), I get the "Library not
>> Registered" message when I call GetTypeInfo(), which I guess is what
>> you were expecting.
>>
>> As far as Tim's question goes, I looked up the ID for the BasicUHI
>> typelib in the registry, and the path there does indeed point to the
>> BasicUHI file.
>>
>> I've got some ideas for things I can test, for the moment, but if you
>> have any ideas or additional information, I would love to hear it.
>>
>> Thanks again!
>>
>> Alec
>>
>>
>> On Wed, Sep 10, 2008 at 10:35 PM, Alec Munro <alecmunro at gmail.com> wrote:
>>> Thanks Mark, Tim!
>>>
>>> Lots of excellent information there for someone as COM-confused as I am. :)
>>>
>>> I'll try out your suggestions tomorrow, to see what I find, but for
>>> now, here's a little more information on our situation.
>>>
>>> We use a third-party solution for testing our products. This solution
>>> requires us to create a COM dll with a class implementing specific
>>> interfaces (at least I1, any others are optional). The third party
>>> defines the interfaces, we just provide implementations for them. The
>>> third party provides us with their testing solution itself, as well as
>>> an SDK for developing the dll.
>>>
>>> In order to use the dll, we do register it manually using regsvr32.
>>> Any other registrations I would assume are performed by the installers
>>> from the third party.
>>>
>>> We've been using this for several years now (I only took over
>>> maintenance on the project 6 months ago), but this is the first time
>>> we are trying to access the dll ourselves, so it's fairly new
>>> territory.
>>>
>>> Hopefully that makes sense, thanks again,
>>>
>>> Alec
>>>
>>> On Wed, Sep 10, 2008 at 8:59 PM, Mark Hammond <mhammond at skippinet.com.au> wrote:
>>>> I'm still a little lost too, but:
>>>>
>>>>>  * I maintain a dll that defines a class that implements I1 and I2. I
>>>>> use regsvr to register this dll.
>>>>>  * I1 and I2 come from a third party, and in order for my dll to work
>>>>> properly, it has to implement them (actually I2 is optional).
>>>>>  * I don't know how I1 and I2 are registered.
>>>>
>>>> I don't understand the above.  If you maintain the class that implements I1
>>>> and I2 how can:
>>>>
>>>> * they come from a 3rd party - either you maintain them, or they do?
>>>> * Surely you know how they are registered?
>>>>
>>>>
>>>>> I do know that if I use
>>>>> the COM browser, I can go to Registered Type Libraries, find one with
>>>>> the typelibname that is associated with my CastTo created objects,
>>>>> which has "Type Library" underneath it, and underneath that are
>>>>> entries for I1, I2, several related interfaces, and a variety of
>>>>> coclasses. Like this:
>>>>>  - TypeLibName 1.0 TypeLibrary
>>>>>  -- IID: {00..00}
>>>>>  -- Type Library
>>>>>  --- I1 - Dispatch
>>>>>  --- I2 - Dispatch
>>>>>  --- IX - Dispatch
>>>>>  --- CX - CoClass
>>>>>
>>>>> I've just spent some time hunting through the registry to determine
>>>>> the differences between I1 and I2.
>>>>
>>>> What I'm guessing is that the GetTypeInfo() and GetTypeLib() calls for the
>>>> other dispatch object are failing.  Below is how we expect things to work,
>>>> demonstratrated using for Internet Explorer:
>>>>
>>>>>>> from win32com.client import Dispatch
>>>>>>> d=Dispatch("InternetExplorer.Application")
>>>>>>> d=d._oleobj_
>>>>>>> d.GetTypeInfoCount()
>>>> 1
>>>>>>> d.GetTypeInfo(0)
>>>> <PyITypeInfo at 0x00D1B038 with obj at 0x0025285C>
>>>>>>> i=d.GetTypeInfo(0)
>>>>>>> tlb, index=i.GetContainingTypeLib()
>>>>>>> tlb.GetLibAttr()
>>>> (IID('{EAB22AC0-30C1-11CF-A7EB-0000C05BAE0B}'), 0, 1, 1, 1, 8)
>>>>>>>
>>>>
>>>> Which is the GUID, version, locale etc information (a TLIBATTR structure)
>>>> for the typelib, and you will find this in the registry.  It sounds like
>>>> this works for one of your objects and fails for the other (well - "fails"
>>>> might mean that the call works, but returns a GID and version info for a
>>>> typelib that isn't registered - ie, attempting to load that typelib would
>>>> fail.)  ie, you would also need the following to work:
>>>>
>>>>>>> guid, lcid, syskind, majver, minver, flags = tlb.GetLibAttr()
>>>>>>> import pythoncom
>>>>>>> pythoncom.LoadRegTypeLib(guid, majver, minver, lcid)
>>>> <PyITypeLib at 0x00D1B038 with obj at 0x00254530>
>>>>>>> _.GetDocumentation(1)
>>>> (u'DWebBrowserEvents', u'Web Browser Control Events (old)', 0, None)
>>>>>>>
>>>>
>>>> Hope this helps,
>>>>
>>>> Mark
>>>>
>>>>
>>>
>>
>

From alecmunro at gmail.com  Wed Sep 17 21:28:55 2008
From: alecmunro at gmail.com (Alec Munro)
Date: Wed, 17 Sep 2008 16:28:55 -0300
Subject: [python-win32] Accessing other interfaces of a Dispatch created
	COM object
In-Reply-To: <9819d58b0809171210h159b3402t654bd3e011e4bb@mail.gmail.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>
	<9819d58b0809020944v488ab4a2n5127c71e7c69424b@mail.gmail.com>
	<9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>
	<48C82728.8060206@probo.com>
	<9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>
	<057d01c913a1$4e80dc20$eb829460$@com.au>
	<9819d58b0809101835h1a581fpb67305217b39a40f@mail.gmail.com>
	<9819d58b0809120949v217de44bn8c71ac792dec70fa@mail.gmail.com>
	<9819d58b0809121028w46cf0267l3c517a33289ea595@mail.gmail.com>
	<9819d58b0809171210h159b3402t654bd3e011e4bb@mail.gmail.com>
Message-ID: <9819d58b0809171228m5f71c45fld2b8a62dec388d03@mail.gmail.com>

If you hear a loud slapping sound, that's my hand and forehead. :)

I just recalled that the third-party gives us a test application that
uses the COM component. Including source.

It's in C#, which is new to me, but looks pretty straightforward. I'm
going to figure out how they do, then see if I can do the same thing
through Python.

Sorry I didn't think of this sooner.

Alec

On Wed, Sep 17, 2008 at 4:10 PM, Alec Munro <alecmunro at gmail.com> wrote:
> I'm back.
>
> I've taken a crack at rewriting the DLL to export the methods I need
> through IDispatch. Unfortunately, what I discovered (and I'm probably
> describing this in the wrong way), is that the class is declared
> ATL_NO_VTABLE, and as far as I can tell, the third-party app relies on
> the interfaces being exported through IDispatch. This means that the
> methods suggested here won't work:
>
> http://www.sellsbrothers.com/tools/multidisp/
>
> Or at least the first two.
>
> So I'm back to trying to figure out what might be going wrong in my
> attempts to access the current implementation through Python.
>
> I wonder, is there any way to log IDispatch operations? If so, that
> might tell me what the third-party operation was doing differently.
>
> I would be happy to hear any other suggestions.
>
> Thanks,
>
> Alec
>
> On Fri, Sep 12, 2008 at 2:28 PM, Alec Munro <alecmunro at gmail.com> wrote:
>> (I just noticed my last two responses only went to Mark. They were
>> intended for the Mailing List and Tim as well. Sorry for the
>> oversight.)
>>
>> So I just ran some tests, fruitlessly. Here is what I did:
>>
>>  - Removed all gen_py stuff
>>  - Set registry entries for both interfaces to point to the BASICUHI typelib.
>>  - Restarted
>>  - Use Make_py on BASICUHI typelib
>>
>>  X: Same error
>>
>>  - Use Make_py for other typelib
>>
>>  X: Same error
>>
>>  - Removed all gen_py stuff
>>  - Set registry entries for both interfaces to point to other typelib.
>>  - Restarted
>>  - Use Make_py for other typelib
>>
>>  X: Same error
>>
>>  - Use Make_py for BASICUHI typelib
>>
>>  X: Same error
>>
>> What I'm wondering at this point is how I determine what Library it is
>> that isn't registered. That might yield some clues, but right now I'm
>> pretty much out of ideas, and toying seriously with the idea of
>> rewriting the dll to allow it to export the interfaces I need through
>> IDispatch.
>>
>> Thanks again for your time,
>>
>> Alec
>>
>>
>> On Fri, Sep 12, 2008 at 1:49 PM, Alec Munro <alecmunro at gmail.com> wrote:
>>> Hi Mark,
>>>
>>> Ok, with my first object, everything works as you demonstrated. With
>>> my second object (the first one CastTo(,I2)), I get the "Library not
>>> Registered" message when I call GetTypeInfo(), which I guess is what
>>> you were expecting.
>>>
>>> As far as Tim's question goes, I looked up the ID for the BasicUHI
>>> typelib in the registry, and the path there does indeed point to the
>>> BasicUHI file.
>>>
>>> I've got some ideas for things I can test, for the moment, but if you
>>> have any ideas or additional information, I would love to hear it.
>>>
>>> Thanks again!
>>>
>>> Alec
>>>
>>>
>>> On Wed, Sep 10, 2008 at 10:35 PM, Alec Munro <alecmunro at gmail.com> wrote:
>>>> Thanks Mark, Tim!
>>>>
>>>> Lots of excellent information there for someone as COM-confused as I am. :)
>>>>
>>>> I'll try out your suggestions tomorrow, to see what I find, but for
>>>> now, here's a little more information on our situation.
>>>>
>>>> We use a third-party solution for testing our products. This solution
>>>> requires us to create a COM dll with a class implementing specific
>>>> interfaces (at least I1, any others are optional). The third party
>>>> defines the interfaces, we just provide implementations for them. The
>>>> third party provides us with their testing solution itself, as well as
>>>> an SDK for developing the dll.
>>>>
>>>> In order to use the dll, we do register it manually using regsvr32.
>>>> Any other registrations I would assume are performed by the installers
>>>> from the third party.
>>>>
>>>> We've been using this for several years now (I only took over
>>>> maintenance on the project 6 months ago), but this is the first time
>>>> we are trying to access the dll ourselves, so it's fairly new
>>>> territory.
>>>>
>>>> Hopefully that makes sense, thanks again,
>>>>
>>>> Alec
>>>>
>>>> On Wed, Sep 10, 2008 at 8:59 PM, Mark Hammond <mhammond at skippinet.com.au> wrote:
>>>>> I'm still a little lost too, but:
>>>>>
>>>>>>  * I maintain a dll that defines a class that implements I1 and I2. I
>>>>>> use regsvr to register this dll.
>>>>>>  * I1 and I2 come from a third party, and in order for my dll to work
>>>>>> properly, it has to implement them (actually I2 is optional).
>>>>>>  * I don't know how I1 and I2 are registered.
>>>>>
>>>>> I don't understand the above.  If you maintain the class that implements I1
>>>>> and I2 how can:
>>>>>
>>>>> * they come from a 3rd party - either you maintain them, or they do?
>>>>> * Surely you know how they are registered?
>>>>>
>>>>>
>>>>>> I do know that if I use
>>>>>> the COM browser, I can go to Registered Type Libraries, find one with
>>>>>> the typelibname that is associated with my CastTo created objects,
>>>>>> which has "Type Library" underneath it, and underneath that are
>>>>>> entries for I1, I2, several related interfaces, and a variety of
>>>>>> coclasses. Like this:
>>>>>>  - TypeLibName 1.0 TypeLibrary
>>>>>>  -- IID: {00..00}
>>>>>>  -- Type Library
>>>>>>  --- I1 - Dispatch
>>>>>>  --- I2 - Dispatch
>>>>>>  --- IX - Dispatch
>>>>>>  --- CX - CoClass
>>>>>>
>>>>>> I've just spent some time hunting through the registry to determine
>>>>>> the differences between I1 and I2.
>>>>>
>>>>> What I'm guessing is that the GetTypeInfo() and GetTypeLib() calls for the
>>>>> other dispatch object are failing.  Below is how we expect things to work,
>>>>> demonstratrated using for Internet Explorer:
>>>>>
>>>>>>>> from win32com.client import Dispatch
>>>>>>>> d=Dispatch("InternetExplorer.Application")
>>>>>>>> d=d._oleobj_
>>>>>>>> d.GetTypeInfoCount()
>>>>> 1
>>>>>>>> d.GetTypeInfo(0)
>>>>> <PyITypeInfo at 0x00D1B038 with obj at 0x0025285C>
>>>>>>>> i=d.GetTypeInfo(0)
>>>>>>>> tlb, index=i.GetContainingTypeLib()
>>>>>>>> tlb.GetLibAttr()
>>>>> (IID('{EAB22AC0-30C1-11CF-A7EB-0000C05BAE0B}'), 0, 1, 1, 1, 8)
>>>>>>>>
>>>>>
>>>>> Which is the GUID, version, locale etc information (a TLIBATTR structure)
>>>>> for the typelib, and you will find this in the registry.  It sounds like
>>>>> this works for one of your objects and fails for the other (well - "fails"
>>>>> might mean that the call works, but returns a GID and version info for a
>>>>> typelib that isn't registered - ie, attempting to load that typelib would
>>>>> fail.)  ie, you would also need the following to work:
>>>>>
>>>>>>>> guid, lcid, syskind, majver, minver, flags = tlb.GetLibAttr()
>>>>>>>> import pythoncom
>>>>>>>> pythoncom.LoadRegTypeLib(guid, majver, minver, lcid)
>>>>> <PyITypeLib at 0x00D1B038 with obj at 0x00254530>
>>>>>>>> _.GetDocumentation(1)
>>>>> (u'DWebBrowserEvents', u'Web Browser Control Events (old)', 0, None)
>>>>>>>>
>>>>>
>>>>> Hope this helps,
>>>>>
>>>>> Mark
>>>>>
>>>>>
>>>>
>>>
>>
>

From alecmunro at gmail.com  Wed Sep 17 21:54:30 2008
From: alecmunro at gmail.com (Alec Munro)
Date: Wed, 17 Sep 2008 16:54:30 -0300
Subject: [python-win32] Accessing other interfaces of a Dispatch created
	COM object
In-Reply-To: <9819d58b0809171228m5f71c45fld2b8a62dec388d03@mail.gmail.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>
	<9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>
	<48C82728.8060206@probo.com>
	<9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>
	<057d01c913a1$4e80dc20$eb829460$@com.au>
	<9819d58b0809101835h1a581fpb67305217b39a40f@mail.gmail.com>
	<9819d58b0809120949v217de44bn8c71ac792dec70fa@mail.gmail.com>
	<9819d58b0809121028w46cf0267l3c517a33289ea595@mail.gmail.com>
	<9819d58b0809171210h159b3402t654bd3e011e4bb@mail.gmail.com>
	<9819d58b0809171228m5f71c45fld2b8a62dec388d03@mail.gmail.com>
Message-ID: <9819d58b0809171254n1c59cac0tc0c7a0dd45f266a2@mail.gmail.com>

I've taken a look, and I can't say I had an "aha" moment, or anything
like that, but perhaps I have learned something important.

Setting up the COM object seems very straightforward:

Type t = Type.GetTypeFromCLSID("{e04f970b-53ce-420a-86f8-55374677703d}")

component = (basic.I1)Activator.CreateInstance(t)

component_2 = component as I2

The ID referenced there is the ID of the coclass in the COM DLL.

I've looked into where it gets "basic" from, and the important
property seems to be:

Identity: {35F40DBC-6CB3-4EEC-843B-57BB52CA2A93}\1.0\0\tlbimp

This ID is the CLSID of the Type Library in _basicUHI.tlb

I'll see if I can reproduce this exactly in Python, and if I get any
different results.

Alec

On Wed, Sep 17, 2008 at 4:28 PM, Alec Munro <alecmunro at gmail.com> wrote:
> If you hear a loud slapping sound, that's my hand and forehead. :)
>
> I just recalled that the third-party gives us a test application that
> uses the COM component. Including source.
>
> It's in C#, which is new to me, but looks pretty straightforward. I'm
> going to figure out how they do, then see if I can do the same thing
> through Python.
>
> Sorry I didn't think of this sooner.
>
> Alec
>
> On Wed, Sep 17, 2008 at 4:10 PM, Alec Munro <alecmunro at gmail.com> wrote:
>> I'm back.
>>
>> I've taken a crack at rewriting the DLL to export the methods I need
>> through IDispatch. Unfortunately, what I discovered (and I'm probably
>> describing this in the wrong way), is that the class is declared
>> ATL_NO_VTABLE, and as far as I can tell, the third-party app relies on
>> the interfaces being exported through IDispatch. This means that the
>> methods suggested here won't work:
>>
>> http://www.sellsbrothers.com/tools/multidisp/
>>
>> Or at least the first two.
>>
>> So I'm back to trying to figure out what might be going wrong in my
>> attempts to access the current implementation through Python.
>>
>> I wonder, is there any way to log IDispatch operations? If so, that
>> might tell me what the third-party operation was doing differently.
>>
>> I would be happy to hear any other suggestions.
>>
>> Thanks,
>>
>> Alec
>>
>> On Fri, Sep 12, 2008 at 2:28 PM, Alec Munro <alecmunro at gmail.com> wrote:
>>> (I just noticed my last two responses only went to Mark. They were
>>> intended for the Mailing List and Tim as well. Sorry for the
>>> oversight.)
>>>
>>> So I just ran some tests, fruitlessly. Here is what I did:
>>>
>>>  - Removed all gen_py stuff
>>>  - Set registry entries for both interfaces to point to the BASICUHI typelib.
>>>  - Restarted
>>>  - Use Make_py on BASICUHI typelib
>>>
>>>  X: Same error
>>>
>>>  - Use Make_py for other typelib
>>>
>>>  X: Same error
>>>
>>>  - Removed all gen_py stuff
>>>  - Set registry entries for both interfaces to point to other typelib.
>>>  - Restarted
>>>  - Use Make_py for other typelib
>>>
>>>  X: Same error
>>>
>>>  - Use Make_py for BASICUHI typelib
>>>
>>>  X: Same error
>>>
>>> What I'm wondering at this point is how I determine what Library it is
>>> that isn't registered. That might yield some clues, but right now I'm
>>> pretty much out of ideas, and toying seriously with the idea of
>>> rewriting the dll to allow it to export the interfaces I need through
>>> IDispatch.
>>>
>>> Thanks again for your time,
>>>
>>> Alec
>>>
>>>
>>> On Fri, Sep 12, 2008 at 1:49 PM, Alec Munro <alecmunro at gmail.com> wrote:
>>>> Hi Mark,
>>>>
>>>> Ok, with my first object, everything works as you demonstrated. With
>>>> my second object (the first one CastTo(,I2)), I get the "Library not
>>>> Registered" message when I call GetTypeInfo(), which I guess is what
>>>> you were expecting.
>>>>
>>>> As far as Tim's question goes, I looked up the ID for the BasicUHI
>>>> typelib in the registry, and the path there does indeed point to the
>>>> BasicUHI file.
>>>>
>>>> I've got some ideas for things I can test, for the moment, but if you
>>>> have any ideas or additional information, I would love to hear it.
>>>>
>>>> Thanks again!
>>>>
>>>> Alec
>>>>
>>>>
>>>> On Wed, Sep 10, 2008 at 10:35 PM, Alec Munro <alecmunro at gmail.com> wrote:
>>>>> Thanks Mark, Tim!
>>>>>
>>>>> Lots of excellent information there for someone as COM-confused as I am. :)
>>>>>
>>>>> I'll try out your suggestions tomorrow, to see what I find, but for
>>>>> now, here's a little more information on our situation.
>>>>>
>>>>> We use a third-party solution for testing our products. This solution
>>>>> requires us to create a COM dll with a class implementing specific
>>>>> interfaces (at least I1, any others are optional). The third party
>>>>> defines the interfaces, we just provide implementations for them. The
>>>>> third party provides us with their testing solution itself, as well as
>>>>> an SDK for developing the dll.
>>>>>
>>>>> In order to use the dll, we do register it manually using regsvr32.
>>>>> Any other registrations I would assume are performed by the installers
>>>>> from the third party.
>>>>>
>>>>> We've been using this for several years now (I only took over
>>>>> maintenance on the project 6 months ago), but this is the first time
>>>>> we are trying to access the dll ourselves, so it's fairly new
>>>>> territory.
>>>>>
>>>>> Hopefully that makes sense, thanks again,
>>>>>
>>>>> Alec
>>>>>
>>>>> On Wed, Sep 10, 2008 at 8:59 PM, Mark Hammond <mhammond at skippinet.com.au> wrote:
>>>>>> I'm still a little lost too, but:
>>>>>>
>>>>>>>  * I maintain a dll that defines a class that implements I1 and I2. I
>>>>>>> use regsvr to register this dll.
>>>>>>>  * I1 and I2 come from a third party, and in order for my dll to work
>>>>>>> properly, it has to implement them (actually I2 is optional).
>>>>>>>  * I don't know how I1 and I2 are registered.
>>>>>>
>>>>>> I don't understand the above.  If you maintain the class that implements I1
>>>>>> and I2 how can:
>>>>>>
>>>>>> * they come from a 3rd party - either you maintain them, or they do?
>>>>>> * Surely you know how they are registered?
>>>>>>
>>>>>>
>>>>>>> I do know that if I use
>>>>>>> the COM browser, I can go to Registered Type Libraries, find one with
>>>>>>> the typelibname that is associated with my CastTo created objects,
>>>>>>> which has "Type Library" underneath it, and underneath that are
>>>>>>> entries for I1, I2, several related interfaces, and a variety of
>>>>>>> coclasses. Like this:
>>>>>>>  - TypeLibName 1.0 TypeLibrary
>>>>>>>  -- IID: {00..00}
>>>>>>>  -- Type Library
>>>>>>>  --- I1 - Dispatch
>>>>>>>  --- I2 - Dispatch
>>>>>>>  --- IX - Dispatch
>>>>>>>  --- CX - CoClass
>>>>>>>
>>>>>>> I've just spent some time hunting through the registry to determine
>>>>>>> the differences between I1 and I2.
>>>>>>
>>>>>> What I'm guessing is that the GetTypeInfo() and GetTypeLib() calls for the
>>>>>> other dispatch object are failing.  Below is how we expect things to work,
>>>>>> demonstratrated using for Internet Explorer:
>>>>>>
>>>>>>>>> from win32com.client import Dispatch
>>>>>>>>> d=Dispatch("InternetExplorer.Application")
>>>>>>>>> d=d._oleobj_
>>>>>>>>> d.GetTypeInfoCount()
>>>>>> 1
>>>>>>>>> d.GetTypeInfo(0)
>>>>>> <PyITypeInfo at 0x00D1B038 with obj at 0x0025285C>
>>>>>>>>> i=d.GetTypeInfo(0)
>>>>>>>>> tlb, index=i.GetContainingTypeLib()
>>>>>>>>> tlb.GetLibAttr()
>>>>>> (IID('{EAB22AC0-30C1-11CF-A7EB-0000C05BAE0B}'), 0, 1, 1, 1, 8)
>>>>>>>>>
>>>>>>
>>>>>> Which is the GUID, version, locale etc information (a TLIBATTR structure)
>>>>>> for the typelib, and you will find this in the registry.  It sounds like
>>>>>> this works for one of your objects and fails for the other (well - "fails"
>>>>>> might mean that the call works, but returns a GID and version info for a
>>>>>> typelib that isn't registered - ie, attempting to load that typelib would
>>>>>> fail.)  ie, you would also need the following to work:
>>>>>>
>>>>>>>>> guid, lcid, syskind, majver, minver, flags = tlb.GetLibAttr()
>>>>>>>>> import pythoncom
>>>>>>>>> pythoncom.LoadRegTypeLib(guid, majver, minver, lcid)
>>>>>> <PyITypeLib at 0x00D1B038 with obj at 0x00254530>
>>>>>>>>> _.GetDocumentation(1)
>>>>>> (u'DWebBrowserEvents', u'Web Browser Control Events (old)', 0, None)
>>>>>>>>>
>>>>>>
>>>>>> Hope this helps,
>>>>>>
>>>>>> Mark
>>>>>>
>>>>>>
>>>>>
>>>>
>>>
>>
>

From mhammond at skippinet.com.au  Thu Sep 18 02:14:51 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Thu, 18 Sep 2008 10:14:51 +1000
Subject: [python-win32] VARIANT as byref parameter
In-Reply-To: <48D14FD8.8050207@probo.com>
References: <CHILKAT-MID-ab8ba63a-9749-63d8-73bb-9dbad913fc79@Entwicklung12.TM>
	<48D14FD8.8050207@probo.com>
Message-ID: <01d101c91923$9591a630$c0b4f290$@com.au>

> Well, that implies that the buffer is an in/out variant, not an output.
> I'm not sure how to construct a buffer Pythoncom will correctly
> translate in this instance.  Mark, maybe?

I'm not sure - but I'm not even sure it is legal automation to do that:

> LONG lValue;
> lValue = plQueryNote->GetVarSize();
> LPBYTE buff = new BYTE[lValue];
> VARIANT varData;
> varData.vt = VT_BYREF;
> varData.byref = buff;
> plQueryNote->ReadVarBlock(&varData,0,lValue);

Isn't VT_BYREF a modifier?  It looks to me like the variant in this snippet
isn't legal.  What I'd really like to see is an example in VBScript/VB or
any other IDispatch aware language that isn't c++ ;)

Cheers,

MArk


From mhammond at skippinet.com.au  Thu Sep 18 02:35:03 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Thu, 18 Sep 2008 10:35:03 +1000
Subject: [python-win32] Win32 Python
	ExtensionsOutlook.ApplicationQuestion
In-Reply-To: <48D112C2.5020208@timgolden.me.uk>
References: <54ee9f9e0809170404m456d2d64m6d66fbf0612339e1@mail.gmail.com>	<48D0EC40.8010905@timgolden.me.uk>	<1FD8FD0C19027145B941FAB2BC40C7B302593624@extxmb17.nam.nsroot.net>	<48D10511.4060503@timgolden.me.uk><1FD8FD0C19027145B941FAB2BC40C7B30259362C@extxmb17.nam.nsroot.net>	<48D10D0C.3020109@timgolden.me.uk>
	<1FD8FD0C19027145B941FAB2BC40C7B302593632@extxmb17.nam.nsroot.net>
	<48D112C2.5020208@timgolden.me.uk>
Message-ID: <01ef01c91926$669963b0$33cc2b10$@com.au>

> > I just discovered that when I do the exact same thing in Outlook
> > 2003,
> > the problem disappears. I believe it's some sort of settings issue in
> > my
> > version of Outlook 2000.  You wouldn't know what that might be would
> > you?  If not, I'm sure I can dig it up by comparing.

I'm guessing it is simply a bug on Outlook 2000 that caused the data to
internally round-trip via an encoding and decoding.  I'd be very surprised
if there was a setting to cause this behaviour (otoh, I've been very
surprised before :)

FWIW, note that for the spambayes project, we used the (more painful)
low-level MAPI interfaces directly and didn't strike this problem - so the
bug is probably in the simple MAPI layer.

Cheers,

Mark


From keston.obendorf at intel.com  Thu Sep 18 02:07:43 2008
From: keston.obendorf at intel.com (Obendorf, Keston)
Date: Wed, 17 Sep 2008 17:07:43 -0700
Subject: [python-win32] COM: Parameterless functions seen as properties
In-Reply-To: <009601c91787$a25c94d0$e715be70$@com.au>
References: <12A5C15467D5B94F8E0FF265D9498ADD03AC5794@orsmsx419.amr.corp.intel.com>
	<009601c91787$a25c94d0$e715be70$@com.au>
Message-ID: <12A5C15467D5B94F8E0FF265D9498ADD03B2C1F1@orsmsx419.amr.corp.intel.com>

Thanks.

 

I don't think the issue is that GetTypeInfo() is not returning anything,
but it's returning something that cannot be used properly.

 

I wrote another implementation of the parameterless function, this time
in C++. The implementation in C++ works just fine:

 

retStr = Dispatch("c++implementation").getNLType() 

 

--causes no exceptions and works as intended.

 

I've been stepping through the CDispatch creation process, which on the
account of both C# and C++ implementations seems to go OK. In both
run-throughs, dynamic is able to get both typeinfo and typecomp. I judge
that this happens because in both cases the _olerepr_ of the CDispatch
object is a LazyDispatchItem not DispatchItem, which only happens if
both typeinfo and typecomp are defined. 

 

The implementations diverge, however, in _getattr_, specifically when
calling __LazyMap__. The C++ implementation returns a 1, whereas the C#
implementation returns None. This is due to the fact that in
_LazyAddAttr_ , typecomp.Bind returns (0, None) for all invoke types for
the C# implementation whereas the C++ implementation binds to
INVOKE_FUNC. 

 

(This leads me to conclude that the typecomp for the C# implementation
is bad in some way. That is probably more an issue for a C# mailing
list. I don't think it's Python's fault.)

 

_getattr_ allows recovery from this issue with a last ditch effort,
which you reference in your post as the second thing that could be going
on.  Unfortunately, it sets the invoke type by default to
pythoncom.INVOKE_PROPERTYGET. C# does not seem to care that its function
is not a property, and returns an evaluation of the function rather than
throwing an exception. I don't know if there is a way to prevent this
behavior. 

 

I feel like there's not much information in this posting for all its
words. =(   I figured out HOW it is failing, but not exactly WHY. It
looks like this is a problem in the way C# handles IDispatch and
ITypeInfo.

 

 

P.S. I am new to this whole mailing-list thing, so I don't know exactly
what e-mail addresses to send this post to. Let's hope it winds up in
the right place. 

 

Thanks, Keston

 

________________________________

From: Mark Hammond [mailto:skippy.hammond at gmail.com] On Behalf Of Mark
Hammond
Sent: Monday, September 15, 2008 4:06 PM
To: Obendorf, Keston; python-win32 at python.org
Subject: RE: [python-win32] COM: Parameterless functions seen as
properties

 

It looks like 2 things are going on:

 

* Your object isn't returning typelib info to Python when it asks.  If
your IDispatch returned something from GetTypeInfo() we could use, this
wouldn't happen.

 

* Without a typelib, Python can't tell if a reference to 'a.foo' is
going to end up being a method call or not (ie, from Python's POV,
a.foo() is just a property fetch for 'foo' and then calling the result.)
When making a property reference, win32com checks for the
ERRORS_BAD_CONTEXT values defined in win32com\client\dynamic.py - if it
sees one of those errors, it assumes it really hit a method and handles
it accordgingly.  But - in your case, your object is succeeding!  The
simplest thing to do here is to have your object refuse to work as a
property using one of the standard error codes, and Python will then try
it as a method.

 

Mark

 

From: python-win32-bounces+skippy.hammond=gmail.com at python.org
[mailto:python-win32-bounces+skippy.hammond=gmail.com at python.org] On
Behalf Of Obendorf, Keston
Sent: Tuesday, 16 September 2008 8:19 AM
To: python-win32 at python.org
Subject: [python-win32] COM: Parameterless functions seen as properties

 

I'm doing COM development in both Python and C#. Things are working
well, but I'm running into an issue where trying to get parameterless
function returns the return value of that function, which Python then
tries to call, causing an error.

 

 

The C# class is an implementation of an IDL-defined interface.

 

//IDL 

interface IService : IDispatch

{

      [id(101)] HRESULT foo([out, retval] BSTR * pstr);

}

 

//C#

[ComImport]

[Guid("/*GUID of IService*/")]

interface IService{}

 

 

[ComVisible(true)]

[ProgId("bar")]

[Guid("Valid GUID")]

public class bar : IService

{

    public bar(){}

 

    public string foo()

    {

        return "foo"; 

    }

}

 

Now, getting a Dispatch of "bar" in Python is where the issue lies.

 

from win32com.client import Dispatch

bar = Dispatch("bar")

str = bar.foo()

 

>>TypeError: 'unicode' object is not callable

 

This is only a problem with the C# class. If the interface is
implemented in Pythong, the above code will work.

 

Now, this issue can be fixed by adding in _FlagAsMethod before calling
bar.foo()

 

bar._FlagAsMethod("foo")

 

I really don't want to have to do that every time I try to call a
parameterless method of a C# COM-object from Python. That's a huge
headache, especially because I am trying to keep the language any one
component is written in a non-issue. 

 

I've started to dive into the source code of dynamic.py and other Client
modules to try and figure out how I can make this distinction at
Dispatch-time, but it's rough going understanding everything that's
going on in there. Does anyone have any pointers or a straight up
solution to this problem? 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080917/5e9c7e23/attachment-0001.htm>

From mhammond at skippinet.com.au  Thu Sep 18 02:45:10 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Thu, 18 Sep 2008 10:45:10 +1000
Subject: [python-win32] Accessing other interfaces of a Dispatch created
	COM object
In-Reply-To: <9819d58b0809171254n1c59cac0tc0c7a0dd45f266a2@mail.gmail.com>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>
	<9819d58b0809101222m77c0648dh6e0752fc83a75ec2@mail.gmail.com>
	<48C82728.8060206@probo.com>
	<9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>
	<057d01c913a1$4e80dc20$eb829460$@com.au>
	<9819d58b0809101835h1a581fpb67305217b39a40f@mail.gmail.com>
	<9819d58b0809120949v217de44bn8c71ac792dec70fa@mail.gmail.com>
	<9819d58b0809121028w46cf0267l3c517a33289ea595@mail.gmail.com>
	<9819d58b0809171210h159b3402t654bd3e011e4bb@mail.gmail.com>
	<9819d58b0809171228m5f71c45fld2b8a62dec388d03@mail.gmail.com>
	<9819d58b0809171254n1c59cac0tc0c7a0dd45f266a2@mail.gmail.com>
Message-ID: <01f901c91927$d0ee0ad0$72ca2070$@com.au>

> I've taken a look, and I can't say I had an "aha" moment, or anything
> like that, but perhaps I have learned something important.
> 
> Setting up the COM object seems very straightforward:
> 
> Type t = Type.GetTypeFromCLSID("{e04f970b-53ce-420a-86f8-
> 55374677703d}")
> 
> component = (basic.I1)Activator.CreateInstance(t)
> 
> component_2 = component as I2

Note that many compiled languages only need the typelib at compile time -
once built, they don't need the typelib to be installed and registered on
the target system.  This is in contrast to Python, where the typelib is
needed at runtime (unless makepy has been successfully run - the typelib is
then needed at makepy time - which is still at "runtime" unless py2exe etc
is involved).  Indeed, the typelib may not even need to be registered on the
build system - just a reference to it on disk by the build process would be
enough.

So - it still sounds to me that one of the 2 typelibs you rely on aren't
registered, and that your other compiled applications don't notice this.
Have you tried asking the people who supply the binaries you are using?

Mark


From mhammond at skippinet.com.au  Thu Sep 18 02:48:43 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Thu, 18 Sep 2008 10:48:43 +1000
Subject: [python-win32] COM: Parameterless functions seen as properties
In-Reply-To: <12A5C15467D5B94F8E0FF265D9498ADD03B2C1F1@orsmsx419.amr.corp.intel.com>
References: <12A5C15467D5B94F8E0FF265D9498ADD03AC5794@orsmsx419.amr.corp.intel.com>
	<009601c91787$a25c94d0$e715be70$@com.au>
	<12A5C15467D5B94F8E0FF265D9498ADD03B2C1F1@orsmsx419.amr.corp.intel.com>
Message-ID: <020301c91928$530df660$f929e320$@com.au>

The difference in behaviour of ITypeComp.Bind() is interesting, and I agree
is the main problem here.  However, from what you have said, I'm surprised
that a makepy-backed implementation isn't working for you.  The first thing
we do when getting an object is ask for its type info so we can get the
CLSID of the object.  We then lookup the CLSID in our makepy/gencache map,
and if we find it, the entire process you outline is skipped - we already
know the dispid and flags for each of the attributes.  Thus, given your
objects implement most of the typeinfo stuff correctly, I'm surprised this
isn't working for you.

 

Have you tried running makepy on your typelib?

 

Cheers,

 

Mark

 

From: Obendorf, Keston [mailto:keston.obendorf at intel.com] 
Sent: Thursday, 18 September 2008 10:08 AM
To: mhammond at skippinet.com.au; python-win32 at python.org
Subject: RE: [python-win32] COM: Parameterless functions seen as properties

 

Thanks.

 

I don't think the issue is that GetTypeInfo() is not returning anything, but
it's returning something that cannot be used properly.

 

I wrote another implementation of the parameterless function, this time in
C++. The implementation in C++ works just fine:

 

retStr = Dispatch("c++implementation").getNLType() 

 

--causes no exceptions and works as intended.

 

I've been stepping through the CDispatch creation process, which on the
account of both C# and C++ implementations seems to go OK. In both
run-throughs, dynamic is able to get both typeinfo and typecomp. I judge
that this happens because in both cases the _olerepr_ of the CDispatch
object is a LazyDispatchItem not DispatchItem, which only happens if both
typeinfo and typecomp are defined. 

 

The implementations diverge, however, in _getattr_, specifically when
calling __LazyMap__. The C++ implementation returns a 1, whereas the C#
implementation returns None. This is due to the fact that in _LazyAddAttr_ ,
typecomp.Bind returns (0, None) for all invoke types for the C#
implementation whereas the C++ implementation binds to INVOKE_FUNC. 

 

(This leads me to conclude that the typecomp for the C# implementation is
bad in some way. That is probably more an issue for a C# mailing list. I
don't think it's Python's fault.)

 

_getattr_ allows recovery from this issue with a last ditch effort, which
you reference in your post as the second thing that could be going on.
Unfortunately, it sets the invoke type by default to
pythoncom.INVOKE_PROPERTYGET. C# does not seem to care that its function is
not a property, and returns an evaluation of the function rather than
throwing an exception. I don't know if there is a way to prevent this
behavior. 

 

I feel like there's not much information in this posting for all its words.
=(   I figured out HOW it is failing, but not exactly WHY. It looks like
this is a problem in the way C# handles IDispatch and ITypeInfo.

 

 

P.S. I am new to this whole mailing-list thing, so I don't know exactly what
e-mail addresses to send this post to. Let's hope it winds up in the right
place. 

 

Thanks, Keston

 

  _____  

From: Mark Hammond [mailto:skippy.hammond at gmail.com] On Behalf Of Mark
Hammond
Sent: Monday, September 15, 2008 4:06 PM
To: Obendorf, Keston; python-win32 at python.org
Subject: RE: [python-win32] COM: Parameterless functions seen as properties

 

It looks like 2 things are going on:

 

* Your object isn't returning typelib info to Python when it asks.  If your
IDispatch returned something from GetTypeInfo() we could use, this wouldn't
happen.

 

* Without a typelib, Python can't tell if a reference to 'a.foo' is going to
end up being a method call or not (ie, from Python's POV, a.foo() is just a
property fetch for 'foo' and then calling the result.)  When making a
property reference, win32com checks for the ERRORS_BAD_CONTEXT values
defined in win32com\client\dynamic.py - if it sees one of those errors, it
assumes it really hit a method and handles it accordgingly.  But - in your
case, your object is succeeding!  The simplest thing to do here is to have
your object refuse to work as a property using one of the standard error
codes, and Python will then try it as a method.

 

Mark

 

From: python-win32-bounces+skippy.hammond=gmail.com at python.org
[mailto:python-win32-bounces+skippy.hammond=gmail.com at python.org] On Behalf
Of Obendorf, Keston
Sent: Tuesday, 16 September 2008 8:19 AM
To: python-win32 at python.org
Subject: [python-win32] COM: Parameterless functions seen as properties

 

I'm doing COM development in both Python and C#. Things are working well,
but I'm running into an issue where trying to get parameterless function
returns the return value of that function, which Python then tries to call,
causing an error.

 

 

The C# class is an implementation of an IDL-defined interface.

 

//IDL 

interface IService : IDispatch

{

      [id(101)] HRESULT foo([out, retval] BSTR * pstr);

}

 

//C#

[ComImport]

[Guid("/*GUID of IService*/")]

interface IService{}

 

 

[ComVisible(true)]

[ProgId("bar")]

[Guid("Valid GUID")]

public class bar : IService

{

    public bar(){}

 

    public string foo()

    {

        return "foo"; 

    }

}

 

Now, getting a Dispatch of "bar" in Python is where the issue lies.

 

from win32com.client import Dispatch

bar = Dispatch("bar")

str = bar.foo()

 

>>TypeError: 'unicode' object is not callable

 

This is only a problem with the C# class. If the interface is implemented in
Pythong, the above code will work.

 

Now, this issue can be fixed by adding in _FlagAsMethod before calling
bar.foo()

 

bar._FlagAsMethod("foo")

 

I really don't want to have to do that every time I try to call a
parameterless method of a C# COM-object from Python. That's a huge headache,
especially because I am trying to keep the language any one component is
written in a non-issue. 

 

I've started to dive into the source code of dynamic.py and other Client
modules to try and figure out how I can make this distinction at
Dispatch-time, but it's rough going understanding everything that's going on
in there. Does anyone have any pointers or a straight up solution to this
problem? 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080918/23a6df21/attachment-0001.htm>

From keston.obendorf at intel.com  Thu Sep 18 22:32:06 2008
From: keston.obendorf at intel.com (Obendorf, Keston)
Date: Thu, 18 Sep 2008 13:32:06 -0700
Subject: [python-win32] COM: Parameterless functions seen as properties
In-Reply-To: <020301c91928$530df660$f929e320$@com.au>
References: <12A5C15467D5B94F8E0FF265D9498ADD03AC5794@orsmsx419.amr.corp.intel.com>
	<009601c91787$a25c94d0$e715be70$@com.au>
	<12A5C15467D5B94F8E0FF265D9498ADD03B2C1F1@orsmsx419.amr.corp.intel.com>
	<020301c91928$530df660$f929e320$@com.au>
Message-ID: <12A5C15467D5B94F8E0FF265D9498ADD03B2C5FF@orsmsx419.amr.corp.intel.com>

Yes I have tried makepy. I haven't actually had any success with makepy
in the past except when running it on type libraries created with the
MIDL compiler. The gen_py modules created from tlbs exported from .NET
assemblies just don't work. They don't have definitions for any of the
methods of the classes, just the interfaces, and attempting to call any
of the interface methods on an instance grabbed by Dispatch fails with
an Attribute Error. 

 

Just an aside on how I'm getting the .tlb. When C# builds it generates a
.dll. I then use REGASM with the /tlb to register the .dll and export a
tlb from the dll and register that as well. 

 

I think I'll go poke around on the MSDN forums for awhile. 

 

_Keston 

 

________________________________

From: Mark Hammond [mailto:mhammond at skippinet.com.au] 
Sent: Wednesday, September 17, 2008 5:49 PM
To: Obendorf, Keston; python-win32 at python.org
Subject: RE: [python-win32] COM: Parameterless functions seen as
properties

 

The difference in behaviour of ITypeComp.Bind() is interesting, and I
agree is the main problem here.  However, from what you have said, I'm
surprised that a makepy-backed implementation isn't working for you.
The first thing we do when getting an object is ask for its type info so
we can get the CLSID of the object.  We then lookup the CLSID in our
makepy/gencache map, and if we find it, the entire process you outline
is skipped - we already know the dispid and flags for each of the
attributes.  Thus, given your objects implement most of the typeinfo
stuff correctly, I'm surprised this isn't working for you.

 

Have you tried running makepy on your typelib?

 

Cheers,

 

Mark

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080918/3d6b4d8a/attachment.htm>

From juanmpubill at hotmail.com  Thu Sep 18 20:39:40 2008
From: juanmpubill at hotmail.com (Juan Manuel Pubill)
Date: Thu, 18 Sep 2008 18:39:40 +0000
Subject: [python-win32] Pythonwin telit
Message-ID: <BLU134-W5475C9C04756B5E90F2595BC4F0@phx.gbl>


Hola quer?a saber si hay posibilidad de conseguir alguna versi?n m?s nueva que la que distribuye Telit del pythonwin porque es poco estable y tiene varios bugs.
Sin m?s muchas gracias
_________________________________________________________________
Get more from your digital life.  Find out how.
http://www.windowslive.com/default.html?ocid=TXT_TAGLM_WL_Home2_082008
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080918/3a63329e/attachment.htm>

From timr at probo.com  Fri Sep 19 02:05:03 2008
From: timr at probo.com (Tim Roberts)
Date: Thu, 18 Sep 2008 17:05:03 -0700
Subject: [python-win32] Pythonwin telit
In-Reply-To: <BLU134-W5475C9C04756B5E90F2595BC4F0@phx.gbl>
References: <BLU134-W5475C9C04756B5E90F2595BC4F0@phx.gbl>
Message-ID: <48D2ECAF.9010106@probo.com>

Juan Manuel Pubill wrote:
>
> Hola quer?a saber si hay posibilidad de conseguir alguna versi?n m?s
> nueva que la que distribuye Telit del pythonwin porque es poco estable
> y tiene varios bugs.
>

BabelFish translates this as:
> Hello it wanted to know if there is possibility of securing some
> version newer than the one than it distributes Telit of pythonwin
> because he is little stable and it has several bugs. Immediately thank
> you very much

You can always fetch the latest version directly from the source,
http://sourceforge.net/projects/pywin32.

However, there haven't been any significant bugs in quite a long time. 
What version are you using, and what "bugs" do you see?

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From alecmunro at gmail.com  Fri Sep 19 16:32:53 2008
From: alecmunro at gmail.com (Alec Munro)
Date: Fri, 19 Sep 2008 11:32:53 -0300
Subject: [python-win32] Accessing other interfaces of a Dispatch created
	COM object
In-Reply-To: <01f901c91927$d0ee0ad0$72ca2070$@com.au>
References: <9819d58b0808280556g46cbd1cbj85f32cc82987615a@mail.gmail.com>
	<9819d58b0809101523u5811d094kfb01109987e9491d@mail.gmail.com>
	<057d01c913a1$4e80dc20$eb829460$@com.au>
	<9819d58b0809101835h1a581fpb67305217b39a40f@mail.gmail.com>
	<9819d58b0809120949v217de44bn8c71ac792dec70fa@mail.gmail.com>
	<9819d58b0809121028w46cf0267l3c517a33289ea595@mail.gmail.com>
	<9819d58b0809171210h159b3402t654bd3e011e4bb@mail.gmail.com>
	<9819d58b0809171228m5f71c45fld2b8a62dec388d03@mail.gmail.com>
	<9819d58b0809171254n1c59cac0tc0c7a0dd45f266a2@mail.gmail.com>
	<01f901c91927$d0ee0ad0$72ca2070$@com.au>
Message-ID: <9819d58b0809190732w5d860ae5qd0f67c1172835b65@mail.gmail.com>

Hi Mark,

I am able to run MakePy successfully, at least as far as I can tell.
Does that affect your conclusion?

Is there any way I can figure out what library it is looking for when
I call the following (from your earlier example)?

>>> d.GetTypeInfo(0)

As that also raises the "library not registered" exception.

I've been in some contact with the vendor who supplies us with this,
mostly when I started trying to use Python for this purpose. I've
recently updated them on my progress, and I'm waiting for a response.

Thanks again.
Alec

On Wed, Sep 17, 2008 at 9:45 PM, Mark Hammond <mhammond at skippinet.com.au> wrote:
>> I've taken a look, and I can't say I had an "aha" moment, or anything
>> like that, but perhaps I have learned something important.
>>
>> Setting up the COM object seems very straightforward:
>>
>> Type t = Type.GetTypeFromCLSID("{e04f970b-53ce-420a-86f8-
>> 55374677703d}")
>>
>> component = (basic.I1)Activator.CreateInstance(t)
>>
>> component_2 = component as I2
>
> Note that many compiled languages only need the typelib at compile time -
> once built, they don't need the typelib to be installed and registered on
> the target system.  This is in contrast to Python, where the typelib is
> needed at runtime (unless makepy has been successfully run - the typelib is
> then needed at makepy time - which is still at "runtime" unless py2exe etc
> is involved).  Indeed, the typelib may not even need to be registered on the
> build system - just a reference to it on disk by the build process would be
> enough.
>
> So - it still sounds to me that one of the 2 typelibs you rely on aren't
> registered, and that your other compiled applications don't notice this.
> Have you tried asking the people who supply the binaries you are using?
>
> Mark
>
>

From gagsl-py2 at yahoo.com.ar  Sat Sep 20 04:10:55 2008
From: gagsl-py2 at yahoo.com.ar (Gabriel Genellina)
Date: Fri, 19 Sep 2008 23:10:55 -0300
Subject: [python-win32] Pythonwin telit
References: <BLU134-W5475C9C04756B5E90F2595BC4F0@phx.gbl>
Message-ID: <op.uhrcchg6x6zn5v@gabriel2.softlabbsas.com.ar>

En Thu, 18 Sep 2008 15:39:40 -0300, Juan Manuel Pubill  
<juanmpubill at hotmail.com> escribi?:

> Hola quer?a saber si hay posibilidad de conseguir alguna versi?n m?s  
> nueva que la que distribuye Telit del pythonwin porque es poco estable y  
> tiene varios bugs.
> Sin m?s muchas gracias

Lista de Python en castellano:
python-es at aditel.org ==  
http://dir.gmane.org/gmane.comp.python.general.castellano

-- 
Gabriel Genellina


From amicitas at gmail.com  Sat Sep 20 20:36:57 2008
From: amicitas at gmail.com (mir amicitas)
Date: Sat, 20 Sep 2008 11:36:57 -0700
Subject: [python-win32] Drag and drop of files (CF_HDROP)
Message-ID: <55e194140809201136r5466d5d9te1acdf9edb431e26@mail.gmail.com>

I have been trying to implement drag and drop of files in python.  My
application will written in PyQt4 but they I want to use the native drag and
drop for windows as PyQt only supports mime based dragging, which does not
work for some of the windows applications that I want to connect with.

I have made a certain amount of progress getting this to work using pywin32,
however I have gotten stuck.

So far I have been able to implement dragging of text (CF_TEXT or
CF_UNICODETEXT).  This basically took reimplementing the IDropSource &
IDataObject objects, wrapping them using win32com.server.util.wrap() and
then calling pythoncom.DoDragDrop().  For some of this I have been able to
follow allong with the code in testClipboard.py that is included with the
pywin32 installaiton.

The next thing that I want to do is to use CF_HDROP as the format type with
a list of file locations.  I am getting stuck at the point when I actually
add the list of file locations to the pythoncom.STGMEDIUM() object.

>From what I can gather this data needs to be placed into a DROPFILES
structure (msdn page on CF_HDROP &
DROPFILES<http://msdn.microsoft.com/en-us/library/bb776902%28VS.85%29.aspx>).
So my questing is how to I create this structure in pywin32?


Any help on this would be appretiated, particularly since I can't find
documentation for a lot of this stuff (I have resorted to greping the source
code to find stuff).  I am totaly new to win32 programming so this has been
a bit more that I meant to bite off.

If anyone knows of an easyer way to achive this I would be happy to hear of
it.


Thanks,
-- amicitas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080920/0a389c9f/attachment.htm>

From mail at timgolden.me.uk  Sun Sep 21 11:25:57 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Sun, 21 Sep 2008 10:25:57 +0100
Subject: [python-win32] Drag and drop of files (CF_HDROP)
In-Reply-To: <55e194140809201136r5466d5d9te1acdf9edb431e26@mail.gmail.com>
References: <55e194140809201136r5466d5d9te1acdf9edb431e26@mail.gmail.com>
Message-ID: <48D61325.9090308@timgolden.me.uk>

mir amicitas wrote:
> I have been trying to implement drag and drop of files in python.  My
> application will written in PyQt4 but they I want to use the native drag and
> drop for windows as PyQt only supports mime based dragging, which does not
> work for some of the windows applications that I want to connect with.
> 
> I have made a certain amount of progress getting this to work using pywin32,
> however I have gotten stuck.
> 
> So far I have been able to implement dragging of text (CF_TEXT or
> CF_UNICODETEXT).  This basically took reimplementing the IDropSource &
> IDataObject objects, wrapping them using win32com.server.util.wrap() and
> then calling pythoncom.DoDragDrop().  For some of this I have been able to
> follow allong with the code in testClipboard.py that is included with the
> pywin32 installaiton.
> 
> The next thing that I want to do is to use CF_HDROP as the format type with
> a list of file locations.  I am getting stuck at the point when I actually
> add the list of file locations to the pythoncom.STGMEDIUM() object.
> 
>>From what I can gather this data needs to be placed into a DROPFILES
> structure (msdn page on CF_HDROP &
> DROPFILES<http://msdn.microsoft.com/en-us/library/bb776902%28VS.85%29.aspx>).
> So my questing is how to I create this structure in pywin32?
> 
> 
> Any help on this would be appretiated, particularly since I can't find
> documentation for a lot of this stuff (I have resorted to greping the source
> code to find stuff).  I am totaly new to win32 programming so this has been
> a bit more that I meant to bite off.


Can you have a look at this thread from earlier this
year to see if it helps to answer you question:

http://mail.python.org/pipermail/python-win32/2008-April/007409.html

TJG

From amicitas at gmail.com  Sun Sep 21 22:36:02 2008
From: amicitas at gmail.com (mir amicitas)
Date: Sun, 21 Sep 2008 13:36:02 -0700
Subject: [python-win32] Drag and drop of files (CF_HDROP)
In-Reply-To: <48D61325.9090308@timgolden.me.uk>
References: <55e194140809201136r5466d5d9te1acdf9edb431e26@mail.gmail.com>
	<48D61325.9090308@timgolden.me.uk>
Message-ID: <55e194140809211336y144fc291t1ddead337356bc68@mail.gmail.com>

>
>
> Can you have a look at this thread from earlier this
> year to see if it helps to answer you question:
>
> http://mail.python.org/pipermail/python-win32/2008-April/007409.html
>
> TJG


Thanks Tim for sending along that thread.
This is close to what I need (and provides some useful information).  This
example however takes the data out of the STGMEDIUM object, rather than
putting the data into it.  In doing so the data object does not have to be
defined.

The relevant code from the example is:

   def Drop (self, data_object, key_state, point, effect):
     data = data_object.GetData ((15, None, 1, -1, 1))
     n_files = shell.DragQueryFileW (data.data_handle, -1)

     filenames = [
       shell.DragQueryFileW (data.data_handle, n_file) \
         for n_file in range (n_files)
     ]


So my question is how can I make the data object so that I can place it into
the data object.

Here is the snippet of code from my project showing what I want to do.

class TestDataObjectFile:
    _com_interfaces_ = [pythoncom.IID_IDataObject]
    _public_methods_ = IDataObject_Methods
    def __init__(self, file_path):
        global num_do_objects
        num_do_objects += 1
        self.file_path = file_path
        self.supported_fe = []

        cf = win32con.CF_HDROP
        fe = cf, None, pythoncom.DVASPECT_CONTENT, -1,
pythoncom.TYMED_HGLOBAL
        self.supported_fe.append(fe)

    def __del__(self):
        global num_do_objects
        num_do_objects -= 1

    def _query_interface_(self, iid):
        if iid==pythoncom.IID_IEnumFORMATETC:
            return NewEnum(self.supported_fe, iid=iid)

    def GetData(self, fe):
        print 'GetData'
        ret_stg = None
        cf, target, aspect, index, tymed  = fe

        print cf, win32con.CF_HDROP

        if aspect & pythoncom.DVASPECT_CONTENT and \
           tymed==pythoncom.TYMED_HGLOBAL:
            if cf == win32con.CF_HDROP:
                ret_stg = pythoncom.STGMEDIUM()


                ret_stg.set(pythoncom.TYMED_HGLOBAL, self.file_path)

        if ret_stg is None:
            print 'No ret_stg.'
            raise COMException(hresult=winerror.E_NOTIMPL)
        return ret_stg

    . . .

----

The point where I am stuck is the lines in getData:
ret_stg = pythoncom.STGMEDIUM()
ret_stg.set(pythoncom.TYMED_HGLOBAL, self.file_path)

The problem is that what needs to be placed into ret_stg is not the
file_names (as I have done above),  but a data structure.

So my question is how do I get the filenames into the data object?

Thanks again for your help with this.


- Novi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080921/b6b0551e/attachment.htm>

From rwupole at msn.com  Mon Sep 22 01:14:00 2008
From: rwupole at msn.com (Roger Upole)
Date: Sun, 21 Sep 2008 19:14:00 -0400
Subject: [python-win32]  Re: Drag and drop of files (CF_HDROP)
Message-ID: <BLU0-SMTP2351F4B43170858512626BA3480@phx.gbl>

You can use struct.pack to create a buffer containing
a DROPFILES struct:

        """
                typedef struct _DROPFILES {
                    DWORD pFiles;
                    POINT pt;
                    BOOL fNC;
                    BOOL fWide; } DROPFILES, *LPDROPFILES;
        """
        fname_buf='\0'.join(fnames)+'\0\0'
        fmt="lllll%ss" %len(fname_buf)
        df=struct.pack(fmt, 20, 0, 0, 0, 0, fname_buf)
        ret_stg.set(pythoncom.TYMED_HGLOBAL, df)

               Roger


From imageguy1206 at gmail.com  Mon Sep 22 03:10:25 2008
From: imageguy1206 at gmail.com (geoff)
Date: Sun, 21 Sep 2008 22:10:25 -0300
Subject: [python-win32] Virtual print driver - similiar to winpdf
Message-ID: <3ee0b280809211810j7d516500o9e93c9673ee82e2b@mail.gmail.com>

Would creating a virtual print driver be possible with pywin32,
similiar to the windows pdf driver ?
WE have an old application that outputs data to a dotmatrix printer.
WE would like to extend this application and integrate it with another
application and the simplest way to do this would be to capture the
print output, extract some information from the print output and then
forward that information to a database application.

Any suggestions.

I have found a couple of websites about print drivers (google is my
friend !) but thought I might ask here for some advice and guidance.

g.

From amicitas at gmail.com  Mon Sep 22 07:18:38 2008
From: amicitas at gmail.com (mir amicitas)
Date: Sun, 21 Sep 2008 22:18:38 -0700
Subject: [python-win32] Drag and drop of files (CF_HDROP)
In-Reply-To: <BLU0-SMTP2351F4B43170858512626BA3480@phx.gbl>
References: <BLU0-SMTP2351F4B43170858512626BA3480@phx.gbl>
Message-ID: <55e194140809212218n7bd9633ev31ef83ee7ebb5fb3@mail.gmail.com>

Thank you for you help, that works perfectly.  I can now drag files from my
program (yay!).

One more question . . .

In the Microsoft specification for DoDragDrop there are 4 arguments (data,
source, allowed effects, used effect).  However in the pythoncom
implementation only 3 arguments are allowed.  Is there a reason for this?

Here is the page on that function from ms:
http://msdn.microsoft.com/en-us/library/ms678486(VS.85).aspx

- novi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080921/72665681/attachment.htm>

From rwupole at msn.com  Mon Sep 22 10:35:24 2008
From: rwupole at msn.com (Roger Upole)
Date: Mon, 22 Sep 2008 04:35:24 -0400
Subject: [python-win32]  Re: Drag and drop of files (CF_HDROP)
Message-ID: <BLU0-SMTP267BDCF36F0926951F88F8A34B0@phx.gbl>

mir amicitas" wrote:
> Thank you for you help, that works perfectly.  I can now drag files from 
> my
> program (yay!).
>
> One more question . . .
>
> In the Microsoft specification for DoDragDrop there are 4 arguments (data,
> source, allowed effects, used effect).  However in the pythoncom
> implementation only 3 arguments are allowed.  Is there a reason for this?
>
> Here is the page on that function from ms:
> http://msdn.microsoft.com/en-us/library/ms678486(VS.85).aspx
>
> - novi

The last arg is output, and will be returned from the function.

        Roger


From timr at probo.com  Mon Sep 22 19:41:57 2008
From: timr at probo.com (Tim Roberts)
Date: Mon, 22 Sep 2008 10:41:57 -0700
Subject: [python-win32] Virtual print driver - similiar to winpdf
In-Reply-To: <3ee0b280809211810j7d516500o9e93c9673ee82e2b@mail.gmail.com>
References: <3ee0b280809211810j7d516500o9e93c9673ee82e2b@mail.gmail.com>
Message-ID: <48D7D8E5.7070106@probo.com>

geoff wrote:
> Would creating a virtual print driver be possible with pywin32,
> similiar to the windows pdf driver ?
> WE have an old application that outputs data to a dotmatrix printer.
> WE would like to extend this application and integrate it with another
> application and the simplest way to do this would be to capture the
> print output, extract some information from the print output and then
> forward that information to a database application.
>
> Any suggestions.
>
> I have found a couple of websites about print drivers (google is my
> friend !) but thought I might ask here for some advice and guidance.
>   

This is trickier than you might guess.  Does the application actually
write to the printer using a Windows printer driver?  Some old
applications that just used a printer for logging wrote directly to the
parallel ports.  That kind of thing is almost impossible to trap.  If it
does, how does it pick the printer?  If it doesn't present you with a
dialog, then it might be picking the printer by name.  That means you
can't just install your own driver, unless you use the same name.  If it
does present you with a dialog, can you use the Generic / Text-Only
printer driver, and have it dump to a text file?  (That's one of the
standard printer drivers built-in to Windows.)

If you really need a driver, you are not going to be able to do this
from Python.  There are a sample printer drivers in the WDK (Windows
Driver Kit).  It is not a project to be undertaken lightly.

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From imageguy1206 at gmail.com  Mon Sep 22 21:20:47 2008
From: imageguy1206 at gmail.com (geoff)
Date: Mon, 22 Sep 2008 15:20:47 -0400
Subject: [python-win32] Virtual print driver - similiar to winpdf
In-Reply-To: <48D7D8E5.7070106@probo.com>
References: <3ee0b280809211810j7d516500o9e93c9673ee82e2b@mail.gmail.com>
	<48D7D8E5.7070106@probo.com>
Message-ID: <3ee0b280809221220q277e3c9ap262d84da3bce85a0@mail.gmail.com>

Thanks for you guidance Tim.  After reviewing I am coming to the same
conclusion.

> This is trickier than you might guess.  Does the application actually
> write to the printer using a Windows printer driver?  Some old
> applications that just used a printer for logging wrote directly to the
> parallel ports.  That kind of thing is almost impossible to trap.  If it
> does, how does it pick the printer?

The application is quite "configurable".  Printers can be configured
so the application to write to either a port (including UNC path), a
Windows print driver or directly to a file that is specificed within
the application.  We typically use the "generic/text" driver to
define the "windows" printers, however, just pointing the app to
"LPT1:" works just fine too.

If it doesn't present you with a
> dialog, then it might be picking the printer by name.  That means you
> can't just install your own driver, unless you use the same name.  If it
> does present you with a dialog, can you use the Generic / Text-Only
> printer driver, and have it dump to a text file?  (That's one of the
> standard printer drivers built-in to Windows.)

I have thought of the Windows printer defined as "file" however, I
don't want the user to be prompted for the file name each time and
haven't found a way to suppress this.

> If you really need a driver, you are not going to be able to do this
> from Python.  There are a sample printer drivers in the WDK (Windows
> Driver Kit).  It is not a project to be undertaken lightly.

Yes, I am starting to understand that :-(

What we are trying to do is capture the printing "dynamically" while
the user is running the application ( -- think of invoice printing in
a business).  The app output is text, with some printer control codes
embedded to change fonts, colors, newlines, etc, but other than that
it is just ascii text printed line-by-line sent to the specificied
port or file.

The problem we see with the "print to a file" approach is we would
like the system to react immediately when the user "prints" and having
some sort of "polling" to watch the file isn't the best solution as it
introduces a delay/latency.

One option I was considering is using Windows printer defined as a
"Standard TCP/IP Printer Port".  Our thinking is to manually configure
a printer in Windows and then use the LPR or RAW method of this print
monitor to redirect the output to our program via TCP/IP.  Haven't
tested this , however, the thinking is that this would tell us
immediately when the user is printing and our utility could re-act
accordingly.

Any thoughts or comments or guidance would be truly appreciated.

g.

From python at kareta.de  Tue Sep 23 00:02:41 2008
From: python at kareta.de (=?ISO-8859-1?Q?J=FCrgen_Kareta?=)
Date: Tue, 23 Sep 2008 00:02:41 +0200
Subject: [python-win32] Virtual print driver - similiar to winpdf
In-Reply-To: <3ee0b280809221220q277e3c9ap262d84da3bce85a0@mail.gmail.com>
References: <3ee0b280809211810j7d516500o9e93c9673ee82e2b@mail.gmail.com>	<48D7D8E5.7070106@probo.com>
	<3ee0b280809221220q277e3c9ap262d84da3bce85a0@mail.gmail.com>
Message-ID: <48D81601.3000308@kareta.de>


> Thanks for you guidance Tim.  After reviewing I am coming to the same
> conclusion.
>
>   
>> This is trickier than you might guess.  Does the application actually
>> write to the printer using a Windows printer driver?  Some old
>> applications that just used a printer for logging wrote directly to the
>> parallel ports.  That kind of thing is almost impossible to trap.  If it
>> does, how does it pick the printer?
>>     
>
> The application is quite "configurable".  Printers can be configured
> so the application to write to either a port (including UNC path), a
> Windows print driver or directly to a file that is specificed within
> the application.  We typically use the "generic/text" driver to
> define the "windows" printers, however, just pointing the app to
> "LPT1:" works just fine too.
>
> If it doesn't present you with a
>   
>> dialog, then it might be picking the printer by name.  That means you
>> can't just install your own driver, unless you use the same name.  If it
>> does present you with a dialog, can you use the Generic / Text-Only
>> printer driver, and have it dump to a text file?  (That's one of the
>> standard printer drivers built-in to Windows.)
>>     
>
> I have thought of the Windows printer defined as "file" however, I
> don't want the user to be prompted for the file name each time and
> haven't found a way to suppress this.
>
>   
>> If you really need a driver, you are not going to be able to do this
>> from Python.  There are a sample printer drivers in the WDK (Windows
>> Driver Kit).  It is not a project to be undertaken lightly.
>>     
>
> Yes, I am starting to understand that :-(
>
> What we are trying to do is capture the printing "dynamically" while
> the user is running the application ( -- think of invoice printing in
> a business).  The app output is text, with some printer control codes
> embedded to change fonts, colors, newlines, etc, but other than that
> it is just ascii text printed line-by-line sent to the specificied
> port or file.
>
> The problem we see with the "print to a file" approach is we would
> like the system to react immediately when the user "prints" and having
> some sort of "polling" to watch the file isn't the best solution as it
> introduces a delay/latency.
>
> One option I was considering is using Windows printer defined as a
> "Standard TCP/IP Printer Port".  Our thinking is to manually configure
> a printer in Windows and then use the LPR or RAW method of this print
> monitor to redirect the output to our program via TCP/IP.  Haven't
> tested this , however, the thinking is that this would tell us
> immediately when the user is printing and our utility could re-act
> accordingly.
>
> Any thoughts or comments or guidance would be truly appreciated.
>
> g.
> _______________________________________________
> python-win32 mailing list
> python-win32 at python.org
> http://mail.python.org/mailman/listinfo/python-win32
>
>
>   
Hi Geoff,

redmon: http://pages.cs.wisc.edu/~ghost/redmon/ might be an option for 
you. I use it for years now without any problems. In one case I provide 
printing capabilities inside citrix apps for  vpn users and in an other 
case I extended an application with fax functionality. One can either 
use a temporary file or send the data directly to a python script. It 
also works nice when the python script is installed as a windows service.

Regards,
J?rgen


From patrick at raptr.com  Tue Sep 23 04:25:30 2008
From: patrick at raptr.com (Patrick Li)
Date: Mon, 22 Sep 2008 19:25:30 -0700
Subject: [python-win32] idle detection on system?
Message-ID: <2b5a48740809221925u1254a23fi160173e4bbeb0696@mail.gmail.com>

Hi,
Does anyone know how I can detect if a system is idle?  By idle I am
referring to a user being away from the computer (no keyboard or mouse
movement) for X minutes.  Tried searching around but I haven't found
anything that mentions how I can hook into these two events.

Thanks in advance,
Patrick
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080922/dfb74aab/attachment.htm>

From imageguy1206 at gmail.com  Tue Sep 23 05:21:31 2008
From: imageguy1206 at gmail.com (geoff)
Date: Mon, 22 Sep 2008 23:21:31 -0400
Subject: [python-win32] Virtual print driver - similiar to winpdf
In-Reply-To: <48D81601.3000308@kareta.de>
References: <3ee0b280809211810j7d516500o9e93c9673ee82e2b@mail.gmail.com>
	<48D7D8E5.7070106@probo.com>
	<3ee0b280809221220q277e3c9ap262d84da3bce85a0@mail.gmail.com>
	<48D81601.3000308@kareta.de>
Message-ID: <3ee0b280809222021h4702109m1e40885c91f2fac6@mail.gmail.com>

> Hi Geoff,
>
> redmon: http://pages.cs.wisc.edu/~ghost/redmon/ might be an option for you.
> I use it for years now without any problems. In one case I provide printing
> capabilities inside citrix apps for  vpn users and in an other case I
> extended an application with fax functionality. One can either use a
> temporary file or send the data directly to a python script. It also works
> nice when the python script is installed as a windows service.
>
Thanks I will give that a try.

g.

From mail at timgolden.me.uk  Tue Sep 23 09:50:19 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Tue, 23 Sep 2008 08:50:19 +0100
Subject: [python-win32] idle detection on system?
In-Reply-To: <2b5a48740809221925u1254a23fi160173e4bbeb0696@mail.gmail.com>
References: <2b5a48740809221925u1254a23fi160173e4bbeb0696@mail.gmail.com>
Message-ID: <48D89FBB.8050406@timgolden.me.uk>

Patrick Li wrote:
> Hi,
> Does anyone know how I can detect if a system is idle?  By idle I am
> referring to a user being away from the computer (no keyboard or mouse
> movement) for X minutes.  Tried searching around but I haven't found
> anything that mentions how I can hook into these two events.


I knew there had been some discussion about this in the
past, but my attempts at searching the archives were
frustrated by the huge number of posts about "IDLE"!

In short: you can use the Scheduled Tasks API to trigger
something when the system has been idle for x minutes:

http://mail.python.org/pipermail/python-list/2005-May/323848.html

Alternatively, the GetLastInputInfo API might be what you want:

http://msdn.microsoft.com/en-us/library/ms646302(VS.85).aspx

TJG

From andrea.gavana at gmail.com  Tue Sep 23 13:03:08 2008
From: andrea.gavana at gmail.com (Andrea Gavana)
Date: Tue, 23 Sep 2008 12:03:08 +0100
Subject: [python-win32] Invalid or corrupted installer...
Message-ID: <d5ff27200809230403y6bed6fdfw993d87bccad9ba89@mail.gmail.com>

Hi All,

    I am having few problems in downloading and installing the release
212 of PyWin32 (and all the other releases as well) from SourceForge.
Whatever executable file I try to download from SourceForge (I need
the pywin32-212.win32-py2.5.exe for my Python version), the download
remains stuck at 4.4/4.7 MB (the pywin32-212.win32-py2.5.exe file size
is 5.3 MB) for a couple of minutes, and then it leaves me with a
corrupted installation file which cannot be run.
Is anyone else experiencing the same issues?

I have tried from 3 different computers here at work but with the same
results, and also from 4 different mirrors at SourceForge and from
another mirror (indipendent from SourceForge) here:

http://gforge.inria.fr/frs/?group_id=79&release_id=1150

This is Windows XP SP2, Python 2.5.2.

Thank you for all your suggestions.

Andrea.

"Imagination Is The Only Weapon In The War Against Reality."
http://xoomer.alice.it/infinity77/

From mail at timgolden.me.uk  Tue Sep 23 13:12:46 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Tue, 23 Sep 2008 12:12:46 +0100
Subject: [python-win32] Invalid or corrupted installer...
In-Reply-To: <d5ff27200809230403y6bed6fdfw993d87bccad9ba89@mail.gmail.com>
References: <d5ff27200809230403y6bed6fdfw993d87bccad9ba89@mail.gmail.com>
Message-ID: <48D8CF2E.6030903@timgolden.me.uk>

Andrea Gavana wrote:
> Hi All,
> 
>     I am having few problems in downloading and installing the release
> 212 of PyWin32 (and all the other releases as well) from SourceForge.
> Whatever executable file I try to download from SourceForge (I need
> the pywin32-212.win32-py2.5.exe for my Python version), the download
> remains stuck at 4.4/4.7 MB (the pywin32-212.win32-py2.5.exe file size
> is 5.3 MB) for a couple of minutes, and then it leaves me with a
> corrupted installation file which cannot be run.
> Is anyone else experiencing the same issues?


Not sure why there might be a problem. To help you out,
I've just dropped a copy here:

http://timgolden.me.uk/python/downloads/pywin32-212.win32-py2.5.exe

TJG


From andrea.gavana at gmail.com  Tue Sep 23 13:34:02 2008
From: andrea.gavana at gmail.com (Andrea Gavana)
Date: Tue, 23 Sep 2008 12:34:02 +0100
Subject: [python-win32] Invalid or corrupted installer...
In-Reply-To: <48D8CF2E.6030903@timgolden.me.uk>
References: <d5ff27200809230403y6bed6fdfw993d87bccad9ba89@mail.gmail.com>
	<48D8CF2E.6030903@timgolden.me.uk>
Message-ID: <d5ff27200809230434oeb6c330pa0df72609ca2dfe7@mail.gmail.com>

Hi Tim,

On Tue, Sep 23, 2008 at 12:12 PM, Tim Golden wrote:
> Andrea Gavana wrote:
>>
>> Hi All,
>>
>>    I am having few problems in downloading and installing the release
>> 212 of PyWin32 (and all the other releases as well) from SourceForge.
>> Whatever executable file I try to download from SourceForge (I need
>> the pywin32-212.win32-py2.5.exe for my Python version), the download
>> remains stuck at 4.4/4.7 MB (the pywin32-212.win32-py2.5.exe file size
>> is 5.3 MB) for a couple of minutes, and then it leaves me with a
>> corrupted installation file which cannot be run.
>> Is anyone else experiencing the same issues?
>
>
> Not sure why there might be a problem. To help you out,
> I've just dropped a copy here:
>
> http://timgolden.me.uk/python/downloads/pywin32-212.win32-py2.5.exe

Thank you for your help, I appreciate it. Unfortunately I am still
having the same issue, even with the download you kindly provided. IE
downloaded only 4.7 MB of the installer, and when I double-click it I
get the "Invalid or corrupted file" message. At this point, it may be
something due to our company firewall/internet management, although I
was able to download other files (bigger, smaller, from SourceForge
and not). I am 110% puzzled and a bit angry with our internet policy
here (why don't they let me use Firefox?), also because I use the
win32* extensions for all my applications and PythonWin as editor for
my files :-(

Thank you again.

Andrea.

"Imagination Is The Only Weapon In The War Against Reality."
http://xoomer.alice.it/infinity77/

From mdriscoll at co.marshall.ia.us  Tue Sep 23 15:12:09 2008
From: mdriscoll at co.marshall.ia.us (Mike Driscoll)
Date: Tue, 23 Sep 2008 08:12:09 -0500
Subject: [python-win32] Virtual print driver - similiar to winpdf
In-Reply-To: <3ee0b280809221220q277e3c9ap262d84da3bce85a0@mail.gmail.com>
References: <3ee0b280809211810j7d516500o9e93c9673ee82e2b@mail.gmail.com>	<48D7D8E5.7070106@probo.com>
	<3ee0b280809221220q277e3c9ap262d84da3bce85a0@mail.gmail.com>
Message-ID: <48D8EB29.60502@co.marshall.ia.us>

geoff wrote:
> Thanks for you guidance Tim.  After reviewing I am coming to the same
> conclusion.
>
>   
>> This is trickier than you might guess.  Does the application actually
>> write to the printer using a Windows printer driver?  Some old
>> applications that just used a printer for logging wrote directly to the
>> parallel ports.  That kind of thing is almost impossible to trap.  If it
>> does, how does it pick the printer?
>>     
>
> The application is quite "configurable".  Printers can be configured
> so the application to write to either a port (including UNC path), a
> Windows print driver or directly to a file that is specificed within
> the application.  We typically use the "generic/text" driver to
> define the "windows" printers, however, just pointing the app to
> "LPT1:" works just fine too.
>
> If it doesn't present you with a
>   
>> dialog, then it might be picking the printer by name.  That means you
>> can't just install your own driver, unless you use the same name.  If it
>> does present you with a dialog, can you use the Generic / Text-Only
>> printer driver, and have it dump to a text file?  (That's one of the
>> standard printer drivers built-in to Windows.)
>>     
>
> I have thought of the Windows printer defined as "file" however, I
> don't want the user to be prompted for the file name each time and
> haven't found a way to suppress this.
>
>   
>> If you really need a driver, you are not going to be able to do this
>> from Python.  There are a sample printer drivers in the WDK (Windows
>> Driver Kit).  It is not a project to be undertaken lightly.
>>     
>
> Yes, I am starting to understand that :-(
>
> What we are trying to do is capture the printing "dynamically" while
> the user is running the application ( -- think of invoice printing in
> a business).  The app output is text, with some printer control codes
> embedded to change fonts, colors, newlines, etc, but other than that
> it is just ascii text printed line-by-line sent to the specificied
> port or file.
>
> The problem we see with the "print to a file" approach is we would
> like the system to react immediately when the user "prints" and having
> some sort of "polling" to watch the file isn't the best solution as it
> introduces a delay/latency.
>
> One option I was considering is using Windows printer defined as a
> "Standard TCP/IP Printer Port".  Our thinking is to manually configure
> a printer in Windows and then use the LPR or RAW method of this print
> monitor to redirect the output to our program via TCP/IP.  Haven't
> tested this , however, the thinking is that this would tell us
> immediately when the user is printing and our utility could re-act
> accordingly.
>
> Any thoughts or comments or guidance would be truly appreciated.
>
> g.
>
>   

You might take a look at Pyla. It's a print to fax library. I think it's 
kind of what you're talking about. I've been trying to hack it off and 
on for a while now as we only need a little of its functionality here at 
work. There isn't much for docs, so you'll have to read the source. 
Here's a link: http://www.teamsw.it/pyla/

Mike

From andrea.gavana at gmail.com  Tue Sep 23 17:01:39 2008
From: andrea.gavana at gmail.com (Andrea Gavana)
Date: Tue, 23 Sep 2008 16:01:39 +0100
Subject: [python-win32] Invalid or corrupted installer...
In-Reply-To: <48D903B3.4090203@timgolden.me.uk>
References: <d5ff27200809230403y6bed6fdfw993d87bccad9ba89@mail.gmail.com>
	<48D8CF2E.6030903@timgolden.me.uk>
	<d5ff27200809230434oeb6c330pa0df72609ca2dfe7@mail.gmail.com>
	<48D903B3.4090203@timgolden.me.uk>
Message-ID: <d5ff27200809230801l66be4c33k275b4e111c43455e@mail.gmail.com>

Hi Tim,

On Tue, Sep 23, 2008 at 3:56 PM, Tim Golden wrote:
> Also, I take it that if you can't use Firefox, you can't use wget either?

No, our firewall blocks *everything* that is not internet explorer (!)

> http://users.ugent.be/~bpuype/wget/
>
> or Python urrllib?

The same as above, unfortunately. In any case, our IT has managed to
download it few minutes ago using an external connection and now I am
back to work. No explanation was provided for the reason of the
corrupted download (!)

In any case, thank you so much to Tim and Harald for helping me!

Andrea.

"Imagination Is The Only Weapon In The War Against Reality."
http://xoomer.alice.it/infinity77/

From mail at timgolden.me.uk  Tue Sep 23 17:07:06 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Tue, 23 Sep 2008 16:07:06 +0100
Subject: [python-win32] Invalid or corrupted installer...
In-Reply-To: <d5ff27200809230801l66be4c33k275b4e111c43455e@mail.gmail.com>
References: <d5ff27200809230403y6bed6fdfw993d87bccad9ba89@mail.gmail.com>	
	<48D8CF2E.6030903@timgolden.me.uk>	
	<d5ff27200809230434oeb6c330pa0df72609ca2dfe7@mail.gmail.com>	
	<48D903B3.4090203@timgolden.me.uk>
	<d5ff27200809230801l66be4c33k275b4e111c43455e@mail.gmail.com>
Message-ID: <48D9061A.8010204@timgolden.me.uk>

Andrea Gavana wrote:
... No explanation was provided for the reason of the
> corrupted download (!)


There used to be a general warning on sites which offered
tgz for download that IE would do something weird with them.
I never experienced the problem myself, and in any case that
was years (and versions) ago so I assumed that whatever it
was had been fixed long before. Maybe not...

TJG

From p.f.moore at gmail.com  Tue Sep 23 17:36:06 2008
From: p.f.moore at gmail.com (Paul Moore)
Date: Tue, 23 Sep 2008 16:36:06 +0100
Subject: [python-win32] Invalid or corrupted installer...
In-Reply-To: <48D9061A.8010204@timgolden.me.uk>
References: <d5ff27200809230403y6bed6fdfw993d87bccad9ba89@mail.gmail.com>
	<48D8CF2E.6030903@timgolden.me.uk>
	<d5ff27200809230434oeb6c330pa0df72609ca2dfe7@mail.gmail.com>
	<48D903B3.4090203@timgolden.me.uk>
	<d5ff27200809230801l66be4c33k275b4e111c43455e@mail.gmail.com>
	<48D9061A.8010204@timgolden.me.uk>
Message-ID: <79990c6b0809230836p1c0a40cj365d8848d3f05dfe@mail.gmail.com>

2008/9/23 Tim Golden <mail at timgolden.me.uk>:
> Andrea Gavana wrote:
> ... No explanation was provided for the reason of the
>>
>> corrupted download (!)
>
>
> There used to be a general warning on sites which offered
> tgz for download that IE would do something weird with them.
> I never experienced the problem myself, and in any case that
> was years (and versions) ago so I assumed that whatever it
> was had been fixed long before. Maybe not...

Sorry, that was meant to go to the list:

What tends to happen is that something decides to automatically
uncompress, or "fix" the name, so you end up with a .tgz file which
actually only contains .tar data, or a .tar file which is gzipped as
well.

Renaming the file usually fixes this for me.

Paul

From imageguy1206 at gmail.com  Tue Sep 23 20:11:50 2008
From: imageguy1206 at gmail.com (geoff)
Date: Tue, 23 Sep 2008 14:11:50 -0400
Subject: [python-win32] Virtual print driver - similiar to winpdf
In-Reply-To: <48D8EB29.60502@co.marshall.ia.us>
References: <3ee0b280809211810j7d516500o9e93c9673ee82e2b@mail.gmail.com>
	<48D7D8E5.7070106@probo.com>
	<3ee0b280809221220q277e3c9ap262d84da3bce85a0@mail.gmail.com>
	<48D8EB29.60502@co.marshall.ia.us>
Message-ID: <3ee0b280809231111u3651a3f9n90ae1506f1e00226@mail.gmail.com>

> You might take a look at Pyla. It's a print to fax library. I think it's
> kind of what you're talking about. I've been trying to hack it off and on
> for a while now as we only need a little of its functionality here at work.
> There isn't much for docs, so you'll have to read the source. Here's a link:
> http://www.teamsw.it/pyla/
>
> Mike
>
Thanks Mike.
I think I am going to dig deeper into using  a Windows printer defined
as using the "Standard TCPIP Port" and my own program based on
SocketServer to grab to  output when it it created.  I set up a very
crude test last night and I think this will do exactly what i need -
ie can react immediately when the user is printing something, complete
access to the datastream and of course pure python solution for easy
maintenance  :-) .

keep you posted.

g.

From mail at timgolden.me.uk  Tue Sep 23 21:57:48 2008
From: mail at timgolden.me.uk (Tim Golden)
Date: Tue, 23 Sep 2008 20:57:48 +0100
Subject: [python-win32] Virtual print driver - similiar to winpdf
In-Reply-To: <3ee0b280809231111u3651a3f9n90ae1506f1e00226@mail.gmail.com>
References: <3ee0b280809211810j7d516500o9e93c9673ee82e2b@mail.gmail.com>	<48D7D8E5.7070106@probo.com>	<3ee0b280809221220q277e3c9ap262d84da3bce85a0@mail.gmail.com>	<48D8EB29.60502@co.marshall.ia.us>
	<3ee0b280809231111u3651a3f9n90ae1506f1e00226@mail.gmail.com>
Message-ID: <48D94A3C.5030304@timgolden.me.uk>

geoff wrote:
>> You might take a look at Pyla. It's a print to fax library. I think it's
>> kind of what you're talking about. I've been trying to hack it off and on
>> for a while now as we only need a little of its functionality here at work.
>> There isn't much for docs, so you'll have to read the source. Here's a link:
>> http://www.teamsw.it/pyla/
>>
>> Mike
>>
> Thanks Mike.
> I think I am going to dig deeper into using  a Windows printer defined
> as using the "Standard TCPIP Port" and my own program based on
> SocketServer to grab to  output when it it created.  I set up a very
> crude test last night and I think this will do exactly what i need -
> ie can react immediately when the user is printing something, complete
> access to the datastream and of course pure python solution for easy
> maintenance  :-) .

Sounds like a technique that's worth knowing about. Do keep
us posted, even if it turns out to be a dead-end.

TJG

From amicitas at gmail.com  Wed Sep 24 09:46:19 2008
From: amicitas at gmail.com (mir amicitas)
Date: Wed, 24 Sep 2008 00:46:19 -0700
Subject: [python-win32] Drag and drop of files (CF_HDROP)
In-Reply-To: <BLU0-SMTP267BDCF36F0926951F88F8A34B0@phx.gbl>
References: <BLU0-SMTP267BDCF36F0926951F88F8A34B0@phx.gbl>
Message-ID: <55e194140809240046oa8c2e81v2496a90b208760cf@mail.gmail.com>

>The last arg is output, and will be returned from the function.
>


In that case I am confused about the output coming out of DoDragDrop().

The output that I get for different cases:
drag canceled: 0
copy succeed: 1
move succeed: 0

I expect that move succeed would give me a value of 2
(shellcon.DROPEFFEC_MOVE).

With the output above I cannot tell weather a move succeded or the drop got
canceled.

This is not critical functionality for me (I don't need to know the end
result).  At this point I am just curious.

Below I am including the dragging code that I put together, just in case it
is useful to anyone. (if anyone happens to notice something horribly wrong
with it let me know though.)

- Novi


-------------------------------------------

import pythoncom
import win32con
import winerror
import win32clipboard

from win32com.server.util import NewEnum, wrap
from win32com.server.exception import COMException
from win32com.shell import shellcon

num_do_objects = 0

def WrapCOMObject(ob, iid=None):
    return wrap(ob, iid=iid, useDispatcher = 0)

def PackDROPFILESStructure(file_name_array):
    import struct
    """
    typedef struct _DROPFILES {
        DWORD pFiles;
        POINT pt;
        BOOL fNC;
        BOOL fWide; } DROPFILES, *LPDROPFILES;
    """

    file_name_buffer='\0'.join(file_name_array)+'\0\0'
    fmt="lllll%ss" %len(file_name_buffer)
    dropfiles=struct.pack(fmt, 20, 0, 0, 0, 0, file_name_buffer)

    return dropfiles

def RegisterClipboardFormats(qmimedata):
    """
    Here we extract all of the formats from the qmimedata object and
    register them.  This will give us an ID for each type.
    This will return a dictionary with the ID's and descriptions for
    each mime type.
    """
    mime_formats = qmimedata.formats()
    format_dictionary = dict()
    for format in mime_formats:
        id = win32clipboard.RegisterClipboardFormat(str(format))
        format_dictionary[id] = format

    return format_dictionary

def GetClipboardFormats(qmimedata):
    """
    What I want to do here is to check if any of the mime data
    in the QMimeData object can/should be converted into a standard
    windows clipboard format.
    If it can I will extract the data here.

    This essentially does what QWindowsMime probably does, but
    since QWindowsMime is not in PyQt4 I have to do this myself.
    """

    format_dictionary = dict()
    if qmimedata.hasText():
        id = win32con.CF_TEXT
        data = str(qmimedata.text())
        format_dictionary[id] = data

        id = win32con.CF_UNICODETEXT
        data = unicode(qmimedata.text())
        format_dictionary[id] = data

    if qmimedata.hasUrls():
        files = []
        for url in qmimedata.urls():
            file_name = url.toLocalFile()
            if file_name:
                files.append(str(file_name))

        if files:
            id = win32con.CF_HDROP
            format_dictionary[id] = files

    return format_dictionary

class IDropSource:
    _com_interfaces_ = [pythoncom.IID_IDropSource]
    _public_methods_ = ["QueryContinueDrag", "GiveFeedback"]
    def __init__(self):
        pass

    def __del__(self):
        pass

    def QueryContinueDrag(self, is_escape_pressed, modifier_key_state):
        if is_escape_pressed == True:
            return winerror.DRAGDROP_S_CANCEL
        if (modifier_key_state & win32con.MK_LBUTTON) == False and \
                (modifier_key_state & win32con.MK_RBUTTON) == False:
            return winerror.DRAGDROP_S_DROP

        return winerror.S_OK

    def GiveFeedback(self, effect):
        return winerror.DRAGDROP_S_USEDEFAULTCURSORS

class IDataObject:
    _com_interfaces_ = [pythoncom.IID_IDataObject]
    _public_methods_ = ["GetData", "GetDataHere", "QueryGetData" \
                            ,"GetCanonicalFormatEtc", "SetData",
"EnumFormatEtc" \
                            ,"DAdvise", "DUnadvise", "EnumDAdvise"]
    def __init__(self, qmimedata):
        global num_do_objects
        num_do_objects += 1
        self.qmimedata = qmimedata

        self.mime_formats = RegisterClipboardFormats(qmimedata)
        self.clipboard_formats = GetClipboardFormats(qmimedata)

        self.supported_formatetc = []
        for format in self.mime_formats:
            formatetc = format, None, pythoncom.DVASPECT_CONTENT, -1,
pythoncom.TYMED_HGLOBAL
            self.supported_formatetc.append(formatetc)
        for format in self.clipboard_formats:
            formatetc = format, None, pythoncom.DVASPECT_CONTENT, -1,
pythoncom.TYMED_HGLOBAL
            self.supported_formatetc.append(formatetc)

    def __del__(self):
        global num_do_objects
        num_do_objects -= 1

    def _query_interface_(self, iid):
        if iid==pythoncom.IID_IEnumFORMATETC:
            return NewEnum(self.supported_formatetc, iid=iid)

    def GetData(self, formatetc_in):
        #print 'GetData'
        ret_stg = None
        cf_in, target_in, aspect_in, index_in, tymed_in  = formatetc_in

        # Loop through the supported formats.
        # We do this as opposed to directly searching mime_formats &
clipboard_formats
        # mostly just to make sure the order is consistant.
        for formatetc in self.supported_formatetc:
            cf, target, aspect, index, tymed  = formatetc
            if (aspect_in == aspect) and \
                    (tymed_in & tymed) and \
                    (cf_in == cf):

                ret_stg = pythoncom.STGMEDIUM()

                if self.mime_formats.has_key(cf):
                    data = self.qmimedata.data(self.mime_formats[cf])
                    ret_stg.set(tymed, data)
                elif self.clipboard_formats.has_key(cf):
                    data = self.clipboard_formats[cf]
                    if cf == win32con.CF_TEXT or cf_in ==
win32con.CF_UNICODETEXT:
                        ret_stg.set(tymed, data)
                    elif cf == win32con.CF_HDROP:
                        dropfiles = PackDROPFILESStructure(data)
                        ret_stg.set(tymed, dropfiles)
                    else:
                        print "hystrixNtDrag.IDataObject.GetData:"
                        print "    Programming Error: id should be supported
but isn't"



        if ret_stg is None:
            raise COMException(hresult=winerror.E_NOTIMPL)

        return ret_stg


    def GetDataHere(self, formatetc):
        #print 'GetDataHere'
        raise COMException(hresult=winerror.E_NOTIMPL)

    def QueryGetData(self, formatetc_in):
        #print 'QueryGetData'
        cf_in, target_in, aspect_in, index_in, tymed_in  = formatetc_in
        for formatetc in self.supported_formatetc:
            cf, target, aspect, index, tymed  = formatetc
            if (aspect_in & aspect) and \
                    (tymed_in & tymed) and \
                    (cf_in == cf):
                return winerror.S_OK

        raise COMException(hresult=winerror.DV_E_FORMATETC)
        return winerror.DV_E_FORMATETC

    def GetCanonicalFormatEtc(self, formatetc):
        #print 'GetCanonicalFormatEtc'
        RaiseCOMException(winerror.DATA_S_SAMEFORMATETC)
        # return fe

    def SetData(self, formatetc, medium, fRelease):
        #print 'SetData'
        raise COMException(hresult=winerror.E_NOTIMPL)

    def EnumFormatEtc(self, direction):
        #print 'EnumFormatEtc'
        if direction != pythoncom.DATADIR_GET:
            raise COMException(hresult=winerror.E_NOTIMPL)
        return NewEnum(self.supported_formatetc,
iid=pythoncom.IID_IEnumFORMATETC)

    def DAdvise(self, formatetc, flags, sink):
        #print 'DAdvise'
        raise COMException(hresult=winerror.E_NOTIMPL)

    def DUnadvise(self, connection):
        #print 'DUnadvise'
        raise COMException(hresult=winerror.E_NOTIMPL)

    def EnumDAdvise(self):
        #print 'EnumDAdvise'
        raise COMException(hresult=winerror.E_NOTIMPL)

def DoDragDrop(qmimedata, allow_copy=True, allow_move=False):

    source = IDropSource()
    source = WrapCOMObject(source, iid=pythoncom.IID_IDropSource)

    data_object = IDataObject(qmimedata)
    data_object = WrapCOMObject(data_object, iid=pythoncom.IID_IDataObject)

    effect_mask = 0
    if allow_copy:
        effect_mask += shellcon.DROPEFFECT_COPY
    if allow_move:
        effect_mask += shellcon.DROPEFFECT_MOVE

    value = pythoncom.DoDragDrop(data_object, source, effect_mask)
    print value
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080924/7cdd9535/attachment-0001.htm>

From waxingfervor at hotmail.com  Wed Sep 24 14:47:00 2008
From: waxingfervor at hotmail.com (kurt munson)
Date: Wed, 24 Sep 2008 12:47:00 +0000
Subject: [python-win32] Controlling Excel via COM: errors with conditional
	formatting
Message-ID: <BAY130-W45C171BA1A64CFC03CB358DB450@phx.gbl>


I  have written a Python program to control MS Excel via win32com.client. This allows me to create and control an Excel spreadsheet.I want to use Excel's conditional formatting to color certain cells, but I can't get it to work.Here's the code I use:
 

import win32com.clientxl= win32com.client.Dispatch("Excel.Application")............channamesSheet.Cells(6,3).FormatConditions.Add() Type:=1, Operator:=4, Formula1:="=C5"channamesSheet.Cells(6,3).FormatConditions(1).Interior.ColorIndex = 3I got this code from a record and replay macro in Excel, then swapped out the xl constants for numeric values (1 and 4).This gives me a syntax error when running.I have tried both .Add and .Add(), since sometimes these () are necessary.What am I doing wrong?
_________________________________________________________________
Get more out of the Web. Learn 10 hidden secrets of Windows Live.
http://windowslive.com/connect/post/jamiethomson.spaces.live.com-Blog-cns!550F681DAD532637!5295.entry?ocid=TXT_TAGLM_WL_domore_092008
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080924/ce11bace/attachment.htm>

From Solomon.Zewdie.Altek at zf.com  Wed Sep 24 15:00:04 2008
From: Solomon.Zewdie.Altek at zf.com (Solomon.Zewdie.Altek at zf.com)
Date: Wed, 24 Sep 2008 15:00:04 +0200
Subject: [python-win32] Controlling Excel via COM: errors with
	conditionalformatting
References: <BAY130-W45C171BA1A64CFC03CB358DB450@phx.gbl>
Message-ID: <8A13418356E45E4784744371A8B4DD7971FFC2@frds00501.emea.zf-world.com>

How about this?:
 
        channamesSheet.Cells(6, 3).Font.FontSize = 10
        channamesSheet.Cells(6, 3).Font.ColorIndex = "blue"
        channamesSheet.Cells(6, 3).Value = "something..."
        channamesSheet.Cells(6, 3).Font.Name = "Arial"
        ...
 
I am not 100% sure whether it works, but i would try that!
 
Regards,
Solomon Z. 
 
 
Von: python-win32-bounces+solomon.zewdie.altek=zf.com at python.org im Auftrag von kurt munson
Gesendet: Mi 24.09.2008 14:47
An: python-win32 at python.org
Betreff: [python-win32] Controlling Excel via COM: errors with conditionalformatting


I  have written a Python program to control MS Excel via win32com.client. This allows me to create and control an Excel spreadsheet.

I want to use Excel's conditional formatting to color certain cells, but I can't get it to work.

Here's the code I use:
 


	import win32com.client
	xl= win32com.client.Dispatch("Excel.Application")
	....
	....
	....
	channamesSheet.Cells(6,3).FormatConditions.Add() Type:=1, Operator:=4, Formula1:="=C5"
	channamesSheet.Cells(6,3).FormatConditions(1).Interior.ColorIndex = 3



I got this code from a record and replay macro in Excel, then swapped out the xl constants for numeric values (1 and 4).

This gives me a syntax error when running.

I have tried both .Add and .Add(), since sometimes these () are necessary.

What am I doing wrong?

________________________________

Get more out of the Web. Learn 10 hidden secrets of Windows Live. Learn Now <http://windowslive.com/connect/post/jamiethomson.spaces.live.com-Blog-cns!550F681DAD532637!5295.entry?ocid=TXT_TAGLM_WL_getmore_092008>  
 
 

From Solomon.Zewdie.Altek at zf.com  Wed Sep 24 15:07:07 2008
From: Solomon.Zewdie.Altek at zf.com (Solomon.Zewdie.Altek at zf.com)
Date: Wed, 24 Sep 2008 15:07:07 +0200
Subject: [python-win32] Controlling Excel via COM: errors
	withconditionalformatting
References: <BAY130-W45C171BA1A64CFC03CB358DB450@phx.gbl>
	<8A13418356E45E4784744371A8B4DD7971FFC2@frds00501.emea.zf-world.com>
Message-ID: <8A13418356E45E4784744371A8B4DD7971FFC3@frds00501.emea.zf-world.com>

correction:
 
...ColorIndex = 3
(i think the value, that ColorIndex gets is an integer not a string)

 
________________________________

Von: python-win32-bounces+solomon.zewdie.altek=zf.com at python.org im Auftrag von Solomon.Zewdie.Altek at zf.com
Gesendet: Mi 24.09.2008 15:00
An: waxingfervor at hotmail.com; python-win32 at python.org
Betreff: Re: [python-win32] Controlling Excel via COM: errors withconditionalformatting



How about this?:

        channamesSheet.Cells(6, 3).Font.FontSize = 10
        channamesSheet.Cells(6, 3).Font.ColorIndex = "blue"
        channamesSheet.Cells(6, 3).Value = "something..."
        channamesSheet.Cells(6, 3).Font.Name = "Arial"
        ...

I am not 100% sure whether it works, but i would try that!

Regards,
Solomon Z.


Von: python-win32-bounces+solomon.zewdie.altek=zf.com at python.org im Auftrag von kurt munson
Gesendet: Mi 24.09.2008 14:47
An: python-win32 at python.org
Betreff: [python-win32] Controlling Excel via COM: errors with conditionalformatting


I  have written a Python program to control MS Excel via win32com.client. This allows me to create and control an Excel spreadsheet.

I want to use Excel's conditional formatting to color certain cells, but I can't get it to work.

Here's the code I use:



        import win32com.client
        xl= win32com.client.Dispatch("Excel.Application")
        ....
        ....
        ....
        channamesSheet.Cells(6,3).FormatConditions.Add() Type:=1, Operator:=4, Formula1:="=C5"
        channamesSheet.Cells(6,3).FormatConditions(1).Interior.ColorIndex = 3



I got this code from a record and replay macro in Excel, then swapped out the xl constants for numeric values (1 and 4).

This gives me a syntax error when running.

I have tried both .Add and .Add(), since sometimes these () are necessary.

What am I doing wrong?

________________________________

Get more out of the Web. Learn 10 hidden secrets of Windows Live. Learn Now <http://windowslive.com/connect/post/jamiethomson.spaces.live.com-Blog-cns!550F681DAD532637!5295.entry?ocid=TXT_TAGLM_WL_getmore_092008> 


_______________________________________________
python-win32 mailing list
python-win32 at python.org
http://mail.python.org/mailman/listinfo/python-win32



From mhammond at skippinet.com.au  Wed Sep 24 15:12:22 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Wed, 24 Sep 2008 23:12:22 +1000
Subject: [python-win32] Drag and drop of files (CF_HDROP)
In-Reply-To: <55e194140809240046oa8c2e81v2496a90b208760cf@mail.gmail.com>
References: <BLU0-SMTP267BDCF36F0926951F88F8A34B0@phx.gbl>
	<55e194140809240046oa8c2e81v2496a90b208760cf@mail.gmail.com>
Message-ID: <020801c91e47$30e512c0$92af3840$@com.au>

An interesting read is http://msdn.microsoft.com/en-us/library/bb776904.aspx
- it shows that lots of interesting state is communicated via SetData calls
with custom formats.  The IAsyncOperation interface is also an interesting
distraction.

 

Cheers,

 

Mark

 

From: python-win32-bounces+skippy.hammond=gmail.com at python.org
[mailto:python-win32-bounces+skippy.hammond=gmail.com at python.org] On Behalf
Of mir amicitas
Sent: Wednesday, 24 September 2008 5:46 PM
To: python-win32 at python.org
Subject: Re: [python-win32] Drag and drop of files (CF_HDROP)

 

>The last arg is output, and will be returned from the function.
>


In that case I am confused about the output coming out of DoDragDrop().

The output that I get for different cases:
drag canceled: 0
copy succeed: 1
move succeed: 0

I expect that move succeed would give me a value of 2
(shellcon.DROPEFFEC_MOVE).

With the output above I cannot tell weather a move succeded or the drop got
canceled.  

This is not critical functionality for me (I don't need to know the end
result).  At this point I am just curious.

Below I am including the dragging code that I put together, just in case it
is useful to anyone. (if anyone happens to notice something horribly wrong
with it let me know though.)

- Novi


-------------------------------------------

import pythoncom
import win32con
import winerror
import win32clipboard

from win32com.server.util import NewEnum, wrap
from win32com.server.exception import COMException
from win32com.shell import shellcon

num_do_objects = 0

def WrapCOMObject(ob, iid=None):
    return wrap(ob, iid=iid, useDispatcher = 0)

def PackDROPFILESStructure(file_name_array):
    import struct
    """
    typedef struct _DROPFILES {
        DWORD pFiles;
        POINT pt;
        BOOL fNC;
        BOOL fWide; } DROPFILES, *LPDROPFILES;
    """

    file_name_buffer='\0'.join(file_name_array)+'\0\0'
    fmt="lllll%ss" %len(file_name_buffer)
    dropfiles=struct.pack(fmt, 20, 0, 0, 0, 0, file_name_buffer)

    return dropfiles

def RegisterClipboardFormats(qmimedata):
    """
    Here we extract all of the formats from the qmimedata object and
    register them.  This will give us an ID for each type.
    This will return a dictionary with the ID's and descriptions for
    each mime type.
    """
    mime_formats = qmimedata.formats()
    format_dictionary = dict()
    for format in mime_formats:
        id = win32clipboard.RegisterClipboardFormat(str(format))
        format_dictionary[id] = format

    return format_dictionary

def GetClipboardFormats(qmimedata):
    """
    What I want to do here is to check if any of the mime data
    in the QMimeData object can/should be converted into a standard
    windows clipboard format.  
    If it can I will extract the data here.

    This essentially does what QWindowsMime probably does, but
    since QWindowsMime is not in PyQt4 I have to do this myself.
    """

    format_dictionary = dict()
    if qmimedata.hasText():
        id = win32con.CF_TEXT
        data = str(qmimedata.text())
        format_dictionary[id] = data

        id = win32con.CF_UNICODETEXT
        data = unicode(qmimedata.text())
        format_dictionary[id] = data

    if qmimedata.hasUrls():
        files = []
        for url in qmimedata.urls():
            file_name = url.toLocalFile()
            if file_name: 
                files.append(str(file_name))

        if files:
            id = win32con.CF_HDROP
            format_dictionary[id] = files 

    return format_dictionary

class IDropSource:
    _com_interfaces_ = [pythoncom.IID_IDropSource]
    _public_methods_ = ["QueryContinueDrag", "GiveFeedback"]
    def __init__(self):
        pass

    def __del__(self):
        pass

    def QueryContinueDrag(self, is_escape_pressed, modifier_key_state):
        if is_escape_pressed == True:
            return winerror.DRAGDROP_S_CANCEL
        if (modifier_key_state & win32con.MK_LBUTTON) == False and \
                (modifier_key_state & win32con.MK_RBUTTON) == False:
            return winerror.DRAGDROP_S_DROP

        return winerror.S_OK

    def GiveFeedback(self, effect):
        return winerror.DRAGDROP_S_USEDEFAULTCURSORS

class IDataObject:
    _com_interfaces_ = [pythoncom.IID_IDataObject]
    _public_methods_ = ["GetData", "GetDataHere", "QueryGetData" \
                            ,"GetCanonicalFormatEtc", "SetData",
"EnumFormatEtc" \
                            ,"DAdvise", "DUnadvise", "EnumDAdvise"]
    def __init__(self, qmimedata):
        global num_do_objects
        num_do_objects += 1
        self.qmimedata = qmimedata
        
        self.mime_formats = RegisterClipboardFormats(qmimedata)
        self.clipboard_formats = GetClipboardFormats(qmimedata)

        self.supported_formatetc = []
        for format in self.mime_formats:
            formatetc = format, None, pythoncom.DVASPECT_CONTENT, -1,
pythoncom.TYMED_HGLOBAL
            self.supported_formatetc.append(formatetc)
        for format in self.clipboard_formats:
            formatetc = format, None, pythoncom.DVASPECT_CONTENT, -1,
pythoncom.TYMED_HGLOBAL
            self.supported_formatetc.append(formatetc)

    def __del__(self):
        global num_do_objects
        num_do_objects -= 1

    def _query_interface_(self, iid):
        if iid==pythoncom.IID_IEnumFORMATETC:
            return NewEnum(self.supported_formatetc, iid=iid)

    def GetData(self, formatetc_in):
        #print 'GetData'
        ret_stg = None
        cf_in, target_in, aspect_in, index_in, tymed_in  = formatetc_in

        # Loop through the supported formats.
        # We do this as opposed to directly searching mime_formats &
clipboard_formats
        # mostly just to make sure the order is consistant.
        for formatetc in self.supported_formatetc:
            cf, target, aspect, index, tymed  = formatetc
            if (aspect_in == aspect) and \
                    (tymed_in & tymed) and \
                    (cf_in == cf):

                ret_stg = pythoncom.STGMEDIUM()

                if self.mime_formats.has_key(cf):
                    data = self.qmimedata.data(self.mime_formats[cf])
                    ret_stg.set(tymed, data)
                elif self.clipboard_formats.has_key(cf):
                    data = self.clipboard_formats[cf]
                    if cf == win32con.CF_TEXT or cf_in ==
win32con.CF_UNICODETEXT:
                        ret_stg.set(tymed, data)
                    elif cf == win32con.CF_HDROP:
                        dropfiles = PackDROPFILESStructure(data)
                        ret_stg.set(tymed, dropfiles)
                    else:
                        print "hystrixNtDrag.IDataObject.GetData:"
                        print "    Programming Error: id should be supported
but isn't"



        if ret_stg is None:
            raise COMException(hresult=winerror.E_NOTIMPL)

        return ret_stg
    

    def GetDataHere(self, formatetc):
        #print 'GetDataHere'
        raise COMException(hresult=winerror.E_NOTIMPL)

    def QueryGetData(self, formatetc_in):
        #print 'QueryGetData'
        cf_in, target_in, aspect_in, index_in, tymed_in  = formatetc_in
        for formatetc in self.supported_formatetc:
            cf, target, aspect, index, tymed  = formatetc
            if (aspect_in & aspect) and \
                    (tymed_in & tymed) and \
                    (cf_in == cf):
                return winerror.S_OK
            
        raise COMException(hresult=winerror.DV_E_FORMATETC)
        return winerror.DV_E_FORMATETC

    def GetCanonicalFormatEtc(self, formatetc):
        #print 'GetCanonicalFormatEtc'
        RaiseCOMException(winerror.DATA_S_SAMEFORMATETC)
        # return fe

    def SetData(self, formatetc, medium, fRelease):
        #print 'SetData'
        raise COMException(hresult=winerror.E_NOTIMPL)

    def EnumFormatEtc(self, direction):
        #print 'EnumFormatEtc'
        if direction != pythoncom.DATADIR_GET:
            raise COMException(hresult=winerror.E_NOTIMPL)
        return NewEnum(self.supported_formatetc,
iid=pythoncom.IID_IEnumFORMATETC)

    def DAdvise(self, formatetc, flags, sink):
        #print 'DAdvise'
        raise COMException(hresult=winerror.E_NOTIMPL)

    def DUnadvise(self, connection):
        #print 'DUnadvise'
        raise COMException(hresult=winerror.E_NOTIMPL)

    def EnumDAdvise(self):
        #print 'EnumDAdvise'
        raise COMException(hresult=winerror.E_NOTIMPL)

def DoDragDrop(qmimedata, allow_copy=True, allow_move=False):

    source = IDropSource()
    source = WrapCOMObject(source, iid=pythoncom.IID_IDropSource)

    data_object = IDataObject(qmimedata)
    data_object = WrapCOMObject(data_object, iid=pythoncom.IID_IDataObject)

    effect_mask = 0
    if allow_copy:
        effect_mask += shellcon.DROPEFFECT_COPY
    if allow_move:
        effect_mask += shellcon.DROPEFFECT_MOVE

    value = pythoncom.DoDragDrop(data_object, source, effect_mask)
    print value

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-win32/attachments/20080924/f200b952/attachment-0001.htm>

From timr at probo.com  Wed Sep 24 19:35:09 2008
From: timr at probo.com (Tim Roberts)
Date: Wed, 24 Sep 2008 10:35:09 -0700
Subject: [python-win32] Controlling Excel via COM: errors with
 conditional formatting
In-Reply-To: <BAY130-W45C171BA1A64CFC03CB358DB450@phx.gbl>
References: <BAY130-W45C171BA1A64CFC03CB358DB450@phx.gbl>
Message-ID: <48DA7A4D.4010706@probo.com>

kurt munson wrote:
> I  have written a Python program to control MS Excel via
> win32com.client. This allows me to create and control an Excel
> spreadsheet.
>
> I want to use Excel's conditional formatting to color certain cells,
> but I can't get it to work.
>
> Here's the code I use:
>
>     |import win32com.client
>     xl= win32com.client.Dispatch("Excel.Application")
>     ....
>     ....
>     ....
>     channamesSheet.Cells(6,3).FormatConditions.Add() Type:=1,
>     Operator:=4, Formula1:="=C5"
>     channamesSheet.Cells(6,3).FormatConditions(1).Interior.ColorIndex = 3|
>
> I got this code from a record and replay macro in Excel, then swapped
> out the xl constants for numeric values (1 and 4).
>
> This gives me a syntax error when running.
>
> I have tried both .Add and .Add(), since sometimes these () are necessary.
>
> What am I doing wrong?

What you are doing wrong is writing Python scripts as if it were Visual
Basic.  Solomon gave you the code, but you need to understand that
Python and Visual Basic are different languages.  When Excel records a
macro, it records it in Visual Basic.  The "Arg:=Value" syntax is a
Visual Basic syntax.  The "sometimes these () are necessary" comment
comes from Visual Basic.

By the way, after you use Dispatch, all of Excel's "xl" constants are
available to you.
    import win32com.client
    xl = win32com.client.dispatch( "Excel.Application" )
    print win32com.clients.constants.xlLandscape

-- 
Tim Roberts, timr at probo.com
Providenza & Boekelheide, Inc.


From Solomon.Zewdie.Altek at zf.com  Mon Sep 29 16:41:20 2008
From: Solomon.Zewdie.Altek at zf.com (Solomon.Zewdie.Altek at zf.com)
Date: Mon, 29 Sep 2008 16:41:20 +0200
Subject: [python-win32] error: 'CoInitialize not found'
Message-ID: <8A13418356E45E4784744371A8B4DD7971FFCA@frds00501.emea.zf-world.com>

Hi,
is someone among you guys out there familier with this kind of error message or rather knows the cause and the solution?:
 
"  File "F:\2_4_2\lib\site-packages\win32com\client\dynamic.py", line 78, in _GetGoodDispatch 
pywintypes.com_error: (-2147221008, 'CoInitialize not found.', None, None) "
 
 
What i am trying to do is: 
- instancing a class B (which i would like to use as an excel api) in class A
- creating an object of this class(B) and than
- using its Methods(like opening excel files, etc...)
 
 
Code in Class A:
...
   classB_obj = classB (InputFile,True,1)
   classB_obj.startExcel()
...
 
Code in Class B:
..
    def __init__(self, filename=None, visible, sheet):

       self.xlapp = win32com.client.dynamic.Dispatch("Excel.Application")
       self.workbook = xlFile = self.xlapp.Workbooks.Open(filename)
       self.xlapp.visible = visible
       self.worksheet = self.xlapp.Sheets(sheet) 
       ...
 
    def startExcel(self):
       ...
 
 
Unfortunately it doesn't work.
I have tested Class B stand a lone and it works....
So the problem seems to be somehow in the interaction between class A and Class B.
 
 
I am thankful for every suggestion and help!!!
 
Solomon
 
 

From mhammond at skippinet.com.au  Tue Sep 30 08:50:45 2008
From: mhammond at skippinet.com.au (Mark Hammond)
Date: Tue, 30 Sep 2008 16:50:45 +1000
Subject: [python-win32] Attributes on exceptions
Message-ID: <048801c922c8$ef1a4690$cd4ed3b0$@com.au>

It is way past time that the pywin32 exception objects grew attributes
instead of only being available via indexing.  In other words, instead of
writing:

  except win32api.error, exc:
    if exc[0] == ERROR_CODE: ...

you can say:

  except win32api.error, exc:
    if exc.winerror == ERROR_CODE: ...

Code that uses exception attributes will also port easier to py3k - but note
that I'm not proposing anyone *must* use exception attributes - any changes
will be backwards compatible.

So - a few questions:

* I've come up with the following for the names:

win32api.error, exc:
  exc.winerror == exc[0] - 'winerror' as that is what python 2.5 and later
use for the windows error code in its builtin exceptions.
  exc.funcname == exc[1] - 'funcname' as its close to 'filename', used by
python builtin exceptions.
  exc.strerror == exc[2] - 'strerror' again as that is what Python builtin
WindowsError exception uses, for example.

pythoncom.error, exc:
  exc.hresult == exc[0] - not 'winerror' as its not a windows error code -
it's a hresult.
  exc.strerror == exc[1] - strerror to be consistent.
  exc.excepinfo == exc[2] - named after the win32 EXCEPINFO struct.  Remains
a tuple without attributes.
  exc.argerror == exc[3] - no good reason for this name!

do they sound reasonable?  Speak now or forever hold your peace :)

* The implementation has caused a subtle change in the way an exception is
*printed* - eg, consider now:
  >>> win32api.CloseHandle(1)
  ...
  error: (6, 'CloseHandle', 'The handle is invalid.')
versus in the new world:
  >>> win32api.CloseHandle(1)
  ...
  pywintypes.error: (6, 'CloseHandle', 'The handle is invalid.')

Note the 'pywintypes.' prefix on the output; a com error will also be
identified as coming from 'pywintypes'.  The class name hasn't changed, just
the way the class is instantiated has (I'm cheating by defining the classes
in .py code embedded in a string in pywintypes!)

Does anyone see a problem with that?

Thanks,

Mark


From haraldarminmassa at gmail.com  Tue Sep 30 09:29:45 2008
From: haraldarminmassa at gmail.com (Harald Armin Massa)
Date: Tue, 30 Sep 2008 09:29:45 +0200
Subject: [python-win32] Attributes on exceptions
In-Reply-To: <048801c922c8$ef1a4690$cd4ed3b0$@com.au>
References: <048801c922c8$ef1a4690$cd4ed3b0$@com.au>
Message-ID: <7be3f35d0809300029i381592descb507c9bb617098c@mail.gmail.com>

Mark,

>  exc.excepinfo == exc[2] - named after the win32 EXCEPINFO struct.  Remains
> a tuple without attributes.

Would'nt it be a good point to use a named tuple, at least on Pythons
where that is available?

> Does anyone see a problem with that?
Not me.

Harald



-- 
GHUM Harald Massa
persuadere et programmare
Harald Armin Massa
Spielberger Stra?e 49
70435 Stuttgart
0173/9409607
no fx, no carrier pigeon
-
EuroPython 2009 will take place in Birmingham - Stay tuned!

From arve.knudsen at gmail.com  Tue Sep 30 11:17:14 2008
From: arve.knudsen at gmail.com (Arve Knudsen)
Date: Tue, 30 Sep 2008 11:17:14 +0200
Subject: [python-win32] Extending DLL search path?
Message-ID: <a0d6258d0809300217s75bbd99dw19fe00db0cbe7851@mail.gmail.com>

Hi

I have a C++ program which embeds Python and is itself modularized
into several DLLs. The Python part (extensions) of the application
again links to these DLLs, the problem is the containing catalog (also
containing the executable itself) is not on the PATH, so the DLLs
aren't resolved when loading the Python extensions. Is there any smart
way of extending the DLL search path at runtime?

Thanks,
Arve

From rdahlstrom at directedge.com  Tue Sep 30 13:18:46 2008
From: rdahlstrom at directedge.com (Dahlstrom, Roger)
Date: Tue, 30 Sep 2008 07:18:46 -0400
Subject: [python-win32] error: 'CoInitialize not found'
In-Reply-To: <8A13418356E45E4784744371A8B4DD7971FFCA@frds00501.emea.zf-world.com>
References: <8A13418356E45E4784744371A8B4DD7971FFCA@frds00501.emea.zf-world.com>
Message-ID: <DE02840141C15249B2FAE5A61B40EA1C014B56C4@jc1wsmail6.global.knight.com>

If you didn't get the answer already, you need to call pythoncom.CoInitialize() prior to the code that is throwing an error,
probably put it right after init. 

-----Original Message-----
From: python-win32-bounces+rdahlstrom=directedge.com at python.org [mailto:python-win32-bounces+rdahlstrom=directedge.com at python.org]
On Behalf Of Solomon.Zewdie.Altek at zf.com
Sent: Monday, September 29, 2008 10:41 AM
To: python-win32 at python.org
Subject: [python-win32] error: 'CoInitialize not found'

Hi,
is someone among you guys out there familier with this kind of error message or rather knows the cause and the solution?:
 
"  File "F:\2_4_2\lib\site-packages\win32com\client\dynamic.py", line 78, in _GetGoodDispatch 
pywintypes.com_error: (-2147221008, 'CoInitialize not found.', None, None) "
 
 
What i am trying to do is: 
- instancing a class B (which i would like to use as an excel api) in class A
- creating an object of this class(B) and than
- using its Methods(like opening excel files, etc...)
 
 
Code in Class A:
...
   classB_obj = classB (InputFile,True,1)
   classB_obj.startExcel()
...
 
Code in Class B:
..
    def __init__(self, filename=None, visible, sheet):

       self.xlapp = win32com.client.dynamic.Dispatch("Excel.Application")
       self.workbook = xlFile = self.xlapp.Workbooks.Open(filename)
       self.xlapp.visible = visible
       self.worksheet = self.xlapp.Sheets(sheet) 
       ...
 
    def startExcel(self):
       ...
 
 
Unfortunately it doesn't work.
I have tested Class B stand a lone and it works....
So the problem seems to be somehow in the interaction between class A and Class B.
 
 
I am thankful for every suggestion and help!!!
 
Solomon
 
 
_______________________________________________
python-win32 mailing list
python-win32 at python.org
http://mail.python.org/mailman/listinfo/python-win32


DISCLAIMER:
This e-mail, and any attachments thereto, is intended only for use by the addressee(s) named herein and
may contain legally privileged and/or confidential information. If you are not the intended recipient
of this e-mail, you are hereby notified that any dissemination, distribution or copying of this e-mail, and 
any attachments thereto, is strictly prohibited. If you have received this in error, please immediately notify 
me and permanently delete the original and any copy of any e-mail and any printout thereof. 
E-mail transmission cannot be guaranteed to be secure or error-free. The sender therefore does not accept 
liability for any errors or omissions in the contents of this message which arise as a result of e-mail transmission.

NOTICE REGARDING PRIVACY AND CONFIDENTIALITY
Direct Edge ECN LLC may, at its discretion, monitor and review the content of all e-mail communications.

www.directedge.com

From Solomon.Zewdie.Altek at zf.com  Tue Sep 30 13:34:17 2008
From: Solomon.Zewdie.Altek at zf.com (Solomon.Zewdie.Altek at zf.com)
Date: Tue, 30 Sep 2008 13:34:17 +0200
Subject: [python-win32] error: 'CoInitialize not found'
In-Reply-To: <DE02840141C15249B2FAE5A61B40EA1C014B56C4@jc1wsmail6.global.knight.com>
Message-ID: <8A13418356E45E4784744371A8B4DD7971B0CD@frds00501.emea.zf-world.com>

Yes, I found the similar solution at google and it works now!!!
But tnx, anyhow!



-----Urspr?ngliche Nachricht-----
Von: Dahlstrom, Roger [mailto:rdahlstrom at directedge.com] 
Gesendet: Dienstag, 30. September 2008 13:19
An: Zewdie Solomon EXT FRD ALTEK; python-win32 at python.org
Betreff: RE: [python-win32] error: 'CoInitialize not found'


If you didn't get the answer already, you need to call pythoncom.CoInitialize() prior to the code that is throwing an error, probably put it right after init. 

-----Original Message-----
From: python-win32-bounces+rdahlstrom=directedge.com at python.org [mailto:python-win32-bounces+rdahlstrom=directedge.com at python.org]
On Behalf Of Solomon.Zewdie.Altek at zf.com
Sent: Monday, September 29, 2008 10:41 AM
To: python-win32 at python.org
Subject: [python-win32] error: 'CoInitialize not found'

Hi,
is someone among you guys out there familier with this kind of error message or rather knows the cause and the solution?:
 
"  File "F:\2_4_2\lib\site-packages\win32com\client\dynamic.py", line 78, in _GetGoodDispatch 
pywintypes.com_error: (-2147221008, 'CoInitialize not found.', None, None) "
 
 
What i am trying to do is: 
- instancing a class B (which i would like to use as an excel api) in class A
- creating an object of this class(B) and than
- using its Methods(like opening excel files, etc...)
 
 
Code in Class A:
...
   classB_obj = classB (InputFile,True,1)
   classB_obj.startExcel()
...
 
Code in Class B:
..
    def __init__(self, filename=None, visible, sheet):

       self.xlapp = win32com.client.dynamic.Dispatch("Excel.Application")
       self.workbook = xlFile = self.xlapp.Workbooks.Open(filename)
       self.xlapp.visible = visible
       self.worksheet = self.xlapp.Sheets(sheet) 
       ...
 
    def startExcel(self):
       ...
 
 
Unfortunately it doesn't work.
I have tested Class B stand a lone and it works....
So the problem seems to be somehow in the interaction between class A and Class B.
 
 
I am thankful for every suggestion and help!!!
 
Solomon
 
 
_______________________________________________
python-win32 mailing list
python-win32 at python.org http://mail.python.org/mailman/listinfo/python-win32


DISCLAIMER:
This e-mail, and any attachments thereto, is intended only for use by the addressee(s) named herein and may contain legally privileged and/or confidential information. If you are not the intended recipient of this e-mail, you are hereby notified that any dissemination, distribution or copying of this e-mail, and 
any attachments thereto, is strictly prohibited. If you have received this in error, please immediately notify 
me and permanently delete the original and any copy of any e-mail and any printout thereof. 
E-mail transmission cannot be guaranteed to be secure or error-free. The sender therefore does not accept 
liability for any errors or omissions in the contents of this message which arise as a result of e-mail transmission.

NOTICE REGARDING PRIVACY AND CONFIDENTIALITY
Direct Edge ECN LLC may, at its discretion, monitor and review the content of all e-mail communications.

www.directedge.com

From le.dahut at laposte.net  Tue Sep 30 15:41:10 2008
From: le.dahut at laposte.net (le dahut)
Date: Tue, 30 Sep 2008 15:41:10 +0200
Subject: [python-win32] NT service and login prompt
Message-ID: <48E22C76.2090908@laposte.net>

Hello,
Is it possible to tell windows to wait a python service has started 
before the login prompt is displayed ?