From jeff at jab.org  Sat Nov  3 22:44:19 2007
From: jeff at jab.org (Jeff Breidenbach)
Date: Sat, 3 Nov 2007 14:44:19 -0700
Subject: [Mailman-Developers] Improving the archives
In-Reply-To: <0673D829-3875-4CC7-9175-D234F880D874@astro.princeton.edu>
References: <0673D829-3875-4CC7-9175-D234F880D874@astro.princeton.edu>
Message-ID: <e03b90ae0711031444u7f3997e2pcfbb402496d6be5c@mail.gmail.com>

>but if you can trust yourself to generate them, consecutive
>integers provide minimal, order-preserving, perfect hashing, too!

Hmm.... this sounds pretty sensible to me.

Jeff

From stephen at xemacs.org  Sun Nov  4 01:42:47 2007
From: stephen at xemacs.org (Stephen J. Turnbull)
Date: Sun, 04 Nov 2007 09:42:47 +0900
Subject: [Mailman-Developers] Improving the archives
In-Reply-To: <0673D829-3875-4CC7-9175-D234F880D874@astro.princeton.edu>
References: <0673D829-3875-4CC7-9175-D234F880D874@astro.princeton.edu>
Message-ID: <87d4uqlu2g.fsf@uwakimon.sk.tsukuba.ac.jp>

Craig Loomis writes:

 >    Globally unique IDs, hashed IDs, etc., are very appealing from  
 > various CS-y and techie points of view, but are simply not memorable  
 > to humans or knowable by dumb external programs. I think as much, or  
 > more, effort should be put into delivering a straightforwardly useable  
 > naming scheme as goes into making an arbitrary message recoverable  
 > from anywhere.  Basically, "friendly URLs" should be a primary  
 > requirement, not an optional afterthought for careless geeks like me  
 > to get wrong later....

Friendly URLs *are* a primary requirement.  The point is that to make
them *reliable* as well, either a globally unique ID is needed, or
individual site admins must suffer through hard-to-document
constraints on what they can do with their archives.  Note that the
system you describe based on the post_id member demonstrates the value
of a unique ID.

"Sufficient reliability" is not a tough requirement for an individual
admin to achieve, as you have demonstrated.  It's much more exacting
for the Mailman developers, who need to satisfy both sites with
different needs *and* archivers with different features.

 >    As an aside on other discussions, can you get away without using  
 > Message-ID or Date?

No.  Not all recipients of the messages get them through the list.
Once again, Mailman developers have to consider that situation, while
in your situation you may not need to worry about it.

From barry at list.org  Sun Nov 18 19:00:48 2007
From: barry at list.org (Barry Warsaw)
Date: Sun, 18 Nov 2007 13:00:48 -0500
Subject: [Mailman-Developers] [Mailman-checkins] [Branch
	~mailman-coders/mailman/3.0]
In-Reply-To: <20071107020714.29820.58368.launchpad@canonical@forster.canonical.com>
References: <20071107020714.29820.58368.launchpad@canonical@forster.canonical.com>
Message-ID: <A5B6388B-992B-4072-9826-449221EFA589@list.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Nov 6, 2007, at 9:07 PM, noreply at launchpad.net wrote:

> ------------------------------------------------------------
> revno: 6573
> committer: Mark Sapiro <msapiro at value.net>
> branch nick: 3.0
> timestamp: Tue 2007-11-06 18:03:28 -0800
> message:
>   - Scrubber.py
>     Fixed an issue where an implicit text/plain part without any  
> headers gets
>     lost.
>
>     Moved the cleansing of the filename extension to a place where  
> it is
>     guaranteed to be a string as opposed to an empty list.
> modified:
>   Mailman/Handlers/Scrubber.py
>
> === modified file 'Mailman/Handlers/Scrubber.py'
> --- a/Mailman/Handlers/Scrubber.py	2007-10-31 21:38:51 +0000
> +++ b/Mailman/Handlers/Scrubber.py	2007-11-07 02:03:28 +0000
> @@ -46,7 +46,7 @@
>  # Path characters for common platforms
>  pre = re.compile(r'[/\\:]')
>  # All other characters to strip out of Content-Disposition: filenames
> -# (essentially anything that isn't an alphanum, dot, slash, or  
> underscore.
> +# (essentially anything that isn't an alphanum, dot, dash, or  
> underscore).
>  sre = re.compile(r'[^-\w.]')
>  # Regexp to strip out leading dots
>  dre = re.compile(r'^\.*')
> @@ -265,7 +265,7 @@
>          # If the message isn't a multipart, then we'll strip it  
> out as an
>          # attachment that would have to be separately downloaded.   
> Pipermail
>          # will transform the url into a hyperlink.
> -        elif part and not part.is_multipart():
> +        elif part._payload and not part.is_multipart():

I'm not sure it's a good idea to use this non-public attribute.   
'part.get_payload()' would probably be a better test.

- -Barry



-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iD8DBQFHQH3Q2YZpQepbvXERAh5kAJ9vkjyCoAV10aYaNEDE34i9KNKQxgCgjF3p
c7yN8r7OSuvkTjVUXa75zKA=
=V12k
-----END PGP SIGNATURE-----

From barry at list.org  Sun Nov 18 19:28:26 2007
From: barry at list.org (Barry Warsaw)
Date: Sun, 18 Nov 2007 13:28:26 -0500
Subject: [Mailman-Developers] [Mailman-checkins] [Branch
	~mailman-coders/mailman/2.1]
In-Reply-To: <20071005014611.28380.40188.launchpad@canonical@forster.canonical.com>
References: <20071005014611.28380.40188.launchpad@canonical@forster.canonical.com>
Message-ID: <7104340B-1E29-4C96-A6AB-3B663134A8A0@list.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Oct 4, 2007, at 9:46 PM, noreply at launchpad.net wrote:

> ------------------------------------------------------------
> revno: 992
> committer: Mark Sapiro <msapiro at value.net>
> branch nick: 2.1
> timestamp: Thu 2007-10-04 18:35:10 -0700
> message:
>   /cygdrive/c/MM_bzr/log.txt
> modified:
>   Mailman/versions.py
>
> === modified file 'Mailman/versions.py'
> --- a/Mailman/versions.py	2005-12-30 18:50:08 +0000
> +++ b/Mailman/versions.py	2007-10-05 01:35:10 +0000
> @@ -483,10 +483,11 @@
>              # blow away the original timestamp and request id.   
> This means the
>              # request will live a little longer than it possibly  
> should have,
>              # but that's no big deal.
> +            import email
>              for p in v:
>                  author, text = p[2]
>                  reason = p[3]
> -                msg = Message.OutgoingMessage(text)
> +                msg = email.message_from_string(text,  
> Message.Message)
>                  l.HoldMessage(msg, reason)
>              del r[k]
>          elif k == 'add_member':

Should the import happen at the module global level?

- -Barry

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iD8DBQFHQIRK2YZpQepbvXERAjRQAJ9/yhfNVNoxMY+V3jS9p7rZka5QmQCgrNpS
ZXyQ0KC915TCjkwM7ECdeOc=
=M8JK
-----END PGP SIGNATURE-----

From msapiro at value.net  Sun Nov 18 19:30:05 2007
From: msapiro at value.net (Mark Sapiro)
Date: Sun, 18 Nov 2007 10:30:05 -0800
Subject: [Mailman-Developers] [Mailman-checkins]
	[Branch	~mailman-coders/mailman/3.0]
In-Reply-To: <A5B6388B-992B-4072-9826-449221EFA589@list.org>
References: <20071107020714.29820.58368.launchpad@canonical@forster.canonical.com>
	<A5B6388B-992B-4072-9826-449221EFA589@list.org>
Message-ID: <474084AD.9070305@value.net>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Barry Warsaw wrote:
> On Nov 6, 2007, at 9:07 PM, noreply at launchpad.net wrote:
> 
>> ------------------------------------------------------------
>> revno: 6573
>> committer: Mark Sapiro <msapiro at value.net>
<snip>
>>          # If the message isn't a multipart, then we'll strip it  
>> out as an
>>          # attachment that would have to be separately downloaded.   
>> Pipermail
>>          # will transform the url into a hyperlink.
>> -        elif part and not part.is_multipart():
>> +        elif part._payload and not part.is_multipart():
> 
> I'm not sure it's a good idea to use this non-public attribute.   
> 'part.get_payload()' would probably be a better test.


I think Barry is right, but I did it as I did because of a concern
expressed by Tokio at
<http://mail.python.org/pipermail/mailman-users/2006-December/054961.html>.

I'll gladly change it to get_payload()

- --
Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)

iD8DBQFHQIStVVuXXpU7hpMRAlVhAJ9M5OX2iT3Blt5QzQ2UPBuRY+1XUACg4rHg
MZywGERykk1M7brgkGDtUcA=
=bXYJ
-----END PGP SIGNATURE-----

From barry at list.org  Sun Nov 18 19:37:53 2007
From: barry at list.org (Barry Warsaw)
Date: Sun, 18 Nov 2007 13:37:53 -0500
Subject: [Mailman-Developers] [Mailman-checkins]
	[Branch	~mailman-coders/mailman/3.0]
In-Reply-To: <474084AD.9070305@value.net>
References: <20071107020714.29820.58368.launchpad@canonical@forster.canonical.com>
	<A5B6388B-992B-4072-9826-449221EFA589@list.org>
	<474084AD.9070305@value.net>
Message-ID: <9E11C301-4B42-498E-A2FE-402C70470D4E@list.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Nov 18, 2007, at 1:30 PM, Mark Sapiro wrote:

> I think Barry is right, but I did it as I did because of a concern
> expressed by Tokio at
> <http://mail.python.org/pipermail/mailman-users/2006-December/ 
> 054961.html>.
>
> I'll gladly change it to get_payload()

Tokio has a point (and it's something I want to improve in the next  
version of the email package), but with decode=False, get_payload()  
isn't too bad.

- -Barry

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iD8DBQFHQIaB2YZpQepbvXERAp7rAJ47da4MPipdBcdY3N0oAB5L6AtTSgCgmQZh
RwjUFOjNg/LXnSvURzoIiEo=
=cQSq
-----END PGP SIGNATURE-----

From mandre at it.dcs.ch  Sun Nov 18 15:55:05 2007
From: mandre at it.dcs.ch (Mathias Andre)
Date: Sun, 18 Nov 2007 15:55:05 +0100
Subject: [Mailman-Developers] Problem with Proxying and Utils.GetRequestURI
Message-ID: <20071118145505.GA27542@eso.org>

Dear all, 

In my company we're currently running Mailman 2.1.9.
The setup is as follow:

We have a dedicated Mailman server list.example.com 
For several reason we have been asked to make use of our main web server to
proxy the request to the Mailman server (this allow us to use one single
certificate SSL) so on our main web server www.example.com we have the
following Apache configuration:

ProxyPass        /lists    http://list.example.com/cgi-bin/mailman/
ProxyPassReverse /lists    http://list.example.com/cgi-bin/mailman/

So we can acess the Mailman web interfaces via for example:
http://www.example.com/lists/listinfo/mailman

This works just fine, we 'renamed' all the lists using fix_url and everything
went ok, all the links are fine.


Well, except for one: the ACTION= entry in all the login forms (admin,
admindb...)

After a bit of investigation I identified the source of the 'problem', in
Mailman/Cgi/Auth.py in def loginpage():

As it's a frontpage the actionurl comes from Utils.GetRequestURI(url), which
in turns uses the REQUEST_URI environment variable.
In our case this variable is set to the *local* Apache path, ie the path on
list.example.com instead of the path on www.example.com, so we have
/cgi-bin/mailman//admin/mailman
instead of:
/lists/admin/mailman

So the web browser then tries to send the form's data to
www.example.com/cgi-bin/mailman//admin/mailman which obviously doesn't exist.

We temporarily fixed our problem by forcing loginpage() to always return
actionurl = url whether or not frontpage is True or not.
This doesn't seem to have broken anything else but obviously I would like to
find a cleaner solution.

So I have a couple of questions: 
Are we doing something wrong in our setup? Why does Mailman GetRequestURI if
frontpage is True? 


I hope that someone will be able to enlighten us :)

Many thanks in advance for this, and for the great piece of software!

Mathias Andre



-- 
  *    Mathias Andre                   Phone : +49 (0)89 3200 6760
*E S*  European Southern Observatory   Fax   : +49 (0)89 3200 6380
  O    Karl Schwarzschildstr. 2        Office 213/6
  *    D-85748 Garching
       Germany
"For every problem, there is a solution that is simple, neat, and wrong."
H. L. Mencken


From barry at list.org  Sun Nov 18 22:44:23 2007
From: barry at list.org (Barry Warsaw)
Date: Sun, 18 Nov 2007 16:44:23 -0500
Subject: [Mailman-Developers] [Mailman-checkins] [Branch
	~mailman-coders/mailman/2.1]
In-Reply-To: <20071118214021.30727.41814.launchpad@canonical@forster.canonical.com>
References: <20071118214021.30727.41814.launchpad@canonical@forster.canonical.com>
Message-ID: <A12D51C2-B23B-4356-AFF1-1E148FFBA640@list.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Mark,

On Nov 18, 2007, at 4:40 PM, noreply at launchpad.net wrote:

> === modified file 'Mailman/Gui/Privacy.py'
> --- a/Mailman/Gui/Privacy.py	2005-12-30 18:50:08 +0000
> +++ b/Mailman/Gui/Privacy.py	2007-11-18 20:01:26 +0000
> @@ -1,17 +1,17 @@
> -# Copyright (C) 2001-2005 by the Free Software Foundation, Inc.
> +# Copyright (C) 2001-2007 by the Free Software Foundation, Inc.
>  #
>  # This program is free software; you can redistribute it and/or
>  # modify it under the terms of the GNU General Public License
>  # as published by the Free Software Foundation; either version 2
>  # of the License, or (at your option) any later version.
> -#
> +#
>  # This program is distributed in the hope that it will be useful,
>  # but WITHOUT ANY WARRANTY; without even the implied warranty of
>  # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>  # GNU General Public License for more details.
> -#
> +#
>  # You should have received a copy of the GNU General Public License
> -# along with this program; if not, write to the Free Software
> +# along with this program; if not, write to the Free Software
>  # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  
> 02110-1301,
>  # USA.
>
> @@ -59,7 +59,7 @@
>                              _('Confirm'),
>                              _('Require approval'),
>                              _('Confirm and approve')),
> -                           0,
> +                           0,
>                             _('What steps are required for  
> subscription?<br>'),
>                             _('''None - no verification steps (<em>Not
>                             Recommended </em>)<br>
> @@ -67,7 +67,7 @@
>                             Require approval - require list  
> administrator
>                             Approval for subscriptions <br>
>                             Confirm and approve - both confirm and  
> approve
> -
> +
>                             <p>(*) when someone requests a  
> subscription,
>                             Mailman sends them a notice with a unique
>                             subscription request number that they  
> must reply to
> @@ -88,7 +88,7 @@
>                             Require approval - require list  
> administrator
>                             approval for subscriptions <br>
>                             Confirm and approve - both confirm and  
> approve
> -
> +
>                             <p>(*) when someone requests a  
> subscription,
>                             Mailman sends them a notice with a unique
>                             subscription request number that they  
> must reply to

These changes affect all translations because they change the source  
string.  Since translators will get the new pot file soon, this might  
be okay, but I just wanted to point that out.

Cheers,
- -Barry

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iD8DBQFHQLI32YZpQepbvXERAlbIAJ9285htmRgaa/bEhdECBJGidBXUfQCglMI5
As32OIHFWg9UvNkk/jhkKTQ=
=mIJw
-----END PGP SIGNATURE-----

From msapiro at value.net  Mon Nov 19 03:28:20 2007
From: msapiro at value.net (Mark Sapiro)
Date: Sun, 18 Nov 2007 18:28:20 -0800
Subject: [Mailman-Developers] [Mailman-checkins] [Branch
	~mailman-coders/mailman/2.1]
In-Reply-To: <A12D51C2-B23B-4356-AFF1-1E148FFBA640@list.org>
References: <20071118214021.30727.41814.launchpad@canonical@forster.canonical.com>
	<A12D51C2-B23B-4356-AFF1-1E148FFBA640@list.org>
Message-ID: <4740F4C4.2010503@value.net>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Barry Warsaw wrote:
> 
> On Nov 18, 2007, at 4:40 PM, noreply at launchpad.net wrote:
> 
>> === modified file 'Mailman/Gui/Privacy.py'
>> --- a/Mailman/Gui/Privacy.py	2005-12-30 18:50:08 +0000
>> +++ b/Mailman/Gui/Privacy.py	2007-11-18 20:01:26 +0000
<snip>
>> @@ -67,7 +67,7 @@
>>                             Require approval - require list  
>> administrator
>>                             Approval for subscriptions <br>
>>                             Confirm and approve - both confirm and  
>> approve
>> -
>> +
>>                             <p>(*) when someone requests a  
>> subscription,
>>                             Mailman sends them a notice with a unique
>>                             subscription request number that they  
>> must reply to
>> @@ -88,7 +88,7 @@
>>                             Require approval - require list  
>> administrator
>>                             approval for subscriptions <br>
>>                             Confirm and approve - both confirm and  
>> approve
>> -
>> +
>>                             <p>(*) when someone requests a  
>> subscription,
>>                             Mailman sends them a notice with a unique
>>                             subscription request number that they  
>> must reply to
> 
> These changes affect all translations because they change the source  
> string.  Since translators will get the new pot file soon, this might  
> be okay, but I just wanted to point that out.


Thanks Barry,

I'm trying to do too many things at once. I didn't consider the effect
on the translations. I have backed out the changes that were just
gratuitous trailing whitespace removal in i18n'd messages and
regenerated the mailman.pot.

- --
Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)

iD8DBQFHQPTEVVuXXpU7hpMRAipDAJwNJShkw7pl3gyUdazD2uK+SHF6jQCcDVsi
EJXHEzStb+LLOwExSbtDg4A=
=E/PX
-----END PGP SIGNATURE-----

From msapiro at value.net  Mon Nov 19 04:01:25 2007
From: msapiro at value.net (Mark Sapiro)
Date: Sun, 18 Nov 2007 19:01:25 -0800
Subject: [Mailman-Developers] Problem with Proxying and
	Utils.GetRequestURI
In-Reply-To: <20071118145505.GA27542@eso.org>
References: <20071118145505.GA27542@eso.org>
Message-ID: <4740FC85.4020206@value.net>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Mathias Andre wrote:
> 
> We have a dedicated Mailman server list.example.com 
> For several reason we have been asked to make use of our main web server to
> proxy the request to the Mailman server (this allow us to use one single
> certificate SSL) so on our main web server www.example.com we have the
> following Apache configuration:
> 
> ProxyPass        /lists    http://list.example.com/cgi-bin/mailman/
> ProxyPassReverse /lists    http://list.example.com/cgi-bin/mailman/
> 
> So we can acess the Mailman web interfaces via for example:
> http://www.example.com/lists/listinfo/mailman
> 
> This works just fine, we 'renamed' all the lists using fix_url and everything
> went ok, all the links are fine.
> 
> 
> Well, except for one: the ACTION= entry in all the login forms (admin,
> admindb...)
> 
> After a bit of investigation I identified the source of the 'problem', in
> Mailman/Cgi/Auth.py in def loginpage():
> 
> As it's a frontpage the actionurl comes from Utils.GetRequestURI(url), which
> in turns uses the REQUEST_URI environment variable.


Actually, that's backwards. actionurl comes from
Utils.GetRequestURI(url) if frontpage is false.


> In our case this variable is set to the *local* Apache path, ie the path on
> list.example.com instead of the path on www.example.com, so we have
> /cgi-bin/mailman//admin/mailman
> instead of:
> /lists/admin/mailman


I think that's all the local web server can do. I don't think there's
anything in the CGI environment that could get the original URI path as
sent to the proxy server.


> So the web browser then tries to send the form's data to
> www.example.com/cgi-bin/mailman//admin/mailman which obviously doesn't exist.
> 
> We temporarily fixed our problem by forcing loginpage() to always return
> actionurl = url whether or not frontpage is True or not.
> This doesn't seem to have broken anything else but obviously I would like to
> find a cleaner solution.


I think you'll find it breaks one thing. Namely if someone tries to go
to a sub-page, e.g
<http://www.example.com/lists/admin/listname/privacy/sender>, without
having logged in, they should be taken to that page upon posting the
password. In your case, it doesn't work anyway, but in general, always
setting frontpage true will cause the login page to always return to the
default <http://www.example.com/lists/admin/listname> page.


> So I have a couple of questions: 
> Are we doing something wrong in our setup? Why does Mailman GetRequestURI if
> frontpage is True? 


Actually, false, but the reason is as I explain above. The only time
loginpage() is called with frontpage true is in response to the logout
request so you go to the default page if you (or someone else) log in again.

- --
Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)

iD8DBQFHQPyFVVuXXpU7hpMRAhzMAKC5MIUqPx9Vzf4dtF8CxvTeVL03GQCgi/jz
/ZSUa896yQdE1gd+tILGfKc=
=LYws
-----END PGP SIGNATURE-----

From msapiro at value.net  Mon Nov 19 04:38:39 2007
From: msapiro at value.net (Mark Sapiro)
Date: Sun, 18 Nov 2007 19:38:39 -0800
Subject: [Mailman-Developers] [Mailman-checkins]
	[Branch	~mailman-coders/mailman/2.1]
In-Reply-To: <7104340B-1E29-4C96-A6AB-3B663134A8A0@list.org>
References: <20071005014611.28380.40188.launchpad@canonical@forster.canonical.com>
	<7104340B-1E29-4C96-A6AB-3B663134A8A0@list.org>
Message-ID: <4741053F.5050801@value.net>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Barry Warsaw wrote:
> On Oct 4, 2007, at 9:46 PM, noreply at launchpad.net wrote:
> 
>> ------------------------------------------------------------
>> revno: 992
>> committer: Mark Sapiro <msapiro at value.net>
>> branch nick: 2.1
>> timestamp: Thu 2007-10-04 18:35:10 -0700
>> message:
>>   /cygdrive/c/MM_bzr/log.txt
>> modified:
>>   Mailman/versions.py
> 
>> === modified file 'Mailman/versions.py'
>> --- a/Mailman/versions.py	2005-12-30 18:50:08 +0000
>> +++ b/Mailman/versions.py	2007-10-05 01:35:10 +0000
>> @@ -483,10 +483,11 @@
>>              # blow away the original timestamp and request id.   
>> This means the
>>              # request will live a little longer than it possibly  
>> should have,
>>              # but that's no big deal.
>> +            import email
>>              for p in v:
>>                  author, text = p[2]
>>                  reason = p[3]
>> -                msg = Message.OutgoingMessage(text)
>> +                msg = email.message_from_string(text,  
>> Message.Message)
>>                  l.HoldMessage(msg, reason)
>>              del r[k]
>>          elif k == 'add_member':
> 
> Should the import happen at the module global level?


I thought about that. Our standards of course say yes, but this
particular import only occurs when updating a Mailman 1.0.x list which I
hope happens very infrequently. However I don't really feel strongly
either way.

- --
Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)

iD8DBQFHQQU/VVuXXpU7hpMRAgKkAKDu+323UJuX81GEFrykXkh+SfxdNgCgx0Sy
YtFAMJL5vdUfNV538VNl0uY=
=aqO7
-----END PGP SIGNATURE-----

From barry at list.org  Mon Nov 19 14:06:28 2007
From: barry at list.org (Barry Warsaw)
Date: Mon, 19 Nov 2007 08:06:28 -0500
Subject: [Mailman-Developers] [Mailman-checkins] [Branch
	~mailman-coders/mailman/2.1]
In-Reply-To: <4740F4C4.2010503@value.net>
References: <20071118214021.30727.41814.launchpad@canonical@forster.canonical.com>
	<A12D51C2-B23B-4356-AFF1-1E148FFBA640@list.org>
	<4740F4C4.2010503@value.net>
Message-ID: <64C90A51-0533-4E43-B065-C6140AB37C46@list.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Nov 18, 2007, at 9:28 PM, Mark Sapiro wrote:

> Thanks Barry,
>
> I'm trying to do too many things at once. I didn't consider the effect
> on the translations. I have backed out the changes that were just
> gratuitous trailing whitespace removal in i18n'd messages and
> regenerated the mailman.pot.

You rock Mark!
- -Barry

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iD8DBQFHQYpU2YZpQepbvXERAkLxAJ46JqBkXw0ZesiH1aOCN8r2PcebyQCgjUiY
vr/H8IFFSBcQWS01P6fd9nI=
=bJML
-----END PGP SIGNATURE-----

From barry at list.org  Mon Nov 19 14:24:39 2007
From: barry at list.org (Barry Warsaw)
Date: Mon, 19 Nov 2007 08:24:39 -0500
Subject: [Mailman-Developers] [Mailman-checkins]
	[Branch	~mailman-coders/mailman/2.1]
In-Reply-To: <4741053F.5050801@value.net>
References: <20071005014611.28380.40188.launchpad@canonical@forster.canonical.com>
	<7104340B-1E29-4C96-A6AB-3B663134A8A0@list.org>
	<4741053F.5050801@value.net>
Message-ID: <57D7F97B-807F-41F7-A94B-FA88034727EC@list.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Nov 18, 2007, at 10:38 PM, Mark Sapiro wrote:

>>> === modified file 'Mailman/versions.py'
>>> --- a/Mailman/versions.py	2005-12-30 18:50:08 +0000
>>> +++ b/Mailman/versions.py	2007-10-05 01:35:10 +0000
>>> @@ -483,10 +483,11 @@
>>>              # blow away the original timestamp and request id.
>>> This means the
>>>              # request will live a little longer than it possibly
>>> should have,
>>>              # but that's no big deal.
>>> +            import email
>>>              for p in v:
>>>                  author, text = p[2]
>>>                  reason = p[3]
>>> -                msg = Message.OutgoingMessage(text)
>>> +                msg = email.message_from_string(text,
>>> Message.Message)
>>>                  l.HoldMessage(msg, reason)
>>>              del r[k]
>>>          elif k == 'add_member':
>>
>> Should the import happen at the module global level?
>
>
> I thought about that. Our standards of course say yes, but this
> particular import only occurs when updating a Mailman 1.0.x list  
> which I
> hope happens very infrequently. However I don't really feel strongly
> either way.

Ultimately it's a style issue.  I'm nearly positive that it will make  
no significant different in performance.  Generally I like to have  
function scope imports only when necessary (e.g. to avoid circular  
import problems).  But I'll leave it up to you to change or not.

Cheers,
- -Barry

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iD8DBQFHQY6X2YZpQepbvXERAlNpAJ91ZKrl0gWY5r+l0OoPhwasBY0YHACgo/38
1QGhKpAIyf8gPgY0mCJIm6o=
=z8gz
-----END PGP SIGNATURE-----

From msapiro at value.net  Mon Nov 19 18:14:26 2007
From: msapiro at value.net (Mark Sapiro)
Date: Mon, 19 Nov 2007 09:14:26 -0800
Subject: [Mailman-Developers] [Mailman-checkins]
	[Branch	~mailman-coders/mailman/2.1]
In-Reply-To: <57D7F97B-807F-41F7-A94B-FA88034727EC@list.org>
References: <20071005014611.28380.40188.launchpad@canonical@forster.canonical.com>
	<7104340B-1E29-4C96-A6AB-3B663134A8A0@list.org>
	<4741053F.5050801@value.net>
	<57D7F97B-807F-41F7-A94B-FA88034727EC@list.org>
Message-ID: <4741C472.90705@value.net>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Barry Warsaw wrote:
> On Nov 18, 2007, at 10:38 PM, Mark Sapiro wrote:
> 
>> I thought about that. Our standards of course say yes, but this
>> particular import only occurs when updating a Mailman 1.0.x list which I
>> hope happens very infrequently. However I don't really feel strongly
>> either way.
> 
> Ultimately it's a style issue.  I'm nearly positive that it will make no
> significant different in performance.  Generally I like to have function
> scope imports only when necessary (e.g. to avoid circular import
> problems).  But I'll leave it up to you to change or not.


I've moved the email import, and also an import of _BounceInfo. I'm
ready to commit this, but I want to run it by you first to see if I'm
missing anything with regard to moving the import of _BounceInfo.

=== modified file 'Mailman/versions.py'
- --- Mailman/versions.py 2007-10-05 02:50:56 +0000
+++ Mailman/versions.py 2007-11-19 17:02:10 +0000
@@ -33,11 +33,14 @@
 """


+import email
+
 from types import ListType, StringType

 from Mailman import mm_cfg
 from Mailman import Utils
 from Mailman import Message
+from Mailman.Bouncer import _BounceInfo
 from Mailman.MemberAdaptor import UNKNOWN
 from Mailman.Logging.Syslog import syslog

@@ -419,7 +422,6 @@
     # Go through all the keys in bounce_info.  If the key is not a
member, or
     # if the data is not a _BounceInfo instance, chuck the bounce info.
 We're
     # doing things differently now.
- -    from Mailman.Bouncer import _BounceInfo
     for m in mlist.bounce_info.keys():
         if not mlist.isMember(m) or not isinstance(mlist.getBounceInfo(m),
                                                    _BounceInfo):
@@ -483,7 +485,6 @@
             # blow away the original timestamp and request id.  This
means the
             # request will live a little longer than it possibly should
have,
             # but that's no big deal.
- -            import email
             for p in v:
                 author, text = p[2]
                 reason = p[3]


- --
Mark Sapiro <msapiro at value.net>       The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)

iD8DBQFHQcRxVVuXXpU7hpMRArZAAKD8R2Cif0NNC1X89rqyHn0hEjKibQCfRsSf
wx9HFj66t+cnR7BtyE16s64=
=C90I
-----END PGP SIGNATURE-----

From barry at list.org  Mon Nov 19 22:26:04 2007
From: barry at list.org (Barry Warsaw)
Date: Mon, 19 Nov 2007 16:26:04 -0500
Subject: [Mailman-Developers] Shameless plug
Message-ID: <F69A3DF8-0257-4AFE-BE80-954D06BBFF33@list.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi folks,

I apologize in advance for the shameless plug.  GNU Mailman is up for  
the 2007 Antonio Pizzigati Prize for software in the public  
interest.  I sure would appreciate it if you take the time to vote  
for GNU Mailman:

http://forums.pizzigatiprize.org/index.php

Vote in the 'Vote' forum.  Leave comments in the 'GNU Mailman' forum.

Yes, you need to register with the site in order to vote.

Thanks for your support!
- -Barry

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iD8DBQFHQf9t2YZpQepbvXERAorYAKCXoVhrBI8eZu+uRxaYlIdyCjAj/ACglrYh
RQstgR9SujgyTqsMVSh2q/s=
=oIKy
-----END PGP SIGNATURE-----

From barry at list.org  Sat Nov 24 03:30:29 2007
From: barry at list.org (Barry Warsaw)
Date: Fri, 23 Nov 2007 21:30:29 -0500
Subject: [Mailman-Developers] [Mailman-checkins]
	[Branch	~mailman-coders/mailman/2.1]
In-Reply-To: <4741C472.90705@value.net>
References: <20071005014611.28380.40188.launchpad@canonical@forster.canonical.com>
	<7104340B-1E29-4C96-A6AB-3B663134A8A0@list.org>
	<4741053F.5050801@value.net>
	<57D7F97B-807F-41F7-A94B-FA88034727EC@list.org>
	<4741C472.90705@value.net>
Message-ID: <318CEF25-46C1-4166-9C20-F035BA35A4FD@list.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Nov 19, 2007, at 12:14 PM, Mark Sapiro wrote:

> Barry Warsaw wrote:
>> On Nov 18, 2007, at 10:38 PM, Mark Sapiro wrote:
>>
>>> I thought about that. Our standards of course say yes, but this
>>> particular import only occurs when updating a Mailman 1.0.x list  
>>> which I
>>> hope happens very infrequently. However I don't really feel strongly
>>> either way.
>>
>> Ultimately it's a style issue.  I'm nearly positive that it will  
>> make no
>> significant different in performance.  Generally I like to have  
>> function
>> scope imports only when necessary (e.g. to avoid circular import
>> problems).  But I'll leave it up to you to change or not.
>
>
> I've moved the email import, and also an import of _BounceInfo. I'm
> ready to commit this, but I want to run it by you first to see if I'm
> missing anything with regard to moving the import of _BounceInfo.

Hi Mark,

I think it should be fine, as long as it doesn't create a circular  
import.  I didn't run the code, but from grepping it looks like we  
should be okay.  The only import of Mailman.versions is at file scope  
afaict.

- -Barry

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iD8DBQFHR4zG2YZpQepbvXERAut0AJ40NPdqbfI3Z+EQsLibny4atEGjLQCgkw20
JLr8bvq5WFV/O45Uo7JoCX8=
=FfSi
-----END PGP SIGNATURE-----

From barry at list.org  Mon Nov 26 15:26:11 2007
From: barry at list.org (Barry Warsaw)
Date: Mon, 26 Nov 2007 09:26:11 -0500
Subject: [Mailman-Developers] lost email
Message-ID: <CEC951C2-BB78-45C9-883C-9DE0D3423A5C@list.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi folks,

I lost my entire raid array over the weekend.  Fortunately, I was  
able to make a backup of most stuff on November 20th, and managed to  
save a few emails out of my local cache up to November 23rd (thank  
you, thank you, thank you Mail.app!).  Depending on what upstream  
servers bounced or delayed, I may not have gotten anything from the  
23rd until today.

Everything's back up again now, but if you sent me an important email  
over the Thanksgiving holiday, assume I lost it and please resend!

Cheers,
- -Barry

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iD8DBQFHSteD2YZpQepbvXERAqtoAJ9aJPGRcQ/An5YS66sydkJp5vSj0ACfRXYE
8KPXA/n5BIyj8BcaArBuYv8=
=G0GI
-----END PGP SIGNATURE-----

From mark at msapiro.net  Mon Nov 26 23:02:33 2007
From: mark at msapiro.net (Mark Sapiro)
Date: Mon, 26 Nov 2007 14:02:33 -0800
Subject: [Mailman-Developers] [Mailman-checkins]
	[Branch	~mailman-coders/mailman/2.1]
In-Reply-To: <318CEF25-46C1-4166-9C20-F035BA35A4FD@list.org>
References: <20071005014611.28380.40188.launchpad@canonical@forster.canonical.com>
	<7104340B-1E29-4C96-A6AB-3B663134A8A0@list.org>
	<4741053F.5050801@value.net>
	<57D7F97B-807F-41F7-A94B-FA88034727EC@list.org>
	<4741C472.90705@value.net>
	<318CEF25-46C1-4166-9C20-F035BA35A4FD@list.org>
Message-ID: <474B4279.4080407@msapiro.net>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Barry Warsaw wrote:
> On Nov 19, 2007, at 12:14 PM, Mark Sapiro wrote:
> 
>> I've moved the email import, and also an import of _BounceInfo. I'm
>> ready to commit this, but I want to run it by you first to see if I'm
>> missing anything with regard to moving the import of _BounceInfo.
> 
> Hi Mark,
> 
> I think it should be fine, as long as it doesn't create a circular
> import.  I didn't run the code, but from grepping it looks like we
> should be okay.  The only import of Mailman.versions is at file scope
> afaict.


I have now actually tested the code thanks to Tokio's sibling list
patch, and it works without problems (and the move of the imports has
been committed).

- --
Mark Sapiro <mark at msapiro.net>        The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (MingW32)

iD8DBQFHS0J5VVuXXpU7hpMRAk+7AJsG50NFD51ekFa+/BTyHubSSofVggCeLRCW
0+C7TwOptmT/3/znMVoY/XI=
=Vd32
-----END PGP SIGNATURE-----

From barry at list.org  Tue Nov 27 05:07:28 2007
From: barry at list.org (Barry Warsaw)
Date: Mon, 26 Nov 2007 23:07:28 -0500
Subject: [Mailman-Developers] [Mailman-checkins]
	[Branch	~mailman-coders/mailman/2.1]
In-Reply-To: <474B4279.4080407@msapiro.net>
References: <20071005014611.28380.40188.launchpad@canonical@forster.canonical.com>
	<7104340B-1E29-4C96-A6AB-3B663134A8A0@list.org>
	<4741053F.5050801@value.net>
	<57D7F97B-807F-41F7-A94B-FA88034727EC@list.org>
	<4741C472.90705@value.net>
	<318CEF25-46C1-4166-9C20-F035BA35A4FD@list.org>
	<474B4279.4080407@msapiro.net>
Message-ID: <93A158DD-D846-444D-A0DA-8CFC8A2CFD86@list.org>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Nov 26, 2007, at 5:02 PM, Mark Sapiro wrote:

> I have now actually tested the code thanks to Tokio's sibling list
> patch, and it works without problems (and the move of the imports has
> been committed).

Awesome, thanks!
- -Barry

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)

iD8DBQFHS5gA2YZpQepbvXERAmpkAKCC0OLVQg2/0/SVUlNLmIcWiMh+QQCeJlQS
Z82hlRrJp5s5z9QG9eyJ3yw=
=Bray
-----END PGP SIGNATURE-----