From Jack.Jansen@cwi.nl Thu Feb 6 16:46:18 1997 From: Jack.Jansen@cwi.nl (Jack Jansen) Date: Thu, 06 Feb 1997 17:46:18 +0100 Subject: [PYTHONMAC-SIG] Motorola C compiler, anyone? Message-ID: <9702061646.AA01009=jack@snelboot.cwi.nl> I was looking at where Python/PPC was spending its time, and at some point I looked at the machine code generated by the MetroWerks C compiler. I must say I was rather appalled: even with all optimizations on it still contained silly things as branches-over-branches and such. I already knew that the MW optimizer isn't exactly state of the art, but I wasn't aware that it was so bad. Apparently the Motorola C compiler, which is available as a plugin for the MW environment, does a much better job. Hence my question: has anyone tried compiling Python/PPC with the Motorola compiler, and what were the results? I'm interested in both performance and porting ease, since the combination of these two is what'll probably influence my decision to get the mcc compiler... -- Jack Jansen | ++++ stop the execution of Mumia Abu-Jamal ++++ Jack.Jansen@cwi.nl | ++++ if you agree copy these lines to your sig ++++ http://www.cwi.nl/~jack | see http://www.xs4all.nl/~tank/spg-l/sigaction.htm _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From sdm7g@virginia.edu Thu Feb 6 17:19:29 1997 From: sdm7g@virginia.edu (Steven D. Majewski) Date: Thu, 6 Feb 1997 12:19:29 -0500 (EST) Subject: [PYTHONMAC-SIG] Motorola C compiler, anyone? In-Reply-To: <9702061646.AA01009=jack@snelboot.cwi.nl> Message-ID: I don't have the Motorola compiler, but I have the (beta?) latest version of Apple's MrC compiler, which they ship as both a MPW and CodeWarrior plugin. It is supposed to produce better code than MetroWerks. I will try to give it a try over the weekend ( IF I get my copy of CW11 shipped soon! ). I have not bothered to try it out, as non of my current projects are in a debugged and final enough state that I would think about trying another compiler. :-( ---| Steven D. Majewski (804-982-0831) |--- ---| Department of Molecular Physiology and Biological Physics |--- ---| University of Virginia Health Sciences Center |--- ---| P.O. Box 10011 Charlottesville, VA 22906-0011 |--- By doing just a little every day, you can gradually let the task completely overwhelm you. _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From just@knoware.nl Thu Feb 6 19:15:26 1997 From: just@knoware.nl (Just van Rossum) Date: Thu, 6 Feb 1997 20:15:26 +0100 Subject: [PYTHONMAC-SIG] Beware of CW 11! Message-ID: At 12:19 PM 2/6/97, Steven D. Majewski wrote: >( IF I get my copy of CW11 shipped soon! ). I tried CW11 recently, and a lot has changed. All ANSI libs are in other places and called different names, SIOUX is in a folder called "obsolete" and and and... in other words: Python won't compile. Even after changing some libraries and accesspaths I couldn't get it to work. Guido said that HE couldn't get it to work either (which made me sigh with relief...), and thought that it was mainly GUSI's problem. So we might have to wait for a GUSI update. hmmm. Just _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From sdm7g@virginia.edu Thu Feb 6 19:28:52 1997 From: sdm7g@virginia.edu (Steven D. Majewski) Date: Thu, 6 Feb 1997 14:28:52 -0500 (EST) Subject: [PYTHONMAC-SIG] Re: Beware of CW 11! In-Reply-To: Message-ID: On Thu, 6 Feb 1997, Just van Rossum wrote: > At 12:19 PM 2/6/97, Steven D. Majewski wrote: > >( IF I get my copy of CW11 shipped soon! ). > > I tried CW11 recently, and a lot has changed. All ANSI libs are in other > places and called different names, SIOUX is in a folder called "obsolete" > and and and... in other words: Python won't compile. Even after changing > some libraries and accesspaths I couldn't get it to work. Guido said that > HE couldn't get it to work either (which made me sigh with relief...), and > thought that it was mainly GUSI's problem. So we might have to wait for a > GUSI update. > What I have *heard* is that SIOUX has been merged into the MSL - it's no longer a separate library. If you are using the ANSI libs, you need either SIOUX or console.stubs, so I guess they just decided to make SIOUX the default. ( Of course, since I haven't gotten mine yet, I can't verify what I've heard, but you might just try taking the SIOUX library reference out of the project and see if it links with just the ANSI libs. ) [There was a thread on comp.sys.mac.programmer.codewarrior about this.] ---| Steven D. Majewski (804-982-0831) |--- ---| Department of Molecular Physiology and Biological Physics |--- ---| University of Virginia Health Sciences Center |--- ---| P.O. Box 10011 Charlottesville, VA 22906-0011 |--- By doing just a little every day, you can gradually let the task completely overwhelm you. _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From Jack.Jansen@cwi.nl Fri Feb 7 10:13:06 1997 From: Jack.Jansen@cwi.nl (Jack Jansen) Date: Fri, 07 Feb 1997 11:13:06 +0100 Subject: [PYTHONMAC-SIG] Re: Beware of CW 11! In-Reply-To: Message by "Steven D. Majewski" , Thu, 6 Feb 1997 14:28:52 -0500 (EST) , Message-ID: <9702071013.AA07312=jack@snelboot.cwi.nl> > What I have *heard* is that SIOUX has been merged into the MSL - it's > no longer a separate library. If you are using the ANSI libs, you > need either SIOUX or console.stubs, so I guess they just decided > to make SIOUX the default. ( Of course, since I haven't gotten > mine yet, I can't verify what I've heard, but you might just try > taking the SIOUX library reference out of the project and see if > it links with just the ANSI libs. ) I'm actually quite happy with this change: it means that we can probably roll our own changes and those from GUSI into a single library with selected parts of MSL (which is apparently still shipped in source form). That would forestall all sorts of errors (like linking libc before GUSI). Also, there's a couple of mods that I would like to make to sioux (a python-specific menubar, clicking on a line to allow you to retype/edit the line, possibly clicking on an error to send you to that line in your editor of choice) that would make the Python environment a lot more friendly (even if still far from a real IDE). Matthias said that he has his CW11, so I'll just wait another week or so for him to release his new version of CWGUSI before I start hacking on it. -- Jack Jansen | ++++ stop the execution of Mumia Abu-Jamal ++++ Jack.Jansen@cwi.nl | ++++ if you agree copy these lines to your sig ++++ http://www.cwi.nl/~jack | see http://www.xs4all.nl/~tank/spg-l/sigaction.htm _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From ptaney@usgs.gov Fri Feb 7 18:26:48 1997 From: ptaney@usgs.gov (Paul Taney, Computer Specialist, Reston, VA ) Date: Fri, 07 Feb 1997 13:26:48 -0500 Subject: [PYTHONMAC-SIG] Re: Motorola C compiler, anyone? Message-ID: <199702071826.NAA02057@mailrvares.er.usgs.gov> yesterday I wrote: > I think the recent Apple developer disks have the Motorola compiler. Sorry. I searched the last 7 disks or so. All they're giving away is an optimized set of C routines called LIBMOTO [consisting of cos, exp, log, log10, memcpy, memmove, pow, sin, sqrt, strcmp, strlen and strncmp] and Apple's mr_c. That's on ETO #22. But searching at mot.com for "C compiler" returns these (and other) hits. Hope this helps. paul Paul Taney, ptaney@usgs.gov @ All have we got it seem we have lost smail: ms431, reston,va 22092 @ We must have 'ready paid the cost. voice: 703-648-5598 @ -- Marley %<---------------------------------------- C/C++ SDK, PowerPC Edition for Mac OS http://www.mot.com/SPS/PowerPC/library/fact_sheet/mac_c_ts.html C/C++ SDK, PowerPC Edition for Mac OS http://www.mot.com/SPS/PowerPC/library/fact_sheet/mac_c.html PowerPC 601 Tools http://www.mot.com/SPS/PowerPC/library/press_releases/601_Tools_release.ht ml Motorola PowerPC Software Products http://www.mot.com/SPS/PowerPC/products/software/sw_prod.html optimize.txt http://www.mot.com/SPS/HPESD/aesop/680X0/gen/optimize.txt Motorola C/C++ Compiler for MacOS Benchmarks http://www.mot.com/SPS/PowerPC/library/technical_papers/performance/mac_bm /bm_top.html Motorola White Papers http://www.mot.com/Index/white_papers.html _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From sdm7g@virginia.edu Fri Feb 7 19:05:44 1997 From: sdm7g@virginia.edu (Steven D. Majewski) Date: Fri, 7 Feb 1997 14:05:44 -0500 (EST) Subject: [PYTHONMAC-SIG] Re: Motorola C compiler, anyone? In-Reply-To: <199702071826.NAA02057@mailrvares.er.usgs.gov> Message-ID: Looking at the moto web site, I see that the Plug-in version of the compiler ( If you already have Code Warrior ) is only $99. Not bad if it really *is* significantly better. ---| Steven D. Majewski (804-982-0831) |--- ---| Department of Molecular Physiology and Biological Physics |--- ---| University of Virginia Health Sciences Center |--- ---| P.O. Box 10011 Charlottesville, VA 22906-0011 |--- By doing just a little every day, you can gradually let the task completely overwhelm you. _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From Jack.Jansen@cwi.nl Fri Feb 7 21:38:03 1997 From: Jack.Jansen@cwi.nl (Jack Jansen) Date: Fri, 07 Feb 1997 22:38:03 +0100 Subject: [PYTHONMAC-SIG] Re: Motorola C compiler, anyone? In-Reply-To: Message by "Steven D. Majewski" , Fri, 7 Feb 1997 14:05:44 -0500 (EST) , Message-ID: <9702072138.AA14320=jack@snelboot.cwi.nl> Recently, "Steven D. Majewski" said: > > Looking at the moto web site, I see that the Plug-in version of > the compiler ( If you already have Code Warrior ) is only $99. > > Not bad if it really *is* significantly better. That was approximately my line of reasoning:-) At the moment I'm investigating using MrC (which is freely downloadable from the apple developers site). Compilation was little problem, but I ran into a problem linking. The Apple development group was *very* speedy in replying to my message, but it'll be sometime next week before I get another chance to look at this. Oh yeah, in the mean time I do have a tip for anyone who compiles Python him/herself: add libmoto before the standard metrowerks libraries. The speedup (according to Pystone) is a few percent, but since it's so little work it's worth it... -- Jack Jansen | ++++ stop the execution of Mumia Abu-Jamal ++++ Jack.Jansen@cwi.nl | ++++ if you agree copy these lines to your sig ++++ http://www.cwi.nl/~jack | see http://www.xs4all.nl/~tank/spg-l/sigaction.htm _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From Jack.Jansen@cwi.nl Wed Feb 12 15:09:36 1997 From: Jack.Jansen@cwi.nl (Jack Jansen) Date: Wed, 12 Feb 1997 16:09:36 +0100 Subject: [PYTHONMAC-SIG] Calling any toolbox routine from Python Message-ID: <9702121509.AA03618=jack@snelboot.cwi.nl> Folks, there is a way to make it possible to call *any* toolbox routine from Python, at least on PPC and CFM68K. Since the toolboxes are completely in shared libraries on those architectures it should be doable to create an interface that allows python programmers access to them. The interface, however, would be *very* error-prone (no typechecking whatsoever) and probably rather convoluted (for instance, I think that the only reasonable way to implement return parameters passed by address is to pass in an a 1-element array of the right type). The advantage, of course, is that I hope everyone on the list will quickly create Python wrappers with decent interfaces (and possibly type-checking) around their favorite toolboxes and make them available. The disadvantage, as stated above, is that the interface will be clumsy and error-prone, thereby possibly giving MacPython a bad name. What do y'all think? If I did the interface, would you use it? -- Jack Jansen | ++++ stop the execution of Mumia Abu-Jamal ++++ Jack.Jansen@cwi.nl | ++++ if you agree copy these lines to your sig ++++ http://www.cwi.nl/~jack | see http://www.xs4all.nl/~tank/spg-l/sigaction.htm _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From guido@CNRI.Reston.Va.US Wed Feb 12 15:46:16 1997 From: guido@CNRI.Reston.Va.US (Guido van Rossum) Date: Wed, 12 Feb 1997 10:46:16 -0500 Subject: [PYTHONMAC-SIG] Calling any toolbox routine from Python In-Reply-To: Your message of "Wed, 12 Feb 1997 16:09:36 +0100." <9702121509.AA03618=jack@snelboot.cwi.nl> References: <9702121509.AA03618=jack@snelboot.cwi.nl> Message-ID: <199702121546.KAA14616@monty> > there is a way to make it possible to call *any* toolbox routine from > Python, at least on PPC and CFM68K. Since the toolboxes are completely > in shared libraries on those architectures it should be doable to > create an interface that allows python programmers access to them. > > The interface, however, would be *very* error-prone (no typechecking > whatsoever) and probably rather convoluted (for instance, I think that > the only reasonable way to implement return parameters passed by > address is to pass in an a 1-element array of the right type). > > The advantage, of course, is that I hope everyone on the list will > quickly create Python wrappers with decent interfaces (and possibly > type-checking) around their favorite toolboxes and make them > available. > > The disadvantage, as stated above, is that the interface will be > clumsy and error-prone, thereby possibly giving MacPython a bad name. > > What do y'all think? If I did the interface, would you use it? Doesn't Visual Basic have a similar escape into the C world? I believe I've seen include files that declared a bunch of entry points with their address and the type of the arguments. It seemed obvious that by specifying the wrong address or type you could blow up the system but of course the include files that are part of the distribution can be trusted in the same sense that you can trust Python extension modules. Can't you put a little more work in the interface so that this approach would work? Then you don't have to provide all the interface files (though they could possibly be generated by a spin-off from bgen) but someone else can at least write interface files that will make it relatively safe to use. Of course you can't ever hope to make the type check foolproof, but at least some checks (like nr. of args) are easy. For return types I would do the standard python thing -- turn them into extra return values. --Guido van Rossum (home page: http://www.python.org/~guido/) _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From Jack.Jansen@cwi.nl Wed Feb 12 16:14:55 1997 From: Jack.Jansen@cwi.nl (Jack Jansen) Date: Wed, 12 Feb 1997 17:14:55 +0100 Subject: [PYTHONMAC-SIG] Calling any toolbox routine from Python In-Reply-To: Message by Guido van Rossum , Wed, 12 Feb 1997 10:46:16 -0500 , <199702121546.KAA14616@monty> Message-ID: <9702121614.AA04233=jack@snelboot.cwi.nl> > Doesn't Visual Basic have a similar escape into the C world? I > believe I've seen include files that declared a bunch of entry points > with their address and the type of the arguments. It seemed obvious > that by specifying the wrong address or type you could blow up the > system but of course the include files that are part of the > distribution can be trusted in the same sense that you can trust > Python extension modules. Yes, that's the other option but it is a bit more work. The "luxury" version I was thinking of would have something like FooBar = dllcaller('FooBar', formatstring) where formatstring would somehow describe the arguments and return values. dllcaller would then return a callable object with the correct calling sequence, and it would perform argument checking. -- Jack Jansen | ++++ stop the execution of Mumia Abu-Jamal ++++ Jack.Jansen@cwi.nl | ++++ if you agree copy these lines to your sig ++++ http://www.cwi.nl/~jack | see http://www.xs4all.nl/~tank/spg-l/sigaction.htm _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From sdm7g@virginia.edu Wed Feb 12 17:15:54 1997 From: sdm7g@virginia.edu (Steven D. Majewski) Date: Wed, 12 Feb 1997 12:15:54 -0500 (EST) Subject: [PYTHONMAC-SIG] Calling any toolbox routine from Python In-Reply-To: <199702121546.KAA14616@monty> Message-ID: My thoughts on a related issue: I've just finished testing a shared library interface to our data acquisition hardware for Xlisp-Stat. I'm also going to do Python and Java bindings, so I've been trying to think of the easiest way to maintain these separate interfaces in one library. [1] The "brute force" method would be to include wrapper/entry-points for all of the above, and use some symbol naming convention to select the appropriate symbols, but I'ld like to avoid having to deal with the wrapper code in the library. [2] Mac Runtime for Java does some undocumented tricks(*) -- you can get some clues if you disassemble the class interfaces for some of the com.apple.MacOS classes. They have class definitions for a bunch of Mac native data types: point, rect, graphics-port, window, etc. and use those in other interface definitions. Java native methods have both a native library and a class file to define their interface. The class files have a representation of the method signature. Is there a way to use this sort of signature to create safe wrappers at runtime ? [3] The above sounds rather like what SOM already does. Write SOM classes and add SOM import support to Java, Python & Xlisp ? ( I believe ILU has a similar interface, but (1) no Mac ILU port yet; (2) SOM is supported on Mac (3) SOM is probably more "lightweight". I was thinking about experimenting with SOM a bit -- and to do this I was thinking about adding a sort of "raw" call interface to Python. I'm still thinking ( in between getting the "brute force" method above done to have something right away for my current problems ) about how to do it or if there is a better and higher level way to package things. If I do a low level interface, I wasn't very concerned with making it safe -- it is planned mostly as a tool to figure out a better/safer way to do it. [ (*) Undocumented MRJ tricks: The latest Class Wrangler has a "show information" button that will display the class interface definitions. Or you can use MOCHA, a free class file disassembler. A large part of the Mac toolbox is wrapped in com.apple.MacOS package. I have been told by Apple that these are experimental and mostly undocumented. ] ---| Steven D. Majewski (804-982-0831) |--- ---| Department of Molecular Physiology and Biological Physics |--- ---| University of Virginia Health Sciences Center |--- ---| P.O. Box 10011 Charlottesville, VA 22906-0011 |--- By doing just a little every day, you can gradually let the task completely overwhelm you. _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From jstrout@ucsd.edu Wed Feb 12 17:16:07 1997 From: jstrout@ucsd.edu (Joseph Strout) Date: Wed, 12 Feb 1997 09:16:07 -0800 (PST) Subject: [PYTHONMAC-SIG] Calling any toolbox routine from Python In-Reply-To: <9702121614.AA04233=jack@snelboot.cwi.nl> Message-ID: On Wed, 12 Feb 1997, Jack Jansen wrote: > Yes, that's the other option but it is a bit more work. The "luxury" > version I was thinking of would have something like > FooBar = dllcaller('FooBar', formatstring) > where formatstring would somehow describe the arguments and return values. > dllcaller would then return a callable object with the correct calling > sequence, and it would perform argument checking. Either way sounds great to me. I remember when I was searching for a way for my brothers-in-law to write games in Python, I got stymied because I couldn't plot color icons from within Python without writing a custom C module and compiling it in. I finally gave up on Python, and gave my bros Cocoa instead (which is extremely cool in its own way). So anything that lets us get at arbitrary Toolbox calls is a big plus, and the sooner the better. Put big "DANGER! DANGER!" warnings on it; the intrepid will manage to use it properly. The only issue I can think of is how certain Toolbox things (like windows?) might interact with the SIOUX console; if you have any info on that, better include it in the docs. Keep up the good work, Jack! -- Joe ,------------------------------------------------------------------. | Joseph J. Strout Department of Neuroscience, UCSD | | jstrout@ucsd.edu http://www-acs.ucsd.edu/~jstrout/ | `------------------------------------------------------------------' _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From sdm7g@virginia.edu Wed Feb 12 18:14:23 1997 From: sdm7g@virginia.edu (Steven D. Majewski) Date: Wed, 12 Feb 1997 13:14:23 -0500 (EST) Subject: [PYTHONMAC-SIG] Calling any toolbox routine from Python In-Reply-To: Message-ID: I seem to be getting only half of this discussion -- I'm seeing replies to Jack's messages, but not Jack's messages. ---| Steven D. Majewski (804-982-0831) |--- ---| Department of Molecular Physiology and Biological Physics |--- ---| University of Virginia Health Sciences Center |--- ---| P.O. Box 10011 Charlottesville, VA 22906-0011 |--- By doing just a little every day, you can gradually let the task completely overwhelm you. _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From guzdial@cc.gatech.edu Thu Feb 13 19:04:03 1997 From: guzdial@cc.gatech.edu (Mark Guzdial) Date: Thu, 13 Feb 1997 14:04:03 -0500 (EST) Subject: [PYTHONMAC-SIG] Controlling Netscape from Python -- portably? Message-ID: I'd like to be able to tell Netscape what URL to look at, from Python. I'm asking here because I have at least a clue how to do that on a Mac: Netscape understands a bunch of AppleEvents and I could tell it to do a getURL command. I'm not EXACTLY sure how to do it from Mac Python, but I'm hoping that somebody here does. But more generally, I think that there are ways of doing this on UNIX and Windows boxes. Does anyone know of a *portable* Python module that would let me control Netscape via whatever inter-application communication mechanisms exist on each platform? Thanks! Mark -------------------------- Mark Guzdial : Georgia Tech : College of Computing : Atlanta, GA 30332-0280 (404) 894-5618 : Fax (404) 894-0673 : guzdial@cc.gatech.edu http://www.cc.gatech.edu/gvu/people/Faculty/Mark.Guzdial.html _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From Jack.Jansen@cwi.nl Thu Feb 13 22:05:27 1997 From: Jack.Jansen@cwi.nl (Jack Jansen) Date: Thu, 13 Feb 1997 23:05:27 +0100 Subject: [PYTHONMAC-SIG] Controlling Netscape from Python -- portably? In-Reply-To: Message by guzdial@cc.gatech.edu (Mark Guzdial) , Thu, 13 Feb 1997 14:04:03 -0500 (EST) , Message-ID: <9702132205.AA16665=jack@snelboot.cwi.nl> Have a look at the files in ftp://ftp.cwi.nl/pub/jack/python/mac/nsremote/. I did that module last year, but I haven't used it since, so I don't know whether it still works. Somehow I forgot to put it in the Demo folder for the 1.4 distribution, I'll do so for the next release. -- Jack Jansen | ++++ stop the execution of Mumia Abu-Jamal ++++ Jack.Jansen@cwi.nl | ++++ if you agree copy these lines to your sig ++++ http://www.cwi.nl/~jack | see http://www.xs4all.nl/~tank/spg-l/sigaction.htm _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From Jack.Jansen@cwi.nl Mon Feb 17 16:46:58 1997 From: Jack.Jansen@cwi.nl (Jack Jansen) Date: Mon, 17 Feb 1997 17:46:58 +0100 Subject: [PYTHONMAC-SIG] U vraagt, wij draaien: calling C routines from Python Message-ID: <9702171646.AA19428=jack@snelboot.cwi.nl> Folks, as there seemed to be quite a lot of interest in a way to call random toolbox routines and I had a bit of spare time anyway I went and implemented a first try. The calldll module allows you to give a sort-of prototype for a C routine, and then lets you call it from Python dynamically. It is currently PowerPC only, I would definitely like it if any of the cfm68k users looked at porting it to that platform. The distribution contains the dynamically loadable module for Python 1.4, and also the source, the project file and (minimal) test and example code. (as an aside: if you want to do the cfm68k port: the relevant info on 68K C calling sequences is best described in IM: PowerPC software, of all places:-) Download the code via http://www.cwi.nl/~jack/macpython.html or ftp://ftp.cwi.nl/pub/jack/python/mac, as usual. Don't use the module in production yet (hehe:-), but please give it a test run and tell me which features are missing, and whether there's anything I should change in the interface (the parameter specification comes to mind). Guido: I expect that with this module you'll have Grail supporting Netscape plugins by the end of the week, is that reasonable? :-) I am now off to try and cure the terrible headache those 20-odd lines in cdc_call() have caused, -- Jack Jansen | ++++ stop the execution of Mumia Abu-Jamal ++++ Jack.Jansen@cwi.nl | ++++ if you agree copy these lines to your sig ++++ http://www.cwi.nl/~jack | see http://www.xs4all.nl/~tank/spg-l/sigaction.htm _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From guzdial@cc.gatech.edu Thu Feb 20 17:39:33 1997 From: guzdial@cc.gatech.edu (Mark Guzdial) Date: Thu, 20 Feb 1997 12:39:33 -0500 (EST) Subject: [PYTHONMAC-SIG] MacPython IDE? Message-ID: I was looking through my archive of PythonMac-SIG email, and came across this: 15 Sep 1996, Just van Rossum -- Well, I have an attempt at such a thing going, but it's not *near* beta. Integrated editor (supports collapsing!) and run command is there, Jack's working on a graphical debugger, so somehow I believe there'll be something very nice at some point. I don't know yet how to do syntax highlighting efficiently. Could Just or Jack post a report on current status? Sounds exciting and was curious to hear if this was something that we might play with yet. (Actually, what I was looking for in the archive is the BBEdit extension to run Python scripts. I found Just's message, but can't find the archive. Is it posted somewhere? Thanks!) Mark -------------------------- Mark Guzdial : Georgia Tech : College of Computing : Atlanta, GA 30332-0280 (404) 894-5618 : Fax (404) 894-0673 : guzdial@cc.gatech.edu http://www.cc.gatech.edu/gvu/people/Faculty/Mark.Guzdial.html _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From just@knoware.nl Thu Feb 20 18:23:55 1997 From: just@knoware.nl (Just van Rossum) Date: Thu, 20 Feb 1997 19:23:55 +0100 Subject: [PYTHONMAC-SIG] MacPython IDE? Message-ID: Hi, At 12:39 PM 2/20/97, Mark Guzdial wrote: >I was looking through my archive of PythonMac-SIG email, and came across this: >15 Sep 1996, Just van Rossum -- > > Well, I have an attempt at such a thing going, but it's not *near* > beta. Integrated editor (supports collapsing!) and run > command is there, Jack's working on a graphical debugger, so > somehow I believe there'll be something very nice at some > point. I don't know yet how to do syntax highlighting > efficiently. > >Could Just or Jack post a report on current status? Sounds exciting and >was curious to hear if this was something that we might play with yet. Update: I have a new attempt at such a thing going, but it's not ready yet... Integrated editor (does not support collapsing :-(, but works ;-) and run command is there as well. Jack has a graphical debugger (TWIT), so somehow I believe there'll be something very nice at some point. Syntax highlighting should somehow be possible (I'm using waste). There is a nice (doubleclickable) traceback window. I'll try and put something together soon-ish. >(Actually, what I was looking for in the archive is the BBEdit extension to >run Python scripts. I found Just's message, but can't find the archive. >Is it posted somewhere? Thanks!) It's included in the 1.4 distribution: $(PYTHON):Mac:Tools:BBPy: Just _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From boyle@pcmdi.llnl.gov Thu Feb 20 19:30:18 1997 From: boyle@pcmdi.llnl.gov (James Boyle) Date: Thu, 20 Feb 1997 11:30:18 -0800 (PST) Subject: [PYTHONMAC-SIG] netCDF module for the Mac? Message-ID: <199702201930.LAA10024@cobra.llnl.gov> I would like to develop some code which reads netCDF files. I have used the netCDF modules developed by Konrad Hinsen, Bill Noon and Kyle Schaum on my Unix box but I have not added a module to the Mac side of Python. I see that the Numpy stuff is implemented using shared libraries and I guess I should take an analogous path for the netCDF stuff. But I need a little direction on where to find the path. I have played around a little with CodeWarrior but have not done any really serious development with it. Thanks for any help. Jim _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From Jack.Jansen@cwi.nl Fri Feb 21 09:52:06 1997 From: Jack.Jansen@cwi.nl (Jack Jansen) Date: Fri, 21 Feb 1997 10:52:06 +0100 Subject: [PYTHONMAC-SIG] netCDF module for the Mac? In-Reply-To: Message by James Boyle , Thu, 20 Feb 1997 11:30:18 -0800 (PST) , <199702201930.LAA10024@cobra.llnl.gov> Message-ID: <9702210952.AA04084=jack@snelboot.cwi.nl> > I would like to develop some code which reads netCDF files. > [...] > I have played around a little with CodeWarrior but have not done any > really serious development with it. Most of the info you need for building extension modules is in Mac:Demo:plugins.html in the 1.4 distribution. Some of the info is out-of-date (i.e. it's much better to get the Mac source distribution than twiddle with the machine-independent source distribution). -- Jack Jansen | ++++ stop the execution of Mumia Abu-Jamal ++++ Jack.Jansen@cwi.nl | ++++ if you agree copy these lines to your sig ++++ http://www.cwi.nl/~jack | see http://www.xs4all.nl/~tank/spg-l/sigaction.htm _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From billpy@mail.demon.net Sun Feb 23 11:22:00 1997 From: billpy@mail.demon.net (Bill Bedford) Date: Sun, 23 Feb 1997 11:22:00 +0000 Subject: [PYTHONMAC-SIG] Standalone templates. In-Reply-To: <9702210952.AA04084=jack@snelboot.cwi.nl> References: Message by James Boyle , Thu, 20 Feb 1997 11:30:18 -0800 (PST) , <199702201930.LAA10024@cobra.llnl.gov> Message-ID: Would some one be kind enough to post me compiled copies of the four standalone templates from the mac source distribution? Thanks a lot. _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________ From jbregman@mail.arc.nasa.gov Fri Feb 28 23:45:34 1997 From: jbregman@mail.arc.nasa.gov (Jesse Bregman) Date: Fri, 28 Feb 1997 15:45:34 -0800 (PST) Subject: [PYTHONMAC-SIG] serial communications Message-ID: I know this must be possible, but I haven't found information anywhere about how to do serial communications using Python. I am used to Basic's method of treating the serial port as a file, then just reading or writing to it as if it were any other file. Does anyone know of a similarly easy way to do this with Python? Thanks in advance. -Jesse Bregman jbregman@mail.arc.nasa.gov _______________ PYTHONMAC-SIG - SIG on Python for the Apple Macintosh send messages to: pythonmac-sig@python.org administrivia to: pythonmac-sig-request@python.org _______________