[sage-devel] Re: gdb, doctests, strange behavior...
On Mon, 12 Nov 2007 06:47:06 -, Robert Miller [EMAIL PROTECTED] wrote: (1) Do you or do you not actually have the same problem when you build the same code under Linux? You don't say above. I hadn't had the chance to try yet, but I have now. I can't get it to give any error in linux at all. All the tests pass, in and out of gdb mode. I tried several times. Turns out I was testing the wrong branch. I can reproduce the problem in linux, and here I actually get a useful-ish backtrace, since gdb ^^ I just want to make the general observation that developing Sage only on OSX is not a very sensible thing to do, just as you just noted. E.g., the valgrind tool exists only in Linux, and is potentially very useful. So OS X users (like I am this month :-) ), really should at least install Linux too and use it for subtle debuging issues. So you're definitely doing the right thing. has enough information. The relevant line in my .pyx file is a simple sage_free call. It looks like an array overflow problem or something. I guess this is all good news, since it seems as if gdb in os x 10.5 could get me this information too if the build had finished. (gdb) r Starting program: /home/hd/sage/local/bin/ python .doctest_binary_code.py [Thread debugging using libthread_db enabled] [New Thread -1209636192 (LWP 5346)] *** glibc detected *** free(): invalid next size (fast): 0x087836d0 *** Program received signal SIGABRT, Aborted. [Switching to Thread -1209636192 (LWP 5346)] 0xe410 in __kernel_vsyscall () (gdb) bt #0 0xe410 in __kernel_vsyscall () #1 0xb7e8f9a1 in raise () from /lib/tls/i686/cmov/libc.so.6 #2 0xb7e912b9 in abort () from /lib/tls/i686/cmov/libc.so.6 #3 0xb7ec387a in __fsetlocking () from /lib/tls/i686/cmov/libc.so.6 #4 0xb7ec9fd4 in malloc_usable_size () from /lib/tls/i686/cmov/ libc.so.6 #5 0xb7eca34a in free () from /lib/tls/i686/cmov/libc.so.6 #6 0xb427429d in __pyx_tp_dealloc_4sage_6coding_11binary_code_PartitionStack (o=0xb4c4c5cc) at sage/coding/binary_code.c:3413 #7 0x0808106b in insertdict (mp=0x6, key=0xb7e498c0, hash=-724865314, value=0x8137720) at Objects/dictobject.c:412 #8 0x08081486 in PyDict_SetItem (op=0xb7e28824, key=0xb7e498c0, value=0x8137720) at Objects/dictobject.c:637 #9 0x0808573a in PyObject_GenericSetAttr (obj=0xb7e2605c, name=0xb7e498c0, value=0x8137720) at Objects/object.c:1427 #10 0x08084e0c in PyObject_SetAttr (v=0xb7e2605c, name=0xb7e498c0, value=0x8137720) at Objects/object.c:1183 #11 0x08084f7b in PyObject_SetAttrString (v=0xb7e2605c, name=0x811bb0b _, w=0x8137720) at Objects/object.c:1097 #12 0x080e8144 in sys_displayhook (self=0x0, o=0xb7de6440) at Python/ sysmodule.c:105 #13 0x0805aa1c in PyObject_Call (func=0x14e2, arg=0x0, kw=0x0) at Objects/abstract.c:1860 #14 0x080bc8f3 in PyEval_CallObjectWithKeywords (func=0x0, arg=0xb42a9b6c, kw=0x0) at Python/ceval.c:3433 #15 0x080c23ca in PyEval_EvalFrameEx (f=0x8d43b5c, throwflag=0) at Python/ceval.c:1530 #16 0x080c420e in PyEval_EvalCodeEx (co=0xb4c3ef50, globals=0xb429e9bc, locals=0xb429e9bc, args=0x0, argcount=0, kws=0x0, kwcount=0, defs=0x0, defcount=0, closure=0x0) at Python/ceval.c: 2831 #17 0x080c2b66 in PyEval_EvalFrameEx (f=0x877e1fc, throwflag=0) at Python/ceval.c:494 #18 0x080c420e in PyEval_EvalCodeEx (co=0xb42ed9b0, globals=0xb42e8b54, locals=0x0, args=0x8d421d0, argcount=4, kws=0x8d421e0, kwcount=0, defs=0x0, defcount=0, closure=0x0) at Python/ceval.c: 2831 #19 0x080c2367 in PyEval_EvalFrameEx (f=0x8d4207c, throwflag=0) at Python/ceval.c:3660 #20 0x080c420e in PyEval_EvalCodeEx (co=0xb42edbf0, globals=0xb42e8b54, locals=0x0, args=0x8c1b3b0, argcount=2, kws=0x8c1b3b8, kwcount=0, defs=0xb42ff1c8, defcount=3, closure=0x0) at Python/ ceval.c:2831 #21 0x080c2367 in PyEval_EvalFrameEx (f=0x8c1b244, throwflag=0) at Python/ceval.c:3660 #22 0x080c420e in PyEval_EvalCodeEx (co=0xb42f1800, globals=0xb42e8b54, locals=0x0, args=0x816b260, argcount=1, kws=0x816b264, kwcount=3, defs=0xb42f4278, defcount=9, closure=0x0) at Python/ ceval.c:2831 #23 0x080c2367 in PyEval_EvalFrameEx (f=0x816b124, throwflag=0) at Python/ceval.c:3660 #24 0x080c420e in PyEval_EvalCodeEx (co=0xb7e01188, globals=0xb7e3eacc, locals=0xb7e3eacc, args=0x0, argcount=0, kws=0x0, kwcount=0, defs=0x0, defcount=0, closure=0x0) at Python/ceval.c: 2831 #25 0x080c4396 in PyEval_EvalCode (co=0x0, globals=0x0, locals=0x0) at Python/ceval.c:494 #26 0x080e2c6e in PyRun_FileExFlags (fp=0x8161008, filename=0xbf8f739a .doctest_binary_code.py, start=0, globals=0x0, locals=0x0, closeit=1, flags=0xbf8f6b88) at Python/pythonrun.c:1271 #27 0x080e2fa9 in PyRun_SimpleFileExFlags (fp=value optimized out, filename=0xbf8f739a .doctest_binary_code.py, closeit=1, flags=0xbf8f6b88) at Python/pythonrun.c:877 #28 0x08056d0f in Py_Main (argc=1, argv=0xbf8f6c24) at Modules/main.c: 523 #29 0xb7e7bea2 in __libc_start_main () from
[sage-devel] Re: My talk with martin Albrecht
Maybe I am too late with this, but I have just a couple of comments. I like the clean slide layout with no unnecessary information like name of the speaker and day of the talk etc., it provides for a very high signal to noise ratio. The slides feel balanced. On the other hand, there is too much text on each slide for my taste. Research has found that it is harder for people to absorb information when the same information is presented to you in spoken and written form at the same time. If you instead provide two forms that complement each other, like spoken words and a diagram, then it becomes easier to understand again. I loved the statistics in the end, I think they give one a good idea of how much of each ingredient is present in Sage. I had no idea that Scipy contributed so much, for instance. However, if you make a separate legend, like you do, then it becomes very hard to quickly see which color is which. Especially since some of these colors are going to be very hard to distinguish with a beamer. Instead, one should try to *minimize eye sweeps*, and present the information as close as possible. So preferably inside the pie chart, or at least directly around it. This is just a minor thing. I think the slides are really good. --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: gdb, doctests, strange behavior...
On Nov 12, 9:20 am, William Stein [EMAIL PROTECTED] wrote: On Mon, 12 Nov 2007 06:47:06 -, Robert Miller [EMAIL PROTECTED] wrote: (1) Do you or do you not actually have the same problem when you build the same code under Linux? You don't say above. Hello, I hadn't had the chance to try yet, but I have now. I can't get it to give any error in linux at all. All the tests pass, in and out of gdb mode. I tried several times. If you want to send me a patch/bundle against 2.8.12 or so and instructions on how to reproduce this and I will take a look. Turns out I was testing the wrong branch. I can reproduce the problem in linux, and here I actually get a useful-ish backtrace, since gdb ^^ I just want to make the general observation that developing Sage only on OSX is not a very sensible thing to do, just as you just noted. E.g., the valgrind tool exists only in Linux, That should change in the near future. and is potentially very useful. So OS X users (like I am this month :-) ), really should at least install Linux too and use it for subtle debuging issues. So you're definitely doing the right thing. Mastering more than one OS is a must in my book ;) has enough information. The relevant line in my .pyx file is a simple sage_free call. It looks like an array overflow problem or something. I guess this is all good news, since it seems as if gdb in os x 10.5 could get me this information too if the build had finished. I have seen the issue with gdb printing loads of output at startup of sage, but I didn't know that it also happened on 10.4 also. I might have a look on either how to get around this or actually solve the problem. Cheers, Michael SNIP -- William Stein Associate Professor of Mathematics University of Washingtonhttp://wstein.org --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: Givaro 3.2.6 build problem on MacOS X 10.5
On Nov 5, 2007, at 09:55 , Jean-Guillaume Dumas wrote: Ralf-Philipp Weinmann wrote: Dear Dr. Dumas, I've encountered a build problem in Givaro 3.2.6 on MacOS X 10.5. The uint type used in src/kernel/zpz/givzpz32std.inl for example is not available unless sys/types.h is included. The following patch fixes the problem for me: --- src/kernel/system/givbasictype.h.ORIG2007-11-01 15:17:57.0 +0100 +++ src/kernel/system/givbasictype.h2007-11-01 15:18:33.0 +0100 @@ -11,6 +11,9 @@ #include givaro/givconfig.h #include stdlib.h // for size_t +#ifdef MACOSX +# include sys/types.h // needed on MacOS X 10.5 for uint type +#endif // -- Neutral type: definition of zero and one class Neutral { --- Kindest Regards, Ralf-Philipp Weinmann Dear Ralf-Philipp Weinmann, the uint32 type should be defined in givconfig.h by : typedef unsigned __GIVARO_INT32 uint32; the latter being defined at the configure in the file givaro-config.h. Therefore I suspect a problem at the configure time. Can you check that your givconfig.h and givaro-config.h are correct. Dear Dr. Dumas, it's actually not a problem with the 'uint32' type but with the 'uint' type that givzpz32std.inl and givzpz32uns.inl use: ./src/src/kernel/zpz/givzpz32std.inl:429: uint i_begin=0; ./src/src/kernel/zpz/givzpz32uns.inl:441: uint i_begin=0; A suggestion by Michael Abshoff was to replace the uint type with a uint32 in givzpz32{std,uns}.inl, actually. I'm strongly in favor of that. The above patch is not necessary then. I also have to apologize for sending you a non-working patch. The correct way to check for MacOS X is not to check whether MACOSX is defined, but rather one has to check for __DARWIN__. Kindest Regards, Ralf-Philipp Weinmann --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: gdb, doctests, strange behavior...
If you want to send me a patch/bundle against 2.8.12 or so and instructions on how to reproduce this and I will take a look. Thanks for the offer but I figured out what the problem with my code was last night, as soon as I started using valgrind. ... I have seen the issue with gdb printing loads of output at startup of sage, but I didn't know that it also happened on 10.4 also. I might have a look on either how to get around this or actually solve the problem. Just to be accurate, I wasn't using 10.4 at all, so I have no idea what would happen there. This was only on 10.5. William, Michael, thanks for the help! It turns out I was using the wrong variable as an endpoint for a loop index, and it was just some bad read/writes during those loops. --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: cython __hash__ question
On Nov 13, 2007 12:32 AM, Joel B. Mohler [EMAIL PROTECTED] wrote: Questions: Should inheriting __hash__ implementations in a cython class work? Just a trivial remark, yes, it actually does work as the following simple example shows: {{{ %cython cdef class A: def __hash__(self): return 2 cdef class B(A): pass }}} {{{ a = A() b = B() }}} {{{ hash(a) /// 2 }}} {{{ hash(b) /// 2 }}} Would C++/C differences matter (there's a smattering of both in these classes)? A patch against 2.8.12 is at http://kiwistrawberry.us/research/hash-patch-2_8_12.patch if you want to see actual code. As for the implementation, I'll note that one of my concerns is that I was going to have to write really nasty low level code to make this fast, but it turns out that my implementation is already faster than the current 2.8.12 and the code is quite decent (i.e. more pythonic than c-like). Actual speed comparisons vary greatly amongst classes, but it's clear that there's more optimizations to do around the corners in the poly classes (in particular some of the __getitem__ implementations are pretty slow). By the way, one of the coding sprint projects, led by Willem Jan Palenstijn (Leiden, NL) implements modifying doctest so it prints out the time for each block of tests (both wall and cpu); this might be useful for testing the impact of your patch, a little... I don't think he's posting anything yet though. William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: plotQt.c with Qt 4.3.2 + GUI for pari/gp
On Nov 13, 2007 1:35 AM, Téragone [EMAIL PROTECTED] wrote: I'm also working on a GUI for pari/gp. I started with mathGuide from which I removed the Python plugin Wow, thanks for pointing out mathGuide, which I had never heard about before. Since this is in English, I translation of the web page http://www.math.uni-siegen.de/ring/mathGUIde is here: http://babelfish.altavista.com/babelfish/trurl_pagecontent?lp=de_enurl=http%3A%2F%2Fwww.math.uni-siegen.de%2Fring%2FmathGUIde%2Findex.html It's interesting how that project is fairly similar to Sage (http://sagemath.org) in many ways. E.g., see this screenshot which looks very much like the screen shot for mathGUIde: http://sagemath.org/screen_shots/.html/sage-screenshot.png . I call gp in application mode, not in library mode. I use most of the code of gp.c and I redirect the output to a global string. Maybe not the best way, but it's easy to code and it work. Interestingly the Sage GP GUI also uses GP rather than the PARI C library for the GUI (though Sage also uses PARI). Anyway, here is a screenshot from Sage that actually looks a lot like yours below: http://sagemath.org/screen_shots/.html/sage-pari.png For now, my new code is source compatible only. All the modifications have been include in #ifdef/#endif so the original gp compile and work with no modification. Still many hours before it will work correctly. I use Qt so it must be possible to port it on many OS. I only tested it on Linux. Here a screenshot : http://pages.videotron.com/teragone/gui.png Please, let me know if you have interest for that. I'm excited that there is so much enthusiasm and energy for open source projects / guis, etc., related to mathematical software right now. I wonder if you've solved any problems I don't know how to solve (and conversely) related to such things? -- William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: plotQt.c with Qt 4.3.2 + GUI for pari/gp
I tried emailing the author of mathguide several times a few years ago, with no response. Of course, my email was in English, which might have been the problem. In any case, I did not get the impression that it was open source. Either I am wrong or I presume this guy got permission from the mathguide author to distribute a derived work based on his code? On Nov 12, 2007 9:33 PM, William Stein [EMAIL PROTECTED] wrote: On Nov 13, 2007 1:35 AM, Téragone [EMAIL PROTECTED] wrote: I'm also working on a GUI for pari/gp. I started with mathGuide from which I removed the Python plugin Wow, thanks for pointing out mathGuide, which I had never heard about before. Since this is in English, I translation of the web page http://www.math.uni-siegen.de/ring/mathGUIde is here: http://babelfish.altavista.com/babelfish/trurl_pagecontent?lp=de_enurl=http%3A%2F%2Fwww.math.uni-siegen.de%2Fring%2FmathGUIde%2Findex.html It's interesting how that project is fairly similar to Sage (http://sagemath.org) in many ways. E.g., see this screenshot which looks very much like the screen shot for mathGUIde: http://sagemath.org/screen_shots/.html/sage-screenshot.png . I call gp in application mode, not in library mode. I use most of the code of gp.c and I redirect the output to a global string. Maybe not the best way, but it's easy to code and it work. Interestingly the Sage GP GUI also uses GP rather than the PARI C library for the GUI (though Sage also uses PARI). Anyway, here is a screenshot from Sage that actually looks a lot like yours below: http://sagemath.org/screen_shots/.html/sage-pari.png For now, my new code is source compatible only. All the modifications have been include in #ifdef/#endif so the original gp compile and work with no modification. Still many hours before it will work correctly. I use Qt so it must be possible to port it on many OS. I only tested it on Linux. Here a screenshot : http://pages.videotron.com/teragone/gui.png Please, let me know if you have interest for that. I'm excited that there is so much enthusiasm and energy for open source projects / guis, etc., related to mathematical software right now. I wonder if you've solved any problems I don't know how to solve (and conversely) related to such things? -- William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: plotQt.c with Qt 4.3.2 + GUI for pari/gp
On Nov 13, 2007 2:50 AM, David Joyner [EMAIL PROTECTED] wrote: I tried emailing the author of mathguide several times a few years ago, with no response. Of course, my email was in English, which might have been the problem. In any case, I did not get the impression that it was open source. Either I am wrong or I presume this guy got permission from the mathguide author to distribute a derived work based on his code? It's probably open source I guess, since the web page says (translated) Freeware is and with all sources is spread. Perhaps a German speaker could look at http://www.math.uni-siegen.de/ring/mathGUIde and confirm this. William On Nov 12, 2007 9:33 PM, William Stein [EMAIL PROTECTED] wrote: On Nov 13, 2007 1:35 AM, Téragone [EMAIL PROTECTED] wrote: I'm also working on a GUI for pari/gp. I started with mathGuide from which I removed the Python plugin Wow, thanks for pointing out mathGuide, which I had never heard about before. Since this is in English, I translation of the web page http://www.math.uni-siegen.de/ring/mathGUIde is here: http://babelfish.altavista.com/babelfish/trurl_pagecontent?lp=de_enurl=http%3A%2F%2Fwww.math.uni-siegen.de%2Fring%2FmathGUIde%2Findex.html It's interesting how that project is fairly similar to Sage (http://sagemath.org) in many ways. E.g., see this screenshot which looks very much like the screen shot for mathGUIde: http://sagemath.org/screen_shots/.html/sage-screenshot.png . I call gp in application mode, not in library mode. I use most of the code of gp.c and I redirect the output to a global string. Maybe not the best way, but it's easy to code and it work. Interestingly the Sage GP GUI also uses GP rather than the PARI C library for the GUI (though Sage also uses PARI). Anyway, here is a screenshot from Sage that actually looks a lot like yours below: http://sagemath.org/screen_shots/.html/sage-pari.png For now, my new code is source compatible only. All the modifications have been include in #ifdef/#endif so the original gp compile and work with no modification. Still many hours before it will work correctly. I use Qt so it must be possible to port it on many OS. I only tested it on Linux. Here a screenshot : http://pages.videotron.com/teragone/gui.png Please, let me know if you have interest for that. I'm excited that there is so much enthusiasm and energy for open source projects / guis, etc., related to mathematical software right now. I wonder if you've solved any problems I don't know how to solve (and conversely) related to such things? -- William -- William Stein Associate Professor of Mathematics University of Washington http://wstein.org --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] #962: automatic precision extension for long decimal literals does very strange things
Hello, I made an initial version of a patch to fix number #962 and posted it http://www.sagetrac.org/sage_trac/ticket/962 . Could people look at it to see if the changes it makes are sane? I don't have a whole lot of familiarity with these things so it'd definitely be good to have additional input. Since these changes will break a lot of doctests, it'll be good to get it right (or a close approximation) the first time. Thanks, Mike --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: plotQt.c with Qt 4.3.2 + GUI for pari/gp
On 11/12/07, William Stein [EMAIL PROTECTED] wrote: I'm excited that there is so much enthusiasm and energy for open source projects / guis, etc., related to mathematical software right now. I wonder if you've solved any problems I don't know how to solve (and conversely) related to such things? I also stumbled accross Reinteract, which is similar to the Sage notebook, only it's running on a GUI (Gtk in this case): http://blog.fishsoup.net/2007/11/10/reinteract-better-interactive-python/ didier --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: plotQt.c with Qt 4.3.2 + GUI for pari/gp
The original Webpage says [mathGUIde] ist Freeware und wird mit allen Quellen verbreitet which means that it's freeware and is distributed with all the sources. So it doesn't state any licensing details e.g. whether you may modify the source etc. Maybe I should download it and look into the sourcefiles? There could be some hidden licensing information. Fabio On Nov 13, 2007 3:53 AM, William Stein [EMAIL PROTECTED] wrote: On Nov 13, 2007 2:50 AM, David Joyner [EMAIL PROTECTED] wrote: I tried emailing the author of mathguide several times a few years ago, with no response. Of course, my email was in English, which might have been the problem. In any case, I did not get the impression that it was open source. Either I am wrong or I presume this guy got permission from the mathguide author to distribute a derived work based on his code? It's probably open source I guess, since the web page says (translated) Freeware is and with all sources is spread. Perhaps a German speaker could look at http://www.math.uni-siegen.de/ring/mathGUIde and confirm this. William On Nov 12, 2007 9:33 PM, William Stein [EMAIL PROTECTED] wrote: On Nov 13, 2007 1:35 AM, Téragone [EMAIL PROTECTED] wrote: I'm also working on a GUI for pari/gp. I started with mathGuide from which I removed the Python plugin Wow, thanks for pointing out mathGuide, which I had never heard about before. Since this is in English, I translation of the web page http://www.math.uni-siegen.de/ring/mathGUIde is here: http://babelfish.altavista.com/babelfish/trurl_pagecontent?lp=de_enurl=http%3A%2F%2Fwww.math.uni-siegen.de%2Fring%2FmathGUIde%2Findex.html It's interesting how that project is fairly similar to Sage (http://sagemath.org) in many ways. E.g., see this screenshot which looks very much like the screen shot for mathGUIde: http://sagemath.org/screen_shots/.html/sage-screenshot.png . I call gp in application mode, not in library mode. I use most of the code of gp.c and I redirect the output to a global string. Maybe not the best way, but it's easy to code and it work. Interestingly the Sage GP GUI also uses GP rather than the PARI C library for the GUI (though Sage also uses PARI). Anyway, here is a screenshot from Sage that actually looks a lot like yours below: http://sagemath.org/screen_shots/.html/sage-pari.png For now, my new code is source compatible only. All the modifications have been include in #ifdef/#endif so the original gp compile and work with no modification. Still many hours before it will work correctly. I use Qt so it must be possible to port it on many OS. I only tested it on Linux. Here a screenshot : http://pages.videotron.com/teragone/gui.png Please, let me know if you have interest for that. I'm excited that there is so much enthusiasm and energy for open source projects / guis, etc., related to mathematical software right now. I wonder if you've solved any problems I don't know how to solve (and conversely) related to such things? -- William -- William Stein Associate Professor of Mathematics University of Washington http://wstein.org --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---