Menu
▾
▴
synopsis-devel
You can subscribe to this list here.
| 2000 |
Jan
|
Feb
|
Mar
(4) |
Apr
|
May
(2) |
Jun
(3) |
Jul
|
Aug
(9) |
Sep
|
Oct
|
Nov
|
Dec
(1) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2001 |
Jan
(29) |
Feb
(12) |
Mar
(19) |
Apr
(11) |
May
(10) |
Jun
(7) |
Jul
(3) |
Aug
(8) |
Sep
|
Oct
|
Nov
|
Dec
|
| 2002 |
Jan
(3) |
Feb
(7) |
Mar
(2) |
Apr
(3) |
May
|
Jun
(4) |
Jul
(1) |
Aug
(4) |
Sep
(2) |
Oct
(2) |
Nov
(13) |
Dec
(15) |
| 2003 |
Jan
(10) |
Feb
(9) |
Mar
(11) |
Apr
(10) |
May
(17) |
Jun
(4) |
Jul
(24) |
Aug
(5) |
Sep
(11) |
Oct
|
Nov
|
Dec
|
|
From: Stefan S. <se...@sy...> - 2003-09-20 19:26:55
|
hi there, Synopsis is moving out of sf.net, to http://synopsis.fresco.org. The cvs repo is already in place (http://synopsis.fresco.org/viewcvs for a cgi interface, and :pserver:ano...@sy...:/cvs/synopsis as CVSROOT for direct (read-only) cvs access. Other services are following, such as the mailing lists. I'll do the batch subscription using the subscriber list from sf.net. Unfortunately that won't preserve the passwords. A monthly reminder will tell you your new password so you can reset it. I'll send out another mail once we are done. In the meantime, sorry for the inconvenience. Regards, Stefan |
|
From: David A. <da...@bo...> - 2003-09-13 01:05:30
|
Stefan Seefeld <se...@sy...> writes: > hi David, > > I would guess sf.net had some trouble with their exim configuration. > My response (and other mails that day) got refused. I'm sending > it again: > > > > Hi David, > > I'm sorry I never got back to you concerning an earlier bug report about > synopsis running on boost code. Stephen has spent some serious time > getting synopsis up to speed with boost, and there are portions > of the synopsis code I never touched. Now he seems to be busy elsewhere > so I have to make myself familiar with his work. > > I'll try to help if I can... Thanks so much Stefan! -- Dave Abrahams Boost Consulting www.boost-consulting.com |
|
From: Stefan S. <se...@sy...> - 2003-09-12 21:48:01
|
hi David, I would guess sf.net had some trouble with their exim configuration. My response (and other mails that day) got refused. I'm sending it again: Hi David, I'm sorry I never got back to you concerning an earlier bug report about synopsis running on boost code. Stephen has spent some serious time getting synopsis up to speed with boost, and there are portions of the synopsis code I never touched. Now he seems to be busy elsewhere so I have to make myself familiar with his work. I'll try to help if I can... Regards, Stefan |
|
From: David A. <da...@bo...> - 2003-09-08 03:18:01
|
-- Dave Abrahams Boost Consulting www.boost-consulting.com |
|
From: Stefan S. <se...@sy...> - 2003-09-07 18:40:35
|
David Abrahams wrote: >>Hmm, I should try to use 'AC_PREREQ(2.56)', rename 'configure.in' >>to 'configure.ac', or somesuch. > > > AC_PREREQ worked. ok, fix committed (together with your patch). > BTW, the way you're detecting Python isn't so scalable (obviously). > Maybe you should try extracting the version by invoking python and > doing something. yes, using distutils will solve that (and some other problems). >>I'm looking into ways to replace the toplevel configure stuff >>by python's distutils to solve a couple of other problems I'm >>running into. > > > Careful: distutils doesn't really work with C++ well. Yeah, I have to extend it with my own commands to build C/C++ modules. The ones provided by distutils (notably 'build_ext') are only useful for the most simple C/C++ extensions :-( I'm looking into this right now. As the distutils are not developed (or even actively maintained) any more, I'm even considering to try to port it (i.e., distutils) over to scons, but that's a different story...:-) Regards, Stefan |
|
From: David A. <da...@bo...> - 2003-09-07 18:14:52
|
Stefan Seefeld <se...@sy...> writes: > Hi David, > > David Abrahams wrote: >> The enclosed patch supports Python 2.3 > > thanks a lot ! > >> BTW, this version of configure.in uses AC_HELP_MESSAGE which is not >> supported in the stable version of autoconf (2.13); the autoconf >> wrapper is supposed to invoke a more recent version "depending on the >> contents of configure.in/configure.ac", but I don't know what, >> precisely, it depends on. > > Hmm, I should try to use 'AC_PREREQ(2.56)', rename 'configure.in' > to 'configure.ac', or somesuch. AC_PREREQ worked. BTW, the way you're detecting Python isn't so scalable (obviously). Maybe you should try extracting the version by invoking python and doing something. > As I don't have an old autoconf version around, may I ask you to > try the above and tell me whether that is enough to make your > 'autoconf wrapper' choose the right version ? > > I'm looking into ways to replace the toplevel configure stuff > by python's distutils to solve a couple of other problems I'm > running into. Careful: distutils doesn't really work with C++ well. -- Dave Abrahams Boost Consulting www.boost-consulting.com |
|
From: David A. <da...@bo...> - 2003-09-07 18:11:41
|
-- Dave Abrahams Boost Consulting www.boost-consulting.com |
|
From: Stefan S. <se...@sy...> - 2003-09-07 17:37:15
|
Hi David, David Abrahams wrote: > The enclosed patch supports Python 2.3 thanks a lot ! > BTW, this version of configure.in uses AC_HELP_MESSAGE which is not > supported in the stable version of autoconf (2.13); the autoconf > wrapper is supposed to invoke a more recent version "depending on the > contents of configure.in/configure.ac", but I don't know what, > precisely, it depends on. Hmm, I should try to use 'AC_PREREQ(2.56)', rename 'configure.in' to 'configure.ac', or somesuch. As I don't have an old autoconf version around, may I ask you to try the above and tell me whether that is enough to make your 'autoconf wrapper' choose the right version ? I'm looking into ways to replace the toplevel configure stuff by python's distutils to solve a couple of other problems I'm running into. Thanks, Stefan |
|
From: David A. <da...@bo...> - 2003-09-07 16:53:02
|
The enclosed patch supports Python 2.3 BTW, this version of configure.in uses AC_HELP_MESSAGE which is not supported in the stable version of autoconf (2.13); the autoconf wrapper is supposed to invoke a more recent version "depending on the contents of configure.in/configure.ac", but I don't know what, precisely, it depends on. It seems to fail instead on Cygwin: autoconf: Undefined macros: configure.in:129: AC_HELP_STRING([--enable-ctool],[Configure with ctool support]), Maybe you need to add something to configure.in? -- Dave Abrahams Boost Consulting www.boost-consulting.com |
|
From: Stefan S. <se...@sy...> - 2003-08-10 00:53:56
|
Nathaniel Smith wrote: > Does it work to use Synopsis's C++ parser with C code? It seems like > it should, but I seem to recall some warning note in the source > somewhere about it not working... Hi there, this is just a little note to let you know that I discovered a nice little C parser (http://ctool.sf.net) which I intend to use as a backend for synopsis. Doing this will provide support for C to synopsis' current AST. The ctool looks to me quite powerful, so I'm tempted to try to make it support the sxr frontend, too. In fact, it may become more interesting to make the AST module more complete, i.e. support non-declaration nodes. Right now the sxr tool is only accessible in a C++-only tool that actually bypasses the Synopsis.AST code. May be it is time for this to change. Regards, Stefan |
|
From: Stefan S. <se...@sy...> - 2003-08-01 00:09:54
|
yang qu wrote: > Hi! > I am trying to document some C++ programs. There are > some headers that are in the directory "../include". > As introduced in the manual, I use the following > command to setup the path. > -Wp,-I,../include > But, the Synopsis keeps telling me that those header > files can't be found. What's wrong in my operation? > thank you! It's a bug in synopsis' parameter handling. You can use either 'synopsis -I ../include ...' or 'synopsis -Wp,-I../include ...', both will work with the current version. I'll see whether I can come up with a quick fix for the problem. Regards, Stefan |
|
From: yang qu <qu...@ya...> - 2003-07-31 22:28:59
|
Hi! I am trying to document some C++ programs. There are some headers that are in the directory "../include". As introduced in the manual, I use the following command to setup the path. -Wp,-I,../include But, the Synopsis keeps telling me that those header files can't be found. What's wrong in my operation? thank you! --yang __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com |
|
From: Stefan S. <se...@sy...> - 2003-07-24 23:22:02
|
yang qu wrote: > HI! > I installed the Synopsis on a RH8.0 using gcc3.2. the > first step was to run the ./configure, and the second > was to run the make. it seemed the compilation went > well. however, when I tried the demo, I got the > following message: > Traceback (most recent call last): > File "./synopsis", line 29, in ? > from Synopsis.Core import Type, AST, Util > ImportError: No module named Synopsis.Core you either need to set your PYTHONPATH to point to the 'sources' or you have to install synopsis with 'make install'. Regards, Stefan |
|
From: yang qu <qu...@ya...> - 2003-07-24 22:57:26
|
HI!
I installed the Synopsis on a RH8.0 using gcc3.2. the
first step was to run the ./configure, and the second
was to run the make. it seemed the compilation went
well. however, when I tried the demo, I got the
following message:
Traceback (most recent call last):
File "./synopsis", line 29, in ?
from Synopsis.Core import Type, AST, Util
ImportError: No module named Synopsis.Core
when did I do wrong?
thank you!
best regards
yang
__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
http://sitebuilder.yahoo.com
|
|
From: David A. <da...@bo...> - 2003-07-18 23:59:48
|
Any suggestions?
dave@penguin /cygdrive/c/src/Synopsis/demo/Boost
$ make boostdemo2
../../bin/synopsis -P boost.synopsis "-Wc,formatter=HTML Formatter - No Frames"
C++ Parser: Parsing boost/boost/python.hpp
Finding info for 'g++-3.2' ... Failed
Finding info for 'gcc' ... Found
Finding info for 'g++' ... Found
Finding info for 'c++' ... Found
Finding info for 'cc' ... Failed
Finding info for 'c++' ... Found
Finding info for 'gcc' ... Found
Finding info for 'g++' ... Found
Finding info for 'g++-2.95' ... Failed
Finding info for 'g++-3.0' ... Failed
Finding info for 'g++-3.1' ... Failed
Warning: The specified compiler (g++-3.2) could not be found.
You may want to retry with the full pathname of the compiler
or with it in your path. If you don't have this compiler, you
will need to modify the C++ Parser part of your config file.
Warning: Falling back to compiler 'g++'
Combined AST: Linking Combined_AST
C++ Parser: Parsing boost/boost/python.hpp
Segmentation Violation caught while processing boost/python/return_opaque_pointer.hpp at line 52
make: *** [boostdemo2] Error 255
dave@penguin /cygdrive/c/src/Synopsis/demo/Boost
$ make boostdemo
../../bin/synopsis -P boost.synopsis
Combined AST: Linking Combined_AST
Traceback (most recent call last):
File "../../bin/synopsis", line 337, in ?
main()
File "../../bin/synopsis", line 263, in main
if project_file: return do_project_file(project_file)
File "../../bin/synopsis", line 231, in do_project_file
ex.get_output(ex.get_output_names()[0][0])
File "/cygdrive/c/src/Synopsis/Synopsis/Core/Executor.py", line 435, in get_output
ast = self.__input_exec.get_output(name)
File "/cygdrive/c/src/Synopsis/Synopsis/Core/Executor.py", line 410, in get_output
ast = self.prepare_output(name, 1)
File "/cygdrive/c/src/Synopsis/Synopsis/Core/Executor.py", line 384, in prepare_output
ast = exec_obj.get_output(name)
File "/cygdrive/c/src/Synopsis/Synopsis/Core/Executor.py", line 274, in get_output
module.resolve([], ast, self.__action.config())
File "Synopsis/Linker/Linker.py", line 323, in resolve
File "Synopsis/Linker/XRefCompiler.py", line 149, in execute
File "Synopsis/Linker/XRefCompiler.py", line 65, in do_compile
ValueError: unpack list of wrong size
--
Dave Abrahams
Boost Consulting
www.boost-consulting.com
|
|
From: Stefan S. <se...@sy...> - 2003-07-16 12:40:24
|
Hi Arieh Arieh Schneier wrote: > For example for the 3 files attached (taken from the python example and > broken into 3 files) I tried, > > synopsis -p Python -o test1.syn test1.py > synopsis -p Python -o test2.syn test2.py > synopsis -p Python -o test3.syn test3.py > synopsis -f HTML -o foobar test1.syn test2.syn test3.syn > > but when looking at the inheritance graph it does not combine them into > 1 graph, rather it has 3 seperate ones. as far as the parsing goes, it is correct, as the types you derive from are in the local scope (because you imported them into it). What is needed is a way to tell the linker to treat these local types as aliases, i.e. replace them by the types they are referring to. That should be easy to add as a linker-specific module. I'll look into it (though I'll be out of town for the rest of the week). Regards, Stefan |
|
From: Arieh S. <ar...@vp...> - 2003-07-16 00:29:30
|
Stefan Seefeld wrote: > Arieh Schneier wrote: > >> I am however having problems with the inheritance graph. The project >> is quite large so all classes have been placed in seperate files, >> however I would like to have the inheritance graph combine them >> (maybe not all into 1 huge graph but at least something). At the >> moment the only way I have managed to do this is by cating all the >> files into 1 large file and running synopsis over that file. > > > similar to compilers, synopsis can generate 'object files' which then are > linked together when the final processing (here formatting) is done. > Instead of concatenating the source files together, you can generate > one 'object > file' per processed source file, and only then format all of them > together, > somewhat like this: > > synopsis -p Python -o foo.syn foo.py > synopsis -p Python -o bar.syn bar.py > synopsis -f HTML -o foobar foo.syn bar.syn I have tried that already, however that doesnt seem to work. For example for the 3 files attached (taken from the python example and broken into 3 files) I tried, synopsis -p Python -o test1.syn test1.py synopsis -p Python -o test2.syn test2.py synopsis -p Python -o test3.syn test3.py synopsis -f HTML -o foobar test1.syn test2.syn test3.syn but when looking at the inheritance graph it does not combine them into 1 graph, rather it has 3 seperate ones. Thankyou in advance. Arieh. |
|
From: Arieh S. <ar...@vp...> - 2003-07-14 00:42:25
|
Hi, I am very new to Synopsis. I have been looking at it for about a week now. I have had great success with it and am thinking of using it to document a project done partly in python and partly in c. I am however having problems with the inheritance graph. The project is quite large so all classes have been placed in seperate files, however I would like to have the inheritance graph combine them (maybe not all into 1 huge graph but at least something). At the moment the only way I have managed to do this is by cating all the files into 1 large file and running synopsis over that file. Does anyone have any suggestions on how I can do this without having to combine all the files into 1? Thankyou in advance. Arieh Schneier |
|
From: Martin K. <mar...@gm...> - 2003-07-11 18:20:13
|
Am 2003.07.11 17:25 schrieb(en) Stefan Seefeld: > Martin Klaffenboeck wrote: >> Hello, >> >> I'm trying to port Synopsis to FreeBSD. >> Now I have the following error on building Synopsis-0.5: >> >> c++ -O1 -ggdb -fno-inline -I.. -I../gc/include -I../occ -fPIC - >> Wall - >> c -o ../occ/member.o ../occ/member.cc >> c++ -O1 -ggdb -fno-inline -I.. -I../gc/include -I../occ -fPIC - >> Wall - >> c -o ../occ/cbodywalk.o ../occ/cbodywalk.cc >> make: don't know how to make w. Stop >> gmake[4]: *** [../gc/.libs/libgc.a] Fehler 2 > > [...] > > does 'make -C Synopsis/Parser/C++/gc/' work ? It seems the error is in > that sub-target. Yes, that works. But it doesn't satisfy, because I cannot port this way, to port means in this case to modify the Makefile. Let me have a look. Martin -- If you've got an idea and need help, or just need general encouragement, write me a message. ;-) |
|
From: Stefan S. <se...@sy...> - 2003-07-11 15:24:07
|
Martin Klaffenboeck wrote: > Hello, > > I'm trying to port Synopsis to FreeBSD. > Now I have the following error on building Synopsis-0.5: > > c++ -O1 -ggdb -fno-inline -I.. -I../gc/include -I../occ -fPIC -Wall - > c -o ../occ/member.o ../occ/member.cc > c++ -O1 -ggdb -fno-inline -I.. -I../gc/include -I../occ -fPIC -Wall - > c -o ../occ/cbodywalk.o ../occ/cbodywalk.cc > make: don't know how to make w. Stop > gmake[4]: *** [../gc/.libs/libgc.a] Fehler 2 [...] does 'make -C Synopsis/Parser/C++/gc/' work ? It seems the error is in that sub-target. The gc build system uses automake/libtool, which I'm not familiar with. I'v no idea what the 'w' stands for make attemps to build. Regards, Stefan |
|
From: David A. <da...@bo...> - 2003-07-11 12:27:36
|
Stefan Seefeld <se...@sy...> writes: > David Abrahams wrote: > >> http://synopsis.sourceforge.net/design.html >> ------------------------------------------- >> THe AST diagram here could be very useful, but where are "Type", >> "Modifier", etc.? It should probably be put in the User Guide. > > I finally updated two of the diagrams. The AST diagram is now > more accurate, though still incomplete. It's hard to balance > to keep the amount of detail at a managable level. Cool; nice work! -- Dave Abrahams Boost Consulting www.boost-consulting.com |
|
From: Martin K. <mar...@gm...> - 2003-07-11 11:46:50
|
Hello, I'm trying to port Synopsis to FreeBSD. Now I have the following error on building Synopsis-0.5: c++ -O1 -ggdb -fno-inline -I.. -I../gc/include -I../occ -fPIC -Wall - c -o ../occ/member.o ../occ/member.cc c++ -O1 -ggdb -fno-inline -I.. -I../gc/include -I../occ -fPIC -Wall - c -o ../occ/cbodywalk.o ../occ/cbodywalk.cc make: don't know how to make w. Stop gmake[4]: *** [../gc/.libs/libgc.a] Fehler 2 gmake[4]: Leaving directory `/var/tmp/root/ports/Synopsis/work/ Synopsis/Synopsis/Parser/C++/syn' gmake[3]: *** [all] Fehler 1 gmake[3]: Leaving directory `/var/tmp/root/ports/Synopsis/work/ Synopsis/Synopsis/Parser/C++' gmake[2]: *** [C++] Fehler 2 gmake[2]: Leaving directory `/var/tmp/root/ports/Synopsis/work/ Synopsis/Synopsis/Parser' gmake[1]: *** [Parser] Fehler 2 gmake[1]: Leaving directory `/var/tmp/root/ports/Synopsis/work/ Synopsis/Synopsis' gmake: *** [Synopsis] Fehler 2 *** Error code 2 Stop in /root/ports/Synopsis. So, what do I have to do to get this working? Thanks, Martin -- If you've got an idea and need help, or just need general encouragement, write me a message. ;-) |
|
From: Stefan S. <se...@sy...> - 2003-07-11 04:42:27
|
David Abrahams wrote: > http://synopsis.sourceforge.net/design.html > ------------------------------------------- > > THe AST diagram here could be very useful, but where are "Type", > "Modifier", etc.? It should probably be put in the User Guide. I finally updated two of the diagrams. The AST diagram is now more accurate, though still incomplete. It's hard to balance to keep the amount of detail at a managable level. Stephen, please let me know whether you'd like to have some smaller diagrams for the manual, I can easily generate them. Or else I could add them myself, but as the manual is your work, I'm reluctant to make changes. Regards, Stefan |
|
From: Stefan S. <se...@sy...> - 2003-07-08 12:25:15
|
Larry Evans wrote: > I'm using Ptree::ToString quite a bit, and now I'm running > out of memory. As shown in the cvs occ/buffer.cc, the statement: > > delete [] buf; > > is commented out. I could understand some justification for this > if gc were being used, but even then, why not just delete the > buffer since any previous value of buf returned by ToString would > be invalid. I did try to reconfigure with --enable-GC, but I'm > still running out of memory. > > Any help appreciated. This is Stephen's area, so I would much prefer if he could comment on this, as he has done lots of modifications and enhancements to the occ code. Hope he'll find the time to respond... Stefan |
|
From: Larry E. <jca...@pr...> - 2003-07-08 01:08:53
|
I'm using Ptree::ToString quite a bit, and now I'm running out of memory. As shown in the cvs occ/buffer.cc, the statement: delete [] buf; is commented out. I could understand some justification for this if gc were being used, but even then, why not just delete the buffer since any previous value of buf returned by ToString would be invalid. I did try to reconfigure with --enable-GC, but I'm still running out of memory. Any help appreciated. |
11 messages has been excluded from this view by a project administrator.