opencxx-users

Re: [Opencxx-users] Testing rel 2.8 under Cygwin

[Grigorenko D. <pos...@na...>]

Hello

> Grigorenko Dmitriy wrote:
> > Come filenames (aux.cc, aux.h and others) were changed in rel_2_8 branch
> > only (not in the main branch).
> > Is it right?
>
> Correct.
I`m sorry, I`m not sure that I`ve understood you. Is it correct, that come
files in the MAIN branch and in the HEAD have deprecated names?

> >  $ libtool -dlopen /home/Dima/opencxx/opencxx/libocc_mop.la
> > --mode=execute /home/Dima/opencxx/occ --private--external-driver -E
> > <tnested-class.ii --private--output tnested-class.occ.tmp
> >  exec: /home/Dima/opencxx/occ: not found
> > Why is it sow?
>
> Perhaps the file should be 'occ.exe' ? Most likely '$(EXEEXT)' or
> @EXEEXT@ is missing somewhere (in tester2.in ???).
No, there is no occ at all.
Cygwin recognizes executible exe-files correct.
If we have a.exe, we can type ./a to run it.

> I think it is doable, but I more and more seriously think that this
> plugins functionality is not really worth the effort that we need to put
> into maintaining it. I am just about to raise this subject on the
> mailing list, so stay tuned and pls. let me (and others) know what you
> think.
Is there anything that is possible only in plugin mode? I do not know any
examle.
In case it is not useful we can refuse this option.
In any case our program do not use plugins.

> Thanks!
You are welcome!
Dmitriy Grigorenko

[Opencxx-users] Next tasks (was: OpenC++ 2.8 is out!)

[<se...@in...>]

On Mon, 2004-08-30 at 09:04, Grzegorz Jakacki wrote:
> Hi,
> HUGE THANKS go to Stefan Seefeld, a veteran OpenC++ user and reuser (see 
> breathtaking project Synopsis), who contributed a lot of ideas, design 
> and code to this release.

What do we do with synopsis,
- must we try to merge,
  1) Stephan, do you have suggestions.
  2) Synopsis seams to have reformat source files,
     do we have a standard for preferred format.
     Stephan, have you used an utility(indent).
  3) Header files have changed from .h to .hh, must we do it.

> Release 2.8 is definitely a step toward modularization of OpenC++.
> This promises benefits, but also poses new challenges.
> Personally, I look at the OpenC++ project not only as a useful tool
> and library, but also as a perfect spot to get in touch with
> compiler technology and honest workout in OO-design and software
> engineering. I hope that the new challenges will continue to attract
> contributors and that soon we can arrive at fully functional,
> multiplatform,

Can we create a branch( branch-281).
- do we merge first, and branch from there.

Dmitriy can you help with with the Cygwin compatibility.

> highly reusable OpenC++ Frontend Library.
> Would that qualify for '3.0' ? ;-)

Thanks for your great work Grzegorz.

Re: [Opencxx-users] symbol lookup

[Grzegorz J. <ja...@ac...>]

Hi!

Stefan Seefeld wrote:
> as far as I understand, symbol lookup in opencxx works by traversing a 
> linked list
> of 'Environment' objects. However, the C++ lookup rules seem to suggest 
> at least
> a tree, if not a (directed acyclic) graph. Just think about multiple 
> inheritance
> as the most obvious example.
> 
> Am I missing something ? 

I don't pretend to understand all the details of how it works, but I 
think there is more to it than you describe. Look at 
'SearchBaseOrUsing()', which to my understanding handles the 'forks' in 
the acyclic graph.

Also note that OpenC++ is non-validating, consequently it rightfully 
ignores, say, ambiguities in multiple inheritance.

On the other hand, lookup implementation is missing several features 
(e.g. ADL).

> Does symbol lookup currently work in such cases ?

Why not check empirically?... :-)

Best regards
Grzegorz

[Opencxx-users] OpenC++ 2.8 is out!

[Grzegorz J. <ja...@ac...>]

Hi,

OpenC++ 2.8 is out. This release encompasses a number of changes that 
are being introduced in a broader effort to refactor OpenC++ so that 
parts of it are reusable as independent libraries.

Major changes in this release include:

* separation of lexer&parser into self-contained library,

* move out of gc, which from now on is an external dependency,
   not a part of OpenC++,

* refactored and simplified test harness

* prototype of script based driver (occ2) which is meant
   to replace platform-sensitive logic in occ executable.

As usual, several bugs were fixed too.

2.8 was tested on Debian and RedHat (CF), it is also very close to 
working out-of-the-box on Cygwin, and reasonably close on XP.

This release would not be possible without large and small contributions 
by many volunteers, in particular:

   Yann Dirson
   Dmitri Grigorenko
   Brian Kahne
   Jean-Sebastien Mouret
   Asen Kovachev
   Marc Ordinas i Llopis
   Horacio Pena
   Ferdinand Prantl
   Shishir Ramam
   Stefan Seefeld
   Gilles Seguin
   Alexandre Tolmos

(Hope I did not miss anybody; if this is the case, offended are 
encouraged to challenge me for a duel immediately.)

Big thanks to all contributors!

HUGE THANKS go to Stefan Seefeld, a veteran OpenC++ user and reuser (see 
breathtaking project Synopsis), who contributed a lot of ideas, design 
and code to this release.

Release 2.8 is definitely a step toward modularisation of OpenC++. This 
promises benefits, but also poses new challenges. Personally, I look at 
the OpenC++ project not only as a useful tool and library, but also as a 
perfect spot to get in touch with compiler technology and honest workout 
in OO-design and software engineering. I hope that the new challenges 
will continue to attract contributors and that soon we can arrive at 
fully functional, multiplatform, highly reusable OpenC++ Frontend 
Library. Would that qualify for '3.0' ? ;-)

Best regards
Grzegorz

Re: [Opencxx-users] Testing rel 2.8 under Cygwin

[Grzegorz J. <ja...@ac...>]

Hi Dimitry,

Good to hear from you.

Grigorenko Dmitriy wrote:
> Come filenames (aux.cc, aux.h and others) were changed in rel_2_8 branch 
> only (not in the main branch).
> Is it right?

Correct.

> I`ve checked opencxx rel_2_8 under Cygwin.
> I typed
>  ./bootstrap
> and got
>  You should update your `aclocal.m4' by running aclocal.
> Is it correct?
> Then I typed
>  ./aclocal; ./configure-for-testing
> You can see log file cft.log in the attachment.
> Then I tried
>  make check
> The log file is mkck.log.
> And there is a tester2.log file in the attachment too.
> More detail information is following
>  Dima@home <mailto:Dima@home> 
> ~/opencxx/testsuite/testsuite.dir/comp/tnested/compilation
>  $ libtool -dlopen /home/Dima/opencxx/opencxx/libocc_mop.la 
> --mode=execute /home /Dima/opencxx/occ --private--external-driver -E 
> <tnested-class.ii --private--output tnested-class.occ.tmp
>  libtool: Couldn't find configure.ac nor configure.in file
> As far as it was discussed here there is an error in the Cygwin 
> automake/autoconf tools.
>  
> The solution is to create an empty configure.in files. I`ve patched 
> tester2 for it.
> Log file is tester2_2.log
> More detail information is following:
>  Dima@home <mailto:Dima@home> 
> ~/opencxx/testsuite/testsuite.dir/comp/tnested/compilation
>  $ libtool -dlopen /home/Dima/opencxx/opencxx/libocc_mop.la 
> --mode=execute /home/Dima/opencxx/occ --private--external-driver -E 
> <tnested-class.ii --private--output tnested-class.occ.tmp
>  exec: /home/Dima/opencxx/occ: not found
> Why is it sow?

Perhaps the file should be 'occ.exe' ? Most likely '$(EXEEXT)' or 
@EXEEXT@ is missing somewhere (in tester2.in ???).

> I tried to run occ2 ang got:
>  occ2: unknown option `--private--external-driver'

This is strange.

> I`ve run make. Log file is mk.log
> After this I tried make check and got the results into tester2_3.log
> I`ve solved this problem by
>  chmod +x occ2
> It works (on some tests)!
 > You can see final tester2_f.log file.

I have seen it. Thank you!
Plugins fail, the rest works.

> More information about failed tests:
>  Dima@home <mailto:Dima@home> 
> ~/opencxx/testsuite/testsuite.dir/comp/typeidt/compilation-via-plugin
>  $ libtool -dlopen /home/Dima/opencxx/opencxx/libocc_mop.la 
> --mode=execute /home/Dima/opencxx/occ --private--external-driver 
> --private--libtool-plugins -S TestClass <typeidt-test.ii 
> --private--output typeidt-test.occ.tmp
>  lt_dlopen(TestClass-init.la) failed: can't open the module
> I do not know how to fix it. May be it is inpossable in Cygwin/WinXP OS. 
> Is not it?

I think it is doable, but I more and more seriously think that this 
plugins functionality is not really worth the effort that we need to put 
into maintaining it. I am just about to raise this subject on the 
mailing list, so stay tuned and pls. let me (and others) know what you 
think.

> I am not sure with how my changes have to be made. So I did not commit 
> anything.



> There is tester2.in patch in the attachment.

Unfortunately it is alredy too late for 2.8, but I have already applied 
to MAIN. I believe it is important to build some momentum around Cygwin 
and .NET ports, so that OpenC++ becomes easier accessible for Windows 
users. Having said that your help is very valuable.

Thanks!
Grzegorz

[Opencxx-users] symbol lookup

[Stefan S. <se...@sy...>]

hi there,

as far as I understand, symbol lookup in opencxx works by traversing a linked list
of 'Environment' objects. However, the C++ lookup rules seem to suggest at least
a tree, if not a (directed acyclic) graph. Just think about multiple inheritance
as the most obvious example.

Am I missing something ? Does symbol lookup currently work in such cases ?

Thanks,
		Stefan

Re: [Opencxx-users] Version 2.8, ready for release

[Grzegorz J. <ja...@ac...>]

Gilles J. Seguin wrote:
> Hello,
> 
> What else needs to be done.

That's about it. I am testing now, hopefully I can release tomorrow.

(I hold on with thanks and credits until tests pass... :-) )

Regards
Grzegorz

> 
> Gilles J. Seguin
> 
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
> 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
> Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
> http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
> _______________________________________________
> Opencxx-users mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opencxx-users

[Opencxx-users] Version 2.8, ready for release

[<se...@in...>]

Hello,

What else needs to be done.

Gilles J. Seguin

Re: [Opencxx-users] Rel. 2.8 about to ship

[Grzegorz J. <ja...@ac...>]

Gilles J. Seguin wrote:
> On Mon, 2004-08-23 at 09:32, Grzegorz Jakacki wrote:
> 
>>Gilles J. Seguin wrote:
>>
>>>On Sun, 2004-08-22 at 11:15, Grzegorz Jakacki wrote:
>>>
>>>
>>>>Hi,
>>>>
>>>>I made (hopefully final) changes to rel_2_8 branch. I tested on
>>>>Debian and RedHat, I could not make plugins functional on RedHat,
>>>>but otherwise things look fine.
>>>
>>>
>>>Could you be more specific on the plugins for RedHat. thanks
>>
>>The only RedHat I can test on is x86-linux2 on CompileFarm (Red Hat 
>>Linux 3.2.2-5). This is what I got:
>>
>><log>
>>...
>>Testing comp/tnested/compilation-via-plugin ...
>>+ cd testsuite.dir/comp/tnested/compilation-via-plugin
>>+ /home/users/j/ja/jakacki/tmp/test-opencxx-22221/opencxx/occ2 -I/h
>>ome/users/j/ja/jakacki/local/gc6.2/include -L/home/users/j/ja/jakac
>>ki/local/gc6.2/i686-pc-linux-gnu/x86-linux2.cf.sourceforge.net/lib
> 
> 
> why have added this library path.
> The thing does not work there, but it work on my computer.

Perhaps you have some version of OpenC++ installed system-wide?

>>-m /home/users/j/ja/jakacki/tmp/test-opencxx-22221/opencxx/testsuit
>>e/comp/tnested-class.mc
>>/home/users/j/ja/jakacki/tmp/test-opencxx-22221/opencxx/testsuite/c
>>omp/tnested-class.mc:9: a metaclass should include at least one not
>>-inlined virtual function: NestedClass
>>
>>g++: /usr/lib/gcc-lib/i386-linux/2.95.2/crtbeginS.o: No such file o
>>r directory
> 
> 
> You are suppose to be in 3.2.2 not 2.95.2,
> add -v option to g++

Good point. Looks like some kind of misconfigured gcc. Anyway, I am not 
fixing it before 2.8.

BR
Grzegorz

> 
> good luck with this,
> 
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
> 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
> Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
> http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
> _______________________________________________
> Opencxx-users mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opencxx-users

Re: [Opencxx-users] examples directory

[Grzegorz J. <ja...@ac...>]

Gilles J. Seguin wrote:
> On Mon, 2004-08-23 at 09:48, Grzegorz Jakacki wrote:
> 
>>Gilles J. Seguin wrote:
>>
>>>I have committed "added using namespace Opencxx" to
>>>verbose/VerboseClass.mc verbose2/VerboseClass2.mc
>>>
>>>"added rule to build SyncClass.o" to wrapper/Makefile
>>>in opencxx/examples directory.
>>>
>>>still have
>>>occ2 -c WrapperClass.cc -o WrapperClass.o
>>>occ2 -o SyncClass.exe WrapperClass.o SyncClass.o -locc -lltdl
>>>occ2: don't know how to handle file `WrapperClass.o'
>>
>>I fixed compilation problem, but there is still something wrong. I am 
>>leaving it for now.
>>
>>
>>>we can forgot that the examples directory does not work
>>>from a build directory for release 2.8, will make it work later.
> 
> 
> I have done it.  At first I was thinking, it was necessary
> because "make dist" was including everything CVS directory included
> from the examples directory.
> Somebody 

Me.

> has put examples in EXTRA_DIST, do not do that.

Thanks for improvig this.

>>That's right, examples will run seamlesly only after occ is installed. 
>>But I think there is no harm in leaving it that way. Making examples 
>>work with uninstalled occ is a big effort and it will definitely pollute 
>>makefiles with things like 'libtool -dlopen @abs_top_builddir@/libocc.la 
>>occ' etc., while examples should be clean and simple.
> 
> 
> I have to reread the documentation.  Since we are doing more relinking
> because of "make check".  configure utility will be better with
> --disable-fast-install by default.  Or force relinking at install time.

OK, but my strong suggestion is to freeze rel_2_8 now, release, merge to 
trunk before weekend and continue the development in trunk or in 
sandboxes. Otherwise 2.8 will never ship.

BR
Grzegorz

>>Moreover, examples can also be used to validate if installation went
>>ok. Maybe later we can make examples run under 'make installcheck' ?
> 
> 
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
> 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
> Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
> http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
> _______________________________________________
> Opencxx-users mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opencxx-users

Re: [Opencxx-users] Rel. 2.8 about to ship

[<se...@in...>]

On Mon, 2004-08-23 at 09:32, Grzegorz Jakacki wrote:
> Gilles J. Seguin wrote:
> > On Sun, 2004-08-22 at 11:15, Grzegorz Jakacki wrote:
> > 
> >>Hi,
> >>
> >>I made (hopefully final) changes to rel_2_8 branch. I tested on
> >>Debian and RedHat, I could not make plugins functional on RedHat,
> >>but otherwise things look fine.
> > 
> > 
> > Could you be more specific on the plugins for RedHat. thanks
> 
> The only RedHat I can test on is x86-linux2 on CompileFarm (Red Hat 
> Linux 3.2.2-5). This is what I got:
> 
> <log>
> ...
> Testing comp/tnested/compilation-via-plugin ...
> + cd testsuite.dir/comp/tnested/compilation-via-plugin
> + /home/users/j/ja/jakacki/tmp/test-opencxx-22221/opencxx/occ2 -I/h
> ome/users/j/ja/jakacki/local/gc6.2/include -L/home/users/j/ja/jakac
> ki/local/gc6.2/i686-pc-linux-gnu/x86-linux2.cf.sourceforge.net/lib

why have added this library path.
The thing does not work there, but it work on my computer.

> -m /home/users/j/ja/jakacki/tmp/test-opencxx-22221/opencxx/testsuit
> e/comp/tnested-class.mc
> /home/users/j/ja/jakacki/tmp/test-opencxx-22221/opencxx/testsuite/c
> omp/tnested-class.mc:9: a metaclass should include at least one not
> -inlined virtual function: NestedClass
> 
> g++: /usr/lib/gcc-lib/i386-linux/2.95.2/crtbeginS.o: No such file o
> r directory

You are suppose to be in 3.2.2 not 2.95.2,
add -v option to g++

good luck with this,

Re: [Opencxx-users] examples directory

[<se...@in...>]

On Mon, 2004-08-23 at 09:48, Grzegorz Jakacki wrote:
> Gilles J. Seguin wrote:
> > I have committed "added using namespace Opencxx" to
> > verbose/VerboseClass.mc verbose2/VerboseClass2.mc
> > 
> > "added rule to build SyncClass.o" to wrapper/Makefile
> > in opencxx/examples directory.
> > 
> > still have
> > occ2 -c WrapperClass.cc -o WrapperClass.o
> > occ2 -o SyncClass.exe WrapperClass.o SyncClass.o -locc -lltdl
> > occ2: don't know how to handle file `WrapperClass.o'
> 
> I fixed compilation problem, but there is still something wrong. I am 
> leaving it for now.
> 
> > we can forgot that the examples directory does not work
> > from a build directory for release 2.8, will make it work later.

I have done it.  At first I was thinking, it was necessary
because "make dist" was including everything CVS directory included
from the examples directory.
Somebody has put examples in EXTRA_DIST, do not do that.

> That's right, examples will run seamlesly only after occ is installed. 
> But I think there is no harm in leaving it that way. Making examples 
> work with uninstalled occ is a big effort and it will definitely pollute 
> makefiles with things like 'libtool -dlopen @abs_top_builddir@/libocc.la 
> occ' etc., while examples should be clean and simple.

I have to reread the documentation.  Since we are doing more relinking
because of "make check".  configure utility will be better with
--disable-fast-install by default.  Or force relinking at install time.

> Moreover, examples can also be used to validate if installation went
> ok. Maybe later we can make examples run under 'make installcheck' ?

Re: [Opencxx-users] examples directory

[Grzegorz J. <ja...@ac...>]

Gilles J. Seguin wrote:
> I have committed "added using namespace Opencxx" to
> verbose/VerboseClass.mc verbose2/VerboseClass2.mc
> 
> "added rule to build SyncClass.o" to wrapper/Makefile
> in opencxx/examples directory.
> 
> still have
> occ2 -c WrapperClass.cc -o WrapperClass.o
> occ2 -o SyncClass.exe WrapperClass.o SyncClass.o -locc -lltdl
> occ2: don't know how to handle file `WrapperClass.o'

I fixed compilation problem, but there is still something wrong. I am 
leaving it for now.

> we can forgot that the examples directory does not work
> from a build directory for release 2.8, will make it work later.

That's right, examples will run seamlesly only after occ is installed. 
But I think there is no harm in leaving it that way. Making examples 
work with uninstalled occ is a big effort and it will definitely pollute 
makefiles with things like 'libtool -dlopen @abs_top_builddir@/libocc.la 
occ' etc., while examples should be clean and simple. Moreover, examples 
can also be used to validate if installation went ok. Maybe later we can 
make examples run under 'make installcheck' ?

Best regards
Grzegorz

> 
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
> 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
> Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
> http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
> _______________________________________________
> Opencxx-users mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opencxx-users

Re: [Opencxx-users] Rel. 2.8 about to ship

[Grzegorz J. <ja...@ac...>]

Gilles J. Seguin wrote:
> On Sun, 2004-08-22 at 11:15, Grzegorz Jakacki wrote:
> 
>>Hi,
>>
>>I made (hopefully final) changes to rel_2_8 branch. I tested on
>>Debian and RedHat, I could not make plugins functional on RedHat,
>>but otherwise things look fine.
> 
> 
> Could you be more specific on the plugins for RedHat. thanks

The only RedHat I can test on is x86-linux2 on CompileFarm (Red Hat 
Linux 3.2.2-5). This is what I got:

<log>
...
Testing comp/tnested/compilation-via-plugin ...
+ cd testsuite.dir/comp/tnested/compilation-via-plugin
+ /home/users/j/ja/jakacki/tmp/test-opencxx-22221/opencxx/occ2 -I/h
ome/users/j/ja/jakacki/local/gc6.2/include -L/home/users/j/ja/jakac
ki/local/gc6.2/i686-pc-linux-gnu/x86-linux2.cf.sourceforge.net/lib
-m /home/users/j/ja/jakacki/tmp/test-opencxx-22221/opencxx/testsuit
e/comp/tnested-class.mc
/home/users/j/ja/jakacki/tmp/test-opencxx-22221/opencxx/testsuite/c
omp/tnested-class.mc:9: a metaclass should include at least one not
-inlined virtual function: NestedClass

g++: /usr/lib/gcc-lib/i386-linux/2.95.2/crtbeginS.o: No such file o
r directory
g++: /usr/lib/gcc-lib/i386-linux/2.95.2/crtendS.o: No such file or
directory
occ2: command failed: libtool --tag=CXX --mode=link --silent g++ -a
void-version -module -rpath /home/users/j/ja/jakacki/tmp/test-openc
xx-22221/opencxx/testsuite/testsuite.dir/comp/tnested/compilation-v
ia-plugin NestedClass-init.lo -o NestedClass-init.la
g++: /usr/lib/gcc-lib/i386-linux/2.95.2/crtbeginS.o: No such file o
r directory
g++: /usr/lib/gcc-lib/i386-linux/2.95.2/crtendS.o: No such file or
directory
occ2: command failed: libtool --tag=CXX --mode=link --silent g++ -a
void-version -module -rpath /home/users/j/ja/jakacki/tmp/test-openc
xx-22221/opencxx/testsuite/testsuite.dir/comp/tnested/compilation-v
ia-plugin tnested-class.lo -o NestedClass.la
FAIL: comp/tnested/compilation-via-plugin
</log>

Looks very strange, but other tests worked, so I decided to give up on 
this issue.

BR
Grzegorz

>>I would appreciate if somebody could test on Cygwin (Gilles, could 
>>you?). If that works, I will release.
> 
> 
> Sorry, I can not
> 
> 
>>Rel. 2.8 is a first release after cleaning and refactoring campaign 
>>started several months ago. This works goes on and HEAD is already
>>ahead rel_2_8. After the release I would like to open a discussion
>>on current project goals, priorities and code sharing with Synopsis.
> 
> 
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
> 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
> Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
> http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
> _______________________________________________
> Opencxx-users mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opencxx-users

[Opencxx-users] examples directory

[<se...@in...>]

I have committed "added using namespace Opencxx" to
verbose/VerboseClass.mc verbose2/VerboseClass2.mc

"added rule to build SyncClass.o" to wrapper/Makefile
in opencxx/examples directory.

still have
occ2 -c WrapperClass.cc -o WrapperClass.o
occ2 -o SyncClass.exe WrapperClass.o SyncClass.o -locc -lltdl
occ2: don't know how to handle file `WrapperClass.o'

we can forgot that the examples directory does not work
from a build directory for release 2.8, will make it work later.

Re: [Opencxx-users] doc/reference.pdf

[<se...@in...>]

On Sun, 2004-08-22 at 11:06, Grzegorz Jakacki wrote:
> Gilles J. Seguin wrote:
> > Bonjour,
> > 
> > How is 'doc/reference.pdf' produce ?
> > How do we modify it ?
> 
> We can't. We don't have source of this file.
> 
> > The file is not synchronize with the HTML ones.
> 
> Where exactly?

OK, corrections done in the HTML files where not reflected in
the pdf file.

> I would like to abandond this PDF file (because we cannot maintain it). 
> If there is any information there that is not in HTML, we should move 
> it. If there are any figures or anything that cannot be easily copied, 
> let me know and I will extract it with Acrobat suite and send you.

Re: [Opencxx-users] Rel. 2.8 about to ship

[<se...@in...>]

On Sun, 2004-08-22 at 11:15, Grzegorz Jakacki wrote:
> Hi,
> 
> I made (hopefully final) changes to rel_2_8 branch. I tested on
> Debian and RedHat, I could not make plugins functional on RedHat,
> but otherwise things look fine.

Could you be more specific on the plugins for RedHat. thanks

> I would appreciate if somebody could test on Cygwin (Gilles, could 
> you?). If that works, I will release.

Sorry, I can not

> Rel. 2.8 is a first release after cleaning and refactoring campaign 
> started several months ago. This works goes on and HEAD is already
> ahead rel_2_8. After the release I would like to open a discussion
> on current project goals, priorities and code sharing with Synopsis.

[Opencxx-users] Rel. 2.8 about to ship

[Grzegorz J. <ja...@ac...>]

Hi,

I made (hopefully final) changes to rel_2_8 branch. I tested on Debian 
and RedHat, I could not make plugins functional on RedHat, but otherwise 
things look fine.

I would appreciate if somebody could test on Cygwin (Gilles, could 
you?). If that works, I will release.

Rel. 2.8 is a first release after cleaning and refactoring campaign 
started several months ago. This works goes on and HEAD is already ahead 
rel_2_8. After the release I would like to open a discussion on current 
project goals, priorities and code sharing with Synopsis.

BR
Grzegorz

Re: [Opencxx-users] doc/reference.pdf

[Grzegorz J. <ja...@ac...>]

Gilles J. Seguin wrote:
> Bonjour,
> 
> How is 'doc/reference.pdf' produce ?
> How do we modify it ?

We can't. We don't have source of this file.

> The file is not synchronize with the HTML ones.

Where exactly?

I would like to abandond this PDF file (because we cannot maintain it). 
If there is any information there that is not in HTML, we should move 
it. If there are any figures or anything that cannot be easily copied, 
let me know and I will extract it with Acrobat suite and send you.

[...]
> ps: what are the plan with release of version 2.8

See the next e-mail.

BR
Grzegorz

Re: [Opencxx-users] unit testing

[Grzegorz J. <ja...@ac...>]

Stefan Seefeld wrote:
> hi there,
> 
> I'v just checked in some preliminary unit test for the OpenCxx module
> into synopsis. You can check it out from the (subversion) repository
> at http://synopsis.fresco.org/viewsvn/synopsis-Synopsis/ or a snapshot
> at http://synopsis.fresco.org/download/ starting tomorrow.

I had a look. Cool.

[...]
> I'd suggest we work together to set up something similar for the opencxx 
> project.
> Please note that the above applets make use of some important changes I 
> made
> to the API, specifically to make the code more modular and unit-testable.

Absoutely, I would like to get down to it after 2.8 release.

BR
Grzegorz

[Opencxx-users] doc/reference.pdf

[<se...@in...>]

Bonjour,

How is 'doc/reference.pdf' produce ?
How do we modify it ?
The file is not synchronize with the HTML ones.

suggestions,

gilles

ps: what are the plan with release of version 2.8

[Opencxx-users] unit testing

[Stefan S. <se...@sy...>]

hi there,

I'v just checked in some preliminary unit test for the OpenCxx module
into synopsis. You can check it out from the (subversion) repository
at http://synopsis.fresco.org/viewsvn/synopsis-Synopsis/ or a snapshot
at http://synopsis.fresco.org/download/ starting tomorrow.

Here is what I'm doing:

the OpenCxx tests rely on 'resources' that are set up first. These
are little testing applets built specifically for the unit tests.
Right now there are three: one for the Lexer, one for the Parser,
and one for the Encoding class.

For each there is a test suite, consisting of an input (source) file,
as well as an expected output file.

If you run 'python setup.py test -s OpenCxx.Encoding' the OpenCxx.Encoding
test suite will be performed (i.e. first the 'Encoding' applet will be
compiled and then all source files pertaining to this suite will be processed,
and the output compared.

Right now there's only one source file per test suite (read: each suite contains
exactly one test), but that's just because I'm still setting everything up.

You can find the sources for these testing applets at
http://synopsis.fresco.org/viewsvn/synopsis-Synopsis/trunk/tests/OpenCxx/src/ .

I'd suggest we work together to set up something similar for the opencxx project.
Please note that the above applets make use of some important changes I made
to the API, specifically to make the code more modular and unit-testable.

Best regards,
		Stefan

Re: [Opencxx-users] understanding TypeInfo

[Grzegorz J. <ja...@ac...>]

Stefan Seefeld wrote:
> Grzegorz Jakacki wrote:
> 
> 
>> Encoding no longer depends on TypeInfo, offending methods have been 
>> promoted to EncodingUtil.
> 
> 
> yeah, I noticed. But is that really a solution ? 

Strictly speaking not. However the details of encoding were not 
encapsulated even before EncodingUtil, since e.g. TypeInfo knew about it.

> The Encoding
> class should encapsulate the representation of declarator type/names,
> and moving half of it out into 'EncodingUtil' breaks that encapsulation,
> i.e. even though EncodingUtil is now a separate piece of code, it
> very much depends on the intrinsics of the Encoding class. In fact
> even the TypeInfo class depends on a large degree on Encoding, so I
> wonder whether we shouldn't try to refactor these classes. 

Yes.

> Another
> issue is that of TypeInfo's expanded vs. non-expanded form...

What's unclear about it?

Best regards
Grzegorz

>>>> Typeinfo --- this class is intended to represent types, however the
>>>
>>>
>>>
>>> representation is not unique, because:
>>>
>>>>  - typedefs are expanded lazily (thus many operations need
>>>>    an environment, to be able to look up typedefs definitions),
>>>>  - dereferences ('*') are applied lazily.
>>>
>>>
>>>
>>>
>>> yeah. Is this lazyness really necessary ? Does it gain performance ?
>>> I'm just wondering because without it (i.e. putting 'Normalize' into
>>> the constructor), most methods could become 'const'...
>>> Or alternatively the inner representation (encoding) could become 
>>> 'mutable',
>>> such that normalizing a TypeInfo keeps the invariants intact and thus 
>>> can
>>> semantically be considered a no-op.
>>
>>
>>
>> This is a profound question.
>>
>> In theory this laziness is not necessary. However in practice it may 
>> be very useful (for the users, not implementors). 'gcc' is very much 
>> praised for expanding typedefs lazily, which results in much more 
>> readable error messages (especially if one uses lots of templates). My 
>> suggestion is not to remove it unless it is really really really 
>> troublesome to maintain it.
> 
>  >
>  > The solution with 'mutable' has a drawback: the const operations will
>  > actually change the observable state of the object, because
>  > 'FullTypeName()' and 'MakePtree()' will not necesarily return the same
>  > result after normalization.
> 
> Speaking of which, what does 'MakePtree' actually generate ? Or, to
> push the question further, what do the two encodings of a declarator
> contain ? What is a 'type encoding', and what is a 'name encoding' ?
> Why isn't one enough ? A declarator either declares a type or an instance
> (let's count function as an instance for now), and in the first case
> I need the type encoding, in the second the name encoding, no ?
> What am I missing ?
> 
> Thanks for any clarifications,
>         Stefan
> 
> 
> -------------------------------------------------------
> SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
> 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
> Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
> http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
> _______________________________________________
> Opencxx-users mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opencxx-users

Re: [Opencxx-users] Missing ".so" in .la file

[Grzegorz J. <ja...@ac...>]

Thanks, Gilles. I rebuilt everything and it is ok, the problem must have 
stemed from mixing files generated by different versions of libtool.

BR
Grzegorz

Gilles J. Seguin wrote:
> On Mon, 2004-08-16 at 07:37, Grzegorz Jakacki wrote:
> 
>>Hi,
>>
>>This is a fragment of some libocc_mop.la file created by
>>libtool 1.5.6:
> 
> 
> I have the same version here.
> And I do have the .so
> 
> If you cd in .libs directory, does the files have it.
> change directory back in directory opencxx where are the *.la .
> $ rm *.la
> $ make
> Suppose to relink.
> Does the command show
> -o .libs/libocc_mop.so.0.0.0
> 
> When you do ./configure add the -C option.
> $ cd top_builddir
> $ rm config.cache
> $ $(top_srcdir)/configure -C  #may be other things also
> $ grep libltdl_cv_shlibext * config.cache
> libltdl_cv_shlibext=${libltdl_cv_shlibext=.so}
> 
> hope this help
> 
> 
>>       # The name that we can dlopen(3).
>>       dlname='libocc_mop.0'
>>
>>       # Names of this library.
>>       library_names='libocc_mop.0.0.0 libocc_mop.0 libocc_mop'
>>
>>       # The name of the static archive.
>>       old_library='libocc_mop.a'
>>
>>Why no '.so' in dlname and library_names ???
>>
>>Thanks
>>Grzegorz
> 
> 
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
> 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
> Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
> http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
> _______________________________________________
> Opencxx-users mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opencxx-users

Re: [Opencxx-users] Missing ".so" in .la file

[<se...@in...>]

On Mon, 2004-08-16 at 07:37, Grzegorz Jakacki wrote:
> Hi,
> 
> This is a fragment of some libocc_mop.la file created by
> libtool 1.5.6:

I have the same version here.
And I do have the .so

If you cd in .libs directory, does the files have it.
change directory back in directory opencxx where are the *.la .
$ rm *.la
$ make
Suppose to relink.
Does the command show
-o .libs/libocc_mop.so.0.0.0

When you do ./configure add the -C option.
$ cd top_builddir
$ rm config.cache
$ $(top_srcdir)/configure -C  #may be other things also
$ grep libltdl_cv_shlibext * config.cache
libltdl_cv_shlibext=${libltdl_cv_shlibext=.so}

hope this help

>        # The name that we can dlopen(3).
>        dlname='libocc_mop.0'
> 
>        # Names of this library.
>        library_names='libocc_mop.0.0.0 libocc_mop.0 libocc_mop'
> 
>        # The name of the static archive.
>        old_library='libocc_mop.a'
> 
> Why no '.so' in dlname and library_names ???
> 
> Thanks
> Grzegorz