Ed Smith-Rowland wrote:
All,
FWIW, I would like to add my support for creating a branch for gpc
with the eventual goal
of integrating Pascal into mainline. I would bootstrap and test this
branch, report bugs and
do my best to help with solutions although I'm new at this.
I think both projects would benefit. I'll venture some predictions:
GCC
1. Another language would test the middle and back ends more and
expose more bugs.
There would be the process, structure, and motivation to fix them
because gpascal is integrated.
2. I know of one person from the gpc world who participates in gcc and
understands the gcc-4+ internals.
I'm sure more would follow.
3. It would be cool to have more languages!
--enable-languages=c,c++,java,objc,obj-
c++,fortran,treelang,ada,pascal, Mwahahahah!!!
plus modula !
4. There are multi-language Pascal-C, etc systems that GCC would be
left out of without gpascal.
GPC
1. The later versions of gcc are getting faster. Moving gpc off of
gcc-3
and into gcc-4 would speed things up. GCC-3.4.6 is the end of the
line for GCC-3.
2. By targeting integration rather than proximity, all the recent work
in
getting gpascal to work with gcc-4 would be less likely to bitrot.
3. GPC would get much wider exposure. It would probably eventually
ship
along with the rest of gcc in OS distributions including *ahem*
MacOSX.
4. I've seen questions on the gpc mailing list about the availability
of OpenMP for Pascal.
I don't think OpenMP treats Pascal but gcc has OpenMP integrated
into the C, C++,
and Fortran frontends and has a common OMP runtime.
If OMP was wanted as an extension or became part of OpenMP then
your best bet is GCC.
5. I believe that more people would join the GPC maintenance effort.
I get the sense that there are wounds from old battles. Based on the
replies I believe that many of the complaints
(perceived and/or real): poor response to bug reports, second class
citizen feelings - would fall by the wayside if
a branch with the goal of full integration were initiated.
This would require that friendly dialog would ensue and that new
people are eased
into a different culture in a welcoming manner. This would require a
change in workflow
and project structure for gpascal. But I think it is worth it.
I really hope that these issues and others can be resolved for the
benefit of all.
This is an excellent résumé of the situation.
Thanks,
Adriaan van Os
