Mail Archives: cygwin/2004/07/15/09:57:52
At 06:24 AM 7/15/2004, Gerrit P. Haase wrote:
>Robert wrote:
>
> > Who's maintaining GCC for Cygwin?
> > Ours is getting old. Gcc 3.4.1 came out a couple of weeks ago.
>
>What is wrong with 3.3.x release series? Are there any serious bugs?
>Are there issues (for you)? Why do you need 3.4.x? Gfortran isn't
>included, precompiled headers do not work on Windows/Cygwin, important
>bugfixes are backported to 3.3.x.
>
>
> > I could go ahead and compile it, but I don't know where the patches are to
> > make it use -mno-cygwin.
>
>All the stuff is in the CVS repository. Check the sources (or the
>patchfile included with the Cygwin release of GCC).
>
> > Mingw is using Gcc 3.4.0 as a candidate. I seen our version of 3.4.0, as
>
>In the 3.4.0 release there was a serious bug in C++ and some Java build
>issues. I have not tried to build 3.4.1 yet, sorry.
AFAICT everyone who has tried has been successful in building and testing
the standard gcc-3.4.1 and 3.4.2 on cygwin, but shows well over 100
testsuite failures due to non-support of pch. At least 3 different people
have posted the results on gcc-testsuite. I don't count problems such as
my failure to build it in 32-bit mode on x64. C++ in my own current
project doesn't build with 3.3.4, but is good with 3.4.x. I agree that my
project is not "important" nor trendy enough to backport fixes to
3.3.x. According to testsuite, Java is slightly better in 3.4.x. I don't
know that anyone has tackled the cygwin patches; inclusion of pch will no
doubt make it much more time consuming. I haven't seen anyone post
testsuite results for that mingw 3.4.0 version, so your evidence that mingw
is ahead of cygwin is lacking.
Tim Prince
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/
- Raw text -