From: Randy W. S. <RandyS@ThePierianSpring.org> - 2002-09-13 04:15:48
|
> -----Original Message----- From: > min...@li... > [mailto:min...@li...]On Behalf Of > Sternbach, William [IT] Sent: Wednesday, September 11, 2002 11:12 AM > To: 'min...@li...' Subject: [Mingw-users] Ioannis > departure from this list. > > > I agree with Ioannis. > > I have reported issues with gcc 2.95.3 in the past, mainly that it can > no longer compile the very important "Perl Compiler" source code which > is written in "C" and is used by half of the websites on the internet. > > gcc 2.95.2 can compile it successfully. Someone tampering with the > windows.h header file since gcc 2.95.2 caused a problem which prevents > this huge "Perl Compiler" source code to compile. I still have been > unable to use gcc 2.95.3 (let alone gcc 3.2) because of this bug. > > I send multiple Emails to this list about 6 months ago about this bug, > but all of my requests were ignored. > > The owners of mingw need to take a more aggressive position to support > Mingw gcc. They need to be more concerned with fixing every bug than > to ignore the bugs. I know gcc is shareware, but the owner of Mingw > needs to manage all bug reports better. He needs to make sure that all > bugs are fixed (not ignored). > > I think Ioannis was upset because when he reported a bug in Mingw gcc > that was important to him, the owners of Mingw spent more time telling > him the bug wasn't important. What they should have done was spend > that time trying to fix the bug. > > I enjoy using gcc 2.95.2, but I wish I could use the newer versions if > this bug compiling the "Perl Source Code" could be fixed. But I don't > even think anyone is even thinking about this bug anymore. > > What a shame. > > - Bill > This may be a little off-topic from the intent of your message, but I'm curious as to what bug your referring to. I've successfully compiled perl 5.6 with both gcc 2.95.3-6 and gcc 3.1, and perl 5.8 with gcc 2.95.3-6, gcc 3.1, gcc 3.2 (as well as msvc 6). I have not run in to any bugs. The compiler emitted a few warnings, but I saw no problems when investigating the source of the warnings. The only warning I'm not sure about and have not had a chance to investigate involves the STRUCT2PTR/STRUCT2RAWPTR macros. IRC, the warning occured in perlhost.h and was something like 'Invalid pointer offset to POD member', but perl passed all tests, and I have not experienced any problems. Since the functions where the warning occured are pretty fundamental to the functioning of perl on windows, I'm fairly certain I would have run into a segfault if there were any real problem. Randy |