This is the mail archive of the
mailing list for the GCC project.
Re: libjava build failure on branch for sparc-sun-solaris2.7
- To: jsturm at one-point dot com
- Subject: Re: libjava build failure on branch for sparc-sun-solaris2.7
- From: Mark Mitchell <mark at codesourcery dot com>
- Date: Tue, 22 May 2001 23:55:07 -0700
- Cc: jbuck at synopsys dot COM, gcc at gcc dot gnu dot org, java at gcc dot gnu dot org
- Organization: CodeSourcery, LLC
- References: <3B0AF3AF.email@example.com>
>>>>> "Jeff" == Jeff Sturm <firstname.lastname@example.org> writes:
Jeff> I saw this too. It turns out that Solaris <sys/socket.h>
Jeff> #define bind __xnet_bind
Wa, ha, ha, ha.
Now we see just how important it will be to fix V3 to allow users to
define the _XOPEN_FOO and _POSIX_FOO things they are used to on each
and every platform.
Unfortunately, this functionality was not designed into V3. It will
be a lot of work, because you must handle all the feature macros that
users are allowed to define on each and every different operating
system. There is simply no way to avoid lots of ugly
platform-dependent #ifdefs. And, you want to be sufficiently clever
that you can use the same binary libstdc++-v3.so, independent of which
#defines are turned on. (Hey, the system vendors do it, so can we.)
Short term, I'd say you'll have to do the `#undef' thing, grotesque
though it is.
Sorry for the breakage. If you do not feel like fixing it, it is
certainly my regression, and I will fix it. Put it in GNATS, mark it
high priority. But, I would be very grateful if you would fix it. :-)
Mark Mitchell email@example.com
CodeSourcery, LLC http://www.codesourcery.com