This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: cpplib: More interface cleanups


Hi Franz,

Franz Sirl wrote:-

> Hmm, maybe you forgot something to checkin?

Hmmm, maybe, rth checked in c-pragma.c which I forgot, but I could
swear CVS showed me nothing left to checkin when I forgot that.

I can tell you I've run at least 3 bootstraps since that patch, and
g++ doesn't report the streams of regression errors in any of them for
me.

> Looking at gcc-testresults it seems like a lot of linux people have
> this problem now (~920 g++ FAILs).  Hmm, I just notice that
> gcc.dg/pack-test-* is failing too and that this happens in your
> report too. Why do these tests fail on your machine? They are
> certainly regressions.

I don't know why those pack-tests fail; but I didn't think I was the
first to have them failing.  This is from my log:-

Executing on host: /usr/src/gcc2/build/gcc/xgcc -B/usr/src/gcc2/build/gcc/ /usr/src/gcc2/gcc/testsuite/gcc.dg/pack-test-1.c    -ansi -pedantic-errors -S  -o pack-test-1.s    (timeout = 300)
/usr/src/gcc2/gcc/testsuite/gcc.dg/pack-test-1.c: In function `s1':
/usr/src/gcc2/gcc/testsuite/gcc.dg/pack-test-1.c:44: size of array `test_0' is negative
/usr/src/gcc2/gcc/testsuite/gcc.dg/pack-test-1.c:45: size of array `test_1' is negative
.... similarly.


Maybe something has gone wrong with the #pragma callback.  I'll have a
look.

Meanwhile, this is what I get from cvs:-

bash-2.04$ cvs update
? build
M gcc/c-decl.c
M gcc/objc/lang-options.h
M gcc/objc/lang-specs.h
M gcc/objc/objc-act.c
bash-2.04$ 

"build" is my build directory.  The only difference between my tree
and CVS is the objective C patch I just posted, and I'm sure that has
no relevance.  Is there some way that my local CVS might think it's in
sync, but in reality isn't?  Something is very odd here...  Any ideas
on what to try next?  Is my understanding of CVS wrong?

Could you do me a favour, and see if each file is updated as
incdicated in that patch in your local tree?  I can only think that my
CVS and the server are different, but claiming they're the same for
some reason.

Cheers,

Neil.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]