This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: patch for flow.c: flow_delete_insn
At 19:45 10.05.00, Richard Henderson wrote:
>On Wed, May 10, 2000 at 12:01:35PM +0200, Franz Sirl wrote:
> > /home/fsirl/cvsx/gccm/gcc/cppinit.c:323: number of bb notes in insn chain
> > (8) != n_basic_blocks (7)
>
>Yes, it's amazing the things enable checking finds. Why I
>don't just leave it on, I don't know. Probably I just forget
>to consistently add the thing to the configure line...
Well, time for my repeat-every-few-months statement :-)):
Enable --enable-checking=tree,gc,misc by default in CVS and only disable it
on release branches, just like we handle the WARN_CFLAGS. I did some some
measurements a while ago, the bootstrap/check time only increases a few
percent and is worth it IMHO.
I don't suggest the 'gcac' and 'rtl' checking options here, because they
lead to an enormous increase of compile time and required memory, eg. 'rtl'
needs ~400MB swap on an 128MB G4/400MHz PPC machine and a few files
(insn-recog.c and insn-extract.c being the worst) take hours to compile...
Thanks for the fix.
Franz.