This is the mail archive of the gcc-bugs@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: real comparison failure, i386



  In message <199902250137.UAA07121@blastula.phys.columbia.edu>you write:
  > ------- =_aaaaaaaaaa0
  > Content-Type: text/plain; charset="us-ascii"
  > Content-ID: <7114.919906621.1@blastula.phys.columbia.edu>
  > 
  > I am seeing a comparison failure with current CVS on x86.  This is a
  > real bug.  It appears that register allocation made different choices
  > in stage2 than stage3.  Files affected are:
  > 
  > c-common.o c-typeck.o caller-save.o calls.o cpphash.o cse.o expr.o flow.o
  > g++.o g77.o gcc.o gcj.o gcov.o gcse.o genattrtab.o genextract.o global.o
  > i386.o loop.o print-tree.o profile.o real.o recog.o reg-stack.o regclass.o
  > reload.o reload1.o sbitmap.o sched.o toplev.o varasm.o ch/actions.o
  > ch/expr.o ch/satisfy.o cp/pt.o cp/search.o cp/typeck.o f/target.o
  > java/decl.o java/expr.o java/gjavah.o java/jcf-dump.o java/jcf-parse.o
  > java/verify.o
  > 
  > I've attached a diff of the asm dumps for loop.o.
When did you last update?  I checked in a fix for a bug which sounds awful
familiar this afternoon (last change to regclass.c).

I've done, hmmm, 5 3-stages on my 686 box today.

jeff


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