This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: gcc bug?
- To: "Chris Cochrane" <chris_cochrane at brooktrout dot com>
- Subject: Re: gcc bug?
- From: Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>
- Date: Tue, 15 Jun 1999 11:14:10 +0200
- Cc: <egcs-bugs at egcs dot cygnus dot com>,<bug-gcc at gnu dot org>
At 19:07 14.06.99 , Chris Cochrane wrote:
>Hi,
>I can't believe I've found a gcc bug at this stage in its life,
>but can't explain the misbehavior we see any other way. Perhaps
>someone can show me the light. The first function in the following
>file fails as described below...Chris
Hmm, for which values does it fail? Can you turn this in an example that
aborts on error condition? I can't see anything suspicious here, except
maybe the bc 12,2,.L2, a blt (I expected a beq). But I couldn't get it to
fail on some testruns I tried.
funcFails:
stwu 1,-16(1)
mflr 0
stw 0,20(1)
addis 9,0,num_lines@ha
lhz 0,num_lines@l(9)
subfc 0,0,3
li 0,0
adde 0,0,0
subfic 3,3,1
subfe 3,3,3
neg 3,3
and. 11,0,3
bc 12,2,.L2
li 3,1
li 4,2
li 5,3
bl assert
.L2:
lwz 0,20(1)
mtlr 0
la 1,16(1)
blr
gcc turns the code into 2 Scc's and the results are and'ed together.
Franz.