This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] Problem in right-assocify
- From: Andreas Jaeger <aj at suse dot de>
- To: gcc-patches at gcc dot gnu dot org
- Date: Fri, 29 Nov 2002 10:36:11 +0100
- Subject: Re: [tree-ssa] Problem in right-assocify
- References: <u8ptsoiwdt.fsf@gromit.moeb>
Andreas Jaeger <aj@suse.de> writes:
> Looking at this warning:
> /cvs/gcc-tree-ssa-20020619-branch/gcc/tree-simple.c:880: warning: unused variable `rhs'
>
> I noticed that either the comment is wrong and this variable is really
> unused - or there's a typo. Could somebody please check whether the
> patch below is the right think to do - and then approve it or tell me
> what to do?
The patch is definitly wrong - bootstrap fails with it. But the
question remains whether the comment is right or should be updated and
why rhs is unused,
Andreas
>
> The problem that caught me is that the comment speaks about rhs chain
> but we go through the lhs:
> /* Walk through the rhs chain from there until we find something
> with a different code. In this case, c. */
> for (q = &TREE_OPERAND (lhs, 1); TREE_CODE (*q) == code;
>
> Andreas
>
> 2002-11-29 Andreas Jaeger <aj@suse.de>
>
> * tree-simple.c (right_assocify_expr): Really walk rhs chain.
>
> ============================================================
> Index: gcc/tree-simple.c
> --- gcc/tree-simple.c 27 Nov 2002 00:43:33 -0000 1.1.4.19
> +++ gcc/tree-simple.c 29 Nov 2002 08:53:51 -0000
> @@ -898,7 +898,7 @@ right_assocify_expr (top)
>
> /* Walk through the rhs chain from there until we find something
> with a different code. In this case, c. */
> - for (q = &TREE_OPERAND (lhs, 1); TREE_CODE (*q) == code;
> + for (q = &TREE_OPERAND (rhs, 1); TREE_CODE (*q) == code;
> q = &TREE_OPERAND (*q, 1))
> TREE_TYPE (*q) = TREE_TYPE (cur);
Andreas
--
Andreas Jaeger
SuSE Labs aj@suse.de
private aj@arthur.inka.de
http://www.suse.de/~aj