Re: [tree-ssa] SSA->normal shuffle

[Andrew MacLeod <amacleod at redhat dot com>]

On Fri, 2003-11-21 at 16:12, law@redhat.com wrote:
> In message <1069425614.11003.893.camel@p4>, Andrew MacLeod writes:
>  >On Fri, 2003-11-21 at 09:35, Andrew MacLeod wrote:
>  >
>  >> Anwyay, here's the patch. It bootstrapped on x86 and caused no
>  >> differences in testruns. Im checking it in momentarily.
>  >
>  >btw, in case no one has noticed, a make check-gcc on a stage 1 compiler
>  >has one less failure than a make check on a boostrapped stage3 compiler.
>  >The bootstrapped compiler fails:
>  >
>  >< FAIL: gcc.dg/20020425-1.c (test for excess errors)
>  >---
>  >> PASS: gcc.dg/20020425-1.c (test for excess errors)
>  >
>  >with a segmentation fault I beleive. 
>  >
>  >I dont know how long this has been going on, but I think its been a very
>  >long time. I seem to vaguely recall tripping over it in passing a few
>  >months ago, but promptly forgot about it.
> Well, this one is dependent on things like how much memory/swap space you
> have.  So I doubt it's directly related to your change:
> 

Oh, I assumed it wasn't. It does the same thing without my change. Just
thought I would mention it :-)

> /* PR c/2161: parser stack overflow.  */
> /* { dg-do compile } */
>  
> #define ONE     else if (0) { }
> #define TEN     ONE ONE ONE ONE ONE ONE ONE ONE ONE ONE
> #define HUN     TEN TEN TEN TEN TEN TEN TEN TEN TEN TEN
> #define THOU    HUN HUN HUN HUN HUN HUN HUN HUN HUN HUN
>  
> void foo()
> {
>   if (0) { }
>   /* 11,000 else if's.  */
>   THOU THOU THOU THOU THOU THOU THOU THOU THOU THOU THOU
> }
> 
yuck.

Its probably because our bootstrapped compiler is using more stack space
or something on each recursive function call. Or something like that :-)

Andrew