Re: PATCH: [3.3 branch] IA64 bootstrap failure

[Jakub Jelinek <jakub at redhat dot com>]

On Tue, Jul 08, 2003 at 11:17:37PM -0700, H. J. Lu wrote:
> What do you mean by "bug"? As I said, the file is NOT miscompiled. It is
> just different. Since compiler sources are different, comparison of
> generated .o files don't mean much here. We have several choices:
> 
> 1. Check each comparison failure to see if it is really miscompiled. Or
> 2. Add stage4 and compare against stage3, instead of stage2. Or
> 3. Make sure stage1 and stage2 have the same source files.

Quoting Andreas Schwab on this:
"Running the same (stage2) compiler command on the command line produces
output that is the same as the output from the stage1 compiler.  So it
indeed looks like it's dependent on the environment size."

This certainly does not seem like a problem between stage1 and stage2
compilers, but that stage2 compiler on the same input and same options
generates different output depending on some environment it is run on,
phase of the moon or whatever.
Compiler output should be IMHO always just function of the
source and command line arguments, maybe some explicitely
listed environment variables, but that's it.

	Jakub