Re: [3.3/3.4 PATCH] fix broken test g++.dg/template/spec10.C

[Mike Stump <mrs at apple dot com>]

On Monday, October 6, 2003, at 02:43 PM, Bob Wilson wrote:
> This test is broken on both the mainline and the 3.3 branch because the
> expected exit value is 0x3000 instead of zero.  I don't know why this 
> doesn't
> show up for all the targets, but I noticed it when testing the 
> xtensa-elf
> target with a simulator.  I guess the test harness doesn't always pay
> attention to the exit value.

Please don't say that, this isn't true.

>   I also fixed up the indentation.
>
> OK for mainline?
> Is this appropriate for the branch, and if so, will someone approve it?
>
> 2003-10-06  Bob Wilson  <bob.wilson@acm.org>
>
>         * g++.dg/template/spec10.C: Set exit value to zero on success. 
>  Fix
>         indentation.

This OK to me.  I checked it out and the change doesn't destroy 
anything that was being tested for in the testcase nor anything from 
the PR, and is `more' correct.

The simulator (or deja, if done in there) should do status&255 on the 
return value from main before passing back to the environment if they 
want to match UNIX more closely.  This is ultimately why you saw the 
problem and others don't.

I think it is trivially safe for 3.3 as well.