g++.old-deja/g++.mike/p7325.C - suspected bogus test case

[Mike Stump]

On Saturday, June 26, 2004, at 11:09  AM, Zack Weinberg wrote:
>> I believe it used to work and pass, and someone broke it and didn't
>> fix it.
>
> It has been XFAILed since the creation of the EGCS CVS repository,
> except for a short window in late 1997-early 1998.

I believe it used to work and pass and then someone broke it, honestly.

>> You want to get rid of a test case that found a regression instead of
>> enticing someone to fix the compiler?
>
> No, I want to get rid of a test case that fails to test anything
> meaningful.

It tests something meaningful.  I am sorry you don't value a compiler 
that can reuse stack space intelligently.  If you did, you'd just fix 
the bug and we'd move on.  I'm sorry you don't value stack space enough 
to want to remove just about the _only_ testcase for this `feature'.

> Let me reiterate that: it does not test anything meaningful.

You can repeat that a hundred times, this will not make it true.

> It currently passes or fails depending on details of the platform ABI.

I understand the test case isn't as reliable as you'd like.  I'd 
welcome a modification that makes it fail more often, if you'd like.

> Worse, the compiler is within its rights to extend the lifetimes of
> the temporaries because it notices that their addresses have been
> taken.

No, you are confused.  This too would be a bug that would need fixing.  
You are thinking of a different test case under different circumstances 
and arguing about something unrelated and completely different to try 
and support your point.  I reject that.

> The only way to test this reliably

Is this test case not 100% reliable for some target triplets?