Re: PATCH: Reguire __cxa_atexit for ext/{new, throw}_allocator/deallocate_global.cc

[Rainer Orth <ro at CeBiTec dot Uni-Bielefeld dot DE>]

Dave Korn <dave.korn.cygwin@googlemail.com> writes:

>   Shouldn't this be xfailed on solaris, rather than skipped on all
> non-cxa_atexit targets?

no, this way lies madness ;-(  I'm strongly opposed to added any
target-based xfails or skips if there's a general feature that can be
tested for.  If it's not __cxa_atexit, suggest another name for the
feature test, and we can check for cygwin there in some way.  A
prominent example (which I'll attack soon) is the large list of
different target lists used to decide if the target has pthread
support.  This is both inconsistent and a nightmare to maintain, e.g. if
adding a new target.

>   The FAIL in this case is indicating an ABI violation w.r.t order of dtors,
> which could equally happen on non-cxa_atexit systems that implement some
> variant scheme (such as e.g. get_exit_frame_monitor hooks!) and it seems
> useful to me if the testsuite flags that up, regardless of the internal
> implementation details of how the system handles (or, in this case, fails to
> handle correctly) finalisation.

Exactly my suggestion: just give an appropriate name and add your target
(or a runtime test) to the feature test.

	Rainer

-- 
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University