[cfe-commits] r73286 - in /cfe/trunk/test: ./ CXX/ CXX/basic/ CXX/basic/basic.align/ CXX/basic/basic.def.odr/ CXX/basic/basic.def/ CXX/basic/basic.life/ CXX/basic/basic.link/ CXX/basic/basic.lookup/ CXX/basic/basic.lookup/basic.lookup.argdep/ CXX

[Daniel Dunbar]

My theory was that we probably don't want to move everything into this
structure. I think it works well for small unit tests which directly
target some part of the standard, but we have many tests that wouldn't
fit well into this structure.

Would it be bad to maintain both styles, using the CXX one where it
makes sense (w/o Lex/Sema/etc. distinction) and using the existing
structure when it doesn't?

 - Daniel

On Fri, Jun 12, 2009 at 11:26 PM, Douglas Gregor<dgregor at apple.com> wrote:
>
> On Jun 12, 2009, at 11:19 PM, Sebastian Redl wrote:
>
>>
>> On Sat, 13 Jun 2009 06:06:57 -0000, Douglas Gregor <dgregor at apple.com>
>> wrote:
>>> Author: dgregor
>>> Date: Sat Jun 13 01:06:53 2009
>>> New Revision: 73286
>>>
>>> URL: http://llvm.org/viewvc/llvm-project?rev=73286&view=rev
>>> Log:
>>> Add the structure for a C++ test suite that mimics the structure of
>>> the
>> C++
>>> standard
>>
>> How do we distinguish between Lex/Parse/Sema/CG tests within this
>> structure? Do we create subdirectories or include it in the file
>> name, e.g.
>
>
> Definitely not subdirectories; we want an exact structural match
> between the directories and the standard, so that we can hook up the
> results of crawling the test directories to the C++ status table. (For
> example, to show which paragraphs we think we've covered with our
> tests).
>
> Personally, I don't see any point in distinguishing Lex/Parse/Sema/CG,
> although we've been doing it for a while. Still, we can tell what kind
> of test we have based on the RUN line anyway.
>
>        - Doug
> _______________________________________________
> cfe-commits mailing list
> cfe-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>