Re: PING *objc, NeXT runtime, testsuite changes*

[IainS <developer at sandoe-acoustics dot co dot uk>]

On 16 Feb 2010, at 17:33, Dominique Dhumieres wrote:

> > 1/ you must be sure to add the additional shared headers directory.
> Yes, I have done it.
Hm. I guess something's out of sync - I can't repeat your result.

refreshed patches vs. current trunk (assuming that you've applied  
version b of the PR43061 fix separately).

Attachment: objc-c++-shared.tar
Description: Unix tar archive

Attachment: target-supports-vs156812-diff.txt
Description: Text document

Attachment: objc-torture-pt1-vs156812-diff.txt
Description: Text document

Attachment: objc-torture-pt2-vs156812-diff.txt
Description: Text document

Attachment: objc-dg-vs156812-diff.txt
Description: Text document

Attachment: obj-c++-dg-vs156812-diff.txt
Description: Text document

the headline results I get are:

                === objc Summary ===

# of expected passes            6145
# of unexpected failures        31
# of expected failures          43
# of unresolved testcases       4
# of unsupported tests          52

                === obj-c++ Summary ===

# of expected passes            1753
# of unexpected failures        30
# of unexpected successes       11
# of expected failures          63
# of unresolved testcases       8
# of unsupported tests          16


> > 2/ there will be some reduction in the number of tests run in the
> > torture section - ...
>
> Should not they go to the count of "unsupported tests"?

that's not the way the old style torture tests work - if the trivial  
test fails they simply vanish.
I agree it's a cause for potential worry.

I guess they could all be re-cast as dg-torture - it's tempting.

Iain