Re: Small test case fixes / observations (php-5.2.6_rc4)

On 2008-04-07 11:22, Josie Messa wrote:

On Sun, Apr 6, 2008 at 10:44 PM, Christian Hoffmann <hoffie@gentoo.org> wrote:

while building and testing a minimal version of PHP and as such disabling the session extension, I discovered that ext/standard/tests/general_functions/bug44394_2.phpt fails without the session extension being enabled. I created a patch to fix that [1]. On a sidenote, some tests seem to break on amd64 systems: ext/standard/tests/array/array_slice_variation6.phpt: Apparently there is a real bug (or difference in behavior between 32bit and 64bit) in the code here, and the test revealed it. See [2] ext/standard/tests/array/array_slice_variation2.phpt ext/standard/tests/array/array_slice_variation3.phpt: Iteration 7 is different on 32bit and 64bit [3][4]. I don't know if this behavior is intended, for now I simply added a skipif for non-32bit [5][6].

Ah, I wrote these tests on a 32 bit machine so I didn't pick up any changes in behaviour. I don't think a SKIPIF section should be added as I wouldn't expect there to be a difference here at all, and adding a SKIPIF section means that the problem will get ignored, I'd raise a bug instead. Ok, thanks for looking into it. I filed bug 44664 [1] and would be happy if the session fix could be committed (as I still think it is correct to skip the test here).

[1] http://bugs.php.net/bug.php?id=44664 -- Christian Hoffmann

Attachment: [application/pgp-signature] OpenPGP digital signature signature.asc