Re: svn: /php/php-src/ branches/PHP_5_4/ext/standard/tests/streams/bug49936.phpt trunk/ext/standard/tests/streams/bug49936.phpt

On 11/13/2011 10:05 AM, Pierre Joye wrote: > On Sun, Nov 13, 2011 at 4:24 PM, Rasmus Lerdorf <rasmus@php.net> wrote: >> On 11/13/2011 06:11 AM, Ferenc Kovacs wrote: >>> Hi Rasmus, >>> >>> this change ( >>> >>> http://ci.qa.php.net/job/php-src-5.4-matrix-tests/architecture=x86-64,os=linux-debian-6.0/109/changes) >>> broke the build on ci.qa.php.net: >>> >>> http://ci.qa.php.net/job/php-src-5.4-matrix-tests/architecture=x86-64,os=linux-debian-6.0/109/testReport/ext_standard_tests_streams_bug49936/phpt/Bug__49936__crash_with_ftp_stream_in_php_stream_context_get_option___/ >>> >>> as you can see, it was passing until that point >>> >>> http://ci.qa.php.net/job/php-src-5.4-matrix-tests/architecture=x86-64,os=linux-debian-6.0/109/testReport/ext_standard_tests_streams_bug49936/phpt/Bug__49936__crash_with_ftp_stream_in_php_stream_context_get_option___/history/? >>> >>> I'm also seeing that result in my local test environment (debian lenny), >>> could you take a second look? >> >> I guess older systems have different error messages then. We may just >> need to delete the test. If you run it on a more recent Linux you will >> see that it passes now. > > I would not delete the tests but make it message independent, as the > message is important the key part is to test this condition. Yeah, but there is also a different number of messages, so it isn't just sticking a %s in the message string part. We are going to end up with an expected output regex that basically accepts anything the way this test is written. -Rasmus