Re: Re: [GSOC] phpruntests - GSoC report 1

Hi Georg, Here is some feedback about phpruntests actual version. All tests were done using PHP 5.3.0RC2 on Linux. - Error message about 'TEST_PHP_EXECUTABLE' has typos errors : * "the -p otion" * trailing \n missing * double space : "the environmental" - Executing the script leads to an error on bash : Command line : /path/php-5.3.0RC2/sapi/cli/php run-tests.php -p /path/php-5.3.0RC2/sapi/cli/php /path/php-5.3.0RC2/Zend/tests/ Output: sh: -v: command not found This version of run-tests requires PHP 5.3 or higher. The PHP executable that you have supplied using either the command line option -p or the environmental variable TEST_PHP_EXECUTABLE does not fulfill this requirement. I think the error is in configuration/preconditions/rtIsPhpVersionCorrect.php with the exec. In my machine, var_dump($config->getSetting('PhpExecutable')) returned NULL. Using the following code line seems to work: $exec = escapeshellcmd($config->getSetting('rtPhpCgiExecutableSetting').' -v'); - Executing the following line leads to an error: /path/php-5.3.0RC2/sapi/cli/php run-tests.php /path/php-5.3.0RC2/Zend/tests/bug36513.phpt The test file %1 is not a valid phpt test file. If I made a mistake on syntax, the error message is not very helpful ... .phpt file exists: -rwxr-xr-x 1 1005 1005 475 2008-05-24 17:55 /path/php-5.3.0RC2/Zend/tests/bug36513.phpt I would be happy to help you with more feedback asap, ie when I would be able to execute the script ;) Olivier P.S. I removed php-qa mailing list, as I think this feedback is strictly related to this gsoc project. Correct me if I'm wrong. 2009/5/12 Georg Gradwohl <georg.gradwohl@gmail.com> > > Hi Olivier, > > > So you're working on this project : > http://wiki.php.net/gsoc/2009/runtests ? > > > > Is it based on the phpruntests Zoe talked about a few weeks ago ? > (available > > here : http://cvs.php.net/viewvc.cgi/phpruntests/src/) > > yes exactly. > > > If so, is there something planned for the last abstract point : <<Another > > focus is to offer a rich variant of reporting-features and the > possibility to > > store and compare results. >> > > Any drafts or "scope statement" made for this tool so far ? If not, I > would be > > glad to offer my help on this topic. > > We don't have a concrete concept for this point so far. At the moment we're > focusing on the basic implementation (means that all existing testcases can > be executed correctly) and the multithreading which will improve the > performance. The report-feature is going to be implemented after these two > main-steps. But I'll be happy to get back on your offer at the right time. > > Zoe, Stefan, please correct me if I said something wrong. ;) > > g2 > > >