Sorry, I've seen these tests failing for a long time but I just didn't
want to bother you guys.. :)

But I'll make it my sole purpose in life to make sure you know about
these from now on. Your wish is my law.. ;)


On Thu, 2008-02-07 at 23:29 +0100, Ulf Wendel wrote:
> Nuno Lopes schrieb:
> > The same for the gcov machine: 
> >
> > Nuno
> Thanks for the hint!
> Every problem report is welcome. It would not be honest to promise 
> immediate fixes nor to promise that my spontanous reaction will always 
> be "great hint" and not a "oh, no - a bug, how boring, go away", but at 
> the end of the day its nothing but "THANKS!".
> Andrey and I had to focus on other issues during the last weeks 
> (Connector C++ and PDO testing). Looks like we can spend a few days on 
> ext/mysql, ext/mysqli and mysqlnd right now. Strike!
> Another good news is that, from what I know, Jim Winstead will work on 
> libmysql in the near future. Why is this good news? Well, not only is 
> Jim a fantastic developer but libmysql issues will no longer bounce back 
> for fixing to Andrey, who has been a short resource in the past.
> Looks like we found out today why the following tests fail with mysqlnd 
> (and fixed them internally):
> mysqli_stmt_affected_rows.phpt
> mysqli_debug_mysqlnd_control_string.phpt
> mysqli_debug_control_string.phpt
> mysqli_debug_append.phpt
> mysqli_connect_oo_defaults.phpt
> So, what's my feeling about how critical the remaining test failures 
> are? Looks like 50% of the test failures are caused by wrong (not 
> portable) tests (my fault). Andrey has fixed the only severe mysqlnd bug 
> I did know (since months) yesterday (pconnect causing a crash). All 
> other issues I know about are very minor issues.
> We have committed tests even if they fail and we won't fix them for 
> weeks/months. Is that OK? Some QA people prefer having the tests in the 
> bug system only until the issue gets fixed in order to not get 
> distracted by tests which are known to fail.
> Once again: please, please continue poking us (using mail, blogs, 
> whatever you want!) as you did in the past and do now. If we don't know 
> what's wrong, we can't listen and try to help.
> Ulf
> PS: We have increased ext/mysql and ext/mysqli tests loc by 500% during 
> the last year - no surprise we see more tests failing. For example, BIT 
> never worked fine. Now one test "documents" that BIT doesn't work 
> perfect. Has our quality slipped? The test helped us to improve on small 
> BIT values already but its still failing and I'm against hiding the test 
> failures.

PHP CVS Mailing List (
To unsubscribe, visit:

Reply via email to