RE: On recent test failures

Hi, Sean et al.:

CTIC says OK to rolling back to a previous version with an older
well-known CSS parser. Abel and Miguel will work on the CSS thing in
order to deserve being included in the final version these days (right
after Abel recovers from the 2nd computer crash in a month :-( ).

Regards,
  
Nacho


-----Mensaje original-----
De: public-mobileok-checker-request@w3.org
[mailto:public-mobileok-checker-request@w3.org] En nombre de Sean Owen
Enviado el: domingo, 23 de marzo de 2008 22:10
Para: public-mobileok-checker
Asunto: On recent test failures


OK, I checked in several small fixes for some small corner cases of
the test harness that I noticed generated failures.

I notice there are still a fair number of test failures due to the new
CSS parser -- shall we roll this back then? at this point I don't mind
sticking with a version that works as we expect rather than stick to
the latest and greatest.


I note that once that is fixed, it may turn up some other spurious
test failures due to my recent changes, which make the order of
certain elements in <moki>, like inline style blocks, more consistent,
and possibly different from the current ordering.


I've also checked in changes that load resources in parallel.
Curiously it makes Java 5 on OS X hang -- so stuck that it has to be
"kill -9"ed which makes me believe it is a Java bug on OS X. It
doesn't happen on Java 6. I'd be curious to hear if anyone else sees
this. I hope not.


I'm pushing to wrap up open issues assigned to me soon since I'll
really truly have to wrap up work on this from my end in the next
month or so. See my open note on the <object> issue. Join me in
pushing through the set of known issues, either fixing or closing
them, so we can soon be ready to release a final version.


Sean

Received on Monday, 24 March 2008 09:19:25 UTC