[webkit-qt] qt5 version updated for bot / development
oszi at inf.u-szeged.hu
Wed May 2 02:41:26 PDT 2012
The latest Qt5 update for WebKit trunk buildbots was a month ago.
I tried to do an update 1.5 weeks ago, but unfortunately a159ca80d73f07d617d6edd9b3e59bda25449a1c
in qtbase made almost all layout tests fail with Qt5-WK1 again. (Qt5-WK2 and Qt4-WK1 works same as
with the latest pinned Qt5 hash.)
Pierre started debugging this problem, but I don't think if we should do the update
on the bots before it is fixed, because update would mean 0% test coverage on Qt5-WK1.
Osztrogonac Csaba írta:
> Hi All,
> I updated the Qt5 on the WK1/WK2 buildbots, WK2 EWS and on the perf bots.
> The new Qt5 hash is 51ca55ff29f53b909c3614b56609e067d2066124 (Thu Apr 5
> 17:08:12 2012 +1000)
> You can find the updated build script here:
> Osztrogonac Csaba írta:
>> I updated the Qt5 on the Qt5 WK1/WK2 buildbots, WK2 EWS and on the WK2
>> perf bot.
>> The new Qt5 hash is 74f7b7f74c0e9ce99123fc836fd22f1f0b5688fe (Thu Mar
>> 29 16:17:33 2012 +0200)
>> But unfortunately there are two regressions:
>> - [Qt] WK1/WK2 layout test results are diverging -
>> After update WK1 tests started to fail, so now the test coverage on
>> Qt5-WK1 is 0% :(
> Many thanks for Pierre for fixing it. (
> http://trac.webkit.org/changeset/113321 )
> And you _need_ to update WebKit testfonts, because it was updated due to
> this bug too.
> Without updating testfonts you won't be able run layout tests with Qt5-WK1.
>> - [Qt][WK2] REGRESSION, fix a failing API test in qmltests -
>> After update a Qt5-WK2 API test started to fail
> This bug is still valid, one API test is still failing. Any volunteer to
> fix it?
> PASS :
> FAIL! : qmltests::WebViewPreferences::test_standardFontFamilyChanged()
> Compared values are not the same
> Actual (): 'Bitstream Vera Serif'
> Expected (): Bitstream Vera Serif
More information about the webkit-qt