And once more:
http://cr.openjdk.java.net/~yan/8170979/webrev.04
Added:
a port of JDK-8138872;
a change in javaclient/shared/screenshots/GoldenImageManager.java
to allow it find old golden images.
In the index.html there's a list of 10 files to remove after the patch
applied. Webrev script includes only 4 of them in the patch.
I tried to commit the change, jcheck has nothing against it.
We use this state of the tests (plus a closed part, and
our own svn storage for images) to run in ATR.
If not to analyze.
Thank you!
-yan
On 12/29/2016 04:38 PM, Yuri Nesterenko wrote:
And again:
http://cr.openjdk.java.net/~yan/8170979/webrev.03
(Discussion in the bug)
Unfortunately, webrev doesn't process properly
files deleted by 'hg rm' .
-yan
On 12/22/2016 12:07 PM, Yuri Nesterenko wrote:
A new version is here:
http://cr.openjdk.java.net/~yan/8170979/webrev.01
(trailing spaces removed).
See also discussion in the bug report:
https://bugs.openjdk.java.net/browse/JDK-8170979
thank you, Kevin!
-yan
On 12/21/2016 12:19 PM, Yuri Nesterenko wrote:
A gentle reminder.
Without this change, we have to make All Test Run out of
the makeshift workspace.
Thanks,
-yan
On 12/09/2016 02:02 PM, Yuri Nesterenko wrote:
Colleagues,
please review this test bug fix:
http://cr.openjdk.java.net/~yan/8170979/webrev.00/
Bug: https://bugs.openjdk.java.net/browse/JDK-8170979
The changes are mostly in tests/functional/.
This is a merge of current testbase in
http://hg.openjdk.java.net/openjfx/9-dev/tests and whatever
we have internally. The tests should now compile and work
with Jigsaw (including b148, --add-opens), recognize already
not too new JDK directory structure, do not use Builders etc.
Small number of tests was excluded and even dropped for various
reasons.
Several suites were tested with b142-b146 and PIT build for
the recent Jigsaw update (re: --add-opens).
There are still some reservations and unanswered questions.
For instance, a hard monotonous work of creation and maintaining
of the so called golden images is up to a tester: I don't know a
good place to share them between independent teams.
A couple of the suites still won't start without a minor
reconfiguration beforehand, not always properly documented.
We intend to use this state of the tests for the next massive test run
starting from jdk9 b150.
