Thu 06 Aug 2009 07:31:30 AM UTC, comment #12:
Can you open new bugs for specific problems? The bug in the original submission was fixed.
|
Wed 05 Aug 2009 09:42:53 PM UTC, comment #11:
Makes me laugh that we ever considered closing this bug report. I am re-opening it as a submission post for specific TextField bugs. Here's one:
<l1m5> http://megavideo.com/?v=047TK8ND
<l1m5> open the video, click play
<l1m5> then highlight the icons on the bottom
<l1m5> until you select switch video to 4:3 nothing is centered, after you mouse over switch video to 4:3 the "4:3" doesn't go away but everything is then centered
|
Tue 28 Jul 2009 09:52:21 AM UTC, comment #10:
Now I've fixed the latest TextField bug again (it was an unrelated breakage caused by commit 11318), the testcases for this bug seem to be restored.
The BBC SWF works as well as it did before (dragging the words to the blank has never worked properly), so this bug is closed. Good work!
|
Tue 28 Jul 2009 07:59:23 AM UTC, comment #9:
I suspect the textfield width bug might be quite recent, and is the same bug that causes the testsuite failures.
|
Tue 28 Jul 2009 07:45:01 AM UTC, comment #8:
I confirm everything is fixed except the BBC SWF, where it seems the width of the textfield bounds is wrong.
|
Tue 21 Jul 2009 06:34:43 PM UTC, comment #7:
correction
sorry, "BLANK" space in tests 3 and 6 of BBC test swf is still wrong. Not sure if this is related to TextField, but you can be sure an investigation is ongoing!
|
Tue 21 Jul 2009 06:30:38 PM UTC, comment #6:
DeviceFontTest and EmbeddedFontTest are now, I think, back to behaving in a manner that is pre-TextField fix. That is, they are restored to their original behavior.
Also, the BBC test swf bwy submitted as bug example is working correctly.
|
Thu 16 Jul 2009 10:51:43 AM UTC, comment #5:
Not a security issue at the moment, as the memory corruption is fixed.
|
Thu 16 Jul 2009 06:07:00 AM UTC, comment #4:
The BBC SWF looks better again with revno 11627, except for one screen, which is probably the same bug that's making all the textfields empty in those two new tests.
|
Wed 15 Jul 2009 05:21:13 PM UTC, comment #3:
Er, this was in progress. Simultaneous editing isn't quite perfect.
|
Wed 15 Jul 2009 05:17:37 PM UTC, comment #2:
Yes, segfaults are more important to fix.
There are two new tests in misc-ming.all (DeviceFontTest and EmbeddedFontTest) that might help with some implementation. Getting the tested actionscript values right is important, but unfortunately it doesn't test that the text is actually there. The results are the same even though most of the text has disappeared. So currently the only way of testing is to run them and look at them.
Any new ideas on how to test what text is there (e.g. with pixel checking) would be useful.
|
Wed 15 Jul 2009 05:00:50 PM UTC, comment #1:
I noticed the incorrect text placement after fixing the segfaults and failed test cases. I felt submitting fixes to segfaults in a hurry was more important.
I have since fixed much of the corrupted functionality, and submitted to trunk.
Not sure about DeviceFontTest, or why it got worse when I tried to improve TextField. I will look at that.
|
Wed 15 Jul 2009 02:42:11 PM UTC, original submission:
The text on the SWF is in the wrong position after the TextField changes. The text also scrolls unpredictably sometimes, when in fact all the TextFields should be static. I don't know if it's related to the memory errors that valgrind shows, but it is certainly worth fixing them first, otherwise it's hard to tell what is a program logic error and what's caused by memory corruption.
gtk-gnash -j 416 -k 234 -U 'http://news.bbc.co.uk/2/hi/uk_news/magazine/8151641.stm' -P 'flashvars=path=http://news.bbc.co.uk/nol/shared/spl/hi/quiz/09/07/14/122120/quiz.xml' -P 'height=234' -P 'id=undefined_emb' -P 'name=undefined' -P 'src=http://news.bbc.co.uk/nol/shared/spl/hi/quiz/player/quizplayer.swf' -P 'type=application/x-shockwave-flash' -P 'width=416' 'http://news.bbc.co.uk/nol/shared/spl/hi/quiz/player/quizplayer.swf'
The commit that changed this behaviour was 11255 (segfaults) or 11256. It looks right in 11254.
But also see misc-ming.all/DeviceFontTest.swf, which seems to be more broken after revision 11254 (it wasn't exactly perfect before).
|