flex-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Frédéric THOMAS <webdoubl...@hotmail.com>
Subject Re: Apache Flex 4.9.1
Date Sun, 17 Feb 2013 16:44:51 GMT
Hi, sorry for the delay, I rerun the all test suite with your fix and it 

For the 2 failed tests relative to my fix, I had to regenerate the images, 
not sure I can to include that in the release though as a vote is started.

For the components/Charts/Bubble/Properties/BubbleProperties 
BubbleSeries_multipleAxes_disabledDays, I tried to play with 
maxColorVariance and numColorVariances without success.


-----Message d'origine----- 
From: Alex Harui
Sent: Saturday, February 16, 2013 4:30 PM
To: dev@flex.apache.org ; Frédéric THOMAS
Subject: Re: Apache Flex 4.9.1

Are the difference significant or just a few pixels? Are you running on mac 
or win?

Sent via the PANTECH Discover, an AT&T 4G LTE smartphone.

Frédéric THOMAS <webdoublefx@hotmail.com> wrote:

Good to hear.
Since your fix on LangPacks, the text is now rendered but there are
differences in bitmap compare, execpt for the 2
spark/styles/local/dropShadowVisible failed tests, it should then be a
windows/mac rendrering difference only.

-----Message d'origine-----
From: Alex Harui
Sent: Saturday, February 16, 2013 7:41 AM
To: dev@flex.apache.org
Subject: Re: Apache Flex 4.9.1

On 2/15/13 7:29 AM, "Frédéric THOMAS" <webdoublefx@hotmail.com> wrote:

> Hi Alex,
> For the 2 last fail tests[1], it comes from the PopUpManagerImpl.as
> changes
> I did, the tests are broken but I still thing the fix is good, if I'm not
> wrong, the fact is the view bounds taken in account are out of what is
> visible, so, the image compare fails because it takes in account the +1
> width and height pixels of the new popup blocker size.
If it only affects Mustella and not the appearance to the customer, then cut
new baseline pngs and check everything in.
> It's not obvious to debug mustella, if needed I'll do it but if you can
> simply update the PopUpManagerImpl.as to include my fix (line 767:
> g.drawRect(screen.x-1, screen.y-1, screen.width+2, screen.height+2);),
> build
> the framework, run mini_run.sh tests/spark/styles/local/ and do an
> ImageDiffAir on the 2 bad images, you might be able to determinate in a
> more
> intuitive way than me if we need to simply re-generate those 2 baselines
> or
> try to fix the UnitTester/CompareBitmap classes it takes in account what
> is
> visible only.
BTW, all tests passed for me, even LangPacks, so unfortunately, you'll have
to debug the tests to find out why you are getting failures.  What kinds of
failures did you get?  Just bitmap failures?  If so, what did they look

Alex Harui
Flex SDK Team
Adobe Systems, Inc.

View raw message