[OpenJDK 2D-Dev] 2 small xrender fixes

Clemens Eisserer linuxhippy at gmail.com
Wed Jun 23 14:50:36 UTC 2010

Hi Phil,

> I thought I sent this email last week but I just found a very patient
> thunderbird compose window :-(
No problem, thanks for the reply

Until next week I am (again) really busy, after that I have plenty of
time to investigate the issue :)

I also have another fix which resolves some problems (disables an
optimization) with transformed-blits on intel-GPUs, and it would be
great to have all those fixes done in one go.
Is it ok to reduce transformation-precision if Graphics has

Thanks, Clemens

> For the record, anyone with an openjdk user id should be able to
> upload webrevs to cr.openjdk.java.net.
> As to the fixes : The Jules one looks fine.
> I can create you a bug ID.
> The whitespace one, I'm not aware that T2K will do this.
> I find that it sets it width and height to zero for whitespace.
> I instrumented the rasteriser and ran  with code like this.
>   public void paint(Graphics g) {
>       Graphics2D g2d = (Graphics2D)g;
>       g2d.setRenderingHint(
>                RenderingHints.KEY_TEXT_ANTIALIASING,
>                RenderingHints.VALUE_TEXT_ANTIALIAS_ON);
>       Font f = new Font("Dialog", Font.PLAIN, 20);
>       g.setFont(f);
>       g.drawString("A C", 10, 20);
> Are you running the published JDK7 build 97 ?
> What is the exact platform and architecture?
> Can you send me the test you are using?
> Perhaps its even something in the exact font being used.
> Can you send a pointer to that?
> -phil.
> Andrew John Hughes wrote:
>> On 15 June 2010 14:40, Clemens Eisserer <linuxhippy at gmail.com> wrote:
>>> Hi,
>>> I would like to get two (very small) changes to the xrender-pipeline in:
>>> 1. deals with the
>>> fact that the closed-source
>>> font rasterizer seems to generate width=65535 for whitespaces, which
>>> resuts in software fallbacks.
>>> Granted with height=0 rendering should be completly omited by the X
>>> server, but as it seems it isnt.
>>> 2.
>>> synchronize-native-cairo-ops synchronizes access to Jules's native
>>> functions.
>>> There seems to be a race somewhere in cairo and I wasn't able to hunt
>>> it down, so synchronizing is the best I can do for now.
>>> Do I need to create two different bug-reports? Is OpenJDK Bugzilla
>>> enough?
>> You need a Sun bug ID (i.e. bugs.sun.com) for each, but it's quicker
>> if an Oracle developer creates one.  Just ask for one when the patch
>> is approved.
>> It would be easier to read these if there were posted as patches or
>> webrevs, not zips.
>> Thanks,

More information about the 2d-dev mailing list