[rfc][icedtea-web] Fix PR854: Resizing an applet several times causes 100% CPU load

[Adam Domurad]

On 05/15/2013 05:59 AM, Jiri Vanek wrote:
> On 05/14/2013 10:49 PM, Adam Domurad wrote:
>> Nasty bug with a simple fix. This has been reported since icedtea-web 1.1
>>
>> See http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=854. The
>> reproducer is very easy to set
>> up. Placing a print statement reveals that after two resize's, all of
>> icedtea-web's worker threads
>> are busy.
>>
>> The waiting before was too fragile it seems. The version in the patch
>> shouldn't dead-lock in any
>> code-path. Everything works as expected with this patch (there really
>> is no possible harm, too).
>>
>> This should go into any branches that will be released again, IMHO.
>> From the bug comments it seems
>> this is quite a visible issue.
>>
>> Thanks,
>> -Adam
>
> This is nice! Please accompany this with automated reproducer and go on.
>
> ok for head and 1.4
>
> J.

Here's a reproducer. Without patch, fails to resize more than once (gets 
stalled).

ChangeLog:
2013-05-16  Adam Domurad  <adomurad at redhat.com>

         Reproducer for PR854.
         * 
tests/reproducers/simple/ResizeApplet/resources/ResizeApplet.html:
         Resizes applet from Javascript.
         * tests/reproducers/simple/ResizeApplet/srcs/ResizeApplet.java:
         Simple applet with a few helper methods.
         * 
tests/reproducers/simple/ResizeApplet/testcases/ResizeAppletTests.java:
         Test applet resizing.

Cheers.
-Adam
-------------- next part --------------
A non-text attachment was scrubbed...
Name: resize-applet-reproducer.patch
Type: text/x-patch
Size: 8520 bytes
Desc: not available
Url : http://mail.openjdk.java.net/pipermail/distro-pkg-dev/attachments/20130516/82889651/resize-applet-reproducer.patch