[rfc] [icedtea-web] Fix EDT hanging on OpenJDK 11

[Laurent Bourgès]

Hi Martin,

Jiri pushed this patch yesterday.

Le mer. 3 oct. 2018 à 10:31, Martin Balao <mbalao at redhat.com> a écrit :

> Hi Laurent,
>
> On Thu, Sep 27, 2018 at 1:19 PM, Laurent Bourgès <
> bourges.laurent at gmail.com> wrote:
>
>> Mario,
>>
>> >     1. Why is EDT_DAEMON_THREAD_POOL needed?
>>> >
>>> >
>>> > This is critical on OpenJDK 11 as SequencedEvent processing is buggy
>>> > (still not fixed).
>>> > - This specific Thread pool uses a specific factory to create threads
>>> in
>>> > main thread group (main thread).
>>> >
>>> >     public Thread newThread(Runnable r) {
>>> >             final Thread t = new Thread(MAIN_GROUP, r, ...
>>> > - Such particular thread then calls SwingUtilities.invokeXxx() methods,
>>> > making EDT/AWT use the proper main AppContext & EventQueue.
>>> >
>>> >     EDT_DAEMON_THREAD_POOL.submit(new Runnable() {
>>> >                 public void run() {
>>> >                     SwingUtilities.invokeLater(doRun);
>>> >                 }
>>> >             });
>>> >
>>> > This pattern allows to transfer any Swing / AWT invocations from any
>>> > running Thread (in or out the main Thread group) into the proper
>>> > AppContext, like a "priviledged" ITW GUI action.
>>>
>>> Yeah, but it's still ending up in the EDT, I'm puzzled as to what is
>>> different here than directly using invokeAndWait/invokeLater.
>>>
>>
>> It is related to Swing/Awt internals that associate awt.AppContext to one
>> ThreadGroup:
>> in ITW, several thread groups are in use (application, security, main
>> ...) so each thread group can have its own awt.AppContext, EventQueue and
>> EventDispatcherThread as a consequence.
>>
>
> What would be the use-case that requires this "bridge"?
>

Typically all ITW internal dialogs (Security prompts, java console,
splashscreen from netx) but never any swing/awt code from the application:
it stays in its sandbox with its own jnlp classloader, security manager &
AppContext.


> I suggest to revert this part of the changeset and further discuss it. In
> case that something were broken by reverting only this part, I suggest to
> put the whole patch on hold.
>

Let's test the patch integrated in ITW 1.7 and see if any security issue
happens or let's discuss how to improve this patch.

Jiri could also do a rollback.

Laurent
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/distro-pkg-dev/attachments/20181003/9ffb6e81/attachment-0001.html>