Monocle in 8u25

[ngalarneau at ABINITIO.COM]

Hi Benjamin,

Your progress sounds very good!

In your email below, you said: "Also I thought that events with 
Platform.runLater() will be processed first-in-first-out, but it seems 
that this is not the case."

Kevin, on an email on Nov. 5th to this list, said that the only exception 
to Platform.runLater() Runnables being processed in order was in the case 
of a showAndWait() being done. That blocks the current thread, which stops 
the current Runnable from completing while a new thread takes the next 
event/Runnable from the queue. In that case later Runnables would be 
processed before the earlier one completes.

Does that explain what you have seen?


Neil




From:   Benjamin Gudehus <hastebrot at gmail.com>
To:     ngalarneau at abinitio.com, 
Cc:     Tom Eugelink <tbee at tbee.org>, "openjfx-dev at openjdk.java.net" 
<openjfx-dev at openjdk.java.net>
Date:   11/15/2014 10:50 AM
Subject:        Re: Monocle in 8u25



Small update:

Regarding (2) "requires interactions to be run withing the JavaFX thread": 
Nearly all integration tests for TestFX pass now. However the interactions 
freeze when using Glass robot, Node.startDragAndDrop() and 
ClipboardContent.putString() together. I guess the timing with 
Platform.runLater() needs some work.

Regarding (3) "retrieve screenshot images in headless mode": I forgot to 
use "-Dprism.order=sw". Screenshots work in headless mode.

--Benjamin

On Sat, Nov 15, 2014 at 1:48 AM, Benjamin Gudehus <hastebrot at gmail.com> 
wrote:
You can find the PR for headless support in TestFX here:
https://github.com/TestFX/TestFX/pull/159

There are some issues with the Glass robot. We need this robot in order to 
simulate user interactions in headless mode. 

One problem is, that the Glass robot (1) requires an JavaFX application to 
be set up. TestFX initializes the robots first and then launches the 
Application. I worked-around this problem, by initializing the Glass robot 
just before the first user interaction is made.

Another problem is, that the Glass robot (2) requires interactions to be 
run within the JavaFX thread. From the integration tests I've wrote it 
looks that we need to wait for the events in the JavaFX thread to complete 
for certain interactions, e.g. we need to wait after mouseMove() in order 
to do mouseClick(). With the AWT robot we only needed to wait for the 
events triggered by interactions to complete (let's call them 
post-interaction events). Now it seems that we also need to wait for the 
interactions itself to complete, which then will also trigger the 
post-interaction events. I'll need to visualize different interaction 
events, post-interaction events and the event waiting, and decide when we 
need to use the semaphore solution we currently use to wait for 
post-interaction events. Also I thought that events with 
Platform.runLater() will be processed first-in-first-out, but it seems 
that this is not the case.

Yet, another problem is, that I did not manage to (3) retrieve screenshot 
images in headless mode with the Glass robot. Although it works in normal 
mode. Maybe the JavaFX event robot can retrieve the screenshots. 
Interestingly the HeadlessScreen updates it's framebuffer (
https://github.com/TestFX/Monocle/blob/0ab3986f4c/src/main/java/com/sun/glass/ui/monocle/headless/HeadlessScreen.java#L117
).

Regards,
Benjamin





 
NOTICE from Ab Initio: This email (including any attachments) may contain 
information that is subject to confidentiality obligations or is legally 
privileged, and sender does not waive confidentiality or privilege. If 
received in error, please notify the sender, delete this email, and make 
no further use, disclosure, or distribution.