Bug 52362 – Java applet canvas view is not rendered when overlapping with Flash canvas

• Component: Java
• Status: RESOLVED WONTFIX

[Raphael Maes]

During recent months, all WebKit builds on OSX 10.6 and running in 64 bit mode have shown this problem.
When there is an overlap (of even 1 pix) between a Java Applet canvas and Flash-player canvas, the applet fails to render it's view. A blank area is shown. All GUI elements are available, just not visible. There is no issue when the 2 canvas area's don't overlap.

IT only affects 64bit Mac machines.
- Adobe Flash-player v10.1.x
- Java v1.6.0_22
- Java preference mode is: Run applets in Browser process
- A crippled work-around is to run the applets in their own process. This however introdues other side effects (2 contexts are introduced, a need to double -click to get focus, etc...).

[Alexey Proskuryakov]

Do you have an example page or a test case for this?

[Raphael Maes]

Created attachment 78895 [details]
Java applet running in Flash canvas - OK (Java Logo is rendered)

[Raphael Maes]

Created attachment 78896 [details]
Java applet running in Flash canvas - Not OK (applet canvas is not rendered and shows a blank area)

[Raphael Maes]

URL to view the applet behaviour: http://delano-eqa.tst.agfa.com/PageMaster

 - In the logon dialog box : Click the link "Check my system". This wil show a floating window executing different system checks.

 - The line "Java Runtime Enviroment" should get a green 'V' sign after loading the (large) Java applet. The applet canvas will show the Java Logo (OK) or a blank area (NOK). See also the attached screenshots for an example.
 - Should the Java line receive a red 'X' i.o. a green 'V' , please click the "Restart Test" button.

[Alexey Proskuryakov]

Created attachment 78902 [details]
different results with a nightly

Thank you!

This doesn't quite match what I'm seeing, even though there seems to be a regression of some kind. Here are my results.

Safari/WebKit 5.0.3 on Mac OS X 10.6.6, Java 1.6.0_22 running within browser process:
- Many times, the applet failed to load (red X near the test).
- When it did load, it rendered as blank - no Java logo. I.e., exactly your "not OK" screenshot.

Nightly r75294:
- The applet never loads, and a partially visible <object> fallback text was displayed (see attachment).

[Alexey Proskuryakov]

The results I'm seeing aren't surprising, as there have been many changes to <object> loading in nightlies. Andy (CC'ed) knows the details better than I do.

It is surprising that we seem to be getting different results in shipping Safari/WebKit. Or did I misunderstand you somehow?

[Raphael Maes]

Indeed, we can confirm your observation about the difference between the WebKit nightly build and the release Safar 5.0.3 / WebKit version.

[Alexey Proskuryakov]

Please clarify what this bug should be about. If it is about issues in shipping Safari/WebKit, then we need a separate bug for the regression in nightlies.

[Raphael Maes]

The blank canvas is first and foremost an issue with the released Safari/WEbKit. (We have logged this issue also in Apple's Radarweb tracking system ae: 8846411)

As you have observed ,  the recent nightly WebKit builds show different behaviour. Here the Applet doesn't get as far as in the Safari/WebKit version. (It fails to start). Should we open a separate case for this or are you guy's doing this ?

Thanks for the quick response.

[Alexey Proskuryakov]

Thanks!

It would be slightly better if you filed a bug about the regression.

[Ahmad Saleem]

Java applets are not support anymore and also NPAPI support is also gone. Marking this as "RESOLVED WONTFIX".