| Can p2 touchpoint actions be conditional? (Sun JVM vs. Apple JVM) [message #132099] |
Fri, 29 May 2009 16:54  |
Eclipse User |
|
|
|
Hi,
We are using a touchpoint action to add a VM arg that is needed as a
workaround when running on a Sun JVM. It seems that these options
may cause the Apple JVM to crash, so we would want to invoke the
touchpoint action ONLY WHEN running on a specific platform (ideally
depending on the current JVM, optionally based on the OS).
Question: is a concept of conditional actions for p2 touchpoints realized
or planned?
For the records, the JVM options in question are:
-XX:+UnlockDiagnosticVMOptions
-XX:+UnsyncloadClass
Running WITHOUT these options can cause class-loading deadlocks on Sun JVM
(as discussed in https://bugs.eclipse.org/bugs/show_bug.cgi?id=212262#c13 ),
running WITH these options seems to cause SIGBUS on Apple JVM.
Has anybody seen this?
thanks
Stephan
|
|
|
| Re: Can p2 touchpoint actions be conditional? (Sun JVM vs. Apple JVM) [message #132152 is a reply to message #132099] |
Mon, 01 June 2009 12:41  |
| Eclipse User |
|
|
|
Hi Stephan,
Filtering is done at the IU level so you can creat an OS specific IUFragment
to do the JVM option adding.
With that said, I think instruction level filtering might be interesting so
why not open a bug to track this idea.
-Simon
"Stephan Herrmann" <stephan@cs.tu-berlin.de> wrote in message
news:gvpi1b$kt7$1@build.eclipse.org...
> Hi,
>
> We are using a touchpoint action to add a VM arg that is needed as a
> workaround when running on a Sun JVM. It seems that these options
> may cause the Apple JVM to crash, so we would want to invoke the
> touchpoint action ONLY WHEN running on a specific platform (ideally
> depending on the current JVM, optionally based on the OS).
>
> Question: is a concept of conditional actions for p2 touchpoints realized
> or planned?
>
> For the records, the JVM options in question are:
> -XX:+UnlockDiagnosticVMOptions
> -XX:+UnsyncloadClass
> Running WITHOUT these options can cause class-loading deadlocks on Sun JVM
> (as discussed in
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=212262#c13 ),
> running WITH these options seems to cause SIGBUS on Apple JVM.
> Has anybody seen this?
>
> thanks
> Stephan
>
>
|
|
|
Powered by
FUDForum. Page generated in 0.24442 seconds
Home