Mark Phippard <MarkP@softlanding.com> wrote on 04/26/2005 02:25:53 PM:
> Doing some more testing for Subclipse. This is on Windows with 1.2 RC1
> with the JavaHL RC2 patch applied.
> I am updating Subclipse to use the new JavaHL merge method signature so
> that we can support the --dry-run option. When I try to run this method
> get a crash. Here are the relevant details:
> An unrecoverable stack overflow has occurred.
> An unexpected exception has been detected in native code outside the VM.
> Unexpected Signal : EXCEPTION_STACK_OVERFLOW (0xc00000fd) occurred at
> Current Java thread:
> at org.tigris.subversion.javahl.SVNClient.merge(Native Method)
> - locked <0x14ca0988> (a java.lang.Class)
This is starting to look like a fairly serious bug. I reverted the
Subclipse code to use the older JavaHL merge method, which does not
support --dry-run, and it still crashes the JVM. Also, I am now using
brane's RC2 binaries.
If svn 1.2 is released with this, we would have to remove the merge option
from Subclipse since it crashes the JVM.
Also, FWIW, I took a look at the JavaHL tests and it does not look like
there are any tests that run merge. That could explain why this problem
does not show up in general.
It sounds like Patrick Mayweg is out for a while, is there anyone else
that could look at this? Maybe it is something simple like the last
JavaHL crasher in RC1.
The info is the same as last time:
An unrecoverable stack overflow has occurred.
An unexpected exception has been detected in native code outside the VM.
Unexpected Signal : EXCEPTION_STACK_OVERFLOW (0xc00000fd) occurred at
Current Java thread:
at org.tigris.subversion.javahl.SVNClient.merge(Native Method)
- locked <0x14c76a00> (a java.lang.Class)
Scanned for SoftLanding Systems, Inc. by IBM Email Security Management Services powered by MessageLabs.
To unsubscribe, e-mail: email@example.com
For additional commands, e-mail: firstname.lastname@example.org
Received on Thu Apr 28 16:06:08 2005