I think I found one potential problem in the JavaHL merge code. See the
patch at the end of this message. I am not really a C or C++ programmer
so I do not know for sure, but that area of the code is different than all
of the others, so I assume it is a bug.
That being said, I do not think this is the cause of the bug I reported.
Why? Well, there was another new merge signature added in 1.2, one that
uses the peg revision syntax. It looks like the native C++ code even
calls a different svn client API. svn_client_merge_peg instead of
svn_client_merge. I modified the Subclipse code to call this JavaHL
method instead (even though we could not use this for real in Subclipse)
and it gets the EXACT same error. This other method in JavaHL does not
have the same problem as the patch below.
So while I think the attached patch may be valid, it seems like there is
still some other problem at work here.
Comments?
Thanks
Mark
Index: javahl/native/SVNClient.cpp
===================================================================
--- javahl/native/SVNClient.cpp (revision 14452)
+++ javahl/native/SVNClient.cpp (working copy)
@@ -1125,7 +1125,7 @@
return;
}
- Path srcPath2 = path2;
+ Path srcPath2(path2);
Err = srcPath2.error_occured();
if(Err != NULL)
{
_____________________________________________________________________________
Scanned for SoftLanding Systems, Inc. by IBM Email Security Management Services powered by MessageLabs.
_____________________________________________________________________________
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Apr 28 17:37:31 2005