7 Nov
2006
7 Nov
'06
1:11 a.m.
On Monday, November 06, 2006 5:03 PM +0100 MichalČihař michal@cihar.com wrote:
Exactly... because svn things commit went wrong. But it actually didn't.
FWIW, I've seen a couple of crashes in the 1.4.0 Win32 client after a lengthy server transaction completed successfully. They seem to be in the APR hash that holds translation handles for console output translation. For instance, following a successful commit the svn.exe client crashed preparing the "successfully committed" message to send to the console. I haven't yet got a repeatable one that would let me properly debug it. What I've figured out has been from post-mortem dumps.