Why is it so buggy?

Nov 24, 2009 at 8:28 PM

Last night I attempted to import source to my codeplex project via SVN.

The import failed.  The remote connection was closed because it timed out.

So I decided to import things one at a time.

I import a directory.  Whoops, it lost part of the directory structure, so I use the repo browser in TortoiseSVN to create the right structure.  Then I try to rename the directory into the appropriate location.  Whoops, can't do that because of the directory rename sorting bug.  OK, just delete the existing files in the wrong place and add them back in the right place in two separate steps.  Clunky, but I can put up with that on the occasional directory rename.

Try to import more directories.  Another timeout and the import failed.  Well, sort of failed.  It committed most of the files, but not all of them.  Do an update.  TortoiseSVN says I'm up to date, but it didn't pull down the files that were partially committed.  OK, check out fresh and I've got the files.  Then make more changes.  Do an update.  It says I'm up to date, but it didn't give me the changes.

I'm sorry, but I've never had this much trouble with a subversion server.  When I was at the MVP Summit last year, people reported that they use the subversion bridge with codeplex all the time and have no difficulties.  Either they only do very mild operations on their projects, or codeplex has gone to sh*t since that report when it comes to using subversion.

Just what is going on here?  Why is the SVN bridge so buggy?