RE: [PATCH] fix for issue #2867: svn cleanup dies on obstructed subdirs

> One rule is that we can't printf() from within a library. We can only
> return errors (we can chain errors together, delay their return till
> after the work is done, whatever, but simply printing is a no-no in a
> library).

Definitely (I made a note to this effect). This was a debug thing and
shouldn't have been in the patch.
 
> The other thing that worries me is, are you sure that this is the only
> circumstance in which SVN_ERR_WC_NOT_DIRECTORY can be returned?

Good question. I suppose I'd have to dive into svn_wc_cleanup2() to
find out?

> Third: a reproduction case would greatly help. (If there was a thread
> preceding your message, my apologies, I don't have it at hand.)

http://subversion.tigris.org/issues/show_bug.cgi?id=2867

I guess the fundamental question is - should an obstruction be
considered a fatal error during svn cleanup? It isn't for status (or
status -u finally) it is of course for update and commit.

If it is, then I believe a -force option for cleanup is in order.

Eric

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-01-21 17:09:52 CET