Re: update/405 question

Hi Justin,

On Fri, Feb 18, 2005 at 04:39:09PM -0500, Justin R. Miller wrote:
> Occasionally, my users will go to perform a backup, which is really 
> just a 'sitecopy -uo <site>' after a previous run was cancelled or 
> otherwise aborted.  Their local sitecopy status cache will be out of 
> sync with the remote version, and when the collection update is 
> attempted, the subdirectories on the server will have already been 
> created so sitecopy fails with a bunch of 405 Method Not Allowed HTTP 
> errors.

The previous run was aborted with Ctrl-C etc locally?  sitecopy's
handling of that is not very graceful; the ideal fix would be to make it
more robust.

> Then, sitecopy will exit with errors and throw my app out of whack 
> because, well, it exited with errors.  However, if only sitecopy would 
> try to do the update without the failed mkdir things would be fine.
> 
> My questions is this:
> 
> I see in the past (via the archives) that the 405 was changed from a 
> 409 and that is according to spec.  However, is there a less fatal way 
> to handle this situation that could be implemented in sitecopy?  Can't 
> sitecopy just see the directories as already existing and then start 
> uploading the content?

No, because it is as an error as far as sitecopy is concerned.  You can
use "sitecopy --keep-going --update" and it should carry on past the
errors; then you could use --catchup to tell sitecopy that the state was
in synch. 

joe
_______________________________________________
sitecopy maillist  -  [email protected]
http://dav.lyra.org/mailman/listinfo/sitecopy