You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Alexander Thomas <al...@collab.net> on 2005/08/02 13:30:15 UTC

[Patch] Fix svn issue #2269 - warn and skip for nonexistent file

Patch to fix svn issue #2269 - warn and skip for nonexistent file
on 'svn rm' and exit with failure.

Patch is still incomplete respect to the expected behavior and also its
a stupid patch, if implemented will fail almost all test. :-). This
patch is only meant for discussion and suggestion and not for review.

What I tried to do is, when a target is nonexistent
svn_wc_adm_open_anchor() will return with an error. Its a very common
function used in many svn commands, so this behavioral change will
trigger almost all test to fail.


Advance Thanks

Regards
-Alexander Thomas (AT)






Re: [Patch] Fix svn issue #2269 - warn and skip for nonexistent file

Posted by Julian Foad <ju...@btopenworld.com>.
Alexander Thomas wrote:
> Patch to fix svn issue #2269 - warn and skip for nonexistent file
> on 'svn rm' and exit with failure.

Thanks for working on this.

> Patch is still incomplete respect to the expected behavior and also its
> a stupid patch, if implemented will fail almost all test. :-). This
> patch is only meant for discussion and suggestion and not for review.

OK, well, what's good or interesting about this patch?  I haven't formed an 
opinion, I'm just asking because you're only saying negative things about it. 
What do you want to discuss?  Do you understand what behaviour is required in 
enough detail to implement it?

> What I tried to do is, when a target is nonexistent
> svn_wc_adm_open_anchor() will return with an error. Its a very common
> function used in many svn commands, so this behavioral change will
> trigger almost all test to fail.

Alexander, you would probably get a response if you tell us what you are 
thinking and ask specific questions.  Do you think this is the best way to 
address the problem, and all the other callers of that function can be updated 
in some predictable way, or should the tests be failing because they are wrong, 
or what?  Are you just throwing out some half-baked ideas?  Have you hit a 
conceptual problem, or a practical problem, or are you still working on this? 
Tell, and ask, or propose ideas, and we'll answer or discuss.

- Julian

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org