Re: [PATCH] Issue #2105 - Add '--no-ignore' option to 'svn add' and 'svn import' (v5)

Julian Foad <julianfoad@btopenworld.com> writes:
> Oh, this is awkward. I see now that we already have parameters named
> "no_ignore" in public APIs, and we can't just make them go away. Does
> this mean that we should be hesitant about calling other parameters
> "do_ignore", which is very similar in name but has the opposite sense?
> I don't know. Somebody help us to decide, please!

Ooops!

Although I agreed with you earlier that a positive-sense parameter was
better, in this case I think consistency should come first. Let's
stick with 'no_ignore' after all. Sorry for the confusion Ramaswamy!

Internally, we should do whatever sense reversals make the code come
out clearest, of course.

> I don't know which less generic error we should raise, and I'm not
> sure that we should do so. Karl said he would just use
> "svntest.Failure", and you did that, and I am happy with that. The
> comment has now been discussed and Karl effectively made a decision
> (that we should not raise a less generic error), so please now delete
> the comment.

Note that 'raise svntest.Failure' can take a string argument
describing the nature of the failure. See other tests for examples of
this.

-Karl

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Jun 2 17:39:13 2005