Re: [PATCH] Issue #1295 (take 2, with log)

On Fri, 9 May 2003, [UTF-8] Branko Čibej wrote:

> int cmp = ((item1->state_flags & SVN_CLIENT_COMMIT_ITEM_UNMARKED)
> - (item2->state_flags & SVN_CLIENT_COMMIT_ITEM_UNMARKED));
> if (cmp == 0)
> cmp = svn_path_compare_paths (item1->url, item2->url);
> return cmp;

I guess that will work too. Depends which one you find more readable. Any
suggestions as to which one?

> The docstring for this function should explain exactly what kind of
> input it expects, what it'll parse and when it'll fail. I'm also a bit
> worried about its silently doing nothing if it encounters something it
> can't parse; I think that's a potential source of many surprises and

If it gets something it can't parse, it aborts and whatever was parsed is
checked in. I figured that's what people would want.

I can certainly make it return an integer of success/failure on parsing
and abort the checkin if there is an error.

Then just wrap the call in SVN_ERR...

> "error" reports; we want to avoid that, so giving a bit of though to the
> failure cases would be good here.
>
> And we really want tests for this feature (in commit-tests.py).

I'd really rather not touch Python. I know Perl, but not Python and this
would take me a considerable amount of time to learn.

> Modifying the commit list in a message passed in with -F should work, right?

I don't know. The code is pretty hard to follow. Not even sure how -F
works.

L8r,
Mark G.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri May 9 09:47:18 2003