win32text and excluding patterns

[Steve Borho]

On Wed, Apr 15, 2009 at 8:06 PM, Mark Hammond <mhammond at skippinet.com.au> wrote:
>> Right, right now people can't rely on the ordering. But people rely on
>> ** filters being applied to _all_ files. What you are asking for is
>> exceptions to that, right? It is hard to get both.
>
> I don't think that is quite correct - IIUC, people currently can not rely on
> a ** pattern _and_ a specific pattern - there currently is no guarantee
> which one is run.
>
> In other words, people can not currently rely on ** filters being applied to
> a file even when that file matches another pattern, as hg currently doesn't
> promise that.  So I don't see a backwards compatibility issue - it is just
> making the behaviour deterministic.
>
> Am I misunderstanding something?
>
>> I think you should ask the core/crew developers instead ;-)
>
> :)  I thought this is where they hung out ;)  I'm afraid I haven't been here
> long enough to know exactly who owns what parts of the puzzle.  Who is the
> nominal 'owner' of win32text?
>
> Could one of the core/crew developers please speak up and let me know if I
> changed localrepo's _filter() function as Dirkjan suggested (ie, longest
> filters get highest priority), is there a reasonable chance it would be
> accepted?

FWIW, these kinds of cricket noises from dev-land usually indicates a lack
of resistance to your idea.  Feel free to submit a patch that resolves the
problem to your satisfaction and be prepared to argue on the merits.

--
Steve Borho