hg grep users, feedback wanted

Steve Borho steve at borho.org
Tue May 18 17:08:07 UTC 2010


On Tue, May 18, 2010 at 12:00 PM, Masklinn <masklinn at masklinn.net> wrote:
> On 2010-05-18, at 18:50 , Steve Borho wrote:
>>> But isn't that also the behavior of the ranged version? As I said, I
>>> have absolutely no issue with the change proposed for -rN, what I
>>> don't get is why grep -rN would change, but grep -rN:M wouldn't.
>>
>> What behavior do you propose for -rN:M?  It doesn't really fit the
>> manifest scan model, near as I can tell.
>
> I think I'd expect the same change of behavior you proposed for -rN:
> for each revision k in the N:M range, instead of scanning only the
> files changed in k it would scan all the files as they existed at
> `k`.
>
> Of course, that would be equivalent to doing a full scan of `N` and
> then replicating the current behavior for `N+1:M` (excluding the files
> matched during the initial full scan).
>
> Basically, I believe -rN:N should still return the same result as -rN.

This makes the old -rN behavior unreachable, but I see the utility of it.

--
Steve Borho



More information about the Mercurial mailing list