[PATCH] [RFC] convert: fix --datesort

[Patrick Mézard]

Matt Mackall a écrit :
> On Thu, Dec 20, 2007 at 11:03:47AM +0300, Kirill Smelkov wrote:
>> ?? ?????????????????? ???? 19 ?????????????? 2007 04:51 Matt Mackall ??????????????(a):
>>> On Wed, Dec 19, 2007 at 01:52:22AM +0300, Maxim Dounin wrote:
>>>> Hello!
>>>>
>>>> On Thu, 13 Dec 2007, Kirill Smelkov wrote:
>>>>> # HG changeset patch
>>>>> # User Kirill Smelkov <kirr at mns.spb.ru>
>>>>> # Date 1197567565 -10800
>>>>> # Node ID 7a1570348d6df8f59289eddaee9452393033b078
>>>>> # Parent  eb2f94f6aeabff1638e0ad3c46fcbe80bf4e8fc2
>>>>> [RFC] convert: fix --datesort
>>>>>
>>>>> The problem is that previously commit.date was used for sorting, but
>>>>> it's a string like "1 Jan xxx 2007", so it it wrong to use it for
>>>>> sorting.
>>>>>
>>>>> Another problem is that why we are using depth for sorting -- I have no
>>>>> clear answer -- it seems to be plain wrong.
>>>> As far as I understand, depth is here to make sure parent changeset
>>>> always commited before it's child regardless of dates specified.
>>>>
>>>> I don't think dropping this invariant is a good idea.
>>> Indeed, and it fails tests in our test suite.
>>>
>>> But never fear, I already fixed it and pushed it to mainline.
>> I agree that commits habe to go after their parents, but sorting on depth is 
>> wrong too.
> 
> Sounds like we'll need to do something smarter with toposort then.

49c69e1e4aa2 should fix this in crew and crew-stable by replacing the depths-based toposort with a straightforward graph traversal with configurable revision selection.

--
Patrick Mézard