[PATCH 3 of 5] preserve order of items in the list of special help topcis

[Adrian Buehlmann]

On 26.05.2008 20:39, Johannes Stezenbach wrote:
> On Mon, May 26, 2008, Adrian Buehlmann wrote:
>> On 24.05.2008 23:11, Johannes Stezenbach wrote:
>>> # HG changeset patch
>>> # User Johannes Stezenbach <js at sig21.net>
>>> # Date 1211658427 -7200
>>> # Node ID fa4fc9c14e654aa2be3e40ba11a75639ba47d61e
>>> # Parent  27401a7fa73a84d109719bbc95fb3fd4dced89a2
>>> preserve order of items in the list of special help topcis
>> Hmmm, so here you move the part about "revs" which you first added to help.py
>> in patch 2 to a different place inside help.py.
>>
>> So I gather this patch here depends on patch 2.
>>
>> But it would be a lot easier to review these both patches if you would place
>> the text in the right order in patch 2 already.
>>
>> Could you please reorganize these patches accordingly?
> 
> Sure. (Patch 3 is a consequence of patch 2 which to my surprise
> inadvertently reordered the text in the man page.)

Well, I had that surprise myself when I tested your patch 2 and
then looked into why things in hg.1.html were in wrong order after patch 2,
just discovering that patch 3 fixes that damage of patch 2.

Exporting your "trace of surprise" to this list is not what
I expect from a good patch series. The perfect patch series drives
the Mercurial files in a steady, observable way.

>> Maybe it would make sense to change the order of patch 2 and 3. First doing
>> the "preserve ordering" thing (this patch here) and then the "revisions help text
>> moving" (your patch no 2).
>>
>> Maybe patch 2 and 3 could be patchbombed as a an independent series? Are they
>> unrelated to patch 1, 4 and 5?
> 
> Logically they are unrelated, but all except patch 2 make changes
> to commands.py so there is potential for conflict...

Ok. I haven't looked into those other patches yet.

> I'm busy today, will resend the patches asap.

There is no need to hurry.