[PATCH 1 of 2] phases: add a phase and phasestr method to file context

[Sean Farley]

On Wed, Oct 17, 2012 at 1:18 PM, Pierre-Yves David
<pierre-yves.david at logilab.fr> wrote:
> On Wed, Oct 17, 2012 at 01:08:35PM -0500, Sean Farley wrote:
>> On Wed, Oct 17, 2012 at 12:49 PM, Pierre-Yves David
>> <pierre-yves.david at logilab.fr> wrote:
>> > On Wed, Oct 17, 2012 at 10:38:40AM -0500, Sean Farley wrote:
>> >> On Wed, Oct 17, 2012 at 10:33 AM, Pierre-Yves David
>> >> <pierre-yves.david at logilab.fr> wrote:
>> >> > On Wed, Oct 17, 2012 at 10:28:20AM -0500, Sean Farley wrote:
>> >> >> # HG changeset patch
>> >> >> # User Sean Farley <sean.michael.farley at gmail.com>
>> >> >> # Date 1350425390 18000
>> >> >> # Node ID 072adc076e641990cf0000ece34e9545ebaaa0b3
>> >> >> # Parent  72c234081ae1350220132c69750f5a093902a1e7
>> >> >> phases: add a phase and phasestr method to file context
>> >> >
>> >> > Hey? we already have them!
>> >>
>> >> But not for filecxt :-( ... unless I am missing something about
>> >> wanting to call 'ctx.phasestr()' without having to test whether it is
>> >> a changectx or filectx?
>> >
>> > Not that filectx is very deceptive. It usually point to the **first revision**
>> > that introduce a content. This revision may be a different revision that the
>> > one you care about. With a differrent phase (and different visibility)
>> >
>> > :-(
>>
>> Hmm, then I'm confused how functions like
>>
>> def date(self):
>>     return self._changectx.date()
>>
>> work?
>
> badly
>
>> If self._changectx points to the first revision then wouldn't date() be
>> possibly incorrect?
>
> The subtility is about what first revision mean. By "first revision" do you mean:
>
> - First of my ancestors that installed the content I have in revision X
>
> - First revision to have the content in the repo ever ?
>
>   This may be a revision in a totaly unrelated branches, with a **more recent**
>   date, difference user, etc.
>
>   It just happen to have the very same content than you.
>
>> Or is it more subtle than that, i.e. is the caller of the function supposed
>> to update self._changectx somehow?
>
> The current state is: Broken in a way few people notice and care.

:-o