bug#22470: 25.1.50; vc-annotate (HG) does not display changes within one day (colors)

[Dmitry Gutov <dgutov@yandex.ru>]

On 01/26/2016 10:18 PM, Uwe Brauer wrote:

> Well I hoped one could generate coloring which reflects changes within
> the day, but display the time format deleting the hours and minutes but
> it seems not to be possible.

Not possible right now, no.

>      > Finally, at a very quick glance that ahg thing seems to colorize
>      > based on revision number, not on age, which is a different scheme.
>
> By revision number do you mean the changeset or the internal local
> revision number which hg assigns to its commits but git does not? If it
> is the former, maybe that would be a more appropriate model for
> vc-annotate?

Whatever revision-identifying numbers or strings appear in the output.

>      > So personally I see no bug here.
>
> I did not use the word bug, I was asked to send a bug report in order to
> find out whether there is a bug.

You were asked to file the bug for a different reason: you said there's 
a difference in behavior here between emasc-25 and master, which would 
be a problem. Was the version from one of the branches too old?

Another problematic thing is visible on your screenshot here: 
http://lists.gnu.org/archive/html/help-gnu-emacs/2016-01/pngrALEdUf1mc.png

Revisions from two consecutive days are still colored the same. At the 
time, I thought it to be a bug, but it depends on how big is the range 
of time between the oldest and the newest revisions in the buffer: after 
all, the annotation palette is limited.

If the commits were made in N most recent days, one can to tweak the 
precision (by customizing vc-annotate-display-mode), but you seem to 
have figured it out already.

Glenn, don't you think it's a problem that we can't distinguish between 
older commits made on consecutive days? I'd say it's an argument in 
favor of adopting aHg's scheme, or at least adding it as an option.