blame improvements in hgweb (was Re: Future development on hgweb and blame)

[Denis Laxalde]

Gregory Szorc a écrit :
> Speaking for Mozilla, we care about improving the interface to blame
>  most. It is surprisingly common how often our engineers need to
> perform "code archeology" to see how something worked 5+ years ago.
> The hgweb interface to blame is used most frequently for that because
> it is much richer than the terminal-based interface. The hgweb
> interface to blame today is very tedious and requires a lot of extra
> cursor/mouse movements and clicks to access information. Making it
> faster to access data (namely blame data for previous/next changeset
> and changeset metadata/diff) would be huge productivity wins. Most of
> the improvements on BlamePlan center around this theme. The "advanced
> html interface" would be the culmination of that work. But the
> smaller scoped improvements would be huge wins on their own.

I just sent a few patches concerning the annotate view in hgweb. These
address the following items from the "blame plan"
(https://www.mercurial-scm.org/wiki/BlamePlan):

* "hgweb links to changeset": add links to the "diff" and
"rev" view aside of the "annotate" in annotate table;

* "hgweb links to ancestor's blame": insert a new column
on the left of the current revision's one in annotate view with a link
to the ancestor revision(s) which modified the file.

The patches mostly touch the gitweb templates (used by Mozilla AFAICT). 
Let me know if that'd be useful.

-- 
Denis Laxalde
Logilab         http://www.logilab.fr