Offering hg log -G by default?

Matt Mackall mpm at selenic.com
Sat Dec 17 10:50:57 CST 2011


On Sat, 2011-12-17 at 11:44 +0100, Dirkjan Ochtman wrote:
> On Fri, Dec 16, 2011 at 23:58, Matt Mackall <mpm at selenic.com> wrote:
> > If you examine the implementation of log and graphlog you'll discover
> > they're basically completely different. The prerequisite for making log
> > -G a core feature is unifying the implementations.
> >
> > And that probably means adding a no-graph mode to graphlog that exactly
> > matches the core log output for all our weird little options without too
> > many serious performance regressions.
> 
> Yup. I spent some time looking at this a few years ago. log supports
> many options that graphlog doesn't support, some of which can't be
> usefully supported by graphlog. There's also the thing about file
> logs, and their graphs, which can be somewhat weird. So it would
> probably be good to have, but it's definitely a non-trivial amount of
> work.

Patrick and Alexander moved this -a lot- closer around the 1.9 sprint.
The biggest remaining problem, as I recall, was filelog emulation, which
I subsequently added support for to revsets.

-- 
Mathematics is the supreme nostalgia of our time.




More information about the Mercurial-devel mailing list