[PATCH 3 of 3] chg: detect chg started by chg

Yuya Nishihara yuya at tcha.org
Thu Feb 25 10:22:05 EST 2016


On Thu, 25 Feb 2016 14:07:41 +0000, Jun Wu wrote:
> On 02/25/2016 01:41 PM, Yuya Nishihara wrote:
> > Can this message be merged with the timeout error? Since we have a
> > lock file, the result is dead lock, which is much safer than
> > executing chg recursively.
> >
> > I want to avoid future headache caused by introducing an environment
> > variable that is always set.
> 
> I think this is better because:
> 
> 1. time out can be caused by other issues like bad extensions or config
> 2. people can get feedback without waiting
> 
> I'm not sure how the environment variable can cause issues. Could you
> give an example?

Yep, it's better, but I can't say it's worth polluting environment variables.
My concern is the CHGMARK variable is visible to child processes. It might
be problem when simulating the hg behavior.

Silly example:

  $ hg --config='alias.x=!echo $CHGMARK' x


More information about the Mercurial-devel mailing list