Excerpts from Augie Fackler's message of 2016-12-28 18:17:00 -0500: > A thought: it's (notionally) optional for clients of the cmdserver to > ignore the pager operation, right? That is, it's totally valid to drop > the "start a pager now" message on the floor and act like no pager is > in play. Maybe it's worth having some mechanism for that.
The compatibility issue is more about "ui.system", less about pager, as it's the format change of the "S" channel. That said, I think the chg protocol should be considered like "internal API", and chg and hg should probably always be in a same package. So it seems fine to just ignore old chg client and optimize for code cleanness - I tend to just do it without introducing new capability layer to make my life easier. > +1. If you want to talk more about pager stuff let me know - I'm not > at work again until the 9th of January, but may be able to find time > here or there to chat before then if you need. I think the path for pager to move forward is clear. I'm more interested in rough comment on "the new changelog" direction [1], like whether it is a direction that will be loved by everyone, are there better alternatives, etc. [1] https://www.mercurial-scm.org/pipermail/mercurial-devel/2016-December/091846.html _______________________________________________ Mercurial-devel mailing list Mercurial-devel@mercurial-scm.org https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel
