Universality [LWN.net]

Universality

Posted Mar 4, 2026 20:48 UTC (Wed) by nix (subscriber, #2304)
In reply to: Universality by quotemstr
Parent article: Magit and Majutsu: discoverable version-control

Honestly, the problem really is that vc's UI is so different from magit's transient-mode-based one that you'd need to completely rewrite vc's user interface, or throw away what makes magit's so usable -- and both of these would face severe resistance from one camp or the other. Breaking compatibility and existing user workflows like that is very much not the Emacs way.

I think magit's UI is far superior and more discoverable, but then of course it is: it's decades newer and written with the benefit of a lot more experience with various sorts of Emacsy UIs than the old, old vc, which was originally built for RCS and thus has a massive impedance mismatch with modern VCSes to overcome as well.

to post comments

Universality

Posted Mar 4, 2026 21:12 UTC (Wed) by quotemstr (subscriber, #45331) [Link] (1 responses)

Perhaps you're right about magit vs. vc, but is jj so different from git that we need to have a magit for magit?

Universality

Posted Mar 5, 2026 14:48 UTC (Thu) by nix (subscriber, #2304) [Link]

Probably not, but again it might be quite difficult to make magit selectively ignore the staging area in its UI, etc -- after all, one rationale for having a magit in the first place was that it didn't need piles of complexity and compromises induced by needing to abstract over differences between multiple version-control systems. We already have one of those -- vc -- and as a direct result it's rather anaemic.