3 ms·

I didn't like it when Tarsius took over the project. I still believe that a lot of changes were unneeded. I think that some changes made it worse (and for a l

I didn't like it when Tarsius took over the project. I still believe that a lot of changes were unneeded. I think that some changes made it worse (and for a long time I was using my own fork from before he took over). All this said, Magit is still the best interface to Git I know, and I would regret to see it become unmaintained.

School-Cotton 4y ago
What were some of the changes you didn’t like?
- crabbone 4y ago
  I remember that my discontent that set off the whole thing was the interaction in magit-refs mode. It used to be the case that pressing RET on a branch would switch to that branch, while what it does today: shows the HEAD commit on that branch, which is, kind of worthless. But, I don't think this is how it worked immediately after the takeover. I think it would check out the head commit w/o switching branches putting you in a "detached head" state for no reason. I would've been probably content with a few changes I didn't like, but Tartius' response to my complaint was what set it all aflame. It turned into pointing fingers and name calling pretty quickly. Then there were useless additions of how pushing worked (the process used to be simpler than it is today, and would do the right thing w/o your intervention). Then Tartius decided to add a new way to diff two branches, where instead of just calling diff, you'd have to set up one local branch as a remote for another local branch (I think, this nonsense still exists). But the problem was more general than individual points of discomfort. Magit was started being flooded with pointless features that'd get in the way of using the remaining useful parts. I generally believe that a huge chunk of Git is worthless. Eg, I'd never regret completely throwing away everything that has to do with merges, subtree commands, few other things... Magit was very good at focusing on useful features and ignoring the worthless ones. That gave an "opinionated", but a conceptually easier model of the program. Since the takeover, the general direction was to try to express every feature present in Git in Magit, which set it on the path of becoming just as unnecessary complex as the original. Except, for some reason not git-grep, which I had to add myself. Go figure.

School-Cotton 4y ago

What were some of the changes you didn’t like?

crabbone 4y ago
I remember that my discontent that set off the whole thing was the interaction in magit-refs mode. It used to be the case that pressing RET on a branch would switch to that branch, while what it does today: shows the HEAD commit on that branch, which is, kind of worthless. But, I don't think this is how it worked immediately after the takeover. I think it would check out the head commit w/o switching branches putting you in a "detached head" state for no reason. I would've been probably content with a few changes I didn't like, but Tartius' response to my complaint was what set it all aflame. It turned into pointing fingers and name calling pretty quickly. Then there were useless additions of how pushing worked (the process used to be simpler than it is today, and would do the right thing w/o your intervention). Then Tartius decided to add a new way to diff two branches, where instead of just calling diff, you'd have to set up one local branch as a remote for another local branch (I think, this nonsense still exists). But the problem was more general than individual points of discomfort. Magit was started being flooded with pointless features that'd get in the way of using the remaining useful parts. I generally believe that a huge chunk of Git is worthless. Eg, I'd never regret completely throwing away everything that has to do with merges, subtree commands, few other things... Magit was very good at focusing on useful features and ignoring the worthless ones. That gave an "opinionated", but a conceptually easier model of the program. Since the takeover, the general direction was to try to express every feature present in Git in Magit, which set it on the path of becoming just as unnecessary complex as the original. Except, for some reason not git-grep, which I had to add myself. Go figure.