Comments on: A Treatise On The History Of Distributed Version Control

By: Jakub Narebski

Jakub Narebski — Fri, 25 Jan 2013 15:48:08 +0000

@btubbs: Git doesn’t allow to “partially type commands if you’ve given enough characters to disambiguate” (and I am not sure if it could be considered to be added, as adding new command could make abbreviation no longer work), but it has tab-completion of commands, options and parameters, and does support aliased.

By: btubbs

btubbs — Tue, 22 Jan 2013 21:07:56 +0000

I admit to being envious of the curation done on the commit history for things like the Linux kernel. That’s a good example of where an understandable history is hugely important.

Nice job making this post about mutable vs immutable histories instead of Git vs Mercurial.

I’ve gotten value in the past from being able to drill down exactly and “hg blame” the changeset that introduced a particular bug. Some of the examples above threaten to subvert that, because commits from several users would be lumped into one.

But I don’t do it that often, and I think I’d probably get more value from the curated history if I had to choose.

I do feel that the Mercurial CLI is a lot more human friendly than Git’s. There seem to be a lot fewer obscure terms, commands, and command options to learn. And though it’s a trivial difference, I really like Mercurial’s feature of letting you partially type commands if you’ve given enough characters to disambiguate what you want to do. I’ll have to see what git-extras offers.

By: Norbert Kéri

Norbert Kéri — Sun, 20 Jan 2013 23:34:32 +0000

Very nice article. I have started off working with distributed version control with mercurial, and for the current job I’m using git, and I strongly relate to everything you said. Both of them can do what the other can, but because they follow a different a mentality, some things are easier in one, and some are harder than in the other.