Re: In support of Jonas Bernoulli's Magit

[raman]

Richard Stallman <address@hidden> writes:

I use both vc and magit -- so this perspective below might help where
each of vc/magit win:

For simple vc tasks -- and that typically equates to many things one
traditionally did in all VC systems, vc-git is what I use --- examples:

1. Check-in  current file,
2. Diff current version against a given version etc.

Reason: These are all one-shot kbd commands starting with C-x v

Magit:

There are many things that Git lets you do that are at the power-user
end of the spectrum --- and magit actually makes those doable, whereas
the git commandline would never encourage  you to venture even close.

Examples:

1. Selectively commiting  diff hunks in a file with many changes 

2. Selectively reverting specific diff hunks

3. Applying git patches 

4. Examining past log entries -- especially commits with many files,

5. From (4) jumping to a given file/diff hunk

These above I dont even know where I'd start with vc -- they may well be
possible.

Bottom-line: When one is going to perform a bunch of git tasks, magit
wins; when one if focusing on writing code and using the vc system as a
helper that makes sure you dont lose work, pressing C-x v   to invoke
the vc keymap means you dont switch attention away from what you are
doing.

Also -- for users coming to Git, magit may have a slight win since the
magit buffer shows you a bunch of things you can do at a given point --
whereas with VC, you need to have been using Emacs/VC for a while, or at
least have read the info pages in order to know what is possible.

> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > > Maybe so.  But it doesn't need to do everything Magit does.  It only
>   > > needs to be good enough to make git easy to use in Emacs.
>
>   > Then I think we already have that in VC
>
> The reason I think this isn't so is that repeatedly I've seen people
> post on this list for help dealing with git, and each time people say
> to use Magit.
>
> I have not done a scientific study, which would include investigating
> whether the recommendation actually helped those people, and what
> fraction of Emacs users think VC is just fine for using git.
> I have simply supposed that the advice that was given is good advice
> since the people who give it are expert Emacs users.

--