Re: Version control tools

[Urs Liska]

Am 08.01.2014 11:24, schrieb Simon Bailey:
> hi,
>
> On Wed, Jan 8, 2014 at 10:20 AM, <address@hidden
> <mailto:address@hidden>> wrote:
>
>     Joram:
>      > Simon Bailey:
>      > > so, my follow up question: why use git as a single user?
>      >
>      > I always use git, event git-svn for svn repos.
>      > Comparing to CVS and SVN, I see these advantages:
>      >
>      > - easy branching
>
>
> what for when working with scores?

Did you see 
http://lilypondblog.org/2014/01/why-use-version-control-for-engraving-scores/ 
?
There's a section about this specific question. The remaining sections 
are rather tool-agnostic and apply equally to SVM.
Bottom line is: You can use branches to keep master in a consistent 
state. You can work at different spots in parallel, keeping the whole 
consistent.

>      > - I am a heavy user of git rebase
>
>
> which does what?
>
>      > - being able to commit, branch, diff etc. without internet connection
>
>
> how often do you branch or diff when working on scores?
>
>      >   that way I can work on a train, for example
>
>
> hmm. i do a lot of work on trains. however, i have a data plan on my
> phone which i use for tethering if i need a connection.
>
>      > - this makes the commit often & early paradigm easy
>
>
> i understand this when working on software programs. but on music scores?

It's exactly the same :-)
Why would you "commit often and early" in software development? mostly 
to have fine-grained control on the content of a commit and not to have 
a big blob of unrelated changes in a unit. That's exactly the same when 
you write scores.

>     With svn/cvs you need more-or-less a permanent connection to
>     "the server". With git I can work off-line and update (both ways)
>     when connected.
>
>
> i only need the connection with SVN if i'm actually committing. with git
> i can commit offline, but then when i get a data connection i have to
> commit again to the repo? that's two steps compared to one with SVN. i
> still don't get it. especially as i generally commit when i reach
> certain milestones or at the end of a working day.

A Git commit isn't about integrating your work with the main line 
(remote repository) but about tailoring a coherent changeset. You may 
enclose your day's work in a commit but you can (as the other extreme) 
enclose a singly typo fix in a commit. With this you have finegrained 
control over your project history.
Each of these commits is a separate unit and step that can be inspected, 
reverted or modified individually later.
When you contact the server again you don't "commit" (in Git speech) but 
"push" your local commits to the server.
Obviously that's not what you're doing so far, so you don't miss it.
But I'm sure it is superior to be able to "tailor" the commit history 
locally instead of creating commits depending on the availability of an 
internet connection.

> don't get me wrong, i'm really trying to find a reason to like git. but
> nothing seems different to SVN so far, except for the fact that i have
> to commit TWICE...
>
> Urs also pointed out:
> On Tue, Jan 7, 2014 at 2:51 PM, Urs Liska <address@hidden
> <mailto:address@hidden>> wrote:
>
>     Ah, finally one idea about your question, not based on experience
>     but on randomly read statements: How can you use branches with SVN?
>     If it's correct that Git branches are conceptually different through
>     being so exceptionally light-weight then I'd think this _is_ an
>     advantage even for a single user.
>
>
> $ svn copy <repo>/trunk <repo>/branch/my-branch
>
> from http://svnbook.red-bean.com/en/1.8/svn.branchmerge.using.html :
> <<<
> Subversion's repository has a special design. When you copy a directory,
> you don't need to worry about the repository growing huge—Subversion
> doesn't actually duplicate any data. Instead, it creates a new directory
> entry that points to an existing tree. If you're an experienced Unix
> user, you'll recognize this as the same concept behind a hard link. As
> further changes are made to files and directories beneath the copied
> directory, Subversion continues to employ this hard link concept where
> it can. It duplicates data only when it is necessary to disambiguate
> different versions of objects.
>  >>>
>
> sounds fairly lightweight to me too... ;) (this is also the same
> workflow for tags).

Hm, don't know. As said, my comment was based on arbitrary reading.
Obviously (from skimming through that linked page) the branching concept 
between SVN and Git is _very_ different.
I can't really judge that, but as far as I've read this area is one of 
the major improvements of Git over the older VCSs.

Urs

> regards,
> sb
> --
> Do not meddle in the affairs of trombonists, for they are subtle and
> quick to anger.
>
>
> _______________________________________________
> lilypond-user mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/lilypond-user


--
Urs Liska
www.openlilylib.org