Re: [Pan-users] [Semi-OT] git question (was Re: Handy new feature...)

[Duncan]

Rhialto posted on Thu, 15 Sep 2011 15:22:03 +0200 as excerpted:

> On Thu 15 Sep 2011 at 08:02:41 -0500, Ron Johnson wrote:
>> $ rm -rf pan2 && git clone git://github.com/judgefudge/pan2
>> 
>> What's the command(s) for us non-developers who "just" want to get all
>> changes?  (Similar to "wget -m".)
> 
> This should be enough:
> 
> $ cd pan2 && git pull
> 
> (where pan2 is the directory where the ".git" directory lives)

Yes.

Expanding a bit[1], git has two repository types, "bare" and default.

In default mode, a clone (and further pulls) place a hidden .git subdir 
in the working dir.  This contains the sources pack-files and metadata 
necessary for history tracking, remote repo links, etc.  A pull both 
updates the sources from the remote, and does a checkout of the new 
current head sources into the working directory.

In bare mode, the pack-files and metadata that would normally be found in 
the .git subdir are placed directly in the target dir itself, and there 
is no working dir.  This mode is for for public repositories, etc, that 
are intended to be pulled/pushed from/to, but not actually worked in, 
directly.  The required space is far less since there's only the packed 
copy of the data, the expanded sources (the checkout, in git terms) are 
omitted.  But this mode requires an extra command line option to setup, 
while the default mode above is just that, default, "the way it works 
unless you tell git to do it otherwise", thus making it simpler for users 
who just want to pull sources from upstream, and periodically update and 
build them.

---
[1] Expanding a bit:  Heh, shouldn't surprise anyone on the list to see 
that in a "Duncan" post.  Perhaps the surprising bit would be if it's 
/just/ "a bit"! =:^)

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman