Re: Git gives you plenty of rope ...

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, Sep 10, 2007 at 11:08:00PM +0100, Johannes Schindelin wrote:
> From: Johannes Schindelin <Johannes.Schindelin@xxxxxx>
> Date: Mon, 10 Sep 2007 23:08:00 +0100 (BST)
> To: Martin Langhoff <martin.langhoff@xxxxxxxxx>
> cc: Ralf Baechle <ralf@xxxxxxxxxxxxxx>, git@xxxxxxxxxxxxxxx
> Subject: Re: Git gives you plenty of rope ...
> Content-Type: TEXT/PLAIN; charset=US-ASCII
> 
> Hi,
> 
> On Tue, 11 Sep 2007, Martin Langhoff wrote:
> 
> > On 9/11/07, Johannes Schindelin <Johannes.Schindelin@xxxxxx> wrote:
> > > On Mon, 10 Sep 2007, Ralf Baechle wrote:
> > >
> > > >   1) In a repository have .git/objects/info/alternates point back to 
> > > >      the repository's object directory.
> > >
> > > What exactly did you think it would do?
> > 
> > Recently, I have been through several machines at work adding a 
> > canonical "/var/cache/git" under which we add bare repos that track the 
> > "master" repos. After that, I add the paths to those repos to alternates 
> > of any git checkout in the machine, and git repack -a -d all of them.
> 
> That actually makes some kind of sense, as a use case which triggers that 
> bug.
> 
> > Both git gc (as an end-user script) and us -- we all need to be a bit 
> > more friendly to newbies. ;-)
> 
> Well, I am opposed to writing "Attention: the contents of this cup are 
> hot, because after all you would complain if your coffee was not hot" on 
> it; Give Darwin a chance, will ya?
> 
> However, I can see that git-repack (and just that, maybe even only for 
> some options) does the (potentially expensive) check if alternates points 
> to itself (you could use make_absolute_path() to normalise it, and the 
> result of getcwd(), and then compare them).

There also is the case of chaining alternates like:

repo1 -> repo2 -> repo3

Afair the git documentation doesn't explicitly allow/forbid this sort of
configuration.  But if yes, there would also be larger loops such as
repo1 -> repo2 -> repo3 -> repo2.

  Ralf
-
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]

  Powered by Linux