when does git start caring about symlinks?

> I have a question on how symlinks work and at what point git starts caring
> about them.  If dirA/repoA/.git has no symlinks and I copy it to
> dirB/repoB/.git (ie. cp -rp /dirA/repoA/.git /dirB/repoB/.git), but /dirB is
> a symlink to /x/dirB does that mean that repoB contains symlinks (I suspect
> that repoB may be made up of all symlinks at this point)?  In other words,
> if I parallel test repoA and repoB am I running a true parallel test or are
> the repos different because repoA has no symlinks and repoB has symlinks?

>"Nguyen Thai Ngoc Duy" <pclouds@gmail.com> wrote in message 
> >news:AANLkTi=O0pZ97kRt0jGfy20znfvfp3UTydTBn_aMBxE+@mail.gmail.com...
>On Wed, Nov 24, 2010 at 6:50 AM, Neal Kreitzinger <neal@rsss.com> wrote:
>> I have a question on how symlinks work and at what point git starts 
>> caring
>> about them. If dirA/repoA/.git has no symlinks and I copy it to
>> dirB/repoB/.git (ie. cp -rp /dirA/repoA/.git /dirB/repoB/.git), but /dirB 
>> is
>> a symlink to /x/dirB does that mean that repoB contains symlinks (I 
>> suspect
>> that repoB may be made up of all symlinks at this point)? In other words,
>> if I parallel test repoA and repoB am I running a true parallel test or 
>> are
>> the repos different because repoA has no symlinks and repoB has symlinks?
>
>There should be no difference between repoA and repoB until you make
>changes. Symlinks outside worktree do not matter. Symlinks inside .git
>dir may cause problems when you start updating repos. But I don't
>think recent git creates symlinks. There are other forms of symlinks
>in .git dir though: .git as a file that points to real .git dir, or
>.git/info/alternates comes to mind.
>-- 
>Duy