Remove redundant git commands.

Hi,
in #1511112: git_deploy slows down module install/enable massively we are trying to make this module faster.
Besides adding caches, one way to make it faster would be to optimize the git commands.
However, I currently am a bit confused by the purpose of the commands.
Could you shed some light?

        exec("$git rev-parse --abbrev-ref HEAD 2>&1", $branch);
        $tag_found = FALSE;
        if ($branch) {
          $branch = $branch[0];

This produces $branch === 'HEAD' in case we are not on a branch. So the if ($branch) is always true.

          // Now try to find a tag.
          exec("$git rev-list --topo-order --max-count=1 HEAD 2>&1", $last_tag_hash);

Doesn't this simply give us the hash of HEAD? It does not really care which tag.

            exec("$git describe  --tags $last_tag_hash[0] 2>&1", $last_tag);

So why not simply "$git describe --tags HEAD?

Also, why not use the --match <pattern> parameter of git describe?

And what if the first tag found by describe is something custom, which prevents a correct version resolution?

Maybe something like this?
git log --topo-order --simplify-by-decoration --format=format:%d HEAD
The output is like this:

 (tag: 7.x-2.x-1511112-53-combined)
 (tag: 7.x-2.x-1511112-50-static-cache-per-git-dir)
 (HEAD, tag: xxxx, origin/HEAD, origin/7.x-2.x, 7.x-2.x)
 (tag: 7.x-2.2)
 (tag: 7.x-2.1)
 (tag: 7.x-2.0)

The goal would be to use as few exec() as possible, but also not let one exec() become too expensive.

I post this as a support request. The goal of this issue is to understand the design choices for the existing code. We will see where it goes from there.