git-diff & cg-diff behavior difference

git-diff & cg-diff behavior difference

[Vincent Legoll]

I did not find something relevant in the docs, so I'll ask here.

git-diff works from everywhere in the tree, whereas cg-diff
does not, is that intentional ?

steps to reproduce:
--------------------
mkdir tmp
cd tmp
touch titi.txt
mkdir tutu
touch tutu/toto.txt
cg-init -I
cg-add -r .
cg-commit -C -m INIT
echo "tata" > titi.txt
git-diff
cg-diff
cd tutu
git-diff
cg-diff
--------------------

The 2 git-diff & 1st cg-diff outputs are the same, whereas the 2nd
cg-diff one is empty...

Quite a minor thing, but this could be added in the cg-diff man-page as a
difference of behavior.

-- 

Re: git-diff & cg-diff behavior difference

[Vincent Legoll]

On 12/18/06, Martin Langhoff <martin.langhoff@gmail.com> wrote:
> I'd say that it's a bug that cg-diff doesn't work in a subdir.
> However, there is a much more important difference between cg-diff and
> git-diff: cg-diff is equivalent to git-diff HEAD because cogito hides
> the index from you.

That's what I wanted to know: if it was a deliberate behavior, a bug on
my side, or on git's...

> If you are developing in a (mostly) straight line, and not dealing
> much with merges and rebases, that's a great simplification. If you
> are coming from a different SCM it's probably a good thing too.

That's me and me !

> If/when you want to use git's smarts for merging and rebasing, you'll
> definitely want to know a lot more about the index, and you'll
> probably stop using cogito.

That will be me some times...

> (Actually, even though I know and love the index I use cogito all the
> time, and switch to git commands for merging and rebasing. Works
> great... and yet Linus is in a merge-centric role, so changing the
> defaults to ignore the index most of the time is a bit of a hot issue

That's no problem with me, if behavior is properly documented...

-- 

Re: git-diff & cg-diff behavior difference

[Vincent Legoll]

On 12/19/06, Vincent Legoll <vincent.legoll@gmail.com> wrote:
> That's what I wanted to know: if it was a deliberate behavior, a bug on
> my side, or on git's...

Errr, I meant cogito's...

--