Re: [RFC] git blame-tree

From: Piotr Krukowiecki <piotr.krukowiecki@gmail.com>
Date: Mar 2, 2011, 18:39 UTC
Message-ID: <AANLkTi=m_WTohMfJZxTqObRT3rhhtxo=QfnDJCHO=U0K@mail.gmail.com>
In-Reply-To: <20110302180722.GA20287@sigill.intra.peff.net>

On Wed, Mar 2, 2011 at 7:07 PM, Jeff King <peff@peff.net> wrote:

Show 22 quoted lines

> On Wed, Mar 02, 2011 at 06:51:57PM +0100, Piotr Krukowiecki wrote:
>
>> On Wed, Mar 2, 2011 at 6:16 PM, Jeff King <peff@peff.net> wrote:
>> > I considered making it a special mode of "git blame" when blame is fed a
>> > directory instead of a file. But the implementations aren't shared at
>> > all (nor do I think they need to be; blame-tree is _way_ simpler). And I
>>
>> git blame dir/file.c
>>   "Show what revision and author last modified each line of a file"
>>
>> git blame dir/
>>   "Show what revision and author last modified each file"
>
> Right, I think we are agreeing.
>
>> This makes sense to me (the user).  I don't understand the
>> implementation thing. I don't see a difference between those two
>> commands. Even more, if I'm educated  Unix user I might know
>> directories are also files.
>
> I mean the implementations are very different, so there was not much
> point in putting the code into builtin/blame.c.

Ah, ok.

Show 14 quoted lines

>> > didn't want to steal that concept in case somebody can think of a more
>> > content-level way of blaming a whole tree that makes sense (obviously
>> > just showing the concatenation of the blames of each file is one way,
>> > but I don't know how useful that would be). If we want to go that way,
>> > we can always catch the special case in blame and just exec blame-tree.
>>
>> Still can be in git-blame command, no?
>
> Right. What I meant was that we don't have to make the decision now. If
> people like blame-tree, we can later magically turn:
>
>  git blame dir
>
> into "git blame-tree dir". So I think we are just agreeing.

I hope nobody likes "blame-dir" :)

Show 18 quoted lines

>> > The initial set of interesting files we come up with is gotten by
>> > looking at the tree of the first pending object after parsing the rev
>> > options (defaulting to HEAD). Which sounds a little flaky to me, but
>> > does what you want in practice. I'd be curious if somebody can come up
>> > with a counterexample where the ability to manually specify the source
>> > tree would be more useful.
>>
>> Same argument as for normal blame: I want to know who modified files at
>> the state of commit X (if I understand the question correctly).
>
> Yeah, that's what it does now. Specifically I was wondering about more
> elaborate examples, like:
>
>  git blame-tree dir branch1 branch2
>
> It will traverse using both branch1 and branch2, but get the initial
> list of files from branch1. I guess we could also union those trees or
> something.

I'd expect this to be something like union. Currently I can only think about following case:

Some files were changed in branch1, some in branch2, some in both.
Show me how the files are changed. For example:
  file1 changed in branch1 in commit1
  file2 changed in branch2 in commit2
  file3 changed in branch1 in commit3 and in branch2 in commit4

If file was not changed since branch creation then don't show it (optionally).

But maybe this is more like a diff or log than a blame. Maybe there's already such mode - I could not find it.

$ git init
Initialized empty Git repository in /tmp/a/.git/
$ echo a > a
$ echo b > b
$ echo c > c
$ git add .
$ git commit -a -m new
[master (root-commit) af5d319] new
 3 files changed, 3 insertions(+), 0 deletions(-)
 create mode 100644 a
 create mode 100644 b
 create mode 100644 c
$ git branch branch1
$ echo trunk1 > a
$ git commit -a -m trunk1
[master 2dc7f47] trunk1
 1 files changed, 1 insertions(+), 1 deletions(-)
$ echo trunk2 > b
$ git commit -a -m trunk1
[master 736fcd2] trunk1
 1 files changed, 1 insertions(+), 1 deletions(-)
$ git checkout branch1
Switched to branch 'branch1'
$ echo branch1 > c
$ git commit -a -m branch1
[branch1 52e371d] branch1
 1 files changed, 1 insertions(+), 1 deletions(-)
$ echo branch2 > b
$ git commit -a -m branch2
[branch1 9fed07c] branch2
 1 files changed, 1 insertions(+), 1 deletions(-)
$ git diff --stat branch1 master
 a |    2 +-
 b |    2 +-
 c |    2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)

I would like to see output like this:
 a   2dc7f47  (master)
 b   736fcd2  (master)
 b   9fed07c  (branch1)
 c   52e371d  (branch1)

Not sure how useful it would be. Just an idea.

Show 5 quoted lines

> But I expect most calls to be:
>
>  git blame-tree dir commit
>
> and that's it.

Me too.

-- 
Piotrek