# bisect and gitk happy together

2 messages from 2006-06-12 to 2006-06-12. Participants: Martin Langhoff, Linus Torvalds.
Thread: https://gitlist.dev/t/4482

## Martin Langhoff, 2006-06-12 11:41

Subject: bisect and gitk happy together
Message-ID: <46a038f90606120441p74dd4872y441fe04470f4acd5@mail.gmail.com>
URL: https://gitlist.dev/e/46a038f90606120441p74dd4872y441fe04470f4acd5%40mail.gmail.com

```
I was using git-bisect earlier today, and at the exact point where it
told be about the bad commit, I opened gitk, which was showing all the
bad and good commits. It is great!

Two "user" notes, however:

 - git-bisect visualise wasn't as useful as just a plain gitk. (This
may be because I was working with ~60 commits in a medium-sized
project).

 - gitk didn't show the bad commit tagged specially, even if
git-bisect had just identified it. Of course I could find it, but I
had all the other good/bad commits well labelled. And not the one I
was looking for. Odd.

In any case, the bisect + gitk combo saved the day. I'm too ashamed to
tell what the bug actually was, though ;-)


martin

```

## Linus Torvalds, 2006-06-12 15:10

Subject: Re: bisect and gitk happy together
Message-ID: <Pine.LNX.4.64.0606120807350.5498@g5.osdl.org>
URL: https://gitlist.dev/e/Pine.LNX.4.64.0606120807350.5498%40g5.osdl.org
In-Reply-To: <46a038f90606120441p74dd4872y441fe04470f4acd5@mail.gmail.com>

```


On Mon, 12 Jun 2006, Martin Langhoff wrote:
> 
> - git-bisect visualise wasn't as useful as just a plain gitk. (This
> may be because I was working with ~60 commits in a medium-sized
> project).

Definitely. Try just firing up gitk when you're bisecting a kernel archive 
with thousands of commits, and complex history..

That's when "git bisect visualize" really helps: when git bisect has 
already narrowed down the list of commits from "5 years" to "1 week", but 
you still have maybe a hundred-odd commits to go.

I agree that just plain "gitk" is actually nicer if you want to see the 
whole context. It's just that often the context is pretty damn confusing ;)

> - gitk didn't show the bad commit tagged specially, even if
> git-bisect had just identified it. Of course I could find it, but I
> had all the other good/bad commits well labelled. And not the one I
> was looking for. Odd.

It should be the head of the "bisect" branch, and naturally tagged that 
way.

			Linus

```