# stgit bisect feature request

3 messages from 2006-09-05 to 2006-09-06. Participants: Torgil Svensson, Karl Hasselström, Catalin Marinas.
Thread: https://gitlist.dev/t/5492

## Torgil Svensson, 2006-09-05 23:49

Subject: stgit bisect feature request
Message-ID: <e7bda7770609051649j56a92085i8d8a73ab12004c43@mail.gmail.com>
URL: https://gitlist.dev/e/e7bda7770609051649j56a92085i8d8a73ab12004c43%40mail.gmail.com

```
I think i'd be nice if stgit wraps bisect and you could write
$ stg bisect start
$ stg bisect bad
$ stg bisect good ORIG_HEAD

and after bisect checks out the files it pushes the current stack as
if you had used stg pull at this point

//Torgil

```

## Karl Hasselström, 2006-09-06 08:31

Subject: Re: stgit bisect feature request
Message-ID: <20060906083153.GA19001@diana.vm.bytemark.co.uk>
URL: https://gitlist.dev/e/20060906083153.GA19001%40diana.vm.bytemark.co.uk
In-Reply-To: <e7bda7770609051649j56a92085i8d8a73ab12004c43@mail.gmail.com>

```
On 2006-09-06 01:49:15 +0200, Torgil Svensson wrote:

> I think i'd be nice if stgit wraps bisect and you could write
>
> $ stg bisect start
> $ stg bisect bad
> $ stg bisect good ORIG_HEAD
>
> and after bisect checks out the files it pushes the current stack as
> if you had used stg pull at this point

Yes, this sounds very useful for bisecting when you can't test the old
revisions as-is, e.g. because you have to patch for an unrelated bug,
or because you need a patch to trigger the problem.

-- 
Karl Hasselström, kha@treskal.com
      www.treskal.com/kalle

```

## Catalin Marinas, 2006-09-06 08:36

Subject: Re: stgit bisect feature request
Message-ID: <tnxfyf5qtmj.fsf@arm.com>
URL: https://gitlist.dev/e/tnxfyf5qtmj.fsf%40arm.com
In-Reply-To: <e7bda7770609051649j56a92085i8d8a73ab12004c43@mail.gmail.com>

```
"Torgil Svensson" <torgil.svensson@gmail.com> wrote:
> I think i'd be nice if stgit wraps bisect and you could write
> $ stg bisect start
> $ stg bisect bad
> $ stg bisect good ORIG_HEAD
>
> and after bisect checks out the files it pushes the current stack as
> if you had used stg pull at this point

It would be nice but there could be conflicts when pushing patches,
especially if you do the bisect over a larger range of commits. It
gets difficult if some StGIT patches rely on features not present some
of the commits you are checking.

Have you tried this idea manually (or with some scripts) and found it
feasible?

-- 
Catalin

```