[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: git bisect; finding exact commits of bugs
From: |
Patrick McCarty |
Subject: |
Re: git bisect; finding exact commits of bugs |
Date: |
Fri, 27 Nov 2009 21:54:49 -0800 |
On Fri, Nov 27, 2009 at 9:27 PM, Graham Percival
<address@hidden> wrote:
> If it doesn't exist already, is it worth writing something for the
> CG about using git bisect to find the exact commit where a
> regression was introduced?
There doesn't appear to be anything in the CG about `git bisect'. The
man page for this command appears to have an easy-to-understand (for
me) example, under the heading "Basic bisect commands: start, bad,
good". But maybe we should write an even simpler example?
> If it does exist, would it be worth trying to convince a frog or
> two to spend time doing this? I heartily admit that it's not
> glamorous work, but if any regression issues could have the commit
> identified, I imagine it would be one less task for whoever's
> working on a patch for said regression.
Sure, I think it would helpful to find the problematic commits,
especially for regressions.
For #907, I only suspected three different commits, and was lucky that
I found the exact commit by reverting each in turn.
-Patrick