Re: [BUG] bisecting miscounts revisions left to test

From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Mar 17, 2007, 00:50 UTC
Message-ID: <Pine.LNX.4.63.0703170139110.22628@wbgn013.biozentrum.uni-wuerzburg.de>
In-Reply-To: <20070316161421.GA24584@lala>

Hi,

On Fri, 16 Mar 2007, Uwe Kleine-König wrote:

Show 7 quoted lines

> zeisberg@cassiopeia:~/gsrc/linux-2.6$ git bisect good
> Bisecting: 2 revisions left to test after this
> [e7b0d26a86943370c04d6833c6edba2a72a6e240] sysfs: reinstate exclusion between method calls and attribute unregistration
> 
> zeisberg@cassiopeia:~/gsrc/linux-2.6$ git bisect good
> Bisecting: 2 revisions left to test after this
> [b810cdfcf91d76f603fd48023aede48ced8e6bed] Merge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6

The problem is that after the first git-bisect good, it looks like this:

g1 - b2 - b1 - M - B
             /
     g2 - b3

where both g1 and g2 are good, and B is bad. The bisection point is b1. (Which makes sense, absolutely.)

After the second git-bisect good, it looks like this:

     g3 - M - B
        /
g2 - b3

and the bisection point is M.

Your problem, of course, is that it is not possible to pick a commit _exactly_ in the middle. So, the number is just an _expected_ number of revisions to test. I am not sure if it is necessary to make this clearer, though, as I fully expect that you just said "What the heck" and continued bisecting just fine.

Ciao, Dscho

P.S.: if Momo's turtle thinks that your name contains no non-ASCIIs, why should I?