Hi Michael, sorry for the delay. On Sat, Dec 24, 2016 at 03:17:57PM +0100, Guido Günther wrote: > Hi Michael, > On Fri, Dec 16, 2016 at 04:10:36PM +0100, Michael Biebl wrote: > > Package: git-buildpackage > > Version: 0.8.8 > > Severity: normal > > > > Hi, > > > > today I ran "gbp pq export" in our Debian systemd git repository. > > This generated quite a bit of churn, in the form of > > > > -index f2d8bf5..7f2d731 100644 > > +index f2d8bf57f..7f2d73199 100644 > > That's bad. > > > > > > > This seems to be a result of using git 2.11.0 [1]: > > > > The default abbreviation length, which has historically been 7, now > > scales as the repository grows, using the approximate number of > > objects in the repository and a bit of math around the birthday > > paradox. The logic suggests to use 12 hexdigits for the Linux > > kernel, and 9 to 10 for Git itself. > > > > I wonder, whether gbp pq should setup a default core.abbrev=12 i.e. > > something which is reasonably large and constant. > > That would be a problem for people that use the same repo for upstream > work. It would be best to have gbp pq set the options on the command > line of the invoked git commands but --abbrev=7 doesn't do that while > > git config core.abbrev 7 > > in the repo works. I'll have to take a closer look.
Using 'git -c core.abbrev=7 diff …'. works. I have this hard coded at the moment to not introduce even more config options. Should you guys need this configuratble let me know. I've uploaded a version to experimental since I first want 0.8.10 to migrate to testing. Cheers, -- Guido