This is an attempt at using a reachability test to replace the use of
orig-base, to decide when it is safe to accept a rebase.
This implementation passes the testcase posted earlier by Karl (resent
here as 2nd patch in this series), BUT fails the 4th test of t2100.
That is, it fails to deal with the case of a rewinding commit occuring
in the upstream branch, and being first git-fetch'd before being
stg-pull'd in fetch-rebase mode. In this case, the former upstream
commit will really be lost.
In this case, however, it is exactly what we want, but I'm still
undecided about how to deal with this best. Possibly insane ideas for
now include:
- parsing the remote.*.fetch lines to detect the leading + (just
kidding ;)
- using ORIG_HEAD, but then, how do we decide when it is valid to do
so ?
I'll now switch to some real-life activity to see if that helps to
find a better solution.
--
Yann Dirson <ydirson@xxxxxxxxxx> |
Debian-related: <dirson@xxxxxxxxxx> | Support Debian GNU/Linux:
| Freedom, Power, Stability, Gratis
http://ydirson.free.fr/ | Check <http://www.debian.org/>
-
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
