Re: Missing changes in merges from emacs-25 to master

[martin rudalics <rudalics@gmx.at>]

[...]
 > which in turn is a parent of:
 >
 > commit 1b76d9168336ede8976b980aeaed64ae2908501a
 > Merge: bca4930 2e5a89f
 > Author: John Wiegley <johnw@newartisans.com>
 > Date:   Mon Jan 18 22:59:51 2016 -0800
 >
 >      -
 >
 > which in turn is a parent of:
 >
 > commit b895c72059521fec064ff27b4cfcfa4104081c4e
 > Merge: badcd38 1b76d91
 > Author: Vincent Belaïche <vincentb1@users.sourceforge.net>
 > Date:   Wed Jan 20 08:30:51 2016 +0100
 >
 >      Merge branch 'master' of git.sv.gnu.org:/srv/git/emacs
 >
 > which leads via a long thin graph to the merge commit
 > 82b089783e71b2aeef950eaecfe4cbc0735e64a2 that you identified as losing
 > the change.

Thanks for the forensics.  I have no idea which tools to use to trace
such a path.  One thing that stupefies me is that in

commit fbce4757a874cc43806eb41b8637538b101c3c69
Author: Alan Mackenzie <acm@muc.de>
Date:   Sun Jan 24 20:30:39 2016 +0000

which IIUC follows the commits you cited above, the "correct" version is
still there why it should have been already lost at least on January 20.

martin