Hi Thomas, If I understand correctly, you want to save the 'M's for later and get back to a clean master. There's a couple of things you can do.
You could do 'git stash' to squirrel away the changes until a later 'git stash pop'. The problem with this is, you may forget to pop. You could also commit those changes to another branch and revisit the branch later (git checkout -b another-branch && git commit --all). Even if the changes are unfinished, you can always refactor the commit history at a later date. If you don't care about the changes, you can totally discard them with 'git reset --hard branch-name'. Once you have dealt with your changes, you can make a new branch with 'git checkout -b branch-name' and the new branch will diverge from where you were previously. Hope that helps! -- Edd Barrett On 27 Dec 2018, 08:56, at 08:56, Thomas Schmitt <scdbac...@gmx.net> wrote: >Hi, > >i tried to create a branch for my fix in lib/driver/gnu_linux.c but >now "git diff" shows unintended deviations: > > $ git checkout master > M include/cdio/types.h > M lib/iso9660/iso9660_private.h > Switched to branch 'master' > Your branch is up-to-date with 'origin/master'. > > $ git pull > ... > 15 files changed, 334 insertions(+), 195 deletions(-) > create mode 100644 test/driver/cdtext.c > > $ git checkout -b trackno-higher-one > M include/cdio/types.h > M lib/iso9660/iso9660_private.h > Switched to a new branch 'trackno-higher-one' > >The "M" action is unintended. I want the current original state of >"master" as foundation of a new branch "trackno-higher-one". > >The diff of types.h and iso9660_private.h seems to be about an >experiment >to silence "riddling error messages which appear if <stdbool.h> is >included >after <cdio/iso9660.h>": > From: "Thomas Schmitt" <scdbac...@gmx.net> > To: libcdio-devel@gnu.org > Date: Sun, 15 Jul 2018 11:54:53 +0200 > Message-Id: <32030804139729892...@scdbackup.webframe.org> > > >What would a git expert do now ? > >I could revert the pending changes and thus lose the obviously >uncommitted >experiment. But how would i preserve such an experiment from a >different >branch while i want to make a change to the remote git repo ? > >Can i simply keep the changed files, commit only my change to file >gnu_linux.c, and push my new branch "trackno-higher-one" without >interference by those changes ? > > >Have a nice day :) > >Thomas