Matthieu Moy <> writes:

> Ramkumar Ramachandra <> writes:
>> --- a/
>> +++ b/
>> @@ -44,6 +44,7 @@ merge_args= edit=
>>  curr_branch=$(git symbolic-ref -q HEAD)
>>  curr_branch_short="${curr_branch#refs/heads/}"
>>  rebase=$(git config --bool branch.$curr_branch_short.rebase)
>> +autostash=$(git config --bool rebase.autostash)
>>  if test -z "$rebase"
>>  then
>>      rebase=$(git config --bool pull.rebase)
>> @@ -203,6 +204,7 @@ test true = "$rebase" && {
>>                      die "$(gettext "updating an unborn branch with changes 
>> added to the index")"
>>              fi
>>      else
>> +            test true = "$autostash" ||
>>              require_clean_work_tree "pull with rebase" "Please commit or 
>> stash them."
> Trivial, indeed!
> It would be nice to have an --autostash command-line option too, and the
> error message in "require_clean_work_tree" could suggest using it. That
> would make the feature easily discoverable.

I would actually suggest doing this the other way around.  --autostash
option, followed by configuration if needed.

I actually have this slight suspicion that rebase.autostash and
pull.autostash may want to be separate variables.  Attempting to
rebase locally when your working tree is in half-baked state is bad
enough but may be warranted in some workflows, but doing so when
integrating with other people's work is a practice in a different
league. It would be quite sensible for somebody to want to allow the
former (i.e. allows autostash with rebase.autostash=true for local
rebase) while retaining the safety of the latter (i.e. stop pulling
with pull.autostash=false).

> Perhaps this patch could mention "pull --rebase" in the doc, like (config.txt)
> -     ends.  This means that you can run rebase on a dirty worktree.
> +     ends.  This means that you can run rebase or `git pull --rebase` on a 
> dirty worktree.
> (or perhaps it's obvious enough and not needed)
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to
More majordomo info at

Reply via email to