Matthieu Moy <Matthieu.Moy@xxxxxxxxxxxxxxx> writes:
>> Confirmed: Git 2.1.4 accepts this, 2.6 doesn't:
>>
>> Warning: the command isn't recognized in the following line:
>> - # pick dbafac11052a0075233bdcf0b71f54d1503aa82d test
>>
>> You can fix this with 'git rebase --edit-todo'.
>> Or you can abort the rebase with 'git rebase --abort'.
>>
>> I didn't bisect, but I guess this was introduced in the series
>> introducing this check on the todolist before starting the bisection.
>
> Indeed:
>
> 804098bb30a5339cccb0be981a3e876245aa0ae5 is the first bad commit
Yup, before that series, expand_todo_ids -> transfom_todo_ids ended
up reading each line with "while read -r command rest" loop and the
we did not honor the usual "# at the beginning line is the comment"
convention, which I think was a bug. With that commit, a separate
step in check_bad_cmd_and_sha1 uses a similar looking "while read"
loop but forgets to take '#' into account.
I know you alluded to preprocess what is fed to stripspace, but I
wonder if we can remove the misguided call to stripspace in the
first place and do something like the attached instead.
git-rebase--interactive.sh | 3 +--
1 file changed, 1 insertion(+), 2 deletions(-)
diff --git a/git-rebase--interactive.sh b/git-rebase--interactive.sh
index f01637b..a64f77a 100644
--- a/git-rebase--interactive.sh
+++ b/git-rebase--interactive.sh
@@ -886,7 +886,6 @@ check_commit_sha () {
# from the todolist in stdin
check_bad_cmd_and_sha () {
retval=0
- git stripspace --strip-comments |
(
while read -r line
do
@@ -896,7 +895,7 @@ check_bad_cmd_and_sha () {
sha1=$2
case $command in
- ''|noop|x|"exec")
+ '#'*|''|noop|x|"exec")
# Doesn't expect a SHA-1
;;
pick|p|drop|d|reword|r|edit|e|squash|s|fixup|f)
--
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