Hi Johannes,
On Mon, 09 Jul 2018 at 10:52:13 +0300, Johannes Schindelin wrote:
Hi Brian,
On Sun, 8 Jul 2018, brian m. carlson wrote:
On Sun, Jul 08, 2018 at 09:41:11PM +0300, Daniel Harding wrote:
Signed-off-by: Daniel Harding <dharding@xxxxxxxxxxxxx>
I think maybe, as you suggested, a separate test for this would be
beneficial. It might be as simple as modifying 'script-from-scratch' by
doing "sed 's/#/>/'".
It might be even simpler if you come up with a new "fake editor" to merely
copy the todo list, then run a rebase without overridden
commentChar, then one with overridden commentChar, then pipe the todo list
of the first through that `sed` call:
write_script copy-todo-list.sh <<-\EOF &&
cp "$1" todo-list.copy
EOF
test_config sequence.editor \""$PWD"/copy-todo-list.sh\" &&
git rebase -r <base> &&
sed "s/#/%/" <todo-list.copy >expect &&
test_config core.commentChar % &&
git rebase -r <base> &&
test_cmp expect todo-list.copy
Indeed, as I thought about it more, using a "no-op" todo editor seemed
like a good approach. Thanks for giving me a head start - I'll play
with that and try to get a new patch with an improved test posted in the
next couple of days.
One question about my original patch - there I had replaced a "grep -v"
call with a "git stripspace" call in the 'generate correct todo list'
test. Is relying on "git stripspace" in a test acceptable, or should
external text manipulation tools like grep, sed etc. be preferred?
Thanks,
Daniel Harding
