Mercurial > public > mercurial-scm > hg
diff tests/test-split.t @ 41560:66399f2e92aa
commit: if interactive, look elsewhere for whitespace settings (BC)
Previously, when doing `commit -i`, we respected `diff.ignorews` and other
whitespace-related settings, which is probably unexpected. The primary reason
for this is to support hgext.record's commandline options, it's probably
accidental that the `[diff]` settings were also considered. See comments on
issue6042 and D5490. This can cause problems (issue5839, issue6042).
It is assumed by the author that the `[diff]` section is primarily for *viewing*
diffs, and that it is unlikely what people intend when attempting to commit or
revert.
With this change, if a user wants the behavior, they can clone their `[diff]`
settings to `commands.commit.interactive.<setting>`. This is thus a mild BC
change, but one I suspect is not going to be relied on by anyone.
Note: while doing a partial commit/revert, we do not know what command the user
is actually running. This means that the split extension, which ends up calling
into this code, will respect the `commands.commit.interactive.<setting>`
settings, and not a hypothetical `commands.split.interactive.<setting>`. This
*also* means that setting `commands.commit.interactive.ignoreblanklines`, for
example, will still cause issue5839. Considering the highly unlikely chance that
a user actually sets `commands.commit.interactive.<setting>`, the author deems
this risk acceptable.
Differential Revision: https://phab.mercurial-scm.org/D5834
author | Kyle Lippincott <spectral@google.com> |
---|---|
date | Thu, 31 Jan 2019 14:47:34 -0800 |
parents | 3a01ce246ece |
children | 7da6307cc07a |
line wrap: on
line diff
--- a/tests/test-split.t Thu Jan 31 14:29:24 2019 -0800 +++ b/tests/test-split.t Thu Jan 31 14:47:34 2019 -0800 @@ -26,6 +26,8 @@ > [diff] > git=1 > unified=0 + > [commands] + > commit.interactive.unified=0 > [alias] > glog=log -G -T '{rev}:{node|short} {desc} {bookmarks}\n' > EOF @@ -105,10 +107,10 @@ This function splits a bit strangely primarily to avoid changing the behavior of the test after a bug was fixed with how split/commit --interactive handled -`diff.unified=0`: when there were no context lines, it kept only the last diff -hunk. When running split, this meant that runsplit was always recording three commits, -one for each diff hunk, in reverse order (the base commit was the last diff hunk -in the file). +`commands.commit.interactive.unified=0`: when there were no context lines, +it kept only the last diff hunk. When running split, this meant that runsplit +was always recording three commits, one for each diff hunk, in reverse order +(the base commit was the last diff hunk in the file). $ runsplit() { > cat > $TESTTMP/messages <<EOF > split 1