Rubén Justo <rjusto@xxxxxxxxx> writes:
> Ensure strict matching of context lines when applying with
> `--whitespace=fix` combined with `--no-ignore-whitespace`.
>
> Signed-off-by: Rubén Justo <rjusto@xxxxxxxxx>
> ---
> apply.c | 3 ++-
> t/t4124-apply-ws-rule.sh | 27 +++++++++++++++++++++++++++
> 2 files changed, 29 insertions(+), 1 deletion(-)
>
> diff --git a/apply.c b/apply.c
> index 63e58086f1..0cb9d38e5a 100644
> --- a/apply.c
> +++ b/apply.c
> @@ -2596,7 +2596,8 @@ static int match_fragment(struct apply_state *state,
> goto out;
> }
>
> - if (state->ws_error_action != correct_ws_error) {
> + if (state->ws_error_action != correct_ws_error ||
> + state->ws_ignore_action == ignore_ws_none) {
> ret = 0;
> goto out;
> }
Hmph, if we are correcting for whitespace violations, even if
whitespace fuzz is not allowed externally, wouldn't the issue that
c1beba5b (git-apply --whitespace=fix: fix whitespace fuzz introduced
by previous run, 2008-01-30) corrected still apply? IOW, isn't this
change introducing a regression when an input touches a file with a
change with broken whitespaces, and then touches the same file to
replace the broken whitespace lines with something else?
