On Thu, Aug 18 2022, Michael J Gruber wrote:
> `error_resolve_conflict()` checks the untranslated action_name
> parameter, so pass it as is.
>
> Suggested-by: Johannes Schindelin <Johannes.Schindelin@xxxxxx>
> Signed-off-by: Michael J Gruber <git@xxxxxxxxx>
> ---
> sequencer.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/sequencer.c b/sequencer.c
> index 51d75dfbe1..8b32b239b9 100644
> --- a/sequencer.c
> +++ b/sequencer.c
> @@ -537,7 +537,7 @@ static struct tree *empty_tree(struct repository *r)
> static int error_dirty_index(struct repository *repo, struct replay_opts *opts)
> {
> if (repo_read_index_unmerged(repo))
> - return error_resolve_conflict(_(action_name(opts)));
> + return error_resolve_conflict(action_name(opts));
>
> error(_("your local changes would be overwritten by %s."),
> _(action_name(opts)));
> @@ -3753,7 +3753,7 @@ static int do_reset(struct repository *r,
> init_checkout_metadata(&unpack_tree_opts.meta, name, &oid, NULL);
>
> if (repo_read_index_unmerged(r)) {
> - ret = error_resolve_conflict(_(action_name(opts)));
> + ret = error_resolve_conflict(action_name(opts));
> goto cleanup;
> }
Perhaps we should have the error_resolve_conflict() function take a
"enum replay_action" instead? We could just do this more isolated
change, but perhaps that "while-we're-at-it" would be acceptable to
reduce the risk of running with this particular set of scissors.
Then we could note in a comment in that function that we do not want to
translate the string we'd get from action_name()...
