Bert Wesarg <bert.wesarg@xxxxxxxxxxxxxx> writes:
> When 46af44b07d (pull --rebase=<type>: allow single-letter abbreviations
> for the type, 2018-08-04) landed in Git, it had the side effect that
> not only 'pull --rebase=<type>' accepted the single-letter abbreviations
> but also the 'pull.rebase' and 'branch.<name>.rebase' configurations.
>
> Secondly, 'git remote rename' did not honor these single-letter
> abbreviations when reading the 'branch.*.rebase' configurations.
Hmph, do you mean s/Secondly/However/ instead?
> The only functional change is the handling of the `branch_info::rebase`
> value. Before it was an unsigned enum, thus the truth value could be
> checked with `branch_info::rebase != 0`. But `enum rebase_type` is
> signed, thus the truth value must now be checked with
> `branch_info::rebase >= REBASE_TRUE`.
I think there is another hidden one, but I do not know offhand the
implications of the change. It could well be benign.
> /**
> * Parses the value of --rebase. If value is a false value, returns
> * REBASE_FALSE. If value is a true value, returns REBASE_TRUE. If value is
> @@ -45,22 +37,9 @@ enum rebase_type {
> static enum rebase_type parse_config_rebase(const char *key, const char *value,
> int fatal)
> {
> - int v = git_parse_maybe_bool(value);
> -
> - if (!v)
> - return REBASE_FALSE;
> - else if (v > 0)
> - return REBASE_TRUE;
> - else if (!strcmp(value, "preserve") || !strcmp(value, "p"))
> - return REBASE_PRESERVE;
> - else if (!strcmp(value, "merges") || !strcmp(value, "m"))
> - return REBASE_MERGES;
> - else if (!strcmp(value, "interactive") || !strcmp(value, "i"))
> - return REBASE_INTERACTIVE;
> - /*
> - * Please update _git_config() in git-completion.bash when you
> - * add new rebase modes.
> - */
I see all of the above, including the "Please update" comment, has
become rebase_parse_value(), which is very good.
> diff --git a/builtin/remote.c b/builtin/remote.c
> index 96bbe828fe..2830c4ab33 100644
> --- a/builtin/remote.c
> +++ b/builtin/remote.c
> @@ -6,6 +6,7 @@
> ...
> - enum {
> - NO_REBASE, NORMAL_REBASE, INTERACTIVE_REBASE, REBASE_MERGES
> - } rebase;
> + enum rebase_type rebase;
Good to see the duplicate go.
> @@ -305,17 +304,8 @@ static int config_read_branches(const char *key, const char *value, void *cb)
> space = strchr(value, ' ');
> }
> string_list_append(&info->merge, xstrdup(value));
> - } else {
> - int v = git_parse_maybe_bool(value);
> - if (v >= 0)
> - info->rebase = v;
> - else if (!strcmp(value, "preserve"))
> - info->rebase = NORMAL_REBASE;
> - else if (!strcmp(value, "merges"))
> - info->rebase = REBASE_MERGES;
> - else if (!strcmp(value, "interactive"))
> - info->rebase = INTERACTIVE_REBASE;
> - }
> + } else
> + info->rebase = rebase_parse_value(value);
Here, we never had info->rebase == REBASE_INVALID. The field was
left intact when the configuration file had a rebase type that is
not known to this version of git. Now it has become possible that
info->rebase to be REBASE_INVALID. Would the code after this part
returns be prepared to handle it, and if so how? At least I think
it deserves a comment here, or in rebase_parse_value(), to say (1)
that unknown rebase value is treated as false for most of the code
that do not need to differentiate between false and unknown, and (2)
that assigning a negative value to REBASE_INVALID and always
checking if the value is the same or greater than REBASE_TRUE helps
to maintain the convention.
> diff --git a/rebase.h b/rebase.h
> new file mode 100644
> index 0000000000..cc723d4748
> --- /dev/null
> +++ b/rebase.h
> @@ -0,0 +1,15 @@
> +#ifndef REBASE_H
> +#define REBASE_H
> +
> +enum rebase_type {
> + REBASE_INVALID = -1,
> + REBASE_FALSE = 0,
> + REBASE_TRUE,
> + REBASE_PRESERVE,
> + REBASE_MERGES,
> + REBASE_INTERACTIVE
> +};
> +
> +enum rebase_type rebase_parse_value(const char *value);
> +
> +#endif /* REBASE */
