René Scharfe <l.s.r@xxxxxx> writes: > Am 24.09.21 um 21:24 schrieb Junio C Hamano: >> Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> writes: >> >>> Since b92891f9783 (parseopt: add PARSE_OPT_NO_INTERNAL_HELP, >>> 2009-03-08) parse_options() has handled "-h" unless told not to, so >>> when show-ref was migrated to parse_options() in >>> 69932bc6117 (show-ref: migrate to parse-options, 2009-06-20) the >>> custom "-h" handling that was retained did nothing. >>> >>> The option was then hidden in e62b3935056 (Show usage string for 'git >>> show-ref -h', 2009-11-09), but that OPT_BOOLEAN didn't do >>> anything. Let's just remove this dead code. >>> >>> Reported-by: Ignacy Gawedzki <ignacy.gawedzki@xxxxxxxxxxxxxxxxxxxxxxx> >>> Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> >>> --- >>> builtin/show-ref.c | 2 -- >>> 1 file changed, 2 deletions(-) >> >> While this is a fine no-op, I am not sure this "fixes" complaint >> in Ignacy's report. "git show-ref -h" would (and should) show the >> short-help, no? >> > > It would, but -h is not a no-op without this patch. The option is > equivalent to --heads as long as it's not the only argument. E.g. it > has an effect in "git show-ref -h HEA." or "git show-ref -hh". Ah, so this actively breaks the command? Yeah, thanks---I smelled something fishy in the change.
