From: "Ævar Arnfjörð Bjarmason" <avarab@gmail.com>
To: "SZEDER Gábor" <szeder.dev@gmail.com>
Cc: git@vger.kernel.org, Junio C Hamano <gitster@pobox.com>
Subject: Re: [PATCH v2 09/20] parse-options: add support for parsing subcommands
Date: Fri, 19 Aug 2022 19:33:46 +0200 [thread overview]
Message-ID: <220819.86bksg6s5l.gmgdl@evledraar.gmail.com> (raw)
In-Reply-To: <20220819160411.1791200-10-szeder.dev@gmail.com>
On Fri, Aug 19 2022, SZEDER Gábor wrote:
> +static enum parse_opt_result parse_subcommand(const char *arg,
> + const struct option *options)
> +{
> + for (; options->type != OPTION_END; options++)
> + if (options->type == OPTION_SUBCOMMAND &&
> + !strcmp(options->long_name, arg)) {
> + *(parse_opt_subcommand_fn **)options->value = options->subcommand_fn;
Nit: Maybe do the cast by assigning to an intermediate variable? Would
make this less dense, and since we already have {}-braces...
> + return PARSE_OPT_SUBCOMMAND;
> + }
> +
> + return PARSE_OPT_UNKNOWN;
> +}
> +
> static void check_typos(const char *arg, const struct option *options)
> {
> if (strlen(arg) < 3)
> @@ -442,6 +457,7 @@ static void check_typos(const char *arg, const struct option *options)
> static void parse_options_check(const struct option *opts)
> {
> char short_opts[128];
> + void *subcommand_value = NULL;
>
> memset(short_opts, '\0', sizeof(short_opts));
> for (; opts->type != OPTION_END; opts++) {
> @@ -489,6 +505,14 @@ static void parse_options_check(const struct option *opts)
> "Are you using parse_options_step() directly?\n"
> "That case is not supported yet.");
> break;
> + case OPTION_SUBCOMMAND:
> + if (!opts->value || !opts->subcommand_fn)
> + optbug(opts, "OPTION_SUBCOMMAND needs a value and a subcommand function");
Better split into two IMO:
if (!opts->value)
optbug(opts, "OPTION_SUBCOMMAND needs a value");
if (!opts->subcommand_fn)
optbug(opts, "OPTION_SUBCOMMAND needs a subcommand_fn");
Then if we have extra checks we don't need to keep adding to a dense ||
and amend an existing string.
> +static int has_subcommands(const struct option *options)
> +{
> + for (; options->type != OPTION_END; options++)
> + if (options->type == OPTION_SUBCOMMAND)
> + return 1;
> + return 0;
> +}
I wondered if parse_options_check() couldn't take a "int
*has_subcommands", since we already loop through the options to check it
as part of the checks there (but we'd need to re-arrange them).
Not for optimization, just wondering if it would make the code flow
clearer, maybe not.
> + if (ctx->has_subcommands) {
> + if (flags & PARSE_OPT_STOP_AT_NON_OPTION)
> + BUG("subcommands are incompatible with PARSE_OPT_STOP_AT_NON_OPTION");
> + if (!(flags & PARSE_OPT_SUBCOMMAND_OPTIONAL)) {
> + if (flags & PARSE_OPT_KEEP_UNKNOWN_OPT)
> + BUG("subcommands are incompatible with PARSE_OPT_KEEP_UNKNOWN_OPT unless in combination with PARSE_OPT_SUBCOMMAND_OPTIONAL");
> + if (flags & PARSE_OPT_KEEP_DASHDASH)
> + BUG("subcommands are incompatible with PARSE_OPT_KEEP_DASHDASH unless in combination with PARSE_OPT_SUBCOMMAND_OPTIONAL");
MM, very long lines. Maybe something like:
BUG("%s flag cannot be combined with %s, unless %s is also provided", ...);
> + }
> + }
> if ((flags & PARSE_OPT_KEEP_UNKNOWN_OPT) &&
> (flags & PARSE_OPT_STOP_AT_NON_OPTION) &&
> !(flags & PARSE_OPT_ONE_SHOT))
> @@ -589,6 +634,7 @@ static int show_gitcomp(const struct option *opts, int show_all)
> int nr_noopts = 0;
>
> for (; opts->type != OPTION_END; opts++) {
> + const char *prefix = "--";
Ah, here we have a prefix variable, but it's more of an "infix"... :)
> + return PARSE_OPT_DONE;
> + error(_("unknown subcommand: `%s'"), arg);
> + usage_with_options(usagestr, options);
ditto earlier patch comment about usage_msg_opt() (also applies below)
> + case PARSE_OPT_NON_OPTION:
> + /* Impossible. */
We could just skip this comment as...
> + BUG("parse_subcommand() cannot return these");
...this makes it clear.
> case PARSE_OPT_DONE:
> + if (ctx.has_subcommands &&
> + !(flags & PARSE_OPT_SUBCOMMAND_OPTIONAL)) {
> + error(_("need a subcommand"));
> + usage_with_options(usagestr, options);
ditto.
> +typedef int parse_opt_subcommand_fn(int argc, const char **argv,
> + const char *prefix);
We usually define function typedefs like:
typedef int (*parse_opt_subcommand_fn)(...);
But I see that...
> enum parse_opt_type type;
> @@ -145,6 +155,7 @@ struct option {
> intptr_t defval;
> parse_opt_ll_cb *ll_callback;
> intptr_t extra;
> + parse_opt_subcommand_fn *subcommand_fn;
You're doing this consistently with parse_opt_ll_cb etc., fair enough.
> +#define OPT_SUBCOMMAND_F(l, v, fn, f) { \
> + .type = OPTION_SUBCOMMAND, \
> + .long_name = (l), \
> + .value = (v), \
> + .flags = (f), \
Nice to see this form
> + .subcommand_fn = (fn) }
Almost all other such decls end with:
.foo = bar, \
}
I.e. we put the closing "}" at the start of the line.
> +#define OPT_SUBCOMMAND(l, v, fn) OPT_SUBCOMMAND_F((l), (v), (fn), 0)
I think this would be more readable as:
#define OPT_SUBCOMMAND(l, v, fn) \
OPT_SUBCOMMAND_F((l), (v), (fn), 0)
Which both aligns nicely, and isn't using a \t in the middle of the line
(which in any case we'll end up aligning with nothing else, as the names
are all different lengths, if other things continue using this pattern.
> enum parse_opt_flags flags;
> + unsigned has_subcommands;
Maybe make it: "unsigned has_subcommands:1"?
> + printf("fn: subcmd_one\n");
s/printf/puts/g here
> + print_args(argc, argv);
> + return 0;
> +}
> +
> +static int subcmd_two(int argc, const char **argv, const char *prefix)
> +{
> + printf("fn: subcmd_two\n");
ditto.
next prev parent reply other threads:[~2022-08-19 18:02 UTC|newest]
Thread overview: 97+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-07-25 12:38 [PATCH 00/20] parse-options: handle subcommands SZEDER Gábor
2022-07-25 12:38 ` [PATCH 01/20] git.c: update NO_PARSEOPT markings SZEDER Gábor
2022-07-25 14:31 ` Ævar Arnfjörð Bjarmason
2022-08-02 17:37 ` SZEDER Gábor
2022-08-02 21:00 ` Junio C Hamano
2022-08-03 13:11 ` Ævar Arnfjörð Bjarmason
2022-08-03 21:34 ` SZEDER Gábor
2022-08-04 7:47 ` Ævar Arnfjörð Bjarmason
2022-08-11 21:35 ` Junio C Hamano
2022-08-12 15:28 ` SZEDER Gábor
2022-08-12 16:46 ` Junio C Hamano
2022-07-26 19:55 ` SZEDER Gábor
2022-07-25 12:38 ` [PATCH 02/20] t3301-notes.sh: check that default operation mode doesn't take arguments SZEDER Gábor
2022-07-25 12:38 ` [PATCH 03/20] t5505-remote.sh: check the behavior without a subcommand SZEDER Gábor
2022-07-25 14:37 ` Ævar Arnfjörð Bjarmason
2022-07-25 12:38 ` [PATCH 04/20] t0040-parse-options: test parse_options() with various 'parse_opt_flags' SZEDER Gábor
2022-07-25 14:38 ` Ævar Arnfjörð Bjarmason
2022-08-12 15:04 ` SZEDER Gábor
2022-07-25 12:38 ` [PATCH 05/20] api-parse-options.txt: fix description of OPT_CMDMODE SZEDER Gábor
2022-07-25 12:38 ` [PATCH 06/20] parse-options: PARSE_OPT_KEEP_UNKNOWN only applies to --options SZEDER Gábor
2022-07-25 12:38 ` [PATCH 07/20] parse-options: clarify the limitations of PARSE_OPT_NODASH SZEDER Gábor
2022-07-25 12:38 ` [PATCH 08/20] parse-options: drop leading space from '--git-completion-helper' output SZEDER Gábor
2022-07-25 12:38 ` [PATCH 09/20] parse-options: add support for parsing subcommands SZEDER Gábor
2022-07-25 14:43 ` Ævar Arnfjörð Bjarmason
2022-07-25 19:29 ` SZEDER Gábor
2022-07-25 19:41 ` Ævar Arnfjörð Bjarmason
2022-07-25 21:02 ` SZEDER Gábor
2022-08-12 15:15 ` SZEDER Gábor
2022-07-25 17:37 ` Junio C Hamano
2022-07-25 12:38 ` [PATCH 10/20] builtin/bundle.c: let parse-options parse subcommands SZEDER Gábor
2022-07-25 12:38 ` [PATCH 11/20] builtin/commit-graph.c: " SZEDER Gábor
2022-07-25 12:38 ` [PATCH 12/20] builtin/gc.c: let parse-options parse 'git maintenance's subcommands SZEDER Gábor
2022-07-25 12:38 ` [PATCH 13/20] builtin/hook.c: let parse-option parse subcommands SZEDER Gábor
2022-07-25 12:38 ` [PATCH 14/20] builtin/multi-pack-index.c: let parse-options " SZEDER Gábor
2022-07-25 12:38 ` [PATCH 15/20] builtin/notes.c: " SZEDER Gábor
2022-07-25 16:49 ` Junio C Hamano
2022-07-25 12:38 ` [PATCH 16/20] builtin/reflog.c: " SZEDER Gábor
2022-07-25 12:38 ` [PATCH 17/20] builtin/remote.c: " SZEDER Gábor
2022-07-25 12:38 ` [PATCH 18/20] builtin/sparse-checkout.c: " SZEDER Gábor
2022-07-25 12:38 ` [PATCH 19/20] builtin/stash.c: " SZEDER Gábor
2022-07-25 12:38 ` [PATCH 20/20] builtin/worktree.c: " SZEDER Gábor
2022-07-25 13:15 ` [PATCH 00/20] parse-options: handle subcommands Derrick Stolee
2022-07-25 16:00 ` SZEDER Gábor
2022-07-25 16:08 ` Derrick Stolee
2022-07-25 17:13 ` Ævar Arnfjörð Bjarmason
2022-07-25 17:56 ` Junio C Hamano
2022-07-26 15:42 ` Johannes Schindelin
2022-07-26 18:02 ` Ævar Arnfjörð Bjarmason
2022-08-19 16:03 ` [PATCH v2 " SZEDER Gábor
2022-08-19 16:03 ` [PATCH v2 01/20] git.c: update NO_PARSEOPT markings SZEDER Gábor
2022-08-19 16:03 ` [PATCH v2 02/20] t3301-notes.sh: check that default operation mode doesn't take arguments SZEDER Gábor
2022-08-19 16:03 ` [PATCH v2 03/20] t5505-remote.sh: check the behavior without a subcommand SZEDER Gábor
2022-08-19 16:03 ` [PATCH v2 04/20] t0040-parse-options: test parse_options() with various 'parse_opt_flags' SZEDER Gábor
2022-08-19 17:23 ` Ævar Arnfjörð Bjarmason
2022-08-20 11:14 ` SZEDER Gábor
2022-08-19 18:18 ` Junio C Hamano
2022-08-20 10:31 ` SZEDER Gábor
2022-08-20 21:27 ` Junio C Hamano
2022-08-19 16:03 ` [PATCH v2 05/20] api-parse-options.txt: fix description of OPT_CMDMODE SZEDER Gábor
2022-08-19 16:03 ` [PATCH v2 06/20] parse-options: PARSE_OPT_KEEP_UNKNOWN only applies to --options SZEDER Gábor
2022-08-19 16:03 ` [PATCH v2 07/20] parse-options: clarify the limitations of PARSE_OPT_NODASH SZEDER Gábor
2022-08-19 16:03 ` [PATCH v2 08/20] parse-options: drop leading space from '--git-completion-helper' output SZEDER Gábor
2022-08-19 17:30 ` Ævar Arnfjörð Bjarmason
2022-08-19 18:35 ` SZEDER Gábor
2022-08-19 16:04 ` [PATCH v2 09/20] parse-options: add support for parsing subcommands SZEDER Gábor
2022-08-19 17:33 ` Ævar Arnfjörð Bjarmason [this message]
2022-08-19 19:03 ` Ævar Arnfjörð Bjarmason
2022-08-19 16:04 ` [PATCH v2 10/20] builtin/bundle.c: let parse-options parse subcommands SZEDER Gábor
2022-08-19 17:50 ` Ævar Arnfjörð Bjarmason
2022-08-19 16:04 ` [PATCH v2 11/20] builtin/commit-graph.c: " SZEDER Gábor
2022-08-19 17:53 ` Ævar Arnfjörð Bjarmason
2022-08-19 17:56 ` Ævar Arnfjörð Bjarmason
2022-08-19 18:22 ` SZEDER Gábor
2022-08-19 16:04 ` [PATCH v2 12/20] builtin/gc.c: let parse-options parse 'git maintenance's subcommands SZEDER Gábor
2022-08-19 20:59 ` Junio C Hamano
2022-08-19 16:04 ` [PATCH v2 13/20] builtin/hook.c: let parse-options parse subcommands SZEDER Gábor
2022-08-19 16:04 ` [PATCH v2 14/20] builtin/multi-pack-index.c: " SZEDER Gábor
2022-08-19 17:57 ` Ævar Arnfjörð Bjarmason
2022-08-19 16:04 ` [PATCH v2 15/20] builtin/notes.c: " SZEDER Gábor
2022-08-19 18:01 ` Ævar Arnfjörð Bjarmason
2022-08-21 17:56 ` SZEDER Gábor
2022-08-19 16:04 ` [PATCH v2 16/20] builtin/reflog.c: " SZEDER Gábor
2022-08-19 18:08 ` Ævar Arnfjörð Bjarmason
2022-08-19 16:04 ` [PATCH v2 17/20] builtin/remote.c: " SZEDER Gábor
2022-08-19 16:04 ` [PATCH v2 18/20] builtin/sparse-checkout.c: " SZEDER Gábor
2022-08-19 16:04 ` [PATCH v2 19/20] builtin/stash.c: " SZEDER Gábor
2022-08-19 19:06 ` Ævar Arnfjörð Bjarmason
2022-08-20 10:27 ` SZEDER Gábor
2022-08-19 16:04 ` [PATCH v2 20/20] builtin/worktree.c: " SZEDER Gábor
2022-09-05 18:50 ` [PATCH 0/5] parse-options: minor cleanups for handling subcommands SZEDER Gábor
2022-09-05 18:50 ` [PATCH 1/5] t0040-parse-options: remove leftover debugging SZEDER Gábor
2022-09-05 18:50 ` [PATCH 2/5] test-parse-options.c: don't use for loop initial declaration SZEDER Gábor
2022-09-05 18:50 ` [PATCH 3/5] test-parse-options.c: fix style of comparison with zero SZEDER Gábor
2022-09-05 18:50 ` [PATCH 4/5] notes: simplify default operation mode arguments check SZEDER Gábor
2022-09-05 18:50 ` [PATCH 5/5] notes, remote: show unknown subcommands between `' SZEDER Gábor
2022-09-07 19:12 ` [PATCH 0/5] parse-options: minor cleanups for handling subcommands Junio C Hamano
2022-09-07 21:22 ` SZEDER Gábor
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=220819.86bksg6s5l.gmgdl@evledraar.gmail.com \
--to=avarab@gmail.com \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=szeder.dev@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).