git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [RFC PATCH 08/11] rebase -i: use struct rebase_options to parse args

From
PWPhillip Wood <phillip.wood123@gmail.com>
Date
Apr 10, 2019, 19:16 UTC
Message-ID
<1413d9ef-7164-c3e6-629d-8388d3177e1b@gmail.com>
In-Reply-To
<7ba0dc1a-b6dd-4ef6-81ee-495e32fc67cf@gmail.com>
Hi Alban

sorry for the slow reply, I think you're probably off-list for a while by now

On 21/03/2019 21:13, Alban Gruin wrote:
Show 97 quoted lines
> Hi Phillip,
> 
> It’s nice to see your work on this on the list.
> 
> Le 19/03/2019 à 20:03, Phillip Wood a écrit :
>> From: Phillip Wood <phillip.wood@dunelm.org.uk>
>>
>> In order to run `rebase -i` without forking `rebase--interactive` it
>> will be convenient to use the same structure when parsing the options in
>> cmd_rebase() and cmd_rebase__interactive().
>>
>> Signed-off-by: Phillip Wood <phillip.wood@dunelm.org.uk>
>> ---
>>   builtin/rebase.c | 203 ++++++++++++++++++++++++++---------------------
>>   1 file changed, 112 insertions(+), 91 deletions(-)
>>
>> diff --git a/builtin/rebase.c b/builtin/rebase.c
>> index c93f2aa629..33a2495032 100644
>> --- a/builtin/rebase.c
>> +++ b/builtin/rebase.c
>> @@ -50,6 +50,73 @@ enum rebase_type {
>>   	REBASE_PRESERVE_MERGES
>>   };
>>   
>> +struct rebase_options {
>> +	enum rebase_type type;
>> +	const char *state_dir;
>> +	struct commit *upstream;
>> +	const char *upstream_name;
>> +	const char *upstream_arg;
>> +	char *head_name;
>> +	struct object_id orig_head;
>> +	struct commit *onto;
>> +	const char *onto_name;
>> +	const char *revisions;
>> +	const char *switch_to;
>> +	int root;
>> +	struct object_id *squash_onto;
>> +	struct commit *restrict_revision;
>> +	int dont_finish_rebase;
>> +	enum {
>> +		REBASE_NO_QUIET = 1<<0,
>> +		REBASE_VERBOSE = 1<<1,
>> +		REBASE_DIFFSTAT = 1<<2,
>> +		REBASE_FORCE = 1<<3,
>> +		REBASE_INTERACTIVE_EXPLICIT = 1<<4,
>> +	} flags;
>> +	struct argv_array git_am_opts;
>> +	const char *action;
>> +	int signoff;
>> +	int allow_rerere_autoupdate;
>> +	int keep_empty;
>> +	int autosquash;
>> +	char *gpg_sign_opt;
>> +	int autostash;
>> +	char *cmd;
>> +	int allow_empty_message;
>> +	int rebase_merges, rebase_cousins;
>> +	char *strategy, *strategy_opts;
>> +	struct strbuf git_format_patch_opt;
>> +	int reschedule_failed_exec;
>> +};
>> +
>> +#define REBASE_OPTIONS_INIT {			  	\
>> +		.type = REBASE_UNSPECIFIED,	  	\
>> +		.flags = REBASE_NO_QUIET, 		\
>> +		.git_am_opts = ARGV_ARRAY_INIT,		\
>> +		.git_format_patch_opt = STRBUF_INIT	\
>> +	}
>> +
>> +static struct replay_opts get_replay_opts(const struct rebase_options *opts)
>> +{
>> +	struct replay_opts replay = REPLAY_OPTS_INIT;
>> +
>> +	sequencer_init_config(&replay);
>> +
>> +	replay.action = REPLAY_INTERACTIVE_REBASE;
>> +	replay.signoff = opts->signoff;
>> +	replay.allow_ff = !(opts->flags & REBASE_FORCE);
>> +	if (opts->allow_rerere_autoupdate)
>> +		replay.allow_rerere_auto = opts->allow_rerere_autoupdate;
>> +	replay.allow_empty = 1;
>> +	replay.allow_empty_message = opts->allow_empty_message;
>> +	replay.verbose = opts->flags & REBASE_VERBOSE;
>> +	replay.reschedule_failed_exec = opts->reschedule_failed_exec;
>> +	replay.gpg_sign = xstrdup_or_null(opts->gpg_sign_opt);
>> +	replay.strategy = opts->strategy;
>> +
>> +	return replay;
>> +}
>> +
> 
> I wonder if `struct rebase_options` and `struct replay_options` could be
> merged, or at least have `replay_options` used in `rebase_options`,
> instead of converting one to the other.  I think it would make things
> simpler and cleaner, but I don’t know how hard it would be, or if my
> assumption is correct.

I did consider doing that, but there are a few subtle differences in the way the options are stored in each struct (eg the gpg option has a leading -S in struct rebase_options but not struct replay_options) and I think it would be a bit of a faff to align them as it would mean messing with the code uses them and the code that reads/writes the various state files (we cannot change the on-disk format without breaking things - sometimes users start a rebase with one version of git bundled with something like tig and then continue on the command-line with a different version). I'd prefer to leave it as a future cleanup once this series has been merged.

Best Wishes
Phillip
> 
> Cheers,
> Alban
>
Previous: Alban GruinNext: Phillip Wood
Message 16 of 40 in “rebase -i run without forking rebase--interactive”
  1. 00/11 rebase -i run without forking rebase--interactivePhillip Wood, Mar 19, 2019
  2. 02/11 rebase: rename write_basic_state()Phillip Wood, Mar 19, 2019
  3. 01/11 sequencer: always discard index after checkoutPhillip Wood, Mar 19, 2019
  4. Duy NguyenMar 20, 2019
  5. Phillip WoodMar 21, 2019
  6. 03/11 rebase: use OPT_RERERE_AUTOUPDATE()Phillip Wood, Mar 19, 2019
  7. 05/11 rebase -i: remove duplicationPhillip Wood, Mar 19, 2019
  8. 06/11 rebase -i: use struct commit when parsing optionsPhillip Wood, Mar 19, 2019
  9. 04/11 rebase -i: combine rebase--interactive.c with rebase.cPhillip Wood, Mar 19, 2019
  10. 07/11 rebase -i: use struct object_id for squash_ontoPhillip Wood, Mar 19, 2019
  11. 08/11 rebase -i: use struct rebase_options to parse argsPhillip Wood, Mar 19, 2019
  12. Junio C HamanoMar 21, 2019
  13. Phillip WoodMar 21, 2019
  14. Junio C HamanoMar 22, 2019
  15. Alban GruinMar 21, 2019
  16. Phillip WoodApr 10, 2019
  17. 10/11 rebase: use a common action enumPhillip Wood, Mar 19, 2019
  18. Ævar Arnfjörð BjarmasonMar 19, 2019
  19. Phillip WoodMar 21, 2019
  20. 09/11 rebase -i: use struct rebase_options in do_interactive_rebase()Phillip Wood, Mar 19, 2019
  21. 11/11 rebase -i: run without forking rebase--interactivePhillip Wood, Mar 19, 2019
  22. Josh SteadmonMar 20, 2019
  23. Ævar Arnfjörð BjarmasonMar 20, 2019
  24. Phillip WoodMar 21, 2019
  25. Junio C HamanoMar 21, 2019
  26. 00/12 Run rebase -i without forking rebase--interactivePhillip Wood, Apr 17, 2019
  27. 01/12 sequencer: always discard index after checkoutPhillip Wood, Apr 17, 2019
  28. 03/12 rebase: rename write_basic_state()Phillip Wood, Apr 17, 2019
  29. 04/12 rebase: use OPT_RERERE_AUTOUPDATE()Phillip Wood, Apr 17, 2019
  30. 02/12 rebase: don't translate trace stringsPhillip Wood, Apr 17, 2019
  31. Junio C HamanoApr 19, 2019
  32. Phillip WoodApr 25, 2019
  33. 05/12 rebase -i: combine rebase--interactive.c with rebase.cPhillip Wood, Apr 17, 2019
  34. 06/12 rebase -i: remove duplicationPhillip Wood, Apr 17, 2019
  35. 08/12 rebase -i: use struct object_id for squash_ontoPhillip Wood, Apr 17, 2019
  36. 07/12 rebase -i: use struct commit when parsing optionsPhillip Wood, Apr 17, 2019
  37. 09/12 rebase -i: use struct rebase_options to parse argsPhillip Wood, Apr 17, 2019
  38. 10/12 rebase -i: use struct rebase_options in do_interactive_rebase()Phillip Wood, Apr 17, 2019
  39. 12/12 rebase -i: run without forking rebase--interactivePhillip Wood, Apr 17, 2019
  40. 11/12 rebase: use a common action enumPhillip Wood, Apr 17, 2019

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.