From: Jens Lehmann <jens.lehmann@web.de>
Date: Thu, 19 May 2011 18:21:52 GMT
Subject: Re: [RFC PATCH] unpack-trees: add check_worktree flag to enable dry-run functionality
Message-ID: <4DD55FC0.4000505@web.de>
In-Reply-To: <7vei3u7e25.fsf@alter.siamese.dyndns.org>

Am 19.05.2011 20:14, schrieb Junio C Hamano:
> Jens Lehmann <Jens.Lehmann@web.de> writes:
> 
>>>> +	if (opts.update && dry_run)
>>>> +		opts.update = 0;
>>>
>>> ... this hunk must go, right?
>>
>> But this is the "don't update the work tree when -n is used together
>> with -u" part, so it is needed, no? With this patch applied first and
>> opts.check_worktree set to 1 inside that if() added there all tests
>> succeed.
> 
> I would say the natural way to do your "dry-run" would be to change the
> inner guts of unpack_trees() codepath that currently does
> 
> 	if (opts.update) {
> 		if (do something to the work tree and get non-zero on failure)
> 			die("... cannot update '%s'", path);
> 	}
> 
> with your "-n" work to
> 
> 	if (opts.update) {
>         	if (opts.dry_run) {
> 			if (would the work tree operation fail?)
> 				say("... update would fail because ... '%s'", path);
> 		} else {
> 			if (do something to the work tree and get non-zero on failure)
> 				die("... cannot update '%s'", path);
> 		}
> 	}
> 
> and that was why I thought you would want to keep the original value of
> opts.update. I wouldn't think of a good way to make the code that kicks
> in when both update and dry_run are set if you clear update that early in
> the codepath.

Ah, now I get it ... Thanks, will add a dry_run flag that keeps update
from changing anything.