Re: [PATCHv12 0/7] Expose submodule parallelism to the user

From: Junio C Hamano <gitster@pobox.com>
Date: Feb 18, 2016, 23:20 UTC
Message-ID: <xmqqio1lehbp.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <CAGZ79kZwgBJQ2bkvKSuT+ULiSPL9x5WTcwVJG0FATNv-D_-AhQ@mail.gmail.com>

Stefan Beller <sbeller@google.com> writes:

Show 7 quoted lines

> On Thu, Feb 18, 2016 at 3:12 PM, Stefan Beller <sbeller@google.com> wrote:
>>> Unless you count "I want to write differently from what was
>>> suggested" is a desirable thing to do, I do not see a point in
>>> favouring the above that uses an extra variable and skip_prefix()
>>> over what I gave you as "how about" patch.  But whatever.
>>
>> The skip_prefix was there before, so it stuck there.

Sorry, but I thought this "parsing update strategy" was all new code.

Show 5 quoted lines

>> Also it seems a bit more high level to me hence easier to read,
>> (though I am biased). I'll use your suggestion.
>
> and it doesn't crash when passing in value == NULL.
> (We don't do that currently, just a side observation)

Hmph. If you pass str==NULL with prefix="!" to what we have below, I would think the first iteration would try to read from *str and do a bizarre thing.

static inline int skip_prefix(const char *str, const char *prefix,
			      const char **out)
{
	do {
		if (!*prefix) {
			*out = str;
			return 1;
		}
	} while (*str++ == *prefix++);
	return 0;
}

Puzzled.