Re: [PATCH] gpg-interface: fix for gpgsm v2.3

From: Todd Zullinger <tmz@pobox.com>
Date: Feb 3, 2022, 22:07 UTC
Message-ID: <YfxSCbmatoSZlTwB@pobox.com>
In-Reply-To: <xmqqsfszr5q6.fsf@gitster.g>

Junio C Hamano wrote:

Show 10 quoted lines

> Todd Zullinger <tmz@pobox.com> writes:
> 
>>     -	ret |= !strstr(gpg_status.buf, "\n[GNUPG:] SIG_CREATED ");
>>     +	string_list_split_in_place(&lines, gpg_status.buf, '\n', -1);
>>     +	ret |= !unsorted_string_list_has_string(&lines, "[GNUPG:] SIG_CREATED ");
> 
> Is "SIG_CREATED " supposed to be at the end of that line?  I thought
> that has_string() asks for an exact match, and unfortunately(?)
> there is not the string_list_has_string_that_has_this_prefix()
> function.  So...

By default, yes. The string_list struct uses strcmp() if no cmp function is given. That's why the previous chunk has:

    struct string_list lines = { .cmp = starts_with };

There aren't any similar uses in the code, which is just one of the reasons I wasn't confident that it was a good idea or even a good implementation.

-- 
Todd