From: Michal Sojka <sojkam1@fel.cvut.cz>
To: Jani Nikula <jani@nikula.org>, notmuch@notmuchmail.org,
	David Bremner <david@tethera.net>
Subject: Re: [PATCH] test: Improve tests for the date/time parser module
In-Reply-To: <87lifn47qn.fsf@nikula.org>
References: <cover.1347484177.git.jani@nikula.org>
	<24186aafbdcb967b8f66c2390c928f3788ab6cbf.1347484177.git.jani@nikula.org>
	<87zk4e1f5k.fsf@steelpick.2x.cz> <87lifn47qn.fsf@nikula.org>
User-Agent: Notmuch/0.14+23~g9d68aca (http://notmuchmail.org) Emacs/24.2.1
	(x86_64-pc-linux-gnu)
Date: Thu, 04 Oct 2012 13:02:28 +0200
Message-ID: <87obkifqkb.fsf@steelpick.2x.cz>
MIME-Version: 1.0
Content-Type: text/plain
Precedence: list

On Wed, Oct 03 2012, Jani Nikula wrote:
> On Tue, 25 Sep 2012, Michal Sojka <sojkam1@fel.cvut.cz> wrote:
>> This patch reworks date/time parser library test program to make it
>> easier to to write the actual tests. It also modifies the notmuch test
>> script and adds several new tests to it.
>

[...]

>>  	case 'n':
>> -	    /* specify now in seconds since epoch */
>> -	    now = (time_t) strtol (optarg, NULL, 10);
>> +	    memset (&tm, 0, sizeof (tm));
>> +	    char *parsed = strptime (optarg, format, &tm);
>
> One of the problems with strptime is that it doesn't support time zones,
> which is why I chose not to use it here. (You can specify %z in the
> format to ignore it, but it looks like it's ignored no matter
> what. *shrug*) Combined with mktime below, you introduce possible TZ and
> DST variations in the tests, which can be problematic. So perhaps we
> should keep the reference time as a timestamp here.

I didn't pay much attention to time zone issues when writing this little
program, so you might be true. But note that test-lib.sh sets TZ=UTC and
I hope that this should eliminate the problems. I think that it is
better to have human understandable values in test scripts.

-Michal