Re: [PATCH] getdate add a week when the wday is the same as the current one

[Jim Meyering]

Giuseppe Scrivano wrote:
> what do you think of the following patch?  When the same day as the
> current one is provided then a week is added.  It doesn't change the
> behaviour for days in the past.
>
> It closes coreutils #25406.
>
> diff --git a/lib/getdate.y b/lib/getdate.y
> index 877b264..8154fd9 100644
> --- a/lib/getdate.y
> +++ b/lib/getdate.y
> @@ -1435,7 +1435,7 @@ get_date (struct timespec *result, char const *p, 
> struct timespec const *now)
>        if (pc.days_seen && ! pc.dates_seen)
>         {
>           tm.tm_mday += ((pc.day_number - tm.tm_wday + 7) % 7
> -                        + 7 * (pc.day_ordinal - (0 < pc.day_ordinal)));
> +                   + 7 * (pc.day_ordinal + (pc.day_ordinal > 0 && tm.tm_wday 
> == pc.day_number) - (0 < pc.day_ordinal)));
>           tm.tm_isdst = -1;
>           Start = mktime (&tm);
>           if (Start == (time_t) -1)

Nice!
It certainly fixes the bug:

    Today is Friday, May 1:

        $ ./date +%a.%F
        Fri.2009-05-01

    Before, "next friday" would give today's date:

        $ date -d 'next fri' +%a.%F
        Fri.2009-05-01

    With your fix, it does what we expect:

        $ ./date -d 'next fri' +%a.%F
        Fri.2009-05-08

and doesn't appear to cause any regression.
At least all of coreutils' tests still pass.
This highlights that we need a new test or two
to go along with your fix.

Do you feel like adding a test along these lines to gnulib's
tests/test-getdate.c?

Also, for extra credit you can write a ChangeLog entry
and send in "git format-patch" style output, as outlined in

  http://git.sv.gnu.org/cgit/coreutils.git/plain/HACKING

Well, I thought better of it, and have done that latter for you ;-)
I didn't want to wait, since I'm releasing coreutils-7.3 today.

I adjusted the patch not to add a line longer than 80 and to use
"0 < pc.day_ordinal" rather than the equivalent
"pc.day_ordinal > 0".

I expect to push this shortly.
Thank you!

Jim

>From 372c3b4a79def664b8fa73316c35caf39bf93e2c Mon Sep 17 00:00:00 2001
From: Giuseppe Scrivano <address@hidden>
Date: Fri, 1 May 2009 09:23:20 +0200
Subject: [PATCH] getdate: correctly interpret "next monday" when run on a Monday

* lib/getdate.y (get_date): Correct the calculation of tm_mday so
that e.g., "next tues" (when run on a tuesday) results in a date
that is one week in the future, and not today's date.
Reported by Tom Broadhurst http://savannah.gnu.org/bugs/?25406
and earlier by Martin Bernreuther.
---
 lib/getdate.y |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/lib/getdate.y b/lib/getdate.y
index 877b264..1964e9a 100644
--- a/lib/getdate.y
+++ b/lib/getdate.y
@@ -1,7 +1,7 @@
 %{
 /* Parse a string into an internal time stamp.

-   Copyright (C) 1999, 2000, 2002, 2003, 2004, 2005, 2006, 2007, 2008
+   Copyright (C) 1999, 2000, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009
    Free Software Foundation, Inc.

    This program is free software: you can redistribute it and/or modify
@@ -1435,7 +1435,10 @@ get_date (struct timespec *result, char const *p, struct 
timespec const *now)
       if (pc.days_seen && ! pc.dates_seen)
        {
          tm.tm_mday += ((pc.day_number - tm.tm_wday + 7) % 7
-                        + 7 * (pc.day_ordinal - (0 < pc.day_ordinal)));
+                        + 7 * (pc.day_ordinal
+                               + (0 < pc.day_ordinal
+                                  && tm.tm_wday == pc.day_number)
+                               - (0 < pc.day_ordinal)));
          tm.tm_isdst = -1;
          Start = mktime (&tm);
          if (Start == (time_t) -1)
--
1.6.3.rc3.212.g8c698