Re: [Nmh-workers] Date Parsing Problems.

[Ralph Corderoy]

Hi again,

> So are all ADJUST_NUMERIC_ONLY_TZ_OFFSETS_WRT_DST's effects cancelled
> out?  Setting it to zero passes all tests and suggests it can be
> deleted along with the code it wraps?

Changing ADJUST_NUMERIC_ONLY_TZ_OFFSETS_WRT_DST from 1 to 0 passes all
tests because it's tested with #ifdef!  :-)

To test if it was doing anything, I deleted it locally, and all the code
it wraps, and the ripples, e.g. a function's unused parameter.  One test
then fails: test/mhical/test-mhical.  I'm not sure the test's correct?

It takes

    BEGIN:VCALENDAR
    VERSION:2.0
    PRODID:test-mhical
    BEGIN:VTIMEZONE
    TZID:Eastern Standard Time
    BEGIN:STANDARD
    DTSTART:16010101T020000
    TZOFFSETFROM:-0400
    TZOFFSETTO:-0500
    RRULE:FREQ=YEARLY;INTERVAL=1;BYDAY=1SU;BYMONTH=11
    END:STANDARD
    BEGIN:DAYLIGHT
    DTSTART:16010101T020000
    TZOFFSETFROM:-0500
    TZOFFSETTO:-0400
    RRULE:FREQ=YEARLY;INTERVAL=1;BYDAY=2SU;BYMONTH=3
    END:DAYLIGHT
    END:VTIMEZONE
    BEGIN:VEVENT
    DTSTAMP:20150101T000000Z
    DTSTART;TZID=Eastern Standard Time:20140309T015959
    DTEND;TZID=Eastern Standard Time:20140309T023000
    Summary: EST to EDT
    END:VEVENT
    END:VCALENDAR

and wants this output

    Summary: EST to EDT
    At: Sun, 09 Mar 2014 01:59 -0500
    To: Sun, 09 Mar 2014 03:30

but the time of the `To' line comes out wrong now.

    To: Sun, 09 Mar 2014 04:30

Firstly, is that a valid VCALENDAR?  Is the event 31 minutes long?  Does

    DTEND;TZID=Eastern Standard Time:20140309T023000

mean that 02:30:00 time is EST?  Only that's in the hour that's jumped
over so date(1) isn't happy it exists, but perhaps VCALENDAR defines
that's OK.  Here's DTSTART.

    $ fmt='+%Y-%m-%d %T %z %Z'
    $
    $ TZ=EST5EDT /bin/date -d '2014-03-09T01:59:00 EST' "$fmt"
    2014-03-09 01:59:00 -0500 EST

And then the time 31 minutes later.

    $ TZ=EST5EDT /bin/date -d '2014-03-09T01:59:00 EST + 31 minutes' "$fmt"
    2014-03-09 03:30:00 -0400 EDT

That matches the expect, zoneless, result.

The lexer sets the zone to -05:00 for both EST and EDT, but sets the DST
flag for the latter.

    {nl}("est"|"EST")               SETZONE(-500);
    {nl}("edt"|"EDT")               SETDST(); SETZONE(-500);

Would it be better to ditch DST from the calculations, and make EDT -400
above?  Then it wouldn't need adjusting again on formatting?

The only other use for the DST flag seems to be mh-format(5)'s %(szone)
function.

    Function    Argument Return   Description
    szone       date     integer  timezone explicit?
                                  (1=explicit,0=implicit,-1=unknown)

I don't think 0, implicit, is ever returned.

    $ try() { uip/fmttest -date -format '%(putnum(szone{text}))' "$@"; }
    $
    $ try 'Sun, 23 Apr 2017 12:00:00 +0100'
    1
    $ try 'Sun, 23 Apr 2017 12:00:00 EST'
    1
    $ try 'Sun, 23 Apr 2017 12:00:00 foo'
    -1
    $ try 'Sun, 23 Apr 2017 12:00:00'
    -1

There's only one bit in the masks.

    #define TW_SZONE 0x0004 /* how timezone was determined    */
    #define TW_SZNIL 0x0000 /*   not given                    */
    #define TW_SZEXP 0x0004 /*   explicitly given             */

-- 
Cheers, Ralph.
https://plus.google.com/+RalphCorderoy