Skip Menu |

This queue is for tickets about the DateTime-Format-Natural CPAN distribution.

Report information
The Basics
Id: 66091
Status: open
Priority: 0/
Queue: DateTime-Format-Natural
People
Owner: Nobody in particular
Requestors: PHILKIME [...] cpan.org
Cc:
AdminCc:
Bug Information
Severity: Important
Broken in: 0.93
Fixed in: (no value)

History Show all quoted text
  Wed Feb 23 12:39:28 2011 PHILKIME [...] cpan.org - Ticket created
Subject: No way to distinguish between empty month/day and "1"?
There doesn't seem to be a way to distinguish between: 2008 2008-01-01 I am trying to parse dates which may or may not have month/day (in bibliographies, very common) and then I need to set individual day/month/year fields. Currently, I can't find any DateTime::Format module which will not set day/month if they are not present, even though a bare year is, for example, a valid ISO8601 date. DateTime::Incomplete will perhaps do this but I can't use an object of this type as a base for DateTime::Format::Natural: DateTime::Format::Natural->new(datetime => DateTime::Incomplete->new); just dies with The 'datetime' parameter ("DateTime::Incomplete=HASH(0x1028bf0a0)") to DateTime::Format::Natural::new did not pass the 'valid object' callback .... This inability to distinguish between an empty month/day and 01/01 is quite a problem when splitting dates into components where the presence/absence of day/mont matters.
  Fri Feb 25 18:56:09 2011 schubiger [...] cpan.org - Correspondence added
Subject: Re: [rt.cpan.org #66091] No way to distinguish between empty month/day and "1"?
Date: Sat, 26 Feb 2011 00:55:52 +0100
To: Philip Kime via RT <bug-DateTime-Format-Natural [...] rt.cpan.org>
From: Steven Schubiger <schubiger [...] cpan.org>
Philip Kime via RT <bug-DateTime-Format-Natural@rt.cpan.org> wrote: Show quoted text
> There doesn't seem to be a way to distinguish between: > > 2008 > 2008-01-01
The resulting dates are indistinguishable, agreed. Show quoted text
> I am trying to parse dates which may or may not have month/day (in bibliographies, very > common) and then I need to set individual day/month/year fields. Currently, I can't find any > DateTime::Format module which will not set day/month if they are not present, even though a > bare year is, for example, a valid ISO8601 date. DateTime::Incomplete will perhaps do this
DateTime::Format::Natural keeps an internal counter of units having been set or modified; furthermore, it also invokes DateTime's truncate method to finally reset time components. As a rough idea, I could imagine exposing the counter data via an accessor to the user in order to allow for it being checkable. For what it's worth, call `dateparse -T` with a recent version, enter the two date strings listed and you should see a summary of altered units for each string successfully parsed. Show quoted text
> but I can't use an object of this type as a base for DateTime::Format::Natural:
Show quoted text
> DateTime::Format::Natural->new(datetime => DateTime::Incomplete->new); > > just dies with > > The 'datetime' parameter ("DateTime::Incomplete=HASH(0x1028bf0a0)") to > DateTime::Format::Natural::new did not pass the 'valid object' callback ....
This happens because not a DateTime object was provided; DateTime::Incomplete's to_datetime method returns one (according to its documentation).
  Fri Feb 25 18:56:09 2011 The RT System itself - Status changed from 'new' to 'open'