bug#6132: bug in date: --rfc-3339=TIMESPEC option doesn't print 'T'

[F. Alexander Njemz]

On 05/07/2010 05:55 PM, Eric Blake wrote:
> On 05/07/2010 07:25 AM, F. Alexander Njemz wrote:
>    
> > Hi,
> >
> > as evidenced here [1] the slightly ambiguous wording in the RFC 3339 [2]
> > seems to be the reason that the date utility using the
> > --rfc3339=TIMESPEC switch formats the date with a space instead of a 'T'.
> >
> > I was confused as well, so I took the liberty to mail the authors of the
> > RFC in question. I attached the reply from Mr. Klyne.
> >
> > There is also a patch attached.
> >      
> Thanks for the report, and also for the patch.  However, I'm reluctant
> to create a date format that we cannot then reparse.  So before we apply
> this patch, we would first need a patch to gnulib's getdate.y that
> allows parsing of an rfc3339 style with a 'T' instead of a space.
>
>    


Hi Mr. Blake, hi mailing list,

I'm looking into getdate.y.
Mr. Newman has replied as well (see attachment).

I don't think Mr. Newman's mail applies to the date command, since IMHO 
date returns a
single field.

What do you think?

Regards,

F. Alexander Njemz
> --- Begin Message ---
>
>
>
> Subject: 
>
> Re: rfc3339 is the T mandatory?
>
>
>
>
> Date: 
>
> Fri, 07 May 2010 14:53:27 -0700
>
>
>
>
> If you're representing the date and time as a single field, then the "T" is 
> necessary for compliance with RFC 3339 as Graham stated.
>
> If you're using a multi-field representation where date and time are 
> separate fields, then the field delimiter depends on your higher level 
> field-encoding format.  In this case you would reference the full-date and 
> full-time syntax elements from section 5.6 separately (instead of 
> referencing the date-time syntax) and would be compliant with RFC 3339 if 
> you were compliant with those elements.
>
>                 - Chris
>
> --On May 7, 2010 13:05:41 +0100 Graham Klyne <address@hidden> wrote:
> > In short: "yes"
> >
> > Per section 5.5, the intent in this draft was to specify a timestamp
> > format using elements from and compatible with 8601, but eliminating as
> > far as reasonable any variations that could make timestamp data harder to
> > process. This includes making the 'T' mandatory in date+time values.
> >
> > # g
> > --
> >
> > F. Alexander Njemz wrote:
> > > Dear Sirs,
> > >
> > > there seems to be some confusion about the interpretation of the
> > > following passages
> > > of RFC 3339:
> > >     in section 5.6:
> > >         NOTE: ISO 8601 defines date and time separated by "T".
> > >         Applications using this syntax may choose, for the sake of
> > >         readability, to specify a full-date and full-time separated by
> > >         (say) a space character.
> > >     and in Appendix A:
> > >         ISO 8601 states that the "T" may be omitted under some
> > >         circumstances.  This grammar requires the "T" to avoid ambiguity.
> > >
> > > So, what I would like to ask is if the "T" is mandatory for RFC 3339
> > > conformance or not?
> > >
> > > Thanks in advance.
> > >
> > > Regards,
> > >
> > > F. Alexander Njemz
>
>
> --- End Message ---