Re: Aligning NPT syntax with RTSP

On Fri, 29 Jan 2010 01:21:48 +0100, Silvia Pfeiffer  
<silviapfeiffer1@gmail.com> wrote:

> On Fri, Jan 29, 2010 at 10:05 AM, Philip Jägenstedt <philipj@opera.com>  
> wrote:
>> On Wed, 27 Jan 2010 23:28:05 +0100, Silvia Pfeiffer
>> <silviapfeiffer1@gmail.com> wrote:
>>
>>> On Wed, Jan 27, 2010 at 11:45 PM, Philip Jägenstedt <philipj@opera.com>
>>> wrote:
>>>>
>>>> On Wed, 27 Jan 2010 13:32:11 +0100, Silvia Pfeiffer
>>>> <silviapfeiffer1@gmail.com> wrote:
>>>>
>>>>> On Wed, Jan 27, 2010 at 10:33 PM, Philip Jägenstedt  
>>>>> <philipj@opera.com>
>>>>> wrote:
>>>>>>
>>>>>> On Wed, 27 Jan 2010 12:23:25 +0100, Silvia Pfeiffer
>>>>>> <silviapfeiffer1@gmail.com> wrote:
>>>>>>
>>>>>>> On Wed, Jan 27, 2010 at 10:09 PM, Philip Jägenstedt
>>>>>>> <philipj@opera.com>
>>>>>>> wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> http://www.w3.org/2008/WebVideo/Fragments/WD-media-fragments-spec/#npttime
>>>>>>>>
>>>>>>>> This is what I mentioned in the teleconf. As it is, '0.' would  
>>>>>>>> not be
>>>>>>>> a
>>>>>>>> valid production of npttime but it is a valid production of  
>>>>>>>> npt-sec
>>>>>>>> from
>>>>>>>> RTSP [1]. The same is true of '00:00:00.'. The difference is in
>>>>>>>> digits
>>>>>>>> after
>>>>>>>> the decimal point.
>>>>>>>
>>>>>>> We currently have:
>>>>>>>
>>>>>>> npttime    ::=          ( 1*DIGIT [ "." 1*DIGIT ] [ timeunit ] ) |
>>>>>>>                      ( 1*DIGIT ":" 2DIGIT ":" 2DIGIT [ "."  
>>>>>>> 1*DIGIT] )
>>>>>>>
>>>>>>> which I think you are proposing to change to
>>>>>>>
>>>>>>> npttime    ::=          ( 1*DIGIT [ "." *DIGIT ] [ timeunit ] ) |
>>>>>>>                      ( 1*DIGIT ":" 2DIGIT ":" 2DIGIT [ "." *DIGIT]  
>>>>>>> )
>>>>>>>
>>>>>>> Correct?
>>>>>>
>>>>>> To be precise, I'm suggesting referring to the definition in  
>>>>>> RFC2326,
>>>>>> noy
>>>>>> copying it. The effect is the same of course.
>>>>>
>>>>> Could do ...  otoh if the RTP spec changes this, we are not
>>>>> dependent... and it's really short.
>>>>
>>>> RFCs can never change, but I have no objections to copying as long as
>>>> there
>>>> is a (normative?) reference to RFC2326 for readers to follow.
>>>
>>> Yeah, that's right. I guess the only reason then is not to have to go
>>> and read another document.
>>> I'm happy to add a normative reference to RFC2326 and still leaven the
>>> two-liner in there.
>>
>> I tried making the edit and it turns out RFC2326 is a bit quirky on the
>> HHMMSS format:
>>
>>   npt-hhmmss   =   npt-hh ":" npt-mm ":" npt-ss [ "." *DIGIT ]
>>   npt-hh       =   1*DIGIT     ; any positive number
>>   npt-mm       =   1*2DIGIT    ; 0-59
>>   npt-ss       =   1*2DIGIT    ; 0-59
>>
>> It allows 1-digit minute or second, which we currently do not. Do we  
>> really
>> want this?
>
> I think you may be misreading this. npt-mm and npt-ss say "use one
> 2DIGIT", i.e. there are two digits always, IMO.

Following the references back to ABNF in  
http://www.ietf.org/rfc/rfc2068.txt:

*rule
      The character "*" preceding an element indicates repetition. The
      full form is "<n>*<m>element" indicating at least <n> and at most
      <m> occurrences of element. Default values are 0 and infinity so
      that "*(element)" allows any number, including zero; "1*element"
      requires at least one; and "1*2element" allows one or two.

Also, the RTSP spec proclaims that "The syntax conforms to ISO 8601." I'm  
unable to figure out if ISO 8601 allows single-digit hours and minutes or  
not. Is anyone familar with RTSP so that they can check if implementations  
actually support this, or if it might just be a typo?

-- 
Philip Jägenstedt
Core Developer
Opera Software

Received on Friday, 29 January 2010 00:54:42 UTC