Re: [MMUSIC] A bug in SDP Syntax (CRLF)?

["Federico Montesino Pouzols" <fedemp@suidzer0.org>]

>>        An implementor would have to choose between strictly
>>applying the ABNF or being tolerant with single LFs.
>>
>>         Should not -at least- a note be added to the ABNF clarifying
>>this exception?
>
> No. A correct implementation will use CRLF in all cases, and the ABNF
> reflects this. Using a single LF to terminate lines is a bug.
>
> Colin
>

   Then I infer that if I am doing a new implementation and I want it
to be compliant with the specification I must always use CR LF and
consider single LFs as a mistake. All the lines in the SDP
descriptions generated by the new implementation must end with CR LF
and my new parser must not accept descriptions with lines ending with
a single LF?

   This new implementation would follow the ABNF, however it would
also cowardly ignore the _"although parsers SHOULD be tolerant and also
accept records terminated with a single newline character"_ in section
6.

   Using a LF to terminate lines may not be strictly correct, but the
specification recommends (SHOULD) parsers to be tolerant about it. So
I think it would be clearer for implementors if the ABNF syntax
mentioned this exception as a note or whatever.

   From my point of view and my understanding of the draft, an
implementation that in all cases only accepts CR LF is correct in some
sense but would be incorrect at the same time, since it ignores the
quoted sentence in the draft and RFC. In fact, it would reject a lot
of SDP descriptions and could not interoperate with many tools out
there.





_______________________________________________
mmusic mailing list
mmusic@ietf.org
https://www1.ietf.org/mailman/listinfo/mmusic