Re: [stir] Proposal for update of erratum #6519 - ppt syntax

[Roman Shpount <[e-mail]>]

On Fri, Apr 23, 2021 at 1:03 PM Christer Holmberg <
[e-mail]> wrote:

> >That was a genuine question -- the language we use drive the way we
> think.  Handcoding a parser as a state machine vs using a LL(1) parser vs
> an LR parser introduce different difficulties when
> >writing them, and I was wondering if your insistence was because of that.
>
> I have many years of parser implementation experience. Having said that,
> this specific issue is not related to something I am currently
> implementing, but more in general.
>
> But, if I am the only one having that issue, I am fine to drop it in order
> to make progress. I guess you'd agree that SIP is not the first choice if
> looking for clean and proper syntax :)
>
>
I am currently maintaining 3 different SIP stacks (2 with handwritten
parsers and one with a rule-based parser). I also got a STIR
implementation, even though it is not as mature as I would like.

I typically test my parsers using input fuzzer to see where they produce
different results between each other or other open-source parsers. My main
problem is that if we define ident-type as "ppt" EQUAL token /
quoted-string, then something like ppt="\s\h\a\k\e\n" is potentially legal.
I do not want to update parsers that do not support such syntax. We can put
some language that prohibits this, but having an ABNF rule and text saying
that the value is token surrounded by quotes archives the same goal much
cleaner.

Best Regards,
_____________
Roman Shpount