ietf-openproxy
[Top] [All Lists]

Re: Decision on Rules Language

2003-09-12 13:01:05

On Fri, 12 Sep 2003, Markus Hofmann wrote:

As mentioned earlier, we'd assume that the "P" syntax as specified
so far is accepted and we won't have to start discussions on syntax
modifications.

We would need to add the following items that will have minor affect
on P syntax. Again, I expect no Bikeshed problems, but I think I have
to mention these for correctness sake.

        1. Fault/error detection/recovery mechanism. This will
           most likely be something simple that resembles
           boolean expressions while having core benefits of
           "exceptions" mechanism common to modern languages:

                {
                        // piece of any code
                } or {
                        // piece of code executed if the above
                        // piece fails for any reason
                } or {
                        // piece of code executed if the above
                        // two pieces fail for any reason
                }

           In other words, the "expression value" of a piece
           of code is "true" unless the code fails for any
           reason. Actually, I do not think this changes
           existing syntax, but that's an important missing
           feature.

        2. Constants:

                - boolean true/false
                - "strings" (already used but not documented)
                - numbers (decimal only or both decimal and hex?)
                - other (IP addresses: '127.0.0.1' or '1::h:5')?

                Note that XML-based language would have to
                document this as well

        3. Comments:

                - none supported
                - /* block comments */
                - // line comments
                - block and line comments

           I like line comments a lot, but would probably vote
           against them in P because they introduce a notion
           of a "line", which may be a bad idea if we are
           going to use P expressions in interesting contexts
           like HTTP headers where there are no lines.

        4. International support (i18n)

           Can we just say that everything is UTF-8 encoded?
           Does Unicode define "space characters"? Or do we have
           to say explicitly what those [ASCII] characters are?


If you have a strong opinion regarding any of the above or know more
about UTF/Unicode, please let me know.

We're already late with the rules document, so we'd expect a first
WG document on "P" to be published by 9/20. Alex, can you please
take the responsibility to do so.

Sure. I will wait for objections until Tuesday night before
submitting.

Andre indicated that he'd be available to help and write text as
well, if this would be helpful, please get in touch with him.

Done. Thanks Andre for not holding this against me. I am sure P will
reuse most of your work.

Alex.

<Prev in Thread] Current Thread [Next in Thread>