[Date Index][Thread Index]
[Date Prev][Date Next][Thread Prev][Thread Next]

Re: WML 2.0.1

On Thu, 25 May 2000, Franz Knipp wrote:

FK> >   * Undefined tags are now parsed like defined tags, it means that it is
FK> IMO undefined tags shouldn't be parsed and altered at all. If there's a
FK> mp4h command inside, this should be expanded as usual, but nothing else.

In order to not altering the tags it is reqired to parse them,
how else should one now where the tag ends!?

FK> A tag ends with a right angle bracket - nested tags have to be possible
FK> (of course).

Especially to support this, a good parser is _essential_.

FK> >   * This implies that left angle brackets inside ePerl code must be
FK> >     protected.
FK> This shouldn't be necessary at all. I like WML because of its
FK> transparency in the sense, that it leaves the files untouched where
FK> there's no WML construct (especially in pass 1-4) - this philosophy has to
FK> take place in all passes! 
FK> So, therefore I've the following suggestion: let's us speak about the
FK> grammar of mp4h before producing three other versions in sequence with
FK> changing syntax...
FK> This will also simplify the job of Denis Barbier who has to check all the
FK> include files before publishing a new version.
FK> What do you think about it?

You are right, as it's obvious now, that this is a serious flaw
in the current version of mp4h, it should be discussed first,
and then implemented. Already too bad, that 2.0.0 is `buggy',
one should try to fix that in not too many new versions.

To me it appears very desirable, if it was not necessary
to tamper ePerl code in WML files by protecting brackets.

To get rid of the need to protect brackets, I suggest to
implement the <: and :> tags into the mp4h Parser, so ePerl
code would not be evaluated except constants, which is 
everything in "" quotes or in so-called "here" documents (<<).

On the other hand I see that it's a Problem that Perl-Semantics
would get into the mp4h parser, which is not desirable if mp4h
shall be usable as a standalone product or if, unlikely enough,
Perl should ever change the ways of using constants.

Denis and the others, do you think, that approach is worth
any further thoughts?

regards, Matthias

http://www.guxhagen.net/matthb/       mailto:matthb@guxhagen.net

Website META Language (WML)                www.engelschall.com/sw/wml/
Official Support Mailing List                   sw-wml@engelschall.com
Automated List Manager                       majordomo@engelschall.com