Re[2]: CEDET, DL & parsing thoughts (was Release plans)

["Eric M. Ludlam" <eric@siege-engine.com>]

>>> "Lennart Borgman (gmail)" <lennart.borgman@gmail.com> seems to think that:
>Eric M. Ludlam wrote:
>> This is in effect what is in CEDET/Semantic now but without the DL.  I
>> had made a replacement for flex, but more Emacs Lisp centric, and
>> David Ponce ported bison into Emacs Lisp directly.  This bison port
>> supports incremental parsing, full parsing, reparsing, and is quite
>> fast, though not nearly as fast as actual flex/bison/c code.
>> 
>> I would assume the concepts in David Ponce's wisent parser generator
>> could be back-ported into Bison if desired.
>
>So (if I understand currectly from the little I know about this), a bit
>sadly, if there were sutiable flex and bison dlls these could be used
>instead (if Davids specials were backported too) and would be quite a
>bit faster.

I can imagine how it *might* be done, though to be honest, it may be
that having "eval" actions in the real bison generated parser might be
all it takes to slow it down.  I would not claim this as an area of
expertise.

>This does not relate to the general question of loading dll:s. This is a
>special case that I suppose Richard would approve. (I have no idea, but
>I guess there are no flex and bison dlls availabe, or are there?)

It would work like this:

1) lex/grammar file -> flex & bison -> c-files
2) compile flex-bison output + generic interface -> dl
3) load dl, bind generated parser commands to fcns
4) bind those fcns to generic CEDET/Semantic calls
5) Semantic infrastructure calls into the dll for parsing goodness.

>However for new languages all this requires also writing language
>specific bison grammars. Is this perhaps be a big job that just a few
>person have insight in how to do?

My goal was to build a tool that would make it easy for anyone to
write a parser that would feed into an infrastructure that provides
the complex specialy functionality.

This is true now for what is in CEDET.  There are already 6 grammars
that work pretty well that enthusiasts of a particular language have
written.  (Erlang, python, csharp, javascript, 2 php parsers, and
ruby.)

>> When I started, I really wanted to have a single generic parsing
>> infrastructure that could do indentation, coloring, and tagging.
>
>nxml-mode does that, but on its own of course.

It is my understanding that only nxml, and that new javascript mode do
this, making the undertaking a statistically less likely occurance in
the current infrastructure.

>> Once CEDET is merged into Emacs, I hope to examine some of the speed
>> issues with others who know more what Emacs' internals are like.  (As
>> an FYI, all of CEDET's papers should now be in order for this.)
>
>Great. Then we can hope for that it is easy to get started using CEDET.

I agree.
Eric

-- 
          Eric Ludlam:                       eric@siege-engine.com
   Siege: www.siege-engine.com          Emacs: http://cedet.sourceforge.net