|
Boost : |
From: Joel de Guzman (joel_at_[hidden])
Date: 2007-04-14 18:50:39
ebo_at_[hidden] wrote:
> Joel de Guzman <joel_at_[hidden]> said:
>
>> ebo_at_[hidden] wrote:
>>
>> EBo! So nice to see you every now and then.
>
> Joel! You always make me feel so welcome ;-)
>
>>> Something I was playing around with a couple of years ago and never took the
>>> project to full production was a user extensible machine-tool parser which
>>> kept the rules and the grammar in persistent dynamically loadable objects. I
>>> was able to load new commands on the fly, or actually change the grammar rules
>>> themselves...
>>> ...
>>> Would something like that be interesting?
>> Yeah, that's cool! I'm not sure it fits the criteria I outlined though.
>
> I'll review the criteria tomorrow...
>
>> Seems to be too complex for the cookbook?
>
> How many lines/pages of code and description is reasonable? The entire
> framework for the above is something like 1,750 lines of code, but that
> includes the dynamic parser and 28 RS-274 M and G codes. Stripped down for a
> minimal example could be a fraction of that. Anyway, maybe we can come up
> with something appropriate.
I'm thinking more in the tune of a few lines of grammar here and there
-- small enough to digest in a few minutes. Something like... if you
want to do something like this, you write something like this.
Example: "I want to parse pascal style strings where the first byte
is the length (N) of the string, followed by N characters. How do
you do it?"
Of course that's a very simple example. We have more elaborate examples
like "How do you write a C-like language and a virtual machine?". We
have that. Now, I want some more in-betweens.
Regards,
-- Joel de Guzman http://www.boost-consulting.com http://spirit.sf.net
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk