parser update...

Jeremy Huntwork jhuntwork at
Tue Mar 7 04:42:43 PST 2006

George Makrydakis wrote:
> only, so that we have more flexibility in both levels. Work in the C++ 
> version is looking good, class  - like "package" management is what I 
> have in mind; the format of the book  is oriented by itself on such a 
> path. The work will eventually surface, but this time I want as much as 
> features to be complete as possible before posting, no need to rush.

Except, I had in mind that we'd be working together a little bit more on 
this. We already have a repository with a place for concept code - and 
the code, even though I did do some work on it yesterday, is still in 
its very early stages. So, there's no harm done if we completely 
restructure it according to whatever ideas you have.

My hope was that we'd discuss features, concepts, and coding methods 
here, and then the code would go into the repository. Until such time as 
you have an account (which I'm sure would happen soon enough if we 
actually started working together like this), any code snippets/patches 
you present I can drop in.

If we do it this way, instead of one of us working on a complete concept 
on his own (and privately at that!) there's a much better chance of more 
people following what's going on and in turn being able to jump in and help.

As of right now, the C++ code that is in the repository is set up to 
parse a file and separate all <...> tags from anything else, as per our 
discussion. It can also recognize a call to a SYSTEM entity file 
declaration and pull that in, parse and add entities to a linked list. 
The next thing I was going to do is add the code to recognize 
xi:includes and parse those.

If you can see a better way to do *any* of those things - or a better 
way to structure the code - I have no objections to altering what's 
there for something better. And the code isn't so far along that we 
can't change it, even drastically so, if we need to. We have the 
resources available, this list and the repository, let's use them please.


More information about the alfs-discuss mailing list