[antlr-interest] strategy doubt..

Sriram Durbha cintyram at yahoo.com
Thu Oct 24 08:56:17 PDT 2002


this is similar to c++ and c but diferent in a way;

suppose i want to allow programmers of my scical language to use the
ty[pes defined in a different design unit/compilation unit;

in c++ we can do that but we cannot do for types, i mean supose i have
a class type declared in a header file, in another file i have to
include this file.. but i want to allow the programmer to not include
this file but  put it in a known location like current working
directory ; it might look very unsage and loose , but i think it will
be powerful feature also.. 
even if it is not to be implemented in the end ,  the  problem still is
interesting ;

when ever such a syntax feature is deteced , the appropriate type node
has to be added to the ast , but since the information on the node is
contained in a different file, we have to open that file , look for the
type's information and then build a tree; also if we can maintain a
database like structure which always contains such frequently used
userdefined types, we can indexinto it and do it real fast;


any opinions on this will be greatfully appreciated
thankyou
ram


__________________________________________________
Do you Yahoo!?
Y! Web Hosting - Let the expert host your web site
http://webhosting.yahoo.com/

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 



More information about the antlr-interest mailing list