Re: dynamic lex library package?

bothner@xenon.stanford.edu (Per
M. Bothner) writes:
[looking for a dynamic lexer library]


I've got a regular expression class in C++ that does something like this.
An example of how I've used it:


        RegExpr re ;


        re |= RegExpr ("[1-9][0-9]*" , 10) ;
        re |= RegExpr ("0[0-7]*" , 8) ;
        re |= RegExpr ("0[xX][0-9A-Fa-f]+" , 16) ;


use:


        const char * str ;
        char * endStr ;


        int base = re.match (str , &endStr) ;


"base" will contain the base of the numeric string, or -1 if there is no
match. "endStr" will point to the first character *not* part of the
matched string.


The class normally builds a NFA, then builds the DFA by lazy evaluation
(in "match()"). However, there is a function to force it to build the
complete DFA, as well as functions to access the resulting DFA (say to
dump the tables to disk).


I have received several requests (via email) to make the class public, but
there is no anonymous ftp access at my site. Also, I do not presently
have the time to do any maintenance on the classs. (There is at present
at least one known bug: the regular expression will never match an empty
string.) At present, it runs with the Zortech compiler (under MS-DOS).


If someone who has an anonymous ftp on their site is willing to take it
over, I'll try to clean it up, get it to compiler under Sun C++ and g++
(the two compilers I have here) and mail them the sources. All I ask is
that my name rest attached. (I'm a consultant, and I need the publicity
:-)).
--
James Kanze GABI Software, SARL
email: kanze@us-es.sel.de 8 rue du Faisan
                                                                                67000 Strasbourg
                                                                                France
--

Return to the comp.compilers page.
Search the comp.compilers archives again.