README
1re2c
2----
3
4Version 0.9.1
5Originally written by Peter Bumbulis (peterr@csg.uwaterloo.ca)
6Currently maintained by Brian Young (bayoung@acm.org)
7
8The re2c distribution can be found at:
9
10 http://www.tildeslash.org/re2c/index.html
11
12The source distribution is available from:
13
14 http://www.tildeslash.org/re2c/re2c-0.9.1.tar.gz
15
16This distribution is a cleaned up version of the 0.5 release
17maintained by me (Brian Young). Several bugs were fixed as well
18as code cleanup for warning free compilation. It has been developed
19and tested with egcs 1.0.2 and gcc 2.7.2.3 on Linux x86. Peter
20Bumbulis' original release can be found at:
21
22 ftp://csg.uwaterloo.ca/pub/peterr/re2c.0.5.tar.gz
23
24re2c is a great tool for writing fast and flexible lexers. It has
25served many people well for many years and it deserves to be
26maintained more actively. re2c is on the order of 2-3 times faster
27than a flex based scanner, and its input model is much more
28flexible.
29
30Patches and requests for features will be entertained. Areas of
31particular interest to me are porting (a Solaris and an NT
32version will be forthcoming) and wide character support. Note
33that the code is already quite portable and should be buildable
34on any platform with minor makefile changes.
35
36Peter's original version 0.5 ANNOUNCE and README follows.
37
38Brian
39
40--
41
42re2c is a tool for generating C-based recognizers from regular
43expressions. re2c-based scanners are efficient: for programming
44languages, given similar specifications, an re2c-based scanner is
45typically almost twice as fast as a flex-based scanner with little or no
46increase in size (possibly a decrease on cisc architectures). Indeed,
47re2c-based scanners are quite competitive with hand-crafted ones.
48
49Unlike flex, re2c does not generate complete scanners: the user must
50supply some interface code. While this code is not bulky (about 50-100
51lines for a flex-like scanner; see the man page and examples in the
52distribution) careful coding is required for efficiency (and
53correctness). One advantage of this arrangement is that the generated
54code is not tied to any particular input model. For example, re2c
55generated code can be used to scan data from a null-byte terminated
56buffer as illustrated below.
57
58Given the following source
59
60 #define NULL ((char*) 0)
61 char *scan(char *p){
62 char *q;
63 #define YYCTYPE char
64 #define YYCURSOR p
65 #define YYLIMIT p
66 #define YYMARKER q
67 #define YYFILL(n)
68 /*!re2c
69 [0-9]+ {return YYCURSOR;}
70 [\000-\377] {return NULL;}
71 */
72 }
73
74re2c will generate
75
76 /* Generated by re2c on Sat Apr 16 11:40:58 1994 */
77 #line 1 "simple.re"
78 #define NULL ((char*) 0)
79 char *scan(char *p){
80 char *q;
81 #define YYCTYPE char
82 #define YYCURSOR p
83 #define YYLIMIT p
84 #define YYMARKER q
85 #define YYFILL(n)
86 {
87 YYCTYPE yych;
88 unsigned int yyaccept;
89 goto yy0;
90 yy1: ++YYCURSOR;
91 yy0:
92 if((YYLIMIT - YYCURSOR) < 2) YYFILL(2);
93 yych = *YYCURSOR;
94 if(yych <= '/') goto yy4;
95 if(yych >= ':') goto yy4;
96 yy2: yych = *++YYCURSOR;
97 goto yy7;
98 yy3:
99 #line 10
100 {return YYCURSOR;}
101 yy4: yych = *++YYCURSOR;
102 yy5:
103 #line 11
104 {return NULL;}
105 yy6: ++YYCURSOR;
106 if(YYLIMIT == YYCURSOR) YYFILL(1);
107 yych = *YYCURSOR;
108 yy7: if(yych <= '/') goto yy3;
109 if(yych <= '9') goto yy6;
110 goto yy3;
111 }
112 #line 12
113
114 }
115
116Note that most compilers will perform dead-code elimination to remove
117all YYCURSOR, YYLIMIT comparisions.
118
119re2c was developed for a particular project (constructing a fast REXX
120scanner of all things!) and so while it has some rough edges, it should
121be quite usable. More information about re2c can be found in the
122(admittedly skimpy) man page; the algorithms and heuristics used are
123described in an upcoming LOPLAS article (included in the distribution).
124Probably the best way to find out more about re2c is to try the supplied
125examples. re2c is written in C++, and is currently being developed
126under Linux using gcc 2.5.8.
127
128Peter
129
130--
131
132re2c is distributed with no warranty whatever. The code is certain to
133contain errors. Neither the author nor any contributor takes
134responsibility for any consequences of its use.
135
136re2c is in the public domain. The data structures and algorithms used
137in re2c are all either taken from documents available to the general
138public or are inventions of the author. Programs generated by re2c may
139be distributed freely. re2c itself may be distributed freely, in source
140or binary, unchanged or modified. Distributors may charge whatever fees
141they can obtain for re2c.
142
143If you do make use of re2c, or incorporate it into a larger project an
144acknowledgement somewhere (documentation, research report, etc.) would
145be appreciated.
146
147Please send bug reports and feedback (including suggestions for
148improving the distribution) to
149
150 peterr@csg.uwaterloo.ca
151
152Include a small example and the banner from parser.y with bug reports.
153
154