Debian Codegraph @

tools/spgen/source/spgen_lexer.mll

raw
1 (*
2 * This file is part of Coccinelle, lincensed under the terms of the GPL v2.
3 * See copyright.txt in the Coccinelle source code for more information.
4 * The Coccinelle source code can be obtained at http://coccinelle.lip6.fr
5 *)
6
7 (* Lexer/parser for spgen config file *)
8
9 {
10
11 (* ------------------------------------------------------------------------- *)
12 (* TYPES *)
13
14 (* the config syntax has pretty low complexity and is very lenient, so we just
15 * interpret the tokens directly and skip the lexical analysis. *)
16
17 type attribute =
18 | Org of string * string list | Report of string * string list
19
20 type part =
21 | Description of string
22 | Limitations of string list
23 | Keywords of string
24 | Confidence of string
25 | Comments of string
26 | Options of string
27 | Authors of string list (* author, affiliation, license in one *)
28 | Url of string
29 | Rule of (string * string option) * attribute list
30
31 (* the table is for line numbers to produce more meaningful error messages *)
32 let table = ref (Array.make 0 (0,0))
33 exception Eof
34
35 let get_start_line lexbuf =
36 match !table.(Lexing.lexeme_start lexbuf) with (l,_) -> l
37
38 let get_column lexbuf =
39 match !table.(Lexing.lexeme_start lexbuf) with (_,c) -> c
40
41 let get_position_str lexbuf =
42 "around line " ^ (string_of_int (get_start_line lexbuf))
43 ^ ", column " ^ (string_of_int (get_column lexbuf))
44
45 let error prob lexbuf =
46 let pos = get_position_str lexbuf in failwith ("Config error: " ^ prob ^ pos)
47
48 let illegal = error "Illegal syntax "
49
50 let split_list delim = Str.split (Str.regexp (" *"^delim^" *"))
51
52 }
53
54 (* ------------------------------------------------------------------------- *)
55 (* LEXING RULES *)
56
57 let sp = [' ' '\t']
58 let ws = ['\n' '\r' '\012' '\013' ' ' '\t']*
59 let notws = [^ '\n' '\r' '\012' '\013' ' ' '\t']*
60 let notnl = [^ '\n' '\r' '\012' '\013']*
61 let equal = (sp* "=" sp*)
62 let colon = (sp* ":" sp*)
63 let percent = (sp* "%" sp*)
64 let letter = ['A'-'Z' 'a'-'z' '_']
65 let number = ['0'-'9']*
66 let cname = letter (letter | number)*
67
68 rule token = parse
69 | ("description"|"d") equal (notnl as d) { Description d }
70 | ("keywords" |"k") equal (notnl as v) { Keywords v }
71 | ("confidence" |"c") equal (notws as c) { Confidence c }
72 | ("comments" |"m") equal (notnl as m) { Comments m }
73 | ("options" |"o") equal (notnl as o) { Options o }
74 | ("url" |"u") equal (notnl as u) { Url u }
75 | ("limitations"|"l") equal (notnl as l) { Limitations (split_list "|" l) }
76 | ("author"|"authors"|"a") equal (notnl as a) { Authors (split_list "|" a) }
77
78 (* for naming unnamed rules. e.g. 8:name = ... for the rule on line 8. *)
79 | (number as oldrule) colon (cname as newrule) equal {
80 Rule ((oldrule, Some newrule), cocci_rule lexbuf)
81 }
82
83 (* standard rules with org and report messages *)
84 | (cname as rulenm) equal {
85 Rule ((rulenm, None), cocci_rule lexbuf)
86 }
87
88 (* comments and whitespace are skipped *)
89 | ws | "//" [^ '\n']* { token lexbuf }
90 | "/*" { comment lexbuf }
91 | eof { raise Eof}
92 | _ { illegal lexbuf }
93
94 and comment = parse
95 | "*/" {token lexbuf}
96 | eof { error "unclosed comment, needs \"*/\"! " lexbuf}
97 | _ {comment lexbuf }
98
99 and cocci_rule = parse
100 | ws* "org" colon {
101 let (a,b) = format_string lexbuf in Org (a,b) :: (cocci_rule lexbuf)
102 }
103 | ws* "report" colon {
104 let (a,b) = format_string lexbuf in Report (a,b) :: (cocci_rule lexbuf)
105 }
106 | _ { [] }
107
108 (* possibly formatted org and report messages *)
109 and format_string = parse
110 | "\"" (notnl as msg) "\"" percent "(" (notnl as lst) ")" {
111 (msg, split_list "," lst)
112 }
113 | "\"" (notnl as msg) "\"" percent (notnl as lst) { (msg,[lst]) }
114 | "\"" (notnl as msg) "\"" { (msg, []) }
115 | _ { illegal lexbuf }