1 (* Copyright (C) 2005, HELM Team.
3 * This file is part of HELM, an Hypertextual, Electronic
4 * Library of Mathematics, developed at the Computer Science
5 * Department, University of Bologna, Italy.
7 * HELM is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License
9 * as published by the Free Software Foundation; either version 2
10 * of the License, or (at your option) any later version.
12 * HELM is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 * GNU General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with HELM; if not, write to the Free Software
19 * Foundation, Inc., 59 Temple Place - Suite 330, Boston,
22 * For details, see the HELM World-Wide-Web page,
23 * http://helm.cs.unibo.it/
28 exception Error of int * int * string
30 let regexp number = xml_digit+
32 (* ZACK: breaks unicode's binder followed by an ascii letter without blank *)
33 (* let regexp ident_letter = xml_letter *)
35 let regexp ident_letter = [ 'a' - 'z' 'A' - 'Z' ]
37 let regexp ligature_char = [ "'`~!?@*()[]<>-+=|:;.,/\"" ]
38 let regexp ligature = ligature_char ligature_char+
40 let regexp ident_decoration = '\'' | '?' | '`'
41 let regexp ident_cont = ident_letter | xml_digit | '_'
42 let regexp ident = ident_letter ident_cont* ident_decoration*
44 let regexp tex_token = '\\' ident
46 let regexp delim_begin = "\\["
47 let regexp delim_end = "\\]"
49 let regexp qkeyword = "'" ident "'"
51 let regexp implicit = '?'
52 let regexp placeholder = '%'
53 let regexp meta = implicit number
55 let regexp csymbol = '\'' ident
57 let regexp begin_group = "@{" | "${"
58 let regexp end_group = '}'
59 let regexp wildcard = "$_"
60 let regexp ast_ident = "@" ident
61 let regexp ast_csymbol = "@" csymbol
62 let regexp meta_ident = "$" ident
63 let regexp meta_anonymous = "$_"
64 let regexp qstring = '"' [^ '"']* '"'
66 let regexp begincomment = "(**" xml_blank
67 let regexp beginnote = "(*"
68 let regexp endcomment = "*)"
69 (* let regexp comment_char = [^'*'] | '*'[^')']
70 let regexp note = "|+" ([^'*'] | "**") comment_char* "+|" *)
75 "over"; "atop"; "frac";
80 [ "hbox"; "hvbox"; "hovbox"; "vbox";
82 "list0"; "list1"; "sep";
84 "term"; "ident"; "number"
87 let level2_meta_keywords =
88 [ "if"; "then"; "else";
89 "fold"; "left"; "right"; "rec";
92 "anonymous"; "ident"; "number"; "term"; "fresh"
95 (* (string, unit) Hashtbl.t, to exploit multiple bindings *)
96 let level2_ast_keywords = Hashtbl.create 23
98 List.iter (fun k -> Hashtbl.add level2_ast_keywords k ())
99 [ "CProp"; "Prop"; "Type"; "Set"; "let"; "rec"; "corec"; "match";
100 "with"; "in"; "and"; "to"; "as"; "on" ]
102 let add_level2_ast_keyword k = Hashtbl.add level2_ast_keywords k ()
103 let remove_level2_ast_keyword k = Hashtbl.remove level2_ast_keywords k
105 (* (string, int) Hashtbl.t, with multiple bindings.
106 * int is the unicode codepoint *)
107 let ligatures = Hashtbl.create 23
110 (fun (ligature, symbol) -> Hashtbl.add ligatures ligature symbol)
111 [ ("->", <:unicode<to>>); ("=>", <:unicode<Rightarrow>>);
112 ("<=", <:unicode<leq>>); (">=", <:unicode<geq>>);
113 ("<>", <:unicode<neq>>); (":=", <:unicode<def>>);
117 ("cic:/" | "theory:/") (* schema *)
118 ident ('/' ident)* (* path *)
119 ('.' ident)+ (* ext *)
120 ("#xpointer(" number ('/' number)+ ")")? (* xpointer *)
122 let error lexbuf msg =
123 let begin_cnum, end_cnum = Ulexing.loc lexbuf in
124 raise (Error (begin_cnum, end_cnum, msg))
125 let error_at_end lexbuf msg =
126 let begin_cnum, end_cnum = Ulexing.loc lexbuf in
127 raise (Error (begin_cnum, end_cnum, msg))
129 let return_with_loc token begin_cnum end_cnum =
130 (* TODO handle line/column numbers *)
131 let flocation_begin =
132 { Lexing.pos_fname = "";
133 Lexing.pos_lnum = -1; Lexing.pos_bol = -1;
134 Lexing.pos_cnum = begin_cnum }
136 let flocation_end = { flocation_begin with Lexing.pos_cnum = end_cnum } in
137 (token, (flocation_begin, flocation_end))
139 let return lexbuf token =
140 let begin_cnum, end_cnum = Ulexing.loc lexbuf in
141 return_with_loc token begin_cnum end_cnum
143 let return_lexeme lexbuf name = return lexbuf (name, Ulexing.utf8_lexeme lexbuf)
145 let return_symbol lexbuf s = return lexbuf ("SYMBOL", s)
146 let return_eoi lexbuf = return lexbuf ("EOI", "")
148 let remove_quotes s = String.sub s 1 (String.length s - 2)
151 let tok_func stream =
152 (* let lexbuf = Ulexing.from_utf8_stream stream in *)
153 (** XXX Obj.magic rationale.
155 * camlp4 constraints the tok_func field of Token.glexer to have type:
156 * Stream.t char -> (Stream.t 'te * flocation_function)
157 * In order to use ulex we have (in theory) to instantiate a new lexbuf each
158 * time a char Stream.t is passed, destroying the previous lexbuf which may
159 * have consumed a character from the old stream which is lost forever :-(
161 * Instead of passing to camlp4 a char Stream.t we pass a lexbuf, casting it to
162 * char Stream.t with Obj.magic where needed.
164 let lexbuf = Obj.magic stream in
165 Token.make_stream_and_flocation
170 | Ulexing.Error -> error_at_end lexbuf "Unexpected character"
171 | Ulexing.InvalidCodepoint p ->
172 error_at_end lexbuf (sprintf "Invalid code point: %d" p))
175 Token.tok_func = tok_func;
176 Token.tok_using = (fun _ -> ());
177 Token.tok_removing = (fun _ -> ());
178 Token.tok_match = Token.default_match;
179 Token.tok_text = Token.lexer_text;
180 Token.tok_comm = None;
183 let expand_macro lexbuf =
185 Ulexing.utf8_sub_lexeme lexbuf 1 (Ulexing.lexeme_length lexbuf - 1)
188 ("SYMBOL", Utf8Macro.expand macro)
189 with Utf8Macro.Macro_not_found _ -> "SYMBOL", Ulexing.utf8_lexeme lexbuf
191 let remove_quotes s = String.sub s 1 (String.length s - 2)
192 let remove_left_quote s = String.sub s 1 (String.length s - 1)
194 let rec level2_pattern_token_group counter buffer =
197 if (counter > 0) then
198 Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
199 snd (Ulexing.loc lexbuf)
201 Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
202 ignore (level2_pattern_token_group (counter + 1) buffer lexbuf) ;
203 level2_pattern_token_group counter buffer lexbuf
205 Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
206 level2_pattern_token_group counter buffer lexbuf
208 let read_unparsed_group token_name lexbuf =
209 let buffer = Buffer.create 16 in
210 let begin_cnum, _ = Ulexing.loc lexbuf in
211 let end_cnum = level2_pattern_token_group 0 buffer lexbuf in
212 return_with_loc (token_name, Buffer.contents buffer) begin_cnum end_cnum
214 let rec level2_meta_token =
216 | xml_blank+ -> level2_meta_token lexbuf
218 let s = Ulexing.utf8_lexeme lexbuf in
220 if List.mem s level2_meta_keywords then
221 return lexbuf ("", s)
223 return lexbuf ("IDENT", s)
225 | "@{" -> read_unparsed_group "UNPARSED_AST" lexbuf
227 return lexbuf ("UNPARSED_AST",
228 remove_left_quote (Ulexing.utf8_lexeme lexbuf))
230 return lexbuf ("UNPARSED_AST",
231 remove_left_quote (Ulexing.utf8_lexeme lexbuf))
232 | eof -> return_eoi lexbuf
234 let rec comment_token acc depth =
237 let acc = acc ^ Ulexing.utf8_lexeme lexbuf in
238 comment_token acc (depth + 1) lexbuf
240 let acc = acc ^ Ulexing.utf8_lexeme lexbuf in
243 else comment_token acc (depth - 1) lexbuf
245 let acc = acc ^ Ulexing.utf8_lexeme lexbuf in
246 comment_token acc depth lexbuf
248 (** @param k continuation to be invoked when no ligature has been found *)
249 let rec ligatures_token k =
252 let lexeme = Ulexing.utf8_lexeme lexbuf in
253 (match Hashtbl.find_all ligatures lexeme with
254 | [] -> (* ligature not found, rollback and try default lexer *)
255 Ulexing.rollback lexbuf;
257 | ligs -> (* ligatures found, use the default one *)
258 let default_lig = List.hd (List.rev ligs) in
259 return_symbol lexbuf default_lig)
260 | eof -> return_eoi lexbuf
261 | _ -> (* not a ligature, rollback and try default lexer *)
262 Ulexing.rollback lexbuf;
265 and level2_ast_token =
267 | xml_blank+ -> ligatures_token level2_ast_token lexbuf
268 | meta -> return lexbuf ("META", Ulexing.utf8_lexeme lexbuf)
269 | implicit -> return lexbuf ("IMPLICIT", "")
270 | placeholder -> return lexbuf ("PLACEHOLDER", "")
272 let lexeme = Ulexing.utf8_lexeme lexbuf in
273 if Hashtbl.mem level2_ast_keywords lexeme then
274 return lexbuf ("", lexeme)
276 return lexbuf ("IDENT", lexeme)
277 | number -> return lexbuf ("NUMBER", Ulexing.utf8_lexeme lexbuf)
278 | tex_token -> return lexbuf (expand_macro lexbuf)
279 | uri -> return lexbuf ("URI", Ulexing.utf8_lexeme lexbuf)
281 return lexbuf ("QSTRING", remove_quotes (Ulexing.utf8_lexeme lexbuf))
283 return lexbuf ("CSYMBOL", remove_left_quote (Ulexing.utf8_lexeme lexbuf))
284 | "${" -> read_unparsed_group "UNPARSED_META" lexbuf
285 | "@{" -> read_unparsed_group "UNPARSED_AST" lexbuf
286 | '(' -> return lexbuf ("LPAREN", "")
287 | ')' -> return lexbuf ("RPAREN", "")
289 return lexbuf ("UNPARSED_META",
290 remove_left_quote (Ulexing.utf8_lexeme lexbuf))
291 | meta_anonymous -> return lexbuf ("UNPARSED_META", "anonymous")
293 let comment = comment_token (Ulexing.utf8_lexeme lexbuf) 0 lexbuf in
295 Ulexing.utf8_sub_lexeme lexbuf 2 (Ulexing.lexeme_length lexbuf - 4)
297 return lexbuf ("NOTE", comment) *)
298 ligatures_token level2_ast_token lexbuf
299 | begincomment -> return lexbuf ("BEGINCOMMENT","")
300 | endcomment -> return lexbuf ("ENDCOMMENT","")
301 | eof -> return_eoi lexbuf
302 | _ -> return_symbol lexbuf (Ulexing.utf8_lexeme lexbuf)
304 and level1_pattern_token =
306 | xml_blank+ -> ligatures_token level1_pattern_token lexbuf
307 | number -> return lexbuf ("NUMBER", Ulexing.utf8_lexeme lexbuf)
309 let s = Ulexing.utf8_lexeme lexbuf in
311 if List.mem s level1_keywords then
312 return lexbuf ("", s)
314 return lexbuf ("IDENT", s)
316 | tex_token -> return lexbuf (expand_macro lexbuf)
318 return lexbuf ("QKEYWORD", remove_quotes (Ulexing.utf8_lexeme lexbuf))
319 | '(' -> return lexbuf ("LPAREN", "")
320 | ')' -> return lexbuf ("RPAREN", "")
321 | eof -> return_eoi lexbuf
322 | _ -> return_symbol lexbuf (Ulexing.utf8_lexeme lexbuf)
324 let level1_pattern_token = ligatures_token level1_pattern_token
325 let level2_ast_token = ligatures_token level2_ast_token
327 (* API implementation *)
329 let level1_pattern_lexer = mk_lexer level1_pattern_token
330 let level2_ast_lexer = mk_lexer level2_ast_token
331 let level2_meta_lexer = mk_lexer level2_meta_token