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 endcomment = "*)"
68 let regexp comment_char = [^'*'] | '*'[^')']
69 let regexp note = "(*" ([^'*'] | "**") comment_char* "*)"
74 "over"; "atop"; "frac";
79 [ "hbox"; "hvbox"; "hovbox"; "vbox";
81 "list0"; "list1"; "sep";
83 "term"; "ident"; "number"
86 let level2_meta_keywords =
87 [ "if"; "then"; "else";
88 "fold"; "left"; "right"; "rec";
91 "anonymous"; "ident"; "number"; "term"; "fresh"
94 (* (string, unit) Hashtbl.t, to exploit multiple bindings *)
95 let level2_ast_keywords = Hashtbl.create 23
97 List.iter (fun k -> Hashtbl.add level2_ast_keywords k ())
98 [ "CProp"; "Prop"; "Type"; "Set"; "let"; "rec"; "corec"; "using"; "match";
99 "with"; "in"; "and"; "to"; "as"; "on"; "names" ]
101 let add_level2_ast_keyword k = Hashtbl.add level2_ast_keywords k ()
102 let remove_level2_ast_keyword k = Hashtbl.remove level2_ast_keywords k
104 (* (string, int) Hashtbl.t, with multiple bindings.
105 * int is the unicode codepoint *)
106 let ligatures = Hashtbl.create 23
109 (fun (ligature, symbol) -> Hashtbl.add ligatures ligature symbol)
110 [ ("->", <:unicode<to>>); ("=>", <:unicode<Rightarrow>>);
111 ("<=", <:unicode<leq>>); (">=", <:unicode<geq>>);
112 ("<>", <:unicode<neq>>); (":=", <:unicode<def>>);
116 ("cic:/" | "theory:/") (* schema *)
117 ident ('/' ident)* (* path *)
118 ('.' ident)+ (* ext *)
119 ("#xpointer(" number ('/' number)+ ")")? (* xpointer *)
121 let error lexbuf msg =
122 let begin_cnum, end_cnum = Ulexing.loc lexbuf in
123 raise (Error (begin_cnum, end_cnum, msg))
124 let error_at_end lexbuf msg =
125 let begin_cnum, end_cnum = Ulexing.loc lexbuf in
126 raise (Error (begin_cnum, end_cnum, msg))
128 let return_with_loc token begin_cnum end_cnum =
129 (* TODO handle line/column numbers *)
130 let flocation_begin =
131 { Lexing.pos_fname = "";
132 Lexing.pos_lnum = -1; Lexing.pos_bol = -1;
133 Lexing.pos_cnum = begin_cnum }
135 let flocation_end = { flocation_begin with Lexing.pos_cnum = end_cnum } in
136 (token, (flocation_begin, flocation_end))
138 let return lexbuf token =
139 let begin_cnum, end_cnum = Ulexing.loc lexbuf in
140 return_with_loc token begin_cnum end_cnum
142 let return_lexeme lexbuf name = return lexbuf (name, Ulexing.utf8_lexeme lexbuf)
144 let return_symbol lexbuf s = return lexbuf ("SYMBOL", s)
145 let return_eoi lexbuf = return lexbuf ("EOI", "")
147 let remove_quotes s = String.sub s 1 (String.length s - 2)
149 let mk_lexer ?(use_ligatures = false) token =
150 let tok_func stream =
151 (* let lexbuf = Ulexing.from_utf8_stream stream in *)
152 (** XXX Obj.magic rationale.
154 * camlp4 constraints the tok_func field of Token.glexer to have type:
155 * Stream.t char -> (Stream.t 'te * flocation_function)
156 * In order to use ulex we have (in theory) to instantiate a new lexbuf each
157 * time a char Stream.t is passed, destroying the previous lexbuf which may
158 * have consumed a character from the old stream which is lost forever :-(
160 * Instead of passing to camlp4 a char Stream.t we pass a lexbuf, casting it to
161 * char Stream.t with Obj.magic where needed.
163 let lexbuf = Obj.magic stream in
164 Token.make_stream_and_flocation
169 | Ulexing.Error -> error_at_end lexbuf "Unexpected character"
170 | Ulexing.InvalidCodepoint p ->
171 error_at_end lexbuf (sprintf "Invalid code point: %d" p))
174 Token.tok_func = tok_func;
175 Token.tok_using = (fun _ -> ());
176 Token.tok_removing = (fun _ -> ());
177 Token.tok_match = Token.default_match;
178 Token.tok_text = Token.lexer_text;
179 Token.tok_comm = None;
182 let expand_macro lexbuf =
184 Ulexing.utf8_sub_lexeme lexbuf 1 (Ulexing.lexeme_length lexbuf - 1)
187 ("SYMBOL", Utf8Macro.expand macro)
188 with Utf8Macro.Macro_not_found _ -> "SYMBOL", Ulexing.utf8_lexeme lexbuf
190 let remove_quotes s = String.sub s 1 (String.length s - 2)
191 let remove_left_quote s = String.sub s 1 (String.length s - 1)
193 let rec level2_pattern_token_group counter buffer =
196 if (counter > 0) then
197 Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
198 snd (Ulexing.loc lexbuf)
200 Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
201 ignore (level2_pattern_token_group (counter + 1) buffer lexbuf) ;
202 level2_pattern_token_group counter buffer lexbuf
204 Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
205 level2_pattern_token_group counter buffer lexbuf
207 let read_unparsed_group token_name lexbuf =
208 let buffer = Buffer.create 16 in
209 let begin_cnum, _ = Ulexing.loc lexbuf in
210 let end_cnum = level2_pattern_token_group 0 buffer lexbuf in
211 return_with_loc (token_name, Buffer.contents buffer) begin_cnum end_cnum
213 let rec level2_meta_token =
215 | xml_blank+ -> level2_meta_token lexbuf
217 let s = Ulexing.utf8_lexeme lexbuf in
219 if List.mem s level2_meta_keywords then
220 return lexbuf ("", s)
222 return lexbuf ("IDENT", s)
224 | "@{" -> read_unparsed_group "UNPARSED_AST" lexbuf
226 return lexbuf ("UNPARSED_AST",
227 remove_left_quote (Ulexing.utf8_lexeme lexbuf))
229 return lexbuf ("UNPARSED_AST",
230 remove_left_quote (Ulexing.utf8_lexeme lexbuf))
231 | eof -> return_eoi lexbuf
233 (** @param k continuation to be invoked when no ligature has been found *)
234 let rec ligatures_token k =
237 let lexeme = Ulexing.utf8_lexeme lexbuf in
238 (match Hashtbl.find_all ligatures lexeme with
239 | [] -> (* ligature not found, rollback and try default lexer *)
240 Ulexing.rollback lexbuf;
242 | ligs -> (* ligatures found, use the default one *)
243 let default_lig = List.hd (List.rev ligs) in
244 return_symbol lexbuf default_lig)
245 | eof -> return_eoi lexbuf
246 | _ -> (* not a ligature, rollback and try default lexer *)
247 Ulexing.rollback lexbuf;
250 and level2_ast_token =
252 | xml_blank+ -> ligatures_token level2_ast_token lexbuf
253 | meta -> return lexbuf ("META", Ulexing.utf8_lexeme lexbuf)
254 | implicit -> return lexbuf ("IMPLICIT", "")
255 | placeholder -> return lexbuf ("PLACEHOLDER", "")
257 let lexeme = Ulexing.utf8_lexeme lexbuf in
258 if Hashtbl.mem level2_ast_keywords lexeme then
259 return lexbuf ("", lexeme)
261 return lexbuf ("IDENT", lexeme)
262 | number -> return lexbuf ("NUMBER", Ulexing.utf8_lexeme lexbuf)
263 | tex_token -> return lexbuf (expand_macro lexbuf)
264 | uri -> return lexbuf ("URI", Ulexing.utf8_lexeme lexbuf)
266 return lexbuf ("QSTRING", remove_quotes (Ulexing.utf8_lexeme lexbuf))
268 return lexbuf ("CSYMBOL", remove_left_quote (Ulexing.utf8_lexeme lexbuf))
269 | "${" -> read_unparsed_group "UNPARSED_META" lexbuf
270 | "@{" -> read_unparsed_group "UNPARSED_AST" lexbuf
271 | '(' -> return lexbuf ("LPAREN", "")
272 | ')' -> return lexbuf ("RPAREN", "")
274 return lexbuf ("UNPARSED_META",
275 remove_left_quote (Ulexing.utf8_lexeme lexbuf))
276 | meta_anonymous -> return lexbuf ("UNPARSED_META", "anonymous")
279 Ulexing.utf8_sub_lexeme lexbuf 2 (Ulexing.lexeme_length lexbuf - 4)
281 return lexbuf ("NOTE", comment)
282 | begincomment -> return lexbuf ("BEGINCOMMENT","")
283 | endcomment -> return lexbuf ("ENDCOMMENT","")
284 | eof -> return_eoi lexbuf
285 | _ -> return_symbol lexbuf (Ulexing.utf8_lexeme lexbuf)
287 and level1_pattern_token =
289 | xml_blank+ -> ligatures_token level1_pattern_token lexbuf
290 | number -> return lexbuf ("NUMBER", Ulexing.utf8_lexeme lexbuf)
292 let s = Ulexing.utf8_lexeme lexbuf in
294 if List.mem s level1_keywords then
295 return lexbuf ("", s)
297 return lexbuf ("IDENT", s)
299 | tex_token -> return lexbuf (expand_macro lexbuf)
301 return lexbuf ("QKEYWORD", remove_quotes (Ulexing.utf8_lexeme lexbuf))
302 | '(' -> return lexbuf ("LPAREN", "")
303 | ')' -> return lexbuf ("RPAREN", "")
304 | eof -> return_eoi lexbuf
305 | _ -> return_symbol lexbuf (Ulexing.utf8_lexeme lexbuf)
307 let level1_pattern_token = ligatures_token level1_pattern_token
308 let level2_ast_token = ligatures_token level2_ast_token
310 (* API implementation *)
312 let level1_pattern_lexer = mk_lexer level1_pattern_token
313 let level2_ast_lexer = mk_lexer level2_ast_token
314 let level2_meta_lexer = mk_lexer level2_meta_token