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/
30 exception Error of int * int * string
32 let regexp number = xml_digit+
33 let regexp utf8_blank = " " | "\r\n" | "\n" | "\t" | [160] (* this is a nbsp *)
34 let regexp floatwithunit =
35 [ '0' - '9' ] + ["."] [ '0' - '9' ] + ([ 'a' - 'z' ] + | "" )
36 let regexp color = "#" [ '0' - '9' 'a' - 'f' 'A' - 'F' ] [ '0' - '9' 'a' - 'f'
37 'A' - 'F' ] [ '0' - '9' 'a' - 'f' 'A' - 'F' ] [ '0' - '9' 'a' - 'f' 'A' - 'F' ]
38 [ '0' - '9' 'a' - 'f' 'A' - 'F' ] [ '0' - '9' 'a' - 'f' 'A' - 'F' ]
40 (* ZACK: breaks unicode's binder followed by an ascii letter without blank *)
41 (* let regexp ident_letter = xml_letter *)
43 let regexp ident_letter = [ 'a' - 'z' 'A' - 'Z' ]
45 (* must be in sync with "is_ligature_char" below *)
46 let regexp ligature_char = [ "'`~!?@*()[]<>-+=|:;.,/\"" ]
47 let regexp ligature = ligature_char ligature_char+
49 let is_ligature_char =
50 (* must be in sync with "regexp ligature_char" above *)
51 let chars = "'`~!?@*()[]<>-+=|:;.,/\"" in
54 ignore (String.index chars char);
56 with Not_found -> false))
58 let regexp we_proved = "we" utf8_blank+ "proved"
59 let regexp we_have = "we" utf8_blank+ "have"
60 let regexp let_rec = "let" utf8_blank+ "rec"
61 let regexp let_corec = "let" utf8_blank+ "corec"
62 let regexp ident_decoration = '\'' | '?' | '`'
63 let regexp ident_cont = ident_letter | xml_digit | '_'
64 let regexp ident = ident_letter ident_cont* ident_decoration*
66 let regexp tex_token = '\\' ident
68 let regexp delim_begin = "\\["
69 let regexp delim_end = "\\]"
71 let regexp qkeyword = "'" ident "'"
73 let regexp implicit = '?'
74 let regexp placeholder = '%'
75 let regexp meta = implicit number
77 let regexp csymbol = '\'' ident
79 let regexp begin_group = "@{" | "${"
80 let regexp end_group = '}'
81 let regexp wildcard = "$_"
82 let regexp ast_ident = "@" ident
83 let regexp ast_csymbol = "@" csymbol
84 let regexp meta_ident = "$" ident
85 let regexp meta_anonymous = "$_"
86 let regexp qstring = '"' [^ '"']* '"'
88 let regexp begincomment = "(**" utf8_blank
89 let regexp beginnote = "(*"
90 let regexp endcomment = "*)"
91 (* let regexp comment_char = [^'*'] | '*'[^')']
92 let regexp note = "|+" ([^'*'] | "**") comment_char* "+|" *)
97 "over"; "atop"; "frac";
101 let level1_keywords =
102 [ "hbox"; "hvbox"; "hovbox"; "vbox";
104 "list0"; "list1"; "sep";
106 "term"; "ident"; "number"; "mstyle"
109 let level2_meta_keywords =
110 [ "if"; "then"; "elCicNotationParser.se";
111 "fold"; "left"; "right"; "rec";
114 "anonymous"; "ident"; "number"; "term"; "fresh"
117 (* (string, unit) Hashtbl.t, to exploit multiple bindings *)
118 let initial_level2_ast_keywords () = Hashtbl.create 23;;
120 let level2_ast_keywords = ref (initial_level2_ast_keywords ())
122 let initialize_keywords () =
123 List.iter (fun k -> Hashtbl.add !level2_ast_keywords k ())
124 [ "CProp"; "Prop"; "Type"; "Set"; "let"; "match";
125 "with"; "in"; "and"; "to"; "as"; "on"; "return"; "done" ]
128 let _ = initialize_keywords ();;
130 let add_level2_ast_keyword k = Hashtbl.add !level2_ast_keywords k ()
131 let remove_level2_ast_keyword k = Hashtbl.remove !level2_ast_keywords k
133 (* (string, int) Hashtbl.t, with multiple bindings.
134 * int is the unicode codepoint *)
135 let ligatures = Hashtbl.create 23
139 (fun (ligature, symbol) -> Hashtbl.add ligatures ligature symbol)
140 [ ("->", <:unicode<to>>); ("=>", <:unicode<Rightarrow>>);
141 ("<=", <:unicode<leq>>); (">=", <:unicode<geq>>);
142 ("<>", <:unicode<neq>>); (":=", <:unicode<def>>);
143 ("==", <:unicode<equiv>>);
146 let regexp uri_step = [ 'a' - 'z' 'A' - 'Z' '0' - '9' '_' '-' ''' ]+
149 ("cic:/" | "theory:/") (* schema *)
150 (* ident ('/' ident)* |+ path +| *)
151 uri_step ('/' uri_step)* (* path *)
152 ('.' ident)+ (* ext *)
153 ("#xpointer(" number ('/' number)+ ")")? (* xpointer *)
155 let error lexbuf msg =
156 let begin_cnum, end_cnum = Ulexing.loc lexbuf in
157 raise (Error (begin_cnum, end_cnum, msg))
158 let error_at_end lexbuf msg =
159 let begin_cnum, end_cnum = Ulexing.loc lexbuf in
160 raise (Error (begin_cnum, end_cnum, msg))
162 let return_with_loc token begin_cnum end_cnum =
163 let flocation = HExtlib.floc_of_loc (begin_cnum,end_cnum) in
166 let return lexbuf token =
167 let begin_cnum, end_cnum = Ulexing.loc lexbuf in
168 return_with_loc token begin_cnum end_cnum
170 let return_lexeme lexbuf name = return lexbuf (name, Ulexing.utf8_lexeme lexbuf)
172 let return_symbol lexbuf s = return lexbuf ("SYMBOL", s)
173 let return_eoi lexbuf = return lexbuf ("EOI", "")
175 let remove_quotes s = String.sub s 1 (String.length s - 2)
178 let tok_func stream =
179 (* let lexbuf = Ulexing.from_utf8_stream stream in *)
180 (** XXX Obj.magic rationale.
182 * camlp5 constraints the tok_func field of Token.glexer to have type:
183 * Stream.t char -> (Stream.t 'te * flocation_function)
184 * In order to use ulex we have (in theory) to instantiate a new lexbuf each
185 * time a char Stream.t is passed, destroying the previous lexbuf which may
186 * have consumed a character from the old stream which is lost forever :-(
188 * Instead of passing to camlp5 a char Stream.t we pass a lexbuf, casting it to
189 * char Stream.t with Obj.magic where needed.
191 let lexbuf = Obj.magic stream in
192 Token.make_stream_and_location
197 | Ulexing.Error -> error_at_end lexbuf "Unexpected character"
198 | Ulexing.InvalidCodepoint p ->
199 error_at_end lexbuf (sprintf "Invalid code point: %d" p))
202 Token.tok_func = tok_func;
203 Token.tok_using = (fun _ -> ());
204 Token.tok_removing = (fun _ -> ());
205 Token.tok_match = Token.default_match;
206 Token.tok_text = Token.lexer_text;
207 Token.tok_comm = None;
210 let expand_macro lexbuf =
212 Ulexing.utf8_sub_lexeme lexbuf 1 (Ulexing.lexeme_length lexbuf - 1)
215 ("SYMBOL", Utf8Macro.expand macro)
216 with Utf8Macro.Macro_not_found _ -> "SYMBOL", Ulexing.utf8_lexeme lexbuf
218 let remove_quotes s = String.sub s 1 (String.length s - 2)
219 let remove_left_quote s = String.sub s 1 (String.length s - 1)
221 let rec level2_pattern_token_group counter buffer =
224 if (counter > 0) then
225 Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
226 snd (Ulexing.loc lexbuf)
228 Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
229 ignore (level2_pattern_token_group (counter + 1) buffer lexbuf) ;
230 level2_pattern_token_group counter buffer lexbuf
232 Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
233 level2_pattern_token_group counter buffer lexbuf
235 let read_unparsed_group token_name lexbuf =
236 let buffer = Buffer.create 16 in
237 let begin_cnum, _ = Ulexing.loc lexbuf in
238 let end_cnum = level2_pattern_token_group 0 buffer lexbuf in
239 return_with_loc (token_name, Buffer.contents buffer) begin_cnum end_cnum
241 let rec level2_meta_token =
243 | utf8_blank+ -> level2_meta_token lexbuf
245 let s = Ulexing.utf8_lexeme lexbuf in
247 if List.mem s level2_meta_keywords then
248 return lexbuf ("", s)
250 return lexbuf ("IDENT", s)
252 | "@{" -> read_unparsed_group "UNPARSED_AST" lexbuf
254 return lexbuf ("UNPARSED_AST",
255 remove_left_quote (Ulexing.utf8_lexeme lexbuf))
257 return lexbuf ("UNPARSED_AST",
258 remove_left_quote (Ulexing.utf8_lexeme lexbuf))
259 | eof -> return_eoi lexbuf
261 let rec comment_token acc depth =
264 let acc = acc ^ Ulexing.utf8_lexeme lexbuf in
265 comment_token acc (depth + 1) lexbuf
267 let acc = acc ^ Ulexing.utf8_lexeme lexbuf in
270 else comment_token acc (depth - 1) lexbuf
272 let acc = acc ^ Ulexing.utf8_lexeme lexbuf in
273 comment_token acc depth lexbuf
275 (** @param k continuation to be invoked when no ligature has been found *)
276 let rec ligatures_token k =
279 let lexeme = Ulexing.utf8_lexeme lexbuf in
280 (match List.rev (Hashtbl.find_all ligatures lexeme) with
281 | [] -> (* ligature not found, rollback and try default lexer *)
282 Ulexing.rollback lexbuf;
284 | default_lig :: _ -> (* ligatures found, use the default one *)
285 return_symbol lexbuf default_lig)
286 | eof -> return_eoi lexbuf
287 | _ -> (* not a ligature, rollback and try default lexer *)
288 Ulexing.rollback lexbuf;
291 and level2_ast_token =
293 | let_rec -> return lexbuf ("LETREC","")
294 | let_corec -> return lexbuf ("LETCOREC","")
295 | we_proved -> return lexbuf ("WEPROVED","")
296 | we_have -> return lexbuf ("WEHAVE","")
297 | utf8_blank+ -> ligatures_token level2_ast_token lexbuf
299 let s = Ulexing.utf8_lexeme lexbuf in
300 return lexbuf ("META", String.sub s 1 (String.length s - 1))
301 | implicit -> return lexbuf ("IMPLICIT", "")
302 | placeholder -> return lexbuf ("PLACEHOLDER", "")
304 let lexeme = Ulexing.utf8_lexeme lexbuf in
305 if Hashtbl.mem !level2_ast_keywords lexeme then
306 return lexbuf ("", lexeme)
308 return lexbuf ("IDENT", lexeme)
309 | number -> return lexbuf ("NUMBER", Ulexing.utf8_lexeme lexbuf)
310 | tex_token -> return lexbuf (expand_macro lexbuf)
311 | uri -> return lexbuf ("URI", Ulexing.utf8_lexeme lexbuf)
313 return lexbuf ("QSTRING", remove_quotes (Ulexing.utf8_lexeme lexbuf))
315 return lexbuf ("CSYMBOL", remove_left_quote (Ulexing.utf8_lexeme lexbuf))
316 | "${" -> read_unparsed_group "UNPARSED_META" lexbuf
317 | "@{" -> read_unparsed_group "UNPARSED_AST" lexbuf
318 | '(' -> return lexbuf ("LPAREN", "")
319 | ')' -> return lexbuf ("RPAREN", "")
321 return lexbuf ("UNPARSED_META",
322 remove_left_quote (Ulexing.utf8_lexeme lexbuf))
323 | meta_anonymous -> return lexbuf ("UNPARSED_META", "anonymous")
325 let _comment = comment_token (Ulexing.utf8_lexeme lexbuf) 0 lexbuf in
327 Ulexing.utf8_sub_lexeme lexbuf 2 (Ulexing.lexeme_length lexbuf - 4)
329 return lexbuf ("NOTE", comment) *)
330 ligatures_token level2_ast_token lexbuf
331 | begincomment -> return lexbuf ("BEGINCOMMENT","")
332 | endcomment -> return lexbuf ("ENDCOMMENT","")
333 | eof -> return_eoi lexbuf
334 | _ -> return_symbol lexbuf (Ulexing.utf8_lexeme lexbuf)
336 and level1_pattern_token =
338 | utf8_blank+ -> ligatures_token level1_pattern_token lexbuf
339 | number -> return lexbuf ("NUMBER", Ulexing.utf8_lexeme lexbuf)
341 let s = Ulexing.utf8_lexeme lexbuf in
343 if List.mem s level1_keywords then
344 return lexbuf ("", s)
346 return lexbuf ("IDENT", s)
348 | color -> return lexbuf ("COLOR", Ulexing.utf8_lexeme lexbuf)
350 return lexbuf ("FLOATWITHUNIT", Ulexing.utf8_lexeme lexbuf)
351 | tex_token -> return lexbuf (expand_macro lexbuf)
353 return lexbuf ("QKEYWORD", remove_quotes (Ulexing.utf8_lexeme lexbuf))
354 | '(' -> return lexbuf ("LPAREN", "")
355 | ')' -> return lexbuf ("RPAREN", "")
356 | eof -> return_eoi lexbuf
357 | _ -> return_symbol lexbuf (Ulexing.utf8_lexeme lexbuf)
359 let level1_pattern_token = ligatures_token level1_pattern_token
360 let level2_ast_token = ligatures_token level2_ast_token
362 (* API implementation *)
364 let initial_level1_pattern_lexer () = mk_lexer level1_pattern_token
365 let initial_level2_ast_lexer () = mk_lexer level2_ast_token
366 let initial_level2_meta_lexer () = mk_lexer level2_meta_token
369 let level1_pattern_lexer_ref = ref (initial_level1_pattern_lexer ())
370 let level2_ast_lexer_ref = ref (initial_level2_ast_lexer ())
371 let level2_meta_lexer_ref = ref (initial_level2_meta_lexer ())
373 let level1_pattern_lexer () = !level1_pattern_lexer_ref
374 let level2_ast_lexer () = !level2_ast_lexer_ref
375 let level2_meta_lexer () = !level2_meta_lexer_ref
377 let lookup_ligatures lexeme =
380 then [ Utf8Macro.expand (String.sub lexeme 1 (String.length lexeme - 1)) ]
381 else List.rev (Hashtbl.find_all ligatures lexeme)
382 with Invalid_argument _ | Utf8Macro.Macro_not_found _ -> []
385 let history = ref [];;
389 (!level2_ast_keywords,!level1_pattern_lexer_ref,
390 !level2_ast_lexer_ref,!level2_meta_lexer_ref) :: !history;
391 level2_ast_keywords := initial_level2_ast_keywords ();
392 initialize_keywords ();
393 level1_pattern_lexer_ref := initial_level1_pattern_lexer ();
394 level2_ast_lexer_ref := initial_level2_ast_lexer ();
395 level2_meta_lexer_ref := initial_level2_meta_lexer ();
401 | (kwd,pl,al,ml) :: tl ->
402 level2_ast_keywords := kwd;
403 level1_pattern_lexer_ref := pl;
404 level2_ast_lexer_ref := al;
405 level2_meta_lexer_ref := ml;