let regexp number = xml_digit+
-let regexp ident_decoration = '\'' | '!' | '?' | '`'
-let regexp ident_cont = xml_letter | xml_digit | '_'
-let regexp ident = xml_letter ident_cont* ident_decoration*
+ (* ZACK: breaks unicode's binder followed by an ascii letter without blank *)
+(* let regexp ident_letter = xml_letter *)
+
+let regexp ident_letter = [ 'a' - 'z' 'A' - 'Z' ]
+
+ (* must be in sync with "is_ligature_char" below *)
+let regexp ligature_char = [ "'`~!?@*()[]<>-+=|:;.,/\"" ]
+let regexp ligature = ligature_char ligature_char+
+
+let is_ligature_char =
+ (* must be in sync with "regexp ligature_char" above *)
+ let chars = "'`~!?@*()[]<>-+=|:;.,/\"" in
+ (fun char ->
+ (try
+ ignore (String.index chars char);
+ true
+ with Not_found -> false))
+
+let regexp ident_decoration = '\'' | '?' | '`'
+let regexp ident_cont = ident_letter | xml_digit | '_'
+let regexp ident = ident_letter ident_cont* ident_decoration*
let regexp tex_token = '\\' ident
-let regexp keyword = '"' ident '"'
+let regexp delim_begin = "\\["
+let regexp delim_end = "\\]"
+
+let regexp qkeyword = "'" ident "'"
let regexp implicit = '?'
+let regexp placeholder = '%'
let regexp meta = implicit number
+let regexp csymbol = '\'' ident
+
+let regexp begin_group = "@{" | "${"
+let regexp end_group = '}'
+let regexp wildcard = "$_"
+let regexp ast_ident = "@" ident
+let regexp ast_csymbol = "@" csymbol
+let regexp meta_ident = "$" ident
+let regexp meta_anonymous = "$_"
+let regexp qstring = '"' [^ '"']* '"'
+
+let regexp begincomment = "(**" xml_blank
+let regexp beginnote = "(*"
+let regexp endcomment = "*)"
+(* let regexp comment_char = [^'*'] | '*'[^')']
+let regexp note = "|+" ([^'*'] | "**") comment_char* "+|" *)
+
+let level1_layouts =
+ [ "sub"; "sup";
+ "below"; "above";
+ "over"; "atop"; "frac";
+ "sqrt"; "root"
+ ]
+
+let level1_keywords =
+ [ "hbox"; "hvbox"; "hovbox"; "vbox";
+ "break";
+ "list0"; "list1"; "sep";
+ "opt";
+ "term"; "ident"; "number"
+ ] @ level1_layouts
+
+let level2_meta_keywords =
+ [ "if"; "then"; "else";
+ "fold"; "left"; "right"; "rec";
+ "fail";
+ "default";
+ "anonymous"; "ident"; "number"; "term"; "fresh"
+ ]
+
+ (* (string, unit) Hashtbl.t, to exploit multiple bindings *)
+let level2_ast_keywords = Hashtbl.create 23
+let _ =
+ List.iter (fun k -> Hashtbl.add level2_ast_keywords k ())
+ [ "CProp"; "Prop"; "Type"; "Set"; "let"; "rec"; "corec"; "match";
+ "with"; "in"; "and"; "to"; "as"; "on"; "return" ]
+
+let add_level2_ast_keyword k = Hashtbl.add level2_ast_keywords k ()
+let remove_level2_ast_keyword k = Hashtbl.remove level2_ast_keywords k
+
+ (* (string, int) Hashtbl.t, with multiple bindings.
+ * int is the unicode codepoint *)
+let ligatures = Hashtbl.create 23
+let _ =
+ List.iter
+ (fun (ligature, symbol) -> Hashtbl.add ligatures ligature symbol)
+ [ ("->", <:unicode<to>>); ("=>", <:unicode<Rightarrow>>);
+ ("<=", <:unicode<leq>>); (">=", <:unicode<geq>>);
+ ("<>", <:unicode<neq>>); (":=", <:unicode<def>>);
+ ]
+
+let regexp uri_step = [ 'a' - 'z' 'A' - 'Z' '0' - '9' '_' '-' ]+
+
let regexp uri =
("cic:/" | "theory:/") (* schema *)
- ident ('/' ident)* (* path *)
+(* ident ('/' ident)* |+ path +| *)
+ uri_step ('/' uri_step)* (* path *)
('.' ident)+ (* ext *)
("#xpointer(" number ('/' number)+ ")")? (* xpointer *)
let begin_cnum, end_cnum = Ulexing.loc lexbuf in
raise (Error (begin_cnum, end_cnum, msg))
-let return lexbuf token =
- let begin_cnum, end_cnum = Ulexing.loc lexbuf in
+let return_with_loc token begin_cnum end_cnum =
(* TODO handle line/column numbers *)
let flocation_begin =
{ Lexing.pos_fname = "";
let flocation_end = { flocation_begin with Lexing.pos_cnum = end_cnum } in
(token, (flocation_begin, flocation_end))
+let return lexbuf token =
+ let begin_cnum, end_cnum = Ulexing.loc lexbuf in
+ return_with_loc token begin_cnum end_cnum
+
+let return_lexeme lexbuf name = return lexbuf (name, Ulexing.utf8_lexeme lexbuf)
+
+let return_symbol lexbuf s = return lexbuf ("SYMBOL", s)
+let return_eoi lexbuf = return lexbuf ("EOI", "")
+
let remove_quotes s = String.sub s 1 (String.length s - 2)
let mk_lexer token =
let tok_func stream =
- let lexbuf = Ulexing.from_utf8_stream stream in
+(* let lexbuf = Ulexing.from_utf8_stream stream in *)
+(** XXX Obj.magic rationale.
+ * The problem.
+ * camlp4 constraints the tok_func field of Token.glexer to have type:
+ * Stream.t char -> (Stream.t 'te * flocation_function)
+ * In order to use ulex we have (in theory) to instantiate a new lexbuf each
+ * time a char Stream.t is passed, destroying the previous lexbuf which may
+ * have consumed a character from the old stream which is lost forever :-(
+ * The "solution".
+ * Instead of passing to camlp4 a char Stream.t we pass a lexbuf, casting it to
+ * char Stream.t with Obj.magic where needed.
+ *)
+ let lexbuf = Obj.magic stream in
Token.make_stream_and_flocation
(fun () ->
try
("SYMBOL", Utf8Macro.expand macro)
with Utf8Macro.Macro_not_found _ -> "SYMBOL", Ulexing.utf8_lexeme lexbuf
-let keyword lexbuf = "KEYWORD", remove_quotes (Ulexing.utf8_lexeme lexbuf)
+let remove_quotes s = String.sub s 1 (String.length s - 2)
+let remove_left_quote s = String.sub s 1 (String.length s - 1)
-(** {2 Level 1 lexer} *)
+let rec level2_pattern_token_group counter buffer =
+ lexer
+ | end_group ->
+ if (counter > 0) then
+ Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
+ snd (Ulexing.loc lexbuf)
+ | begin_group ->
+ Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
+ ignore (level2_pattern_token_group (counter + 1) buffer lexbuf) ;
+ level2_pattern_token_group counter buffer lexbuf
+ | _ ->
+ Buffer.add_string buffer (Ulexing.utf8_lexeme lexbuf) ;
+ level2_pattern_token_group counter buffer lexbuf
-let rec level1_token = lexer
- | xml_blank+ -> level1_token lexbuf
- | ident -> return lexbuf ("IDENT", Ulexing.utf8_lexeme lexbuf)
- | number -> return lexbuf ("NUMBER", Ulexing.utf8_lexeme lexbuf)
- | keyword -> return lexbuf (keyword lexbuf)
- | tex_token -> return lexbuf (expand_macro lexbuf)
- | eof -> return lexbuf ("EOI", "")
- | _ -> return lexbuf ("SYMBOL", Ulexing.utf8_lexeme lexbuf)
+let read_unparsed_group token_name lexbuf =
+ let buffer = Buffer.create 16 in
+ let begin_cnum, _ = Ulexing.loc lexbuf in
+ let end_cnum = level2_pattern_token_group 0 buffer lexbuf in
+ return_with_loc (token_name, Buffer.contents buffer) begin_cnum end_cnum
-(** {2 Level 2 lexer} *)
+let rec level2_meta_token =
+ lexer
+ | xml_blank+ -> level2_meta_token lexbuf
+ | ident ->
+ let s = Ulexing.utf8_lexeme lexbuf in
+ begin
+ if List.mem s level2_meta_keywords then
+ return lexbuf ("", s)
+ else
+ return lexbuf ("IDENT", s)
+ end
+ | "@{" -> read_unparsed_group "UNPARSED_AST" lexbuf
+ | ast_ident ->
+ return lexbuf ("UNPARSED_AST",
+ remove_left_quote (Ulexing.utf8_lexeme lexbuf))
+ | ast_csymbol ->
+ return lexbuf ("UNPARSED_AST",
+ remove_left_quote (Ulexing.utf8_lexeme lexbuf))
+ | eof -> return_eoi lexbuf
-let rec level2_token = lexer
- | xml_blank+ -> level2_token lexbuf
+let rec comment_token acc depth =
+ lexer
+ | beginnote ->
+ let acc = acc ^ Ulexing.utf8_lexeme lexbuf in
+ comment_token acc (depth + 1) lexbuf
+ | endcomment ->
+ let acc = acc ^ Ulexing.utf8_lexeme lexbuf in
+ if depth = 0
+ then acc
+ else comment_token acc (depth - 1) lexbuf
+ | _ ->
+ let acc = acc ^ Ulexing.utf8_lexeme lexbuf in
+ comment_token acc depth lexbuf
+
+ (** @param k continuation to be invoked when no ligature has been found *)
+let rec ligatures_token k =
+ lexer
+ | ligature ->
+ let lexeme = Ulexing.utf8_lexeme lexbuf in
+ (match List.rev (Hashtbl.find_all ligatures lexeme) with
+ | [] -> (* ligature not found, rollback and try default lexer *)
+ Ulexing.rollback lexbuf;
+ k lexbuf
+ | default_lig :: _ -> (* ligatures found, use the default one *)
+ return_symbol lexbuf default_lig)
+ | eof -> return_eoi lexbuf
+ | _ -> (* not a ligature, rollback and try default lexer *)
+ Ulexing.rollback lexbuf;
+ k lexbuf
+
+and level2_ast_token =
+ lexer
+ | xml_blank+ -> ligatures_token level2_ast_token lexbuf
| meta -> return lexbuf ("META", Ulexing.utf8_lexeme lexbuf)
- | implicit -> return lexbuf ("IMPLICIT", Ulexing.utf8_lexeme lexbuf)
- | ident -> return lexbuf ("IDENT", Ulexing.utf8_lexeme lexbuf)
+ | implicit -> return lexbuf ("IMPLICIT", "")
+ | placeholder -> return lexbuf ("PLACEHOLDER", "")
+ | ident ->
+ let lexeme = Ulexing.utf8_lexeme lexbuf in
+ if Hashtbl.mem level2_ast_keywords lexeme then
+ return lexbuf ("", lexeme)
+ else
+ return lexbuf ("IDENT", lexeme)
| number -> return lexbuf ("NUMBER", Ulexing.utf8_lexeme lexbuf)
- | keyword -> return lexbuf (keyword lexbuf)
| tex_token -> return lexbuf (expand_macro lexbuf)
| uri -> return lexbuf ("URI", Ulexing.utf8_lexeme lexbuf)
- | eof -> return lexbuf ("EOI", "")
- | _ -> return lexbuf ("SYMBOL", Ulexing.utf8_lexeme lexbuf)
+ | qstring ->
+ return lexbuf ("QSTRING", remove_quotes (Ulexing.utf8_lexeme lexbuf))
+ | csymbol ->
+ return lexbuf ("CSYMBOL", remove_left_quote (Ulexing.utf8_lexeme lexbuf))
+ | "${" -> read_unparsed_group "UNPARSED_META" lexbuf
+ | "@{" -> read_unparsed_group "UNPARSED_AST" lexbuf
+ | '(' -> return lexbuf ("LPAREN", "")
+ | ')' -> return lexbuf ("RPAREN", "")
+ | meta_ident ->
+ return lexbuf ("UNPARSED_META",
+ remove_left_quote (Ulexing.utf8_lexeme lexbuf))
+ | meta_anonymous -> return lexbuf ("UNPARSED_META", "anonymous")
+ | beginnote ->
+ let comment = comment_token (Ulexing.utf8_lexeme lexbuf) 0 lexbuf in
+(* let comment =
+ Ulexing.utf8_sub_lexeme lexbuf 2 (Ulexing.lexeme_length lexbuf - 4)
+ in
+ return lexbuf ("NOTE", comment) *)
+ ligatures_token level2_ast_token lexbuf
+ | begincomment -> return lexbuf ("BEGINCOMMENT","")
+ | endcomment -> return lexbuf ("ENDCOMMENT","")
+ | eof -> return_eoi lexbuf
+ | _ -> return_symbol lexbuf (Ulexing.utf8_lexeme lexbuf)
+
+and level1_pattern_token =
+ lexer
+ | xml_blank+ -> ligatures_token level1_pattern_token lexbuf
+ | number -> return lexbuf ("NUMBER", Ulexing.utf8_lexeme lexbuf)
+ | ident ->
+ let s = Ulexing.utf8_lexeme lexbuf in
+ begin
+ if List.mem s level1_keywords then
+ return lexbuf ("", s)
+ else
+ return lexbuf ("IDENT", s)
+ end
+ | tex_token -> return lexbuf (expand_macro lexbuf)
+ | qkeyword ->
+ return lexbuf ("QKEYWORD", remove_quotes (Ulexing.utf8_lexeme lexbuf))
+ | '(' -> return lexbuf ("LPAREN", "")
+ | ')' -> return lexbuf ("RPAREN", "")
+ | eof -> return_eoi lexbuf
+ | _ -> return_symbol lexbuf (Ulexing.utf8_lexeme lexbuf)
+
+let level1_pattern_token = ligatures_token level1_pattern_token
+let level2_ast_token = ligatures_token level2_ast_token
(* API implementation *)
-let syntax_pattern_lexer = mk_lexer level1_token
-let ast_pattern_lexer = mk_lexer level2_token
+let level1_pattern_lexer = mk_lexer level1_pattern_token
+let level2_ast_lexer = mk_lexer level2_ast_token
+let level2_meta_lexer = mk_lexer level2_meta_token
+
+let lookup_ligatures lexeme =
+ try
+ if lexeme.[0] = '\\'
+ then [ Utf8Macro.expand (String.sub lexeme 1 (String.length lexeme - 1)) ]
+ else List.rev (Hashtbl.find_all ligatures lexeme)
+ with Invalid_argument _ | Utf8Macro.Macro_not_found _ as exn -> []
projects / helm.git / blobdiff