src/Pure/General/symbol_pos.ML
author wenzelm
Thu Mar 19 15:22:53 2009 +0100 (2009-03-19)
changeset 30586 9674f64a0702
parent 30573 49899f26fbd1
child 32784 1a5dde5079ac
permissions -rw-r--r--
moved basic change_prompt, scan_string, scan_alt_string, scan_quoted to symbol_pos.ML;
scan_comment: recovered change_prompt;
moved read_antiq to outer_lex.ML;
     1 (*  Title:      Pure/General/symbol_pos.ML
     2     Author:     Makarius
     3 
     4 Symbols with explicit position information.
     5 *)
     6 
     7 signature BASIC_SYMBOL_POS =
     8 sig
     9   type T = Symbol.symbol * Position.T
    10   val symbol: T -> Symbol.symbol
    11   val $$$ : Symbol.symbol -> T list -> T list * T list
    12   val ~$$$ : Symbol.symbol -> T list -> T list * T list
    13 end
    14 
    15 signature SYMBOL_POS =
    16 sig
    17   include BASIC_SYMBOL_POS
    18   val content: T list -> string
    19   val untabify_content: T list -> string
    20   val is_eof: T -> bool
    21   val stopper: T Scan.stopper
    22   val !!! : string -> (T list -> 'a) -> T list -> 'a
    23   val change_prompt: ('a -> 'b) -> 'a -> 'b
    24   val scan_pos: T list -> Position.T * T list
    25   val scan_string: T list -> (Position.T * (T list * Position.T)) * T list
    26   val scan_alt_string: T list -> (Position.T * (T list * Position.T)) * T list
    27   val scan_quoted: T list -> T list * T list
    28   val scan_comment: (string -> (T list -> T list * T list) -> T list -> T list * T list) ->
    29     T list -> T list * T list
    30   val scan_comment_body: (string -> (T list -> T list * T list) -> T list -> T list * T list) ->
    31     T list -> T list * T list
    32   val source: Position.T -> (Symbol.symbol, 'a) Source.source ->
    33     (T, Position.T * (Symbol.symbol, 'a) Source.source) Source.source
    34   type text = string
    35   val implode: T list -> text
    36   val range: T list -> Position.range
    37   val implode_range: Position.T -> Position.T -> T list -> text * Position.range
    38   val explode: text * Position.T -> T list
    39 end;
    40 
    41 structure Symbol_Pos: SYMBOL_POS =
    42 struct
    43 
    44 (* type T *)
    45 
    46 type T = Symbol.symbol * Position.T;
    47 
    48 fun symbol ((s, _): T) = s;
    49 
    50 
    51 (* content *)
    52 
    53 val content = implode o map symbol;
    54 
    55 
    56 val tab_width = (8: int);
    57 
    58 fun untabify ("\t", pos) =
    59       (case Position.column_of pos of
    60         SOME n => Symbol.spaces (tab_width - ((n - 1) mod tab_width))
    61       | NONE => Symbol.space)
    62   | untabify (s, _) = s;
    63 
    64 val untabify_content = implode o map untabify;
    65 
    66 
    67 (* stopper *)
    68 
    69 fun mk_eof pos = (Symbol.eof, pos);
    70 val eof = mk_eof Position.none;
    71 
    72 val is_eof = Symbol.is_eof o symbol;
    73 
    74 val stopper =
    75   Scan.stopper (fn [] => eof | inp => mk_eof (List.last inp |-> Position.advance)) is_eof;
    76 
    77 
    78 (* basic scanners *)
    79 
    80 fun !!! text scan =
    81   let
    82     fun get_pos [] = " (past end-of-text!)"
    83       | get_pos ((_, pos) :: _) = Position.str_of pos;
    84 
    85     fun err (syms, msg) =
    86       text ^ get_pos syms ^ " at " ^ Symbol.beginning 10 (map symbol syms) ^
    87       (case msg of NONE => "" | SOME s => "\n" ^ s);
    88   in Scan.!! err scan end;
    89 
    90 fun change_prompt scan = Scan.prompt "# " scan;
    91 
    92 fun $$$ s = Scan.one (fn x => symbol x = s) >> single;
    93 fun ~$$$ s = Scan.one (fn x => symbol x <> s) >> single;
    94 
    95 val scan_pos = Scan.ahead (Scan.one (K true)) >> (fn (_, pos): T => pos);
    96 
    97 
    98 (* Isabelle strings *)
    99 
   100 local
   101 
   102 val char_code =
   103   Scan.one (Symbol.is_ascii_digit o symbol) --
   104   Scan.one (Symbol.is_ascii_digit o symbol) --
   105   Scan.one (Symbol.is_ascii_digit o symbol) :|--
   106   (fn (((a, pos), (b, _)), (c, _)) =>
   107     let val (n, _) = Library.read_int [a, b, c]
   108     in if n <= 255 then Scan.succeed [(chr n, pos)] else Scan.fail end);
   109 
   110 fun scan_str q =
   111   $$$ "\\" |-- !!! "bad escape character in string" ($$$ q || $$$ "\\" || char_code) ||
   112   Scan.one (fn (s, _) => s <> q andalso s <> "\\" andalso Symbol.is_regular s) >> single;
   113 
   114 fun scan_strs q =
   115   (scan_pos --| $$$ q) -- !!! "missing quote at end of string"
   116     (change_prompt ((Scan.repeat (scan_str q) >> flat) -- ($$$ q |-- scan_pos)));
   117 
   118 in
   119 
   120 val scan_string = scan_strs "\"";
   121 val scan_alt_string = scan_strs "`";
   122 
   123 val scan_quoted = Scan.trace (scan_string || scan_alt_string) >> #2;
   124 
   125 end;
   126 
   127 
   128 (* ML-style comments *)
   129 
   130 local
   131 
   132 val scan_cmt =
   133   Scan.depend (fn (d: int) => $$$ "(" @@@ $$$ "*" >> pair (d + 1)) ||
   134   Scan.depend (fn 0 => Scan.fail | d => $$$ "*" @@@ $$$ ")" >> pair (d - 1)) ||
   135   Scan.lift ($$$ "*" --| Scan.ahead (~$$$ ")")) ||
   136   Scan.lift (Scan.one (fn (s, _) => s <> "*" andalso Symbol.is_regular s)) >> single;
   137 
   138 val scan_body = change_prompt (Scan.pass 0 (Scan.repeat scan_cmt >> flat));
   139 
   140 in
   141 
   142 fun scan_comment cut =
   143   $$$ "(" @@@ $$$ "*" @@@ cut "missing end of comment" (scan_body @@@ $$$ "*" @@@ $$$ ")");
   144 
   145 fun scan_comment_body cut =
   146   $$$ "(" |-- $$$ "*" |-- cut "missing end of comment" (scan_body --| $$$ "*" --| $$$ ")");
   147 
   148 end;
   149 
   150 
   151 (* source *)
   152 
   153 fun source pos =
   154   Source.source' pos Symbol.stopper (Scan.bulk (Scan.depend (fn pos =>
   155     Scan.one Symbol.not_eof >> (fn s => (Position.advance s pos, (s, pos)))))) NONE;
   156 
   157 
   158 (* compact representation -- with Symbol.DEL padding *)
   159 
   160 type text = string;
   161 
   162 fun pad [] = []
   163   | pad [(s, _)] = [s]
   164   | pad ((s1, pos1) :: (rest as (s2, pos2) :: _)) =
   165       let
   166         val end_pos1 = Position.advance s1 pos1;
   167         val d = Int.max (0, Position.distance_of end_pos1 pos2);
   168       in s1 :: replicate d Symbol.DEL @ pad rest end;
   169 
   170 val implode = implode o pad;
   171 
   172 fun range (syms as (_, pos) :: _) =
   173       let val pos' = List.last syms |-> Position.advance
   174       in Position.range pos pos' end
   175   | range [] = Position.no_range;
   176 
   177 fun implode_range pos1 pos2 syms =
   178   let val syms' = (("", pos1) :: syms @ [("", pos2)])
   179   in (implode syms', range syms') end;
   180 
   181 fun explode (str, pos) =
   182   fold_map (fn s => fn p => ((s, p), (Position.advance s p)))
   183     (Symbol.explode str) (Position.reset_range pos)
   184   |> #1 |> filter_out (fn (s, _) => s = Symbol.DEL);
   185 
   186 end;
   187 
   188 structure Basic_Symbol_Pos: BASIC_SYMBOL_POS = Symbol_Pos;   (*not open by default*)
   189