aboutsummaryrefslogtreecommitdiff
path: root/lib/lexer.ml
blob: a91bfdbaee94ff18db51856ef0066dbacd3419eb (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
(** 
    Lexer using sedlex
 *)

open Tokens

exception UnclosedQuote of { content : string; line : int }
exception LexError of Lexing.position * string
exception EOF

let pp_pos out { Lexing.pos_lnum; pos_cnum; pos_bol; _ } =
  Format.fprintf out "line %d:%d" pos_lnum (pos_cnum - pos_bol)

(* Extract the location name from the pattern *)
let location_name = Str.regexp {|.* \(.*\)|}

module type Encoding = sig
  val lexeme : Sedlexing.lexbuf -> string
end

(* The comment system is terrible. The same symbol can be used for :
     - starting a comment
     - inequality operation
   In order to manage this, I try to identify the context in a very basic way,
   using a flag True False for determining the token to send.
*)
module Bucket = Ephemeron.K1.Bucket

type bucket = (Sedlexing.lexbuf, int) Bucket.t

let is_expression : bucket = Bucket.make ()

let incr_level lexbuf =
  match Bucket.find is_expression lexbuf with
  | None -> Bucket.add is_expression lexbuf 1
  | Some v -> Bucket.add is_expression lexbuf (v + 1)

let decr_level lexbuf =
  match Bucket.find is_expression lexbuf with
  | None -> ()
  | Some v ->
      if v > 1 then Bucket.add is_expression lexbuf (v - 1)
      else Bucket.remove is_expression lexbuf

let build_ident lexbuf =
  let id =
    Sedlexing.lexeme lexbuf |> Idents.of_uchars |> String.uppercase_ascii
  in
  try
    let value = Hashtbl.find Idents.keyword_table id in
    let _ = match value with IF | ELIF -> incr_level lexbuf | _ -> () in
    value
  with Not_found -> IDENT id

let incr_level lexbuf =
  match Bucket.find is_expression lexbuf with
  | None -> Bucket.add is_expression lexbuf 1
  | Some v -> Bucket.add is_expression lexbuf (v + 1)

let wait_balance : (Buffer.t -> Lexbuf.t -> 'a) -> Lexbuf.t -> 'a =
 fun rule lexbuf ->
  try[@warning "-52"]
    let token = rule (Buffer.create 256) lexbuf in
    token
  with Failure "lexing: empty token" ->
    let position, _ = Lexbuf.positions lexbuf in
    let line = position.Lexing.pos_lnum and content = Lexbuf.content lexbuf in
    (raise (UnclosedQuote { line; content }) [@warning "+52"])

let space = [%sedlex.regexp? ' ' | '\t']
let eol = [%sedlex.regexp? '\r' | '\n' | "\r\n"]
let coma = [%sedlex.regexp? ',']
let digit = [%sedlex.regexp? '0' .. '9']
let letters = [%sedlex.regexp? 'a' .. 'z' | 'A' .. 'Z' | '_']
let spaces = [%sedlex.regexp? Plus space]
let ident = [%sedlex.regexp? ('$' | letters), Star (digit | letters)]
let location_ident = [%sedlex.regexp? letters | digit]
let location_prefix = [%sedlex.regexp? '!' | '$' | '#' | '^']
let location = [%sedlex.regexp? Opt location_prefix, Plus location_ident]

let rec read_long_string level buf buffer =
  let lexbuf = Lexbuf.buffer buffer in
  match%sedlex lexbuf with
  | '{' ->
      Buffer.add_string buf (Sedlexing.Utf8.lexeme lexbuf);
      read_long_string (level + 1) buf buffer
  | '}' -> (
      match level with
      | 0 -> Buffer.contents buf
      | _ ->
          Buffer.add_string buf (Sedlexing.Utf8.lexeme lexbuf);
          read_long_string (level - 1) buf buffer)
  | eol ->
      Buffer.add_string buf (Lexbuf.content buffer);
      read_long_string level buf buffer
  | any ->
      Buffer.add_string buf (Lexbuf.content buffer);
      read_long_string level buf buffer
  | _ -> raise Not_found

let rec read_dquoted_string buf buffer =
  let lexbuf = Lexbuf.buffer buffer in
  match%sedlex lexbuf with
  | "\"\"" ->
      Buffer.add_char buf '"';
      read_dquoted_string buf buffer
  | '"' -> Buffer.contents buf
  | any ->
      Buffer.add_string buf (Sedlexing.Utf8.lexeme lexbuf);
      read_dquoted_string buf buffer
  | _ -> raise Not_found

let rec read_quoted_string buf buffer =
  let lexbuf = Lexbuf.buffer buffer in
  match%sedlex lexbuf with
  | "''" ->
      Buffer.add_char buf '\'';
      read_quoted_string buf buffer
  | '\'' -> Buffer.contents buf
  | eol ->
      Buffer.add_string buf (Lexbuf.content buffer);
      read_quoted_string buf buffer
  | any ->
      Buffer.add_string buf (Lexbuf.content buffer);
      read_quoted_string buf buffer
  | _ -> raise Not_found

let rec skip_comment buffer =
  let lexbuf = Lexbuf.buffer buffer in
  match%sedlex lexbuf with
  | '{' ->
      let _ = wait_balance (read_long_string 0) buffer in
      skip_comment buffer
  | '\'' ->
      let _ = wait_balance read_quoted_string buffer in
      skip_comment buffer
  | '"' ->
      let _ = wait_balance read_dquoted_string buffer in
      skip_comment buffer
  | eol ->
      (* Ugly hack used in order to put the eol in the front of the next
         parsing. *)
      Sedlexing.rollback lexbuf;
      COMMENT
  | any -> skip_comment buffer
  | _ -> raise Not_found

(** Main lexer *)
let rec token : Lexbuf.t -> token =
 fun buffer ->
  let lexbuf = Lexbuf.buffer buffer in
  match%sedlex lexbuf with
  | 0Xfeff ->
      Lexbuf.start buffer;
      (* Ignore the BOM *)
      token buffer
  | '#', Star space, location ->
      let _start_pos, end_pos = Sedlexing.lexing_positions lexbuf in

      (* Extract the location name *)
      let ident = Lexbuf.content buffer in
      let () =
        match Str.string_match location_name ident 0 with
        | false -> ()
        | true -> Sedlexing.set_filename lexbuf (Str.matched_group 1 ident)
      in

      (* Restart the line number (new location here) *)
      Sedlexing.start lexbuf;
      Sedlexing.set_position lexbuf { end_pos with Lexing.pos_lnum = 1 };

      LOCATION_START ident
  | '-', Plus '-', Star (Sub (any, ('\r' | '\n'))) ->
      Bucket.remove is_expression lexbuf;
      LOCATION_END
  | Plus digit -> INTEGER (Lexbuf.content buffer)
  | '+' -> PLUS
  | '-' -> MINUS
  | "+=" -> INCR
  | "-=" -> DECR
  | "*=" -> MULT_EQUAL
  | '/' -> DIV
  | '*' -> STAR
  | ':' ->
      (* We are leaving the block, the comment will be handled again *)
      decr_level lexbuf;
      COLUMN
  | '[' -> L_BRACKET
  | ']' -> R_BRACKET
  | '(' ->
      incr_level lexbuf;
      L_PAREN
  | ')' ->
      decr_level lexbuf;
      R_PAREN
  | '<' -> LT
  | '>' -> GT
  | coma -> COMA
  | '=' ->
      incr_level lexbuf;
      EQUAL
  | ident -> build_ident lexbuf
  | eol ->
      Bucket.add is_expression lexbuf 0;
      EOL
  | '&' ->
      Bucket.add is_expression lexbuf 0;
      AMPERSAND
  | '!' -> (
      match Bucket.find is_expression lexbuf with
      | Some i when i <> 0 -> EXCLAMATION
      | _ -> skip_comment buffer)
  | spaces -> token buffer
  | '\'' -> LITERAL (wait_balance read_quoted_string buffer)
  | '"' -> LITERAL (wait_balance read_dquoted_string buffer)
  | '{' -> LITERAL (wait_balance (read_long_string 0) buffer)
  | eof -> raise EOF
  | _ ->
      let position = fst @@ Sedlexing.lexing_positions lexbuf in
      let tok = Lexbuf.content buffer in

      let msg =
        Format.asprintf "Unexpected character %S at %a" tok pp_pos position
      in

      raise @@ LexError (position, msg)

let rec discard buffer =
  let lexbuf = Lexbuf.buffer buffer in
  match%sedlex lexbuf with
  | '\'' ->
      ignore (wait_balance read_quoted_string buffer);
      discard buffer
  | '"' ->
      ignore (wait_balance read_dquoted_string buffer);
      discard buffer
  | '{' ->
      ignore (wait_balance (read_long_string 0) buffer);
      discard buffer
  | eof -> raise EOF
  | '-', Plus '-', Star (Sub (any, ('\r' | '\n'))) ->
      Bucket.remove is_expression lexbuf;
      ()
  | '!' ->
      ignore @@ skip_comment buffer;
      discard buffer
  | any -> discard buffer
  | _ -> raise EOF