Lexer.x 100 KB
Newer Older
1
-----------------------------------------------------------------------------
2
-- (c) The University of Glasgow, 2006
3
--
4
-- GHC's lexer for Haskell 2010 [1].
5
--
6 7
-- This is a combination of an Alex-generated lexer [2] from a regex
-- definition, with some hand-coded bits. [3]
8 9
--
-- Completely accurate information about token-spans within the source
Ian Lynagh's avatar
Ian Lynagh committed
10 11
-- file is maintained.  Every token has a start and end RealSrcLoc
-- attached to it.
12
--
13 14 15 16 17
-- References:
-- [1] https://www.haskell.org/onlinereport/haskell2010/haskellch2.html
-- [2] http://www.haskell.org/alex/
-- [3] https://ghc.haskell.org/trac/ghc/wiki/Commentary/Compiler/Parser
--
18 19 20 21 22 23 24 25 26 27
-----------------------------------------------------------------------------

--   ToDo / known bugs:
--    - parsing integers is a bit slow
--    - readRational is a bit slow
--
--   Known bugs, that were also in the previous version:
--    - M... should be 3 tokens, not 1.
--    - pragma-end should be only valid in a pragma

28
--   qualified operator NOTES.
Ian Lynagh's avatar
Ian Lynagh committed
29
--
30 31 32 33 34 35 36 37 38
--   - If M.(+) is a single lexeme, then..
--     - Probably (+) should be a single lexeme too, for consistency.
--       Otherwise ( + ) would be a prefix operator, but M.( + ) would not be.
--     - But we have to rule out reserved operators, otherwise (..) becomes
--       a different lexeme.
--     - Should we therefore also rule out reserved operators in the qualified
--       form?  This is quite difficult to achieve.  We don't do it for
--       qualified varids.

39 40 41 42

-- -----------------------------------------------------------------------------
-- Alex "Haskell code fragment top"

43
{
44
-- XXX The above flags turn off warnings in the generated code:
45
{-# LANGUAGE BangPatterns #-}
Alan Zimmerman's avatar
Alan Zimmerman committed
46
{-# LANGUAGE DeriveDataTypeable #-}
47 48 49 50 51 52 53 54
{-# OPTIONS_GHC -fno-warn-unused-matches #-}
{-# OPTIONS_GHC -fno-warn-unused-binds #-}
{-# OPTIONS_GHC -fno-warn-unused-imports #-}
{-# OPTIONS_GHC -fno-warn-missing-signatures #-}
-- But alex still generates some code that causes the "lazy unlifted bindings"
-- warning, and old compilers don't know about it so we can't easily turn
-- it off, so for now we use the sledge hammer:
{-# OPTIONS_GHC -w #-}
55

56 57
{-# OPTIONS_GHC -funbox-strict-fields #-}

58
module Lexer (
59
   Token(..), SourceText, lexer, pragState, mkPState, PState(..),
Ian Lynagh's avatar
Ian Lynagh committed
60
   P(..), ParseResult(..), getSrcLoc,
61
   getPState, getDynFlags, withThisPackage,
62
   failLocMsgP, failSpanMsgP, srcParseFail,
Ian Lynagh's avatar
Ian Lynagh committed
63
   getMessages,
64
   popContext, pushCurrentContext, setLastToken, setSrcLoc,
vivian's avatar
vivian committed
65
   activeContext, nextIsEOF,
simonpj@microsoft.com's avatar
simonpj@microsoft.com committed
66
   getLexState, popLexState, pushLexState,
67
   extension, bangPatEnabled, datatypeContextsEnabled,
68
   traditionalRecordSyntaxEnabled,
69 70
   explicitForallEnabled,
   inRulePrag,
cactus's avatar
cactus committed
71 72 73
   explicitNamespacesEnabled,
   patternSynonymsEnabled,
   sccProfilingOn, hpcEnabled,
Jedai's avatar
Jedai committed
74
   addWarning,
Alan Zimmerman's avatar
Alan Zimmerman committed
75 76
   lexTokenStream,
   addAnnotation
77 78
  ) where

rodlogic's avatar
rodlogic committed
79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95
-- base
import Control.Applicative
import Control.Monad
import Data.Bits
import Data.Char
import Data.List
import Data.Maybe
import Data.Ratio
import Data.Word

-- bytestring
import Data.ByteString (ByteString)

-- containers
import Data.Map (Map)
import qualified Data.Map as Map

Alan Zimmerman's avatar
Alan Zimmerman committed
96 97 98 99
-- data/typeable
import Data.Data
import Data.Typeable

rodlogic's avatar
rodlogic committed
100
-- compiler/utils
101
import Bag
102 103 104 105
import Outputable
import StringBuffer
import FastString
import UniqFM
rodlogic's avatar
rodlogic committed
106 107 108 109
import Util             ( readRational )

-- compiler/main
import ErrUtils
110
import DynFlags
rodlogic's avatar
rodlogic committed
111 112 113

-- compiler/basicTypes
import SrcLoc
114
import Module
Ian Lynagh's avatar
Ian Lynagh committed
115
import BasicTypes       ( InlineSpec(..), RuleMatchInfo(..), FractionalLit(..) )
116

rodlogic's avatar
rodlogic committed
117 118
-- compiler/parser
import Ctype
Alan Zimmerman's avatar
Alan Zimmerman committed
119 120

import ApiAnnotation
121 122
}

123 124 125
-- -----------------------------------------------------------------------------
-- Alex "Character set macros"

126 127
-- NB: The logic behind these definitions is also reflected in basicTypes/Lexeme.hs
-- Any changes here should likely be reflected there.
128 129 130 131
$unispace    = \x05 -- Trick Alex into handling Unicode. See alexGetByte.
$nl          = [\n\r\f]
$whitechar   = [$nl\v\ $unispace]
$white_no_nl = $whitechar # \n -- TODO #8424
132
$tab         = \t
133 134

$ascdigit  = 0-9
135
$unidigit  = \x03 -- Trick Alex into handling Unicode. See alexGetByte.
136
$decdigit  = $ascdigit -- for now, should really be $digit (ToDo)
137 138 139
$digit     = [$ascdigit $unidigit]

$special   = [\(\)\,\;\[\]\`\{\}]
140 141 142
$ascsymbol = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~\:]
$unisymbol = \x04 -- Trick Alex into handling Unicode. See alexGetByte.
$symbol    = [$ascsymbol $unisymbol] # [$special \_\"\']
143

144
$unilarge  = \x01 -- Trick Alex into handling Unicode. See alexGetByte.
145
$asclarge  = [A-Z]
146 147
$large     = [$asclarge $unilarge]

148
$unismall  = \x02 -- Trick Alex into handling Unicode. See alexGetByte.
149
$ascsmall  = [a-z]
150 151
$small     = [$ascsmall $unismall \_]

152 153
$unigraphic = \x06 -- Trick Alex into handling Unicode. See alexGetByte.
$graphic   = [$small $large $symbol $digit $special $unigraphic \"\']
154

155
$binit     = 0-1
Ian Lynagh's avatar
Ian Lynagh committed
156
$octit     = 0-7
157
$hexit     = [$decdigit A-F a-f]
158 159
$idchar    = [$small $large $digit \']

160 161
$pragmachar = [$small $large $digit]

162 163
$docsym    = [\| \^ \* \$]

164

165 166 167 168 169 170 171 172
-- -----------------------------------------------------------------------------
-- Alex "Regular expression macros"

@varid     = $small $idchar*          -- variable identifiers
@conid     = $large $idchar*          -- constructor identifiers

@varsym    = ($symbol # \:) $symbol*  -- variable (operator) symbol
@consym    = \: $symbol*              -- constructor (operator) symbol
173

174
@decimal     = $decdigit+
175
@binary      = $binit+
176 177 178 179 180
@octal       = $octit+
@hexadecimal = $hexit+
@exponent    = [eE] [\-\+]? @decimal

@qual = (@conid \.)+
181 182 183 184
@qvarid = @qual @varid
@qconid = @qual @conid
@qvarsym = @qual @varsym
@qconsym = @qual @consym
185 186 187

@floating_point = @decimal \. @decimal @exponent? | @decimal @exponent

188 189 190 191 192
-- normal signed numerical literals can only be explicitly negative,
-- not explicitly positive (contrast @exponent)
@negative = \-
@signed = @negative ?

193 194 195 196

-- -----------------------------------------------------------------------------
-- Alex "Identifier"

197 198
haskell :-

199 200 201 202 203

-- -----------------------------------------------------------------------------
-- Alex "Rules"

-- everywhere: skip whitespace
Ian Lynagh's avatar
Ian Lynagh committed
204
$white_no_nl+ ;
205
$tab+         { warn Opt_WarnTabs (text "Tab character") }
206 207 208 209 210

-- Everywhere: deal with nested comments.  We explicitly rule out
-- pragmas, "{-#", so that we don't accidentally treat them as comments.
-- (this can happen even though pragmas will normally take precedence due to
-- longest-match, because pragmas aren't valid in every state, but comments
211 212 213 214
-- are). We also rule out nested Haddock comments, if the -haddock flag is
-- set.

"{-" / { isNormalComment } { nested_comment lexToken }
215 216 217 218

-- Single-line comments are a bit tricky.  Haskell 98 says that two or
-- more dashes followed by a symbol should be parsed as a varsym, so we
-- have to exclude those.
219 220

-- Since Haddock comments aren't valid in every state, we need to rule them
Ian Lynagh's avatar
Ian Lynagh committed
221
-- out here.
222 223 224 225 226 227 228

-- The following two rules match comments that begin with two dashes, but
-- continue with a different character. The rules test that this character
-- is not a symbol (in which case we'd have a varsym), and that it's not a
-- space followed by a Haddock comment symbol (docsym) (in which case we'd
-- have a Haddock comment). The rules then munch the rest of the line.

Jedai's avatar
Jedai committed
229
"-- " ~[$docsym \#] .* { lineCommentToken }
230
"--" [^$symbol \ ] .* { lineCommentToken }
231 232 233

-- Next, match Haddock comments if no -haddock flag

Jedai's avatar
Jedai committed
234
"-- " [$docsym \#] .* / { ifExtension (not . haddockEnabled) } { lineCommentToken }
235 236 237 238 239 240 241

-- Now, when we've matched comments that begin with 2 dashes and continue
-- with a different character, we need to match comments that begin with three
-- or more dashes (which clearly can't be Haddock comments). We only need to
-- make sure that the first non-dash character isn't a symbol, and munch the
-- rest of the line.

242
"---"\-* ~$symbol .* { lineCommentToken }
243 244 245 246

-- Since the previous rules all match dashes followed by at least one
-- character, we also need to match a whole line filled with just dashes.

Jedai's avatar
Jedai committed
247
"--"\-* / { atEOL } { lineCommentToken }
248 249 250 251

-- We need this rule since none of the other single line comment rules
-- actually match this case.

Jedai's avatar
Jedai committed
252
"-- " / { atEOL } { lineCommentToken }
253 254 255 256 257 258 259 260 261 262 263

-- 'bol' state: beginning of a line.  Slurp up all the whitespace (including
-- blank lines) until we find a non-whitespace character, then do layout
-- processing.
--
-- One slight wibble here: what if the line begins with {-#? In
-- theory, we have to lex the pragma to see if it's one we recognise,
-- and if it is, then we backtrack and do_bol, otherwise we treat it
-- as a nested comment.  We don't bother with this: if the line begins
-- with {-#, then we'll assume it's a pragma we know about and go for do_bol.
<bol> {
Ian Lynagh's avatar
Ian Lynagh committed
264 265 266 267 268
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
  ^\# pragma .* \n                      ; -- GCC 3.3 CPP generated, apparently
  ^\# \! .* \n                          ; -- #!, for scripts
  ()                                    { do_bol }
269 270 271 272 273
}

-- after a layout keyword (let, where, do, of), we begin a new layout
-- context if the curly brace is missing.
-- Careful! This stuff is quite delicate.
274
<layout, layout_do, layout_if> {
Ian Lynagh's avatar
Ian Lynagh committed
275 276 277 278
  \{ / { notFollowedBy '-' }            { hopefully_open_brace }
        -- we might encounter {-# here, but {- has been handled already
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
279 280
}

281 282 283 284 285 286
-- after an 'if', a vertical bar starts a layout context for MultiWayIf
<layout_if> {
  \| / { notFollowedBySymbol }          { new_layout_context True ITvbar }
  ()                                    { pop }
}

287
-- do is treated in a subtly different way, see new_layout_context
288 289
<layout>    ()                          { new_layout_context True  ITvocurly }
<layout_do> ()                          { new_layout_context False ITvocurly }
290 291 292 293

-- after a new layout context which was found to be to the left of the
-- previous context, we have generated a '{' token, and we now need to
-- generate a matching '}' token.
Ian Lynagh's avatar
Ian Lynagh committed
294
<layout_left>  ()                       { do_layout_left }
295

Ian Lynagh's avatar
Ian Lynagh committed
296
<0,option_prags> \n                     { begin bol }
297

298 299
"{-#" $whitechar* $pragmachar+ / { known_pragma linePrags }
                                { dispatch_pragmas linePrags }
300 301 302

-- single-line line pragmas, of the form
--    # <line> "<file>" <extra-stuff> \n
303
<line_prag1> @decimal                   { setLine line_prag1a }
Ian Lynagh's avatar
Ian Lynagh committed
304 305
<line_prag1a> \" [$graphic \ ]* \"      { setFile line_prag1b }
<line_prag1b> .*                        { pop }
306 307 308

-- Haskell-style line pragmas, of the form
--    {-# LINE <line> "<file>" #-}
309
<line_prag2> @decimal                   { setLine line_prag2a }
Ian Lynagh's avatar
Ian Lynagh committed
310 311
<line_prag2a> \" [$graphic \ ]* \"      { setFile line_prag2b }
<line_prag2b> "#-}"|"-}"                { pop }
312 313
   -- NOTE: accept -} at the end of a LINE pragma, for compatibility
   -- with older versions of GHC which generated these.
314

315
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
316
  "{-#" $whitechar* $pragmachar+
317 318 319 320 321
        $whitechar+ $pragmachar+ / { known_pragma twoWordPrags }
                                 { dispatch_pragmas twoWordPrags }

  "{-#" $whitechar* $pragmachar+ / { known_pragma oneWordPrags }
                                 { dispatch_pragmas oneWordPrags }
322

323
  -- We ignore all these pragmas, but don't generate a warning for them
324 325
  "{-#" $whitechar* $pragmachar+ / { known_pragma ignoredPrags }
                                 { dispatch_pragmas ignoredPrags }
326 327

  -- ToDo: should only be valid inside a pragma:
Ian Lynagh's avatar
Ian Lynagh committed
328
  "#-}"                          { endPrag }
329 330
}

331
<option_prags> {
332 333 334
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
                                   { dispatch_pragmas fileHeaderPrags }

Ian Lynagh's avatar
Ian Lynagh committed
335
  "-- #"                           { multiline_doc_comment }
David Waern's avatar
David Waern committed
336 337
}

338 339
<0> {
  -- In the "0" mode we ignore these pragmas
340
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
341 342 343
                     { nested_comment lexToken }
}

David Waern's avatar
David Waern committed
344
<0> {
Jedai's avatar
Jedai committed
345
  "-- #" .* { lineCommentToken }
346
}
347

348
<0,option_prags> {
349 350
  "{-#"  { warnThen Opt_WarnUnrecognisedPragmas (text "Unrecognised pragma")
                    (nested_comment lexToken) }
351 352
}

353 354
-- '0' state: ordinary lexemes

355 356
-- Haddock comments

357
<0,option_prags> {
David Waern's avatar
David Waern committed
358 359
  "-- " $docsym      / { ifExtension haddockEnabled } { multiline_doc_comment }
  "{-" \ ? $docsym   / { ifExtension haddockEnabled } { nested_doc_comment }
360 361
}

362 363
-- "special" symbols

364
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
365 366
  "[:" / { ifExtension parrEnabled }    { token ITopabrack }
  ":]" / { ifExtension parrEnabled }    { token ITcpabrack }
367
}
Ian Lynagh's avatar
Ian Lynagh committed
368

369
<0> {
370 371 372
  "[|"        / { ifExtension thEnabled } { token ITopenExpQuote }
  "[||"       / { ifExtension thEnabled } { token ITopenTExpQuote }
  "[e|"       / { ifExtension thEnabled } { token ITopenExpQuote }
373
  "[e||"      / { ifExtension thEnabled } { token ITopenTExpQuote }
374 375 376 377 378 379 380 381 382
  "[p|"       / { ifExtension thEnabled } { token ITopenPatQuote }
  "[d|"       / { ifExtension thEnabled } { layout_token ITopenDecQuote }
  "[t|"       / { ifExtension thEnabled } { token ITopenTypQuote }
  "|]"        / { ifExtension thEnabled } { token ITcloseQuote }
  "||]"       / { ifExtension thEnabled } { token ITcloseTExpQuote }
  \$ @varid   / { ifExtension thEnabled } { skip_one_varid ITidEscape }
  "$$" @varid / { ifExtension thEnabled } { skip_two_varid ITidTyEscape }
  "$("        / { ifExtension thEnabled } { token ITparenEscape }
  "$$("       / { ifExtension thEnabled } { token ITparenTyEscape }
383

384 385 386 387
-- For backward compatibility, accept the old dollar syntax
  "[$" @varid "|"  / { ifExtension qqEnabled }
                     { lex_quasiquote_tok }

388
  "[" @varid "|"  / { ifExtension qqEnabled }
389
                     { lex_quasiquote_tok }
390 391

  -- qualified quasi-quote (#5555)
392 393
  "[" @qvarid "|"  / { ifExtension qqEnabled }
                     { lex_qquasiquote_tok }
394 395
}

396
<0> {
ross's avatar
ross committed
397
  "(|" / { ifExtension arrowsEnabled `alexAndPred` notFollowedBySymbol }
Ian Lynagh's avatar
Ian Lynagh committed
398
                                        { special IToparenbar }
399 400 401
  "|)" / { ifExtension arrowsEnabled }  { special ITcparenbar }
}

402
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
403
  \? @varid / { ifExtension ipEnabled } { skip_one_varid ITdupipvarid }
404 405
}

406
<0> {
407
  "(#" / { ifExtension unboxedTuplesEnabled }
Ian Lynagh's avatar
Ian Lynagh committed
408 409 410 411 412
         { token IToubxparen }
  "#)" / { ifExtension unboxedTuplesEnabled }
         { token ITcubxparen }
}

413
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
414 415 416 417 418 419 420 421 422 423
  \(                                    { special IToparen }
  \)                                    { special ITcparen }
  \[                                    { special ITobrack }
  \]                                    { special ITcbrack }
  \,                                    { special ITcomma }
  \;                                    { special ITsemi }
  \`                                    { special ITbackquote }

  \{                                    { open_brace }
  \}                                    { close_brace }
424 425
}

426
<0,option_prags> {
427 428
  @qvarid                       { idtoken qvarid }
  @qconid                       { idtoken qconid }
Ian Lynagh's avatar
Ian Lynagh committed
429
  @varid                        { varid }
430
  @conid                        { idtoken conid }
431 432
}

433
<0> {
434 435
  @qvarid "#"+      / { ifExtension magicHashEnabled } { idtoken qvarid }
  @qconid "#"+      / { ifExtension magicHashEnabled } { idtoken qconid }
436
  @varid "#"+       / { ifExtension magicHashEnabled } { varid }
437
  @conid "#"+       / { ifExtension magicHashEnabled } { idtoken conid }
438 439
}

440 441
-- ToDo: - move `var` and (sym) into lexical syntax?
--       - remove backquote from $special?
442
<0> {
443 444
  @qvarsym                                         { idtoken qvarsym }
  @qconsym                                         { idtoken qconsym }
445 446
  @varsym                                          { varsym }
  @consym                                          { consym }
447 448
}

449 450
-- For the normal boxed literals we need to be careful
-- when trying to be close to Haskell98
451
<0> {
452
  -- Normal integral literals (:: Num a => a, from Integer)
453
  @decimal                                                               { tok_num positive 0 0 decimal }
454
  0[bB] @binary                / { ifExtension binaryLiteralsEnabled }   { tok_num positive 2 2 binary }
455 456 457
  0[oO] @octal                                                           { tok_num positive 2 2 octal }
  0[xX] @hexadecimal                                                     { tok_num positive 2 2 hexadecimal }
  @negative @decimal           / { ifExtension negativeLiteralsEnabled } { tok_num negative 1 1 decimal }
458 459
  @negative 0[bB] @binary      / { ifExtension negativeLiteralsEnabled `alexAndPred`
                                   ifExtension binaryLiteralsEnabled }   { tok_num negative 3 3 binary }
460 461
  @negative 0[oO] @octal       / { ifExtension negativeLiteralsEnabled } { tok_num negative 3 3 octal }
  @negative 0[xX] @hexadecimal / { ifExtension negativeLiteralsEnabled } { tok_num negative 3 3 hexadecimal }
462 463

  -- Normal rational literals (:: Fractional a => a, from Rational)
464 465
  @floating_point                                                        { strtoken tok_float }
  @negative @floating_point    / { ifExtension negativeLiteralsEnabled } { strtoken tok_float }
466 467
}

468
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
469
  -- Unboxed ints (:: Int#) and words (:: Word#)
470 471
  -- It's simpler (and faster?) to give separate cases to the negatives,
  -- especially considering octal/hexadecimal prefixes.
Ian Lynagh's avatar
Ian Lynagh committed
472
  @decimal                     \# / { ifExtension magicHashEnabled } { tok_primint positive 0 1 decimal }
473 474
  0[bB] @binary                \# / { ifExtension magicHashEnabled `alexAndPred`
                                      ifExtension binaryLiteralsEnabled } { tok_primint positive 2 3 binary }
Ian Lynagh's avatar
Ian Lynagh committed
475 476 477
  0[oO] @octal                 \# / { ifExtension magicHashEnabled } { tok_primint positive 2 3 octal }
  0[xX] @hexadecimal           \# / { ifExtension magicHashEnabled } { tok_primint positive 2 3 hexadecimal }
  @negative @decimal           \# / { ifExtension magicHashEnabled } { tok_primint negative 1 2 decimal }
478 479
  @negative 0[bB] @binary      \# / { ifExtension magicHashEnabled `alexAndPred`
                                      ifExtension binaryLiteralsEnabled } { tok_primint negative 3 4 binary }
Ian Lynagh's avatar
Ian Lynagh committed
480
  @negative 0[oO] @octal       \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 octal }
481
  @negative 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 hexadecimal }
482

Ian Lynagh's avatar
Ian Lynagh committed
483
  @decimal                     \# \# / { ifExtension magicHashEnabled } { tok_primword 0 2 decimal }
484 485
  0[bB] @binary                \# \# / { ifExtension magicHashEnabled `alexAndPred`
                                         ifExtension binaryLiteralsEnabled } { tok_primword 2 4 binary }
Ian Lynagh's avatar
Ian Lynagh committed
486 487 488
  0[oO] @octal                 \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 octal }
  0[xX] @hexadecimal           \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 hexadecimal }

489 490
  -- Unboxed floats and doubles (:: Float#, :: Double#)
  -- prim_{float,double} work with signed literals
491 492
  @signed @floating_point \# / { ifExtension magicHashEnabled } { init_strtoken 1 tok_primfloat }
  @signed @floating_point \# \# / { ifExtension magicHashEnabled } { init_strtoken 2 tok_primdouble }
493 494 495 496 497 498
}

-- Strings and chars are lexed by hand-written code.  The reason is
-- that even if we recognise the string or char here in the regex
-- lexer, we would still have to parse the string afterward in order
-- to convert it to a String.
499
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
500 501
  \'                            { lex_char_tok }
  \"                            { lex_string_tok }
502 503
}

504 505 506 507

-- -----------------------------------------------------------------------------
-- Alex "Haskell code fragment bottom"

508
{
509 510 511

type SourceText = String -- Note [literal source text] in HsLit

512 513 514
-- -----------------------------------------------------------------------------
-- The token type

515
data Token
Ian Lynagh's avatar
Ian Lynagh committed
516
  = ITas                        -- Haskell keywords
517 518 519 520 521 522 523 524
  | ITcase
  | ITclass
  | ITdata
  | ITdefault
  | ITderiving
  | ITdo
  | ITelse
  | IThiding
525
  | ITforeign
526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541
  | ITif
  | ITimport
  | ITin
  | ITinfix
  | ITinfixl
  | ITinfixr
  | ITinstance
  | ITlet
  | ITmodule
  | ITnewtype
  | ITof
  | ITqualified
  | ITthen
  | ITtype
  | ITwhere

Ian Lynagh's avatar
Ian Lynagh committed
542
  | ITforall                    -- GHC extension keywords
543 544 545 546
  | ITexport
  | ITlabel
  | ITdynamic
  | ITsafe
547
  | ITinterruptible
548 549 550
  | ITunsafe
  | ITstdcallconv
  | ITccallconv
551
  | ITcapiconv
552
  | ITprimcallconv
thoughtpolice's avatar
thoughtpolice committed
553
  | ITjavascriptcallconv
554
  | ITmdo
555
  | ITfamily
556
  | ITrole
557 558 559
  | ITgroup
  | ITby
  | ITusing
cactus's avatar
cactus committed
560
  | ITpattern
561

Ian Lynagh's avatar
Ian Lynagh committed
562
  -- Pragmas
563
  | ITinline_prag InlineSpec RuleMatchInfo
Ian Lynagh's avatar
Ian Lynagh committed
564 565
  | ITspec_prag                 -- SPECIALISE
  | ITspec_inline_prag Bool     -- SPECIALISE INLINE (or NOINLINE)
566 567
  | ITsource_prag
  | ITrules_prag
Ian Lynagh's avatar
Ian Lynagh committed
568
  | ITwarning_prag
569 570 571
  | ITdeprecated_prag
  | ITline_prag
  | ITscc_prag
andy@galois.com's avatar
andy@galois.com committed
572
  | ITgenerated_prag
573
  | ITcore_prag                 -- hdaume: core annotations
574
  | ITunpack_prag
575
  | ITnounpack_prag
576
  | ITann_prag
577
  | ITclose_prag
578 579 580
  | IToptions_prag String
  | ITinclude_prag String
  | ITlanguage_prag
581 582
  | ITvect_prag
  | ITvect_scalar_prag
583
  | ITnovect_prag
584
  | ITminimal_prag
585 586 587
  | IToverlappable_prag         -- instance overlap mode
  | IToverlapping_prag          -- instance overlap mode
  | IToverlaps_prag             -- instance overlap mode
588
  | ITincoherent_prag           -- instance overlap mode
589
  | ITctype
590

Ian Lynagh's avatar
Ian Lynagh committed
591
  | ITdotdot                    -- reserved symbols
592 593 594 595
  | ITcolon
  | ITdcolon
  | ITequal
  | ITlam
596
  | ITlcase
597 598 599 600 601
  | ITvbar
  | ITlarrow
  | ITrarrow
  | ITat
  | ITtilde
batterseapower's avatar
batterseapower committed
602
  | ITtildehsh
603 604 605 606 607 608
  | ITdarrow
  | ITminus
  | ITbang
  | ITstar
  | ITdot

Ian Lynagh's avatar
Ian Lynagh committed
609
  | ITbiglam                    -- GHC-extension symbols
610

Ian Lynagh's avatar
Ian Lynagh committed
611
  | ITocurly                    -- special symbols
612 613 614 615
  | ITccurly
  | ITvocurly
  | ITvccurly
  | ITobrack
Ian Lynagh's avatar
Ian Lynagh committed
616 617
  | ITopabrack                  -- [:, for parallel arrays with -XParallelArrays
  | ITcpabrack                  -- :], for parallel arrays with -XParallelArrays
618 619 620 621 622 623 624 625 626
  | ITcbrack
  | IToparen
  | ITcparen
  | IToubxparen
  | ITcubxparen
  | ITsemi
  | ITcomma
  | ITunderscore
  | ITbackquote
dreixel's avatar
dreixel committed
627
  | ITsimpleQuote               --  '
628

Ian Lynagh's avatar
Ian Lynagh committed
629
  | ITvarid   FastString        -- identifiers
630 631 632 633 634 635 636
  | ITconid   FastString
  | ITvarsym  FastString
  | ITconsym  FastString
  | ITqvarid  (FastString,FastString)
  | ITqconid  (FastString,FastString)
  | ITqvarsym (FastString,FastString)
  | ITqconsym (FastString,FastString)
637 638
  | ITprefixqvarsym (FastString,FastString)
  | ITprefixqconsym (FastString,FastString)
639

Ian Lynagh's avatar
Ian Lynagh committed
640
  | ITdupipvarid   FastString   -- GHC extension: implicit param: ?x
641

642 643 644
  | ITchar       SourceText Char        -- Note [literal source text] in HsLit
  | ITstring     SourceText FastString  -- Note [literal source text] in HsLit
  | ITinteger    SourceText Integer     -- Note [literal source text] in HsLit
645
  | ITrational   FractionalLit
646

647 648 649 650
  | ITprimchar   SourceText Char        -- Note [literal source text] in HsLit
  | ITprimstring SourceText ByteString  -- Note [literal source text] in HsLit
  | ITprimint    SourceText Integer     -- Note [literal source text] in HsLit
  | ITprimword   SourceText Integer     -- Note [literal source text] in HsLit
651 652
  | ITprimfloat  FractionalLit
  | ITprimdouble FractionalLit
653

654
  -- Template Haskell extension tokens
Ian Lynagh's avatar
Ian Lynagh committed
655 656 657 658 659
  | ITopenExpQuote              --  [| or [e|
  | ITopenPatQuote              --  [p|
  | ITopenDecQuote              --  [d|
  | ITopenTypQuote              --  [t|
  | ITcloseQuote                --  |]
660 661
  | ITopenTExpQuote             --  [||
  | ITcloseTExpQuote            --  ||]
Ian Lynagh's avatar
Ian Lynagh committed
662 663
  | ITidEscape   FastString     --  $x
  | ITparenEscape               --  $(
664 665
  | ITidTyEscape   FastString   --  $$x
  | ITparenTyEscape             --  $$(
Ian Lynagh's avatar
Ian Lynagh committed
666
  | ITtyQuote                   --  ''
667 668 669 670 671 672 673 674
  | ITquasiQuote (FastString,FastString,RealSrcSpan)
    -- ITquasiQuote(quoter, quote, loc)
    -- represents a quasi-quote of the form
    -- [quoter| quote |]
  | ITqQuasiQuote (FastString,FastString,FastString,RealSrcSpan)
    -- ITqQuasiQuote(Qual, quoter, quote, loc)
    -- represents a qualified quasi-quote of the form
    -- [Qual.quoter| quote |]
675 676 677 678

  -- Arrow notation extension
  | ITproc
  | ITrec
Ian Lynagh's avatar
Ian Lynagh committed
679 680 681 682 683 684
  | IToparenbar                 --  (|
  | ITcparenbar                 --  |)
  | ITlarrowtail                --  -<
  | ITrarrowtail                --  >-
  | ITLarrowtail                --  -<<
  | ITRarrowtail                --  >>-
685

Ian Lynagh's avatar
Ian Lynagh committed
686 687
  | ITunknown String            -- Used when the lexer can't make sense of it
  | ITeof                       -- end of file token
688 689 690 691 692 693 694

  -- Documentation annotations
  | ITdocCommentNext  String     -- something beginning '-- |'
  | ITdocCommentPrev  String     -- something beginning '-- ^'
  | ITdocCommentNamed String     -- something beginning '-- $'
  | ITdocSection      Int String -- a section heading
  | ITdocOptions      String     -- doc options (prune, ignore-exports, etc)
David Waern's avatar
David Waern committed
695
  | ITdocOptionsOld   String     -- doc options declared "-- # ..."-style
Jedai's avatar
Jedai committed
696 697
  | ITlineComment     String     -- comment starting by "--"
  | ITblockComment    String     -- comment in {- -}
698

699
  deriving Show
700

Alan Zimmerman's avatar
Alan Zimmerman committed
701 702 703
instance Outputable Token where
  ppr x = text (show x)

704 705 706 707 708 709 710
-- the bitmap provided as the third component indicates whether the
-- corresponding extension keyword is valid under the extension options
-- provided to the compiler; if the extension corresponding to *any* of the
-- bits set in the bitmap is enabled, the keyword is valid (this setup
-- facilitates using a keyword in two different extensions that can be
-- activated independently)
--
711
reservedWordsFM :: UniqFM (Token, ExtsBitmap)
712
reservedWordsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739
    map (\(x, y, z) -> (mkFastString x, (y, z)))
        [( "_",              ITunderscore,    0 ),
         ( "as",             ITas,            0 ),
         ( "case",           ITcase,          0 ),
         ( "class",          ITclass,         0 ),
         ( "data",           ITdata,          0 ),
         ( "default",        ITdefault,       0 ),
         ( "deriving",       ITderiving,      0 ),
         ( "do",             ITdo,            0 ),
         ( "else",           ITelse,          0 ),
         ( "hiding",         IThiding,        0 ),
         ( "if",             ITif,            0 ),
         ( "import",         ITimport,        0 ),
         ( "in",             ITin,            0 ),
         ( "infix",          ITinfix,         0 ),
         ( "infixl",         ITinfixl,        0 ),
         ( "infixr",         ITinfixr,        0 ),
         ( "instance",       ITinstance,      0 ),
         ( "let",            ITlet,           0 ),
         ( "module",         ITmodule,        0 ),
         ( "newtype",        ITnewtype,       0 ),
         ( "of",             ITof,            0 ),
         ( "qualified",      ITqualified,     0 ),
         ( "then",           ITthen,          0 ),
         ( "type",           ITtype,          0 ),
         ( "where",          ITwhere,         0 ),

740 741 742
         ( "forall",         ITforall,        xbit ExplicitForallBit .|.
                                              xbit InRulePragBit),
         ( "mdo",            ITmdo,           xbit RecursiveDoBit),
743 744 745
             -- See Note [Lexing type pseudo-keywords]
         ( "family",         ITfamily,        0 ),
         ( "role",           ITrole,          0 ),
746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767
         ( "pattern",        ITpattern,       xbit PatternSynonymsBit),
         ( "group",          ITgroup,         xbit TransformComprehensionsBit),
         ( "by",             ITby,            xbit TransformComprehensionsBit),
         ( "using",          ITusing,         xbit TransformComprehensionsBit),

         ( "foreign",        ITforeign,       xbit FfiBit),
         ( "export",         ITexport,        xbit FfiBit),
         ( "label",          ITlabel,         xbit FfiBit),
         ( "dynamic",        ITdynamic,       xbit FfiBit),
         ( "safe",           ITsafe,          xbit FfiBit .|.
                                              xbit SafeHaskellBit),
         ( "interruptible",  ITinterruptible, xbit InterruptibleFfiBit),
         ( "unsafe",         ITunsafe,        xbit FfiBit),
         ( "stdcall",        ITstdcallconv,   xbit FfiBit),
         ( "ccall",          ITccallconv,     xbit FfiBit),
         ( "capi",           ITcapiconv,      xbit CApiFfiBit),
         ( "prim",           ITprimcallconv,  xbit FfiBit),
         ( "javascript",     ITjavascriptcallconv, xbit FfiBit),

         ( "rec",            ITrec,           xbit ArrowsBit .|.
                                              xbit RecursiveDoBit),
         ( "proc",           ITproc,          xbit ArrowsBit)
768 769
     ]

770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785
{-----------------------------------
Note [Lexing type pseudo-keywords]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

One might think that we wish to treat 'family' and 'role' as regular old
varids whenever -XTypeFamilies and -XRoleAnnotations are off, respectively.
But, there is no need to do so. These pseudo-keywords are not stolen syntax:
they are only used after the keyword 'type' at the top-level, where varids are
not allowed. Furthermore, checks further downstream (TcTyClsDecls) ensure that
type families and role annotations are never declared without their extensions
on. In fact, by unconditionally lexing these pseudo-keywords as special, we
can get better error messages.

Also, note that these are included in the `varid` production in the parser --
a key detail to make all this work.
-------------------------------------}
786

787
reservedSymsFM :: UniqFM (Token, ExtsBitmap -> Bool)
788
reservedSymsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
789 790 791 792 793 794 795 796 797 798 799 800
    map (\ (x,y,z) -> (mkFastString x,(y,z)))
      [ ("..",  ITdotdot,   always)
        -- (:) is a reserved op, meaning only list cons
       ,(":",   ITcolon,    always)
       ,("::",  ITdcolon,   always)
       ,("=",   ITequal,    always)
       ,("\\",  ITlam,      always)
       ,("|",   ITvbar,     always)
       ,("<-",  ITlarrow,   always)
       ,("->",  ITrarrow,   always)
       ,("@",   ITat,       always)
       ,("~",   ITtilde,    always)
801
       ,("~#",  ITtildehsh, magicHashEnabled)
Ian Lynagh's avatar
Ian Lynagh committed
802 803 804 805 806
       ,("=>",  ITdarrow,   always)
       ,("-",   ITminus,    always)
       ,("!",   ITbang,     always)

        -- For data T (a::*) = MkT
807
       ,("*", ITstar, always) -- \i -> kindSigsEnabled i || tyFamEnabled i)
Ian Lynagh's avatar
Ian Lynagh committed
808
        -- For 'forall a . t'
809
       ,(".", ITdot,  always) -- \i -> explicitForallEnabled i || inRulePrag i)
Ian Lynagh's avatar
Ian Lynagh committed
810 811 812 813 814

       ,("-<",  ITlarrowtail, arrowsEnabled)
       ,(">-",  ITrarrowtail, arrowsEnabled)
       ,("-<<", ITLarrowtail, arrowsEnabled)
       ,(">>-", ITRarrowtail, arrowsEnabled)
815

Ian Lynagh's avatar
Ian Lynagh committed
816 817
       ,("∷",   ITdcolon, unicodeSyntaxEnabled)
       ,("⇒",   ITdarrow, unicodeSyntaxEnabled)
818
       ,("∀",   ITforall, unicodeSyntaxEnabled)
Ian Lynagh's avatar
Ian Lynagh committed
819 820
       ,("→",   ITrarrow, unicodeSyntaxEnabled)
       ,("←",   ITlarrow, unicodeSyntaxEnabled)
821 822 823 824 825 826 827 828

       ,("⤙",   ITlarrowtail, \i -> unicodeSyntaxEnabled i && arrowsEnabled i)
       ,("⤚",   ITrarrowtail, \i -> unicodeSyntaxEnabled i && arrowsEnabled i)
       ,("⤛",   ITLarrowtail, \i -> unicodeSyntaxEnabled i && arrowsEnabled i)
       ,("⤜",   ITRarrowtail, \i -> unicodeSyntaxEnabled i && arrowsEnabled i)

       ,("★", ITstar, unicodeSyntaxEnabled)

Simon Marlow's avatar
Simon Marlow committed
829 830 831
        -- ToDo: ideally, → and ∷ should be "specials", so that they cannot
        -- form part of a large operator.  This would let us have a better
        -- syntax for kinds: ɑ∷*→* would be a legal kind signature. (maybe).
832 833 834 835 836
       ]

-- -----------------------------------------------------------------------------
-- Lexer actions

Ian Lynagh's avatar
Ian Lynagh committed
837
type Action = RealSrcSpan -> StringBuffer -> Int -> P (RealLocated Token)
838

839
special :: Token -> Action
twanvl's avatar
twanvl committed
840
special tok span _buf _len = return (L span tok)
841

842
token, layout_token :: Token -> Action
twanvl's avatar
twanvl committed
843 844
token t span _buf _len = return (L span t)
layout_token t span _buf _len = pushLexState layout >> return (L span t)
845

846 847
idtoken :: (StringBuffer -> Int -> Token) -> Action
idtoken f span buf len = return (L span $! (f buf len))
848

849
skip_one_varid :: (FastString -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
850
skip_one_varid f span buf len
851
  = return (L span $! f (lexemeToFastString (stepOn buf) (len-1)))
852

853 854 855 856
skip_two_varid :: (FastString -> Token) -> Action
skip_two_varid f span buf len
  = return (L span $! f (lexemeToFastString (stepOn (stepOn buf)) (len-2)))

857
strtoken :: (String -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
858
strtoken f span buf len =
859
  return (L span $! (f $! lexemeToString buf len))
860

861
init_strtoken :: Int -> (String -> Token) -> Action
862
-- like strtoken, but drops the last N character(s)
Ian Lynagh's avatar
Ian Lynagh committed
863
init_strtoken drop f span buf len =
864
  return (L span $! (f $! lexemeToString buf (len-drop)))
865 866

begin :: Int -> Action
867
begin code _span _str _len = do pushLexState code; lexToken
868 869

pop :: Action
Ian Lynagh's avatar
Ian Lynagh committed
870 871
pop _span _buf _len = do _ <- popLexState
                         lexToken
872

873 874 875 876 877 878 879 880 881 882 883
hopefully_open_brace :: Action
hopefully_open_brace span buf len
 = do relaxed <- extension relaxedLayout
      ctx <- getContext
      (AI l _) <- getInput
      let offset = srcLocCol l
          isOK = relaxed ||
                 case ctx of
                 Layout prev_off : _ -> prev_off < offset
                 _                   -> True
      if isOK then pop_and open_brace span buf len
Ian Lynagh's avatar
Ian Lynagh committed
884
              else failSpanMsgP (RealSrcSpan span) (text "Missing block")
885

886
pop_and :: Action -> Action
Ian Lynagh's avatar
Ian Lynagh committed
887 888
pop_and act span buf len = do _ <- popLexState
                              act span buf len
889

890
{-# INLINE nextCharIs #-}
Ian Lynagh's avatar
Ian Lynagh committed
891
nextCharIs :: StringBuffer -> (Char -> Bool) -> Bool
892 893
nextCharIs buf p = not (atEnd buf) && p (currentChar buf)

894 895 896 897
{-# INLINE nextCharIsNot #-}
nextCharIsNot :: StringBuffer -> (Char -> Bool) -> Bool
nextCharIsNot buf p = not (nextCharIs buf p)

898
notFollowedBy :: Char -> AlexAccPred ExtsBitmap
Ian Lynagh's avatar
Ian Lynagh committed
899
notFollowedBy char _ _ _ (AI _ buf)
900
  = nextCharIsNot buf (== char)
901

902
notFollowedBySymbol :: AlexAccPred ExtsBitmap
903
notFollowedBySymbol _ _ _ (AI _ buf)
904
  = nextCharIsNot buf (`elem` "!#$%&*+./<=>?@\\^|-~")
ross's avatar
ross committed
905

906 907 908 909 910
-- We must reject doc comments as being ordinary comments everywhere.
-- In some cases the doc comment will be selected as the lexeme due to
-- maximal munch, but not always, because the nested comment rule is
-- valid in all states, but the doc-comment rules are only valid in
-- the non-layout states.
911
isNormalComment :: AlexAccPred ExtsBitmap
912
isNormalComment bits _ _ (AI _ buf)
913
  | haddockEnabled bits = notFollowedByDocOrPragma
914
  | otherwise           = nextCharIsNot buf (== '#')
915 916
  where
    notFollowedByDocOrPragma
917
       = afterOptionalSpace buf (\b -> nextCharIsNot b (`elem` "|^*$#"))
918

919 920 921 922 923
afterOptionalSpace :: StringBuffer -> (StringBuffer -> Bool) -> Bool
afterOptionalSpace buf p
    = if nextCharIs buf (== ' ')
      then p (snd (nextChar buf))
      else p buf
924

925
atEOL :: AlexAccPred ExtsBitmap
926
atEOL _ _ _ (AI _ buf) = atEnd buf || currentChar buf == '\n'
927

928
ifExtension :: (ExtsBitmap -> Bool) -> AlexAccPred ExtsBitmap
929 930
ifExtension pred bits _ _ _ = pred bits

931 932 933
multiline_doc_comment :: Action
multiline_doc_comment span buf _len = withLexedDocType (worker "")
  where
934
    worker commentAcc input docType oneLine = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
935
      Just ('\n', input')
936 937 938 939 940 941
        | oneLine -> docCommentEnd input commentAcc docType buf span
        | otherwise -> case checkIfCommentLine input' of
          Just input -> worker ('\n':commentAcc) input docType False
          Nothing -> docCommentEnd input commentAcc docType buf span
      Just (c, input) -> worker (c:commentAcc) input docType oneLine
      Nothing -> docCommentEnd input commentAcc docType buf span
Ian Lynagh's avatar
Ian Lynagh committed
942

943 944
    checkIfCommentLine input = check (dropNonNewlineSpace input)
      where
945 946 947
        check input = case alexGetChar' input of
          Just ('-', input) -> case alexGetChar' input of
            Just ('-', input) -> case alexGetChar' input of
948 949 950 951 952
              Just (c, _) | c /= '-' -> Just input
              _ -> Nothing
            _ -> Nothing
          _ -> Nothing

953
        dropNonNewlineSpace input = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
954
          Just (c, input')
955 956 957 958
            | isSpace c && c /= '\n' -> dropNonNewlineSpace input'
            | otherwise -> input
          Nothing -> input

Jedai's avatar
Jedai committed
959 960 961 962 963
lineCommentToken :: Action
lineCommentToken span buf len = do
  b <- extension rawTokenStreamEnabled
  if b then strtoken ITlineComment span buf len else lexToken

964 965 966 967
{-
  nested comments require traversing by hand, they can't be parsed
  using regular expressions.
-}
Ian Lynagh's avatar
Ian Lynagh committed
968
nested_comment :: P (RealLocated Token) -> Action
Alan Zimmerman's avatar
Alan Zimmerman committed
969
nested_comment cont span buf len = do
970
  input <- getInput
Alan Zimmerman's avatar
Alan Zimmerman committed
971
  go (reverse $ drop 2 $ lexemeToString buf len) (1::Int) input
972
  where
Alan Zimmerman's avatar
Alan Zimmerman committed
973 974 975 976 977 978
    go commentAcc 0 input = do
      setInput input
      b <- extension rawTokenStreamEnabled
      if b
        then docCommentEnd input commentAcc ITblockComment buf span
        else cont
979
    go commentAcc n input = case alexGetChar' input of
980
      Nothing -> errBrace input span
981
      Just ('-',input) -> case alexGetChar' input of
982
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
983 984
        Just ('\125',input) -> go commentAcc (n-1) input
        Just (_,_)          -> go ('-':commentAcc) n input
985
      Just ('\123',input) -> case alexGetChar' input of
986
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
987 988 989
        Just ('-',input) -> go ('-':'\123':commentAcc) (n+1) input
        Just (_,_)       -> go ('\123':commentAcc) n input
      Just (c,input) -> go (c:commentAcc) n input
990 991 992 993

nested_doc_comment :: Action
nested_doc_comment span buf _len = withLexedDocType (go "")
  where
994
    go commentAcc input docType _ = case alexGetChar' input of
995
      Nothing -> errBrace input span
996
      Just ('-',input) -> case alexGetChar' input of
997
        Nothing -> errBrace input span
twanvl's avatar
twanvl committed
998
        Just ('\125',input) ->
999
          docCommentEnd input commentAcc docType buf span
twanvl's avatar
twanvl committed
1000
        Just (_,_) -> go ('-':commentAcc) input docType False
1001
      Just ('\123', input) -> case alexGetChar' input of
1002 1003 1004 1005 1006
        Nothing  -> errBrace input span
        Just ('-',input) -> do
          setInput input
          let cont = do input <- getInput; go commentAcc input docType False
          nested_comment cont span buf _len
twanvl's avatar
twanvl committed
1007
        Just (_,_) -> go ('\123':commentAcc) input docType False
1008 1009
      Just (c,input) -> go (c:commentAcc) input docType False

Ian Lynagh's avatar
Ian Lynagh committed
1010 1011
withLexedDocType :: (AlexInput -> (String -> Token) -> Bool -> P (RealLocated Token))
                 -> P (RealLocated Token)
1012
withLexedDocType lexDocComment = do
1013
  input@(AI _ buf) <- getInput
1014 1015 1016 1017
  case prevChar buf ' ' of
    '|' -> lexDocComment input ITdocCommentNext False
    '^' -> lexDocComment input ITdocCommentPrev False
    '$' -> lexDocComment input ITdocCommentNamed False
David Waern's avatar
David Waern committed
1018 1019
    '*' -> lexDocSection 1 input
    '#' -> lexDocComment input ITdocOptionsOld False
Ian Lynagh's avatar
Ian Lynagh committed
1020
    _ -> panic "withLexedDocType: Bad doc type"
Ian Lynagh's avatar
Ian Lynagh committed
1021
 where
1022
    lexDocSection n input = case alexGetChar' input of
1023
      Just ('*', input) -> lexDocSection (n+1) input
twanvl's avatar
twanvl committed
1024
      Just (_,   _)     -> lexDocComment input (ITdocSection n) True
1025 1026
      Nothing -> do setInput input; lexToken -- eof reached, lex it normally

Simon Marlow's avatar