Lexer.x 105 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
{-# LANGUAGE BangPatterns #-}
45 46

-- See Note [Warnings in code generated by Alex] in compiler/parser/Lexer.x
47 48
{-# OPTIONS_GHC -fno-warn-unused-matches #-}
{-# OPTIONS_GHC -fno-warn-unused-binds #-}
49
{-# OPTIONS_GHC -fno-warn-unused-imports #-}
50
{-# OPTIONS_GHC -fno-warn-tabs #-}
51
{-# OPTIONS_GHC -fno-warn-missing-signatures #-}
52

53 54
{-# OPTIONS_GHC -funbox-strict-fields #-}

55
module Lexer (
Alan Zimmerman's avatar
Alan Zimmerman committed
56
   Token(..), lexer, pragState, mkPState, PState(..),
Ian Lynagh's avatar
Ian Lynagh committed
57
   P(..), ParseResult(..), getSrcLoc,
58
   getPState, getDynFlags, withThisPackage,
59
   failLocMsgP, failSpanMsgP, srcParseFail,
Ian Lynagh's avatar
Ian Lynagh committed
60
   getMessages,
61
   popContext, pushCurrentContext, setLastToken, setSrcLoc,
vivian's avatar
vivian committed
62
   activeContext, nextIsEOF,
simonpj@microsoft.com's avatar
simonpj@microsoft.com committed
63
   getLexState, popLexState, pushLexState,
64
   extension, bangPatEnabled, datatypeContextsEnabled,
65
   traditionalRecordSyntaxEnabled,
66 67
   explicitForallEnabled,
   inRulePrag,
cactus's avatar
cactus committed
68 69 70
   explicitNamespacesEnabled,
   patternSynonymsEnabled,
   sccProfilingOn, hpcEnabled,
Jedai's avatar
Jedai committed
71
   addWarning,
Alan Zimmerman's avatar
Alan Zimmerman committed
72
   lexTokenStream,
Alan Zimmerman's avatar
Alan Zimmerman committed
73
   addAnnotation,AddAnn,mkParensApiAnn
74 75
  ) where

rodlogic's avatar
rodlogic committed
76
-- base
77
#if __GLASGOW_HASKELL__ < 709
rodlogic's avatar
rodlogic committed
78
import Control.Applicative
79
#endif
rodlogic's avatar
rodlogic committed
80
import Control.Monad
quchen's avatar
quchen committed
81 82 83
#if __GLASGOW_HASKELL__ > 710
import Control.Monad.Fail
#endif
rodlogic's avatar
rodlogic committed
84
import Data.Bits
85
import Data.Char
rodlogic's avatar
rodlogic committed
86 87 88 89
import Data.List
import Data.Maybe
import Data.Word

90 91 92
-- ghc-boot
import qualified GHC.LanguageExtensions as LangExt

rodlogic's avatar
rodlogic committed
93 94 95 96 97 98 99 100
-- bytestring
import Data.ByteString (ByteString)

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

-- 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
Alan Zimmerman's avatar
Alan Zimmerman committed
115 116
import BasicTypes     ( InlineSpec(..), RuleMatchInfo(..), FractionalLit(..),
                        SourceText )
117

rodlogic's avatar
rodlogic committed
118 119
-- compiler/parser
import Ctype
Alan Zimmerman's avatar
Alan Zimmerman committed
120 121

import ApiAnnotation
122 123
}

124 125 126
-- -----------------------------------------------------------------------------
-- Alex "Character set macros"

127 128
-- NB: The logic behind these definitions is also reflected in basicTypes/Lexeme.hs
-- Any changes here should likely be reflected there.
129 130 131 132
$unispace    = \x05 -- Trick Alex into handling Unicode. See alexGetByte.
$nl          = [\n\r\f]
$whitechar   = [$nl\v\ $unispace]
$white_no_nl = $whitechar # \n -- TODO #8424
133
$tab         = \t
134 135

$ascdigit  = 0-9
136
$unidigit  = \x03 -- Trick Alex into handling Unicode. See alexGetByte.
137
$decdigit  = $ascdigit -- for now, should really be $digit (ToDo)
138 139 140
$digit     = [$ascdigit $unidigit]

$special   = [\(\)\,\;\[\]\`\{\}]
141 142 143
$ascsymbol = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~\:]
$unisymbol = \x04 -- Trick Alex into handling Unicode. See alexGetByte.
$symbol    = [$ascsymbol $unisymbol] # [$special \_\"\']
144

145
$unilarge  = \x01 -- Trick Alex into handling Unicode. See alexGetByte.
146
$asclarge  = [A-Z]
147 148
$large     = [$asclarge $unilarge]

149
$unismall  = \x02 -- Trick Alex into handling Unicode. See alexGetByte.
150
$ascsmall  = [a-z]
151 152
$small     = [$ascsmall $unismall \_]

153 154
$unigraphic = \x06 -- Trick Alex into handling Unicode. See alexGetByte.
$graphic   = [$small $large $symbol $digit $special $unigraphic \"\']
155

156
$binit     = 0-1
Ian Lynagh's avatar
Ian Lynagh committed
157
$octit     = 0-7
158
$hexit     = [$decdigit A-F a-f]
159 160 161 162

$suffix    = \x07 -- Trick Alex into handling Unicode. See alexGetByte.
-- TODO #10196. Only allow modifier letters in the suffix of an identifier.
$idchar    = [$small $large $digit $suffix \']
163

164 165
$pragmachar = [$small $large $digit]

166 167
$docsym    = [\| \^ \* \$]

168

169 170 171 172 173 174 175 176
-- -----------------------------------------------------------------------------
-- 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
177

178
@decimal     = $decdigit+
179
@binary      = $binit+
180 181 182 183 184
@octal       = $octit+
@hexadecimal = $hexit+
@exponent    = [eE] [\-\+]? @decimal

@qual = (@conid \.)+
185 186 187 188
@qvarid = @qual @varid
@qconid = @qual @conid
@qvarsym = @qual @varsym
@qconsym = @qual @consym
189 190 191

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

192 193 194 195 196
-- normal signed numerical literals can only be explicitly negative,
-- not explicitly positive (contrast @exponent)
@negative = \-
@signed = @negative ?

197 198 199 200

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

201 202
haskell :-

203 204 205 206 207

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

-- everywhere: skip whitespace
Ian Lynagh's avatar
Ian Lynagh committed
208
$white_no_nl+ ;
209
$tab          { warnTab }
210 211 212 213 214

-- 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
215 216 217 218
-- are). We also rule out nested Haddock comments, if the -haddock flag is
-- set.

"{-" / { isNormalComment } { nested_comment lexToken }
219 220 221 222

-- 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.
223 224

-- Since Haddock comments aren't valid in every state, we need to rule them
Ian Lynagh's avatar
Ian Lynagh committed
225
-- out here.
226 227 228 229 230 231 232

-- 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
233
"-- " ~[$docsym \#] .* { lineCommentToken }
234
"--" [^$symbol \ ] .* { lineCommentToken }
235 236 237

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

Jedai's avatar
Jedai committed
238
"-- " [$docsym \#] .* / { ifExtension (not . haddockEnabled) } { lineCommentToken }
239 240 241 242 243 244 245

-- 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.

246
"---"\-* ~$symbol .* { lineCommentToken }
247 248 249 250

-- 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
251
"--"\-* / { atEOL } { lineCommentToken }
252 253 254 255

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

Jedai's avatar
Jedai committed
256
"-- " / { atEOL } { lineCommentToken }
257 258 259 260 261 262 263 264 265 266 267

-- '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
268
  \n                                    ;
Adam Gundry's avatar
Adam Gundry committed
269 270
  ^\# line                              { begin line_prag1 }
  ^\# / { followedByDigit }             { begin line_prag1 }
Ian Lynagh's avatar
Ian Lynagh committed
271 272 273
  ^\# pragma .* \n                      ; -- GCC 3.3 CPP generated, apparently
  ^\# \! .* \n                          ; -- #!, for scripts
  ()                                    { do_bol }
274 275 276 277 278
}

-- 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.
279
<layout, layout_do, layout_if> {
Ian Lynagh's avatar
Ian Lynagh committed
280 281 282 283
  \{ / { notFollowedBy '-' }            { hopefully_open_brace }
        -- we might encounter {-# here, but {- has been handled already
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
284 285
}

286 287 288 289 290 291
-- after an 'if', a vertical bar starts a layout context for MultiWayIf
<layout_if> {
  \| / { notFollowedBySymbol }          { new_layout_context True ITvbar }
  ()                                    { pop }
}

292
-- do is treated in a subtly different way, see new_layout_context
293 294
<layout>    ()                          { new_layout_context True  ITvocurly }
<layout_do> ()                          { new_layout_context False ITvocurly }
295 296 297 298

-- 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
299
<layout_left>  ()                       { do_layout_left }
300

Ian Lynagh's avatar
Ian Lynagh committed
301
<0,option_prags> \n                     { begin bol }
302

303 304
"{-#" $whitechar* $pragmachar+ / { known_pragma linePrags }
                                { dispatch_pragmas linePrags }
305 306 307

-- single-line line pragmas, of the form
--    # <line> "<file>" <extra-stuff> \n
308
<line_prag1> @decimal                   { setLine line_prag1a }
Ian Lynagh's avatar
Ian Lynagh committed
309 310
<line_prag1a> \" [$graphic \ ]* \"      { setFile line_prag1b }
<line_prag1b> .*                        { pop }
311 312 313

-- Haskell-style line pragmas, of the form
--    {-# LINE <line> "<file>" #-}
314
<line_prag2> @decimal                   { setLine line_prag2a }
Ian Lynagh's avatar
Ian Lynagh committed
315 316
<line_prag2a> \" [$graphic \ ]* \"      { setFile line_prag2b }
<line_prag2b> "#-}"|"-}"                { pop }
317 318
   -- NOTE: accept -} at the end of a LINE pragma, for compatibility
   -- with older versions of GHC which generated these.
319

320
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
321
  "{-#" $whitechar* $pragmachar+
322 323 324 325 326
        $whitechar+ $pragmachar+ / { known_pragma twoWordPrags }
                                 { dispatch_pragmas twoWordPrags }

  "{-#" $whitechar* $pragmachar+ / { known_pragma oneWordPrags }
                                 { dispatch_pragmas oneWordPrags }
327

328
  -- We ignore all these pragmas, but don't generate a warning for them
329 330
  "{-#" $whitechar* $pragmachar+ / { known_pragma ignoredPrags }
                                 { dispatch_pragmas ignoredPrags }
331 332

  -- ToDo: should only be valid inside a pragma:
Ian Lynagh's avatar
Ian Lynagh committed
333
  "#-}"                          { endPrag }
334 335
}

336
<option_prags> {
337 338 339
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
                                   { dispatch_pragmas fileHeaderPrags }

Ian Lynagh's avatar
Ian Lynagh committed
340
  "-- #"                           { multiline_doc_comment }
David Waern's avatar
David Waern committed
341 342
}

343 344
<0> {
  -- In the "0" mode we ignore these pragmas
345
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
346 347 348
                     { nested_comment lexToken }
}

David Waern's avatar
David Waern committed
349
<0> {
Jedai's avatar
Jedai committed
350
  "-- #" .* { lineCommentToken }
351
}
352

353
<0,option_prags> {
354 355
  "{-#"  { warnThen Opt_WarnUnrecognisedPragmas (text "Unrecognised pragma")
                    (nested_comment lexToken) }
356 357
}

358 359
-- '0' state: ordinary lexemes

360 361
-- Haddock comments

362
<0,option_prags> {
David Waern's avatar
David Waern committed
363 364
  "-- " $docsym      / { ifExtension haddockEnabled } { multiline_doc_comment }
  "{-" \ ? $docsym   / { ifExtension haddockEnabled } { nested_doc_comment }
365 366
}

367 368
-- "special" symbols

369
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
370 371
  "[:" / { ifExtension parrEnabled }    { token ITopabrack }
  ":]" / { ifExtension parrEnabled }    { token ITcpabrack }
372
}
Ian Lynagh's avatar
Ian Lynagh committed
373

374
<0> {
375 376 377 378 379 380 381 382 383
  "[|"        / { ifExtension thQuotesEnabled } { token (ITopenExpQuote NoE) }
  "[||"       / { ifExtension thQuotesEnabled } { token (ITopenTExpQuote NoE) }
  "[e|"       / { ifExtension thQuotesEnabled } { token (ITopenExpQuote HasE) }
  "[e||"      / { ifExtension thQuotesEnabled } { token (ITopenTExpQuote HasE) }
  "[p|"       / { ifExtension thQuotesEnabled } { token ITopenPatQuote }
  "[d|"       / { ifExtension thQuotesEnabled } { layout_token ITopenDecQuote }
  "[t|"       / { ifExtension thQuotesEnabled } { token ITopenTypQuote }
  "|]"        / { ifExtension thQuotesEnabled } { token ITcloseQuote }
  "||]"       / { ifExtension thQuotesEnabled } { token ITcloseTExpQuote }
384 385 386 387
  \$ @varid   / { ifExtension thEnabled } { skip_one_varid ITidEscape }
  "$$" @varid / { ifExtension thEnabled } { skip_two_varid ITidTyEscape }
  "$("        / { ifExtension thEnabled } { token ITparenEscape }
  "$$("       / { ifExtension thEnabled } { token ITparenTyEscape }
388

389
  "[" @varid "|"  / { ifExtension qqEnabled }
390
                     { lex_quasiquote_tok }
391 392

  -- qualified quasi-quote (#5555)
393 394
  "[" @qvarid "|"  / { ifExtension qqEnabled }
                     { lex_qquasiquote_tok }
395 396
}

397
<0> {
ross's avatar
ross committed
398
  "(|" / { ifExtension arrowsEnabled `alexAndPred` notFollowedBySymbol }
Ian Lynagh's avatar
Ian Lynagh committed
399
                                        { special IToparenbar }
400 401 402
  "|)" / { ifExtension arrowsEnabled }  { special ITcparenbar }
}

403
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
404
  \? @varid / { ifExtension ipEnabled } { skip_one_varid ITdupipvarid }
405 406
}

Adam Gundry's avatar
Adam Gundry committed
407 408 409 410 411
<0> {
  "#" @varid / { ifExtension overloadedLabelsEnabled }
               { skip_one_varid ITlabelvarid }
}

412
<0> {
413
  "(#" / { ifExtension unboxedTuplesEnabled }
Ian Lynagh's avatar
Ian Lynagh committed
414 415 416 417 418
         { token IToubxparen }
  "#)" / { ifExtension unboxedTuplesEnabled }
         { token ITcubxparen }
}

419
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
420 421 422 423 424 425 426 427 428 429
  \(                                    { special IToparen }
  \)                                    { special ITcparen }
  \[                                    { special ITobrack }
  \]                                    { special ITcbrack }
  \,                                    { special ITcomma }
  \;                                    { special ITsemi }
  \`                                    { special ITbackquote }

  \{                                    { open_brace }
  \}                                    { close_brace }
430 431
}

432
<0,option_prags> {
433 434
  @qvarid                       { idtoken qvarid }
  @qconid                       { idtoken qconid }
Ian Lynagh's avatar
Ian Lynagh committed
435
  @varid                        { varid }
436
  @conid                        { idtoken conid }
437 438
}

439
<0> {
440 441
  @qvarid "#"+      / { ifExtension magicHashEnabled } { idtoken qvarid }
  @qconid "#"+      / { ifExtension magicHashEnabled } { idtoken qconid }
442
  @varid "#"+       / { ifExtension magicHashEnabled } { varid }
443
  @conid "#"+       / { ifExtension magicHashEnabled } { idtoken conid }
444 445
}

446 447
-- ToDo: - move `var` and (sym) into lexical syntax?
--       - remove backquote from $special?
448
<0> {
449 450
  @qvarsym                                         { idtoken qvarsym }
  @qconsym                                         { idtoken qconsym }
451 452
  @varsym                                          { varsym }
  @consym                                          { consym }
453 454
}

455 456
-- For the normal boxed literals we need to be careful
-- when trying to be close to Haskell98
457
<0> {
458
  -- Normal integral literals (:: Num a => a, from Integer)
459
  @decimal                                                               { tok_num positive 0 0 decimal }
460
  0[bB] @binary                / { ifExtension binaryLiteralsEnabled }   { tok_num positive 2 2 binary }
461 462 463
  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 }
464 465
  @negative 0[bB] @binary      / { ifExtension negativeLiteralsEnabled `alexAndPred`
                                   ifExtension binaryLiteralsEnabled }   { tok_num negative 3 3 binary }
466 467
  @negative 0[oO] @octal       / { ifExtension negativeLiteralsEnabled } { tok_num negative 3 3 octal }
  @negative 0[xX] @hexadecimal / { ifExtension negativeLiteralsEnabled } { tok_num negative 3 3 hexadecimal }
468 469

  -- Normal rational literals (:: Fractional a => a, from Rational)
470 471
  @floating_point                                                        { strtoken tok_float }
  @negative @floating_point    / { ifExtension negativeLiteralsEnabled } { strtoken tok_float }
472 473
}

474
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
475
  -- Unboxed ints (:: Int#) and words (:: Word#)
476 477
  -- It's simpler (and faster?) to give separate cases to the negatives,
  -- especially considering octal/hexadecimal prefixes.
Ian Lynagh's avatar
Ian Lynagh committed
478
  @decimal                     \# / { ifExtension magicHashEnabled } { tok_primint positive 0 1 decimal }
479 480
  0[bB] @binary                \# / { ifExtension magicHashEnabled `alexAndPred`
                                      ifExtension binaryLiteralsEnabled } { tok_primint positive 2 3 binary }
Ian Lynagh's avatar
Ian Lynagh committed
481 482 483
  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 }
484 485
  @negative 0[bB] @binary      \# / { ifExtension magicHashEnabled `alexAndPred`
                                      ifExtension binaryLiteralsEnabled } { tok_primint negative 3 4 binary }
Ian Lynagh's avatar
Ian Lynagh committed
486
  @negative 0[oO] @octal       \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 octal }
487
  @negative 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 hexadecimal }
488

Ian Lynagh's avatar
Ian Lynagh committed
489
  @decimal                     \# \# / { ifExtension magicHashEnabled } { tok_primword 0 2 decimal }
490 491
  0[bB] @binary                \# \# / { ifExtension magicHashEnabled `alexAndPred`
                                         ifExtension binaryLiteralsEnabled } { tok_primword 2 4 binary }
Ian Lynagh's avatar
Ian Lynagh committed
492 493 494
  0[oO] @octal                 \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 octal }
  0[xX] @hexadecimal           \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 hexadecimal }

495 496
  -- Unboxed floats and doubles (:: Float#, :: Double#)
  -- prim_{float,double} work with signed literals
497 498
  @signed @floating_point \# / { ifExtension magicHashEnabled } { init_strtoken 1 tok_primfloat }
  @signed @floating_point \# \# / { ifExtension magicHashEnabled } { init_strtoken 2 tok_primdouble }
499 500 501 502 503 504
}

-- 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.
505
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
506 507
  \'                            { lex_char_tok }
  \"                            { lex_string_tok }
508 509
}

510 511 512 513

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

514
{
515

516 517 518
-- -----------------------------------------------------------------------------
-- The token type

519
data Token
Ian Lynagh's avatar
Ian Lynagh committed
520
  = ITas                        -- Haskell keywords
521 522 523 524 525 526 527 528
  | ITcase
  | ITclass
  | ITdata
  | ITdefault
  | ITderiving
  | ITdo
  | ITelse
  | IThiding
529
  | ITforeign
530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545
  | ITif
  | ITimport
  | ITin
  | ITinfix
  | ITinfixl
  | ITinfixr
  | ITinstance
  | ITlet
  | ITmodule
  | ITnewtype
  | ITof
  | ITqualified
  | ITthen
  | ITtype
  | ITwhere

546
  | ITforall            IsUnicodeSyntax -- GHC extension keywords
547 548 549 550
  | ITexport
  | ITlabel
  | ITdynamic
  | ITsafe
551
  | ITinterruptible
552 553 554
  | ITunsafe
  | ITstdcallconv
  | ITccallconv
555
  | ITcapiconv
556
  | ITprimcallconv
thoughtpolice's avatar
thoughtpolice committed
557
  | ITjavascriptcallconv
558
  | ITmdo
559
  | ITfamily
560
  | ITrole
561 562 563
  | ITgroup
  | ITby
  | ITusing
cactus's avatar
cactus committed
564
  | ITpattern
Facundo Domínguez's avatar
Facundo Domínguez committed
565
  | ITstatic
566

Alan Zimmerman's avatar
Alan Zimmerman committed
567 568 569 570 571 572 573 574
  -- Pragmas, see  note [Pragma source text] in BasicTypes
  | ITinline_prag       SourceText InlineSpec RuleMatchInfo
  | ITspec_prag         SourceText                -- SPECIALISE
  | ITspec_inline_prag  SourceText Bool    -- SPECIALISE INLINE (or NOINLINE)
  | ITsource_prag       SourceText
  | ITrules_prag        SourceText
  | ITwarning_prag      SourceText
  | ITdeprecated_prag   SourceText
575
  | ITline_prag
Alan Zimmerman's avatar
Alan Zimmerman committed
576 577 578 579 580 581
  | ITscc_prag          SourceText
  | ITgenerated_prag    SourceText
  | ITcore_prag         SourceText         -- hdaume: core annotations
  | ITunpack_prag       SourceText
  | ITnounpack_prag     SourceText
  | ITann_prag          SourceText
582
  | ITclose_prag
583 584 585
  | IToptions_prag String
  | ITinclude_prag String
  | ITlanguage_prag
Alan Zimmerman's avatar
Alan Zimmerman committed
586 587 588 589 590 591 592 593 594
  | ITvect_prag         SourceText
  | ITvect_scalar_prag  SourceText
  | ITnovect_prag       SourceText
  | ITminimal_prag      SourceText
  | IToverlappable_prag SourceText  -- instance overlap mode
  | IToverlapping_prag  SourceText  -- instance overlap mode
  | IToverlaps_prag     SourceText  -- instance overlap mode
  | ITincoherent_prag   SourceText  -- instance overlap mode
  | ITctype             SourceText
595

Ian Lynagh's avatar
Ian Lynagh committed
596
  | ITdotdot                    -- reserved symbols
597
  | ITcolon
598
  | ITdcolon            IsUnicodeSyntax
599 600
  | ITequal
  | ITlam
601
  | ITlcase
602
  | ITvbar
603 604
  | ITlarrow            IsUnicodeSyntax
  | ITrarrow            IsUnicodeSyntax
605 606
  | ITat
  | ITtilde
batterseapower's avatar
batterseapower committed
607
  | ITtildehsh
608
  | ITdarrow            IsUnicodeSyntax
609 610 611 612
  | ITminus
  | ITbang
  | ITdot

Ian Lynagh's avatar
Ian Lynagh committed
613
  | ITbiglam                    -- GHC-extension symbols
614

Ian Lynagh's avatar
Ian Lynagh committed
615
  | ITocurly                    -- special symbols
616 617 618 619
  | ITccurly
  | ITvocurly
  | ITvccurly
  | ITobrack
Ian Lynagh's avatar
Ian Lynagh committed
620 621
  | ITopabrack                  -- [:, for parallel arrays with -XParallelArrays
  | ITcpabrack                  -- :], for parallel arrays with -XParallelArrays
622 623 624 625 626 627 628 629 630
  | ITcbrack
  | IToparen
  | ITcparen
  | IToubxparen
  | ITcubxparen
  | ITsemi
  | ITcomma
  | ITunderscore
  | ITbackquote
dreixel's avatar
dreixel committed
631
  | ITsimpleQuote               --  '
632

Ian Lynagh's avatar
Ian Lynagh committed
633
  | ITvarid   FastString        -- identifiers
634 635 636 637 638 639 640 641
  | ITconid   FastString
  | ITvarsym  FastString
  | ITconsym  FastString
  | ITqvarid  (FastString,FastString)
  | ITqconid  (FastString,FastString)
  | ITqvarsym (FastString,FastString)
  | ITqconsym (FastString,FastString)

Ian Lynagh's avatar
Ian Lynagh committed
642
  | ITdupipvarid   FastString   -- GHC extension: implicit param: ?x
Adam Gundry's avatar
Adam Gundry committed
643
  | ITlabelvarid   FastString   -- Overloaded label: #x
644

645 646 647
  | ITchar     SourceText Char       -- Note [Literal source text] in BasicTypes
  | ITstring   SourceText FastString -- Note [Literal source text] in BasicTypes
  | ITinteger  SourceText Integer    -- Note [Literal source text] in BasicTypes
Alan Zimmerman's avatar
Alan Zimmerman committed
648
  | ITrational FractionalLit
649

650 651 652 653
  | ITprimchar   SourceText Char     -- Note [Literal source text] in BasicTypes
  | ITprimstring SourceText ByteString -- Note [Literal source text] @BasicTypes
  | ITprimint    SourceText Integer  -- Note [Literal source text] in BasicTypes
  | ITprimword   SourceText Integer  -- Note [Literal source text] in BasicTypes
654 655
  | ITprimfloat  FractionalLit
  | ITprimdouble FractionalLit
656

657
  -- Template Haskell extension tokens
658
  | ITopenExpQuote HasE         --  [| or [e|
Ian Lynagh's avatar
Ian Lynagh committed
659 660 661 662
  | ITopenPatQuote              --  [p|
  | ITopenDecQuote              --  [d|
  | ITopenTypQuote              --  [t|
  | ITcloseQuote                --  |]
663
  | ITopenTExpQuote HasE        --  [|| or [e||
664
  | ITcloseTExpQuote            --  ||]
Ian Lynagh's avatar
Ian Lynagh committed
665 666
  | ITidEscape   FastString     --  $x
  | ITparenEscape               --  $(
667 668
  | ITidTyEscape   FastString   --  $$x
  | ITparenTyEscape             --  $$(
Ian Lynagh's avatar
Ian Lynagh committed
669
  | ITtyQuote                   --  ''
670 671 672 673 674 675 676 677
  | 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 |]
678 679 680 681

  -- Arrow notation extension
  | ITproc
  | ITrec
682 683 684 685 686 687
  | IToparenbar                  --  (|
  | ITcparenbar                  --  |)
  | ITlarrowtail IsUnicodeSyntax --  -<
  | ITrarrowtail IsUnicodeSyntax --  >-
  | ITLarrowtail IsUnicodeSyntax --  -<<
  | ITRarrowtail IsUnicodeSyntax --  >>-
688

689 690
  | ITunknown String             -- Used when the lexer can't make sense of it
  | ITeof                        -- end of file token
691 692 693 694 695 696 697

  -- 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
698
  | ITdocOptionsOld   String     -- doc options declared "-- # ..."-style
Jedai's avatar
Jedai committed
699 700
  | ITlineComment     String     -- comment starting by "--"
  | ITblockComment    String     -- comment in {- -}
701

702
  deriving Show
703

Alan Zimmerman's avatar
Alan Zimmerman committed
704 705 706
instance Outputable Token where
  ppr x = text (show x)

Alan Zimmerman's avatar
Alan Zimmerman committed
707

708 709 710 711 712 713 714
-- 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)
--
715
reservedWordsFM :: UniqFM (Token, ExtsBitmap)
716
reservedWordsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743
    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 ),

744 745
         ( "forall",         ITforall NormalSyntax,
                                              xbit ExplicitForallBit .|.
746 747
                                              xbit InRulePragBit),
         ( "mdo",            ITmdo,           xbit RecursiveDoBit),
748 749 750
             -- See Note [Lexing type pseudo-keywords]
         ( "family",         ITfamily,        0 ),
         ( "role",           ITrole,          0 ),
751
         ( "pattern",        ITpattern,       xbit PatternSynonymsBit),
Facundo Domínguez's avatar
Facundo Domínguez committed
752
         ( "static",         ITstatic,        0 ),
753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773
         ( "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)
774 775
     ]

776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791
{-----------------------------------
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.
-------------------------------------}
792

793
reservedSymsFM :: UniqFM (Token, ExtsBitmap -> Bool)
794
reservedSymsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
795
    map (\ (x,y,z) -> (mkFastString x,(y,z)))
796
      [ ("..",  ITdotdot,              always)
Ian Lynagh's avatar
Ian Lynagh committed
797
        -- (:) is a reserved op, meaning only list cons
798 799 800 801 802 803 804 805 806 807 808 809 810
       ,(":",   ITcolon,               always)
       ,("::",  ITdcolon NormalSyntax, always)
       ,("=",   ITequal,               always)
       ,("\\",  ITlam,                 always)
       ,("|",   ITvbar,                always)
       ,("<-",  ITlarrow NormalSyntax, always)
       ,("->",  ITrarrow NormalSyntax, always)
       ,("@",   ITat,                  always)
       ,("~",   ITtilde,               always)
       ,("~#",  ITtildehsh,            magicHashEnabled)
       ,("=>",  ITdarrow NormalSyntax, always)
       ,("-",   ITminus,               always)
       ,("!",   ITbang,                always)
Ian Lynagh's avatar
Ian Lynagh committed
811 812

        -- For 'forall a . t'
813
       ,(".", ITdot,  always) -- \i -> explicitForallEnabled i || inRulePrag i)
Ian Lynagh's avatar
Ian Lynagh committed
814

815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834
       ,("-<",  ITlarrowtail NormalSyntax, arrowsEnabled)
       ,(">-",  ITrarrowtail NormalSyntax, arrowsEnabled)
       ,("-<<", ITLarrowtail NormalSyntax, arrowsEnabled)
       ,(">>-", ITRarrowtail NormalSyntax, arrowsEnabled)

       ,("∷",   ITdcolon UnicodeSyntax, unicodeSyntaxEnabled)
       ,("⇒",   ITdarrow UnicodeSyntax, unicodeSyntaxEnabled)
       ,("∀",   ITforall UnicodeSyntax, unicodeSyntaxEnabled)
       ,("→",   ITrarrow UnicodeSyntax, unicodeSyntaxEnabled)
       ,("←",   ITlarrow UnicodeSyntax, unicodeSyntaxEnabled)

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

Simon Marlow's avatar
Simon Marlow committed
835 836 837
        -- 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).
838 839 840 841 842
       ]

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

Ian Lynagh's avatar
Ian Lynagh committed
843
type Action = RealSrcSpan -> StringBuffer -> Int -> P (RealLocated Token)
844

845
special :: Token -> Action
twanvl's avatar
twanvl committed
846
special tok span _buf _len = return (L span tok)
847

848
token, layout_token :: Token -> Action
twanvl's avatar
twanvl committed
849 850
token t span _buf _len = return (L span t)
layout_token t span _buf _len = pushLexState layout >> return (L span t)
851

852 853
idtoken :: (StringBuffer -> Int -> Token) -> Action
idtoken f span buf len = return (L span $! (f buf len))
854

855
skip_one_varid :: (FastString -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
856
skip_one_varid f span buf len
857
  = return (L span $! f (lexemeToFastString (stepOn buf) (len-1)))
858

859 860 861 862
skip_two_varid :: (FastString -> Token) -> Action
skip_two_varid f span buf len
  = return (L span $! f (lexemeToFastString (stepOn (stepOn buf)) (len-2)))

863
strtoken :: (String -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
864
strtoken f span buf len =
865
  return (L span $! (f $! lexemeToString buf len))
866

867
init_strtoken :: Int -> (String -> Token) -> Action
868
-- like strtoken, but drops the last N character(s)
Ian Lynagh's avatar
Ian Lynagh committed
869
init_strtoken drop f span buf len =
870
  return (L span $! (f $! lexemeToString buf (len-drop)))
871 872

begin :: Int -> Action
873
begin code _span _str _len = do pushLexState code; lexToken
874 875

pop :: Action
Ian Lynagh's avatar
Ian Lynagh committed
876 877
pop _span _buf _len = do _ <- popLexState
                         lexToken
878

879 880 881 882 883 884 885 886 887 888 889
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
890
              else failSpanMsgP (RealSrcSpan span) (text "Missing block")
891

892
pop_and :: Action -> Action
Ian Lynagh's avatar
Ian Lynagh committed
893 894
pop_and act span buf len = do _ <- popLexState
                              act span buf len
895

896
{-# INLINE nextCharIs #-}
Ian Lynagh's avatar
Ian Lynagh committed
897
nextCharIs :: StringBuffer -> (Char -> Bool) -> Bool
898 899
nextCharIs buf p = not (atEnd buf) && p (currentChar buf)

900 901 902 903
{-# INLINE nextCharIsNot #-}
nextCharIsNot :: StringBuffer -> (Char -> Bool) -> Bool
nextCharIsNot buf p = not (nextCharIs buf p)

904
notFollowedBy :: Char -> AlexAccPred ExtsBitmap
Ian Lynagh's avatar
Ian Lynagh committed
905
notFollowedBy char _ _ _ (AI _ buf)
906
  = nextCharIsNot buf (== char)
907

908
notFollowedBySymbol :: AlexAccPred ExtsBitmap
909
notFollowedBySymbol _ _ _ (AI _ buf)
910
  = nextCharIsNot buf (`elem` "!#$%&*+./<=>?@\\^|-~")
ross's avatar
ross committed
911

Adam Gundry's avatar
Adam Gundry committed
912 913 914 915
followedByDigit :: AlexAccPred ExtsBitmap
followedByDigit _ _ _ (AI _ buf)
  = afterOptionalSpace buf (\b -> nextCharIs b (`elem` ['0'..'9']))

916 917 918 919 920
-- 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.
921
isNormalComment :: AlexAccPred ExtsBitmap
922
isNormalComment bits _ _ (AI _ buf)
923
  | haddockEnabled bits = notFollowedByDocOrPragma
924
  | otherwise           = nextCharIsNot buf (== '#')
925 926
  where
    notFollowedByDocOrPragma
927
       = afterOptionalSpace buf (\b -> nextCharIsNot b (`elem` "|^*$#"))
928

929 930 931 932 933
afterOptionalSpace :: StringBuffer -> (StringBuffer -> Bool) -> Bool
afterOptionalSpace buf p
    = if nextCharIs buf (== ' ')
      then p (snd (nextChar buf))
      else p buf
934

935
atEOL :: AlexAccPred ExtsBitmap
936
atEOL _ _ _ (AI _ buf) = atEnd buf || currentChar buf == '\n'
937

938
ifExtension :: (ExtsBitmap -> Bool) -> AlexAccPred ExtsBitmap
939 940
ifExtension pred bits _ _ _ = pred bits

941 942 943
multiline_doc_comment :: Action
multiline_doc_comment span buf _len = withLexedDocType (worker "")
  where
944
    worker commentAcc input docType oneLine = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
945
      Just ('\n', input')
946 947 948 949 950 951
        | 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
952

953 954
    checkIfCommentLine input = check (dropNonNewlineSpace input)
      where
955 956 957
        check input = case alexGetChar' input of
          Just ('-', input) -> case alexGetChar' input of
            Just ('-', input) -> case alexGetChar' input of
958 959 960 961 962
              Just (c, _) | c /= '-' -> Just input
              _ -> Nothing
            _ -> Nothing
          _ -> Nothing

963
        dropNonNewlineSpace input = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
964
          Just (c, input')
965 966 967 968
            | isSpace c && c /= '\n' -> dropNonNewlineSpace input'
            | otherwise -> input
          Nothing -> input

Jedai's avatar
Jedai committed
969 970 971 972 973
lineCommentToken :: Action
lineCommentToken span buf len = do
  b <- extension rawTokenStreamEnabled
  if b then strtoken ITlineComment span buf len else lexToken

974 975 976 977
{-
  nested comments require traversing by hand, they can't be parsed
  using regular expressions.
-}
Ian Lynagh's avatar
Ian Lynagh committed
978
nested_comment :: P (RealLocated Token) -> Action
Alan Zimmerman's avatar
Alan Zimmerman committed
979
nested_comment cont span buf len = do
980
  input <- getInput
981
  go (reverse $ lexemeToString buf len) (1::Int) input
982
  where
Alan Zimmerman's avatar
Alan Zimmerman committed
983 984 985 986 987 988
    go commentAcc 0 input = do
      setInput input
      b <- extension rawTokenStreamEnabled
      if b
        then docCommentEnd input commentAcc ITblockComment buf span
        else cont
989
    go commentAcc n input = case alexGetChar' input of
990
      Nothing -> errBrace input span
991
      Just ('-',input) -> case alexGetChar' input of
992
        Nothing  -> errBrace input span
993
        Just ('\125',input) -> go ('\125':'-':commentAcc) (n-1) input -- '}'
Jedai's avatar
Jedai committed
994
        Just (_,_)          -> go ('-':commentAcc) n input
995
      Just ('\123',input) -> case alexGetChar' input of  -- '{' char
996
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
997 998 999
        Just ('-',input) -> go ('-':'\123':commentAcc) (n+1) input
        Just (_,_)       -> go ('\123':commentAcc) n input
      Just (c,input) -> go (c:commentAcc) n input
1000 1001 1002 1003

nested_doc_comment :: Action
nested_doc_comment span buf _len = withLexedDocType (go "")
  where
1004
    go commentAcc input docType _ = case alexGetChar' input of
1005
      Nothing -> errBrace input span
1006
      Just ('-',input) -> case alexGetChar' input of
1007
        Nothing -> errBrace input span
twanvl's avatar
twanvl committed
1008
        Just ('\125',input) ->
1009
          docCommentEnd input commentAcc docType buf span
twanvl's avatar
twanvl committed
1010
        Just (_,_) -> go ('-':commentAcc) input docType False
1011
      Just ('\123', input) -> case alexGetChar' input of
1012 1013 1014 1015 1016
        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
1017
        Just (_,_) -> go ('\123':commentAcc) input docType False
1018 1019
      Just (c,input) -> go (c:commentAcc) input docType False

Ian Lynagh's avatar
Ian Lynagh committed
1020 1021
withLexedDocType :: (AlexInput -> (String -> Token) -> Bool -> P (RealLocated Token))
                 -> P (RealLocated Token)
1022
withLexedDocType lexDocComment = do