Lexer.x 108 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,
73 74
   addAnnotation,AddAnn,addAnnsAt,mkParensApiAnn,
   moveAnnotations
75 76
  ) where

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

88 89 90
-- ghc-boot
import qualified GHC.LanguageExtensions as LangExt

rodlogic's avatar
rodlogic committed
91 92 93 94 95 96 97 98
-- bytestring
import Data.ByteString (ByteString)

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

-- compiler/utils
99
import Bag
100 101 102 103
import Outputable
import StringBuffer
import FastString
import UniqFM
rodlogic's avatar
rodlogic committed
104 105 106 107
import Util             ( readRational )

-- compiler/main
import ErrUtils
108
import DynFlags
rodlogic's avatar
rodlogic committed
109 110 111

-- compiler/basicTypes
import SrcLoc
112
import Module
Alan Zimmerman's avatar
Alan Zimmerman committed
113 114
import BasicTypes     ( InlineSpec(..), RuleMatchInfo(..), FractionalLit(..),
                        SourceText )
115

rodlogic's avatar
rodlogic committed
116 117
-- compiler/parser
import Ctype
Alan Zimmerman's avatar
Alan Zimmerman committed
118 119

import ApiAnnotation
120 121
}

122 123 124
-- -----------------------------------------------------------------------------
-- Alex "Character set macros"

125 126
-- NB: The logic behind these definitions is also reflected in basicTypes/Lexeme.hs
-- Any changes here should likely be reflected there.
127 128 129 130
$unispace    = \x05 -- Trick Alex into handling Unicode. See alexGetByte.
$nl          = [\n\r\f]
$whitechar   = [$nl\v\ $unispace]
$white_no_nl = $whitechar # \n -- TODO #8424
131
$tab         = \t
132 133

$ascdigit  = 0-9
134
$unidigit  = \x03 -- Trick Alex into handling Unicode. See alexGetByte.
135
$decdigit  = $ascdigit -- for now, should really be $digit (ToDo)
136 137 138
$digit     = [$ascdigit $unidigit]

$special   = [\(\)\,\;\[\]\`\{\}]
139 140 141
$ascsymbol = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~\:]
$unisymbol = \x04 -- Trick Alex into handling Unicode. See alexGetByte.
$symbol    = [$ascsymbol $unisymbol] # [$special \_\"\']
142

143
$unilarge  = \x01 -- Trick Alex into handling Unicode. See alexGetByte.
144
$asclarge  = [A-Z]
145 146
$large     = [$asclarge $unilarge]

147
$unismall  = \x02 -- Trick Alex into handling Unicode. See alexGetByte.
148
$ascsmall  = [a-z]
149 150
$small     = [$ascsmall $unismall \_]

151 152
$unigraphic = \x06 -- Trick Alex into handling Unicode. See alexGetByte.
$graphic   = [$small $large $symbol $digit $special $unigraphic \"\']
153

154
$binit     = 0-1
Ian Lynagh's avatar
Ian Lynagh committed
155
$octit     = 0-7
156
$hexit     = [$decdigit A-F a-f]
157 158 159 160

$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 \']
161

162 163
$pragmachar = [$small $large $digit]

164 165
$docsym    = [\| \^ \* \$]

166

167 168 169 170 171 172 173 174
-- -----------------------------------------------------------------------------
-- 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
175

176
@decimal     = $decdigit+
177
@binary      = $binit+
178 179 180 181 182
@octal       = $octit+
@hexadecimal = $hexit+
@exponent    = [eE] [\-\+]? @decimal

@qual = (@conid \.)+
183 184 185 186
@qvarid = @qual @varid
@qconid = @qual @conid
@qvarsym = @qual @varsym
@qconsym = @qual @consym
187 188 189

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

190 191 192 193 194
-- normal signed numerical literals can only be explicitly negative,
-- not explicitly positive (contrast @exponent)
@negative = \-
@signed = @negative ?

195 196 197 198

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

199 200
haskell :-

201 202 203 204 205

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

-- everywhere: skip whitespace
Ian Lynagh's avatar
Ian Lynagh committed
206
$white_no_nl+ ;
207
$tab          { warnTab }
208 209 210 211 212

-- 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
213 214 215 216
-- are). We also rule out nested Haddock comments, if the -haddock flag is
-- set.

"{-" / { isNormalComment } { nested_comment lexToken }
217 218 219 220

-- 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.
221 222

-- Since Haddock comments aren't valid in every state, we need to rule them
Ian Lynagh's avatar
Ian Lynagh committed
223
-- out here.
224 225 226 227 228 229 230

-- 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
231
"-- " ~[$docsym \#] .* { lineCommentToken }
232
"--" [^$symbol \ ] .* { lineCommentToken }
233 234 235

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

Jedai's avatar
Jedai committed
236
"-- " [$docsym \#] .* / { ifExtension (not . haddockEnabled) } { lineCommentToken }
237 238 239 240 241 242 243

-- 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.

244
"---"\-* ~$symbol .* { lineCommentToken }
245 246 247 248

-- 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
249
"--"\-* / { atEOL } { lineCommentToken }
250 251 252 253

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

Jedai's avatar
Jedai committed
254
"-- " / { atEOL } { lineCommentToken }
255 256 257 258 259 260 261 262 263 264 265

-- '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
266
  \n                                    ;
Adam Gundry's avatar
Adam Gundry committed
267 268
  ^\# line                              { begin line_prag1 }
  ^\# / { followedByDigit }             { begin line_prag1 }
Ian Lynagh's avatar
Ian Lynagh committed
269 270 271
  ^\# pragma .* \n                      ; -- GCC 3.3 CPP generated, apparently
  ^\# \! .* \n                          ; -- #!, for scripts
  ()                                    { do_bol }
272 273 274 275 276
}

-- 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.
277
<layout, layout_do, layout_if> {
Ian Lynagh's avatar
Ian Lynagh committed
278 279 280 281
  \{ / { notFollowedBy '-' }            { hopefully_open_brace }
        -- we might encounter {-# here, but {- has been handled already
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
282 283
}

284 285 286 287 288 289
-- after an 'if', a vertical bar starts a layout context for MultiWayIf
<layout_if> {
  \| / { notFollowedBySymbol }          { new_layout_context True ITvbar }
  ()                                    { pop }
}

290
-- do is treated in a subtly different way, see new_layout_context
291 292
<layout>    ()                          { new_layout_context True  ITvocurly }
<layout_do> ()                          { new_layout_context False ITvocurly }
293 294 295 296

-- 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
297
<layout_left>  ()                       { do_layout_left }
298

Ian Lynagh's avatar
Ian Lynagh committed
299
<0,option_prags> \n                     { begin bol }
300

301 302
"{-#" $whitechar* $pragmachar+ / { known_pragma linePrags }
                                { dispatch_pragmas linePrags }
303 304 305

-- single-line line pragmas, of the form
--    # <line> "<file>" <extra-stuff> \n
306
<line_prag1> @decimal                   { setLine line_prag1a }
Ian Lynagh's avatar
Ian Lynagh committed
307 308
<line_prag1a> \" [$graphic \ ]* \"      { setFile line_prag1b }
<line_prag1b> .*                        { pop }
309 310 311

-- Haskell-style line pragmas, of the form
--    {-# LINE <line> "<file>" #-}
312
<line_prag2> @decimal                   { setLine line_prag2a }
Ian Lynagh's avatar
Ian Lynagh committed
313 314
<line_prag2a> \" [$graphic \ ]* \"      { setFile line_prag2b }
<line_prag2b> "#-}"|"-}"                { pop }
315 316
   -- NOTE: accept -} at the end of a LINE pragma, for compatibility
   -- with older versions of GHC which generated these.
317

318
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
319
  "{-#" $whitechar* $pragmachar+
320 321 322 323 324
        $whitechar+ $pragmachar+ / { known_pragma twoWordPrags }
                                 { dispatch_pragmas twoWordPrags }

  "{-#" $whitechar* $pragmachar+ / { known_pragma oneWordPrags }
                                 { dispatch_pragmas oneWordPrags }
325

326
  -- We ignore all these pragmas, but don't generate a warning for them
327 328
  "{-#" $whitechar* $pragmachar+ / { known_pragma ignoredPrags }
                                 { dispatch_pragmas ignoredPrags }
329 330

  -- ToDo: should only be valid inside a pragma:
Ian Lynagh's avatar
Ian Lynagh committed
331
  "#-}"                          { endPrag }
332 333
}

334
<option_prags> {
335 336 337
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
                                   { dispatch_pragmas fileHeaderPrags }

Ian Lynagh's avatar
Ian Lynagh committed
338
  "-- #"                           { multiline_doc_comment }
David Waern's avatar
David Waern committed
339 340
}

341 342
<0> {
  -- In the "0" mode we ignore these pragmas
343
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
344 345 346
                     { nested_comment lexToken }
}

David Waern's avatar
David Waern committed
347
<0> {
Jedai's avatar
Jedai committed
348
  "-- #" .* { lineCommentToken }
349
}
350

351
<0,option_prags> {
352 353
  "{-#"  { warnThen Opt_WarnUnrecognisedPragmas (text "Unrecognised pragma")
                    (nested_comment lexToken) }
354 355
}

356 357
-- '0' state: ordinary lexemes

358 359
-- Haddock comments

360
<0,option_prags> {
David Waern's avatar
David Waern committed
361 362
  "-- " $docsym      / { ifExtension haddockEnabled } { multiline_doc_comment }
  "{-" \ ? $docsym   / { ifExtension haddockEnabled } { nested_doc_comment }
363 364
}

365 366
-- "special" symbols

367
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
368 369
  "[:" / { ifExtension parrEnabled }    { token ITopabrack }
  ":]" / { ifExtension parrEnabled }    { token ITcpabrack }
370
}
Ian Lynagh's avatar
Ian Lynagh committed
371

372
<0> {
373 374 375 376 377 378 379 380 381
  "[|"        / { 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 }
382 383 384 385
  \$ @varid   / { ifExtension thEnabled } { skip_one_varid ITidEscape }
  "$$" @varid / { ifExtension thEnabled } { skip_two_varid ITidTyEscape }
  "$("        / { ifExtension thEnabled } { token ITparenEscape }
  "$$("       / { ifExtension thEnabled } { token ITparenTyEscape }
386

387
  "[" @varid "|"  / { ifExtension qqEnabled }
388
                     { lex_quasiquote_tok }
389 390

  -- qualified quasi-quote (#5555)
391 392
  "[" @qvarid "|"  / { ifExtension qqEnabled }
                     { lex_qquasiquote_tok }
393 394
}

eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
395 396 397 398 399 400 401 402
  -- See Note [Lexing type applications]
<0> {
    [^ $idchar \) ] ^
  "@"
    / { ifExtension typeApplicationEnabled `alexAndPred` notFollowedBySymbol }
    { token ITtypeApp }
}

403
<0> {
ross's avatar
ross committed
404
  "(|" / { ifExtension arrowsEnabled `alexAndPred` notFollowedBySymbol }
Ian Lynagh's avatar
Ian Lynagh committed
405
                                        { special IToparenbar }
406 407 408
  "|)" / { ifExtension arrowsEnabled }  { special ITcparenbar }
}

409
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
410
  \? @varid / { ifExtension ipEnabled } { skip_one_varid ITdupipvarid }
411 412
}

Adam Gundry's avatar
Adam Gundry committed
413 414 415 416 417
<0> {
  "#" @varid / { ifExtension overloadedLabelsEnabled }
               { skip_one_varid ITlabelvarid }
}

418
<0> {
419
  "(#" / { ifExtension unboxedTuplesEnabled }
Ian Lynagh's avatar
Ian Lynagh committed
420 421 422 423 424
         { token IToubxparen }
  "#)" / { ifExtension unboxedTuplesEnabled }
         { token ITcubxparen }
}

425
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
426 427 428 429 430 431 432 433 434 435
  \(                                    { special IToparen }
  \)                                    { special ITcparen }
  \[                                    { special ITobrack }
  \]                                    { special ITcbrack }
  \,                                    { special ITcomma }
  \;                                    { special ITsemi }
  \`                                    { special ITbackquote }

  \{                                    { open_brace }
  \}                                    { close_brace }
436 437
}

438
<0,option_prags> {
439 440
  @qvarid                       { idtoken qvarid }
  @qconid                       { idtoken qconid }
Ian Lynagh's avatar
Ian Lynagh committed
441
  @varid                        { varid }
442
  @conid                        { idtoken conid }
443 444
}

445
<0> {
446 447
  @qvarid "#"+      / { ifExtension magicHashEnabled } { idtoken qvarid }
  @qconid "#"+      / { ifExtension magicHashEnabled } { idtoken qconid }
448
  @varid "#"+       / { ifExtension magicHashEnabled } { varid }
449
  @conid "#"+       / { ifExtension magicHashEnabled } { idtoken conid }
450 451
}

452 453
-- ToDo: - move `var` and (sym) into lexical syntax?
--       - remove backquote from $special?
454
<0> {
455 456
  @qvarsym                                         { idtoken qvarsym }
  @qconsym                                         { idtoken qconsym }
457 458
  @varsym                                          { varsym }
  @consym                                          { consym }
459 460
}

461 462
-- For the normal boxed literals we need to be careful
-- when trying to be close to Haskell98
463
<0> {
464
  -- Normal integral literals (:: Num a => a, from Integer)
465
  @decimal                                                               { tok_num positive 0 0 decimal }
466
  0[bB] @binary                / { ifExtension binaryLiteralsEnabled }   { tok_num positive 2 2 binary }
467 468 469
  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 }
470 471
  @negative 0[bB] @binary      / { ifExtension negativeLiteralsEnabled `alexAndPred`
                                   ifExtension binaryLiteralsEnabled }   { tok_num negative 3 3 binary }
472 473
  @negative 0[oO] @octal       / { ifExtension negativeLiteralsEnabled } { tok_num negative 3 3 octal }
  @negative 0[xX] @hexadecimal / { ifExtension negativeLiteralsEnabled } { tok_num negative 3 3 hexadecimal }
474 475

  -- Normal rational literals (:: Fractional a => a, from Rational)
476 477
  @floating_point                                                        { strtoken tok_float }
  @negative @floating_point    / { ifExtension negativeLiteralsEnabled } { strtoken tok_float }
478 479
}

480
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
481
  -- Unboxed ints (:: Int#) and words (:: Word#)
482 483
  -- It's simpler (and faster?) to give separate cases to the negatives,
  -- especially considering octal/hexadecimal prefixes.
Ian Lynagh's avatar
Ian Lynagh committed
484
  @decimal                     \# / { ifExtension magicHashEnabled } { tok_primint positive 0 1 decimal }
485 486
  0[bB] @binary                \# / { ifExtension magicHashEnabled `alexAndPred`
                                      ifExtension binaryLiteralsEnabled } { tok_primint positive 2 3 binary }
Ian Lynagh's avatar
Ian Lynagh committed
487 488 489
  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 }
490 491
  @negative 0[bB] @binary      \# / { ifExtension magicHashEnabled `alexAndPred`
                                      ifExtension binaryLiteralsEnabled } { tok_primint negative 3 4 binary }
Ian Lynagh's avatar
Ian Lynagh committed
492
  @negative 0[oO] @octal       \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 octal }
493
  @negative 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 hexadecimal }
494

Ian Lynagh's avatar
Ian Lynagh committed
495
  @decimal                     \# \# / { ifExtension magicHashEnabled } { tok_primword 0 2 decimal }
496 497
  0[bB] @binary                \# \# / { ifExtension magicHashEnabled `alexAndPred`
                                         ifExtension binaryLiteralsEnabled } { tok_primword 2 4 binary }
Ian Lynagh's avatar
Ian Lynagh committed
498 499 500
  0[oO] @octal                 \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 octal }
  0[xX] @hexadecimal           \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 hexadecimal }

501 502
  -- Unboxed floats and doubles (:: Float#, :: Double#)
  -- prim_{float,double} work with signed literals
503 504
  @signed @floating_point \# / { ifExtension magicHashEnabled } { init_strtoken 1 tok_primfloat }
  @signed @floating_point \# \# / { ifExtension magicHashEnabled } { init_strtoken 2 tok_primdouble }
505 506 507 508 509 510
}

-- 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.
511
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
512 513
  \'                            { lex_char_tok }
  \"                            { lex_string_tok }
514 515
}

eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541
-- Note [Lexing type applications]
-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-- The desired syntax for type applications is to prefix the type application
-- with '@', like this:
--
--   foo @Int @Bool baz bum
--
-- This, of course, conflicts with as-patterns. The conflict arises because
-- expressions and patterns use the same parser, and also because we want
-- to allow type patterns within expression patterns.
--
-- Disambiguation is accomplished by requiring *something* to appear betwen
-- type application and the preceding token. This something must end with
-- a character that cannot be the end of the variable bound in an as-pattern.
-- Currently (June 2015), this means that the something cannot end with a
-- $idchar or a close-paren. (The close-paren is necessary if the as-bound
-- identifier is symbolic.)
--
-- Note that looking for whitespace before the '@' is insufficient, because
-- of this pathological case:
--
--   foo {- hi -}@Int
--
-- This design is predicated on the fact that as-patterns are generally
-- whitespace-free, and also that this whole thing is opt-in, with the
-- TypeApplications extension.
542 543 544 545

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

546
{
547

548 549 550
-- -----------------------------------------------------------------------------
-- The token type

551
data Token
Ian Lynagh's avatar
Ian Lynagh committed
552
  = ITas                        -- Haskell keywords
553 554 555 556 557 558 559 560
  | ITcase
  | ITclass
  | ITdata
  | ITdefault
  | ITderiving
  | ITdo
  | ITelse
  | IThiding
561
  | ITforeign
562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577
  | ITif
  | ITimport
  | ITin
  | ITinfix
  | ITinfixl
  | ITinfixr
  | ITinstance
  | ITlet
  | ITmodule
  | ITnewtype
  | ITof
  | ITqualified
  | ITthen
  | ITtype
  | ITwhere

578
  | ITforall            IsUnicodeSyntax -- GHC extension keywords
579 580 581 582
  | ITexport
  | ITlabel
  | ITdynamic
  | ITsafe
583
  | ITinterruptible
584 585 586
  | ITunsafe
  | ITstdcallconv
  | ITccallconv
587
  | ITcapiconv
588
  | ITprimcallconv
thoughtpolice's avatar
thoughtpolice committed
589
  | ITjavascriptcallconv
590
  | ITmdo
591
  | ITfamily
592
  | ITrole
593 594 595
  | ITgroup
  | ITby
  | ITusing
cactus's avatar
cactus committed
596
  | ITpattern
Facundo Domínguez's avatar
Facundo Domínguez committed
597
  | ITstatic
598

Alan Zimmerman's avatar
Alan Zimmerman committed
599 600 601 602 603 604 605 606
  -- 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
607
  | ITline_prag
Alan Zimmerman's avatar
Alan Zimmerman committed
608 609 610 611 612 613
  | ITscc_prag          SourceText
  | ITgenerated_prag    SourceText
  | ITcore_prag         SourceText         -- hdaume: core annotations
  | ITunpack_prag       SourceText
  | ITnounpack_prag     SourceText
  | ITann_prag          SourceText
614
  | ITclose_prag
615 616 617
  | IToptions_prag String
  | ITinclude_prag String
  | ITlanguage_prag
Alan Zimmerman's avatar
Alan Zimmerman committed
618 619 620 621 622 623 624 625 626
  | 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
627

Ian Lynagh's avatar
Ian Lynagh committed
628
  | ITdotdot                    -- reserved symbols
629
  | ITcolon
630
  | ITdcolon            IsUnicodeSyntax
631 632
  | ITequal
  | ITlam
633
  | ITlcase
634
  | ITvbar
635 636
  | ITlarrow            IsUnicodeSyntax
  | ITrarrow            IsUnicodeSyntax
637 638
  | ITat
  | ITtilde
batterseapower's avatar
batterseapower committed
639
  | ITtildehsh
640
  | ITdarrow            IsUnicodeSyntax
641 642 643 644
  | ITminus
  | ITbang
  | ITdot

Ian Lynagh's avatar
Ian Lynagh committed
645
  | ITbiglam                    -- GHC-extension symbols
646

Ian Lynagh's avatar
Ian Lynagh committed
647
  | ITocurly                    -- special symbols
648 649 650 651
  | ITccurly
  | ITvocurly
  | ITvccurly
  | ITobrack
Ian Lynagh's avatar
Ian Lynagh committed
652 653
  | ITopabrack                  -- [:, for parallel arrays with -XParallelArrays
  | ITcpabrack                  -- :], for parallel arrays with -XParallelArrays
654 655 656 657 658 659 660 661 662
  | ITcbrack
  | IToparen
  | ITcparen
  | IToubxparen
  | ITcubxparen
  | ITsemi
  | ITcomma
  | ITunderscore
  | ITbackquote
dreixel's avatar
dreixel committed
663
  | ITsimpleQuote               --  '
664

Ian Lynagh's avatar
Ian Lynagh committed
665
  | ITvarid   FastString        -- identifiers
666 667 668 669 670 671 672 673
  | 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
674
  | ITdupipvarid   FastString   -- GHC extension: implicit param: ?x
Adam Gundry's avatar
Adam Gundry committed
675
  | ITlabelvarid   FastString   -- Overloaded label: #x
676

677 678 679
  | 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
680
  | ITrational FractionalLit
681

682 683 684 685
  | 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
686 687
  | ITprimfloat  FractionalLit
  | ITprimdouble FractionalLit
688

689
  -- Template Haskell extension tokens
690
  | ITopenExpQuote HasE         --  [| or [e|
Ian Lynagh's avatar
Ian Lynagh committed
691 692 693 694
  | ITopenPatQuote              --  [p|
  | ITopenDecQuote              --  [d|
  | ITopenTypQuote              --  [t|
  | ITcloseQuote                --  |]
695
  | ITopenTExpQuote HasE        --  [|| or [e||
696
  | ITcloseTExpQuote            --  ||]
Ian Lynagh's avatar
Ian Lynagh committed
697 698
  | ITidEscape   FastString     --  $x
  | ITparenEscape               --  $(
699 700
  | ITidTyEscape   FastString   --  $$x
  | ITparenTyEscape             --  $$(
Ian Lynagh's avatar
Ian Lynagh committed
701
  | ITtyQuote                   --  ''
702 703 704 705 706 707 708 709
  | 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 |]
710 711 712 713

  -- Arrow notation extension
  | ITproc
  | ITrec
714 715 716 717 718 719
  | IToparenbar                  --  (|
  | ITcparenbar                  --  |)
  | ITlarrowtail IsUnicodeSyntax --  -<
  | ITrarrowtail IsUnicodeSyntax --  >-
  | ITLarrowtail IsUnicodeSyntax --  -<<
  | ITRarrowtail IsUnicodeSyntax --  >>-
720

eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
721 722 723 724 725 726 727
  -- type application '@' (lexed differently than as-pattern '@',
  -- due to checking for preceding whitespace)
  | ITtypeApp


  | ITunknown String            -- Used when the lexer can't make sense of it
  | ITeof                       -- end of file token
728 729 730 731 732 733 734

  -- 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
735
  | ITdocOptionsOld   String     -- doc options declared "-- # ..."-style
Jedai's avatar
Jedai committed
736 737
  | ITlineComment     String     -- comment starting by "--"
  | ITblockComment    String     -- comment in {- -}
738

739
  deriving Show
740

Alan Zimmerman's avatar
Alan Zimmerman committed
741 742 743
instance Outputable Token where
  ppr x = text (show x)

Alan Zimmerman's avatar
Alan Zimmerman committed
744

745 746 747 748 749 750 751
-- 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)
--
752
reservedWordsFM :: UniqFM (Token, ExtsBitmap)
753
reservedWordsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780
    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 ),

781 782
         ( "forall",         ITforall NormalSyntax,
                                              xbit ExplicitForallBit .|.
783 784
                                              xbit InRulePragBit),
         ( "mdo",            ITmdo,           xbit RecursiveDoBit),
785 786 787
             -- See Note [Lexing type pseudo-keywords]
         ( "family",         ITfamily,        0 ),
         ( "role",           ITrole,          0 ),
788
         ( "pattern",        ITpattern,       xbit PatternSynonymsBit),
Facundo Domínguez's avatar
Facundo Domínguez committed
789
         ( "static",         ITstatic,        0 ),
790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810
         ( "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)
811 812
     ]

813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828
{-----------------------------------
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.
-------------------------------------}
829

830
reservedSymsFM :: UniqFM (Token, ExtsBitmap -> Bool)
831
reservedSymsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
832
    map (\ (x,y,z) -> (mkFastString x,(y,z)))
833
      [ ("..",  ITdotdot,              always)
Ian Lynagh's avatar
Ian Lynagh committed
834
        -- (:) is a reserved op, meaning only list cons
835 836 837 838 839 840 841 842 843 844 845 846 847
       ,(":",   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
848 849

        -- For 'forall a . t'
850
       ,(".", ITdot,  always) -- \i -> explicitForallEnabled i || inRulePrag i)
Ian Lynagh's avatar
Ian Lynagh committed
851

852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871
       ,("-<",  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
872 873 874
        -- 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).
875 876 877 878 879
       ]

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

Ian Lynagh's avatar
Ian Lynagh committed
880
type Action = RealSrcSpan -> StringBuffer -> Int -> P (RealLocated Token)
881

882
special :: Token -> Action
twanvl's avatar
twanvl committed
883
special tok span _buf _len = return (L span tok)
884

885
token, layout_token :: Token -> Action
twanvl's avatar
twanvl committed
886 887
token t span _buf _len = return (L span t)
layout_token t span _buf _len = pushLexState layout >> return (L span t)
888

889 890
idtoken :: (StringBuffer -> Int -> Token) -> Action
idtoken f span buf len = return (L span $! (f buf len))
891

892
skip_one_varid :: (FastString -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
893
skip_one_varid f span buf len
894
  = return (L span $! f (lexemeToFastString (stepOn buf) (len-1)))
895

896 897 898 899
skip_two_varid :: (FastString -> Token) -> Action
skip_two_varid f span buf len
  = return (L span $! f (lexemeToFastString (stepOn (stepOn buf)) (len-2)))

900
strtoken :: (String -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
901
strtoken f span buf len =
902
  return (L span $! (f $! lexemeToString buf len))
903

904
init_strtoken :: Int -> (String -> Token) -> Action
905
-- like strtoken, but drops the last N character(s)
Ian Lynagh's avatar
Ian Lynagh committed
906
init_strtoken drop f span buf len =
907
  return (L span $! (f $! lexemeToString buf (len-drop)))
908 909

begin :: Int -> Action
910
begin code _span _str _len = do pushLexState code; lexToken
911 912

pop :: Action
Ian Lynagh's avatar
Ian Lynagh committed
913 914
pop _span _buf _len = do _ <- popLexState
                         lexToken
915

916 917 918 919 920 921 922 923 924 925 926
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
927
              else failSpanMsgP (RealSrcSpan span) (text "Missing block")
928

929
pop_and :: Action -> Action
Ian Lynagh's avatar
Ian Lynagh committed
930 931
pop_and act span buf len = do _ <- popLexState
                              act span buf len
932

933
{-# INLINE nextCharIs #-}
Ian Lynagh's avatar
Ian Lynagh committed
934
nextCharIs :: StringBuffer -> (Char -> Bool) -> Bool
935 936
nextCharIs buf p = not (atEnd buf) && p (currentChar buf)

937 938 939 940
{-# INLINE nextCharIsNot #-}
nextCharIsNot :: StringBuffer -> (Char -> Bool) -> Bool
nextCharIsNot buf p = not (nextCharIs buf p)

941
notFollowedBy :: Char -> AlexAccPred ExtsBitmap
Ian Lynagh's avatar
Ian Lynagh committed
942
notFollowedBy char _ _ _ (AI _ buf)
943
  = nextCharIsNot buf (== char)
944

945
notFollowedBySymbol :: AlexAccPred ExtsBitmap
946
notFollowedBySymbol _ _ _ (AI _ buf)
947
  = nextCharIsNot buf (`elem` "!#$%&*+./<=>?@\\^|-~")
ross's avatar
ross committed
948

Adam Gundry's avatar
Adam Gundry committed
949 950 951 952
followedByDigit :: AlexAccPred ExtsBitmap
followedByDigit _ _ _ (AI _ buf)
  = afterOptionalSpace buf (\b -> nextCharIs b (`elem` ['0'..'9']))

953 954 955 956 957
-- 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.
958
isNormalComment :: AlexAccPred ExtsBitmap
959
isNormalComment bits _ _ (AI _ buf)
960
  | haddockEnabled bits = notFollowedByDocOrPragma
961
  | otherwise           = nextCharIsNot buf (== '#')
962 963
  where
    notFollowedByDocOrPragma
964
       = afterOptionalSpace buf (\b -> nextCharIsNot b (`elem` "|^*$#"))
965

966 967 968 969 970
afterOptionalSpace :: StringBuffer -> (StringBuffer -> Bool) -> Bool
afterOptionalSpace buf p
    = if nextCharIs buf (== ' ')
      then p (snd (nextChar buf))
      else p buf
971

972
atEOL :: AlexAccPred ExtsBitmap
973
atEOL _ _ _ (AI _ buf) = atEnd buf || currentChar buf == '\n'
974

975
ifExtension :: (ExtsBitmap -> Bool) -> AlexAccPred ExtsBitmap
976 977
ifExtension pred bits _ _ _ = pred bits

978 979 980
multiline_doc_comment :: Action
multiline_doc_comment span buf _len = withLexedDocType (worker "")
  where
981
    worker commentAcc input docType oneLine = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
982
      Just ('\n', input')
983 984 985 986 987 988
        | 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
989

990 991
    checkIfCommentLine input = check (dropNonNewlineSpace input)
      where
992 993 994
        check input = case alexGetChar' input of
          Just ('-', input) -> case alexGetChar' input of
            Just ('-', input) -> case alexGetChar' input of
995 996 997 998 999
              Just (c, _) | c /= '-' -> Just input
              _ -> Nothing
            _ -> Nothing
          _ -> Nothing

1000
        dropNonNewlineSpace input = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
1001
          Just (c, input')
1002 1003 1004 1005
            | isSpace c && c /= '\n' -> dropNonNewlineSpace input'
            | otherwise -> input
          Nothing -> input

Jedai's avatar
Jedai committed
1006 1007 1008 1009 1010
lineCommentToken :: Action
lineCommentToken span buf len = do
  b <- extension rawTokenStreamEnabled
  if b then strtoken ITlineComment span buf len else lexToken

1011 1012 1013 1014
{-
  nested comments require traversing by hand, they can't be parsed
  using regular expressions.
-}
Ian Lynagh's avatar
Ian Lynagh committed
1015
nested_comment :: P (RealLocated Token) -> Action
Alan Zimmerman's avatar
Alan Zimmerman committed
1016
nested_comment cont span buf len = do
1017
  input <- getInput
1018
  go (reverse $ lexemeToString buf len) (1::Int) input
1019
  where
Alan Zimmerman's avatar
Alan Zimmerman committed
1020 1021 1022 1023 1024 1025
    go commentAcc 0 input = do
      setInput input
      b <- extension rawTokenStreamEnabled
      if b
        then docCommentEnd input commentAcc ITblockComment buf span
        else cont
1026
    go commentAcc n input = case alexGetChar' input of
1027
      Nothing -> errBrace input span
1028
      Just ('-',input) -> case alexGetChar' input of
1029
        Nothing  -> errBrace input span
1030
        Just ('\125',input) -> go ('\125':'-':commentAcc) (n-1) input -- '}'
Jedai's avatar
Jedai committed
1031
        Just (_,_)          -> go ('-':commentAcc) n input
1032
      Just ('\123',input) -> case alexGetChar' input of  -- '{' char
1033
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
1034 1035 1036
        Just ('-',input) -> go ('-':'\123':commentAcc) (n+1) input
        Just (_,_)       -> go ('\123':commentAcc) n input
      Just (c,input) -> go (c:commentAcc) n input
1037 1038 1039 1040

nested_doc_comment :: Action
nested_doc_comment span buf _len = withLexedDocType (go "")
  where
1041
    go commentAcc input docType _ = case alexGetChar' input of
1042
      Nothing -> errBrace input span
1043
      Just ('-',input) -> case alexGetChar' input of
1044
        Nothing -> errBrace input span
twanvl's avatar
twanvl committed
1045
        Just ('\125',input) ->
1046
          docCommentEnd input commentAcc docType buf span
twanvl's avatar
twanvl committed
1047
        Just (_,_) -> go ('-':commentAcc) input docType False
1048
      Just ('\123', input) -> case alexGetChar' input of
1049 1050 1051 1052 1053
        Nothing  -> errBrace input span
        Just ('-',input) -> do