Lexer.x 130 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
-- References:
-- [1] https://www.haskell.org/onlinereport/haskell2010/haskellch2.html
-- [2] http://www.haskell.org/alex/
16
-- [3] https://gitlab.haskell.org/ghc/ghc/wikis/commentary/compiler/parser
17
--
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
{-# LANGUAGE LambdaCase #-}
46
{-# LANGUAGE MultiWayIf #-}
47

48
{-# OPTIONS_GHC -funbox-strict-fields #-}
49
{-# OPTIONS_GHC -Wno-incomplete-uni-patterns #-}
50

Sylvain Henry's avatar
Sylvain Henry committed
51
module GHC.Parser.Lexer (
52
   Token(..), lexer, lexerDbg, pragState, mkPState, mkPStatePure, PState(..),
Vladislav Zavialov's avatar
Vladislav Zavialov committed
53 54 55 56
   P(..), ParseResult(..), mkParserFlags, mkParserFlags', ParserFlags(..),
   appendWarning,
   appendError,
   allocateComments,
57
   MonadP(..),
Sylvain Henry's avatar
Sylvain Henry committed
58
   getRealSrcLoc, getPState, withHomeUnitId,
59
   failMsgP, failLocMsgP, srcParseFail,
60
   getErrorMessages, getMessages,
61
   popContext, pushModuleContext, setLastToken, setSrcLoc,
vivian's avatar
vivian committed
62
   activeContext, nextIsEOF,
simonpj@microsoft.com's avatar
simonpj@microsoft.com committed
63
   getLexState, popLexState, pushLexState,
64
   ExtBits(..),
Vladislav Zavialov's avatar
Vladislav Zavialov committed
65
   xtest,
Alan Zimmerman's avatar
Alan Zimmerman committed
66
   lexTokenStream,
Vladislav Zavialov's avatar
Vladislav Zavialov committed
67 68
   AddAnn(..),mkParensApiAnn,
   addAnnsAt,
69
   commentToAnnotation
70 71
  ) where

72
import GHC.Prelude
73

rodlogic's avatar
rodlogic committed
74 75 76
-- base
import Control.Monad
import Data.Bits
77
import Data.Char
rodlogic's avatar
rodlogic committed
78 79 80 81
import Data.List
import Data.Maybe
import Data.Word

82
import GHC.Data.EnumSet as EnumSet
83

84 85 86
-- ghc-boot
import qualified GHC.LanguageExtensions as LangExt

rodlogic's avatar
rodlogic committed
87 88 89 90 91 92 93 94
-- bytestring
import Data.ByteString (ByteString)

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

-- compiler/utils
95 96 97 98
import GHC.Data.Bag
import GHC.Utils.Outputable
import GHC.Data.StringBuffer
import GHC.Data.FastString
Sylvain Henry's avatar
Sylvain Henry committed
99
import GHC.Types.Unique.FM
100
import GHC.Utils.Misc ( readRational, readHexRational )
rodlogic's avatar
rodlogic committed
101 102

-- compiler/main
103
import GHC.Utils.Error
Sylvain Henry's avatar
Sylvain Henry committed
104
import GHC.Driver.Session as DynFlags
rodlogic's avatar
rodlogic committed
105 106

-- compiler/basicTypes
Sylvain Henry's avatar
Sylvain Henry committed
107
import GHC.Types.SrcLoc
Sylvain Henry's avatar
Sylvain Henry committed
108
import GHC.Unit
Sylvain Henry's avatar
Sylvain Henry committed
109 110 111
import GHC.Types.Basic ( InlineSpec(..), RuleMatchInfo(..),
                         IntegralLit(..), FractionalLit(..),
                         SourceText(..) )
112

rodlogic's avatar
rodlogic committed
113
-- compiler/parser
Sylvain Henry's avatar
Sylvain Henry committed
114
import GHC.Parser.CharClass
Alan Zimmerman's avatar
Alan Zimmerman committed
115

Sylvain Henry's avatar
Sylvain Henry committed
116
import GHC.Parser.Annotation
117 118
}

119 120 121
-- -----------------------------------------------------------------------------
-- Alex "Character set macros"

122
-- NB: The logic behind these definitions is also reflected in "GHC.Utils.Lexeme"
123
-- Any changes here should likely be reflected there.
124
$unispace    = \x05 -- Trick Alex into handling Unicode. See [Unicode in Alex].
125 126 127
$nl          = [\n\r\f]
$whitechar   = [$nl\v\ $unispace]
$white_no_nl = $whitechar # \n -- TODO #8424
128
$tab         = \t
129 130

$ascdigit  = 0-9
131
$unidigit  = \x03 -- Trick Alex into handling Unicode. See [Unicode in Alex].
132
$decdigit  = $ascdigit -- for now, should really be $digit (ToDo)
133 134 135
$digit     = [$ascdigit $unidigit]

$special   = [\(\)\,\;\[\]\`\{\}]
136
$ascsymbol = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~\:]
137
$unisymbol = \x04 -- Trick Alex into handling Unicode. See [Unicode in Alex].
138
$symbol    = [$ascsymbol $unisymbol] # [$special \_\"\']
139

140
$unilarge  = \x01 -- Trick Alex into handling Unicode. See [Unicode in Alex].
141
$asclarge  = [A-Z]
142 143
$large     = [$asclarge $unilarge]

144
$unismall  = \x02 -- Trick Alex into handling Unicode. See [Unicode in Alex].
145
$ascsmall  = [a-z]
146 147
$small     = [$ascsmall $unismall \_]

148
$unigraphic = \x06 -- Trick Alex into handling Unicode. See [Unicode in Alex].
149
$graphic   = [$small $large $symbol $digit $special $unigraphic \"\']
150

151
$binit     = 0-1
Ian Lynagh's avatar
Ian Lynagh committed
152
$octit     = 0-7
153
$hexit     = [$decdigit A-F a-f]
154

155
$uniidchar = \x07 -- Trick Alex into handling Unicode. See [Unicode in Alex].
156
$idchar    = [$small $large $digit $uniidchar \']
157

158 159
$pragmachar = [$small $large $digit]

160 161
$docsym    = [\| \^ \* \$]

162

163 164 165 166 167 168 169 170
-- -----------------------------------------------------------------------------
-- 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
171

172 173 174 175 176 177 178 179
-- See Note [Lexing NumericUnderscores extension] and #14473
@numspc       = _*                   -- numeric spacer (#14473)
@decimal      = $decdigit(@numspc $decdigit)*
@binary       = $binit(@numspc $binit)*
@octal        = $octit(@numspc $octit)*
@hexadecimal  = $hexit(@numspc $hexit)*
@exponent     = @numspc [eE] [\-\+]? @decimal
@bin_exponent = @numspc [pP] [\-\+]? @decimal
180 181

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

187 188 189 190 191
-- QualifiedDo needs to parse "M.do" not as a variable, so as to keep the
-- layout rules.
@qdo    = @qual "do"
@qmdo   = @qual "mdo"

192 193
@floating_point = @numspc @decimal \. @decimal @exponent? | @numspc @decimal @exponent
@hex_floating_point = @numspc @hexadecimal \. @hexadecimal @bin_exponent? | @numspc @hexadecimal @bin_exponent
194

195 196 197 198 199
-- normal signed numerical literals can only be explicitly negative,
-- not explicitly positive (contrast @exponent)
@negative = \-
@signed = @negative ?

200 201 202 203

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

204 205
haskell :-

206 207 208 209 210

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

-- everywhere: skip whitespace
Ian Lynagh's avatar
Ian Lynagh committed
211
$white_no_nl+ ;
212
$tab          { warnTab }
213 214 215 216 217

-- 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
218 219 220 221
-- are). We also rule out nested Haddock comments, if the -haddock flag is
-- set.

"{-" / { isNormalComment } { nested_comment lexToken }
222 223 224 225

-- 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.
226 227

-- Since Haddock comments aren't valid in every state, we need to rule them
Ian Lynagh's avatar
Ian Lynagh committed
228
-- out here.
229 230 231 232 233 234 235

-- 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.

236
"-- " ~$docsym .* { lineCommentToken }
237
"--" [^$symbol \ ] .* { lineCommentToken }
238 239 240

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

241
"-- " $docsym .* / { alexNotPred (ifExtension HaddockBit) } { lineCommentToken }
242 243 244 245 246 247 248

-- 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.

249
"---"\-* ~$symbol .* { lineCommentToken }
250 251 252 253

-- 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
254
"--"\-* / { atEOL } { lineCommentToken }
255 256 257 258

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

Jedai's avatar
Jedai committed
259
"-- " / { atEOL } { lineCommentToken }
260 261 262 263 264 265 266 267 268 269 270

-- '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
271
  \n                                    ;
Adam Gundry's avatar
Adam Gundry committed
272 273
  ^\# line                              { begin line_prag1 }
  ^\# / { followedByDigit }             { begin line_prag1 }
Ian Lynagh's avatar
Ian Lynagh committed
274 275 276
  ^\# pragma .* \n                      ; -- GCC 3.3 CPP generated, apparently
  ^\# \! .* \n                          ; -- #!, for scripts
  ()                                    { do_bol }
277 278 279 280 281
}

-- 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.
282
<layout, layout_do, layout_if> {
Ian Lynagh's avatar
Ian Lynagh committed
283 284 285 286
  \{ / { notFollowedBy '-' }            { hopefully_open_brace }
        -- we might encounter {-# here, but {- has been handled already
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
287 288
}

289 290
-- after an 'if', a vertical bar starts a layout context for MultiWayIf
<layout_if> {
291
  \| / { notFollowedBySymbol }          { new_layout_context True dontGenerateSemic ITvbar }
292 293 294
  ()                                    { pop }
}

295
-- do is treated in a subtly different way, see new_layout_context
296 297
<layout>    ()                          { new_layout_context True  generateSemic ITvocurly }
<layout_do> ()                          { new_layout_context False generateSemic ITvocurly }
298 299 300 301

-- 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
302
<layout_left>  ()                       { do_layout_left }
303

Ian Lynagh's avatar
Ian Lynagh committed
304
<0,option_prags> \n                     { begin bol }
305

306 307
"{-#" $whitechar* $pragmachar+ / { known_pragma linePrags }
                                { dispatch_pragmas linePrags }
308 309 310

-- single-line line pragmas, of the form
--    # <line> "<file>" <extra-stuff> \n
311 312 313 314 315
<line_prag1> {
  @decimal $white_no_nl+ \" [$graphic \ ]* \"  { setLineAndFile line_prag1a }
  ()                                           { failLinePrag1 }
}
<line_prag1a> .*                               { popLinePrag1 }
316 317 318

-- Haskell-style line pragmas, of the form
--    {-# LINE <line> "<file>" #-}
319 320 321 322
<line_prag2> {
  @decimal $white_no_nl+ \" [$graphic \ ]* \"  { setLineAndFile line_prag2a }
}
<line_prag2a> "#-}"|"-}"                       { pop }
323 324
   -- NOTE: accept -} at the end of a LINE pragma, for compatibility
   -- with older versions of GHC which generated these.
325

Rufflewind's avatar
Rufflewind committed
326 327 328 329
-- Haskell-style column pragmas, of the form
--    {-# COLUMN <column> #-}
<column_prag> @decimal $whitechar* "#-}" { setColumn }

330
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
331
  "{-#" $whitechar* $pragmachar+
332 333 334 335 336
        $whitechar+ $pragmachar+ / { known_pragma twoWordPrags }
                                 { dispatch_pragmas twoWordPrags }

  "{-#" $whitechar* $pragmachar+ / { known_pragma oneWordPrags }
                                 { dispatch_pragmas oneWordPrags }
337

338
  -- We ignore all these pragmas, but don't generate a warning for them
339 340
  "{-#" $whitechar* $pragmachar+ / { known_pragma ignoredPrags }
                                 { dispatch_pragmas ignoredPrags }
341 342

  -- ToDo: should only be valid inside a pragma:
Ian Lynagh's avatar
Ian Lynagh committed
343
  "#-}"                          { endPrag }
344 345
}

346
<option_prags> {
347 348
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
                                   { dispatch_pragmas fileHeaderPrags }
David Waern's avatar
David Waern committed
349 350
}

351 352
<0> {
  -- In the "0" mode we ignore these pragmas
353
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
354 355 356
                     { nested_comment lexToken }
}

357
<0,option_prags> {
358 359
  "{-#"  { warnThen Opt_WarnUnrecognisedPragmas (text "Unrecognised pragma")
                    (nested_comment lexToken) }
360 361
}

362 363
-- '0' state: ordinary lexemes

364 365
-- Haddock comments

366
<0,option_prags> {
367 368
  "-- " $docsym      / { ifExtension HaddockBit } { multiline_doc_comment }
  "{-" \ ? $docsym   / { ifExtension HaddockBit } { nested_doc_comment }
369 370
}

371 372
-- "special" symbols

373
<0> {
374 375 376 377 378 379 380 381 382

  -- Don't check ThQuotesBit here as the renamer can produce a better
  -- error message than the lexer (see the thQuotesEnabled check in rnBracket).
  "[|"  { token (ITopenExpQuote NoE NormalSyntax) }
  "[||" { token (ITopenTExpQuote NoE) }
  "|]"  { token (ITcloseQuote NormalSyntax) }
  "||]" { token ITcloseTExpQuote }

  -- Check ThQuotesBit here as to not steal syntax.
383 384 385 386 387 388 389
  "[e|"       / { ifExtension ThQuotesBit } { token (ITopenExpQuote HasE NormalSyntax) }
  "[e||"      / { ifExtension ThQuotesBit } { token (ITopenTExpQuote HasE) }
  "[p|"       / { ifExtension ThQuotesBit } { token ITopenPatQuote }
  "[d|"       / { ifExtension ThQuotesBit } { layout_token ITopenDecQuote }
  "[t|"       / { ifExtension ThQuotesBit } { token ITopenTypQuote }

  "[" @varid "|"  / { ifExtension QqBit }   { lex_quasiquote_tok }
390 391

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

  $unigraphic -- ⟦
    / { ifCurrentChar '⟦' `alexAndPred`
396 397
        ifExtension UnicodeSyntaxBit `alexAndPred`
        ifExtension ThQuotesBit }
398 399 400
    { token (ITopenExpQuote NoE UnicodeSyntax) }
  $unigraphic -- ⟧
    / { ifCurrentChar '⟧' `alexAndPred`
401 402
        ifExtension UnicodeSyntaxBit `alexAndPred`
        ifExtension ThQuotesBit }
403
    { token (ITcloseQuote UnicodeSyntax) }
404 405
}

406
<0> {
407 408 409 410 411 412 413
  "(|"
    / { ifExtension ArrowsBit `alexAndPred`
        notFollowedBySymbol }
    { special (IToparenbar NormalSyntax) }
  "|)"
    / { ifExtension ArrowsBit }
    { special (ITcparenbar NormalSyntax) }
414 415 416

  $unigraphic -- ⦇
    / { ifCurrentChar '⦇' `alexAndPred`
417 418
        ifExtension UnicodeSyntaxBit `alexAndPred`
        ifExtension ArrowsBit }
419 420 421
    { special (IToparenbar UnicodeSyntax) }
  $unigraphic -- ⦈
    / { ifCurrentChar '⦈' `alexAndPred`
422 423
        ifExtension UnicodeSyntaxBit `alexAndPred`
        ifExtension ArrowsBit }
424
    { special (ITcparenbar UnicodeSyntax) }
425 426
}

427
<0> {
428
  \? @varid / { ifExtension IpBit } { skip_one_varid ITdupipvarid }
429 430
}

Adam Gundry's avatar
Adam Gundry committed
431
<0> {
432
  "#" @varid / { ifExtension OverloadedLabelsBit } { skip_one_varid ITlabelvarid }
Adam Gundry's avatar
Adam Gundry committed
433 434
}

435
<0> {
436 437
  "(#" / { ifExtension UnboxedTuplesBit `alexOrPred`
           ifExtension UnboxedSumsBit }
Ian Lynagh's avatar
Ian Lynagh committed
438
         { token IToubxparen }
439 440
  "#)" / { ifExtension UnboxedTuplesBit `alexOrPred`
           ifExtension UnboxedSumsBit }
Ian Lynagh's avatar
Ian Lynagh committed
441 442 443
         { token ITcubxparen }
}

444
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
445 446 447 448 449 450 451 452 453 454
  \(                                    { special IToparen }
  \)                                    { special ITcparen }
  \[                                    { special ITobrack }
  \]                                    { special ITcbrack }
  \,                                    { special ITcomma }
  \;                                    { special ITsemi }
  \`                                    { special ITbackquote }

  \{                                    { open_brace }
  \}                                    { close_brace }
455 456
}

457
<0,option_prags> {
458 459
  @qdo                                      { qdo_token ITdo }
  @qmdo    / { ifExtension RecursiveDoBit } { qdo_token ITmdo }
460 461
  @qvarid                       { idtoken qvarid }
  @qconid                       { idtoken qconid }
Ian Lynagh's avatar
Ian Lynagh committed
462
  @varid                        { varid }
463
  @conid                        { idtoken conid }
464 465
}

466
<0> {
467 468 469 470
  @qvarid "#"+      / { ifExtension MagicHashBit } { idtoken qvarid }
  @qconid "#"+      / { ifExtension MagicHashBit } { idtoken qconid }
  @varid "#"+       / { ifExtension MagicHashBit } { varid }
  @conid "#"+       / { ifExtension MagicHashBit } { idtoken conid }
471 472
}

473 474 475 476 477 478 479 480 481
-- Operators classified into prefix, suffix, tight infix, and loose infix.
-- See Note [Whitespace-sensitive operator parsing]
<0> {
  @varsym / { precededByClosingToken `alexAndPred` followedByOpeningToken } { varsym_tight_infix }
  @varsym / { followedByOpeningToken }  { varsym_prefix }
  @varsym / { precededByClosingToken }  { varsym_suffix }
  @varsym                               { varsym_loose_infix }
}

482 483
-- ToDo: - move `var` and (sym) into lexical syntax?
--       - remove backquote from $special?
484
<0> {
485 486
  @qvarsym                                         { idtoken qvarsym }
  @qconsym                                         { idtoken qconsym }
487
  @consym                                          { consym }
488 489
}

490 491
-- For the normal boxed literals we need to be careful
-- when trying to be close to Haskell98
492 493 494 495 496 497 498 499 500 501

-- Note [Lexing NumericUnderscores extension] (#14473)
--
-- NumericUnderscores extension allows underscores in numeric literals.
-- Multiple underscores are represented with @numspc macro.
-- To be simpler, we have only the definitions with underscores.
-- And then we have a separate function (tok_integral and tok_frac)
-- that validates the literals.
-- If extensions are not enabled, check that there are no underscores.
--
502
<0> {
503
  -- Normal integral literals (:: Num a => a, from Integer)
504 505 506 507 508 509 510 511 512
  @decimal                                                                   { tok_num positive 0 0 decimal }
  0[bB] @numspc @binary                / { ifExtension BinaryLiteralsBit }   { tok_num positive 2 2 binary }
  0[oO] @numspc @octal                                                       { tok_num positive 2 2 octal }
  0[xX] @numspc @hexadecimal                                                 { tok_num positive 2 2 hexadecimal }
  @negative @decimal                   / { ifExtension NegativeLiteralsBit } { tok_num negative 1 1 decimal }
  @negative 0[bB] @numspc @binary      / { ifExtension NegativeLiteralsBit `alexAndPred`
                                           ifExtension BinaryLiteralsBit }   { tok_num negative 3 3 binary }
  @negative 0[oO] @numspc @octal       / { ifExtension NegativeLiteralsBit } { tok_num negative 3 3 octal }
  @negative 0[xX] @numspc @hexadecimal / { ifExtension NegativeLiteralsBit } { tok_num negative 3 3 hexadecimal }
513 514

  -- Normal rational literals (:: Fractional a => a, from Rational)
515 516 517 518 519 520
  @floating_point                                                            { tok_frac 0 tok_float }
  @negative @floating_point            / { ifExtension NegativeLiteralsBit } { tok_frac 0 tok_float }
  0[xX] @numspc @hex_floating_point    / { ifExtension HexFloatLiteralsBit } { tok_frac 0 tok_hex_float }
  @negative 0[xX] @numspc @hex_floating_point
                                       / { ifExtension HexFloatLiteralsBit `alexAndPred`
                                           ifExtension NegativeLiteralsBit } { tok_frac 0 tok_hex_float }
521 522
}

523
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
524
  -- Unboxed ints (:: Int#) and words (:: Word#)
525 526
  -- It's simpler (and faster?) to give separate cases to the negatives,
  -- especially considering octal/hexadecimal prefixes.
527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543
  @decimal                          \# / { ifExtension MagicHashBit }        { tok_primint positive 0 1 decimal }
  0[bB] @numspc @binary             \# / { ifExtension MagicHashBit `alexAndPred`
                                           ifExtension BinaryLiteralsBit }   { tok_primint positive 2 3 binary }
  0[oO] @numspc @octal              \# / { ifExtension MagicHashBit }        { tok_primint positive 2 3 octal }
  0[xX] @numspc @hexadecimal        \# / { ifExtension MagicHashBit }        { tok_primint positive 2 3 hexadecimal }
  @negative @decimal                \# / { ifExtension MagicHashBit }        { tok_primint negative 1 2 decimal }
  @negative 0[bB] @numspc @binary   \# / { ifExtension MagicHashBit `alexAndPred`
                                           ifExtension BinaryLiteralsBit }   { tok_primint negative 3 4 binary }
  @negative 0[oO] @numspc @octal    \# / { ifExtension MagicHashBit }        { tok_primint negative 3 4 octal }
  @negative 0[xX] @numspc @hexadecimal \#
                                       / { ifExtension MagicHashBit }        { tok_primint negative 3 4 hexadecimal }

  @decimal                       \# \# / { ifExtension MagicHashBit }        { tok_primword 0 2 decimal }
  0[bB] @numspc @binary          \# \# / { ifExtension MagicHashBit `alexAndPred`
                                           ifExtension BinaryLiteralsBit }   { tok_primword 2 4 binary }
  0[oO] @numspc @octal           \# \# / { ifExtension MagicHashBit }        { tok_primword 2 4 octal }
  0[xX] @numspc @hexadecimal     \# \# / { ifExtension MagicHashBit }        { tok_primword 2 4 hexadecimal }
Ian Lynagh's avatar
Ian Lynagh committed
544

545 546
  -- Unboxed floats and doubles (:: Float#, :: Double#)
  -- prim_{float,double} work with signed literals
547 548
  @signed @floating_point           \# / { ifExtension MagicHashBit }        { tok_frac 1 tok_primfloat }
  @signed @floating_point        \# \# / { ifExtension MagicHashBit }        { tok_frac 2 tok_primdouble }
549 550 551 552 553 554
}

-- 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.
555
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
556 557
  \'                            { lex_char_tok }
  \"                            { lex_string_tok }
558 559
}

560 561 562 563 564 565 566 567 568 569 570
-- Note [Whitespace-sensitive operator parsing]
-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-- In accord with GHC Proposal #229 https://github.com/ghc-proposals/ghc-proposals/blob/master/proposals/0229-whitespace-bang-patterns.rst
-- we classify operator occurrences into four categories:
--
--     a ! b   -- a loose infix occurrence
--     a!b     -- a tight infix occurrence
--     a !b    -- a prefix occurrence
--     a! b    -- a suffix occurrence
--
-- The rules are a bit more elaborate than simply checking for whitespace, in
Brian Wignall's avatar
Brian Wignall committed
571
-- order to accommodate the following use cases:
572 573 574 575 576 577 578
--
--     f (!a) = ...    -- prefix occurrence
--     g (a !)         -- loose infix occurrence
--     g (! a)         -- loose infix occurrence
--
-- The precise rules are as follows:
--
579 580 581
--  * Identifiers, literals, and opening brackets (, (#, (|, [, [|, [||, [p|,
--    [e|, [t|, {, ⟦, ⦇, are considered "opening tokens". The function
--    followedByOpeningToken tests whether the next token is an opening token.
582
--
583
--  * Identifiers, literals, and closing brackets ), #), |), ], |], }, ⟧, ⦈,
584 585
--    are considered "closing tokens". The function precededByClosingToken tests
--    whether the previous token is a closing token.
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
586
--
587 588
--  * Whitespace, comments, separators, and other tokens, are considered
--    neither opening nor closing.
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
589
--
590 591
--  * Any unqualified operator occurrence is classified as prefix, suffix, or
--    tight/loose infix, based on preceding and following tokens:
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
592
--
593 594 595 596 597 598 599
--       precededByClosingToken | followedByOpeningToken | Occurrence
--      ------------------------+------------------------+------------
--       False                  | True                   | prefix
--       True                   | False                  | suffix
--       True                   | True                   | tight infix
--       False                  | False                  | loose infix
--      ------------------------+------------------------+------------
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
600
--
601 602
-- A loose infix occurrence is always considered an operator. Other types of
-- occurrences may be assigned a special per-operator meaning override:
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
603
--
604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636
--   Operator |  Occurrence   | Token returned
--  ----------+---------------+------------------------------------------
--    !       |  prefix       | ITbang
--            |               |   strictness annotation or bang pattern,
--            |               |   e.g.  f !x = rhs, data T = MkT !a
--            |  not prefix   | ITvarsym "!"
--            |               |   ordinary operator or type operator,
--            |               |   e.g.  xs ! 3, (! x), Int ! Bool
--  ----------+---------------+------------------------------------------
--    ~       |  prefix       | ITtilde
--            |               |   laziness annotation or lazy pattern,
--            |               |   e.g.  f ~x = rhs, data T = MkT ~a
--            |  not prefix   | ITvarsym "~"
--            |               |   ordinary operator or type operator,
--            |               |   e.g.  xs ~ 3, (~ x), Int ~ Bool
--  ----------+---------------+------------------------------------------
--    $  $$   |  prefix       | ITdollar, ITdollardollar
--            |               |   untyped or typed Template Haskell splice,
--            |               |   e.g.  $(f x), $$(f x), $$"str"
--            |  not prefix   | ITvarsym "$", ITvarsym "$$"
--            |               |   ordinary operator or type operator,
--            |               |   e.g.  f $ g x, a $$ b
--  ----------+---------------+------------------------------------------
--    @       |  prefix       | ITtypeApp
--            |               |   type application, e.g.  fmap @Maybe
--            |  tight infix  | ITat
--            |               |   as-pattern, e.g.  f p@(a,b) = rhs
--            |  suffix       | parse error
--            |               |   e.g. f p@ x = rhs
--            |  loose infix  | ITvarsym "@"
--            |               |   ordinary operator or type operator,
--            |               |   e.g.  f @ g, (f @)
--  ----------+---------------+------------------------------------------
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
637
--
638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667
-- Also, some of these overrides are guarded behind language extensions.
-- According to the specification, we must determine the occurrence based on
-- surrounding *tokens* (see the proposal for the exact rules). However, in
-- the implementation we cheat a little and do the classification based on
-- characters, for reasons of both simplicity and efficiency (see
-- 'followedByOpeningToken' and 'precededByClosingToken')
--
-- When an operator is subject to a meaning override, it is mapped to special
-- token: ITbang, ITtilde, ITat, ITdollar, ITdollardollar. Otherwise, it is
-- returned as ITvarsym.
--
-- For example, this is how we process the (!):
--
--    precededByClosingToken | followedByOpeningToken | Token
--   ------------------------+------------------------+-------------
--    False                  | True                   | ITbang
--    True                   | False                  | ITvarsym "!"
--    True                   | True                   | ITvarsym "!"
--    False                  | False                  | ITvarsym "!"
--   ------------------------+------------------------+-------------
--
-- And this is how we process the (@):
--
--    precededByClosingToken | followedByOpeningToken | Token
--   ------------------------+------------------------+-------------
--    False                  | True                   | ITtypeApp
--    True                   | False                  | parse error
--    True                   | True                   | ITat
--    False                  | False                  | ITvarsym "@"
--   ------------------------+------------------------+-------------
668 669 670 671

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

672
{
673

674 675 676
-- -----------------------------------------------------------------------------
-- The token type

677
data Token
Ian Lynagh's avatar
Ian Lynagh committed
678
  = ITas                        -- Haskell keywords
679 680 681 682 683
  | ITcase
  | ITclass
  | ITdata
  | ITdefault
  | ITderiving
684
  | ITdo (Maybe FastString)
685 686
  | ITelse
  | IThiding
687
  | ITforeign
688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703
  | ITif
  | ITimport
  | ITin
  | ITinfix
  | ITinfixl
  | ITinfixr
  | ITinstance
  | ITlet
  | ITmodule
  | ITnewtype
  | ITof
  | ITqualified
  | ITthen
  | ITtype
  | ITwhere

704
  | ITforall            IsUnicodeSyntax -- GHC extension keywords
705 706 707 708
  | ITexport
  | ITlabel
  | ITdynamic
  | ITsafe
709
  | ITinterruptible
710 711 712
  | ITunsafe
  | ITstdcallconv
  | ITccallconv
713
  | ITcapiconv
714
  | ITprimcallconv
thoughtpolice's avatar
thoughtpolice committed
715
  | ITjavascriptcallconv
716
  | ITmdo (Maybe FastString)
717
  | ITfamily
718
  | ITrole
719 720 721
  | ITgroup
  | ITby
  | ITusing
Gergő Érdi's avatar
Gergő Érdi committed
722
  | ITpattern
Facundo Domínguez's avatar
Facundo Domínguez committed
723
  | ITstatic
Ryan Scott's avatar
Ryan Scott committed
724 725
  | ITstock
  | ITanyclass
Ryan Scott's avatar
Ryan Scott committed
726
  | ITvia
727

Edward Z. Yang's avatar
Edward Z. Yang committed
728 729 730 731 732 733
  -- Backpack tokens
  | ITunit
  | ITsignature
  | ITdependency
  | ITrequires

734
  -- Pragmas, see  note [Pragma source text] in "GHC.Types.Basic"
Alan Zimmerman's avatar
Alan Zimmerman committed
735 736 737 738 739 740 741
  | 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
742 743
  | ITline_prag         SourceText  -- not usually produced, see 'UsePosPragsBit'
  | ITcolumn_prag       SourceText  -- not usually produced, see 'UsePosPragsBit'
Alan Zimmerman's avatar
Alan Zimmerman committed
744 745 746 747 748 749
  | ITscc_prag          SourceText
  | ITgenerated_prag    SourceText
  | ITcore_prag         SourceText         -- hdaume: core annotations
  | ITunpack_prag       SourceText
  | ITnounpack_prag     SourceText
  | ITann_prag          SourceText
750
  | ITcomplete_prag     SourceText
751
  | ITclose_prag
752 753 754
  | IToptions_prag String
  | ITinclude_prag String
  | ITlanguage_prag
Alan Zimmerman's avatar
Alan Zimmerman committed
755 756 757 758 759 760
  | 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
761
  | ITcomment_line_prag         -- See Note [Nested comment line pragmas]
762

Ian Lynagh's avatar
Ian Lynagh committed
763
  | ITdotdot                    -- reserved symbols
764
  | ITcolon
765
  | ITdcolon            IsUnicodeSyntax
766 767
  | ITequal
  | ITlam
768
  | ITlcase
769
  | ITvbar
770 771
  | ITlarrow            IsUnicodeSyntax
  | ITrarrow            IsUnicodeSyntax
Krzysztof Gogolewski's avatar
Krzysztof Gogolewski committed
772
  | ITlolly             IsUnicodeSyntax
773
  | ITdarrow            IsUnicodeSyntax
774
  | ITminus
775 776 777 778
  | ITbang     -- Prefix (!) only, e.g. f !x = rhs
  | ITtilde    -- Prefix (~) only, e.g. f ~x = rhs
  | ITat       -- Tight infix (@) only, e.g. f x@pat = rhs
  | ITtypeApp  -- Prefix (@) only, e.g. f @t
779
  | ITstar              IsUnicodeSyntax
780 781
  | ITdot

Ian Lynagh's avatar
Ian Lynagh committed
782
  | ITbiglam                    -- GHC-extension symbols
783

Ian Lynagh's avatar
Ian Lynagh committed
784
  | ITocurly                    -- special symbols
785 786 787 788
  | ITccurly
  | ITvocurly
  | ITvccurly
  | ITobrack
Ian Lynagh's avatar
Ian Lynagh committed
789 790
  | ITopabrack                  -- [:, for parallel arrays with -XParallelArrays
  | ITcpabrack                  -- :], for parallel arrays with -XParallelArrays
791 792 793 794 795 796 797 798 799
  | ITcbrack
  | IToparen
  | ITcparen
  | IToubxparen
  | ITcubxparen
  | ITsemi
  | ITcomma
  | ITunderscore
  | ITbackquote
dreixel's avatar
dreixel committed
800
  | ITsimpleQuote               --  '
801

Ian Lynagh's avatar
Ian Lynagh committed
802
  | ITvarid   FastString        -- identifiers
803 804 805 806 807 808 809 810
  | 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
811
  | ITdupipvarid   FastString   -- GHC extension: implicit param: ?x
Adam Gundry's avatar
Adam Gundry committed
812
  | ITlabelvarid   FastString   -- Overloaded label: #x
813

814 815 816
  | ITchar     SourceText Char       -- Note [Literal source text] in "GHC.Types.Basic"
  | ITstring   SourceText FastString -- Note [Literal source text] in "GHC.Types.Basic"
  | ITinteger  IntegralLit           -- Note [Literal source text] in "GHC.Types.Basic"
Alan Zimmerman's avatar
Alan Zimmerman committed
817
  | ITrational FractionalLit
818

819 820 821 822
  | ITprimchar   SourceText Char     -- Note [Literal source text] in "GHC.Types.Basic"
  | ITprimstring SourceText ByteString -- Note [Literal source text] in "GHC.Types.Basic"
  | ITprimint    SourceText Integer  -- Note [Literal source text] in "GHC.Types.Basic"
  | ITprimword   SourceText Integer  -- Note [Literal source text] in "GHC.Types.Basic"
823 824
  | ITprimfloat  FractionalLit
  | ITprimdouble FractionalLit
825

826
  -- Template Haskell extension tokens
827 828 829 830 831 832 833
  | ITopenExpQuote HasE IsUnicodeSyntax --  [| or [e|
  | ITopenPatQuote                      --  [p|
  | ITopenDecQuote                      --  [d|
  | ITopenTypQuote                      --  [t|
  | ITcloseQuote IsUnicodeSyntax        --  |]
  | ITopenTExpQuote HasE                --  [|| or [e||
  | ITcloseTExpQuote                    --  ||]
834 835
  | ITdollar                            --  prefix $
  | ITdollardollar                      --  prefix $$
836
  | ITtyQuote                           --  ''
837
  | ITquasiQuote (FastString,FastString,PsSpan)
838 839 840
    -- ITquasiQuote(quoter, quote, loc)
    -- represents a quasi-quote of the form
    -- [quoter| quote |]
841
  | ITqQuasiQuote (FastString,FastString,FastString,PsSpan)
842 843 844
    -- ITqQuasiQuote(Qual, quoter, quote, loc)
    -- represents a qualified quasi-quote of the form
    -- [Qual.quoter| quote |]
845 846 847 848

  -- Arrow notation extension
  | ITproc
  | ITrec
849 850 851 852 853 854 855 856 857
  | IToparenbar  IsUnicodeSyntax -- ^ @(|@
  | ITcparenbar  IsUnicodeSyntax -- ^ @|)@
  | ITlarrowtail IsUnicodeSyntax -- ^ @-<@
  | ITrarrowtail IsUnicodeSyntax -- ^ @>-@
  | ITLarrowtail IsUnicodeSyntax -- ^ @-<<@
  | ITRarrowtail IsUnicodeSyntax -- ^ @>>-@

  | ITunknown String             -- ^ Used when the lexer can't make sense of it
  | ITeof                        -- ^ end of file token
858 859

  -- Documentation annotations
860 861 862 863 864 865 866
  | 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)
  | ITlineComment     String     -- ^ comment starting by "--"
  | ITblockComment    String     -- ^ comment in {- -}
867

868
  deriving Show
869

Alan Zimmerman's avatar
Alan Zimmerman committed
870 871 872
instance Outputable Token where
  ppr x = text (show x)

Alan Zimmerman's avatar
Alan Zimmerman committed
873

874

875 876 877 878 879 880 881
-- 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)
--
882
reservedWordsFM :: UniqFM (Token, ExtsBitmap)
883
reservedWordsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
884 885 886 887 888 889 890 891
    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 ),
892
         ( "do",             ITdo Nothing,    0 ),
Ian Lynagh's avatar
Ian Lynagh committed
893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910
         ( "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 ),

911
         ( "forall",         ITforall NormalSyntax, 0),
912
         ( "mdo",            ITmdo Nothing,   xbit RecursiveDoBit),
913 914 915
             -- See Note [Lexing type pseudo-keywords]
         ( "family",         ITfamily,        0 ),
         ( "role",           ITrole,          0 ),
916
         ( "pattern",        ITpattern,       xbit PatternSynonymsBit),
917
         ( "static",         ITstatic,        xbit StaticPointersBit ),
Ryan Scott's avatar
Ryan Scott committed
918 919
         ( "stock",          ITstock,         0 ),
         ( "anyclass",       ITanyclass,      0 ),
Ryan Scott's avatar
Ryan Scott committed
920
         ( "via",            ITvia,           0 ),
921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938
         ( "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),

Edward Z. Yang's avatar
Edward Z. Yang committed
939 940 941 942
         ( "unit",           ITunit,          0 ),
         ( "dependency",     ITdependency,       0 ),
         ( "signature",      ITsignature,     0 ),

943 944 945
         ( "rec",            ITrec,           xbit ArrowsBit .|.
                                              xbit RecursiveDoBit),
         ( "proc",           ITproc,          xbit ArrowsBit)
946 947
     ]

948 949 950 951 952 953 954 955
{-----------------------------------
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
Sylvain Henry's avatar
Sylvain Henry committed
956
not allowed. Furthermore, checks further downstream (GHC.Tc.TyCl) ensure that
957 958 959 960 961 962 963
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.
-------------------------------------}
964

965
reservedSymsFM :: UniqFM (Token, IsUnicodeSyntax, ExtsBitmap)
966
reservedSymsFM = listToUFM $
967 968
    map (\ (x,w,y,z) -> (mkFastString x,(w,y,z)))
      [ ("..",  ITdotdot,                   NormalSyntax,  0 )
Ian Lynagh's avatar
Ian Lynagh committed
969
        -- (:) is a reserved op, meaning only list cons
970 971 972 973 974 975 976 977 978 979 980
       ,(":",   ITcolon,                    NormalSyntax,  0 )
       ,("::",  ITdcolon NormalSyntax,      NormalSyntax,  0 )
       ,("=",   ITequal,                    NormalSyntax,  0 )
       ,("\\",  ITlam,                      NormalSyntax,  0 )
       ,("|",   ITvbar,                     NormalSyntax,  0 )
       ,("<-",  ITlarrow NormalSyntax,      NormalSyntax,  0 )
       ,("->",  ITrarrow NormalSyntax,      NormalSyntax,  0 )
       ,("=>",  ITdarrow NormalSyntax,      NormalSyntax,  0 )
       ,("-",   ITminus,                    NormalSyntax,  0 )

       ,("*",   ITstar NormalSyntax,        NormalSyntax,  xbit StarIsTypeBit)
981

Ian Lynagh's avatar
Ian Lynagh committed
982
        -- For 'forall a . t'
983 984 985 986 987 988 989 990 991 992 993 994 995
       ,(".",   ITdot,                      NormalSyntax,  0 )

       ,("-<",  ITlarrowtail NormalSyntax,  NormalSyntax,  xbit ArrowsBit)
       ,(">-",  ITrarrowtail NormalSyntax,  NormalSyntax,  xbit ArrowsBit)
       ,("-<<", ITLarrowtail NormalSyntax,  NormalSyntax,  xbit ArrowsBit)
       ,(">>-", ITRarrowtail NormalSyntax,  NormalSyntax,  xbit ArrowsBit)

       ,("∷",   ITdcolon UnicodeSyntax,     UnicodeSyntax, 0 )
       ,("⇒",   ITdarrow UnicodeSyntax,     UnicodeSyntax, 0 )
       ,("∀",   ITforall UnicodeSyntax,     UnicodeSyntax, 0 )
       ,("→",   ITrarrow UnicodeSyntax,     UnicodeSyntax, 0 )
       ,("←",   ITlarrow UnicodeSyntax,     UnicodeSyntax, 0 )

Krzysztof Gogolewski's avatar
Krzysztof Gogolewski committed
996 997 998
       ,("#->", ITlolly NormalSyntax, NormalSyntax, 0)
       ,("⊸",   ITlolly UnicodeSyntax, UnicodeSyntax, 0)

999 1000 1001 1002 1003 1004
       ,("⤙",   ITlarrowtail UnicodeSyntax, UnicodeSyntax, xbit ArrowsBit)
       ,("⤚",   ITrarrowtail UnicodeSyntax, UnicodeSyntax, xbit ArrowsBit)
       ,("⤛",   ITLarrowtail UnicodeSyntax, UnicodeSyntax, xbit ArrowsBit)
       ,("⤜",   ITRarrowtail UnicodeSyntax, UnicodeSyntax, xbit ArrowsBit)

       ,("★",   ITstar UnicodeSyntax,       UnicodeSyntax, xbit StarIsTypeBit)
1005

Simon Marlow's avatar
Simon Marlow committed
1006 1007 1008
        -- 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).
1009 1010 1011 1012 1013
       ]

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

1014
type Action = PsSpan -> StringBuffer -> Int -> P (PsLocated Token)
1015

1016
special :: Token -> Action
twanvl's avatar
twanvl committed
1017
special tok span _buf _len = return (L span tok)
1018

1019
token, layout_token :: Token -> Action
twanvl's avatar
twanvl committed
1020 1021
token t span _buf _len = return (L span t)
layout_token t span _buf _len = pushLexState layout >> return (L span t)
1022

1023 1024
idtoken :: (StringBuffer -> Int -> Token) -> Action
idtoken f span buf len = return (L span $! (f buf len))