Lexer.x 129 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(..),
58
   getRealSrcLoc, getPState, withThisPackage,
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
@floating_point = @numspc @decimal \. @decimal @exponent? | @numspc @decimal @exponent
@hex_floating_point = @numspc @hexadecimal \. @hexadecimal @bin_exponent? | @numspc @hexadecimal @bin_exponent
189

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.

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

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

236
"-- " $docsym .* / { alexNotPred (ifExtension HaddockBit) } { 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
-- after an 'if', a vertical bar starts a layout context for MultiWayIf
<layout_if> {
286
  \| / { notFollowedBySymbol }          { new_layout_context True dontGenerateSemic ITvbar }
287 288 289
  ()                                    { pop }
}

290
-- do is treated in a subtly different way, see new_layout_context
291 292
<layout>    ()                          { new_layout_context True  generateSemic ITvocurly }
<layout_do> ()                          { new_layout_context False generateSemic 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
306 307 308 309 310
<line_prag1> {
  @decimal $white_no_nl+ \" [$graphic \ ]* \"  { setLineAndFile line_prag1a }
  ()                                           { failLinePrag1 }
}
<line_prag1a> .*                               { popLinePrag1 }
311 312 313

-- Haskell-style line pragmas, of the form
--    {-# LINE <line> "<file>" #-}
314 315 316 317
<line_prag2> {
  @decimal $white_no_nl+ \" [$graphic \ ]* \"  { setLineAndFile line_prag2a }
}
<line_prag2a> "#-}"|"-}"                       { pop }
318 319
   -- NOTE: accept -} at the end of a LINE pragma, for compatibility
   -- with older versions of GHC which generated these.
320

Rufflewind's avatar
Rufflewind committed
321 322 323 324
-- Haskell-style column pragmas, of the form
--    {-# COLUMN <column> #-}
<column_prag> @decimal $whitechar* "#-}" { setColumn }

325
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
326
  "{-#" $whitechar* $pragmachar+
327 328 329 330 331
        $whitechar+ $pragmachar+ / { known_pragma twoWordPrags }
                                 { dispatch_pragmas twoWordPrags }

  "{-#" $whitechar* $pragmachar+ / { known_pragma oneWordPrags }
                                 { dispatch_pragmas oneWordPrags }
332

333
  -- We ignore all these pragmas, but don't generate a warning for them
334 335
  "{-#" $whitechar* $pragmachar+ / { known_pragma ignoredPrags }
                                 { dispatch_pragmas ignoredPrags }
336 337

  -- ToDo: should only be valid inside a pragma:
Ian Lynagh's avatar
Ian Lynagh committed
338
  "#-}"                          { endPrag }
339 340
}

341
<option_prags> {
342 343
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
                                   { dispatch_pragmas fileHeaderPrags }
David Waern's avatar
David Waern committed
344 345
}

346 347
<0> {
  -- In the "0" mode we ignore these pragmas
348
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
349 350 351
                     { nested_comment lexToken }
}

352
<0,option_prags> {
353 354
  "{-#"  { warnThen Opt_WarnUnrecognisedPragmas (text "Unrecognised pragma")
                    (nested_comment lexToken) }
355 356
}

357 358
-- '0' state: ordinary lexemes

359 360
-- Haddock comments

361
<0,option_prags> {
362 363
  "-- " $docsym      / { ifExtension HaddockBit } { multiline_doc_comment }
  "{-" \ ? $docsym   / { ifExtension HaddockBit } { nested_doc_comment }
364 365
}

366 367
-- "special" symbols

368
<0> {
369 370 371 372 373 374 375 376 377 378 379
  "[|"        / { ifExtension ThQuotesBit } { token (ITopenExpQuote NoE NormalSyntax) }
  "[||"       / { ifExtension ThQuotesBit } { token (ITopenTExpQuote NoE) }
  "[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 }
  "|]"        / { ifExtension ThQuotesBit } { token (ITcloseQuote NormalSyntax) }
  "||]"       / { ifExtension ThQuotesBit } { token ITcloseTExpQuote }

  "[" @varid "|"  / { ifExtension QqBit }   { lex_quasiquote_tok }
380 381

  -- qualified quasi-quote (#5555)
382
  "[" @qvarid "|"  / { ifExtension QqBit }  { lex_qquasiquote_tok }
383 384 385

  $unigraphic -- ⟦
    / { ifCurrentChar '⟦' `alexAndPred`
386 387
        ifExtension UnicodeSyntaxBit `alexAndPred`
        ifExtension ThQuotesBit }
388 389 390
    { token (ITopenExpQuote NoE UnicodeSyntax) }
  $unigraphic -- ⟧
    / { ifCurrentChar '⟧' `alexAndPred`
391 392
        ifExtension UnicodeSyntaxBit `alexAndPred`
        ifExtension ThQuotesBit }
393
    { token (ITcloseQuote UnicodeSyntax) }
394 395
}

396
<0> {
397 398 399 400 401 402 403
  "(|"
    / { ifExtension ArrowsBit `alexAndPred`
        notFollowedBySymbol }
    { special (IToparenbar NormalSyntax) }
  "|)"
    / { ifExtension ArrowsBit }
    { special (ITcparenbar NormalSyntax) }
404 405 406

  $unigraphic -- ⦇
    / { ifCurrentChar '⦇' `alexAndPred`
407 408
        ifExtension UnicodeSyntaxBit `alexAndPred`
        ifExtension ArrowsBit }
409 410 411
    { special (IToparenbar UnicodeSyntax) }
  $unigraphic -- ⦈
    / { ifCurrentChar '⦈' `alexAndPred`
412 413
        ifExtension UnicodeSyntaxBit `alexAndPred`
        ifExtension ArrowsBit }
414
    { special (ITcparenbar UnicodeSyntax) }
415 416
}

417
<0> {
418
  \? @varid / { ifExtension IpBit } { skip_one_varid ITdupipvarid }
419 420
}

Adam Gundry's avatar
Adam Gundry committed
421
<0> {
422
  "#" @varid / { ifExtension OverloadedLabelsBit } { skip_one_varid ITlabelvarid }
Adam Gundry's avatar
Adam Gundry committed
423 424
}

425
<0> {
426 427
  "(#" / { ifExtension UnboxedTuplesBit `alexOrPred`
           ifExtension UnboxedSumsBit }
Ian Lynagh's avatar
Ian Lynagh committed
428
         { token IToubxparen }
429 430
  "#)" / { ifExtension UnboxedTuplesBit `alexOrPred`
           ifExtension UnboxedSumsBit }
Ian Lynagh's avatar
Ian Lynagh committed
431 432 433
         { token ITcubxparen }
}

434
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
435 436 437 438 439 440 441 442 443 444
  \(                                    { special IToparen }
  \)                                    { special ITcparen }
  \[                                    { special ITobrack }
  \]                                    { special ITcbrack }
  \,                                    { special ITcomma }
  \;                                    { special ITsemi }
  \`                                    { special ITbackquote }

  \{                                    { open_brace }
  \}                                    { close_brace }
445 446
}

447
<0,option_prags> {
448 449
  @qvarid                       { idtoken qvarid }
  @qconid                       { idtoken qconid }
Ian Lynagh's avatar
Ian Lynagh committed
450
  @varid                        { varid }
451
  @conid                        { idtoken conid }
452 453
}

454
<0> {
455 456 457 458
  @qvarid "#"+      / { ifExtension MagicHashBit } { idtoken qvarid }
  @qconid "#"+      / { ifExtension MagicHashBit } { idtoken qconid }
  @varid "#"+       / { ifExtension MagicHashBit } { varid }
  @conid "#"+       / { ifExtension MagicHashBit } { idtoken conid }
459 460
}

461 462 463 464 465 466 467 468 469
-- 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 }
}

470 471
-- ToDo: - move `var` and (sym) into lexical syntax?
--       - remove backquote from $special?
472
<0> {
473 474
  @qvarsym                                         { idtoken qvarsym }
  @qconsym                                         { idtoken qconsym }
475
  @consym                                          { consym }
476 477
}

478 479
-- For the normal boxed literals we need to be careful
-- when trying to be close to Haskell98
480 481 482 483 484 485 486 487 488 489

-- 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.
--
490
<0> {
491
  -- Normal integral literals (:: Num a => a, from Integer)
492 493 494 495 496 497 498 499 500
  @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 }
501 502

  -- Normal rational literals (:: Fractional a => a, from Rational)
503 504 505 506 507 508
  @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 }
509 510
}

511
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
512
  -- Unboxed ints (:: Int#) and words (:: Word#)
513 514
  -- It's simpler (and faster?) to give separate cases to the negatives,
  -- especially considering octal/hexadecimal prefixes.
515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531
  @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
532

533 534
  -- Unboxed floats and doubles (:: Float#, :: Double#)
  -- prim_{float,double} work with signed literals
535 536
  @signed @floating_point           \# / { ifExtension MagicHashBit }        { tok_frac 1 tok_primfloat }
  @signed @floating_point        \# \# / { ifExtension MagicHashBit }        { tok_frac 2 tok_primdouble }
537 538 539 540 541 542
}

-- 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.
543
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
544 545
  \'                            { lex_char_tok }
  \"                            { lex_string_tok }
546 547
}

548 549 550 551 552 553 554 555 556 557 558
-- 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
559
-- order to accommodate the following use cases:
560 561 562 563 564 565 566
--
--     f (!a) = ...    -- prefix occurrence
--     g (a !)         -- loose infix occurrence
--     g (! a)         -- loose infix occurrence
--
-- The precise rules are as follows:
--
567 568 569
--  * Identifiers, literals, and opening brackets (, (#, (|, [, [|, [||, [p|,
--    [e|, [t|, {, ⟦, ⦇, are considered "opening tokens". The function
--    followedByOpeningToken tests whether the next token is an opening token.
570
--
571
--  * Identifiers, literals, and closing brackets ), #), |), ], |], }, ⟧, ⦈,
572 573
--    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
574
--
575 576
--  * Whitespace, comments, separators, and other tokens, are considered
--    neither opening nor closing.
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
577
--
578 579
--  * 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
580
--
581 582 583 584 585 586 587
--       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
588
--
589 590
-- 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
591
--
592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624
--   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
625
--
626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655
-- 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 "@"
--   ------------------------+------------------------+-------------
656 657 658 659

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

660
{
661

662 663 664
-- -----------------------------------------------------------------------------
-- The token type

665
data Token
Ian Lynagh's avatar
Ian Lynagh committed
666
  = ITas                        -- Haskell keywords
667 668 669 670 671 672 673 674
  | ITcase
  | ITclass
  | ITdata
  | ITdefault
  | ITderiving
  | ITdo
  | ITelse
  | IThiding
675
  | ITforeign
676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691
  | ITif
  | ITimport
  | ITin
  | ITinfix
  | ITinfixl
  | ITinfixr
  | ITinstance
  | ITlet
  | ITmodule
  | ITnewtype
  | ITof
  | ITqualified
  | ITthen
  | ITtype
  | ITwhere

692
  | ITforall            IsUnicodeSyntax -- GHC extension keywords
693 694 695 696
  | ITexport
  | ITlabel
  | ITdynamic
  | ITsafe
697
  | ITinterruptible
698 699 700
  | ITunsafe
  | ITstdcallconv
  | ITccallconv
701
  | ITcapiconv
702
  | ITprimcallconv
thoughtpolice's avatar
thoughtpolice committed
703
  | ITjavascriptcallconv
704
  | ITmdo
705
  | ITfamily
706
  | ITrole
707 708 709
  | ITgroup
  | ITby
  | ITusing
Gergő Érdi's avatar
Gergő Érdi committed
710
  | ITpattern
Facundo Domínguez's avatar
Facundo Domínguez committed
711
  | ITstatic
Ryan Scott's avatar
Ryan Scott committed
712 713
  | ITstock
  | ITanyclass
Ryan Scott's avatar
Ryan Scott committed
714
  | ITvia
715

Edward Z. Yang's avatar
Edward Z. Yang committed
716 717 718 719 720 721
  -- Backpack tokens
  | ITunit
  | ITsignature
  | ITdependency
  | ITrequires

Alan Zimmerman's avatar
Alan Zimmerman committed
722 723 724 725 726 727 728 729
  -- 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
730 731
  | ITline_prag         SourceText  -- not usually produced, see 'UsePosPragsBit'
  | ITcolumn_prag       SourceText  -- not usually produced, see 'UsePosPragsBit'
Alan Zimmerman's avatar
Alan Zimmerman committed
732 733 734 735 736 737
  | ITscc_prag          SourceText
  | ITgenerated_prag    SourceText
  | ITcore_prag         SourceText         -- hdaume: core annotations
  | ITunpack_prag       SourceText
  | ITnounpack_prag     SourceText
  | ITann_prag          SourceText
738
  | ITcomplete_prag     SourceText
739
  | ITclose_prag
740 741 742
  | IToptions_prag String
  | ITinclude_prag String
  | ITlanguage_prag
Alan Zimmerman's avatar
Alan Zimmerman committed
743 744 745 746 747 748
  | 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
749
  | ITcomment_line_prag         -- See Note [Nested comment line pragmas]
750

Ian Lynagh's avatar
Ian Lynagh committed
751
  | ITdotdot                    -- reserved symbols
752
  | ITcolon
753
  | ITdcolon            IsUnicodeSyntax
754 755
  | ITequal
  | ITlam
756
  | ITlcase
757
  | ITvbar
758 759 760
  | ITlarrow            IsUnicodeSyntax
  | ITrarrow            IsUnicodeSyntax
  | ITdarrow            IsUnicodeSyntax
761
  | ITminus
762 763 764 765
  | 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
766
  | ITstar              IsUnicodeSyntax
767 768
  | ITdot

Ian Lynagh's avatar
Ian Lynagh committed
769
  | ITbiglam                    -- GHC-extension symbols
770

Ian Lynagh's avatar
Ian Lynagh committed
771
  | ITocurly                    -- special symbols
772 773 774 775
  | ITccurly
  | ITvocurly
  | ITvccurly
  | ITobrack
Ian Lynagh's avatar
Ian Lynagh committed
776 777
  | ITopabrack                  -- [:, for parallel arrays with -XParallelArrays
  | ITcpabrack                  -- :], for parallel arrays with -XParallelArrays
778 779 780 781 782 783 784 785 786
  | ITcbrack
  | IToparen
  | ITcparen
  | IToubxparen
  | ITcubxparen
  | ITsemi
  | ITcomma
  | ITunderscore
  | ITbackquote
dreixel's avatar
dreixel committed
787
  | ITsimpleQuote               --  '
788

Ian Lynagh's avatar
Ian Lynagh committed
789
  | ITvarid   FastString        -- identifiers
790 791 792 793 794 795 796 797
  | 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
798
  | ITdupipvarid   FastString   -- GHC extension: implicit param: ?x
Adam Gundry's avatar
Adam Gundry committed
799
  | ITlabelvarid   FastString   -- Overloaded label: #x
800

801 802
  | ITchar     SourceText Char       -- Note [Literal source text] in BasicTypes
  | ITstring   SourceText FastString -- Note [Literal source text] in BasicTypes
803
  | ITinteger  IntegralLit           -- Note [Literal source text] in BasicTypes
Alan Zimmerman's avatar
Alan Zimmerman committed
804
  | ITrational FractionalLit
805

806 807 808 809
  | 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
810 811
  | ITprimfloat  FractionalLit
  | ITprimdouble FractionalLit
812

813
  -- Template Haskell extension tokens
814 815 816 817 818 819 820
  | ITopenExpQuote HasE IsUnicodeSyntax --  [| or [e|
  | ITopenPatQuote                      --  [p|
  | ITopenDecQuote                      --  [d|
  | ITopenTypQuote                      --  [t|
  | ITcloseQuote IsUnicodeSyntax        --  |]
  | ITopenTExpQuote HasE                --  [|| or [e||
  | ITcloseTExpQuote                    --  ||]
821 822
  | ITdollar                            --  prefix $
  | ITdollardollar                      --  prefix $$
823
  | ITtyQuote                           --  ''
824
  | ITquasiQuote (FastString,FastString,PsSpan)
825 826 827
    -- ITquasiQuote(quoter, quote, loc)
    -- represents a quasi-quote of the form
    -- [quoter| quote |]
828
  | ITqQuasiQuote (FastString,FastString,FastString,PsSpan)
829 830 831
    -- ITqQuasiQuote(Qual, quoter, quote, loc)
    -- represents a qualified quasi-quote of the form
    -- [Qual.quoter| quote |]
832 833 834 835

  -- Arrow notation extension
  | ITproc
  | ITrec
836 837 838 839 840 841 842 843 844
  | 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
845 846

  -- Documentation annotations
847 848 849 850 851 852 853
  | 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 {- -}
854

855
  deriving Show
856

Alan Zimmerman's avatar
Alan Zimmerman committed
857 858 859
instance Outputable Token where
  ppr x = text (show x)

Alan Zimmerman's avatar
Alan Zimmerman committed
860

861 862 863 864 865 866 867
-- 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)
--
868
reservedWordsFM :: UniqFM (Token, ExtsBitmap)
869
reservedWordsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896
    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 ),

897
         ( "forall",         ITforall NormalSyntax, 0),
898
         ( "mdo",            ITmdo,           xbit RecursiveDoBit),
899 900 901
             -- See Note [Lexing type pseudo-keywords]
         ( "family",         ITfamily,        0 ),
         ( "role",           ITrole,          0 ),
902
         ( "pattern",        ITpattern,       xbit PatternSynonymsBit),
903
         ( "static",         ITstatic,        xbit StaticPointersBit ),
Ryan Scott's avatar
Ryan Scott committed
904 905
         ( "stock",          ITstock,         0 ),
         ( "anyclass",       ITanyclass,      0 ),
Ryan Scott's avatar
Ryan Scott committed
906
         ( "via",            ITvia,           0 ),
907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924
         ( "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
925 926 927 928
         ( "unit",           ITunit,          0 ),
         ( "dependency",     ITdependency,       0 ),
         ( "signature",      ITsignature,     0 ),

929 930 931
         ( "rec",            ITrec,           xbit ArrowsBit .|.
                                              xbit RecursiveDoBit),
         ( "proc",           ITproc,          xbit ArrowsBit)
932 933
     ]

934 935 936 937 938 939 940 941
{-----------------------------------
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
942
not allowed. Furthermore, checks further downstream (GHC.Tc.TyCl) ensure that
943 944 945 946 947 948 949
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.
-------------------------------------}
950

951
reservedSymsFM :: UniqFM (Token, IsUnicodeSyntax, ExtsBitmap)
952
reservedSymsFM = listToUFM $
953 954
    map (\ (x,w,y,z) -> (mkFastString x,(w,y,z)))
      [ ("..",  ITdotdot,                   NormalSyntax,  0 )
Ian Lynagh's avatar
Ian Lynagh committed
955
        -- (:) is a reserved op, meaning only list cons
956 957 958 959 960 961 962 963 964 965 966
       ,(":",   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)
967

Ian Lynagh's avatar
Ian Lynagh committed
968
        -- For 'forall a . t'
969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987
       ,(".",   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 )

       ,("⤙",   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)
988

Simon Marlow's avatar
Simon Marlow committed
989 990 991
        -- 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).
992 993 994 995 996
       ]

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

997
type Action = PsSpan -> StringBuffer -> Int -> P (PsLocated Token)
998

999
special :: Token -> Action
twanvl's avatar
twanvl committed
1000
special tok span _buf _len = return (L span tok)
1001

1002
token, layout_token :: Token -> Action
twanvl's avatar
twanvl committed
1003 1004
token t span _buf _len = return (L span t)
layout_token t span _buf _len = pushLexState layout >> return (L span t)
1005

1006 1007
idtoken :: (StringBuffer -> Int -> Token) -> Action
idtoken f span buf len = return (L span $! (f buf len))
1008

1009
skip_one_varid :: (FastString -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
1010
skip_one_varid f span buf len
1011
  = return (L span $! f (lexemeToFastString (stepOn buf) (len-1)))
1012

1013 1014 1015 1016
skip_two_varid :: (FastString -> Token) -> Action
skip_two_varid f span buf len
  = return (L span $! f (lexemeToFastString (stepOn (stepOn buf)) (len-2)))

1017
strtoken :: (String -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
1018
strtoken f span buf len =
1019
  return (L span $! (f $! lexemeToString buf len))
1020 1021

begin :: Int -> Action
1022
begin code _span _str _len = do pushLexState code; lexToken
1023 1024

pop :: Action
Ian Lynagh's avatar
Ian Lynagh committed