Lexer.x 95.5 KB
Newer Older
1
-----------------------------------------------------------------------------
2
-- (c) The University of Glasgow, 2006
3
--
4
-- GHC's lexer for Haskell 2010 [1].
5
--
6 7
-- This is a combination of an Alex-generated lexer [2] from a regex
-- definition, with some hand-coded bits. [3]
8 9
--
-- Completely accurate information about token-spans within the source
Ian Lynagh's avatar
Ian Lynagh committed
10 11
-- file is maintained.  Every token has a start and end RealSrcLoc
-- attached to it.
12
--
13 14 15 16 17
-- References:
-- [1] https://www.haskell.org/onlinereport/haskell2010/haskellch2.html
-- [2] http://www.haskell.org/alex/
-- [3] https://ghc.haskell.org/trac/ghc/wiki/Commentary/Compiler/Parser
--
18 19 20 21 22 23 24 25 26 27
-----------------------------------------------------------------------------

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

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

39 40 41 42

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

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

55 56
{-# OPTIONS_GHC -funbox-strict-fields #-}

57
module Lexer (
58
   Token(..), lexer, pragState, mkPState, PState(..),
Ian Lynagh's avatar
Ian Lynagh committed
59
   P(..), ParseResult(..), getSrcLoc,
60
   getPState, getDynFlags, withThisPackage,
61
   failLocMsgP, failSpanMsgP, srcParseFail,
Ian Lynagh's avatar
Ian Lynagh committed
62
   getMessages,
63
   popContext, pushCurrentContext, setLastToken, setSrcLoc,
vivian's avatar
vivian committed
64
   activeContext, nextIsEOF,
simonpj@microsoft.com's avatar
simonpj@microsoft.com committed
65
   getLexState, popLexState, pushLexState,
66
   extension, bangPatEnabled, datatypeContextsEnabled,
67
   traditionalRecordSyntaxEnabled,
68 69
   explicitForallEnabled,
   inRulePrag,
cactus's avatar
cactus committed
70 71 72
   explicitNamespacesEnabled,
   patternSynonymsEnabled,
   sccProfilingOn, hpcEnabled,
Jedai's avatar
Jedai committed
73 74
   addWarning,
   lexTokenStream
75 76
  ) where

77 78
import Bag
import ErrUtils
79 80 81 82 83
import Outputable
import StringBuffer
import FastString
import SrcLoc
import UniqFM
84
import DynFlags
85
import Module
86
import Ctype
Ian Lynagh's avatar
Ian Lynagh committed
87 88
import BasicTypes       ( InlineSpec(..), RuleMatchInfo(..), FractionalLit(..) )
import Util             ( readRational )
89

90
import Control.Applicative
91
import Control.Monad
Simon Marlow's avatar
Simon Marlow committed
92
import Data.Bits
93
import Data.ByteString (ByteString)
94
import Data.Char
95
import Data.List
Ian Lynagh's avatar
Ian Lynagh committed
96
import Data.Maybe
97 98
import Data.Map (Map)
import qualified Data.Map as Map
Simon Marlow's avatar
Simon Marlow committed
99
import Data.Ratio
100
import Data.Word
101 102
}

103 104 105 106 107 108 109 110

-- -----------------------------------------------------------------------------
-- Alex "Character set macros"

$unispace    = \x05 -- Trick Alex into handling Unicode. See alexGetByte.
$nl          = [\n\r\f]
$whitechar   = [$nl\v\ $unispace]
$white_no_nl = $whitechar # \n -- TODO #8424
111
$tab         = \t
112 113

$ascdigit  = 0-9
114
$unidigit  = \x03 -- Trick Alex into handling Unicode. See alexGetByte.
115
$decdigit  = $ascdigit -- for now, should really be $digit (ToDo)
116 117 118
$digit     = [$ascdigit $unidigit]

$special   = [\(\)\,\;\[\]\`\{\}]
119 120 121
$ascsymbol = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~\:]
$unisymbol = \x04 -- Trick Alex into handling Unicode. See alexGetByte.
$symbol    = [$ascsymbol $unisymbol] # [$special \_\"\']
122

123
$unilarge  = \x01 -- Trick Alex into handling Unicode. See alexGetByte.
124
$asclarge  = [A-Z]
125 126
$large     = [$asclarge $unilarge]

127
$unismall  = \x02 -- Trick Alex into handling Unicode. See alexGetByte.
128
$ascsmall  = [a-z]
129 130
$small     = [$ascsmall $unismall \_]

131 132
$unigraphic = \x06 -- Trick Alex into handling Unicode. See alexGetByte.
$graphic   = [$small $large $symbol $digit $special $unigraphic \"\']
133

134
$binit     = 0-1
Ian Lynagh's avatar
Ian Lynagh committed
135
$octit     = 0-7
136
$hexit     = [$decdigit A-F a-f]
137 138
$idchar    = [$small $large $digit \']

139 140
$pragmachar = [$small $large $digit]

141 142
$docsym    = [\| \^ \* \$]

143

144 145 146 147 148 149 150 151
-- -----------------------------------------------------------------------------
-- 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
152

153
@decimal     = $decdigit+
154
@binary      = $binit+
155 156 157 158 159
@octal       = $octit+
@hexadecimal = $hexit+
@exponent    = [eE] [\-\+]? @decimal

@qual = (@conid \.)+
160 161 162 163
@qvarid = @qual @varid
@qconid = @qual @conid
@qvarsym = @qual @varsym
@qconsym = @qual @consym
164 165 166

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

167 168 169 170 171
-- normal signed numerical literals can only be explicitly negative,
-- not explicitly positive (contrast @exponent)
@negative = \-
@signed = @negative ?

172 173 174 175

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

176 177
haskell :-

178 179 180 181 182

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

-- everywhere: skip whitespace
Ian Lynagh's avatar
Ian Lynagh committed
183
$white_no_nl+ ;
184
$tab+         { warn Opt_WarnTabs (text "Tab character") }
185 186 187 188 189

-- 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
190 191 192 193
-- are). We also rule out nested Haddock comments, if the -haddock flag is
-- set.

"{-" / { isNormalComment } { nested_comment lexToken }
194 195 196 197

-- 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.
198 199

-- Since Haddock comments aren't valid in every state, we need to rule them
Ian Lynagh's avatar
Ian Lynagh committed
200
-- out here.
201 202 203 204 205 206 207

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

Jedai's avatar
Jedai committed
208
"-- " ~[$docsym \#] .* { lineCommentToken }
209
"--" [^$symbol \ ] .* { lineCommentToken }
210 211 212

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

Jedai's avatar
Jedai committed
213
"-- " [$docsym \#] .* / { ifExtension (not . haddockEnabled) } { lineCommentToken }
214 215 216 217 218 219 220

-- 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.

221
"---"\-* ~$symbol .* { lineCommentToken }
222 223 224 225

-- 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
226
"--"\-* / { atEOL } { lineCommentToken }
227 228 229 230

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

Jedai's avatar
Jedai committed
231
"-- " / { atEOL } { lineCommentToken }
232 233 234 235 236 237 238 239 240 241 242

-- '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
243 244 245 246 247
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
  ^\# pragma .* \n                      ; -- GCC 3.3 CPP generated, apparently
  ^\# \! .* \n                          ; -- #!, for scripts
  ()                                    { do_bol }
248 249 250 251 252
}

-- 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.
253
<layout, layout_do, layout_if> {
Ian Lynagh's avatar
Ian Lynagh committed
254 255 256 257
  \{ / { notFollowedBy '-' }            { hopefully_open_brace }
        -- we might encounter {-# here, but {- has been handled already
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
258 259
}

260 261 262 263 264 265
-- after an 'if', a vertical bar starts a layout context for MultiWayIf
<layout_if> {
  \| / { notFollowedBySymbol }          { new_layout_context True ITvbar }
  ()                                    { pop }
}

266
-- do is treated in a subtly different way, see new_layout_context
267 268
<layout>    ()                          { new_layout_context True  ITvocurly }
<layout_do> ()                          { new_layout_context False ITvocurly }
269 270 271 272

-- 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
273
<layout_left>  ()                       { do_layout_left }
274

Ian Lynagh's avatar
Ian Lynagh committed
275
<0,option_prags> \n                     { begin bol }
276

277 278
"{-#" $whitechar* $pragmachar+ / { known_pragma linePrags }
                                { dispatch_pragmas linePrags }
279 280 281

-- single-line line pragmas, of the form
--    # <line> "<file>" <extra-stuff> \n
282
<line_prag1> @decimal                   { setLine line_prag1a }
Ian Lynagh's avatar
Ian Lynagh committed
283 284
<line_prag1a> \" [$graphic \ ]* \"      { setFile line_prag1b }
<line_prag1b> .*                        { pop }
285 286 287

-- Haskell-style line pragmas, of the form
--    {-# LINE <line> "<file>" #-}
288
<line_prag2> @decimal                   { setLine line_prag2a }
Ian Lynagh's avatar
Ian Lynagh committed
289 290
<line_prag2a> \" [$graphic \ ]* \"      { setFile line_prag2b }
<line_prag2b> "#-}"|"-}"                { pop }
291 292
   -- NOTE: accept -} at the end of a LINE pragma, for compatibility
   -- with older versions of GHC which generated these.
293

294
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
295
  "{-#" $whitechar* $pragmachar+
296 297 298 299 300
        $whitechar+ $pragmachar+ / { known_pragma twoWordPrags }
                                 { dispatch_pragmas twoWordPrags }

  "{-#" $whitechar* $pragmachar+ / { known_pragma oneWordPrags }
                                 { dispatch_pragmas oneWordPrags }
301

302
  -- We ignore all these pragmas, but don't generate a warning for them
303 304
  "{-#" $whitechar* $pragmachar+ / { known_pragma ignoredPrags }
                                 { dispatch_pragmas ignoredPrags }
305 306

  -- ToDo: should only be valid inside a pragma:
Ian Lynagh's avatar
Ian Lynagh committed
307
  "#-}"                          { endPrag }
308 309
}

310
<option_prags> {
311 312 313
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
                                   { dispatch_pragmas fileHeaderPrags }

Ian Lynagh's avatar
Ian Lynagh committed
314
  "-- #"                           { multiline_doc_comment }
David Waern's avatar
David Waern committed
315 316
}

317 318
<0> {
  -- In the "0" mode we ignore these pragmas
319
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
320 321 322
                     { nested_comment lexToken }
}

David Waern's avatar
David Waern committed
323
<0> {
Jedai's avatar
Jedai committed
324
  "-- #" .* { lineCommentToken }
325
}
326

327
<0,option_prags> {
328 329
  "{-#"  { warnThen Opt_WarnUnrecognisedPragmas (text "Unrecognised pragma")
                    (nested_comment lexToken) }
330 331
}

332 333
-- '0' state: ordinary lexemes

334 335
-- Haddock comments

336
<0,option_prags> {
David Waern's avatar
David Waern committed
337 338
  "-- " $docsym      / { ifExtension haddockEnabled } { multiline_doc_comment }
  "{-" \ ? $docsym   / { ifExtension haddockEnabled } { nested_doc_comment }
339 340
}

341 342
-- "special" symbols

343
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
344 345
  "[:" / { ifExtension parrEnabled }    { token ITopabrack }
  ":]" / { ifExtension parrEnabled }    { token ITcpabrack }
346
}
Ian Lynagh's avatar
Ian Lynagh committed
347

348
<0> {
349 350 351
  "[|"        / { ifExtension thEnabled } { token ITopenExpQuote }
  "[||"       / { ifExtension thEnabled } { token ITopenTExpQuote }
  "[e|"       / { ifExtension thEnabled } { token ITopenExpQuote }
352
  "[e||"      / { ifExtension thEnabled } { token ITopenTExpQuote }
353 354 355 356 357 358 359 360 361
  "[p|"       / { ifExtension thEnabled } { token ITopenPatQuote }
  "[d|"       / { ifExtension thEnabled } { layout_token ITopenDecQuote }
  "[t|"       / { ifExtension thEnabled } { token ITopenTypQuote }
  "|]"        / { ifExtension thEnabled } { token ITcloseQuote }
  "||]"       / { ifExtension thEnabled } { token ITcloseTExpQuote }
  \$ @varid   / { ifExtension thEnabled } { skip_one_varid ITidEscape }
  "$$" @varid / { ifExtension thEnabled } { skip_two_varid ITidTyEscape }
  "$("        / { ifExtension thEnabled } { token ITparenEscape }
  "$$("       / { ifExtension thEnabled } { token ITparenTyEscape }
362

363 364 365 366
-- For backward compatibility, accept the old dollar syntax
  "[$" @varid "|"  / { ifExtension qqEnabled }
                     { lex_quasiquote_tok }

367
  "[" @varid "|"  / { ifExtension qqEnabled }
368
                     { lex_quasiquote_tok }
369 370

  -- qualified quasi-quote (#5555)
371 372
  "[" @qvarid "|"  / { ifExtension qqEnabled }
                     { lex_qquasiquote_tok }
373 374
}

375
<0> {
ross's avatar
ross committed
376
  "(|" / { ifExtension arrowsEnabled `alexAndPred` notFollowedBySymbol }
Ian Lynagh's avatar
Ian Lynagh committed
377
                                        { special IToparenbar }
378 379 380
  "|)" / { ifExtension arrowsEnabled }  { special ITcparenbar }
}

381
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
382
  \? @varid / { ifExtension ipEnabled } { skip_one_varid ITdupipvarid }
383 384
}

385
<0> {
386
  "(#" / { ifExtension unboxedTuplesEnabled }
Ian Lynagh's avatar
Ian Lynagh committed
387 388 389 390 391
         { token IToubxparen }
  "#)" / { ifExtension unboxedTuplesEnabled }
         { token ITcubxparen }
}

392
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
393 394 395 396 397 398 399 400 401 402
  \(                                    { special IToparen }
  \)                                    { special ITcparen }
  \[                                    { special ITobrack }
  \]                                    { special ITcbrack }
  \,                                    { special ITcomma }
  \;                                    { special ITsemi }
  \`                                    { special ITbackquote }

  \{                                    { open_brace }
  \}                                    { close_brace }
403 404
}

405
<0,option_prags> {
406 407
  @qvarid                       { idtoken qvarid }
  @qconid                       { idtoken qconid }
Ian Lynagh's avatar
Ian Lynagh committed
408
  @varid                        { varid }
409
  @conid                        { idtoken conid }
410 411
}

412
<0> {
413 414
  @qvarid "#"+      / { ifExtension magicHashEnabled } { idtoken qvarid }
  @qconid "#"+      / { ifExtension magicHashEnabled } { idtoken qconid }
415
  @varid "#"+       / { ifExtension magicHashEnabled } { varid }
416
  @conid "#"+       / { ifExtension magicHashEnabled } { idtoken conid }
417 418
}

419 420
-- ToDo: - move `var` and (sym) into lexical syntax?
--       - remove backquote from $special?
421
<0> {
422 423
  @qvarsym                                         { idtoken qvarsym }
  @qconsym                                         { idtoken qconsym }
424 425
  @varsym                                          { varsym }
  @consym                                          { consym }
426 427
}

428 429
-- For the normal boxed literals we need to be careful
-- when trying to be close to Haskell98
430
<0> {
431
  -- Normal integral literals (:: Num a => a, from Integer)
432
  @decimal                                                               { tok_num positive 0 0 decimal }
433
  0[bB] @binary                / { ifExtension binaryLiteralsEnabled }   { tok_num positive 2 2 binary }
434 435 436
  0[oO] @octal                                                           { tok_num positive 2 2 octal }
  0[xX] @hexadecimal                                                     { tok_num positive 2 2 hexadecimal }
  @negative @decimal           / { ifExtension negativeLiteralsEnabled } { tok_num negative 1 1 decimal }
437 438
  @negative 0[bB] @binary      / { ifExtension negativeLiteralsEnabled `alexAndPred`
                                   ifExtension binaryLiteralsEnabled }   { tok_num negative 3 3 binary }
439 440
  @negative 0[oO] @octal       / { ifExtension negativeLiteralsEnabled } { tok_num negative 3 3 octal }
  @negative 0[xX] @hexadecimal / { ifExtension negativeLiteralsEnabled } { tok_num negative 3 3 hexadecimal }
441 442

  -- Normal rational literals (:: Fractional a => a, from Rational)
443 444
  @floating_point                                                        { strtoken tok_float }
  @negative @floating_point    / { ifExtension negativeLiteralsEnabled } { strtoken tok_float }
445 446
}

447
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
448
  -- Unboxed ints (:: Int#) and words (:: Word#)
449 450
  -- It's simpler (and faster?) to give separate cases to the negatives,
  -- especially considering octal/hexadecimal prefixes.
Ian Lynagh's avatar
Ian Lynagh committed
451
  @decimal                     \# / { ifExtension magicHashEnabled } { tok_primint positive 0 1 decimal }
452 453
  0[bB] @binary                \# / { ifExtension magicHashEnabled `alexAndPred`
                                      ifExtension binaryLiteralsEnabled } { tok_primint positive 2 3 binary }
Ian Lynagh's avatar
Ian Lynagh committed
454 455 456
  0[oO] @octal                 \# / { ifExtension magicHashEnabled } { tok_primint positive 2 3 octal }
  0[xX] @hexadecimal           \# / { ifExtension magicHashEnabled } { tok_primint positive 2 3 hexadecimal }
  @negative @decimal           \# / { ifExtension magicHashEnabled } { tok_primint negative 1 2 decimal }
457 458
  @negative 0[bB] @binary      \# / { ifExtension magicHashEnabled `alexAndPred`
                                      ifExtension binaryLiteralsEnabled } { tok_primint negative 3 4 binary }
Ian Lynagh's avatar
Ian Lynagh committed
459
  @negative 0[oO] @octal       \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 octal }
460
  @negative 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 hexadecimal }
461

Ian Lynagh's avatar
Ian Lynagh committed
462
  @decimal                     \# \# / { ifExtension magicHashEnabled } { tok_primword 0 2 decimal }
463 464
  0[bB] @binary                \# \# / { ifExtension magicHashEnabled `alexAndPred`
                                         ifExtension binaryLiteralsEnabled } { tok_primword 2 4 binary }
Ian Lynagh's avatar
Ian Lynagh committed
465 466 467
  0[oO] @octal                 \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 octal }
  0[xX] @hexadecimal           \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 hexadecimal }

468 469
  -- Unboxed floats and doubles (:: Float#, :: Double#)
  -- prim_{float,double} work with signed literals
470 471
  @signed @floating_point \# / { ifExtension magicHashEnabled } { init_strtoken 1 tok_primfloat }
  @signed @floating_point \# \# / { ifExtension magicHashEnabled } { init_strtoken 2 tok_primdouble }
472 473 474 475 476 477
}

-- 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.
478
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
479 480
  \'                            { lex_char_tok }
  \"                            { lex_string_tok }
481 482
}

483 484 485 486

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

487 488 489 490
{
-- -----------------------------------------------------------------------------
-- The token type

491
data Token
Ian Lynagh's avatar
Ian Lynagh committed
492
  = ITas                        -- Haskell keywords
493 494 495 496 497 498 499 500
  | ITcase
  | ITclass
  | ITdata
  | ITdefault
  | ITderiving
  | ITdo
  | ITelse
  | IThiding
501
  | ITforeign
502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517
  | ITif
  | ITimport
  | ITin
  | ITinfix
  | ITinfixl
  | ITinfixr
  | ITinstance
  | ITlet
  | ITmodule
  | ITnewtype
  | ITof
  | ITqualified
  | ITthen
  | ITtype
  | ITwhere

Ian Lynagh's avatar
Ian Lynagh committed
518
  | ITforall                    -- GHC extension keywords
519 520 521 522
  | ITexport
  | ITlabel
  | ITdynamic
  | ITsafe
523
  | ITinterruptible
524 525 526
  | ITunsafe
  | ITstdcallconv
  | ITccallconv
527
  | ITcapiconv
528
  | ITprimcallconv
thoughtpolice's avatar
thoughtpolice committed
529
  | ITjavascriptcallconv
530
  | ITmdo
531
  | ITfamily
532
  | ITrole
533 534 535
  | ITgroup
  | ITby
  | ITusing
cactus's avatar
cactus committed
536
  | ITpattern
537

Ian Lynagh's avatar
Ian Lynagh committed
538
  -- Pragmas
539
  | ITinline_prag InlineSpec RuleMatchInfo
Ian Lynagh's avatar
Ian Lynagh committed
540 541
  | ITspec_prag                 -- SPECIALISE
  | ITspec_inline_prag Bool     -- SPECIALISE INLINE (or NOINLINE)
542 543
  | ITsource_prag
  | ITrules_prag
Ian Lynagh's avatar
Ian Lynagh committed
544
  | ITwarning_prag
545 546 547
  | ITdeprecated_prag
  | ITline_prag
  | ITscc_prag
andy@galois.com's avatar
andy@galois.com committed
548
  | ITgenerated_prag
549
  | ITcore_prag                 -- hdaume: core annotations
550
  | ITunpack_prag
551
  | ITnounpack_prag
552
  | ITann_prag
553
  | ITclose_prag
554 555 556
  | IToptions_prag String
  | ITinclude_prag String
  | ITlanguage_prag
557 558
  | ITvect_prag
  | ITvect_scalar_prag
559
  | ITnovect_prag
560
  | ITminimal_prag
561 562 563
  | IToverlappable_prag         -- instance overlap mode
  | IToverlapping_prag          -- instance overlap mode
  | IToverlaps_prag             -- instance overlap mode
564
  | ITincoherent_prag           -- instance overlap mode
565
  | ITctype
566

Ian Lynagh's avatar
Ian Lynagh committed
567
  | ITdotdot                    -- reserved symbols
568 569 570 571
  | ITcolon
  | ITdcolon
  | ITequal
  | ITlam
572
  | ITlcase
573 574 575 576 577
  | ITvbar
  | ITlarrow
  | ITrarrow
  | ITat
  | ITtilde
batterseapower's avatar
batterseapower committed
578
  | ITtildehsh
579 580 581 582 583 584
  | ITdarrow
  | ITminus
  | ITbang
  | ITstar
  | ITdot

Ian Lynagh's avatar
Ian Lynagh committed
585
  | ITbiglam                    -- GHC-extension symbols
586

Ian Lynagh's avatar
Ian Lynagh committed
587
  | ITocurly                    -- special symbols
588 589 590 591
  | ITccurly
  | ITvocurly
  | ITvccurly
  | ITobrack
Ian Lynagh's avatar
Ian Lynagh committed
592 593
  | ITopabrack                  -- [:, for parallel arrays with -XParallelArrays
  | ITcpabrack                  -- :], for parallel arrays with -XParallelArrays
594 595 596 597 598 599 600 601 602
  | ITcbrack
  | IToparen
  | ITcparen
  | IToubxparen
  | ITcubxparen
  | ITsemi
  | ITcomma
  | ITunderscore
  | ITbackquote
dreixel's avatar
dreixel committed
603
  | ITsimpleQuote               --  '
604

Ian Lynagh's avatar
Ian Lynagh committed
605
  | ITvarid   FastString        -- identifiers
606 607 608 609 610 611 612
  | ITconid   FastString
  | ITvarsym  FastString
  | ITconsym  FastString
  | ITqvarid  (FastString,FastString)
  | ITqconid  (FastString,FastString)
  | ITqvarsym (FastString,FastString)
  | ITqconsym (FastString,FastString)
613 614
  | ITprefixqvarsym (FastString,FastString)
  | ITprefixqconsym (FastString,FastString)
615

Ian Lynagh's avatar
Ian Lynagh committed
616
  | ITdupipvarid   FastString   -- GHC extension: implicit param: ?x
617 618 619 620

  | ITchar       Char
  | ITstring     FastString
  | ITinteger    Integer
621
  | ITrational   FractionalLit
622 623

  | ITprimchar   Char
624
  | ITprimstring ByteString
625
  | ITprimint    Integer
Ian Lynagh's avatar
Ian Lynagh committed
626
  | ITprimword   Integer
627 628
  | ITprimfloat  FractionalLit
  | ITprimdouble FractionalLit
629

630
  -- Template Haskell extension tokens
Ian Lynagh's avatar
Ian Lynagh committed
631 632 633 634 635
  | ITopenExpQuote              --  [| or [e|
  | ITopenPatQuote              --  [p|
  | ITopenDecQuote              --  [d|
  | ITopenTypQuote              --  [t|
  | ITcloseQuote                --  |]
636 637
  | ITopenTExpQuote             --  [||
  | ITcloseTExpQuote            --  ||]
Ian Lynagh's avatar
Ian Lynagh committed
638 639
  | ITidEscape   FastString     --  $x
  | ITparenEscape               --  $(
640 641
  | ITidTyEscape   FastString   --  $$x
  | ITparenTyEscape             --  $$(
Ian Lynagh's avatar
Ian Lynagh committed
642
  | ITtyQuote                   --  ''
643 644 645 646 647 648 649 650
  | ITquasiQuote (FastString,FastString,RealSrcSpan)
    -- ITquasiQuote(quoter, quote, loc)
    -- represents a quasi-quote of the form
    -- [quoter| quote |]
  | ITqQuasiQuote (FastString,FastString,FastString,RealSrcSpan)
    -- ITqQuasiQuote(Qual, quoter, quote, loc)
    -- represents a qualified quasi-quote of the form
    -- [Qual.quoter| quote |]
651 652 653 654

  -- Arrow notation extension
  | ITproc
  | ITrec
Ian Lynagh's avatar
Ian Lynagh committed
655 656 657 658 659 660
  | IToparenbar                 --  (|
  | ITcparenbar                 --  |)
  | ITlarrowtail                --  -<
  | ITrarrowtail                --  >-
  | ITLarrowtail                --  -<<
  | ITRarrowtail                --  >>-
661

Ian Lynagh's avatar
Ian Lynagh committed
662 663
  | ITunknown String            -- Used when the lexer can't make sense of it
  | ITeof                       -- end of file token
664 665 666 667 668 669 670

  -- Documentation annotations
  | ITdocCommentNext  String     -- something beginning '-- |'
  | ITdocCommentPrev  String     -- something beginning '-- ^'
  | ITdocCommentNamed String     -- something beginning '-- $'
  | ITdocSection      Int String -- a section heading
  | ITdocOptions      String     -- doc options (prune, ignore-exports, etc)
David Waern's avatar
David Waern committed
671
  | ITdocOptionsOld   String     -- doc options declared "-- # ..."-style
Jedai's avatar
Jedai committed
672 673
  | ITlineComment     String     -- comment starting by "--"
  | ITblockComment    String     -- comment in {- -}
674

675
  deriving Show
676 677 678 679 680 681 682 683

-- 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)
--
684
reservedWordsFM :: UniqFM (Token, ExtsBitmap)
685
reservedWordsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712
    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 ),

713 714 715
         ( "forall",         ITforall,        xbit ExplicitForallBit .|.
                                              xbit InRulePragBit),
         ( "mdo",            ITmdo,           xbit RecursiveDoBit),
716 717 718
             -- See Note [Lexing type pseudo-keywords]
         ( "family",         ITfamily,        0 ),
         ( "role",           ITrole,          0 ),
719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740
         ( "pattern",        ITpattern,       xbit PatternSynonymsBit),
         ( "group",          ITgroup,         xbit TransformComprehensionsBit),
         ( "by",             ITby,            xbit TransformComprehensionsBit),
         ( "using",          ITusing,         xbit TransformComprehensionsBit),

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

         ( "rec",            ITrec,           xbit ArrowsBit .|.
                                              xbit RecursiveDoBit),
         ( "proc",           ITproc,          xbit ArrowsBit)
741 742
     ]

743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758
{-----------------------------------
Note [Lexing type pseudo-keywords]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

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

Also, note that these are included in the `varid` production in the parser --
a key detail to make all this work.
-------------------------------------}
759

760
reservedSymsFM :: UniqFM (Token, ExtsBitmap -> Bool)
761
reservedSymsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
762 763 764 765 766 767 768 769 770 771 772 773
    map (\ (x,y,z) -> (mkFastString x,(y,z)))
      [ ("..",  ITdotdot,   always)
        -- (:) is a reserved op, meaning only list cons
       ,(":",   ITcolon,    always)
       ,("::",  ITdcolon,   always)
       ,("=",   ITequal,    always)
       ,("\\",  ITlam,      always)
       ,("|",   ITvbar,     always)
       ,("<-",  ITlarrow,   always)
       ,("->",  ITrarrow,   always)
       ,("@",   ITat,       always)
       ,("~",   ITtilde,    always)
774
       ,("~#",  ITtildehsh, magicHashEnabled)
Ian Lynagh's avatar
Ian Lynagh committed
775 776 777 778 779
       ,("=>",  ITdarrow,   always)
       ,("-",   ITminus,    always)
       ,("!",   ITbang,     always)

        -- For data T (a::*) = MkT
780
       ,("*", ITstar, always) -- \i -> kindSigsEnabled i || tyFamEnabled i)
Ian Lynagh's avatar
Ian Lynagh committed
781
        -- For 'forall a . t'
782
       ,(".", ITdot,  always) -- \i -> explicitForallEnabled i || inRulePrag i)
Ian Lynagh's avatar
Ian Lynagh committed
783 784 785 786 787

       ,("-<",  ITlarrowtail, arrowsEnabled)
       ,(">-",  ITrarrowtail, arrowsEnabled)
       ,("-<<", ITLarrowtail, arrowsEnabled)
       ,(">>-", ITRarrowtail, arrowsEnabled)
788

Ian Lynagh's avatar
Ian Lynagh committed
789 790
       ,("∷",   ITdcolon, unicodeSyntaxEnabled)
       ,("⇒",   ITdarrow, unicodeSyntaxEnabled)
791
       ,("∀",   ITforall, unicodeSyntaxEnabled)
Ian Lynagh's avatar
Ian Lynagh committed
792 793
       ,("→",   ITrarrow, unicodeSyntaxEnabled)
       ,("←",   ITlarrow, unicodeSyntaxEnabled)
794 795 796 797 798 799 800 801

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

       ,("★", ITstar, unicodeSyntaxEnabled)

Simon Marlow's avatar
Simon Marlow committed
802 803 804
        -- 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).
805 806 807 808 809
       ]

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

Ian Lynagh's avatar
Ian Lynagh committed
810
type Action = RealSrcSpan -> StringBuffer -> Int -> P (RealLocated Token)
811

812
special :: Token -> Action
twanvl's avatar
twanvl committed
813
special tok span _buf _len = return (L span tok)
814

815
token, layout_token :: Token -> Action
twanvl's avatar
twanvl committed
816 817
token t span _buf _len = return (L span t)
layout_token t span _buf _len = pushLexState layout >> return (L span t)
818

819 820
idtoken :: (StringBuffer -> Int -> Token) -> Action
idtoken f span buf len = return (L span $! (f buf len))
821

822
skip_one_varid :: (FastString -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
823
skip_one_varid f span buf len
824
  = return (L span $! f (lexemeToFastString (stepOn buf) (len-1)))
825

826 827 828 829
skip_two_varid :: (FastString -> Token) -> Action
skip_two_varid f span buf len
  = return (L span $! f (lexemeToFastString (stepOn (stepOn buf)) (len-2)))

830
strtoken :: (String -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
831
strtoken f span buf len =
832
  return (L span $! (f $! lexemeToString buf len))
833

834
init_strtoken :: Int -> (String -> Token) -> Action
835
-- like strtoken, but drops the last N character(s)
Ian Lynagh's avatar
Ian Lynagh committed
836
init_strtoken drop f span buf len =
837
  return (L span $! (f $! lexemeToString buf (len-drop)))
838 839

begin :: Int -> Action
840
begin code _span _str _len = do pushLexState code; lexToken
841 842

pop :: Action
Ian Lynagh's avatar
Ian Lynagh committed
843 844
pop _span _buf _len = do _ <- popLexState
                         lexToken
845

846 847 848 849 850 851 852 853 854 855 856
hopefully_open_brace :: Action
hopefully_open_brace span buf len
 = do relaxed <- extension relaxedLayout
      ctx <- getContext
      (AI l _) <- getInput
      let offset = srcLocCol l
          isOK = relaxed ||
                 case ctx of
                 Layout prev_off : _ -> prev_off < offset
                 _                   -> True
      if isOK then pop_and open_brace span buf len
Ian Lynagh's avatar
Ian Lynagh committed
857
              else failSpanMsgP (RealSrcSpan span) (text "Missing block")
858

859
pop_and :: Action -> Action
Ian Lynagh's avatar
Ian Lynagh committed
860 861
pop_and act span buf len = do _ <- popLexState
                              act span buf len
862

863
{-# INLINE nextCharIs #-}
Ian Lynagh's avatar
Ian Lynagh committed
864
nextCharIs :: StringBuffer -> (Char -> Bool) -> Bool
865 866
nextCharIs buf p = not (atEnd buf) && p (currentChar buf)

867 868 869 870
{-# INLINE nextCharIsNot #-}
nextCharIsNot :: StringBuffer -> (Char -> Bool) -> Bool
nextCharIsNot buf p = not (nextCharIs buf p)

871
notFollowedBy :: Char -> AlexAccPred ExtsBitmap
Ian Lynagh's avatar
Ian Lynagh committed
872
notFollowedBy char _ _ _ (AI _ buf)
873
  = nextCharIsNot buf (== char)
874

875
notFollowedBySymbol :: AlexAccPred ExtsBitmap
876
notFollowedBySymbol _ _ _ (AI _ buf)
877
  = nextCharIsNot buf (`elem` "!#$%&*+./<=>?@\\^|-~")
ross's avatar
ross committed
878

879 880 881 882 883
-- We must reject doc comments as being ordinary comments everywhere.
-- In some cases the doc comment will be selected as the lexeme due to
-- maximal munch, but not always, because the nested comment rule is
-- valid in all states, but the doc-comment rules are only valid in
-- the non-layout states.
884
isNormalComment :: AlexAccPred ExtsBitmap
885
isNormalComment bits _ _ (AI _ buf)
886
  | haddockEnabled bits = notFollowedByDocOrPragma
887
  | otherwise           = nextCharIsNot buf (== '#')
888 889
  where
    notFollowedByDocOrPragma
890
       = afterOptionalSpace buf (\b -> nextCharIsNot b (`elem` "|^*$#"))
891

892 893 894 895 896
afterOptionalSpace :: StringBuffer -> (StringBuffer -> Bool) -> Bool
afterOptionalSpace buf p
    = if nextCharIs buf (== ' ')
      then p (snd (nextChar buf))
      else p buf
897

898
atEOL :: AlexAccPred ExtsBitmap
899
atEOL _ _ _ (AI _ buf) = atEnd buf || currentChar buf == '\n'
900

901
ifExtension :: (ExtsBitmap -> Bool) -> AlexAccPred ExtsBitmap
902 903
ifExtension pred bits _ _ _ = pred bits

904 905 906
multiline_doc_comment :: Action
multiline_doc_comment span buf _len = withLexedDocType (worker "")
  where
907
    worker commentAcc input docType oneLine = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
908
      Just ('\n', input')
909 910 911 912 913 914
        | oneLine -> docCommentEnd input commentAcc docType buf span
        | otherwise -> case checkIfCommentLine input' of
          Just input -> worker ('\n':commentAcc) input docType False
          Nothing -> docCommentEnd input commentAcc docType buf span
      Just (c, input) -> worker (c:commentAcc) input docType oneLine
      Nothing -> docCommentEnd input commentAcc docType buf span
Ian Lynagh's avatar
Ian Lynagh committed
915

916 917
    checkIfCommentLine input = check (dropNonNewlineSpace input)
      where
918 919 920
        check input = case alexGetChar' input of
          Just ('-', input) -> case alexGetChar' input of
            Just ('-', input) -> case alexGetChar' input of
921 922 923 924 925
              Just (c, _) | c /= '-' -> Just input
              _ -> Nothing
            _ -> Nothing
          _ -> Nothing

926
        dropNonNewlineSpace input = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
927
          Just (c, input')
928 929 930 931
            | isSpace c && c /= '\n' -> dropNonNewlineSpace input'
            | otherwise -> input
          Nothing -> input

Jedai's avatar
Jedai committed
932 933 934 935 936
lineCommentToken :: Action
lineCommentToken span buf len = do
  b <- extension rawTokenStreamEnabled
  if b then strtoken ITlineComment span buf len else lexToken

937 938 939 940
{-
  nested comments require traversing by hand, they can't be parsed
  using regular expressions.
-}
Ian Lynagh's avatar
Ian Lynagh committed
941
nested_comment :: P (RealLocated Token) -> Action
942
nested_comment cont span _str _len = do
943
  input <- getInput
Jedai's avatar
Jedai committed
944
  go "" (1::Int) input
945
  where
Jedai's avatar
Jedai committed
946 947 948 949 950
    go commentAcc 0 input = do setInput input
                               b <- extension rawTokenStreamEnabled
                               if b
                                 then docCommentEnd input commentAcc ITblockComment _str span
                                 else cont
951
    go commentAcc n input = case alexGetChar' input of
952
      Nothing -> errBrace input span
953
      Just ('-',input) -> case alexGetChar' input of
954
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
955 956
        Just ('\125',input) -> go commentAcc (n-1) input
        Just (_,_)          -> go ('-':commentAcc) n input
957
      Just ('\123',input) -> case alexGetChar' input of
958
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
959 960 961
        Just ('-',input) -> go ('-':'\123':commentAcc) (n+1) input
        Just (_,_)       -> go ('\123':commentAcc) n input
      Just (c,input) -> go (c:commentAcc) n input
962 963 964 965

nested_doc_comment :: Action
nested_doc_comment span buf _len = withLexedDocType (go "")
  where
966
    go commentAcc input docType _ = case alexGetChar' input of
967
      Nothing -> errBrace input span
968
      Just ('-',input) -> case alexGetChar' input of
969
        Nothing -> errBrace input span
twanvl's avatar
twanvl committed
970
        Just ('\125',input) ->
971
          docCommentEnd input commentAcc docType buf span
twanvl's avatar
twanvl committed
972
        Just (_,_) -> go ('-':commentAcc) input docType False
973
      Just ('\123', input) -> case alexGetChar' input of
974 975 976 977 978
        Nothing  -> errBrace input span
        Just ('-',input) -> do
          setInput input
          let cont = do input <- getInput; go commentAcc input docType False
          nested_comment cont span buf _len
twanvl's avatar
twanvl committed
979
        Just (_,_) -> go ('\123':commentAcc) input docType False
980 981
      Just (c,input) -> go (c:commentAcc) input docType False

Ian Lynagh's avatar
Ian Lynagh committed
982 983
withLexedDocType :: (AlexInput -> (String -> Token) -> Bool -> P (RealLocated Token))
                 -> P (RealLocated Token)
984
withLexedDocType lexDocComment = do
985
  input@(AI _ buf) <- getInput
986 987 988 989
  case prevChar buf ' ' of
    '|' -> lexDocComment input ITdocCommentNext False
    '^' -> lexDocComment input ITdocCommentPrev False
    '$' -> lexDocComment input ITdocCommentNamed False
David Waern's avatar
David Waern committed
990 991
    '*' -> lexDocSection 1 input
    '#' -> lexDocComment input ITdocOptionsOld False
Ian Lynagh's avatar
Ian Lynagh committed
992
    _ -> panic "withLexedDocType: Bad doc type"
Ian Lynagh's avatar
Ian Lynagh committed
993
 where
994
    lexDocSection n input = case alexGetChar' input of
995
      Just ('*', input) -> lexDocSection (n+1) input
twanvl's avatar
twanvl committed
996
      Just (_,   _)     -> lexDocComment input (ITdocSection n) True
997 998
      Nothing -> do setInput input; lexToken -- eof reached, lex it normally

999 1000 1001
-- RULES pragmas turn on the forall and '.' keywords, and we turn them
-- off again at the end of the pragma.
rulePrag :: Action
1002
rulePrag span _buf _len = do
1003
  setExts (.|. xbit InRulePragBit)
1004 1005 1006
  return (L span ITrules_prag)

endPrag :: Action
1007
endPrag span _buf _len = do
1008
  setExts (.&. complement (xbit InRulePragBit))