Lexer.x 90.4 KB
Newer Older
1
-----------------------------------------------------------------------------
2
-- (c) The University of Glasgow, 2006
3 4 5 6 7 8 9
--
-- GHC's lexer.
--
-- This is a combination of an Alex-generated lexer from a regex
-- definition, with some hand-coded bits.
--
-- 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 18 19 20 21 22
--
-----------------------------------------------------------------------------

--   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

23
--   qualified operator NOTES.
Ian Lynagh's avatar
Ian Lynagh committed
24
--
25 26 27 28 29 30 31 32 33
--   - 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.

34
{
35
-- XXX The above flags turn off warnings in the generated code:
36
{-# LANGUAGE BangPatterns #-}
37 38 39 40 41 42 43 44
{-# 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 #-}
45

46 47
{-# OPTIONS_GHC -funbox-strict-fields #-}

48
module Lexer (
49
   Token(..), lexer, pragState, mkPState, PState(..),
Ian Lynagh's avatar
Ian Lynagh committed
50
   P(..), ParseResult(..), getSrcLoc,
51
   getPState, getDynFlags, withThisPackage,
52
   failLocMsgP, failSpanMsgP, srcParseFail,
Ian Lynagh's avatar
Ian Lynagh committed
53
   getMessages,
54
   popContext, pushCurrentContext, setLastToken, setSrcLoc,
vivian's avatar
vivian committed
55
   activeContext, nextIsEOF,
simonpj@microsoft.com's avatar
simonpj@microsoft.com committed
56
   getLexState, popLexState, pushLexState,
57
   extension, bangPatEnabled, datatypeContextsEnabled,
58
   traditionalRecordSyntaxEnabled,
59
   typeLiteralsEnabled,
ian@well-typed.com's avatar
ian@well-typed.com committed
60
   explicitNamespacesEnabled, sccProfilingOn, hpcEnabled,
Jedai's avatar
Jedai committed
61 62
   addWarning,
   lexTokenStream
63 64
  ) where

65 66
import Bag
import ErrUtils
67 68 69 70 71
import Outputable
import StringBuffer
import FastString
import SrcLoc
import UniqFM
72
import DynFlags
73
import Module
74
import Ctype
Ian Lynagh's avatar
Ian Lynagh committed
75 76
import BasicTypes       ( InlineSpec(..), RuleMatchInfo(..), FractionalLit(..) )
import Util             ( readRational )
77

78
import Control.Monad
Simon Marlow's avatar
Simon Marlow committed
79
import Data.Bits
80
import Data.ByteString (ByteString)
81
import Data.Char
82
import Data.List
Ian Lynagh's avatar
Ian Lynagh committed
83
import Data.Maybe
84 85
import Data.Map (Map)
import qualified Data.Map as Map
Simon Marlow's avatar
Simon Marlow committed
86
import Data.Ratio
87
import Data.Word
88 89
}

90
$unispace    = \x05 -- Trick Alex into handling Unicode. See alexGetChar.
91
$whitechar   = [\ \n\r\f\v $unispace]
92
$white_no_nl = $whitechar # \n
93
$tab         = \t
94 95

$ascdigit  = 0-9
96
$unidigit  = \x03 -- Trick Alex into handling Unicode. See alexGetChar.
97
$decdigit  = $ascdigit -- for now, should really be $digit (ToDo)
98 99 100
$digit     = [$ascdigit $unidigit]

$special   = [\(\)\,\;\[\]\`\{\}]
101
$ascsymbol = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~]
102
$unisymbol = \x04 -- Trick Alex into handling Unicode. See alexGetChar.
103 104
$symbol    = [$ascsymbol $unisymbol] # [$special \_\:\"\']

105
$unilarge  = \x01 -- Trick Alex into handling Unicode. See alexGetChar.
106
$asclarge  = [A-Z]
107 108
$large     = [$asclarge $unilarge]

109
$unismall  = \x02 -- Trick Alex into handling Unicode. See alexGetChar.
110
$ascsmall  = [a-z]
111 112
$small     = [$ascsmall $unismall \_]

113
$unigraphic = \x06 -- Trick Alex into handling Unicode. See alexGetChar.
114
$graphic   = [$small $large $symbol $digit $special $unigraphic \:\"\']
115

Ian Lynagh's avatar
Ian Lynagh committed
116
$octit     = 0-7
117
$hexit     = [$decdigit A-F a-f]
118 119 120 121
$symchar   = [$symbol \:]
$nl        = [\n\r]
$idchar    = [$small $large $digit \']

122 123
$pragmachar = [$small $large $digit]

124 125
$docsym    = [\| \^ \* \$]

126 127 128 129 130 131
@varid     = $small $idchar*
@conid     = $large $idchar*

@varsym    = $symbol $symchar*
@consym    = \: $symchar*

132
@decimal     = $decdigit+
133 134 135 136 137 138 139 140 141
@octal       = $octit+
@hexadecimal = $hexit+
@exponent    = [eE] [\-\+]? @decimal

-- we support the hierarchical module name extension:
@qual = (@conid \.)+

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

142 143 144 145 146
-- normal signed numerical literals can only be explicitly negative,
-- not explicitly positive (contrast @exponent)
@negative = \-
@signed = @negative ?

147 148 149
haskell :-

-- everywhere: skip whitespace and comments
Ian Lynagh's avatar
Ian Lynagh committed
150
$white_no_nl+ ;
151
$tab+         { warn Opt_WarnTabs (text "Tab character") }
152 153 154 155 156

-- 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
157 158 159 160
-- are). We also rule out nested Haddock comments, if the -haddock flag is
-- set.

"{-" / { isNormalComment } { nested_comment lexToken }
161 162 163 164

-- 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.
165 166

-- Since Haddock comments aren't valid in every state, we need to rule them
Ian Lynagh's avatar
Ian Lynagh committed
167
-- out here.
168 169 170 171 172 173 174

-- 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
175 176
"-- " ~[$docsym \#] .* { lineCommentToken }
"--" [^$symbol : \ ] .* { lineCommentToken }
177 178 179

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

Jedai's avatar
Jedai committed
180
"-- " [$docsym \#] .* / { ifExtension (not . haddockEnabled) } { lineCommentToken }
181 182 183 184 185 186 187

-- 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.

Jedai's avatar
Jedai committed
188
"---"\-* [^$symbol :] .* { lineCommentToken }
189 190 191 192

-- 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
193
"--"\-* / { atEOL } { lineCommentToken }
194 195 196 197

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

Jedai's avatar
Jedai committed
198
"-- " / { atEOL } { lineCommentToken }
199 200 201 202 203 204 205 206 207 208 209

-- '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
210 211 212 213 214
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
  ^\# pragma .* \n                      ; -- GCC 3.3 CPP generated, apparently
  ^\# \! .* \n                          ; -- #!, for scripts
  ()                                    { do_bol }
215 216 217 218 219 220
}

-- 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.
<layout, layout_do> {
Ian Lynagh's avatar
Ian Lynagh committed
221 222 223 224
  \{ / { notFollowedBy '-' }            { hopefully_open_brace }
        -- we might encounter {-# here, but {- has been handled already
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
225 226 227
}

-- do is treated in a subtly different way, see new_layout_context
Ian Lynagh's avatar
Ian Lynagh committed
228 229
<layout>    ()                          { new_layout_context True }
<layout_do> ()                          { new_layout_context False }
230 231 232 233

-- 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
234
<layout_left>  ()                       { do_layout_left }
235

Ian Lynagh's avatar
Ian Lynagh committed
236
<0,option_prags> \n                     { begin bol }
237

238 239
"{-#" $whitechar* $pragmachar+ / { known_pragma linePrags }
                                { dispatch_pragmas linePrags }
240 241 242

-- single-line line pragmas, of the form
--    # <line> "<file>" <extra-stuff> \n
Ian Lynagh's avatar
Ian Lynagh committed
243 244 245
<line_prag1> $decdigit+                 { setLine line_prag1a }
<line_prag1a> \" [$graphic \ ]* \"      { setFile line_prag1b }
<line_prag1b> .*                        { pop }
246 247 248

-- Haskell-style line pragmas, of the form
--    {-# LINE <line> "<file>" #-}
Ian Lynagh's avatar
Ian Lynagh committed
249 250 251
<line_prag2> $decdigit+                 { setLine line_prag2a }
<line_prag2a> \" [$graphic \ ]* \"      { setFile line_prag2b }
<line_prag2b> "#-}"|"-}"                { pop }
252 253
   -- NOTE: accept -} at the end of a LINE pragma, for compatibility
   -- with older versions of GHC which generated these.
254

255
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
256
  "{-#" $whitechar* $pragmachar+
257 258 259 260 261
        $whitechar+ $pragmachar+ / { known_pragma twoWordPrags }
                                 { dispatch_pragmas twoWordPrags }

  "{-#" $whitechar* $pragmachar+ / { known_pragma oneWordPrags }
                                 { dispatch_pragmas oneWordPrags }
262

263
  -- We ignore all these pragmas, but don't generate a warning for them
264 265
  "{-#" $whitechar* $pragmachar+ / { known_pragma ignoredPrags }
                                 { dispatch_pragmas ignoredPrags }
266 267

  -- ToDo: should only be valid inside a pragma:
Ian Lynagh's avatar
Ian Lynagh committed
268
  "#-}"                          { endPrag }
269 270
}

271
<option_prags> {
272 273 274
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
                                   { dispatch_pragmas fileHeaderPrags }

Ian Lynagh's avatar
Ian Lynagh committed
275
  "-- #"                           { multiline_doc_comment }
David Waern's avatar
David Waern committed
276 277
}

278 279
<0> {
  -- In the "0" mode we ignore these pragmas
280
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
281 282 283
                     { nested_comment lexToken }
}

David Waern's avatar
David Waern committed
284
<0> {
Jedai's avatar
Jedai committed
285
  "-- #" .* { lineCommentToken }
286
}
287

288
<0,option_prags> {
289 290
  "{-#"  { warnThen Opt_WarnUnrecognisedPragmas (text "Unrecognised pragma")
                    (nested_comment lexToken) }
291 292
}

293 294
-- '0' state: ordinary lexemes

295 296
-- Haddock comments

297
<0,option_prags> {
David Waern's avatar
David Waern committed
298 299
  "-- " $docsym      / { ifExtension haddockEnabled } { multiline_doc_comment }
  "{-" \ ? $docsym   / { ifExtension haddockEnabled } { nested_doc_comment }
300 301
}

302 303
-- "special" symbols

304
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
305 306
  "[:" / { ifExtension parrEnabled }    { token ITopabrack }
  ":]" / { ifExtension parrEnabled }    { token ITcpabrack }
307
}
Ian Lynagh's avatar
Ian Lynagh committed
308

309
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
310 311 312 313 314 315 316 317
  "[|"      / { ifExtension thEnabled } { token ITopenExpQuote }
  "[e|"     / { ifExtension thEnabled } { token ITopenExpQuote }
  "[p|"     / { ifExtension thEnabled } { token ITopenPatQuote }
  "[d|"     / { ifExtension thEnabled } { layout_token ITopenDecQuote }
  "[t|"     / { ifExtension thEnabled } { token ITopenTypQuote }
  "|]"      / { ifExtension thEnabled } { token ITcloseQuote }
  \$ @varid / { ifExtension thEnabled } { skip_one_varid ITidEscape }
  "$("      / { ifExtension thEnabled } { token ITparenEscape }
318

319 320 321 322
-- For backward compatibility, accept the old dollar syntax
  "[$" @varid "|"  / { ifExtension qqEnabled }
                     { lex_quasiquote_tok }

323
  "[" @varid "|"  / { ifExtension qqEnabled }
324
                     { lex_quasiquote_tok }
325 326 327 328

  -- qualified quasi-quote (#5555)
  "[" @qual @varid "|"  / { ifExtension qqEnabled }
                          { lex_qquasiquote_tok }
329 330
}

331
<0> {
ross's avatar
ross committed
332
  "(|" / { ifExtension arrowsEnabled `alexAndPred` notFollowedBySymbol }
Ian Lynagh's avatar
Ian Lynagh committed
333
                                        { special IToparenbar }
334 335 336
  "|)" / { ifExtension arrowsEnabled }  { special ITcparenbar }
}

337
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
338
  \? @varid / { ifExtension ipEnabled } { skip_one_varid ITdupipvarid }
339 340
}

341
<0> {
342
  "(#" / { ifExtension unboxedTuplesEnabled }
Ian Lynagh's avatar
Ian Lynagh committed
343 344 345 346 347
         { token IToubxparen }
  "#)" / { ifExtension unboxedTuplesEnabled }
         { token ITcubxparen }
}

348
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
349 350 351 352 353 354 355 356 357 358
  \(                                    { special IToparen }
  \)                                    { special ITcparen }
  \[                                    { special ITobrack }
  \]                                    { special ITcbrack }
  \,                                    { special ITcomma }
  \;                                    { special ITsemi }
  \`                                    { special ITbackquote }

  \{                                    { open_brace }
  \}                                    { close_brace }
359 360
}

361
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
362 363 364 365
  @qual @varid                  { idtoken qvarid }
  @qual @conid                  { idtoken qconid }
  @varid                        { varid }
  @conid                        { idtoken conid }
366 367
}

368
<0> {
369 370 371 372
  @qual @varid "#"+ / { ifExtension magicHashEnabled } { idtoken qvarid }
  @qual @conid "#"+ / { ifExtension magicHashEnabled } { idtoken qconid }
  @varid "#"+       / { ifExtension magicHashEnabled } { varid }
  @conid "#"+       / { ifExtension magicHashEnabled } { idtoken conid }
373 374
}

375 376
-- ToDo: - move `var` and (sym) into lexical syntax?
--       - remove backquote from $special?
377
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
378 379
  @qual @varsym                                    { idtoken qvarsym }
  @qual @consym                                    { idtoken qconsym }
380 381
  @varsym                                          { varsym }
  @consym                                          { consym }
382 383
}

384 385
-- For the normal boxed literals we need to be careful
-- when trying to be close to Haskell98
386
<0> {
387
  -- Normal integral literals (:: Num a => a, from Integer)
Ian Lynagh's avatar
Ian Lynagh committed
388 389 390
  @decimal           { tok_num positive 0 0 decimal }
  0[oO] @octal       { tok_num positive 2 2 octal }
  0[xX] @hexadecimal { tok_num positive 2 2 hexadecimal }
391 392

  -- Normal rational literals (:: Fractional a => a, from Rational)
Ian Lynagh's avatar
Ian Lynagh committed
393
  @floating_point    { strtoken tok_float }
394 395
}

396
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
397
  -- Unboxed ints (:: Int#) and words (:: Word#)
398 399
  -- It's simpler (and faster?) to give separate cases to the negatives,
  -- especially considering octal/hexadecimal prefixes.
Ian Lynagh's avatar
Ian Lynagh committed
400 401 402 403 404
  @decimal                     \# / { ifExtension magicHashEnabled } { tok_primint positive 0 1 decimal }
  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 }
  @negative 0[oO] @octal       \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 octal }
405
  @negative 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 hexadecimal }
406

Ian Lynagh's avatar
Ian Lynagh committed
407 408 409 410
  @decimal                     \# \# / { ifExtension magicHashEnabled } { tok_primword 0 2 decimal }
  0[oO] @octal                 \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 octal }
  0[xX] @hexadecimal           \# \# / { ifExtension magicHashEnabled } { tok_primword 2 4 hexadecimal }

411 412
  -- Unboxed floats and doubles (:: Float#, :: Double#)
  -- prim_{float,double} work with signed literals
413 414
  @signed @floating_point \# / { ifExtension magicHashEnabled } { init_strtoken 1 tok_primfloat }
  @signed @floating_point \# \# / { ifExtension magicHashEnabled } { init_strtoken 2 tok_primdouble }
415 416 417 418 419 420
}

-- 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.
421
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
422 423
  \'                            { lex_char_tok }
  \"                            { lex_string_tok }
424 425 426 427 428 429
}

{
-- -----------------------------------------------------------------------------
-- The token type

430
data Token
Ian Lynagh's avatar
Ian Lynagh committed
431
  = ITas                        -- Haskell keywords
432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454
  | ITcase
  | ITclass
  | ITdata
  | ITdefault
  | ITderiving
  | ITdo
  | ITelse
  | IThiding
  | ITif
  | ITimport
  | ITin
  | ITinfix
  | ITinfixl
  | ITinfixr
  | ITinstance
  | ITlet
  | ITmodule
  | ITnewtype
  | ITof
  | ITqualified
  | ITthen
  | ITtype
  | ITwhere
Ian Lynagh's avatar
Ian Lynagh committed
455
  | ITscc                       -- ToDo: remove (we use {-# SCC "..." #-} now)
456

Ian Lynagh's avatar
Ian Lynagh committed
457
  | ITforall                    -- GHC extension keywords
458 459 460 461 462
  | ITforeign
  | ITexport
  | ITlabel
  | ITdynamic
  | ITsafe
463
  | ITinterruptible
464 465 466
  | ITunsafe
  | ITstdcallconv
  | ITccallconv
467
  | ITcapiconv
468
  | ITprimcallconv
469
  | ITmdo
470
  | ITfamily
471 472 473
  | ITgroup
  | ITby
  | ITusing
474

Ian Lynagh's avatar
Ian Lynagh committed
475
  -- Pragmas
476
  | ITinline_prag InlineSpec RuleMatchInfo
Ian Lynagh's avatar
Ian Lynagh committed
477 478
  | ITspec_prag                 -- SPECIALISE
  | ITspec_inline_prag Bool     -- SPECIALISE INLINE (or NOINLINE)
479 480
  | ITsource_prag
  | ITrules_prag
Ian Lynagh's avatar
Ian Lynagh committed
481
  | ITwarning_prag
482 483 484
  | ITdeprecated_prag
  | ITline_prag
  | ITscc_prag
andy@galois.com's avatar
andy@galois.com committed
485
  | ITgenerated_prag
486
  | ITcore_prag                 -- hdaume: core annotations
487
  | ITunpack_prag
488
  | ITnounpack_prag
489
  | ITann_prag
490
  | ITclose_prag
491 492 493
  | IToptions_prag String
  | ITinclude_prag String
  | ITlanguage_prag
494 495
  | ITvect_prag
  | ITvect_scalar_prag
496
  | ITnovect_prag
497
  | ITctype
498

Ian Lynagh's avatar
Ian Lynagh committed
499
  | ITdotdot                    -- reserved symbols
500 501 502 503
  | ITcolon
  | ITdcolon
  | ITequal
  | ITlam
504
  | ITlcase
505 506 507 508 509
  | ITvbar
  | ITlarrow
  | ITrarrow
  | ITat
  | ITtilde
batterseapower's avatar
batterseapower committed
510
  | ITtildehsh
511 512 513 514 515 516
  | ITdarrow
  | ITminus
  | ITbang
  | ITstar
  | ITdot

Ian Lynagh's avatar
Ian Lynagh committed
517
  | ITbiglam                    -- GHC-extension symbols
518

Ian Lynagh's avatar
Ian Lynagh committed
519
  | ITocurly                    -- special symbols
520 521 522 523
  | ITccurly
  | ITvocurly
  | ITvccurly
  | ITobrack
Ian Lynagh's avatar
Ian Lynagh committed
524 525
  | ITopabrack                  -- [:, for parallel arrays with -XParallelArrays
  | ITcpabrack                  -- :], for parallel arrays with -XParallelArrays
526 527 528 529 530 531 532 533 534
  | ITcbrack
  | IToparen
  | ITcparen
  | IToubxparen
  | ITcubxparen
  | ITsemi
  | ITcomma
  | ITunderscore
  | ITbackquote
dreixel's avatar
dreixel committed
535
  | ITsimpleQuote               --  '
536

Ian Lynagh's avatar
Ian Lynagh committed
537
  | ITvarid   FastString        -- identifiers
538 539 540 541 542 543 544
  | ITconid   FastString
  | ITvarsym  FastString
  | ITconsym  FastString
  | ITqvarid  (FastString,FastString)
  | ITqconid  (FastString,FastString)
  | ITqvarsym (FastString,FastString)
  | ITqconsym (FastString,FastString)
545 546
  | ITprefixqvarsym (FastString,FastString)
  | ITprefixqconsym (FastString,FastString)
547

Ian Lynagh's avatar
Ian Lynagh committed
548
  | ITdupipvarid   FastString   -- GHC extension: implicit param: ?x
549 550 551 552

  | ITchar       Char
  | ITstring     FastString
  | ITinteger    Integer
553
  | ITrational   FractionalLit
554 555

  | ITprimchar   Char
556
  | ITprimstring ByteString
557
  | ITprimint    Integer
Ian Lynagh's avatar
Ian Lynagh committed
558
  | ITprimword   Integer
559 560
  | ITprimfloat  FractionalLit
  | ITprimdouble FractionalLit
561

562
  -- Template Haskell extension tokens
Ian Lynagh's avatar
Ian Lynagh committed
563 564 565 566 567 568 569 570
  | ITopenExpQuote              --  [| or [e|
  | ITopenPatQuote              --  [p|
  | ITopenDecQuote              --  [d|
  | ITopenTypQuote              --  [t|
  | ITcloseQuote                --  |]
  | ITidEscape   FastString     --  $x
  | ITparenEscape               --  $(
  | ITtyQuote                   --  ''
571 572 573 574 575 576 577 578
  | 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 |]
579 580 581 582

  -- Arrow notation extension
  | ITproc
  | ITrec
Ian Lynagh's avatar
Ian Lynagh committed
583 584 585 586 587 588
  | IToparenbar                 --  (|
  | ITcparenbar                 --  |)
  | ITlarrowtail                --  -<
  | ITrarrowtail                --  >-
  | ITLarrowtail                --  -<<
  | ITRarrowtail                --  >>-
589

Ian Lynagh's avatar
Ian Lynagh committed
590 591
  | ITunknown String            -- Used when the lexer can't make sense of it
  | ITeof                       -- end of file token
592 593 594 595 596 597 598

  -- 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
599
  | ITdocOptionsOld   String     -- doc options declared "-- # ..."-style
Jedai's avatar
Jedai committed
600 601
  | ITlineComment     String     -- comment starting by "--"
  | ITblockComment    String     -- comment in {- -}
602

603
  deriving Show
604 605 606 607 608 609 610 611

-- 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)
--
Ian Lynagh's avatar
Ian Lynagh committed
612
reservedWordsFM :: UniqFM (Token, Int)
613
reservedWordsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659
    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 ),
         ( "_scc_",          ITscc,           0 ),            -- ToDo: remove

         ( "forall",         ITforall,        bit explicitForallBit .|.
                                              bit inRulePragBit),
         ( "mdo",            ITmdo,           bit recursiveDoBit),
         ( "family",         ITfamily,        bit tyFamBit),
         ( "group",          ITgroup,         bit transformComprehensionsBit),
         ( "by",             ITby,            bit transformComprehensionsBit),
         ( "using",          ITusing,         bit transformComprehensionsBit),

         ( "foreign",        ITforeign,       bit ffiBit),
         ( "export",         ITexport,        bit ffiBit),
         ( "label",          ITlabel,         bit ffiBit),
         ( "dynamic",        ITdynamic,       bit ffiBit),
         ( "safe",           ITsafe,          bit ffiBit .|.
                                              bit safeHaskellBit),
         ( "interruptible",  ITinterruptible, bit interruptibleFfiBit),
         ( "unsafe",         ITunsafe,        bit ffiBit),
         ( "stdcall",        ITstdcallconv,   bit ffiBit),
         ( "ccall",          ITccallconv,     bit ffiBit),
660
         ( "capi",           ITcapiconv,      bit cApiFfiBit),
Ian Lynagh's avatar
Ian Lynagh committed
661 662
         ( "prim",           ITprimcallconv,  bit ffiBit),

663 664
         ( "rec",            ITrec,           bit arrowsBit .|. 
                                              bit recursiveDoBit),
Ian Lynagh's avatar
Ian Lynagh committed
665
         ( "proc",           ITproc,          bit arrowsBit)
666 667
     ]

Ian Lynagh's avatar
Ian Lynagh committed
668
reservedSymsFM :: UniqFM (Token, Int -> Bool)
669
reservedSymsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
670 671 672 673 674 675 676 677 678 679 680 681
    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)
682
       ,("~#",  ITtildehsh, magicHashEnabled)
Ian Lynagh's avatar
Ian Lynagh committed
683 684 685 686 687
       ,("=>",  ITdarrow,   always)
       ,("-",   ITminus,    always)
       ,("!",   ITbang,     always)

        -- For data T (a::*) = MkT
688
       ,("*", ITstar, always) -- \i -> kindSigsEnabled i || tyFamEnabled i)
Ian Lynagh's avatar
Ian Lynagh committed
689
        -- For 'forall a . t'
690
       ,(".", ITdot,  always) -- \i -> explicitForallEnabled i || inRulePrag i)
Ian Lynagh's avatar
Ian Lynagh committed
691 692 693 694 695

       ,("-<",  ITlarrowtail, arrowsEnabled)
       ,(">-",  ITrarrowtail, arrowsEnabled)
       ,("-<<", ITLarrowtail, arrowsEnabled)
       ,(">>-", ITRarrowtail, arrowsEnabled)
696

Ian Lynagh's avatar
Ian Lynagh committed
697 698
       ,("∷",   ITdcolon, unicodeSyntaxEnabled)
       ,("⇒",   ITdarrow, unicodeSyntaxEnabled)
Ian Lynagh's avatar
Ian Lynagh committed
699 700
       ,("∀",   ITforall, \i -> unicodeSyntaxEnabled i &&
                                explicitForallEnabled i)
Ian Lynagh's avatar
Ian Lynagh committed
701 702
       ,("→",   ITrarrow, unicodeSyntaxEnabled)
       ,("←",   ITlarrow, unicodeSyntaxEnabled)
703 704 705 706 707 708 709 710

       ,("⤙",   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
711 712 713
        -- 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).
714 715 716 717 718
       ]

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

Ian Lynagh's avatar
Ian Lynagh committed
719
type Action = RealSrcSpan -> StringBuffer -> Int -> P (RealLocated Token)
720

721
special :: Token -> Action
twanvl's avatar
twanvl committed
722
special tok span _buf _len = return (L span tok)
723

724
token, layout_token :: Token -> Action
twanvl's avatar
twanvl committed
725 726
token t span _buf _len = return (L span t)
layout_token t span _buf _len = pushLexState layout >> return (L span t)
727

728 729
idtoken :: (StringBuffer -> Int -> Token) -> Action
idtoken f span buf len = return (L span $! (f buf len))
730

731
skip_one_varid :: (FastString -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
732
skip_one_varid f span buf len
733
  = return (L span $! f (lexemeToFastString (stepOn buf) (len-1)))
734

735
strtoken :: (String -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
736
strtoken f span buf len =
737
  return (L span $! (f $! lexemeToString buf len))
738

739
init_strtoken :: Int -> (String -> Token) -> Action
740
-- like strtoken, but drops the last N character(s)
Ian Lynagh's avatar
Ian Lynagh committed
741
init_strtoken drop f span buf len =
742
  return (L span $! (f $! lexemeToString buf (len-drop)))
743 744

begin :: Int -> Action
745
begin code _span _str _len = do pushLexState code; lexToken
746 747

pop :: Action
Ian Lynagh's avatar
Ian Lynagh committed
748 749
pop _span _buf _len = do _ <- popLexState
                         lexToken
750

751 752 753 754 755 756 757 758 759 760 761
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
762
              else failSpanMsgP (RealSrcSpan span) (text "Missing block")
763

764
pop_and :: Action -> Action
Ian Lynagh's avatar
Ian Lynagh committed
765 766
pop_and act span buf len = do _ <- popLexState
                              act span buf len
767

768
{-# INLINE nextCharIs #-}
Ian Lynagh's avatar
Ian Lynagh committed
769
nextCharIs :: StringBuffer -> (Char -> Bool) -> Bool
770 771
nextCharIs buf p = not (atEnd buf) && p (currentChar buf)

772 773 774 775
{-# INLINE nextCharIsNot #-}
nextCharIsNot :: StringBuffer -> (Char -> Bool) -> Bool
nextCharIsNot buf p = not (nextCharIs buf p)

Ian Lynagh's avatar
Ian Lynagh committed
776
notFollowedBy :: Char -> AlexAccPred Int
Ian Lynagh's avatar
Ian Lynagh committed
777
notFollowedBy char _ _ _ (AI _ buf)
778
  = nextCharIsNot buf (== char)
779

Ian Lynagh's avatar
Ian Lynagh committed
780
notFollowedBySymbol :: AlexAccPred Int
781
notFollowedBySymbol _ _ _ (AI _ buf)
782
  = nextCharIsNot buf (`elem` "!#$%&*+./<=>?@\\^|-~")
ross's avatar
ross committed
783

784 785 786 787 788
-- 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.
Ian Lynagh's avatar
Ian Lynagh committed
789
isNormalComment :: AlexAccPred Int
790
isNormalComment bits _ _ (AI _ buf)
791
  | haddockEnabled bits = notFollowedByDocOrPragma
792
  | otherwise           = nextCharIsNot buf (== '#')
793 794
  where
    notFollowedByDocOrPragma
795
       = afterOptionalSpace buf (\b -> nextCharIsNot b (`elem` "|^*$#"))
796

797 798 799 800 801
afterOptionalSpace :: StringBuffer -> (StringBuffer -> Bool) -> Bool
afterOptionalSpace buf p
    = if nextCharIs buf (== ' ')
      then p (snd (nextChar buf))
      else p buf
802

Ian Lynagh's avatar
Ian Lynagh committed
803
atEOL :: AlexAccPred Int
804
atEOL _ _ _ (AI _ buf) = atEnd buf || currentChar buf == '\n'
805

Ian Lynagh's avatar
Ian Lynagh committed
806
ifExtension :: (Int -> Bool) -> AlexAccPred Int
807 808
ifExtension pred bits _ _ _ = pred bits

809 810 811
multiline_doc_comment :: Action
multiline_doc_comment span buf _len = withLexedDocType (worker "")
  where
812
    worker commentAcc input docType oneLine = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
813
      Just ('\n', input')
814 815 816 817 818 819
        | 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
820

821 822
    checkIfCommentLine input = check (dropNonNewlineSpace input)
      where
823 824 825
        check input = case alexGetChar' input of
          Just ('-', input) -> case alexGetChar' input of
            Just ('-', input) -> case alexGetChar' input of
826 827 828 829 830
              Just (c, _) | c /= '-' -> Just input
              _ -> Nothing
            _ -> Nothing
          _ -> Nothing

831
        dropNonNewlineSpace input = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
832
          Just (c, input')
833 834 835 836
            | isSpace c && c /= '\n' -> dropNonNewlineSpace input'
            | otherwise -> input
          Nothing -> input

Jedai's avatar
Jedai committed
837 838 839 840 841
lineCommentToken :: Action
lineCommentToken span buf len = do
  b <- extension rawTokenStreamEnabled
  if b then strtoken ITlineComment span buf len else lexToken

842 843 844 845
{-
  nested comments require traversing by hand, they can't be parsed
  using regular expressions.
-}
Ian Lynagh's avatar
Ian Lynagh committed
846
nested_comment :: P (RealLocated Token) -> Action
847
nested_comment cont span _str _len = do
848
  input <- getInput
Jedai's avatar
Jedai committed
849
  go "" (1::Int) input
850
  where
Jedai's avatar
Jedai committed
851 852 853 854 855
    go commentAcc 0 input = do setInput input
                               b <- extension rawTokenStreamEnabled
                               if b
                                 then docCommentEnd input commentAcc ITblockComment _str span
                                 else cont
856
    go commentAcc n input = case alexGetChar' input of
857
      Nothing -> errBrace input span
858
      Just ('-',input) -> case alexGetChar' input of
859
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
860 861
        Just ('\125',input) -> go commentAcc (n-1) input
        Just (_,_)          -> go ('-':commentAcc) n input
862
      Just ('\123',input) -> case alexGetChar' input of
863
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
864 865 866
        Just ('-',input) -> go ('-':'\123':commentAcc) (n+1) input
        Just (_,_)       -> go ('\123':commentAcc) n input
      Just (c,input) -> go (c:commentAcc) n input
867 868 869 870

nested_doc_comment :: Action
nested_doc_comment span buf _len = withLexedDocType (go "")
  where
871
    go commentAcc input docType _ = case alexGetChar' input of
872
      Nothing -> errBrace input span
873
      Just ('-',input) -> case alexGetChar' input of
874
        Nothing -> errBrace input span
twanvl's avatar
twanvl committed
875
        Just ('\125',input) ->
876
          docCommentEnd input commentAcc docType buf span
twanvl's avatar
twanvl committed
877
        Just (_,_) -> go ('-':commentAcc) input docType False
878
      Just ('\123', input) -> case alexGetChar' input of
879 880 881 882 883
        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
884
        Just (_,_) -> go ('\123':commentAcc) input docType False
885 886
      Just (c,input) -> go (c:commentAcc) input docType False

Ian Lynagh's avatar
Ian Lynagh committed
887 888
withLexedDocType :: (AlexInput -> (String -> Token) -> Bool -> P (RealLocated Token))
                 -> P (RealLocated Token)
889
withLexedDocType lexDocComment = do
890
  input@(AI _ buf) <- getInput
891 892 893 894
  case prevChar buf ' ' of
    '|' -> lexDocComment input ITdocCommentNext False
    '^' -> lexDocComment input ITdocCommentPrev False
    '$' -> lexDocComment input ITdocCommentNamed False
David Waern's avatar
David Waern committed
895 896
    '*' -> lexDocSection 1 input
    '#' -> lexDocComment input ITdocOptionsOld False
Ian Lynagh's avatar
Ian Lynagh committed
897
    _ -> panic "withLexedDocType: Bad doc type"
Ian Lynagh's avatar
Ian Lynagh committed
898
 where
899
    lexDocSection n input = case alexGetChar' input of
900
      Just ('*', input) -> lexDocSection (n+1) input
twanvl's avatar
twanvl committed
901
      Just (_,   _)     -> lexDocComment input (ITdocSection n) True
902 903
      Nothing -> do setInput input; lexToken -- eof reached, lex it normally

904 905 906
-- RULES pragmas turn on the forall and '.' keywords, and we turn them
-- off again at the end of the pragma.
rulePrag :: Action
907
rulePrag span _buf _len = do
908
  setExts (.|. bit inRulePragBit)
909 910 911
  return (L span ITrules_prag)

endPrag :: Action
912
endPrag span _buf _len = do
913 914 915
  setExts (.&. complement (bit inRulePragBit))
  return (L span ITclose_prag)

916 917 918 919
-- docCommentEnd
-------------------------------------------------------------------------------
-- This function is quite tricky. We can't just return a new token, we also
-- need to update the state of the parser. Why? Because the token is longer
Ian Lynagh's avatar
Ian Lynagh committed
920
-- than what was lexed by Alex, and the lexToken function doesn't know this, so
921
-- it writes the wrong token length to the parser state. This function is
Ian Lynagh's avatar
Ian Lynagh committed
922
-- called afterwards, so it can just update the state.
923 924

docCommentEnd :: AlexInput -> String -> (String -> Token) -> StringBuffer ->
Ian Lynagh's avatar
Ian Lynagh committed
925
                 RealSrcSpan -> P (RealLocated Token)
926 927
docCommentEnd input commentAcc docType buf span = do
  setInput input
928
  let (AI loc nextBuf) = input
929
      comment = reverse commentAcc
Ian Lynagh's avatar
Ian Lynagh committed
930
      span' = mkRealSrcSpan (realSrcSpanStart span) loc
931
      last_len = byteDiff buf nextBuf
Ian Lynagh's avatar
Ian Lynagh committed
932

933
  span `seq` setLastToken span' last_len
934
  return (L span' (docType comment))
Ian Lynagh's avatar
Ian Lynagh committed
935

Ian Lynagh's avatar
Ian Lynagh committed
936 937
errBrace :: AlexInput -> RealSrcSpan -> P a
errBrace (AI end _) span = failLocMsgP (realSrcSpanStart span) end "unterminated `{-'"
Ian Lynagh's avatar
Ian Lynagh committed
938

939
open_brace, close_brace :: Action
Ian Lynagh's avatar
Ian Lynagh committed
940
open_brace span _str _len = do
941 942
  ctx <- getContext
  setContext (NoLayout:ctx)
943
  return (L span ITocurly)
Ian Lynagh's avatar
Ian Lynagh committed
944
close_brace span _str _len = do
945
  popContext
946
  return (L span ITccurly)
947

Ian Lynagh's avatar
Ian Lynagh committed
948
qvarid, qconid :: StringBuffer -> Int -> Token
949 950
qvarid buf len = ITqvarid $! splitQualName buf len False
qconid buf len = ITqconid $! splitQualName buf len False
951

952
splitQualName :: StringBuffer -> Int -> Bool -> (FastString,FastString)
953 954 955
-- takes a StringBuffer and a length, and returns the module name
-- and identifier parts of a qualified name.  Splits at the *last* dot,
-- because of hierarchical module names.
956
splitQualName orig_buf len parens = split orig_buf orig_buf
957
  where
958
    split buf dot_buf
Ian Lynagh's avatar
Ian Lynagh committed
959 960 961
        | orig_buf `byteDiff` buf >= len  = done dot_buf
        | c == '.'                        = found_dot buf'
        | otherwise                       = split buf' dot_buf
962 963
      where
       (c,buf') = nextChar buf
Ian Lynagh's avatar
Ian Lynagh committed
964

965 966 967
    -- careful, we might get names like M....
    -- so, if the character after the dot is not upper-case, this is
    -- the end of the qualifier part.
968
    found_dot buf -- buf points after the '.'
Ian Lynagh's avatar
Ian Lynagh committed
969 970
        | isUpper c    = split buf' buf
        | otherwise    = done buf
971 972 973 974
      where
       (c,buf') = nextChar buf

    done dot_buf =
Ian Lynagh's avatar
Ian Lynagh committed
975 976
        (lexemeToFastString orig_buf (qual_size - 1),
         if parens -- Prelude.(+)
977 978
            then lexemeToFastString (stepOn dot_buf) (len - qual_size - 2)
            else lexemeToFastString dot_buf (len - qual_size))
979
      where
Ian Lynagh's avatar
Ian Lynagh committed
980
        qual_size = orig_buf `byteDiff` dot_buf
981

Ian Lynagh's avatar
Ian Lynagh committed
982 983
varid :: Action
varid span buf len =
984
  case lookupUFM reservedWordsFM fs of
985 986 987 988 989 990 991 992 993
    Just (ITcase, _) -> do
      lambdaCase <- extension lambdaCaseEnabled
      keyword <- if lambdaCase
                 then do
                   lastTk <- getLastTk
                   return $ case lastTk of
                     Just ITlam -> ITlcase
                     _          -> ITcase
                 else
Simon Marlow's avatar
Simon Marlow committed
994
                   return ITcase
995 996 997 998 999 1000 1001 1002
      maybe_layout keyword
      return $ L span keyword
    Just (keyword, 0) -> do
      maybe_layout keyword
      return $ L span keyword
    Just (keyword, exts) -> do
      extsEnabled <- extension $ \i -> exts .&. i /= 0
      if extsEnabled
Simon Marlow's avatar
Simon Marlow committed
1003 1004
        then do
          maybe_layout keyword