Lexer.x 92.9 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,
60 61
   explicitForallEnabled,
   inRulePrag,
cactus's avatar
cactus committed
62 63 64
   explicitNamespacesEnabled,
   patternSynonymsEnabled,
   sccProfilingOn, hpcEnabled,
Jedai's avatar
Jedai committed
65 66
   addWarning,
   lexTokenStream
67 68
  ) where

69 70
import Bag
import ErrUtils
71 72 73 74 75
import Outputable
import StringBuffer
import FastString
import SrcLoc
import UniqFM
76
import DynFlags
77
import Module
78
import Ctype
Ian Lynagh's avatar
Ian Lynagh committed
79 80
import BasicTypes       ( InlineSpec(..), RuleMatchInfo(..), FractionalLit(..) )
import Util             ( readRational )
81

82
import Control.Monad
Simon Marlow's avatar
Simon Marlow committed
83
import Data.Bits
84
import Data.ByteString (ByteString)
85
import Data.Char
86
import Data.List
Ian Lynagh's avatar
Ian Lynagh committed
87
import Data.Maybe
88 89
import Data.Map (Map)
import qualified Data.Map as Map
Simon Marlow's avatar
Simon Marlow committed
90
import Data.Ratio
91
import Data.Word
92 93
}

94
$unispace    = \x05 -- Trick Alex into handling Unicode. See alexGetChar.
95
$whitechar   = [\ \n\r\f\v $unispace]
96
$white_no_nl = $whitechar # \n
97
$tab         = \t
98 99

$ascdigit  = 0-9
100
$unidigit  = \x03 -- Trick Alex into handling Unicode. See alexGetChar.
101
$decdigit  = $ascdigit -- for now, should really be $digit (ToDo)
102 103 104
$digit     = [$ascdigit $unidigit]

$special   = [\(\)\,\;\[\]\`\{\}]
105
$ascsymbol = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~]
106
$unisymbol = \x04 -- Trick Alex into handling Unicode. See alexGetChar.
107 108
$symbol    = [$ascsymbol $unisymbol] # [$special \_\:\"\']

109
$unilarge  = \x01 -- Trick Alex into handling Unicode. See alexGetChar.
110
$asclarge  = [A-Z]
111 112
$large     = [$asclarge $unilarge]

113
$unismall  = \x02 -- Trick Alex into handling Unicode. See alexGetChar.
114
$ascsmall  = [a-z]
115 116
$small     = [$ascsmall $unismall \_]

117
$unigraphic = \x06 -- Trick Alex into handling Unicode. See alexGetChar.
118
$graphic   = [$small $large $symbol $digit $special $unigraphic \:\"\']
119

Ian Lynagh's avatar
Ian Lynagh committed
120
$octit     = 0-7
121
$hexit     = [$decdigit A-F a-f]
122 123 124 125
$symchar   = [$symbol \:]
$nl        = [\n\r]
$idchar    = [$small $large $digit \']

126 127
$pragmachar = [$small $large $digit]

128 129
$docsym    = [\| \^ \* \$]

130 131 132 133 134 135
@varid     = $small $idchar*
@conid     = $large $idchar*

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

136
@decimal     = $decdigit+
137 138 139 140 141 142 143 144 145
@octal       = $octit+
@hexadecimal = $hexit+
@exponent    = [eE] [\-\+]? @decimal

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

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

146 147 148 149 150
-- normal signed numerical literals can only be explicitly negative,
-- not explicitly positive (contrast @exponent)
@negative = \-
@signed = @negative ?

151 152 153
haskell :-

-- everywhere: skip whitespace and comments
Ian Lynagh's avatar
Ian Lynagh committed
154
$white_no_nl+ ;
155
$tab+         { warn Opt_WarnTabs (text "Tab character") }
156 157 158 159 160

-- 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
161 162 163 164
-- are). We also rule out nested Haddock comments, if the -haddock flag is
-- set.

"{-" / { isNormalComment } { nested_comment lexToken }
165 166 167 168

-- 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.
169 170

-- Since Haddock comments aren't valid in every state, we need to rule them
Ian Lynagh's avatar
Ian Lynagh committed
171
-- out here.
172 173 174 175 176 177 178

-- 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
179 180
"-- " ~[$docsym \#] .* { lineCommentToken }
"--" [^$symbol : \ ] .* { lineCommentToken }
181 182 183

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

Jedai's avatar
Jedai committed
184
"-- " [$docsym \#] .* / { ifExtension (not . haddockEnabled) } { lineCommentToken }
185 186 187 188 189 190 191

-- 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
192
"---"\-* [^$symbol :] .* { lineCommentToken }
193 194 195 196

-- 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
197
"--"\-* / { atEOL } { lineCommentToken }
198 199 200 201

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

Jedai's avatar
Jedai committed
202
"-- " / { atEOL } { lineCommentToken }
203 204 205 206 207 208 209 210 211 212 213

-- '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
214 215 216 217 218
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
  ^\# pragma .* \n                      ; -- GCC 3.3 CPP generated, apparently
  ^\# \! .* \n                          ; -- #!, for scripts
  ()                                    { do_bol }
219 220 221 222 223
}

-- 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.
224
<layout, layout_do, layout_if> {
Ian Lynagh's avatar
Ian Lynagh committed
225 226 227 228
  \{ / { notFollowedBy '-' }            { hopefully_open_brace }
        -- we might encounter {-# here, but {- has been handled already
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
229 230
}

231 232 233 234 235 236
-- after an 'if', a vertical bar starts a layout context for MultiWayIf
<layout_if> {
  \| / { notFollowedBySymbol }          { new_layout_context True ITvbar }
  ()                                    { pop }
}

237
-- do is treated in a subtly different way, see new_layout_context
238 239
<layout>    ()                          { new_layout_context True  ITvocurly }
<layout_do> ()                          { new_layout_context False ITvocurly }
240 241 242 243

-- 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
244
<layout_left>  ()                       { do_layout_left }
245

Ian Lynagh's avatar
Ian Lynagh committed
246
<0,option_prags> \n                     { begin bol }
247

248 249
"{-#" $whitechar* $pragmachar+ / { known_pragma linePrags }
                                { dispatch_pragmas linePrags }
250 251 252

-- single-line line pragmas, of the form
--    # <line> "<file>" <extra-stuff> \n
Ian Lynagh's avatar
Ian Lynagh committed
253 254 255
<line_prag1> $decdigit+                 { setLine line_prag1a }
<line_prag1a> \" [$graphic \ ]* \"      { setFile line_prag1b }
<line_prag1b> .*                        { pop }
256 257 258

-- Haskell-style line pragmas, of the form
--    {-# LINE <line> "<file>" #-}
Ian Lynagh's avatar
Ian Lynagh committed
259 260 261
<line_prag2> $decdigit+                 { setLine line_prag2a }
<line_prag2a> \" [$graphic \ ]* \"      { setFile line_prag2b }
<line_prag2b> "#-}"|"-}"                { pop }
262 263
   -- NOTE: accept -} at the end of a LINE pragma, for compatibility
   -- with older versions of GHC which generated these.
264

265
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
266
  "{-#" $whitechar* $pragmachar+
267 268 269 270 271
        $whitechar+ $pragmachar+ / { known_pragma twoWordPrags }
                                 { dispatch_pragmas twoWordPrags }

  "{-#" $whitechar* $pragmachar+ / { known_pragma oneWordPrags }
                                 { dispatch_pragmas oneWordPrags }
272

273
  -- We ignore all these pragmas, but don't generate a warning for them
274 275
  "{-#" $whitechar* $pragmachar+ / { known_pragma ignoredPrags }
                                 { dispatch_pragmas ignoredPrags }
276 277

  -- ToDo: should only be valid inside a pragma:
Ian Lynagh's avatar
Ian Lynagh committed
278
  "#-}"                          { endPrag }
279 280
}

281
<option_prags> {
282 283 284
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
                                   { dispatch_pragmas fileHeaderPrags }

Ian Lynagh's avatar
Ian Lynagh committed
285
  "-- #"                           { multiline_doc_comment }
David Waern's avatar
David Waern committed
286 287
}

288 289
<0> {
  -- In the "0" mode we ignore these pragmas
290
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
291 292 293
                     { nested_comment lexToken }
}

David Waern's avatar
David Waern committed
294
<0> {
Jedai's avatar
Jedai committed
295
  "-- #" .* { lineCommentToken }
296
}
297

298
<0,option_prags> {
299 300
  "{-#"  { warnThen Opt_WarnUnrecognisedPragmas (text "Unrecognised pragma")
                    (nested_comment lexToken) }
301 302
}

303 304
-- '0' state: ordinary lexemes

305 306
-- Haddock comments

307
<0,option_prags> {
David Waern's avatar
David Waern committed
308 309
  "-- " $docsym      / { ifExtension haddockEnabled } { multiline_doc_comment }
  "{-" \ ? $docsym   / { ifExtension haddockEnabled } { nested_doc_comment }
310 311
}

312 313
-- "special" symbols

314
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
315 316
  "[:" / { ifExtension parrEnabled }    { token ITopabrack }
  ":]" / { ifExtension parrEnabled }    { token ITcpabrack }
317
}
Ian Lynagh's avatar
Ian Lynagh committed
318

319
<0> {
320 321 322
  "[|"        / { ifExtension thEnabled } { token ITopenExpQuote }
  "[||"       / { ifExtension thEnabled } { token ITopenTExpQuote }
  "[e|"       / { ifExtension thEnabled } { token ITopenExpQuote }
323
  "[e||"      / { ifExtension thEnabled } { token ITopenTExpQuote }
324 325 326 327 328 329 330 331 332
  "[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 }
333

334 335 336 337
-- For backward compatibility, accept the old dollar syntax
  "[$" @varid "|"  / { ifExtension qqEnabled }
                     { lex_quasiquote_tok }

338
  "[" @varid "|"  / { ifExtension qqEnabled }
339
                     { lex_quasiquote_tok }
340 341 342 343

  -- qualified quasi-quote (#5555)
  "[" @qual @varid "|"  / { ifExtension qqEnabled }
                          { lex_qquasiquote_tok }
344 345
}

346
<0> {
ross's avatar
ross committed
347
  "(|" / { ifExtension arrowsEnabled `alexAndPred` notFollowedBySymbol }
Ian Lynagh's avatar
Ian Lynagh committed
348
                                        { special IToparenbar }
349 350 351
  "|)" / { ifExtension arrowsEnabled }  { special ITcparenbar }
}

352
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
353
  \? @varid / { ifExtension ipEnabled } { skip_one_varid ITdupipvarid }
354 355
}

356
<0> {
357
  "(#" / { ifExtension unboxedTuplesEnabled }
Ian Lynagh's avatar
Ian Lynagh committed
358 359 360 361 362
         { token IToubxparen }
  "#)" / { ifExtension unboxedTuplesEnabled }
         { token ITcubxparen }
}

363
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
364 365 366 367 368 369 370 371 372 373
  \(                                    { special IToparen }
  \)                                    { special ITcparen }
  \[                                    { special ITobrack }
  \]                                    { special ITcbrack }
  \,                                    { special ITcomma }
  \;                                    { special ITsemi }
  \`                                    { special ITbackquote }

  \{                                    { open_brace }
  \}                                    { close_brace }
374 375
}

376
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
377 378 379
  @qual @varid                  { idtoken qvarid }
  @qual @conid                  { idtoken qconid }
  @varid                        { varid }
380
  @conid                        { idtoken conid }
381 382
}

383
<0> {
384 385 386
  @qual @varid "#"+ / { ifExtension magicHashEnabled } { idtoken qvarid }
  @qual @conid "#"+ / { ifExtension magicHashEnabled } { idtoken qconid }
  @varid "#"+       / { ifExtension magicHashEnabled } { varid }
387
  @conid "#"+       / { ifExtension magicHashEnabled } { idtoken conid }
388 389
}

390 391
-- ToDo: - move `var` and (sym) into lexical syntax?
--       - remove backquote from $special?
392
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
393 394
  @qual @varsym                                    { idtoken qvarsym }
  @qual @consym                                    { idtoken qconsym }
395 396
  @varsym                                          { varsym }
  @consym                                          { consym }
397 398
}

399 400
-- For the normal boxed literals we need to be careful
-- when trying to be close to Haskell98
401
<0> {
402
  -- Normal integral literals (:: Num a => a, from Integer)
403 404 405 406 407 408
  @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 }
  @negative @decimal           / { ifExtension negativeLiteralsEnabled } { tok_num negative 1 1 decimal }
  @negative 0[oO] @octal       / { ifExtension negativeLiteralsEnabled } { tok_num negative 3 3 octal }
  @negative 0[xX] @hexadecimal / { ifExtension negativeLiteralsEnabled } { tok_num negative 3 3 hexadecimal }
409 410

  -- Normal rational literals (:: Fractional a => a, from Rational)
411 412
  @floating_point                                                        { strtoken tok_float }
  @negative @floating_point    / { ifExtension negativeLiteralsEnabled } { strtoken tok_float }
413 414
}

415
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
416
  -- Unboxed ints (:: Int#) and words (:: Word#)
417 418
  -- It's simpler (and faster?) to give separate cases to the negatives,
  -- especially considering octal/hexadecimal prefixes.
Ian Lynagh's avatar
Ian Lynagh committed
419 420 421 422 423
  @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 }
424
  @negative 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 hexadecimal }
425

Ian Lynagh's avatar
Ian Lynagh committed
426 427 428 429
  @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 }

430 431
  -- Unboxed floats and doubles (:: Float#, :: Double#)
  -- prim_{float,double} work with signed literals
432 433
  @signed @floating_point \# / { ifExtension magicHashEnabled } { init_strtoken 1 tok_primfloat }
  @signed @floating_point \# \# / { ifExtension magicHashEnabled } { init_strtoken 2 tok_primdouble }
434 435 436 437 438 439
}

-- 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.
440
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
441 442
  \'                            { lex_char_tok }
  \"                            { lex_string_tok }
443 444 445 446 447 448
}

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

449
data Token
Ian Lynagh's avatar
Ian Lynagh committed
450
  = ITas                        -- Haskell keywords
451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474
  | 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
475
  | ITforall                    -- GHC extension keywords
476 477 478 479 480
  | ITforeign
  | ITexport
  | ITlabel
  | ITdynamic
  | ITsafe
481
  | ITinterruptible
482 483 484
  | ITunsafe
  | ITstdcallconv
  | ITccallconv
485
  | ITcapiconv
486
  | ITprimcallconv
thoughtpolice's avatar
thoughtpolice committed
487
  | ITjavascriptcallconv
488
  | ITmdo
489
  | ITfamily
490
  | ITrole
491 492 493
  | ITgroup
  | ITby
  | ITusing
cactus's avatar
cactus committed
494
  | ITpattern
495

Ian Lynagh's avatar
Ian Lynagh committed
496
  -- Pragmas
497
  | ITinline_prag InlineSpec RuleMatchInfo
Ian Lynagh's avatar
Ian Lynagh committed
498 499
  | ITspec_prag                 -- SPECIALISE
  | ITspec_inline_prag Bool     -- SPECIALISE INLINE (or NOINLINE)
500 501
  | ITsource_prag
  | ITrules_prag
Ian Lynagh's avatar
Ian Lynagh committed
502
  | ITwarning_prag
503 504 505
  | ITdeprecated_prag
  | ITline_prag
  | ITscc_prag
andy@galois.com's avatar
andy@galois.com committed
506
  | ITgenerated_prag
507
  | ITcore_prag                 -- hdaume: core annotations
508
  | ITunpack_prag
509
  | ITnounpack_prag
510
  | ITann_prag
511
  | ITclose_prag
512 513 514
  | IToptions_prag String
  | ITinclude_prag String
  | ITlanguage_prag
515 516
  | ITvect_prag
  | ITvect_scalar_prag
517
  | ITnovect_prag
518
  | ITminimal_prag
519
  | ITctype
520

Ian Lynagh's avatar
Ian Lynagh committed
521
  | ITdotdot                    -- reserved symbols
522 523 524 525
  | ITcolon
  | ITdcolon
  | ITequal
  | ITlam
526
  | ITlcase
527 528 529 530 531
  | ITvbar
  | ITlarrow
  | ITrarrow
  | ITat
  | ITtilde
batterseapower's avatar
batterseapower committed
532
  | ITtildehsh
533 534 535 536 537 538
  | ITdarrow
  | ITminus
  | ITbang
  | ITstar
  | ITdot

Ian Lynagh's avatar
Ian Lynagh committed
539
  | ITbiglam                    -- GHC-extension symbols
540

Ian Lynagh's avatar
Ian Lynagh committed
541
  | ITocurly                    -- special symbols
542 543 544 545
  | ITccurly
  | ITvocurly
  | ITvccurly
  | ITobrack
Ian Lynagh's avatar
Ian Lynagh committed
546 547
  | ITopabrack                  -- [:, for parallel arrays with -XParallelArrays
  | ITcpabrack                  -- :], for parallel arrays with -XParallelArrays
548 549 550 551 552 553 554 555 556
  | ITcbrack
  | IToparen
  | ITcparen
  | IToubxparen
  | ITcubxparen
  | ITsemi
  | ITcomma
  | ITunderscore
  | ITbackquote
dreixel's avatar
dreixel committed
557
  | ITsimpleQuote               --  '
558

Ian Lynagh's avatar
Ian Lynagh committed
559
  | ITvarid   FastString        -- identifiers
560 561 562 563 564 565 566
  | ITconid   FastString
  | ITvarsym  FastString
  | ITconsym  FastString
  | ITqvarid  (FastString,FastString)
  | ITqconid  (FastString,FastString)
  | ITqvarsym (FastString,FastString)
  | ITqconsym (FastString,FastString)
567 568
  | ITprefixqvarsym (FastString,FastString)
  | ITprefixqconsym (FastString,FastString)
569

Ian Lynagh's avatar
Ian Lynagh committed
570
  | ITdupipvarid   FastString   -- GHC extension: implicit param: ?x
571 572 573 574

  | ITchar       Char
  | ITstring     FastString
  | ITinteger    Integer
575
  | ITrational   FractionalLit
576 577

  | ITprimchar   Char
578
  | ITprimstring ByteString
579
  | ITprimint    Integer
Ian Lynagh's avatar
Ian Lynagh committed
580
  | ITprimword   Integer
581 582
  | ITprimfloat  FractionalLit
  | ITprimdouble FractionalLit
583

584
  -- Template Haskell extension tokens
Ian Lynagh's avatar
Ian Lynagh committed
585 586 587 588 589
  | ITopenExpQuote              --  [| or [e|
  | ITopenPatQuote              --  [p|
  | ITopenDecQuote              --  [d|
  | ITopenTypQuote              --  [t|
  | ITcloseQuote                --  |]
590 591
  | ITopenTExpQuote             --  [||
  | ITcloseTExpQuote            --  ||]
Ian Lynagh's avatar
Ian Lynagh committed
592 593
  | ITidEscape   FastString     --  $x
  | ITparenEscape               --  $(
594 595
  | ITidTyEscape   FastString   --  $$x
  | ITparenTyEscape             --  $$(
Ian Lynagh's avatar
Ian Lynagh committed
596
  | ITtyQuote                   --  ''
597 598 599 600 601 602 603 604
  | 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 |]
605 606 607 608

  -- Arrow notation extension
  | ITproc
  | ITrec
Ian Lynagh's avatar
Ian Lynagh committed
609 610 611 612 613 614
  | IToparenbar                 --  (|
  | ITcparenbar                 --  |)
  | ITlarrowtail                --  -<
  | ITrarrowtail                --  >-
  | ITLarrowtail                --  -<<
  | ITRarrowtail                --  >>-
615

Ian Lynagh's avatar
Ian Lynagh committed
616 617
  | ITunknown String            -- Used when the lexer can't make sense of it
  | ITeof                       -- end of file token
618 619 620 621 622 623 624

  -- 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
625
  | ITdocOptionsOld   String     -- doc options declared "-- # ..."-style
Jedai's avatar
Jedai committed
626 627
  | ITlineComment     String     -- comment starting by "--"
  | ITblockComment    String     -- comment in {- -}
628

629
  deriving Show
630 631 632 633 634 635 636 637

-- 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)
--
638
reservedWordsFM :: UniqFM (Token, ExtsBitmap)
639
reservedWordsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666
    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 ),

667 668 669
         ( "forall",         ITforall,        xbit ExplicitForallBit .|.
                                              xbit InRulePragBit),
         ( "mdo",            ITmdo,           xbit RecursiveDoBit),
670 671 672
             -- See Note [Lexing type pseudo-keywords]
         ( "family",         ITfamily,        0 ),
         ( "role",           ITrole,          0 ),
673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694
         ( "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)
695 696
     ]

697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712
{-----------------------------------
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.
-------------------------------------}
713

714
reservedSymsFM :: UniqFM (Token, ExtsBitmap -> Bool)
715
reservedSymsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
716 717 718 719 720 721 722 723 724 725 726 727
    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)
728
       ,("~#",  ITtildehsh, magicHashEnabled)
Ian Lynagh's avatar
Ian Lynagh committed
729 730 731 732 733
       ,("=>",  ITdarrow,   always)
       ,("-",   ITminus,    always)
       ,("!",   ITbang,     always)

        -- For data T (a::*) = MkT
734
       ,("*", ITstar, always) -- \i -> kindSigsEnabled i || tyFamEnabled i)
Ian Lynagh's avatar
Ian Lynagh committed
735
        -- For 'forall a . t'
736
       ,(".", ITdot,  always) -- \i -> explicitForallEnabled i || inRulePrag i)
Ian Lynagh's avatar
Ian Lynagh committed
737 738 739 740 741

       ,("-<",  ITlarrowtail, arrowsEnabled)
       ,(">-",  ITrarrowtail, arrowsEnabled)
       ,("-<<", ITLarrowtail, arrowsEnabled)
       ,(">>-", ITRarrowtail, arrowsEnabled)
742

Ian Lynagh's avatar
Ian Lynagh committed
743 744
       ,("∷",   ITdcolon, unicodeSyntaxEnabled)
       ,("⇒",   ITdarrow, unicodeSyntaxEnabled)
745
       ,("∀",   ITforall, unicodeSyntaxEnabled)
Ian Lynagh's avatar
Ian Lynagh committed
746 747
       ,("→",   ITrarrow, unicodeSyntaxEnabled)
       ,("←",   ITlarrow, unicodeSyntaxEnabled)
748 749 750 751 752 753 754 755

       ,("⤙",   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
756 757 758
        -- 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).
759 760 761 762 763
       ]

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

Ian Lynagh's avatar
Ian Lynagh committed
764
type Action = RealSrcSpan -> StringBuffer -> Int -> P (RealLocated Token)
765

766
special :: Token -> Action
twanvl's avatar
twanvl committed
767
special tok span _buf _len = return (L span tok)
768

769
token, layout_token :: Token -> Action
twanvl's avatar
twanvl committed
770 771
token t span _buf _len = return (L span t)
layout_token t span _buf _len = pushLexState layout >> return (L span t)
772

773 774
idtoken :: (StringBuffer -> Int -> Token) -> Action
idtoken f span buf len = return (L span $! (f buf len))
775

776
skip_one_varid :: (FastString -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
777
skip_one_varid f span buf len
778
  = return (L span $! f (lexemeToFastString (stepOn buf) (len-1)))
779

780 781 782 783
skip_two_varid :: (FastString -> Token) -> Action
skip_two_varid f span buf len
  = return (L span $! f (lexemeToFastString (stepOn (stepOn buf)) (len-2)))

784
strtoken :: (String -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
785
strtoken f span buf len =
786
  return (L span $! (f $! lexemeToString buf len))
787

788
init_strtoken :: Int -> (String -> Token) -> Action
789
-- like strtoken, but drops the last N character(s)
Ian Lynagh's avatar
Ian Lynagh committed
790
init_strtoken drop f span buf len =
791
  return (L span $! (f $! lexemeToString buf (len-drop)))
792 793

begin :: Int -> Action
794
begin code _span _str _len = do pushLexState code; lexToken
795 796

pop :: Action
Ian Lynagh's avatar
Ian Lynagh committed
797 798
pop _span _buf _len = do _ <- popLexState
                         lexToken
799

800 801 802 803 804 805 806 807 808 809 810
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
811
              else failSpanMsgP (RealSrcSpan span) (text "Missing block")
812

813
pop_and :: Action -> Action
Ian Lynagh's avatar
Ian Lynagh committed
814 815
pop_and act span buf len = do _ <- popLexState
                              act span buf len
816

817
{-# INLINE nextCharIs #-}
Ian Lynagh's avatar
Ian Lynagh committed
818
nextCharIs :: StringBuffer -> (Char -> Bool) -> Bool
819 820
nextCharIs buf p = not (atEnd buf) && p (currentChar buf)

821 822 823 824
{-# INLINE nextCharIsNot #-}
nextCharIsNot :: StringBuffer -> (Char -> Bool) -> Bool
nextCharIsNot buf p = not (nextCharIs buf p)

825
notFollowedBy :: Char -> AlexAccPred ExtsBitmap
Ian Lynagh's avatar
Ian Lynagh committed
826
notFollowedBy char _ _ _ (AI _ buf)
827
  = nextCharIsNot buf (== char)
828

829
notFollowedBySymbol :: AlexAccPred ExtsBitmap
830
notFollowedBySymbol _ _ _ (AI _ buf)
831
  = nextCharIsNot buf (`elem` "!#$%&*+./<=>?@\\^|-~")
ross's avatar
ross committed
832

833 834 835 836 837
-- 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.
838
isNormalComment :: AlexAccPred ExtsBitmap
839
isNormalComment bits _ _ (AI _ buf)
840
  | haddockEnabled bits = notFollowedByDocOrPragma
841
  | otherwise           = nextCharIsNot buf (== '#')
842 843
  where
    notFollowedByDocOrPragma
844
       = afterOptionalSpace buf (\b -> nextCharIsNot b (`elem` "|^*$#"))
845

846 847 848 849 850
afterOptionalSpace :: StringBuffer -> (StringBuffer -> Bool) -> Bool
afterOptionalSpace buf p
    = if nextCharIs buf (== ' ')
      then p (snd (nextChar buf))
      else p buf
851

852
atEOL :: AlexAccPred ExtsBitmap
853
atEOL _ _ _ (AI _ buf) = atEnd buf || currentChar buf == '\n'
854

855
ifExtension :: (ExtsBitmap -> Bool) -> AlexAccPred ExtsBitmap
856 857
ifExtension pred bits _ _ _ = pred bits

858 859 860
multiline_doc_comment :: Action
multiline_doc_comment span buf _len = withLexedDocType (worker "")
  where
861
    worker commentAcc input docType oneLine = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
862
      Just ('\n', input')
863 864 865 866 867 868
        | 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
869

870 871
    checkIfCommentLine input = check (dropNonNewlineSpace input)
      where
872 873 874
        check input = case alexGetChar' input of
          Just ('-', input) -> case alexGetChar' input of
            Just ('-', input) -> case alexGetChar' input of
875 876 877 878 879
              Just (c, _) | c /= '-' -> Just input
              _ -> Nothing
            _ -> Nothing
          _ -> Nothing

880
        dropNonNewlineSpace input = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
881
          Just (c, input')
882 883 884 885
            | isSpace c && c /= '\n' -> dropNonNewlineSpace input'
            | otherwise -> input
          Nothing -> input

Jedai's avatar
Jedai committed
886 887 888 889 890
lineCommentToken :: Action
lineCommentToken span buf len = do
  b <- extension rawTokenStreamEnabled
  if b then strtoken ITlineComment span buf len else lexToken

891 892 893 894
{-
  nested comments require traversing by hand, they can't be parsed
  using regular expressions.
-}
Ian Lynagh's avatar
Ian Lynagh committed
895
nested_comment :: P (RealLocated Token) -> Action
896
nested_comment cont span _str _len = do
897
  input <- getInput
Jedai's avatar
Jedai committed
898
  go "" (1::Int) input
899
  where
Jedai's avatar
Jedai committed
900 901 902 903 904
    go commentAcc 0 input = do setInput input
                               b <- extension rawTokenStreamEnabled
                               if b
                                 then docCommentEnd input commentAcc ITblockComment _str span
                                 else cont
905
    go commentAcc n input = case alexGetChar' input of
906
      Nothing -> errBrace input span
907
      Just ('-',input) -> case alexGetChar' input of
908
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
909 910
        Just ('\125',input) -> go commentAcc (n-1) input
        Just (_,_)          -> go ('-':commentAcc) n input
911
      Just ('\123',input) -> case alexGetChar' input of
912
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
913 914 915
        Just ('-',input) -> go ('-':'\123':commentAcc) (n+1) input
        Just (_,_)       -> go ('\123':commentAcc) n input
      Just (c,input) -> go (c:commentAcc) n input
916 917 918 919

nested_doc_comment :: Action
nested_doc_comment span buf _len = withLexedDocType (go "")
  where
920
    go commentAcc input docType _ = case alexGetChar' input of
921
      Nothing -> errBrace input span
922
      Just ('-',input) -> case alexGetChar' input of
923
        Nothing -> errBrace input span
twanvl's avatar
twanvl committed
924
        Just ('\125',input) ->
925
          docCommentEnd input commentAcc docType buf span
twanvl's avatar
twanvl committed
926
        Just (_,_) -> go ('-':commentAcc) input docType False
927
      Just ('\123', input) -> case alexGetChar' input of
928 929 930 931 932
        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
933
        Just (_,_) -> go ('\123':commentAcc) input docType False
934 935
      Just (c,input) -> go (c:commentAcc) input docType False

Ian Lynagh's avatar
Ian Lynagh committed
936 937
withLexedDocType :: (AlexInput -> (String -> Token) -> Bool -> P (RealLocated Token))
                 -> P (RealLocated Token)
938
withLexedDocType lexDocComment = do
939
  input@(AI _ buf) <- getInput
940 941 942 943
  case prevChar buf ' ' of
    '|' -> lexDocComment input ITdocCommentNext False
    '^' -> lexDocComment input ITdocCommentPrev False
    '$' -> lexDocComment input ITdocCommentNamed False
David Waern's avatar
David Waern committed
944 945
    '*' -> lexDocSection 1 input
    '#' -> lexDocComment input ITdocOptionsOld False
Ian Lynagh's avatar
Ian Lynagh committed
946
    _ -> panic "withLexedDocType: Bad doc type"
Ian Lynagh's avatar
Ian Lynagh committed
947
 where
948
    lexDocSection n input = case alexGetChar' input of
949
      Just ('*', input) -> lexDocSection (n+1) input
twanvl's avatar
twanvl committed
950
      Just (_,   _)     -> lexDocComment input (ITdocSection n) True
951 952
      Nothing -> do setInput input; lexToken -- eof reached, lex it normally

953 954 955
-- RULES pragmas turn on the forall and '.' keywords, and we turn them
-- off again at the end of the pragma.
rulePrag :: Action
956
rulePrag span _buf _len = do
957
  setExts (.|. xbit InRulePragBit)
958 959 960
  return (L span ITrules_prag)

endPrag :: Action
961
endPrag span _buf _len = do
962
  setExts (.&. complement (xbit InRulePragBit))
963 964
  return (L span ITclose_prag)

965 966 967 968
-- 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
969
-- than what was lexed by Alex, and the lexToken function doesn't know this, so
970
-- it writes the wrong token length to the parser state. This function is
Ian Lynagh's avatar
Ian Lynagh committed
971
-- called afterwards, so it can just update the state.
972 973

docCommentEnd :: AlexInput -> String -> (String -> Token) -> StringBuffer ->
Ian Lynagh's avatar
Ian Lynagh committed
974
                 RealSrcSpan -> P (RealLocated Token)
975 976
docCommentEnd input commentAcc docType buf span = do
  setInput input
977
  let (AI loc nextBuf) = input
978
      comment = reverse commentAcc
Ian Lynagh's avatar
Ian Lynagh committed
979
      span' = mkRealSrcSpan (realSrcSpanStart span) loc
980
      last_len = byteDiff buf nextBuf
Ian Lynagh's avatar
Ian Lynagh committed
981

982
  span `seq` setLastToken span' last_len
983
  return (L span' (docType comment))
Ian Lynagh's avatar
Ian Lynagh committed
984

Ian Lynagh's avatar
Ian Lynagh committed
985 986
errBrace :: AlexInput -> RealSrcSpan -> P a
errBrace (AI end _) span = failLocMsgP (realSrcSpanStart span) end "unterminated `{-'"
Ian Lynagh's avatar
Ian Lynagh committed
987

988
open_brace, close_brace :: Action
Ian Lynagh's avatar
Ian Lynagh committed
989
open_brace span _str _len = do
990 991
  ctx <- getContext
  setContext (NoLayout:ctx)
992
  return (L span ITocurly)
Ian Lynagh's avatar
Ian Lynagh committed
993
close_brace span _str _len = do
994
  popContext
995
  return (L span ITccurly)
996

Ian Lynagh's avatar
Ian Lynagh committed
997
qvarid, qconid :: StringBuffer -> Int -> Token
998 999
qvarid buf len = ITqvarid $! splitQualName buf len False
qconid buf len = ITqconid $! splitQualName buf len False
1000

1001
splitQualName :: StringBuffer -> Int -> Bool -> (FastString,FastString)
1002 1003 1004
-- 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.
1005
splitQualName orig_buf len parens = split orig_buf orig_buf
1006
  where