Lexer.x 93 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,
ian@well-typed.com's avatar
ian@well-typed.com committed
62
   explicitNamespacesEnabled, sccProfilingOn, hpcEnabled,
Jedai's avatar
Jedai committed
63 64
   addWarning,
   lexTokenStream
65 66
  ) where

67 68
import Bag
import ErrUtils
69 70 71 72 73
import Outputable
import StringBuffer
import FastString
import SrcLoc
import UniqFM
74
import DynFlags
75
import Module
76
import Ctype
Ian Lynagh's avatar
Ian Lynagh committed
77 78
import BasicTypes       ( InlineSpec(..), RuleMatchInfo(..), FractionalLit(..) )
import Util             ( readRational )
79

80
import Control.Monad
Simon Marlow's avatar
Simon Marlow committed
81
import Data.Bits
82
import Data.ByteString (ByteString)
83
import Data.Char
84
import Data.List
Ian Lynagh's avatar
Ian Lynagh committed
85
import Data.Maybe
86 87
import Data.Map (Map)
import qualified Data.Map as Map
Simon Marlow's avatar
Simon Marlow committed
88
import Data.Ratio
89
import Data.Word
90 91
}

92
$unispace    = \x05 -- Trick Alex into handling Unicode. See alexGetChar.
93
$whitechar   = [\ \n\r\f\v $unispace]
94
$white_no_nl = $whitechar # \n
95
$tab         = \t
96 97

$ascdigit  = 0-9
98
$unidigit  = \x03 -- Trick Alex into handling Unicode. See alexGetChar.
99
$decdigit  = $ascdigit -- for now, should really be $digit (ToDo)
100 101 102
$digit     = [$ascdigit $unidigit]

$special   = [\(\)\,\;\[\]\`\{\}]
103
$ascsymbol = [\!\#\$\%\&\*\+\.\/\<\=\>\?\@\\\^\|\-\~]
104
$unisymbol = \x04 -- Trick Alex into handling Unicode. See alexGetChar.
105 106
$symbol    = [$ascsymbol $unisymbol] # [$special \_\:\"\']

107
$unilarge  = \x01 -- Trick Alex into handling Unicode. See alexGetChar.
108
$asclarge  = [A-Z]
109 110
$large     = [$asclarge $unilarge]

111
$unismall  = \x02 -- Trick Alex into handling Unicode. See alexGetChar.
112
$ascsmall  = [a-z]
113 114
$small     = [$ascsmall $unismall \_]

115
$unigraphic = \x06 -- Trick Alex into handling Unicode. See alexGetChar.
116
$graphic   = [$small $large $symbol $digit $special $unigraphic \:\"\']
117

Ian Lynagh's avatar
Ian Lynagh committed
118
$octit     = 0-7
119
$hexit     = [$decdigit A-F a-f]
120 121 122 123
$symchar   = [$symbol \:]
$nl        = [\n\r]
$idchar    = [$small $large $digit \']

124 125
$pragmachar = [$small $large $digit]

126 127
$docsym    = [\| \^ \* \$]

128 129 130 131 132 133
@varid     = $small $idchar*
@conid     = $large $idchar*

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

134
@decimal     = $decdigit+
135 136 137 138 139 140 141 142 143
@octal       = $octit+
@hexadecimal = $hexit+
@exponent    = [eE] [\-\+]? @decimal

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

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

144 145 146 147 148
-- normal signed numerical literals can only be explicitly negative,
-- not explicitly positive (contrast @exponent)
@negative = \-
@signed = @negative ?

149 150 151
haskell :-

-- everywhere: skip whitespace and comments
Ian Lynagh's avatar
Ian Lynagh committed
152
$white_no_nl+ ;
153
$tab+         { warn Opt_WarnTabs (text "Tab character") }
154 155 156 157 158

-- 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
159 160 161 162
-- are). We also rule out nested Haddock comments, if the -haddock flag is
-- set.

"{-" / { isNormalComment } { nested_comment lexToken }
163 164 165 166

-- 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.
167 168

-- Since Haddock comments aren't valid in every state, we need to rule them
Ian Lynagh's avatar
Ian Lynagh committed
169
-- out here.
170 171 172 173 174 175 176

-- 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
177 178
"-- " ~[$docsym \#] .* { lineCommentToken }
"--" [^$symbol : \ ] .* { lineCommentToken }
179 180 181

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

Jedai's avatar
Jedai committed
182
"-- " [$docsym \#] .* / { ifExtension (not . haddockEnabled) } { lineCommentToken }
183 184 185 186 187 188 189

-- 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
190
"---"\-* [^$symbol :] .* { lineCommentToken }
191 192 193 194

-- 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
195
"--"\-* / { atEOL } { lineCommentToken }
196 197 198 199

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

Jedai's avatar
Jedai committed
200
"-- " / { atEOL } { lineCommentToken }
201 202 203 204 205 206 207 208 209 210 211

-- '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
212 213 214 215 216
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
  ^\# pragma .* \n                      ; -- GCC 3.3 CPP generated, apparently
  ^\# \! .* \n                          ; -- #!, for scripts
  ()                                    { do_bol }
217 218 219 220 221
}

-- 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.
222
<layout, layout_do, layout_if> {
Ian Lynagh's avatar
Ian Lynagh committed
223 224 225 226
  \{ / { notFollowedBy '-' }            { hopefully_open_brace }
        -- we might encounter {-# here, but {- has been handled already
  \n                                    ;
  ^\# (line)?                           { begin line_prag1 }
227 228
}

229 230 231 232 233 234
-- after an 'if', a vertical bar starts a layout context for MultiWayIf
<layout_if> {
  \| / { notFollowedBySymbol }          { new_layout_context True ITvbar }
  ()                                    { pop }
}

235
-- do is treated in a subtly different way, see new_layout_context
236 237
<layout>    ()                          { new_layout_context True  ITvocurly }
<layout_do> ()                          { new_layout_context False ITvocurly }
238 239 240 241

-- 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
242
<layout_left>  ()                       { do_layout_left }
243

Ian Lynagh's avatar
Ian Lynagh committed
244
<0,option_prags> \n                     { begin bol }
245

246 247
"{-#" $whitechar* $pragmachar+ / { known_pragma linePrags }
                                { dispatch_pragmas linePrags }
248 249 250

-- single-line line pragmas, of the form
--    # <line> "<file>" <extra-stuff> \n
Ian Lynagh's avatar
Ian Lynagh committed
251 252 253
<line_prag1> $decdigit+                 { setLine line_prag1a }
<line_prag1a> \" [$graphic \ ]* \"      { setFile line_prag1b }
<line_prag1b> .*                        { pop }
254 255 256

-- Haskell-style line pragmas, of the form
--    {-# LINE <line> "<file>" #-}
Ian Lynagh's avatar
Ian Lynagh committed
257 258 259
<line_prag2> $decdigit+                 { setLine line_prag2a }
<line_prag2a> \" [$graphic \ ]* \"      { setFile line_prag2b }
<line_prag2b> "#-}"|"-}"                { pop }
260 261
   -- NOTE: accept -} at the end of a LINE pragma, for compatibility
   -- with older versions of GHC which generated these.
262

263
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
264
  "{-#" $whitechar* $pragmachar+
265 266 267 268 269
        $whitechar+ $pragmachar+ / { known_pragma twoWordPrags }
                                 { dispatch_pragmas twoWordPrags }

  "{-#" $whitechar* $pragmachar+ / { known_pragma oneWordPrags }
                                 { dispatch_pragmas oneWordPrags }
270

271
  -- We ignore all these pragmas, but don't generate a warning for them
272 273
  "{-#" $whitechar* $pragmachar+ / { known_pragma ignoredPrags }
                                 { dispatch_pragmas ignoredPrags }
274 275

  -- ToDo: should only be valid inside a pragma:
Ian Lynagh's avatar
Ian Lynagh committed
276
  "#-}"                          { endPrag }
277 278
}

279
<option_prags> {
280 281 282
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
                                   { dispatch_pragmas fileHeaderPrags }

Ian Lynagh's avatar
Ian Lynagh committed
283
  "-- #"                           { multiline_doc_comment }
David Waern's avatar
David Waern committed
284 285
}

286 287
<0> {
  -- In the "0" mode we ignore these pragmas
288
  "{-#"  $whitechar* $pragmachar+ / { known_pragma fileHeaderPrags }
289 290 291
                     { nested_comment lexToken }
}

David Waern's avatar
David Waern committed
292
<0> {
Jedai's avatar
Jedai committed
293
  "-- #" .* { lineCommentToken }
294
}
295

296
<0,option_prags> {
297 298
  "{-#"  { warnThen Opt_WarnUnrecognisedPragmas (text "Unrecognised pragma")
                    (nested_comment lexToken) }
299 300
}

301 302
-- '0' state: ordinary lexemes

303 304
-- Haddock comments

305
<0,option_prags> {
David Waern's avatar
David Waern committed
306 307
  "-- " $docsym      / { ifExtension haddockEnabled } { multiline_doc_comment }
  "{-" \ ? $docsym   / { ifExtension haddockEnabled } { nested_doc_comment }
308 309
}

310 311
-- "special" symbols

312
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
313 314
  "[:" / { ifExtension parrEnabled }    { token ITopabrack }
  ":]" / { ifExtension parrEnabled }    { token ITcpabrack }
315
}
Ian Lynagh's avatar
Ian Lynagh committed
316

317
<0> {
318 319 320 321 322 323 324 325 326 327 328 329
  "[|"        / { ifExtension thEnabled } { token ITopenExpQuote }
  "[||"       / { ifExtension thEnabled } { token ITopenTExpQuote }
  "[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 }
  "||]"       / { 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 }
330

331 332 333 334
-- For backward compatibility, accept the old dollar syntax
  "[$" @varid "|"  / { ifExtension qqEnabled }
                     { lex_quasiquote_tok }

335
  "[" @varid "|"  / { ifExtension qqEnabled }
336
                     { lex_quasiquote_tok }
337 338 339 340

  -- qualified quasi-quote (#5555)
  "[" @qual @varid "|"  / { ifExtension qqEnabled }
                          { lex_qquasiquote_tok }
341 342
}

343
<0> {
ross's avatar
ross committed
344
  "(|" / { ifExtension arrowsEnabled `alexAndPred` notFollowedBySymbol }
Ian Lynagh's avatar
Ian Lynagh committed
345
                                        { special IToparenbar }
346 347 348
  "|)" / { ifExtension arrowsEnabled }  { special ITcparenbar }
}

349
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
350
  \? @varid / { ifExtension ipEnabled } { skip_one_varid ITdupipvarid }
351 352
}

353
<0> {
354
  "(#" / { ifExtension unboxedTuplesEnabled }
Ian Lynagh's avatar
Ian Lynagh committed
355 356 357 358 359
         { token IToubxparen }
  "#)" / { ifExtension unboxedTuplesEnabled }
         { token ITcubxparen }
}

360
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
361 362 363 364 365 366 367 368 369 370
  \(                                    { special IToparen }
  \)                                    { special ITcparen }
  \[                                    { special ITobrack }
  \]                                    { special ITcbrack }
  \,                                    { special ITcomma }
  \;                                    { special ITsemi }
  \`                                    { special ITbackquote }

  \{                                    { open_brace }
  \}                                    { close_brace }
371 372
}

373
<0,option_prags> {
Ian Lynagh's avatar
Ian Lynagh committed
374 375 376
  @qual @varid                  { idtoken qvarid }
  @qual @conid                  { idtoken qconid }
  @varid                        { varid }
377
  @conid                        { idtoken conid }
378 379
}

380
<0> {
381 382 383
  @qual @varid "#"+ / { ifExtension magicHashEnabled } { idtoken qvarid }
  @qual @conid "#"+ / { ifExtension magicHashEnabled } { idtoken qconid }
  @varid "#"+       / { ifExtension magicHashEnabled } { varid }
384
  @conid "#"+       / { ifExtension magicHashEnabled } { idtoken conid }
385 386
}

387 388
-- ToDo: - move `var` and (sym) into lexical syntax?
--       - remove backquote from $special?
389
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
390 391
  @qual @varsym                                    { idtoken qvarsym }
  @qual @consym                                    { idtoken qconsym }
392 393
  @varsym                                          { varsym }
  @consym                                          { consym }
394 395
}

396 397
-- For the normal boxed literals we need to be careful
-- when trying to be close to Haskell98
398
<0> {
399
  -- Normal integral literals (:: Num a => a, from Integer)
400 401 402 403 404 405
  @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 }
406 407

  -- Normal rational literals (:: Fractional a => a, from Rational)
408 409
  @floating_point                                                        { strtoken tok_float }
  @negative @floating_point    / { ifExtension negativeLiteralsEnabled } { strtoken tok_float }
410 411
}

412
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
413
  -- Unboxed ints (:: Int#) and words (:: Word#)
414 415
  -- It's simpler (and faster?) to give separate cases to the negatives,
  -- especially considering octal/hexadecimal prefixes.
Ian Lynagh's avatar
Ian Lynagh committed
416 417 418 419 420
  @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 }
421
  @negative 0[xX] @hexadecimal \# / { ifExtension magicHashEnabled } { tok_primint negative 3 4 hexadecimal }
422

Ian Lynagh's avatar
Ian Lynagh committed
423 424 425 426
  @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 }

427 428
  -- Unboxed floats and doubles (:: Float#, :: Double#)
  -- prim_{float,double} work with signed literals
429 430
  @signed @floating_point \# / { ifExtension magicHashEnabled } { init_strtoken 1 tok_primfloat }
  @signed @floating_point \# \# / { ifExtension magicHashEnabled } { init_strtoken 2 tok_primdouble }
431 432 433 434 435 436
}

-- 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.
437
<0> {
Ian Lynagh's avatar
Ian Lynagh committed
438 439
  \'                            { lex_char_tok }
  \"                            { lex_string_tok }
440 441 442 443 444 445
}

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

446
data Token
Ian Lynagh's avatar
Ian Lynagh committed
447
  = ITas                        -- Haskell keywords
448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470
  | 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
471
  | ITscc                       -- ToDo: remove (we use {-# SCC "..." #-} now)
472

Ian Lynagh's avatar
Ian Lynagh committed
473
  | ITforall                    -- GHC extension keywords
474 475 476 477 478
  | ITforeign
  | ITexport
  | ITlabel
  | ITdynamic
  | ITsafe
479
  | ITinterruptible
480 481 482
  | ITunsafe
  | ITstdcallconv
  | ITccallconv
483
  | ITcapiconv
484
  | ITprimcallconv
thoughtpolice's avatar
thoughtpolice committed
485
  | ITjavascriptcallconv
486
  | ITmdo
487
  | ITfamily
488
  | ITrole
489 490 491
  | ITgroup
  | ITby
  | ITusing
492

Ian Lynagh's avatar
Ian Lynagh committed
493
  -- Pragmas
494
  | ITinline_prag InlineSpec RuleMatchInfo
Ian Lynagh's avatar
Ian Lynagh committed
495 496
  | ITspec_prag                 -- SPECIALISE
  | ITspec_inline_prag Bool     -- SPECIALISE INLINE (or NOINLINE)
497 498
  | ITsource_prag
  | ITrules_prag
Ian Lynagh's avatar
Ian Lynagh committed
499
  | ITwarning_prag
500 501 502
  | ITdeprecated_prag
  | ITline_prag
  | ITscc_prag
andy@galois.com's avatar
andy@galois.com committed
503
  | ITgenerated_prag
504
  | ITcore_prag                 -- hdaume: core annotations
505
  | ITunpack_prag
506
  | ITnounpack_prag
507
  | ITann_prag
508
  | ITclose_prag
509 510 511
  | IToptions_prag String
  | ITinclude_prag String
  | ITlanguage_prag
512 513
  | ITvect_prag
  | ITvect_scalar_prag
514
  | ITnovect_prag
515
  | ITminimal_prag
516
  | ITctype
517

Ian Lynagh's avatar
Ian Lynagh committed
518
  | ITdotdot                    -- reserved symbols
519 520 521 522
  | ITcolon
  | ITdcolon
  | ITequal
  | ITlam
523
  | ITlcase
524 525 526 527 528
  | ITvbar
  | ITlarrow
  | ITrarrow
  | ITat
  | ITtilde
batterseapower's avatar
batterseapower committed
529
  | ITtildehsh
530 531 532 533 534 535
  | ITdarrow
  | ITminus
  | ITbang
  | ITstar
  | ITdot

Ian Lynagh's avatar
Ian Lynagh committed
536
  | ITbiglam                    -- GHC-extension symbols
537

Ian Lynagh's avatar
Ian Lynagh committed
538
  | ITocurly                    -- special symbols
539 540 541 542
  | ITccurly
  | ITvocurly
  | ITvccurly
  | ITobrack
Ian Lynagh's avatar
Ian Lynagh committed
543 544
  | ITopabrack                  -- [:, for parallel arrays with -XParallelArrays
  | ITcpabrack                  -- :], for parallel arrays with -XParallelArrays
545 546 547 548 549 550 551 552 553
  | ITcbrack
  | IToparen
  | ITcparen
  | IToubxparen
  | ITcubxparen
  | ITsemi
  | ITcomma
  | ITunderscore
  | ITbackquote
dreixel's avatar
dreixel committed
554
  | ITsimpleQuote               --  '
555

Ian Lynagh's avatar
Ian Lynagh committed
556
  | ITvarid   FastString        -- identifiers
557 558 559 560 561 562 563
  | ITconid   FastString
  | ITvarsym  FastString
  | ITconsym  FastString
  | ITqvarid  (FastString,FastString)
  | ITqconid  (FastString,FastString)
  | ITqvarsym (FastString,FastString)
  | ITqconsym (FastString,FastString)
564 565
  | ITprefixqvarsym (FastString,FastString)
  | ITprefixqconsym (FastString,FastString)
566

Ian Lynagh's avatar
Ian Lynagh committed
567
  | ITdupipvarid   FastString   -- GHC extension: implicit param: ?x
568 569 570 571

  | ITchar       Char
  | ITstring     FastString
  | ITinteger    Integer
572
  | ITrational   FractionalLit
573 574

  | ITprimchar   Char
575
  | ITprimstring ByteString
576
  | ITprimint    Integer
Ian Lynagh's avatar
Ian Lynagh committed
577
  | ITprimword   Integer
578 579
  | ITprimfloat  FractionalLit
  | ITprimdouble FractionalLit
580

581
  -- Template Haskell extension tokens
Ian Lynagh's avatar
Ian Lynagh committed
582 583 584 585 586
  | ITopenExpQuote              --  [| or [e|
  | ITopenPatQuote              --  [p|
  | ITopenDecQuote              --  [d|
  | ITopenTypQuote              --  [t|
  | ITcloseQuote                --  |]
587 588
  | ITopenTExpQuote             --  [||
  | ITcloseTExpQuote            --  ||]
Ian Lynagh's avatar
Ian Lynagh committed
589 590
  | ITidEscape   FastString     --  $x
  | ITparenEscape               --  $(
591 592
  | ITidTyEscape   FastString   --  $$x
  | ITparenTyEscape             --  $$(
Ian Lynagh's avatar
Ian Lynagh committed
593
  | ITtyQuote                   --  ''
594 595 596 597 598 599 600 601
  | 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 |]
602 603 604 605

  -- Arrow notation extension
  | ITproc
  | ITrec
Ian Lynagh's avatar
Ian Lynagh committed
606 607 608 609 610 611
  | IToparenbar                 --  (|
  | ITcparenbar                 --  |)
  | ITlarrowtail                --  -<
  | ITrarrowtail                --  >-
  | ITLarrowtail                --  -<<
  | ITRarrowtail                --  >>-
612

Ian Lynagh's avatar
Ian Lynagh committed
613 614
  | ITunknown String            -- Used when the lexer can't make sense of it
  | ITeof                       -- end of file token
615 616 617 618 619 620 621

  -- 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
622
  | ITdocOptionsOld   String     -- doc options declared "-- # ..."-style
Jedai's avatar
Jedai committed
623 624
  | ITlineComment     String     -- comment starting by "--"
  | ITblockComment    String     -- comment in {- -}
625

626
  deriving Show
627 628 629 630 631 632 633 634

-- 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
635
reservedWordsFM :: UniqFM (Token, Int)
636
reservedWordsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666
    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 ),

         ( "forall",         ITforall,        bit explicitForallBit .|.
                                              bit inRulePragBit),
         ( "mdo",            ITmdo,           bit recursiveDoBit),
667 668 669
             -- See Note [Lexing type pseudo-keywords]
         ( "family",         ITfamily,        0 ),
         ( "role",           ITrole,          0 ),
Ian Lynagh's avatar
Ian Lynagh committed
670 671 672 673 674 675 676 677 678 679 680 681 682 683
         ( "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),
684
         ( "capi",           ITcapiconv,      bit cApiFfiBit),
Ian Lynagh's avatar
Ian Lynagh committed
685
         ( "prim",           ITprimcallconv,  bit ffiBit),
thoughtpolice's avatar
thoughtpolice committed
686
         ( "javascript",     ITjavascriptcallconv, bit ffiBit),
Ian Lynagh's avatar
Ian Lynagh committed
687

688
         ( "rec",            ITrec,           bit arrowsBit .|.
689
                                              bit recursiveDoBit),
Ian Lynagh's avatar
Ian Lynagh committed
690
         ( "proc",           ITproc,          bit arrowsBit)
691 692
     ]

693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708
{-----------------------------------
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.
-------------------------------------}
709

Ian Lynagh's avatar
Ian Lynagh committed
710
reservedSymsFM :: UniqFM (Token, Int -> Bool)
711
reservedSymsFM = listToUFM $
Ian Lynagh's avatar
Ian Lynagh committed
712 713 714 715 716 717 718 719 720 721 722 723
    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)
724
       ,("~#",  ITtildehsh, magicHashEnabled)
Ian Lynagh's avatar
Ian Lynagh committed
725 726 727 728 729
       ,("=>",  ITdarrow,   always)
       ,("-",   ITminus,    always)
       ,("!",   ITbang,     always)

        -- For data T (a::*) = MkT
730
       ,("*", ITstar, always) -- \i -> kindSigsEnabled i || tyFamEnabled i)
Ian Lynagh's avatar
Ian Lynagh committed
731
        -- For 'forall a . t'
732
       ,(".", ITdot,  always) -- \i -> explicitForallEnabled i || inRulePrag i)
Ian Lynagh's avatar
Ian Lynagh committed
733 734 735 736 737

       ,("-<",  ITlarrowtail, arrowsEnabled)
       ,(">-",  ITrarrowtail, arrowsEnabled)
       ,("-<<", ITLarrowtail, arrowsEnabled)
       ,(">>-", ITRarrowtail, arrowsEnabled)
738

Ian Lynagh's avatar
Ian Lynagh committed
739 740
       ,("∷",   ITdcolon, unicodeSyntaxEnabled)
       ,("⇒",   ITdarrow, unicodeSyntaxEnabled)
741
       ,("∀",   ITforall, unicodeSyntaxEnabled)
Ian Lynagh's avatar
Ian Lynagh committed
742 743
       ,("→",   ITrarrow, unicodeSyntaxEnabled)
       ,("←",   ITlarrow, unicodeSyntaxEnabled)
744 745 746 747 748 749 750 751

       ,("⤙",   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
752 753 754
        -- 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).
755 756 757 758 759
       ]

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

Ian Lynagh's avatar
Ian Lynagh committed
760
type Action = RealSrcSpan -> StringBuffer -> Int -> P (RealLocated Token)
761

762
special :: Token -> Action
twanvl's avatar
twanvl committed
763
special tok span _buf _len = return (L span tok)
764

765
token, layout_token :: Token -> Action
twanvl's avatar
twanvl committed
766 767
token t span _buf _len = return (L span t)
layout_token t span _buf _len = pushLexState layout >> return (L span t)
768

769 770
idtoken :: (StringBuffer -> Int -> Token) -> Action
idtoken f span buf len = return (L span $! (f buf len))
771

772
skip_one_varid :: (FastString -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
773
skip_one_varid f span buf len
774
  = return (L span $! f (lexemeToFastString (stepOn buf) (len-1)))
775

776 777 778 779
skip_two_varid :: (FastString -> Token) -> Action
skip_two_varid f span buf len
  = return (L span $! f (lexemeToFastString (stepOn (stepOn buf)) (len-2)))

780
strtoken :: (String -> Token) -> Action
Ian Lynagh's avatar
Ian Lynagh committed
781
strtoken f span buf len =
782
  return (L span $! (f $! lexemeToString buf len))
783

784
init_strtoken :: Int -> (String -> Token) -> Action
785
-- like strtoken, but drops the last N character(s)
Ian Lynagh's avatar
Ian Lynagh committed
786
init_strtoken drop f span buf len =
787
  return (L span $! (f $! lexemeToString buf (len-drop)))
788 789

begin :: Int -> Action
790
begin code _span _str _len = do pushLexState code; lexToken
791 792

pop :: Action
Ian Lynagh's avatar
Ian Lynagh committed
793 794
pop _span _buf _len = do _ <- popLexState
                         lexToken
795

796 797 798 799 800 801 802 803 804 805 806
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
807
              else failSpanMsgP (RealSrcSpan span) (text "Missing block")
808

809
pop_and :: Action -> Action
Ian Lynagh's avatar
Ian Lynagh committed
810 811
pop_and act span buf len = do _ <- popLexState
                              act span buf len
812

813
{-# INLINE nextCharIs #-}
Ian Lynagh's avatar
Ian Lynagh committed
814
nextCharIs :: StringBuffer -> (Char -> Bool) -> Bool
815 816
nextCharIs buf p = not (atEnd buf) && p (currentChar buf)

817 818 819 820
{-# INLINE nextCharIsNot #-}
nextCharIsNot :: StringBuffer -> (Char -> Bool) -> Bool
nextCharIsNot buf p = not (nextCharIs buf p)

Ian Lynagh's avatar
Ian Lynagh committed
821
notFollowedBy :: Char -> AlexAccPred Int
Ian Lynagh's avatar
Ian Lynagh committed
822
notFollowedBy char _ _ _ (AI _ buf)
823
  = nextCharIsNot buf (== char)
824

Ian Lynagh's avatar
Ian Lynagh committed
825
notFollowedBySymbol :: AlexAccPred Int
826
notFollowedBySymbol _ _ _ (AI _ buf)
827
  = nextCharIsNot buf (`elem` "!#$%&*+./<=>?@\\^|-~")
ross's avatar
ross committed
828

829 830 831 832 833
-- 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
834
isNormalComment :: AlexAccPred Int
835
isNormalComment bits _ _ (AI _ buf)
836
  | haddockEnabled bits = notFollowedByDocOrPragma
837
  | otherwise           = nextCharIsNot buf (== '#')
838 839
  where
    notFollowedByDocOrPragma
840
       = afterOptionalSpace buf (\b -> nextCharIsNot b (`elem` "|^*$#"))
841

842 843 844 845 846
afterOptionalSpace :: StringBuffer -> (StringBuffer -> Bool) -> Bool
afterOptionalSpace buf p
    = if nextCharIs buf (== ' ')
      then p (snd (nextChar buf))
      else p buf
847

Ian Lynagh's avatar
Ian Lynagh committed
848
atEOL :: AlexAccPred Int
849
atEOL _ _ _ (AI _ buf) = atEnd buf || currentChar buf == '\n'
850

Ian Lynagh's avatar
Ian Lynagh committed
851
ifExtension :: (Int -> Bool) -> AlexAccPred Int
852 853
ifExtension pred bits _ _ _ = pred bits

854 855 856
multiline_doc_comment :: Action
multiline_doc_comment span buf _len = withLexedDocType (worker "")
  where
857
    worker commentAcc input docType oneLine = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
858
      Just ('\n', input')
859 860 861 862 863 864
        | 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
865

866 867
    checkIfCommentLine input = check (dropNonNewlineSpace input)
      where
868 869 870
        check input = case alexGetChar' input of
          Just ('-', input) -> case alexGetChar' input of
            Just ('-', input) -> case alexGetChar' input of
871 872 873 874 875
              Just (c, _) | c /= '-' -> Just input
              _ -> Nothing
            _ -> Nothing
          _ -> Nothing

876
        dropNonNewlineSpace input = case alexGetChar' input of
Ian Lynagh's avatar
Ian Lynagh committed
877
          Just (c, input')
878 879 880 881
            | isSpace c && c /= '\n' -> dropNonNewlineSpace input'
            | otherwise -> input
          Nothing -> input

Jedai's avatar
Jedai committed
882 883 884 885 886
lineCommentToken :: Action
lineCommentToken span buf len = do
  b <- extension rawTokenStreamEnabled
  if b then strtoken ITlineComment span buf len else lexToken

887 888 889 890
{-
  nested comments require traversing by hand, they can't be parsed
  using regular expressions.
-}
Ian Lynagh's avatar
Ian Lynagh committed
891
nested_comment :: P (RealLocated Token) -> Action
892
nested_comment cont span _str _len = do
893
  input <- getInput
Jedai's avatar
Jedai committed
894
  go "" (1::Int) input
895
  where
Jedai's avatar
Jedai committed
896 897 898 899 900
    go commentAcc 0 input = do setInput input
                               b <- extension rawTokenStreamEnabled
                               if b
                                 then docCommentEnd input commentAcc ITblockComment _str span
                                 else cont
901
    go commentAcc n input = case alexGetChar' input of
902
      Nothing -> errBrace input span
903
      Just ('-',input) -> case alexGetChar' input of
904
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
905 906
        Just ('\125',input) -> go commentAcc (n-1) input
        Just (_,_)          -> go ('-':commentAcc) n input
907
      Just ('\123',input) -> case alexGetChar' input of
908
        Nothing  -> errBrace input span
Jedai's avatar
Jedai committed
909 910 911
        Just ('-',input) -> go ('-':'\123':commentAcc) (n+1) input
        Just (_,_)       -> go ('\123':commentAcc) n input
      Just (c,input) -> go (c:commentAcc) n input
912 913 914 915

nested_doc_comment :: Action
nested_doc_comment span buf _len = withLexedDocType (go "")
  where
916
    go commentAcc input docType _ = case alexGetChar' input of
917
      Nothing -> errBrace input span
918
      Just ('-',input) -> case alexGetChar' input of
919
        Nothing -> errBrace input span
twanvl's avatar
twanvl committed
920
        Just ('\125',input) ->
921
          docCommentEnd input commentAcc docType buf span
twanvl's avatar
twanvl committed
922
        Just (_,_) -> go ('-':commentAcc) input docType False
923
      Just ('\123', input) -> case alexGetChar' input of
924 925 926 927 928
        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
929
        Just (_,_) -> go ('\123':commentAcc) input docType False
930 931
      Just (c,input) -> go (c:commentAcc) input docType False

Ian Lynagh's avatar
Ian Lynagh committed
932 933
withLexedDocType :: (AlexInput -> (String -> Token) -> Bool -> P (RealLocated Token))
                 -> P (RealLocated Token)
934
withLexedDocType lexDocComment = do
935
  input@(AI _ buf) <- getInput
936 937 938 939
  case prevChar buf ' ' of
    '|' -> lexDocComment input ITdocCommentNext False
    '^' -> lexDocComment input ITdocCommentPrev False
    '$' -> lexDocComment input ITdocCommentNamed False
David Waern's avatar
David Waern committed
940 941
    '*' -> lexDocSection 1 input
    '#' -> lexDocComment input ITdocOptionsOld False
Ian Lynagh's avatar
Ian Lynagh committed
942
    _ -> panic "withLexedDocType: Bad doc type"
Ian Lynagh's avatar
Ian Lynagh committed
943
 where
944
    lexDocSection n input = case alexGetChar' input of
945
      Just ('*', input) -> lexDocSection (n+1) input
twanvl's avatar
twanvl committed
946
      Just (_,   _)     -> lexDocComment input (ITdocSection n) True
947 948
      Nothing -> do setInput input; lexToken -- eof reached, lex it normally

949 950 951
-- RULES pragmas turn on the forall and '.' keywords, and we turn them
-- off again at the end of the pragma.
rulePrag :: Action
952
rulePrag span _buf _len = do
953
  setExts (.|. bit inRulePragBit)
954 955 956
  return (L span ITrules_prag)

endPrag :: Action
957
endPrag span _buf _len = do
958 959 960
  setExts (.&. complement (bit inRulePragBit))
  return (L span ITclose_prag)

961 962 963 964
-- 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
965
-- than what was lexed by Alex, and the lexToken function doesn't know this, so
966
-- it writes the wrong token length to the parser state. This function is
Ian Lynagh's avatar
Ian Lynagh committed
967
-- called afterwards, so it can just update the state.
968 969

docCommentEnd :: AlexInput -> String -> (String -> Token) -> StringBuffer ->
Ian Lynagh's avatar
Ian Lynagh committed
970
                 RealSrcSpan -> P (RealLocated Token)
971 972
docCommentEnd input commentAcc docType buf span = do
  setInput input
973
  let (AI loc nextBuf) = input
974
      comment = reverse commentAcc
Ian Lynagh's avatar
Ian Lynagh committed
975
      span' = mkRealSrcSpan (realSrcSpanStart span) loc
976
      last_len = byteDiff buf nextBuf
Ian Lynagh's avatar
Ian Lynagh committed
977