2 % (c) The GRASP/AQUA Project, Glasgow University, 1992-1998
4 \section[Lexical analysis]{Lexical analysis}
6 --------------------------------------------------------
8 There's a known bug in here:
10 If an interface file ends prematurely, Lex tries to
11 do headFS of an empty FastString.
13 An example that provokes the error is
15 f _:_ _forall_ [a] <<<END OF FILE>>>
16 --------------------------------------------------------
19 {-# OPTIONS -#include "ctypes.h" #-}
26 Token(..), lexer, ParseResult(..), PState(..),
30 P, thenP, thenP_, returnP, mapP, failP, failMsgP,
31 getSrcLocP, getSrcFile,
32 layoutOn, layoutOff, pushContext, popContext
35 #include "HsVersions.h"
37 import Char ( ord, isSpace )
38 import List ( isSuffixOf )
40 import IdInfo ( InlinePragInfo(..), CprInfo(..) )
41 import Name ( isLowerISO, isUpperISO )
42 import PrelMods ( mkTupNameStr, mkUbxTupNameStr )
43 import CmdLineOpts ( opt_IgnoreIfacePragmas, opt_HiVersion, opt_NoHiCheck )
44 import Demand ( Demand(..) {- instance Read -} )
45 import UniqFM ( UniqFM, listToUFM, lookupUFM)
46 import BasicTypes ( NewOrData(..) )
47 import SrcLoc ( SrcLoc, incSrcLine, srcLocFile, srcLocLine,
48 replaceSrcLine, mkSrcLoc )
50 import Maybes ( MaybeErr(..) )
51 import ErrUtils ( Message )
59 #if __GLASGOW_HASKELL__ >= 303
66 import PrelRead ( readRational__ ) -- Glasgow non-std
69 %************************************************************************
71 \subsection{Data types}
73 %************************************************************************
75 The token data type, fairly un-interesting except from one
76 constructor, @ITidinfo@, which is used to lazily lex id info (arity,
77 strictness, unfolding etc).
79 The Idea/Observation here is that the renamer needs to scan through
80 all of an interface file before it can continue. But only a fraction
81 of the information contained in the file turns out to be useful, so
82 delaying as much as possible of the scanning and parsing of an
83 interface file Makes Sense (Heap profiles of the compiler
84 show a reduction in heap usage by at least a factor of two,
87 Hence, the interface file lexer spots when value declarations are
88 being scanned and return the @ITidinfo@ and @ITtype@ constructors
89 for the type and any other id info for that binding (unfolding, strictness
90 etc). These constructors are applied to the result of lexing these sub-chunks.
92 The lexing of the type and id info is all done lazily, of course, so
93 the scanning (and subsequent parsing) will be done *only* on the ids the
94 renamer finds out that it is interested in. The rest will just be junked.
95 Laziness, you know it makes sense :-)
99 = ITas -- Haskell keywords
124 | ITforall -- GHC extension keywords
131 | ITinterface -- interface keywords
139 | ITccall (Bool,Bool,Bool) -- (is_dyn, is_casm, may_gc)
154 | ITunfold InlinePragInfo
155 | ITstrict ([Demand], Bool)
157 | ITcprinfo (CprInfo)
161 | ITspecialise_prag -- Pragmas
169 | ITdotdot -- reserved symbols
183 | ITbiglam -- GHC-extension symbols
185 | ITocurly -- special symbols
199 | ITvarid FAST_STRING -- identifiers
200 | ITconid FAST_STRING
201 | ITvarsym FAST_STRING
202 | ITconsym FAST_STRING
203 | ITqvarid (FAST_STRING,FAST_STRING)
204 | ITqconid (FAST_STRING,FAST_STRING)
205 | ITqvarsym (FAST_STRING,FAST_STRING)
206 | ITqconsym (FAST_STRING,FAST_STRING)
208 | ITpragma StringBuffer
211 | ITstring FAST_STRING
213 | ITrational Rational
216 | ITprimstring FAST_STRING
218 | ITprimfloat Rational
219 | ITprimdouble Rational
220 | ITlitlit FAST_STRING
222 | ITunknown String -- Used when the lexer can't make sense of it
223 | ITeof -- end of file token
224 deriving Text -- debugging
227 -----------------------------------------------------------------------------
231 pragmaKeywordsFM = listToUFM $
232 map (\ (x,y) -> (_PK_ x,y))
233 [( "SPECIALISE", ITspecialise_prag ),
234 ( "SPECIALIZE", ITspecialise_prag ),
235 ( "SOURCE", ITsource_prag ),
236 ( "INLINE", ITinline_prag ),
237 ( "NOINLINE", ITnoinline_prag ),
238 ( "LINE", ITline_prag ),
239 ( "RULES", ITrules_prag ),
240 ( "RULEZ", ITrules_prag ) -- american spelling :-)
243 haskellKeywordsFM = listToUFM $
244 map (\ (x,y) -> (_PK_ x,y))
245 [( "_", ITunderscore ),
248 ( "class", ITclass ),
250 ( "default", ITdefault ),
251 ( "deriving", ITderiving ),
254 ( "hiding", IThiding ),
256 ( "import", ITimport ),
258 ( "infix", ITinfix ),
259 ( "infixl", ITinfixl ),
260 ( "infixr", ITinfixr ),
261 ( "instance", ITinstance ),
263 ( "module", ITmodule ),
264 ( "newtype", ITnewtype ),
266 ( "qualified", ITqualified ),
273 ghcExtensionKeywordsFM = listToUFM $
274 map (\ (x,y) -> (_PK_ x,y))
275 [ ( "forall", ITforall ),
276 ( "foreign", ITforeign ),
277 ( "export", ITexport ),
278 ( "label", ITlabel ),
279 ( "dynamic", ITdynamic ),
280 ( "unsafe", ITunsafe ),
281 ("_ccall_", ITccall (False, False, False)),
282 ("_ccall_GC_", ITccall (False, False, True)),
283 ("_casm_", ITccall (False, True, False)),
284 ("_casm_GC_", ITccall (False, True, True)),
286 -- interface keywords
287 ("__interface", ITinterface),
288 ("__export", IT__export),
289 ("__depends", ITdepends),
290 ("__forall", IT__forall),
291 ("__letrec", ITletrec),
292 ("__coerce", ITcoerce),
293 ("__inline_me", ITinlineMe),
294 ("__inline_call", ITinlineCall),
295 ("__depends", ITdepends),
296 ("__DEFAULT", ITdefaultbranch),
298 ("__integer", ITinteger_lit),
299 ("__float", ITfloat_lit),
300 ("__rational", ITrational_lit),
301 ("__addr", ITaddr_lit),
302 ("__litlit", ITlit_lit),
303 ("__string", ITstring_lit),
308 ("__P", ITspecialise),
311 ("__u", ITunfold NoInlinePragInfo),
313 ("__ccall", ITccall (False, False, False)),
314 ("__ccall_GC", ITccall (False, False, True)),
315 ("__dyn_ccall", ITccall (True, False, False)),
316 ("__dyn_ccall_GC", ITccall (True, False, True)),
317 ("__casm", ITccall (False, True, False)),
318 ("__dyn_casm", ITccall (True, True, False)),
319 ("__casm_GC", ITccall (False, True, True)),
320 ("__dyn_casm_GC", ITccall (True, True, True)),
326 haskellKeySymsFM = listToUFM $
327 map (\ (x,y) -> (_PK_ x,y))
340 ,(".", ITdot) -- sadly, for 'forall a . t'
344 -----------------------------------------------------------------------------
349 - (glaexts) lexing an interface file or -fglasgow-exts
350 - (bol) pointer to beginning of line (for column calculations)
351 - (buf) pointer to beginning of token
352 - (buf) pointer to current char
353 - (atbol) flag indicating whether we're at the beginning of a line
356 lexer :: (Token -> P a) -> P a
357 lexer cont buf s@(PState{
359 glasgow_exts = glaexts,
365 -- first, start a new lexeme and lose all the whitespace
366 = tab line bol atbol (stepOverLexeme buf)
368 line = srcLocLine loc
370 tab y bol atbol buf = --trace ("tab: " ++ show (I# y) ++ " : " ++ show (currentChar buf)) $
371 case currentChar# buf of
374 if bufferExhausted (stepOn buf)
375 then cont ITeof buf s'
376 else trace "lexer: misplaced NUL?" $
377 tab y bol atbol (stepOn buf)
379 '\n'# -> let buf' = stepOn buf
380 in tab (y +# 1#) (currentIndex# buf') 1# buf'
382 -- find comments. This got harder in Haskell 98.
383 '-'# -> let trundle n =
384 let next = lookAhead# buf n in
385 if next `eqChar#` '-'# then trundle (n +# 1#)
386 else if is_symbol next || n <# 2#
388 else case untilChar# (stepOnBy# buf n) '\n'# of
389 { buf' -> tab y bol atbol (stepOverLexeme buf')
393 -- comments and pragmas. We deal with LINE pragmas here,
394 -- and throw out any unrecognised pragmas as comments. Any
395 -- pragmas we know about are dealt with later (after any layout
396 -- processing if necessary).
398 '{'# | lookAhead# buf 1# `eqChar#` '-'# ->
399 if lookAhead# buf 2# `eqChar#` '#'# then
400 if lookAhead# buf 3# `eqChar#` '#'# then is_a_token else
401 case expandWhile# is_space (setCurrentPos# buf 3#) of { buf1->
402 case expandWhile# is_ident (stepOverLexeme buf1) of { buf2->
403 let lexeme = lexemeToFastString buf2 in
404 case lookupUFM pragmaKeywordsFM lexeme of
405 Just ITline_prag -> line_prag (lexer cont) buf2 s'
406 Just other -> is_a_token
407 Nothing -> skip_to_end (stepOnBy# buf 2#)
410 else skip_to_end (stepOnBy# buf 2#)
412 skip_to_end buf = nested_comment (lexer cont) buf s'
414 -- tabs have been expanded beforehand
415 c | is_space c -> tab y bol atbol (stepOn buf)
416 | otherwise -> is_a_token
418 where s' = s{loc = replaceSrcLine loc y,
422 is_a_token | atbol /=# 0# = lexBOL cont buf s'
423 | otherwise = lexToken cont glaexts buf s'
425 -- {-# LINE .. #-} pragmas. yeuch.
427 case expandWhile# is_space buf of { buf1 ->
428 case scanNumLit 0 (stepOverLexeme buf1) of { (line,buf2) ->
429 -- subtract one: the line number refers to the *following* line.
430 let real_line = line - 1 in
431 case fromInteger real_line of { i@(I# l) ->
432 case expandWhile# is_space buf2 of { buf3 ->
433 case currentChar# buf3 of
435 case untilEndOfString# (stepOn (stepOverLexeme buf3)) of { buf4 ->
436 let file = lexemeToFastString buf4 in
437 \s@PState{loc=loc} -> skipToEnd buf4 s{loc = mkSrcLoc file i}
439 other -> \s@PState{loc=loc} -> skipToEnd buf3 s{loc = replaceSrcLine loc l}
442 skipToEnd buf = nested_comment cont buf
444 nested_comment :: P a -> P a
445 nested_comment cont buf = loop buf
448 case currentChar# buf of
449 '\NUL'# | bufferExhausted (stepOn buf) ->
450 lexError "unterminated `{-'" buf
452 '-'# | lookAhead# buf 1# `eqChar#` '}'# ->
453 cont (stepOnBy# buf 2#)
455 '{'# | lookAhead# buf 1# `eqChar#` '-'# ->
456 nested_comment (nested_comment cont) (stepOnBy# buf 2#)
458 '\n'# -> \ s@PState{loc=loc} ->
459 let buf' = stepOn buf in
460 nested_comment cont buf'
461 s{loc = incSrcLine loc, bol = currentIndex# buf',
464 _ -> nested_comment cont (stepOn buf)
466 -- When we are lexing the first token of a line, check whether we need to
467 -- insert virtual semicolons or close braces due to layout.
469 lexBOL :: (Token -> P a) -> P a
470 lexBOL cont buf s@(PState{
472 glasgow_exts = glaexts,
477 if need_close_curly then
478 --trace ("col = " ++ show (I# col) ++ ", layout: inserting '}'") $
479 cont ITvccurly buf s{atbol = 1#, context = tail ctx}
480 else if need_semi_colon then
481 --trace ("col = " ++ show (I# col) ++ ", layout: inserting ';'") $
482 cont ITsemi buf s{atbol = 0#}
484 lexToken cont glaexts buf s{atbol = 0#}
486 col = currentIndex# buf -# bol
499 Layout n -> col ==# n
502 lexToken :: (Token -> P a) -> Int# -> P a
503 lexToken cont glaexts buf =
506 case currentChar# buf of
508 -- special symbols ----------------------------------------------------
509 '('# | flag glaexts && lookAhead# buf 1# `eqChar#` '#'#
510 -> cont IToubxparen (setCurrentPos# buf 2#)
512 -> cont IToparen (incLexeme buf)
514 ')'# -> cont ITcparen (incLexeme buf)
515 '['# -> cont ITobrack (incLexeme buf)
516 ']'# -> cont ITcbrack (incLexeme buf)
517 ','# -> cont ITcomma (incLexeme buf)
518 ';'# -> cont ITsemi (incLexeme buf)
520 '}'# -> \ s@PState{context = ctx} ->
522 (_:ctx') -> cont ITccurly (incLexeme buf) s{context=ctx'}
523 _ -> lexError "too many '}'s" buf s
526 -> case lookAhead# buf 1# of
527 ')'# -> cont ITcubxparen (setCurrentPos# buf 2#)
528 '-'# -> case lookAhead# buf 2# of
529 '}'# -> cont ITclose_prag (setCurrentPos# buf 3#)
530 _ -> lex_sym cont (incLexeme buf)
531 _ -> lex_sym cont (incLexeme buf)
533 '`'# | flag glaexts && lookAhead# buf 1# `eqChar#` '`'#
534 -> lex_cstring cont (setCurrentPos# buf 2#)
536 -> cont ITbackquote (incLexeme buf)
538 '{'# -> -- look for "{-##" special iface pragma
539 case lookAhead# buf 1# of
540 '-'# -> case lookAhead# buf 2# of
541 '#'# -> case lookAhead# buf 3# of
544 = doDiscard False (stepOnBy# (stepOverLexeme buf) 4#) in
545 cont (ITpragma lexeme) buf'
546 _ -> lex_prag cont (setCurrentPos# buf 3#)
547 _ -> cont ITocurly (incLexeme buf)
548 _ -> (layoutOff `thenP_` cont ITocurly) (incLexeme buf)
550 -- strings/characters -------------------------------------------------
551 '\"'#{-"-} -> lex_string cont glaexts "" (incLexeme buf)
552 '\''# -> lex_char (char_end cont) glaexts (incLexeme buf)
554 -- strictness and cpr pragmas and __scc treated specially.
555 '_'# | flag glaexts ->
556 case lookAhead# buf 1# of
557 '_'# -> case lookAhead# buf 2# of
559 lex_demand cont (stepOnUntil (not . isSpace)
560 (stepOnBy# buf 3#)) -- past __S
562 lex_cpr cont (stepOnUntil (not . isSpace)
563 (stepOnBy# buf 3#)) -- past __M
565 case prefixMatch (stepOnBy# buf 3#) "cc" of
566 Just buf' -> lex_scc cont (stepOverLexeme buf')
567 Nothing -> lex_id cont glaexts buf
568 _ -> lex_id cont glaexts buf
569 _ -> lex_id cont glaexts buf
571 -- Hexadecimal and octal constants
572 '0'# | (ch `eqChar#` 'x'# || ch `eqChar#` 'X'#) && is_hexdigit ch2
573 -> readNum (after_lexnum cont glaexts) buf' is_hexdigit 16 hex
574 | (ch `eqChar#` 'o'# || ch `eqChar#` 'O'#) && is_octdigit ch2
575 -> readNum (after_lexnum cont glaexts) buf' is_octdigit 8 oct_or_dec
576 where ch = lookAhead# buf 1#
577 ch2 = lookAhead# buf 2#
578 buf' = setCurrentPos# buf 2#
581 if bufferExhausted (stepOn buf) then
584 trace "lexIface: misplaced NUL?" $
585 cont (ITunknown "\NUL") (stepOn buf)
587 c | is_digit c -> lex_num cont glaexts 0 buf
588 | is_symbol c -> lex_sym cont buf
589 | is_upper c -> lex_con cont glaexts buf
590 | is_ident c -> lex_id cont glaexts buf
592 -- Int# is unlifted, and therefore faster than Bool for flags.
598 -------------------------------------------------------------------------------
602 = case expandWhile# is_space buf of { buf1 ->
603 case expandWhile# is_ident (stepOverLexeme buf1) of { buf2 ->
604 let lexeme = lexemeToFastString buf2 in
605 case lookupUFM pragmaKeywordsFM lexeme of
606 Just kw -> cont kw (mergeLexemes buf buf2)
607 Nothing -> panic "lex_prag"
610 -------------------------------------------------------------------------------
613 lex_string cont glaexts s buf
614 = case currentChar# buf of
616 let buf' = incLexeme buf; s' = mkFastString (reverse s) in
617 case currentChar# buf' of
618 '#'# | flag glaexts -> cont (ITprimstring s') (incLexeme buf')
619 _ -> cont (ITstring s') buf'
621 -- ignore \& in a string, deal with string gaps
622 '\\'# | next_ch `eqChar#` '&'#
623 -> lex_string cont glaexts s (setCurrentPos# buf 2#)
625 -> lex_stringgap cont glaexts s buf'
627 where next_ch = lookAhead# buf 1#
628 buf' = setCurrentPos# buf 2#
630 _ -> lex_char (lex_next_string cont s) glaexts buf
632 lex_stringgap cont glaexts s buf
633 = let buf' = incLexeme buf in
634 case currentChar# buf of
635 '\n'# -> \st@PState{loc = loc} -> lex_stringgap cont glaexts s buf'
636 st{loc = incSrcLine loc}
637 '\\'# -> lex_string cont glaexts s buf'
638 c | is_space c -> lex_stringgap cont glaexts s buf'
639 other -> charError buf'
641 lex_next_string cont s glaexts c buf = lex_string cont glaexts (c:s) buf
643 lex_char :: (Int# -> Char -> P a) -> Int# -> P a
644 lex_char cont glaexts buf
645 = case currentChar# buf of
646 '\\'# -> lex_escape (cont glaexts) (incLexeme buf)
647 c | is_any c -> cont glaexts (C# c) (incLexeme buf)
648 other -> charError buf
650 char_end cont glaexts c buf
651 = case currentChar# buf of
652 '\''# -> let buf' = incLexeme buf in
653 case currentChar# buf' of
655 -> cont (ITprimchar c) (incLexeme buf')
656 _ -> cont (ITchar c) buf'
660 = let buf' = incLexeme buf in
661 case currentChar# buf of
662 'a'# -> cont '\a' buf'
663 'b'# -> cont '\b' buf'
664 'f'# -> cont '\f' buf'
665 'n'# -> cont '\n' buf'
666 'r'# -> cont '\r' buf'
667 't'# -> cont '\t' buf'
668 'v'# -> cont '\v' buf'
669 '\\'# -> cont '\\' buf'
670 '"'# -> cont '\"' buf'
671 '\''# -> cont '\'' buf'
672 '^'# -> let c = currentChar# buf' in
673 if c `geChar#` '@'# && c `leChar#` '_'#
674 then cont (C# (chr# (ord# c -# ord# '@'#))) (incLexeme buf')
677 'x'# -> readNum (after_charnum cont) buf' is_hexdigit 16 hex
678 'o'# -> readNum (after_charnum cont) buf' is_octdigit 8 oct_or_dec
680 -> readNum (after_charnum cont) buf is_digit 10 oct_or_dec
682 _ -> case [ (c,buf2) | (p,c) <- silly_escape_chars,
683 Just buf2 <- [prefixMatch buf p] ] of
684 (c,buf2):_ -> cont c buf2
687 after_charnum cont i buf
688 = let int = fromInteger i in
689 if i >= 0 && i <= 255
690 then cont (chr int) buf
693 readNum cont buf is_digit base conv = read buf 0
695 = case currentChar# buf of { c ->
697 then read (incLexeme buf) (i*base + (toInteger (I# (conv c))))
703 || (c `geChar#` 'a'# && c `leChar#` 'h'#)
704 || (c `geChar#` 'A'# && c `leChar#` 'H'#)
706 hex c | is_digit c = ord# c -# ord# '0'#
707 | otherwise = ord# (to_lower c) -# ord# 'a'# +# 10#
708 oct_or_dec c = ord# c -# ord# '0'#
710 is_octdigit c = c `geChar#` '0'# && c `leChar#` '7'#
713 | c `geChar#` 'A'# && c `leChar#` 'Z'#
714 = chr# (ord# c -# (ord# 'A'# -# ord# 'a'#))
717 charError buf = lexError "error in character literal" buf
719 silly_escape_chars = [
756 -------------------------------------------------------------------------------
758 lex_demand cont buf =
759 case read_em [] buf of { (ls,buf') ->
760 case currentChar# buf' of
761 'B'# -> cont (ITstrict (ls, True )) (incLexeme buf')
762 _ -> cont (ITstrict (ls, False)) buf'
765 -- code snatched from Demand.lhs
767 case currentChar# buf of
768 'L'# -> read_em (WwLazy False : acc) (stepOn buf)
769 'A'# -> read_em (WwLazy True : acc) (stepOn buf)
770 'S'# -> read_em (WwStrict : acc) (stepOn buf)
771 'P'# -> read_em (WwPrim : acc) (stepOn buf)
772 'E'# -> read_em (WwEnum : acc) (stepOn buf)
773 ')'# -> (reverse acc, stepOn buf)
774 'U'# -> do_unpack DataType True acc (stepOnBy# buf 2#)
775 'u'# -> do_unpack DataType False acc (stepOnBy# buf 2#)
776 'N'# -> do_unpack NewType True acc (stepOnBy# buf 2#)
777 'n'# -> do_unpack NewType False acc (stepOnBy# buf 2#)
778 _ -> (reverse acc, buf)
780 do_unpack new_or_data wrapper_unpacks acc buf
781 = case read_em [] buf of
782 (stuff, rest) -> read_em (WwUnpack new_or_data wrapper_unpacks stuff : acc) rest
785 case read_em [] buf of { (cpr_inf,buf') ->
786 ASSERT ( null (tail cpr_inf) )
787 cont (ITcprinfo $ head cpr_inf) buf'
790 -- code snatched from lex_demand above
792 case currentChar# buf of
793 '-'# -> read_em (NoCPRInfo : acc) (stepOn buf)
794 '('# -> do_unpack acc (stepOn buf)
795 ')'# -> (reverse acc, stepOn buf)
796 _ -> (reverse acc, buf)
799 = case read_em [] buf of
800 (stuff, rest) -> read_em ((CPRInfo stuff) : acc) rest
804 case currentChar# buf of
805 'C'# -> cont ITsccAllCafs (incLexeme buf)
806 other -> cont ITscc buf
808 -----------------------------------------------------------------------------
811 lex_num :: (Token -> P a) -> Int# -> Integer -> P a
812 lex_num cont glaexts acc buf =
813 case scanNumLit acc buf of
815 case currentChar# buf' of
816 '.'# | is_digit (lookAhead# buf' 1#) ->
817 -- this case is not optimised at all, as the
818 -- presence of floating point numbers in interface
819 -- files is not that common. (ToDo)
820 case expandWhile# is_digit (incLexeme buf') of
821 buf2 -> -- points to first non digit char
823 let l = case currentChar# buf2 of
829 = let buf3 = incLexeme buf2 in
830 case currentChar# buf3 of
831 '-'# -> expandWhile# is_digit (incLexeme buf3)
832 '+'# -> expandWhile# is_digit (incLexeme buf3)
833 x | is_digit x -> expandWhile# is_digit buf3
836 v = readRational__ (lexemeToString l)
838 in case currentChar# l of -- glasgow exts only
839 '#'# | flag glaexts -> let l' = incLexeme l in
840 case currentChar# l' of
841 '#'# -> cont (ITprimdouble v) (incLexeme l')
842 _ -> cont (ITprimfloat v) l'
843 _ -> cont (ITrational v) l
845 _ -> after_lexnum cont glaexts acc' buf'
847 after_lexnum cont glaexts i buf
848 = case currentChar# buf of
849 '#'# | flag glaexts -> cont (ITprimint i) (incLexeme buf)
850 _ -> cont (ITinteger i) buf
852 -----------------------------------------------------------------------------
853 -- C "literal literal"s (i.e. things like ``NULL'', ``stdout'' etc.)
855 -- we lexemeToFastString on the bit between the ``''s, but include the
856 -- quotes in the full lexeme.
858 lex_cstring cont buf =
859 case expandUntilMatch (stepOverLexeme buf) "\'\'" of
860 buf' -> cont (ITlitlit (lexemeToFastString
861 (setCurrentPos# buf' (negateInt# 2#))))
862 (mergeLexemes buf buf')
864 ------------------------------------------------------------------------------
867 is_ident, is_symbol, is_any, is_upper, is_digit :: Char# -> Bool
869 {-# INLINE is_ctype #-}
870 #if __GLASGOW_HASKELL__ >= 303
871 is_ctype :: Word8 -> Char# -> Bool
872 is_ctype mask = \c ->
873 (indexWord8OffAddr (``char_types'' :: Addr) (ord (C# c)) .&. mask) /= 0
875 is_ctype :: Int -> Char# -> Bool
876 is_ctype (I# mask) = \c ->
877 let (A# ctype) = ``char_types'' :: Addr
878 flag_word = int2Word# (ord# (indexCharOffAddr# ctype (ord# c)))
880 (flag_word `and#` (int2Word# mask)) `neWord#` (int2Word# 0#)
883 is_ident = is_ctype 1
884 is_symbol = is_ctype 2
886 is_space = is_ctype 8
887 is_upper = is_ctype 16
888 is_digit = is_ctype 32
890 -----------------------------------------------------------------------------
891 -- identifiers, symbols etc.
893 lex_id cont glaexts buf =
894 case expandWhile# is_ident buf of { buf1 ->
896 case (if flag glaexts
897 then expandWhile# (eqChar# '#'#) buf1 -- slurp trailing hashes
898 else buf1) of { buf' ->
900 let lexeme = lexemeToFastString buf' in
902 case _scc_ "Lex.haskellKeyword" lookupUFM haskellKeywordsFM lexeme of {
903 Just kwd_token -> --trace ("hkeywd: "++_UNPK_(lexeme)) $
907 let var_token = cont (mk_var_token lexeme) buf' in
909 if not (flag glaexts)
913 case lookupUFM ghcExtensionKeywordsFM lexeme of {
914 Just kwd_token -> cont kwd_token buf';
920 case expandWhile# is_symbol buf of
921 buf' -> case lookupUFM haskellKeySymsFM lexeme of {
922 Just kwd_token -> --trace ("keysym: "++unpackFS lexeme) $
923 cont kwd_token buf' ;
924 Nothing -> --trace ("sym: "++unpackFS lexeme) $
925 cont (mk_var_token lexeme) buf'
927 where lexeme = lexemeToFastString buf'
930 lex_con cont glaexts buf =
931 case expandWhile# is_ident buf of { buf1 ->
932 case slurp_trailing_hashes buf1 glaexts of { buf' ->
934 case currentChar# buf' of
939 just_a_conid = --trace ("con: "++unpackFS lexeme) $
940 cont (ITconid lexeme) buf'
941 lexeme = lexemeToFastString buf'
942 munch = lex_qid cont glaexts lexeme (incLexeme buf') just_a_conid
945 lex_qid cont glaexts mod buf just_a_conid =
946 case currentChar# buf of
947 '['# -> -- Special case for []
948 case lookAhead# buf 1# of
949 ']'# -> cont (ITqconid (mod,SLIT("[]"))) (setCurrentPos# buf 2#)
952 '('# -> -- Special case for (,,,)
953 -- This *is* necessary to deal with e.g. "instance C PrelBase.(,,)"
954 case lookAhead# buf 1# of
955 '#'# | flag glaexts -> case lookAhead# buf 2# of
956 ','# -> lex_ubx_tuple cont mod (setCurrentPos# buf 3#)
959 ')'# -> cont (ITqconid (mod,SLIT("()"))) (setCurrentPos# buf 2#)
960 ','# -> lex_tuple cont mod (setCurrentPos# buf 2#) just_a_conid
963 '-'# -> case lookAhead# buf 1# of
964 '>'# -> cont (ITqconid (mod,SLIT("->"))) (setCurrentPos# buf 2#)
965 _ -> lex_id3 cont glaexts mod buf just_a_conid
966 _ -> lex_id3 cont glaexts mod buf just_a_conid
968 lex_id3 cont glaexts mod buf just_a_conid
969 | is_symbol (currentChar# buf) =
971 start_new_lexeme = stepOverLexeme buf
973 case expandWhile# is_symbol start_new_lexeme of { buf' ->
975 lexeme = lexemeToFastString buf'
976 -- real lexeme is M.<sym>
977 new_buf = mergeLexemes buf buf'
979 case lookupUFM haskellKeySymsFM lexeme of {
980 Just kwd_token -> just_a_conid; -- avoid M.:: etc.
981 Nothing -> cont (mk_qvar_token mod lexeme) new_buf
986 start_new_lexeme = stepOverLexeme buf
988 case expandWhile# is_ident start_new_lexeme of { buf1 ->
993 case slurp_trailing_hashes buf1 glaexts of { buf' ->
996 lexeme = lexemeToFastString buf'
997 new_buf = mergeLexemes buf buf'
999 case _scc_ "Lex.haskellKeyword" lookupUFM haskellKeywordsFM lexeme of {
1000 Just kwd_token -> just_a_conid; -- avoid M.where etc.
1003 then case lookupUFM ghcExtensionKeywordsFM lexeme of {
1004 Just kwd_token -> just_a_conid;
1005 Nothing -> cont (mk_qvar_token mod lexeme) new_buf }
1010 slurp_trailing_hashes buf glaexts
1011 | flag glaexts = expandWhile# (`eqChar#` '#'#) buf
1016 | is_upper f = ITconid pk_str
1017 -- _[A-Z] is treated as a constructor in interface files.
1018 | f `eqChar#` '_'# && not (_NULL_ tl)
1019 && (case _HEAD_ tl of { C# g -> is_upper g }) = ITconid pk_str
1020 | is_ident f = ITvarid pk_str
1021 | f `eqChar#` ':'# = ITconsym pk_str
1022 | otherwise = ITvarsym pk_str
1024 (C# f) = _HEAD_ pk_str
1027 mk_qvar_token m token =
1028 case mk_var_token token of
1029 ITconid n -> ITqconid (m,n)
1030 ITvarid n -> ITqvarid (m,n)
1031 ITconsym n -> ITqconsym (m,n)
1032 ITvarsym n -> ITqvarsym (m,n)
1033 _ -> ITunknown (show token)
1036 ----------------------------------------------------------------------------
1037 Horrible stuff for dealing with M.(,,,)
1040 lex_tuple cont mod buf back_off =
1044 case currentChar# buf of
1045 ','# -> go (n+1) (stepOn buf)
1046 ')'# -> cont (ITqconid (mod, snd (mkTupNameStr n))) (stepOn buf)
1049 lex_ubx_tuple cont mod buf back_off =
1053 case currentChar# buf of
1054 ','# -> go (n+1) (stepOn buf)
1055 '#'# -> case lookAhead# buf 1# of
1056 ')'# -> cont (ITqconid (mod, snd (mkUbxTupNameStr n)))
1062 -----------------------------------------------------------------------------
1063 doDiscard rips along really fast, looking for a '#-}',
1064 indicating the end of the pragma we're skipping
1067 doDiscard inStr buf =
1068 case currentChar# buf of
1070 case lookAhead# buf 1# of { '#'# ->
1071 case lookAhead# buf 2# of { '-'# ->
1072 case lookAhead# buf 3# of { '}'# ->
1073 (lexemeToBuffer buf, stepOverLexeme (setCurrentPos# buf 4#));
1074 _ -> doDiscard inStr (incLexeme buf) };
1075 _ -> doDiscard inStr (incLexeme buf) };
1076 _ -> doDiscard inStr (incLexeme buf) }
1079 odd_slashes buf flg i# =
1080 case lookAhead# buf i# of
1081 '\\'# -> odd_slashes buf (not flg) (i# -# 1#)
1084 case lookAhead# buf (negateInt# 1#) of --backwards, actually
1085 '\\'# -> -- escaping something..
1086 if odd_slashes buf True (negateInt# 2#) then
1087 -- odd number of slashes, " is escaped.
1088 doDiscard inStr (incLexeme buf)
1090 -- even number of slashes, \ is escaped.
1091 doDiscard (not inStr) (incLexeme buf)
1092 _ -> case inStr of -- forced to avoid build-up
1093 True -> doDiscard False (incLexeme buf)
1094 False -> doDiscard True (incLexeme buf)
1095 _ -> doDiscard inStr (incLexeme buf)
1099 -----------------------------------------------------------------------------
1110 data PState = PState {
1112 glasgow_exts :: Int#,
1115 context :: [LayoutContext]
1118 type P a = StringBuffer -- Input string
1123 returnP a buf s = POk s a
1125 thenP :: P a -> (a -> P b) -> P b
1126 m `thenP` k = \ buf s ->
1128 POk s1 a -> k a buf s1
1129 PFailed err -> PFailed err
1131 thenP_ :: P a -> P b -> P b
1132 m `thenP_` k = m `thenP` \_ -> k
1134 mapP :: (a -> P b) -> [a] -> P [b]
1135 mapP f [] = returnP []
1138 mapP f as `thenP` \bs ->
1141 failP :: String -> P a
1142 failP msg buf s = PFailed (text msg)
1144 failMsgP :: Message -> P a
1145 failMsgP msg buf s = PFailed msg
1147 lexError :: String -> P a
1148 lexError str buf s@PState{ loc = loc }
1149 = failMsgP (hcat [ppr loc, text ": ", text str]) buf s
1151 getSrcLocP :: P SrcLoc
1152 getSrcLocP buf s@(PState{ loc = loc }) = POk s loc
1154 getSrcFile :: P FAST_STRING
1155 getSrcFile buf s@(PState{ loc = loc }) = POk s (srcLocFile loc)
1157 getContext :: P [LayoutContext]
1158 getContext buf s@(PState{ context = ctx }) = POk s ctx
1160 pushContext :: LayoutContext -> P ()
1161 pushContext ctxt buf s@(PState{ context = ctx }) = POk s{context = ctxt:ctx} ()
1165 This special case in layoutOn is to handle layout contexts with are
1166 indented the same or less than the current context. This is illegal
1167 according to the Haskell spec, so we have to arrange to close the
1168 current context. eg.
1173 after the first 'where', the sequence of events is:
1175 - layout system inserts a ';' (column 0)
1176 - parser begins a new context at column 0
1177 - parser shifts ';' (legal empty declaration)
1178 - parser sees 'class': parse error (we're still in the inner context)
1180 trouble is, by the time we know we need a new context, the lexer has
1181 already generated the ';'. Hacky solution is as follows: since we
1182 know the column of the next token (it's the column number of the new
1183 context), we set the ACTUAL column number of the new context to this
1184 numer plus one. Hence the next time the lexer is called, a '}' will
1185 be generated to close the new context straight away. Furthermore, we
1186 have to set the atbol flag so that the ';' that the parser shifted as
1187 part of the new context is re-generated.
1189 when the new context is *less* indented than the current one:
1191 f = f where g = g where
1194 - current context: column 12.
1195 - on seeing 'h' (column 0), the layout system inserts '}'
1196 - parser starts a new context, column 0
1197 - parser sees '}', uses it to close new context
1198 - we still need to insert another '}' followed by a ';',
1199 hence the atbol trick.
1204 layoutOn buf s@(PState{ bol = bol, context = ctx }) =
1205 let offset = lexemeIndex buf -# bol in
1207 Layout prev_off : _ | prev_off >=# offset ->
1208 --trace ("layout on, column: " ++ show (I# offset)) $
1209 POk s{ context = Layout (offset +# 1#) : ctx, atbol = 1# } ()
1211 --trace ("layout on, column: " ++ show (I# offset)) $
1212 POk s{ context = Layout offset : ctx } ()
1215 layoutOff buf s@(PState{ context = ctx }) =
1216 POk s{ context = NoLayout:ctx } ()
1219 popContext = \ buf s@(PState{ context = ctx }) ->
1221 (_:tl) -> POk s{ context = tl } ()
1222 [] -> panic "Lex.popContext: empty context"
1225 Note that if the name of the file we're processing ends
1226 with `hi-boot', we accept it on faith as having the right
1227 version. This is done so that .hi-boot files that comes
1228 with hsc don't have to be updated before every release,
1229 *and* it allows us to share .hi-boot files with versions
1230 of hsc that don't have .hi version checking (e.g., ghc-2.10's)
1232 If the version number is 0, the checking is also turned off.
1233 (needed to deal with GHC.hi only!)
1235 Once we can assume we're compiling with a version of ghc that
1236 supports interface file checking, we can drop the special
1239 checkVersion :: Maybe Integer -> P ()
1240 checkVersion mb@(Just v) buf s@(PState{loc = loc})
1241 | (v==0) || (v == fromInt opt_HiVersion) || opt_NoHiCheck = POk s ()
1242 | otherwise = PFailed (ifaceVersionErr mb loc ([]::[Token]){-Todo-})
1243 checkVersion mb@Nothing buf s@(PState{loc = loc})
1244 | "hi-boot" `isSuffixOf` (_UNPK_ (srcLocFile loc)) = POk s ()
1245 | otherwise = PFailed (ifaceVersionErr mb loc ([]::[Token]){-Todo-})
1247 -----------------------------------------------------------------
1249 ifaceParseErr :: StringBuffer -> SrcLoc -> Message
1251 = hsep [ppr l, ptext SLIT("Interface file parse error; on input `"),
1252 text (lexemeToString s), char '\'']
1254 ifaceVersionErr hi_vers l toks
1255 = hsep [ppr l, ptext SLIT("Interface file version error;"),
1256 ptext SLIT("Expected"), int opt_HiVersion,
1257 ptext SLIT("found "), pp_version]
1261 Nothing -> ptext SLIT("pre ghc-3.02 version")
1262 Just v -> ptext SLIT("version") <+> integer v