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 --------------------------------------------------------
22 ifaceParseErr, srcParseErr,
25 Token(..), lexer, ParseResult(..), PState(..),
26 checkVersion, ExtFlags(..), mkPState,
29 P, thenP, thenP_, returnP, mapP, failP, failMsgP,
30 getSrcLocP, setSrcLocP, getSrcFile,
31 layoutOn, layoutOff, pushContext, popContext
34 #include "HsVersions.h"
36 import Char ( isSpace, toUpper )
37 import List ( isSuffixOf )
39 import PrelNames ( mkTupNameStr )
40 import CmdLineOpts ( opt_HiVersion, opt_NoHiCheck )
41 import ForeignCall ( Safety(..) )
42 import NewDemand ( StrictSig(..), Demand(..), Demands(..),
43 DmdResult(..), mkTopDmdType, evalDmd, lazyDmd )
44 import UniqFM ( listToUFM, lookupUFM )
45 import BasicTypes ( Boxity(..) )
46 import SrcLoc ( SrcLoc, incSrcLine, srcLocFile, srcLocLine,
47 replaceSrcLine, mkSrcLoc )
49 import ErrUtils ( Message )
56 import Char ( chr, ord )
57 import PrelRead ( readRational__ ) -- Glasgow non-std
58 import PrelBits ( Bits(..) ) -- non-std
61 %************************************************************************
63 \subsection{Data types}
65 %************************************************************************
67 The token data type, fairly un-interesting except from one
68 constructor, @ITidinfo@, which is used to lazily lex id info (arity,
69 strictness, unfolding etc).
71 The Idea/Observation here is that the renamer needs to scan through
72 all of an interface file before it can continue. But only a fraction
73 of the information contained in the file turns out to be useful, so
74 delaying as much as possible of the scanning and parsing of an
75 interface file Makes Sense (Heap profiles of the compiler
76 show a reduction in heap usage by at least a factor of two,
79 Hence, the interface file lexer spots when value declarations are
80 being scanned and return the @ITidinfo@ and @ITtype@ constructors
81 for the type and any other id info for that binding (unfolding, strictness
82 etc). These constructors are applied to the result of lexing these sub-chunks.
84 The lexing of the type and id info is all done lazily, of course, so
85 the scanning (and subsequent parsing) will be done *only* on the ids the
86 renamer finds out that it is interested in. The rest will just be junked.
87 Laziness, you know it makes sense :-)
91 = ITas -- Haskell keywords
115 | ITscc -- ToDo: remove (we use {-# SCC "..." #-} now)
117 | ITforall -- GHC extension keywords
129 | ITinterface -- interface keywords
137 | ITccall (Bool,Bool,Safety) -- (is_dyn, is_casm, may_gc)
164 | ITspecialise_prag -- Pragmas
174 | ITdotdot -- reserved symbols
189 | ITbiglam -- GHC-extension symbols
191 | ITocurly -- special symbols
193 | ITocurlybar -- {|, for type applications
194 | ITccurlybar -- |}, for type applications
197 | ITopabrack -- [:, for parallel arrays with -fparr
198 | ITcpabrack -- :], for parallel arrays with -fparr
209 | ITvarid FAST_STRING -- identifiers
210 | ITconid FAST_STRING
211 | ITvarsym FAST_STRING
212 | ITconsym FAST_STRING
213 | ITqvarid (FAST_STRING,FAST_STRING)
214 | ITqconid (FAST_STRING,FAST_STRING)
215 | ITqvarsym (FAST_STRING,FAST_STRING)
216 | ITqconsym (FAST_STRING,FAST_STRING)
218 | ITdupipvarid FAST_STRING -- GHC extension: implicit param: ?x
219 | ITsplitipvarid FAST_STRING -- GHC extension: implicit param: %x
221 | ITpragma StringBuffer
224 | ITstring FAST_STRING
226 | ITrational Rational
229 | ITprimstring FAST_STRING
231 | ITprimfloat Rational
232 | ITprimdouble Rational
233 | ITlitlit FAST_STRING
235 | ITunknown String -- Used when the lexer can't make sense of it
236 | ITeof -- end of file token
237 deriving Show -- debugging
240 -----------------------------------------------------------------------------
244 pragmaKeywordsFM = listToUFM $
245 map (\ (x,y) -> (_PK_ x,y))
246 [( "SPECIALISE", ITspecialise_prag ),
247 ( "SPECIALIZE", ITspecialise_prag ),
248 ( "SOURCE", ITsource_prag ),
249 ( "INLINE", ITinline_prag ),
250 ( "NOINLINE", ITnoinline_prag ),
251 ( "NOTINLINE", ITnoinline_prag ),
252 ( "LINE", ITline_prag ),
253 ( "RULES", ITrules_prag ),
254 ( "RULEZ", ITrules_prag ), -- american spelling :-)
255 ( "SCC", ITscc_prag ),
256 ( "DEPRECATED", ITdeprecated_prag )
259 haskellKeywordsFM = listToUFM $
260 map (\ (x,y) -> (_PK_ x,y))
261 [( "_", ITunderscore ),
264 ( "class", ITclass ),
266 ( "default", ITdefault ),
267 ( "deriving", ITderiving ),
270 ( "hiding", IThiding ),
272 ( "import", ITimport ),
274 ( "infix", ITinfix ),
275 ( "infixl", ITinfixl ),
276 ( "infixr", ITinfixr ),
277 ( "instance", ITinstance ),
279 ( "module", ITmodule ),
280 ( "newtype", ITnewtype ),
282 ( "qualified", ITqualified ),
285 ( "where", ITwhere ),
286 ( "_scc_", ITscc ) -- ToDo: remove
289 isSpecial :: Token -> Bool
290 -- If we see M.x, where x is a keyword, but
291 -- is special, we treat is as just plain M.x,
293 isSpecial ITas = True
294 isSpecial IThiding = True
295 isSpecial ITqualified = True
296 isSpecial ITforall = True
297 isSpecial ITexport = True
298 isSpecial ITlabel = True
299 isSpecial ITdynamic = True
300 isSpecial ITsafe = True
301 isSpecial ITunsafe = True
302 isSpecial ITwith = True
303 isSpecial ITccallconv = True
304 isSpecial ITstdcallconv = True
307 -- IMPORTANT: Keep this in synch with ParseIface.y's var_fs production! (SUP)
308 ghcExtensionKeywordsFM = listToUFM $
309 map (\ (x,y) -> (_PK_ x,y))
310 [ ( "forall", ITforall ),
311 ( "foreign", ITforeign ),
312 ( "export", ITexport ),
313 ( "label", ITlabel ),
314 ( "dynamic", ITdynamic ),
315 ( "safe", ITunsafe ),
316 ( "unsafe", ITunsafe ),
318 ( "stdcall", ITstdcallconv),
319 ( "ccall", ITccallconv),
320 ( "dotnet", ITdotnet),
321 ("_ccall_", ITccall (False, False, PlayRisky)),
322 ("_ccall_GC_", ITccall (False, False, PlaySafe)),
323 ("_casm_", ITccall (False, True, PlayRisky)),
324 ("_casm_GC_", ITccall (False, True, PlaySafe)),
326 -- interface keywords
327 ("__interface", ITinterface),
328 ("__export", IT__export),
329 ("__depends", ITdepends),
330 ("__forall", IT__forall),
331 ("__letrec", ITletrec),
332 ("__coerce", ITcoerce),
333 ("__inline_me", ITinlineMe),
334 ("__inline_call", ITinlineCall),
335 ("__depends", ITdepends),
336 ("__DEFAULT", ITdefaultbranch),
338 ("__integer", ITinteger_lit),
339 ("__float", ITfloat_lit),
340 ("__int64", ITint64_lit),
341 ("__word", ITword_lit),
342 ("__word64", ITword64_lit),
343 ("__rational", ITrational_lit),
344 ("__addr", ITaddr_lit),
345 ("__label", ITlabel_lit),
346 ("__litlit", ITlit_lit),
347 ("__string", ITstring_lit),
350 ("__fuall", ITfuall),
352 ("__P", ITspecialise),
355 ("__D", ITdeprecated),
358 ("__ccall", ITccall (False, False, PlayRisky)),
359 ("__ccall_GC", ITccall (False, False, PlaySafe)),
360 ("__dyn_ccall", ITccall (True, False, PlayRisky)),
361 ("__dyn_ccall_GC", ITccall (True, False, PlaySafe)),
362 ("__casm", ITccall (False, True, PlayRisky)),
363 ("__dyn_casm", ITccall (True, True, PlayRisky)),
364 ("__casm_GC", ITccall (False, True, PlaySafe)),
365 ("__dyn_casm_GC", ITccall (True, True, PlaySafe)),
371 haskellKeySymsFM = listToUFM $
372 map (\ (x,y) -> (_PK_ x,y))
386 ,(".", ITdot) -- sadly, for 'forall a . t'
390 -----------------------------------------------------------------------------
395 - (exts) lexing a source with extensions, eg, an interface file or
397 - (bol) pointer to beginning of line (for column calculations)
398 - (buf) pointer to beginning of token
399 - (buf) pointer to current char
400 - (atbol) flag indicating whether we're at the beginning of a line
403 lexer :: (Token -> P a) -> P a
404 lexer cont buf s@(PState{
412 -- first, start a new lexeme and lose all the whitespace
414 tab line bol atbol (stepOverLexeme buf)
416 line = srcLocLine loc
418 tab y bol atbol buf = --trace ("tab: " ++ show (I# y) ++ " : " ++ show (currentChar buf)) $
419 case currentChar# buf of
422 if bufferExhausted (stepOn buf)
423 then cont ITeof buf s'
424 else trace "lexer: misplaced NUL?" $
425 tab y bol atbol (stepOn buf)
427 '\n'# -> let buf' = stepOn buf
428 in tab (y +# 1#) (currentIndex# buf') 1# buf'
430 -- find comments. This got harder in Haskell 98.
431 '-'# -> let trundle n =
432 let next = lookAhead# buf n in
433 if next `eqChar#` '-'# then trundle (n +# 1#)
434 else if is_symbol next || n <# 2#
437 (stepOnUntilChar# (stepOnBy# buf n) '\n'#)
440 -- comments and pragmas. We deal with LINE pragmas here,
441 -- and throw out any unrecognised pragmas as comments. Any
442 -- pragmas we know about are dealt with later (after any layout
443 -- processing if necessary).
444 '{'# | lookAhead# buf 1# `eqChar#` '-'# ->
445 if lookAhead# buf 2# `eqChar#` '#'# then
446 if lookAhead# buf 3# `eqChar#` '#'# then is_a_token else
447 case expandWhile# is_space (setCurrentPos# buf 3#) of { buf1->
448 case expandWhile# is_ident (stepOverLexeme buf1) of { buf2->
449 let lexeme = mkFastString -- ToDo: too slow
450 (map toUpper (lexemeToString buf2)) in
451 case lookupUFM pragmaKeywordsFM lexeme of
452 -- ignore RULES pragmas when -fglasgow-exts is off
453 Just ITrules_prag | not (glaExtsEnabled exts) ->
454 skip_to_end (stepOnBy# buf 2#) s'
456 line_prag skip_to_end buf2 s'
457 Just other -> is_a_token
458 Nothing -> skip_to_end (stepOnBy# buf 2#) s'
461 else skip_to_end (stepOnBy# buf 2#) s'
463 skip_to_end = skipNestedComment (lexer cont)
465 -- special GHC extension: we grok cpp-style #line pragmas
466 '#'# | lexemeIndex buf ==# bol -> -- the '#' must be in column 0
467 let buf1 | lookAhead# buf 1# `eqChar#` 'l'# &&
468 lookAhead# buf 2# `eqChar#` 'i'# &&
469 lookAhead# buf 3# `eqChar#` 'n'# &&
470 lookAhead# buf 4# `eqChar#` 'e'# = stepOnBy# buf 5#
471 | otherwise = stepOn buf
473 case expandWhile# is_space buf1 of { buf2 ->
474 if is_digit (currentChar# buf2)
475 then line_prag next_line buf2 s'
479 next_line buf = lexer cont (stepOnUntilChar# buf '\n'#)
481 -- tabs have been expanded beforehand
482 c | is_space c -> tab y bol atbol (stepOn buf)
483 | otherwise -> is_a_token
485 where s' = s{loc = replaceSrcLine loc y,
489 is_a_token | atbol /=# 0# = lexBOL cont buf s'
490 | otherwise = lexToken cont exts buf s'
492 -- {-# LINE .. #-} pragmas. yeuch.
493 line_prag cont buf s@PState{loc=loc} =
494 case expandWhile# is_space buf of { buf1 ->
495 case scanNumLit 0 (stepOverLexeme buf1) of { (line,buf2) ->
496 -- subtract one: the line number refers to the *following* line.
497 let real_line = line - 1 in
498 case fromInteger real_line of { i@(I# l) ->
499 -- ToDo, if no filename then we skip the newline.... d'oh
500 case expandWhile# is_space buf2 of { buf3 ->
501 case currentChar# buf3 of
503 case untilEndOfString# (stepOn (stepOverLexeme buf3)) of { buf4 ->
505 file = lexemeToFastString buf4
506 new_buf = stepOn (stepOverLexeme buf4)
508 if nullFastString file
509 then cont new_buf s{loc = replaceSrcLine loc l}
510 else cont new_buf s{loc = mkSrcLoc file i}
512 _other -> cont (stepOverLexeme buf3) s{loc = replaceSrcLine loc l}
515 skipNestedComment :: P a -> P a
516 skipNestedComment cont buf state = skipNestedComment' (loc state) cont buf state
518 skipNestedComment' :: SrcLoc -> P a -> P a
519 skipNestedComment' orig_loc cont buf = loop buf
522 case currentChar# buf of
523 '-'# | lookAhead# buf 1# `eqChar#` '}'# -> cont (stepOnBy# buf 2#)
525 '{'# | lookAhead# buf 1# `eqChar#` '-'# ->
527 (skipNestedComment' orig_loc cont)
530 '\n'# -> \ s@PState{loc=loc} ->
531 let buf' = stepOn buf in
532 loop buf' s{loc = incSrcLine loc,
533 bol = currentIndex# buf',
536 -- pass the original SrcLoc to lexError so that the error is
537 -- reported at the line it was originally on, not the line at
538 -- the end of the file.
539 '\NUL'# | bufferExhausted (stepOn buf) ->
540 \s -> lexError "unterminated `{-'" buf s{loc=orig_loc} -- -}
542 _ -> loop (stepOn buf)
544 -- When we are lexing the first token of a line, check whether we need to
545 -- insert virtual semicolons or close braces due to layout.
547 lexBOL :: (Token -> P a) -> P a
548 lexBOL cont buf s@(PState{
555 if need_close_curly then
556 --trace ("col = " ++ show (I# col) ++ ", layout: inserting '}'") $
557 cont ITvccurly buf s{atbol = 1#, context = tail ctx}
558 else if need_semi_colon then
559 --trace ("col = " ++ show (I# col) ++ ", layout: inserting ';'") $
560 cont ITsemi buf s{atbol = 0#}
562 lexToken cont exts buf s{atbol = 0#}
564 col = currentIndex# buf -# bol
577 Layout n -> col ==# n
580 lexToken :: (Token -> P a) -> Int# -> P a
581 lexToken cont exts buf =
582 -- trace "lexToken" $
583 case currentChar# buf of
585 -- special symbols ----------------------------------------------------
586 '('# | glaExtsEnabled exts && lookAhead# buf 1# `eqChar#` '#'#
587 -> cont IToubxparen (setCurrentPos# buf 2#)
589 -> cont IToparen (incLexeme buf)
591 ')'# -> cont ITcparen (incLexeme buf)
592 '['# | parrEnabled exts && lookAhead# buf 1# `eqChar#` ':'# ->
593 cont ITopabrack (setCurrentPos# buf 2#)
595 cont ITobrack (incLexeme buf)
596 ']'# -> cont ITcbrack (incLexeme buf)
597 ','# -> cont ITcomma (incLexeme buf)
598 ';'# -> cont ITsemi (incLexeme buf)
599 '}'# -> \ s@PState{context = ctx} ->
601 (_:ctx') -> cont ITccurly (incLexeme buf) s{context=ctx'}
602 _ -> lexError "too many '}'s" buf s
603 '|'# -> case lookAhead# buf 1# of
604 '}'# | glaExtsEnabled exts -> cont ITccurlybar
605 (setCurrentPos# buf 2#)
606 _ -> lex_sym cont (incLexeme buf)
607 ':'# -> case lookAhead# buf 1# of
608 ']'# | parrEnabled exts -> cont ITcpabrack
609 (setCurrentPos# buf 2#)
610 _ -> lex_sym cont (incLexeme buf)
613 '#'# -> case lookAhead# buf 1# of
614 ')'# | glaExtsEnabled exts
615 -> cont ITcubxparen (setCurrentPos# buf 2#)
616 '-'# -> case lookAhead# buf 2# of
617 '}'# -> cont ITclose_prag (setCurrentPos# buf 3#)
618 _ -> lex_sym cont (incLexeme buf)
619 _ -> lex_sym cont (incLexeme buf)
621 '`'# | glaExtsEnabled exts && lookAhead# buf 1# `eqChar#` '`'#
622 -> lex_cstring cont (setCurrentPos# buf 2#)
624 -> cont ITbackquote (incLexeme buf)
626 '{'# -> -- look for "{-##" special iface pragma -- for Emacs: -}
627 case lookAhead# buf 1# of
628 '|'# | glaExtsEnabled exts
629 -> cont ITocurlybar (setCurrentPos# buf 2#)
630 '-'# -> case lookAhead# buf 2# of
631 '#'# -> case lookAhead# buf 3# of
635 (\ cont lexeme buf' -> cont (ITpragma lexeme) buf')
637 (stepOnBy# (stepOverLexeme buf) 4#)
638 _ -> lex_prag cont (setCurrentPos# buf 3#)
639 _ -> cont ITocurly (incLexeme buf)
640 _ -> (layoutOff `thenP_` cont ITocurly) (incLexeme buf)
642 -- strings/characters -------------------------------------------------
643 '\"'#{-"-} -> lex_string cont exts [] (incLexeme buf)
644 '\''# -> lex_char (char_end cont) exts (incLexeme buf)
646 -- strictness and cpr pragmas and __scc treated specially.
647 '_'# | glaExtsEnabled exts ->
648 case lookAhead# buf 1# of
649 '_'# -> case lookAhead# buf 2# of
651 lex_demand cont (stepOnUntil (not . isSpace)
652 (stepOnBy# buf 3#)) -- past __S
654 cont ITcprinfo (stepOnBy# buf 3#) -- past __M
657 case prefixMatch (stepOnBy# buf 3#) "cc" of
658 Just buf' -> lex_scc cont (stepOverLexeme buf')
659 Nothing -> lex_id cont exts buf
660 _ -> lex_id cont exts buf
661 _ -> lex_id cont exts buf
663 -- Hexadecimal and octal constants
664 '0'# | (ch `eqChar#` 'x'# || ch `eqChar#` 'X'#) && is_hexdigit ch2
665 -> readNum (after_lexnum cont exts) buf' is_hexdigit 16 hex
666 | (ch `eqChar#` 'o'# || ch `eqChar#` 'O'#) && is_octdigit ch2
667 -> readNum (after_lexnum cont exts) buf' is_octdigit 8 oct_or_dec
668 where ch = lookAhead# buf 1#
669 ch2 = lookAhead# buf 2#
670 buf' = setCurrentPos# buf 2#
673 if bufferExhausted (stepOn buf) then
676 trace "lexIface: misplaced NUL?" $
677 cont (ITunknown "\NUL") (stepOn buf)
679 '?'# | glaExtsEnabled exts && is_lower (lookAhead# buf 1#) ->
680 lex_ip ITdupipvarid cont (incLexeme buf)
681 '%'# | glaExtsEnabled exts && is_lower (lookAhead# buf 1#) ->
682 lex_ip ITsplitipvarid cont (incLexeme buf)
683 c | is_digit c -> lex_num cont exts 0 buf
684 | is_symbol c -> lex_sym cont buf
685 | is_upper c -> lex_con cont exts buf
686 | is_ident c -> lex_id cont exts buf
687 | otherwise -> lexError "illegal character" buf
689 -- Int# is unlifted, and therefore faster than Bool for flags.
695 -------------------------------------------------------------------------------
699 = case expandWhile# is_space buf of { buf1 ->
700 case expandWhile# is_ident (stepOverLexeme buf1) of { buf2 ->
701 let lexeme = mkFastString (map toUpper (lexemeToString buf2)) in
702 case lookupUFM pragmaKeywordsFM lexeme of
703 Just kw -> cont kw (mergeLexemes buf buf2)
704 Nothing -> panic "lex_prag"
707 -------------------------------------------------------------------------------
710 lex_string cont exts s buf
711 = case currentChar# buf of
713 let buf' = incLexeme buf
714 s' = mkFastStringNarrow (map chr (reverse s))
715 in case currentChar# buf' of
716 '#'# | glaExtsEnabled exts -> if all (<= 0xFF) s
717 then cont (ITprimstring s') (incLexeme buf')
718 else lexError "primitive string literal must contain only characters <= \'\\xFF\'" buf'
719 _ -> cont (ITstring s') buf'
721 -- ignore \& in a string, deal with string gaps
722 '\\'# | next_ch `eqChar#` '&'#
723 -> lex_string cont exts s buf'
725 -> lex_stringgap cont exts s (incLexeme buf)
727 where next_ch = lookAhead# buf 1#
728 buf' = setCurrentPos# buf 2#
730 _ -> lex_char (lex_next_string cont s) exts buf
732 lex_stringgap cont exts s buf
733 = let buf' = incLexeme buf in
734 case currentChar# buf of
735 '\n'# -> \st@PState{loc = loc} -> lex_stringgap cont exts s buf'
736 st{loc = incSrcLine loc}
737 '\\'# -> lex_string cont exts s buf'
738 c | is_space c -> lex_stringgap cont exts s buf'
739 other -> charError buf'
741 lex_next_string cont s exts c buf = lex_string cont exts (c:s) buf
743 lex_char :: (Int# -> Int -> P a) -> Int# -> P a
744 lex_char cont exts buf
745 = case currentChar# buf of
746 '\\'# -> lex_escape (cont exts) (incLexeme buf)
747 c | is_any c -> cont exts (I# (ord# c)) (incLexeme buf)
748 other -> charError buf
750 char_end cont exts c buf
751 = case currentChar# buf of
752 '\''# -> let buf' = incLexeme buf in
753 case currentChar# buf' of
754 '#'# | glaExtsEnabled exts
755 -> cont (ITprimchar c) (incLexeme buf')
756 _ -> cont (ITchar c) buf'
760 = let buf' = incLexeme buf in
761 case currentChar# buf of
762 'a'# -> cont (ord '\a') buf'
763 'b'# -> cont (ord '\b') buf'
764 'f'# -> cont (ord '\f') buf'
765 'n'# -> cont (ord '\n') buf'
766 'r'# -> cont (ord '\r') buf'
767 't'# -> cont (ord '\t') buf'
768 'v'# -> cont (ord '\v') buf'
769 '\\'# -> cont (ord '\\') buf'
770 '"'# -> cont (ord '\"') buf'
771 '\''# -> cont (ord '\'') buf'
772 '^'# -> let c = currentChar# buf' in
773 if c `geChar#` '@'# && c `leChar#` '_'#
774 then cont (I# (ord# c -# ord# '@'#)) (incLexeme buf')
777 'x'# -> readNum (after_charnum cont) buf' is_hexdigit 16 hex
778 'o'# -> readNum (after_charnum cont) buf' is_octdigit 8 oct_or_dec
780 -> readNum (after_charnum cont) buf is_digit 10 oct_or_dec
782 _ -> case [ (c,buf2) | (p,c) <- silly_escape_chars,
783 Just buf2 <- [prefixMatch buf p] ] of
784 (c,buf2):_ -> cont (ord c) buf2
787 after_charnum cont i buf
788 = if i >= 0 && i <= 0x10FFFF
789 then cont (fromInteger i) buf
792 readNum cont buf is_digit base conv = read buf 0
794 = case currentChar# buf of { c ->
796 then read (incLexeme buf) (i*base + (toInteger (I# (conv c))))
802 || (c `geChar#` 'a'# && c `leChar#` 'f'#)
803 || (c `geChar#` 'A'# && c `leChar#` 'F'#)
805 hex c | is_digit c = ord# c -# ord# '0'#
806 | otherwise = ord# (to_lower c) -# ord# 'a'# +# 10#
807 oct_or_dec c = ord# c -# ord# '0'#
809 is_octdigit c = c `geChar#` '0'# && c `leChar#` '7'#
812 | c `geChar#` 'A'# && c `leChar#` 'Z'#
813 = chr# (ord# c -# (ord# 'A'# -# ord# 'a'#))
816 charError buf = lexError "error in character literal" buf
818 silly_escape_chars = [
855 -------------------------------------------------------------------------------
857 lex_demand cont buf =
858 case read_em [] buf of { (ls,buf') ->
859 case currentChar# buf' of
860 'b'# -> cont (ITstrict (StrictSig (mkTopDmdType ls BotRes))) (incLexeme buf')
861 'm'# -> cont (ITstrict (StrictSig (mkTopDmdType ls RetCPR))) (incLexeme buf')
862 _ -> cont (ITstrict (StrictSig (mkTopDmdType ls TopRes))) buf'
866 case currentChar# buf of
867 'T'# -> read_em (Top : acc) (stepOn buf)
868 'L'# -> read_em (lazyDmd : acc) (stepOn buf)
869 'A'# -> read_em (Abs : acc) (stepOn buf)
870 'V'# -> read_em (evalDmd : acc) (stepOn buf) -- Temporary, until
871 -- we've recompiled prelude etc
872 'C'# -> do_unary Call acc (stepOnBy# buf 2#) -- Skip 'C('
874 'U'# -> do_seq1 Eval acc (stepOnBy# buf 1#)
875 'D'# -> do_seq1 Defer acc (stepOnBy# buf 1#)
876 'S'# -> do_seq1 (Box . Eval) acc (stepOnBy# buf 1#)
878 _ -> (reverse acc, buf)
881 = case currentChar# buf of
882 '('# -> do_seq2 fn acc (stepOnBy# buf 1#)
883 _ -> read_em (fn (Poly Abs) : acc) buf
886 = case read_em [] buf of { (dmds, buf) ->
887 case currentChar# buf of
888 ')'# -> read_em (fn (Prod dmds) : acc)
890 '*'# -> ASSERT( length dmds == 1 )
891 read_em (fn (Poly (head dmds)) : acc)
892 (stepOnBy# buf 2#) -- Skip '*)'
896 = case read_em [] buf of
897 ([dmd], rest) -> read_em (fn dmd : acc) (stepOn rest) -- Skip ')'
901 case currentChar# buf of
902 'C'# -> cont ITsccAllCafs (incLexeme buf)
903 other -> cont ITscc buf
905 -----------------------------------------------------------------------------
908 lex_num :: (Token -> P a) -> Int# -> Integer -> P a
909 lex_num cont exts acc buf =
910 case scanNumLit acc buf of
912 case currentChar# buf' of
913 '.'# | is_digit (lookAhead# buf' 1#) ->
914 -- this case is not optimised at all, as the
915 -- presence of floating point numbers in interface
916 -- files is not that common. (ToDo)
917 case expandWhile# is_digit (incLexeme buf') of
918 buf2 -> -- points to first non digit char
920 let l = case currentChar# buf2 of
926 = let buf3 = incLexeme buf2 in
927 case currentChar# buf3 of
928 '-'# | is_digit (lookAhead# buf 1#)
929 -> expandWhile# is_digit (incLexeme buf3)
930 '+'# | is_digit (lookAhead# buf 1#)
931 -> expandWhile# is_digit (incLexeme buf3)
932 x | is_digit x -> expandWhile# is_digit buf3
935 v = readRational__ (lexemeToString l)
937 in case currentChar# l of -- glasgow exts only
938 '#'# | glaExtsEnabled exts -> let l' = incLexeme l in
939 case currentChar# l' of
940 '#'# -> cont (ITprimdouble v) (incLexeme l')
941 _ -> cont (ITprimfloat v) l'
942 _ -> cont (ITrational v) l
944 _ -> after_lexnum cont exts acc' buf'
946 after_lexnum cont exts i buf
947 = case currentChar# buf of
948 '#'# | glaExtsEnabled exts -> cont (ITprimint i) (incLexeme buf)
949 _ -> cont (ITinteger i) buf
951 -----------------------------------------------------------------------------
952 -- C "literal literal"s (i.e. things like ``NULL'', ``stdout'' etc.)
954 -- we lexemeToFastString on the bit between the ``''s, but include the
955 -- quotes in the full lexeme.
957 lex_cstring cont buf =
958 case expandUntilMatch (stepOverLexeme buf) "\'\'" of
959 Just buf' -> cont (ITlitlit (lexemeToFastString
960 (setCurrentPos# buf' (negateInt# 2#))))
961 (mergeLexemes buf buf')
962 Nothing -> lexError "unterminated ``" buf
964 -----------------------------------------------------------------------------
965 -- identifiers, symbols etc.
967 lex_ip ip_constr cont buf =
968 case expandWhile# is_ident buf of
969 buf' -> cont (ip_constr (tailFS lexeme)) buf'
970 where lexeme = lexemeToFastString buf'
972 lex_id cont exts buf =
973 let buf1 = expandWhile# is_ident buf in
976 case (if glaExtsEnabled exts
977 then expandWhile# (eqChar# '#'#) buf1 -- slurp trailing hashes
978 else buf1) of { buf' ->
980 let lexeme = lexemeToFastString buf' in
982 case _scc_ "Lex.haskellKeyword" lookupUFM haskellKeywordsFM lexeme of {
983 Just kwd_token -> --trace ("hkeywd: "++_UNPK_(lexeme)) $
987 let var_token = cont (ITvarid lexeme) buf' in
989 if not (glaExtsEnabled exts)
993 case lookupUFM ghcExtensionKeywordsFM lexeme of {
994 Just kwd_token -> cont kwd_token buf';
1000 -- trace "lex_sym" $
1001 case expandWhile# is_symbol buf of
1002 buf' -> case lookupUFM haskellKeySymsFM lexeme of {
1003 Just kwd_token -> --trace ("keysym: "++unpackFS lexeme) $
1004 cont kwd_token buf' ;
1005 Nothing -> --trace ("sym: "++unpackFS lexeme) $
1006 cont (mk_var_token lexeme) buf'
1008 where lexeme = lexemeToFastString buf'
1011 -- lex_con recursively collects components of a qualified identifer.
1012 -- The argument buf is the StringBuffer representing the lexeme
1013 -- identified so far, where the next character is upper-case.
1015 lex_con cont exts buf =
1016 -- trace ("con: "{-++unpackFS lexeme-}) $
1017 let empty_buf = stepOverLexeme buf in
1018 case expandWhile# is_ident empty_buf of { buf1 ->
1019 case slurp_trailing_hashes buf1 exts of { con_buf ->
1021 let all_buf = mergeLexemes buf con_buf
1023 con_lexeme = lexemeToFastString con_buf
1024 mod_lexeme = lexemeToFastString (decLexeme buf)
1025 all_lexeme = lexemeToFastString all_buf
1028 | emptyLexeme buf = cont (ITconid con_lexeme) all_buf
1029 | otherwise = cont (ITqconid (mod_lexeme,con_lexeme)) all_buf
1032 case currentChar# all_buf of
1033 '.'# -> maybe_qualified cont exts all_lexeme
1034 (incLexeme all_buf) just_a_conid
1039 maybe_qualified cont exts mod buf just_a_conid =
1040 -- trace ("qid: "{-++unpackFS lexeme-}) $
1041 case currentChar# buf of
1042 '['# -> -- Special case for []
1043 case lookAhead# buf 1# of
1044 ']'# -> cont (ITqconid (mod,SLIT("[]"))) (setCurrentPos# buf 2#)
1047 '('# -> -- Special case for (,,,)
1048 -- This *is* necessary to deal with e.g. "instance C PrelBase.(,,)"
1049 case lookAhead# buf 1# of
1050 '#'# | glaExtsEnabled exts -> case lookAhead# buf 2# of
1051 ','# -> lex_ubx_tuple cont mod (setCurrentPos# buf 3#)
1054 ')'# -> cont (ITqconid (mod,SLIT("()"))) (setCurrentPos# buf 2#)
1055 ','# -> lex_tuple cont mod (setCurrentPos# buf 2#) just_a_conid
1058 '-'# -> case lookAhead# buf 1# of
1059 '>'# -> cont (ITqconid (mod,SLIT("(->)"))) (setCurrentPos# buf 2#)
1060 _ -> lex_id3 cont exts mod buf just_a_conid
1062 _ -> lex_id3 cont exts mod buf just_a_conid
1065 lex_id3 cont exts mod buf just_a_conid
1066 | is_upper (currentChar# buf) =
1067 lex_con cont exts buf
1069 | is_symbol (currentChar# buf) =
1071 start_new_lexeme = stepOverLexeme buf
1073 -- trace ("lex_id31 "{-++unpackFS lexeme-}) $
1074 case expandWhile# is_symbol start_new_lexeme of { buf' ->
1076 lexeme = lexemeToFastString buf'
1077 -- real lexeme is M.<sym>
1078 new_buf = mergeLexemes buf buf'
1080 cont (mk_qvar_token mod lexeme) new_buf
1081 -- wrong, but arguably morally right: M... is now a qvarsym
1086 start_new_lexeme = stepOverLexeme buf
1088 -- trace ("lex_id32 "{-++unpackFS lexeme-}) $
1089 case expandWhile# is_ident start_new_lexeme of { buf1 ->
1094 case slurp_trailing_hashes buf1 exts of { buf' ->
1097 lexeme = lexemeToFastString buf'
1098 new_buf = mergeLexemes buf buf'
1099 is_a_qvarid = cont (mk_qvar_token mod lexeme) new_buf
1101 case _scc_ "Lex.haskellKeyword" lookupUFM haskellKeywordsFM lexeme of {
1102 Nothing -> is_a_qvarid ;
1104 Just kwd_token | isSpecial kwd_token -- special ids (as, qualified, hiding) shouldn't be
1105 -> is_a_qvarid -- recognised as keywords here.
1107 -> just_a_conid -- avoid M.where etc.
1110 slurp_trailing_hashes buf exts
1111 | glaExtsEnabled exts = expandWhile# (`eqChar#` '#'#) buf
1116 | is_upper f = ITconid pk_str
1117 | is_ident f = ITvarid pk_str
1118 | f `eqChar#` ':'# = ITconsym pk_str
1119 | otherwise = ITvarsym pk_str
1121 (C# f) = _HEAD_ pk_str
1122 -- tl = _TAIL_ pk_str
1124 mk_qvar_token m token =
1125 -- trace ("mk_qvar ") $
1126 case mk_var_token token of
1127 ITconid n -> ITqconid (m,n)
1128 ITvarid n -> ITqvarid (m,n)
1129 ITconsym n -> ITqconsym (m,n)
1130 ITvarsym n -> ITqvarsym (m,n)
1131 _ -> ITunknown (show token)
1134 ----------------------------------------------------------------------------
1135 Horrible stuff for dealing with M.(,,,)
1138 lex_tuple cont mod buf back_off =
1142 case currentChar# buf of
1143 ','# -> go (n+1) (stepOn buf)
1144 ')'# -> cont (ITqconid (mod, snd (mkTupNameStr Boxed n))) (stepOn buf)
1147 lex_ubx_tuple cont mod buf back_off =
1151 case currentChar# buf of
1152 ','# -> go (n+1) (stepOn buf)
1153 '#'# -> case lookAhead# buf 1# of
1154 ')'# -> cont (ITqconid (mod, snd (mkTupNameStr Unboxed n)))
1160 -----------------------------------------------------------------------------
1161 'lexPragma' rips along really fast, looking for a '##-}',
1162 indicating the end of the pragma we're skipping
1165 lexPragma cont contf inStr buf =
1166 case currentChar# buf of
1167 '#'# | inStr ==# 0# ->
1168 case lookAhead# buf 1# of { '#'# ->
1169 case lookAhead# buf 2# of { '-'# ->
1170 case lookAhead# buf 3# of { '}'# ->
1171 contf cont (lexemeToBuffer buf)
1172 (stepOverLexeme (setCurrentPos# buf 4#));
1173 _ -> lexPragma cont contf inStr (incLexeme buf) };
1174 _ -> lexPragma cont contf inStr (incLexeme buf) };
1175 _ -> lexPragma cont contf inStr (incLexeme buf) }
1179 odd_slashes buf flg i# =
1180 case lookAhead# buf i# of
1181 '\\'# -> odd_slashes buf (not flg) (i# -# 1#)
1184 not_inStr = if inStr ==# 0# then 1# else 0#
1186 case lookAhead# buf (negateInt# 1#) of --backwards, actually
1187 '\\'# -> -- escaping something..
1188 if odd_slashes buf True (negateInt# 2#)
1189 then -- odd number of slashes, " is escaped.
1190 lexPragma cont contf inStr (incLexeme buf)
1191 else -- even number of slashes, \ is escaped.
1192 lexPragma cont contf not_inStr (incLexeme buf)
1193 _ -> lexPragma cont contf not_inStr (incLexeme buf)
1195 '\''# | inStr ==# 0# ->
1196 case lookAhead# buf 1# of { '"'# ->
1197 case lookAhead# buf 2# of { '\''# ->
1198 lexPragma cont contf inStr (setCurrentPos# buf 3#);
1199 _ -> lexPragma cont contf inStr (incLexeme buf) };
1200 _ -> lexPragma cont contf inStr (incLexeme buf) }
1202 -- a sign that the input is ill-formed, since pragmas are
1203 -- assumed to always be properly closed (in .hi files).
1204 '\NUL'# -> trace "lexPragma: unexpected end-of-file" $
1205 cont (ITunknown "\NUL") buf
1207 _ -> lexPragma cont contf inStr (incLexeme buf)
1211 -----------------------------------------------------------------------------
1222 data PState = PState {
1224 extsBitmap :: Int#, -- bitmap that determines permitted extensions
1227 context :: [LayoutContext]
1230 type P a = StringBuffer -- Input string
1235 returnP a buf s = POk s a
1237 thenP :: P a -> (a -> P b) -> P b
1238 m `thenP` k = \ buf s ->
1240 POk s1 a -> k a buf s1
1241 PFailed err -> PFailed err
1243 thenP_ :: P a -> P b -> P b
1244 m `thenP_` k = m `thenP` \_ -> k
1246 mapP :: (a -> P b) -> [a] -> P [b]
1247 mapP f [] = returnP []
1250 mapP f as `thenP` \bs ->
1253 failP :: String -> P a
1254 failP msg buf s = PFailed (text msg)
1256 failMsgP :: Message -> P a
1257 failMsgP msg buf s = PFailed msg
1259 lexError :: String -> P a
1260 lexError str buf s@PState{ loc = loc }
1261 = failMsgP (hcat [ppr loc, text ": ", text str]) buf s
1263 getSrcLocP :: P SrcLoc
1264 getSrcLocP buf s@(PState{ loc = loc }) = POk s loc
1266 -- use a temporary SrcLoc for the duration of the argument
1267 setSrcLocP :: SrcLoc -> P a -> P a
1268 setSrcLocP new_loc p buf s =
1269 case p buf s{ loc=new_loc } of
1271 PFailed e -> PFailed e
1273 getSrcFile :: P FAST_STRING
1274 getSrcFile buf s@(PState{ loc = loc }) = POk s (srcLocFile loc)
1276 pushContext :: LayoutContext -> P ()
1277 pushContext ctxt buf s@(PState{ context = ctx }) = POk s{context = ctxt:ctx} ()
1281 This special case in layoutOn is to handle layout contexts with are
1282 indented the same or less than the current context. This is illegal
1283 according to the Haskell spec, so we have to arrange to close the
1284 current context. eg.
1289 after the first 'where', the sequence of events is:
1291 - layout system inserts a ';' (column 0)
1292 - parser begins a new context at column 0
1293 - parser shifts ';' (legal empty declaration)
1294 - parser sees 'class': parse error (we're still in the inner context)
1296 trouble is, by the time we know we need a new context, the lexer has
1297 already generated the ';'. Hacky solution is as follows: since we
1298 know the column of the next token (it's the column number of the new
1299 context), we set the ACTUAL column number of the new context to this
1300 numer plus one. Hence the next time the lexer is called, a '}' will
1301 be generated to close the new context straight away. Furthermore, we
1302 have to set the atbol flag so that the ';' that the parser shifted as
1303 part of the new context is re-generated.
1305 when the new context is *less* indented than the current one:
1307 f = f where g = g where
1310 - current context: column 12.
1311 - on seeing 'h' (column 0), the layout system inserts '}'
1312 - parser starts a new context, column 0
1313 - parser sees '}', uses it to close new context
1314 - we still need to insert another '}' followed by a ';',
1315 hence the atbol trick.
1317 There's also a special hack in here to deal with
1324 i.e. the inner context is at the same indentation level as the outer
1325 context. This is strictly illegal according to Haskell 98, but
1326 there's a lot of existing code using this style and it doesn't make
1327 any sense to disallow it, since empty 'do' lists don't make sense.
1330 layoutOn :: Bool -> P ()
1331 layoutOn strict buf s@(PState{ bol = bol, context = ctx }) =
1332 let offset = lexemeIndex buf -# bol in
1335 | if strict then prev_off >=# offset else prev_off ># offset ->
1336 --trace ("layout on, column: " ++ show (I# offset)) $
1337 POk s{ context = Layout (offset +# 1#) : ctx, atbol = 1# } ()
1339 --trace ("layout on, column: " ++ show (I# offset)) $
1340 POk s{ context = Layout offset : ctx } ()
1343 layoutOff buf s@(PState{ context = ctx }) =
1344 POk s{ context = NoLayout:ctx } ()
1347 popContext = \ buf s@(PState{ context = ctx, loc = loc }) ->
1349 (_:tl) -> POk s{ context = tl } ()
1350 [] -> PFailed (srcParseErr buf loc)
1353 Note that if the name of the file we're processing ends
1354 with `hi-boot', we accept it on faith as having the right
1355 version. This is done so that .hi-boot files that comes
1356 with hsc don't have to be updated before every release,
1357 *and* it allows us to share .hi-boot files with versions
1358 of hsc that don't have .hi version checking (e.g., ghc-2.10's)
1360 If the version number is 0, the checking is also turned off.
1361 (needed to deal with GHC.hi only!)
1363 Once we can assume we're compiling with a version of ghc that
1364 supports interface file checking, we can drop the special
1367 checkVersion :: Maybe Integer -> P ()
1368 checkVersion mb@(Just v) buf s@(PState{loc = loc})
1369 | (v==0) || (v == fromInt opt_HiVersion) || opt_NoHiCheck = POk s ()
1370 | otherwise = PFailed (ifaceVersionErr mb loc ([]::[Token]){-Todo-})
1371 checkVersion mb@Nothing buf s@(PState{loc = loc})
1372 | "hi-boot" `isSuffixOf` (_UNPK_ (srcLocFile loc)) = POk s ()
1373 | otherwise = PFailed (ifaceVersionErr mb loc ([]::[Token]){-Todo-})
1376 -- for reasons of efficiency, flags indicating language extensions (eg,
1377 -- -fglasgow-exts or -fparr) are represented by a bitmap stored in an unboxed
1380 glaExtsBit, ffiBit, parrBit :: Int
1382 ffiBit = 1 -- FIXME: not used yet; still part of `glaExtsBit'
1385 glaExtsEnabled, ffiEnabled, parrEnabled :: Int# -> Bool
1386 glaExtsEnabled flags = testBit (I# flags) glaExtsBit
1387 ffiEnabled flags = testBit (I# flags) ffiBit
1388 parrEnabled flags = testBit (I# flags) parrBit
1390 -- convenient record-based bitmap for the interface to the rest of the world
1392 data ExtFlags = ExtFlags {
1393 glasgowExtsEF :: Bool,
1394 -- ffiEF :: Bool, -- commented out to avoid warnings
1395 parrEF :: Bool -- while not used yet
1398 -- create a parse state
1400 mkPState :: SrcLoc -> ExtFlags -> PState
1401 mkPState loc exts = PState {
1403 extsBitmap = case bitmap of {I# bits -> bits},
1409 bitmap = glaExtsBit `setBitIf` glasgowExtsEF exts
1410 -- .|. ffiBit `setBitIf` ffiEF exts
1411 .|. parrBit `setBitIf` parrEF exts
1413 b `setBitIf` cond | cond = bit b
1417 -----------------------------------------------------------------
1419 ifaceParseErr :: StringBuffer -> SrcLoc -> Message
1421 = hsep [ppr l, ptext SLIT("Interface file parse error; on input `"),
1422 text (lexemeToString s), char '\'']
1424 ifaceVersionErr hi_vers l toks
1425 = hsep [ppr l, ptext SLIT("Interface file version error;"),
1426 ptext SLIT("Expected"), int opt_HiVersion,
1427 ptext SLIT("found "), pp_version]
1431 Nothing -> ptext SLIT("pre ghc-3.02 version")
1432 Just v -> ptext SLIT("version") <+> integer v
1434 -----------------------------------------------------------------------------
1436 srcParseErr :: StringBuffer -> SrcLoc -> Message
1440 then ptext SLIT(": parse error (possibly incorrect indentation)")
1441 else hcat [ptext SLIT(": parse error on input "),
1442 char '`', text token, char '\'']
1445 token = lexemeToString s