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(..),
29 P, thenP, thenP_, returnP, mapP, failP, failMsgP,
30 getSrcLocP, getSrcFile,
31 layoutOn, layoutOff, pushContext, popContext
34 #include "HsVersions.h"
36 import Char ( isSpace, toUpper )
37 import List ( isSuffixOf )
39 import IdInfo ( InlinePragInfo(..) )
40 import PrelNames ( mkTupNameStr )
41 import CmdLineOpts ( opt_HiVersion, opt_NoHiCheck )
42 import Demand ( Demand(..) {- instance Read -} )
43 import UniqFM ( listToUFM, lookupUFM )
44 import BasicTypes ( NewOrData(..), Boxity(..) )
45 import SrcLoc ( SrcLoc, incSrcLine, srcLocFile, srcLocLine,
46 replaceSrcLine, mkSrcLoc )
48 import ErrUtils ( Message )
56 import PrelRead ( readRational__ ) -- Glasgow non-std
59 %************************************************************************
61 \subsection{Data types}
63 %************************************************************************
65 The token data type, fairly un-interesting except from one
66 constructor, @ITidinfo@, which is used to lazily lex id info (arity,
67 strictness, unfolding etc).
69 The Idea/Observation here is that the renamer needs to scan through
70 all of an interface file before it can continue. But only a fraction
71 of the information contained in the file turns out to be useful, so
72 delaying as much as possible of the scanning and parsing of an
73 interface file Makes Sense (Heap profiles of the compiler
74 show a reduction in heap usage by at least a factor of two,
77 Hence, the interface file lexer spots when value declarations are
78 being scanned and return the @ITidinfo@ and @ITtype@ constructors
79 for the type and any other id info for that binding (unfolding, strictness
80 etc). These constructors are applied to the result of lexing these sub-chunks.
82 The lexing of the type and id info is all done lazily, of course, so
83 the scanning (and subsequent parsing) will be done *only* on the ids the
84 renamer finds out that it is interested in. The rest will just be junked.
85 Laziness, you know it makes sense :-)
89 = ITas -- Haskell keywords
115 | ITforall -- GHC extension keywords
125 | ITinterface -- interface keywords
133 | ITccall (Bool,Bool,Bool) -- (is_dyn, is_casm, may_gc)
152 | ITunfold InlinePragInfo
153 | ITstrict ([Demand], Bool)
160 | ITspecialise_prag -- Pragmas
169 | ITdotdot -- reserved symbols
183 | ITbiglam -- GHC-extension symbols
185 | ITocurly -- special symbols
187 | ITocurlybar -- {|, for type applications
188 | ITccurlybar -- |}, for type applications
201 | ITvarid FAST_STRING -- identifiers
202 | ITconid FAST_STRING
203 | ITvarsym FAST_STRING
204 | ITconsym FAST_STRING
205 | ITqvarid (FAST_STRING,FAST_STRING)
206 | ITqconid (FAST_STRING,FAST_STRING)
207 | ITqvarsym (FAST_STRING,FAST_STRING)
208 | ITqconsym (FAST_STRING,FAST_STRING)
210 | ITipvarid FAST_STRING -- GHC extension: implicit param: ?x
212 | ITpragma StringBuffer
215 | ITstring FAST_STRING
217 | ITrational Rational
220 | ITprimstring FAST_STRING
222 | ITprimfloat Rational
223 | ITprimdouble Rational
224 | ITlitlit FAST_STRING
226 | ITunknown String -- Used when the lexer can't make sense of it
227 | ITeof -- end of file token
228 deriving Show -- debugging
231 -----------------------------------------------------------------------------
235 pragmaKeywordsFM = listToUFM $
236 map (\ (x,y) -> (_PK_ x,y))
237 [( "SPECIALISE", ITspecialise_prag ),
238 ( "SPECIALIZE", ITspecialise_prag ),
239 ( "SOURCE", ITsource_prag ),
240 ( "INLINE", ITinline_prag ),
241 ( "NOINLINE", ITnoinline_prag ),
242 ( "NOTINLINE", ITnoinline_prag ),
243 ( "LINE", ITline_prag ),
244 ( "RULES", ITrules_prag ),
245 ( "RULEZ", ITrules_prag ), -- american spelling :-)
246 ( "DEPRECATED", ITdeprecated_prag )
249 haskellKeywordsFM = listToUFM $
250 map (\ (x,y) -> (_PK_ x,y))
251 [( "_", ITunderscore ),
254 ( "class", ITclass ),
256 ( "default", ITdefault ),
257 ( "deriving", ITderiving ),
260 ( "hiding", IThiding ),
262 ( "import", ITimport ),
264 ( "infix", ITinfix ),
265 ( "infixl", ITinfixl ),
266 ( "infixr", ITinfixr ),
267 ( "instance", ITinstance ),
269 ( "module", ITmodule ),
270 ( "newtype", ITnewtype ),
272 ( "qualified", ITqualified ),
275 ( "where", ITwhere ),
279 isSpecial :: Token -> Bool
280 -- If we see M.x, where x is a keyword, but
281 -- is special, we treat is as just plain M.x,
283 isSpecial ITas = True
284 isSpecial IThiding = True
285 isSpecial ITqualified = True
286 isSpecial ITforall = True
287 isSpecial ITexport = True
288 isSpecial ITlabel = True
289 isSpecial ITdynamic = True
290 isSpecial ITunsafe = True
291 isSpecial ITwith = True
292 isSpecial ITccallconv = True
293 isSpecial ITstdcallconv = True
296 -- IMPORTANT: Keep this in synch with ParseIface.y's var_fs production! (SUP)
297 ghcExtensionKeywordsFM = listToUFM $
298 map (\ (x,y) -> (_PK_ x,y))
299 [ ( "forall", ITforall ),
300 ( "foreign", ITforeign ),
301 ( "export", ITexport ),
302 ( "label", ITlabel ),
303 ( "dynamic", ITdynamic ),
304 ( "unsafe", ITunsafe ),
306 ( "stdcall", ITstdcallconv),
307 ( "ccall", ITccallconv),
308 ("_ccall_", ITccall (False, False, False)),
309 ("_ccall_GC_", ITccall (False, False, True)),
310 ("_casm_", ITccall (False, True, False)),
311 ("_casm_GC_", ITccall (False, True, True)),
313 -- interface keywords
314 ("__interface", ITinterface),
315 ("__export", IT__export),
316 ("__depends", ITdepends),
317 ("__forall", IT__forall),
318 ("__letrec", ITletrec),
319 ("__coerce", ITcoerce),
320 ("__inline_me", ITinlineMe),
321 ("__inline_call", ITinlineCall),
322 ("__depends", ITdepends),
323 ("__DEFAULT", ITdefaultbranch),
325 ("__integer", ITinteger_lit),
326 ("__float", ITfloat_lit),
327 ("__int64", ITint64_lit),
328 ("__word", ITword_lit),
329 ("__word64", ITword64_lit),
330 ("__rational", ITrational_lit),
331 ("__addr", ITaddr_lit),
332 ("__label", ITlabel_lit),
333 ("__litlit", ITlit_lit),
334 ("__string", ITstring_lit),
337 ("__fuall", ITfuall),
339 ("__P", ITspecialise),
342 ("__D", ITdeprecated),
343 ("__U", ITunfold NoInlinePragInfo),
345 ("__ccall", ITccall (False, False, False)),
346 ("__ccall_GC", ITccall (False, False, True)),
347 ("__dyn_ccall", ITccall (True, False, False)),
348 ("__dyn_ccall_GC", ITccall (True, False, True)),
349 ("__casm", ITccall (False, True, False)),
350 ("__dyn_casm", ITccall (True, True, False)),
351 ("__casm_GC", ITccall (False, True, True)),
352 ("__dyn_casm_GC", ITccall (True, True, True)),
358 haskellKeySymsFM = listToUFM $
359 map (\ (x,y) -> (_PK_ x,y))
372 ,(".", ITdot) -- sadly, for 'forall a . t'
376 -----------------------------------------------------------------------------
381 - (glaexts) lexing an interface file or -fglasgow-exts
382 - (bol) pointer to beginning of line (for column calculations)
383 - (buf) pointer to beginning of token
384 - (buf) pointer to current char
385 - (atbol) flag indicating whether we're at the beginning of a line
388 lexer :: (Token -> P a) -> P a
389 lexer cont buf s@(PState{
391 glasgow_exts = glaexts,
397 -- first, start a new lexeme and lose all the whitespace
399 tab line bol atbol (stepOverLexeme buf)
401 line = srcLocLine loc
403 tab y bol atbol buf = -- trace ("tab: " ++ show (I# y) ++ " : " ++ show (currentChar buf)) $
404 case currentChar# buf of
407 if bufferExhausted (stepOn buf)
408 then cont ITeof buf s'
409 else trace "lexer: misplaced NUL?" $
410 tab y bol atbol (stepOn buf)
412 '\n'# -> let buf' = stepOn buf
413 in tab (y +# 1#) (currentIndex# buf') 1# buf'
415 -- find comments. This got harder in Haskell 98.
416 '-'# -> let trundle n =
417 let next = lookAhead# buf n in
418 if next `eqChar#` '-'# then trundle (n +# 1#)
419 else if is_symbol next || n <# 2#
422 (stepOnUntilChar# (stepOnBy# buf n) '\n'#)
425 -- comments and pragmas. We deal with LINE pragmas here,
426 -- and throw out any unrecognised pragmas as comments. Any
427 -- pragmas we know about are dealt with later (after any layout
428 -- processing if necessary).
429 '{'# | lookAhead# buf 1# `eqChar#` '-'# ->
430 if lookAhead# buf 2# `eqChar#` '#'# then
431 if lookAhead# buf 3# `eqChar#` '#'# then is_a_token else
432 case expandWhile# is_space (setCurrentPos# buf 3#) of { buf1->
433 case expandWhile# is_ident (stepOverLexeme buf1) of { buf2->
434 let lexeme = mkFastString -- ToDo: too slow
435 (map toUpper (lexemeToString buf2)) in
436 case lookupUFM pragmaKeywordsFM lexeme of
438 line_prag skip_to_end buf2 s'
439 Just other -> is_a_token
440 Nothing -> skip_to_end (stepOnBy# buf 2#) s'
443 else skip_to_end (stepOnBy# buf 2#) s'
445 skip_to_end = nested_comment (lexer cont)
447 -- special GHC extension: we grok cpp-style #line pragmas
448 '#'# | lexemeIndex buf ==# bol -> -- the '#' must be in column 0
449 case expandWhile# is_space (stepOn buf) of { buf1 ->
450 if is_digit (currentChar# buf1)
451 then line_prag next_line buf1 s'
455 next_line buf = lexer cont (stepOnUntilChar# buf '\n'#)
457 -- tabs have been expanded beforehand
458 c | is_space c -> tab y bol atbol (stepOn buf)
459 | otherwise -> is_a_token
461 where s' = s{loc = replaceSrcLine loc y,
465 is_a_token | atbol /=# 0# = lexBOL cont buf s'
466 | otherwise = lexToken cont glaexts buf s'
468 -- {-# LINE .. #-} pragmas. yeuch.
469 line_prag cont buf s@PState{loc=loc} =
470 case expandWhile# is_space buf of { buf1 ->
471 case scanNumLit 0 (stepOverLexeme buf1) of { (line,buf2) ->
472 -- subtract one: the line number refers to the *following* line.
473 let real_line = line - 1 in
474 case fromInteger real_line of { i@(I# l) ->
475 -- ToDo, if no filename then we skip the newline.... d'oh
476 case expandWhile# is_space buf2 of { buf3 ->
477 case currentChar# buf3 of
479 case untilEndOfString# (stepOn (stepOverLexeme buf3)) of { buf4 ->
481 file = lexemeToFastString buf4
482 new_buf = stepOn (stepOverLexeme buf4)
484 if nullFastString file
485 then cont new_buf s{loc = replaceSrcLine loc l}
486 else cont new_buf s{loc = mkSrcLoc file i}
488 _other -> cont (stepOverLexeme buf3) s{loc = replaceSrcLine loc l}
491 nested_comment :: P a -> P a
492 nested_comment cont buf = loop buf
495 case currentChar# buf of
496 '\NUL'# | bufferExhausted (stepOn buf) ->
497 lexError "unterminated `{-'" buf -- -}
498 '-'# | lookAhead# buf 1# `eqChar#` '}'# ->
499 cont (stepOnBy# buf 2#)
501 '{'# | lookAhead# buf 1# `eqChar#` '-'# ->
502 nested_comment (nested_comment cont) (stepOnBy# buf 2#)
504 '\n'# -> \ s@PState{loc=loc} ->
505 let buf' = stepOn buf in
506 nested_comment cont buf'
507 s{loc = incSrcLine loc, bol = currentIndex# buf',
510 _ -> nested_comment cont (stepOn buf)
512 -- When we are lexing the first token of a line, check whether we need to
513 -- insert virtual semicolons or close braces due to layout.
515 lexBOL :: (Token -> P a) -> P a
516 lexBOL cont buf s@(PState{
518 glasgow_exts = glaexts,
523 if need_close_curly then
524 --trace ("col = " ++ show (I# col) ++ ", layout: inserting '}'") $
525 cont ITvccurly buf s{atbol = 1#, context = tail ctx}
526 else if need_semi_colon then
527 --trace ("col = " ++ show (I# col) ++ ", layout: inserting ';'") $
528 cont ITsemi buf s{atbol = 0#}
530 lexToken cont glaexts buf s{atbol = 0#}
532 col = currentIndex# buf -# bol
545 Layout n -> col ==# n
548 lexToken :: (Token -> P a) -> Int# -> P a
549 lexToken cont glaexts buf =
550 -- trace "lexToken" $
551 case currentChar# buf of
553 -- special symbols ----------------------------------------------------
554 '('# | flag glaexts && lookAhead# buf 1# `eqChar#` '#'#
555 -> cont IToubxparen (setCurrentPos# buf 2#)
557 -> cont IToparen (incLexeme buf)
559 ')'# -> cont ITcparen (incLexeme buf)
560 '['# -> cont ITobrack (incLexeme buf)
561 ']'# -> cont ITcbrack (incLexeme buf)
562 ','# -> cont ITcomma (incLexeme buf)
563 ';'# -> cont ITsemi (incLexeme buf)
564 '}'# -> \ s@PState{context = ctx} ->
566 (_:ctx') -> cont ITccurly (incLexeme buf) s{context=ctx'}
567 _ -> lexError "too many '}'s" buf s
568 '|'# -> case lookAhead# buf 1# of
569 '}'# | flag glaexts -> cont ITccurlybar
570 (setCurrentPos# buf 2#)
571 _ -> lex_sym cont (incLexeme buf)
574 '#'# -> case lookAhead# buf 1# of
575 ')'# | flag glaexts -> cont ITcubxparen (setCurrentPos# buf 2#)
576 '-'# -> case lookAhead# buf 2# of
577 '}'# -> cont ITclose_prag (setCurrentPos# buf 3#)
578 _ -> lex_sym cont (incLexeme buf)
579 _ -> lex_sym cont (incLexeme buf)
581 '`'# | flag glaexts && lookAhead# buf 1# `eqChar#` '`'#
582 -> lex_cstring cont (setCurrentPos# buf 2#)
584 -> cont ITbackquote (incLexeme buf)
586 '{'# -> -- look for "{-##" special iface pragma
587 case lookAhead# buf 1# of
589 -> cont ITocurlybar (setCurrentPos# buf 2#)
590 '-'# -> case lookAhead# buf 2# of
591 '#'# -> case lookAhead# buf 3# of
594 = doDiscard 0# (stepOnBy# (stepOverLexeme buf) 4#) in
595 cont (ITpragma lexeme) buf'
596 _ -> lex_prag cont (setCurrentPos# buf 3#)
597 _ -> cont ITocurly (incLexeme buf)
598 _ -> (layoutOff `thenP_` cont ITocurly) (incLexeme buf)
600 -- strings/characters -------------------------------------------------
601 '\"'#{-"-} -> lex_string cont glaexts [] (incLexeme buf)
602 '\''# -> lex_char (char_end cont) glaexts (incLexeme buf)
604 -- strictness and cpr pragmas and __scc treated specially.
605 '_'# | flag glaexts ->
606 case lookAhead# buf 1# of
607 '_'# -> case lookAhead# buf 2# of
609 lex_demand cont (stepOnUntil (not . isSpace)
610 (stepOnBy# buf 3#)) -- past __S
612 cont ITcprinfo (stepOnBy# buf 3#) -- past __M
615 case prefixMatch (stepOnBy# buf 3#) "cc" of
616 Just buf' -> lex_scc cont (stepOverLexeme buf')
617 Nothing -> lex_id cont glaexts buf
618 _ -> lex_id cont glaexts buf
619 _ -> lex_id cont glaexts buf
621 -- Hexadecimal and octal constants
622 '0'# | (ch `eqChar#` 'x'# || ch `eqChar#` 'X'#) && is_hexdigit ch2
623 -> readNum (after_lexnum cont glaexts) buf' is_hexdigit 16 hex
624 | (ch `eqChar#` 'o'# || ch `eqChar#` 'O'#) && is_octdigit ch2
625 -> readNum (after_lexnum cont glaexts) buf' is_octdigit 8 oct_or_dec
626 where ch = lookAhead# buf 1#
627 ch2 = lookAhead# buf 2#
628 buf' = setCurrentPos# buf 2#
631 if bufferExhausted (stepOn buf) then
634 trace "lexIface: misplaced NUL?" $
635 cont (ITunknown "\NUL") (stepOn buf)
637 '?'# | flag glaexts && is_lower (lookAhead# buf 1#) ->
638 lex_ip cont (incLexeme buf)
639 c | is_digit c -> lex_num cont glaexts 0 buf
640 | is_symbol c -> lex_sym cont buf
641 | is_upper c -> lex_con cont glaexts buf
642 | is_ident c -> lex_id cont glaexts buf
643 | otherwise -> lexError "illegal character" buf
645 -- Int# is unlifted, and therefore faster than Bool for flags.
651 -------------------------------------------------------------------------------
655 = case expandWhile# is_space buf of { buf1 ->
656 case expandWhile# is_ident (stepOverLexeme buf1) of { buf2 ->
657 let lexeme = mkFastString (map toUpper (lexemeToString buf2)) in
658 case lookupUFM pragmaKeywordsFM lexeme of
659 Just kw -> cont kw (mergeLexemes buf buf2)
660 Nothing -> panic "lex_prag"
663 -------------------------------------------------------------------------------
666 lex_string cont glaexts s buf
667 = case currentChar# buf of
669 let buf' = incLexeme buf; s' = mkFastStringInt (reverse s) in
670 case currentChar# buf' of
671 '#'# | flag glaexts -> if all (<= 0xFF) s
672 then cont (ITprimstring s') (incLexeme buf')
673 else lexError "primitive string literal must contain only characters <= '\xFF'" buf'
674 _ -> cont (ITstring s') buf'
676 -- ignore \& in a string, deal with string gaps
677 '\\'# | next_ch `eqChar#` '&'#
678 -> lex_string cont glaexts s buf'
680 -> lex_stringgap cont glaexts s (incLexeme buf)
682 where next_ch = lookAhead# buf 1#
683 buf' = setCurrentPos# buf 2#
685 _ -> lex_char (lex_next_string cont s) glaexts buf
687 lex_stringgap cont glaexts s buf
688 = let buf' = incLexeme buf in
689 case currentChar# buf of
690 '\n'# -> \st@PState{loc = loc} -> lex_stringgap cont glaexts s buf'
691 st{loc = incSrcLine loc}
692 '\\'# -> lex_string cont glaexts s buf'
693 c | is_space c -> lex_stringgap cont glaexts s buf'
694 other -> charError buf'
696 lex_next_string cont s glaexts c buf = lex_string cont glaexts (c:s) buf
698 lex_char :: (Int# -> Int -> P a) -> Int# -> P a
699 lex_char cont glaexts buf
700 = case currentChar# buf of
701 '\\'# -> lex_escape (cont glaexts) (incLexeme buf)
702 c | is_any c -> cont glaexts (I# (ord# c)) (incLexeme buf)
703 other -> charError buf
705 char_end cont glaexts c buf
706 = case currentChar# buf of
707 '\''# -> let buf' = incLexeme buf in
708 case currentChar# buf' of
710 -> cont (ITprimchar c) (incLexeme buf')
711 _ -> cont (ITchar c) buf'
715 = let buf' = incLexeme buf in
716 case currentChar# buf of
717 'a'# -> cont (ord '\a') buf'
718 'b'# -> cont (ord '\b') buf'
719 'f'# -> cont (ord '\f') buf'
720 'n'# -> cont (ord '\n') buf'
721 'r'# -> cont (ord '\r') buf'
722 't'# -> cont (ord '\t') buf'
723 'v'# -> cont (ord '\v') buf'
724 '\\'# -> cont (ord '\\') buf'
725 '"'# -> cont (ord '\"') buf'
726 '\''# -> cont (ord '\'') buf'
727 '^'# -> let c = currentChar# buf' in
728 if c `geChar#` '@'# && c `leChar#` '_'#
729 then cont (I# (ord# c -# ord# '@'#)) (incLexeme buf')
732 'x'# -> readNum (after_charnum cont) buf' is_hexdigit 16 hex
733 'o'# -> readNum (after_charnum cont) buf' is_octdigit 8 oct_or_dec
735 -> readNum (after_charnum cont) buf is_digit 10 oct_or_dec
737 _ -> case [ (c,buf2) | (p,c) <- silly_escape_chars,
738 Just buf2 <- [prefixMatch buf p] ] of
739 (c,buf2):_ -> cont (ord c) buf2
742 after_charnum cont i buf
743 = if i >= 0 && i <= 0x7FFFFFFF
744 then cont (fromInteger i) buf
747 readNum cont buf is_digit base conv = read buf 0
749 = case currentChar# buf of { c ->
751 then read (incLexeme buf) (i*base + (toInteger (I# (conv c))))
757 || (c `geChar#` 'a'# && c `leChar#` 'f'#)
758 || (c `geChar#` 'A'# && c `leChar#` 'F'#)
760 hex c | is_digit c = ord# c -# ord# '0'#
761 | otherwise = ord# (to_lower c) -# ord# 'a'# +# 10#
762 oct_or_dec c = ord# c -# ord# '0'#
764 is_octdigit c = c `geChar#` '0'# && c `leChar#` '7'#
767 | c `geChar#` 'A'# && c `leChar#` 'Z'#
768 = chr# (ord# c -# (ord# 'A'# -# ord# 'a'#))
771 charError buf = lexError "error in character literal" buf
773 silly_escape_chars = [
810 -------------------------------------------------------------------------------
812 lex_demand cont buf =
813 case read_em [] buf of { (ls,buf') ->
814 case currentChar# buf' of
815 'B'# -> cont (ITstrict (ls, True )) (incLexeme buf')
816 _ -> cont (ITstrict (ls, False)) buf'
819 -- code snatched from Demand.lhs
821 case currentChar# buf of
822 'L'# -> read_em (WwLazy False : acc) (stepOn buf)
823 'A'# -> read_em (WwLazy True : acc) (stepOn buf)
824 'S'# -> read_em (WwStrict : acc) (stepOn buf)
825 'P'# -> read_em (WwPrim : acc) (stepOn buf)
826 'E'# -> read_em (WwEnum : acc) (stepOn buf)
827 ')'# -> (reverse acc, stepOn buf)
828 'U'# -> do_unpack DataType True acc (stepOnBy# buf 2#)
829 'u'# -> do_unpack DataType False acc (stepOnBy# buf 2#)
830 'N'# -> do_unpack NewType True acc (stepOnBy# buf 2#)
831 'n'# -> do_unpack NewType False acc (stepOnBy# buf 2#)
832 _ -> (reverse acc, buf)
834 do_unpack new_or_data wrapper_unpacks acc buf
835 = case read_em [] buf of
836 (stuff, rest) -> read_em (WwUnpack new_or_data wrapper_unpacks stuff : acc) rest
841 case currentChar# buf of
842 'C'# -> cont ITsccAllCafs (incLexeme buf)
843 other -> cont ITscc buf
845 -----------------------------------------------------------------------------
848 lex_num :: (Token -> P a) -> Int# -> Integer -> P a
849 lex_num cont glaexts acc buf =
850 case scanNumLit acc buf of
852 case currentChar# buf' of
853 '.'# | is_digit (lookAhead# buf' 1#) ->
854 -- this case is not optimised at all, as the
855 -- presence of floating point numbers in interface
856 -- files is not that common. (ToDo)
857 case expandWhile# is_digit (incLexeme buf') of
858 buf2 -> -- points to first non digit char
860 let l = case currentChar# buf2 of
866 = let buf3 = incLexeme buf2 in
867 case currentChar# buf3 of
868 '-'# -> expandWhile# is_digit (incLexeme buf3)
869 '+'# -> expandWhile# is_digit (incLexeme buf3)
870 x | is_digit x -> expandWhile# is_digit buf3
873 v = readRational__ (lexemeToString l)
875 in case currentChar# l of -- glasgow exts only
876 '#'# | flag glaexts -> let l' = incLexeme l in
877 case currentChar# l' of
878 '#'# -> cont (ITprimdouble v) (incLexeme l')
879 _ -> cont (ITprimfloat v) l'
880 _ -> cont (ITrational v) l
882 _ -> after_lexnum cont glaexts acc' buf'
884 after_lexnum cont glaexts i buf
885 = case currentChar# buf of
886 '#'# | flag glaexts -> cont (ITprimint i) (incLexeme buf)
887 _ -> cont (ITinteger i) buf
889 -----------------------------------------------------------------------------
890 -- C "literal literal"s (i.e. things like ``NULL'', ``stdout'' etc.)
892 -- we lexemeToFastString on the bit between the ``''s, but include the
893 -- quotes in the full lexeme.
895 lex_cstring cont buf =
896 case expandUntilMatch (stepOverLexeme buf) "\'\'" of
897 Just buf' -> cont (ITlitlit (lexemeToFastString
898 (setCurrentPos# buf' (negateInt# 2#))))
899 (mergeLexemes buf buf')
900 Nothing -> lexError "unterminated ``" buf
902 -----------------------------------------------------------------------------
903 -- identifiers, symbols etc.
906 case expandWhile# is_ident buf of
907 buf' -> cont (ITipvarid lexeme) buf'
908 where lexeme = lexemeToFastString buf'
910 lex_id cont glaexts buf =
911 let buf1 = expandWhile# is_ident buf in
914 case (if flag glaexts
915 then expandWhile# (eqChar# '#'#) buf1 -- slurp trailing hashes
916 else buf1) of { buf' ->
918 let lexeme = lexemeToFastString buf' in
920 case _scc_ "Lex.haskellKeyword" lookupUFM haskellKeywordsFM lexeme of {
921 Just kwd_token -> --trace ("hkeywd: "++_UNPK_(lexeme)) $
925 let var_token = cont (ITvarid lexeme) buf' in
927 if not (flag glaexts)
931 case lookupUFM ghcExtensionKeywordsFM lexeme of {
932 Just kwd_token -> cont kwd_token buf';
939 case expandWhile# is_symbol buf of
940 buf' -> case lookupUFM haskellKeySymsFM lexeme of {
941 Just kwd_token -> --trace ("keysym: "++unpackFS lexeme) $
942 cont kwd_token buf' ;
943 Nothing -> --trace ("sym: "++unpackFS lexeme) $
944 cont (mk_var_token lexeme) buf'
946 where lexeme = lexemeToFastString buf'
949 lex_con cont glaexts buf =
950 -- trace ("con: "{-++unpackFS lexeme-}) $
951 case expandWhile# is_ident buf of { buf1 ->
952 case slurp_trailing_hashes buf1 glaexts of { buf' ->
954 case currentChar# buf' of
959 just_a_conid = cont (ITconid lexeme) buf'
960 lexeme = lexemeToFastString buf'
961 munch = lex_qid cont glaexts lexeme (incLexeme buf') just_a_conid
964 lex_qid cont glaexts mod buf just_a_conid =
965 -- trace ("quid: "{-++unpackFS lexeme-}) $
966 case currentChar# buf of
967 '['# -> -- Special case for []
968 case lookAhead# buf 1# of
969 ']'# -> cont (ITqconid (mod,SLIT("[]"))) (setCurrentPos# buf 2#)
972 '('# -> -- Special case for (,,,)
973 -- This *is* necessary to deal with e.g. "instance C PrelBase.(,,)"
974 case lookAhead# buf 1# of
975 '#'# | flag glaexts -> case lookAhead# buf 2# of
976 ','# -> lex_ubx_tuple cont mod (setCurrentPos# buf 3#)
979 ')'# -> cont (ITqconid (mod,SLIT("()"))) (setCurrentPos# buf 2#)
980 ','# -> lex_tuple cont mod (setCurrentPos# buf 2#) just_a_conid
983 '-'# -> case lookAhead# buf 1# of
984 '>'# -> cont (ITqconid (mod,SLIT("(->)"))) (setCurrentPos# buf 2#)
985 _ -> lex_id3 cont glaexts mod buf just_a_conid
986 _ -> lex_id3 cont glaexts mod buf just_a_conid
988 lex_id3 cont glaexts mod buf just_a_conid
989 | is_symbol (currentChar# buf) =
991 start_new_lexeme = stepOverLexeme buf
993 -- trace ("lex_id31 "{-++unpackFS lexeme-}) $
994 case expandWhile# is_symbol start_new_lexeme of { buf' ->
996 lexeme = lexemeToFastString buf'
997 -- real lexeme is M.<sym>
998 new_buf = mergeLexemes buf buf'
1000 cont (mk_qvar_token mod lexeme) new_buf
1001 -- wrong, but arguably morally right: M... is now a qvarsym
1006 start_new_lexeme = stepOverLexeme buf
1008 -- trace ("lex_id32 "{-++unpackFS lexeme-}) $
1009 case expandWhile# is_ident start_new_lexeme of { buf1 ->
1014 case slurp_trailing_hashes buf1 glaexts of { buf' ->
1017 lexeme = lexemeToFastString buf'
1018 new_buf = mergeLexemes buf buf'
1019 is_a_qvarid = cont (mk_qvar_token mod lexeme) new_buf
1021 case _scc_ "Lex.haskellKeyword" lookupUFM haskellKeywordsFM lexeme of {
1022 Nothing -> is_a_qvarid ;
1024 Just kwd_token | isSpecial kwd_token -- special ids (as, qualified, hiding) shouldn't be
1025 -> is_a_qvarid -- recognised as keywords here.
1027 -> just_a_conid -- avoid M.where etc.
1030 slurp_trailing_hashes buf glaexts
1031 | flag glaexts = expandWhile# (`eqChar#` '#'#) buf
1036 | is_upper f = ITconid pk_str
1037 | is_ident f = ITvarid pk_str
1038 | f `eqChar#` ':'# = ITconsym pk_str
1039 | otherwise = ITvarsym pk_str
1041 (C# f) = _HEAD_ pk_str
1042 -- tl = _TAIL_ pk_str
1044 mk_qvar_token m token =
1045 -- trace ("mk_qvar ") $
1046 case mk_var_token token of
1047 ITconid n -> ITqconid (m,n)
1048 ITvarid n -> ITqvarid (m,n)
1049 ITconsym n -> ITqconsym (m,n)
1050 ITvarsym n -> ITqvarsym (m,n)
1051 _ -> ITunknown (show token)
1054 ----------------------------------------------------------------------------
1055 Horrible stuff for dealing with M.(,,,)
1058 lex_tuple cont mod buf back_off =
1062 case currentChar# buf of
1063 ','# -> go (n+1) (stepOn buf)
1064 ')'# -> cont (ITqconid (mod, snd (mkTupNameStr Boxed n))) (stepOn buf)
1067 lex_ubx_tuple cont mod buf back_off =
1071 case currentChar# buf of
1072 ','# -> go (n+1) (stepOn buf)
1073 '#'# -> case lookAhead# buf 1# of
1074 ')'# -> cont (ITqconid (mod, snd (mkTupNameStr Unboxed n)))
1080 -----------------------------------------------------------------------------
1081 doDiscard rips along really fast, looking for a '##-}',
1082 indicating the end of the pragma we're skipping
1085 doDiscard inStr buf =
1086 case currentChar# buf of
1087 '#'# | inStr ==# 0# ->
1088 case lookAhead# buf 1# of { '#'# ->
1089 case lookAhead# buf 2# of { '-'# ->
1090 case lookAhead# buf 3# of { '}'# ->
1091 (lexemeToBuffer buf, stepOverLexeme (setCurrentPos# buf 4#));
1092 _ -> doDiscard inStr (incLexeme buf) };
1093 _ -> doDiscard inStr (incLexeme buf) };
1094 _ -> doDiscard inStr (incLexeme buf) }
1098 odd_slashes buf flg i# =
1099 case lookAhead# buf i# of
1100 '\\'# -> odd_slashes buf (not flg) (i# -# 1#)
1103 not_inStr = if inStr ==# 0# then 1# else 0#
1105 case lookAhead# buf (negateInt# 1#) of --backwards, actually
1106 '\\'# -> -- escaping something..
1107 if odd_slashes buf True (negateInt# 2#)
1108 then -- odd number of slashes, " is escaped.
1109 doDiscard inStr (incLexeme buf)
1110 else -- even number of slashes, \ is escaped.
1111 doDiscard not_inStr (incLexeme buf)
1112 _ -> doDiscard not_inStr (incLexeme buf)
1114 '\''# | inStr ==# 0# ->
1115 case lookAhead# buf 1# of { '"'# ->
1116 case lookAhead# buf 2# of { '\''# ->
1117 doDiscard inStr (setCurrentPos# buf 3#);
1118 _ -> doDiscard inStr (incLexeme buf) };
1119 _ -> doDiscard inStr (incLexeme buf) }
1121 _ -> doDiscard inStr (incLexeme buf)
1125 -----------------------------------------------------------------------------
1136 data PState = PState {
1138 glasgow_exts :: Int#,
1141 context :: [LayoutContext]
1144 type P a = StringBuffer -- Input string
1149 returnP a buf s = POk s a
1151 thenP :: P a -> (a -> P b) -> P b
1152 m `thenP` k = \ buf s ->
1154 POk s1 a -> k a buf s1
1155 PFailed err -> PFailed err
1157 thenP_ :: P a -> P b -> P b
1158 m `thenP_` k = m `thenP` \_ -> k
1160 mapP :: (a -> P b) -> [a] -> P [b]
1161 mapP f [] = returnP []
1164 mapP f as `thenP` \bs ->
1167 failP :: String -> P a
1168 failP msg buf s = PFailed (text msg)
1170 failMsgP :: Message -> P a
1171 failMsgP msg buf s = PFailed msg
1173 lexError :: String -> P a
1174 lexError str buf s@PState{ loc = loc }
1175 = failMsgP (hcat [ppr loc, text ": ", text str]) buf s
1177 getSrcLocP :: P SrcLoc
1178 getSrcLocP buf s@(PState{ loc = loc }) = POk s loc
1180 getSrcFile :: P FAST_STRING
1181 getSrcFile buf s@(PState{ loc = loc }) = POk s (srcLocFile loc)
1183 getContext :: P [LayoutContext]
1184 getContext buf s@(PState{ context = ctx }) = POk s ctx
1186 pushContext :: LayoutContext -> P ()
1187 pushContext ctxt buf s@(PState{ context = ctx }) = POk s{context = ctxt:ctx} ()
1191 This special case in layoutOn is to handle layout contexts with are
1192 indented the same or less than the current context. This is illegal
1193 according to the Haskell spec, so we have to arrange to close the
1194 current context. eg.
1199 after the first 'where', the sequence of events is:
1201 - layout system inserts a ';' (column 0)
1202 - parser begins a new context at column 0
1203 - parser shifts ';' (legal empty declaration)
1204 - parser sees 'class': parse error (we're still in the inner context)
1206 trouble is, by the time we know we need a new context, the lexer has
1207 already generated the ';'. Hacky solution is as follows: since we
1208 know the column of the next token (it's the column number of the new
1209 context), we set the ACTUAL column number of the new context to this
1210 numer plus one. Hence the next time the lexer is called, a '}' will
1211 be generated to close the new context straight away. Furthermore, we
1212 have to set the atbol flag so that the ';' that the parser shifted as
1213 part of the new context is re-generated.
1215 when the new context is *less* indented than the current one:
1217 f = f where g = g where
1220 - current context: column 12.
1221 - on seeing 'h' (column 0), the layout system inserts '}'
1222 - parser starts a new context, column 0
1223 - parser sees '}', uses it to close new context
1224 - we still need to insert another '}' followed by a ';',
1225 hence the atbol trick.
1227 There's also a special hack in here to deal with
1234 i.e. the inner context is at the same indentation level as the outer
1235 context. This is strictly illegal according to Haskell 98, but
1236 there's a lot of existing code using this style and it doesn't make
1237 any sense to disallow it, since empty 'do' lists don't make sense.
1240 layoutOn :: Bool -> P ()
1241 layoutOn strict buf s@(PState{ bol = bol, context = ctx }) =
1242 let offset = lexemeIndex buf -# bol in
1245 | if strict then prev_off >=# offset else prev_off ># offset ->
1246 --trace ("layout on, column: " ++ show (I# offset)) $
1247 POk s{ context = Layout (offset +# 1#) : ctx, atbol = 1# } ()
1249 --trace ("layout on, column: " ++ show (I# offset)) $
1250 POk s{ context = Layout offset : ctx } ()
1253 layoutOff buf s@(PState{ context = ctx }) =
1254 POk s{ context = NoLayout:ctx } ()
1257 popContext = \ buf s@(PState{ context = ctx, loc = loc }) ->
1259 (_:tl) -> POk s{ context = tl } ()
1260 [] -> PFailed (srcParseErr buf loc)
1263 Note that if the name of the file we're processing ends
1264 with `hi-boot', we accept it on faith as having the right
1265 version. This is done so that .hi-boot files that comes
1266 with hsc don't have to be updated before every release,
1267 *and* it allows us to share .hi-boot files with versions
1268 of hsc that don't have .hi version checking (e.g., ghc-2.10's)
1270 If the version number is 0, the checking is also turned off.
1271 (needed to deal with GHC.hi only!)
1273 Once we can assume we're compiling with a version of ghc that
1274 supports interface file checking, we can drop the special
1277 checkVersion :: Maybe Integer -> P ()
1278 checkVersion mb@(Just v) buf s@(PState{loc = loc})
1279 | (v==0) || (v == fromInt opt_HiVersion) || opt_NoHiCheck = POk s ()
1280 | otherwise = PFailed (ifaceVersionErr mb loc ([]::[Token]){-Todo-})
1281 checkVersion mb@Nothing buf s@(PState{loc = loc})
1282 | "hi-boot" `isSuffixOf` (_UNPK_ (srcLocFile loc)) = POk s ()
1283 | otherwise = PFailed (ifaceVersionErr mb loc ([]::[Token]){-Todo-})
1285 -----------------------------------------------------------------
1287 ifaceParseErr :: StringBuffer -> SrcLoc -> Message
1289 = hsep [ppr l, ptext SLIT("Interface file parse error; on input `"),
1290 text (lexemeToString s), char '\'']
1292 ifaceVersionErr hi_vers l toks
1293 = hsep [ppr l, ptext SLIT("Interface file version error;"),
1294 ptext SLIT("Expected"), int opt_HiVersion,
1295 ptext SLIT("found "), pp_version]
1299 Nothing -> ptext SLIT("pre ghc-3.02 version")
1300 Just v -> ptext SLIT("version") <+> integer v
1302 -----------------------------------------------------------------------------
1304 srcParseErr :: StringBuffer -> SrcLoc -> Message
1308 then ptext SLIT(": parse error (possibly incorrect indentation)")
1309 else hcat [ptext SLIT(": parse error on input "),
1310 char '`', text token, char '\'']
1313 token = lexemeToString s