diff options
Diffstat (limited to 'compiler/c2nim/clex.nim')
| -rw-r--r-- | compiler/c2nim/clex.nim | 100 |
1 files changed, 66 insertions, 34 deletions
diff --git a/compiler/c2nim/clex.nim b/compiler/c2nim/clex.nim index f949b97cb..3934eea63 100644 --- a/compiler/c2nim/clex.nim +++ b/compiler/c2nim/clex.nim @@ -103,7 +103,7 @@ type inDirective: bool proc getTok*(L: var TLexer, tok: var TToken) -proc PrintTok*(tok: TToken) +proc printTok*(tok: TToken) proc `$`*(tok: TToken): string # implementation @@ -138,7 +138,7 @@ proc lexMessagePos(L: var TLexer, msg: TMsgKind, pos: int, arg = "") = var info = newLineInfo(L.fileIdx, L.linenumber, pos - L.lineStart) msgs.GlobalError(info, msg, arg) -proc TokKindToStr*(k: TTokKind): string = +proc tokKindToStr*(k: TTokKind): string = case k of pxEof: result = "[EOF]" of pxInvalid: result = "[invalid]" @@ -211,9 +211,9 @@ proc `$`(tok: TToken): string = of pxSymbol, pxInvalid, pxStarComment, pxLineComment, pxStrLit: result = tok.s of pxIntLit, pxInt64Lit: result = $tok.iNumber of pxFloatLit: result = $tok.fNumber - else: result = TokKindToStr(tok.xkind) + else: result = tokKindToStr(tok.xkind) -proc PrintTok(tok: TToken) = +proc printTok(tok: TToken) = writeln(stdout, $tok) proc matchUnderscoreChars(L: var TLexer, tok: var TToken, chars: TCharSet) = @@ -223,12 +223,12 @@ proc matchUnderscoreChars(L: var TLexer, tok: var TToken, chars: TCharSet) = while true: if buf[pos] in chars: add(tok.s, buf[pos]) - Inc(pos) + inc(pos) else: break if buf[pos] == '_': add(tok.s, '_') - Inc(pos) + inc(pos) L.bufPos = pos proc isFloatLiteral(s: string): bool = @@ -239,7 +239,7 @@ proc isFloatLiteral(s: string): bool = proc getNumber2(L: var TLexer, tok: var TToken) = var pos = L.bufpos + 2 # skip 0b tok.base = base2 - var xi: biggestInt = 0 + var xi: BiggestInt = 0 var bits = 0 while true: case L.buf[pos] @@ -264,7 +264,7 @@ proc getNumber2(L: var TLexer, tok: var TToken) = proc getNumber8(L: var TLexer, tok: var TToken) = var pos = L.bufpos + 1 # skip 0 tok.base = base8 - var xi: biggestInt = 0 + var xi: BiggestInt = 0 var bits = 0 while true: case L.buf[pos] @@ -289,7 +289,7 @@ proc getNumber8(L: var TLexer, tok: var TToken) = proc getNumber16(L: var TLexer, tok: var TToken) = var pos = L.bufpos + 2 # skip 0x tok.base = base16 - var xi: biggestInt = 0 + var xi: BiggestInt = 0 var bits = 0 while true: case L.buf[pos] @@ -315,19 +315,34 @@ proc getNumber16(L: var TLexer, tok: var TToken) = else: tok.xkind = pxIntLit L.bufpos = pos +proc getFloating(L: var TLexer, tok: var TToken) = + matchUnderscoreChars(L, tok, {'0'..'9'}) + if L.buf[L.bufpos] in {'e', 'E'}: + add(tok.s, L.buf[L.bufpos]) + inc(L.bufpos) + if L.buf[L.bufpos] in {'+', '-'}: + add(tok.s, L.buf[L.bufpos]) + inc(L.bufpos) + matchUnderscoreChars(L, tok, {'0'..'9'}) + proc getNumber(L: var TLexer, tok: var TToken) = tok.base = base10 - matchUnderscoreChars(L, tok, {'0'..'9'}) - if (L.buf[L.bufpos] == '.') and (L.buf[L.bufpos + 1] in {'0'..'9'}): - add(tok.s, '.') + if L.buf[L.bufpos] == '.': + add(tok.s, "0.") inc(L.bufpos) - matchUnderscoreChars(L, tok, {'e', 'E', '+', '-', '0'..'9'}) + getFloating(L, tok) + else: + matchUnderscoreChars(L, tok, {'0'..'9'}) + if L.buf[L.bufpos] == '.': + add(tok.s, '.') + inc(L.bufpos) + getFloating(L, tok) try: if isFloatLiteral(tok.s): tok.fnumber = parseFloat(tok.s) tok.xkind = pxFloatLit else: - tok.iNumber = ParseInt(tok.s) + tok.iNumber = parseInt(tok.s) if (tok.iNumber < low(int32)) or (tok.iNumber > high(int32)): tok.xkind = pxInt64Lit else: @@ -339,10 +354,10 @@ proc getNumber(L: var TLexer, tok: var TToken) = # ignore type suffix: while L.buf[L.bufpos] in {'A'..'Z', 'a'..'z'}: inc(L.bufpos) -proc HandleCRLF(L: var TLexer, pos: int): int = +proc handleCRLF(L: var TLexer, pos: int): int = case L.buf[pos] - of CR: result = nimlexbase.HandleCR(L, pos) - of LF: result = nimlexbase.HandleLF(L, pos) + of CR: result = nimlexbase.handleCR(L, pos) + of LF: result = nimlexbase.handleLF(L, pos) else: result = pos proc escape(L: var TLexer, tok: var TToken, allowEmpty=false) = @@ -382,6 +397,23 @@ proc escape(L: var TLexer, tok: var TToken, allowEmpty=false) = xi = (xi shl 3) or (ord(L.buf[L.bufpos]) - ord('0')) inc(L.bufpos) add(tok.s, chr(xi)) + of 'x': + var xi = 0 + inc(L.bufpos) + while true: + case L.buf[L.bufpos] + of '0'..'9': + xi = `shl`(xi, 4) or (ord(L.buf[L.bufpos]) - ord('0')) + inc(L.bufpos) + of 'a'..'f': + xi = `shl`(xi, 4) or (ord(L.buf[L.bufpos]) - ord('a') + 10) + inc(L.bufpos) + of 'A'..'F': + xi = `shl`(xi, 4) or (ord(L.buf[L.bufpos]) - ord('A') + 10) + inc(L.bufpos) + else: + break + add(tok.s, chr(xi)) elif not allowEmpty: lexMessage(L, errInvalidCharacterConstant) @@ -405,7 +437,7 @@ proc getString(L: var TLexer, tok: var TToken) = while true: case buf[pos] of '\"': - Inc(pos) + inc(pos) break of CR: pos = nimlexbase.HandleCR(L, pos) @@ -427,7 +459,7 @@ proc getString(L: var TLexer, tok: var TToken) = pos = L.bufpos else: add(tok.s, buf[pos]) - Inc(pos) + inc(pos) L.bufpos = pos tok.xkind = pxStrLit @@ -438,7 +470,7 @@ proc getSymbol(L: var TLexer, tok: var TToken) = var c = buf[pos] if c notin SymChars: break add(tok.s, c) - Inc(pos) + inc(pos) L.bufpos = pos tok.xkind = pxSymbol @@ -475,7 +507,7 @@ proc scanStarComment(L: var TLexer, tok: var TToken) = while true: case buf[pos] of CR, LF: - pos = HandleCRLF(L, pos) + pos = handleCRLF(L, pos) buf = L.buf add(tok.s, "\n#") # skip annoying stars as line prefix: (eg. @@ -511,12 +543,12 @@ proc skip(L: var TLexer, tok: var TToken) = if L.inDirective: while buf[pos] in {' ', '\t'}: inc(pos) if buf[pos] in {CR, LF}: - pos = HandleCRLF(L, pos) + pos = handleCRLF(L, pos) buf = L.buf of ' ', Tabulator: - Inc(pos) # newline is special: + inc(pos) # newline is special: of CR, LF: - pos = HandleCRLF(L, pos) + pos = handleCRLF(L, pos) buf = L.buf if L.inDirective: tok.xkind = pxNewLine @@ -559,13 +591,13 @@ proc getTok(L: var TLexer, tok: var TToken) = of 'b', 'B': getNumber2(L, tok) of '1'..'7': getNumber8(L, tok) else: getNumber(L, tok) - elif c in {'1'..'9'}: + elif c in {'1'..'9'} or (c == '.' and L.buf[L.bufpos+1] in {'0'..'9'}): getNumber(L, tok) else: case c of ';': tok.xkind = pxSemicolon - Inc(L.bufpos) + inc(L.bufpos) of '/': if L.buf[L.bufpos + 1] == '/': scanLineComment(L, tok) @@ -580,9 +612,9 @@ proc getTok(L: var TLexer, tok: var TToken) = inc(L.bufpos) of ',': tok.xkind = pxComma - Inc(L.bufpos) + inc(L.bufpos) of '(': - Inc(L.bufpos) + inc(L.bufpos) tok.xkind = pxParLe of '*': inc(L.bufpos) @@ -592,13 +624,13 @@ proc getTok(L: var TLexer, tok: var TToken) = else: tok.xkind = pxStar of ')': - Inc(L.bufpos) + inc(L.bufpos) tok.xkind = pxParRi of '[': - Inc(L.bufpos) + inc(L.bufpos) tok.xkind = pxBracketLe of ']': - Inc(L.bufpos) + inc(L.bufpos) tok.xkind = pxBracketRi of '.': inc(L.bufpos) @@ -608,10 +640,10 @@ proc getTok(L: var TLexer, tok: var TToken) = else: tok.xkind = pxDot of '{': - Inc(L.bufpos) + inc(L.bufpos) tok.xkind = pxCurlyLe of '}': - Inc(L.bufpos) + inc(L.bufpos) tok.xkind = pxCurlyRi of '+': inc(L.bufpos) @@ -752,4 +784,4 @@ proc getTok(L: var TLexer, tok: var TToken) = tok.s = $c tok.xkind = pxInvalid lexMessage(L, errInvalidToken, c & " (\\" & $(ord(c)) & ')') - Inc(L.bufpos) + inc(L.bufpos) |