diff options
Diffstat (limited to 'lualibs-lpeg.lua')
| -rw-r--r-- | lualibs-lpeg.lua | 106 |
1 files changed, 74 insertions, 32 deletions
diff --git a/lualibs-lpeg.lua b/lualibs-lpeg.lua index 323c73b..7be86d3 100644 --- a/lualibs-lpeg.lua +++ b/lualibs-lpeg.lua @@ -79,13 +79,22 @@ patterns.endofstring = endofstring patterns.beginofstring = alwaysmatched patterns.alwaysmatched = alwaysmatched -local digit, sign = R('09'), S('+-') +local sign = S('+-') +local zero = P('0') +local digit = R('09') +local octdigit = R("07") +local lowercase = R("az") +local uppercase = R("AZ") +local underscore = P("_") +local hexdigit = digit + lowercase + uppercase local cr, lf, crlf = P("\r"), P("\n"), P("\r\n") local newline = crlf + S("\r\n") -- cr + lf local escaped = P("\\") * anything local squote = P("'") local dquote = P('"') local space = P(" ") +local period = P(".") +local comma = P(",") local utfbom_32_be = P('\000\000\254\255') local utfbom_32_le = P('\255\254\000\000') @@ -137,29 +146,14 @@ patterns.nonwhitespace = nonwhitespace local stripper = spacer^0 * C((spacer^0 * nonspacer^1)^0) -- from example by roberto ------ collapser = Cs(spacer^0/"" * ((spacer^1 * P(-1) / "") + (spacer^1/" ") + P(1))^0) +----- collapser = Cs(spacer^0/"" * ((spacer^1 * endofstring / "") + (spacer^1/" ") + P(1))^0) local collapser = Cs(spacer^0/"" * nonspacer^0 * ((spacer^0/" " * nonspacer^1)^0)) patterns.stripper = stripper patterns.collapser = collapser -patterns.digit = digit -patterns.sign = sign -patterns.cardinal = sign^0 * digit^1 -patterns.integer = sign^0 * digit^1 -patterns.unsigned = digit^0 * P('.') * digit^1 -patterns.float = sign^0 * patterns.unsigned -patterns.cunsigned = digit^0 * P(',') * digit^1 -patterns.cfloat = sign^0 * patterns.cunsigned -patterns.number = patterns.float + patterns.integer -patterns.cnumber = patterns.cfloat + patterns.integer -patterns.oct = P("0") * R("07")^1 -patterns.octal = patterns.oct -patterns.HEX = P("0x") * R("09","AF")^1 -patterns.hex = P("0x") * R("09","af")^1 -patterns.hexadecimal = P("0x") * R("09","AF","af")^1 -patterns.lowercase = R("az") -patterns.uppercase = R("AZ") +patterns.lowercase = lowercase +patterns.uppercase = uppercase patterns.letter = patterns.lowercase + patterns.uppercase patterns.space = space patterns.tab = P("\t") @@ -167,12 +161,12 @@ patterns.spaceortab = patterns.space + patterns.tab patterns.newline = newline patterns.emptyline = newline^1 patterns.equal = P("=") -patterns.comma = P(",") -patterns.commaspacer = P(",") * spacer^0 -patterns.period = P(".") +patterns.comma = comma +patterns.commaspacer = comma * spacer^0 +patterns.period = period patterns.colon = P(":") patterns.semicolon = P(";") -patterns.underscore = P("_") +patterns.underscore = underscore patterns.escaped = escaped patterns.squote = squote patterns.dquote = dquote @@ -187,7 +181,33 @@ patterns.singlequoted = squote * patterns.nosquote * squote patterns.doublequoted = dquote * patterns.nodquote * dquote patterns.quoted = patterns.doublequoted + patterns.singlequoted -patterns.propername = R("AZ","az","__") * R("09","AZ","az", "__")^0 * P(-1) +patterns.digit = digit +patterns.octdigit = octdigit +patterns.hexdigit = hexdigit +patterns.sign = sign +patterns.cardinal = digit^1 +patterns.integer = sign^-1 * digit^1 +patterns.unsigned = digit^0 * period * digit^1 +patterns.float = sign^-1 * patterns.unsigned +patterns.cunsigned = digit^0 * comma * digit^1 +patterns.cfloat = sign^-1 * patterns.cunsigned +patterns.number = patterns.float + patterns.integer +patterns.cnumber = patterns.cfloat + patterns.integer +patterns.oct = zero * octdigit^1 +patterns.octal = patterns.oct +patterns.HEX = zero * P("X") * (digit+uppercase)^1 +patterns.hex = zero * P("x") * (digit+lowercase)^1 +patterns.hexadecimal = zero * S("xX") * hexdigit^1 + +patterns.hexafloat = sign^-1 + * zero * S("xX") + * (hexdigit^0 * period * hexdigit^1 + hexdigit^1 * period * hexdigit^0 + hexdigit^1) + * (S("pP") * sign^-1 * hexdigit^1)^-1 +patterns.decafloat = sign^-1 + * (digit^0 * period * digit^1 + digit^1 * period * digit^0 + digit^1) + * S("eE") * sign^-1 * digit^1 + +patterns.propername = (uppercase + lowercase + underscore) * (uppercase + lowercase + underscore + digit)^0 * endofstring patterns.somecontent = (anything - newline - space)^1 -- (utf8char - newline - space)^1 patterns.beginline = #(1-newline) @@ -421,7 +441,7 @@ function lpeg.replacer(one,two,makefunction,isutf) -- in principle we should sor end end -function lpeg.finder(lst,makefunction) +function lpeg.finder(lst,makefunction) -- beware: slower than find with 'patternless finds' local pattern if type(lst) == "table" then pattern = P(false) @@ -456,8 +476,8 @@ local splitters_f, splitters_s = { }, { } function lpeg.firstofsplit(separator) -- always return value local splitter = splitters_f[separator] if not splitter then - separator = P(separator) - splitter = C((1 - separator)^0) + local pattern = P(separator) + splitter = C((1 - pattern)^0) splitters_f[separator] = splitter end return splitter @@ -466,13 +486,35 @@ end function lpeg.secondofsplit(separator) -- nil if not split local splitter = splitters_s[separator] if not splitter then - separator = P(separator) - splitter = (1 - separator)^0 * separator * C(anything^0) + local pattern = P(separator) + splitter = (1 - pattern)^0 * pattern * C(anything^0) splitters_s[separator] = splitter end return splitter end +local splitters_s, splitters_p = { }, { } + +function lpeg.beforesuffix(separator) -- nil if nothing but empty is ok + local splitter = splitters_s[separator] + if not splitter then + local pattern = P(separator) + splitter = C((1 - pattern)^0) * pattern * endofstring + splitters_s[separator] = splitter + end + return splitter +end + +function lpeg.afterprefix(separator) -- nil if nothing but empty is ok + local splitter = splitters_p[separator] + if not splitter then + local pattern = P(separator) + splitter = pattern * C(anything^0) + splitters_p[separator] = splitter + end + return splitter +end + function lpeg.balancer(left,right) left, right = P(left), P(right) return P { left * ((1 - left - right) + V(1))^0 * right } @@ -832,9 +874,9 @@ end -- moved here (before util-str) -local digit = R("09") -local period = P(".") -local zero = P("0") +----- digit = R("09") +----- period = P(".") +----- zero = P("0") local trailingzeros = zero^0 * -digit -- suggested by Roberto R local case_1 = period * trailingzeros / "" local case_2 = period * (digit - trailingzeros)^1 * (trailingzeros / "") |