From 510cb140d2e8baed13b8b27daa02f3ad2f5df3e1 Mon Sep 17 00:00:00 2001 From: Marius Date: Sat, 11 Jun 2011 18:00:15 +0300 Subject: beta 2011.06.11 16:45 --- tex/generic/context/luatex-fonts-merged.lua | 2693 ++++++++++++++------------- tex/generic/context/luatex-fonts.lua | 2 +- tex/generic/context/luatex-plain.tex | 25 + 3 files changed, 1404 insertions(+), 1316 deletions(-) create mode 100644 tex/generic/context/luatex-plain.tex (limited to 'tex/generic') diff --git a/tex/generic/context/luatex-fonts-merged.lua b/tex/generic/context/luatex-fonts-merged.lua index c90043534..0cfbf2a0c 100644 --- a/tex/generic/context/luatex-fonts-merged.lua +++ b/tex/generic/context/luatex-fonts-merged.lua @@ -1,6 +1,6 @@ -- merged file : luatex-fonts-merged.lua -- parent file : luatex-fonts.lua --- merge date : 06/09/11 12:49:16 +-- merge date : 06/11/11 16:45:35 do -- begin closure to overcome local limits and interference @@ -127,7 +127,7 @@ end -- closure do -- begin closure to overcome local limits and interference -if not modules then modules = { } end modules ['l-lpeg'] = { +if not modules then modules = { } end modules ['l-table'] = { version = 1.001, comment = "companion to luat-lib.mkiv", author = "Hans Hagen, PRAGMA-ADE, Hasselt NL", @@ -135,643 +135,945 @@ if not modules then modules = { } end modules ['l-lpeg'] = { license = "see context related readme files" } -local lpeg = require("lpeg") - -local type = type - --- Beware, we predefine a bunch of patterns here and one reason for doing so --- is that we get consistent behaviour in some of the visualizers. - -lpeg.patterns = lpeg.patterns or { } -- so that we can share -local patterns = lpeg.patterns - -local P, R, S, V, match = lpeg.P, lpeg.R, lpeg.S, lpeg.V, lpeg.match -local Ct, C, Cs, Cc = lpeg.Ct, lpeg.C, lpeg.Cs, lpeg.Cc -local lpegtype = lpeg.type - -local utfcharacters = string.utfcharacters -local utfgmatch = unicode and unicode.utf8.gmatch +local type, next, tostring, tonumber, ipairs, table, string = type, next, tostring, tonumber, ipairs, table, string +local concat, sort, insert, remove = table.concat, table.sort, table.insert, table.remove +local format, find, gsub, lower, dump, match = string.format, string.find, string.gsub, string.lower, string.dump, string.match +local getmetatable, setmetatable = getmetatable, setmetatable +local getinfo = debug.getinfo -local anything = P(1) -local endofstring = P(-1) -local alwaysmatched = P(true) +-- Starting with version 5.2 Lua no longer provide ipairs, which makes +-- sense. As we already used the for loop and # in most places the +-- impact on ConTeXt was not that large; the remaining ipairs already +-- have been replaced. In a similar fashio we also hardly used pairs. +-- +-- Just in case, we provide the fallbacks as discussed in Programming +-- in Lua (http://www.lua.org/pil/7.3.html): -patterns.anything = anything -patterns.endofstring = endofstring -patterns.beginofstring = alwaysmatched -patterns.alwaysmatched = alwaysmatched +if not ipairs then -local digit, sign = R('09'), S('+-') -local cr, lf, crlf = P("\r"), P("\n"), P("\r\n") -local newline = crlf + cr + lf -local escaped = P("\\") * anything -local squote = P("'") -local dquote = P('"') -local space = P(" ") + -- for k, v in ipairs(t) do ... end + -- for k=1,#t do local v = t[k] ... end -local utfbom_32_be = P('\000\000\254\255') -local utfbom_32_le = P('\255\254\000\000') -local utfbom_16_be = P('\255\254') -local utfbom_16_le = P('\254\255') -local utfbom_8 = P('\239\187\191') -local utfbom = utfbom_32_be + utfbom_32_le - + utfbom_16_be + utfbom_16_le - + utfbom_8 -local utftype = utfbom_32_be / "utf-32-be" + utfbom_32_le / "utf-32-le" - + utfbom_16_be / "utf-16-be" + utfbom_16_le / "utf-16-le" - + utfbom_8 / "utf-8" + alwaysmatched / "unknown" + local function iterate(a,i) + i = i + 1 + local v = a[i] + if v ~= nil then + return i, v --, nil + end + end -local utf8next = R("\128\191") + function ipairs(a) + return iterate, a, 0 + end -patterns.utf8one = R("\000\127") -patterns.utf8two = R("\194\223") * utf8next -patterns.utf8three = R("\224\239") * utf8next * utf8next -patterns.utf8four = R("\240\244") * utf8next * utf8next * utf8next -patterns.utfbom = utfbom -patterns.utftype = utftype +end -local utf8char = patterns.utf8one + patterns.utf8two + patterns.utf8three + patterns.utf8four -local validutf8char = utf8char^0 * endofstring * Cc(true) + Cc(false) +if not pairs then -patterns.utf8 = utf8char -patterns.utf8char = utf8char -patterns.validutf8 = validutf8char -patterns.validutf8char = validutf8char + -- for k, v in pairs(t) do ... end + -- for k, v in next, t do ... end -patterns.digit = digit -patterns.sign = sign -patterns.cardinal = sign^0 * digit^1 -patterns.integer = sign^0 * digit^1 -patterns.float = sign^0 * digit^0 * P('.') * digit^1 -patterns.cfloat = sign^0 * digit^0 * P(',') * digit^1 -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.letter = patterns.lowercase + patterns.uppercase -patterns.space = space -patterns.tab = P("\t") -patterns.spaceortab = patterns.space + patterns.tab -patterns.eol = S("\n\r") -patterns.spacer = S(" \t\f\v") -- + string.char(0xc2, 0xa0) if we want utf (cf mail roberto) -patterns.newline = newline -patterns.emptyline = newline^1 -patterns.nonspacer = 1 - patterns.spacer -patterns.whitespace = patterns.eol + patterns.spacer -patterns.nonwhitespace = 1 - patterns.whitespace -patterns.equal = P("=") -patterns.comma = P(",") -patterns.commaspacer = P(",") * patterns.spacer^0 -patterns.period = P(".") -patterns.colon = P(":") -patterns.semicolon = P(";") -patterns.underscore = P("_") -patterns.escaped = escaped -patterns.squote = squote -patterns.dquote = dquote -patterns.nosquote = (escaped + (1-squote))^0 -patterns.nodquote = (escaped + (1-dquote))^0 -patterns.unsingle = (squote/"") * patterns.nosquote * (squote/"") -patterns.undouble = (dquote/"") * patterns.nodquote * (dquote/"") -patterns.unquoted = patterns.undouble + patterns.unsingle -- more often undouble -patterns.unspacer = ((patterns.spacer^1)/"")^0 + function pairs(t) + return next, t -- , nil + end -patterns.somecontent = (anything - newline - space)^1 -- (utf8char - newline - space)^1 -patterns.beginline = #(1-newline) +end -local unquoted = Cs(patterns.unquoted * endofstring) -- not C +-- Also, unpack has been moved to the table table, and for compatiility +-- reasons we provide both now. -function string.unquoted(str) - return match(unquoted,str) or str +if not table.unpack then + table.unpack = _G.unpack +elseif not unpack then + _G.unpack = table.unpack end ---~ print(string.unquoted("test")) ---~ print(string.unquoted([["t\"est"]])) ---~ print(string.unquoted([["t\"est"x]])) ---~ print(string.unquoted("\'test\'")) +-- extra functions, some might go (when not used) -function lpeg.anywhere(pattern) --slightly adapted from website - return P { P(pattern) + 1 * V(1) } -- why so complex? +function table.strip(tab) + local lst, l = { }, 0 + for i=1,#tab do + local s = gsub(tab[i],"^%s*(.-)%s*$","%1") + if s == "" then + -- skip this one + else + l = l + 1 + lst[l] = s + end + end + return lst end -function lpeg.splitter(pattern, action) - return (((1-P(pattern))^1)/action+1)^0 +function table.keys(t) + local keys, k = { }, 0 + for key, _ in next, t do + k = k + 1 + keys[k] = key + end + return keys end -local splitters_s, splitters_m = { }, { } +local function compare(a,b) + local ta, tb = type(a), type(b) -- needed, else 11 < 2 + if ta == tb then + return a < b + else + return tostring(a) < tostring(b) + end +end -local function splitat(separator,single) - local splitter = (single and splitters_s[separator]) or splitters_m[separator] - if not splitter then - separator = P(separator) - local other = C((1 - separator)^0) - if single then - local any = anything - splitter = other * (separator * C(any^0) + "") -- ? - splitters_s[separator] = splitter +local function sortedkeys(tab) + local srt, category, s = { }, 0, 0 -- 0=unknown 1=string, 2=number 3=mixed + for key,_ in next, tab do + s = s + 1 + srt[s] = key + if category == 3 then + -- no further check else - splitter = other * (separator * other)^0 - splitters_m[separator] = splitter + local tkey = type(key) + if tkey == "string" then + category = (category == 2 and 3) or 1 + elseif tkey == "number" then + category = (category == 1 and 3) or 2 + else + category = 3 + end end end - return splitter + if category == 0 or category == 3 then + sort(srt,compare) + else + sort(srt) + end + return srt end -lpeg.splitat = splitat +local function sortedhashkeys(tab) -- fast one + local srt, s = { }, 0 + for key,_ in next, tab do + if key then + s= s + 1 + srt[s] = key + end + end + sort(srt) + return srt +end ---~ local p = splitat("->",false) print(match(p,"oeps->what->more")) -- oeps what more ---~ local p = splitat("->",true) print(match(p,"oeps->what->more")) -- oeps what->more ---~ local p = splitat("->",false) print(match(p,"oeps")) -- oeps ---~ local p = splitat("->",true) print(match(p,"oeps")) -- oeps +table.sortedkeys = sortedkeys +table.sortedhashkeys = sortedhashkeys -local cache = { } +local function nothing() end -function lpeg.split(separator,str) - local c = cache[separator] - if not c then - c = Ct(splitat(separator)) - cache[separator] = c +local function sortedhash(t) + if t then + local n, s = 0, sortedkeys(t) -- the robust one + local function kv(s) + n = n + 1 + local k = s[n] + return k, t[k] + end + return kv, s + else + return nothing end - return match(c,str) end -function string.split(str,separator) - local c = cache[separator] - if not c then - c = Ct(splitat(separator)) - cache[separator] = c - end - return match(c,str) -end +table.sortedhash = sortedhash +table.sortedpairs = sortedhash -local spacing = patterns.spacer^0 * newline -- sort of strip -local empty = spacing * Cc("") -local nonempty = Cs((1-spacing)^1) * spacing^-1 -local content = (empty + nonempty)^1 - -patterns.textline = content - ---~ local linesplitter = Ct(content^0) ---~ ---~ function string.splitlines(str) ---~ return match(linesplitter,str) ---~ end - -local linesplitter = Ct(splitat(newline)) - -patterns.linesplitter = linesplitter - -function string.splitlines(str) - return match(linesplitter,str) -end - -local utflinesplitter = utfbom^-1 * Ct(splitat(newline)) - -patterns.utflinesplitter = utflinesplitter - -function string.utfsplitlines(str) - return match(utflinesplitter,str) +function table.append(t, list) + local n = #t + for i=1,#list do + n = n + 1 + t[n] = list[i] + end + return t end ---~ lpeg.splitters = cache -- no longer public - -local cache = { } - -function lpeg.checkedsplit(separator,str) - local c = cache[separator] - if not c then - separator = P(separator) - local other = C((1 - separator)^1) - c = Ct(separator^0 * other * (separator^1 * other)^0) - cache[separator] = c +function table.prepend(t, list) + local nl = #list + local nt = nl + #t + for i=#t,1,-1 do + t[nt] = t[i] + nt = nt - 1 end - return match(c,str) + for i=1,#list do + t[i] = list[i] + end + return t end -function string.checkedsplit(str,separator) - local c = cache[separator] - if not c then - separator = P(separator) - local other = C((1 - separator)^1) - c = Ct(separator^0 * other * (separator^1 * other)^0) - cache[separator] = c +function table.merge(t, ...) -- first one is target + t = t or { } + local lst = { ... } + for i=1,#lst do + for k, v in next, lst[i] do + t[k] = v + end end - return match(c,str) + return t end ---~ from roberto's site: - -local f1 = string.byte - -local function f2(s) local c1, c2 = f1(s,1,2) return c1 * 64 + c2 - 12416 end -local function f3(s) local c1, c2, c3 = f1(s,1,3) return (c1 * 64 + c2) * 64 + c3 - 925824 end -local function f4(s) local c1, c2, c3, c4 = f1(s,1,4) return ((c1 * 64 + c2) * 64 + c3) * 64 + c4 - 63447168 end - -local utf8byte = patterns.utf8one/f1 + patterns.utf8two/f2 + patterns.utf8three/f3 + patterns.utf8four/f4 - -patterns.utf8byte = utf8byte - ---~ local str = " a b c d " - ---~ local s = lpeg.stripper(lpeg.R("az")) print("["..lpeg.match(s,str).."]") ---~ local s = lpeg.keeper(lpeg.R("az")) print("["..lpeg.match(s,str).."]") ---~ local s = lpeg.stripper("ab") print("["..lpeg.match(s,str).."]") ---~ local s = lpeg.keeper("ab") print("["..lpeg.match(s,str).."]") - -local cache = { } - -function lpeg.stripper(str) - if type(str) == "string" then - local s = cache[str] - if not s then - s = Cs(((S(str)^1)/"" + 1)^0) - cache[str] = s +function table.merged(...) + local tmp, lst = { }, { ... } + for i=1,#lst do + for k, v in next, lst[i] do + tmp[k] = v end - return s - else - return Cs(((str^1)/"" + 1)^0) end + return tmp end -local cache = { } - -function lpeg.keeper(str) - if type(str) == "string" then - local s = cache[str] - if not s then - s = Cs((((1-S(str))^1)/"" + 1)^0) - cache[str] = s +function table.imerge(t, ...) + local lst, nt = { ... }, #t + for i=1,#lst do + local nst = lst[i] + for j=1,#nst do + nt = nt + 1 + t[nt] = nst[j] end - return s - else - return Cs((((1-str)^1)/"" + 1)^0) end + return t end -function lpeg.frontstripper(str) -- or pattern (yet undocumented) - return (P(str) + P(true)) * Cs(P(1)^0) +function table.imerged(...) + local tmp, ntmp, lst = { }, 0, {...} + for i=1,#lst do + local nst = lst[i] + for j=1,#nst do + ntmp = ntmp + 1 + tmp[ntmp] = nst[j] + end + end + return tmp end -function lpeg.endstripper(str) -- or pattern (yet undocumented) - return Cs((1 - P(str) * P(-1))^0) +local function fastcopy(old,metatabletoo) -- fast one + if old then + local new = { } + for k,v in next, old do + if type(v) == "table" then + new[k] = fastcopy(v,metatabletoo) -- was just table.copy + else + new[k] = v + end + end + if metatabletoo then + -- optional second arg + local mt = getmetatable(old) + if mt then + setmetatable(new,mt) + end + end + return new + else + return { } + end end --- Just for fun I looked at the used bytecode and --- p = (p and p + pp) or pp gets one more (testset). +-- todo : copy without metatable -function lpeg.replacer(one,two) - if type(one) == "table" then - local no = #one - if no > 0 then - local p - for i=1,no do - local o = one[i] - local pp = P(o[1]) / o[2] - if p then - p = p + pp - else - p = pp - end +local function copy(t, tables) -- taken from lua wiki, slightly adapted + tables = tables or { } + local tcopy = {} + if not tables[t] then + tables[t] = tcopy + end + for i,v in next, t do -- brrr, what happens with sparse indexed + if type(i) == "table" then + if tables[i] then + i = tables[i] + else + i = copy(i, tables) end - return Cs((p + 1)^0) end - else - two = two or "" - return Cs((P(one)/two + 1)^0) + if type(v) ~= "table" then + tcopy[i] = v + elseif tables[v] then + tcopy[i] = tables[v] + else + tcopy[i] = copy(v, tables) + end end + local mt = getmetatable(t) + if mt then + setmetatable(tcopy,mt) + end + return tcopy end -local splitters_f, splitters_s = { }, { } +table.fastcopy = fastcopy +table.copy = copy -function lpeg.firstofsplit(separator) -- always return value - local splitter = splitters_f[separator] - if not splitter then - separator = P(separator) - splitter = C((1 - separator)^0) - splitters_f[separator] = splitter +function table.derive(parent) + local child = { } + if parent then + setmetatable(child,{ __index = parent }) end - return splitter + return child 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) - splitters_s[separator] = splitter +function table.tohash(t,value) + local h = { } + if t then + if value == nil then value = true end + for _, v in next, t do -- no ipairs here + h[v] = value + end end - return splitter + return h end -function lpeg.balancer(left,right) - left, right = P(left), P(right) - return P { left * ((1 - left - right) + V(1))^0 * right } +function table.fromhash(t) + local hsh, h = { }, 0 + for k, v in next, t do -- no ipairs here + if v then + h = h + 1 + hsh[h] = k + end + end + return hsh end ---~ print(1,match(lpeg.firstofsplit(":"),"bc:de")) ---~ print(2,match(lpeg.firstofsplit(":"),":de")) -- empty ---~ print(3,match(lpeg.firstofsplit(":"),"bc")) ---~ print(4,match(lpeg.secondofsplit(":"),"bc:de")) ---~ print(5,match(lpeg.secondofsplit(":"),"bc:")) -- empty ---~ print(6,match(lpeg.secondofsplit(":",""),"bc")) ---~ print(7,match(lpeg.secondofsplit(":"),"bc")) ---~ print(9,match(lpeg.secondofsplit(":","123"),"bc")) - ---~ -- slower: ---~ ---~ function lpeg.counter(pattern) ---~ local n, pattern = 0, (lpeg.P(pattern)/function() n = n + 1 end + lpeg.anything)^0 ---~ return function(str) n = 0 ; lpegmatch(pattern,str) ; return n end ---~ end +local noquotes, hexify, handle, reduce, compact, inline, functions -local nany = utf8char/"" - -function lpeg.counter(pattern) - pattern = Cs((P(pattern)/" " + nany)^0) - return function(str) - return #match(pattern,str) - end -end - -if utfgmatch then +local reserved = table.tohash { -- intercept a language inconvenience: no reserved words as key + 'and', 'break', 'do', 'else', 'elseif', 'end', 'false', 'for', 'function', 'if', + 'in', 'local', 'nil', 'not', 'or', 'repeat', 'return', 'then', 'true', 'until', 'while', +} - function lpeg.count(str,what) -- replaces string.count - if type(what) == "string" then - local n = 0 - for _ in utfgmatch(str,what) do - n = n + 1 - end - return n - else -- 4 times slower but still faster than / function - return #match(Cs((P(what)/" " + nany)^0),str) +local function simple_table(t) + if #t > 0 then + local n = 0 + for _,v in next, t do + n = n + 1 end - end - -else - - local cache = { } - - function lpeg.count(str,what) -- replaces string.count - if type(what) == "string" then - local p = cache[what] - if not p then - p = Cs((P(what)/" " + nany)^0) - cache[p] = p + if n == #t then + local tt, nt = { }, 0 + for i=1,#t do + local v = t[i] + local tv = type(v) + if tv == "number" then + nt = nt + 1 + if hexify then + tt[nt] = format("0x%04X",v) + else + tt[nt] = tostring(v) -- tostring not needed + end + elseif tv == "boolean" then + nt = nt + 1 + tt[nt] = tostring(v) + elseif tv == "string" then + nt = nt + 1 + tt[nt] = format("%q",v) + else + tt = nil + break + end end - return #match(p,str) - else -- 4 times slower but still faster than / function - return #match(Cs((P(what)/" " + nany)^0),str) + return tt end end - -end - -local patterns_escapes = { -- also defines in l-string - ["%"] = "%%", - ["."] = "%.", - ["+"] = "%+", ["-"] = "%-", ["*"] = "%*", - ["["] = "%[", ["]"] = "%]", - ["("] = "%)", [")"] = "%)", - -- ["{"] = "%{", ["}"] = "%}" - -- ["^"] = "%^", ["$"] = "%$", -} - -local simple_escapes = { -- also defines in l-string - ["-"] = "%-", - ["."] = "%.", - ["?"] = ".", - ["*"] = ".*", -} - -local p = Cs((S("-.+*%()[]") / patterns_escapes + anything)^0) -local s = Cs((S("-.+*%()[]") / simple_escapes + anything)^0) - -function string.escapedpattern(str,simple) - return match(simple and s or p,str) + return nil end --- utf extensies - -lpeg.UP = lpeg.P - -if utfcharacters then - - function lpeg.US(str) - local p - for uc in utfcharacters(str) do - if p then - p = p + P(uc) - else - p = P(uc) - end - end - return p - end - - -elseif utfgmatch then +-- Because this is a core function of mkiv I moved some function calls +-- inline. +-- +-- twice as fast in a test: +-- +-- local propername = lpeg.P(lpeg.R("AZ","az","__") * lpeg.R("09","AZ","az", "__")^0 * lpeg.P(-1) ) - function lpeg.US(str) - local p - for uc in utfgmatch(str,".") do - if p then - p = p + P(uc) - else - p = P(uc) - end - end - return p - end +-- problem: there no good number_to_string converter with the best resolution -else +local function dummy() end - function lpeg.US(str) - local p - local f = function(uc) - if p then - p = p + P(uc) +local function do_serialize(root,name,depth,level,indexed) + if level > 0 then + depth = depth .. " " + if indexed then + handle(format("%s{",depth)) + else + local tn = type(name) + if tn == "number" then -- or find(k,"^%d+$") then + if hexify then + handle(format("%s[0x%04X]={",depth,name)) + else + handle(format("%s[%s]={",depth,name)) + end + elseif tn == "string" then + if noquotes and not reserved[name] and find(name,"^%a[%w%_]*$") then + handle(format("%s%s={",depth,name)) + else + handle(format("%s[%q]={",depth,name)) + end + elseif tn == "boolean" then + handle(format("%s[%s]={",depth,tostring(name))) else - p = P(uc) + handle(format("%s{",depth)) end end - match((utf8char/f)^0,str) - return p - end - -end - -local range = Cs(utf8byte) * (Cs(utf8byte) + Cc(false)) - -local utfchar = unicode and unicode.utf8 and unicode.utf8.char - -function lpeg.UR(str,more) - local first, last - if type(str) == "number" then - first = str - last = more or first - else - first, last = match(range,str) - if not last then - return P(str) - end end - if first == last then - return P(str) - elseif utfchar and last - first < 8 then -- a somewhat arbitrary criterium - local p - for i=first,last do - if p then - p = p + P(utfchar(i)) - else - p = P(utfchar(i)) + -- we could check for k (index) being number (cardinal) + if root and next(root) then + local first, last = nil, 0 -- #root cannot be trusted here (will be ok in 5.2 when ipairs is gone) + if compact then + -- NOT: for k=1,#root do (we need to quit at nil) + for k,v in ipairs(root) do -- can we use next? + if not first then first = k end + last = last + 1 end end - return p -- nil when invalid range - else - local f = function(b) - return b >= first and b <= last - end - return utf8byte / f -- nil when invalid range - end -end - ---~ lpeg.print(lpeg.R("ab","cd","gh")) ---~ lpeg.print(lpeg.P("a","b","c")) ---~ lpeg.print(lpeg.S("a","b","c")) - ---~ print(lpeg.count("äáàa",lpeg.P("á") + lpeg.P("à"))) ---~ print(lpeg.count("äáàa",lpeg.UP("áà"))) ---~ print(lpeg.count("äáàa",lpeg.US("àá"))) ---~ print(lpeg.count("äáàa",lpeg.UR("aá"))) ---~ print(lpeg.count("äáàa",lpeg.UR("àá"))) ---~ print(lpeg.count("äáàa",lpeg.UR(0x0000,0xFFFF))) - -function lpeg.oneof(list,...) -- lpeg.oneof("elseif","else","if","then") - if type(list) ~= "table" then - list = { list, ... } - end - -- sort(list) -- longest match first - local p = P(list[1]) - for l=2,#list do - p = p + P(list[l]) - end - return p -end - -function lpeg.is_lpeg(p) - return p and lpegtype(p) == "pattern" -end - ---~ Cf(Ct("") * (Cg(C(...) * "=" * Cs(...)))^0, rawset) - -end -- closure - -do -- begin closure to overcome local limits and interference - -if not modules then modules = { } end modules ['l-boolean'] = { - version = 1.001, - comment = "companion to luat-lib.mkiv", - author = "Hans Hagen, PRAGMA-ADE, Hasselt NL", - copyright = "PRAGMA ADE / ConTeXt Development Team", - license = "see context related readme files" -} - -local type, tonumber = type, tonumber - -boolean = boolean or { } -local boolean = boolean - --- function boolean.tonumber(b) --- return b and 1 or 0 -- test and test and return or return --- end - -function boolean.tonumber(b) - if b then return 1 else return 0 end -- test and return or return -end + local sk = sortedkeys(root) + for i=1,#sk do + local k = sk[i] + local v = root[k] + --~ if v == root then + -- circular + --~ else + local t, tk = type(v), type(k) + if compact and first and tk == "number" and k >= first and k <= last then + if t == "number" then + if hexify then + handle(format("%s 0x%04X,",depth,v)) + else + handle(format("%s %s,",depth,v)) -- %.99g + end + elseif t == "string" then + if reduce and tonumber(v) then + handle(format("%s %s,",depth,v)) + else + handle(format("%s %q,",depth,v)) + end + elseif t == "table" then + if not next(v) then + handle(format("%s {},",depth)) + elseif inline then -- and #t > 0 + local st = simple_table(v) + if st then + handle(format("%s { %s },",depth,concat(st,", "))) + else + do_serialize(v,k,depth,level+1,true) + end + else + do_serialize(v,k,depth,level+1,true) + end + elseif t == "boolean" then + handle(format("%s %s,",depth,tostring(v))) + elseif t == "function" then + if functions then + handle(format('%s loadstring(%q),',depth,dump(v))) + else + handle(format('%s "function",',depth)) + end + else + handle(format("%s %q,",depth,tostring(v))) + end + elseif k == "__p__" then -- parent + if false then + handle(format("%s __p__=nil,",depth)) + end + elseif t == "number" then + if tk == "number" then -- or find(k,"^%d+$") then + if hexify then + handle(format("%s [0x%04X]=0x%04X,",depth,k,v)) + else + handle(format("%s [%s]=%s,",depth,k,v)) -- %.99g + end + elseif tk == "boolean" then + if hexify then + handle(format("%s [%s]=0x%04X,",depth,tostring(k),v)) + else + handle(format("%s [%s]=%s,",depth,tostring(k),v)) -- %.99g + end + elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then + if hexify then + handle(format("%s %s=0x%04X,",depth,k,v)) + else + handle(format("%s %s=%s,",depth,k,v)) -- %.99g + end + else + if hexify then + handle(format("%s [%q]=0x%04X,",depth,k,v)) + else + handle(format("%s [%q]=%s,",depth,k,v)) -- %.99g + end + end + elseif t == "string" then + if reduce and tonumber(v) then + if tk == "number" then -- or find(k,"^%d+$") then + if hexify then + handle(format("%s [0x%04X]=%s,",depth,k,v)) + else + handle(format("%s [%s]=%s,",depth,k,v)) + end + elseif tk == "boolean" then + handle(format("%s [%s]=%s,",depth,tostring(k),v)) + elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then + handle(format("%s %s=%s,",depth,k,v)) + else + handle(format("%s [%q]=%s,",depth,k,v)) + end + else + if tk == "number" then -- or find(k,"^%d+$") then + if hexify then + handle(format("%s [0x%04X]=%q,",depth,k,v)) + else + handle(format("%s [%s]=%q,",depth,k,v)) + end + elseif tk == "boolean" then + handle(format("%s [%s]=%q,",depth,tostring(k),v)) + elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then + handle(format("%s %s=%q,",depth,k,v)) + else + handle(format("%s [%q]=%q,",depth,k,v)) + end + end + elseif t == "table" then + if not next(v) then + if tk == "number" then -- or find(k,"^%d+$") then + if hexify then + handle(format("%s [0x%04X]={},",depth,k)) + else + handle(format("%s [%s]={},",depth,k)) + end + elseif tk == "boolean" then + handle(format("%s [%s]={},",depth,tostring(k))) + elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then + handle(format("%s %s={},",depth,k)) + else + handle(format("%s [%q]={},",depth,k)) + end + elseif inline then + local st = simple_table(v) + if st then + if tk == "number" then -- or find(k,"^%d+$") then + if hexify then + handle(format("%s [0x%04X]={ %s },",depth,k,concat(st,", "))) + else + handle(format("%s [%s]={ %s },",depth,k,concat(st,", "))) + end + elseif tk == "boolean" then -- or find(k,"^%d+$") then + handle(format("%s [%s]={ %s },",depth,tostring(k),concat(st,", "))) + elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then + handle(format("%s %s={ %s },",depth,k,concat(st,", "))) + else + handle(format("%s [%q]={ %s },",depth,k,concat(st,", "))) + end + else + do_serialize(v,k,depth,level+1) + end + else + do_serialize(v,k,depth,level+1) + end + elseif t == "boolean" then + if tk == "number" then -- or find(k,"^%d+$") then + if hexify then + handle(format("%s [0x%04X]=%s,",depth,k,tostring(v))) + else + handle(format("%s [%s]=%s,",depth,k,tostring(v))) + end + elseif tk == "boolean" then -- or find(k,"^%d+$") then + handle(format("%s [%s]=%s,",depth,tostring(k),tostring(v))) + elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then + handle(format("%s %s=%s,",depth,k,tostring(v))) + else + handle(format("%s [%q]=%s,",depth,k,tostring(v))) + end + elseif t == "function" then + if functions then + local f = getinfo(v).what == "C" and dump(dummy) or dump(v) + -- local f = getinfo(v).what == "C" and dump(function(...) return v(...) end) or dump(v) + if tk == "number" then -- or find(k,"^%d+$") then + if hexify then + handle(format("%s [0x%04X]=loadstring(%q),",depth,k,f)) + else + handle(format("%s [%s]=loadstring(%q),",depth,k,f)) + end + elseif tk == "boolean" then + handle(format("%s [%s]=loadstring(%q),",depth,tostring(k),f)) + elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then + handle(format("%s %s=loadstring(%q),",depth,k,f)) + else + handle(format("%s [%q]=loadstring(%q),",depth,k,f)) + end + end + else + if tk == "number" then -- or find(k,"^%d+$") then + if hexify then + handle(format("%s [0x%04X]=%q,",depth,k,tostring(v))) + else + handle(format("%s [%s]=%q,",depth,k,tostring(v))) + end + elseif tk == "boolean" then -- or find(k,"^%d+$") then + handle(format("%s [%s]=%q,",depth,tostring(k),tostring(v))) + elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then + handle(format("%s %s=%q,",depth,k,tostring(v))) + else + handle(format("%s [%q]=%q,",depth,k,tostring(v))) + end + end + --~ end + end + end + if level > 0 then + handle(format("%s},",depth)) + end +end + +-- replacing handle by a direct t[#t+1] = ... (plus test) is not much +-- faster (0.03 on 1.00 for zapfino.tma) + +local function serialize(_handle,root,name,specification) -- handle wins + local tname = type(name) + if type(specification) == "table" then + noquotes = specification.noquotes + hexify = specification.hexify + handle = _handle or specification.handle or print + reduce = specification.reduce or false + functions = specification.functions + compact = specification.compact + inline = specification.inline and compact + if functions == nil then + functions = true + end + if compact == nil then + compact = true + end + if inline == nil then + inline = compact + end + else + noquotes = false + hexify = false + handle = _handle or print + reduce = false + compact = true + inline = true + functions = true + end + if tname == "string" then + if name == "return" then + handle("return {") + else + handle(name .. "={") + end + elseif tname == "number" then + if hexify then + handle(format("[0x%04X]={",name)) + else + handle("[" .. name .. "]={") + end + elseif tname == "boolean" then + if name then + handle("return {") + else + handle("{") + end + else + handle("t={") + end + if root then + -- The dummy access will initialize a table that has a delayed initialization + -- using a metatable. (maybe explicitly test for metatable) + if getmetatable(root) then -- todo: make this an option, maybe even per subtable + local dummy = root._w_h_a_t_e_v_e_r_ + root._w_h_a_t_e_v_e_r_ = nil + end + -- Let's forget about empty tables. + if next(root) then + do_serialize(root,name,"",0) + end + end + handle("}") +end + +--~ name: +--~ +--~ true : return { } +--~ false : { } +--~ nil : t = { } +--~ string : string = { } +--~ 'return' : return { } +--~ number : [number] = { } + +function table.serialize(root,name,specification) + local t, n = { }, 0 + local function flush(s) + n = n + 1 + t[n] = s + end + serialize(flush,root,name,specification) + return concat(t,"\n") +end + +table.tohandle = serialize + +-- sometimes tables are real use (zapfino extra pro is some 85M) in which +-- case a stepwise serialization is nice; actually, we could consider: +-- +-- for line in table.serializer(root,name,reduce,noquotes) do +-- ...(line) +-- end +-- +-- so this is on the todo list + +local maxtab = 2*1024 + +function table.tofile(filename,root,name,specification) + local f = io.open(filename,'w') + if f then + if maxtab > 1 then + local t, n = { }, 0 + local function flush(s) + n = n + 1 + t[n] = s + if n > maxtab then + f:write(concat(t,"\n"),"\n") -- hm, write(sometable) should be nice + t, n = { }, 0 -- we could recycle t if needed + end + end + serialize(flush,root,name,specification) + f:write(concat(t,"\n"),"\n") + else + local function flush(s) + f:write(s,"\n") + end + serialize(flush,root,name,specification) + end + f:close() + io.flush() + end +end -function toboolean(str,tolerant) - if tolerant then - local tstr = type(str) - if tstr == "string" then - return str == "true" or str == "yes" or str == "on" or str == "1" or str == "t" - elseif tstr == "number" then - return tonumber(str) ~= 0 - elseif tstr == "nil" then - return false +local function flattened(t,f,depth) + if f == nil then + f = { } + depth = 0xFFFF + elseif tonumber(f) then + -- assume then only two arguments are given + depth = f + f = { } + elseif not depth then + depth = 0xFFFF + end + for k, v in next, t do + if type(k) ~= "number" then + if depth > 0 and type(v) == "table" then + flattened(v,f,depth-1) + else + f[k] = v + end + end + end + local n = #f + for k=1,#t do + local v = t[k] + if depth > 0 and type(v) == "table" then + flattened(v,f,depth-1) + n = #f else - return str + n = n + 1 + f[n] = v + end + end + return f +end + +table.flattened = flattened + +local function unnest(t,f) -- only used in mk, for old times sake + if not f then -- and only relevant for token lists + f = { } + end + for i=1,#t do + local v = t[i] + if type(v) == "table" then + if type(v[1]) == "table" then + unnest(v,f) + else + f[#f+1] = v + end + else + f[#f+1] = v + end + end + return f +end + +function table.unnest(t) -- bad name + return unnest(t) +end + +local function are_equal(a,b,n,m) -- indexed + if a and b and #a == #b then + n = n or 1 + m = m or #a + for i=n,m do + local ai, bi = a[i], b[i] + if ai==bi then + -- same + elseif type(ai)=="table" and type(bi)=="table" then + if not are_equal(ai,bi) then + return false + end + else + return false + end end - elseif str == "true" then return true - elseif str == "false" then - return false else - return str + return false end end -string.toboolean = toboolean - -function string.is_boolean(str,default) - if type(str) == "string" then - if str == "true" or str == "yes" or str == "on" or str == "t" then - return true - elseif str == "false" or str == "no" or str == "off" or str == "f" then +local function identical(a,b) -- assumes same structure + for ka, va in next, a do + local vb = b[ka] + if va == vb then + -- same + elseif type(va) == "table" and type(vb) == "table" then + if not identical(va,vb) then + return false + end + else return false end end - return default + return true end -end -- closure +table.identical = identical +table.are_equal = are_equal -do -- begin closure to overcome local limits and interference +-- maybe also make a combined one -if not modules then modules = { } end modules ['l-math'] = { - version = 1.001, - comment = "companion to luat-lib.mkiv", - author = "Hans Hagen, PRAGMA-ADE, Hasselt NL", - copyright = "PRAGMA ADE / ConTeXt Development Team", - license = "see context related readme files" -} +function table.compact(t) + if t then + for k,v in next, t do + if not next(v) then + t[k] = nil + end + end + end +end -local floor, sin, cos, tan = math.floor, math.sin, math.cos, math.tan +function table.contains(t, v) + if t then + for i=1, #t do + if t[i] == v then + return i + end + end + end + return false +end -if not math.round then - function math.round(x) return floor(x + 0.5) end +function table.count(t) + local n = 0 + for k, v in next, t do + n = n + 1 + end + return n end -if not math.div then - function math.div(n,m) return floor(n/m) end +function table.swapped(t,s) -- hash + local n = { } + if s then +--~ for i=1,#s do +--~ n[i] = s[i] +--~ end + for k, v in next, s do + n[k] = v + end + end +--~ for i=1,#t do +--~ local ti = t[i] -- don't ask but t[i] can be nil +--~ if ti then +--~ n[ti] = i +--~ end +--~ end + for k, v in next, t do + n[v] = k + end + return n end -if not math.mod then - function math.mod(n,m) return n % m end +function table.reversed(t) + if t then + local tt, tn = { }, #t + if tn > 0 then + local ttn = 0 + for i=tn,1,-1 do + ttn = ttn + 1 + tt[ttn] = t[i] + end + end + return tt + end end -local pipi = 2*math.pi/360 +function table.sequenced(t,sep,simple) -- hash only + local s, n = { }, 0 + for k, v in sortedhash(t) do + if simple then + if v == true then + n = n + 1 + s[n] = k + elseif v and v~= "" then + n = n + 1 + s[n] = k .. "=" .. tostring(v) + end + else + n = n + 1 + s[n] = k .. "=" .. tostring(v) + end + end + return concat(s, sep or " | ") +end -if not math.sind then - function math.sind(d) return sin(d*pipi) end - function math.cosd(d) return cos(d*pipi) end - function math.tand(d) return tan(d*pipi) end +function table.print(t,...) + if type(t) ~= "table" then + print(tostring(t)) + else + table.tohandle(print,t,...) + end end -if not math.odd then - function math.odd (n) return n % 2 == 0 end - function math.even(n) return n % 2 ~= 0 end +-- -- -- obsolete but we keep them for a while and might comment them later -- -- -- + +-- roughly: copy-loop : unpack : sub == 0.9 : 0.4 : 0.45 (so in critical apps, use unpack) + +function table.sub(t,i,j) + return { unpack(t,i,j) } +end + +-- slower than #t on indexed tables (#t only returns the size of the numerically indexed slice) + +function table.is_empty(t) + return not t or not next(t) +end + +function table.has_one_entry(t) + return t and not next(t,next(t)) +end + +-- new + +function table.loweredkeys(t) -- maybe utf + local l = { } + for k, v in next, t do + l[lower(k)] = v + end + return l end end -- closure do -- begin closure to overcome local limits and interference -if not modules then modules = { } end modules ['l-table'] = { +if not modules then modules = { } end modules ['l-lpeg'] = { version = 1.001, comment = "companion to luat-lib.mkiv", author = "Hans Hagen, PRAGMA-ADE, Hasselt NL", @@ -779,928 +1081,689 @@ if not modules then modules = { } end modules ['l-table'] = { license = "see context related readme files" } -local type, next, tostring, tonumber, ipairs, table, string = type, next, tostring, tonumber, ipairs, table, string -local concat, sort, insert, remove = table.concat, table.sort, table.insert, table.remove -local format, find, gsub, lower, dump, match = string.format, string.find, string.gsub, string.lower, string.dump, string.match -local getmetatable, setmetatable = getmetatable, setmetatable -local getinfo = debug.getinfo +local lpeg = require("lpeg") + +local type = type +local byte, char = string.byte, string.char + +-- Beware, we predefine a bunch of patterns here and one reason for doing so +-- is that we get consistent behaviour in some of the visualizers. --- Starting with version 5.2 Lua no longer provide ipairs, which makes --- sense. As we already used the for loop and # in most places the --- impact on ConTeXt was not that large; the remaining ipairs already --- have been replaced. In a similar fashio we also hardly used pairs. --- --- Just in case, we provide the fallbacks as discussed in Programming --- in Lua (http://www.lua.org/pil/7.3.html): +lpeg.patterns = lpeg.patterns or { } -- so that we can share +local patterns = lpeg.patterns -if not ipairs then +local P, R, S, V, match = lpeg.P, lpeg.R, lpeg.S, lpeg.V, lpeg.match +local Ct, C, Cs, Cc = lpeg.Ct, lpeg.C, lpeg.Cs, lpeg.Cc +local lpegtype = lpeg.type - -- for k, v in ipairs(t) do ... end - -- for k=1,#t do local v = t[k] ... end +local utfcharacters = string.utfcharacters +local utfgmatch = unicode and unicode.utf8.gmatch - local function iterate(a,i) - i = i + 1 - local v = a[i] - if v ~= nil then - return i, v --, nil - end - end +local anything = P(1) +local endofstring = P(-1) +local alwaysmatched = P(true) - function ipairs(a) - return iterate, a, 0 - end +patterns.anything = anything +patterns.endofstring = endofstring +patterns.beginofstring = alwaysmatched +patterns.alwaysmatched = alwaysmatched -end +local digit, sign = R('09'), S('+-') +local cr, lf, crlf = P("\r"), P("\n"), P("\r\n") +local newline = crlf + cr + lf +local escaped = P("\\") * anything +local squote = P("'") +local dquote = P('"') +local space = P(" ") -if not pairs then +local utfbom_32_be = P('\000\000\254\255') +local utfbom_32_le = P('\255\254\000\000') +local utfbom_16_be = P('\255\254') +local utfbom_16_le = P('\254\255') +local utfbom_8 = P('\239\187\191') +local utfbom = utfbom_32_be + utfbom_32_le + + utfbom_16_be + utfbom_16_le + + utfbom_8 +local utftype = utfbom_32_be / "utf-32-be" + utfbom_32_le / "utf-32-le" + + utfbom_16_be / "utf-16-be" + utfbom_16_le / "utf-16-le" + + utfbom_8 / "utf-8" + alwaysmatched / "unknown" - -- for k, v in pairs(t) do ... end - -- for k, v in next, t do ... end +local utf8next = R("\128\191") - function pairs(t) - return next, t -- , nil - end +patterns.utf8one = R("\000\127") +patterns.utf8two = R("\194\223") * utf8next +patterns.utf8three = R("\224\239") * utf8next * utf8next +patterns.utf8four = R("\240\244") * utf8next * utf8next * utf8next +patterns.utfbom = utfbom +patterns.utftype = utftype -end +local utf8char = patterns.utf8one + patterns.utf8two + patterns.utf8three + patterns.utf8four +local validutf8char = utf8char^0 * endofstring * Cc(true) + Cc(false) --- Also, unpack has been moved to the table table, and for compatiility --- reasons we provide both now. +patterns.utf8 = utf8char +patterns.utf8char = utf8char +patterns.validutf8 = validutf8char +patterns.validutf8char = validutf8char -if not table.unpack then - table.unpack = _G.unpack -elseif not unpack then - _G.unpack = table.unpack -end +patterns.digit = digit +patterns.sign = sign +patterns.cardinal = sign^0 * digit^1 +patterns.integer = sign^0 * digit^1 +patterns.float = sign^0 * digit^0 * P('.') * digit^1 +patterns.cfloat = sign^0 * digit^0 * P(',') * digit^1 +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.letter = patterns.lowercase + patterns.uppercase +patterns.space = space +patterns.tab = P("\t") +patterns.spaceortab = patterns.space + patterns.tab +patterns.eol = S("\n\r") +patterns.spacer = S(" \t\f\v") -- + char(0xc2, 0xa0) if we want utf (cf mail roberto) +patterns.newline = newline +patterns.emptyline = newline^1 +patterns.nonspacer = 1 - patterns.spacer +patterns.whitespace = patterns.eol + patterns.spacer +patterns.nonwhitespace = 1 - patterns.whitespace +patterns.equal = P("=") +patterns.comma = P(",") +patterns.commaspacer = P(",") * patterns.spacer^0 +patterns.period = P(".") +patterns.colon = P(":") +patterns.semicolon = P(";") +patterns.underscore = P("_") +patterns.escaped = escaped +patterns.squote = squote +patterns.dquote = dquote +patterns.nosquote = (escaped + (1-squote))^0 +patterns.nodquote = (escaped + (1-dquote))^0 +patterns.unsingle = (squote/"") * patterns.nosquote * (squote/"") +patterns.undouble = (dquote/"") * patterns.nodquote * (dquote/"") +patterns.unquoted = patterns.undouble + patterns.unsingle -- more often undouble +patterns.unspacer = ((patterns.spacer^1)/"")^0 --- extra functions, some might go (when not used) +patterns.somecontent = (anything - newline - space)^1 -- (utf8char - newline - space)^1 +patterns.beginline = #(1-newline) -function table.strip(tab) - local lst, l = { }, 0 - for i=1,#tab do - local s = gsub(tab[i],"^%s*(.-)%s*$","%1") - if s == "" then - -- skip this one - else - l = l + 1 - lst[l] = s - end - end - return lst -end +local unquoted = Cs(patterns.unquoted * endofstring) -- not C -function table.keys(t) - local keys, k = { }, 0 - for key, _ in next, t do - k = k + 1 - keys[k] = key - end - return keys +function string.unquoted(str) + return match(unquoted,str) or str end -local function compare(a,b) - local ta, tb = type(a), type(b) -- needed, else 11 < 2 - if ta == tb then - return a < b - else - return tostring(a) < tostring(b) - end -end +-- more efficient: -local function sortedkeys(tab) - local srt, category, s = { }, 0, 0 -- 0=unknown 1=string, 2=number 3=mixed - for key,_ in next, tab do - s = s + 1 - srt[s] = key - if category == 3 then - -- no further check - else - local tkey = type(key) - if tkey == "string" then - category = (category == 2 and 3) or 1 - elseif tkey == "number" then - category = (category == 1 and 3) or 2 - else - category = 3 - end - end - end - if category == 0 or category == 3 then - sort(srt,compare) - else - sort(srt) - end - return srt -end +local unquoted = ( + squote * Cs(1 - P(-2)) * squote + + dquote * Cs(1 - P(-2)) * dquote +) -local function sortedhashkeys(tab) -- fast one - local srt, s = { }, 0 - for key,_ in next, tab do - if key then - s= s + 1 - srt[s] = key - end - end - sort(srt) - return srt +function string.unquoted(str) + return match(unquoted,str) or str end -table.sortedkeys = sortedkeys -table.sortedhashkeys = sortedhashkeys +patterns.unquoted = unquoted -local function nothing() end +--~ print(string.unquoted("test")) +--~ print(string.unquoted([["t\"est"]])) +--~ print(string.unquoted([["t\"est"x]])) +--~ print(string.unquoted("\'test\'")) -local function sortedhash(t) - if t then - local n, s = 0, sortedkeys(t) -- the robust one - local function kv(s) - n = n + 1 - local k = s[n] - return k, t[k] - end - return kv, s - else - return nothing - end +function lpeg.anywhere(pattern) --slightly adapted from website + return P { P(pattern) + 1 * V(1) } -- why so complex? end -table.sortedhash = sortedhash -table.sortedpairs = sortedhash - -function table.append(t, list) - local n = #t - for i=1,#list do - n = n + 1 - t[n] = list[i] - end - return t +function lpeg.splitter(pattern, action) + return (((1-P(pattern))^1)/action+1)^0 end -function table.prepend(t, list) - local nl = #list - local nt = nl + #t - for i=#t,1,-1 do - t[nt] = t[i] - nt = nt - 1 - end - for i=1,#list do - t[i] = list[i] - end - return t +function lpeg.tsplitter(pattern, action) + return Ct((((1-P(pattern))^1)/action+1)^0) end -function table.merge(t, ...) -- first one is target - t = t or { } - local lst = { ... } - for i=1,#lst do - for k, v in next, lst[i] do - t[k] = v +-- probleem: separator can be lpeg and that does not hash too well, but +-- it's quite okay as the key is then not garbage collected + +local splitters_s, splitters_m, splitters_t = { }, { }, { } + +local function splitat(separator,single) + local splitter = (single and splitters_s[separator]) or splitters_m[separator] + if not splitter then + separator = P(separator) + local other = C((1 - separator)^0) + if single then + local any = anything + splitter = other * (separator * C(any^0) + "") -- ? + splitters_s[separator] = splitter + else + splitter = other * (separator * other)^0 + splitters_m[separator] = splitter end end - return t + return splitter end -function table.merged(...) - local tmp, lst = { }, { ... } - for i=1,#lst do - for k, v in next, lst[i] do - tmp[k] = v - end +local function tsplitat(separator) + local splitter = splitters_t[separator] + if not splitter then + splitter = Ct(splitat(separator)) + splitters_t[separator] = splitter end - return tmp + return splitter end -function table.imerge(t, ...) - local lst, nt = { ... }, #t - for i=1,#lst do - local nst = lst[i] - for j=1,#nst do - nt = nt + 1 - t[nt] = nst[j] - end - end - return t -end +lpeg.splitat = splitat +lpeg.tsplitat = tsplitat -function table.imerged(...) - local tmp, ntmp, lst = { }, 0, {...} - for i=1,#lst do - local nst = lst[i] - for j=1,#nst do - ntmp = ntmp + 1 - tmp[ntmp] = nst[j] - end +--~ local p = splitat("->",false) print(match(p,"oeps->what->more")) -- oeps what more +--~ local p = splitat("->",true) print(match(p,"oeps->what->more")) -- oeps what->more +--~ local p = splitat("->",false) print(match(p,"oeps")) -- oeps +--~ local p = splitat("->",true) print(match(p,"oeps")) -- oeps + +local cache = { } + +function lpeg.split(separator,str) + local c = cache[separator] + if not c then + c = tsplitat(separator) + cache[separator] = c end - return tmp + return match(c,str) end -local function fastcopy(old,metatabletoo) -- fast one - if old then - local new = { } - for k,v in next, old do - if type(v) == "table" then - new[k] = fastcopy(v,metatabletoo) -- was just table.copy - else - new[k] = v - end - end - if metatabletoo then - -- optional second arg - local mt = getmetatable(old) - if mt then - setmetatable(new,mt) - end - end - return new - else - return { } +function string.split(str,separator) + local c = cache[separator] + if not c then + c = tsplitat(separator) + cache[separator] = c end + return match(c,str) end --- todo : copy without metatable +local spacing = patterns.spacer^0 * newline -- sort of strip +local empty = spacing * Cc("") +local nonempty = Cs((1-spacing)^1) * spacing^-1 +local content = (empty + nonempty)^1 -local function copy(t, tables) -- taken from lua wiki, slightly adapted - tables = tables or { } - local tcopy = {} - if not tables[t] then - tables[t] = tcopy - end - for i,v in next, t do -- brrr, what happens with sparse indexed - if type(i) == "table" then - if tables[i] then - i = tables[i] - else - i = copy(i, tables) - end - end - if type(v) ~= "table" then - tcopy[i] = v - elseif tables[v] then - tcopy[i] = tables[v] - else - tcopy[i] = copy(v, tables) - end - end - local mt = getmetatable(t) - if mt then - setmetatable(tcopy,mt) - end - return tcopy +patterns.textline = content + +--~ local linesplitter = Ct(content^0) +--~ +--~ function string.splitlines(str) +--~ return match(linesplitter,str) +--~ end + +local linesplitter = tsplitat(newline) + +patterns.linesplitter = linesplitter + +function string.splitlines(str) + return match(linesplitter,str) end -table.fastcopy = fastcopy -table.copy = copy +local utflinesplitter = utfbom^-1 * tsplitat(newline) -function table.derive(parent) - local child = { } - if parent then - setmetatable(child,{ __index = parent }) - end - return child +patterns.utflinesplitter = utflinesplitter + +function string.utfsplitlines(str) + return match(utflinesplitter,str) end -function table.tohash(t,value) - local h = { } - if t then - if value == nil then value = true end - for _, v in next, t do -- no ipairs here - h[v] = value - end +--~ lpeg.splitters = cache -- no longer public + +local cache = { } + +function lpeg.checkedsplit(separator,str) + local c = cache[separator] + if not c then + separator = P(separator) + local other = C((1 - separator)^1) + c = Ct(separator^0 * other * (separator^1 * other)^0) + cache[separator] = c end - return h + return match(c,str) end -function table.fromhash(t) - local hsh, h = { }, 0 - for k, v in next, t do -- no ipairs here - if v then - h = h + 1 - hsh[h] = k - end +function string.checkedsplit(str,separator) + local c = cache[separator] + if not c then + separator = P(separator) + local other = C((1 - separator)^1) + c = Ct(separator^0 * other * (separator^1 * other)^0) + cache[separator] = c end - return hsh + return match(c,str) end -local noquotes, hexify, handle, reduce, compact, inline, functions +--~ from roberto's site: -local reserved = table.tohash { -- intercept a language inconvenience: no reserved words as key - 'and', 'break', 'do', 'else', 'elseif', 'end', 'false', 'for', 'function', 'if', - 'in', 'local', 'nil', 'not', 'or', 'repeat', 'return', 'then', 'true', 'until', 'while', -} +local function f2(s) local c1, c2 = byte(s,1,2) return c1 * 64 + c2 - 12416 end +local function f3(s) local c1, c2, c3 = byte(s,1,3) return (c1 * 64 + c2) * 64 + c3 - 925824 end +local function f4(s) local c1, c2, c3, c4 = byte(s,1,4) return ((c1 * 64 + c2) * 64 + c3) * 64 + c4 - 63447168 end -local function simple_table(t) - if #t > 0 then - local n = 0 - for _,v in next, t do - n = n + 1 - end - if n == #t then - local tt, nt = { }, 0 - for i=1,#t do - local v = t[i] - local tv = type(v) - if tv == "number" then - nt = nt + 1 - if hexify then - tt[nt] = format("0x%04X",v) - else - tt[nt] = tostring(v) -- tostring not needed - end - elseif tv == "boolean" then - nt = nt + 1 - tt[nt] = tostring(v) - elseif tv == "string" then - nt = nt + 1 - tt[nt] = format("%q",v) - else - tt = nil - break - end - end - return tt - end - end - return nil -end +local utf8byte = patterns.utf8one/byte + patterns.utf8two/f2 + patterns.utf8three/f3 + patterns.utf8four/f4 --- Because this is a core function of mkiv I moved some function calls --- inline. --- --- twice as fast in a test: --- --- local propername = lpeg.P(lpeg.R("AZ","az","__") * lpeg.R("09","AZ","az", "__")^0 * lpeg.P(-1) ) +patterns.utf8byte = utf8byte --- problem: there no good number_to_string converter with the best resolution +--~ local str = " a b c d " -local function dummy() end +--~ local s = lpeg.stripper(lpeg.R("az")) print("["..lpeg.match(s,str).."]") +--~ local s = lpeg.keeper(lpeg.R("az")) print("["..lpeg.match(s,str).."]") +--~ local s = lpeg.stripper("ab") print("["..lpeg.match(s,str).."]") +--~ local s = lpeg.keeper("ab") print("["..lpeg.match(s,str).."]") -local function do_serialize(root,name,depth,level,indexed) - if level > 0 then - depth = depth .. " " - if indexed then - handle(format("%s{",depth)) - else - local tn = type(name) - if tn == "number" then -- or find(k,"^%d+$") then - if hexify then - handle(format("%s[0x%04X]={",depth,name)) - else - handle(format("%s[%s]={",depth,name)) - end - elseif tn == "string" then - if noquotes and not reserved[name] and find(name,"^%a[%w%_]*$") then - handle(format("%s%s={",depth,name)) - else - handle(format("%s[%q]={",depth,name)) - end - elseif tn == "boolean" then - handle(format("%s[%s]={",depth,tostring(name))) - else - handle(format("%s{",depth)) - end +local cache = { } + +function lpeg.stripper(str) + if type(str) == "string" then + local s = cache[str] + if not s then + s = Cs(((S(str)^1)/"" + 1)^0) + cache[str] = s + end + return s + else + return Cs(((str^1)/"" + 1)^0) + end +end + +local cache = { } + +function lpeg.keeper(str) + if type(str) == "string" then + local s = cache[str] + if not s then + s = Cs((((1-S(str))^1)/"" + 1)^0) + cache[str] = s end + return s + else + return Cs((((1-str)^1)/"" + 1)^0) end - -- we could check for k (index) being number (cardinal) - if root and next(root) then - local first, last = nil, 0 -- #root cannot be trusted here (will be ok in 5.2 when ipairs is gone) - if compact then - -- NOT: for k=1,#root do (we need to quit at nil) - for k,v in ipairs(root) do -- can we use next? - if not first then first = k end - last = last + 1 - end - end - local sk = sortedkeys(root) - for i=1,#sk do - local k = sk[i] - local v = root[k] - --~ if v == root then - -- circular - --~ else - local t, tk = type(v), type(k) - if compact and first and tk == "number" and k >= first and k <= last then - if t == "number" then - if hexify then - handle(format("%s 0x%04X,",depth,v)) - else - handle(format("%s %s,",depth,v)) -- %.99g - end - elseif t == "string" then - if reduce and tonumber(v) then - handle(format("%s %s,",depth,v)) - else - handle(format("%s %q,",depth,v)) - end - elseif t == "table" then - if not next(v) then - handle(format("%s {},",depth)) - elseif inline then -- and #t > 0 - local st = simple_table(v) - if st then - handle(format("%s { %s },",depth,concat(st,", "))) - else - do_serialize(v,k,depth,level+1,true) - end - else - do_serialize(v,k,depth,level+1,true) - end - elseif t == "boolean" then - handle(format("%s %s,",depth,tostring(v))) - elseif t == "function" then - if functions then - handle(format('%s loadstring(%q),',depth,dump(v))) - else - handle(format('%s "function",',depth)) - end - else - handle(format("%s %q,",depth,tostring(v))) - end - elseif k == "__p__" then -- parent - if false then - handle(format("%s __p__=nil,",depth)) - end - elseif t == "number" then - if tk == "number" then -- or find(k,"^%d+$") then - if hexify then - handle(format("%s [0x%04X]=0x%04X,",depth,k,v)) - else - handle(format("%s [%s]=%s,",depth,k,v)) -- %.99g - end - elseif tk == "boolean" then - if hexify then - handle(format("%s [%s]=0x%04X,",depth,tostring(k),v)) - else - handle(format("%s [%s]=%s,",depth,tostring(k),v)) -- %.99g - end - elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then - if hexify then - handle(format("%s %s=0x%04X,",depth,k,v)) - else - handle(format("%s %s=%s,",depth,k,v)) -- %.99g - end - else - if hexify then - handle(format("%s [%q]=0x%04X,",depth,k,v)) - else - handle(format("%s [%q]=%s,",depth,k,v)) -- %.99g - end - end - elseif t == "string" then - if reduce and tonumber(v) then - if tk == "number" then -- or find(k,"^%d+$") then - if hexify then - handle(format("%s [0x%04X]=%s,",depth,k,v)) - else - handle(format("%s [%s]=%s,",depth,k,v)) - end - elseif tk == "boolean" then - handle(format("%s [%s]=%s,",depth,tostring(k),v)) - elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then - handle(format("%s %s=%s,",depth,k,v)) - else - handle(format("%s [%q]=%s,",depth,k,v)) - end - else - if tk == "number" then -- or find(k,"^%d+$") then - if hexify then - handle(format("%s [0x%04X]=%q,",depth,k,v)) - else - handle(format("%s [%s]=%q,",depth,k,v)) - end - elseif tk == "boolean" then - handle(format("%s [%s]=%q,",depth,tostring(k),v)) - elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then - handle(format("%s %s=%q,",depth,k,v)) - else - handle(format("%s [%q]=%q,",depth,k,v)) - end - end - elseif t == "table" then - if not next(v) then - if tk == "number" then -- or find(k,"^%d+$") then - if hexify then - handle(format("%s [0x%04X]={},",depth,k)) - else - handle(format("%s [%s]={},",depth,k)) - end - elseif tk == "boolean" then - handle(format("%s [%s]={},",depth,tostring(k))) - elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then - handle(format("%s %s={},",depth,k)) - else - handle(format("%s [%q]={},",depth,k)) - end - elseif inline then - local st = simple_table(v) - if st then - if tk == "number" then -- or find(k,"^%d+$") then - if hexify then - handle(format("%s [0x%04X]={ %s },",depth,k,concat(st,", "))) - else - handle(format("%s [%s]={ %s },",depth,k,concat(st,", "))) - end - elseif tk == "boolean" then -- or find(k,"^%d+$") then - handle(format("%s [%s]={ %s },",depth,tostring(k),concat(st,", "))) - elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then - handle(format("%s %s={ %s },",depth,k,concat(st,", "))) - else - handle(format("%s [%q]={ %s },",depth,k,concat(st,", "))) - end - else - do_serialize(v,k,depth,level+1) - end - else - do_serialize(v,k,depth,level+1) - end - elseif t == "boolean" then - if tk == "number" then -- or find(k,"^%d+$") then - if hexify then - handle(format("%s [0x%04X]=%s,",depth,k,tostring(v))) - else - handle(format("%s [%s]=%s,",depth,k,tostring(v))) - end - elseif tk == "boolean" then -- or find(k,"^%d+$") then - handle(format("%s [%s]=%s,",depth,tostring(k),tostring(v))) - elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then - handle(format("%s %s=%s,",depth,k,tostring(v))) - else - handle(format("%s [%q]=%s,",depth,k,tostring(v))) - end - elseif t == "function" then - if functions then - local f = getinfo(v).what == "C" and dump(dummy) or dump(v) - -- local f = getinfo(v).what == "C" and dump(function(...) return v(...) end) or dump(v) - if tk == "number" then -- or find(k,"^%d+$") then - if hexify then - handle(format("%s [0x%04X]=loadstring(%q),",depth,k,f)) - else - handle(format("%s [%s]=loadstring(%q),",depth,k,f)) - end - elseif tk == "boolean" then - handle(format("%s [%s]=loadstring(%q),",depth,tostring(k),f)) - elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then - handle(format("%s %s=loadstring(%q),",depth,k,f)) - else - handle(format("%s [%q]=loadstring(%q),",depth,k,f)) - end - end - else - if tk == "number" then -- or find(k,"^%d+$") then - if hexify then - handle(format("%s [0x%04X]=%q,",depth,k,tostring(v))) - else - handle(format("%s [%s]=%q,",depth,k,tostring(v))) - end - elseif tk == "boolean" then -- or find(k,"^%d+$") then - handle(format("%s [%s]=%q,",depth,tostring(k),tostring(v))) - elseif noquotes and not reserved[k] and find(k,"^%a[%w%_]*$") then - handle(format("%s %s=%q,",depth,k,tostring(v))) +end + +function lpeg.frontstripper(str) -- or pattern (yet undocumented) + return (P(str) + P(true)) * Cs(P(1)^0) +end + +function lpeg.endstripper(str) -- or pattern (yet undocumented) + return Cs((1 - P(str) * P(-1))^0) +end + +-- Just for fun I looked at the used bytecode and +-- p = (p and p + pp) or pp gets one more (testset). + +function lpeg.replacer(one,two) + if type(one) == "table" then + local no = #one + if no > 0 then + local p + for i=1,no do + local o = one[i] + local pp = P(o[1]) / o[2] + if p then + p = p + pp else - handle(format("%s [%q]=%q,",depth,k,tostring(v))) + p = pp end end - --~ end + return Cs((p + 1)^0) end + else + two = two or "" + return Cs((P(one)/two + 1)^0) end - if level > 0 then - handle(format("%s},",depth)) +end + +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) + splitters_f[separator] = splitter + end + return splitter +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) + splitters_s[separator] = splitter end + return splitter end --- replacing handle by a direct t[#t+1] = ... (plus test) is not much --- faster (0.03 on 1.00 for zapfino.tma) +function lpeg.balancer(left,right) + left, right = P(left), P(right) + return P { left * ((1 - left - right) + V(1))^0 * right } +end -local function serialize(_handle,root,name,specification) -- handle wins - local tname = type(name) - if type(specification) == "table" then - noquotes = specification.noquotes - hexify = specification.hexify - handle = _handle or specification.handle or print - reduce = specification.reduce or false - functions = specification.functions - compact = specification.compact - inline = specification.inline and compact - if functions == nil then - functions = true - end - if compact == nil then - compact = true - end - if inline == nil then - inline = compact - end - else - noquotes = false - hexify = false - handle = _handle or print - reduce = false - compact = true - inline = true - functions = true +--~ print(1,match(lpeg.firstofsplit(":"),"bc:de")) +--~ print(2,match(lpeg.firstofsplit(":"),":de")) -- empty +--~ print(3,match(lpeg.firstofsplit(":"),"bc")) +--~ print(4,match(lpeg.secondofsplit(":"),"bc:de")) +--~ print(5,match(lpeg.secondofsplit(":"),"bc:")) -- empty +--~ print(6,match(lpeg.secondofsplit(":",""),"bc")) +--~ print(7,match(lpeg.secondofsplit(":"),"bc")) +--~ print(9,match(lpeg.secondofsplit(":","123"),"bc")) + +--~ -- slower: +--~ +--~ function lpeg.counter(pattern) +--~ local n, pattern = 0, (lpeg.P(pattern)/function() n = n + 1 end + lpeg.anything)^0 +--~ return function(str) n = 0 ; lpegmatch(pattern,str) ; return n end +--~ end + +local nany = utf8char/"" + +function lpeg.counter(pattern) + pattern = Cs((P(pattern)/" " + nany)^0) + return function(str) + return #match(pattern,str) end - if tname == "string" then - if name == "return" then - handle("return {") - else - handle(name .. "={") - end - elseif tname == "number" then - if hexify then - handle(format("[0x%04X]={",name)) - else - handle("[" .. name .. "]={") - end - elseif tname == "boolean" then - if name then - handle("return {") - else - handle("{") +end + +if utfgmatch then + + function lpeg.count(str,what) -- replaces string.count + if type(what) == "string" then + local n = 0 + for _ in utfgmatch(str,what) do + n = n + 1 + end + return n + else -- 4 times slower but still faster than / function + return #match(Cs((P(what)/" " + nany)^0),str) end - else - handle("t={") end - if root then - -- The dummy access will initialize a table that has a delayed initialization - -- using a metatable. (maybe explicitly test for metatable) - if getmetatable(root) then -- todo: make this an option, maybe even per subtable - local dummy = root._w_h_a_t_e_v_e_r_ - root._w_h_a_t_e_v_e_r_ = nil - end - -- Let's forget about empty tables. - if next(root) then - do_serialize(root,name,"",0) + +else + + local cache = { } + + function lpeg.count(str,what) -- replaces string.count + if type(what) == "string" then + local p = cache[what] + if not p then + p = Cs((P(what)/" " + nany)^0) + cache[p] = p + end + return #match(p,str) + else -- 4 times slower but still faster than / function + return #match(Cs((P(what)/" " + nany)^0),str) end end - handle("}") + end ---~ name: ---~ ---~ true : return { } ---~ false : { } ---~ nil : t = { } ---~ string : string = { } ---~ 'return' : return { } ---~ number : [number] = { } +local patterns_escapes = { -- also defines in l-string + ["%"] = "%%", + ["."] = "%.", + ["+"] = "%+", ["-"] = "%-", ["*"] = "%*", + ["["] = "%[", ["]"] = "%]", + ["("] = "%)", [")"] = "%)", + -- ["{"] = "%{", ["}"] = "%}" + -- ["^"] = "%^", ["$"] = "%$", +} -function table.serialize(root,name,specification) - local t, n = { }, 0 - local function flush(s) - n = n + 1 - t[n] = s - end - serialize(flush,root,name,specification) - return concat(t,"\n") +local simple_escapes = { -- also defines in l-string + ["-"] = "%-", + ["."] = "%.", + ["?"] = ".", + ["*"] = ".*", +} + +local p = Cs((S("-.+*%()[]") / patterns_escapes + anything)^0) +local s = Cs((S("-.+*%()[]") / simple_escapes + anything)^0) + +function string.escapedpattern(str,simple) + return match(simple and s or p,str) end -table.tohandle = serialize +-- utf extensies --- sometimes tables are real use (zapfino extra pro is some 85M) in which --- case a stepwise serialization is nice; actually, we could consider: --- --- for line in table.serializer(root,name,reduce,noquotes) do --- ...(line) --- end --- --- so this is on the todo list +lpeg.UP = lpeg.P -local maxtab = 2*1024 +if utfcharacters then -function table.tofile(filename,root,name,specification) - local f = io.open(filename,'w') - if f then - if maxtab > 1 then - local t, n = { }, 0 - local function flush(s) - n = n + 1 - t[n] = s - if n > maxtab then - f:write(concat(t,"\n"),"\n") -- hm, write(sometable) should be nice - t, n = { }, 0 -- we could recycle t if needed - end + function lpeg.US(str) + local p + for uc in utfcharacters(str) do + if p then + p = p + P(uc) + else + p = P(uc) end - serialize(flush,root,name,specification) - f:write(concat(t,"\n"),"\n") - else - local function flush(s) - f:write(s,"\n") + end + return p + end + + +elseif utfgmatch then + + function lpeg.US(str) + local p + for uc in utfgmatch(str,".") do + if p then + p = p + P(uc) + else + p = P(uc) end - serialize(flush,root,name,specification) end - f:close() - io.flush() + return p + end + +else + + function lpeg.US(str) + local p + local f = function(uc) + if p then + p = p + P(uc) + else + p = P(uc) + end + end + match((utf8char/f)^0,str) + return p end + end -local function flattened(t,f,depth) - if f == nil then - f = { } - depth = 0xFFFF - elseif tonumber(f) then - -- assume then only two arguments are given - depth = f - f = { } - elseif not depth then - depth = 0xFFFF +local range = Cs(utf8byte) * (Cs(utf8byte) + Cc(false)) + +local utfchar = unicode and unicode.utf8 and unicode.utf8.char + +function lpeg.UR(str,more) + local first, last + if type(str) == "number" then + first = str + last = more or first + else + first, last = match(range,str) + if not last then + return P(str) + end end - for k, v in next, t do - if type(k) ~= "number" then - if depth > 0 and type(v) == "table" then - flattened(v,f,depth-1) + if first == last then + return P(str) + elseif utfchar and last - first < 8 then -- a somewhat arbitrary criterium + local p + for i=first,last do + if p then + p = p + P(utfchar(i)) else - f[k] = v + p = P(utfchar(i)) end end - end - local n = #f - for k=1,#t do - local v = t[k] - if depth > 0 and type(v) == "table" then - flattened(v,f,depth-1) - n = #f - else - n = n + 1 - f[n] = v + return p -- nil when invalid range + else + local f = function(b) + return b >= first and b <= last end + return utf8byte / f -- nil when invalid range end - return f end -table.flattened = flattened +--~ lpeg.print(lpeg.R("ab","cd","gh")) +--~ lpeg.print(lpeg.P("a","b","c")) +--~ lpeg.print(lpeg.S("a","b","c")) -local function unnest(t,f) -- only used in mk, for old times sake - if not f then -- and only relevant for token lists - f = { } +--~ print(lpeg.count("äáàa",lpeg.P("á") + lpeg.P("à"))) +--~ print(lpeg.count("äáàa",lpeg.UP("áà"))) +--~ print(lpeg.count("äáàa",lpeg.US("àá"))) +--~ print(lpeg.count("äáàa",lpeg.UR("aá"))) +--~ print(lpeg.count("äáàa",lpeg.UR("àá"))) +--~ print(lpeg.count("äáàa",lpeg.UR(0x0000,0xFFFF))) + +function lpeg.oneof(list,...) -- lpeg.oneof("elseif","else","if","then") + if type(list) ~= "table" then + list = { list, ... } end - for i=1,#t do - local v = t[i] - if type(v) == "table" then - if type(v[1]) == "table" then - unnest(v,f) - else - f[#f+1] = v - end - else - f[#f+1] = v - end + -- sort(list) -- longest match first + local p = P(list[1]) + for l=2,#list do + p = p + P(list[l]) end - return f + return p end -function table.unnest(t) -- bad name - return unnest(t) +function lpeg.is_lpeg(p) + return p and lpegtype(p) == "pattern" end -local function are_equal(a,b,n,m) -- indexed - if a and b and #a == #b then - n = n or 1 - m = m or #a - for i=n,m do - local ai, bi = a[i], b[i] - if ai==bi then - -- same - elseif type(ai)=="table" and type(bi)=="table" then - if not are_equal(ai,bi) then - return false - end +-- For the moment here, but it might move to utilities: + +local sort, fastcopy, sortedpairs = table.sort, table.fastcopy, table.sortedpairs -- dependency! + +function lpeg.append(list,pp) + local p = pp + if #list > 0 then + list = fastcopy(list) + sort(list) + for l=1,#list do + if p then + p = P(list[l]) + p else - return false + p = P(list[l]) end end - return true else - return false - end -end - -local function identical(a,b) -- assumes same structure - for ka, va in next, a do - local vb = b[ka] - if va == vb then - -- same - elseif type(va) == "table" and type(vb) == "table" then - if not identical(va,vb) then - return false + for k, v in sortedpairs(list) do + if p then + p = P(k)/v + p + else + p = P(k)/v end - else - return false end end - return true + return p end -table.identical = identical -table.are_equal = are_equal +--~ Cf(Ct("") * (Cg(C(...) * "=" * Cs(...)))^0, rawset) --- maybe also make a combined one +end -- closure -function table.compact(t) - if t then - for k,v in next, t do - if not next(v) then - t[k] = nil - end - end - end -end +do -- begin closure to overcome local limits and interference -function table.contains(t, v) - if t then - for i=1, #t do - if t[i] == v then - return i - end - end - end - return false -end +if not modules then modules = { } end modules ['l-boolean'] = { + version = 1.001, + comment = "companion to luat-lib.mkiv", + author = "Hans Hagen, PRAGMA-ADE, Hasselt NL", + copyright = "PRAGMA ADE / ConTeXt Development Team", + license = "see context related readme files" +} -function table.count(t) - local n = 0 - for k, v in next, t do - n = n + 1 - end - return n -end +local type, tonumber = type, tonumber -function table.swapped(t,s) -- hash - local n = { } - if s then ---~ for i=1,#s do ---~ n[i] = s[i] ---~ end - for k, v in next, s do - n[k] = v - end - end ---~ for i=1,#t do ---~ local ti = t[i] -- don't ask but t[i] can be nil ---~ if ti then ---~ n[ti] = i ---~ end ---~ end - for k, v in next, t do - n[v] = k - end - return n -end +boolean = boolean or { } +local boolean = boolean -function table.reversed(t) - if t then - local tt, tn = { }, #t - if tn > 0 then - local ttn = 0 - for i=tn,1,-1 do - ttn = ttn + 1 - tt[ttn] = t[i] - end - end - return tt - end +function boolean.tonumber(b) + if b then return 1 else return 0 end -- test and return or return end -function table.sequenced(t,sep,simple) -- hash only - local s, n = { }, 0 - for k, v in sortedhash(t) do - if simple then - if v == true then - n = n + 1 - s[n] = k - elseif v and v~= "" then - n = n + 1 - s[n] = k .. "=" .. tostring(v) - end +function toboolean(str,tolerant) + if tolerant then + local tstr = type(str) + if tstr == "string" then + return str == "true" or str == "yes" or str == "on" or str == "1" or str == "t" + elseif tstr == "number" then + return tonumber(str) ~= 0 + elseif tstr == "nil" then + return false else - n = n + 1 - s[n] = k .. "=" .. tostring(v) + return str end + elseif str == "true" then + return true + elseif str == "false" then + return false + else + return str end - return concat(s, sep or " | ") end -function table.print(t,...) - if type(t) ~= "table" then - print(tostring(t)) - else - table.tohandle(print,t,...) +string.toboolean = toboolean + +function string.is_boolean(str,default) + if type(str) == "string" then + if str == "true" or str == "yes" or str == "on" or str == "t" then + return true + elseif str == "false" or str == "no" or str == "off" or str == "f" then + return false + end end + return default end --- -- -- obsolete but we keep them for a while and might comment them later -- -- -- +end -- closure --- roughly: copy-loop : unpack : sub == 0.9 : 0.4 : 0.45 (so in critical apps, use unpack) +do -- begin closure to overcome local limits and interference -function table.sub(t,i,j) - return { unpack(t,i,j) } +if not modules then modules = { } end modules ['l-math'] = { + version = 1.001, + comment = "companion to luat-lib.mkiv", + author = "Hans Hagen, PRAGMA-ADE, Hasselt NL", + copyright = "PRAGMA ADE / ConTeXt Development Team", + license = "see context related readme files" +} + +local floor, sin, cos, tan = math.floor, math.sin, math.cos, math.tan + +if not math.round then + function math.round(x) return floor(x + 0.5) end end --- slower than #t on indexed tables (#t only returns the size of the numerically indexed slice) +if not math.div then + function math.div(n,m) return floor(n/m) end +end -function table.is_empty(t) - return not t or not next(t) +if not math.mod then + function math.mod(n,m) return n % m end end -function table.has_one_entry(t) - return t and not next(t,next(t)) +local pipi = 2*math.pi/360 + +if not math.sind then + function math.sind(d) return sin(d*pipi) end + function math.cosd(d) return cos(d*pipi) end + function math.tand(d) return tan(d*pipi) end +end + +if not math.odd then + function math.odd (n) return n % 2 == 0 end + function math.even(n) return n % 2 ~= 0 end end end -- closure @@ -2958,6 +3021,7 @@ local next, tostring, rawget = next, tostring, rawget local format, match, lower, gsub = string.format, string.match, string.lower, string.gsub local utfbyte = utf.byte local sort, insert, concat, sortedkeys, serialize, fastcopy = table.sort, table.insert, table.concat, table.sortedkeys, table.serialize, table.fastcopy +local derivetable = table.derive local trace_defining = false trackers.register("fonts.defining", function(v) trace_defining = v end) local trace_scaling = false trackers.register("fonts.scaling" , function(v) trace_scaling = v end) @@ -3139,10 +3203,9 @@ function constructors.scale(tfmdata,specification) local mathparameters = tfmdata.mathparameters or { } -- local targetcharacters = { } - local targetdescriptions = table.derive(descriptions) - local targetparameters = table.derive(parameters) - -- local targetmathparameters = table.fastcopy(mathparameters) -- happens elsewhere - local targetproperties = table.derive(properties) + local targetdescriptions = derivetable(descriptions) + local targetparameters = derivetable(parameters) + local targetproperties = derivetable(properties) local targetgoodies = goodies -- we need to loop so no metatable target.characters = targetcharacters target.descriptions = targetdescriptions @@ -4827,7 +4890,7 @@ local getn = table.getn local lpegmatch = lpeg.match local reversed, concat, remove = table.reversed, table.concat, table.remove local ioflush = io.flush -local fastcopy, tohash = table.fastcopy, table.tohash +local fastcopy, tohash, derivetable = table.fastcopy, table.tohash, table.derive local allocate = utilities.storage.allocate local registertracker = trackers.register @@ -6476,9 +6539,9 @@ local function copytotfm(data,cache_id) if data then local metadata = data.metadata local resources = data.resources - local properties = table.derive(data.properties) - local descriptions = table.derive(data.descriptions) - local goodies = table.derive(data.goodies) + local properties = derivetable(data.properties) + local descriptions = derivetable(data.descriptions) + local goodies = derivetable(data.goodies) local characters = { } local parameters = { } local mathparameters = { } diff --git a/tex/generic/context/luatex-fonts.lua b/tex/generic/context/luatex-fonts.lua index 23d33f26b..1d844911d 100644 --- a/tex/generic/context/luatex-fonts.lua +++ b/tex/generic/context/luatex-fonts.lua @@ -122,10 +122,10 @@ else -- version 1.0 there will be an official api defined. loadmodule('l-string.lua') + loadmodule('l-table.lua') loadmodule('l-lpeg.lua') loadmodule('l-boolean.lua') loadmodule('l-math.lua') - loadmodule('l-table.lua') loadmodule('l-file.lua') loadmodule('l-io.lua') diff --git a/tex/generic/context/luatex-plain.tex b/tex/generic/context/luatex-plain.tex new file mode 100644 index 000000000..e47ad58ad --- /dev/null +++ b/tex/generic/context/luatex-plain.tex @@ -0,0 +1,25 @@ +%D \module +%D [ file=luatex-plain, +%D version=2009.12.01, +%D title=\LUATEX\ Macros, +%D subtitle=Plain Format, +%D author=Hans Hagen, +%D date=\currentdate, +%D copyright={PRAGMA ADE \& \CONTEXT\ Development Team}] + +\input plain + +\directlua {tex.enableprimitives('', tex.extraprimitives())} + +\pdfoutput=1 + +\everyjob \expandafter {% + \the\everyjob + \input luatex-basics\relax + \input luatex-fonts\relax + \input luatex-mplib\relax +} + +\edef\fmtversion{\fmtversion+luatex} + +\dump -- cgit v1.2.3