--[[ Copyright (c) 2010 Peter "Corsix" Cawley Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. --]] local lfs = require("lfs") local TH = require("TH") --! Layer which handles the loading of localised text. class "Strings" ---@type Strings local Strings = _G["Strings"] function Strings:Strings(app) self.app = app end local utf8conv local cp437conv local function id(...) return ... end function Strings:init() -- Load (but do not execute) everything from the language directory -- Note that files are loaded with loadfile_envcall self.language_chunks = {} local pathsep = package.config:sub(1,1) local path = self.app:getFullPath({"Lua", "languages"}, true) for file in lfs.dir(path) do if file:match("%.lua$") then local result, err = loadfile_envcall(path .. file) if not result then print("Error loading languages" .. pathsep .. file .. ":\n" .. tostring(err)) else self.language_chunks[result] = "languages" .. pathsep .. file end end end -- Build the language table from Language() calls -- Every file in the languages folder should have a call to Language() near -- the start of the file which gives the names for the language. These names -- are used to link the user's choice of language to a file, and to link the -- names given to Inherit() to a file. The first name is used as the official -- name for this language, the others may be abbreviations or such. self.languages = {} self.languages_english = {} self.language_to_chunk = {} self.chunk_to_font = {} self.chunk_to_names = {} self.language_to_lang_code = {} self.languages_with_arabic_numerals = {} for chunk, filename in pairs(self.language_chunks) do -- To allow the file to set global variables without causing an error, it -- is given an infinite table as an environment. Reading a non-existent -- key from an infinite table returns another infinite table, rather than -- the default value of nil. local infinite_table_mt infinite_table_mt = { __index = function(_, _) return setmetatable({}, infinite_table_mt) end } -- To abort evaluation of the chunk after Language() is called, a unique -- marker is used as an error message. The other alternative would be to -- do complex co-routine calls. A freshly made empty table is a suitable -- unique marker. local good_error_marker = {} local env = setmetatable({ utf8 = id, cp437 = cp437conv, pairs = pairs, Language = function(...) local names = {...} -- Use the first name for display purposes (case-dependent!). if names[1] ~= "original_strings" then self.languages[#self.languages + 1] = names[1] -- Also save the second name for tooltips and internal purposes. assert(names[2], filename .. " does not have an English name.") self.languages_english[names[1]] = names[2] end -- Associate every passed name with this file, case-independently for _, name in pairs(names) do self.language_to_chunk[name:lower()] = chunk self.language_to_lang_code[name:lower()] = names[3] end self.chunk_to_names[chunk] = names error(good_error_marker) end, Font = function(...) self.chunk_to_font[chunk] = ... end, -- Set Inherit and SetSpeechFile to do nothing Inherit = function() end, SetSpeechFile = function() end, Encoding = function() end, IsArabicNumerals = function() end, -- Set LoadStrings to return an infinite table LoadStrings = infinite_table_mt.__index, }, infinite_table_mt) -- Actually run the language file local status, err = pcall(chunk, env) if not status and err ~= good_error_marker and TheApp.good_install_folder then print("Error evaluating " .. filename .. ":\n" .. tostring(err)) end end table.sort(self.languages) end -- String tables are shadowed so that all access to a string table goes -- through a metamethod (which is desirable to catch invalid reads, provide -- "__random", and prevent writes). This means that every string table is in -- fact empty, and has it's keys and values stored in a separate table, called -- the shadow table. The shadows table is used to associate a shadow table -- with a string table. local shadows = setmetatable({}, {__mode = "k"}) -- Metatable which is used for all tables returned by Strings:load() -- The end effect is to raise errors on accesses to non-existent strings -- (unless no_restriction is set to true), to add a special string called -- "__random" to each table (which always resolves to a random string from -- the table), and to prevent editing or adding to a string table. local strings_metatable = function(no_restriction) return { __index = function(t, key) t = shadows[t] local val = t[key] if val ~= nil then return val end if key ~= "__random" then if no_restriction then return nil end error("Non-existent string: " .. tostring(key), 2) end local candidates = {} for _, v in pairs(t) do candidates[#candidates + 1] = v end return candidates[math.random(1, #candidates)] end, __newindex = function(_, _, _) error("String tables are read-only", 2) end, __pairs = function(t) return pairs(shadows[t]) end, __ipairs = function(t) return ipairs(shadows[t]) end, } end -- no_restriction disables errors on access to non-existent strings (for debug purposes) -- no_inheritance disables inheritance except original_strings (for debug purposes) function Strings:load(language, no_restriction, no_inheritance) assert(language ~= "original_strings", "Original strings can not be loaded directly. Please select a proper language.") -- env is the table of globals to execute to the language file in, and hence -- it also stores the resulting strings. local env = {} shadows[env] = {} -- speech_file holds the result of any call to SetSpeechFile() local speech_file local default_encoding = id local encoding = default_encoding local language_called = false local functions; functions = { -- Convert UTF-8 to the file's default encoding utf8 = function(s) if encoding == cp437conv then return utf8conv(s) else return s end end, -- Convert CP437 to the file's default encoding cp437 = function(s) if encoding == cp437conv then return s else return cp437conv(s) end end, ipairs = ipairs, pairs = pairs, -- Calling the Language() function should have no effect any more Language = function() language_called = true end, Font = function() if language_called then error("Font declaration must occur before Language declaration") end end, -- Inherit() should evaluate the named language in the current environment -- NB: Inheritance of any but original_strings disabled when no_inheritance set Inherit = function(lang, ...) if no_inheritance and lang ~= "original_strings" then return end local old_encoding = encoding encoding = default_encoding local old_language_called = language_called language_called = false self:_loadPrivate(lang, env, ...) encoding = old_encoding language_called = old_language_called end, -- Encoding() should set the default encoding for the remainder of the file Encoding = function(new_encoding) if new_encoding == functions.utf8 then encoding = id elseif new_encoding == functions.cp437 then encoding = cp437conv else error("Invalid encoding; expected utf8 or cp437") end end, -- LoadStrings() should return the original game string table LoadStrings = function(filename) return assert(TH.LoadStrings(self.app:readDataFile(filename)), "Cannot load original string file '" .. filename .. "'") end, -- SetSpeechFile() should remember the named file to return to our caller SetSpeechFile = function(...) speech_file = ... end, IsArabicNumerals = function(is_enabled) self.languages_with_arabic_numerals[language] = is_enabled end, _G = env, type = type, } -- The metatable on the environment is set so that the above functions look -- like top-level level globals, so that the environment behaves like an -- infinite table, and so that assignments are merged into the string table. local metatable metatable = { __index = function(t, k) -- Make the functions look like top-level globals local shadow = shadows[t] local value = ((t == env) and functions[k]) or shadow[k] if value == nil then value = setmetatable({}, metatable) shadow[k] = value shadows[value] = {} end return value end, __newindex = function(t, k, v) local ty = type(v) if ty ~= "table" then -- non-table values cannot be merged if ty == "string" then -- convert from file's default encoding to UTF-8 v = encoding(v) end shadows[t][k] = v else -- v should be merged into t[k] -- Perform t[k][k2] = v2 for each (k2, v2) in v to recursively merge t = t[k] for k2, v2 in pairs(v) do t[k2] = v2 end end end, } metatable.__pairs = strings_metatable(no_restriction).__pairs -- Evaluate the language file setmetatable(env, metatable) self:_loadPrivate(language, env) -- Change the metamethods on every string table to match strings_metatable for k, v in pairs(strings_metatable(no_restriction)) do metatable[k] = v end return env, speech_file end --! Get the Font() declaration of a language, if there was one. function Strings:getFont(language) local chunk = self.language_to_chunk[language:lower()] return chunk and self.chunk_to_font[chunk] end function Strings:getLanguageNames(language) local chunk = self.language_to_chunk[language:lower()] return chunk and self.chunk_to_names[chunk] end function Strings:getLangCode(language) local lang = language or self.app.config.language return self.language_to_lang_code[lang:lower()] end function Strings:isArabicNumerals(language) return self.languages_with_arabic_numerals[language:lower()] end --! Use local language text where possible. --!param string (string) The default, likely English, text --!param table (table) A table of translated text in language code fields --!return (string) The text in the current language if available, or in English, or the default string. function Strings:getLocalisedText(string, table) if string and not table then return string elseif table[self:getLangCode()] then return table[self:getLangCode()] elseif table.en then return table.en else return string end end --! Checks the language we specify is known to the application. --!param language (string) Language to check --!return true if known function Strings:checkLanguageExists(language) return not not self.language_to_chunk[language:lower()] end function Strings:_loadPrivate(language, env, ...) local chunk = self.language_to_chunk[language:lower()] if not chunk then -- If selected language could not be found, try to revert to English print(serialize(self.language_to_chunk, {detect_cycles=true, pretty=true})) print("Language '" .. language .. "' could not be found. Reverting to English.") chunk = self.language_to_chunk["english"] if not chunk then -- If english could not be found, raise an error error("Language 'English' could not be found. Please verify your installation.") end end local status, err = pcall(chunk, env, ...) if not status and self.app.good_install_folder then print("Error evaluating " .. self.language_chunks[chunk] .. ":\n" .. tostring(err)) end end function Strings:setupAdviserMessage(messages) local prioTable prioTable = { _priority = 5, -- Default priority tutorial = { _priority = 11}, epidemic = { _priority = 6 }, --staff_advice --earthquake --multiplayer --surgery_requirements --vomit_wave --level_progress --staff_place_advice --room_forbidden_non_reachable_parts --research --boiler_issue --room_requirements --goals warnings = { _priority = 10 }, --placement_info praise = { _priority = 1 }, --information --build_advice --cheats } -- Passes adviser message through string.format process local formatFunc formatFunc = function(format_self, ...) -- After 'format', it is not useful to have indexing magic anymore. return { text = format_self.text:format(...), priority = format_self.priority } end local indexFunc indexFunc = function(index_self, field) -- Since prioTable is a partial table, index_self.table may disappear, prevent infinite recursion. if field == "table" then return nil end local val = {} val.text = index_self.text[field] val.format = formatFunc val.table = index_self.table val.priority = index_self.priority if val.table ~= nil then val.table = val.table[field] if val.table ~= nil and val.table._priority ~= nil then val.priority = val.table._priority end end setmetatable(val, {__index = indexFunc}) return val end -- Initial value local adviserMessage = { text = messages, table = prioTable, priority = prioTable._priority, format = formatFunc } setmetatable(adviserMessage, {__index = indexFunc}) return adviserMessage end -- Primitive system to map UTF-8 characters onto Code Page 437. -- Provided so that language scripts can encode text in a modern and well -- supported manner, and have the text automatically transcoded into the -- encoding which the Theme Hospital fonts use. local codepoints_to_cp437 = { -- Below 0x80 need no translation [0xC7] = 0x80, -- majuscule c-cedilla [0xFC] = 0x81, -- minuscule u-umlaut [0xE9] = 0x82, -- minuscule e-acute [0xE2] = 0x83, -- minuscule a-circumflex [0xE4] = 0x84, -- minuscule a-umlaut [0xE0] = 0x85, -- minuscule a-grave [0xE5] = 0x86, -- minuscule a-ring [0xE7] = 0x87, -- minuscule c-cedilla [0xEA] = 0x88, -- minuscule e-circumflex [0xEB] = 0x89, -- minuscule e-umlaut [0xE8] = 0x8A, -- minuscule e-grave [0xEF] = 0x8B, -- minuscule i-umlaut [0xEE] = 0x8C, -- minuscule i-circumflex [0xEC] = 0x8D, -- minuscule i-grave [0xC4] = 0x8E, -- majuscule a-umlaut [0xC5] = 0x8F, -- majuscule a-ring [0xC9] = 0x90, -- majuscule e-acute [0xE6] = 0x91, -- minuscule ae [0xC6] = 0x91, -- majuscule ae (not in TH fonts - mapped to minuscule ae) [0xF4] = 0x93, -- minuscule o-circumflex [0xF6] = 0x94, -- minuscule o-umlaut [0xF2] = 0x95, -- minuscule o-grave [0xFB] = 0x96, -- minuscule u-circumflex [0xF9] = 0x97, -- minuscule u-grave [0xFF] = 0x98, -- minuscule y-umlaut [0xD6] = 0x99, -- majuscule o-umlaut [0xDC] = 0x9A, -- majuscule u-umlaut -- 0x9B through 0x9F are currency symbols and not present in TH fonts [0xE1] = 0xA0, -- minuscule a-acute [0xED] = 0xA1, -- minuscule i-acute [0xF3] = 0xA2, -- minuscule o-acute [0xFA] = 0xA3, -- minuscule u-acute [0xF1] = 0xA4, -- minuscule n-tilde [0xD1] = 0xA5, -- majuscule n-tilde -- 0xA6 and 0xA7 are ordinal indicators and not present in TH fonts [0xBF] = 0xA8, -- inverted question mark -- 0xA9 through 0xAC are not present in TH fonts [0xA1] = 0xAD, -- inverted exclaimation mark -- 0xAE through 0xE0 are not present in TH fonts [0xDF] = 0xE1, -- eszett / sharp-S / lowercase-beta -- 0xE2 through 0xFF are not present in TH fonts } local cp437_to_codepoints = { [0x91] = 0xE6, -- minuscule ae [0x9B] = 0xA2, -- cent sign [0x9C] = 0xA3, -- pound sign [0x9D] = 0xA5, -- yen sign [0x9E] = 0x20A7, -- peseta sign [0x9F] = 0x192, -- florin sign [0xA6] = 0xAA, -- a ordinal indicator [0xA7] = 0xBA, -- o ordinal indicator [0xA9] = 0x2310, -- negation [0xAA] = 0xAC, -- negation [0xAB] = 0xBD, -- 1/2 [0xAC] = 0xBC, -- 1/4 [0xAE] = 0xAB, -- << guillemets [0xAF] = 0xBB, -- >> guillemets -- 0xB0 through 0xDF omitted [0xE0] = 0x3B1, -- alpha [0xE2] = 0x393, -- gamma [0xE3] = 0x3C0, -- pi [0xE4] = 0x3A3, -- majuscule sigma [0xE5] = 0x3C3, -- minuscule sigma [0xE6] = 0x3BC, -- minuscule mu [0xE7] = 0x3C4, -- tau [0xE8] = 0x3A6, -- majuscule phi [0xE9] = 0x398, -- theta [0xEA] = 0x3A9, -- omega [0xEB] = 0x3B4, -- delta [0xEC] = 0x221E, -- infinity [0xED] = 0x3C6, -- minuscule phi [0xEE] = 0x3B5, -- epsilon [0xEF] = 0x2229, -- set intersection [0xF0] = 0x2261, -- modular congruence / triple equals [0xF1] = 0xB1, -- plus / minus [0xF2] = 0x2265, -- >= [0xF3] = 0x2264, -- <= [0xF4] = 0x2320, -- upper integral sign [0xF5] = 0x2321, -- lower integral sign [0xF6] = 0xF7, -- obelus [0xF7] = 0x2248, -- approximate equality [0xF8] = 0xB0, -- degree symbol [0xF9] = 0x2219, -- bullet [0xFA] = 0xB7, -- interpunct [0xFB] = 0x221A, -- square root [0xFC] = 0x207F, -- exponentiation [0xFD] = 0xB2, -- superscript 2 [0xFE] = 0x25A0, -- filled square [0xFF] = 0xA0, -- non-breaking space } local function utf8encode(codepoint) if codepoint <= 0x7F then return string.char(codepoint) elseif codepoint <= 0x7FF then local sextet = codepoint % 64 codepoint = math.floor((codepoint - sextet) / 64) return string.char(0xC0 + codepoint, 0x80 + sextet) elseif codepoint <= 0xFFFF then local sextet2 = codepoint % 64 codepoint = math.floor((codepoint - sextet2) / 64) local sextet1 = codepoint % 64 codepoint = math.floor((codepoint - sextet2) / 64) return string.char(0xE0 + codepoint, 0x80 + sextet1, 0x80 + sextet2) else local sextet3 = codepoint % 64 codepoint = math.floor((codepoint - sextet3) / 64) local sextet2 = codepoint % 64 codepoint = math.floor((codepoint - sextet2) / 64) local sextet1 = codepoint % 64 codepoint = math.floor((codepoint - sextet2) / 64) return string.char(0xF0 + codepoint, 0x80 + sextet1, 0x80 + sextet2, 0x80 + sextet3) end end local cp437_to_utf8_pattern = {"["} local cp437_to_utf8_replacement = {} for codepoint, cp437 in pairs(codepoints_to_cp437) do if not cp437_to_codepoints[cp437] then cp437_to_utf8_pattern[#cp437_to_utf8_pattern + 1] = string.char(cp437) cp437_to_utf8_replacement[string.char(cp437)] = utf8encode(codepoint) end end for cp437, codepoint in pairs(cp437_to_codepoints) do cp437_to_utf8_pattern[#cp437_to_utf8_pattern + 1] = string.char(cp437) cp437_to_utf8_replacement[string.char(cp437)] = utf8encode(codepoint) end cp437_to_utf8_pattern[#cp437_to_utf8_pattern + 1] = "]" cp437_to_utf8_pattern = table.concat(cp437_to_utf8_pattern) cp437conv = function(s) return (s:gsub(cp437_to_utf8_pattern, cp437_to_utf8_replacement)) end -- Table which maps a single character and a single unicode combining -- diacritical mark to a single unicode codepoint local circumflex = 0x302 local cedilla = 0x327 local umlaut = 0x308 local acute = 0x301 local grave = 0x300 local ring = 0x30A local tilde = 0x303 local combine_diacritical_marks = { a = { [grave] = 0xE0, [acute] = 0xE1, [circumflex] = 0xE2, [umlaut] = 0xE4, [ring] = 0xE5, }, e = { [grave] = 0xE8, [acute] = 0xE9, [circumflex] = 0xEA, [umlaut] = 0xEB, }, i = { [grave] = 0xEC, [acute] = 0xED, [circumflex] = 0xEE, [umlaut] = 0xEF, }, o = { [grave] = 0xF2, [acute] = 0xF3, }, u = { [grave] = 0xF9, [acute] = 0xFA, [circumflex] = 0xFB, [umlaut] = 0xFC, }, c = { [cedilla] = 0xE7, }, n = { [tilde] = 0xF1, }, y = { [umlaut] = 0xFF, }, A = { [umlaut] = 0xC4, [ring] = 0xC5, }, E = { [acute] = 0xC9, }, O = { [umlaut] = 0xD6, }, U = { [umlaut] = 0xDC, }, C = { [cedilla] = 0xC7, }, N = { [tilde] = 0xD1, }, } local function utf8char(c) -- Test for presence of a normal character prior to the utf-8 character local prechar if c:byte() < 128 then prechar = c:sub(1, 1) c = c:sub(2, -1) end -- Extract the codepoint of the utf-8 character local codepoint = 0 local multiplier = 1 for i = #c, 2, -1 do codepoint = codepoint + (c:byte(i) - 128) * multiplier multiplier = multiplier * 2^6 end codepoint = codepoint + (c:byte(1) % 2^(7 - #c)) * multiplier -- If the utf-8 character is a combining diacritical mark, merge it with the -- preceding normal character if prechar and (0x300 <= codepoint and codepoint < 0x370) then if combine_diacritical_marks[prechar] then if combine_diacritical_marks[prechar][codepoint] then codepoint = combine_diacritical_marks[prechar][codepoint] prechar = nil else return prechar end else return prechar end end -- Convert to Code Page 437 return (prechar or "") .. string.char(codepoints_to_cp437[codepoint] or 0x3F) end utf8conv = function(s) -- Pull out each individual utf-8 character and pass it through utf8char -- [\1-\127] picks up a preceding ASCII character to combine diacritics -- [\192-\253] picks up the first byte of a utf-8 character (technically -- only 194 through 244 should be used) -- [\128-\191] picks up the remaining bytes of a utf-8 character return (s:gsub("[\1-\127]?[\192-\253][\128-\191]*", utf8char)) end -- Fix string.upper and string.lower to work with CP437 rather than current locale local lower_to_upper, upper_to_lower = {}, {} local function case(lower, upper) lower = cp437conv(string.char(lower)) upper = cp437conv(string.char(upper)) lower_to_upper[lower] = upper lower_to_upper[upper] = upper upper_to_lower[lower] = lower upper_to_lower[upper] = lower end case(0x87, 0x80) -- c-cedilla case(0x81, 0x9A) -- u-umlaut case(0x82, 0x90) -- e-acute case(0x84, 0x8E) -- a-umlaut case(0x86, 0x8F) -- a-ring case(0x94, 0x99) -- o-umlaut case(0xA4, 0xA5) -- n-tilde local case_pattern = "\195[\128-\191]" -- Unicode range [0xC0, 0xFF] as UTF-8 local orig_upper = string.upper function string.upper(s) -- luacheck: ignore 122 return orig_upper(s:gsub(case_pattern, lower_to_upper)) end local orig_lower = string.lower function string.lower(s) -- luacheck: ignore 122 return orig_lower(s:gsub(case_pattern, upper_to_lower)) end