| 1 | --[[-------------------------------------------------------------------- |
| 2 | |
| 3 | llex.lua: Lua 5.1 lexical analyzer in Lua |
| 4 | This file is part of LuaSrcDiet, based on Yueliang material. |
| 5 | |
| 6 | Copyright (c) 2008 Kein-Hong Man <khman@users.sf.net> |
| 7 | The COPYRIGHT file describes the conditions |
| 8 | under which this software may be distributed. |
| 9 | |
| 10 | See the ChangeLog for more information. |
| 11 | |
| 12 | ----------------------------------------------------------------------]] |
| 13 | |
| 14 | --[[-------------------------------------------------------------------- |
| 15 | -- NOTES: |
| 16 | -- * This is a version of the native 5.1.x lexer from Yueliang 0.4.0, |
| 17 | -- with significant modifications to handle LuaSrcDiet's needs: |
| 18 | -- (1) llex.error is an optional error function handler |
| 19 | -- (2) seminfo for strings include their delimiters and no |
| 20 | -- translation operations are performed on them |
| 21 | -- * ADDED shbang handling has been added to support executable scripts |
| 22 | -- * NO localized decimal point replacement magic |
| 23 | -- * NO limit to number of lines |
| 24 | -- * NO support for compatible long strings (LUA_COMPAT_LSTR) |
| 25 | -- * Please read technotes.txt for more technical details. |
| 26 | ----------------------------------------------------------------------]] |
| 27 | |
| 28 | local base = _G |
| 29 | local string = require "string" |
| 30 | --module "llex" |
| 31 | |
| 32 | local find = string.find |
| 33 | local match = string.match |
| 34 | local sub = string.sub |
| 35 | |
| 36 | ---------------------------------------------------------------------- |
| 37 | -- initialize keyword list, variables |
| 38 | ---------------------------------------------------------------------- |
| 39 | |
| 40 | local kw = {} |
| 41 | for v in string.gmatch([[ |
| 42 | and break do else elseif end false for function if in |
| 43 | local nil not or repeat return then true until while]], "%S+") do |
| 44 | kw[v] = true |
| 45 | end |
| 46 | |
| 47 | -- NOTE: see init() for module variables (externally visible): |
| 48 | -- tok, seminfo, tokln |
| 49 | |
| 50 | local z = '', -- source |
| 51 | sourceid = '', -- name of source |
| 52 | I = 1, -- lexer's position in source |
| 53 | buff = '', |
| 54 | ln = 1, -- line number |
| 55 | tok = {}, -- lexed token list* |
| 56 | seminfo = {}, -- lexed semantic information list* |
| 57 | tokln = {}, -- line numbers for messages* |
| 58 | |
| 59 | ---------------------------------------------------------------------- |
| 60 | -- add information to token listing |
| 61 | ---------------------------------------------------------------------- |
| 62 | |
| 63 | local function addtoken(token, info) |
| 64 | local i = #tok + 1 |
| 65 | tok[i] = token |
| 66 | seminfo[i] = info |
| 67 | tokln[i] = ln |
| 68 | end |
| 69 | |
| 70 | ---------------------------------------------------------------------- |
| 71 | -- handles line number incrementation and end-of-line characters |
| 72 | ---------------------------------------------------------------------- |
| 73 | |
| 74 | local function inclinenumber(i, is_tok) |
| 75 | local sub = sub |
| 76 | local old = sub(z, i, i) |
| 77 | i = i + 1 -- skip '\n' or '\r' |
| 78 | local c = sub(z, i, i) |
| 79 | if (c == "\n" or c == "\r") and (c ~= old) then |
| 80 | i = i + 1 -- skip '\n\r' or '\r\n' |
| 81 | old = old..c |
| 82 | end |
| 83 | if is_tok then addtoken("TK_EOL", old) end |
| 84 | ln = ln + 1 |
| 85 | I = i |
| 86 | return i |
| 87 | end |
| 88 | |
| 89 | ---------------------------------------------------------------------- |
| 90 | -- initialize lexer for given source _z and source name _sourceid |
| 91 | ---------------------------------------------------------------------- |
| 92 | |
| 93 | function init(_z, _sourceid) |
| 94 | z = _z -- source |
| 95 | sourceid = _sourceid -- name of source |
| 96 | I = 1 -- lexer's position in source |
| 97 | ln = 1 -- line number |
| 98 | tok = {} -- lexed token list* |
| 99 | seminfo = {} -- lexed semantic information list* |
| 100 | tokln = {} -- line numbers for messages* |
| 101 | -- (*) externally visible thru' module |
| 102 | -------------------------------------------------------------------- |
| 103 | -- initial processing (shbang handling) |
| 104 | -------------------------------------------------------------------- |
| 105 | local p, _, q, r = find(z, "^(#[^\r\n]*)(\r?\n?)") |
| 106 | if p then -- skip first line |
| 107 | I = I + #q |
| 108 | addtoken("TK_COMMENT", q) |
| 109 | if #r > 0 then inclinenumber(I, true) end |
| 110 | end |
| 111 | end |
| 112 | |
| 113 | ---------------------------------------------------------------------- |
| 114 | -- returns a chunk name or id, no truncation for long names |
| 115 | ---------------------------------------------------------------------- |
| 116 | |
| 117 | function chunkid() |
| 118 | if sourceid and match(sourceid, "^[=@]") then |
| 119 | return sub(sourceid, 2) -- remove first char |
| 120 | end |
| 121 | return "[string]" |
| 122 | end |
| 123 | |
| 124 | ---------------------------------------------------------------------- |
| 125 | -- formats error message and throws error |
| 126 | -- * a simplified version, does not report what token was responsible |
| 127 | ---------------------------------------------------------------------- |
| 128 | |
| 129 | function errorline(s, line) |
| 130 | local e = error or base.error |
| 131 | e(string.format("%s:%d: %s", chunkid(), line or ln, s)) |
| 132 | end |
| 133 | local errorline = errorline |
| 134 | |
| 135 | ------------------------------------------------------------------------ |
| 136 | -- count separators ("=") in a long string delimiter |
| 137 | ------------------------------------------------------------------------ |
| 138 | |
| 139 | local function skip_sep(i) |
| 140 | local sub = sub |
| 141 | local s = sub(z, i, i) |
| 142 | i = i + 1 |
| 143 | local count = #match(z, "=*", i) -- note, take the length |
| 144 | i = i + count |
| 145 | I = i |
| 146 | return (sub(z, i, i) == s) and count or (-count) - 1 |
| 147 | end |
| 148 | |
| 149 | ---------------------------------------------------------------------- |
| 150 | -- reads a long string or long comment |
| 151 | ---------------------------------------------------------------------- |
| 152 | |
| 153 | local function read_long_string(is_str, sep) |
| 154 | local i = I + 1 -- skip 2nd '[' |
| 155 | local sub = sub |
| 156 | local c = sub(z, i, i) |
| 157 | if c == "\r" or c == "\n" then -- string starts with a newline? |
| 158 | i = inclinenumber(i) -- skip it |
| 159 | end |
| 160 | local j = i |
| 161 | while true do |
| 162 | local p, q, r = find(z, "([\r\n%]])", i) -- (long range) |
| 163 | if not p then |
| 164 | errorline(is_str and "unfinished long string" or |
| 165 | "unfinished long comment") |
| 166 | end |
| 167 | i = p |
| 168 | if r == "]" then -- delimiter test |
| 169 | if skip_sep(i) == sep then |
| 170 | buff = sub(z, buff, I) |
| 171 | I = I + 1 -- skip 2nd ']' |
| 172 | return buff |
| 173 | end |
| 174 | i = I |
| 175 | else -- newline |
| 176 | buff = buff.."\n" |
| 177 | i = inclinenumber(i) |
| 178 | end |
| 179 | end--while |
| 180 | end |
| 181 | |
| 182 | ---------------------------------------------------------------------- |
| 183 | -- reads a string |
| 184 | ---------------------------------------------------------------------- |
| 185 | |
| 186 | local function read_string(del) |
| 187 | local i = I |
| 188 | local find = find |
| 189 | local sub = sub |
| 190 | while true do |
| 191 | local p, q, r = find(z, "([\n\r\\\"\'])", i) -- (long range) |
| 192 | if p then |
| 193 | if r == "\n" or r == "\r" then |
| 194 | errorline("unfinished string") |
| 195 | end |
| 196 | i = p |
| 197 | if r == "\\" then -- handle escapes |
| 198 | i = i + 1 |
| 199 | r = sub(z, i, i) |
| 200 | if r == "" then break end -- (EOZ error) |
| 201 | p = find("abfnrtv\n\r", r, 1, true) |
| 202 | ------------------------------------------------------ |
| 203 | if p then -- special escapes |
| 204 | if p > 7 then |
| 205 | i = inclinenumber(i) |
| 206 | else |
| 207 | i = i + 1 |
| 208 | end |
| 209 | ------------------------------------------------------ |
| 210 | elseif find(r, "%D") then -- other non-digits |
| 211 | i = i + 1 |
| 212 | ------------------------------------------------------ |
| 213 | else -- \xxx sequence |
| 214 | local p, q, s = find(z, "^(%d%d?%d?)", i) |
| 215 | i = q + 1 |
| 216 | if s + 1 > 256 then -- UCHAR_MAX |
| 217 | errorline("escape sequence too large") |
| 218 | end |
| 219 | ------------------------------------------------------ |
| 220 | end--if p |
| 221 | else |
| 222 | i = i + 1 |
| 223 | if r == del then -- ending delimiter |
| 224 | I = i |
| 225 | return sub(z, buff, i - 1) -- return string |
| 226 | end |
| 227 | end--if r |
| 228 | else |
| 229 | break -- (error) |
| 230 | end--if p |
| 231 | end--while |
| 232 | errorline("unfinished string") |
| 233 | end |
| 234 | |
| 235 | ------------------------------------------------------------------------ |
| 236 | -- main lexer function |
| 237 | ------------------------------------------------------------------------ |
| 238 | |
| 239 | function llex() |
| 240 | local find = find |
| 241 | local match = match |
| 242 | while true do--outer |
| 243 | local i = I |
| 244 | -- inner loop allows break to be used to nicely section tests |
| 245 | while true do--inner |
| 246 | ---------------------------------------------------------------- |
| 247 | local p, _, r = find(z, "^([_%a][_%w]*)", i) |
| 248 | if p then |
| 249 | I = i + #r |
| 250 | if kw[r] then |
| 251 | addtoken("TK_KEYWORD", r) -- reserved word (keyword) |
| 252 | else |
| 253 | addtoken("TK_NAME", r) -- identifier |
| 254 | end |
| 255 | break -- (continue) |
| 256 | end |
| 257 | ---------------------------------------------------------------- |
| 258 | local p, _, r = find(z, "^(%.?)%d", i) |
| 259 | if p then -- numeral |
| 260 | if r == "." then i = i + 1 end |
| 261 | local _, q, r = find(z, "^%d*[%.%d]*([eE]?)", i) |
| 262 | i = q + 1 |
| 263 | if #r == 1 then -- optional exponent |
| 264 | if match(z, "^[%+%-]", i) then -- optional sign |
| 265 | i = i + 1 |
| 266 | end |
| 267 | end |
| 268 | local _, q = find(z, "^[_%w]*", i) |
| 269 | I = q + 1 |
| 270 | local v = sub(z, p, q) -- string equivalent |
| 271 | if not base.tonumber(v) then -- handles hex test also |
| 272 | errorline("malformed number") |
| 273 | end |
| 274 | addtoken("TK_NUMBER", v) |
| 275 | break -- (continue) |
| 276 | end |
| 277 | ---------------------------------------------------------------- |
| 278 | local p, q, r, t = find(z, "^((%s)[ \t\v\f]*)", i) |
| 279 | if p then |
| 280 | if t == "\n" or t == "\r" then -- newline |
| 281 | inclinenumber(i, true) |
| 282 | else |
| 283 | I = q + 1 -- whitespace |
| 284 | addtoken("TK_SPACE", r) |
| 285 | end |
| 286 | break -- (continue) |
| 287 | end |
| 288 | ---------------------------------------------------------------- |
| 289 | local r = match(z, "^%p", i) |
| 290 | if r then |
| 291 | buff = i |
| 292 | local p = find("-[\"\'.=<>~", r, 1, true) |
| 293 | if p then |
| 294 | -- two-level if block for punctuation/symbols |
| 295 | -------------------------------------------------------- |
| 296 | if p <= 2 then |
| 297 | if p == 1 then -- minus |
| 298 | local c = match(z, "^%-%-(%[?)", i) |
| 299 | if c then |
| 300 | i = i + 2 |
| 301 | local sep = -1 |
| 302 | if c == "[" then |
| 303 | sep = skip_sep(i) |
| 304 | end |
| 305 | if sep >= 0 then -- long comment |
| 306 | addtoken("TK_LCOMMENT", read_long_string(false, sep)) |
| 307 | else -- short comment |
| 308 | I = find(z, "[\n\r]", i) or (#z + 1) |
| 309 | addtoken("TK_COMMENT", sub(z, buff, I - 1)) |
| 310 | end |
| 311 | break -- (continue) |
| 312 | end |
| 313 | -- (fall through for "-") |
| 314 | else -- [ or long string |
| 315 | local sep = skip_sep(i) |
| 316 | if sep >= 0 then |
| 317 | addtoken("TK_LSTRING", read_long_string(true, sep)) |
| 318 | elseif sep == -1 then |
| 319 | addtoken("TK_OP", "[") |
| 320 | else |
| 321 | errorline("invalid long string delimiter") |
| 322 | end |
| 323 | break -- (continue) |
| 324 | end |
| 325 | -------------------------------------------------------- |
| 326 | elseif p <= 5 then |
| 327 | if p < 5 then -- strings |
| 328 | I = i + 1 |
| 329 | addtoken("TK_STRING", read_string(r)) |
| 330 | break -- (continue) |
| 331 | end |
| 332 | r = match(z, "^%.%.?%.?", i) -- .|..|... dots |
| 333 | -- (fall through) |
| 334 | -------------------------------------------------------- |
| 335 | else -- relational |
| 336 | r = match(z, "^%p=?", i) |
| 337 | -- (fall through) |
| 338 | end |
| 339 | end |
| 340 | I = i + #r |
| 341 | addtoken("TK_OP", r) -- for other symbols, fall through |
| 342 | break -- (continue) |
| 343 | end |
| 344 | ---------------------------------------------------------------- |
| 345 | local r = sub(z, i, i) |
| 346 | if r ~= "" then |
| 347 | I = i + 1 |
| 348 | addtoken("TK_OP", r) -- other single-char tokens |
| 349 | break |
| 350 | end |
| 351 | addtoken("TK_EOS", "") -- end of stream, |
| 352 | return -- exit here |
| 353 | ---------------------------------------------------------------- |
| 354 | end--while inner |
| 355 | end--while outer |
| 356 | end |
| 357 | |
| 358 | return { |
| 359 | init = init, |
| 360 | llex = llex |
| 361 | } |