fangle

   1 #! /usr/bin/awk -f
   2 # fangle - fully featured notangle replacement in awk
   3 #
   4 #  Copyright (C) 2009-2010 Sam Liddicott <sam@liddicott.com>
   5 #
   6 #  This program is free software: you can redistribute it and/or modify
   7 #  it under the terms of the GNU General Public License as published by
   8 #  the Free Software Foundation, either version 3 of the License, or
   9 #  (at your option) any later version.
  10 #
  11 #  This program is distributed in the hope that it will be useful,
  12 #  but WITHOUT ANY WARRANTY; without even the implied warranty of
  13 #  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  14 #  GNU General Public License for more details.
  15 #
  16 #  You should have received a copy of the GNU General Public License
  17 #  along with this program.  If not, see <http://www.gnu.org/licenses/>.
  18 # NOTE: Arnold Robbins public domain getopt for awk is also used:
  19 # getopt.awk --- do C library getopt(3) function in awk
  20 #
  21 # Arnold Robbins, arnold@skeeve.com, Public Domain
  22 #
  23 # Initial version: March, 1991
  24 # Revised: May, 1993
  25
  26 function getopt(argc, argv, options,    thisopt, i)
  27 {
  28     if (length(options) == 0)    # no options given
  29         return -1
  30     if (argv[Optind] == "--") {  # all done
  31         Optind++
  32         _opti = 0
  33         return -1
  34     } else if (argv[Optind] !~ /^-[^: \t\n\f\r\v\b]/) {
  35         _opti = 0
  36         return -1
  37     }
  38     if (_opti == 0)
  39         _opti = 2
  40     thisopt = substr(argv[Optind], _opti, 1)
  41     Optopt = thisopt
  42     i = index(options, thisopt)
  43     if (i == 0) {
  44         if (Opterr)
  45             printf("%c -- invalid option\n",
  46                                   thisopt) > "/dev/stderr"
  47         if (_opti >= length(argv[Optind])) {
  48             Optind++
  49             _opti = 0
  50         } else
  51             _opti++
  52         return "?"
  53     }
  54     if (substr(options, i + 1, 1) == ":") {
  55         # get option argument
  56         if (length(substr(argv[Optind], _opti + 1)) > 0)
  57             Optarg = substr(argv[Optind], _opti + 1)
  58         else
  59             Optarg = argv[++Optind]
  60         _opti = 0
  61     } else
  62         Optarg = ""
  63     if (_opti == 0 || _opti >= length(argv[Optind])) {
  64         Optind++
  65         _opti = 0
  66     } else
  67         _opti++
  68     return thisopt
  69 }
  70
  71 function error(message)
  72 {
  73   print "ERROR: " FILENAME ":" FNR " " message > "/dev/stderr";
  74   exit 1;
  75 }
  76 function warning(message)
  77 {
  78   print "WARNING: " FILENAME ":" FNR " " message > "/dev/stderr";
  79   warnings++;
  80 }
  81 function debug_log(message)
  82 {
  83   print "DEBUG: " FILENAME ":" FNR " " message > "/dev/stderr";
  84 }
  85 function new_mode_tracker(context, language, mode) {
  86   context[""] = 0;
  87   context[0, "language"] = language;
  88   context[0, "mode"] = mode;
  89 }
  90 function push_mode_tracker(context, language, mode,
  91   # local vars
  92   top)
  93 {
  94   if (! ("" in context)) {
  95     split("", context);
  96     new_mode_tracker(context, language, mode);
  97     return;
  98   } else {
  99     top = context[""];
 100     if (context[top, "language"] == language && mode=="") mode = context[top, "mode"];
 101     old_top = top;
 102     top++;
 103     context[top, "language"] = language;
 104     context[top, "mode"] = mode;
 105     context[""] = top;
 106   }
 107   return old_top;
 108 }
 109 function dump_mode_tracker(context,
 110   c, d)
 111 {
 112   for(c=0; c <= context[""]; c++) {
 113     printf(" %2d   %s:%s\n", c, context[c, "language"], context[c, "mode"]) > "/dev/stderr";
 114     for(d=1; ( (c, "values", d) in context); d++) {
 115       printf("   %2d %s\n", d, context[c, "values", d]) > "/dev/stderr";
 116     }
 117   }
 118 }
 119 function pop_mode_tracker(context, context_origin)
 120 {
 121   if ( (context_origin) && ("" in context) && context[""] != (1+context_origin)) return 0;
 122   context[""] = context_origin;
 123   return 1;
 124 }
 125 function mode_tracker(context, text, values,
 126   # optional parameters
 127   # local vars
 128   mode, submodes, language,
 129   cindex, c, a, part, item, name, result, new_values, new_mode,
 130   delimiters, terminators)
 131 {
 132   cindex = context[""] + 0;
 133   mode = context[cindex, "mode"];
 134   language = context[cindex, "language" ];
 135   submodes=modes[language, mode, "submodes"];
 136
 137   if ((language, mode, "delimiters") in modes) {
 138     delimiters = modes[language, mode, "delimiters"];
 139     if (length(submodes)>0) submodes = submodes "|";
 140     submodes=submodes delimiters;
 141   } else delimiters="";
 142   if ((language, mode, "terminators") in modes) {
 143     terminators = modes[language, mode, "terminators"];
 144     if (length(submodes)>0) submodes = submodes "|";
 145     submodes=submodes terminators;
 146   } else terminators="";
 147  if (! length(submodes)) return text;
 148   while((cindex >= 0) && length(text)) {
 149     if (match(text, "(" submodes ")", a)) {
 150       if (RLENGTH<1) {
 151         error(sprintf("Internal error, matched zero length submode, should be impossible - likely regex computation error\n" \
 152                 "Language=%s\nmode=%s\nmatch=%s\n", language, mode, submodes));
 153       }
 154       part = substr(text, 1, RSTART -1);
 155       item = item part;
 156       if (match(a[1], "^" terminators "$")) {
 157 #printf("%2d EXIT  MODE [%s] by [%s] [%s]\n", cindex, mode, a[1], text) > "/dev/stderr"
 158         context[cindex, "values", ++context[cindex, "values"]] = item;
 159         delete context[cindex];
 160         context[""] = --cindex;
 161         if (cindex>=0) {
 162           mode = context[cindex, "mode"];
 163           language = context[cindex, "language"];
 164             submodes=modes[language, mode, "submodes"];
 165
 166             if ((language, mode, "delimiters") in modes) {
 167               delimiters = modes[language, mode, "delimiters"];
 168               if (length(submodes)>0) submodes = submodes "|";
 169               submodes=submodes delimiters;
 170             } else delimiters="";
 171             if ((language, mode, "terminators") in modes) {
 172               terminators = modes[language, mode, "terminators"];
 173               if (length(submodes)>0) submodes = submodes "|";
 174               submodes=submodes terminators;
 175             } else terminators="";
 176            if (! length(submodes)) return text;
 177         }
 178         item = item a[1];
 179         text = substr(text, 1 + length(part) + length(a[1]));
 180       }
 181       else if (match(a[1], "^" delimiters "$")) {
 182         if (cindex==0) {
 183           context[cindex, "values", ++context[cindex, "values"]] = item;
 184           item = "";
 185         } else {
 186           item = item a[1];
 187         }
 188         text = substr(text, 1 + length(part) + length(a[1]));
 189       }
 190  else if ((language, a[1], "terminators") in modes) {
 191         #check if new_mode is defined
 192         item = item a[1];
 193 #printf("%2d ENTER MODE [%s] in [%s]\n", cindex, a[1], text) > "/dev/stderr"
 194         text = substr(text, 1 + length(part) + length(a[1]));
 195         context[""] = ++cindex;
 196         context[cindex, "mode"] = a[1];
 197         context[cindex, "language"] = language;
 198         mode = a[1];
 199           submodes=modes[language, mode, "submodes"];
 200
 201           if ((language, mode, "delimiters") in modes) {
 202             delimiters = modes[language, mode, "delimiters"];
 203             if (length(submodes)>0) submodes = submodes "|";
 204             submodes=submodes delimiters;
 205           } else delimiters="";
 206           if ((language, mode, "terminators") in modes) {
 207             terminators = modes[language, mode, "terminators"];
 208             if (length(submodes)>0) submodes = submodes "|";
 209             submodes=submodes terminators;
 210           } else terminators="";
 211          if (! length(submodes)) return text;
 212       } else {
 213         error(sprintf("Submode '%s' set unknown mode in text: %s\nLanguage %s Mode %s\n", a[1], text, language, mode));
 214         text = substr(text, 1 + length(part) + length(a[1]));
 215       }
 216     }
 217 else {
 218       context[cindex, "values", ++context[cindex, "values"]] = item text;
 219       text = "";
 220       item = "";
 221     }
 222   }
 223
 224   context["item"] = item;
 225
 226   if (length(item)) context[cindex, "values", ++context[cindex, "values"]] = item;
 227   return text;
 228 }
 229
 230 function untab(text) {
 231   gsub("[[:space:]]*\xE2\x86\xA4","", text);
 232   return text;
 233 }
 234 function transform_escape(context, text, top,
 235   c, cp, cpl, s, r)
 236 {
 237   for(c = top; c >= 0; c--) {
 238     if ( (context[c, "language"], context[c, "mode"]) in escapes) {
 239       cpl = escapes[context[c, "language"], context[c, "mode"]];
 240       for (cp = 1; cp <= cpl; cp ++) {
 241         s = escapes[context[c, "language"], context[c, "mode"], cp, "s"];
 242         r = escapes[context[c, "language"], context[c, "mode"], cp, "r"];
 243         if (length(s)) {
 244           gsub(s, r, text);
 245         }
 246         if ( (context[c, "language"], context[c, "mode"], cp, "t") in escapes ) {
 247           quotes[src, "t"] = escapes[context[c, "language"], context[c, "mode"], cp, "t"];
 248         }
 249       }
 250     }
 251   }
 252   return text;
 253 }
 254 function dump_escaper(quotes, r, cc) {
 255   for(cc=1; cc<=c; cc++) {
 256     printf("%2d s[%s] r[%s]\n", cc, quotes[cc, "s"], quotes[cc, "r"]) > "/dev/stderr"
 257   }
 258 }
 259 function parse_chunk_args(language, text, values, mode,
 260   # local vars
 261   c, context, rest)
 262 {
 263   split("", context);
 264   new_mode_tracker(context, language, mode);
 265   rest = mode_tracker(context, text, values);
 266   # extract values
 267   for(c=1; c <= context[0, "values"]; c++) {
 268     values[c] = context[0, "values", c];
 269   }
 270   return rest;
 271 }
 272 function new_chunk(chunk_name, opts, args,
 273   # local vars
 274   p, append )
 275 {
 276   # HACK WHILE WE CHANGE TO ( ) for PARAM CHUNKS
 277   gsub("\\(\\)$", "", chunk_name);
 278   if (! (chunk_name in chunk_names)) {
 279     if (debug) print "New chunk " chunk_name;
 280     chunk_names[chunk_name];
 281     for (p in opts) {
 282       chunks[chunk_name, p] = opts[p];
 283       if (debug) print "chunks[" chunk_name "," p "] = " opts[p];
 284     }
 285     for (p in args) {
 286       chunks[chunk_name, "params", p] = args[p];
 287     }
 288     if ("append" in opts) {
 289       append=opts["append"];
 290       if (! (append in chunk_names)) {
 291         warning("Chunk " chunk_name " is appended to chunk " append " which is not defined yet");
 292         new_chunk(append);
 293       }
 294       chunk_include(append, chunk_name);
 295       chunk_line(append, ORS);
 296     }
 297   }
 298   active_chunk = chunk_name;
 299   prime_chunk(chunk_name);
 300 }
 301
 302 function prime_chunk(chunk_name)
 303 {
 304   chunks[chunk_name, "part", ++chunks[chunk_name, "part"] ] = \
 305          chunk_name SUBSEP "chunklet" SUBSEP "" ++chunks[chunk_name, "chunklet"];
 306   chunks[chunk_name, "part", chunks[chunk_name, "part"], "FILENAME"] = FILENAME;
 307   chunks[chunk_name, "part", chunks[chunk_name, "part"], "LINENO"] = FNR + 1;
 308 }
 309
 310 function chunk_line(chunk_name, line){
 311   chunks[chunk_name, "chunklet", chunks[chunk_name, "chunklet"],
 312          ++chunks[chunk_name, "chunklet", chunks[chunk_name, "chunklet"], "line"]  ] = line;
 313 }
 314
 315 function chunk_include(chunk_name, chunk_ref, indent, tail)
 316 {
 317   chunks[chunk_name, "part", ++chunks[chunk_name, "part"] ] = chunk_ref;
 318   chunks[chunk_name, "part", chunks[chunk_name, "part"], "type" ] = part_type_chunk;
 319   chunks[chunk_name, "part", chunks[chunk_name, "part"], "indent" ] = indent_string(indent);
 320   chunks[chunk_name, "part", chunks[chunk_name, "part"], "tail" ] = tail;
 321   prime_chunk(chunk_name);
 322 }
 323
 324 function indent_string(indent) {
 325   return sprintf("%" indent "s", "");
 326 }
 327 function output_chunk_names(   c, prefix, suffix)
 328 {
 329   if (notangle_mode) {
 330     prefix="<<";
 331     suffix=">>";
 332   }
 333   for (c in chunk_names) {
 334     print prefix c suffix "\n";
 335   }
 336 }
 337 function output_chunks(  a)
 338 {
 339   for (a in chunk_names) {
 340     output_chunk(a);
 341   }
 342 }
 343
 344 function output_chunk(chunk) {
 345   newline = 1;
 346   lineno_needed = linenos;
 347
 348   write_chunk(chunk);
 349 }
 350
 351 function write_chunk(chunk_name) {
 352   split("", context);
 353   return write_chunk_r(chunk_name, context);
 354 }
 355
 356 function write_chunk_r(chunk_name, context, indent, tail,
 357   # optional vars
 358   chunk_path, chunk_args,
 359   # local vars
 360   context_origin,
 361   chunk_params, part, max_part, part_line, frag, max_frag, text,
 362   chunklet, only_part, call_chunk_args, new_context)
 363 {
 364   if (debug) debug_log("write_chunk_r(" chunk_name ")");
 365   if (match(chunk_name, "^(.*)\\[([0-9]*)\\]$", chunk_name_parts)) {
 366     chunk_name = chunk_name_parts[1];
 367     only_part = chunk_name_parts[2];
 368   }
 369   context_origin = push_mode_tracker(context, chunks[chunk_name, "language"], "");
 370   split(chunks[chunk_name, "params"], chunk_params, " *; *");
 371   if (! (chunk_name in chunk_names)) {
 372     error(sprintf(_"The root module <<%s>> was not defined.\nUsed by: %s",\
 373                   chunk_name, chunk_path));
 374   }
 375
 376   max_part = chunks[chunk_name, "part"];
 377   for(part = 1; part <= max_part; part++) {
 378     if (! only_part || part == only_part) {
 379       if (linenos && (chunk_name SUBSEP "part" SUBSEP part SUBSEP "FILENAME" in chunks)) {
 380         a_filename = chunks[chunk_name, "part", part, "FILENAME"];
 381         a_lineno = chunks[chunk_name, "part", part, "LINENO"];
 382         if (a_filename != filename || a_lineno != lineno) {
 383           lineno_needed++;
 384         }
 385       }
 386
 387       chunklet = chunks[chunk_name, "part", part];
 388       if (chunks[chunk_name, "part", part, "type"] == part_type_chunk) {
 389         if (match(chunklet, "^([^\\[\\(]*)\\((.*)\\)$", chunklet_parts)) {
 390           chunklet = chunklet_parts[1];
 391         # hack
 392         gsub(sprintf("%c",11), "", chunklet);
 393         gsub(sprintf("%c",11), "", chunklet_parts[2]);
 394           parse_chunk_args("c-like", chunklet_parts[2], call_chunk_args, "(");
 395           for (c in call_chunk_args) {
 396             call_chunk_args[c] = expand_chunk_args(call_chunk_args[c], chunk_params, chunk_args);
 397           }
 398         } else {
 399           split("", call_chunk_args);
 400         }
 401
 402         write_chunk_r(chunklet, context,
 403                     chunks[chunk_name, "part", part, "indent"] indent,
 404                     chunks[chunk_name, "part", part, "tail"],
 405                     chunk_path "\n         " chunk_name,
 406                     call_chunk_args);
 407       } else if (chunklet SUBSEP "line" in chunks) {
 408         max_frag = chunks[chunklet, "line"];
 409         for(frag = 1; frag <= max_frag; frag++) {
 410           if (newline && lineno_needed && ! lineno_suppressed) {
 411             filename = a_filename;
 412             lineno = a_lineno;
 413             print "#line " lineno " \"" filename "\"\n"
 414             lineno_needed = 0;
 415           }
 416
 417           text = chunks[chunklet, frag];
 418
 419           /* check params */
 420           text = expand_chunk_args(text, chunk_params, chunk_args);
 421
 422          if (text == "\n") {
 423             lineno++;
 424             if (part == max_part && frag == max_frag && length(chunk_path)) {
 425               text = "";
 426               break;
 427             } else {
 428               newline = 1;
 429             }
 430           } else if (length(text) || length(tail)) {
 431             if (newline) text = indent text;
 432             newline = 0;
 433           }
 434
 435           text = text tail;
 436           mode_tracker(context, text);
 437           print untab(transform_escape(context, text, context_origin));
 438           if (linenos) {
 439             lineno_suppressed = substr(lastline, length(lastline)) == "\\";
 440           }
 441         }
 442       } else {
 443         # empty last chunklet
 444       }
 445     }
 446   }
 447   if (! pop_mode_tracker(context, context_origin)) {
 448     dump_mode_tracker(context);
 449     error(sprintf(_"Module %s did not close context properly.\nUsed by: %s\n", chunk_name, chunk_path));
 450   }
 451 }
 452 function expand_chunk_args(text, params, args,
 453   p, text_array, next_text, v, t, l)
 454 {
 455   if (split(text, text_array, "\\${")) {
 456     for(p in params) {
 457       v[params[p]]=args[p];
 458     }
 459     text=text_array[1];
 460     for(t=2; t in text_array; t++) {
 461       if (match(text_array[t], "^([a-zA-Z_][a-zA-Z0-9_]*)}", l) &&
 462           l[1] in v)
 463       {
 464         text = text v[l[1]] substr(text_array[t], length(l[1])+2);
 465       } else {
 466         text = text "${" text_array[t];
 467       }
 468     }
 469   }
 470
 471   return text;
 472 }
 473
 474 BEGIN {
 475   ARG_SEPARATOR=sprintf("%c", 11);
 476   part_type_chunk=1;
 477   SUBSEP=",";
 478   modes["c-like", "",  "submodes"]="\\\\|\"|'|{|\\(|\\[";
 479   modes["c-like", "",  "delimiters"]=" *, *";
 480   modes["c-like", "\\", "terminators"]=".";
 481   modes["c-like", "\"", "submodes"]="\\\\";
 482   modes["c-like", "\"", "terminators"]="\"";
 483   escapes["c-like", "\"", ++escapes["c-like", "\""], "s"]="\\\\";
 484   escapes["c-like", "\"",   escapes["c-like", "\""], "r"]="\\\\";
 485   escapes["c-like", "\"", ++escapes["c-like", "\""], "s"]="\"";
 486   escapes["c-like", "\"",   escapes["c-like", "\""], "r"]="\\" "\"";
 487   escapes["c-like", "\"", ++escapes["c-like", "\""], "s"]="\n";
 488   escapes["c-like", "\"",   escapes["c-like", "\""], "r"]="\\n";
 489   modes["c-like", "'", "submodes"]="\\\\";
 490   modes["c-like", "'", "terminators"]="'";
 491   escapes["c-like", "'", ++escapes["c-like", "'"], "s"]="\\\\";
 492   escapes["c-like", "'",   escapes["c-like", "'"], "r"]="\\\\";
 493   escapes["c-like", "'", ++escapes["c-like", "'"], "s"]="'";
 494   escapes["c-like", "'",   escapes["c-like", "'"], "r"]="\\" "'";
 495   escapes["c-like", "'", ++escapes["c-like", "'"], "s"]="\n";
 496   escapes["c-like", "'",   escapes["c-like", "'"], "r"]="\\n";
 497   modes["c-like", "{",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 498   modes["c-like", "{",  "delimiters"]=" *, *";
 499   modes["c-like", "{",  "terminators"]="}";
 500   modes["c-like", "[",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 501   modes["c-like", "[",  "delimiters"]=" *, *";
 502   modes["c-like", "[",  "terminators"]="\\]";
 503   modes["c-like", "(",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 504   modes["c-like", "(",  "delimiters"]=" *, *";
 505   modes["c-like", "(",  "terminators"]="\\)";
 506
 507   modes["c", "",  "submodes"]="\\\\|\"|'|{|\\(|\\[";
 508   modes["c", "",  "delimiters"]=" *, *";
 509   modes["c", "\\", "terminators"]=".";
 510   modes["c", "\"", "submodes"]="\\\\";
 511   modes["c", "\"", "terminators"]="\"";
 512   escapes["c", "\"", ++escapes["c", "\""], "s"]="\\\\";
 513   escapes["c", "\"",   escapes["c", "\""], "r"]="\\\\";
 514   escapes["c", "\"", ++escapes["c", "\""], "s"]="\"";
 515   escapes["c", "\"",   escapes["c", "\""], "r"]="\\" "\"";
 516   escapes["c", "\"", ++escapes["c", "\""], "s"]="\n";
 517   escapes["c", "\"",   escapes["c", "\""], "r"]="\\n";
 518   modes["c", "'", "submodes"]="\\\\";
 519   modes["c", "'", "terminators"]="'";
 520   escapes["c", "'", ++escapes["c", "'"], "s"]="\\\\";
 521   escapes["c", "'",   escapes["c", "'"], "r"]="\\\\";
 522   escapes["c", "'", ++escapes["c", "'"], "s"]="'";
 523   escapes["c", "'",   escapes["c", "'"], "r"]="\\" "'";
 524   escapes["c", "'", ++escapes["c", "'"], "s"]="\n";
 525   escapes["c", "'",   escapes["c", "'"], "r"]="\\n";
 526   modes["c", "{",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 527   modes["c", "{",  "delimiters"]=" *, *";
 528   modes["c", "{",  "terminators"]="}";
 529   modes["c", "[",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 530   modes["c", "[",  "delimiters"]=" *, *";
 531   modes["c", "[",  "terminators"]="\\]";
 532   modes["c", "(",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 533   modes["c", "(",  "delimiters"]=" *, *";
 534   modes["c", "(",  "terminators"]="\\)";
 535   modes["c", "", "submodes"] = modes["c", "", "submodes"] "|" "/\\*";
 536   modes["c", "/*", "terminators"]="\\*/";
 537   modes["c", "", "submodes"] = modes["c", "", "submodes"] "|" "//";
 538   modes["c", "//", "terminators"]="\n";
 539   escapes["c", "//", ++escapes["c", "//"], "s"]="\n";
 540   escapes["c", "//",   escapes["c", "//"], "r"]="\n//";
 541   modes["c", "", "submodes"] = modes["c", "", "submodes"] "|" "#";
 542   modes["c", "#", "submodes" ]="\\\\";
 543   modes["c", "#", "terminators"]="\n";
 544   escapes["c", "#", ++escapes["c", "#"], "s"]="\n";
 545   escapes["c", "#",   escapes["c", "#"], "r"]="\\\\\n";
 546
 547   modes["awk", "",  "submodes"]="\\\\|\"|'|{|\\(|\\[";
 548   modes["awk", "",  "delimiters"]=" *, *";
 549   modes["awk", "\\", "terminators"]=".";
 550   modes["awk", "\"", "submodes"]="\\\\";
 551   modes["awk", "\"", "terminators"]="\"";
 552   escapes["awk", "\"", ++escapes["awk", "\""], "s"]="\\\\";
 553   escapes["awk", "\"",   escapes["awk", "\""], "r"]="\\\\";
 554   escapes["awk", "\"", ++escapes["awk", "\""], "s"]="\"";
 555   escapes["awk", "\"",   escapes["awk", "\""], "r"]="\\" "\"";
 556   escapes["awk", "\"", ++escapes["awk", "\""], "s"]="\n";
 557   escapes["awk", "\"",   escapes["awk", "\""], "r"]="\\n";
 558   modes["awk", "'", "submodes"]="\\\\";
 559   modes["awk", "'", "terminators"]="'";
 560   escapes["awk", "'", ++escapes["awk", "'"], "s"]="\\\\";
 561   escapes["awk", "'",   escapes["awk", "'"], "r"]="\\\\";
 562   escapes["awk", "'", ++escapes["awk", "'"], "s"]="'";
 563   escapes["awk", "'",   escapes["awk", "'"], "r"]="\\" "'";
 564   escapes["awk", "'", ++escapes["awk", "'"], "s"]="\n";
 565   escapes["awk", "'",   escapes["awk", "'"], "r"]="\\n";
 566   modes["awk", "{",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 567   modes["awk", "{",  "delimiters"]=" *, *";
 568   modes["awk", "{",  "terminators"]="}";
 569   modes["awk", "[",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 570   modes["awk", "[",  "delimiters"]=" *, *";
 571   modes["awk", "[",  "terminators"]="\\]";
 572   modes["awk", "(",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 573   modes["awk", "(",  "delimiters"]=" *, *";
 574   modes["awk", "(",  "terminators"]="\\)";
 575   modes["awk", "", "submodes"] = modes["awk", "", "submodes"] "|" "#";
 576   modes["awk", "#", "terminators"]="\n";
 577   escapes["awk", "#", ++escapes["awk", "#"], "s"]="\n";
 578   escapes["awk", "#",   escapes["awk", "#"], "r"]="\n#";
 579   modes["awk", "", "submodes"] = modes["awk", "", "submodes"] "|" "/\\^";
 580   modes["awk", "/^", "terminators"]="/";
 581   modes["perl", "",  "submodes"]="\\\\|\"|'|{|\\(|\\[";
 582   modes["perl", "",  "delimiters"]=" *, *";
 583   modes["perl", "\\", "terminators"]=".";
 584   modes["perl", "\"", "submodes"]="\\\\";
 585   modes["perl", "\"", "terminators"]="\"";
 586   escapes["perl", "\"", ++escapes["perl", "\""], "s"]="\\\\";
 587   escapes["perl", "\"",   escapes["perl", "\""], "r"]="\\\\";
 588   escapes["perl", "\"", ++escapes["perl", "\""], "s"]="\"";
 589   escapes["perl", "\"",   escapes["perl", "\""], "r"]="\\" "\"";
 590   escapes["perl", "\"", ++escapes["perl", "\""], "s"]="\n";
 591   escapes["perl", "\"",   escapes["perl", "\""], "r"]="\\n";
 592   modes["perl", "'", "submodes"]="\\\\";
 593   modes["perl", "'", "terminators"]="'";
 594   escapes["perl", "'", ++escapes["perl", "'"], "s"]="\\\\";
 595   escapes["perl", "'",   escapes["perl", "'"], "r"]="\\\\";
 596   escapes["perl", "'", ++escapes["perl", "'"], "s"]="'";
 597   escapes["perl", "'",   escapes["perl", "'"], "r"]="\\" "'";
 598   escapes["perl", "'", ++escapes["perl", "'"], "s"]="\n";
 599   escapes["perl", "'",   escapes["perl", "'"], "r"]="\\n";
 600   modes["perl", "{",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 601   modes["perl", "{",  "delimiters"]=" *, *";
 602   modes["perl", "{",  "terminators"]="}";
 603   modes["perl", "[",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 604   modes["perl", "[",  "delimiters"]=" *, *";
 605   modes["perl", "[",  "terminators"]="\\]";
 606   modes["perl", "(",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 607   modes["perl", "(",  "delimiters"]=" *, *";
 608   modes["perl", "(",  "terminators"]="\\)";
 609   modes["perl", "", "submodes"] = modes["perl", "", "submodes"] "|" "/\\*";
 610   modes["perl", "/*", "terminators"]="\\*/";
 611   modes["perl", "", "submodes"] = modes["perl", "", "submodes"] "|" "#";
 612   modes["perl", "#", "terminators"]="\n";
 613   escapes["perl", "#", ++escapes["perl", "#"], "s"]="\n";
 614   escapes["perl", "#",   escapes["perl", "#"], "r"]="\n#";
 615   modes["sh", "",  "submodes"]="\\\\|\"|'|{|\\(|\\[|\\$\\(";
 616   modes["sh", "\\", "terminators"]=".";
 617
 618   modes["sh", "\"", "submodes"]="\\\\|\\$\\(";
 619   modes["sh", "\"", "terminators"]="\"";
 620   escapes["sh", "\"", ++escapes["sh", "\""], "s"]="\\\\";
 621   escapes["sh", "\"",   escapes["sh", "\""], "r"]="\\\\";
 622   escapes["sh", "\"", ++escapes["sh", "\""], "s"]="\"";
 623   escapes["sh", "\"",   escapes["sh", "\""], "r"]="\\" "\"";
 624   escapes["sh", "\"", ++escapes["sh", "\""], "s"]="\n";
 625   escapes["sh", "\"",   escapes["sh", "\""], "r"]="\\n";
 626
 627   modes["sh", "'", "terminators"]="'";
 628   escapes["sh", "'", ++escapes["sh", "'"], "s"]="'";
 629   escapes["sh", "'",   escapes["sh", "'"], "r"]="'\\'" "'";
 630   modes["sh", "$(",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 631   modes["sh", "$(",  "delimiters"]=" *, *";
 632   modes["sh", "$(",  "terminators"]="\\)";
 633   escapes["sh", "$(", ++escapes["sh", "$("], "tunnel"]="";
 634   modes["sh", "{",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 635   modes["sh", "{",  "delimiters"]=" *, *";
 636   modes["sh", "{",  "terminators"]="}";
 637   modes["sh", "[",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 638   modes["sh", "[",  "delimiters"]=" *, *";
 639   modes["sh", "[",  "terminators"]="\\]";
 640   modes["sh", "(",  "submodes" ]="\\\\|\"|{|\\(|\\[|'|/\\*";
 641   modes["sh", "(",  "delimiters"]=" *, *";
 642   modes["sh", "(",  "terminators"]="\\)";
 643   modes["sh", "", "submodes"] = modes["sh", "", "submodes"] "|" "#";
 644   modes["sh", "#", "terminators"]="\n";
 645   escapes["sh", "#", ++escapes["sh", "#"], "s"]="\n";
 646   escapes["sh", "#",   escapes["sh", "#"], "r"]="\n#";
 647   escapes["sh", "\"", ++escapes["sh", "\""], "s"]="\\$";
 648   escapes["sh", "\"",   escapes["sh", "\""], "r"]="\\$";
 649   modes["make", "",  "submodes"]="      ";
 650   escapes["make", "", ++escapes["make", ""], "s"]="\\$";
 651   escapes["make", "", escapes["make", ""], "r"]="$$";
 652   modes["make", "       ", "terminators"]="\\n";
 653   escapes["make", "     ", ++escapes["make", "  "], "s"]="\\n";
 654   escapes["make", "     ", escapes["make", "    "], "r"]=" ;\\\n        ";
 655   debug=0;
 656   linenos=0;
 657   notangle_mode=0;
 658   root="*";
 659   tabs = "";
 660
 661   Optind = 1    # skip ARGV[0]
 662   while(getopt(ARGC, ARGV, "R:LdT:hr")!=-1) {
 663     if (Optopt == "R") root = Optarg;
 664     else if (Optopt == "r") root="";
 665     else if (Optopt == "L") linenos = 1;
 666     else if (Optopt == "d") debug = 1;
 667     else if (Optopt == "T") tabs = indent_string(Optarg+0);
 668     else if (Optopt == "h") help();
 669     else if (Optopt == "?") help();
 670   }
 671   for (i=1; i<Optind; i++) { ARGV[i]=""; }
 672 }
 673
 674 #/\n/ {
 675 #  gsub("\n*$","");
 676 #  gsub("\n", " ");
 677 #}
 678 #===
 679 /\xE2\x86\xA6/ {
 680   gsub("\\xE2\\x86\\xA6", "\x09");
 681 }
 682
 683 /\xE2\x80\x98/ {
 684   gsub("\\xE2\\x80\\x98", "`");
 685 }
 686
 687 /\xE2\x89\xA1/ {
 688   if (match($0, "^ *([^[ ]* |)<([^[ ]*)\\[[0-9]*\\][(](.*)[)].*, lang=([^ ]*)>", line)) {
 689     next_chunk_name=line[2];
 690     get_texmacs_chunk_args(line[3], next_chunk_params);
 691     gsub(ARG_SEPARATOR ",? ?", ";", line[3]);
 692     params = "params=" line[3];
 693     if ((line[4])) {
 694       params = params ",language=" line[4]
 695     }
 696     get_tex_chunk_args(params, next_chunk_opts);
 697     new_chunk(next_chunk_name, next_chunk_opts, next_chunk_params);
 698     texmacs_chunking = 1;
 699   } else {
 700     # warning(sprintf("Unexpected chunk match: %s\n", $_))
 701   }
 702   next;
 703 }
 704 /^\\Chunk{/ {
 705   if (match($0, "^\\\\Chunk{ *([^ ,}]*),?(.*)}", line)) {
 706     next_chunk_name = line[1];
 707     get_tex_chunk_args(line[2], next_chunk_opts);
 708   }
 709   next;
 710 }
 711 /^\\begin{lstlisting}|^\\begin{Chunk}/ {
 712   if (match($0, "}.*[[,] *name= *{? *([^], }]*)", line)) {
 713     new_chunk(line[1]);
 714   } else {
 715     new_chunk(next_chunk_name, next_chunk_opts);
 716   }
 717   chunking=1;
 718   next;
 719 }
 720 /^ *\|____________*/ && texmacs_chunking {
 721   active_chunk="";
 722   texmacs_chunking=0;
 723   chunking=0;
 724 }
 725 /^ *\|\/\\/ && texmacs_chunking {
 726   texmacs_chunking=0;
 727   chunking=0;
 728   active_chunk="";
 729 }
 730 texmacs_chunk=0;
 731 /^ *[1-9][0-9]* *\| / {
 732   if (texmacs_chunking) {
 733     chunking=1;
 734     texmacs_chunk=1;
 735     gsub("^ *[1-9][0-9]* *\\| ", "")
 736   }
 737 }
 738 /^ *\.\/\\/ && texmacs_chunking {
 739   next;
 740 }
 741 /^ *__*$/ && texmacs_chunking {
 742   next;
 743 }
 744 texmacs_chunking {
 745   if (! texmacs_chunk) {
 746     # must be a texmacs continued line
 747     chunking=1;
 748     texmacs_chunk=1;
 749   }
 750 }
 751 ! texmacs_chunk {
 752 #  texmacs_chunking=0;
 753   chunking=0;
 754 }
 755 /^[<]<.*[>]>=/ {
 756   if (match($0, "^[<]<(.*)[>]>= *$", line)) {
 757     chunking=1;
 758     notangle_mode=1;
 759     new_chunk(line[1]);
 760     next;
 761   }
 762 }
 763 /^\\[e]nd{lstlisting}|^\\[e]nd{Chunk}/ {
 764   chunking=0;
 765   active_chunk="";
 766   next;
 767 }
 768 /^@ *$/ {
 769   chunking=0;
 770   active_chunk="";
 771 }
 772 ! chunking { next; }
 773 length(active_chunk) {
 774   if (length(tabs)) {
 775     gsub("\t", tabs);
 776   }
 777   chunk = $0;
 778   indent = 0;
 779   while(match(chunk,"(\xC2\xAB)([^\xC2\xBB]*) [^\xC2\xBB]*\xC2\xBB", line) ||
 780         match(chunk,
 781               "([=]<\\\\chunkref{([^}>]*)}(\\(.*\\)|)>|<<([a-zA-Z_][-a-zA-Z0-9_]*)>>)",
 782               line)\
 783   ) {
 784     chunklet = substr(chunk, 1, RSTART - 1);
 785     indent += length(chunklet);
 786     chunk_line(active_chunk, chunklet);
 787     chunk = substr(chunk, RSTART + RLENGTH);
 788     if (substr(line[1], 1, 1) == "=") {
 789       # chunk name up to }
 790           # FILTHY HACK
 791           gsub("\\\\#", "#", line[3]);
 792           gsub("\\\\textbackslash{}", "\\", line[3]);
 793           gsub("\\\\\\^", "^", line[3]);
 794       chunk_include(active_chunk, line[2] line[3], indent);
 795     } else if (substr(line[1], 1, 1) == "<") {
 796       chunk_include(active_chunk, line[4], indent);
 797     } else if (line[1] == "\xC2\xAB") {
 798       chunk_include(active_chunk, line[2], indent);
 799     } else {
 800       error("Unknown chunk fragment: " line[1]);
 801     }
 802   }
 803   chunk_line(active_chunk, chunk);
 804   chunk_line(active_chunk, "\n");
 805 }
 806 END {
 807   if (debug) {
 808     print "------ chunk names "
 809     output_chunk_names();
 810     print "====== chunks"
 811     output_chunks();
 812     print "++++++ debug"
 813     for (a in chunks) {
 814       print a "=" chunks[a];
 815     }
 816   }
 817   ORS="";
 818   if (length(root)) output_chunk(root);
 819   else output_chunk_names();
 820 }
 821 function get_texmacs_chunk_args(text, args,   a, done) {
 822   split(text, args, ARG_SEPARATOR);
 823
 824   done=0
 825   for (a=1; (a in args); a++) if (a>1) {
 826     if (args[a] == "" || substr(args[a], 1, 1) == ")") done=1;
 827     if (done) {
 828       delete args[a];
 829       break;
 830     }
 831
 832     if (substr(args[a], 1, 2) == ", ") args[a]=substr(args[a], 3);
 833     else if (substr(args[a], 1, 1) == ",") args[a]=substr(args[a], 2);
 834   }
 835 }
 836 function get_tex_chunk_args(text, values,
 837   # optional parameters
 838   path, # hierarchical precursors
 839   # local vars
 840   a, name)
 841 {
 842   split("", values);
 843   while(length(text)) {
 844     if (match(text, "^ *}(.*)", a)) {
 845       return a[1];
 846     }
 847     if (! match(text, " *([^,=]*[^,= ]) *(([,=]) *(([^,}]*) *,* *(.*))|)$", a)) {
 848       return text;
 849     }
 850     name=a[1];
 851     if (a[3] == "=") {
 852       if (substr(a[4],1,1) == "{") {
 853         text = get_tex_chunk_args(substr(a[4],2), values, path name SUBSEP);
 854       } else {
 855         values[path name]=a[5];
 856         text = a[6];
 857       }
 858     } else {
 859       values[path name]="";
 860       text = a[2];
 861     }
 862   }
 863   return text;
 864 }