mdw@git.distorted.org.uk Git - sw-tools/blob - src/sw_env.c

   1 /* -*-c-*-
   2  *
   3  * $Id: sw_env.c,v 1.3 2004/04/08 01:52:19 mdw Exp $
   4  *
   5  * Mangling of environment variables
   6  *
   7  * (c) 1999 EBI
   8  */
   9
  10 /*----- Licensing notice --------------------------------------------------*
  11  *
  12  * This file is part of sw-tools.
  13  *
  14  * sw-tools is free software; you can redistribute it and/or modify
  15  * it under the terms of the GNU General Public License as published by
  16  * the Free Software Foundation; either version 2 of the License, or
  17  * (at your option) any later version.
  18  *
  19  * sw-tools is distributed in the hope that it will be useful,
  20  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  21  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  22  * GNU General Public License for more details.
  23  *
  24  * You should have received a copy of the GNU General Public License
  25  * along with sw-tools; if not, write to the Free Software Foundation,
  26  * Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
  27  */
  28
  29 /*----- Header files ------------------------------------------------------*/
  30
  31 #include "config.h"
  32
  33 #include <ctype.h>
  34 #include <errno.h>
  35 #include <stdio.h>
  36 #include <stdlib.h>
  37 #include <string.h>
  38
  39 #include <unistd.h>
  40 #include <sys/wait.h>
  41
  42 #ifndef DECL_ENVIRON
  43   extern char **environ;
  44 #endif
  45
  46 #include <mLib/alloc.h>
  47 #include <mLib/dstr.h>
  48 #include <mLib/env.h>
  49 #include <mLib/report.h>
  50 #include <mLib/sym.h>
  51
  52 #include "sw_env.h"
  53
  54 /*----- Main code ---------------------------------------------------------*/
  55
  56 /* --- @env_error@ --- *
  57  *
  58  * Arguments:   @int e@ = error code
  59  *
  60  * Returns:     String representation of error.
  61  *
  62  * Use:         Transforms an error into something a user can understand.
  63  */
  64
  65 const char *env_error(int e)
  66 {
  67   const char *tab[] = {
  68     "Everything is fine",
  69     "Unexpected end-of-file",
  70     "Bad character in variable name",
  71     "Bad parameter substitution",
  72     "Mismatched quote",
  73     "Missing or spurious `}'",
  74     "<see errno>",
  75     "Internal error"
  76   };
  77   if (e == ENV_SYSTEM)
  78     return (strerror(errno));
  79   else
  80     return (tab[e]);
  81 }
  82
  83 /* --- @peek@ --- *
  84  *
  85  * Arguments:   @FILE *fp@ = stream to read from
  86  *
  87  * Returns:     Next nonwhitespace character.
  88  *
  89  * Use:         Advances the file position past whitespace characters, and
  90  *              returns the following nonwhitespace character.  The character
  91  *              is not `read'.
  92  */
  93
  94 static int peek(FILE *fp)
  95 {
  96   int ch;
  97
  98   do {
  99     ch = getc(fp);
 100     if (ch == EOF)
 101       return (EOF);
 102   } while (isspace((unsigned char)ch));
 103   ungetc(ch, fp);
 104   return (ch);
 105 }
 106
 107 /* --- @env_var@ --- *
 108  *
 109  * Arguments:   @sym_table *t@ = pointer to symbol table
 110  *              @FILE *fp@ = pointer to stream to read from
 111  *              @dstr *d@ = pointer to output variable
 112  *
 113  * Returns:     One of the @ENV_@ constants.
 114  *
 115  * Use:         Scans a variable name from the input stream.
 116  */
 117
 118 int env_var(sym_table *t, FILE *fp, dstr *d)
 119 {
 120   int ch = getc(fp);
 121
 122   if (ch == EOF)
 123     return (ENV_EOF);
 124   if (ch != '_' && !isalpha((unsigned char)ch))
 125     return (ENV_VCHAR);
 126   for (;;) {
 127     DPUTC(d, ch);
 128     ch = getc(fp);
 129     if (ch == EOF || (ch != '_' && !isalnum((unsigned char)ch)))
 130       break;
 131   }
 132   ungetc(ch, fp);
 133   DPUTZ(d);
 134   return (ENV_OK);
 135 }
 136
 137 /* --- @cmdsubst@ --- *
 138  *
 139  * Arguments:   @sym_table *t@ = pointer to symbol table
 140  *              @FILE *fp@ = pointer to stream to read from
 141  *              @dstr *d@ = pointer to output variable
 142  *              @unsigned f@ = interesting flags
 143  *
 144  * Returns:     An @ENV_@ magic code.
 145  *
 146  * Use:         Rips a command line out of the input stream and writes the
 147  *              output of the command to the variable.  The parsing has some
 148  *              bizarre artifacts, but it's fairly serviceable.
 149  */
 150
 151 static int cmdsubst(sym_table *t, FILE *fp, dstr *d, unsigned f)
 152 {
 153   int term = (f & EVF_BACKTICK) ? '`' : ')';
 154   int argc = 1;
 155   size_t l = d->len;
 156   pid_t kid;
 157   int fd[2];
 158   int e;
 159   char **argv;
 160
 161   /* --- Snarfle the arguments --- */
 162
 163   f &= ~EVF_INCSPC;
 164   while (peek(fp) != term) {
 165     if ((e = env_value(t, fp, d, f)) != ENV_OK)
 166       return (e);
 167     DPUTC(d, 0);
 168     argc++;
 169   }
 170   getc(fp);
 171   if (argc == 1) {
 172     d->len = l;
 173     return (ENV_OK);
 174   }
 175
 176   /* --- Make the @argv@ array --- */
 177
 178   {
 179     char *p = d->buf + l;
 180     char *lim = d->buf + d->len;
 181     char **v;
 182
 183     v = argv = xmalloc(argc * sizeof(char *));
 184     while (p < lim) {
 185       *v++ = p;
 186       while (*p) {
 187         p++;
 188         if (p >= lim)
 189           goto done;
 190       }
 191       p++;
 192     }
 193   done:;
 194     *v++ = 0;
 195   }
 196
 197   /* --- Do the fork/exec thing --- */
 198
 199   if (pipe(fd))
 200     goto fail_0;
 201   if ((kid = fork()) < 0)
 202     goto fail_1;
 203
 204   if (kid == 0) {
 205     close(fd[0]);
 206     if (fd[1] != 1) {
 207       dup2(fd[1], 1);
 208       close(fd[1]);
 209     }
 210     environ = env_export(t);
 211     execvp(argv[0], argv);
 212     _exit(127);
 213   }
 214
 215   d->len = l;
 216   close(fd[1]);
 217   for (;;) {
 218     char buf[4096];
 219     ssize_t n = read(fd[0], buf, sizeof(buf));
 220     if (n <= 0)
 221       break;
 222     DPUTM(d, buf, n);
 223   }
 224   close(fd[0]);
 225   waitpid(kid, 0, 0);
 226   l = d->len;
 227   while (l > 0 && d->buf[l - 1] == '\n')
 228     l--;
 229   d->len = l;
 230   free(argv);
 231   return (ENV_OK);
 232
 233 fail_1:
 234   close(fd[0]);
 235   close(fd[1]);
 236 fail_0:
 237   free(argv);
 238   return (ENV_SYSTEM);
 239 }
 240
 241 /* --- @env_value@ --- *
 242  *
 243  * Arguments:   @sym_table *t@ = pointer to symbol table
 244  *              @FILE *fp@ = pointer to stream to read from
 245  *              @dstr *d@ = pointer to output variable
 246  *              @unsigned f@ = various interesting flags
 247  *
 248  * Returns:     0 if OK, @EOF@ if end-of-file encountered, or >0 on error.
 249  *
 250  * Use:         Scans a value from the input stream.  The value read may be
 251  *              quoted in a Bourne-shell sort of a way, and contain Bourne-
 252  *              shell-like parameter substitutions.  Some substitutions
 253  *              aren't available because they're too awkward to implement.
 254  */
 255
 256 int env_value(sym_table *t, FILE *fp, dstr *d, unsigned f)
 257 {
 258   enum { Q_NONE, Q_SINGLE, Q_DOUBLE, Q_BACK } qt = Q_NONE;
 259   int ch;
 260
 261   do {
 262     ch = getc(fp);
 263     if (ch == EOF)
 264       return (ENV_EOF);
 265   } while ((f & EVF_INITSPC) && isspace((unsigned char)ch));
 266
 267   for (;; ch = getc(fp)) {
 268
 269     /* --- Sort out the current character --- */
 270
 271     if (ch == EOF) break;
 272
 273     /* --- A backslash escapes the next character --- */
 274
 275     if (ch == '\\') {
 276       if ((ch = getc(fp)) == EOF) break;
 277       else if (ch != '\n')
 278         DPUTC(d, ch);
 279       continue;
 280     }
 281
 282     /* --- A single quote starts single-quoting, unless quoted --- *
 283      *
 284      * Do the single-quoted snarf here rather than fiddling with anything
 285      * else.
 286      */
 287
 288     if (ch == '\'' && qt == Q_NONE) {
 289       qt = Q_SINGLE;
 290       for (;;) {
 291         if ((ch = getc(fp)) == EOF) goto done;
 292         if (ch == '\'') break;
 293         DPUTC(d, ch);
 294       }
 295       qt = Q_NONE;
 296       continue;
 297     }
 298
 299     /* --- A backtick does the obvious thing --- */
 300
 301     if (ch == '`' && !(f & EVF_BACKTICK)) {
 302       int e;
 303       if ((e = cmdsubst(t, fp, d, f | EVF_BACKTICK)) != ENV_OK)
 304         return (e);
 305       continue;
 306     }
 307
 308     /* --- Handle double-quoted text --- */
 309
 310     if (ch == '\"') {
 311       if (qt == Q_DOUBLE)
 312         qt = Q_NONE;
 313       else if (qt == Q_NONE)
 314         qt = Q_DOUBLE;
 315       else
 316         return (ENV_INTERNAL);
 317       continue;
 318     }
 319
 320     /* --- Handle variable references and similar magic --- */
 321
 322     if (ch == '$') {
 323       size_t l = d->len;
 324       int e;
 325       char *v;
 326       char *vn;
 327
 328       /* --- Read one character ahead --- */
 329
 330       if ((ch = getc(fp)) == EOF) goto done;
 331
 332       /* --- An alphabetic means this is a direct reference --- */
 333
 334       if (ch == '_' || isalpha(ch)) {
 335         ungetc(ch, fp);
 336         if ((e = env_var(t, fp, d)) != ENV_OK) return (e);
 337         d->len = l;
 338         if ((v = env_get(t, d->buf + l)) != 0)
 339           DPUTS(d, v);
 340       }
 341
 342       /* --- A brace means this is a more complex substitution --- */
 343
 344       else if (ch == '{') {
 345         if ((e = env_var(t, fp, d)) != ENV_OK)
 346           return (e);
 347         d->len = l;
 348         v = env_get(t, d->buf + l);
 349
 350       again:
 351         ch = getc(fp);
 352         switch (ch) {
 353
 354           case EOF:
 355             goto done;
 356
 357           case '}':
 358             if (v)
 359               DPUTS(d, v);
 360             ungetc(ch, fp);
 361             break;
 362
 363           case ':':
 364             if (v && !*v)
 365               v = 0;
 366             goto again;                 /* `::'  and `:}' should be errors */
 367
 368           case '+':
 369             if (v)
 370               v = 0;
 371             else
 372               v = "";
 373             /* Drop through hackily */
 374
 375           case '-':
 376             if (v) {
 377               DPUTS(d, v);
 378               l = d->len;
 379             }
 380             if ((e = env_value(t, fp, d, EVF_INCSPC)) != ENV_OK)
 381               return (e);
 382             if (v)
 383               d->len = l;
 384             break;
 385
 386           case '=':
 387             if (v) {
 388               DPUTS(d, v);
 389               l = d->len;
 390               if ((e = env_value(t, fp, d, EVF_INCSPC)) != ENV_OK)
 391                 return (e);
 392               d->len = l;
 393             } else {
 394               vn = xstrdup(d->buf + l);
 395               if ((e = env_value(t, fp, d, EVF_INCSPC)) != ENV_OK)
 396                 return (e);
 397               if (!(f & EVF_SKIP))
 398                 env_put(t, vn, d->buf + l);
 399               free(vn);
 400             }
 401             break;
 402
 403           default:
 404             return (ENV_SUBST);
 405         }
 406         if (getc(fp) != '}')
 407           return (ENV_BRACE);
 408       }
 409
 410       /* --- Handle `$(...)'-style command substitution --- */
 411
 412       else if (ch == '(') {
 413         if ((e = cmdsubst(t, fp, d, f & ~EVF_BACKTICK)) != ENV_OK)
 414           return (e);
 415       }
 416
 417       /* --- No other `$...' tricks implemented yet --- *
 418        *
 419        * Other ideas: `$((...))' arithmetic.
 420        */
 421
 422       else
 423         return (ENV_SUBST);
 424       continue;
 425     }
 426
 427     /* --- At this point, anything else double-quoted is munched --- */
 428
 429     if (qt == Q_DOUBLE) {
 430       DPUTC(d, ch);
 431       continue;
 432     }
 433
 434     /* --- Some characters just aren't allowed unquoted --- *
 435      *
 436      * They're not an error; they just mean I should stop parsing.  They're
 437      * probably interesting to the next level up.
 438      */
 439
 440     switch (ch) {
 441       case '(': case ')':
 442       case '{': case '}':
 443       case ';':
 444         ungetc(ch, fp);
 445         goto done;
 446       case '`':
 447         if (f & EVF_BACKTICK) {
 448           ungetc(ch, fp);
 449           goto done;
 450         }
 451         break;
 452     }
 453
 454     /* --- Whitespace characters --- *
 455      *
 456      * I might snarf them myself anyway, according to flags.  Or I might
 457      * stop, and skip any following whitespace
 458      */
 459
 460     if (isspace((unsigned char)ch) && (f & EVF_INCSPC) == 0) {
 461       do
 462         ch = getc(fp);
 463       while (ch != EOF && isspace((unsigned char)ch));
 464       ungetc(ch, fp);
 465       break;
 466     }
 467
 468     /* --- Get a new character and go around again --- */
 469
 470     DPUTC(d, ch);
 471   }
 472
 473   /* --- Tidying --- */
 474
 475 done:
 476   DPUTZ(d);
 477   return (qt == Q_NONE ? ENV_OK : ENV_QUOTE);
 478 }
 479
 480 /* --- @env_read@ --- *
 481  *
 482  * Arguments:   @sym_table *t@ = pointer to symbol table
 483  *              @FILE *fp@ = file handle to read
 484  *              @unsigned f@ = various flags
 485  *
 486  * Returns:     Zero if OK, @EOF@ for end-of-file, or error code.
 487  *
 488  * Use:         Reads the environment assignment statements in the file.
 489  */
 490
 491 int env_read(sym_table *t, FILE *fp, unsigned f)
 492 {
 493   dstr n = DSTR_INIT, v = DSTR_INIT;
 494   int e = ENV_OK, ch;
 495
 496   for (;;) {
 497     ch = peek(fp);
 498
 499     if (ch == ':') {
 500       getc(fp);
 501       peek(fp);
 502       if ((e = env_value(t, fp, &v, f)) != ENV_OK)
 503         goto done;
 504     }
 505
 506     else if (ch == '#')
 507       do ch = getc(fp); while (ch != EOF && ch != '\n');
 508
 509     else if (peek(fp) == '}' ||
 510         (e = env_var(t, fp, &n)) != ENV_OK)
 511       goto done;
 512
 513     else if (strcmp(n.buf, "include") == 0) {
 514       peek(fp);
 515       if ((e = env_value(t, fp, &v, f)) != ENV_OK)
 516         goto done;
 517       if (!(f & EVF_SKIP))
 518         env_file(t, v.buf);
 519     }
 520
 521     else if (strcmp(n.buf, "arch") == 0) {
 522       peek(fp);
 523       if ((e = env_value(t, fp, &v, f)) != ENV_OK)
 524         goto done;
 525       if (peek(fp) != '{') {
 526         e = ENV_BRACE;
 527         goto done;
 528       }
 529       getc(fp);
 530       e = env_read(t, fp, strcmp(v.buf, ARCH) ? f | EVF_SKIP : f);
 531       if (e != ENV_OK)
 532         goto done;
 533       if (getc(fp) != '}') {
 534         e = ENV_BRACE;
 535         goto done;
 536       }
 537     }
 538
 539     else if (strcmp(n.buf, "unset") == 0) {
 540       peek(fp);
 541       if ((e = env_var(t, fp, &v)) != ENV_OK)
 542         goto done;
 543       env_put(t, v.buf, 0);
 544     }
 545
 546     else {
 547       if (strcmp(n.buf, "set") == 0) {
 548         DRESET(&n);
 549         peek(fp);
 550         if ((e = env_var(t, fp, &n)) != ENV_OK)
 551           goto done;
 552       }
 553       if (peek(fp) == '=') {
 554         getc(fp);
 555         peek(fp);
 556       }
 557       if ((e = env_value(t, fp, &v, f)) != ENV_OK)
 558         goto done;
 559
 560       if (!(f & EVF_SKIP))
 561         env_put(t, n.buf, v.buf);
 562     }
 563
 564     if (peek(fp) == ';')
 565       getc(fp);
 566     DRESET(&n);
 567     DRESET(&v);
 568   }
 569
 570 done:
 571   dstr_destroy(&n);
 572   dstr_destroy(&v);
 573   return (e);
 574 }
 575
 576 /* --- @env_file@ --- *
 577  *
 578  * Arguments:   @sym_table *t@ = pointer to symbol table
 579  *              @const char *name@ = pointer to filename
 580  *
 581  * Returns:     Zero if OK, or an error code.
 582  *
 583  * Use:         Reads a named file of environment assignments.
 584  */
 585
 586 int env_file(sym_table *t, const char *name)
 587 {
 588   FILE *fp;
 589   int e;
 590
 591   if ((fp = fopen(name, "r")) == 0)
 592     return (ENV_SYSTEM);
 593   e = env_read(t, fp, 0);
 594   fclose(fp);
 595   if (e == ENV_EOF)
 596     e = ENV_OK;
 597   else if (e == ENV_OK)
 598     e = ENV_BRACE;
 599   return (e);
 600 }
 601
 602 /*----- That's all, folks -------------------------------------------------*/