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

   1 /* -*-c-*-
   2  *
   3  * $Id: sw_env.c,v 1.2 1999/07/27 13:38:27 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 /*----- Revision history --------------------------------------------------*
  30  *
  31  * $Log: sw_env.c,v $
  32  * Revision 1.2  1999/07/27 13:38:27  mdw
  33  * Cauterized out the low-level environment operations and put them in
  34  * mLib.
  35  *
  36  * Revision 1.1.1.1  1999/06/02  16:53:35  mdw
  37  * Initial import.
  38  *
  39  */
  40
  41 /*----- Header files ------------------------------------------------------*/
  42
  43 #include "config.h"
  44
  45 #include <ctype.h>
  46 #include <errno.h>
  47 #include <stdio.h>
  48 #include <stdlib.h>
  49 #include <string.h>
  50
  51 #include <unistd.h>
  52 #include <sys/wait.h>
  53
  54 #ifndef DECL_ENVIRON
  55   extern char **environ;
  56 #endif
  57
  58 #include <mLib/alloc.h>
  59 #include <mLib/dstr.h>
  60 #include <mLib/env.h>
  61 #include <mLib/report.h>
  62 #include <mLib/sym.h>
  63
  64 #include "sw_env.h"
  65
  66 /*----- Main code ---------------------------------------------------------*/
  67
  68 /* --- @env_error@ --- *
  69  *
  70  * Arguments:   @int e@ = error code
  71  *
  72  * Returns:     String representation of error.
  73  *
  74  * Use:         Transforms an error into something a user can understand.
  75  */
  76
  77 const char *env_error(int e)
  78 {
  79   const char *tab[] = {
  80     "Everything is fine",
  81     "Unexpected end-of-file",
  82     "Bad character in variable name",
  83     "Bad parameter substitution",
  84     "Mismatched quote",
  85     "Missing or spurious `}'",
  86     "<see errno>",
  87     "Internal error"
  88   };
  89   if (e == ENV_SYSTEM)
  90     return (strerror(errno));
  91   else
  92     return (tab[e]);
  93 }
  94
  95 /* --- @peek@ --- *
  96  *
  97  * Arguments:   @FILE *fp@ = stream to read from
  98  *
  99  * Returns:     Next nonwhitespace character.
 100  *
 101  * Use:         Advances the file position past whitespace characters, and
 102  *              returns the following nonwhitespace character.  The character
 103  *              is not `read'.
 104  */
 105
 106 static int peek(FILE *fp)
 107 {
 108   int ch;
 109
 110   do {
 111     ch = getc(fp);
 112     if (ch == EOF)
 113       return (EOF);
 114   } while (isspace((unsigned char)ch));
 115   ungetc(ch, fp);
 116   return (ch);
 117 }
 118
 119 /* --- @env_var@ --- *
 120  *
 121  * Arguments:   @sym_table *t@ = pointer to symbol table
 122  *              @FILE *fp@ = pointer to stream to read from
 123  *              @dstr *d@ = pointer to output variable
 124  *
 125  * Returns:     One of the @ENV_@ constants.
 126  *
 127  * Use:         Scans a variable name from the input stream.
 128  */
 129
 130 int env_var(sym_table *t, FILE *fp, dstr *d)
 131 {
 132   int ch = getc(fp);
 133
 134   if (ch == EOF)
 135     return (ENV_EOF);
 136   if (ch != '_' && !isalpha((unsigned char)ch))
 137     return (ENV_VCHAR);
 138   for (;;) {
 139     DPUTC(d, ch);
 140     ch = getc(fp);
 141     if (ch == EOF || (ch != '_' && !isalnum((unsigned char)ch)))
 142       break;
 143   }
 144   ungetc(ch, fp);
 145   DPUTZ(d);
 146   return (ENV_OK);
 147 }
 148
 149 /* --- @cmdsubst@ --- *
 150  *
 151  * Arguments:   @sym_table *t@ = pointer to symbol table
 152  *              @FILE *fp@ = pointer to stream to read from
 153  *              @dstr *d@ = pointer to output variable
 154  *              @unsigned f@ = interesting flags
 155  *
 156  * Returns:     An @ENV_@ magic code.
 157  *
 158  * Use:         Rips a command line out of the input stream and writes the
 159  *              output of the command to the variable.  The parsing has some
 160  *              bizarre artifacts, but it's fairly serviceable.
 161  */
 162
 163 static int cmdsubst(sym_table *t, FILE *fp, dstr *d, unsigned f)
 164 {
 165   int term = (f & EVF_BACKTICK) ? '`' : ')';
 166   int argc = 1;
 167   size_t l = d->len;
 168   pid_t kid;
 169   int fd[2];
 170   int e;
 171   char **argv;
 172
 173   /* --- Snarfle the arguments --- */
 174
 175   f &= ~EVF_INCSPC;
 176   while (peek(fp) != term) {
 177     if ((e = env_value(t, fp, d, f)) != ENV_OK)
 178       return (e);
 179     DPUTC(d, 0);
 180     argc++;
 181   }
 182   getc(fp);
 183   if (argc == 1) {
 184     d->len = l;
 185     return (ENV_OK);
 186   }
 187
 188   /* --- Make the @argv@ array --- */
 189
 190   {
 191     char *p = d->buf + l;
 192     char *lim = d->buf + d->len;
 193     char **v;
 194
 195     v = argv = xmalloc(argc * sizeof(char *));
 196     while (p < lim) {
 197       *v++ = p;
 198       while (*p) {
 199         p++;
 200         if (p >= lim)
 201           goto done;
 202       }
 203       p++;
 204     }
 205   done:;
 206     *v++ = 0;
 207   }
 208
 209   /* --- Do the fork/exec thing --- */
 210
 211   if (pipe(fd))
 212     goto fail_0;
 213   if ((kid = fork()) < 0)
 214     goto fail_1;
 215
 216   if (kid == 0) {
 217     close(fd[0]);
 218     if (fd[1] != 1) {
 219       dup2(fd[1], 1);
 220       close(fd[1]);
 221     }
 222     environ = env_export(t);
 223     execvp(argv[0], argv);
 224     _exit(127);
 225   }
 226
 227   d->len = l;
 228   close(fd[1]);
 229   for (;;) {
 230     char buf[4096];
 231     ssize_t n = read(fd[0], buf, sizeof(buf));
 232     if (n <= 0)
 233       break;
 234     DPUTM(d, buf, n);
 235   }
 236   close(fd[0]);
 237   waitpid(kid, 0, 0);
 238   l = d->len;
 239   while (l > 0 && d->buf[l - 1] == '\n')
 240     l--;
 241   d->len = l;
 242   free(argv);
 243   return (ENV_OK);
 244
 245 fail_1:
 246   close(fd[0]);
 247   close(fd[1]);
 248 fail_0:
 249   free(argv);
 250   return (ENV_SYSTEM);
 251 }
 252
 253 /* --- @env_value@ --- *
 254  *
 255  * Arguments:   @sym_table *t@ = pointer to symbol table
 256  *              @FILE *fp@ = pointer to stream to read from
 257  *              @dstr *d@ = pointer to output variable
 258  *              @unsigned f@ = various interesting flags
 259  *
 260  * Returns:     0 if OK, @EOF@ if end-of-file encountered, or >0 on error.
 261  *
 262  * Use:         Scans a value from the input stream.  The value read may be
 263  *              quoted in a Bourne-shell sort of a way, and contain Bourne-
 264  *              shell-like parameter substitutions.  Some substitutions
 265  *              aren't available because they're too awkward to implement.
 266  */
 267
 268 int env_value(sym_table *t, FILE *fp, dstr *d, unsigned f)
 269 {
 270   enum { Q_NONE, Q_SINGLE, Q_DOUBLE, Q_BACK } qt = Q_NONE;
 271   int ch;
 272
 273   do {
 274     ch = getc(fp);
 275     if (ch == EOF)
 276       return (ENV_EOF);
 277   } while ((f & EVF_INITSPC) && isspace((unsigned char)ch));
 278
 279   for (;; ch = getc(fp)) {
 280
 281     /* --- Sort out the current character --- */
 282
 283     if (ch == EOF) break;
 284
 285     /* --- A backslash escapes the next character --- */
 286
 287     if (ch == '\\') {
 288       if ((ch = getc(fp)) == EOF) break;
 289       else if (ch != '\n')
 290         DPUTC(d, ch);
 291       continue;
 292     }
 293
 294     /* --- A single quote starts single-quoting, unless quoted --- *
 295      *
 296      * Do the single-quoted snarf here rather than fiddling with anything
 297      * else.
 298      */
 299
 300     if (ch == '\'' && qt == Q_NONE) {
 301       qt = Q_SINGLE;
 302       for (;;) {
 303         if ((ch = getc(fp)) == EOF) goto done;
 304         if (ch == '\'') break;
 305         DPUTC(d, ch);
 306       }
 307       qt = Q_NONE;
 308       continue;
 309     }
 310
 311     /* --- A backtick does the obvious thing --- */
 312
 313     if (ch == '`' && !(f & EVF_BACKTICK)) {
 314       int e;
 315       if ((e = cmdsubst(t, fp, d, f | EVF_BACKTICK)) != ENV_OK)
 316         return (e);
 317       continue;
 318     }
 319
 320     /* --- Handle double-quoted text --- */
 321
 322     if (ch == '\"') {
 323       if (qt == Q_DOUBLE)
 324         qt = Q_NONE;
 325       else if (qt == Q_NONE)
 326         qt = Q_DOUBLE;
 327       else
 328         return (ENV_INTERNAL);
 329       continue;
 330     }
 331
 332     /* --- Handle variable references and similar magic --- */
 333
 334     if (ch == '$') {
 335       size_t l = d->len;
 336       int e;
 337       char *v;
 338       char *vn;
 339
 340       /* --- Read one character ahead --- */
 341
 342       if ((ch = getc(fp)) == EOF) goto done;
 343
 344       /* --- An alphabetic means this is a direct reference --- */
 345
 346       if (ch == '_' || isalpha(ch)) {
 347         ungetc(ch, fp);
 348         if ((e = env_var(t, fp, d)) != ENV_OK) return (e);
 349         d->len = l;
 350         if ((v = env_get(t, d->buf + l)) != 0)
 351           DPUTS(d, v);
 352       }
 353
 354       /* --- A brace means this is a more complex substitution --- */
 355
 356       else if (ch == '{') {
 357         if ((e = env_var(t, fp, d)) != ENV_OK)
 358           return (e);
 359         d->len = l;
 360         v = env_get(t, d->buf + l);
 361
 362       again:
 363         ch = getc(fp);
 364         switch (ch) {
 365
 366           case EOF:
 367             goto done;
 368
 369           case '}':
 370             if (v)
 371               DPUTS(d, v);
 372             ungetc(ch, fp);
 373             break;
 374
 375           case ':':
 376             if (v && !*v)
 377               v = 0;
 378             goto again;                 /* `::'  and `:}' should be errors */
 379
 380           case '+':
 381             if (v)
 382               v = 0;
 383             else
 384               v = "";
 385             /* Drop through hackily */
 386
 387           case '-':
 388             if (v) {
 389               DPUTS(d, v);
 390               l = d->len;
 391             }
 392             if ((e = env_value(t, fp, d, EVF_INCSPC)) != ENV_OK)
 393               return (e);
 394             if (v)
 395               d->len = l;
 396             break;
 397
 398           case '=':
 399             if (v) {
 400               DPUTS(d, v);
 401               l = d->len;
 402               if ((e = env_value(t, fp, d, EVF_INCSPC)) != ENV_OK)
 403                 return (e);
 404               d->len = l;
 405             } else {
 406               vn = xstrdup(d->buf + l);
 407               if ((e = env_value(t, fp, d, EVF_INCSPC)) != ENV_OK)
 408                 return (e);
 409               if (!(f & EVF_SKIP))
 410                 env_put(t, vn, d->buf + l);
 411               free(vn);
 412             }
 413             break;
 414
 415           default:
 416             return (ENV_SUBST);
 417         }
 418         if (getc(fp) != '}')
 419           return (ENV_BRACE);
 420       }
 421
 422       /* --- Handle `$(...)'-style command substitution --- */
 423
 424       else if (ch == '(') {
 425         if ((e = cmdsubst(t, fp, d, f & ~EVF_BACKTICK)) != ENV_OK)
 426           return (e);
 427       }
 428
 429       /* --- No other `$...' tricks implemented yet --- *
 430        *
 431        * Other ideas: `$((...))' arithmetic.
 432        */
 433
 434       else
 435         return (ENV_SUBST);
 436       continue;
 437     }
 438
 439     /* --- At this point, anything else double-quoted is munched --- */
 440
 441     if (qt == Q_DOUBLE) {
 442       DPUTC(d, ch);
 443       continue;
 444     }
 445
 446     /* --- Some characters just aren't allowed unquoted --- *
 447      *
 448      * They're not an error; they just mean I should stop parsing.  They're
 449      * probably interesting to the next level up.
 450      */
 451
 452     switch (ch) {
 453       case '(': case ')':
 454       case '{': case '}':
 455       case ';':
 456         ungetc(ch, fp);
 457         goto done;
 458       case '`':
 459         if (f & EVF_BACKTICK) {
 460           ungetc(ch, fp);
 461           goto done;
 462         }
 463         break;
 464     }
 465
 466     /* --- Whitespace characters --- *
 467      *
 468      * I might snarf them myself anyway, according to flags.  Or I might
 469      * stop, and skip any following whitespace
 470      */
 471
 472     if (isspace((unsigned char)ch) && (f & EVF_INCSPC) == 0) {
 473       do
 474         ch = getc(fp);
 475       while (ch != EOF && isspace((unsigned char)ch));
 476       ungetc(ch, fp);
 477       break;
 478     }
 479
 480     /* --- Get a new character and go around again --- */
 481
 482     DPUTC(d, ch);
 483   }
 484
 485   /* --- Tidying --- */
 486
 487 done:
 488   DPUTZ(d);
 489   return (qt == Q_NONE ? ENV_OK : ENV_QUOTE);
 490 }
 491
 492 /* --- @env_read@ --- *
 493  *
 494  * Arguments:   @sym_table *t@ = pointer to symbol table
 495  *              @FILE *fp@ = file handle to read
 496  *              @unsigned f@ = various flags
 497  *
 498  * Returns:     Zero if OK, @EOF@ for end-of-file, or error code.
 499  *
 500  * Use:         Reads the environment assignment statements in the file.
 501  */
 502
 503 int env_read(sym_table *t, FILE *fp, unsigned f)
 504 {
 505   dstr n = DSTR_INIT, v = DSTR_INIT;
 506   int e = ENV_OK, ch;
 507
 508   for (;;) {
 509     ch = peek(fp);
 510
 511     if (ch == ':') {
 512       getc(fp);
 513       peek(fp);
 514       if ((e = env_value(t, fp, &v, f)) != ENV_OK)
 515         goto done;
 516     }
 517
 518     else if (ch == '#')
 519       do ch = getc(fp); while (ch != EOF && ch != '\n');
 520
 521     else if (peek(fp) == '}' ||
 522         (e = env_var(t, fp, &n)) != ENV_OK)
 523       goto done;
 524
 525     else if (strcmp(n.buf, "include") == 0) {
 526       peek(fp);
 527       if ((e = env_value(t, fp, &v, f)) != ENV_OK)
 528         goto done;
 529       if (!(f & EVF_SKIP))
 530         env_file(t, v.buf);
 531     }
 532
 533     else if (strcmp(n.buf, "arch") == 0) {
 534       peek(fp);
 535       if ((e = env_value(t, fp, &v, f)) != ENV_OK)
 536         goto done;
 537       if (peek(fp) != '{') {
 538         e = ENV_BRACE;
 539         goto done;
 540       }
 541       getc(fp);
 542       e = env_read(t, fp, strcmp(v.buf, ARCH) ? f | EVF_SKIP : f);
 543       if (e != ENV_OK)
 544         goto done;
 545       if (getc(fp) != '}') {
 546         e = ENV_BRACE;
 547         goto done;
 548       }
 549     }
 550
 551     else if (strcmp(n.buf, "unset") == 0) {
 552       peek(fp);
 553       if ((e = env_var(t, fp, &v)) != ENV_OK)
 554         goto done;
 555       env_put(t, v.buf, 0);
 556     }
 557
 558     else {
 559       if (strcmp(n.buf, "set") == 0) {
 560         DRESET(&n);
 561         peek(fp);
 562         if ((e = env_var(t, fp, &n)) != ENV_OK)
 563           goto done;
 564       }
 565       if (peek(fp) == '=') {
 566         getc(fp);
 567         peek(fp);
 568       }
 569       if ((e = env_value(t, fp, &v, f)) != ENV_OK)
 570         goto done;
 571
 572       if (!(f & EVF_SKIP))
 573         env_put(t, n.buf, v.buf);
 574     }
 575
 576     if (peek(fp) == ';')
 577       getc(fp);
 578     DRESET(&n);
 579     DRESET(&v);
 580   }
 581
 582 done:
 583   dstr_destroy(&n);
 584   dstr_destroy(&v);
 585   return (e);
 586 }
 587
 588 /* --- @env_file@ --- *
 589  *
 590  * Arguments:   @sym_table *t@ = pointer to symbol table
 591  *              @const char *name@ = pointer to filename
 592  *
 593  * Returns:     Zero if OK, or an error code.
 594  *
 595  * Use:         Reads a named file of environment assignments.
 596  */
 597
 598 int env_file(sym_table *t, const char *name)
 599 {
 600   FILE *fp;
 601   int e;
 602
 603   if ((fp = fopen(name, "r")) == 0)
 604     return (ENV_SYSTEM);
 605   e = env_read(t, fp, 0);
 606   fclose(fp);
 607   if (e == ENV_EOF)
 608     e = ENV_OK;
 609   else if (e == ENV_OK)
 610     e = ENV_BRACE;
 611   return (e);
 612 }
 613
 614 /*----- That's all, folks -------------------------------------------------*/