dbus-shell.c

00001 /* -*- mode: C; c-file-style: "gnu" -*- */
00002 /* dbus-shell.c Shell command line utility functions.
00003  *
00004  * Copyright (C) 2002, 2003  Red Hat, Inc.
00005  * Copyright (C) 2003 CodeFactory AB
00006  *
00007  * Licensed under the Academic Free License version 2.1
00008  *
00009  * This program is free software; you can redistribute it and/or modify
00010  * it under the terms of the GNU General Public License as published by
00011  * the Free Software Foundation; either version 2 of the License, or
00012  * (at your option) any later version.
00013  *
00014  * This program is distributed in the hope that it will be useful,
00015  * but WITHOUT ANY WARRANTY; without even the implied warranty of
00016  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
00017  * GNU General Public License for more details.
00018  *
00019  * You should have received a copy of the GNU General Public License
00020  * along with this program; if not, write to the Free Software
00021  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
00022  *
00023  */
00024 
00025 #include <string.h>
00026 #include "dbus-internals.h"
00027 #include "dbus-list.h"
00028 #include "dbus-memory.h"
00029 #include "dbus-protocol.h"
00030 #include "dbus-shell.h"
00031 #include "dbus-string.h"
00032 
00033 /* Single quotes preserve the literal string exactly. escape
00034  * sequences are not allowed; not even \' - if you want a '
00035  * in the quoted text, you have to do something like 'foo'\''bar'
00036  *
00037  * Double quotes allow $ ` " \ and newline to be escaped with backslash.
00038  * Otherwise double quotes preserve things literally.
00039  */
00040 
00041 static dbus_bool_t
00042 unquote_string_inplace (char* str, char** end)
00043 {
00044   char* dest;
00045   char* s;
00046   char quote_char;
00047   
00048   dest = s = str;
00049 
00050   quote_char = *s;
00051   
00052   if (!(*s == '"' || *s == '\''))
00053     {
00054       *end = str;
00055       return FALSE;
00056     }
00057 
00058   /* Skip the initial quote mark */
00059   ++s;
00060 
00061   if (quote_char == '"')
00062     {
00063       while (*s)
00064         {
00065           _dbus_assert(s > dest); /* loop invariant */
00066       
00067           switch (*s)
00068             {
00069             case '"':
00070               /* End of the string, return now */
00071               *dest = '\0';
00072               ++s;
00073               *end = s;
00074               return TRUE;
00075 
00076             case '\\':
00077               /* Possible escaped quote or \ */
00078               ++s;
00079               switch (*s)
00080                 {
00081                 case '"':
00082                 case '\\':
00083                 case '`':
00084                 case '$':
00085                 case '\n':
00086                   *dest = *s;
00087                   ++s;
00088                   ++dest;
00089                   break;
00090 
00091                 default:
00092                   /* not an escaped char */
00093                   *dest = '\\';
00094                   ++dest;
00095                   /* ++s already done. */
00096                   break;
00097                 }
00098               break;
00099 
00100             default:
00101               *dest = *s;
00102               ++dest;
00103               ++s;
00104               break;
00105             }
00106 
00107           _dbus_assert(s > dest); /* loop invariant */
00108         }
00109     }
00110   else
00111     {
00112       while (*s)
00113         {
00114           _dbus_assert(s > dest); /* loop invariant */
00115           
00116           if (*s == '\'')
00117             {
00118               /* End of the string, return now */
00119               *dest = '\0';
00120               ++s;
00121               *end = s;
00122               return TRUE;
00123             }
00124           else
00125             {
00126               *dest = *s;
00127               ++dest;
00128               ++s;
00129             }
00130 
00131           _dbus_assert(s > dest); /* loop invariant */
00132         }
00133     }
00134   
00135   /* If we reach here this means the close quote was never encountered */
00136 
00137   *dest = '\0';
00138   
00139   *end = s;
00140   return FALSE;
00141 }
00142 
00167 char*
00168 _dbus_shell_unquote (const char *quoted_string)
00169 {
00170   char *unquoted;
00171   char *end;
00172   char *start;
00173   char *ret;
00174   DBusString retval;
00175 
00176   unquoted = _dbus_strdup (quoted_string);
00177   if (unquoted == NULL)
00178     return NULL;
00179 
00180   start = unquoted;
00181   end = unquoted;
00182   if (!_dbus_string_init (&retval))
00183     {
00184       dbus_free (unquoted);
00185       return NULL;
00186     }
00187 
00188   /* The loop allows cases such as
00189    * "foo"blah blah'bar'woo foo"baz"la la la\'\''foo'
00190    */
00191   while (*start)
00192     {
00193       /* Append all non-quoted chars, honoring backslash escape
00194        */
00195       
00196       while (*start && !(*start == '"' || *start == '\''))
00197         {
00198           if (*start == '\\')
00199             {
00200               /* all characters can get escaped by backslash,
00201                * except newline, which is removed if it follows
00202                * a backslash outside of quotes
00203                */
00204               
00205               ++start;
00206               if (*start)
00207                 {
00208                   if (*start != '\n')
00209                     {
00210                       if (!_dbus_string_append_byte (&retval, *start))
00211                         goto error;
00212                     }
00213                   ++start;
00214                 }
00215             }
00216           else
00217             {
00218               if (!_dbus_string_append_byte (&retval, *start))
00219                 goto error;
00220               ++start;
00221             }
00222         }
00223 
00224       if (*start)
00225         {
00226           if (!unquote_string_inplace (start, &end))
00227             goto error;
00228           else
00229             {
00230               if (!_dbus_string_append (&retval, start))
00231                 goto error;
00232               start = end;
00233             }
00234         }
00235     }
00236 
00237   ret = _dbus_strdup (_dbus_string_get_data (&retval));
00238   if (!ret)
00239     goto error;
00240 
00241   dbus_free (unquoted);
00242   _dbus_string_free (&retval);
00243   
00244   return ret;
00245   
00246  error:
00247   dbus_free (unquoted);
00248   _dbus_string_free (&retval);
00249   return NULL;
00250 }
00251 
00252 /* _dbus_shell_parse_argv() does a semi-arbitrary weird subset of the way
00253  * the shell parses a command line. We don't do variable expansion,
00254  * don't understand that operators are tokens, don't do tilde expansion,
00255  * don't do command substitution, no arithmetic expansion, IFS gets ignored,
00256  * don't do filename globs, don't remove redirection stuff, etc.
00257  *
00258  * READ THE UNIX98 SPEC on "Shell Command Language" before changing
00259  * the behavior of this code.
00260  *
00261  * Steps to parsing the argv string:
00262  *
00263  *  - tokenize the string (but since we ignore operators,
00264  *    our tokenization may diverge from what the shell would do)
00265  *    note that tokenization ignores the internals of a quoted
00266  *    word and it always splits on spaces, not on IFS even
00267  *    if we used IFS. We also ignore "end of input indicator"
00268  *    (I guess this is control-D?)
00269  *
00270  *    Tokenization steps, from UNIX98 with operator stuff removed,
00271  *    are:
00272  * 
00273  *    1) "If the current character is backslash, single-quote or
00274  *        double-quote (\, ' or ") and it is not quoted, it will affect
00275  *        quoting for subsequent characters up to the end of the quoted
00276  *        text. The rules for quoting are as described in Quoting
00277  *        . During token recognition no substitutions will be actually
00278  *        performed, and the result token will contain exactly the
00279  *        characters that appear in the input (except for newline
00280  *        character joining), unmodified, including any embedded or
00281  *        enclosing quotes or substitution operators, between the quote
00282  *        mark and the end of the quoted text. The token will not be
00283  *        delimited by the end of the quoted field."
00284  *
00285  *    2) "If the current character is an unquoted newline character,
00286  *        the current token will be delimited."
00287  *
00288  *    3) "If the current character is an unquoted blank character, any
00289  *        token containing the previous character is delimited and the
00290  *        current character will be discarded."
00291  *
00292  *    4) "If the previous character was part of a word, the current
00293  *        character will be appended to that word."
00294  *
00295  *    5) "If the current character is a "#", it and all subsequent
00296  *        characters up to, but excluding, the next newline character
00297  *        will be discarded as a comment. The newline character that
00298  *        ends the line is not considered part of the comment. The
00299  *        "#" starts a comment only when it is at the beginning of a
00300  *        token. Since the search for the end-of-comment does not
00301  *        consider an escaped newline character specially, a comment
00302  *        cannot be continued to the next line."
00303  *
00304  *    6) "The current character will be used as the start of a new word."
00305  *
00306  *
00307  *  - for each token (word), perform portions of word expansion, namely
00308  *    field splitting (using default whitespace IFS) and quote
00309  *    removal.  Field splitting may increase the number of words.
00310  *    Quote removal does not increase the number of words.
00311  *
00312  *   "If the complete expansion appropriate for a word results in an
00313  *   empty field, that empty field will be deleted from the list of
00314  *   fields that form the completely expanded command, unless the
00315  *   original word contained single-quote or double-quote characters."
00316  *    - UNIX98 spec
00317  *
00318  *
00319  */
00320 
00321 static dbus_bool_t
00322 delimit_token (DBusString *token,
00323                DBusList **retval,
00324                DBusError *error)
00325 {
00326   char *str;
00327 
00328   str = _dbus_strdup (_dbus_string_get_data (token));
00329   if (!str)
00330     {
00331       _DBUS_SET_OOM (error);
00332       return FALSE;
00333     }
00334 
00335   if (!_dbus_list_append (retval, str))
00336     {
00337       dbus_free (str);
00338       _DBUS_SET_OOM (error);
00339       return FALSE;
00340     }
00341 
00342   return TRUE;
00343 }
00344 
00345 static DBusList*
00346 tokenize_command_line (const char *command_line, DBusError *error)
00347 {
00348   char current_quote;
00349   const char *p;
00350   DBusString current_token;
00351   DBusList *retval = NULL;
00352   dbus_bool_t quoted;;
00353 
00354   current_quote = '\0';
00355   quoted = FALSE;
00356   p = command_line;
00357 
00358   if (!_dbus_string_init (&current_token))
00359     {
00360       _DBUS_SET_OOM (error);
00361       return NULL;
00362     }
00363 
00364   while (*p)
00365     {
00366       if (current_quote == '\\')
00367         {
00368           if (*p == '\n')
00369             {
00370               /* we append nothing; backslash-newline become nothing */
00371             }
00372           else
00373             {
00374               if (!_dbus_string_append_byte (&current_token, '\\') || 
00375                   !_dbus_string_append_byte (&current_token, *p))
00376                 {
00377                   _DBUS_SET_OOM (error);
00378                   goto error;
00379                 }
00380             }
00381 
00382           current_quote = '\0';
00383         }
00384       else if (current_quote == '#')
00385         {
00386           /* Discard up to and including next newline */
00387           while (*p && *p != '\n')
00388             ++p;
00389 
00390           current_quote = '\0';
00391           
00392           if (*p == '\0')
00393             break;
00394         }
00395       else if (current_quote)
00396         {
00397           if (*p == current_quote &&
00398               /* check that it isn't an escaped double quote */
00399               !(current_quote == '"' && quoted))
00400             {
00401               /* close the quote */
00402               current_quote = '\0';
00403             }
00404 
00405           /* Everything inside quotes, and the close quote,
00406            * gets appended literally.
00407            */
00408 
00409           if (!_dbus_string_append_byte (&current_token, *p))
00410             {
00411               _DBUS_SET_OOM (error);
00412               goto error;
00413             }
00414         }
00415       else
00416         {
00417           switch (*p)
00418             {
00419             case '\n':
00420               if (!delimit_token (&current_token, &retval, error))
00421                 goto error;
00422 
00423                 _dbus_string_free (&current_token);
00424 
00425                 if (!_dbus_string_init (&current_token))
00426                   {
00427                     _DBUS_SET_OOM (error);
00428                     goto init_error;
00429                   }
00430 
00431               break;
00432 
00433             case ' ':
00434             case '\t':
00435               /* If the current token contains the previous char, delimit
00436                * the current token. A nonzero length
00437                * token should always contain the previous char.
00438                */
00439               if (_dbus_string_get_length (&current_token) > 0)
00440                 {
00441                   if (!delimit_token (&current_token, &retval, error))
00442                     goto error;
00443 
00444                   _dbus_string_free (&current_token);
00445 
00446                   if (!_dbus_string_init (&current_token))
00447                     {
00448                       _DBUS_SET_OOM (error);
00449                       goto init_error;
00450                     }
00451 
00452                 }
00453               
00454               /* discard all unquoted blanks (don't add them to a token) */
00455               break;
00456 
00457 
00458               /* single/double quotes are appended to the token,
00459                * escapes are maybe appended next time through the loop,
00460                * comment chars are never appended.
00461                */
00462               
00463             case '\'':
00464             case '"':
00465               if (!_dbus_string_append_byte (&current_token, *p))
00466                 {
00467                   _DBUS_SET_OOM (error);
00468                   goto error;
00469                 }
00470 
00471               /* FALL THRU */
00472               
00473             case '#':
00474             case '\\':
00475               current_quote = *p;
00476               break;
00477 
00478             default:
00479               /* Combines rules 4) and 6) - if we have a token, append to it,
00480                * otherwise create a new token.
00481                */
00482               if (!_dbus_string_append_byte (&current_token, *p))
00483                 {
00484                   _DBUS_SET_OOM (error);
00485                   goto error;
00486                 }
00487               break;
00488             }
00489         }
00490 
00491       /* We need to count consecutive backslashes mod 2, 
00492        * to detect escaped doublequotes.
00493        */
00494       if (*p != '\\')
00495         quoted = FALSE;
00496       else
00497         quoted = !quoted;
00498 
00499       ++p;
00500     }
00501 
00502   if (!delimit_token (&current_token, &retval, error))
00503     goto error;
00504 
00505   if (current_quote)
00506     {
00507       dbus_set_error_const (error, DBUS_ERROR_INVALID_ARGS, "Unclosed quotes in command line");
00508       goto error;
00509     }
00510 
00511   if (retval == NULL)
00512     {
00513       dbus_set_error_const (error, DBUS_ERROR_INVALID_ARGS, "No tokens found in command line");
00514       goto error;
00515     }
00516  
00517   _dbus_string_free (&current_token);
00518  
00519   return retval;
00520 
00521  error:
00522   _dbus_string_free (&current_token);
00523 
00524  init_error:
00525   if (retval)
00526     {
00527       _dbus_list_foreach (&retval, (DBusForeachFunction) dbus_free, NULL);
00528       _dbus_list_clear (&retval);
00529     }
00530 
00531   return NULL;
00532 }
00533 
00551 dbus_bool_t
00552 _dbus_shell_parse_argv (const char *command_line,
00553                         int        *argcp,
00554                         char     ***argvp,
00555                         DBusError  *error)
00556 {
00557   /* Code based on poptParseArgvString() from libpopt */
00558   int argc = 0;
00559   char **argv = NULL;
00560   DBusList *tokens = NULL;
00561   int i;
00562   DBusList *tmp_list;
00563   
00564   if (!command_line)
00565     {
00566       _dbus_verbose ("Command line is NULL\n");
00567       return FALSE;
00568     }
00569 
00570   tokens = tokenize_command_line (command_line, error);
00571   if (tokens == NULL)
00572     {
00573       _dbus_verbose ("No tokens for command line '%s'\n", command_line);
00574       return FALSE;
00575     }
00576 
00577   /* Because we can't have introduced any new blank space into the
00578    * tokens (we didn't do any new expansions), we don't need to
00579    * perform field splitting. If we were going to honor IFS or do any
00580    * expansions, we would have to do field splitting on each word
00581    * here. Also, if we were going to do any expansion we would need to
00582    * remove any zero-length words that didn't contain quotes
00583    * originally; but since there's no expansion we know all words have
00584    * nonzero length, unless they contain quotes.
00585    * 
00586    * So, we simply remove quotes, and don't do any field splitting or
00587    * empty word removal, since we know there was no way to introduce
00588    * such things.
00589    */
00590 
00591   argc = _dbus_list_get_length (&tokens);
00592   argv = dbus_new (char *, argc + 1);
00593   if (!argv)
00594     {
00595       _DBUS_SET_OOM (error);
00596       goto error;
00597     }
00598 
00599   i = 0;
00600   tmp_list = tokens;
00601   while (tmp_list)
00602     {
00603       argv[i] = _dbus_shell_unquote (tmp_list->data);
00604 
00605       if (!argv[i])
00606         {
00607           int j;
00608           for (j = 0; j < i; j++)
00609             dbus_free(argv[j]);
00610 
00611           dbus_free (argv);
00612           _DBUS_SET_OOM (error);
00613           goto error;
00614         }
00615 
00616       tmp_list = _dbus_list_get_next_link (&tokens, tmp_list);
00617       ++i;
00618     }
00619   argv[argc] = NULL;
00620   
00621   _dbus_list_foreach (&tokens, (DBusForeachFunction) dbus_free, NULL);
00622   _dbus_list_clear (&tokens);
00623   
00624   if (argcp)
00625     *argcp = argc;
00626 
00627   if (argvp)
00628     *argvp = argv;
00629   else
00630     dbus_free_string_array (argv);
00631 
00632   return TRUE;
00633 
00634  error:
00635   _dbus_list_foreach (&tokens, (DBusForeachFunction) dbus_free, NULL);
00636   _dbus_list_clear (&tokens);
00637 
00638   return FALSE;
00639 
00640 }

Generated on Tue Apr 15 15:54:01 2008 for D-Bus by  doxygen 1.5.1