bug-bash
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Associative array indexes not working as expected in Bash 4.0


From: Chet Ramey
Subject: Re: Associative array indexes not working as expected in Bash 4.0
Date: Thu, 26 Feb 2009 16:58:33 -0500
User-agent: Thunderbird 2.0.0.19 (Macintosh/20081209)

AnMaster wrote:
> Configuration Information [Automatically generated, do not change]:
> Machine: x86_64
> OS: linux-gnu
> Compiler: gcc
> Compilation CFLAGS:  -DPROGRAM='bash' -DCONF_HOSTTYPE='x86_64'
> -DCONF_OSTYPE='linux-gnu' -DCONF_MACHTYPE='x86_64-unknown-linux-gnu'
> -DCONF_VENDOR='unknown'
> -DLOCALEDIR='/home/arvid/local/shells/bash/4.0/share/locale' -DPACKAGE='bash'
> -DSHELL -DHAVE_CONFIG_H   -I.  -I. -I./include -I./lib
> -DSYS_BASHRC=/etc/bash/bashrc -DSYS_BASH_LOGOUT=/etc/bash/bash_logout
> -DDEFAULT_PATH_VALUE=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
> -pipe -O1 -g
> uname output: Linux tux.lan 2.6.27-gentoo-r8-1 #1 Sat Jan 31 04:55:36 CET 2009
> x86_64 AMD Sempron(tm) Processor 3300+ AuthenticAMD GNU/Linux
> Machine Type: x86_64-unknown-linux-gnu
> 
> Bash Version: 4.0
> Patch Level: 0
> Release Status: release
> 
> Description:
> It seems that associative array indexes are not treated in a consistent way in
> bash 4.0. I was testing if associative arrays would work for what I needed 
> (the
> strings I plan to use for indexes are be untrusted and can legally contain
> anything except null-bytes, thus I was testing if it was robust with "strange"
> data). See "Repeat-By:" below.
> 
> So the issues are:
> 1. How do you escape the ] in myarray["a]a"] so that bash handles it the same
>    way as foo="a]a"; myarray["$foo"].
> 2. While myarray["a]a"] errors out when assigning, reading with
>    ${myarray["a]a"]} doesn't error out, but doesn't return the correct result
>    either.
> 3. I'm not sure what happened in myarray['a]=test2;#a']="def". Did it eval the
>    single-quoted string? Seems buggy anyway.

This patch should correct all of these associative array subscript parsing
problems.  Let me know how it works for you.

Chet
-- 
``The lyf so short, the craft so long to lerne.'' - Chaucer

Chet Ramey, ITS, CWRU    chet@case.edu    http://cnswww.cns.cwru.edu/~chet/
*** ../bash-4.0/parse.y 2009-01-08 08:29:12.000000000 -0500
--- parse.y     2009-02-25 17:25:56.000000000 -0500
***************
*** 2916,2919 ****
--- 2919,2923 ----
  #define P_COMMAND     0x08    /* parsing a command, so look for comments */
  #define P_BACKQUOTE   0x10    /* parsing a backquoted command substitution */
+ #define P_ARRAYSUB    0x20    /* parsing a [...] array subscript for 
assignment */
  
  /* Lexical state while parsing a grouping construct or $(...). */
***************
*** 3130,3133 ****
--- 3134,3139 ----
              FREE (nestret);
            }
+         else if ((flags & P_ARRAYSUB) && (tflags & LEX_WASDOL) && (ch == '(' 
|| ch == '{' || ch == '['))      /* ) } ] */
+           goto parse_dollar_word;
        }
        /* Parse an old-style command substitution within double quotes as a
***************
*** 3146,3149 ****
--- 3152,3156 ----
        /* check for $(), $[], or ${} inside quoted string. */
        {
+ parse_dollar_word:
          if (open == ch)       /* undo previous increment */
            count--;
***************
*** 4249,4253 ****
                      (token_index == 0 && (parser_state&PST_COMPASSIGN))))
          {
!         ttok = parse_matched_pair (cd, '[', ']', &ttoklen, 0);
          if (ttok == &matched_pair_error)
            return -1;          /* Bail immediately. */
--- 4256,4260 ----
                      (token_index == 0 && (parser_state&PST_COMPASSIGN))))
          {
!         ttok = parse_matched_pair (cd, '[', ']', &ttoklen, P_ARRAYSUB);
          if (ttok == &matched_pair_error)
            return -1;          /* Bail immediately. */
*** ../bash-4.0/arrayfunc.c     2009-01-04 14:32:21.000000000 -0500
--- arrayfunc.c 2009-02-25 07:58:54.000000000 -0500
***************
*** 605,666 ****
  }
  
! /* This function assumes s[i] == '['; returns with s[ret] == ']' if
!    an array subscript is correctly parsed. */
! int
! skipsubscript (s, i)
!      const char *s;
!      int i;
! {
!   int count, c;
! #if defined (HANDLE_MULTIBYTE)
!   mbstate_t state, state_bak;
!   size_t slength, mblength;
! #endif
! 
! #if defined (HANDLE_MULTIBYTE)
!   memset (&state, '\0', sizeof (mbstate_t));
!   slength = strlen (s + i);
! #endif
!   
!   count = 1;
!   while (count)
!     {
!       /* Advance one (possibly multibyte) character in S starting at I. */
! #if defined (HANDLE_MULTIBYTE)
!       if (MB_CUR_MAX > 1)
!       {
!         state_bak = state;
!         mblength = mbrlen (s + i, slength, &state);
! 
!         if (MB_INVALIDCH (mblength))
!           {
!             state = state_bak;
!             i++;
!             slength--;
!           }
!         else if (MB_NULLWCH (mblength))
!           return i;
!         else
!           {
!             i += mblength;
!             slength -= mblength;
!           }
!       }
!       else
! #endif
!       ++i;
! 
!       c = s[i];
! 
!       if (c == 0)
!       break;
!       else if (c == '[')
!       count++;
!       else if (c == ']')
!       count--;
!     }
! 
!   return i;
! }
  
  /* This function is called with SUB pointing to just after the beginning
--- 605,609 ----
  }
  
! /* skipsubscript moved to subst.c to use private functions. 2009/02/24. */
  
  /* This function is called with SUB pointing to just after the beginning
*** ../bash-4.0/subst.c 2009-01-28 14:34:12.000000000 -0500
--- subst.c     2009-02-25 09:18:33.000000000 -0500
***************
*** 223,226 ****
--- 223,227 ----
  static char *extract_delimited_string __P((char *, int *, char *, char *, 
char *, int));
  static char *extract_dollar_brace_string __P((char *, int *, int, int));
+ static int skip_matched_pair __P((const char *, int, int, int, int));
  
  static char *pos_params __P((char *, int, int, int));
***************
*** 1375,1378 ****
--- 1376,1480 ----
  #define CQ_RETURN(x) do { no_longjmp_on_fatal_error = 0; return (x); } while 
(0)
  
+ /* This function assumes s[i] == open; returns with s[ret] == close; used to
+    parse array subscripts.  FLAGS currently unused. */
+ static int
+ skip_matched_pair (string, start, open, close, flags)
+      const char *string;
+      int start, open, close, flags;
+ {
+   int i, pass_next, backq, si, c, count;
+   size_t slen;
+   char *temp, *ss;
+   DECLARE_MBSTATE;
+ 
+   slen = strlen (string + start) + start;
+   no_longjmp_on_fatal_error = 1;
+ 
+   i = start + 1;              /* skip over leading bracket */
+   count = 1;
+   pass_next = backq = 0;
+   ss = (char *)string;
+   while (c = string[i])
+     {
+       if (pass_next)
+       {
+         pass_next = 0;
+         if (c == 0)
+           CQ_RETURN(i);
+         ADVANCE_CHAR (string, slen, i);
+         continue;
+       }
+       else if (c == '\\')
+       {
+         pass_next = 1;
+         i++;
+         continue;
+       }
+       else if (backq)
+       {
+         if (c == '`')
+           backq = 0;
+         ADVANCE_CHAR (string, slen, i);
+         continue;
+       }
+       else if (c == '`')
+       {
+         backq = 1;
+         i++;
+         continue;
+       }
+       else if (c == open)
+       {
+         count++;
+         i++;
+         continue;
+       }
+       else if (c == close)
+       {
+         count--;
+         if (count == 0)
+           break;
+         i++;
+         continue;
+       }
+       else if (c == '\'' || c == '"')
+       {
+         i = (c == '\'') ? skip_single_quoted (ss, slen, ++i)
+                         : skip_double_quoted (ss, slen, ++i);
+         /* no increment, the skip functions increment past the closing quote. 
*/
+       }
+       else if (c == '$' && (string[i+1] == LPAREN || string[i+1] == LBRACE))
+       {
+         si = i + 2;
+         if (string[si] == '\0')
+           CQ_RETURN(si);
+ 
+         if (string[i+1] == LPAREN)
+           temp = extract_delimited_string (ss, &si, "$(", "(", ")", 
SX_NOALLOC|SX_COMMAND); /* ) */
+         else
+           temp = extract_dollar_brace_string (ss, &si, 0, SX_NOALLOC);
+         i = si;
+         if (string[i] == '\0')        /* don't increment i past EOS in loop */
+           break;
+         i++;
+         continue;
+       }
+       else
+       ADVANCE_CHAR (string, slen, i);
+     }
+ 
+   CQ_RETURN(i);
+ }
+ 
+ #if defined (ARRAY_VARS)
+ int
+ skipsubscript (string, start)
+      const char *string;
+      int start;
+ {
+   return (skip_matched_pair (string, start, '[', ']', 0));
+ }
+ #endif
+ 
  /* Skip characters in STRING until we find a character in DELIMS, and return
     the index of that character.  START is the index into string at which we

reply via email to

[Prev in Thread] Current Thread [Next in Thread]