Improve variable handling in wordsplit.

Positional variables ($N and ${N}) are recognized. Variable names in
curly braces follow the same rules as unadorned ones.
This commit also changes memory reallocation strategy in wsplt_assign_var.
If ws_envbuf needs to be expanded, new allocation size is selected as
3/2 of the previous allocation, if that size is less than max(size_t).

2 files changed, 60 insertions, 23 deletions

diff --git a/include/wordsplit.h b/include/wordsplit.h
index 1a047f7..3a7ab25 100644
--- a/include/wordsplit.h
+++ b/include/wordsplit.h
@@ -1,4 +1,4 @@
 /* wordsplit - a word splitter
-   Copyright (C) 2009-2018 Sergey Poznyakoff
+   Copyright (C) 2009-2019 Sergey Poznyakoff
 
    This program is free software; you can redistribute it and/or modify it
@@ -67,7 +67,13 @@ struct wordsplit
 			       environment variables. */
 
-  char **ws_envbuf;
-  size_t ws_envidx;
-  size_t ws_envsiz;
+	/* Temporary storage for environment variables. It is initialized
+	   upon first assignment which occurs during the parsing process
+	   (e.g. ${x:=2}). When this happens, all variables from ws_env are
+	   moved to ws_envbuf first, and the ws_envbuf address is assigned
+	   to ws_env. From this moment on, all variable expansions are served
+	   from ws_envbuf. */
+  char **ws_envbuf;         /* Storage for variables */
+  size_t ws_envidx;         /* Index of first free slot */
+  size_t ws_envsiz;         /* Size of the ws_envbuf array */
   
   int (*ws_getvar) (char **ret, const char *var, size_t len, void *clos);
@@ -198,6 +204,6 @@ struct wordsplit
 #if 0 /* Unused value */
 #define WRDSO_ARGV            0x00000008
-/* Keep backslash in unrecognized escape sequences in words */
 #endif
+/* Keep backslash in unrecognized escape sequences in words */
 #define WRDSO_BSKEEP_WORD     0x00000010
 /* Handle octal escapes in words */
diff --git a/src/wordsplit.c b/src/wordsplit.c
index bad59b1..f94015a 100644
--- a/src/wordsplit.c
+++ b/src/wordsplit.c
@@ -1087,10 +1087,14 @@ wsplt_assign_var (struct wordsplit *wsp, const char *name, size_t namelen,
       else
 	{
-	  wsp->ws_envsiz *= 2;
-	  newenv = realloc (wsp->ws_envbuf,
-			    wsp->ws_envsiz * sizeof (wsp->ws_envbuf[0]));
+	  size_t n = wsp->ws_envsiz;
+	  
+	  if ((size_t) -1 / 3 * 2 / sizeof (wsp->ws_envbuf[0]) <= n)
+	    return _wsplt_nomem (wsp);
+	  n += (n + 1) / 2;
+	  newenv = realloc (wsp->ws_envbuf, n * sizeof (wsp->ws_envbuf[0]));
 	  if (!newenv)
 	    return _wsplt_nomem (wsp);
 	  wsp->ws_envbuf = newenv;
+	  wsp->ws_envsiz = n;
 	  wsp->ws_env = (const char**) wsp->ws_envbuf;
 	}
@@ -1129,4 +1133,27 @@ wsplt_assign_var (struct wordsplit *wsp, const char *name, size_t namelen,
 }
 
+/* Recover from what looked like a variable reference, but turned out
+   not to be one. STR points to first character after '$'. */
+static int
+expvar_recover (struct wordsplit *wsp, const char *str,
+		struct wordsplit_node **ptail, const char **pend, int flg)
+{
+  struct wordsplit_node *newnode;
+
+  if (wsnode_new (wsp, &newnode))
+    return 1;
+  wsnode_insert (wsp, newnode, *ptail, 0);
+  *ptail = newnode;
+  newnode->flags = _WSNF_WORD | flg;
+  newnode->v.word = malloc (3);
+  if (!newnode->v.word)
+    return _wsplt_nomem (wsp);
+  newnode->v.word[0] = '$';
+  newnode->v.word[1] = str[0];
+  newnode->v.word[2] = 0;
+  *pend = str;
+  return 0;
+}
+
 static int
 expvar (struct wordsplit *wsp, const char *str, size_t len,
@@ -1149,5 +1176,10 @@ expvar (struct wordsplit *wsp, const char *str, size_t len,
       *pend = str + i - 1;
     }
-  else if (str[0] == '{')
+  else if (ISDIGIT (str[0]))
+    {
+      i = 1;
+      *pend = str;
+    }
+  else if (str[0] == '{' && (ISVARBEG (str[1]) || ISDIGIT (str[1])))
     {
       str++;
@@ -1181,4 +1213,15 @@ expvar (struct wordsplit *wsp, const char *str, size_t len,
 	      break;
 	    }
+	  else if (ISDIGIT (str[1]))
+	    {
+	      if (!ISDIGIT (str[i]))
+		{
+		  return expvar_recover (wsp, str - 1, ptail, pend, flg);
+		}
+	    }
+	  else if (!ISVARCHR (str[i]))
+	    {
+	      return expvar_recover (wsp, str - 1, ptail, pend, flg);
+	    }
 	}
       if (i == len)
@@ -1187,17 +1230,5 @@ expvar (struct wordsplit *wsp, const char *str, size_t len,
   else
     {
-      if (wsnode_new (wsp, &newnode))
-	return 1;
-      wsnode_insert (wsp, newnode, *ptail, 0);
-      *ptail = newnode;
-      newnode->flags = _WSNF_WORD | flg;
-      newnode->v.word = malloc (3);
-      if (!newnode->v.word)
-	return _wsplt_nomem (wsp);
-      newnode->v.word[0] = '$';
-      newnode->v.word[1] = str[0];
-      newnode->v.word[2] = 0;
-      *pend = str;
-      return 0;
+      return expvar_recover (wsp, str, ptail, pend, flg);
     }
 
@@ -1418,5 +1449,5 @@ static int
 begin_var_p (int c)
 {
-  return c == '{' || ISVARBEG (c);
+  return c == '{' || ISVARBEG (c) || ISDIGIT (c);
 }