Upstream 0f968ceb7bc1a65773979ef419872ce43677c790

Original by Paolo Bonzini  <bonzini@gnu.org> - 2012-04-13

Modified by Martin Briza (original patch didn't apply clearly)

* sed/compile.c (convert_number): Remove default_char argument,

 expect buf to point to it.  Remove maxdigits argument and compute

 it on the fly.

 (normalize_text): Unify calls to convert_number under the convert

 label.  For TEXT_REPLACEMENT add a backslash to the output if

 convert_number returns ch == '&'.

--

diff -rpu sed-4.2.1/sed/compile.c sed-4.2.1-modified/sed/compile.c

--- sed-4.2.1/sed/compile.c	2009-06-17 14:54:43.000000000 +0200

+++ sed-4.2.1-modified/sed/compile.c	2012-06-12 16:46:19.310301149 +0200

@@ -300,20 +300,19 @@ add_then_next(b, ch)

   return inchar();

 }

-static char * convert_number P_((char *, char *, const char *, int, int, int));

+static char * convert_number P_((char *, char *, const char *, int));

 static char *

-convert_number(result, buf, bufend, base, maxdigits, default_char)

+convert_number(result, buf, bufend, base)

   char *result;

   char *buf;

   const char *bufend;

   int base;

-  int maxdigits;

-  int default_char;

 {

   int n = 0;

+  int max = 1;

   char *p;

-  for (p=buf; p < bufend && maxdigits-- > 0; ++p)

+  for (p=buf+1; p < bufend && max <= 255; ++p, max *= base)

     {

       int d = -1;

       switch (*p)

@@ -339,8 +338,8 @@ convert_number(result, buf, bufend, base

 	break;

       n = n * base + d;

     }

-  if (p == buf)

-    *result = default_char;

+  if (p == buf+1)

+    *result = *buf;

   else

     *result = n;

   return p;

@@ -1417,6 +1416,8 @@ normalize_text(buf, len, buftype)

   const char *bufend = buf + len;

   char *p = buf;

   char *q = buf;

+  char ch;

+  int base;

   /* This variable prevents normalizing text within bracket

      subexpressions when conforming to POSIX.  If 0, we

@@ -1464,14 +1465,12 @@ normalize_text(buf, len, buftype)

 	  case 'v': *q++ = '\v'; p++; continue;

 	  case 'd': /* decimal byte */

-	    p = convert_number(q, p+1, bufend, 10, 3, 'd');

-	    q++;

-	    continue;

+            base = 10;

+            goto convert;

 	  case 'x': /* hexadecimal byte */

-	    p = convert_number(q, p+1, bufend, 16, 2, 'x');

-	    q++;

-	    continue;

+            base = 16;

+            goto convert;

 #ifdef REG_PERL

 	  case '0': case '1': case '2': case '3':

@@ -1480,8 +1479,8 @@ normalize_text(buf, len, buftype)

 		&& p+1 < bufend

 		&& p[1] >= '0' && p[1] <= '9')

 	      {

-		p = convert_number(q, p, bufend, 8, 3, *p);

-		q++;

+                base = 8;

+                goto convert;

 	      }

 	    else

 	      {

@@ -1495,8 +1494,8 @@ normalize_text(buf, len, buftype)

 	  case 'o': /* octal byte */

 	    if (!(extended_regexp_flags & REG_PERL))

 	      {

-	        p = convert_number(q, p+1, bufend,  8, 3, 'o');

-		q++;

+                base = 8;

+                goto convert;

 	      }

 	    else

 	      {

@@ -1508,10 +1507,16 @@ normalize_text(buf, len, buftype)

 	    continue;

 #else

 	  case 'o': /* octal byte */

-	    p = convert_number(q, p+1, bufend,  8, 3, 'o');

-	    q++;

-	    continue;

+            base = 8;

 #endif

+convert:

+            p = convert_number(&ch, p, bufend, base);

+

+            /* for an ampersand in a replacement, pass the \ up one level */

+            if (buftype == TEXT_REPLACEMENT && ch == '&')

+              *q++ = '\\';

+            *q++ = ch;

+            continue;

 	  case 'c':

 	    if (++p < bufend)

diff -rpu sed-4.2.1/testsuite/Makefile.am sed-4.2.1-modified/testsuite/Makefile.am

--- sed-4.2.1/testsuite/Makefile.am	2009-06-25 20:55:35.000000000 +0200

+++ sed-4.2.1-modified/testsuite/Makefile.am	2012-06-12 16:45:56.875331158 +0200

@@ -24,7 +24,7 @@ SEDTESTS += \

 	fasts uniq manis khadafy linecnt eval distrib 8to7 y-bracket \

 	y-newline allsub cv-vars classes middle bsd stdin flipcase \

 	insens subwrite writeout readin insert utf8-1 utf8-2 utf8-3 utf8-4 \

-	badenc inplace-hold brackets \

+	badenc inplace-hold brackets amp-escape\

 	help version file quiet \

 	factor binary3 binary2 binary dc

@@ -39,6 +39,7 @@ EXTRA_DIST = \

 	8bit.good 8bit.inp 8bit.sed \

 	8to7.good 8to7.inp 8to7.sed \

 	allsub.good allsub.inp allsub.sed \

+	amp-escape.good amp-escape.inp amp-escape.sed \

 	appquit.good appquit.inp appquit.sed \

 	binary.good binary.inp binary.sed binary2.sed binary3.sed \

 	bkslashes.good bkslashes.inp bkslashes.sed \

diff -rpu sed-4.2.1/testsuite/Makefile.tests sed-4.2.1-modified/testsuite/Makefile.tests

--- sed-4.2.1/testsuite/Makefile.tests	2009-06-25 20:55:35.000000000 +0200

+++ sed-4.2.1-modified/testsuite/Makefile.tests	2012-06-12 16:46:17.195303978 +0200

@@ -21,7 +21,7 @@ SKIP = :>$@.skip; exit 77

 enable sep inclib 8bit 8to7 newjis xabcx dollar noeol bkslashes \

 numsub head madding mac-mf empty xbxcx xbxcx3 recall recall2 xemacs \

 appquit fasts uniq manis linecnt khadafy allsub flipcase space modulo \

-y-bracket y-newline insert brackets::

+y-bracket y-newline insert brackets amp-escape::

 	$(SEDENV) $(SED) -f $(srcdir)/$@.sed \

 		< $(srcdir)/$@.inp | $(TR) -d \\r > $@.out 

 	$(CMP) $(srcdir)/$@.good $@.out