add -z/--null-data

2012-02-05  Paolo Bonzini  <bonzini@gnu.org>
            Jim Hill  <gjthill@gmail.com>

	* autoboot.conf: Change getline to getdelim.
	* doc/sed-in.texi: Document -z/--null-data.
	* doc/sed.texi: Regenerate.
	* sed/execute.c: Change '\n' to buffer_delimiter.
	* sed/sed.c: Add support for -z/--null-data.
	* sed/sed.h: Add buffer_delimiter.
	* sed/utils.c: Change ck_getline to ck_getdelim.
	* sed/utils.h: Change ck_getline to ck_getdelim.
	* NEWS: Document new option.

11 files changed, 77 insertions, 21 deletions

diff --git a/ChangeLog b/ChangeLog
index 53cfb44..bc660cf 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,17 @@
+2012-02-05  Paolo Bonzini  <bonzini@gnu.org>
+            Jim Hill  <gjthill@gmail.com>
+
+	* autoboot.conf: Change getline to getdelim.
+	* doc/sed-in.texi: Document -z/--null-data.
+	* doc/sed.texi: Regenerate.
+	* doc/sed.1: Regenerate.
+	* sed/execute.c: Change '\n' to buffer_delimiter.
+	* sed/sed.c: Add support for -z/--null-data.
+	* sed/sed.h: Add buffer_delimiter.
+	* sed/utils.c: Change ck_getline to ck_getdelim.
+	* sed/utils.h: Change ck_getline to ck_getdelim.
+	* NEWS: Document new option.
+
 2011-11-23  Paolo Bonzini  <bonzini@gnu.org>
 
 	* doc/sed-in.texi: Document how multiline mode affects matching
diff --git a/NEWS b/NEWS
index 7b56799..2e71f46 100644
--- a/NEWS
+++ b/NEWS
@@ -14,6 +14,8 @@ Sed 4.2.2
 * --posix fails for scripts (or fragments as passed to the -e option) that
   end in a backslash, as they are not portable.
 
+* New option -z (--null-data) to separate lines by ASCII NUL characters.
+
 ----------------------------------------------------------------------------
 Sed 4.2.1
 
diff --git a/autoboot.conf b/autoboot.conf
index 436bace..c6919ac 100644
--- a/autoboot.conf
+++ b/autoboot.conf
@@ -18,7 +18,7 @@
 
 # gnulib modules used by this package.
 gnulib_modules="
-    acl alloca btowc c-ctype extensions fwriting getline getopt gettext-h
+    acl alloca btowc c-ctype extensions fwriting getdelim getopt gettext-h
     localcharset mbrlen mbrtowc mbsinit memchr mkostemp obstack pathmax regex
     rename selinux-h stdbool stat-macros ssize_t strerror strverscmp
     unlocked-io verify version-etc-fsf wcrtomb wctob"
diff --git a/doc/sed-in.texi b/doc/sed-in.texi
index bc7320d..ecaabad 100644
--- a/doc/sed-in.texi
+++ b/doc/sed-in.texi
@@ -369,6 +369,16 @@ Buffer both input and output as minimally as practical.
 the likes of @samp{tail -f}, and you wish to see the transformed
 output as soon as possible.)
 
+@item -z
+@itemx --null-data
+@itemx --zero-terminated
+@opindex -z
+@opindex --null-data
+@opindex --zero-terminated
+Treat the input as a set of lines, each terminated by a zero byte
+(the ASCII @samp{NUL} character) instead of a newline.  This option can
+be used with commands like @samp{sort -z} and @samp{find -print0}
+to process arbitrary file names.
 @end table
 
 If no @option{-e}, @option{-f}, @option{--expression}, or @option{--file}
diff --git a/doc/sed.1 b/doc/sed.1
index 374b448..b7a86be 100644
--- a/doc/sed.1
+++ b/doc/sed.1
@@ -1,5 +1,5 @@
 .\" DO NOT MODIFY THIS FILE!  It was generated by help2man 1.28.
-.TH SED "1" "December 2010" "sed 4.2.1" "User Commands"
+.TH SED "1" "February 2012" "sed 4.2.1" "User Commands"
 .SH NAME
 sed \- stream editor for filtering and transforming text
 .SH SYNOPSIS
@@ -38,7 +38,7 @@ follow symlinks when processing in place
 .HP
 \fB\-i[SUFFIX]\fR, \fB\-\-in\-place\fR[=\fISUFFIX\fR]
 .IP
-edit files in place (makes backup if extension supplied)
+edit files in place (makes backup if SUFFIX supplied)
 .HP
 \fB\-l\fR N, \fB\-\-line\-length\fR=\fIN\fR
 .IP
@@ -61,6 +61,10 @@ long stream.
 .IP
 load minimal amounts of data from the input files and flush
 the output buffers more often
+.HP
+\fB\-z\fR, \fB\-\-null\-data\fR
+.IP
+separate lines by NUL characters
 .TP
 \fB\-\-help\fR
 display this help and exit
@@ -374,7 +378,7 @@ General help using GNU software: <http://www.gnu.org/gethelp/>.
 E-mail bug reports to: <bug-sed@gnu.org>.
 Be sure to include the word ``sed'' somewhere in the ``Subject:'' field.
 .SH COPYRIGHT
-Copyright \(co 2010 Free Software Foundation, Inc.
+Copyright \(co 2011 Free Software Foundation, Inc.
 License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>.
 .br
 This is free software: you are free to change and redistribute it.
diff --git a/doc/sed.texi b/doc/sed.texi
index 689113d..d599e74 100644
--- a/doc/sed.texi
+++ b/doc/sed.texi
@@ -370,6 +370,16 @@ Buffer both input and output as minimally as practical.
 the likes of @samp{tail -f}, and you wish to see the transformed
 output as soon as possible.)
 
+@item -z
+@itemx --null-data
+@itemx --zero-terminated
+@opindex -z
+@opindex --null-data
+@opindex --zero-terminated
+Treat the input as a set of lines, each terminated by a zero byte
+(the ASCII @samp{NUL} character) instead of a newline.  This option can
+be used with commands like @samp{sort -z} and @samp{find -print0}
+to process arbitrary file names.
 @end table
 
 If no @option{-e}, @option{-f}, @option{--expression}, or @option{--file}
diff --git a/sed/execute.c b/sed/execute.c
index 076ed0a..6eaf70e 100644
--- a/sed/execute.c
+++ b/sed/execute.c
@@ -472,7 +472,7 @@ line_append(from, to, state)
   struct line *to;
   int state;
 {
-  str_append(to, "\n", 1);
+  str_append(to, &buffer_delimiter, 1);
   str_append(to, from->active, from->length);
   to->chomped = from->chomped;
 
@@ -525,12 +525,12 @@ read_file_line(input)
   static char *b;
   static size_t blen;
 
-  long result = ck_getline (&b, &blen, input->fp);
+  long result = ck_getdelim (&b, &blen, buffer_delimiter, input->fp);
   if (result <= 0)
     return false;
 
   /* Remove the trailing new-line that is left by getline. */
-  if (b[result - 1] == '\n')
+  if (b[result - 1] == buffer_delimiter)
     --result;
   else
     line.chomped = false;
@@ -547,7 +547,7 @@ output_missing_newline(outf)
 {
   if (outf->missing_newline)
     {
-      ck_fwrite("\n", 1, 1, outf->fp);
+      ck_fwrite(&buffer_delimiter, 1, 1, outf->fp);
       outf->missing_newline = false;
     }
 }
@@ -576,7 +576,7 @@ output_line(text, length, nl, outf)
   if (length)
     ck_fwrite(text, 1, length, outf->fp);
   if (nl)
-    ck_fwrite("\n", 1, 1, outf->fp);
+    ck_fwrite(&buffer_delimiter, 1, 1, outf->fp);
   else
     outf->missing_newline = true;
 
@@ -1315,7 +1315,7 @@ do_subst(sub)
 	     for 'g' as to while the third argument is incorrect anyway.  */
 	  line_exchange(&line, &s_accum, true);
 	  if (line.length &&
-	      line.active[line.length - 1] == '\n')
+	      line.active[line.length - 1] == buffer_delimiter)
 	    line.length--;
 	}
       else
@@ -1435,7 +1435,7 @@ execute_program(vec, input)
 
 	    case 'D':
 	      {
-		char *p = memchr(line.active, '\n', line.length);
+		char *p = memchr(line.active, buffer_delimiter, line.length);
 		if (!p)
 		  return -1;
 
@@ -1485,7 +1485,7 @@ execute_program(vec, input)
 		    {
 		      /* Store into pattern space for plain `e' commands */
 		      if (s_accum.length &&
-			  s_accum.active[s_accum.length - 1] == '\n')
+			  s_accum.active[s_accum.length - 1] == buffer_delimiter)
 			s_accum.length--;
 
 		      /* Exchange line and s_accum.  This can be much
@@ -1566,7 +1566,7 @@ execute_program(vec, input)
 	      break;
 
 	    case 'N':
-	      str_append(&line, "\n", 1);
+	      str_append(&line, &buffer_delimiter, 1);
  
               if (test_eof(input) || !read_pattern_space(input, vec, true))
                 {
@@ -1584,7 +1584,7 @@ execute_program(vec, input)
 
 	    case 'P':
 	      {
-		char *p = memchr(line.active, '\n', line.length);
+		char *p = memchr(line.active, buffer_delimiter, line.length);
 		output_line(line.active, p ? p - line.active : line.length,
 			    p ? true : line.chomped, &output_file);
 	      }
@@ -1614,7 +1614,8 @@ execute_program(vec, input)
 		  char *text = NULL;
 		  int result;
 
-		  result = ck_getline (&text, &buflen, cur_cmd->x.fp);
+		  result = ck_getdelim (&text, &buflen, buffer_delimiter,
+				        cur_cmd->x.fp);
 		  if (result != EOF)
 		    {
 		      aq = next_append_slot();
@@ -1657,7 +1658,7 @@ execute_program(vec, input)
 	    case 'W':
 	      if (cur_cmd->x.fp)
 	        {
-		  char *p = memchr(line.active, '\n', line.length);
+		  char *p = memchr(line.active, buffer_delimiter, line.length);
 		  output_line(line.active, p ? p - line.active : line.length,
 			      p ? true : line.chomped, cur_cmd->x.outf);
 	        }
diff --git a/sed/sed.c b/sed/sed.c
index aa085d4..5167a55 100644
--- a/sed/sed.c
+++ b/sed/sed.c
@@ -68,6 +68,9 @@ char *program_name;
 
 int extended_regexp_flags = 0;
 
+/* one-byte buffer delimiter */
+char buffer_delimiter = '\n';
+
 /* If set, fflush(stdout) on every line output. */
 bool unbuffered = false;
 
@@ -163,6 +166,8 @@ Usage: %s [OPTION]... {script-only-if-no-other-script} [input-file]...\n\
   fprintf(out, _("  -u, --unbuffered\n\
                  load minimal amounts of data from the input files and flush\n\
                  the output buffers more often\n"));
+  fprintf(out, _("  -z, --null-data\n\
+                 separate lines by NUL characters\n"));
   fprintf(out, _("      --help     display this help and exit\n"));
   fprintf(out, _("      --version  output version information and exit\n"));
   fprintf(out, _("\n\
@@ -183,9 +188,9 @@ main(argc, argv)
   char **argv;
 {
 #ifdef REG_PERL
-#define SHORTOPTS "bsnrRuEe:f:l:i::V:"
+#define SHORTOPTS "bsnrzRuEe:f:l:i::V:"
 #else
-#define SHORTOPTS "bsnruEe:f:l:i::V:"
+#define SHORTOPTS "bsnrzuEe:f:l:i::V:"
 #endif
 
   static struct option longopts[] = {
@@ -198,6 +203,8 @@ main(argc, argv)
     {"file", 1, NULL, 'f'},
     {"in-place", 2, NULL, 'i'},
     {"line-length", 1, NULL, 'l'},
+    {"null-data", 0, NULL, 'z'},
+    {"zero-terminated", 0, NULL, 'z'},
     {"quiet", 0, NULL, 'n'},
     {"posix", 0, NULL, 'p'},
     {"silent", 0, NULL, 'n'},
@@ -261,6 +268,10 @@ main(argc, argv)
 	  the_program = compile_file(the_program, optarg);
 	  break;
 
+	case 'z':
+	  buffer_delimiter = 0;
+	  break;
+
 	case 'F':
 	  follow_symlinks = true;
 	  break;
diff --git a/sed/sed.h b/sed/sed.h
index f8ccccc..13dcf5e 100644
--- a/sed/sed.h
+++ b/sed/sed.h
@@ -211,6 +211,9 @@ extern void fmt P_ ((const char *line, const char *line_end, int max_length, FIL
 
 extern int extended_regexp_flags;
 
+/* one-byte buffer delimiter */
+extern char buffer_delimiter;
+
 /* If set, fflush(stdout) on every line output,
    and turn off stream buffering on inputs.  */
 extern bool unbuffered;
diff --git a/sed/utils.c b/sed/utils.c
index 7a00f67..cb29b87 100644
--- a/sed/utils.c
+++ b/sed/utils.c
@@ -259,9 +259,10 @@ ck_fread(ptr, size, nmemb, stream)
 }
 
 size_t
-ck_getline(text, buflen, stream)
+ck_getdelim(text, buflen, buffer_delimiter, stream)
   char **text;
   size_t *buflen;
+  char buffer_delimiter;
   FILE *stream;
 {
   ssize_t result;
@@ -270,7 +271,7 @@ ck_getline(text, buflen, stream)
   error = ferror (stream);
   if (!error)
     {
-      result = getline (text, buflen, stream);
+      result = getdelim (text, buflen, buffer_delimiter, stream);
       error = ferror (stream);
     }
 
diff --git a/sed/utils.h b/sed/utils.h
index aef8e65..9132e3d 100644
--- a/sed/utils.h
+++ b/sed/utils.h
@@ -29,7 +29,7 @@ size_t ck_fread P_((VOID *ptr, size_t size, size_t nmemb, FILE *stream));
 void ck_fflush P_((FILE *stream));
 void ck_fclose P_((FILE *stream));
 const char *follow_symlink P_((const char *path));
-size_t ck_getline P_((char **text, size_t *buflen, FILE *stream));
+size_t ck_getdelim P_((char **text, size_t *buflen, char buffer_delimiter, FILE *stream));
 FILE * ck_mkstemp P_((char **p_filename, const char *tmpdir, const char *base,
 		      const char *mode));
 void ck_rename P_((const char *from, const char *to, const char *unlink_if_fail));