diff -ruN -x '*.xml' -x AUTHORS -x BACKLOG -x ABOUT-NLS -x alloca.h -x flags.texi -x atconfig -x '*.gmo' -x config.h -x requests -x version.texi -x compile -x depcomp -x missing -x configure -x stamp-vti -x '*.template' -x POTFILES -x stamp-h1 -x testsuite -x '*.Po' -x INSTALL -x Makefile -x '*-sh' -x '*.m4' -x '*.0' -x '*.2' -x '*.3' -x '*.4' -x '*.guess' -x '*.rpath' -x '*.sub' -x '*.tex' -x '*.status' -x '*~' -x '*.in' -x '*.po' -x '*.pot' -x '*.info' -x '*.1' -x '*.log' wdiff-1.2.2/doc/wdiff.texi wdiff-1.2.2.htmldiff/doc/wdiff.texi
--- wdiff-1.2.2/doc/wdiff.texi	2013-10-21 11:28:04.000000000 +0200
+++ wdiff-1.2.2.htmldiff/doc/wdiff.texi	2020-03-09 14:06:35.193486586 +0100
@@ -350,6 +350,30 @@
 wdiff -l @var{old_file} @var{new_file} | less
 @end example
 
+@item --ignore-markup
+@item -m
+Skip XML or HTML markup. This option causes HTML or XML tags (enclosed
+in angle brackets <...>) to be ignored for the purpose of comparing
+the files.
+
+If none of the options @option{-2} or @option{-3} is given, tags from
+the first file are ignored and tags from the second file are copied to
+the output. If option @option{-3} or both @option{-1} and @option{-2}
+are given, no tags are copied. With only @option{-2}, the tags from
+the first file are copied.
+
+The option @option{-m} changes the default strings that mark the start
+and end of changes to @samp{<del>}, @samp{</del>}, @samp{<ins>} and
+@samp{</ins>}. These can be changed with the options @option{-w},
+@option{-x}, @option{-y} and @option{-z}.
+
+WARNING: @command{wdiff} inserts the markers without checking
+if they are valid in a given context. E.g., in HTML, it is not correct
+to insert @samp{<del>} inside a @samp{<style>} element, but
+@command{wdiff} will do so anyway, if it finds a difference. (In XML
+terms: if the markers themselves are well-formed tags, the output will be
+@emph{well-formed} as well, but not necessarily @emph{valid.})
+
 @item --terminal
 @itemx -t
 Force the production of @code{termcap} strings for emphasising parts of
@@ -387,6 +411,8 @@
 other means of distinguishing where such text starts; in this case the
 default start delete string is @samp{[-}.
 
+But if the option @option{-m} is given, the default is @samp{<del>}.
+
 @item --end-delete @var{argument}
 @itemx -x @var{argument}
 Use @var{argument} as the @dfn{end delete} string.  This string will be
@@ -395,6 +421,8 @@
 distinguishing where such text ends; in this case the default end delete
 string is @option{-]}.
 
+But if the option @option{-m} is given, the default is @samp{</del>}.
+
 @item --start-insert @var{argument}
 @itemx -y @var{argument}
 Use @var{argument} as the @dfn{start insert} string.  This string will
@@ -403,6 +431,8 @@
 other means of distinguishing where such text starts; in this case the
 default start insert string is @samp{@{+}.
 
+But if the option @option{-m} is given, the default is @samp{<ins>}.
+
 @item --end-insert @var{argument}
 @itemx -z @var{argument}
 Use @var{argument} as the @dfn{end insert} string.  This string will be
@@ -411,6 +441,8 @@
 distinguishing where such text ends; in this case the default end insert
 string is @samp{+@}}.
 
+But if the option @option{-m} is given, the default is @samp{</ins>}.
+
 @item --avoid-wraps
 @itemx -n
 Avoid spanning the end of line while showing deleted or inserted text.
diff -ruN -x '*.xml' -x AUTHORS -x BACKLOG -x ABOUT-NLS -x alloca.h -x flags.texi -x atconfig -x '*.gmo' -x config.h -x requests -x version.texi -x compile -x depcomp -x missing -x configure -x stamp-vti -x '*.template' -x POTFILES -x stamp-h1 -x testsuite -x '*.Po' -x INSTALL -x Makefile -x '*-sh' -x '*.m4' -x '*.0' -x '*.2' -x '*.3' -x '*.4' -x '*.guess' -x '*.rpath' -x '*.sub' -x '*.tex' -x '*.status' -x '*~' -x '*.in' -x '*.po' -x '*.pot' -x '*.info' -x '*.1' -x '*.log' wdiff-1.2.2/src/wdiff.c wdiff-1.2.2.htmldiff/src/wdiff.c
--- wdiff-1.2.2/src/wdiff.c	2013-10-21 11:28:04.000000000 +0200
+++ wdiff-1.2.2.htmldiff/src/wdiff.c	2020-03-04 19:59:35.481404468 +0100
@@ -94,6 +94,7 @@
   {"end-insert", 1, NULL, 'z'},
   {"help", 0, NULL, 'h'},
   {"ignore-case", 0, NULL, 'i'},
+  {"ignore-markup", 0, NULL, 'm'},
   {"less-mode", 0, NULL, 'l'},
   {"no-common", 0, NULL, '3'},
   {"no-deleted", 0, NULL, '1'},
@@ -116,6 +117,7 @@
 int inhibit_common;		/* inhibit display of common words */
 int diff_input;			/* expect (unified) diff as input */
 int ignore_case;		/* ignore case in comparisons */
+int ignore_markup;		/* skip XML or HTML tags */
 int show_statistics;		/* if printing summary statistics */
 int no_wrapping;		/* end/restart strings at end of lines */
 int autopager;			/* if calling the pager automatically */
@@ -243,7 +245,14 @@
 
   /* Ensure some default strings.  */
 
-  if (!overstrike)
+  if (ignore_markup)
+    {
+      if (!user_delete_start) user_delete_start = "<del>";
+      if (!user_delete_end) user_delete_end = "</del>";
+      if (!user_insert_start) user_insert_start = "<ins>";
+      if (!user_insert_end) user_insert_end = "</ins>";
+    }
+  else if (!overstrike)
     {
       if (!term_delete_start && !user_delete_start)
 	user_delete_start = "[-";
@@ -361,9 +370,37 @@
   copy_mode = COPY_NORMAL;
 }
 
-/*--------------------------------.
-| Skip over white space on SIDE.  |
-`--------------------------------*/
+/*---------------------------.
+| Skip over markup on SIDE.  |
+`---------------------------*/
+
+static void
+skip_markup (SIDE * side)
+{
+  if (interrupted)
+    longjmp (signal_label, 1);
+
+  while (side->character != '>' && side->character != EOF)
+    {
+      if (side->character == '"')
+	{
+	  do side->character = getc (side->file);
+	  while (side->character != '"' && side->character != EOF);
+	}
+      else if (side->character == '\'')
+	{
+	  do side->character = getc (side->file);
+	  while (side->character != '\'' && side->character != EOF);
+	}
+
+      side->character = getc (side->file);
+    }
+  /* Now side->character is '>' or EOF. */
+}
+
+/*-------------------------------------------.
+| Skip over white space and markup on SIDE.  |
+`-------------------------------------------*/
 
 static void
 skip_whitespace (SIDE * side)
@@ -371,8 +408,18 @@
   if (interrupted)
     longjmp (signal_label, 1);
 
-  while (isspace (side->character))
-    side->character = getc (side->file);
+  if (ignore_markup)
+
+    while (isspace (side->character) || side->character == '<')
+      {
+	if (side->character == '<') skip_markup (side);
+	side->character = getc (side->file);
+      }
+
+  else
+
+    while (isspace (side->character))
+      side->character = getc (side->file);
 }
 
 /*------------------------------------.
@@ -385,90 +432,150 @@
   if (interrupted)
     longjmp (signal_label, 1);
 
-  while (side->character != EOF && !isspace (side->character))
-    side->character = getc (side->file);
+  if (ignore_markup)
+
+    while (side->character != EOF && !isspace (side->character)
+	   && side->character != '<')
+      side->character = getc (side->file);
+
+  else
+
+    while (side->character != EOF && !isspace (side->character))
+      side->character = getc (side->file);
+
   side->position++;
 }
 
-/*-------------------------------------.
-| Copy white space from SIDE to FILE.  |
-`-------------------------------------*/
+/*---------------------------------------------.
+| write one character according to copy_mode.  |
+`---------------------------------------------*/
 
 static void
-copy_whitespace (SIDE * side, FILE * file)
+outc (int c, FILE * file)
+{
+  /* While changing lines, ensure we stop any special display prior
+     to, and restore the special display after.  When copy_mode is
+     anything else than COPY_NORMAL, file is always output_file.  We
+     care underlining whitespace or overstriking it with itself,
+     because "less" understands these things as emphasis requests.  */
+
+  switch (copy_mode)
+    {
+    case COPY_NORMAL:
+      putc (c, file);
+      break;
+
+    case COPY_DELETED:
+      if (c == '\n')
+	{
+	  if (no_wrapping && user_delete_end)
+	    fprintf (file, "%s", user_delete_end);
+# if HAVE_TPUTS
+	  if (term_delete_end)
+	    tputs (term_delete_end, 0, putc_for_tputs);
+# endif
+	  putc ('\n', file);
+# if HAVE_TPUTS
+	  if (term_delete_start)
+	    tputs (term_delete_start, 0, putc_for_tputs);
+# endif
+	  if (no_wrapping && user_delete_start)
+	    fprintf (file, "%s", user_delete_start);
+	}
+      else if (overstrike_for_less)
+	{
+	  putc ('_', file);
+	  putc ('\b', output_file);
+	  putc (c, output_file);
+	}
+      else if (overstrike && !isspace (c) && c != '_')
+	{
+	  putc ('_', file);
+	  putc ('\b', output_file);
+	  putc (c, output_file);
+	}
+      else
+	putc (c, file);
+      break;
+
+    case COPY_INSERTED:
+      if (c == '\n')
+	{
+	  if (no_wrapping && user_insert_end)
+	    fprintf (file, "%s", user_insert_end);
+# if HAVE_TPUTS
+	  if (term_insert_end)
+	    tputs (term_insert_end, 0, putc_for_tputs);
+# endif
+	  putc ('\n', file);
+# if HAVE_TPUTS
+	  if (term_insert_start)
+	    tputs (term_insert_start, 0, putc_for_tputs);
+# endif
+	  if (no_wrapping && user_insert_start)
+	    fprintf (file, "%s", user_insert_start);
+	}
+      else if (overstrike)
+	{
+	  putc (c, file);
+	  putc ('\b', file);
+	  putc (c, file);
+	}
+      else
+	putc (c, file);
+      break;
+    }
+}
+
+/*--------------------------------.
+| Copy markup from SIDE to FILE.  |
+`--------------------------------*/
+
+static void
+copy_markup (SIDE * side, FILE * file)
 {
   if (interrupted)
     longjmp (signal_label, 1);
 
-  while (isspace (side->character))
+  while (side->character != '>' && side->character != EOF)
     {
+      /* At a quote mark, copy until the end of the string "..." or '...'. */
+      if (side->character == '"')
+	do {outc (side->character, file); side->character = getc (side->file);}
+	while (side->character != '"' && side->character != EOF);
+      else if (side->character == '\'')
+	do {outc (side->character, file); side->character = getc (side->file);}
+	while (side->character != '\'' && side->character != EOF);
 
-      /* While changing lines, ensure we stop any special display prior
-         to, and restore the special display after.  When copy_mode is
-         anything else than COPY_NORMAL, file is always output_file.  We
-         care underlining whitespace or overstriking it with itself,
-         because "less" understands these things as emphasis requests.  */
-
-      switch (copy_mode)
+      if (side->character != EOF)
 	{
-	case COPY_NORMAL:
-	  putc (side->character, file);
-	  break;
+	  outc (side->character, file);
+	  side->character = getc (side->file);
+	}
+    }
 
-	case COPY_DELETED:
-	  if (side->character == '\n')
-	    {
-	      if (no_wrapping && user_delete_end)
-		fprintf (output_file, "%s", user_delete_end);
-#if HAVE_TPUTS
-	      if (term_delete_end)
-		tputs (term_delete_end, 0, putc_for_tputs);
-#endif
-	      putc ('\n', output_file);
-#if HAVE_TPUTS
-	      if (term_delete_start)
-		tputs (term_delete_start, 0, putc_for_tputs);
-#endif
-	      if (no_wrapping && user_delete_start)
-		fprintf (output_file, "%s", user_delete_start);
-	    }
-	  else if (overstrike_for_less)
-	    {
-	      putc ('_', output_file);
-	      putc ('\b', output_file);
-	      putc (side->character, output_file);
-	    }
-	  else
-	    putc (side->character, output_file);
-	  break;
+  if (side->character != EOF)
+    outc (side->character, file);
 
-	case COPY_INSERTED:
-	  if (side->character == '\n')
-	    {
-	      if (no_wrapping && user_insert_end)
-		fprintf (output_file, "%s", user_insert_end);
-#if HAVE_TPUTS
-	      if (term_insert_end)
-		tputs (term_insert_end, 0, putc_for_tputs);
-#endif
-	      putc ('\n', output_file);
-#if HAVE_TPUTS
-	      if (term_insert_start)
-		tputs (term_insert_start, 0, putc_for_tputs);
-#endif
-	      if (no_wrapping && user_insert_start)
-		fprintf (output_file, "%s", user_insert_start);
-	    }
-	  else if (overstrike_for_less)
-	    {
-	      putc (side->character, output_file);
-	      putc ('\b', output_file);
-	      putc (side->character, output_file);
-	    }
-	  else
-	    putc (side->character, output_file);
-	  break;
-	}
+  /* Now side->character is '>' or EOF. */
+}
+
+/*------------------------------------------------.
+| Copy white space and markup from SIDE to FILE.  |
+`------------------------------------------------*/
+
+static void
+copy_whitespace (SIDE * side, FILE * file, int with_markup)
+{
+  if (interrupted)
+    longjmp (signal_label, 1);
+
+  while (isspace (side->character) ||
+	 (ignore_markup && side->character == '<'))
+    {
+      if (side->character != '<') outc (side->character, file);
+      else if (with_markup) copy_markup (side, file);
+      else skip_markup (side);
 
       /* Advance to next character.  */
 
@@ -486,40 +593,11 @@
   if (interrupted)
     longjmp (signal_label, 1);
 
-  while (side->character != EOF && !isspace (side->character))
+  while (side->character != EOF &&
+	 !(isspace (side->character) ||
+	  (ignore_markup && side->character == '<')))
     {
-
-      /* In printer mode, act according to copy_mode.  If copy_mode is not
-         COPY_NORMAL, we know that file is necessarily output_file.  */
-
-      if (overstrike)
-	switch (copy_mode)
-	  {
-	  case COPY_NORMAL:
-	    putc (side->character, file);
-	    break;
-
-	  case COPY_DELETED:
-	    putc ('_', output_file);
-
-	    /* Avoid underlining an underscore.  */
-
-	    if (side->character != '_')
-	      {
-		putc ('\b', output_file);
-		putc (side->character, output_file);
-	      }
-	    break;
-
-	  case COPY_INSERTED:
-	    putc (side->character, output_file);
-	    putc ('\b', output_file);
-	    putc (side->character, output_file);
-	    break;
-	  }
-      else
-	putc (side->character, file);
-
+      outc (side->character, file);
       side->character = getc (side->file);
     }
   side->position++;
@@ -819,11 +897,11 @@
 `----------------------------------------------*/
 
 static void
-copy_until_ordinal (SIDE * side, int ordinal)
+copy_until_ordinal (SIDE * side, int ordinal, int with_markup)
 {
   while (side->position < ordinal)
     {
-      copy_whitespace (side, output_file);
+      copy_whitespace (side, output_file, with_markup);
       copy_word (side, output_file);
     }
 }
@@ -910,7 +988,7 @@
 	  if (!inhibit_left)
 	    {
 	      if (!inhibit_common && inhibit_right)
-		copy_until_ordinal (left_side, resync_left);
+		copy_until_ordinal (left_side, resync_left, 1);
 	      else
 		skip_until_ordinal (left_side, resync_left);
 	    }
@@ -920,11 +998,11 @@
 	      if (inhibit_common)
 		skip_until_ordinal (right_side, resync_right);
 	      else
-		copy_until_ordinal (right_side, resync_right);
+		copy_until_ordinal (right_side, resync_right, 1);
 	    }
 
 	  if (!inhibit_common && inhibit_left && inhibit_right)
-	    copy_until_ordinal (right_side, resync_right);
+	    copy_until_ordinal (right_side, resync_right, 0);
 
 	  /* Use separator lines to disambiguate the output.  */
 
@@ -940,10 +1018,18 @@
 
 	  if ((directive == 'd' || directive == 'c') && !inhibit_left)
 	    {
-	      copy_whitespace (left_side, output_file);
+	      copy_whitespace (left_side, output_file,
+			       inhibit_right && !inhibit_common);
 	      start_of_delete ();
 	      copy_word (left_side, output_file);
-	      copy_until_ordinal (left_side, argument[1]);
+	      while (left_side->position < argument[1])
+		{
+		  if (ignore_markup && inhibit_right) end_of_delete ();
+		  copy_whitespace (left_side, output_file,
+				   inhibit_right && !inhibit_common);
+		  if (ignore_markup && inhibit_right) start_of_delete ();
+		  copy_word (left_side, output_file);
+		}
 	      end_of_delete ();
 	    }
 
@@ -951,6 +1037,7 @@
 	     right file is used merely to show common words.  */
 
 	  if (directive == 'a' || directive == 'c')
+	    {
 	    if (inhibit_right)
 	      {
 		if (!inhibit_common && inhibit_left)
@@ -958,12 +1045,19 @@
 	      }
 	    else
 	      {
-		copy_whitespace (right_side, output_file);
+		copy_whitespace (right_side, output_file, !inhibit_common);
 		start_of_insert ();
 		copy_word (right_side, output_file);
-		copy_until_ordinal (right_side, argument[3]);
+		while (right_side->position < argument[3])
+		  {
+		    if (ignore_markup && !inhibit_common) end_of_insert ();
+		    copy_whitespace (right_side, output_file, !inhibit_common);
+		    if (ignore_markup && !inhibit_common) start_of_insert ();
+		    copy_word (right_side, output_file);
+		  }
 		end_of_insert ();
 	      }
+	    }
 	}
     }
 
@@ -980,13 +1074,13 @@
     }
   else if (!inhibit_left && inhibit_right)
     {
-      copy_until_ordinal (left_side, count_total_left);
-      copy_whitespace (left_side, output_file);
+      copy_until_ordinal (left_side, count_total_left, 1);
+      copy_whitespace (left_side, output_file, 1);
     }
   else
     {
-      copy_until_ordinal (right_side, count_total_right);
-      copy_whitespace (right_side, output_file);
+      copy_until_ordinal (right_side, count_total_right, !inhibit_right);
+      copy_whitespace (right_side, output_file, !inhibit_right);
     }
 
   /* Close input files.  */
@@ -1256,6 +1350,7 @@
       fputs (_("  -h, --help                 display this help then exit\n"), stdout);
       fputs (_("  -i, --ignore-case          fold character case while comparing\n"), stdout);
       fputs (_("  -l, --less-mode            variation of printer mode for \"less\"\n"), stdout);
+      fputs (_("  -m, --ignore-markup        skip XML or HTML tags\n"), stdout);
       fputs (_("  -n, --avoid-wraps          do not extend fields through newlines\n"), stdout);
       fputs (_("  -p, --printer              overstrike as for printers\n"), stdout);
       fputs (_("  -s, --statistics           say how many words deleted, inserted etc.\n"), stdout);
@@ -1294,6 +1389,7 @@
 
   diff_input = 0;
   ignore_case = 0;
+  ignore_markup = 0;
   show_statistics = 0;
   no_wrapping = 0;
   autopager = 0;
@@ -1319,7 +1415,7 @@
   count_changed_right = 0;
 
   while (option_char = getopt_long (argc, (char **) argv,
-				    "123CKadhilnpstvw:x:y:z:", longopts,
+				    "123CKadhilmnpstvw:x:y:z:", longopts,
 				    NULL), option_char != EOF)
     switch (option_char)
       {
@@ -1361,6 +1457,10 @@
 	overstrike_for_less = 1;
 	break;
 
+      case 'm':
+	ignore_markup = 1;
+	break;
+
       case 'n':
 	no_wrapping = 1;
 	break;
diff -ruN -x '*.xml' -x AUTHORS -x BACKLOG -x ABOUT-NLS -x alloca.h -x flags.texi -x atconfig -x '*.gmo' -x config.h -x requests -x version.texi -x compile -x depcomp -x missing -x configure -x stamp-vti -x '*.template' -x POTFILES -x stamp-h1 -x testsuite -x '*.Po' -x INSTALL -x Makefile -x '*-sh' -x '*.m4' -x '*.0' -x '*.2' -x '*.3' -x '*.4' -x '*.guess' -x '*.rpath' -x '*.sub' -x '*.tex' -x '*.status' -x '*~' -x '*.in' -x '*.po' -x '*.pot' -x '*.info' -x '*.1' -x '*.log' wdiff-1.2.2/tests/wdiff.at wdiff-1.2.2.htmldiff/tests/wdiff.at
--- wdiff-1.2.2/tests/wdiff.at	2013-10-21 11:28:04.000000000 +0200
+++ wdiff-1.2.2.htmldiff/tests/wdiff.at	2020-03-04 16:56:21.057509319 +0100
@@ -238,3 +238,47 @@
 ], [])
 
 AT_CLEANUP()
+
+AT_SETUP(wdiff with markup)
+dnl      -----------------
+
+AT_DATA(wdiff-x1.xml,
+[<doc>
+ The <b title="<QUICK>">quick</b> brown fox
+ <em>jumps</em> over the lazy dog.
+</doc>
+])
+
+AT_DATA(wdiff-x2.xml,
+[<doc>
+The quick brown fox <strong>jumps</strong> over a <i>lady</i> dog.
+</doc>
+])
+
+AT_CHECK([wdiff -m wdiff-x1.xml wdiff-x2.xml], 1,
+[<doc>
+The quick brown fox <strong>jumps</strong> over <del>the lazy</del> <ins>a</ins> <i><ins>lady</ins></i> dog.
+</doc>
+])
+
+AT_CHECK([wdiff -m -1 wdiff-x1.xml wdiff-x2.xml], 1,
+[<doc>
+The quick brown fox <strong>jumps</strong> over <ins>a</ins> <i><ins>lady</ins></i> dog.
+</doc>
+])
+
+AT_CHECK([wdiff -m -2 wdiff-x1.xml wdiff-x2.xml], 1,
+[<doc>
+ The <b title="<QUICK>">quick</b> brown fox
+ <em>jumps</em> over <del>the</del> <del>lazy</del> dog.
+</doc>
+])
+
+AT_CHECK([wdiff -m -3 wdiff-x1.xml wdiff-x2.xml], 1,
+[
+======================================================================
+ <del>the lazy</del> <ins>a lady</ins>
+======================================================================
+])
+
+AT_CLEANUP()