[PATCH] gnun-validate-html may break on ASCII art

[Ineiev]

Hi,

gnun-validate-html expands the SSIs substituting the include
with m4_include and processing the file with m4.

The intermediate file may confuse m4 when it contains ASCII art,
like /graphics/alternative-ascii.html.

It is suggested that the SSIs may be expanded via recursive
awk script; by the way, configuration time variables for
awk and sed programs are used like the variable for m4 in the
current version of gnun-validate-html is used.

Any ideas?

Index: gnun-validate-html.in
===================================================================
RCS file: 
/sources/trans-coord/trans-coord/gnun/server/gnun/gnun-validate-html.in,v
retrieving revision 1.9
diff -U 2 -r1.9 gnun-validate-html.in
--- gnun-validate-html.in       7 Jan 2010 14:43:04 -0000       1.9
+++ gnun-validate-html.in       18 Aug 2011 15:22:38 -0000
@@ -1,5 +1,5 @@
 #!/bin/bash
 
-# Copyright (C) 2008, 2009, 2010 Free Software Foundation, Inc.
+# Copyright (C) 2008, 2009, 2010, 2011 Free Software Foundation, Inc.
 
 # This file is part of GNUnited Nations.
@@ -21,5 +21,5 @@
 cat <<EOF
 gnun-validate-html (@PACKAGE_NAME@) @PACKAGE_VERSION@
-Copyright (C) 2010 Free Software Foundation, Inc.
+Copyright (C) 2011 Free Software Foundation, Inc.
 You may redistribute copies of @PACKAGE_NAME@
 under the terms of the GNU General Public License.
@@ -72,15 +72,33 @@
 trap "rm -f $TMP1 $TMP2 $TMP3" EXIT
 
+cat > $TMP3 <<"EOF"
+/<!--#include virtual/ {
+  n = split ($0, names, "<!--#include virtual=\"");
+  printf ("%s", names[1]);
+  for (i = 2; i <= n; i++)
+    {
+      m = split (names[i], rest, "-->");
+      name = rest[1];
+      gsub ("\"[^\"]*$", "", name);
+      gsub ("^", root, name);
+      system ("@AWK@ -v script_name='" script_name "' -v root='" \
+               root "' -f '" script_name "' " name);
+      for (j = 2; j <= m; j++) 
+        {
+          printf ("%s", rest[j]);
+          if (j < m)
+            print ("-->");
+        }
+    }
+  next;
+}
+
+{ print; }
+EOF
+
 # Expand input file's #include directives and save the result in
 # $TMP1.
-cat $1 > $TMP1
 
-while true; do
-    grep --quiet '<!--#include virtual' $TMP1 || break
-    sed --in-place \
-      "s/<\!--#include virtual=\"\/\?\(.*\)\" -->/m4_include(\`\1')/g" $TMP1
-    @M4@ -P -EE -I $ROOT $TMP1 > $TMP3
-    cp $TMP3 $TMP1
-done
address@hidden@ -v script_name=$TMP3 -v root="$ROOT" -f $TMP3 $1 > $TMP1
 
 # Execute xmllint on $TMP1 and save its output to $TMP2.
@@ -98,5 +116,5 @@
 # expanded #include directives) and the translator can not easily look
 # up for references in it.
-cat $TMP2 | sed '
+cat $TMP2 | @SED@ '
   /line [[:digit:]]\+/ {
      p
Index: Makefile.am
===================================================================
RCS file: /sources/trans-coord/trans-coord/gnun/server/gnun/Makefile.am,v
retrieving revision 1.17
diff -U 2 -r1.17 Makefile.am
--- Makefile.am 2 Aug 2010 13:54:11 -0000       1.17
+++ Makefile.am 18 Aug 2011 15:22:38 -0000
@@ -1,3 +1,3 @@
-# Copyright (C) 2008, 2009, 2010 Free Software Foundation, Inc.
+# Copyright (C) 2008, 2009, 2010, 2011 Free Software Foundation, Inc.
 
 # This file is part of GNUnited Nations.
@@ -35,4 +35,5 @@
 
 noinst_SCRIPTS = stamp-config.mk
+pkglibexecdir = $(datadir)/libexec
 pkglibexec_SCRIPTS = validate-html-notify
 
@@ -52,5 +53,8 @@
        -e 's|@address@hidden|$(prefix)|g' \
        -e 's|@address@hidden|$(pkglibexecdir)|g' \
-       -e 's|@address@hidden|$(M4)|g'
+       -e 's|@address@hidden|$(AWK)|g' \
+       -e 's|@address@hidden|$(GREP)|g' \
+       -e 's|@address@hidden|$(M4)|g' \
+       -e 's|@address@hidden|$(SED)|g'
 
 gnun-validate-html validate-html-notify: Makefile