for documents served as XML, and for which the root element is not <html>,

the fallback to XHTML is not a great choice. Instead, we can perform a basic xml-wf checking, and report a positive result for these.
author: ot <ot@localhost> 2007-05-07 19:17:58 +0000
committer: ot <ot@localhost> 2007-05-07 19:17:58 +0000
commit: 86eef21345edbca7f8ef532e91d2117c616f6934 (patch)
tree: 2576b8668203caf792ecc305b3774f34998e09cb
parent: ff8ed778f0a58b141cedc75bb802b2a42d1105d4 (diff)
download: markup-validator-86eef21345edbca7f8ef532e91d2117c616f6934.zip
markup-validator-86eef21345edbca7f8ef532e91d2117c616f6934.tar.gz
markup-validator-86eef21345edbca7f8ef532e91d2117c616f6934.tar.bz2
4 files changed, 42 insertions, 16 deletions
diff --git a/htdocs/dev/tests/xhtml1-missing-doctype-and-xmlns.xhtml b/htdocs/dev/tests/xhtml1-missing-doctype-and-xmlns.xhtml
index 1381810..529cfe1 100755
--- a/htdocs/dev/tests/xhtml1-missing-doctype-and-xmlns.xhtml
+++ b/htdocs/dev/tests/xhtml1-missing-doctype-and-xmlns.xhtml
@@ -6,5 +6,6 @@
     <p>This is a simple bogus HTML doc with neither a doctype nor
     an xmlns attribute on its root element.
     </p>
+<img />
   </body>
 </html>
diff --git a/httpd/cgi-bin/check b/httpd/cgi-bin/check
index 67abc59..90b1d60 100755
--- a/httpd/cgi-bin/check
+++ b/httpd/cgi-bin/check
@@ -14,7 +14,7 @@
 # This source code is available under the license at:
 #     http://www.w3.org/Consortium/Legal/copyright-software
 #
-# $Id: check,v 1.517 2007-05-07 16:14:15 ot Exp $
+# $Id: check,v 1.518 2007-05-07 19:17:58 ot Exp $
 
 #
 # Disable buffering on STDOUT!
@@ -180,7 +180,7 @@ Directory not readable (permission denied): @_r
 
   #
   # Strings
-  $VERSION =  q$Revision: 1.517 $;
+  $VERSION =  q$Revision: 1.518 $;
   $VERSION =~ s/Revision: ([\d\.]+) /$1/;
 
   #
@@ -254,6 +254,8 @@ $File->{Charset}->{Override} = ''; # From CGI/user override.
 # Misc simple types.
 $File->{Mode} = 'SGML'; # Default parse mode is SGML.
 
+# By default, perform validation (we may perform only xml-wf in some cases)
+$File->{XMLWF_ONLY} = FALSE;
 #
 # Listrefs.
 $File->{Warnings}   = []; # Warnings...
@@ -588,8 +590,7 @@ if ($File->{Charset}->{Use} eq 'utf-8' &&
 #
 # Override DOCTYPE if user asked for it.
 if ($File->{Opt}->{DOCTYPE}
-    and not $File->{Opt}->{DOCTYPE} =~ /(Inline|detect)/i
-    and $File->{Opt}->{DOCTYPE} ne '1' ) {
+    and not $File->{Opt}->{DOCTYPE} =~ /(Inline|detect)/i) {
   $File = &override_doctype($File);
 }
 
@@ -781,7 +782,7 @@ sub parse (\$) {
 #
 # Force "XML" if type is an XML type and an FPI was not found.
 # Otherwise set the type to be the FPI.
-if (&is_xml($File) and not $File->{DOCTYPE}) {
+if (&is_xml($File) and not $File->{DOCTYPE} and  lc($File->{Root}) ne 'html') {
   $File->{Version} = 'XML';
 } else {
   $File->{Version} = $File->{DOCTYPE} unless $File->{Version};
@@ -962,7 +963,9 @@ sub prep_template ($$) {
   $T->param(tip_uri  => $tip->[0]);
   $T->param(tip_slug => $tip->[1]);
 
-  #
+  # Root Element
+  $T->param(root_element  => $File->{Root});
+  
   # Namespaces...
   $T->param(file_namespace  => $File->{Namespace});
   my %seen_ns = ();
@@ -1102,7 +1105,9 @@ sub report_valid {
   if ($File->{Opt}->{'Outline'}) {
     $T->param(file_outline => $File->{heading_outline});
   }
-
+  if ($File->{XMLWF_ONLY}){
+    $T->param(xmlwf_only => TRUE);
+  }
   my $thispage = self_url_file($File);
   $T->param(file_thispage => $thispage);
 }
@@ -2358,9 +2363,17 @@ sub W3C::Validator::SAXHandler::error
     }
   }
 
+  if ($err->{num} eq '187')
+  # filtering out no "document type declaration; will parse without validation"
+  # if root element is not html and mode is xml...
+  {
+    # since parsing was done without validation, result can only be "well-formed"
+    if ($File->{Mode} eq 'XML' and lc($File->{Root}) ne 'html') {
+      $File->{XMLWF_ONLY} = TRUE;
+    }
+    return;
+  }
 
-  # ...
-  $File->{'Is Valid'} = FALSE if $err->{type} eq 'E';
 
   # Workaround for onsgmls as of 1.5 sometimes allegedly reporting errors
   # beyond EOL.  If you see this warning in your web server logs, please
@@ -2379,12 +2392,14 @@ sub W3C::Validator::SAXHandler::error
     $File->{Templates}->{Error}->param(fatal_parse_extid_msg   => $err->{msg});
   }
 
-  # No DOCTYPE.
+  # No DOCTYPE found! We are falling back to vanilla DTD
   if ($err->{msg} =~ m(prolog can\'t be omitted)) {
-    my $dtd = ($File->{Mode} eq 'XML' ?
-               'XHTML 1.0 Transitional' : 'HTML 4.01 Transitional' );
-
-    add_warning('W09', {W09_dtd => $dtd});
+    if (lc($File->{Root}) eq 'html') { 
+      my $dtd = ($File->{Mode} eq 'XML' ?
+                 'XHTML 1.0 Transitional' : 'HTML 4.01 Transitional' );      
+      add_warning('W09', {W09_dtd => $dtd});
+    }
+    
     return; # Don't report this as a normal error.
   }
 
@@ -2392,6 +2407,8 @@ sub W3C::Validator::SAXHandler::error
   abort_if_error_flagged($File, O_DOCTYPE);
 
   push @{$File->{Errors}}, $err;
+  # ...
+  $File->{'Is Valid'} = FALSE if $err->{type} eq 'E';
 
   if (defined $mess->{aux_message})
   {
diff --git a/share/templates/en_US/table.tmpl b/share/templates/en_US/table.tmpl
index 2f92125..c581233 100644
--- a/share/templates/en_US/table.tmpl
+++ b/share/templates/en_US/table.tmpl
@@ -3,7 +3,8 @@
     <h2 class="tentative_valid">This Page Is <em>Tentatively</em> Valid 
     <TMPL_VAR NAME="file_version" ESCAPE="HTML"></h2>
   <TMPL_ELSE>
-    <h2 class="valid">This Page Is Valid <TMPL_VAR NAME="file_version" ESCAPE="HTML">!</h2>
+    <h2 class="valid">This Page Is <TMPL_IF NAME="xmlwf_only">well-formed<TMPL_ELSE>Valid</TMPL_IF>
+     <TMPL_VAR NAME="file_version" ESCAPE="HTML">!</h2>
   </TMPL_IF>
 <TMPL_ELSE>
   <TMPL_IF NAME="fatal_error">
@@ -73,6 +74,12 @@
     <tr><th>Doctype:</th><td colspan="2"><TMPL_INCLUDE NAME="popup_doctype.tmpl"></td></tr>
   </TMPL_IF>
 </TMPL_IF>
+<TMPL_IF NAME="root_element">
+  <tr>
+    <th>Root Element:</th>
+    <td colspan="2"><TMPL_VAR NAME="root_element" ESCAPE="HTML"></td>
+  </tr>
+</TMPL_IF>
 <TMPL_IF NAME="file_namespace">
     <tr>
       <th>Root Namespace:</th>
diff --git a/share/templates/en_US/valid.tmpl b/share/templates/en_US/valid.tmpl
index f92c268..e683f49 100644
--- a/share/templates/en_US/valid.tmpl
+++ b/share/templates/en_US/valid.tmpl
@@ -34,7 +34,8 @@
     to match the changes we have performed automatically, but 
     <strong>it will not be valid until you make these changes</strong>.
 <TMPL_ELSE>
-    was checked and found to be valid
+    was checked and found to be 
+    <TMPL_IF NAME="xmlwf_only">well-formed<TMPL_ELSE>valid</TMPL_IF>
     <TMPL_VAR NAME="file_version" ESCAPE="HTML">.  This means that the
     resource in question identified itself as
     "<TMPL_VAR NAME="file_version" ESCAPE="HTML">" and that we successfully
author	ot <ot@localhost>	2007-05-07 19:17:58 +0000
committer	ot <ot@localhost>	2007-05-07 19:17:58 +0000
commit	86eef21345edbca7f8ef532e91d2117c616f6934 (patch)
tree	2576b8668203caf792ecc305b3774f34998e09cb
parent	ff8ed778f0a58b141cedc75bb802b2a42d1105d4 (diff)
download	markup-validator-86eef21345edbca7f8ef532e91d2117c616f6934.zip markup-validator-86eef21345edbca7f8ef532e91d2117c616f6934.tar.gz markup-validator-86eef21345edbca7f8ef532e91d2117c616f6934.tar.bz2