diff options
| -rwxr-xr-x | httpd/cgi-bin/check | 54 |
1 files changed, 37 insertions, 17 deletions
diff --git a/httpd/cgi-bin/check b/httpd/cgi-bin/check index 69716ed..911c193 100755 --- a/httpd/cgi-bin/check +++ b/httpd/cgi-bin/check @@ -8,7 +8,7 @@ # This source code is available under the license at: # http://www.w3.org/Consortium/Legal/copyright-software # -# $Id: check,v 1.74 2000-10-21 11:40:45 gerald Exp $ +# $Id: check,v 1.75 2000-12-11 11:16:33 gerald Exp $ # # We need Perl 5.004. @@ -44,6 +44,9 @@ use vars qw($frag $pub_ids $element_uri $file_type); # Cfg hashes. # Paths and file locations my $logfile = '/var/log/httpd/val-svc'; my $base_path = '/usr/local/validator/'; +if ( $ENV{SERVER_PORT} eq "8000" ) { + $base_path = '/home/gerald/validator/'; +} my $html_path = $base_path . 'htdocs/'; my $elem_db = $html_path . 'config/eref.cfg'; my $fpis_db = $html_path . 'config/fpis.cfg'; @@ -71,9 +74,9 @@ my $element_ref = 'http://www.htmlhelp.com/reference/html40/'; # # Strings -$VERSION = q$Revision: 1.74 $; +$VERSION = q$Revision: 1.75 $; $VERSION =~ s/Revision: ([\d\.]+) /$1/; -$DATE = q$Date: 2000-10-21 11:40:45 $; +$DATE = q$Date: 2000-12-11 11:16:33 $; $MAINTAINER = 'gerald@w3.org'; my $notice = ''; # "<p><strong>Note: This service will be ...</strong>"; @@ -89,7 +92,7 @@ my $xhtmlt_doctype = q(<!DOCTYPE HTML PUBLIC "-//W3C//DTD XHTML 1.0 Transitiona # HTML fragments my $lt = "\020"; my $gt = "\021"; -my $gifborder = ' border=0'; +my $gifborder = ' border="0"'; my $xhtmlendtag; # @@ -100,7 +103,7 @@ my @options = qw(weblint pw outline ss sp noatt); # Stopgap to shut -w up. It won't actually fix anything, but it'll keep us # running without warnings until we can fix the problems. my ($validity, %undef_frag, $effective_charset, $catalog, - @fake_errors, $guessed_doctype, $doctype, $line, $col, $type, $msg, $diff, + @fake_errors, $doctype, $line, $col, $type, $msg, $diff, $pos, $indent, $image_uri, $alttext, $gifhw, $pedanticflags, $pedantic_blurb, $level, $prevlevel, $prevdata); @@ -192,7 +195,7 @@ $html40t_doctype <body bgcolor="#FFFFFF" text="#000000" link="#0000ee" vlink="#551a8b"> <h1><a href="http://www.w3.org/"><img src="http://www.w3.org/Icons/WWW/w3c_home" - width=72 height=48 border=0 alt="W3C"></a> + width=72 height=48 border="0" alt="W3C"></a> <a href="/">HTML Validation Service</a> Results</h1> EOF @@ -254,6 +257,7 @@ EOF # inferred by the check_for_doctype function # +my $guessed_doctype = 2; # # Try to extract or guess the DOCTYPE for HTML and XHTML files. if ($File->{Type} eq 'html' or $File->{Type} eq 'xhtml' @@ -269,6 +273,12 @@ if ($doctype =~ /xhtml/i && $guessed_doctype != 2) { } # +# Set document type to MathML if the DOCTYPE was for MathML. +if ($doctype =~ /mathml/i && $guessed_doctype != 2) { + $File->{Type} = 'mathml'; +} + +# # If we find a META element with charset information, we take it into account. foreach my $line (@{$File->{Content}}) { # @@ needs to handle meta elements that span more than one line @@ -347,6 +357,11 @@ if ($File->{Type} eq 'xhtml') { $xmlflags .= '-wno-valid '; } $decl = $xhtmldecl; +} elsif ($File->{Type} eq 'mathml') { + $catalog = "-c $sgmlstuff/CR-MathML2-20001113/catalog"; + $ENV{SP_CHARSET_FIXED} = 'YES'; + $ENV{SP_ENCODING} = 'XML'; + $xmlflags = '-wxml '; } elsif ($File->{Type} eq 'xml' or $guessed_doctype == 2) { # no doctype, with xmlns attr on 1st element $File->{Type} = 'xml'; # @@ probably a better way to do this @@ -363,7 +378,7 @@ if ($File->{Type} eq 'xhtml') { my $command = "$codeconv $sp -E0 $xmlflags $catalog $decl"; -if ( $ENV{SERVER_PORT} eq "8000" ) { +if ( $ENV{SERVER_PORT} eq "8000" ) { # running on the test server? print " <li>nsgmls command line: <code>$command</code>\n"; } @@ -406,7 +421,7 @@ close ESIS or warn "close($temp.esis) returned: $!"; my $fpi; my $version = 'unknown'; -if ($File->{Type} eq 'xhtml') { +if (($File->{Type} eq 'xhtml') || ($File->{Type} eq 'mathml')) { $fpi = $doctype; } elsif ($File->{Type} eq 'xml') { $fpi = 'XML'; @@ -501,6 +516,8 @@ if ( $? || ($guessed_doctype == 1) ) { for ((@fake_errors,@errors)) { next if /^<OSFD>0:[0-9]+:[0-9]+:[^A-Z]/; next if / numbers exceeding 65535 not supported$/; + # @@ could maybe replace the following statement with an SGMLDECL + # setting in the catalog, like mathml2? next if /:W: SGML declaration was not implied$/ && ($File->{Type} =~ /^x(ht)?ml$/); s/^$sp:<OSFD>//g; @@ -614,6 +631,9 @@ if ( $? || ($guessed_doctype == 1) ) { if ( $version eq "unknown" ) { print "\n <p>\n Sorry, I can't validate this document.\n </p>\n"; } + elsif ( $File->{Type} eq 'xml' ) { + print "\n <p>\n Sorry, this document is not well-formed XML.\n </p>\n"; + } else { print "\n <p>\n Sorry, this document does not validate as $version.\n </p>\n\n"; &output_css_validator_blurb( $q->param('uri') ); @@ -637,29 +657,29 @@ else { elsif ( $version =~ /HTML 3\.2</ ) { $image_uri = "http://www.w3.org/Icons/valid-html32"; $alttext = "Valid HTML 3.2!"; - $gifhw = " height=31 width=88"; + $gifhw = ' height="31" width="88"'; } elsif ( $version =~ /HTML 4\.0<\/a> Strict$/ ) { $image_uri = "http://www.w3.org/Icons/valid-html40"; $alttext = "Valid HTML 4.0!"; $gifborder = ""; - $gifhw = " height=31 width=88"; + $gifhw = ' height="31" width="88"'; } elsif ( $version =~ /HTML 4\.0<\/a> / ) { $image_uri = "http://www.w3.org/Icons/valid-html40"; $alttext = "Valid HTML 4.0!"; - $gifhw = " height=31 width=88"; + $gifhw = ' height="31" width="88"'; } elsif ( $version =~ /HTML 4\.01<\/a> Strict$/ ) { $image_uri = "http://www.w3.org/Icons/valid-html401"; $alttext = "Valid HTML 4.01!"; $gifborder = ""; - $gifhw = " height=31 width=88"; + $gifhw = ' height="31" width="88"'; } elsif ( $version =~ /HTML 4\.01<\/a> / ) { $image_uri = "http://www.w3.org/Icons/valid-html401"; $alttext = "Valid HTML 4.01!"; - $gifhw = " height=31 width=88"; + $gifhw = ' height="31" width="88"'; } elsif ( $version =~ /XHTML 1\.0<\/a> / ) { $image_uri = "http://www.w3.org/Icons/valid-xhtml10"; @@ -1013,8 +1033,8 @@ sub output_closing { <address> <a href="http://validator.w3.org/check/referer"><img - src="http://www.w3.org/Icons/valid-html401" height=31 width=88 - align=right border=0 alt="Valid HTML 4.01!"></a> + src="http://www.w3.org/Icons/valid-html401" height="31" width="88" + align=right border="0" alt="Valid HTML 4.01!"></a> <a href="/feedback.html">Gerald Oskoboiny</a><br> Last modified: $DATE </address> @@ -1136,7 +1156,7 @@ sub check_for_doctype { if ($line =~ /<!doctype/i) { # found a doctype my $dttext = join '', @{$file}[$count .. $count + 5]; $dttext =~ s/\n//g; - $dttext =~ s/.*?doctype\s+html\s+public\s*([\'\"])([^\1]+?)\1.*/$2/i; + $dttext =~ s/.*?doctype\s+(html|math)\s+public\s*([\'\"])([^\2]+?)\2.*/$3/i; return 0, $dttext; } @@ -1246,7 +1266,7 @@ sub output_css_validator_blurb { print <<"EOHD"; <p> - If you use <a href="http://www.w3.org/Style/css/">CSS</a> in your document, + If you use <a href="http://www.w3.org/Style/CSS/">CSS</a> in your document, you should also <a href="http://jigsaw.w3.org/css-validator/validator?uri=$uri">check it for validity</a> using the W3C <a |