diff options
| -rwxr-xr-x | htdocs/docs/install.html | 6 | ||||
| -rwxr-xr-x | httpd/cgi-bin/check | 172 | ||||
| -rw-r--r-- | misc/bundle/META.yml | 2 | ||||
| -rw-r--r-- | misc/bundle/Makefile.PL | 2 | ||||
| -rw-r--r-- | misc/bundle/lib/Bundle/W3C/Validator.pm | 2 |
5 files changed, 29 insertions, 155 deletions
diff --git a/htdocs/docs/install.html b/htdocs/docs/install.html index 48a91da..75397c7 100755 --- a/htdocs/docs/install.html +++ b/htdocs/docs/install.html @@ -1,5 +1,5 @@ -<!--#set var="revision" value="\$Id: install.html,v 1.59 2010-03-03 19:13:29 ville Exp $" ---><!--#set var="date" value="\$Date: 2010-03-03 19:13:29 $" +<!--#set var="revision" value="\$Id: install.html,v 1.60 2010-03-03 19:40:59 ville Exp $" +--><!--#set var="date" value="\$Date: 2010-03-03 19:40:59 $" --><!--#set var="title" value="Installation Documentation for The W3C Markup Validation Service" --><!--#set var="relroot" value="../" --><!--#include virtual="../header.html" --> @@ -180,7 +180,7 @@ install Bundle::W3C::Validator <dd> Library to handle URIs and escaping special characters in them. </dd> - <dt><a href="http://search.cpan.org/dist/XML-LibXML/">XML-LibXML</a> >= 1.56</dt> + <dt><a href="http://search.cpan.org/dist/XML-LibXML/">XML-LibXML</a> >= 1.69</dt> <dd> The Perl binding for libxml2, used to check the syntax of XML-based document types. </dd> diff --git a/httpd/cgi-bin/check b/httpd/cgi-bin/check index ddac822..a65f77a 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 1.768 2010-03-03 19:13:29 ville Exp $ +# $Id: check,v 1.769 2010-03-03 19:40:59 ville Exp $ # # We need Perl 5.8.0+. @@ -61,7 +61,7 @@ use JSON 2.00 qw(); use SGML::Parser::OpenSP 0.991 qw(); use URI qw(); use URI::Escape qw(uri_escape); -use XML::LibXML 1.56 qw(); # Need 1.56 for line_numbers() +use XML::LibXML 1.69 qw(); # Need 1.69 for (working) structured errors ############################################################################### #### Constant definitions. #################################################### @@ -191,7 +191,7 @@ EOF # # Strings - $VERSION = q$Revision: 1.768 $; + $VERSION = q$Revision: 1.769 $; $VERSION =~ s/Revision: ([\d\.]+) /$1/; # Read friendly error message file @@ -634,152 +634,33 @@ if (&is_xml($File)) { eval { $xmlparser->parse_string($xml_string); }; $xml_string = undef; - my $xml_parse_errors_line = undef; - my @xmlwf_error_list; - if (ref($@)) { + my $err_obj = $@; + while ($err_obj) { + my $err; + $err->{src} = '...'; # do this with show_open_entities()? + $err->{line} = $err_obj->line(); - # handle a structured error (XML::LibXML::Error object) - # (lib XML::LibXML > 1.66, but will work MUCH better with > 1.69) + # -> column() is available in XML::LibXML >= 1.69_2 + $err->{char} = eval { $err_obj->column() }; + $err->{num} = "libxml2-" . $err_obj->code(); + $err->{type} = "E"; + $err->{msg} = $err_obj->message(); - my $err_obj = $@; - my $num_xmlwf_error = 0; - while ($err_obj) { - my $err; - $err->{src} = '...'; # do this with show_open_entities()? - $err->{line} = $err_obj->line(); + $err_obj = $err_obj->_prev(); - # -> column() is available in XML::LibXML >= 1.69_2 - $err->{char} = eval { $err_obj->column() }; - $err->{num} = "libxml2-" . $err_obj->code(); - $err->{type} = "E"; - $err->{msg} = $err_obj->message(); - - $err_obj = $err_obj->_prev(); - - # The validator will sometimes fail to dereference entities - # files; we're filtering the resulting bogus error for - # non-standalone documents. @@@TODO: is this still needed? - if (!$standalone && - $err->{msg} =~ /Entity '\w+' not defined/) - { - $err = undef; - next; - } - - unshift(@xmlwf_error_list, $err); - $num_xmlwf_error++; + # The validator will sometimes fail to dereference entities + # files; we're filtering the resulting bogus error for + # non-standalone documents. @@@TODO: is this still needed? + if (!$standalone && + $err->{msg} =~ /Entity '\w+' not defined/) + { + $err = undef; + next; } - } - elsif ($@) { - my $xmlwf_errors = $@; - my $xmlwf_error_line = undef; - my $xmlwf_error_col = undef; - my $xmlwf_error_msg = undef; - my $got_error_message = undef; - my $got_quoted_line = undef; - my $num_xmlwf_error = 0; - foreach my $msg_line (split "\n", $xmlwf_errors) { - - $msg_line =~ s{[^\x0d\x0a](:\d+:)}{\n$1}g; - $msg_line =~ s{[^\x0d\x0a]+[\x0d\x0a]$}{}; - - # first we get the actual error message - if (!$got_error_message && - $msg_line =~ /^(:\d+:)( parser error : .*)/) - { - $xmlwf_error_line = $1; - $xmlwf_error_msg = $2; - $xmlwf_error_line =~ s/:(\d+):/$1/; - $xmlwf_error_msg =~ s/ parser error :/XML Parsing Error: /; - $got_error_message = 1; - } - - # then we skip the second line, which shows the context - # (we don't use that) - elsif ($got_error_message && !$got_quoted_line) { - $got_quoted_line = 1; - } - # we now take the third line, with the pointer to the error's - # column - elsif (($msg_line =~ /(\s+)\^/) and - $got_error_message and - $got_quoted_line) - { - $xmlwf_error_col = length($1); - } - - # cleanup for a number of bugs for the column number - if (defined($xmlwf_error_col)) { - if (( my $l = - length($File->{Content}->[$xmlwf_error_line - 1]) - ) < $xmlwf_error_col - ) - { - - # http://bugzilla.gnome.org/show_bug.cgi?id=434196 - #warn("Warning: reported error column larger than line length " . - # "($xmlwf_error_col > $l) in $File->{URI} line " . - # "$xmlwf_error_line, libxml2 bug? Resetting to line length."); - $xmlwf_error_col = $l; - } - elsif ($xmlwf_error_col == 79) { - - # working around an apparent odd limitation of libxml - # which only gives context for lines up to 80 chars - # http://www.w3.org/Bugs/Public/show_bug.cgi?id=4420 - # http://bugzilla.gnome.org/show_bug.cgi?id=424017 - $xmlwf_error_col = "> 80"; - - # non-int line number will trigger the proper behavior - # in report_error - } - } - - # when we have all the info (one full error message), proceed - # and move on to the next error - if ((defined $xmlwf_error_line) and - (defined $xmlwf_error_col) and - (defined $xmlwf_error_msg)) - { - - # Reinitializing for the next batch of 3 lines - $got_error_message = undef; - $got_quoted_line = undef; - - # formatting the error message for output - my $err; - $err->{src} = '...'; # do this with show_open_entities()? - $err->{line} = $xmlwf_error_line; - $err->{char} = $xmlwf_error_col; - $err->{num} = 'xmlwf'; - $err->{type} = "E"; - $err->{msg} = $xmlwf_error_msg; - - # The validator will sometimes fail to dereference entities - # files; we're filtering the resulting bogus error for - # non-standalone documents. @@@TODO: is this still needed? - if (!$standalone && - $err->{msg} =~ /Entity '\w+' not defined/) - { - $xmlwf_error_line = undef; - $xmlwf_error_col = undef; - $xmlwf_error_msg = undef; - next; - } - push(@xmlwf_error_list, $err); - $xmlwf_error_line = undef; - $xmlwf_error_col = undef; - $xmlwf_error_msg = undef; - $num_xmlwf_error++; - - } - } - } - foreach my $errmsg (@xmlwf_error_list) { $File->{'Is Valid'} = FALSE; - push @{$File->{WF_Errors}}, $errmsg; + unshift(@{$File->{WF_Errors}}, $err); } } } @@ -970,13 +851,6 @@ if ($File->{Opt}->{Output} eq 'json') { for my $key (qw(msg expl)) { $msg->{$key} = $json->encode($msg->{$key}) if $msg->{$key}; } - - # Drop non-numeric char indicators from output, e.g. - # "> 80" for some XML parse error ones (see the non-structured - # XML::LibXML code branch in XML preparsing below). - if ($msg->{char} && $msg->{char} !~ /^\d+$/) { - delete($msg->{char}); - } } } } diff --git a/misc/bundle/META.yml b/misc/bundle/META.yml index af3782e..632c198 100644 --- a/misc/bundle/META.yml +++ b/misc/bundle/META.yml @@ -34,7 +34,7 @@ requires: Socket: 0 URI: 0 URI::Escape: 0 - XML::LibXML: 1.56 + XML::LibXML: 1.69 meta-spec: url: http://module-build.sourceforge.net/META-spec-v1.3.html version: 1.3 diff --git a/misc/bundle/Makefile.PL b/misc/bundle/Makefile.PL index 311b1a1..279c432 100644 --- a/misc/bundle/Makefile.PL +++ b/misc/bundle/Makefile.PL @@ -37,7 +37,7 @@ WriteMakefile( "Socket" => 0, "URI" => 0, "URI::Escape" => 0, - "XML::LibXML" => 1.56, + "XML::LibXML" => 1.69, # Optional: "Encode::JIS2K" => 0, diff --git a/misc/bundle/lib/Bundle/W3C/Validator.pm b/misc/bundle/lib/Bundle/W3C/Validator.pm index 43163cb..ac096de 100644 --- a/misc/bundle/lib/Bundle/W3C/Validator.pm +++ b/misc/bundle/lib/Bundle/W3C/Validator.pm @@ -47,7 +47,7 @@ C<perl -MCPAN -e "install Bundle::W3C::Validator"> Socket URI URI::Escape - XML::LibXML 1.56 + XML::LibXML 1.69 =head1 DESCRIPTION |