Chris/webperl
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Adding more error reporting to html validator.

Browse Source
This commit is contained in:
Chris 2011-10-27 22:35:29 +01:00
parent eae1f3fffe
commit e028e5a258
2 changed files with 21 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
Block.pm
View File

@ -298,7 +298,8 @@ sub validate_htmlarea {
return ($tidied, undef);
# If the return from check_xhtml is one or more digits, it is an error count
} elsif($valid =~ /^\d+$/) {
} elsif($valid =~ /^\d+:/) {
$valid =~ s/^\d+://;
return ($tidied, $self -> {"template"} -> replace_langvar("BLOCK_VALIDATE_CHKERRS", "", {"***field***" => $settings -> {"nicename"},
"***error***" => $valid}));

33
HTMLValidator.pm
View File

@ -1,6 +1,6 @@
## @file
# HTML validation and checking functions. This file contains functions to
# support the cleaning and checking of html using a combination of
# support the cleaning and checking of html using a combination of
# HTML::Scrubber to do first-stage cleaning, HTML::Tidy to clear up the
# content as needed, and the W3C validator via the WebService::Validator::HTML::W3C
# to ensure that the xhtml generated by HTML::Tidy is valid.
@ -42,22 +42,22 @@ our $VERSION = 1.0;
# List of tags we are going to let through, lifted from the security
# discussion on http://wiki.moxiecode.com/index.php/TinyMCE:Security
# Several tags removed to make xhtml conformance easier and to remove
# Several tags removed to make xhtml conformance easier and to remove
# deprecated and eyestabbery.
my @allow = ("a", "b", "blockquote", "br", "caption", "col", "colgroup", "comment",
"em", "h1", "h2", "h3", "h4", "h5", "h6", "hr", "img", "li", "ol", "p",
"pre", "small", "span", "strong", "sub", "sup", "table", "tbody", "td",
my @allow = ("a", "b", "blockquote", "br", "caption", "col", "colgroup", "comment",
"em", "h1", "h2", "h3", "h4", "h5", "h6", "hr", "img", "li", "ol", "p",
"pre", "small", "span", "strong", "sub", "sup", "table", "tbody", "td",
"tfoot", "th", "thead", "tr", "tt", "ul");
# Explicit rules for allowed tags, required to provide per-tag tweaks to the filter.
my @rules = (
img => {
src => qr{^(?:http|https)://}i,
alt => 1,
alt => 1,
style => 1,
width => 1,
height => 1,
'*' => 0,
'*' => 0,
},
a => {
href => qr{^(?:http|https)://}i,
@ -131,7 +131,7 @@ sub scrub_html {
default => \@default,
comment => 0,
process => 0);
# fix problems with the parser setup. This is hacky and nasty,
# but from CPAN's bug tracker, this appears to have been present for
# the past 3 years at least.
@ -153,7 +153,7 @@ sub scrub_html {
# HTML::Tidy related code
## @fn $ tidy_html($html, $options)
# Pass a chunk of html through htmltidy. This should produce well-formed xhtml
# Pass a chunk of html through htmltidy. This should produce well-formed xhtml
# that can be passed on to the validator to check.
#
# @param html The string containing html to tidy.
@ -173,13 +173,13 @@ sub tidy_html {
# WebService::Validator::HTML::W3C related code
## @fn @ check_xhtml($xhtml, $options)
# Check that the xhtml is valid by passing it through the W3C validator service.
# If this is unable to contact the validation service, it will return the reason,
# Check that the xhtml is valid by passing it through the W3C validator service.
# If this is unable to contact the validation service, it will return the reason,
# otherwise the number of errors will be returned (0 indicates that the xhtml
# passed validation with no errors)
#
# @param xhtml The xhtml to validate with the W3C validator
# @param options A hash containing options to pass to the validator module.
# @param options A hash containing options to pass to the validator module.
# Currently supports 'timeout' and 'uri'.
# @return The number of errors during validation (0 = valid), or a string
# from the validator module explaining why the validation bombed.
@ -196,11 +196,16 @@ sub check_xhtml {
return 0
if($validator -> is_valid());
my $errs = "";
foreach my $err (@{$validator -> errors}) {
$errs .= $err -> msg." at line ".$err -> line."<br/>";
}
# otherwise, the xhtml is not valid, so return the error count
return $validator -> num_errors();
return $validator -> num_errors().":$errs";
}
# Get here and the validation request fell over, return the 'oh shit' result...
# Get here and the validation request fell over, return the 'oh shit' result...
return $validator -> validator_error();
}