- From: Ville Skytta via cvs-syncmail <cvsmail@w3.org>
- Date: Tue, 15 Sep 2009 19:01:08 +0000
- To: www-validator-cvs@w3.org
Update of /sources/public/validator/httpd/cgi-bin In directory hutz:/tmp/cvs-serv4373 Modified Files: check Log Message: Comment updates. Index: check =================================================================== RCS file: /sources/public/validator/httpd/cgi-bin/check,v retrieving revision 1.690 retrieving revision 1.691 diff -u -d -r1.690 -r1.691 --- check 12 Sep 2009 15:57:50 -0000 1.690 +++ check 15 Sep 2009 19:01:06 -0000 1.691 @@ -974,7 +974,8 @@ # because $File->{Bytes} is not affected by the overrides. This will # most likely be a source of errors about internal/actual charset # differences as long as our transcoding process does not "fix" the - # charset info in XML declaration and meta http-equiv (any others?). + # charset info in XML declaration, meta http-equiv/charset and/or BOM + # (any others?). if($File->{'Direct Input'}) { # sane default when using html5 validator by direct input $req->content_type("text/html; charset=UTF-8"); } @@ -986,7 +987,8 @@ else { # Pass original bytes, Content-Type and charset as-is. # We trust that our and validator.nu's interpretation of line numbers - # is the same (regardless of EOL chars used in the document). + # is the same later when displaying error contexts (regardless of EOL chars + # used in the document). my @content_type = ($File->{ContentType} => undef); push(@content_type, charset => $File->{Charset}->{HTTP}) @@ -998,6 +1000,8 @@ $req->content_language($File->{ContentLang}) if $File->{ContentLang}; # Intentionally using direct header access instead of $req->last_modified + # (the latter takes seconds since epoch, but $File->{Modified} is an already + # formatted string). $req->header('Last-Modified', $File->{Modified}) if $File->{Modified}; # If not in debug mode, gzip the request (LWP >= 5.817)
Received on Tuesday, 15 September 2009 19:01:16 UTC