- From: Frankie <frankieh@iinet.net.au>
- Date: Sat, 30 Aug 2003 14:58:46 +0800
- To: <www-validator@w3.org>
Hi Guys, I wanted to setup a copy of the W3C validator on our site, which is targetted for newbies just learning HTML/XHTML, so they get started on the right track.. Anyway, I have been pointing people to the w3c validator for a long time, and my most common questions are related to the error messages offered by the validator... I actually wrote a small tute with really simple explanations of some of the most common reported validator error messages.. It even has a section on what "foo" and "bar" (and foobar) mean and are used for. (which was a common question) But i decided it might be worth going a step further and creating a newbie version of the validator, where the messages are more verbose and simplier in terminology.. The idea being that if they find a friendly validator, they might use it from the get go, and stop asking me how to fix pages with really bad IE only code. Good practises are better taught early on rather then later on when people are set in their ways. So with that idea in mind, I grabbed the lastest from the w3c CVS and proceeded to install it. My servers specs are as follows: Mandrake Linux 9.0 kernel: 2.4.20 perl: 5.8.0 openSP: 1.50 CGI: 2.810 libwww-perl: 5.65 HTML-Parser: 3.26 URI: 1.25 Set-IntSpan: 1.07 Text-Iconv: 1.2 And a CSV dump of the latest validator. $Id: check,v 1.324 2003/05/25 12:07:56 link Exp $ The validator.conf paths all match the placement of the various libs and stuff. Its being picked up fine by the check script It even sort of works, when you try to validate a page, it goes off and gets it, but always reports that the page is not valid, (even when it is) and it never reports any of the errors it finds. If you select "Show Source" It displays the pages source HTML/XHTML correctly, but never any error messages. No errors are reported in the server error log about this. (The path in validator.conf appears to be correct: SGML_Parser = /usr/bin/onsgmls and also in my %config_opts in the check script itself ). I left all the defaults in place when I compiled openSP.. should I have??? Also, when you try to manually upload a html file, I get this error: The URI you gave me, <upload://index.html>, returned a redirect to <index.html>. I can see in "check" where it is returning this message, but I don't really understand why. Any ideas on how to sort that out? Once its working, I will start writing newbie friendly messages for it. One last thing, the linkchecker works fine, but i get alot of warnings in the error log: [Sat Aug 30 14:28:19 2003] LinkChecker.pl: Illegal character in prototype for W3C::LinkChecker::escape_match : $,\% at /home/httpd/fixit/www/validator/LinkChecker.pl line 954. [Sat Aug 30 14:28:19 2003] LinkChecker.pl: Illegal character in prototype for W3C::LinkChecker::record_redirects : \%,\% at /home/httpd/fixit/www/validator/LinkChecker.pl line 1019. [Sat Aug 30 14:28:19 2003] LinkChecker.pl: Illegal character in prototype for W3C::LinkChecker::is_redirected : $,% at /home/httpd/fixit/www/validator/LinkChecker.pl line 1028. [Sat Aug 30 14:28:19 2003] LinkChecker.pl: Illegal character in prototype for W3C::LinkChecker::get_redirects : $,% at /home/httpd/fixit/www/validator/LinkChecker.pl line 1034. [Sat Aug 30 14:28:19 2003] LinkChecker.pl: Illegal character in prototype for W3C::LinkChecker::anchors_summary : \%,\% at /home/httpd/fixit/www/validator/LinkChecker.pl line 1061. [Sat Aug 30 14:28:19 2003] LinkChecker.pl: W3C::LinkChecker::urize() called too early to check prototype at /home/httpd/fixit/www/validator/LinkChecker.pl line 82. In the end I turned off warnings just to save my error log getting filled. Any tips would be much appreciated. regards Franki The Blaster virus said: "Billy Gates, why do you make this possible?" "Stop making money and fix your software." Valid points don't you think?
Received on Saturday, 30 August 2003 03:08:20 UTC