Note: This site might seem inactive… That’s because it is. Don’t worry though, I’m still coding webpages and stuff! If you’re interested, I suggest you get a translator and head over to Qiwi; or you could just check the latest site we’ve been working on: AD Delhaize Lebbeke. Enjoy!
XHTML content negotiation through PHP
The most obvious one of the many perils of using XHTML properly, is of course the fact that IE cannot handle the application/xhtml+xml MIME type. Unless you don’t mind if your site can’t be visited through that wanna-be–browser, this means that you’ll have to serve up your XHTML pages as application/xhtml+xml where possible, but as text/html to the crappier ones. Stuff like this is called content negotiation.
Through mod_rewrite
RewriteEngine on
RewriteBase /
RewriteCond %{HTTP_ACCEPT} application/xhtml\+xml
RewriteCond %{HTTP_ACCEPT} !application/xhtml\+xml\s*;\s*q=0
RewriteCond %{REQUEST_URI} \.xhtml$
RewriteCond %{THE_REQUEST} HTTP/1\.1
RewriteRule .* - [T=application/xhtml+xml]Indeed, XHTML content can be negotiation through some spiffy rewrite rules in your .htaccess, but because of the complexity of this matter, it’s probably a better idea to use PHP instead if you want top notch accurate results.
Through PHP
Until recently, I was using Simon Jessey’s well-known serving XHTML with the correct MIME type script. There were a couple of things about it that really bugged me:
- The script changes the MIME type to
text/htmlwhenapplication/xhtml+xmlisn’t supported, which is of course fine. But it will also alter the contents of the file you’re serving, making it so that you’re pretty much serving real HTML… What’s the point in that? If you want to use HTML, don’t use XHTML, and vice versa. Whatever it is you’re serving to IE, it will be handled as tag soup anyway. So why even bother removing those sexy/>? (Besides, there’s more XML-only code than just/>— think ofxml:lang, for example.) - Though it flawlessly analyzes the
qvalue for browsers, the script serves XHTML pages astext/htmlto validators, which might make it look as if you’re serving your XHTML astext/htmlall the time. And hey, if it’s a validator, it can “handle”application/xhtml+xml. - I personally don’t really like XML prologs. (Don’t ask.) And the alternating
DOCTYPEtotally swallows as well, for the same reasons I mentioned above (IE doesn’t even have a friggin’ standards mode).
So, I started hacking away in the code, and eventually I came up with my custom PHP XHTML content negotiation script. I added “support” for several validators:
- The W3C Markup Validation Service
- WDG HTML Validator
- The W3C Link Checker
- The HTTP Request and Response Header Viewer
- The Feed Validator for Atom and RSS — just so it would output the correct error message:
Feeds should not be served with the
. Yes, I really am that geeky.application/xhtml+xml;charset=utf-8media type - The Poodle Predictor
- The Leknor.com Gzip Tester
Anyway, here’s the script:
<?php
$mime = 'text/html';
if(strstr($_SERVER['HTTP_USER_AGENT'], 'W3C_Validator') ||
strstr($_SERVER['HTTP_USER_AGENT'], 'WDG_SiteValidator') ||
strstr($_SERVER['HTTP_USER_AGENT'], 'W3C-checklink') ||
strstr($_SERVER['HTTP_USER_AGENT'], 'Web-Sniffer') ||
strstr($_SERVER['HTTP_USER_AGENT'], 'FeedValidator') ||
strstr($_SERVER['HTTP_USER_AGENT'], 'Poodle predictor') ||
strstr($_SERVER['HTTP_USER_AGENT'], 'Leknor.com gzip tester')) {
$mime = 'application/xhtml+xml';
} else {
if(stristr($_SERVER['HTTP_ACCEPT'], 'application/xhtml+xml')) {
if(preg_match("/application\/xhtml\+xml;q=([01]|0\.\d{1,3}|1\.0)/i", $_SERVER['HTTP_ACCEPT'], $matches)) {
$xhtml_q = $matches[1];
if(preg_match("/text\/html;q=q=([01]|0\.\d{1,3}|1\.0)/i", $_SERVER['HTTP_ACCEPT'], $matches)) {
$html_q = $matches[1];
if((float)$xhtml_q >= (float)$html_q) {
$mime = 'application/xhtml+xml';
}
}
} else {
$mime = 'application/xhtml+xml';
}
}
}
header('Content-Type: ' . $mime . ';charset=utf-8');
header('Vary: Accept');
?>As you can see, this is far more accurate compared to the seven-line mod_rewrite solution.
Hey, how about AdSense?
Good question. As we all know, AdSense won’t work with correctly served XHTML. Again, instead of plainly using Simon Jessey’s workaround, I wrote another function that takes one argument (the $mime that was generated by running the previous script), and then outputs the matching code.
<?php
function mj_adsense($mime) {
if($mime == 'application/xhtml+xml') {
echo '<object data="/include/adsense.php" type="text/html"></object>';
} else {
echo '<script type="text/javascript"><!--
google_ad_client = "pub-7821233126901128";
google_ad_width = 468;
google_ad_height = 60;
google_ad_format = "468x60_as";
google_ad_channel ="";
google_ad_type = "text";
google_color_border = "B4D0DC";
google_color_bg = "ECF8FF";
google_color_link = "0000CC";
google_color_url = "008000";
google_color_text = "6F6F6F";
//--></script>
<script type="text/javascript" src="http://pagead2.googlesyndication.com/pagead/show_ads.js"></script>';
}
}
?>So basically, when serving as application/xhtml+xml, an object will be included in the document (in this case: /include/adsense.php, see below for source code). Otherwise, the standard AdSense JavaScript is used.
My /include/adsense.php looks like this:
<?php header('Content-Type: text/html;charset=utf-8'); ?>
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://w3.org/TR/html4/loose.dtd">
<html lang="en">
<head>
<title>AdSense | Math Jazz</title>
<style type="text/css">
body { margin: 0; padding: 0; overflow: none; }
</style>
</head>
<body>
<script type="text/javascript"><!--
google_ad_client = "pub-7821233126901128";
google_ad_width = 468;
google_ad_height = 60;
google_ad_format = "468x60_as";
google_ad_channel ="";
google_ad_type = "text";
google_color_border = "B4D0DC";
google_color_bg = "ECF8FF";
google_color_link = "0000CC";
google_color_url = "008000";
google_color_text = "6F6F6F";
//--></script>
<script type="text/javascript" src="http://pagead2.googlesyndication.com/pagead/show_ads.js"></script>
</body>
</html>Any suggestions as to validators that could be added? Do you know of a better script? Any ideas on how to improve this script? Please, let me know.
Comments (17)
Listed below are the responses for this entry.
Trackbacks & Pingbacks (4)
Listed below are resources on the web that mention this article.
-
- Nexen.net: Négociation de contenu XHTML:
Négociation de contenu XHTML
L’un des problèmes lorsque vous passez votre site en XHTML, c’est que Explorer ne reconnaitra pas le type de contenuapplication/xhtml+xml, et vous devrez vous adapter pour lui servir le même texte entext/html[…]- Trackback made on May 6th, 2005 @ 1:48 pm
-
- Negociação de conteúdo via PHP para servir documentos XHTML como aplicação XML:
Negociação de conteúdo via PHP para servir documentos XHTML como aplicação XML
Mathias publicou no seu site um excelente artigo mostrando um meio de servir um documento XHTML como XML via PHP.- Trackback made on July 8th, 2005 @ 6:03 am
-
- Effair: Billet: WordPress et les types de contenu:
WordPress et les types de contenu
Comme nous le savons tous, MSIE ne gère pas correctement les types MIME, plus particulièrementapplication/xhtml+xml, celui qui sert à déclarer du contenu en XHTML. […]- Pingback made on October 29th, 2005 @ 6:19 pm
-
- Gary Court: Proper HTTP content negotiation in PHP:
[…] Last summer, I was looking into a quick PHP script that does HTTP content negotiation for a small project I was working on. After searching around Google for several hours, I found several scripts that determined if a browser supported proper XHTML mime-types or not, but nothing for proper HTTP content negotiation. So I decided to write one. […]
- Pingback made on January 9th, 2006 @ 8:11 pm
To serve
application/xhtml+xmlto conforming UAs, I use a rewritten version of Keystone Websites’ script.Instead of using PHP’s
$_SERVER['HTTP_ACCEPT'], I useapache_request_headers(), as it is more extensive, and gives a more fullfilling printout of the accepted headers.Excellent work! I’ll definitely use some of this in a few near-future projects. Thanks for making it available.
Yeah, Henrik, me too.
Thanks for the tip, all this sounds very interesting. I shall look in to this. Unfortunately, I can’t access your site at the moment (?), so I can’t view your script’s source.
Yes, I’m really sorry about that, my domain-registrar have forgotten to bill me, so the domain expired, and it won’t be available until a few days. I registered misinterpreted.org for one year, just to have something to play with for the next couple of days, though :)
misinterpreted.org will be up sometime today, but until then, the file is available on Arve Systad’s server.
Google’s Adsense scripts change their JavaScripts all the time. For one thing, they don’t like to be modified, at all (tracking insufficiency). Soon, the modification of Google ads may be well out of the question.
Here I come thinking I am going to be able to read without thinking and I encounter this…
Very cool! You can also have all
.htmlfiles parsed asapplication/xhtml+xmlby editing the Windows registry (Start > Run >regedit.exe>HKEY_something [the first or second folder] >.html). I wouldn’t recommend this, though.Dante, what the fuck?! It is your responsibility as a web developer to serve up your files with the correct MIME type — there’s no way you can assume all your visitors hack their registry like that. Besides, the
.htmlextension makes it look like an ordinary HTML file, which should always be served astext/html. You should’ve put :POuch! Calm down man!
What my previous comment was stating was that you can force all your browsers you use on XP (or at the very least the default browser) to force the type as
application/xhtml+xml. It only effects YOUR BROWSER, not the visitors’.Also, I was using your code but threw away the thought of
application/xhtml+xmlin disgust after I realised I could no longer useinnerHTML. Fuck that, I can’t live withoutinnerHTML; using DOM methods all the time makes scripts run way slower.Nice script, though.
Dante, you make two mistakes (and I make one, by replying so late). (1)
innerHTMLdoes work in XHTML; use a recent nightly of Mozilla and be happy and (2) mapping.htmltoapplication/xhtml+xmlwill not really make it being parsed as XML because Internet Explorer does not know that MIME type.I was using a recent Firefox. The problem is other people who aren’t won’t be able to enjoy the extra effects my scripts with
innerHTMLprovide.Hey Dante, I’m sorry if I sounded like beyatch up there or anything, I didn’t intend that at all. I just didn’t quite understand what exactly you were telling to do in the Register (
HKEY_–what?), and the overall funkiness of your solution amazed me.Yeah, well that was my whole point.
Nice work! I definitely hope to use this in the near future. Actually, I’m trying to figure out how to write it up using a class.
Question though.
I look at Simon Jessey’s version in order to compare it to yours. In his, he calls a function which converts “
/>” to “>” for docs served astext/html. You opted to remove this function. I don’t know where (that’s why I’m asking) but I thought I read that “/>” can break older browsers? I know that adding a space before the slash fixes older browsers from outputting “>” everywhere, but are there other problems older browsers have when fed “/>”, with or without the starting space?Hope I made sense. Thanks.
That sounds crappy! I never heard anything like it before, though. (An intensive search session for information on this subject had no result, probably because Google doesn’t like
/>.) Link, anyone?After all, this doesn’t have much to do with XHTML content negotiation — it’s more an issue of XHTML in general.
Alright, sorry about that comment. I should of done my research before posting. I did an intense Google session as well and the only thing I found was from the W3C XHTML 1.0 spec, which is what I already knew…
LOL!
“XML parsing error”. This is the message display when you are trying to read this page with Firefox 1.0.3. I was obliged to look for my dusty IE to write this comment.
Is it a rock solid solution? :-)
In fact it is, Frédéric! The XML error message you got was caused by John’s latest comment :)