Forum Moderators: phranque
Please believe me when I say that I wouldn't post something as simple as this unless it was a last resort, but as it is I really am about to go out of my mind! It's no exaggeration to say that I've spent about 10 hours a day for the past three days on nothing but this:
Options +FollowSymLinks
RewriteEngine On
RewriteRule ^index\.html$ newpage.htmlI kid you not, it's as simple as that. That is the entirety of an .htaccess file that is located in the same directory as the two html files in question. On my test server at home it works perfectly. But when I try it on the VDS it does not.
Here's the deal: There is a website out there that links to my page, let's say "startpage.html." Upon loading the .htaccess file above into the directory where both html files are located, and clicking on the link on startpage, instead of opening newpage.html, I find myself at index.html. However, from index.html if I click the "home" link (which links to itself), then newpage.html opens correctly.
I'm telling you, I'm about to lose it. If anyone can help point me in the right direction you'd be making me (and everyone who's had to be around me these last three days) very happy indeed. Thanks for reading.
If so, have you tried changing the rewrite into a redirect, pointed at an external site, just to see if that works? (This would expose problems such as other rules, Alias directives, or possibly even scripts interfering.)
Jim
Thanks for the reply. I discovered sometime around the middle of Day 2 that clearing the browser cache made a difference. I thought for sure that was what my probem was, because I could think of no other reason for it not to work.
And if by pointing it at an external site, you mean something like
RewriteRule ^index\.html$ http://mysite/newpage.html
then yes, I have tried that and many variations on it as well; even tried IP address e.g. [123.456.7.8...] but to no avail.
At the moment I have resorted to the ugliest hack you can imagine:
DirectoryIndex newpage.htmlBut after 3 days of this nonsense I don't care how ugly it is. Is there a way to use Conditionals elsewhere besides RewriteRule? Because it would be nice if employing this awful hack didn't have the unfortunate effect of sending users who are already on my site and click the "home" button to newpage! For instance, if the .htaccess file were a perl script, I'd say:
$DirectoryIndex = 'index.html';
if ($referer =~ /startpage/){ $DirectoryIndex = 'newpage.html'; }but maybe I can only declare DirectoryIndex once in .htaccess and not change it?
And I understand that that is basically what RewriteCond and RewriteRule are supposed to do, but I've tried all the stuff that is supposed to work and now I'll try anything at all. Heck, I'll go to each user's computer and physically type in the URL for them if that's what it takes!
You can change the 'old' index.html page URL as well, again as a test:
RewriteRule ^foo\.html$ http://www.google.com [R=301,L]
and then if that works, try:
RewriteRule ^index\.html$ http://www.google.com [R=301,L]
On the other hand, if index.html is already declared as your DirectoryIndex, then either deleting the old index.html file and renaming newpage.html to index.html, or changing DirectoryIndex (as you have done) is the correct thing to do, and the latter is not really an ugly hack.
Compare the DirectoryIndex directives in the httpd.conf files of your test and hosted servers, and you'll likely find the reason that it works on your test server, but not on the hosted server.
It's not entirely clear why you haven't simply renamed the file. If this is part of some larger problem, it might do to back up a step and ask about that instead.
You can get around some problems with internal rewrites such as those done by mod_rewrite, mod_alias, and DirectoryIndex, by using RewriteCond to examine %{THE_REQUEST} to check the URL originally requested by the client when it started the current HTTP transaction. This can be used to avoid rewriting or redirecting a previously-rewritten URL (again, with 'previous' meaning 'within this current HTTP request').
An example is avoining an infinite loop' when trying to redirect requests for "index.html" to "/" -- a problem when "/" is being internally rewritten by DirectoryIndex to "index.html":
RewriteCond %{THE_REQUEST} ^[A-Z]{3,9}\ /index\.html\ HTTP/
RewriteRule ^index\.html$ http://www.example.com/ [R=301,L]
GET /index.html HTTP/1.1
Jim
And of course, you're right. There's more to it than the simple scenario I sketched out. The full scenario is that I want visitors referred from this one specific domain to be taken to a special page on my site, while everyone else continues to see the normal home page. And this, too, I have achieved on the test server (though not from the actual external domain, of course).
It's interesting, though, because just writing that line of perl in my previous post made me realize that maybe I'd caught a bit of tunnel vision these past 3 days, so I dashed off the following:
#!/usr/bin/perl -w
# index.cgi
use strict;
use CGI qw(referer redirect);
my $site = 'http://mysite.com/';
my $newpage = 'newpage.html';
my $home = 'home.html';
my $q = new CGI;
if ($q->referer =~ /thatdomain/){
print redirect($site . $newpage);
}
else { print redirect($site . $home); }
DirectoryIndex index.cgithen rename my old index.html to 'home.html' and I'm done!
Now if I've learned anything these past 3 days (and I have), it's that I shouldn't get too proud of myself before it actually works... but at the least, it's good to know I've got more options than I thought.
And you make a great point about examining the different config files, as that should definitely narrow down what's going on.
By the way, I love this site. I had another question that I asked the one and only other time I posted here, which is: is it possible to view the whole thread while composing a reply? As it is, I can only see my original post but not either of your subsequent replies -- in fact, I'm not even 100% sure if you've replied once or twice, but I think it was twice :) -- which makes it kind of hard to properly respond to what you wrote. Am I missing a button somewhere?
Thank you, Jim. I appreciate you lending a hand.
At best, referrer-based solutions will work 50% of the time, and should only be used for applications (such as hotlink protection) where that is "good enough" to stop the undesired behaviour. For example, the reason that it's 'good enough' to stop image hotlinking is that if a Webmaster codes his page to load an image from your server, and then finds that half of *his* visitors report a broken image on his site because you're blocking referrals from his site, then he'll probably look elsewhere for an image to steal. And even if he doesn't, you've at least cut your bandwidth loss to less that it would have been.
Anyway, decide for yourself how to handle the case of a blank referrer, because it will be a common case, and you'll need your design to 'degrade gracefully' in the case where the referrer is not present in the request.
As to viewing the whole thread when replying, I'm afraid the best way is to pop another window or browser tab. The forum's behaviour of showing only the first post is intentional, and is meant to keep threads here on-topic. We treat the original poster as "the owner" of a thread, and replies should be made to his/her question, and not to subsequent posts that wander off-topic... It does have its limitations, especially in a technical discussion, but since the work-around is simple, the benefits outweigh the negatives.
Jim
# If referrer is my site
RewriteCond %{HTTP_REFERER} ^http://www\.my-site\.com
# serve /my-site-home-page for "/" or "/index.html" requests
RewriteRule ^(index\.html)?$ /my-site-home-page.html [L]
# else serve /not-my-site-home-page.html for "/" or "/index.html" requests
RewriteRule ^(index\.html)?$ /not-my-site-home-page.html [L]
# If referrer is my site
RewriteCond %{HTTP_REFERER} ^http://www\.my-site\.com [OR]
# or if it is blank
RewriteCond %{HTTP_REFERER} ^$
# serve /my-site-home-page for "/" or "/index.html" requests
RewriteRule ^(index\.html)?$ /my-site-home-page.html [L]
# else serve /not-my-site-home-page.html for "/" or "/index.html" requests
RewriteRule ^(index\.html)?$ /not-my-site-home-page.html [L]
Or alternately, using negative logic:
# If referrer is NOT my site (or if referer is blank)
RewriteCond %{HTTP_REFERER} !^http://www\.my-site\.com
# serve /not-my-site-home-page for "/" or "/index.html" requests
RewriteRule ^(index\.html)?$ /not-my-site-home-page.html [L]
# else serve /my-site-home-page.html for "/" or "/index.html" requests
RewriteRule ^(index\.html)?$ /my-site-home-page.html [L]
# If referrer is NOT my site
RewriteCond %{HTTP_REFERER} !^http://www\.my-site\.com
# and if referrer is non-blank
RewriteCond %{HTTP_REFERER} .
# serve /not-my-site-home-page for "/" or "/index.html" requests
RewriteRule ^(index\.html)?$ /not-my-site-home-page.html [L]
# else serve /my-site-home-page.html for "/" or "/index.html" requests
RewriteRule ^(index\.html)?$ /my-site-home-page.html [L]
Jim
1) there is a referer string and it matches
2) there is a referer string and it doesn't match
3) there is no referer string
?
As for the posting thing: I very much applaud that philosophy, and the fact that that amount of thought went into the site's design. Well done.
I should also mention that nothing bad would happen if visitors from that domain aren't properly redirected. They would just see the regular home page like everyone else. I just want to redirect as many of them as possible, and the referer is the only way I know of to do that.
So, you decide what to do with blank referers and implement the code to do what you want. I just wanted to make you aware of the issue... ;)
Jim
