mod_rewrite to absolute path in .htaccess - turning up 404 - apache

I want to map a number of directories in a URL:
www.example.com/manual
www.example.com/login
to directories outside the web root.
My web root is
/www/htdocs/customername/site
the manual I want to redirect to is in
/www/customer/some_other_dir/manual
In mod_alias, this would be equal to
Alias /manual /www/customer/some_other_dir/manual
but as I have access only to .htaccess, I can't use Alias, so I have to use mod_rewrite.
What I have got right now after this question is the following:
RewriteRule ^manual(/(.*))?$ /www/htdocs/customername/manual/$2 [L]
this works in the sense that requests are recognized and redirected properly, but I get a 404 that looks like this (note the absolute path):
The requested URL /www/htdocs/customername/manual/resourcename.htm
was not found on this server.
However, I have checked with PHP: echo file_exists(...) and that file definitely exists.
why would this be? According to the mod_rewrite docs, this is possible, even in a .htaccess file. I understand that when doing mod_rewrite in .htaccess, there will be an automated prefix, but not to absolute paths, will it?
It shouldn't be a rights problem either: It's not in the web root, but within the FTP tree to which only one user, the main FTP account, has access.
I can change the web root in the control panel anytime, but I want this to work the way I described.
This is shared hosting, so I have no access to the error logs.
I just checked, this is not a wrongful 301 redirection, just an internal rewrite.

In .htaccess, you cannot rewrite to files outside the wwwroot.
You need to have a symbolic link within the webroot that points to the location of the manual.
Then in your .htaccess you need the line:
Options +SymLinksIfOwnerMatch
or maybe a little more blindly
Options +FollowSymlinks
Then you can
RewriteRule ^manual(/(.*))?$ /www/htdocs/customername/site/manual/$2 [L]
where manual under site is a link to /www/customer/some_other_dir/manual
You create the symlink on the command line with:
ln -s /www/htdocs/customername/site/manual /www/customer/some_other_dir/manual
But I imagine you're on shared hosting without shell access, so look into creating symbolic links within CPanel,Webmin, or whatever your admin interface is. There are php/cgi scripts that do it as well. Of course, you're still limited to the permissions that the host has given you. If they don't allow you to follow symlinks as a policy, you cannot override that within your .htaccess.

AFAIK mod_rewrite works at the 'protocol' level (meaning on the wire HTTP). So I suspect you are getting HTTP 302 with your directory path in the location.
So I'm afraid you might be stuck unless.. your hosting lets you follow symbolic links; so you can link to that location (assuming you have shell access or this is possible using FTP or your control panel) under your current document root.
Edit: It actually mentions URL-file phase hook in the docs so now I suspect the directory directives aren't allowing enough permissions.

This tells you what you need to know.
The requested URL /www/htdocs/customername/manual/resourcename.htm
was not found on this server.
It interprets RewriteRule ^manual(/(.*))?$ /www/htdocs/customername/manual/$2 [L] to mean rewrite example.com/manual/ as if it were example.com/www/htdocs/customername/manual/.
Try
RewriteRule ^manual(/(.*))?$ /customername/manual/$2 [L]
instead.

Related

Howto use path/directory outside the domain root via symlink without access to domain config?

I am working an a shared hosting plattform which does not allow to edit or access the Apache config file.
The goal is to access the same files from two different domains which point to different domain roots:
test.example.com ---> /test_root/web
public.example.com ---> /public_root/web
Now I would like to access the same files using test.example.com/some/files/... and public.example.com/some/files/...
Of course I could simply copy the files to /test_root/web/some/files/... and to /public_root/web/some/files/... but this obviously just an example. In reality the files are a helpdesk system which should be integreated both into the test- and public-site. Copying the files would include maintaining two different systems, etc.
The goal is, to place the the files somewhere outside the two domain roots and make them available from both domains:
/test_root/web/some/files ---> /path/to/some/files
/public_root/web/some/files ---> /path/to/some/files
I created symlink to achive this (ln -s ...) but this does not work out. When I access on of the domains (e.g. test.example.com/some/files) I only get a blank page without any information what whent wrong.
I assume that the Apache is not configured to follow the symlinks. Without access to the Apache config I can neither check nor fix it.
Adding Symlinks to .../web/.htacess does not make any difference:
// test_root/web/.htaccess
Options +FollowSymLinks
So the question is: How can I make files outside the domain root available within a domain? Is this even possible? Is this possible using symlinks?
Add this in you .htaccess file.
RewriteEngine on
RewriteBase /
RewriteCond %{HTTP_HOST} test.example.com/path/to/files [NC]
RewriteRule ^(.*)$ http://public.example.com/path/to/files [R=301,NC]
Test out this code. /path/to/files must be under /public_root/web/.
Something like /public_root/web/path/to/files/

Generic .htaccess for multiple websites stored in subdirectories

My development environment is set up for using a single host (localhost). I am developing multiple websites on my machine, each stored under its own directory like this:
/var/www/site1
/var/www/site2
...
The document root is set to /var/www on my machine.
I am using URL rewriting for most of these websites and most of the .htaccess files will rewrite a sub-directory to GET parameters in different ways like this:
http://localhost/site1/home/red -> http://localhost/site1/index.php?page=home&p1=red
http://localhost/site2/index/param1/param2/param3 -> http://localhost/site2/index.php?page=index&p1=param1&p2=param2&p3=param3
I also tend to copy some of these websites under different directories and, when I do that, I have to make a lot of changes in the .htaccess files for the website that I'm copying.
I would like to know if there is a way to define a constant that contains the website's root directory (not the host's document root) and how can that be used with the rewrite rule so that I would need to change only one line of code (setting this constant to a different value) when copying a website.
Putting this in a different form, is there a way to perform rewrites that relate to a website root instead of a host / %{HTTP_HOST} (i.e. the "host" for the website being localhost/site1 instead of localhost) and how can this be done?
I have tried removing the host from each request at the beginning of the script and prepending it back at the end of the script, but this does not work with rewrite rules that use the [L] option.
Thank you!
Regards,
Lucian
You could make an htaccess file with rules like this:
RewriteEngine On
RewriteBase /site1/
RewriteRule ^([^/]+)/([^/]+)/([^/]+)/([^/]+) index.php?page=$1&p1=$2&p2=$3&p4=$4 [L,QSA]
And put this in the directory /var/www/site1, and if you want for it to apply to site2, change the RewriteBase and put the rules in /var/www/site2.

How do I configure apache for a custom directory?

Trying to configure apache2 to load example.com/forum/ from a different document root, relative to the site root. Forums are installed somewhere else on the server.
Is there a directory alias command? I've found the alias configuration entry for apache, but had no luck.
Basically, I want example.com to have the same directory its always had, but example.com/forum/ to be hosted somewhere else, on the same server.
I tagged this question with mod_rewrite because I thought maybe it would be the key, here.
Cheers!
Alias is the right way, unless you have some subtlety that you didn't reveal in your question.
# http.conf
Alias /forum /usr/lib/bbs/ # or whatever
The job of Alias is to take the abstract URL coming into your system and map it to a concrete filesystem path. Once it has done that, the request is no longer an URL but a path. If there is no Alias or similar directive handling that URL, then it will get mapped to a conrete path via DocumentRoot.
If this isn't working, you have to debug it further. Are you getting errors when you access /forum? Look in the error log.
It all depends of what you want. You can "hardlink" with real path and it works (so you were right to think it could work with mod_rewrite).
Quick sample (that works on my production domains) to make an internal change (I add a subdirectory):
RewriteRule (.*) %{DOCUMENT_ROOT}/mysubfolder%{REQUEST_FILENAME} [QSA,L]
So you can easily do something like:
RewriteRule ^/forum/(.*) %{DOCUMENT_ROOT}/mysubfolder%{REQUEST_FILENAME} [QSA,L]
And my suggestion would be that if you plan to have more rewrite rules, keep everything homogeneous, i.e.: keep on using only rewrite rules, so use my suggestion above. This way you'll not get a bad mix of Alias, RewriteRules and so on. For nice and clean stuff: keep everything homogeneous.

Changing a file's URL without physically moving it

I have a site, running Linux + Apache.
I have a file in my root directory, let's say file.php.
I want the URL to the file to be "domain.com/newdir/file.php", but I don't want to actually create the newdir and move the file there because it would be a huge hassle to update many many links all over my site.
Is there a way to accomplish this, meaning making the file accessible by the new URL without moving it?
Thank you.
On this site: workwith.me, you can find information about .htaccess and mod_rewrite. For your example you have to make a file called .htaccess and put it in the root directory. The file should contain these directives:
RewriteEngine on
RewriteRule ^newdir/file.php$ /file.php [L]
You can do this for every file you want to rename.
Four possible solutions I can think of:
If your OS supports it, create a symlink:
mkdir /home/foo/htdocs/newdir
ln -s /home/foo/htdocs/file.php home/foo/htdocs/newdir/file.php
... and make sure Apache is configured to follow them:
Options FollowSymLinks
Create an Alias or AliasMatch (probably overkill)
Good old mod_rewrite:
RewriteEngine One
RewriteRule ^newdir/file\.php$ file.php [L]
Ugly: use a custom 404 error page with a PHP script that checks $_SERVER['REQUEST_URI'].
I guess the standard solutions are #1 and #3.

How do I force Apache to simply redirect the user and ignore the directory structure?

Ok, so this problem recently arose and I don't know why it is happening; it's actually two problems in one...
0. My .htaccess file, for reference. (EDITED)
Options -Indexes +FollowSymLinks
RewriteEngine On
RewriteBase /
ErrorDocument 400 /index.php?400
ErrorDocument 401 /index.php?401
ErrorDocument 403 /index.php?403
ErrorDocument 404 /index.php?404
ErrorDocument 410 /index.php?410
ErrorDocument 414 /index.php?414
ErrorDocument 500 /global/500.php
RewriteCond %{HTTP_HOST} !^$ [NC]
RewriteRule .* index.php [L]
RewriteCond %{HTTP_REFERER} !^$
RewriteCond %{HTTP_REFERER} !^https?://(.*\.)?(animuson)\.(biz|com|info|me|net|org|us|ws)/.*$ [NC]
RewriteRule ^.*$ - [F]
1. My 'pictures' folder is following the hard path instead of the redirect.
I have no idea WHY it is doing this. It's really bugging me. The 'pictures' folder is a symbolic link to another place so that I can easily upload files to that folder without having to search through folders and such via my FTP account, but that's the only thing I use it for. However, when I visit http://example.com/pictures my htaccess sees it as accessing that other folder, which is restricted, and throws a 403 error rather than redirecting to index.php and displaying the page like normal.
I figured it has something to do with that specific folder being a symbolic link causing it to act oddly, but I have determined that my rules are not being applied to folders at all. If I visit folders such as 'css' and 'com' which are folders in the web root, it displays a 404 error page and adds the '/' to the end of the URL because it's treating it as a directory. It also does the same 403 error for my 'images' directory which is set up in the same fashion.
So, the question here is how do I modify my RewriteRule to apply to the directories as well? I want everything accessed via the web to be redirected back to index.php while maintaining the full access path in the address bar, why is it not working? (I'm pretty sure it was working fine before.)
Here's a small chart to show the paths they're following...
example.com/pictures -> pictures/ -> /home/animuson/animuson-pictures -> 403
example.com/com -> com/ -> 404
example.com/test -> index.php
example.com/ -> index.php
example.com/images -> images/ -> /home/animuson/animuson-images -> 403
example.com/css -> css/ -> 404
EDIT: Following information added.
Apache is processing the structure of the directory first. It's determining if the path exists based on what was typed into the address bar. If someone types in a folder name that happens to exist, it will redirect the user to the path with the "/" at the end of the URL signifying that it's a directory. For the 'pictures' directory explained above, the user does not have permission to access that folder so it is redirecting them to a 403 Access Denied page rather than simply showing the page that is supposed to be displayed there via the RewriteRule above. My biggest question is why is Apache processing the directory first and how do I make it stop doing that? I would really love an answer to this question.
2. Why is my compression not working? (EDIT: This part is fixed.)
When analyzing my site through a web optimizer, it keeps saying my page isn't using web compression, but I'm almost 100% positive that it was working fine before under the same settings. Can anyone suggest any reasons why it might not be working with this set up or suggest a better way of doing it?
Where is this .htaccess file situated? At the root or in the pictures directory?
1) You're using Options -Indexes which will deny access to directory listings. This is handled by /index.php?403 which in turn will redirect to /403. (I confirmed this by manually going to /index.php?403) I don't see any other rules in the posted .htaccess that are supposed to affect this. So this either happens because either index.php or some other .htaccess file or server rule makes that redirect.
You might also want to check the UNIX file permissions of the directory in question.
2) According to this aptimizer, http://www.websiteoptimization.com/services/analyze/, compression is indeed enabled for html, js and css files, as specified in the rules. My bet is that the optimizer is being stupid and does one of these three things:
1)) Complaining about images not being compressed. (It's generally a bad idea to compress images because they're typically already compressed and the extra CPU load typically isn't worth it since the net gain is so small. So your rules are OK in this regard.)
2)) It might think that DEFLATE doesn't count as compression, and wants you to use GZip.
3)) It might also react to the externally included StatCounter js file, which is not compressed. (And there's not much you can do about that.)
After a while of deliberating on Apache's IRC channel, I was finally able to figure out the real reasoning behind this on a fluke. I just happened to be looking at the directory structure using ls -l and noticed that all of the symbolic links had somehow has their permissions changed to animuson:animuson from the root:root original. I tried to run a simple chown root:root on them and it had no effect, so I deleted them all and recreated them and the problem has gone away. I don't really have any idea why the permissions made any different in this scenario but the solution worked and everything is okay now. I've also added a DirectorySlash Off to my .htaccess file to get rid of the slashes after folders that exist, just to make it look all that much nicer.