How do I write SCRIPT_URI to a header in Apache? - apache

Question

I have a setup where I have servers like this:
load balancer -> Apache -> Tomcat
I would like Apache to write the url that the client used into a header, so I can read that once I hit tomcat.
I've tried to use mod_rewrite and mod_headers so do it, but with no luck.
if I look at http://httpd.apache.org/docs/2.0/mod/mod_rewrite.html then it seems clear that i need the variable called SCRIPT_URI:
SCRIPT_URI=http://en1.engelschall.com/u/rse/
I also looked at this http://www.askapache.com/htaccess/crazy-advanced-mod_rewrite-tutorial.html so figure out how to write headers and have had some succes, but not enough.
I have php installed on the apache server, and if i look at phpinfo() i can see the SCRIPT_URI is there and has a sensible value.
I just can't get it to write it to a header. Here's a simplified version of what I've done:
#load modules
LoadModule rewrite_module modules/mod_rewrite.so
LoadModule headers_module modules/mod_headers.so
#Get the original uri used
RewriteRule .* - [E=INFO_SCRIPT_URI:%{SCRIPT_URI},NE]
RequestHeader set x-orig-uri "%{INFO_SCRIPT_URI}e"
I've tried several other options and both on windows, cygwin and ubuntu linux
Any ideas?

Answer 1

I found a workaround myself, although it's not the clean and easy solution I wanted.
I was able to get the indvidual parts and can use them to build the full URI:
<IfModule !rewrite_module>
LoadModule rewrite_module modules/mod_rewrite.so
</IfModule>
<IfModule !headers_module>
LoadModule headers_module modules/mod_headers.so
</IfModule>
<IfModule rewrite_module>
<IfModule headers_module>
####### INITIAL SETUP #########################
RewriteEngine on
####### SET HEADERS #########################
#get and set the host name
RewriteRule .* - [E=INFO_HTTP_HOST:%{HTTP_HOST},NE]
RequestHeader set x-orig-host "%{INFO_HTTP_HOST}e"
#get and set the host port
RewriteRule .* - [E=INFO_SERVER_PORT:%{SERVER_PORT},NE]
RequestHeader set x-orig-port "%{INFO_SERVER_PORT}e"
#If the uri starts with a slash and some alphanumerics, then make a
#group of that until the first non-alpha (ie. the next slash)
RewriteCond %{REQUEST_URI} ^(/[\w-]+)
#Save the content of the regex match group ( %1 ) in an environment variable
RewriteRule .* - [E=INFO_REQUEST_CONTEXT:%1,NE]
#Set a header with the content of the environment variable
RequestHeader set x-orig-context "%{INFO_REQUEST_CONTEXT}e"
#If the accept-header contains a number after ;version= then make a regex group of that number
RewriteCond %{HTTP_ACCEPT} \+json;version=(\d+)$
#Save the content of the regex match group ( %1 ) in an environment variable
RewriteRule .* - [E=INFO_ACCEPT_VERSION:%1,NE]
#Set a header with the content of the environment variable
RequestHeader set x-orig-accept-version "%{INFO_ACCEPT_VERSION}e"
#If the accept-header contains kasia2. followed by some letters,
#then make a regex group of those letters
RewriteCond %{HTTP_ACCEPT} kasia2.(\w+).*
#Save the content of the regex match group ( %1 ) in an environment variable
RewriteRule .* - [E=INFO_ACCEPT_NAME:%1,NE]
#Set a header with the content of the environment variable
RequestHeader set x-orig-accept-name "%{INFO_ACCEPT_NAME}e"
#If https is on ...
RewriteCond %{HTTPS} on
#...then set the protocol environment variable to "https"
RewriteRule .* - [E=INFO_PROTOCOL:https,NE]
#If https is off ...
RewriteCond %{HTTPS} off
#...then we assume it must be "http"
RewriteRule .* - [E=INFO_PROTOCOL:http,NE]
#Finally, set the protocol header
RequestHeader set x-orig-protocol "%{INFO_PROTOCOL}e"
#Get the request uri and set an environment variable
RewriteRule .* - [E=INFO_REQUEST_URI:%{REQUEST_URI},NE]
#Build the whole original url out of the available parts. SCRIPT_URI is always null, otherwise we could have used that.
RequestHeader set x-orig-url "%{INFO_PROTOCOL}e://%{INFO_HTTP_HOST}e%{INFO_REQUEST_URI}e"
#In addition make an url with only the host and context, for convenience
RequestHeader set x-orig-url-base "%{INFO_PROTOCOL}e://%{INFO_HTTP_HOST}e%{INFO_REQUEST_CONTEXT}e"
</IfModule>
</IfModule>

Answer 2

According to https://serverfault.com/questions/23470/setting-apache-environment-variable-on-mod-rewrite-rewrite-condition, you need to tweak the order of execution by using different configuration sections
Here I added a dummy <Location> section
RewriteRule ^ - [E=FOO:set_by_rewrite_rule]
RequestHeader set x-foo-outside-location %{FOO}e
<Location "/bla">
RequestHeader set x-foo-in-location %{FOO}e
</Location>
and these are the headers that I'm getting
x-foo-in-location='set_by_rewrite_rule'
x-foo-outside-location='(null)'
Note that I get both headers to null if I use <Location "/"> (and that I'm unable to find a solution valid for all request paths)
Note also that Sebastian's answer is getting the env variables in the header without any dummy configuration section... Go figure!