- Configuration is space/tab insensitive, you can do it all in one line like Nginx.
- Blocks are defined inside curvy brackets
- Every definition must be closed by
;(except the curvy brakets)
A server block is defined by the server tag followed by curvy brackets including the configuration of this one server.
server {
//... your configuration here
}
port and l'host can be defined with IPV4 addresses with the listen tag. In this configuration, you have to choose one of each for each server block.
You can't set several port and host for one server. If you don't set one, default settings are set.
The following definitions are allowed:
# localhost = 127.0.0.1
# default settings are host = 0.0.0.0 (any address) port = 80
//...
listen 0.0.0.0:80; # host = 0.0.0.0 # port = 80
//...
listen 0.0.0.0 80; # host = 0.0.0.0 # port = 80
//...
listen 0.0.0.0; # host = 0.0.0.0
listen 80; #port = 80
//...
The first server for a host:port will be the default for this host:port (meaning it will answer to all request that doesn’t belong to an other server)
you can set default error page for error from 300 to 599 (as nginx does). This is going to make a redirection to the choosen page.
error_page 404 /error/404.html;
error_page 501 502 503 /error/50x.html;
🍪 Location block (setup routes with one or multiple of the following rules/configuration (routes wont be using regexp))
Like server blocks, location blocks begin and end with curvy brackets. Each location is defined with a modifier (optional) and an URI specification. Unlike Nginx, . for any char and wildcard * and ? and $any_var are not implemented (nothing including regex).
location [modifier] [URI] {
...
...
}
If no location is found, the webserv will use default settings;
modifier can be :
- none: If no modifiers are present in a location block then the requested URI will be matched against the beginning of the requested URI.
=: The equal sign is used to match a location block exactly against a requested URI.~: The tilde sign is used for case-sensitive regular expression match against a requested URI.~*: The tilde followed by asterisk sign is used for case insensitive regular expression match against a requested URI.
location can be defined by any route that match the requested URI. $ symbol is used at the end of a route to say that nothing can be found after the last char.
The following settings can only be defined inside a location block.
By default GET and POST methods are accepted, DELETE must be allowed.
Both tags allow_methods and limit_methods can be used.
if limit_methods is used, not included methods can't be allowed.
If DELETE is in limited methods, you'll still need to allow it with allow_methods.
Redirection are set with the tag return. (it can be also with error_page). It is set if the page is temporarly (302) or permanently moved (301);
return 301 @newlocation;
return 301;
🍪 define a directory or a file from where the file should be search (for example if url /kapouet is rooted to /tmp/www, url /kapouet/pouic/toto/pouet is /tmp/www/pouic/toto/pouet)
with the root tag, define a directory with absolute path. If no root is choosen, it goes by default on the /www folder of the project.
If no default file is found in this location and you requested a folder, then you can set an auto index with the autoindex tag.
autoindex on;
# OR
autoindex off; #->> default setting
Can be several file after the index tag. If none is set, you'll have to type the page's name by hand.
with CGI_path tag followed with an absolute path to the binary file.
location ~ .php$ {
CGI_path /Users/xxx/.brew/bin/php-cgi;
}