Nginx Location使用详解
官方文档
文档链接：http://wiki.nginx.org/NginxChsHttpCoreModule#location
语法：location [=|~|~*|^~] /uri/ {...}
匹配逻辑：
- Directives with the = prefix that match the query exactly. If found, searching stops.
- All remaining directives with conventional strings. If this match used the ^~ prefix, searching stops.
- Regular expressions, in the order they are defined in the configuration file.
- If #3 yielded a match, that result is used. Otherwise, the match from #2 is used.
测试方法
1. 写个test.php，用来输出规则到浏览器
<?php
echo $_GET['req'].’ -> '.$_GET['rule'];
?>
2. 配置nginx，将规则传到test.php，显示
location /uri/ {
rewrite (/.*) /test.php?req=$1&rule=[str]/uri/ last;
}
location ~* ^/test.php$ {
include fastcgi.conf;
fastcgi_pass 127.0.0.1:9000;
fastcgi_param SCRIPT_FILENAME $document_root/test.php;
}
测试用例与结果
1. 字符串匹配
- 匹配“/uri/”打头的url，大小写敏感
- 示例：
location / {...}或location /uri/ {...} - 多条规则间，取最长匹配的规则
测试用例：
nginx rules：
location / {
rewrite (/.*) /test.php?req=$1&rule=[str]/ last;
}
location /uri/ {
rewrite (/.*) /test.php?req=$1&rule=[str]/uri/ last;
}
location /uri/sub/ {
rewrite (/.*) /test.php?req=$1&rule=[str]/uri/sub/ last;
}
test results：
/uri/ss -> [str]/uri/
/uri/sub/ss -> [str]/uri/sub/
2. 正则匹配
- ~ 表示大小写敏感正则，~* 表示大小写不敏感正则
- 示例：
location ~ ^/test/(.*)$ {...}或location ~* ^/uri/ {...} - 多条正则规则间，取最先匹配的规则
- 优先级高于字符串匹配
测试用例1：
nginx rules：
location ~* ^/uri/ {
rewrite (/.*) /test.php?req=$1&rule=[reg]/uri/ last;
}
location ~* ^/uri/sub/ {
rewrite (/.*) /test.php?req=$1&rule=[reg]/uri/sub/ last;
}
test results：
/uri/xx -> [reg]/uri/
/uri/sub/xx -> [reg]/uri/
测试用例2：
nginx rules:
location /uri/sub/ {
rewrite (/.*) /test.php?req=$1&rule=[str]/uri/sub/ last;
}
location ~* ^/uri/ {
rewrite (/.*) /test.php?req=$1&rule=[reg]/uri/ last;
}
test results：
/uri/sub/xx -> [reg]/uri/
3. ^~ 前缀
- 优先级提升的字符串匹配
- 示例：
location ^~ /uri/ {...} - 多条规则间，取最长匹配的规则
- 匹配成功后，则不再匹配正则
测试用例1：
nginx rules:
location /uri/ {
rewrite (/.*) /test.php?req=$1&rule=[str]/uri/ last;
}
location ^~ /uri/sub/1/ {
rewrite (/.*) /test.php?req=$1&rule=[^~str]/uri/sub/1/ last;
}
location ~* ^/uri/sub/ {
rewrite (/.*) /test.php?req=$1&rule=[reg]/uri/sub/ last;
}
test results:
/uri/xx -> [str]/uri/
/uri/sub/xx -> [reg]/uri/sub/
/uri/sub/1/xx -> [^~str]/uri/sub/1/
测试用例2：
nginx rules:
location ^~ /uri/ {
rewrite (/.*) /test.php?req=$1&rule=[^~str]/uri/ last;
}
location /uri/sub/ {
rewrite (/.*) /test.php?req=$1&rule=[str]/uri/sub/ last;
}
location ~* ^/uri/ {
rewrite (/.*) /test.php?req=$1&rule=[reg]/uri/ last;
}
test results:
/uri/xx -> [^~str]/uri/
/uri/sub/xx -> [reg]/uri/
测试用例3：
nginx rules:
location /uri/ {
rewrite (/.*) /test.php?req=$1&rule=[str]/uri/ last;
}
location ^~ /uri/ {
rewrite (/.*) /test.php?req=$1&rule=[^~str]/uri/ last;
}
test results:
nginx启动报错：duplicate location
4. = 前缀
- 全等匹配
- 示例：
location = /crossdomain.xml {...} - 优先级最高
测试用例：
nginx rules:
location ^~ /uri/sub/ {
rewrite (/.*) /test.php?req=$1&rule=[str]/uri/sub/ last;
}
location ~* ^/uri/sub/ {
rewrite (/.*) /test.php?req=$1&rule=[reg]/uri/sub/ last;
}
location = /uri/sub/equal {
rewrite (/.*) /test.php?req=$1&rule=[equal]/uri/sub/equal last;
}
test results:
/uri/sub/xx -> [str]/uri/sub/
/uri/sub/equal -> [equal]/uri/sub/equal
总结
再看官方文档：
- Directives with the = prefix that match the query exactly. If found, searching stops.
- All remaining directives with conventional strings. If this match used the ^~ prefix, searching stops.
- Regular expressions, in the order they are defined in the configuration file.
- If #3 yielded a match, that result is used. Otherwise, the match from #2 is used.
解读：
- 先匹配 = 前缀，如果匹配到，则停止匹配，取该location
- 匹配所有字符串规则（包括无前缀 和 ^~ 前缀），从中取出最长匹配的一条（可以理解有个变量，依次扫描规则，遇到匹配的则与变量中规则对比，如果长度更大，则顶替变量中的规则。最终变量里的规则就是第 2 步的匹配结果）。这一步中，字符串规则不允许重复，否则启动nginx会报错，而且有无 ^~ 前缀是无影响的。例如同时有两个规则：
location /uri/sub/ {...}和location ^~ /uri/sub/ {...}，启动会报错。 - 如果第二步的结果是 ^~ 前缀的，则不再进行正则匹配，直接取该结果。否则，按规则定义的顺序依次匹配正则，一旦遇到匹配的，则停止匹配，取该正则的location。如果匹配下来没找到匹配的规则，则取 第二步 的结果。 （正则的匹配只有 match 和 not match 两种结果，没有长度的概念，所以是取最先匹配到的）
总结：
- 尽可能地使用正则匹配，字符串匹配规则可以用正则替代，正则匹配应按粒度从细到粗配置；
- 尽量避免 ^~ 前缀，它会使规则变得复杂，结果难以预测；
- 一般预留一个
location / {...}，表示没有匹配到任何规则时的默认行为； - 对于唯一的路径，使用 = 前缀，它总能保证不被覆盖。
（完）
blog comments powered by Disqus