Rewriting uri

Hi Gaelle,

 

 

A string operation would be fastest, but I can't think of a way to handle this easily with just string commands. A regex would provide more flexibility, but cost more in CPU usage. You can use regsub to match the uid=value and replace it with nothing. Here are a few example regexes.

 

 

If you don't know what characters the UID can contain, match any character following uid= that isn't an & or new line:

 

uid=[^&^\n]*&?

 

 

If you do know which characters the UID can contain, limit the character class (this example is alphanumerics only:

 

uid=[a-zA-Z0-9]*&?

 

 

If the uid parameter name can be mixed case, replace "uid" with:

 

 

[uU][iI][dD]

 

 

If you know how long the uid value can be, you can replace * with {0,100} in the above examples, where 0 is the shortest length and 100 is the longest:

 

 

uid=[a-zA-Z0-9]{0,100}&?

 

 

 

Of these examples, the most efficient regex would be uid=[a-zA-Z0-9]{0,100}&?

 

 

Here are some test cases for the URI with the matches in bold:

 

 

?uid=test&param=value

 

?uid=&param=value

 

?param=value&uid=test&param=value

 

?param=value&uid=&param=value

 

?param=value&uid=test

 

?param=value&uid=

 

 

To use the regex, you can use regsub:

 

 

regsub $regex $source_string $replacement_string

 

 

So this command:

 

regsub {uid=[a-zA-Z0-9]{1,100}&?} {?param=value&uid=test&param=value} ""

 

 

Returns:

 

 

?param=value&param=value

 

 

Aaron