nginx location regex - character class and range o

Question

I am trying to setup a regex for the path /s/<4-6 character string here> where I capture the 4-6 character string as $1.

I tried using the following two entries, but both fail

location ~ ^/s/([0-9a-zA-Z]){4,6}+$ { ...

location ~ ^/s/([0-9a-zA-Z]{4,6})+$ { ...

The first one comes up with 'unknown directive' and the second comes up with 'pcre_compile() failed: missing )'

EDIT

The following routes would be served by this location:

/s/1234 (and I would capture '1234' in $1)
/s/12345 (and I would capture '12345' in $1)
/s/123456 (and I would capture '123456' in $1)
/s/abcd (and I would capture 'abcd' in $1)
/s/abcde (and I would capture 'abcde' in $1)
/s/abcdef (and I would capture 'abcdef' in $1)
/s/a1b2c (and I would capture 'a1b2c' in $1)

The following routes would NOT be served by this location:

/s/1
/s/12
/s/123
/s/a
/s/ab
/s/abc
/s/abc1234
/s/12345678

etc...

Answer 1

If you want to capture 4 to 6 characters, why you don't have put the quantifier inside the capture parenthesis?

Something like that perhaps:

location ~ "^/s/([0-9a-zA-Z]{4,6})$" {...

Curly braces are used both in regex and for block control, you must enclose your regex with quotes (single or double) (<-- wiki nginx)