Hi,
we have been experimenting with version 1.5.4 (haven't got around to
the new 1.6.0 yet, but judging from the changelog and the validator
code on beta branch [1], it seems like this hasn't changed since then)
of Rhodecode, and have set up LDAP authentication. Some of our LDAP
usernames are e-mail addresses, so they have "@" characters in them.
These users cannot authenticate in Rhodecode because Rhodecode rejects
the username as invalid because of the supposedly invalid "@"
characters in them (it fails the match against the regular expression
^[a-zA-Z0-9]{1}[a-zA-Z0-9\-\_\.]*$ [1]). That happens even before
checking LDAP.
Is there a compelling reason to outright reject those types of
usernames? Isn't this a common use case enough to be allowed into the
main code of Rhodecode, instead of us having to patch it in our
instance?
[1]
https://bitbucket.org/marcinkuzminski/rhodecode/src/7d29355db6f44802d3af3c37b3139e97201ed30c/rhodecode/model/validators.py?at=beta#cl-108
Best regards,
Augusto Herrmann