[PATCH conductor] BZ-863465 username can contain only characters and numbers

Matt Wagner matt.wagner at redhat.com
Mon Oct 8 15:59:08 UTC 2012


On Mon, Oct 08, 2012 at 05:33:11PM +0200, Tomas Hrcka wrote:
> On Mon, 2012-10-08 at 11:16 -0400, Matt Wagner wrote:
> > On Mon, Oct 08, 2012 at 02:00:14PM +0200, Tomáš Hrčka wrote:
> > > +  validates_format_of :username, :with => /[a-z0-9A-Z]/, :message => I18n.t("users.flash.warning.invalid_username")
> > 
> > I'm going to comment on the bug itself and question whether it's even a
> > valid bug, but I'm not a fan of this particular implementation.
> > 
> > This would keep you from being thrčka ('č' is not "valid"), me from being
> > 'matt.wagner' (it uses a verboten '.'), or someone who speaks Chinese or
> > Russian or any other language with a non-Latin alphabet from even being
> > able to use their name in their username. To me, that is far worse than
> > being able to create a username named '.'.
> 
> You are right, now when I think about it it's all wrong. What about
> validation at least first char is not in .~!@#$%^&*()

What irks me about this "bug" is that there's no clear solution, at
least not that I can see. I think it's really an RFE, but it's an RFE
that's too vague for us to be able to implement. Personally I would vote
that we just close this thing, but that's not my call.

I can't think of a case where '#' or '&' would ever be a reasonable
character in a username. But if some weird customer decides to that they
want to conserve user accounts, there's no reason that 'Tomas & Matt'
can't be a valid username, except that we think it's strange.

I think a '.' or '_' is fine in the middle of a username, but not at
either end. 'matt.wagner' is a normal username, but '.matt.' would be
really weird.

But at the end of the day, I don't know that it's our place to make up
arbitrary username rules. If someone wants to create a user named
'.~!@#$%^&*()' or '☃☃⚔ Fluffy Bunnies ♿♚♚' there is no technical reason
to prevent it.

-- Matt '☃☃⚔ Fluffy Bunnies ♿♚♚' Wagner



More information about the aeolus-devel mailing list