Validating with a scratch code is probably a "giant trap that newbies
could fall into".
Bug: T150824
Change-Id: I5710b151d7682e4cdb0b6a692f7b2c108f051caf
Google authenticator uses a space character as a group seperator.
We can thus expect users to enter such a separator and we should not
fail on that. Might as well trim whitespace too, as that is another
oft occuring user input mistake, that should not affect functionality.
Bug: T150603
Change-Id: I7334ed5dfaf933e61831438e2f86aa979cf9f51b
Extend the token validation failure checks introduced in I4884f6e to the
other interactions where OATHAuthKey::verifyToken is used.
Depends-On: Ia3add8bbbab0307f036e9b77e752c382da3a0d04
Change-Id: Icbe5cdf561c683dc971a099d61cedff311b26b43
We need to pass the db name to getConnection, in addition to wfGetLB.
Also, use core's CentralIdLookup for mapping local user to CentralId
when using a central DB for OATH secret storage.
Change-Id: I12a457633956a9a34dc5302ddcff468e31dd9cef
Add configuration variable for specifying what database the OATH
credentials are stored in, that way wikis that use CentralAuth can
centralize their two-factor authentication data as well.
Bug: T100374
Change-Id: I285e2fe29fee43ddc6c5a6e51823911d43c596f6
Made new class ProxySpecialPage, which acts as a
proxy object to another SpecialPage object that is
determined based on context information other than
the title.
Then Special:OATH has been split into two separate
special page classes (both FormSpecialPages using
HTMLForm) that are routed to by a ProxySpecialPage
object.
In addition, the form for enabling two-factor auth
has been refactored into vform style, with some
better instructions on how to enable two-factor
authentication.
Change-Id: Ib9117cbc9d7f044de9607db81a157e1b472b5ec0
This takes out the actual key information from
OATHUser and puts it into an OATHKey class, which OATHUser
depends on. This allows easily swapping keys in/out from
a user.
Change-Id: Ife5f1bae4ad65b66c5e20017cc43c0576b4aba19
