[PATCH 0/3] Password and passphrase patches

David Shea dshea at redhat.com
Mon May 12 13:53:27 UTC 2014


On 05/09/2014 09:19 PM, Brian C. Lane wrote:
> On Fri, May 09, 2014 at 04:19:19PM -0400, David Shea wrote:
>> The first two patches are for the LUKS passphrase spoke. The conditions checked
>> are the same, but the method is moved from a tooltip in the password entry's
>> secondary icon to a separate GtkLabel. This is because the password entry
>> widgets also have caps-lock-warning set, which also uses the secondary icon,
>> and depending on the order things happened in, the error could get stuck or not
>> set at all. So let's not do that and just put the warnings somewhere else.
>>
>> The third patch allows (but warns about) non-ASCII passwords, since if we let
>> people use non-ASCII characters for the hard drive encryption I guess we don't
>> have much reason to forbid them for login.
>>
>> David Shea (3):
>>    Move more of the passphrase dialog into the glade file.
>>    Use a separate label for passphrase warnings.
>>    Allow non-ASCII characters in passwords (#960837)
>>
>>   pyanaconda/constants.py                       |   2 +
>>   pyanaconda/ui/gui/spokes/lib/passphrase.glade |  68 ++++++++++++---
>>   pyanaconda/ui/gui/spokes/lib/passphrase.py    | 119 +++++++++++++++-----------
>>   pyanaconda/ui/gui/spokes/password.py          |  28 +++++-
>>   pyanaconda/ui/gui/spokes/user.py              |  27 +++++-
>>   pyanaconda/ui/tui/spokes/__init__.py          |  11 ++-
>>   pyanaconda/users.py                           |  11 +--
>>   7 files changed, 188 insertions(+), 78 deletions(-)
>>
>> -- 
>> 1.9.0
> These all look ok to me, but do we really want to make non-ASCII users
> hit done twice? I'd expect that they are already aware of the pitfalls
> and just showing a warning should be enough.
>

That sounds reasonable. In the TUI I'll just print the warning and get 
rid of the YesNoDialog prompt, but in the GUI I realized that there's no 
warning printed for the non-ASCII part if the password is also weak, so 
I'll rework that a little bit.



More information about the anaconda-patches mailing list