[apparmor] [patch] move regexes from aa.py to regex.py

Christian Boltz apparmor at cboltz.de
Tue Nov 11 23:30:11 UTC 2014


Hello,

Am Mittwoch, 12. November 2014 schrieb Kshitij Gupta:
> On Wed, Nov 12, 2014 at 1:06 AM, Christian Boltz wrote:
> > this patch moves the profile parsing regexes from aa.py to a new
> > file, regex.py, and adds an "import" line so that they are still
> > available in aa.py.
> > 
> > This is needed to avoid circular dependencies in the rule classes
> > (aa.py will need to import the rule classes, which also means they
> > can't import something from aa.py)
> > 
> > 
> Its actually a good thing, the rule classes should be as stand-alone 
> as possible (specially not dependent on the aa.py, which is arguably 
> the most unstable piece of code in there).

Indeed, that's a nice side effect ;-)

> > === modified file 'utils/apparmor/aa.py'
> > --- utils/apparmor/aa.py        2014-11-09 00:33:40 +0000
> > +++ utils/apparmor/aa.py        2014-11-10 23:27:53 +0000
> > @@ -40,8 +41,9 @@
> > 
> >                               mode_to_str_user, mode_contains,
> >                               AA_OTHER,
> >                               flatten_mode, owner_flatten_mode)
> > 
> > +from apparmor.regex import *
> 
> Hmm would it be better to import the regex's in a seperate namespace?

I wanted to make this patch as small and simple as possible and 
therefore decided to import the regexes into the main namespace.

Using another namespace would be another patch (with _lots of_ little 
changes). However, I'm more focusing on the rule classes currently so I 
won't do this in the next weeks.

> > Too many empty lines at the end of file. Are they needed?

Not really. I removed them before commiting.


Regards,

Christian Boltz
-- 
> > > Nochmal: Insgesamt macht das PDF einen guten Eindruck!
> > Gell? Bin ja auch stolz wie Oskar :=)
> Zu Recht!
Oh, danke! *erroet* *verbeug*
[> Christian Boltz und David Haller in suse-linux-faq]




More information about the AppArmor mailing list