[apparmor] [patch 1/3] utils: add base and capability rule classes

[Christian Boltz]

Hello,

Am Mittwoch, 3. Dezember 2014 schrieb Steve Beattie:
> This patch adds four classes - two "base" classes and two specific for
> capabilities:

> +    def __init__(self, cap_list, audit=False, deny=False,
...
> +        if cap_list == CapabilityRule.ALL:
> +            self.all_caps = True
> +            self.capability = set()
> +        else:
> +            if type(cap_list) == str:
> +                self.capability = {cap_list}

I still think using split() on the string would make sense ;-)
It also gives us some additional safety and flexibility more or less 
"for free" (for example, maybe someone uses "capability (.*)," as cheap 
parser and hands over the match to __init__())

(Since this is the only to-be-discussed part of the patch, we can 
postpone it for some days and do a small separate patch for it.)

> +            elif type(cap_list) == list and len(cap_list) > 0:
> +                self.capability = set(cap_list)
> +            else:
> +                raise AppArmorBug('Passed unknown object to
> CapabilityRule: %s' % str(cap_list)) 


> +def parse_capability(raw_rule):
> +        '''parse raw_rule and return CapabilityRule'''
> +
> +        matches = RE_PROFILE_CAP.search(raw_rule)

Too much whitespace, please move the code in parse_capability() one 
level to the left.


With the whitespace changed and optionally using split() in __init__() 
for strings,
    Acked-by: Christian Boltz <apparmor at cboltz.de>

Needless to mention that large parts of this patch were originally 
written by me, so another review won't hurt ;-)


Regards,

Christian Boltz
-- 
> Du kennst den Spruch: Nach dem Release.  [Ich sage absichtlich nicht, 
> wie groß der Zeitabstand zum Release sein wird ;-) ]
In zeitlicher Reihenfolge:
Vor Longhorn,
zwischen Longhorn und der Apokalypse,
zwischen der Apokalypse und Hurd,
oder nach Hurd?
[> Christian Boltz und Ratti in fontlinge-devel]