[apparmor] GSoC r13, r14, r15 review

[Christian Boltz]

Hello,

Am Dienstag, 9. Juli 2013 schrieb Kshitij Gupta:
> >> I somehow doubt this is intentional - the test should catch this
> >> exception ;-)
> 
> As it turns this issue has got something to do with the Python3's new
> feature called Exception Chaining.
> Its another of those Python2-3 problem.
> An additional line needs to be incorporated to check if Python2 or 3
> and raise accordingly. :/

:-(

> > It was inside the try except block of test, I got it out to try how
> > exception looked.
> > Turns out \n\t aren't working yet. :-\
> > 
> >> Also note that the last line contains \n\t - this should become a
> >> real line break and tab in the output...
> 
> I did not notice until later, but the AppArmorException uses repr() to
> print the description and hence its not possible to print the nicely
> formatted exception with the line number in new line.

I noticed you removed repr() in r16 - does this fix the issue?
(a short test tells me the answer is probably "no")

> I feel I should use the error(), it will print something like this and
> terminate program:
> 
> ERROR: No severity value present in file: severity_broken.db
>     [Line 14]: CAP_SYS_MODULE
> What would you say? move to error() instead of raising
> AppArmorException?

Can error() be catched with try/except like an exception? For example, 
we'll need this to display a nice error dialog in YaST. 
(If yes, then using error() is OK.)


Regards,

Christian Boltz
-- 
The normal user is happy with openSUSE because: [...]
  - openSUSE isn't a religion (like a few others);
What users blame us is for lack of customization :)
[Nelson Marques in opensuse-factory]