[apparmor] [opensuse-project] Google Summer of Code'13 accepted student
kgupta8592 at gmail.com
Fri Jun 7 00:27:09 UTC 2013
On Jun 7, 2013 2:51 AM, "John Johansen" <john.johansen at canonical.com> wrote:
> On 06/06/2013 01:58 PM, Kshitij Gupta wrote:
> > Hello,
> > @John as suggested I used the configparser module and as it turns out
we do have a problem with using it. Actually in the config files at present
the default section is represented by an empty string (for e.g. in
/etc/apparmor/easyprof.conf ), but configparser needs a none empty section
header and hence raises an Error for the same.
> > I think we should change the format and hence default section headers
should be something like [DEFAULT] instead of an empty string.
> Hrmm, no. easyprof.conf has already shipped, as such unless we have a
really good reason to break backwards compatibility we don't.
> Also easyprof.conf is not the config file you should be looking at, or at
least not initially
> genprof/logprof have used
> logprof.conf and repository.cong
Yes I did try the tools on logprof.conf for which both worked perfectly.
However, I expected the tool to breakdown on the default section case which
it did. There's no reason to have different parsers for different config
Maybe you could work the issue out in AppArmor3. :-)
> > Also, the order in which the config parser writes to output file is
random (expected of a dictionary) and not sorted.
> Does this matter? Generally our configs and policy have been declarative
where order is irrelevant. However if config parser is not the right tool
for the job don't use it. I have never actually used it, I just know it
exists and uses a format that is close if not the same as what we where
using so it is worth looking at. If after looking at it you decide it isn't
the right tool, then just tell us so and don't use it.
No, I too suppose the order does not matter. I was merely stating it.
In my understanding I should work with the previous version, instead of the
one using configparser and will use that, unless somebody gave a solution
for the problem.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the AppArmor