[Qgis-developer] Sandboxing python

[Nathan Woodrow]

And Qt.

On Mon, Oct 17, 2016 at 8:35 PM, Nathan Woodrow <madmanwoo at gmail.com> wrote:

> Yes I have read it, however we don't run on PyPy, we use CPython.
>
> On Mon, Oct 17, 2016 at 8:06 PM, <geodrinx at gmail.com> wrote:
>
>>
>>
>> Inviato da iPhone
>>
>> Il giorno 17 ott 2016, alle ore 11:19, Nathan Woodrow <
>> madmanwoo at gmail.com> ha scritto:
>>
>> For context from here: https://lwn.net/Articles/574215/
>>
>> "I now agree that putting a sandbox in CPython is the wrong design.
>> There are too many ways to escape the untrusted namespace using the various
>> introspection features of the Python language. To guarantee the [safety] of
>> a security product, the code should be [carefully] audited and the code to
>> review must be as small as possible. Using pysandbox, the "code" is the
>> whole Python core which is a really huge code base. For example, the Python
>> and Objects directories of Python 3.4 contain more than 126,000 lines of C
>> code.
>>
>> The security of pysandbox is the security of its weakest part. A single
>> bug is enough to escape the whole sandbox."
>>
>>
>> Correct.  In fact I am talking about this other:
>>
>> pypy.org  and its sandboxing, that is:
>> http://doc.pypy.org/en/latest/sandbox.html?highlight=Sandboxing
>>
>> that reports the interesting issues you can have with other solutions, as
>> CPython, for example.
>> Do you have read it ?  If not, good reading  :)
>>
>>
>>
>>
>> On Mon, Oct 17, 2016 at 7:17 PM, Nathan Woodrow <madmanwoo at gmail.com>
>> wrote:
>>
>>> Honestly, this is getting tiresome.
>>>
>>> If you don't like the approval processes that we have in place currently
>>> you can create a new plugin repo, it's just an XML file exposed via
>>> webserver, there are docs around for it.
>>>
>>> You can give this URL to people and they can install your plugins via
>>> that.
>>>
>>> If you don't want to do that, then you will have to go through the
>>> approval process.  I'm sure there are reasons it took longer than normal,
>>> maybe review those first.
>>>
>>> If you want to follow the sandboxed Python route and see how far you get
>>> fine, however again I suspect you are in for a long road given the complex
>>> nature of that and you would still have to
>>> be able to support what we can in core, etc.
>>>
>>> - Nathan
>>>
>>>
>>> On Mon, Oct 17, 2016 at 6:31 PM, Geo DrinX <geodrinx at gmail.com> wrote:
>>>
>>>>
>>>>
>>>> 2016-10-17 10:19 GMT+02:00 Nathan Woodrow <madmanwoo at gmail.com>:
>>>>
>>>>> Qgis uses CPython. You also have to sandbox Qt, so I suspect you are
>>>>> running into a lot of dead ends.
>>>>>
>>>> Thank you for the suggestion.  We will see who is moving in dead
>>>> ends.   :)
>>>>
>>>> I am the crow's nest of the ship and I am experiencing the arrival of
>>>> the iceberg.
>>>> And I hear the orchestra playing :)
>>>>
>>>>
>>>> Best regards and wishes for a safe journey.   ;)
>>>>
>>>>
>>>> Geo
>>>>
>>>> PS:  and then if you want to exit from the one direction maze you are
>>>> going,  here I am.
>>>> PPS:  in the meantime, take a look of this competition:
>>>> https://goo.gl/WR8LVF
>>>>
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-developer/attachments/20161017/4c309d3c/attachment-0001.html>