[Python-Dev] run time error anlysis of python source code

[Maciej Fijalkowski]

On Sun, Jun 21, 2009 at 5:05 PM, Daniel Diniz<ajaksu at gmail.com> wrote:
> Hi David,
>
> <omega_force2003 at yahoo.com> wrote:
>>
>> It appears that one possibility of investigation into the development of a
>> safety-critical variant of the python language
>
> There is some interesting work related to a safety-critical variant of Python.
>
>> would be to conduct run time
>> error analysis of the source code that is responsible for producing the
>> python language.
>
> There's been some effort into this too, and the Coverity and Klocwork
> based fixes could also be of interest to you.
>
>> Therefore, I will now conduct these run time error
>> analysis of the python source code as if the python environment itself was
>> to be utilized as a FADEC controller within an aircraft engine.
>
> Nice, what tools do you have available for this? Any papers that would
> be a good start on the topic?
>
>> I
>> have already conducted some analysis already and it appears to be some
>> concern with memory management within Python.  I will redouble my efforts
>> and determine if I am correct and as a promise.  I will give my findings to
>> everyone to enjoy if they so want it.  I will also give the correct source
>> if anyone would want it for their own purposes.  The source code that I will
>> be evaluating is the one responsible for the newer variants of Python 3.0.
>
> You want the py3k branch. BTW, take a look at Brett Cannon's work on
> Python security, as well as tav's.

Is py3k branch even passing all tests on all buildbots all the time? I
don't think
svn head is the right thing to look at. Also, it's worth noting that
most big libraries
are 2.x compatible only.