[pypy-dev] Online Python interpreter on top of PyPy
kracethekingmaker at gmail.com
Fri Mar 9 18:32:32 CET 2012
> On Thu, Mar 8, 2012 at 16:12, Maciej Fijalkowski<fijall at gmail.com> wrote:
>>> Sorry, I did some mistake with git stash pop, I apologise. so whole system
>>> was corrupted, Thanks for your time. Now system is up. Kindly try.
> It looks like a nice demo :-) The system has limits of 200MB or 60
> seconds. For a pypy-sandbox, the only thing that needs to be done
> carefully is what occurs if a lot of requests come at the same time.
> Does your demo run only one of them and queue or discard the other
Thanks :-) . System has 175MB and 60 seconds timeout as well. I run the
flask site using gunicorn, the command goes
gunicorn pylive:app --workers=2 --access-logfile
/path/to/pylive/error_log.log --log-level debug --timeout=120 --daemon
Currently total worker space is around 50MB for 2 workers, 175 MB for
single pypy-sandbox , so 50 + 250MB(300MB), so if server receives more
than 2 concurrent request, it should be handled by gunicorn(I might be
wrong). Yes there are chances request will get lost, I am working on
this area, to separate app and sandbox to different system and introduce
queuing system, I will be increasing no of workers as well as memory.
>> you don't correctly escape the output. So a print '<h1>' would do
>> exactly that ;-)
> Indeed, it's annoying when you try to print e.g. 'sys.modules'. But
> it's not really a security bug.
I have fixed parsing html tags.
Yes its not security bug since its impossible to write to particular
location in filesystem and read.
The below code also fails
with open('a.txt', 'w') as f:
f.write('I am hacker')
with open('a.txt', 'r') as f:
> A bientôt,
"Talk is cheap, show me the code" -- Linus Torvalds
More information about the pypy-dev