Hello.
We have just finished probably the smallest sprint ever
in PyPy history. For most of the time it was just me
and Armin pairing together.
We also had a chance to work a bit with people from
the University, but there were definitely not enough
core developers to organize the work in a reasonable
manner. At some point we ended up having two pairs containing
four people each.
Jakub and Bartosz (who were our gentle hosts) worked
on getting PyPy's sandbox integrated with django.
It's still just an example what you can do (ie you
can do much more), but it's already interesting to look
at. The code can be found in user dir. This server (not yet online anywhere, sorry)
is able to run untrusted python code provided by user inside
a fully configurable sandbox.
We also implemented missing peepholer optimizations from
CPython, finding out that some peepholer tests were failing,
just because PyPy is optimizing better :-)
The main part of the sprint was work on JIT (most notable the fifth
generation of the JIT), which was moved
from the obscure directory in Carl's user in svn (which contains
branches these days!) into a PyPy branch. It's still very much
work in progress and a lot of pen and paper or handwaving was
involved, but we were able to implement a lot of basics in record time.
Right now we need a lot of rest after the exhaustive sprint,
but after that, stay tuned for more information about
progressing JIT!
Cheers,
fijal
Friday, February 13, 2009
Wroclaw 2009 sprint progress report
Hello.
We have just finished probably the smallest sprint ever in PyPy history. For most of the time it was just me and Armin pairing together.
We also had a chance to work a bit with people from the University, but there were definitely not enough core developers to organize the work in a reasonable manner. At some point we ended up having two pairs containing four people each.
Jakub and Bartosz (who were our gentle hosts) worked on getting PyPy's sandbox integrated with django. It's still just an example what you can do (ie you can do much more), but it's already interesting to look at. The code can be found in user dir. This server (not yet online anywhere, sorry) is able to run untrusted python code provided by user inside a fully configurable sandbox.
We also implemented missing peepholer optimizations from CPython, finding out that some peepholer tests were failing, just because PyPy is optimizing better :-)
The main part of the sprint was work on JIT (most notable the fifth generation of the JIT), which was moved from the obscure directory in Carl's user in svn (which contains branches these days!) into a PyPy branch. It's still very much work in progress and a lot of pen and paper or handwaving was involved, but we were able to implement a lot of basics in record time.
Right now we need a lot of rest after the exhaustive sprint, but after that, stay tuned for more information about progressing JIT!
Cheers,
fijal
We have just finished probably the smallest sprint ever in PyPy history. For most of the time it was just me and Armin pairing together.
We also had a chance to work a bit with people from the University, but there were definitely not enough core developers to organize the work in a reasonable manner. At some point we ended up having two pairs containing four people each.
Jakub and Bartosz (who were our gentle hosts) worked on getting PyPy's sandbox integrated with django. It's still just an example what you can do (ie you can do much more), but it's already interesting to look at. The code can be found in user dir. This server (not yet online anywhere, sorry) is able to run untrusted python code provided by user inside a fully configurable sandbox.
We also implemented missing peepholer optimizations from CPython, finding out that some peepholer tests were failing, just because PyPy is optimizing better :-)
The main part of the sprint was work on JIT (most notable the fifth generation of the JIT), which was moved from the obscure directory in Carl's user in svn (which contains branches these days!) into a PyPy branch. It's still very much work in progress and a lot of pen and paper or handwaving was involved, but we were able to implement a lot of basics in record time.
Right now we need a lot of rest after the exhaustive sprint, but after that, stay tuned for more information about progressing JIT!
Cheers,
fijal
Wroclaw 2009 sprint progress report
Subscribe to:
Posts (Atom)