lp:~mwhudson/pypy/imported-cli-jit

Branch merges

Propose for merging

No branches dependent on this one.

Related source package recipes

No recipes using this branch. (?)

Create packaging recipe

Related snap packages

Related bugs

Related blueprints

23529. By antocuni on 2010-01-20

tentative checkin: implement WeakValueDictionary in rpython, without using the
implementation in rlib.rweakref. This way it works also with ootype

23528. By antocuni on 2010-01-19

merge r70415 and r70454 from trunk: merge the jit-trace branch and kill a lot
of "we_are_jitted", to make sys.settrace working inside loops

23527. By antocuni on 2010-01-19

don't crash if obj is null

23526. By antocuni on 2010-01-18

make AbstractMethDescr.jitcodes much smaller: instead of storing an entry for
each possible class, we store it only for classes where the method is actually
defined, and then we walk through the superclasses when we need it.

The net result is that the static data is waaay smaller: the exe of
pypy-cli-jit is now 14MB instead of 26MB, and the startup time is 29s instead
of 1m:49s

23525. By antocuni on 2010-01-18

add a new ootype operations 'getsuperclassof', the get the superclass of
objects of type ootype.Class

23524. By antocuni on 2010-01-18

it seems that some benchmarks requires more than 256 input arguments. The
arrays should grow automatically, but for now just set it to a larger number

23523. By antocuni on 2010-01-12

what is not tested is broken...: test&fix for llong_lshift

23522. By antocuni on 2009-12-01

not all the constants needs to be initialized (e.g., instances without fields,
empty lists, etc.). This saves a bit of useless "ldsfld xxx; pop" sequences
during the constant initialization phase

23521. By antocuni on 2009-12-01

aesthetic changes

23520. By antocuni on 2009-12-01

a branch where to keep the cli jit working, while it's being destroyed on trunk :-)