lp:~mwhudson/pypy/imported-builtin-call-speedup

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

13613. By mwh on 2006-12-06

change some "if isinstances" to "assert isinstances" and delete some code.
now a pypy-c-fast is 2-4% faster than pypy-c-multi

13612. By mwh on 2006-12-05

fix a typo that prevented the WaryDictImplementation from being used, and
another that prevented it from working.
comment out a print.

13611. By cfbolz on 2006-12-02

new attempt to use the multidict instead of a new dictionary implementation.
Makes things less bad, but still no net speed gain.

13610. By mwh on 2006-11-29

(cfbolz, mwh)
small rewrites for correctness and translatability.

13609. By cfbolz on 2006-11-28

slightly embarassing: of course this file is need to make anything actually
work

13608. By cfbolz on 2006-11-25

first round of trying to make builtin lookups faster (nearly as fast as a local
lookup) while still behaving correctly when builtins are shadowed.

13607. By cfbolz on 2006-11-25

a branch to experiment with the speedup to calls of unshadowed builtins

13606. By cfbolz on 2006-11-25

use _mixin_ instead of manually creating the effect.

13605. By pedronis on 2006-11-25

match the changed resume_points

13604. By pedronis on 2006-11-25

(arre, arigo around, pedronis)

now that we can return multiple states from a function, try to avoid forcing virtual structs at merge points.

Still force merges when returning from the portal.

Now test_virt_obj_method_call_promote passes, als test_factorial for the tlc only promotes once and produces
a box at the end for the result, which is good.