1. Pypy
  2. Untitled project
  3. pypy

Commits

Author Commit Message Date Builds
Maciej Fijalkowski
close about to be merged branch
Brian Kearns
null merge default past the backout of this branch
Brian Kearns
merge default up to 295a4447f6ba
Antonio Cuni
close to-be-merged branch
Antonio Cuni
hg merge default
Antonio Cuni
bah, more tests in which we need to convince the annotator to see the constructor of VRawBufferValue
Antonio Cuni
fix after the renaming
Antonio Cuni
fix imports after the merge
Antonio Cuni
hg merge default
Antonio Cuni
all_virtuals is no longer a list, but an instance of VirtualCache now
Antonio Cuni
replace all the references to virtual_ref_cache&co. with virtual_ptr_cache&co., to avoid confusions with virtualrefs, which are (obviously :)) another thing
Antonio Cuni
implement bh_new_raw_buffer
Antonio Cuni
now we can implement unpack_arraydescr_size for real
Antonio Cuni
fix
Antonio Cuni
hg merge default. IN-PROGRESS: a couple of things broke because of remove-globals-in-jit, need to fix them
Antonio Cuni
bah, we need to play this dance to fix the annotator in a couple of tests in which VRawBuffer would never be instantiated, thus leading to blocked blocks
Antonio Cuni
fix
Antonio Cuni
we cannot move these methods here, else we get an annotation error
Antonio Cuni
fix translation
Antonio Cuni
fix the tests in test_optimizebasic, which failed because Whatever() compared == INT even when it should not
Antonio Cuni
rpython fix
Antonio Cuni
refactor virtuals_cache and virtuals_int_cache: put them in a separate class because it is easier to handle and to make it rpython
Antonio Cuni
tweak the test to make it failing because of raw virtuals, and fix the code
Antonio Cuni
rpython fix: this can never happen
Antonio Cuni
implement bh_new_raw_buffer for ll cpus: this makes x86 virtualref tests passing
Antonio Cuni
move raw virtuals tests to VirtualMiscTests, so that they can be run also by the x86 backend tests
Antonio Cuni
tentative RPython fix: in ResumeDataDirectReader virtuals_int_cache contains real integers, while in ResumeDataBox reader it contains boxes, so we cannot initialize it by 0
Antonio Cuni
kill support for rawbuffer virtualstate: we do not need to have it for optimizing cffi, and it seems to cause bugs
Antonio Cuni
failing test and fix for forcing raw virtuals after GUARD_NOT_FORCE
Antonio Cuni
apply the raw virtuals optimization only if we are allocating an array of chars. With other types it's harder because the optimizer does not know the size of the items, and it's not useful for cffi anyway
  1. Prev
  2. Next