1. Michael McGlynn [Atlassian]
  2. pypy

Commits

Show all
Author Commit Message Date Builds
Brian Kearns
these imports are still needed
Brian Kearns
add a fast path to this check
Brian Kearns
some more behavioral fixes for datetime
Brian Kearns
use slots for datetime objects, test
Brian Kearns
believe this test_rarithmetic approach should work (but doesn't)
Alex Gaynor
dead import removal + pep8
Alex Gaynor
remove some dead imports, some PEP8, general cleanups
Brian Kearns
try to fix test_rarithmetic again, this time on 64bit
Brian Kearns
fix thread+fork+signals after _signalsenabled dict removal
Brian Kearns
fix the test: this now raises a thread.error
Brian Kearns
make this test failure more obvious
Maciej Fijalkowski
merge
Maciej Fijalkowski
fix
Alex Gaynor
(alex, arigo) Make sure this value is immutable, otherwise traces are suboptimal
Armin Rigo
Kill the _signalsenabled dictionary. Replace it by a value directly on the ExecutionContext. This avoids stm issues.
Armin Rigo
merge heads
Armin Rigo
Don't name the SignalActionFlag explicitly here.
Brian Kearns
wrap the exception in signals_exit
Brian Kearns
mark these unroll safe also
Brian Kearns
try to clean up more on fork
Armin Rigo
Update to cffi/70927696eb9c
Brian Kearns
clean up the thread's signals enabled count on thread exit
Brian Kearns
prevent signals from being disabled twice
Brian Kearns
fix the failing test by still keeping track of mainthreadident so we know if the forking thread should get an enable_signals point for becoming main
Brian Kearns
a failing test for the signals_enabled context manager's fork logic
Brian Kearns
update whatsnew for signal-and-thread branch merge
Brian Kearns
improve the numpypy concatenate test
Brian Kearns
clean up rstruct.ieee 80-bit float packing and improve tests
Brian Kearns
modify this test so it would have failed untranslated without 1de5c08faeca
Brian Kearns
add a modified test of the signals_enabled context manager that works untranslated
  1. Prev
  2. Next