Commits

Author Commit Message Labels Comments Date
Brian Kearns
store next_row on the cursor not the statement
mattip
merge heads
mattip
merge heads
mattip
fix crashes by disallowing astype(numeric_type) for str arrays, allow astype(str) for numeric arrays since it was implemented previously
mattip
add tests that expose crashes
Armin Rigo
Move gc_stack_bottom() around to after the GIL is acquired and the
Armin Rigo
Add a check (for lldebug builds) which points directly to where we forget to close the stack.
Brian Kearns
remove test of cursor.__del__
Brian Kearns
only need to check in_use if it isn't a new statement
Armin Rigo
(fijal, arigo around) try to fix statements going away under our feet.
Maciej Fijalkowski
merge
Maciej Fijalkowski
fix packaging to include .so
Brian Kearns
these seem mistakenly hg added
Brian Kearns
these aren't necessary
Brian Kearns
merge default
Branches
py3k
Maciej Fijalkowski
merge default
Branches
release-2.0-beta2
Brian Kearns
enhance this test to demonstrate flaws in the initial fix
Brian Kearns
take care about next_char's memory in a couple more places
Brian Kearns
a more proper fix than the one applied in c1c610355c04
Maciej Fijalkowski
write some tests
Maciej Fijalkowski
detect BEGIN. (CPython does check of statement type)
Maciej Fijalkowski
We need to keep alive the sql statement for a bit longer, since next_char
Brian Kearns
more sqlite adaptations for cffi
Maciej Fijalkowski
cffi does not return none from externals (but can return NULL),
Carl Meyer
Fix home and user distutils installation schemes. (issue1272)
Brian Kearns
redo more changes lost in sqlite-cffi merge
Brian Kearns
redo another change lost in sqlite-cffi merge
Brian Kearns
merge heads
Brian Kearns
slightly more correct, may matter in obscure or future cases
Maciej Fijalkowski
merge
  1. Prev
  2. Next