]> granicus.if.org Git - python/commit
SF bug [#456252] Python should never stomp on [u]intptr_t.
authorTim Peters <tim.peters@gmail.com>
Wed, 29 Aug 2001 21:37:10 +0000 (21:37 +0000)
committerTim Peters <tim.peters@gmail.com>
Wed, 29 Aug 2001 21:37:10 +0000 (21:37 +0000)
commit79248aa1e4a8f6510b1f0ef95dc9592d51e16d6c
tree8bdd40a12e1a9c279a7cd0741333a4a1e00c2bf6
parent1936745668f44dd121b50fef31a70efc07cc4ebf
SF bug [#456252] Python should never stomp on [u]intptr_t.
pyport.h:  typedef a new Py_intptr_t type.
    DELICATE ASSUMPTION:  That HAVE_UINTPTR_T implies intptr_t is
    available as well as uintptr_t.  If that turns out not to be
    true, things must get uglier (C99 wants both, so I think it's
    an assumption we're *likely* to get away with).
thread_nt.h, PyThread_start_new_thread:  MS _beginthread is documented
    as returning unsigned long; no idea why uintptr_t was being used.
Others:  Always use Py_[u]intptr_t, never [u]intptr_t directly.
Include/pyport.h
Modules/posixmodule.c
Modules/socketmodule.c
PC/msvcrtmodule.c
PC/pyconfig.h
Python/thread_nt.h