Issue #15894: Document why we don't worry about re-acquiring the
global import lock after forking.
This commit is contained in:
parent
1e1e8aa507
commit
e4710cfced
|
@ -202,8 +202,11 @@ _PyImport_ReInitLock(void)
|
|||
if (import_lock_level > 1) {
|
||||
/* Forked as a side effect of import */
|
||||
long me = PyThread_get_thread_ident();
|
||||
PyThread_acquire_lock(import_lock, 0);
|
||||
/* XXX: can the previous line fail? */
|
||||
/* The following could fail if the lock is already held, but forking as
|
||||
a side-effect of an import is a) rare, b) nuts, and c) difficult to
|
||||
do thanks to the lock only being held when doing individual module
|
||||
locks per import. */
|
||||
PyThread_acquire_lock(import_lock, NOWAIT_LOCK);
|
||||
import_lock_thread = me;
|
||||
import_lock_level--;
|
||||
} else {
|
||||
|
|
Loading…
Reference in New Issue