PEP 489: Multi-phase extension module initialization
Known limitations of the current implementation:
- documentation changes are incomplete
- there's a reference leak I haven't tracked down yet
The leak is most visible by running:
./python -m test -R3:3 test_importlib
However, you can also see it by running:
./python -X showrefcount
Importing the array or _testmultiphase modules, and
then deleting them from both sys.modules and the local
namespace shows significant increases in the total
number of active references each cycle. By contrast,
with _testcapi (which continues to use single-phase
initialisation) the global refcounts stabilise after
a couple of cycles.
diff --git a/Python/dynload_next.c b/Python/dynload_next.c
index 85c95b4..c555b38 100644
--- a/Python/dynload_next.c
+++ b/Python/dynload_next.c
@@ -27,8 +27,9 @@
#define LINKOPTIONS NSLINKMODULE_OPTION_BINDNOW| \
NSLINKMODULE_OPTION_RETURN_ON_ERROR|NSLINKMODULE_OPTION_PRIVATE
#endif
-dl_funcptr _PyImport_GetDynLoadFunc(const char *shortname,
- const char *pathname, FILE *fp)
+dl_funcptr _PyImport_FindSharedFuncptr(const char *prefix,
+ const char *shortname,
+ const char *pathname, FILE *fp)
{
dl_funcptr p = NULL;
char funcname[258];
@@ -39,7 +40,7 @@
const char *errString;
char errBuf[512];
- PyOS_snprintf(funcname, sizeof(funcname), "_PyInit_%.200s", shortname);
+ PyOS_snprintf(funcname, sizeof(funcname), "_%20s_%.200s", prefix, shortname);
#ifdef USE_DYLD_GLOBAL_NAMESPACE
if (NSIsSymbolNameDefined(funcname)) {