diff -r f89995a4ec11 Include/codecs.h --- a/Include/codecs.h Sat Sep 06 22:47:58 2014 +0300 +++ b/Include/codecs.h Sun Sep 07 08:26:22 2014 +1000 @@ -49,6 +49,10 @@ PyAPI_FUNC(PyObject *) _PyCodec_Lookup( const char *encoding ); + +PyAPI_FUNC(int) _PyCodec_Forget( + const char *encoding + ); #endif /* Codec registry encoding check API. diff -r f89995a4ec11 Lib/test/test_codecs.py --- a/Lib/test/test_codecs.py Sat Sep 06 22:47:58 2014 +0300 +++ b/Lib/test/test_codecs.py Sun Sep 07 08:26:22 2014 +1000 @@ -2578,6 +2578,14 @@ return _TEST_CODECS.get(codec_name) codecs.register(_get_test_codec) # Returns None, not usable as a decorator +try: + # Issue #22166: Also need to clear the internal cache in CPython + from _codecs import _forget_codec +except ImportError: + def _forget_codec(codec_name): + pass + + class ExceptionChainingTest(unittest.TestCase): def setUp(self): @@ -2603,6 +2611,12 @@ def tearDown(self): _TEST_CODECS.pop(self.codec_name, None) + # Issue #22166: Also pop from caches to avoid appearance of ref leaks + encodings._cache.pop(self.codec_name, None) + try: + _forget_codec(self.codec_name) + except KeyError: + pass def set_codec(self, encode, decode): codec_info = codecs.CodecInfo(encode, decode, diff -r f89995a4ec11 Modules/_codecsmodule.c --- a/Modules/_codecsmodule.c Sat Sep 06 22:47:58 2014 +0300 +++ b/Modules/_codecsmodule.c Sun Sep 07 08:26:22 2014 +1000 @@ -42,6 +42,12 @@ #include #endif +/*[clinic input] +module _codecs +[clinic start generated code]*/ +/*[clinic end generated code: output=da39a3ee5e6b4b0d input=e1390e3da3cb9deb]*/ + + /* --- Registry ----------------------------------------------------------- */ PyDoc_STRVAR(register__doc__, @@ -134,6 +140,53 @@ /* --- Helpers ------------------------------------------------------------ */ +/*[clinic input] +_codecs._forget_codec + + encoding: str + / + +Purge the named codec from the internal codec lookup cache +[clinic start generated code]*/ + +PyDoc_STRVAR(_codecs__forget_codec__doc__, +"_forget_codec($module, encoding, /)\n" +"--\n" +"\n" +"Purge the named codec from the internal codec lookup cache"); + +#define _CODECS__FORGET_CODEC_METHODDEF \ + {"_forget_codec", (PyCFunction)_codecs__forget_codec, METH_VARARGS, _codecs__forget_codec__doc__}, + +static PyObject * +_codecs__forget_codec_impl(PyModuleDef *module, const char *encoding); + +static PyObject * +_codecs__forget_codec(PyModuleDef *module, PyObject *args) +{ + PyObject *return_value = NULL; + const char *encoding; + + if (!PyArg_ParseTuple(args, + "s:_forget_codec", + &encoding)) + goto exit; + return_value = _codecs__forget_codec_impl(module, encoding); + +exit: + return return_value; +} + +static PyObject * +_codecs__forget_codec_impl(PyModuleDef *module, const char *encoding) +/*[clinic end generated code: output=a75e631591702a5c input=18d5d92d0e386c38]*/ +{ + if (_PyCodec_Forget(encoding) < 0) { + return NULL; + }; + Py_RETURN_NONE; +} + static PyObject *codec_tuple(PyObject *unicode, Py_ssize_t len) @@ -1168,6 +1221,7 @@ register_error__doc__}, {"lookup_error", lookup_error, METH_VARARGS, lookup_error__doc__}, + _CODECS__FORGET_CODEC_METHODDEF {NULL, NULL} /* sentinel */ }; diff -r f89995a4ec11 Python/codecs.c --- a/Python/codecs.c Sat Sep 06 22:47:58 2014 +0300 +++ b/Python/codecs.c Sun Sep 07 08:26:22 2014 +1000 @@ -185,6 +185,32 @@ return NULL; } +int _PyCodec_Forget(const char *encoding) +{ + PyInterpreterState *interp; + PyObject *v; + int result; + + interp = PyThreadState_GET()->interp; + if (interp->codec_search_path == NULL) { + return -1; + } + + /* Convert the encoding to a normalized Python string: all + characters are converted to lower case, spaces and hyphens are + replaced with underscores. */ + v = normalizestring(encoding); + if (v == NULL) { + return -1; + } + + /* Drop the named codec from the internal cache */ + result = PyDict_DelItem(interp->codec_search_cache, v); + Py_DECREF(v); + + return result; +} + /* Codec registry encoding check API. */ int PyCodec_KnownEncoding(const char *encoding)