Merge pull request #630 from kmod/gc_fixes

fix two minor bugs

from_cpython/Include/stringobject.h

@@ -106,7 +106,14 @@ PyAPI_FUNC(int) _PyString_CheckInterned(PyObject *) PYSTON_NOEXCEPT;
 /* Macro, trading safety for speed */
 // Pyston changes: these aren't direct macros any more [they potentially could be though]
 #define PyString_AS_STRING(op) PyString_AsString((PyObject*)op)
-#define PyString_GET_SIZE(op)  PyString_Size((PyObject*)op)
+// Note: there are buggy extension modules (unicodedata.c) that rely on the fact that
+// PyString_GET_SIZE does *not* have the same behavior as PyString_Size.  In particular,
+// you can get away with calling PyString_GET_SIZE on a unicode object and getting the
+// length of the unicode string, not the length of the bytes it encodes to in the default
+// encoding.
+// So, set up a different function for those callers to use.
+PyAPI_FUNC(Py_ssize_t) _PyString_SizeMacro(PyObject *) PYSTON_NOEXCEPT;
+#define PyString_GET_SIZE(op)  _PyString_SizeMacro((PyObject*)op)
 //#define PyString_AS_STRING(op) (((PyStringObject *)(op))->ob_sval)
 //#define PyString_GET_SIZE(op)  Py_SIZE(op)
 

src/capi/codecs.cpp

@@ -744,14 +744,10 @@ static int _PyCodecRegistry_Init(void) {
     if (interp->codec_search_path != NULL)
         return 0;
 
-    interp->codec_search_path = PyList_New(0);
-    interp->codec_search_cache = PyDict_New();
-    interp->codec_error_registry = PyDict_New();
-
     // Pyston change: register roots
-    gc::registerPermanentRoot(interp->codec_search_path);
-    gc::registerPermanentRoot(interp->codec_search_cache);
-    gc::registerPermanentRoot(interp->codec_error_registry);
+    interp->codec_search_path = PyGC_AddRoot(PyList_New(0));
+    interp->codec_search_cache = PyGC_AddRoot(PyDict_New());
+    interp->codec_error_registry = PyGC_AddRoot(PyDict_New());
 
     if (interp->codec_error_registry) {
         for (i = 0; i < sizeof(methods) / sizeof(methods[0]); ++i) {

src/runtime/str.cpp

@@ -2349,6 +2349,16 @@ extern "C" Py_ssize_t PyString_Size(PyObject* op) noexcept {
     return len;
 }
 
+extern "C" Py_ssize_t _PyString_SizeMacro(PyObject* op) noexcept {
+    if (PyString_Check(op))
+        return static_cast<BoxedString*>(op)->size();
+
+    if (PyUnicode_Check(op))
+        return Py_SIZE(op);
+
+    RELEASE_ASSERT(0, "Need to verify the behavior of PyString_GET_SIZE on %s objects", op->cls->tp_name);
+}
+
 extern "C" int _PyString_Resize(PyObject** pv, Py_ssize_t newsize) noexcept {
     // This is only allowed to be called when there is only one user of the string (ie a refcount of 1 in CPython)