cpython/Python/importdl.c

251 lines
7.4 KiB
C
Raw Normal View History

/* Support for dynamic loading of extension modules */
#include "Python.h"
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928) We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules. The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings). https://bugs.python.org/issue46541#msg411799 explains the rationale for this change. The core of the change is in: * (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros * Include/internal/pycore_runtime_init.h - added the static initializers for the global strings * Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState * Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config. The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *. The following are not changed (yet): * stop using _Py_IDENTIFIER() in the stdlib modules * (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API * (maybe) intern the strings during runtime init https://bugs.python.org/issue46541
2022-02-08 16:39:07 -04:00
#include "pycore_call.h"
#include "pycore_pystate.h"
#include "pycore_runtime.h"
/* ./configure sets HAVE_DYNAMIC_LOADING if dynamic loading of modules is
supported on this platform. configure will then compile and link in one
of the dynload_*.c files, as appropriate. We will call a function in
those modules to get a function pointer to the module's init function.
*/
#ifdef HAVE_DYNAMIC_LOADING
#include "importdl.h"
#ifdef MS_WINDOWS
extern dl_funcptr _PyImport_FindSharedFuncptrWindows(const char *prefix,
const char *shortname,
PyObject *pathname,
FILE *fp);
#else
extern dl_funcptr _PyImport_FindSharedFuncptr(const char *prefix,
const char *shortname,
const char *pathname, FILE *fp);
#endif
1997-11-22 17:53:48 -04:00
static const char * const ascii_only_prefix = "PyInit";
static const char * const nonascii_prefix = "PyInitU";
/* Get the variable part of a module's export symbol name.
* Returns a bytes instance. For non-ASCII-named modules, the name is
* encoded as per PEP 489.
* The hook_prefix pointer is set to either ascii_only_prefix or
* nonascii_prefix, as appropriate.
*/
static PyObject *
get_encoded_name(PyObject *name, const char **hook_prefix) {
PyObject *tmp;
PyObject *encoded = NULL;
PyObject *modname = NULL;
Py_ssize_t name_len, lastdot;
/* Get the short name (substring after last dot) */
name_len = PyUnicode_GetLength(name);
if (name_len < 0) {
return NULL;
}
lastdot = PyUnicode_FindChar(name, '.', 0, name_len, -1);
if (lastdot < -1) {
return NULL;
} else if (lastdot >= 0) {
tmp = PyUnicode_Substring(name, lastdot + 1, name_len);
if (tmp == NULL)
return NULL;
name = tmp;
/* "name" now holds a new reference to the substring */
} else {
Py_INCREF(name);
}
/* Encode to ASCII or Punycode, as needed */
encoded = PyUnicode_AsEncodedString(name, "ascii", NULL);
if (encoded != NULL) {
*hook_prefix = ascii_only_prefix;
} else {
if (PyErr_ExceptionMatches(PyExc_UnicodeEncodeError)) {
PyErr_Clear();
encoded = PyUnicode_AsEncodedString(name, "punycode", NULL);
if (encoded == NULL) {
goto error;
}
*hook_prefix = nonascii_prefix;
} else {
goto error;
}
}
/* Replace '-' by '_' */
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928) We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules. The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings). https://bugs.python.org/issue46541#msg411799 explains the rationale for this change. The core of the change is in: * (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros * Include/internal/pycore_runtime_init.h - added the static initializers for the global strings * Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState * Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config. The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *. The following are not changed (yet): * stop using _Py_IDENTIFIER() in the stdlib modules * (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API * (maybe) intern the strings during runtime init https://bugs.python.org/issue46541
2022-02-08 16:39:07 -04:00
modname = _PyObject_CallMethod(encoded, &_Py_ID(replace), "cc", '-', '_');
if (modname == NULL)
goto error;
Py_DECREF(name);
Py_DECREF(encoded);
return modname;
error:
Py_DECREF(name);
Py_XDECREF(encoded);
return NULL;
}
PyObject *
_PyImport_LoadDynamicModuleWithSpec(PyObject *spec, FILE *fp)
{
#ifndef MS_WINDOWS
PyObject *pathbytes = NULL;
#endif
PyObject *name_unicode = NULL, *name = NULL, *path = NULL, *m = NULL;
const char *name_buf, *hook_prefix;
const char *oldcontext;
dl_funcptr exportfunc;
PyModuleDef *def;
PyObject *(*p0)(void);
name_unicode = PyObject_GetAttrString(spec, "name");
if (name_unicode == NULL) {
return NULL;
}
if (!PyUnicode_Check(name_unicode)) {
PyErr_SetString(PyExc_TypeError,
"spec.name must be a string");
goto error;
}
name = get_encoded_name(name_unicode, &hook_prefix);
if (name == NULL) {
goto error;
}
name_buf = PyBytes_AS_STRING(name);
path = PyObject_GetAttrString(spec, "origin");
if (path == NULL)
goto error;
if (PySys_Audit("import", "OOOOO", name_unicode, path,
Py_None, Py_None, Py_None) < 0) {
goto error;
}
#ifdef MS_WINDOWS
exportfunc = _PyImport_FindSharedFuncptrWindows(hook_prefix, name_buf,
path, fp);
#else
pathbytes = PyUnicode_EncodeFSDefault(path);
if (pathbytes == NULL)
goto error;
exportfunc = _PyImport_FindSharedFuncptr(hook_prefix, name_buf,
PyBytes_AS_STRING(pathbytes),
fp);
Py_DECREF(pathbytes);
#endif
if (exportfunc == NULL) {
if (!PyErr_Occurred()) {
PyObject *msg;
msg = PyUnicode_FromFormat(
"dynamic module does not define "
"module export function (%s_%s)",
hook_prefix, name_buf);
if (msg == NULL)
goto error;
PyErr_SetImportError(msg, name_unicode, path);
Py_DECREF(msg);
}
goto error;
}
p0 = (PyObject *(*)(void))exportfunc;
/* Package context is needed for single-phase init */
oldcontext = _Py_PackageContext;
_Py_PackageContext = PyUnicode_AsUTF8(name_unicode);
if (_Py_PackageContext == NULL) {
_Py_PackageContext = oldcontext;
goto error;
}
m = p0();
_Py_PackageContext = oldcontext;
if (m == NULL) {
if (!PyErr_Occurred()) {
PyErr_Format(
PyExc_SystemError,
"initialization of %s failed without raising an exception",
name_buf);
}
goto error;
} else if (PyErr_Occurred()) {
PyErr_Clear();
PyErr_Format(
PyExc_SystemError,
"initialization of %s raised unreported exception",
name_buf);
m = NULL;
goto error;
}
if (Py_IS_TYPE(m, NULL)) {
/* This can happen when a PyModuleDef is returned without calling
* PyModuleDef_Init on it
*/
PyErr_Format(PyExc_SystemError,
"init function of %s returned uninitialized object",
name_buf);
m = NULL; /* prevent segfault in DECREF */
goto error;
}
if (PyObject_TypeCheck(m, &PyModuleDef_Type)) {
Py_DECREF(name_unicode);
Py_DECREF(name);
Py_DECREF(path);
return PyModule_FromDefAndSpec((PyModuleDef*)m, spec);
}
/* Fall back to single-phase init mechanism */
if (hook_prefix == nonascii_prefix) {
/* don't allow legacy init for non-ASCII module names */
PyErr_Format(
PyExc_SystemError,
"initialization of %s did not return PyModuleDef",
name_buf);
goto error;
}
/* Remember pointer to module init function. */
def = PyModule_GetDef(m);
if (def == NULL) {
PyErr_Format(PyExc_SystemError,
"initialization of %s did not return an extension "
"module", name_buf);
goto error;
}
def->m_base.m_init = p0;
/* Remember the filename as the __file__ attribute */
if (PyModule_AddObjectRef(m, "__file__", path) < 0) {
PyErr_Clear(); /* Not important enough to report */
}
PyObject *modules = PyImport_GetModuleDict();
if (_PyImport_FixupExtensionObject(m, name_unicode, path, modules) < 0)
goto error;
Py_DECREF(name_unicode);
Py_DECREF(name);
Py_DECREF(path);
return m;
error:
Py_DECREF(name_unicode);
Py_XDECREF(name);
Py_XDECREF(path);
Py_XDECREF(m);
return NULL;
}
#endif /* HAVE_DYNAMIC_LOADING */