RetroSearch Browse

Mon Jun 7 10:11:07 EDT 2004 · https://mail.python.org/pipermail/python-dev/2004-June/045263.html

In the context of http://www.python.org/sf/845802 I'm looking at the
code in Python/import.c, function load_package().

The latter part of the function is this code, which tries to import the
__init__.py file, after building the module object in 'm', and
initializing it's __file__ and __path__ attributes:

        fdp = find_module(name, "__init__", path, buf, sizeof(buf), &fp, NULL);
        if (fdp == NULL) {
                if (PyErr_ExceptionMatches(PyExc_ImportError)) {
                        PyErr_Clear();
                }
                else
                        m = NULL;
                goto cleanup;
        }
        m = load_module(name, fp, buf, fdp->type, NULL);
        if (fp != NULL)
                fclose(fp);
  cleanup:
        Py_XDECREF(path);
        Py_XDECREF(file);
        return m;
}

I do not understand why the function doesn't fail when find_module
returns NULL because of an PyExc_ImportError.  For the above mentioned
bug, when __init__.py is a directory instead of a file, find_module
returns NULL.

Thomas

Home - News ( United States | United Kingdom | Italy | Germany ) - Football scores

Showing content from https://mail.python.org/pipermail/python-dev/2004-June/045263.html below:

[Python-Dev] Python/import.c