DO-NOT-MERGE: bpo-34595: Add %t format to PyUnicode_FromFormatV() by vstinner · Pull Request #9122 · python/cpython (original) (raw)

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service andprivacy statement. We’ll occasionally send you account related emails.

Already on GitHub?Sign in to your account

Conversation12 Commits1 Checks0 Files changed

Conversation

This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.Learn more about bidirectional Unicode characters

[ Show hidden characters]({{ revealButtonHref }})

The %T format of PyUnicode_FromFormatV() now returns the fully
qualified name of an object type (ex: "module.namespace.typename").
Add %t format to PyUnicode_FromFormatV(), and so to
PyUnicode_FromFormat() and PyErr_Format(), to format the "short
name" of an object type: equivalent to "%s" with
_PyType_Name(Py_TYPE(obj)).
Replace %T format with %t format in unicodeobject.c.

https://bugs.python.org/issue34595



PyObject *
_PyType_FullName(PyTypeObject *type)

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In case of non-heap types you can just return PyUnicode_FromString(type->tp_name).

In case of heap types the code can be a tiny bit simpler if inline _PyType_Module() and _PyType_QualName() because you can get rid of increfs/decrefs and NULL checks.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

In case of non-heap types you can just return PyUnicode_FromString(type->tp_name).

Done. This change removes "builtins." when formatting builtins type. So %T of a string becomes "str" (instead of "builtins.str"). IMHO it's the expected behaviour.

In case of heap types the code can be a tiny bit simpler if inline _PyType_Module() and _PyType_QualName() because you can get rid of increfs/decrefs and NULL checks.

I don't think that performance matters here. I prefer to reuse the same code, to make sure that type.qualname, type.module and %T behave the same for heap types.

static PyObject *
type_name(PyTypeObject type, void context)
_PyType_QualName(PyTypeObject *type)

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why type_qualname() and type_module() have been renamed to _PyType_QualName() and _PyType_Module()?

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hum. I moved code in a weird way. I fixed that.

@@ -536,6 +540,9 @@ APIs:
.. [1] For integer specifiers (d, u, ld, li, lu, lld, lli, llu, zd, zi,
zu, i, x): the 0-conversion flag has effect even when a precision is given.

.. [2] The object type fully qualified name is equivalent to:
``f"{type(obj).__module__}.{type(obj).__qualname__}"``.

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Except that the module name is omitted for types in the builtins module (and for non-heap extension types that don't specify the module, but this can be considered as a bug).

@@ -768,7 +768,7 @@ ensure_unicode(PyObject *obj)
{
if (!PyUnicode_Check(obj)) {
PyErr_Format(PyExc_TypeError,
"must be str, not %T", obj);
"must be str, not %t", obj);

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What was used instead of %t/%T before?

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Python 3.7 code:

        PyErr_Format(PyExc_TypeError,
                     "must be str, not %.100s",
                     Py_TYPE(obj)->tp_name);

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thus it was closer to %T. In error messages it is better to use fully qualified names.

The %T format of PyUnicode_FromFormatV() now returns the fully qualified name of an object type (ex: "module.namespace.typename").
Add %t format to PyUnicode_FromFormatV(), and so to PyUnicode_FromFormat() and PyErr_Format(), to format the "short name" of an object type: equivalent to "%s" with _PyType_Name(Py_TYPE(obj)).
Replace %T format with %t format in unicodeobject.c.
Update existing NEWS entry

I rebased my changed and made requested changes.

I propose to replace Py_TYPE(obj)->tp_name with %t in C to mimick Python code which uses type(obj).name or obj.class.name.

If we want to use %T in C code, I suggest to also update the related Python code, especially for "C accelerators" modules like _asyncio/_pickle. I suggest to only start to use %T on a case by base basis.

#Linux-PR_20180911.05 failed: the CI is broken, it failed on apt-get install.

vstinner changed the title~~bpo-34595: Add %t format to PyUnicode_FromFormatV()~~ DO-NOT-MERGE: bpo-34595: Add %t format to PyUnicode_FromFormatV()

Sep 11, 2018