In D4312#79573, @smf wrote:

In D4312#79509, @durin42 wrote:

There's been some good discussion on this. I'm sympathetic to both arguments here, namely: "we could improve bookmarks and make them less bad" and "bookmarks are a dead end and nobody should use them and we shouldn't improve them" (or thereabouts - I'm summarizing complicated positions to less than a sentence, so bear with me.) I think not following up on our plans to at least make plausible incremental improvements to bookmarks serves our users poorly, and this extension merits landing as an experimental extension. We can always spin it back out if we get unhappy with it.

Obviously, I can't say I'm too happy with this. Allowing users to shoot themselves in the foot even more is pretty bad.

I'm +1, but would like someone else to before I push.

There's been some good discussion on this. I'm sympathetic to both arguments here, namely: "we could improve bookmarks and make them less bad" and "bookmarks are a dead end and nobody should use them and we shouldn't improve them" (or thereabouts - I'm summarizing complicated positions to less than a sentence, so bear with me.) I think not following up on our plans to at least make plausible incremental improvements to bookmarks serves our users poorly, and this extension merits landing as an experimental extension. We can always spin it back out if we get unhappy with it.

In D5064#78994, @mharbison72 wrote:

In D5064#78990, @durin42 wrote:

Does anyone strongly object, or do we want to embrace black and the formatting changes it implies?

Is there a fix extension configuration (and I guess precommit hook- I've never used it) for this? If so, maybe it should go on the ContributingChanges page.

Updated. Note that this and its child should go on *stable*, since otherwise the test is broken on FreeBSD 11.2 there.

In D5064#78780, @mjpieters wrote:

In D5064#78545, @av6 wrote:

I look at the changes and see nitpicks at best. On the one hand, black proved better than any linter that we can already write consistent code. On the other, if black were a linter... I'd switch to flake8, which at least is configurable.

The whole point of black is that it is not configurable. Configurable means you still have to argue about style and decide on a configuration.

Very nice! I've been intermittently looking for this for a long time and never found it.

In D5235#78172, @yuja wrote:

+/* Convert a PyInt or PyLong to a long. Returns false if there is an
+ error, in which case an exception will already have been set. */
+static inline bool pylong_to_long(PyObject *pylong, long *out)

Nit: I prefer 0/-1 return value instead of bool since that's the convention
of CPython API.

In D5224#78093, @yuja wrote:

In this case, I suspect something really weird because `PyInt_AS_LONG()` doesn't exist on Python 3 and the docs for Python 2 say it performs no error checking. So the fact that it is setting an exception on Python 3 causes me to raise an eyebrow.

It's #define PyLong_AS_LONG(op) PyLong_AsLong(op), sigh. And the reason
of the lack of PyLong_AS_LONG(op) would be that Python 3 has no bounded
integer type.

+	if (PyInt_Check(value)) {
+		long arg = PyInt_AS_LONG(value);

In this case, we'll probably have to replace PyInt_Check() + PyInt_AS_LONG()
with PyInt_AsLong() + PyErr_Occurred().