* Added test_thread testing for ostream_redirect segfault recreation

* fix: scoped_ostream_redirect str created outside gil

* Moved threading tests into test_iostream. Cleaned up some formatting. Deleted test_thread.{cpp,py}

* CI: few formatting fixes

* CI: yet another formatting fix

* CI: more formatting fixes. Removed unecessary comment

* Ignore 'warning C4702: unreachable code' in MSVC 2015

Co-authored-by: Nick Bridge <nick.bridge.chess@gmail.com>
Co-authored-by: Nick Bridge <nbridge@jumptrading.com>
Co-authored-by: Yannick Jadoul <yannick.jadoul@belgacom.net>

* feat: lazy compile

* refactor: lazy -> only_changed

* refactor: leave the changed function up to the user

* refactor: pass a function, based on @YannickJadoul and @HDembinski's suggestions

* refactor: old -> _old, as it's not intended for users

* docs: slight improvmenent from @rwgk

* docs: Ccache spelling, extra warning about pip caching

Ccache spelling noted by @YannickJadoul

* Add argument names to enum_ methods

* Add test_enum::test_docstring_signatures

* style: clang-tidy: modernize-use-using

* style: more clang-tidy checking

Co-authored-by: Henry Schreiner <henryschreineriii@gmail.com>

* fix: match new extension discovery with changes to classic discovery

Followup to #2638 - this was fixed in 2.6.0, but only for classic Python

* fix: followup to avoid warnings

* Adding missing virtual destructors, to silence clang -Wnon-virtual-dtor warnings.

Tested with clang version 9.0.1-12 under an Ubuntu-like OS.
Originally discovered in the Google-internal environment.

* adding -Wnon-virtual-dtor for GNU|Intel|Clang

* Add __builtins__ to globals argument of `py::exec` and `py::eval` if not present

* Refactor into inline ensure_builtins_in_globals function

* ci: add more jobs

* ci: minor trimdown

Replaced adviced with advised.

This line had two bugs:

1. It declares `py::detail::class_` as `friend`.
2. After fixing that, we would have to change it to
   `template <typename, typename...>`

The first one was introduced ~5 years ago, when a large refactoring was
made, probably as an intermediate step during refactoring.

The second was made when `generic_type` was made to be agnostic with
respect to the order of `py::class_` template parameters.

 

We're removing the declaration altogether, because it was never relied
on. This is what makes me think that it was an intermediate step in
refactoring that shouldn't have ended up in commit history.

* demo kwarg with none(false)

* Reorder and extend tests for arg::none(false) in test_methods_and_attributes.py::test_accepts_none

* Fix arg::none() for keyword arguments

* Add changelog note

* Fix names of no_none_kw test functions

Co-authored-by: Yannick Jadoul <yannick.jadoul@belgacom.net>

* tests/CMakeLists: Robustify against add'l versions of Boost

* refactor: simpler impl

Co-authored-by: Henry Schreiner <henryschreineriii@gmail.com>

* fix: chapters in PDF again

* fix: sections in README

* docs: fix PDF build by adding unicode mappings

* fix: better PDF

* fix: html build issue

* Demonstrate test_factory_constructors.py failure without functional changes from #2335

* Revert "Demonstrate test_factory_constructors.py failure without functional changes from #2335"

This reverts commit ca33a8021fc2a3617c3356b188796528f4594419.

* Fix test crash where registered Python type gets garbage collected

* Clean up some more internal structures when class objects go out of scope

* Reduce length of std::erase_if-in-C++20 comment

* Clean up code for cleaning up type internals

* Move cleaning up of type info in internals to tp_dealloc on pybind11_metaclass

* style: use Black everywhere

* style: minor touchup from review