I confirmed that these 3 platforms share the same signal definitions
from these sources.

Haiku: https://github.com/haiku/haiku/blob/master/headers/posix/signal.h
OpenBSD: https://github.com/openbsd/src/blob/master/sys/sys/signal.h
NetBSD: http://fxr.watson.org/fxr/source/sys/signal.h?v=NETBSD

Changes

- Support multi-threaded testing.
- Support `set_fail_timeout` on all platforms.
- Display an animated progress report.
- Setup all tests with a context logger.
- Give all tests their own separate custom allocators.
- Support tracking test memory usage.
- Display a summary of the failed tests at the end.
- Let users select only specific tests to run.
- Support copying failed tests to the clipboard to run again.
- Support catching SIGINT (CTRL-C) to cancel early.
- Record context in cleanup procs.
- Write all log messages to STDERR for easy redirection.
- Possibly more I've forgotten.

New Options

- `-define:test_threads=N`: Specify thread count.
- `-define:test_thread_memory=B`: Specify initial memory block size in bytes to each thread.
- `-define:test_track_memory=true`: Track the memory usage of individual tests.
- `-define:test_fancy=false`: Disable animated progress report.
- `-define:test_select=package.test_name,...`: Run only select tests.
- `-define:test_clipboard=true`: Copy names of failed tests to the clipboard.
- `-define:test_progress_width=24`: Change the width of the animated progress bars.

I do not have a Windows machine to test the refactored test runner, and
I am unsure if it would even run correctly on Windows without this
disabled.

`T` no longer has a writer assigned to it.

`test_core_cbor.odin` has global state and is run with `odin test`,
so I've set it to use only one thread.

I was under the impression that the default `context.allocator` was not
thread-safe, but I've been told that this is not the case.

`STDERR` might be redirected, and this code signals to the terminal to
show the cursor again. Otherwise, the cursor will be invisible.

This commit also changes the name of `test_select` to `ODIN_TEST_NAMES`,
to better conform with the already-existing `-test-name:<name>` option.

This should fix tests failing on 32-bit platforms.

Adds new option `ODIN_TEST_ALWAYS_REPORT_MEMORY`, for when you always
want to see the memory usage report.

This will automatically calculate how wide the progress bars should be
based on the package with the greatest number of tests.

The progress width is now capped to 100.

It was leaky and required a substantial number of `loc := #caller_location` additions to parts of the core library to make it easier to track down how and where it leaked.

The tests now run fine multi-threaded.

Add a new option `ODIN_TEST_RANDOM_SEED` which is picked from the cycle
counter at startup, if it's not specified by the user.

This number is sent to every test in the `T` struct and reset every test
(just in case).

And fix a few leaks in `core:encoding/hxa` while at it.

Made them run in parallel as well.

Disable benchmark on Windows for a moment.

`-define:ODIN_TEST_NAMES=...` is capable of selecting test by package
and name or name only, with the ability to access packages included by
`-all-packages`.

Provides a helpful info message about the option to change how many
threads are used per run.

This should tidy up the CI output logs a bit.

I was encountering bounds-check error messages being overwritten during
a test, if the test failed for another reason and sent a log message.

The original intent of having this check inside of the above `if` block
was that if a test sent an error message, then it was assumed an
overwrite would be safe, but it's completely possible for a test to fail
for a legitimate reason, then do an unrelated bounds check somewhere
else that would be buried under the animation.

This change will make sure that, no matter what, the progress display
will not trigger a clear if a signal was raised. There's still no
guarantee that bounds-check messages will be printed properly, and it's
best to redirect STDERR.

The only way that can be fixed is if they get a similar hook to
`context.assertion_failure_proc`.

Tested on FreeBSD 14.0 and NetBSD 10.0

OpenBSD is untested, but link names were sourced from:
https://github.com/openbsd/src/blob/master/include/stdio.h

According to this, OpenBSD shares the same layout as NetBSD.

FreeBSD has the same as Darwin in this regard.

… solved."

This reverts commit 21a1ddf.

The test runner was deadlocking when a test raised a signal on FreeBSD.

This is untested on OpenBSD, but I have referenced this file:
https://github.com/openbsd/src/blob/master/include/pthread.h

Add `pthread_testcancel` to `core:sys/unix`

Works well with `-define:ODIN_TEST_LOG_LEVEL=warning`.