All tests pass.

`defer` allows to schedule a cleanup to be executed when current function
returns. It is similar to `try`/`finally` but does not force the cleanup part
to be far away in the end. For example::

   wc = wcfs.join(zurl)    │     wc = wcfs.join(zurl)
   defer(wc.close)         │     try:
                           │        ...
   ...                     │        ...
   ...                     │        ...
   ...                     │     finally:
                           │        wc.close()

For completeness there is `recover` and `panic` that allow to program with
Go-style error handling, for example::

   def _():
      r = recover()
      if r is not None:
         print("recovered. error was: %s" % (r,))
   defer(_)

   ...

   panic("aaa")

But `recover` and `panic` are probably of less utility since they can be
practically natively modelled with `try`/`except`.

If `defer` is used, the function that uses it must be wrapped with `@func` or
`@method` decorators.

The implementation is partly inspired by work of Denis Kolodin:

- https://habr.com/post/191786
- https://stackoverflow.com/a/43028386/9456786

While the second phase of select is running we queue send/recv cases to
corresponding channels. At some point - when some of the cases are
already queued - a peer goroutine might try to send/recv on that
channel. And it will succeed because a waiter was queued to the channel.

At the same time select is continuing its enqueue loop and before enqueuing
to a channel it tries to send/recv there. If that channel became just ready
(i.e. just after select poll phase) the try to send/recv will succeed. This
means that actually 2 select cases could be executed at the same time.

Fix it by carefully checking whether some case already won before trying
to send/recv on a channel.

This fixes the test failures that were demonstrated by previous 2 patches.

Previous select-select test uses 2 channels and recreats them every
iteration. However the bug we just saw in b51b8d5d (select: Run tests
more thoroughly) is of parasitic nature - where misbehaviour depends on
what state has been left there from previous select and whether the race
to get to that state fast enough succeeded.

So add a more secialized test which tries to trigger the parasitic
effects that depend on previous state left by select on the channels.

The test, similarly to b51b8d5d, currently fails.

With more iterations it triggers a bug:

	kirr@deco:~/src/tools/go/pygolang$ py.test -vsx
	============================================== test session starts ===============================================
	platform linux2 -- Python 2.7.15+, pytest-3.6.4, py-1.6.0, pluggy-0.6.0 -- /usr/bin/python2
	cachedir: .pytest_cache
	rootdir: /home/kirr/src/tools/go/pygolang, inifile:
	collected 7 items

	golang/_gopath_test.py::test_import_module PASSED
	golang/_gopath_test.py::test_import_package PASSED
	golang/gcompat_test.py::test_qq PASSED
	golang/golang_test.py::test_go PASSED
	golang/golang_test.py::test_chan PASSED
	golang/golang_test.py::test_select Exception in thread Thread-117:
	Traceback (most recent call last):
	  File "/usr/lib/python2.7/threading.py", line 801, in __bootstrap_inner
	    self.run()
	  File "/usr/lib/python2.7/threading.py", line 754, in run
	    self.__target(*self.__args, **self.__kwargs)
	  File "/home/kirr/src/tools/go/pygolang/golang/golang_test.py", line 320, in _
	    assert (_, _rx) == (1, 'b')
	AssertionError: assert (1, 'xxx1') == (1, 'b')
	  At index 1 diff: 'xxx1' != 'b'
	  Full diff:
	  - (1, 'xxx1')
	  + (1, 'b')

This test was failing when pygolang was not dev-mode installed (`pip install -e .`):

	kirr@deco:~/src/tools/go/pygolang$ py.test
	=========================================== test session starts ============================================
	platform linux2 -- Python 2.7.15+, pytest-3.6.4, py-1.6.0, pluggy-0.6.0
	rootdir: /home/kirr/src/tools/go/pygolang, inifile:
	collected 8 items

	golang/_gopath_test.py ..                                                                            [ 25%]
	golang/gcompat_test.py .                                                                             [ 37%]
	golang/golang_test.py F....                                                                          [100%]

	================================================= FAILURES =================================================
	_________________________________________________ test_go __________________________________________________

	    def test_go():
	        # leaked goroutine behaviour check: done in separate process because we need
	        # to test process termination exit there.
	        subprocess.check_call([sys.executable,
	>           dirname(__file__) + "/golang_test_goleaked.py"])

	golang/golang_test.py:38:
	_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

	popenargs = (['/usr/bin/python2', '/home/kirr/src/tools/go/pygolang/golang/golang_test_goleaked.py'],)
	kwargs = {}, retcode = 1
	cmd = ['/usr/bin/python2', '/home/kirr/src/tools/go/pygolang/golang/golang_test_goleaked.py']

	    def check_call(*popenargs, **kwargs):
	        """Run command with arguments.  Wait for command to complete.  If
	        the exit code was zero then return, otherwise raise
	        CalledProcessError.  The CalledProcessError object will have the
	        return code in the returncode attribute.

	        The arguments are the same as for the Popen constructor.  Example:

	        check_call(["ls", "-l"])
	        """
	        retcode = call(*popenargs, **kwargs)
	        if retcode:
	            cmd = kwargs.get("args")
	            if cmd is None:
	                cmd = popenargs[0]
	>           raise CalledProcessError(retcode, cmd)
	E           CalledProcessError: Command '['/usr/bin/python2', '/home/kirr/src/tools/go/pygolang/golang/golang_test_goleaked.py']' returned non-zero exit status 1

	/usr/lib/python2.7/subprocess.py:190: CalledProcessError
	------------------------------------------- Captured stderr call -------------------------------------------
	Traceback (most recent call last):
	  File "/home/kirr/src/tools/go/pygolang/golang/golang_test_goleaked.py", line 23, in <module>
	    from golang import go, chan
	ImportError: No module named golang
	==================================== 1 failed, 7 passed in 0.10 seconds ====================================

Fix it by injecting top-level pygolang/ into $PYTHONPATH when testing
via external processes.

Fixes 69cef96e (go: Don't allow leaked goroutines to prevent program to exit)

There was no test for @method so far and that's why it went unnoticed.
But on Python3 it breaks on f.func_name:

	In [3]: def f(): pass

	In [4]: f.func_name
	---------------------------------------------------------------------------
	AttributeError                            Traceback (most recent call last)
	<ipython-input-4-662dcbac1531> in <module>()
	----> 1 f.func_name

	AttributeError: 'function' object has no attribute 'func_name'

Fix it by using f.__name__ which works on both py2 and py3.

Add test for @method to make sure it doesn't break unnoticed.

This is the Go behaviour, as demonstratd by the following program:

---- 8< ----
package main

import (
	"fmt"
	"time"
)

func work(w int) {
	for i := 0; ; i++ {
		fmt.Printf("w%d: %d\n", w, i)
		time.Sleep(1*time.Second)
	}
}

func main() {
	for i := 0; i < 100; i++ {
		go work(i)
	}

	time.Sleep(3*time.Second)
	println("main: exit")
}
---- 8< ----

+ corresponding bits in golang.testing package.

py.bench description follows:

"""
py.bench, similarly to py.test, discovers bench_* functions and Bench* classes
and then runs each discovered benchmark via golang.testing.benchmark. Similarly
to `go test -bench`, benchmarking results are printed in Go benchmark format.

For example, running py.bench on the following code::

    def bench_add(b):
        x, y = 1, 2
        for i in xrange(b.N):
            x + y

gives something like::

    $ py.bench --count=3 x.py
    ...
    pymod: bench_add.py
    Benchmarkadd    50000000        0.020 µs/op
    Benchmarkadd    50000000        0.020 µs/op
    Benchmarkadd    50000000        0.020 µs/op
"""

The implementation is based on t/py.bench from Wendelin.core - see
following commits for its history:

lab.nexedi.com/nexedi/wendelin.core/commit/51f252d4
lab.nexedi.com/nexedi/wendelin.core/commit/074ce24d
lab.nexedi.com/nexedi/wendelin.core/commit/ed13c3f9
lab.nexedi.com/nexedi/wendelin.core/commit/fc08766d
lab.nexedi.com/nexedi/wendelin.core/commit/5a1ed45a
lab.nexedi.com/nexedi/wendelin.core/commit/bcab1246

Python3 complains that there is no .im_func & friends of a method.
Workaround it all with compat code.

This patch made its first step as a way to teach qq to also work on
python3. However the differences in str / unicode and escapes in between
py2 / py3 quickly popped out and then it became easier to just handle
whole escaping logic myself.

The implementation is based on

	go123@c0bbd06e

and byproduct of manual handling is that now we don't escape printable
UTF-8 characters.

Go benchmark format is described here:

	https://github.com/golang/proposal/blob/master/design/14313-benchmark-format.md

Additionally we support extension to that format:

	- a line starting with `*** neotest:` denotes start of neotest extension block.
	  The block consists of labels describing e.g. hardware and software on a node.
	  The block ends with a blank line.
	  Labels in the block are not added to benchmarking lines from main stream.
	  The block itself should not contain benchmark lines.

and upon processing benchmark units are normalized to common base,
similarly to

	golang.org/cl/82955
	golang.org/cl/82956
	golang.org/cl/82957

Orignally implemented here:

	neo@502d9477
	neo@a9b10a45
	neo@f5fec740
	neo@916782b6

Not only we can import modules by full path, but now we can also spawn
threads/coroutines and exchange data in between them with the same
primitives and semantic as in Go.

The bulk of new functionality is copied from here:

	kirr/go123@9e1aa6ab

Original commit description follows:

"""
golang: New _Python_ package to provide Go-like features to Python language
- `go` spawns lightweight thread.
- `chan` and `select` provide channels with Go semantic.
- `method` allows to define methods separate from class.
- `gimport` allows to import python modules by full path in a Go workspace.

The focus of first draft was on usage interface and on correctness, not speed.
In particular select should be fully working.

If there is a chance I will maybe try to followup with gevent-based
implementation in the future.
Hide whitespace changes
"""

Module gopath provides way to import python modules by full path in a Go workspace.

For example

    lonet = gopath.gimport('lab.nexedi.com/kirr/go123/xnet/lonet')

will import either

    lab.nexedi.com/kirr/go123/xnet/lonet.py, or
    lab.nexedi.com/kirr/go123/xnet/lonet/__init__.py

located somewhere under $GOPATH.

A simple project to allow python imports to be done by full path in a Go
workspace.