return 0, data, bufio.ErrFinalToken
		}
		// If the token is "STOP", stop the scanning and ignore the rest.
		if string(data[:i]) == "STOP" {
			return i + 1, nil, bufio.ErrFinalToken
		}
		// Otherwise, return the token before the comma.
		return i + 1, data[:i], nil
	}
	const input = "1,2,STOP,4,"
	scanner := bufio.NewScanner(strings.NewReader(input))
	scanner.Split(onComma)

`Stop` methods have not been called.
Earlier versions of Go did not collect unstopped `Timer`s until after
they had fired and never collected unstopped `Ticker`s.

Second, the timer channel associated with a `Timer` or `Ticker` is
now unbuffered, with capacity 0.
The main effect of this change is that Go now guarantees
that for any call to a `Reset` or `Stop` method, no stale values

			break
		}
	}

	// Now look for the first ASCII non-space byte from the end
	stop := len(s)
	for ; stop > start; stop-- {
		c := s[stop-1]
		if c >= utf8.RuneSelf {
			return TrimFunc(s[start:stop], unicode.IsSpace)
		}
		if asciiSpace[c] == 0 {
			break
		}
	}

	// At this point s[start:stop] starts and ends with an ASCII
	// non-space bytes, so we're done. Non-ASCII cases have already

//
// Scanning stops if the function returns an error, in which case some of
// the input may be discarded. If that error is [ErrFinalToken], scanning
// stops with no error. A non-nil token delivered with [ErrFinalToken]
// will be the last token, and a nil token with [ErrFinalToken]
// immediately stops the scanning.
//

and the [go command documentation](/cmd/go#hdr-Build_and_test_caching).

### Go 1.23

Go 1.23 changed the channels created by package time to be unbuffered
(synchronous), which makes correct use of the [`Timer.Stop`](/pkg/time/#Timer.Stop)
and [`Timer.Reset`](/pkg/time/#Timer.Reset) method results much easier.
The [`asynctimerchan` setting](/pkg/time/#NewTimer) disables this change.
There are no runtime metrics for this change,

	testThreadLockFunc = testThreadLock
}

func testThreadLock(t *testing.T) {
	stop := make(chan int)
	go func() {
		// We need the G continue running,
		// so the M has a chance to run this G.
		for {
			select {
			case <-stop:
				return
			case <-time.After(time.Millisecond * 100):
			}
		}
	}()
	defer close(stop)

	for i := 0; i < 1000; i++ {
		if !C.Ctid() {

present to create such a binary.

Most builds both compile source code and invoke the linker to create a
binary. When cgo is involved, the compile step already requires gcc, so
it is not problematic for the link step to require gcc too.

An important exception is builds using a pre-compiled copy of the
standard library. In particular, package net uses cgo on most systems,

// panicking goroutine. Executing a call to recover inside a deferred
// function (but not any function called by it) stops the panicking sequence
// by restoring normal execution and retrieves the error value passed to the
// call of panic. If recover is called outside the deferred function it will
// not stop a panicking sequence. In this case, or when the goroutine is not
// panicking, recover returns nil.
//

# As of CL 464955, the build system uses make.bash -compile-only instead,
# so this script no longer runs in any automated fashion.
#
# Options:
#   -e: stop at first failure

if [ ! -f run.bash ]; then
	echo 'buildall.bash must be run from $GOROOT/src' 1>&2
	exit 1
fi

sete=false
if [ "$1" = "-e" ]; then
	sete=true
	shift
fi

Comments serve as program documentation. There are two forms:
</p>

<ol>
<li>
<i>Line comments</i> start with the character sequence <code>//</code>
and stop at the end of the line.
</li>
<li>
<i>General comments</i> start with the character sequence <code>/*</code>
and stop with the first subsequent character sequence <code>*/</code>.
</li>
</ol>

<p>
A comment cannot start inside a <a href="#Rune_literals">rune</a> or