Revert flutter command to shlock when flock isn't available (#58281)

This reverts the flutter command to use shlock when flock isn't available. It seems that the mkdir method isn't as reliable as we want. I think that this is because the trap isn't always be executed, which is why I think that shlock uses PIDs to help it be more reliable. Unfortunately, that means that we're back to not working over network shares (which is where things were before I moved to the mkdir method, so not really a regression). I did leave in the mkdir method for platforms that have neither flock nor shlock (which should be very few and far between, but still), so at least we'll do some locking there now.

bin/internal/shared.sh

@@ -44,6 +44,8 @@ function _rmlock () {
 function _lock () {
   if hash flock 2>/dev/null; then
     flock --nonblock --exclusive 7 2>/dev/null
+  elif hash shlock 2>/dev/null; then
+    shlock -f "$1" -p $$
   else
     mkdir "$1" 2>/dev/null
   fi
@@ -64,14 +66,20 @@ function _lock () {
 #
 # The first is if the platform doesn't have "flock", for example on macOS. There
 # is not a direct equivalent, so on platforms that don't have flock, we fall
-# back to using mkdir as an atomic operation to create a lock directory. If
-# mkdir is able to create the directory, then the lock is acquired. To determine
-# if we have "flock" available, we use the "hash" shell built-in.
+# back to using trying to use the shlock command, and if that doesn't exist,
+# then we use mkdir as an atomic operation to create a lock directory. If mkdir
+# is able to create the directory, then the lock is acquired. To determine if we
+# have "flock" or "shlock" available, we use the "hash" shell built-in.
 #
-# The second complication is NFS. On NFS, to obtain an exclusive lock you need a
-# file descriptor that is open for writing. Thus, we ignore errors from flock by
-# redirecting all output to /dev/null, since users will typically not care about
-# errors from flock and are more likely to be confused by them than helped.
+# The second complication is on network file shares. On NFS, to obtain an
+# exclusive lock you need a file descriptor that is open for writing. Thus, we
+# ignore errors from flock by redirecting all output to /dev/null, since users
+# will typically not care about errors from flock and are more likely to be
+# confused by them than helped. The "shlock" method doesn't work for network
+# shares, since it is PID-based. The "mkdir" method does work over NFS
+# implementations that support atomic directory creation (which is most of
+# them). The "schlock" and "flock" commands are more reliable than the mkdir
+# method, however, or we would use mkdir in all cases.
 #
 # The upgrade_flutter function calling _wait_for_lock is executed in a subshell
 # with a redirect that pipes the source of this script into file descriptor 7.