Bad remote file systems can change mtime unexpectedly,
and then there is the problem that git rebase and similar
operations like to change the files and then change them back,
modifying the mtimes but not the content.

Avoid spurious Put errors on both of those by checking file
content.

(False positive "modified since last read" make the real ones
difficult to notice.)

We ran for a long time with 10ms kernel resolution,
so 10ms user space resolution here should be fine.
Some systems actually provide 1ms sleeps, which
makes this polling use a bit more cpu than we'd like.
Since the timers are for user-visible things, 10ms should
still be far from noticeable.

Reduces acme's cpu usage on Macs when plumber is missing
(and plumbproc is sleeping waiting for it to appear).

LGTM=aram, r
R=r, aram
https://codereview.appspot.com/99570043

Move libfmt, libutf into subdirectories of lib9.

Add poll-based socket i/o to libthread, so that we can
avoid using multiple procs when possible, thus removing
dependence on crappy pthreads implementations.

Convert samterm, acme to the single-proc libthread.

Bring libcomplete, acme up-to-date w.r.t. Plan 9 distribution.

Also add acme!