tail: fix checking of currently unavailable directories

* src/tail.c (tail_forever_inotify): Handle the case where
tail --follow=name with inotify, is not able to add a watch on
a specified directory.  This may happen due to inotify resource
limits or if the directory is currently missing or inaccessible.
In all these cases, revert to polling which will try to reopen
the file later.  Note inotify returns ENOSPC when it runs out
of resources, and instead we report a particular error message,
lest users think one of their file systems is full.
(main): Document another caveat with using inotify, where we
currently don't recheck directories recreated after the
initial watch is setup.
* tests/tail-2/F-vs-rename: Fix the endless loop triggered by
the above issue.
* tests/tail-2/inotify-hash-abuse: Likewise.
* tests/tail-2/wait: Don't fail in the resource exhaustion case.
* tests/tail-2/F-vs-missing: A new test for this failure mode
which was until now just triggered on older buggy linux kernels
which returned ENOSPC constantly from inotify_add_watch().
* NEWS: Mention the fix.

1 files changed, 15 insertions, 12 deletions

diff --git a/tests/tail-2/inotify-hash-abuse b/tests/tail-2/inotify-hash-abuse
index 65a341211..d45668a69 100755
--- a/tests/tail-2/inotify-hash-abuse
+++ b/tests/tail-2/inotify-hash-abuse
@@ -16,30 +16,33 @@
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 
-if test "$VERBOSE" = yes; then
-  set -x
-  tail --version
-fi
-
-. $srcdir/test-lib.sh
+. "${srcdir=.}/init.sh"; path_prepend_ ../src
+test "$VERBOSE" = yes && tail --version
 
 # 9 is a magic number, related to internal details of tail.c and hash.c
 n=9
 seq $n | xargs touch || framework_failure
 
-debug='---disable-inotify -s .01'
+debug='---disable-inotify'
 debug=
-tail $debug -qF $(seq $n) > out 2>&1 & pid=$!
+tail $debug -s.1 -qF $(seq $n) > out 2>&1 & pid=$!
+
+check_tail_output()
+{
+  local delay="$1"
+  grep "$tail_re" out > /dev/null ||
+    { sleep $delay; return 1; }
+}
 
-# Wait until tail has started...
+# Wait up to 6.3s for tail to start
 echo x > $n
-until grep '^x$' out >/dev/null 2>&1; do :; done
+tail_re='^x$' retry_delay_ check_tail_output .1 7 || fail=1
 
 mv 1 f || fail=1
 
-# Wait for this diagnostic:
+# Wait 6.3s for this diagnostic:
 # tail: `1' has become inaccessible: No such file or directory
-until grep inaccessible out >/dev/null 2>&1; do :; done
+tail_re='inaccessible' retry_delay_ check_tail_output .1 7 || fail=1
 
 # Trigger the bug.  Before the fix, this would provoke the abort.
 echo a > 1 || fail=1