adding tests for clobber when more than 1 instance of file input is runn...

test/globtail/test11.data

@@ -0,0 +1,2 @@
+a.log: a1
+b.log: b1

test/globtail/test11.sh

@@ -0,0 +1,90 @@
+#!/bin/sh
+#
+# Test: multiple instances 
+# Description: running tails to simulate the effect of input {  file { path => ... } file { path => ...  } file { path => ...  } }
+#  checks that the sincedb reflects the processed state
+# 
+
+. $(dirname $0)/framework.sh
+
+test_multi_start() {
+# need a unique arg. doesn't support many tests of same file
+[ $# -eq 0  ] && exit 99
+echo $TAIL
+$TAIL "$TEST_DIR/$1" >>$(mktemp --tmpdir=${TEST_DIR}  multi_XXXXXXXX) 2>&1 &
+[ "$TEST_MULTI_TAIL_PID" ] && export TEST_MULTI_TAIL_PID=$TEST_MULTI_TAIL_PID,$! || export TEST_MULTI_TAIL_PID=$!
+}
+
+test_multi_stop() {
+OLD_TEST_TAIL_PID=$TEST_TAIL_PID
+
+# if no args stop them all
+[ $# -eq 0  ] && { echo "$TEST_MULTI_TAIL_PID" | tr "," "\n" | while read TEST_TAIL_PID; do test_stop 2>/dev/null ;done; export TEST_MULTI_TAIL_PID=; }
+# or
+[ $# -eq 1  ] && { TEST_TAIL_PID=$1; test_stop; }
+
+export TEST_TAIL_PID=$OLD_TEST_TAIL_PID
+}
+
+
+test_init
+
+# so not have to wait for discover
+> $TEST_DIR/a.log
+> $TEST_DIR/b.log
+
+# start 2 threads watch 2 files.
+test_multi_start a.log
+test_multi_start b.log
+
+# misses first logs without a sleep
+sleep 1
+
+# put something in both log files.
+echo a1 >> $TEST_DIR/a.log
+echo b1 >> $TEST_DIR/b.log
+
+# allow both tails to write out their SINCEDB
+sleep 5
+
+# stop everything (instances write out their sincedb also on shutdown?)
+test_multi_stop
+
+# make a aggregate log file for use by test harness
+cat $TEST_DIR/multi_* > $TEST_OUT
+
+# Status: draft
+# if there is a bashism for finding the number jruby is using for dev_major and dev_minor? (is not the same as lstat output)
+# or i just check  the inodes and sizes for a simple test
+# this check wont create accurate lists for tests with files with truncations, and deletes during the tests
+# a more sophisticated would be required to track inode changes etc
+#
+filelist=$(sed -e "s,^${TEST_DIR}/,," -e '/^D, \[/d' $TEST_OUT | awk '{print $1}' | awk -F ":" '{print $1}' | sort -u)
+if [ $(cat $SINCEDB | egrep "^" | wc -l) -ne $(echo "$filelist" | egrep "^" | wc -l)  ]; then
+errs=$errs"ERROR: file count for sincedb not match count of files processed\n"
+errs=$errs"sincedb file count: $(cat $SINCEDB | egrep "^" | wc -l)\n"
+errs=$errs"filelist file count: $(echo "$filelist" | egrep "^" | wc -l)\n"
+fi
+
+# ridiculous subshell f****y
+errs=$errs$(
+echo "$filelist" | while read file; do
+inode=`stat --printf="%i" ${TEST_DIR}/$file`
+disk_size=`stat --printf="%s" ${TEST_DIR}/$file`
+egrep "^$inode " $SINCEDB >/dev/null || { echo "File $file not found in sincedb"; continue; }
+sincedb_size=$(egrep "^$inode" $SINCEDB | awk '{print $4}')
+echo "Filename: $file Inode: $inode DiskSize: $disk_size SinceDB_size: $sincedb_size" 1>&2
+[ $sincedb_size -eq $disk_size ] || { echo "File $file size mismatch $disk_size != "; continue; }
+done
+)
+
+
+[ "$errs" ] && echo -e "Sincedb not match filelist\n${errs}" >> $TEST_OUT
+
+
+
+
+test_done
+
+
+