initialization for time() not complete in certain cases.

[amitmurthy]

Saw this just once:

   _ _   _| |_  __ _   |  Type "help()" to list help topics
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 0.2.0-2508.r2d97566e.dirty
 _/ |\__'_|_|_|\__'_|  |  Commit 2d97566e44 2013-07-09 23:05:21*
|__/                   |  x86_64-linux-gnu

julia> addprocs(10)Master process (id 1) could not connect within 60.0 seconds.
exiting.

Master process (id 1) could not connect within 60.0 seconds.
exiting.
Master process (id 1) could not connect within 60.0 seconds.
exiting.
Master process (id 1) could not connect within 60.0 seconds.
exiting.
Master process (id 1) could not connect within 60.0 seconds.
exiting.
Master process (id 1) could not connect within 60.0 seconds.
exiting.
Master process (id 1) could not connect within 60.0 seconds.
exiting.
Master process (id 1) could not connect within 60.0 seconds.
exiting.
Master process (id 1) could not connect within 60.0 seconds.
exiting.
Master process (id 1) could not connect within 60.0 seconds.
exiting.

when I did an addprocs(10) after launching julia. All 10 workers exited immediately after the call, presumably because the first call to time() in the below code returned 0

        start = time()
        while !haskey(map_pid_wrkr, 1) && (time() - start) < timeout
            sleep(1.0)
        end

        if !haskey(map_pid_wrkr, 1)
            print(STDERR, "Master process (id 1) could not connect within $timeout seconds.\nexiting.\n")
            exit(1)
        end

Since, it happened to all 10 processes, and is not reproducable, I am guessing it is because something on my Linux system delayed some proper initialization required by time() which is just a ccall to clock_now

[JeffBezanson]

clock_now just calls gettimeofday. I'm not aware that it's possible for that to return an incorrect value. It doesn't even return any error codes.

[amitmurthy]

Could have been because I had not done a make clean for quite some time. Will reopen if I see it again.