Precompiling exits early using --inline=no -O0

[Keno]

E.g.:

               _
   _       _ _(_)_     |  A fresh approach to technical computing
  (_)     | (_) (_)    |  Documentation: http://docs.julialang.org
   _ _   _| |_  __ _   |  Type "?help" for help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 0.5.0-dev+4118 (2016-05-16 19:15 UTC)
 _/ |\__'_|_|_|\__'_|  |  master/5f22ab8* (fork: 4 commits, 0 days)
|__/                   |  x86_64-apple-darwin15.3.0

julia> using TerminalExtensions; using Gadfly; using Gallium
INFO: Recompiling stale cache file /Users/kfischer/.julia/lib/v0.5/DataStructures.ji for module DataStructures.
Process 17829 exited with status = 1 (0x00000001)

[vtjnash]

http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap11.html#tag_11_01_03

it's unspecified in POSIX whether we will be terminated and unspecified if there's any way to prevent it. There's a chance we can work around this though.

[vtjnash]

simpler repro:

julia> p = spawn(detach(`./julia -e nothing`), (DevNull, STDOUT, DevNull))

[vtjnash]

hm, it seems there is of course a flag to change this behavior, if we don't like it:

diff --git a/src/unix/tty.c b/src/unix/tty.c
index 32fa37e..3727a83 100644
--- a/src/unix/tty.c
+++ b/src/unix/tty.c
@@ -87,7 +87,7 @@ int uv_tty_init(uv_loop_t* loop, uv_tty_t* tty, int fd, int readable) {
      * slave device.
      */
     if (uv__tty_is_slave(fd) && ttyname_r(fd, path, sizeof(path)) == 0)
-      r = uv__open_cloexec(path, O_RDWR);
+      r = uv__open_cloexec(path, O_RDWR | O_NOCTTY);
     else
       r = -1;

(adding regression flag since this code path was not present in the v0.4 releases)

[Keno]

Some of the tests rely on a detach process getting its own controlling terminal.

[vtjnash]

Which ones? That sounds like a bug in the test.