Quoting Michael Roth (2015-11-13 16:40:04)
> From: Yuri Pudgorodskiy <y...@virtuozzo.com>
>
> This was original behaviour before GLIB gspawn() rework and we rely on
> this behaviour.
>
> Signed-off-by: Yuri Pudgorodskiy <y...@virtuozzo.com>
> Signed-off-by: Denis V. Lunev <d...@openvz.org>
> CC: Michael Roth <mdr...@linux.vnet.ibm.com>
> Signed-off-by: Michael Roth <mdr...@linux.vnet.ibm.com>
> ---
> qga/commands.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/qga/commands.c b/qga/commands.c
> index 7644ca0..559f18f 100644
> --- a/qga/commands.c
> +++ b/qga/commands.c
> @@ -400,7 +400,8 @@ GuestExec *qmp_guest_exec(const char *path,
> argv = guest_exec_get_args(&arglist, true);
> envp = has_env ? guest_exec_get_args(env, false) : NULL;
>
> - flags = G_SPAWN_SEARCH_PATH | G_SPAWN_DO_NOT_REAP_CHILD;
> + flags = G_SPAWN_SEARCH_PATH | G_SPAWN_SEARCH_PATH_FROM_ENVP |
Sorry, just noticed this was only added around 2.33. I'll send updated
PULL that adds a version check.
> + G_SPAWN_DO_NOT_REAP_CHILD;
> if (!has_output) {
> flags |= G_SPAWN_STDOUT_TO_DEV_NULL | G_SPAWN_STDERR_TO_DEV_NULL;
> }
> --
> 1.9.1
>
