2010/5/4 Eric Blake <eblake(a)redhat.com>:
On 05/03/2010 05:15 PM, Matthias Bolte wrote:
> Add an empty body for virCondWaitUntil and move virPipeReadUntilEOF
> out of the '#ifndef WIN32' block, because it compiles fine with MinGW
> in combination with gnulib.
> ---
> src/util/threads-win32.c | 10 +++
> src/util/util.c | 160 +++++++++++++++++++++++-----------------------
> 2 files changed, 90 insertions(+), 80 deletions(-)
>
> diff --git a/src/util/threads-win32.c b/src/util/threads-win32.c
> index a30bccf..e478560 100644
> --- a/src/util/threads-win32.c
> +++ b/src/util/threads-win32.c
> @@ -157,6 +157,16 @@ int virCondWait(virCondPtr c, virMutexPtr m)
> return 0;
> }
>
> +int virCondWaitUntil(virCondPtr c ATTRIBUTE_UNUSED,
> + virMutexPtr m ATTRIBUTE_UNUSED,
> + unsigned long long whenms ATTRIBUTE_UNUSED)
> +{
> + /* FIXME: this function is currently only used by the QEMU driver that
> + * is not compiled on Windows, so it's okay for now to just
> + * miss an implementation */
> + return -1;
Does anyone care if errno is set? Then again, since it is never called,
who cares.
ACK.
Thanks, pushed.
Matthias