On 01/20/2010 03:05 PM, Daniel Veillard wrote:
On Wed, Jan 20, 2010 at 02:29:41AM -0500, Laine Stump wrote:
> + path);
> + goto error;
> + }
> + if (fstat(fd,&st) == -1) {
> + ret = errno;
> + virReportSystemError(NULL, errno, _("stat of '%s'
failed"), path);
>
misses a close(fd); here
> + goto error;
> + }
>
Oops! To make it more difficult to do that again in the future, I've
changed it so that fd is initialized to -1, reset to -1 after close is
called in the non-error path, and I check for it after error:.
ACK, but there is the small fd leak to plug
Revised patch following momentarily.