Re: test harness/test cases for common-src/shm-ring.c?

I searched the internet and found no reference to sem_close failing with 
EBADF.

Can you remove the exit(1); in the error path, sendbackup will exit soon 
after, so a failure to close them should not be a problem.

Jean-Louis

On 08/02/17 09:04 PM, Eric Schnoebelen wrote:
> Jean-Louis Martineau writes:
> - There is no functionality test
> -
> - Can you post the complete debug file?
>
> Sure, but it's very un-interesting.
>
> Hmm, with 3.4 and 3.4.1, EINVAL was returned by sem_*wait(), with
> 3.4.2 I am seeing a EBADF from sem_close().  In both cases, the
> errors keep sendbackup from sending the size line.
>
> Attached is the amdump results, the sendbackup log files, and
> the patch I applied to shm-ring.c to add additional debugging
> information (and create shorter shared memory/semaphore names as
> required by NetBSD's implementation.)
>
> - On 08/02/17 10:54 AM, Eric Schnoebelen wrote:
> - > I'm attempting to port Amanda 3.4.x to NetBSD, and am seeing strange
> - > failures from sendbackup.
> - >
> - > The failures are in shm-ring.c, with "invalid operand" (EINVAL)
> - > in shm_ring_sem_wait.
> - >
> - > Does a simpler test program exist for the functionality in
> - > shm-ring.c?  Debugging as part of sendbackup is more painful
> - > than I have time for.
> - >
> - > Thank you,
> - > 	Eric
>
> --
> Eric Schnoebelen		[email protected]		http://www.cirr.com
> 	You can only live once, but if you do it right, once is enough.