Re: [PATCH 00/14] FIFO bug fixes and code simplifications

[Ken Brown <kbrown at cornell dot edu>]

On 4/17/2019 1:23 PM, Achim Gratz wrote:
> Corinna Vinschen writes:
>> Pushed with v2 of patch 13.  Developer snaps should be up shortly.
> 
> I gave the snapshot some testing today.
> 
> The good news is that the named FIFO branch of our code works correctly
> again and is faster than going through pseudo-file STDIO pipes.
> 
> The bad news is that there is still some problem that seems to hit while
> forking.  I've got an empty stackdump file from sh (which gets used when
> exec'ing due to the way perl implements that) a few times and also one
> error message about a terminated thread due to "Windows WFSO error 6"
> (hilariously the output file was produced correctly in that case).  But
> most of the time the processes in my data pipeline would all have execed
> correctly, but then none of them ever gets runnable again.  So this
> seems to be something of a race around the exec.  If I kill the stalled
> processes and start the same commands again, then everything works as it
> should most of the time.

Thanks for testing.  I have STCs for fork and exec that I used when first 
writing the code, and I forgot to retest those after the recent changes.  I just 
tried, and the fork test succeeds but the exec test fails.  I'll try to debug that.

Ken