why is my NTP client timestamp wrong?

edit

Oops. sigh... It sucks getting old and having your eyes go south (that's my excuse anyway). Thanks, cmaynard.

With that little miscue out of the way, I'm still seeing results that I don't understand. Just for debugging purposes, I've modified my NTP client to populate all 4 timestamps in the message, as follows:

NTPRefTs = 0xde494301;              //  2018 03 06 mzimmers: try different time.

NTPTxBuf->REFTSS = htonl(0);        //  Reference TimeStamp Seconds 
NTPTxBuf->ORGTSS = 1;               //  Origin TimeStamp Seconds
NTPTxBuf->RXTSS = htonl(1);         //  Receive TimeStamp Seconds
NTPTxBuf->TXTSS = htonl(NTPRefTs);  //  Big Endian Transmit TimeStamp Seconds

The fractions are all 0.

As you can see, I'm using two byte orders of the value 1 in the middle timestamps. My layman expectation is that one of these should reveal a timestamp only one second off the timestamp with the value of 0. But they're both in 2036. Can someone explain what's happening here ...(more)

Well, the folks at Cloudshark apparently don't even allow one to view others' pcap files once the 30 day trial period has expired, so I can't view this pcap file. If it's a small enough file, you can try the trick I described before for attaching pcaps to questions, but I haven't tried it on the new Q & A site yet, so I don't know if it will work. Looking at the old question, maybe something was modified so it no longer works there either since I no longer see the images, but at least at one point in the past it worked.

Long ago, Joe McEachern from QA Café (Cloudshark) offered to provide the Wireshark project with its own Cloudshark appliance. If the offer still stood and Wireshark did have its own appliance, that would probably solve all of these annoying problems with ...(more)