GA Joe, and thanks to you and the development team for their efforts. I hope I included all the info that you wanted.
WSJT-x doesn't really "crash"- one of the WSJT Entry Windows simply "disappears". No BSOD, No RTE. There is an error message in a little box: "Something unexpected caused WSJT to close the Radio TCP port unexpectedly. Please try restarting WSJT again. If that does not work restart N1MM." I can't tell what was happening just before the failure, as there is no warning there is going to be a failure. WSJT is controllilng two radios through N1MM using the DX Lab Commander process, as detailed in the N1MM setup instructions. I have not found a way to force a failure. I will note, that version 2.5.4 did not have this problem. The EW that closes may be either EW1 or EW2; there is no way of guessing which is going to crash, and no indication that I've noticed to indicate which window is about to crash. If you then restart WSJT, you will get another error box: "Sub-process error. Failed to close orphaned jt9 process." Going into Task Manager and closing that JT9 function does not restore functionality. If I close all the WSJT windows and all the N1MM+ windows, and try to restart N1MM, I get another error message about a port 13064 not being closed. The quickest and most reliable way to get back to making contacts is to restart the computer, restart N1MM, and restart each WSJT window. Yes, there are other reasons for wanting N1MM. I use it to keep track of multipliers- what do I need, and where what I need is operating. I also tie in Telnet with N1MM to keep me up on multipliers and what bands are open and to where. Rate info, transferring score to one of the Online Scoreboards. Since I don't have a big big display for all this, I also network N1MM to display info that won't fit on my regular display. One other thing I noticed after the last crash- I went in to delete the orphaned JT9 process, and found that there were two JT9 processes, one about 20Mb and one about 180Mb. After I restarted the N1MM and the 2 WSJT EWs, I went into the same process and found only 1 JT9 process, about 18-22Mb. I just restarted the computer and reentered everything again, and now I have 2 jT9 processes, one about 20Mb and one about 194Mb. Took another look- now they are both 14.5Mb. Now both are at 180Mb. Don't know if this is apropos at all, just seems odd. Right now, I'm planning to use 2.5.4 for the WWDigi Test. I hate to give away the advantages of using 2.6.0rc3, but I can't handle the interruptions and lost QSOs. Dennis W1UE On Tue, Aug 23, 2022 at 4:35 PM Joe Taylor <j...@princeton.edu> wrote: > Hi Dennis and Ron, > > Thanks for your reports. This message is off-list, since your needs are > rather specialized -- but I have CC'd other members of our core > development group. > > On 8/23/2022 2:43 PM, Dennis W1UE via wsjt-devel wrote: > > I had high hopes for 2.6.0 rc3. I ran it for 25 hours decode time using > > an SO2R setup. It never stopped working nor crashed.2.6.0 rc1 and rc2- > > the EW just disappeared from the screen and I got an error box related > > to TCP. Since that first stoppage- or crash- I've had 4 more crashes. > > The longest I've operated since then has been about 40 minutes. > > > > So, the bug that is causing the program to suddenly close, has not been > > found. And it was looking so good! > > On 8/23/2022 2:52 PM, Ron WV4P via wsjt-devel wrote: > > Same here, running N1MM with 2 instances it keeps crashing and locking > > everything down when it does. It took a long time for the first > > occurrence but now its lucky to stay up 30 min. > > ... > > > I'll be single op. > > If anyone makes it work with the Max Distance feature they will have a > distinct advantage. Sadly it's back to 2.54 for me, I could not make it > work reliably. I'll run it again this evening to try and get some screen > shots and data that might help the Smart people, before I swap back to the > stable version. > > Good luck, Dennis ! You rocked the last one. > > We'll be happy to help you guys to isolate the problem you're having, > apparently with communication between WSJT-X and N1MM+, but I must > caution you that it may take some time. Moreover, in the immediate > future my time is largely committed to a paper on the EME Echo mode > features in WSJT-X 2.6.0. > > None of our program versions 2.6.0-rc1, -rc2, or -rc3 have intentional > changes to the way communications take place with an external program > such as N1MM+. At present I don't know the details of what happens when > you say it (I suppose you mean WSJT-X) crashes. Any error message? > What was happening just before the failure? Is WSJT-X doing any rig > control? Can you stress-test the situation in a way that reliably > forces a failure? > > These release-candidate versions of WSJT-X 2.6.0 have been bullet-proof > in my own usage patterns, including in the ARRL-DIGI contest back in > June at W2ZQ. But we were using it alone, not with N1MM. Of course, > WSJT-X does not have built-in SO2R features. Are there other reasons > that you want N1MM in addition to WSJT-X for, say, the WW-DIGI contest? > > None of us is a regular user of the N1MM + WSJT-X combination, so we > really need your help -- and possibly that of the N1MM development > group, to track down what's flaky about this two-program kludge. > > -- 73, Joe, K1JT >
_______________________________________________ wsjt-devel mailing list wsjt-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/wsjt-devel
