JS8Call timimg [SOLVED]

[bschnzl]

Greetings -

JS8Call will not key the PTT in a timely fashion after the first transmission. It works as expected the first time it sends. After that, it may key up to 8 seconds late (or more). The audio stops on time - within the JS8 protocol time slot. It may stay keyed up to two seconds into the next slot, but most responding stations still "print".

I use FLRig to key the PTT. I run multiple instances of FLDigi concurrently. I use JS8Call to change frequencies if I have something more pressing. JS8Call changes frequencies late too. Interestingly, JS8Call runs thru another freq change cycle immediately following. This has no effect on the radio if I let it run. Cancelling, and resetting the config to not change freq's lets me keep my schedule diversion.

I have been trying to trouble shoot this on my own for months. I think it started around January 2023. I have been starting it from the command line for some time. I first built this machine in January 2016 - It was a Christmas present from my wife. Its a Dell Precision M6800. I was so impressed I bought a bunch for my company. It is a But I digress...

wsjtx works well... it keys directly.

I had hoped that the next version would fix it, but upstream is confusing. Yeah, thats it.

This is what I get when JS8Call starts:
(I believe the issue is in QMetaObject. The ALSA stderr stuff also shows in FLDigi - but it works well)

[bschnzl]

So...

JS8Call has been removed from Gentoo! I guess I have to do an overlay to continue! C'est la Vie! I'll be able to grab the unauthorized latest version!

But I mention above that it involves QT. That opens a ton of dependencies! While running updates today, I found that QT fails to build! I moved ninja to reference samurai. At least that made the FAIL show up in the last line! I'm getting:

[bschnzl]

Well!!! What do you know!

JS8Call has remained on my Haswell and AlderLake machines, unmolested, through this entire ordeal!

After finishing a world update, including building

[szatox]

Whad do you mean by "js8call got removed from repo"?

[bschnzl]

Well! What do you know!
I last updated JS8Call with portage on Dec 06, 2023 07:32:25. since then it has been unavailable.
I did not save the output at the time.
I just sync'd so...
_________________
Persistance Pays

[szatox]

Are you sure you didn't sync with a broken mirror or something like that? I'm up to date and it's still there.
It's not even masked, which should be the case for 30 days before removal. I don't see it going anytime soon.

[bschnzl]

I know what the rules are...
It dropped from availability right after I posted this thread!
But that is a distraction!
I still have a haswell machine that has the timing issue - or at least it did a week ago.
And QT5 is broke!
_________________
Persistance Pays

[Hu]

I see no commits within the last 12 months to media-radio/js8call in gentoo.git. Why do you think it "has been removed from Gentoo"?

Qt is known to use recent hardware features, but it is not clear to me from your output whether that is truly mandatory from upstream or only that Qt will build with a hard dependency on any features it is configured to be able to use.

As always, if you have a problem, you should show us the errors that demonstrate that a problem exists. "QT5 is broke" is not an error report with which we can work.

[bschnzl]

[szatox]

Since it's marked as solved now, what was the solution actually?

[bschnzl]

Apologies!

This started as a JS8Call notice! The application was exercised to all the limits, and more!

Then QT6 failed to build on Xen. More needs to be said about that, but QT6 built when the machine was booted without Xen.

I booted back into Xen immediately after QT6 built - because I need the functions in the VMs.

JS8Call was started, so I could decode JS8 signals. I sent a query, and it sent exactly right. I finished the update within the next day. Since then, I have had the opportunity to fully exercise JS8Call and it is working flawlessly, as built in December 2023.

The solution is to upgrade from QT5!
_________________
Persistance Pays

[asturm]

[bschnzl]

[Hu]

While js8call is running, what is the output of lsof -c js8call -Pn?

[wjb]

Under some circumstances, you can actually get portage telling you there are no ebuilds that match an installed package even though there is (are?).

For example, one of my own ebuilds from my local overlay was reported as not available when doxygen moved from app-doc/ to app-text/ the other day - my ebuild still depended on app-doc/doxygen, so portage masked it quietly then essentially told me it wasn't there.

As for Qt and Xen, I've had issues with Qt related s/w in VirtualBox where certain processor features (fma and f16c) are needed for a build but are not accessible in the virtual environment.

[bschnzl]

[asturm]

[bschnzl]

Ho Ho!!!

From the lsof - QT5 is upgraded to the latest version provided by the tireless unappreciated coders here! Y'all deserve our thanks.

LSOF shows us that JS8Call only uses the QT5 parts of the system.

Which is to say, QT5 and QT6 are now running together from different packages on Gentoo. They use the same busses. The graphical system has changed. Some packages now run differently than they did before. JS8Call requires a GUI. This has an effect!

I used the term paperwork to refer to all records that tell us what the machine is doing. Gentoo lets us run so very light as to expose the gaps and edges. Everyone should want that, but even Gentoo lets users run rich. Such a wonderful world! Each package runs side by side with all the others - in the same shells, environments, and machines.

That is what makes attacks successful, and plans go awry! Boundaries get crossed.

Welcome to reality!
_________________
Persistance Pays