dialog box construction

  • 1
  • Problem
  • Updated 4 years ago
  • (Edited)
While this is, what I consider, a lower priority than others, I've noticed on startup Where the panadapter is never instantiated. Generally, a shutdown and restart resolved this but this last startup, not only did the panadapter not get instantiated, the RX dialog only got partially built. A subsequent restart did fully start the GUI. While it is annoying, beyond that this is mostly FYI. This is on SSDR (for Windows) 1.4
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes

Posted 4 years ago

  • 1
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9152 Posts
  • 3478 Reply Likes
Try updating your video drivers.

And are you operating on a PC natively running Windows?
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
Tim, I am running very current Win 7 natively. I believe the partially built TX box was due to no panadapter being instantiated therefore no slice receiver therefore no frequency or mode or antenna selector or AGC. So really the issue is startup failed to create the panadapter.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9152 Posts
  • 3478 Reply Likes
I still suspect some unexpected interaction with the video subsystem and .NET, particularly if you did not change the state of the radio hardware in between the GUI restarts.
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
Tim, As I said, the fields missing on the RX dialog/activity fragment are fields that would be populated by the active slice receiver. This points to the root cause being the panadapter is not always instantiated. That happened occasionally with 1.3.8, not sure why. It's happening a lot more with 1.4. 1.3 - once every few months, 1.4 - once every few days? Again, not sure why. I used to think I started SSDR before SDRCAT was completely up. I am actually not sure that makes a difference. Just because CAT invokes API.Init(), I am sure the gui does as well. If I shut SSDR down and restart, all is fine. Actually, occasionally (perhaps in the 1.3.x timeframe) when SSDR was restarted it would come up in perhaps an initial mode of all multiple, not sure if all, panadapters.

Assuming SSDR knows it has not finished a normal startup, is there a log or something that can be used as breadcrumbs to determine why? If not, I'd like that to be a suggestion for 1.5 or 1.4.1, where logging is active during startup. If the startup completed normally the logging is stopped and the log erased but if it fails to start the log could be used as the trail of what occurred, even if the log was encrypted by XOR'ing it with something to hide the inner workings, it could be sent to you guys for evaluation. This would help diagnostics a great deal.

On the issue of you suspecting it is a bad video subsystem. I honestly don't think so. As I said, it doesn't always happen, it just happens more frequently in 1.4. If you are using the same WPF calls and the graphical side of SSDR is largely the same, the issue in not on my machine. All the other apps work fine, the version of .NET is identical, everything is identical except the version of ssdr.

Since I can not desk check the UI code, there is a sequence which open the dialogs/activity fragments (Android term)/ child windows, that appear on the right side of the screen. This is always done in advance of drawing the panadapter. There is, however a greyish background with the greyish very lazy S TM/logo which get overwritten as the panadapter(s) slide out from right to left. It's like sliding a window over another window. I am not sure why it slides out vs. just appearing, but it slides out. I generally just use the one panadapter. On a 15" laptop monitor, more than one is self defeating. So here is my guess, the person writing the GUI would know for sure. Somewhere, I suspect on the radio, there is a persisted state (aren't there new issues with persistence?) such that when the radio is gracefully shut down the state of the GUI is preserved, panadapter(s), slice receiver(s) maybe other stuff. Generally, on startup, that information makes it to the SSDR for drawing purposes. I am guessing there is an issue in that section where sometimes either that information isn't properly persisted, properly recovered. I don't know what else but there are 3 states:
1) everything persisted and recovered correctly - new desktop pretty much like desktop before termination.
2) things either didn't persist didnt recover - no panadapter constructed
3) things either didn't persist or didn't recover - all possible panadapters drawn, at least one horizontal, at least 2 stacked in lower horizontal plane.

Would you please run this by someone who knows the startup behavior of the UI? .NET wouldn't know how to start multiple panadapters. When the multiple panadapters are created at startup there is usually a pause then they roll over right to left and they all have slice receivers at some frequency.

Thanks,
Walt
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
update: Just now a panadapter was instanciated on 17 but there was no slice receiver so no content in the RX dialog box on bottom right (above equalizer).

Does SSDR perform identical termination whether triggered from File->exit or OS shutdown event handler? Generally I File->exit but this last time I told WIndows to reboot. One has to write the exist handler logic. For instance, on the TX dialog, RF power is full Tune power is almost minimum and no profile. I was using a system profile that limits rf power to about 20w. After a File->exit, that box is properly restored. Perhaps the issue is there is no shutdownHandler attached such that shutdown wouldn't persist anything. From msdn.windows.com:

"The shutdown conditions that are described by the ShutdownMode property are specific to an application. In some cases, though, an application may shut down as a result of an external condition. The most common external condition occurs when a user ends the Windows session by the following actions:
  • Logging off

  • Shutting down

  • Restarting

  • Hibernating

To detect when a Windows session ends, you can handle the SessionEnding event"

I believe that is part of the issue.
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
Tim, would you restate what you meant by, "if you did not change the state of the radio hardware in between the GUI restarts"? Do you mean like disconnect the GPS antenna? I think that is the only thing I could do to effect the state of the radio hardware. And I did not remove the GPS antenna.

Thanks,

Walt -kz1f
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9152 Posts
  • 3478 Reply Likes
Plugged or unplugged any thing or power cycled the radio.
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9152 Posts
  • 3478 Reply Likes
Can you take a screen capture of what you are observing?
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
OK, I was ready to take a screenshot. While I didn't expect this to fail (you can see the panadapter widget just starting to emerge on the right, next to the vertical bar of controls. In what I was describing up top the panadapter never gets instantiated.
case 1) the panadapter is instantiated with the slc last active, the RX control reflect slice
case 2) the panadapter is instantiated w no slc even though one existed prior to shutdown, RX control not populated
case 3) the panadapter is not instantiated (very similar to below) and RX control not populated.
it is cases 2 and 3 I was referring to. Case 1 is clearly what should always happen.
Does this help visualize what I was describing up top?

(Edited)
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
Unlike the last screen shot, which never got a response, I have another one, not where you can see a panadapter start to roll out, as above, but never rolls out. I was successful however in eventually getting a panadapter via the 'add panadapter' button in the bottom left. That took between 30 and 60 seconds to appear. However, what is very noteworthy is on any band I move to the slice receiver ALWAYS starts with B. Nowhere is there an A slice. Remember too that when I attempted to move this problem to the API corner as one of the return code possibilities when display pan fails is "Unable to get foundation receiver assignment". Would this explain why I cannot get an A slice? Again, the following picture denote no panadapter being created at all. You can also see where the RX dialog bottom right is not completely built. This makes sense as there is no Panadapter created. Is the, coupled with the inability to create an A slice something that would lead to an answer?
Photo of Tim - W4TME

Tim - W4TME, Customer Experience Manager

  • 9152 Posts
  • 3478 Reply Likes
This sounds like you have an issue with the internal persistence database.  I recommend a reset to factory defaults.  If you have transmit or global profiles, do an export of those without the Preferences option checked before you do the reset.
(Edited)
Photo of Steve Cook

Steve Cook

  • 15 Posts
  • 0 Reply Likes
I have had this happen a few times also I will try to get a capture when this happens to me.
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
It doesn't happen all the time here either. It does seem to happen more than with 1.3.8. Steve, when you see it again, pls try to manually create a new pan via the left most button on the bottom left (as I recall). The tooltip will say create new panadapter. This may look like it fails but 45 seconds or so +/- it likely will create one. If/when that happens see if the slice receiver instantiated is B, not A. Or if you already know A isn't created pls update so the follks at FRS will have a confirmation.

Thanks,

Walt - kz1f
Photo of Eric - KE5DTO

Eric - KE5DTO, Official Rep

  • 718 Posts
  • 211 Reply Likes
Walt and Steve,

Can you tell us more about your PC configuration?  What CPU/OS/RAM/Video card?  
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
Hi Eric,
  Dell XPS 15Z (or Z15) I7 processor, Win7, Video says
Intel HD Graphics Family
NVIDIA GeForce GT 525M
8GB memory

When I was trying to ask about the errors out of display pan .... I did so as that would be a, perhaps, the place where the creation of a pan would fail. One of the possible errors and, since this happens on fresh starts, the likely candidate reason is something to the effect of unable to create a receiver. When I created the pan manually, the receiver that got built was Slice B. Slice A no longer existed. There is also another thread open with the identical problem and more people saying they have experienced this.
(Edited)
Photo of DrTeeth

DrTeeth

  • 1687 Posts
  • 387 Reply Likes
Eric,

AcerAspire 5575G; CPU Intel i5-2450@2.5GHz, Win 7 x64, 8Gb, SSDR using nVidia 630M card (dedicated memory).

No similar prob with 1.3.8 for months' of use.
Photo of Eric - KE5DTO

Eric - KE5DTO, Official Rep

  • 718 Posts
  • 211 Reply Likes
Thanks for the info.  We will add this to our issue detail.
Photo of F4CLB, Jean Louis

F4CLB, Jean Louis

  • 5 Posts
  • 0 Reply Likes
Hello, sorry for my english .........

I have reproduced the same problem using a laptop (Win 8.1 64 celeron 2x1,2ghz, 4GB ram) less efficient than my usual desktop PC (Win 7 64 3.6 ghz cpu i7 12GB Ram)

I end up with the same SmartSDR screen has the pictures, but with audio and operation of the ATU and other

I think this is due to the PC, which loses frames Flex sources, suddenly the program does not have the information for display,

If I open a 2nd receiver and well it works properly

FYI, CPU usage on the i7 10% or less, and 80% on the laptop.


73 Jean Louis / F4CLB
Photo of DrTeeth

DrTeeth

  • 1687 Posts
  • 387 Reply Likes
I never had this issue with 1.3.8 in all the months I was using it. Was using 1.4.0 until 1.4.3 came out and now have the issue approx 1 start in 4-5. It does not seem to make any difference as to which order I close/shut down things; I generally close SSDR then shut down the rig.
(Edited)
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
Guy, would you give Eric your environment, like computer, mem, cpu, video card etc. Refer to his request up a few posts.
Photo of DrTeeth

DrTeeth

  • 1687 Posts
  • 387 Reply Likes
Done Walt.
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
I upgraded to 1.4.11 last night. After the upgrade I started ssdr and it fired up as I'd expect it to. This morning I started it and, once again, it started with no panadapter and no slice information for those dialog boxes on the right. The screen looks just like my post of one month ago. I had left display alone as I opened this site, found this issue, and updated it. Nothing happened, the pan never appeared, so it's not a timing thing. I restarted SSDR and it started fine. I do not believe it is a broken anything. I restarted SSDR, not the radio to test if the radio proper was in a bad state. It appears it wasn't.

Eric, is there any movement, further info on what causes this?

Thanks,

Walt - kz1f
(Edited)
Photo of Eric - KE5DTO

Eric - KE5DTO, Official Rep

  • 718 Posts
  • 211 Reply Likes
Yes, there is.  We have identified what is likely causing this problem, but unfortunately it was too late to get the fix into this release (v1.4.11).  Now that we have identified the problem, we should be able to address this for the next round.  Thanks for your patience on this elusive issue (#1906).
(Edited)
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
Thank you Eric.
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
Eric, good news, I now have onPanadapterAdded notifications.
Do you recall this issue? Here is what I just saw in my trace
received reply  R20|0|Unable to get foundation receiver assignment
request 20 was ....
sending cmd C20|slice create  pan=0x40000000 freq=10.0000006 mode=AM

At one point in this thread I surmised the failure to display a pan was it couldn't get the slice for it.
Photo of Eric - KE5DTO

Eric - KE5DTO, Official Rep

  • 718 Posts
  • 211 Reply Likes
We have fixed this in our latest Alpha release.  It was not an issue in FlexLib, but rather in SmartSDR-Win.
Photo of Walt - KZ1F

Walt - KZ1F

  • 3040 Posts
  • 643 Reply Likes
OK, that's actually very good news indeed. I just thought where I saw it in the log that might be helpful. Although...where I am not using SSDR-win, this is a little intriguing.