Console Error Msg: Digital Filter Absolute Value

[Alant]

I see it on all aircraft that I have tried - 707, and my VC10 and TSR2 WIPs.

[polly]

OK Tks for that. With the breakpoints noted above I can at least see the pathname of the file being parsed immediately before the error logs. Tracing through the parsing process will be a different kettle of offal.

[Alant]

Does this filter actually work in an autopilot?
Perhaps we should hand it back to the original author now we know that it is to blame for the error message.

[polly]

Well, I didn't expect to find the navselector in any of the models you mentioned, or in the UFO. We see the error message point to either <absolute> or <value> tags and point to coherent-noise filter and to gain filter. I thought it worth stepping through the initialisation of a failing filter to try to see if it's a general failure in proprule initialisation.

[wkitty42]

carry on, polly! you may find the actual problem and provide a fix before anyone else can

[polly]

I'm beginning to wonder. I can't see the error with the UFO that mariuszXC is getting and single stepping causes the code to barf ( technical term for segfaulting ) during the xml read.
I've tried git bisecting but anything earlier than 2023 seems unable to build, ( something something std::mutex )
On to hedge hopping with the TSR-2, that travesty of policy still hurts.

Edit
For the 707: I don't see these errors. These:

Code: Select all

 0.48 [ALRT]:terrain    Unexpected property in listener /sim[0]/rendering[0]/static-lod[0]/detailed[0]
    0.48 [ALRT]:terrain    Unexpected property in listener /sim[0]/rendering[0]/static-lod[0]/rough-delta[0]
    0.48 [ALRT]:terrain    Unexpected property in listener /sim[0]/rendering[0]/static-lod[0]/bare-delta[0]


are different, I see them on all A/C in next and have raised a ticket for this issue.

VC10 Same, cannot see this particulare error message, just the LOD listener as 707 above

BAC-TSR2 ditto, only the LOD listener

That all brings me back to the only culprit seems to be navselector.xml; I wonder if it's to do with picking up alii from a separate parameter file ?

[Alant]

Could it be Fgdata "Ground shaking effect" by Stuart Buchanan on 30/12/20, ?
This has connected changes in g-forces.xml, defaults.xml and cockpit-view.xml.

[polly]

@alant , Yes, thanks, I'd forgotton I had deleted the g-forces reference locally while I was tracing navselector.

I sourced a couple of tickets, one to do with the LOD listener error messages and one for this issue:

It doesn't actually say, in the wiki, that <input> stanza may contain anything other than a simple property and, indeed, tracing throgh the configuration calls it looks to me as if there's support for conditions, values etc inside the <enable> stanza but not for the <input> stanza.

I'll look forward to any replies on the ticket but it looks like, "If it hurts when we do this" we shouldn't do it; perhaps g-forces.xml and navselector.xml need rethinking. Thanks, hupeoples (!) for all the help.

[polly]

The ticket https://sourceforge.net/p/flightgear/codetickets/2801/ has an answer, it suggests navselector.xml needs a re-do. I cannot see who is the author, is that Stuart ? The same applies to g-forces.xml regarding the <absolute-value> tag , I suppose.

[MariuszXC]

There is a little more to that. I tried to submit a fix via SourceForge, but its servers seem to be barfing at me at the moment... So here it is, a one-liner actually. It fixes the cause for 'unknown config node absolute' message from Autopilot.

Code: Select all

diff --git a/src/Autopilot/digitalfilter.cxx b/src/Autopilot/digitalfilter.cxx
index a79e057da..5ca7bbb0f 100644
--- a/src/Autopilot/digitalfilter.cxx
+++ b/src/Autopilot/digitalfilter.cxx
@@ -839,6 +839,7 @@ bool CoherentNoiseFilterImplementation::configure(SGPropertyNode& cfg_node,
 
     if (cfg_name == "absolute") {
         _absoluteVal = cfg_node.getBoolValue();
+        return true;
     }
     return false;
 }


[xDraconian]

This has been merged. Thanks @MariuszXC

[MariuszXC]

Great and apologies if there are multiple merge requests. Me and Sourceforge servers didn't have a good time together today.