Board index FlightGear Development Spaceflight

Space Shuttle - Bugfixes

Discussion about development and usage of spacecraft

Re: Space Shuttle - Bugfixes

Postby wlbragg » Fri Jan 03, 2020 8:28 am

My "clean, fresh install" of 2019 seems to be stable. I fixed one cargo hauling bug I had with the AirCrane which turns out to be self inflicted. I identified another bug, that is not a show stopper, that has to do with saved aircraft-data conflicts when switching between locations. I'm pretty sure it is also self inflicted and can be solved with more robust logic.

I am running the Shuttle launch and re-enter over and over and over and I have yet to see any anomalies.

I know there has been many individual issues as of late, including some of my own. But after doing a "clean" install and then working any immediate problems, I have accounted for all my issues, non of which point to fgfs core and so far I am able to consistent launch and re-enter without issue. Maybe I am not going far enough, or stressing the systems enough.

So, can anyone describe a procedure using 2019.2.0 and the dev version of the Shuttle that "currently" will produce an error, consistent or not and describe what they see when they get this error?

I'm still interested to know if you are required to manually put the pitch and roll buttons to auto when starting with the "Launch" variant or start up scenario.
Kansas and Ohio/Midwest scenery development.
KEQA, 3AU, KRCP Airport Layout
Intel i7/GeForce RTX 2070/Max-Q
User avatar
wlbragg
 
Posts: 7587
Joined: Sun Aug 26, 2012 12:31 am
Location: Kansas (Tornado Alley), USA
Callsign: WC2020
Version: next
OS: Win10/Linux/RTX 2070

Re: Space Shuttle - Bugfixes

Postby Thorsten » Fri Jan 03, 2020 8:33 am

That would be nice, but still asks for an explanation of why you've been burning up before and how the glide/rollout loop came to be active during launch - none of which has anything to do with persistent data.

(I don't think there's a procedure to reliably produce an error - pretty much all I've seen and heard are random glitches, and I'd almost say if it is reproducible, it's not what I'm looking for...)
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Space Shuttle - Bugfixes

Postby wlbragg » Fri Jan 03, 2020 9:05 am

why you've been burning up before

Because pitch and roll buttons were set to CSS and either never switched to auto or your supposed to place them in auto. I haven't been told that answer yet. So long as I make sure it is set to auto, I have no issues.

glide/rollout loop came to be active during launch

Caused by an unrelated nasal error that dirtied the nasal logic. That limit.nas error was the cause. The limit.nas error was caused by using --altitude=-9999 as a startup parameter.

pretty much all I've seen and heard are random glitches, and I'd almost say if it is reproducible, it's not what I'm looking for...

Right, and for me they have all been reproducible for the most part once I figured out the cause. Before I figured out the cause they appeard to be random and questionably reproducible. As of now, all explainable. I have not had any "random glitches" with a fresh, clean install once I figured out the causes of what appeared to be random issues. But i have had many that appeared random and sudden at times, but turns out to be caused by pushing the system in a development environment where things are in a constant state of flux along with trying different procedures to startup, such as using startup switches and the likes.
All I am really trying to say is that we need to make sure that a "random glitch" it truly random and not being caused by something as simple as using -9999 in a startup switch, or not realizing the autopilot is not in auto.
I know that doesn't necessarily explain the gear issue you encountered, but if -9999 can cause a limit.nas nasal error early on that put the shuttle in "glide/rollout loop during launch, then?
Kansas and Ohio/Midwest scenery development.
KEQA, 3AU, KRCP Airport Layout
Intel i7/GeForce RTX 2070/Max-Q
User avatar
wlbragg
 
Posts: 7587
Joined: Sun Aug 26, 2012 12:31 am
Location: Kansas (Tornado Alley), USA
Callsign: WC2020
Version: next
OS: Win10/Linux/RTX 2070

Re: Space Shuttle - Bugfixes

Postby Thorsten » Fri Jan 03, 2020 9:17 am

Caused by an unrelated nasal error that dirtied the nasal logic. That limit.nas error was the cause.


I'll try to reproduce that one, I can't figure out in my head how the causal chain should look like...
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Space Shuttle - Bugfixes

Postby Thorsten » Fri Jan 03, 2020 11:53 am

I swear this is driving me NUTS. Bugs appear, bugs disappear, new bugs appear,...
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Space Shuttle - Bugfixes

Postby Hooray » Fri Jan 03, 2020 12:56 pm

Thorsten wrote in Thu Jan 02, 2020 4:52 pm:The only compositor feature that is planned is the CCTV camera once we can render secondary views to canvas.



As far as I remember, that feature (patch) existed long before the compositor did - so I think it's unrelated to it, i.e. could be integrated for either renderer.
Stuart even offered to handle its review because he needed a similar feature for the FG1000, and presumably there are other aircraft/developers in need of rendering exterior views onto some sort of MFD.

Even though it would obviously make more sense in a compositor context. I think I saw commits that would suggest that Fernando recently hooked up compositor buffers to the effects framework - compared to that, the old patch adding camera support to the canvas system by supporting a new "view-camera" is relatively straightforward. I think it's been ~2-3 years by now that the patch first circulated here - the long-term idea back then was to re-implement the built-in view-manager as a dedicated canvas element, so that the existing notation/properties can be used "as is" to control such "canvas cameras": http://wiki.flightgear.org/View_manager

Anyway, I don't think any of this would be specific to the compositor build/mode - the patch really only touched $FG_SRC/Canvas and $SG_SRC/canvas, so it's not like the compositor would be required to make this work.

Here's the original topic: Canvas:View development
Please don't send support requests by PM, instead post your questions on the forum so that all users can contribute and benefit
Thanks & all the best,
Hooray
Help write next month's newsletter !
pui2canvas | MapStructure | Canvas Development | Programming resources
Hooray
 
Posts: 12707
Joined: Tue Mar 25, 2008 9:40 am
Pronouns: THOU

Re: Space Shuttle - Bugfixes

Postby wkitty42 » Fri Jan 03, 2020 2:52 pm

wlbragg wrote in Fri Jan 03, 2020 8:28 am:I'm still interested to know if you are required to manually put the pitch and roll buttons to auto when starting with the "Launch" variant or start up scenario.

since i checked yesterday, mine are defaulting to auto in the launch scenario... my problem is apparently somewhere else in my system... i'm still investigating...
"You get more air close to the ground," said Angalo. "I read that in a book. You get lots of air low down, and not much when you go up."
"Why not?" said Gurder.
"Dunno. It's frightened of heights, I guess."
User avatar
wkitty42
 
Posts: 9146
Joined: Fri Feb 20, 2015 4:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 20.04

Re: Space Shuttle - Bugfixes

Postby wlbragg » Fri Jan 03, 2020 7:21 pm

I swear this is driving me NUTS. Bugs appear, bugs disappear, new bugs appear,...

Well, that certainly doesn't give me a warm fuzzy feeling.

Is this without any changes or is there always a minor change, ie: location, scenario, command switch, etc?

Both of my most mysterious bugs of late were cause by an early nasal "nil" error, which apparently corrupted nasal space in some manner, but maybe not a full corruption like you might normally see where nothing nasal after that point works. One of them (the AirCrane) involved aircraft-saved data and then a location change (I haven't got to the bottom of that one but have an idea).
The others (the shuttle) was the -9999 altitude command line switch setting. Both caused a nasal nil error which affected a completely different logic than could be accounted for with the error message, reminds me of memory corruption.
Point being, a changing environment, so to speak, triggered them to show up.

The only other error (not truly a bug) appeared random as I didn't realize on a clean build or wipe of Shuttle launch aircraft-data that the Shuttle Launch scenario will be broken if the CSS/Auto switch is set as default to CSS. I just verified that switch is set to "CSS" by default after an aircraft-data purge. All that would take for someone to burn up on reentry is to run reentry after an aircraft-data wipe and not set that switch. If you'd been successfully running reentries all this time, clear your cash for some reason or another, come back to reentry scenario and suddenly burn up, it would look totally random. Especially if you put that switch to auto and it got "auto saved" without you making that connection.
By the way, I think we need to set that switch to auto in the "Re-entry" set file or its corresponding nasal block.

Can you identify a procedure or scenario and some steps to take that is more likely than another to generate one of the "random" bugs?

Let me know if there is something you want me to try. I can switch between 2018.3.4 and 2019.2 in seconds. Both clean, recent builds. I also have a pretty good system laid out to build interim versions if needed.
Kansas and Ohio/Midwest scenery development.
KEQA, 3AU, KRCP Airport Layout
Intel i7/GeForce RTX 2070/Max-Q
User avatar
wlbragg
 
Posts: 7587
Joined: Sun Aug 26, 2012 12:31 am
Location: Kansas (Tornado Alley), USA
Callsign: WC2020
Version: next
OS: Win10/Linux/RTX 2070

Re: Space Shuttle - Bugfixes

Postby GinGin » Fri Jan 03, 2020 7:59 pm

Let me know if there is something you want me to try. I can switch between 2018.3.4 and 2019.2 in seconds. Both clean, recent builds. I also have a pretty good system laid out to build interim versions if needed.


That might be a good idea to cross check bugs in 2019 version with the stable 2018 to see if the culprit is due to the problem that arised in 2019.
I am trying to reproduce what I read usually there in 2018, and most of the time, it is not impacted
GinGin
 
Posts: 1580
Joined: Wed Jul 05, 2017 11:41 am
Location: Paris
Callsign: Gingin

Re: Space Shuttle - Bugfixes

Postby wlbragg » Fri Jan 03, 2020 10:21 pm

We just need a consistent enough bug to cross-check.
Kansas and Ohio/Midwest scenery development.
KEQA, 3AU, KRCP Airport Layout
Intel i7/GeForce RTX 2070/Max-Q
User avatar
wlbragg
 
Posts: 7587
Joined: Sun Aug 26, 2012 12:31 am
Location: Kansas (Tornado Alley), USA
Callsign: WC2020
Version: next
OS: Win10/Linux/RTX 2070

Re: Space Shuttle - Bugfixes

Postby Thorsten » Sat Jan 04, 2020 11:15 am

I've tested the high energy entry this morning, and it basically worked right until the touchdown point - then the brakes again didn't. :(

However that's good, because if it's reproducible, I can track it down eventually... Anyone else seeing brakes doing nothing to slow down the Shuttle on the runway?
Thorsten
 
Posts: 12490
Joined: Mon Nov 02, 2009 9:33 am

Re: Space Shuttle - Bugfixes

Postby legoboyvdlp » Sat Jan 04, 2020 12:20 pm

Thorsten this is really obvious but you have updated your nightly to include the last commit by the JSBSIM team (I can get a commit hash if you like)? Because this sounds like the inoperative hydraulics in the Mig21bis where the hydraulics channel would just randomly (but consistently) not work at all. If you'd like to try the Mig21 also and see if you get hydraulic pressure on the gauge (lower right of the cockpit) - because it certainly is reproducible but now has been corrected by the JSBsim team.
User avatar
legoboyvdlp
 
Posts: 7981
Joined: Sat Jul 26, 2014 2:28 am
Location: Northern Ireland
Callsign: G-LEGO
Version: next
OS: Windows 10 HP

Re: Space Shuttle - Bugfixes

Postby legoboyvdlp » Sat Jan 04, 2020 1:48 pm

A few reports from a test:

a) Starting a 3200 entry to KSC (which came in over Mexico, not over the USA as I was expecting - but hopefully that's acceptable) started in CSS pitch and roll. If the idea is the pilot manually flies the scenario I can understand that - but is that what is supposed to happen?

b) The shuttle went immediately far below the nominal path:
Image

(more trajectory differences between versions / computers?)

It essentially cut the corner as it began to trend back towards the line after a few minutes and we did in fact reach Florida - just rather stressed at the apparent impact in the Gulf of Mexico!

c) The end of the dotted line on the trajectory map was also in Mexico as soon as the scenario started - also, despite the trajectory display starting as showing "NO HORIZON" - the horizon was in fact showing, looks like the label is not updated to the actual state of the map

Image

d) the FlightGear HUD with alpha, inclination, and distance remaining never showed up.

e) a strange warning -
Code: Select all
[SGExclusiveThread]  not finished - skipping

which has just started showing up nearly every time I fly.

f) TACANS, MLS, and radio altimeters were not powered up by the scenario.

g) Things such as cockpit detail, ADI simulation detail, MDU update speed, enhanced displays were changed by the scenario.

h) I never saw YAW CONTROL AERO - as late as Mach 1 I saw the yaw jets fire.

i) The brakes worked perfectly as far as I can see. I don't see any custom code for the brakes - so surely any brake error must be within JSBSIM or FlightGear, though?



Some of these, particularly b), e) and h) seem like genuine problems while the others may just be omissions in the scenario code.
User avatar
legoboyvdlp
 
Posts: 7981
Joined: Sat Jul 26, 2014 2:28 am
Location: Northern Ireland
Callsign: G-LEGO
Version: next
OS: Windows 10 HP

Re: Space Shuttle - Bugfixes

Postby GinGin » Sat Jan 04, 2020 5:55 pm

Hey Lego,

b) The shuttle went immediately far below the nominal path:


That's not a big deal, even not anormal depending on flight path angle at REI range ( Combination of Perigee, location of Perigee, REI location etc)

The nominal Line drawn is almost the max thermal line hard boundary.
So you can be a bit to the right, more in the safe zone
Not too much to the right where you will fall under the equilibrium line, where you will never make the target energy wise.

That is where I am usually also when I target a nominal Deorbit burn ( perigee around 20 Nm above the target and REI around 4000 Nm)

Image

Image


d) the FlightGear HUD with alpha, inclination, and distance remaining never showed up.


The one with have by pressing H ?
Did you have coherent informations on the PFD for the range to target and Delta Azimuth ?
If no, could you check in the entry dialog that the terrain is correctly selected or reselected again after scenario start ?
GinGin
 
Posts: 1580
Joined: Wed Jul 05, 2017 11:41 am
Location: Paris
Callsign: Gingin

Re: Space Shuttle - Bugfixes

Postby wkitty42 » Sat Jan 04, 2020 6:14 pm

legoboyvdlp wrote in Sat Jan 04, 2020 1:48 pm:e) a strange warning -
Code: Select all
[SGExclusiveThread]  not finished - skipping

which has just started showing up nearly every time I fly.

i'll copy my post from the mailing list concerning your post there about this...

wkitty42 on the dev mailing list wrote:On 1/4/20 7:49 AM, Jonathan R wrote:
> Hi,
> This error message has started showing up on almost every session of FlightGear:
>
> [SGExclusiveThread] not finished - skipping
>
> I've no idea what it is or where it came from - but it does sound somewhat serious if threads aren't being dealt with properly?


this is part of richard's nasal garbage collection code... it basically says that the collection thread has been triggered but the previous execution has not been completed yet so this execution is being skipped... the next execution should be ok to run and collect new garbage... sometimes, if there's a lot of garbage, this will be seen...

AIUI, this message is benign and nothing to worry about...

"You get more air close to the ground," said Angalo. "I read that in a book. You get lots of air low down, and not much when you go up."
"Why not?" said Gurder.
"Dunno. It's frightened of heights, I guess."
User avatar
wkitty42
 
Posts: 9146
Joined: Fri Feb 20, 2015 4:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 20.04

PreviousNext

Return to Spaceflight

Who is online

Users browsing this forum: No registered users and 4 guests