Board index FlightGear Release candidates

Please test the 3.6 RC

Release candidate testers are encouraged to post their feedback here. Please read the introduction topic for details.
Forum rules
Please read the introduction topic for details.

Re: Please test the 3.6 RC

Postby clrCoda » Tue Oct 20, 2015 8:39 pm

more like this :)

Image

--Ray
Ray St. Marie
clrCoda
 
Posts: 1225
Joined: Wed Apr 07, 2010 12:04 pm

Re: Please test the 3.6 RC

Postby clrCoda » Tue Oct 20, 2015 8:51 pm

It occurs to me that some of us can only render canvas thru Rembrandt and I don't use Rembrandt regularly because of the problem it has rendering complicated hotspots.

If a hotspot is based on the wire frame of an object, for some of us the delay in rendering the hotspot after keying ctrl-c is many many minutes, making hotspots a real PIA. If instead those hotspots are the square kind where one side is down-value and the other side up-value, this is not a problem.

I tend to hybrid planes that have canvas nav displays for the same reason, using a previous pre-canvas nav display. I can see the canvas nav display thru Rembrandt, but not otherwise.

--Ray
Ray St. Marie
clrCoda
 
Posts: 1225
Joined: Wed Apr 07, 2010 12:04 pm

Re: Please test the 3.6 RC

Postby wkitty42 » Tue Oct 20, 2015 9:00 pm

clrCoda wrote in Tue Oct 20, 2015 8:39 pm:more like this :)

the aircraft center inside the sim is fully populated in my just updated 3.7 from git... it has been that way for weeks... the one thing i don't know about is why some of the thumbnails seem to take forever to download or are not cached locally but mine is fully populated as the other pic showed...
"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: 9147
Joined: Fri Feb 20, 2015 4:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 20.04

Re: Please test the 3.6 RC

Postby clrCoda » Tue Oct 20, 2015 9:01 pm

thanks wkitty :) but this is about the release candidate :)
Ray St. Marie
clrCoda
 
Posts: 1225
Joined: Wed Apr 07, 2010 12:04 pm

Re: Please test the 3.6 RC

Postby wkitty42 » Tue Oct 20, 2015 9:02 pm

ahhh... right... sorry 'bout that ;)
"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: 9147
Joined: Fri Feb 20, 2015 4:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 20.04

Re: Please test the 3.6 RC

Postby Hooray » Tue Oct 20, 2015 9:06 pm

clrCoda wrote in Tue Oct 20, 2015 8:51 pm:It occurs to me that some of us can only render canvas thru Rembrandt and I don't use Rembrandt regularly because of the problem it has rendering complicated hotspots.


this is the first time I am hearing about this - so please file a proper bug report, including hardware/os -and driver- information, but also if this is specific to any effects/shaders you are using. Ideally, try it first with all shaders/effects disabled (minimal profile), and please use the map-canvas dialog for testing, so that this is not specific to the aircraft/location you are using.

Basically, everything Canvas related should work - if it doesn't, it's a bug, and we need to know about it - even if just that means that some ATI/AMD drivers may be incompatible with the Canvas currently works.

Under the hood, the Canvas code is using the same RTT/FBO mechanism that other od_gauges are using, so we need to understand what is going on here.

For testing, the tooltips/aircraft center and/or the map-canvas dialog should work.

PUI is the legacy GUI engine and it not working properly is usually due to certain effects/shaders, and can be worked around using a different ui style
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: Please test the 3.6 RC

Postby Richard » Tue Oct 20, 2015 9:33 pm

Soitanen wrote in Tue Oct 20, 2015 8:13 am:AMD Radeon R9 200 Series (infp from Windows device manager).


The builtin bitmapped font also fails to render correctly on my R9 290 (win7,8,10) - sometimes it flickers, sometimes depending on the view angle it appears, but usually it doesn't. However it does appear most of the time without ALS or Rembrandt.

I run the alternate GUI style that uses txf fonts and they work fine, but anything in the bitmapped font when using ALS or Rembrandt generally doesn't appear. It's something to do with the R9 200 series card as I used to have a GTX 460 and that worked fine. I don't get any GLSL errors in the console.
Richard
 
Posts: 810
Joined: Sun Nov 02, 2014 11:17 pm
Version: Git
OS: Win10

Re: Please test the 3.6 RC

Postby clrCoda » Tue Oct 20, 2015 10:07 pm

I pretty much only run with minimal shaders because they are of little use at 30000 feet+ and I'm not looking around the airports for pretty, just taxi, the few minutes I'm at depart or arrive. Performance being more important for my flights than pretty .

seems most things work in the canvas map with the exception of fixes and vors:

Vors only
Image
Fixes and vors
Image

I will do my best to included a bug report. ( under my breath is heard -- so it too can be ignored like every other report I've ever produced -- end attitude ) :)

--Ray

POST EDIT sorry, meant to say first one tests old gui under rembrandt second rembrand off newer gui style.
Ray St. Marie
clrCoda
 
Posts: 1225
Joined: Wed Apr 07, 2010 12:04 pm

Re: Please test the 3.6 RC

Postby clrCoda » Thu Oct 22, 2015 8:05 am

Flightgear still crashes if exit with a dialog open. Adding bug report.
--Ray
Ray St. Marie
clrCoda
 
Posts: 1225
Joined: Wed Apr 07, 2010 12:04 pm

Re: Please test the 3.6 RC

Postby Hooray » Fri Oct 23, 2015 4:51 pm

Thank you for your bug report, and thank you for using the proper channel (the issue tracker) for making bug reports.

Regarding the Canvas issue shown above (missing symbols for VORs/FIXes): Could you please also post a screen shot of the aircraft center - or alternatively, a screen shot of the dialog rendered when running the following Canvas snippet via the Nasal Console (see the debug menu):

http://wiki.flightgear.org/Canvas_Snipp ... ter_Images

Code: Select all
var (width,height) = (320,160);
var title = 'My new Window';

# create a new window, dimensions are WIDTH x HEIGHT, using the dialog decoration (i.e. titlebar)
var window = canvas.Window.new([width,height],"dialog")
 .set('title',title);

# adding a canvas to the new window and setting up background colors/transparency
var myCanvas = window.createCanvas().set("background", canvas.style.getColor("bg_color"));

# Using specific css colors would also be possible:
# myCanvas.set("background", "#ffaac0");

# creating the top-level/root group which will contain all other elements/group
var root = myCanvas.createGroup();

# path is relative to $FG_ROOT (base package)
var path = "Textures/Splash1.png";
# create an image child for the texture
var child = root.createChild("image")
    .setFile(path)
    .setTranslation(100, 10)
    .setSize(130, 130);


The reason being two-fold: You are using the map-canvas dialog on ATI/AMD hardware, which is known to have some driver issues relating to shaders/effects, but also PLIB/PUI (our legacy UI engine, which is probably the reason for you using the legacy UI theme?).

So to ensure that this is unrelated, it would be good to check a single Canvas dialog, without it using any PUI.
Equally, the MapStructure framework is using an instancing/caching scheme for those symbols that are white boxes in your screen shot - the caching scheme works by using osg::Image, so it would be good to see if other OSG images show up correctly for you or not - we can then conclude if the problem is related to raster images in general, if it is specific to PUI interoperability or the MapStructure framework - but also determine if shaders/effects (or your driver version) play a role here.

The snippet, when executed, should end up looking like this:
Image

Thank you
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: Please test the 3.6 RC

Postby clrCoda » Fri Oct 23, 2015 8:05 pm

Hooray, could that be the reason I only see load splash screens using rembrandt on and just a white box where the splash should go when rembrandt off? ( with rembrandt off the spinner gets spread over the entire screen and you can kind of see it spread over the screen if one is looking for it )

Thanks for your above post. I'll do my level best to understand what you said. :)

--Ray
Ray St. Marie
clrCoda
 
Posts: 1225
Joined: Wed Apr 07, 2010 12:04 pm

Re: Please test the 3.6 RC

Postby clrCoda » Fri Oct 23, 2015 8:14 pm

@Hooray,

The pic with the classic style 0 was just for this specific test. Normally I would use style 1 anthrax. I did the two styles simply to demonstrate the problem in both styles.


A bit more difficult to notice from my pics is that they are at different monitor resolutions.
The difference in resolution between the two pics was because for me on Radeon x300 Rembrandt only ever seems to work in full screen mode in either 1024X768 or in my monitors max rez of 1680x1050. I can, after sim load use the shift-10 to window the sim, but can't load rembrandt initially to a window.

Thanks again,
Ray
Ray St. Marie
clrCoda
 
Posts: 1225
Joined: Wed Apr 07, 2010 12:04 pm

Re: Please test the 3.6 RC

Postby Hooray » Fri Oct 23, 2015 8:17 pm

Yes, that could indeed be related.
If you can, could you take a screen shot while loading FG (i.e. showing the white splash screen) ?

In general, this seems to be due to an incompatibility between the way your driver works, and the way FG is using OpenGL.

Unfortunately, FG contains tons of legacy OpenGL code that hasn't been maintained/touched in years. This includes much of the legacy GUI engine (plib pui), the 2D panel/HUD code, but also the splash screen routines.

OSG however, needs to be aware of other GL code, and sometimes, the combination of these features end up crippling the driver - i.e. too much going on at once (legacy code, new code, effects/shaders, legacy code, osg code, and even more legacy code).

It seems your driver cannot deal with the way FG is using OpenGL - once you have checked if the raster image snippet works or not, I would next suggest to try some of the standalone executables, i.e. fgviewer/osgviewer.

In summary, the Canvas effort is all about unifying the 2D rendering back-end to get rid of legacy code in favor of using modern OSG code, which is not just about MFDs or the GUI, but also the HUD, 2D panels and even the splash screen routines used during startup.

Thus it would be good to know if other od_gauge based displays are working correctly or not, e.g. the NavDisplay (non-Canvas), or the hard-coded Map dialog. If those work correctly elsewhere, Canvas can surely be made to work properly, too - even though this may require a few workarounds

So the reason for my asking all these questions is to determine if/where and how other related features work properly.

My suggestion would be to disable the splash screen code during startup, try the minimal startup profile (non rembrandt!!!) and then run the snippet shown above via your nasal console (debug menu) - to ensure that we can check whether OSG images (raster images) work elsewhere or not - ideally, the bug would be related to effects/shaders and/or PUI breaking compatibility with your graphics drivers.
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: Please test the 3.6 RC

Postby clrCoda » Fri Oct 23, 2015 8:20 pm

I shared this earlier in the week,

Image

In a bit I shall do as you asked with that cool nasal script.

--Ray
Ray St. Marie
clrCoda
 
Posts: 1225
Joined: Wed Apr 07, 2010 12:04 pm

Re: Please test the 3.6 RC

Postby Hooray » Fri Oct 23, 2015 8:27 pm

Yes, I saw that, which is why I think it is a PUI/PLIB (legacy OpenGL) related incompatibilty with your graphics drivers - because the icon in the upper left corner is a PNG image (raster image) using osg::Image internally, and it shows up just fine. In other words, I expect other raster images to show up properly just as well - in fact, the underlying Nasal code is using the exact same mechanism that the MapStructure code is using, i.e. texture maps via nested Canvases.

You could also try the "Canvas self test" via the debug menu (configure development extensions) - that should also render a moving map-like dialog using the exact same framework, and if those white boxes show up there as well, it would be unrelated to PUI/PLIB, but might still be related to effects/shaders.

It should look similar to this (assuming you are at KSFO):
http://wiki.flightgear.org/Canvas_MapStructure
Image
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

PreviousNext

Return to Release candidates

Who is online

Users browsing this forum: No registered users and 2 guests