Reply
Thread Tools
Posts: 1,414 | Thanked: 7,547 times | Joined on Aug 2016 @ Estonia
#21
Testing time - I would need some help with testing. In particular, I am interested if the map widget works as intended.

Please use the provided app for testing only - that would then not deplete the map views and would be of great help for development.

So, I am looking for feedback regarding crashes and functionality. Please get the app at https://build.merproject.org/package...pbox-demo-sfos , choose your architecture, get build RPM and install using

Code:
pkcon install-local RPMfilePath
On start, you should see Helsinki map, as shown on the screenshot. You should be able to move, zoom in and out, change orientation (bearing slider), and tilt (pitch slider).

Please let me know how it went and specify
  • Device
  • SFOS version
  • Did it crash?
  • If crashed, please start in terminal.
    • Did it crash in terminal?
    • If crashed in terminal, please paste stdout
  • Is the map displayed properly?
  • Are the names rendered?
  • Does movement work?
  • Does zoom work?
  • Does rotation work?
  • Does tilting work?

If you see that someone already tested for your device/SFOS combination there is probably no need to test more.
Attached Images
 
 

The Following 2 Users Say Thank You to rinigus For This Useful Post:
Posts: 339 | Thanked: 1,623 times | Joined on Oct 2013 @ France
#22
Using mapbox-sfos-0.1+master.20170916171905.daedaf3-10.6.1.jolla.armv7hl.rpm and mapbox-sfos-debuginfo-0.1+master.20170916171905.daedaf3-10.6.1.jolla.armv7hl.rpm
  • Device : Jolla 1
  • SFOS version : not yet updated, still on 2.1.0.11 (cat /etc/sailfish-release)
  • Did it crash? Yes
  • If crashed, please start in terminal.
    • Did it crash in terminal? Yes
    • If crashed in terminal, please paste stdout
      Only a segfault on stdout. Running through gdb on device gives thoses details (devel-su pkcon install gdb; gdb mapbox-sfos; type "r "+ Return):
      Code:
      Program received signal SIGSEGV, Segmentation fault.
      [Switching to LWP 11606]
      0x2a2333a8 in mbgl::RenderSymbolLayer::render(mbgl::PaintParameters&, mbgl::RenderSource*) ()
  • Is the map displayed properly? Before crash map is displayed OK
  • Are the names rendered? Yes
  • Does movement work? Yes
  • Does zoom work? Yes
  • Does rotation work? Yes
  • Does tilting work? Yes

Some notes:
So basically, the apps is working great but is killed after playing with it for some time (can be as low as 5 seconds, to 30 seconds), whatever the user action is (bearing, zooming...).

I have noticed a black square (a few pixels wide) on the top left of the screen. Not sure what it is, as it seems to always be there, no matter the map shown. Can provide a screenshot if needed.



Seeing those labels realign themselves while changing the bearing looks great and is really encouraging !

Please let me know if there is a way to get you more output from gdb (call trace, local variables....). I will update the version on my Fedora desktop to see if I can see the same failure or not.
 

The Following User Says Thank You to Zeta For This Useful Post:
Posts: 1,414 | Thanked: 7,547 times | Joined on Aug 2016 @ Estonia
#23
Zeta, thank you very much for a thorough test!

On OnePlus X, same SFOS, the app is rock solid. I have not had any crashes after I changed OpenGL interface with QML to match the current implementation in QtLocation. Since I don't have any access to J1:

* would you mind to provide backtrace?

I can then search for this crash in Mapbox GL Native issues and maybe we would have to open an issue. We could also try to get the latest version of qt bindings (there seem to be one commit missing when compared to the current QtLocation version).

Let's see the backtrace and move from there.
 
Posts: 33 | Thanked: 29 times | Joined on Aug 2008 @ Netherlands
#24
Hi,

Impressive. Your demo app works flawless on my oneplus one running 2.1.1.24. And smooth.
 

The Following User Says Thank You to wdehoog For This Useful Post:
Posts: 172 | Thanked: 353 times | Joined on Nov 2014
#25
  • Device: Onda v820w V5
  • SFOS version: 2.0.5.6 (i486)
  • Did it crash? Yes
  • If crashed, please start in terminal.
    • Did it crash in terminal? Yes
    • If crashed in terminal, please paste stdout
Code:
[nemo@Sailfish ~]$ mapbox-sfos
[F] unknown:0 - This application failed to start because it could not find or load the Qt platform plugin "wayland"
in "".

Available platform plugins are: minimal.

Reinstalling the application may fix this problem.
Aborted
[nemo@Sailfish ~]$
Reinstalled mapbox with the same result.
 

The Following 2 Users Say Thank You to tmi For This Useful Post:
Posts: 1,414 | Thanked: 7,547 times | Joined on Aug 2016 @ Estonia
#26
@tmi: I will have to add i486 Sailfish 2.0.5.6 as a target. At present, only Qt 5.6 is supported on i486 since I expected that everyone moved from older releases. I'll look into it later today. It would take probably several hours of compilation on OBS though
 

The Following 2 Users Say Thank You to rinigus For This Useful Post:
Posts: 172 | Thanked: 353 times | Joined on Nov 2014
#27
Originally Posted by rinigus View Post
@tmi: I will have to add i486 Sailfish 2.0.5.6 as a target. At present, only Qt 5.6 is supported on i486 since I expected that everyone moved from older releases. I'll look into it later today. It would take probably several hours of compilation on OBS though
ATM 2.0.5.6. is the latest port for Onda v820w, but I'll be upgrading as soon as 2.1.x.x is available. I just don't know when that happens.
Personally, I can very well wait for the SFOS update so you can work on other issues if time is limited.
 

The Following 2 Users Say Thank You to tmi For This Useful Post:
Posts: 1,414 | Thanked: 7,547 times | Joined on Aug 2016 @ Estonia
#28
Originally Posted by tmi View Post
ATM 2.0.5.6. is the latest port for Onda v820w, but I'll be upgrading as soon as 2.1.x.x is available. I just don't know when that happens.
Personally, I can very well wait for the SFOS update so you can work on other issues if time is limited.
OK, looks like 2.0.5.6 would require some additional effort and let's postpone it (see https://build.merproject.org/package...86_latest/i586 for error message). Probably used QML/OpenGL interaction is not compatible with Qt 5.2
 

The Following User Says Thank You to rinigus For This Useful Post:
Posts: 339 | Thanked: 1,623 times | Joined on Oct 2013 @ France
#29
Originally Posted by rinigus View Post
* would you mind to provide backtrace?
It looks like I get differents outputs each time I run it...
The Jolla 1 is easily subject to OOM killer, maybe that would be such case ? Not sure where to look for this, as journalctl reports that no journal exists...

Some of these outputs, I installed the libc missing symbols at some point, but it didn't change much the output...

Code:
Program received signal SIGSEGV, Segmentation fault.
[Switching to LWP 13452]
0x4100bb22 in free () from /lib/libc.so.6
(gdb) bt
#0  0x4100bb22 in free () from /lib/libc.so.6
#1  0x434e1152 in ?? ()
#2  0x434e1152 in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
Code:
Program received signal SIGSEGV, Segmentation fault.
[Switching to LWP 13452]
0x4100bb22 in free () from /lib/libc.so.6
(gdb) bt
#0  0x4100bb22 in free () from /lib/libc.so.6
#1  0x434e1152 in ?? ()
#2  0x434e1152 in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
Code:
Program received signal SIGABRT, Aborted.
[Switching to LWP 14103]
0x40fc7be6 in ?? () from /lib/libc.so.6
(gdb) bt
#0  0x40fc7be6 in ?? () from /lib/libc.so.6
#1  0x40fd7d2e in raise () from /lib/libc.so.6
#2  0x40fdabb4 in abort () from /lib/libc.so.6
#3  0x410009f8 in ?? () from /lib/libc.so.6
#4  0x410009f8 in ?? () from /lib/libc.so.6
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
My last try, I though it wouldn't crash after shaking it for a long time, but it finally crashed after close to 1 minute maybe, with an additional message:
Code:
*** Error in `/usr/bin/mapbox-sfos': malloc(): smallbin double linked list corrupted: 0x2ade60c0 ***
[New LWP 3120]

Program received signal SIGABRT, Aborted.
[Switching to LWP 3120]
0x40fc7be6 in ?? () from /lib/libc.so.6
(gdb) bt
#0  0x40fc7be6 in ?? () from /lib/libc.so.6
#1  0x40fd7d2e in raise () from /lib/libc.so.6
#2  0x40fdabb4 in abort () from /lib/libc.so.6
#3  0x410009f8 in ?? () from /lib/libc.so.6
#4  0x410009f8 in ?? () from /lib/libc.so.6
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

After a reboot (Android VM not started by default), the memory looks like this (only SSH connected, no actions locally); if it help to compare to other platforms :
[nemo@Sailfish ~]$ free
total used free shared buffers cached
Mem: 828692 719220 109472 0 35308 419416
-/+ buffers/cache: 264496 564196
Swap: 627908 0 627908
How to know if the OOM killer is the one that closed the app ?
If not him, then it looks like some memory corruption problem, leading to segmentation fault and corrupted call stack...

Edit : anyone else with a Jolla 1 can confirm/refute what I see on my device ?
 

The Following 2 Users Say Thank You to Zeta For This Useful Post:
Posts: 56 | Thanked: 146 times | Joined on Sep 2013 @ Norway
#30
Testing: mapbox-sfos-0.1+master.20170916171905.daedaf3-10.6.1.jolla.armv7hl.rpm

Device: Jolla C
SFOS version: 2.1.1.26
Did it crash? No
Is the map displayed properly? Yes
Are the names rendered? No
Does movement work? Yes
Does zoom work? Yes
Does rotation work? Yes
Does tilting work? Yes

POI: When changing bearing or pitch is seems than names of POI disappear.
Attached Images
  

Last edited by PamNor; 2017-09-17 at 18:37.
 

The Following 2 Users Say Thank You to PamNor For This Useful Post:
Reply

Thread Tools

 
Forum Jump


All times are GMT. The time now is 16:34.