OpenCPN Beta Version 3.3.1824 Released

[Ptizef]

Originally Posted by Ptizef

I would not conclude two quickly, but it seems that these crashes could be related to the touch screen mode, but more certainly to the target tracks display function.
Jean Pierre

I have no means to simulate so this morning I tested "as if in navigation" :
- GPS and AIS switch on feeding O, a track and a route activated, but the boat not moving
I got several crashes, always the same scenario
- three hours with the two parameters off -> no crash
- Activated both -> crash within half an hour
But now I ould be more focused on the touch screen mode because I got one with only this parameter on and the AIS track display off
Unfortunately, I was never proposed to send a crash report

Jean Pierre

[Ptizef]

I have found a scenario witch creates crash almost each time
No need of GPS/AIS feed, neither waiting time
Just start Opencpn with a route visible and Touch screen mode selected
click or touch a route point to highlight the route and try to move it . If you alternate finger and mouse, you will get very quickly a crash
If you have selected "lock way points ... " the crash will be quasi immediate but in any case the crash will happen
But to reproduce it you will need a touch screen device and may be Windows 8.1, as I was enable to reproduce on my Vista which has a "normal" screen
Jean Pierre

PS: I would like to add that the last tests was done with a home made compilation of the last GIT (3.3.1911) and it happens both with GL and no GL mode

[rgleason]

Pitzef, you get the BETA CRASH KING award for this week!! Excellent work. I challenge someone to take it away from him! Please imagine that a large soft helmet will arrive in the mail....

[bdbcat]

jp...

Thanks for the information.

I am trying to reproduce your situation, and so far no crash. But I'll keep trying. I am running Win8.1 on a tablet, with GPS and AIS tracking coming in by wifi UDP.

If you can build from source (3.3.1911), can you run OCPN inside VS2010 debugger and see where exactly the crash occurs? A stack trace would be very useful.

Dave

[cagney]

Dave and Ptizef

O seems to have a problem with, what seems to be, a malformed saildoc gribfile.
I received this file a while ago http://first-escape.com/sail/doc/GFS...14154505494.7z

The grib-file caused this crash, using the latest git code:



Grib interpolation was off. No difference with "on".
The crash is repeatable with a fresh opencpn.conf and no other gribfile causes a crash. Editing opencpn.conf, deleting any reference to the file also solves the problem.
No problems on a later download from saildocs.

Is it possible to do a sanity check on gribfiles?

Thomas

[seandepagnier]

I probably fixed this 2 days ago but the pull request is not yet merged

Can you try

github.com/seandepagnier/OpenCPN

[Ptizef]

Originally Posted by bdbcat

jp...
If you can build from source (3.3.1911), can you run OCPN inside VS2010 debugger and see where exactly the crash occurs? A stack trace would be very useful.
Dave

Hi Dave
Unfortunately for the moment, there is no development environment on this net book
I compile on my vista and copy the result to it.
... And I am unable to reproduce on the Vista device which has no touch screen

Jean Pierre

[Ptizef]

Oups ! Sorry but I get another crash with compiled 3.3.1911 (last git)
When opening track properties by any mean -> crash
This time reproducible on both machines and I was able to recreate one in debug mode, so you can find attached the crash debug dump
I also sent two crash reports
Jean Pierre

[bdbcat]

jp...

OK, the root cause of the latest crash on Track Properties is clear enough. Resulted from starting OCPN with tracking on, but no initial track actually started. All this comes from the "improvement" in initial track conditions implemented in 3.3.1911.

Fixed in github now.

I don't see how this could be closely related to the Win8.1 touch interface fault. Still trying to reproduce that one. Had my 8.1 touch tablet running all day, pushing and pulling on routes occasionally, GPS and AIS running and tracking. No faults.....

I am beginning to suspect that the touch driver on your netbook is somehow different from the one on my tablet. Touch screens are surely different than mice. For example, the action of a touch on the screen is interpreted and executed when the finger is lifted, not when the finger is first pushed. This is in contrast to a mouse click which is executed on the down-click, and generally nothing happens on the up-click.

Anyway, all this means that the code path is different for touch screens and mice.


Does your netbook operate well and smoothly by touch interface? Is there anything you can tell me about the touch driver? How does this touch screen generate right-clicks? By press-hold?

Can you say positively that the crash only occurs when you touch the screen? That is, if you just leave it running without interaction, will it crash in 2 hours or so, all by itself?



Dave

[Ptizef]

Hi Dave

1) compiled new git : track properties crash fixed OK
I agree that this has nothing to do with the touch screen mode crash as this last one happened in 1824 in witch the track properties worked.

1) I use an ASUS A102 net book . The touch screen has a very basic possibilities to change parameters. I left in in the default state : Double click is generated by double touch, and right click by a long touch
With the touch screen mode activated, it works fine. Panning, zooming (by two fingers as for my smart phone) are very good and smooth. The only black point for me is creating routes. If you have to put a route point outside the screen, the touch to drag creates points and it's not easy to get the route you want . Moving a route point is not easy too.
but that is another story...
I am not very experienced in tablet , but I think it should not be very different. The true difference is that I can mix touches and mouse clicks

3) The crashes happened in navigation always after around 2 hours and always with both AIS tracks and touch screen mode and always after a user action (gragging, touching, clicking etc...)
I never had a crash without. yesterday, I got a crash without AIS track, so I eliminated it and focused on the touch screen mode.
In fact I got as many crashes I wanted, in any conditions as soon as the touch mode was activated (and not the track properties one ! ) juste by clicking, touching , dragging, trying to move route point, but always near the boat. So yesterday evening, I would have answered YES! it is related to the touch mode
Today, after the compilation of the new git, I don't know. I can't for the moment reproduce and I am still testing
Jean Pierre

[Ptizef]

Hi Dave
OpenCPN is running for 15 hours now ... no crash !
I am confused but this is a fact ... and a good news

Jean Pierre

[Ptizef]

Originally Posted by boat_alexandra

I probably fixed this 2 days ago but the pull request is not yet merged

Can you try

github.com/seandepagnier/OpenCPN

Hi Sean

I tried your last commit
The file still crashes.
it seems to have some BMS data and something do not work in interpolate these data

The only way to make it working is to deactivate BMS data
I don't know if BMS data are really useful for us ?
Jean Pierre

[seandepagnier]

Originally Posted by Ptizef

Hi Sean

I tried your last commit
The file still crashes.
it seems to have some BMS data and something do not work in interpolate these data

The only way to make it working is to deactivate BMS data
I don't know if BMS data are really useful for us ?
Jean Pierre

It didn't crash, but didn't interpolate BMS data correctly before. I thought I fixed this. I don't know if it is useful either. Maybe it is ok to disable BMS data since it is a bitfield, but the actual raw data has an undefined value as well. If you can post the grib file in question I'm sure I can correct the crash.

The other crash was when it tried to load an invalid grib file (eg file.bz2 which was not a grib file)

[Ptizef]

Sean

You can get the file here

http://first-escape.com/sail/doc/GFS...14154505494.7z

I had curious behavior with this file
In debug mode it crashed on "delete m_pTimelineSet"
Correcting this permit to work but still crashed in release mode with the same sources !
A the end, only deactivating BMS data allowed it working in any case
There is something wrong in BMS treatment, but I didn't find it
Jean Pierre

[rgleason]

This Grib file crashes instantly when I click on Grib in Win Opencpn v3.3.1824 because Grib is trying to load the most recent file which is this file! I am going to delete the file now.