Board index FlightGear Support Tools ATC-pie

ATC-Pie 1.2.2 crash when handover a strip

ATC-pie is a radar air traffic control simulation program for the FlightGear multi-player network.

ATC-Pie 1.2.2 crash when handover a strip

Postby LoCall » Mon Apr 03, 2017 3:35 am

ATC-Pie 1.2.2
Fedora 25 / full update
qt5 dependencies installed
+ (more a zillion of *qt5*) :mrgreen:

# myTaylor is !rich but myHardDisk.manufacturer is ;
// I am not trying to relaunch a 16 GB conversation. :wink: I am just saying I have installed --all-- the *qt5* rpm-packages except 3 related to lastfm
... so, yes there was a lot... :shock:

I have seen/read the other post with D-ECHO and Myckybadia pkg dep list,
I have read and re read trusted and untrusted the README

In 100% of the cases + on 3 different machines I have the same error message when, in ATC-Pie, I try to handover any strip, I click + drag the strip over any "envelop symbol" in ATClist
and => freeze ... crash / and it jump back in the shell with the following message :

----------------------------------------------
Traceback (most recent call last):
File "/mnt/disk2/1/Zone1/shares/FGFS/Actives/ATC/ATC-Pie/ATC-Pie-1/models/ATCs.py", line 177, in dropMimeData
yesNo_question(self, 'Confirm handover', 'You are about to send your strip to %s.' % atc_callsign, 'Confirm?'):
File "/mnt/disk2/1/Zone1/shares/FGFS/Actives/ATC/ATC-Pie/ATC-Pie-1/GUI/dialog/miscDialogs.py", line 57, in yesNo_question
box = QMessageBox(parent_window)
TypeError: arguments did not match any overloaded call:
QMessageBox(parent: QWidget = None): argument 1 has unexpected type 'AtcNeighboursModel'
QMessageBox(QMessageBox.Icon, str, str, buttons: Union[QMessageBox.StandardButtons, QMessageBox.StandardButton] = QMessageBox.NoButton, parent: QWidget = None, flags: Union[Qt.WindowFlags, Qt.WindowType] = Qt.Dialog|Qt.MSWindowsFixedSizeDialogHint): argument 1 has unexpected type 'AtcNeighboursModel'
Traceback (most recent call last):
File "/mnt/disk2/1/Zone1/shares/FGFS/Actives/ATC/ATC-Pie/ATC-Pie-1/models/ATCs.py", line 177, in dropMimeData
yesNo_question(self, 'Confirm handover', 'You are about to send your strip to %s.' % atc_callsign, 'Confirm?'):
File "/mnt/disk2/1/Zone1/shares/FGFS/Actives/ATC/ATC-Pie/ATC-Pie-1/GUI/dialog/miscDialogs.py", line 57, in yesNo_question
box = QMessageBox(parent_window)
TypeError: arguments did not match any overloaded call:
QMessageBox(parent: QWidget = None): argument 1 has unexpected type 'AtcNeighboursModel'
QMessageBox(QMessageBox.Icon, str, str, buttons: Union[QMessageBox.StandardButtons, QMessageBox.StandardButton] = QMessageBox.NoButton, parent: QWidget = None, flags: Union[Qt.WindowFlags, Qt.WindowType] = Qt.Dialog|Qt.MSWindowsFixedSizeDialogHint): argument 1 has unexpected type 'AtcNeighboursModel'
Aborted (core dumped)
----------------------------------------------

I have read the posts relevant but had no success to find a usable hint
(but like I have said I am still débutante in python)
So i do not know it he following message is meaningful or not.
I already have tried to dive in the python code but a hint of someone more comfortable in this matter will be handful.

Did I miss something / package / ... ?
If it an alteration is really needed in the code, perhaps someone may point me out on a direction with some good hint ...
I will of course come back here and share the result of your suggestions

Any how, Thank you for your time:D, and Thank you Mickybadia for ATC-Pie. :D
Freedom and Chaos are easily taken for each other.
I am afraid if we let grow the Chaos it is going to suffocate the Freedom.
But if we wipe out the chaos putting everything "in adjusted cut",
I know for sure the Freedom is gone.
User avatar
LoCall
 
Posts: 9
Joined: Sat Aug 06, 2016 10:45 pm
Location: UK B'ham
Version: 2017.2.1
OS: Linux

Re: ATC-Pie 1.2.2 crash when handover a strip

Postby mickybadia » Mon Apr 03, 2017 7:21 am

Thank you for spotting this, I see what the problem is. I will solve it tonight (Europe). In the meantime you can untick the "confirm handovers" option to avoid the problem.
mickybadia
 
Posts: 394
Joined: Tue Sep 24, 2013 9:12 am

Re: ATC-Pie 1.2.2 crash when handover a strip

Postby LoCall » Mon Apr 03, 2017 2:22 pm

Thank you MickyBadia,

I appreciate the attention given
and
the "en attente" suggestion to untick the handover option.

I have started, and will continue, regular ATC sessions at Paris Orly.
I prefer as ATC-tool your "ATC-pie".

I have already accumulated some interesting experience, and will continue on this line.

Do I remember correctly, you wrote somewhere about the need to have people to test ATC-Pie and provide an usable feed back ?

If you think it could be of benefit to set up a test protocol, do not be reluctant to get in touch with me, I would be please to share my modest time and resources.

Cheers.
Freedom and Chaos are easily taken for each other.
I am afraid if we let grow the Chaos it is going to suffocate the Freedom.
But if we wipe out the chaos putting everything "in adjusted cut",
I know for sure the Freedom is gone.
User avatar
LoCall
 
Posts: 9
Joined: Sat Aug 06, 2016 10:45 pm
Location: UK B'ham
Version: 2017.2.1
OS: Linux

Re: ATC-Pie 1.2.2 crash when handover a strip

Postby mickybadia » Mon Apr 03, 2017 9:18 pm

Sure, I am always happy to learn more about who is in the regular user base. If you are familiar with Git and want to be discussing/testing the features before they are released in the numbered versions, join the testing list and checkout "dev". Contact me for details if you need any. Your bug report is already fixed there.

LoCall wrote in Mon Apr 03, 2017 2:22 pm:I have started, and will continue, regular ATC sessions at Paris Orly.
I prefer as ATC-tool your "ATC-pie".
I have already accumulated some interesting experience, and will continue on this line.

Quick side note not to go off topic: the included .dat for Orly contains no ground routes so you will not feel the full power of the pie when controlling ground (taxi/parking instructions generation), nor be able to play solo GND, unless you find a better LFPO.dat or make one.
mickybadia
 
Posts: 394
Joined: Tue Sep 24, 2013 9:12 am

Re: ATC-Pie 1.2.2 crash when handover a strip

Postby LoCall » Tue Apr 04, 2017 2:56 am

Thank you for your side note, Yes, of course, I see, it is off topic and, of course, I will not epilogue on it here.

So, in order to close the "opened box" with the shorter answer I can : I am reworking the whole LFPO airport.
LFPO is already amazing but have some errors and as you said some missing.

It is a crazy work. Tons of Source Info, ASE·NASA images, (no-google), ICAO "Airport Best Practices" and similars, visits with the "ufo" and tests.
Blender, Gimp, texturing and re texturing, re tests, WED, Tons of hours, Patience and Tea are my actual best friends . :mrgreen:
And I do not forget the recompilation of the scenario. Already a huge time invested on all that.

If some one wish to epilogue on this, -> new topic and, not here in ATP-pie sub forum.

Back on topic,

I will have a look on to the said dev-list.
But before to go there, I still will to share here without waiting any longer:
The "multimedia / sound " could very likely be the culprit of the tons of the sigsevs I had (in average 1x by 15min )

I was testing which check/options was related with the seg faults, 1 by 1.
Unchecked, then run ATC-pie, ...crash ...test next.

Then, with *no sound* at all ,perfect !

I was able to run ATC-pie 4 hours long with zero crash.
Loaded the more I could in Solo Game with
GND+TWR+DEP+APP ...
+ 18 planes
+ all graphic options
+ additional radar window
+ and never ending [alt+tab]
...4 hours long = ZERO crash !
(of ATC-pie, the Radar supp window closed by itself 3 times in 10 minutes but the rest have ran ok)

These next days I will try to run the same in sub dimensioned computers and I will share back the result.
(let me know if here or in in the dev list but my familiarity level with python is "cousin" not "brother"..)

Hope that give some "useful spot of light" onto the debugging direction.

Cheers.
Freedom and Chaos are easily taken for each other.
I am afraid if we let grow the Chaos it is going to suffocate the Freedom.
But if we wipe out the chaos putting everything "in adjusted cut",
I know for sure the Freedom is gone.
User avatar
LoCall
 
Posts: 9
Joined: Sat Aug 06, 2016 10:45 pm
Location: UK B'ham
Version: 2017.2.1
OS: Linux

Re: ATC-Pie 1.2.2 crash when handover a strip

Postby mickybadia » Tue Apr 04, 2017 9:21 am

I understand you are now reporting on a different issue than the handover one, i.e. a crash every ~15 min unless sounds are muted. If so, please say:
  • commit number or correct tarball version (the one in the topic is invalid);
  • send the full console output from command line to end;
  • if you can, say if this happens when you are NOT touching the program at all, and with any individual sound selection in the notification dock. (+ what type of game or connection status, etc. -- any strictly reproducible scenario is welcome)

Given the possible technicality and the time you are spending on it, it would make sense to join the testing list, but note that serious debugging will only be easy if you switch to Git. No fluency in programming or kinship with Python is required however. Otherwise this type of stuff could be placed as a ticket on the bug tracker.
mickybadia
 
Posts: 394
Joined: Tue Sep 24, 2013 9:12 am

Re: ATC-Pie 1.1.2 crash when handover a strip

Postby LoCall » Tue Apr 04, 2017 8:25 pm

commit number or correct tarball version (the one in the topic is invalid);

yeek, my bad, it is of course a bad typo, sorry. :oops:
CORR 1.1.2.
by the way, it could be good to check into the sourceforge tarball.
It have an error of the same type.
The tarball filename say 1.1.2 but the line 33 of the settings.py say version_string = '1.1.1'
On the other hand the git version is ok (1.1.2)
(checked this 2017/04/04:1600 UTC)

I do not want to go deeper tek here in the forum. I continue in the dev-list.

Just for the other users who may experience the same crash /segmentation fault message,
could we let clear this ugly but working workaround :
- in the ATC-pie menu, disable all SOUNDS ;
- in the ATC-pie options (3 places) disable all kind of confirmations.

send the full console output from command line to end;

ooops...
in my case, python3.5 -vvv ATC-pie ICAO echo more then 10'000 lines in the console...
But I am going to be nice :
- there is no error message in the output > 10'000 lines.
- there are 4 warnings about some hooks around coordinate manipulation at the exact time when the windows are opened.
then and ATC-pie continue as long as an external EVENT to display, trigger the crash with SEGMENTATION FAULT message in the console.

...and yes, this crash happen even if I am NOT interacting with the program.
again, details in the DEV-LIST (even with the never ending 10'000 lines if it please you.)

Cheers
Freedom and Chaos are easily taken for each other.
I am afraid if we let grow the Chaos it is going to suffocate the Freedom.
But if we wipe out the chaos putting everything "in adjusted cut",
I know for sure the Freedom is gone.
User avatar
LoCall
 
Posts: 9
Joined: Sat Aug 06, 2016 10:45 pm
Location: UK B'ham
Version: 2017.2.1
OS: Linux

Re: ATC-Pie 1.1.2 crash when handover a strip

Postby mickybadia » Wed Apr 05, 2017 9:08 am

LoCall wrote in Tue Apr 04, 2017 8:25 pm:CORR 1.1.2.
by the way, it could be good to check into the sourceforge tarball.
It have an error of the same type.
The tarball filename say 1.1.2 but the line 33 of the settings.py say version_string = '1.1.1'
On the other hand the git version is ok (1.1.2)
(checked this 2017/04/04:1600 UTC)

The way I understand it, this statement is incorrect. I have just checked the v1.1.2 tarball currently up for download. Quoting line 34 (not 33):
Code: Select all
version_string = '1.1.2'

Your files may be messed up. Have you at some point been replacing/overwriting files manually over prior downloads and missed a few? Please figure this out first. Download again clean or clone from Git, and let us continue on the list.

LoCall wrote in Tue Apr 04, 2017 8:25 pm:in my case, python3.5 -vvv ATC-pie ICAO echo more then 10'000 lines in the console...

Starting with -vvv is indeed asking for a lot of trace. I was referring to the output from ATC-pie itself first, though I will have a look if you send me that log.
mickybadia
 
Posts: 394
Joined: Tue Sep 24, 2013 9:12 am

Re: ATC-Pie 1.2.2 crash when handover a strip

Postby LoCall » Wed Apr 05, 2017 2:35 pm

Ok MickyBadia let us continue on the dev-list

About the file mess, sorry, I really do not know.
I am very procedural and I do not easily mess up my files system.
But I guess no one is error proof.
In the mean time I have switched to the git version and erased all previous downloads ...to avoid possible mess up...
So the event logs will be of the "new" version.

Toward
Running now the git version + :
- 1. activated abrt
- 2. will after first runs add sys.settrace at the start of your code.
It is going to sort out ton of infos which must be filtered out.
But I am very interested in the result. (feed back not here but in the DEV-LIST)

For the users not intersted in the debugging :

Some may have noted WARNING MESSAGES in the console :
Code (): Select all
QXcbConnection: XCB error: 3 (BadWindow), sequence: 17797, resource id: 9751401, major code: 40 (TranslateCoords), minor code: 0
QXcbConnection: XCB error: 3 (BadWindow), sequence: 59094, resource id: 9104083, major code: 40 (TranslateCoords), minor code: 0
QXcbConnection: XCB error: 3 (BadWindow), sequence: 34653, resource id: 9082201, major code: 40 (TranslateCoords), minor code: 0
QXcbConnection: XCB error: 3 (BadWindow), sequence: 58972, resource id: 9801647, major code: 40 (TranslateCoords), minor code: 0
Reload: colour configuration
QXcbConnection: XCB error: 3 (BadWindow), sequence: 21331, resource id: 8871314, major code: 40 (TranslateCoords), minor code: 0
QXcbConnection: XCB error: 3 (BadWindow), sequence: 7209, resource id: 9878048, major code: 40 (TranslateCoords), minor code: 0
QXcbConnection: XCB error: 3 (BadWindow), sequence: 55329, resource id: 9003647, major code: 40 (TranslateCoords), minor code: 0

there is no reason to worry about.

ATC-pie work very well (considering the SOUND + CONFIRMATION workarounds)

Theses messages does happen at start, NOT at SIGSEV (Crash) time.

to MickyBadia and other DBGrs, AFAIK,
SEGMENTATION FAULT do not likely happen because of python itself but "around" python.
Python use several Cpp extensions, some are for the memory management.
Some nasty problems may be happening exactly there.
Debugging "only the python code" could very well lead to pass by the cause without seeing it.
+
In certain architectures, in RH systems, the use of ctypes in python have already raised problems.
(I have wrote I am not an "expert" in python but I am not a dummy.)
So the segmentation faults are NOT manifesting there but, as said, they can have theses nasty "long roots"
And of course the activated trace logs come up suitably with infos expectedly interesting.
Feed back in few days in DEV-LIST

Cheers.
Freedom and Chaos are easily taken for each other.
I am afraid if we let grow the Chaos it is going to suffocate the Freedom.
But if we wipe out the chaos putting everything "in adjusted cut",
I know for sure the Freedom is gone.
User avatar
LoCall
 
Posts: 9
Joined: Sat Aug 06, 2016 10:45 pm
Location: UK B'ham
Version: 2017.2.1
OS: Linux

Re: ATC-Pie 1.2.2 crash when handover a strip

Postby mickybadia » Mon Apr 17, 2017 4:26 pm

Regarding the topic here, I have bumped to 1.1.3 to include a fix for the crashing handover confirmation.
mickybadia
 
Posts: 394
Joined: Tue Sep 24, 2013 9:12 am


Return to ATC-pie

Who is online

Users browsing this forum: No registered users and 1 guest