Fast Threading Crash

[ianfhunter]

When I reduce my gamespeed to 0.1 - this happens sometimes:

[xcb] Unknown request in queue while dequeuing [xcb] Most likely this is a multi-threaded client and XInitThreads has not been called [xcb] Aborting, sorry about that. python: ../../src/xcb_io.c:179: dequeue_pending_request: Assertion !xcb_xlib_unknown_req_in_deq' failed. Aborted (core dumped)

It's starting to crop up in the intervention interface when I run it at 0.5 :S , so could someone verify it's not just my particular computer giving this error?

First thoughts are that it's clashing threads

[ianfhunter]

Stopped happening at normal speed on intervention, i think it was cuz my cpu was under heavy load. But still, it'd be good to figure this out.

[ianfhunter]

"Something somewhere is trying to do something with the UI or UI objects from the wrong thread. " https://groups.google.com/forum/#!msg/wxpython-users/6rKyY_ICUSc/Idi7L3HrY9QJ

Maybe the fact the engine calls the UI update :s that's way to much to fix.

Or possibly the event threads? More likely I think

[ianfhunter]

Yeah it's back. Occurs when doing inquiries and interventions a lot interchangeably. Good thing interventions was high value :P

[rogersst]

This only happens to me when I set the gamespeed to 0.01. I'm pretty sure it's a performance issue. I think its happening because there isn't enough time between turns for the engine to crunch all the numbers fully. Solution: Don't set the gamespeed so low.

On 28 March 2014 21:55, Ian Hunter notifications@github.com wrote:

Yeah it's back. Occurs when doing inquiries and interventions a lot interchangeably. Good thing interventions was high value :P

Reply to this email directly or view it on GitHubhttps://github.com/ianfhunter/TeamCrab/issues/57#issuecomment-38972445 .

[scottcunningham]

Maybe we could write a function that checks config values on game start to make sure the user doesn't do anything ridiculous?

[ianfhunter]

@rogersst can you pull iteration7 and run the scenario with 7 sites on normal speed.

Once it's up can you open inquiry, intervention,inquiry, intervention about 7 times?

:seven: :seven: :seven:

[ianfhunter]

Made sure it wasn't a memory thing again. It's not

And actually, thinking about it - The inquiry and Interventions aren't really in new threads, they just take control of the current thread.

[rogersst]

Ian, nothing is happening with that?

And scott, I don't think that's our responsibility to be honest. Like, my laptop couldn't run a AAA game on Ultra settings, and quite frankly it's my own fault if anything breaks if I did try to run one. :P

On 28 March 2014 22:01, Ian Hunter notifications@github.com wrote:

Made sure it wasn't a memory thing again. It's not

Reply to this email directly or view it on GitHubhttps://github.com/ianfhunter/TeamCrab/issues/57#issuecomment-38973007 .

[ianfhunter]

@rogersst well, that's a relief. it might just be that my laptop is really hot right now. I'll try in the AM.

I'll put something in the readme for the issue. There's already a small note in the config

[ianfhunter]

Yeah, this is still happening :/ I guess I'll try see if it happens with just interventions, or inquries too

[ianfhunter]

Fixed. So I was looking into xInitThread errors and I was starting to get code like:

        try:
            x11 = ctypes.cdll.LoadLibrary('libX11.so')
            x11.XInitThreads()
            print "XInitThreads"
        except:
            pass
        pygame.display.flip()
        try:
            x11 = ctypes.cdll.LoadLibrary('libX11.so')
            x11.tk_init(&argc, &argv);
            print "XInitThreads"
        except:
            pass

But the actual solution was to change pygame.display.flip() to pygame.display.update()

I don't know why, but I presume it has something to do with the underlying openGL calls - http://www.pygame.org/docs/ref/display.html#comment_pygame_display_update

[scottcunningham]

Great job, detective Hunter. I'm impressed by how you keep finding fixes to the obscure errors caused by PGU. Nice one!

On 31 March 2014 17:29, Ian Hunter notifications@github.com wrote:

Fixed. So I was looking into xInitThread errors and I was starting to get code like:

    try:
        x11 = ctypes.cdll.LoadLibrary('libX11.so')
        x11.XInitThreads()
        print "XInitThreads"
    except:
        pass
    pygame.display.flip()
    try:
        x11 = ctypes.cdll.LoadLibrary('libX11.so')
        x11.tk_init(&argc, &argv);
        print "XInitThreads"
    except:
        pass

But the actual solution was to change pygame.display.flip() to pygame.display.update()

I don't know why, but I presume it has something to do with the underlying openGL calls - http://www.pygame.org/docs/ref/display.html#comment_pygame_display_update

Reply to this email directly or view it on GitHubhttps://github.com/ianfhunter/TeamCrab/issues/57#issuecomment-39109533 .

[scottcunningham]

s/PGU/pygame/ :p

On 31 March 2014 18:13, Scott Cunningham cunninsc@tcd.ie wrote:

Great job, detective Hunter. I'm impressed by how you keep finding fixes to the obscure errors caused by PGU. Nice one!

On 31 March 2014 17:29, Ian Hunter notifications@github.com wrote:

Fixed. So I was looking into xInitThread errors and I was starting to get code like:

    try:
        x11 = ctypes.cdll.LoadLibrary('libX11.so')
        x11.XInitThreads()
        print "XInitThreads"
    except:
        pass
    pygame.display.flip()
    try:
        x11 = ctypes.cdll.LoadLibrary('libX11.so')
        x11.tk_init(&argc, &argv);
        print "XInitThreads"
    except:
        pass

But the actual solution was to change pygame.display.flip() to pygame.display.update()

I don't know why, but I presume it has something to do with the underlying openGL calls - http://www.pygame.org/docs/ref/display.html#comment_pygame_display_update

Reply to this email directly or view it on GitHubhttps://github.com/ianfhunter/TeamCrab/issues/57#issuecomment-39109533 .

[ianfhunter]