No no, I'm fairly sure it's not me :D I beyond-compare'd anything between prop_physics and the engine, no relevant changes on my side. I still think it'll turn out to be an issue with the linux build - eg gcc version can be a factor too, although we're on 4.2.4 so we should be good. If I end up not finding it at all I'll compile a clean OB mod on my linux machine or something and work from there, but hope that won't be necessary.

On 20/02/2011 15:39, Tobias Kammersgaard wrote:
Sounds like you need to check previous revisions of your code to be honest.

- ScarT


On 20 February 2011 14:18, Maarten De Meyer <maar...@off-limits.be <mailto:maar...@off-limits.be>> wrote:

    Found out another piece of interesting information. When we drop
    weapons, we give them the direction of the player to make the drop
    look realistic. This has always worked decently. On our linux
    server, the dropped weapons - also physics objects ofcourse -
    always have orientation 0 0 0, irrespective of the player
    orientation.


    On 19/02/2011 15:01, Tom Edwards wrote:
    Without knowing how the debug code works it's hard to say what
    that means. But it's still definitely a networking problem rather
    than something that's happening on the server. What happens when
    a predicted player walks around the affected object, hugging it
    tightly?

    ------------------------------------------------------------------------
    *From:* Maarten De Meyer <maar...@off-limits.be>
    <mailto:maar...@off-limits.be>
    *To:* hlcoders@list.valvesoftware.com
    <mailto:hlcoders@list.valvesoftware.com>
    *Sent:* Saturday, 19 February, 2011 11:14:39
    *Subject:* Re: [hlcoders] Physics props 90° angle snap issue -
    (linux?)

    vcollide_wireframe moves together with the mesh I'm afraid.

    On 19/02/2011 11:52, Tom Edwards wrote:
    If you look closely, the collision model doesn't change. It's
    only the visual, client-side mesh that's out of place. Try using
    vcollide_wireframe to see what's really going on.||

    ------------------------------------------------------------------------
    *From:* Jonathan Murphy <nuclearfri...@gmail.com>
    <mailto:nuclearfri...@gmail.com>
    *To:* Discussion of Half-Life Programming
    <hlcoders@list.valvesoftware.com>
    <mailto:hlcoders@list.valvesoftware.com>
    *Sent:* Saturday, 19 February, 2011 10:41:28
    *Subject:* Re: [hlcoders] Physics props 90° angle snap issue -
    (linux?)

    In my experience it is unlikely to be a Linux specific issue,
    and more of a prediction issue. Try introducing fake lag on a
    listen server (net_fakelag 200) and seeing if the issue is
    reproducable. That should make it much easier to debug.

    On Sat, Feb 19, 2011 at 9:33 PM, Maarten De Meyer
    <maar...@off-limits.be <mailto:maar...@off-limits.be>> wrote:

        Yes, they are.
        ------------------------------------------------------------------------
        From: Jonathan Murphy
        Sent: zaterdag 19 februari 2011 11:02
        To: Discussion of Half-Life Programming
        Subject: Re: [hlcoders] Physics props 90° angle snap issue -
        (linux?)


        Would be a interesting clue to know if all players are
        seeing exactly the same thing?

        On Sat, Feb 19, 2011 at 7:42 PM, Joel R.
        <joelru...@gmail.com <mailto:joelru...@gmail.com>> wrote:

Have you tested it on a windows dedicated server? Listenservers don't act entirely the same as dedicated
            servers.  It appears like the angle of the physics
            object is updating a bad networked angle value.


            On Sat, Feb 19, 2011 at 2:22 AM, Maarten De Meyer
            <maar...@off-limits.be <mailto:maar...@off-limits.be>>
            wrote:

                As great as Noir Desir's song is, obviously it's not
                what I wanted to show :D
                Here's the real vid with the issue:

                *http://www.youtube.com/watch?v=QhmrP6eSAek*

                sorry for that.


                On 19/02/2011 9:15, Maarten De Meyer wrote:
                Hi list,

                we're facing an issue I don't immediately know
                where to start to debug. Basically, in all of our
                maps, some physics props have an issue where the
                orientation of their model snaps at 90° angles
                instantaneously. This is not only with props, but
                also eg with our vehicles, so I'm looking
                suspicously at vphysics :). In addition, this only
                seems to happen on our linux server. Not 100% sure,
                since it does not happen consistently on all props,
                but we've tried reproducing it a lot on a windows
                listenserver without success, and on our linux
                dedicated server it happens frequently. Here's a
                video of the issue:

                *http://www.youtube.com/watch?v=NrgcRvBJYBE*

                At one point we had a vehicle that stood still with
                a driver in it, and it did a 'snap' at nearly fixed
                intervals of several seconds. Player got out, it
                went away, player stepped in, it was back ( another
                reason for me to look at vphysics with the evil eye ).

                I'm hoping someone has hit this issue before, or
                can at least point me in the right direction, or, a
                method for debugging this. Since I'm not sure if
                it's related to vphysics linux internals, the linux
                build machine/settings, networking tolerances, ...
                I'm not sure where to start looking.

                Thanks for any feedback ( or sympathy :D ),

                Maarten


                _______________________________________________
                To unsubscribe, edit your list preferences, or view the list 
archives, please visit:
                http://list.valvesoftware.com/mailman/listinfo/hlcoders



                _______________________________________________
                To unsubscribe, edit your list preferences, or view
                the list archives, please visit:
                http://list.valvesoftware.com/mailman/listinfo/hlcoders




            _______________________________________________
            To unsubscribe, edit your list preferences, or view the
            list archives, please visit:
            http://list.valvesoftware.com/mailman/listinfo/hlcoders




        _______________________________________________
        To unsubscribe, edit your list preferences, or view the list
        archives, please visit:
        http://list.valvesoftware.com/mailman/listinfo/hlcoders




    _______________________________________________
    To unsubscribe, edit your list preferences, or view the list archives, 
please visit:
    http://list.valvesoftware.com/mailman/listinfo/hlcoders



    _______________________________________________
    To unsubscribe, edit your list preferences, or view the list archives, 
please visit:
    http://list.valvesoftware.com/mailman/listinfo/hlcoders



    _______________________________________________
    To unsubscribe, edit your list preferences, or view the list
    archives, please visit:
    http://list.valvesoftware.com/mailman/listinfo/hlcoders




_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlcoders


_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives, please 
visit:
http://list.valvesoftware.com/mailman/listinfo/hlcoders

Reply via email to