Re: [hlcoders] Loading woes
a quick question is this freezing happening when the server is set to insecure for these users? could be some bug involving the download of the vac security module. On 9/28/07, Jeremy [EMAIL PROTECTED] wrote: -- [ Picked text/plain from multipart/alternative ] We're getting users that have game freezing during the load phase of the game. A number of them have reported having the same problem in TF2. I myself have had this problem in the past as well. Both times I had this problem the only way I could fix it was to reinstall steam and re-download everything. Simply re-downloading the cache files didn't fix it, and at the time I wasn't aware of the blob file delete trick to try. Also clearing source sdk base and re-running to get fresh files didn't fix it. Supposedly one of the recent TF2 updates recently improved the situation for a couple of people, so I'm wondering if there was a load time bug found and fixed in TF2 that still effects other mods? The first time it happened to me it was at the end of the load bar, and steam would eat 100% cpu until I killed the process. The second time it happened to me was at about the 50% mark. What could steam possibly be doing when loading a 3rd party mod? Now that we've released of course we're getting similar reports by some users. Some are only getting it in FF, some are also seeing it in TF2. One thing seems sure to me though, and that is that the problem is unlikely to be in our mod, and seems more likely to be an engine or steam level thing. This is quite a problem as there is really nothing we can do to debug what is going on when this happens during the load phase. In most cases these problems coincided with steam updates. The team made it a habit to run sdk base after steam updates to 'refresh' the engine dlls and such, presumably from the gcf files. This didn't always work though, and I was under the impression that running games from the steam menu should in effect refresh these files as part of the process. Valve, is there any means by which we can debug these sorts of issues? Is there any known load time problems and/or any that have been fixed since the last sdk base update? Any information about this is appreciated. Jeremy -- ___ 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
Re: [hlcoders] Creating a Sphere Trace
Jay Stelly wrote: But it's not a framework that lets you sweep a vcollide against the entire scene - which is probably what you want (and what John wanted as I recall). In fact it's pretty straightforward to extend vphysics to do this, but the current implementation would be incredibly slow if you used any displacements at all because it would have to consider each triangle separately. That would be really nice. Or actually, would have been really nice. And kinda overdue, since the SDK-side framework is there d) physics tests only happen once per frame in static positions - not exactly sure what you are trying to say here but it sounds incorrect to me. The movement in vphysics though not described as sweeps/traces is still computed as continuous motion. It's not like all of the objects move and then do a static test to see if they are inside another object or have passed through another object. The collisions are computed continuously along the path of translation/rotation. It may still be a pain for your player movement strategy, but it certainly isn't evaluated in static positions if that helps to understand it better. I realize that VPhysics fundamentally does continuous collisions. I've played around some with the Bullet physics library some, and an idea of how things work. (Whether this is a correct idea or not is another question.) What I mean was that from game code, you cannot evaluate collisions on a collideable from more than one position in a given frame, or move/sweep them during the frame for movement tests. -John Sheu ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
Re: [hlcoders] Creating a Sphere Trace
-- [ Picked text/plain from multipart/alternative ] Not having a method of tracing a VCollide to the world really limits what we can do with prediction, especially when talking about vehicles. Since VPhysics is essentially unpredictable in the networking sense, we can't run the collision code client-side (unless we limit ourselves to an AABB, which really won't work for proned players, cars, trucks, planes etc..). The ability to sweep a collidable and implement our own physics in game code would fix this... There is a SweepCollideable function in the physics interface, but it is stubbed. We've found a happy medium, which is to have predictable simulation code much like the players and having a Physics Shadow follow the object server side. This works for flying vehicles (since most of the time they don't hit stuff). Any mod working with ground vehicles has a huge problem though, since they're always colliding. Not predicting vehicles is an option, but expect your player base to complain about laggy controls... On 9/29/07, John Sheu [EMAIL PROTECTED] wrote: Jay Stelly wrote: But it's not a framework that lets you sweep a vcollide against the entire scene - which is probably what you want (and what John wanted as I recall). In fact it's pretty straightforward to extend vphysics to do this, but the current implementation would be incredibly slow if you used any displacements at all because it would have to consider each triangle separately. That would be really nice. Or actually, would have been really nice. And kinda overdue, since the SDK-side framework is there d) physics tests only happen once per frame in static positions - not exactly sure what you are trying to say here but it sounds incorrect to me. The movement in vphysics though not described as sweeps/traces is still computed as continuous motion. It's not like all of the objects move and then do a static test to see if they are inside another object or have passed through another object. The collisions are computed continuously along the path of translation/rotation. It may still be a pain for your player movement strategy, but it certainly isn't evaluated in static positions if that helps to understand it better. I realize that VPhysics fundamentally does continuous collisions. I've played around some with the Bullet physics library some, and an idea of how things work. (Whether this is a correct idea or not is another question.) What I mean was that from game code, you cannot evaluate collisions on a collideable from more than one position in a given frame, or move/sweep them during the frame for movement tests. -John Sheu ___ 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
Re: [hlcoders] World shader override
-- [ Picked text/plain from multipart/alternative ] As far as I know you can't override the base shaders, you have to make new ones. You can make some that use lightmaps but the problem is that Hammer won't create lightmaps for materials other than LightmappedGeneric, so you have to constantly hotswap the shaders between compiling your map and running your mod. Maybe you can swap them at runtime by checking a certain material variable (like the way we have custom material proxies)... I dunno. On 9/20/07, Tobias Kammersgaard [EMAIL PROTECTED] wrote: -- [ Picked text/plain from multipart/alternative ] Doing a custom lightmappedgeneric shader could probably give you the effect you wanted. Not too sure though :)! /ProZak On 20/09/2007, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Hey list, I'm wondering if there is an equivalent of the ForceMaterialOverride for the world. Say I wanted to render eg. the entire world ( without the entities, just the brushwork ) in just one color, or better yet shaded but untextured, is there a feasible way to do this? [ I'm not talking post-process, I do mean the world rendering itself ]. ForceMaterialOverride itself works for all entities except the world. Thanks for any info, -- 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
Re: [hlcoders] drag and drop
-- [ Picked text/plain from multipart/alternative ] I would assume this is because the mouse is only updated after your OnTick code, and there is one frame lag there... I had a similar problem in my mod because we were setting the position for a panel in the Paint function of another. The only thing I can think of is to modify the Paint function of the point, if there is even one, or just let it lag and call it a feature. On 9/5/07, Oliver [EMAIL PROTECTED] wrote: -- [ Picked text/plain from multipart/alternative ] CONTEXT: We'd like to create a drag-and-drop interface for our mod. When the user drags off of a panel, the cursor changes and text/image follows the cursor as the user drags (i.e. user clicks and drags off a panel containing text dog; cursor changes to dog cursor and text dog follows cursor around). PROBLEM: The cursor change is no problem. To make the text/image follow the cursor, we create a transparent panel with a label/image, then set that panel's position to the mouse position OnTick. We've tried setting the tick to occur every .1 sec to every .001 sec but the panel still noticeably lags behind the cursor. QUESTION: Is there a way to 'attach' a panel to the mouse position without lag? -- ___ 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
[hlcoders] CBaseClientRenderTargets and rendertarget interface
-- [ Picked text/plain from multipart/alternative ] Just a quick question to see if anyone knows if CBaseClientRenderTargets is no longer how the _rt_WaterReflection, _rt_Camera etc.. rendertargets are being intitalized. We've added our own rendertargets for simple view bluring and I've been trying to clean up our code a bit since people complain of the intialization of our targets causing increases in loading times. Anyway, after creating similar functions to the existing ones in CBaseClientRenderTargets I've discovered that the class never even has the init function called on it. Am I looking in the wrong place for this? Are the other render targets created inside of the engine? Or has somewhere along the lines our initialization of the render targets class been removed, even though the default ones are still available in game. -- ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
Re: [hlcoders] drag and drop
-- [ Picked text/plain from multipart/alternative ] SOLUTION: For anyone still interested, the way to fix this is to update the position in the OnCursorMoved(...) event rather than in OnTick(). That's what the devs do in the built-in drag drop code and it works for custom jobs as well. Remember to catch the OnCursorMoved event for all panels that the mouse could pass through while 'dragging'. If anyone wants more information, let me know. On 9/29/07, Daniel Menard [EMAIL PROTECTED] wrote: -- [ Picked text/plain from multipart/alternative ] I would assume this is because the mouse is only updated after your OnTick code, and there is one frame lag there... I had a similar problem in my mod because we were setting the position for a panel in the Paint function of another. The only thing I can think of is to modify the Paint function of the point, if there is even one, or just let it lag and call it a feature. On 9/5/07, Oliver [EMAIL PROTECTED] wrote: -- [ Picked text/plain from multipart/alternative ] CONTEXT: We'd like to create a drag-and-drop interface for our mod. When the user drags off of a panel, the cursor changes and text/image follows the cursor as the user drags (i.e. user clicks and drags off a panel containing text dog; cursor changes to dog cursor and text dog follows cursor around). PROBLEM: The cursor change is no problem. To make the text/image follow the cursor, we create a transparent panel with a label/image, then set that panel's position to the mouse position OnTick. We've tried setting the tick to occur every .1 sec to every .001 sec but the panel still noticeably lags behind the cursor. QUESTION: Is there a way to 'attach' a panel to the mouse position without lag? -- ___ 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
