| 01:44.07 | *** join/#riftuidev r (~r@c-98-228-123-33.hsd1.in.comcast.net) |
| 01:49.11 | *** join/#riftuidev Cairenn (~Owner@MMOI/Administratrix/Cairenn) |
| 01:49.11 | *** mode/#riftuidev [+o Cairenn] by ChanServ |
| 01:55.34 | Noshei | so SL tomorrow |
| 01:56.05 | Noshei | was today the party day at Trion ZorbaTHut? |
| 02:00.00 | Imhothar | Party day before deployment? That sounds very dangerous |
| 02:00.27 | Noshei | bah, they probably finialied the ship build last week |
| 02:01.07 | Noshei | so party today, before it's off to fix stuff tomorrow |
| 02:06.15 | Imhothar | I'm sure stuff is already being fixed ;) |
| 04:56.26 | *** join/#riftuidev ZorbaTHut (tocohl@maximilian.pavlovian.net) |
| 04:56.26 | *** mode/#riftuidev [+o ZorbaTHut] by ChanServ |
| 04:59.41 | *** join/#riftuidev Noshei_ (~quassel@c-67-161-146-244.hsd1.co.comcast.net) |
| 05:07.54 | Noshei | hey ZorbaTHut, would it be possible to get a buff that shows which mount you are using and the speed the mount is using |
| 05:41.32 | *** join/#riftuidev Imhothar (~miranda@188.174.131.81) |
| 07:24.31 | *** join/#riftuidev Noshei_ (~quassel@c-67-161-146-244.hsd1.co.comcast.net) |
| 07:33.24 | ZorbaTHut | noshei, likely that would be part of a separate API |
| 08:14.46 | ZorbaTHut | my cat keeps putting all his toys in his food dish. cat brain wiring is fascinating |
| 09:36.43 | Baanano | our younger one started dipping them in the water bowl some weeks ago, now our home is filled with wet mice toys |
| 09:37.07 | Baanano | my gf thinks she does that so the older cat can't play with them |
| 10:15.48 | *** join/#riftuidev Cairenn|afk (~Owner@MMOI/Administratrix/Cairenn) |
| 10:15.48 | *** join/#riftuidev Natch (~Natch@c-eccde155.25-4-64736c10.cust.bredbandsbolaget.se) |
| 10:15.48 | *** join/#riftuidev seebs (~seebs@home.seebs.net) |
| 10:15.48 | *** join/#riftuidev ChanServ (ChanServ@services.) |
| 10:15.48 | *** mode/#riftuidev [+ooo Cairenn|afk seebs ChanServ] by cameron.freenode.net |
| 10:16.07 | *** join/#riftuidev Baanano_ (~chatzilla@85.53.171.79) |
| 10:22.11 | *** join/#riftuidev Baanano (~chatzilla@85.53.171.79) |
| 10:22.11 | *** join/#riftuidev bctrainers (~bc@unaffiliated/bctrainers) |
| 10:22.11 | *** join/#riftuidev starlon (~starlon@melody.lyrical.net) |
| 10:26.55 | *** join/#riftuidev Baanano (~chatzilla@85.53.171.79) |
| 10:27.04 | *** join/#riftuidev bctrainers (~bc@unaffiliated/bctrainers) |
| 10:28.44 | *** join/#riftuidev Kaelten (~Kaelten@WoWUIDev/WoWAce/WoWIFA/CurseStaff/kaelten) |
| 10:28.44 | *** mode/#riftuidev [+o Kaelten] by cameron.freenode.net |
| 12:31.11 | *** join/#riftuidev Wildtide (0fe3b94a@gateway/web/freenode/ip.15.227.185.74) |
| 12:31.23 | Wildtide | afternoon all |
| 13:10.47 | starlon | Oh there's people fighting in the city today. |
| 13:10.58 | starlon | hopes he isn't attacked. >.< |
| 13:11.29 | starlon | I've left to get something to drink and come back to my death. |
| 13:11.46 | starlon | In Rift. Sitting at a quest hub or something. |
| 13:11.52 | starlon | Where I thought it was safe. :P |
| 13:36.46 | starlon | So Rift is going down too? :( |
| 13:37.29 | starlon | oh they may be talking about beta |
| 13:40.09 | Wildtide | Rift will be down in about 2.5 hours for the Storm Legion update |
| 13:40.44 | starlon | ah |
| 13:41.02 | Wildtide | have you been away for a while? the city invasions have been none-stop for weeks now. they are getting kinda tiresome |
| 13:41.23 | starlon | nah I don't log in very often |
| 13:43.06 | Wildtide | oh, well you came back at the right time. it's the big release today :) |
| 13:43.22 | starlon | ah damn |
| 13:43.24 | starlon | cool |
| 13:43.46 | Wildtide | we finally get to wear ZorbaTHut's capes :) |
| 13:43.52 | starlon | lol |
| 14:30.36 | *** join/#riftuidev Torhal (~torhal@71.207.224.241) |
| 16:55.54 | *** join/#riftuidev purl (~purl@rikers.org) |
| 16:55.54 | *** topic/#riftuidev is http://pastebin.com for code snippets || PIL: http://www.lua.org/pil || Rift AddOn Policy: http://t.co/7xXd4LL || AddOn Wiki: wiki.riftui.com/ || RiftUI.com || rift.curse.com || this channel is logged, via purl |
| 17:21.12 | Noshei_ | hey Wildtide, so I noticed that there is no mouseover tooltip for the XP bar |
| 17:22.27 | Wildtide | no, there isn't |
| 17:22.54 | Wildtide | i haven't created any fake tooltips anywhere. i probably should, but i've avoided it so far |
| 17:24.14 | Wildtide | sorry, i need to go (was just in the process of logging off). i'll catch you later |
| 17:32.10 | *** join/#riftuidev Torhal (~torhal@c-71-207-224-241.hsd1.al.comcast.net) |
| 18:20.54 | *** join/#riftuidev Imhothar (~miranda@ppp-188-174-131-81.dynamic.mnet-online.de) |
| 18:25.22 | *** join/#riftuidev elninja (~r@98.228.123.33) |
| 18:42.06 | Thorarin | gifv servers! |
| 19:06.24 | *** join/#riftuidev ZorbaTHut (tocohl@maximilian.pavlovian.net) |
| 19:06.25 | *** mode/#riftuidev [+o ZorbaTHut] by ChanServ |
| 19:44.28 | seebs | wb Zorba. |
| 19:45.02 | seebs | I ran patchers on all my machines. Then I re-ran them for the SL patch. Then I re-ran them to pick up the new SL version of RiftPatch, and it's re-patching them again. |
| 19:45.11 | seebs | So I have another hour or so before everything's updated. :P |
| 19:55.07 | Imhothar | sometimes I wish my collegue would just stop committing code |
| 20:03.33 | ZorbaTHut | the good news is that the streaming patcher means you can get into the game a whole lot faster ^^ |
| 20:03.51 | Imhothar | faster on overcrowded servers? |
| 20:04.02 | ZorbaTHut | faster than you would be able to *without* the streaming patcher :V |
| 20:04.16 | seebs | So, Zorba, I have an AWESOME feature idea. |
| 20:04.41 | seebs | Which I wish to mention because I think it would make your life as Temporary Rendering Guy a lot easier if it existed. And all you need is a few months extra development time. |
| 20:05.07 | ZorbaTHut | that's all? :V |
| 20:05.16 | Imhothar | if it's just a few months |
| 20:05.30 | seebs | It's a feature CoH had, called "demo record". You do /demorecord, and the game creates a log of EXACTLY what is happening in the game, down to the server tick counts. And then: You can play back these demos. As often as you want. On any hardware. With any graphics settings. |
| 20:05.53 | seebs | The thing about this being: If you have a /demorecord, you have a *STABLE* performance test that doesn't depend on how many people are around or what's happening. |
| 20:06.03 | seebs | But which you can use to study performance issues that happen during busy fights. |
| 20:06.36 | ZorbaTHut | actually, I've been considering the same thing, but it's difficult to make it work in a way that doesn't introduce security vulnerabilities - it's a great tool for hackers as well |
| 20:06.38 | seebs | Basically, it lets you separate "rendering engine" and "game engine" issues for testing. Also! It lets players compare notes meaningfully. |
| 20:07.25 | seebs | Huh. How so? |
| 20:08.23 | ZorbaTHut | really, for the same reasons that it's helpful for development - any reproducible state is vital for debugging. And hacking is essentially a kind of debugging. |
| 20:09.50 | seebs | Interesting! |
| 20:12.18 | seebs | I guess I wouldn't have thought that the rendering engine would be all that useful for understanding the rest of what was happening. |
| 20:12.25 | seebs | CoH, I think, did not much care about security. |
| 20:13.05 | ZorbaTHut | ah, there's another problem, then - if we're attempting to test performance, we can't extricate the rendering engine and the gameplay engine like that. I'm sure you're aware of how CPU-hungry Rift is. Most of that isn't rendering. |
| 20:13.17 | seebs | Yeah. |
| 20:13.39 | seebs | My theory is that the game's rendering engine isn't what's painfully slow for the most part. Although... I certainly get MUCH worse results at very high resolutions. |
| 20:13.49 | seebs | But when I get 2-3fps during huge battles, the GPU is barely running. |
| 20:14.10 | ZorbaTHut | it's a huge pile of "it's complicated". And yeah, I do need to figure out some way to debug huge battles |
| 20:14.33 | ZorbaTHut | the solution may be to implement demo-record internally, then go PvP a bit on a hacked-up debug/release client :V |
| 20:14.37 | seebs | Well, a separate rendering engine test would still be useful for SOMETHING, which is diagnosing system build problems. |
| 20:15.06 | ZorbaTHut | That said, right now I think there are probably enough easy pickings that I don't need reproducible cases |
| 20:15.14 | seebs | In CoH, /demorecord is what led to one of the players correcting the devs as to how their game mechanics actually work. |
| 20:15.19 | ZorbaTHut | hah, nice |
| 20:15.26 | seebs | And thus, "arcanatime", the actual measure of how long powers take to activate. |
| 20:15.37 | ZorbaTHut | what was the problem? |
| 20:15.56 | seebs | Not really a problem as such, they just didn't understand their own "activation time" and animation system. |
| 20:16.36 | ZorbaTHut | ah, nod |
| 20:16.39 | seebs | They never did quite figure that out. |
| 20:16.46 | seebs | To the end, they balanced powers around recharge, but not activation. |
| 20:16.47 | ZorbaTHut | reminds me of Age of Conan, which actually had a slower weapon swing animation on women on release |
| 20:17.05 | seebs | So at high levels, their entire model of how to balance powers was wrong, because you could have recharge reduced enough that you were waiting on animations. |
| 20:18.06 | seebs | And yeah, for overall performance: The biggest thing it looks to me like Rift needs is ANY way at all to get more work out of the Huge Thread. |
| 20:18.26 | seebs | Because in practice, Rift seems to be totally unable to use more than about 1.4 cores worth of CPU. |
| 20:18.32 | ZorbaTHut | yeah, the problem is that multithreading is a nightmare if it wasn't in to begin with |
| 20:18.36 | seebs | Yeah. |
| 20:18.44 | seebs | Same bin as portability, in my experience. |
| 20:18.46 | ZorbaTHut | that said, if you have a lot of open cores, ramp the cloak detail all the way up - that system will happily span as many cores as you have ^^ |
| 20:18.55 | seebs | Awesome. :P |
| 20:19.06 | seebs | Heh, today is the day that Windows Update has 12+ important updates for everything. |
| 20:19.36 | seebs | You guys have this really weird mix of great engineering and WHAT WERE YOU THINKING?!?!. |
| 20:19.58 | Imhothar | seebs don't forget the task sceduler may think it's a good idea to pack multiple threads on a single core, especially on cpus with e.g. turbo boost |
| 20:20.00 | seebs | I am unable to understand how a game that started development in the 2000s was this poorly prepared for multicore. |
| 20:20.05 | ZorbaTHut | you're more right on that front than you'd expect ;) Rift has a curious history. |
| 20:20.26 | seebs | I mean, there's no WAY that they didn't know that multicore was going to be a factor. |
| 20:20.59 | seebs | On the other hand, all the evidence I can see suggests that your management are more than usually willing to agree to refactoring stuff when the devs think it's needed. |
| 20:22.19 | ZorbaTHut | Yep. Stuff gets refactored all the time. Actually, thanks to your GoofBall crash, I'm currently ripping out and redesigning our entire image caching infrastructure, which turned out to be phenomenally broken |
| 20:22.23 | seebs | At this point, from a technical standpoint, better multicore is by FAR the largest issue affecting the game, I think. In terms of gameplay experience in general, I'd go with community and troll handling. |
| 20:22.28 | seebs | AWESOME! |
| 20:22.29 | ZorbaTHut | one of those "I don't understand how this has ever worked" deals |
| 20:22.33 | Imhothar | and coin lock again |
| 20:22.44 | Imhothar | what the hell causes that |
| 20:23.05 | seebs | Usually, first time from a new IP address or local machine, except first or first in a while or sometimes just because. |
| 20:23.21 | Imhothar | I think it's more like "just because" on my account |
| 20:23.32 | seebs | I am pleased to have been able to provide a good bug case. :) |
| 20:23.55 | seebs | And I'm glad it gives me an excuse to disable goofball, because I cannot play Rift competently with that up. |
| 20:24.39 | ZorbaTHut | I still don't understand why that crash wasn't happening . . . oh . . . *constantly, for everyone* . . . but at least we'll get it fixed before it changes its mind on that point |
| 20:24.57 | seebs | So between 10:30 this morning (central) and now, 12+ important updates for Windows. |
| 20:25.20 | ZorbaTHut | Impressive! my media box will have fun updating next time I turn it on |
| 20:25.23 | seebs | So, I dunno if you were around when this came up. |
| 20:25.33 | seebs | I think this was while you were out. |
| 20:26.06 | seebs | There was some poking around at the rendering engine performance of Rift, and someone who had GPU-debugging tools reported that it looked like Rift was WAY too aggressive at uncaching textures that had recently been in use even when there was plenty of spare video memory. |
| 20:27.00 | ZorbaTHut | wouldn't surprise me, but note that "video memory" and "main memory" are both used by textures, and rift occasionally has main memory problems |
| 20:27.03 | ZorbaTHut | any idea what the tool was? |
| 20:27.08 | ZorbaTHut | or what the textures were? :V |
| 20:27.25 | seebs | I do not recall. I *think* it was someone in this chat, maybe. |
| 20:27.39 | seebs | So how likely is it that you guys will be able to do a native 64-bit build? :) |
| 20:28.03 | ZorbaTHut | I wouldn't put money on it :) though note that we do build it with large-address-aware, so on a 64-bit system it will happily use 4gb of RAM |
| 20:28.05 | seebs | I ask partially because I know that there have been cases in the past where performance was improved dramatically by the nicer instruction set. And also because memory is cheap. |
| 20:28.35 | seebs | That helps. In WoW, midway through WotLK, I would consistently crash out under some circumstances because it didn't have that bit set. |
| 20:29.21 | Imhothar | well a naive 64build usualyl takes more memroy to begin with. It's when you start using techniques that benefit from the huge address space si when you get real improvements |
| 20:29.56 | seebs | Yeah. Although apparently, the instruction set and register set are a LOT nicer and can improve things. |
| 20:32.40 | Imhothar | I'm not 100% sure about the instruction set but if your compiler is using ILP64 you can get problems with stack overflow if you're unlucky |
| 20:32.44 | seebs | I sorta wish that, way back when, the devs had planned originally for cross-platform. Network effects are huge in MMOs, and I know a number of people who are stuck in WoW because they have a friend who is only willing to play things that run on a Mac. :( |
| 20:33.19 | Imhothar | I never understood why people would limit themselves to windows/directx |
| 20:33.22 | ZorbaTHut | unfortunately, cross-platform is also pretty dang expensive. there can be major issues with libraries that don't play nicely on one platform or another, or bizarre bugs that occur only on one platform |
| 20:33.29 | seebs | Yeah. |
| 20:33.54 | seebs | A friend of mine is doing dev in PureBasic (for long and complicated, but sound, reasons). And discovered that in their implementation, OpenGL gives roughly a 10x speedup. |
| 20:33.56 | seebs | This is atypical. |
| 20:34.11 | seebs | WTF. Okay, this is odd. |
| 20:34.24 | ZorbaTHut | I suspect that's more "bug in directx renderer" than anything else |
| 20:34.29 | Imhothar | DirectX suffers in speed from it's instruction implementation |
| 20:34.31 | seebs | Since the updates, ALL my Windows machines now, consistently, pop up a little message saying that "additional log on information may be required". |
| 20:34.37 | seebs | It gets better! |
| 20:34.44 | seebs | On ATI and Intel hardware, it's just slower. |
| 20:35.01 | seebs | On nVidia, it starts out slower, then after a few seconds gets about 30% slower, and then a bit later gets 30% slower again. |
| 20:35.04 | seebs | Then it seems stable. |
| 20:35.10 | ZorbaTHut | well that's pretty dang strange |
| 20:35.27 | Imhothar | I'd say that's the driver screwing up |
| 20:36.36 | seebs | The network thing isn't just me, lots of people hitting it today. |
| 20:39.57 | seebs | And it appears to be OpenDNS related, and my ISP's suggested default nameserver is OpenDNS. |
| 20:49.24 | *** join/#riftuidev seebs (~seebs@74.122.98.108) |
| 20:49.24 | *** mode/#riftuidev [+o seebs] by ChanServ |
| 20:49.37 | seebs | And also apparently, if you mess with QoS settings on ddwrt, it can cause your IRC and AIM connections to die silently. |
| 21:07.19 | ZorbaTHut | hey, luajit 2.0 is officially out |
| 21:08.45 | seebs | Oooh. |
| 21:09.57 | seebs | I keep hoping the Lua SDK I use for iOS/Android will get that someday. Although! They did find the performance problem I was hitting, apparently I have created a test case under which if I touch any object, every object is sent to OpenGL. Separately. |
| 21:26.56 | ZorbaTHut | if it's an open-source sdk, it probably wouldn't be hard to splice it in yourself. luajit is API-compatible. |
| 21:36.03 | seebs | Not open-source, sadly. I'm using Corona. Long story behind that choice, but overall I'm pretty happy with them. |
| 21:36.24 | seebs | There are downsides and upsides, but it works well enough for my purposes for now. |
| 21:41.51 | Baanano | ZorbaTHut, will we have luajit2? |
| 21:42.17 | ZorbaTHut | baanano, you already almost do - it's one of the latest pre-release betas |
| 21:42.23 | ZorbaTHut | but I'll probably get luajit2 itself in for 2.1 |
| 21:42.47 | Baanano | the __pairs metamethod sounds useful to do filtering and sorting |
| 21:43.04 | ZorbaTHut | that's already available, in fact, I use it in some of the API code :) |
| 21:43.15 | Imhothar | that's a Lua, not luajit feautre |
| 21:44.15 | Baanano | oh nice |
| 21:44.40 | Baanano | time to refactor things |
| 21:45.00 | ZorbaTHut | I'm not sure it's 100% guaranteed to continue existing, but as long as luajit continues supporting it - which I suspect it will - it'll stay in |
| 21:46.16 | Imhothar | __metatable is also a nice undocumented thing |
| 21:47.24 | Imhothar | having that field set on a metatable prevents the object from having its metatable changed once assigned |
| 21:47.40 | Baanano | that's the one that prevents changing the metatable, isn't it? |
| 21:48.30 | Imhothar | aye |
| 21:59.31 | Noshei | so ZorbaTHut no way to see your PA XP via the API right now, untill you hit 60 |
| 23:08.04 | seebs | Huh, interesting. Lag problems that were resolved by temporarily disabling tithing and guild banks. |
| 23:27.22 | seebs | ooh, IRC client update. Back in a bit. |
| 23:27.48 | *** join/#riftuidev seebs (~seebs@home.seebs.net) |
| 23:30.30 | *** join/#riftuidev seebs (~seebs@home.seebs.net) |
| 23:30.31 | *** mode/#riftuidev [+o seebs] by ChanServ |