| *** spldart <spldart!~spldart@bzflag/contributor/spldart> has joined #bzflag | 00:14 | |
| *** ChanServ sets mode: +v spldart | 00:14 | |
| Zehra | For those in the 1vs1 league, the official servers should use the same designation system of the lu, in which [name].[league].org is used. | 01:28 |
|---|---|---|
| allejo | I think I'm the only one who uses that format... | 01:28 |
| Zehra | Could this be applied to the 1vs1 servers run? And maybe to be suggested as a guideline/requirement? | 01:30 |
| allejo | I don't think the 1v1 league is active enough to bother | 01:33 |
| Zehra | It's mostly so a change can be made to default server search filters and improve the accuracy of them. | 01:34 |
| Zehra | Although I do understand the inactivity not warranting such a change to be made. | 01:35 |
| allejo | http://bzstats.strayer.de/leagues/1vs1/matchservers/?lang=en. literally all of the active servers have "1vs1" in the name | 01:35 |
| Zehra | Well, it looks the change won't be needed. | 01:38 |
| allejo | judging by http://1vs1.bzflag.net/allmatches.php, looks like the active players in the league have no troubles searching for said servers :) | 01:39 |
| allejo | if anything, the league would need a revival and more activity | 01:40 |
| kierra | was speaking with strayer a while back about a tourney on 1v1....been a while since last one | 01:50 |
| *** Sgeo_ <Sgeo_!~Sgeo@ool-18b982ad.dyn.optonline.net> has joined #bzflag | 02:02 | |
| *** Sgeo <Sgeo!~Sgeo@ool-18b982ad.dyn.optonline.net> has quit IRC (Ping timeout: 240 seconds) | 02:05 | |
| *** spldart <spldart!~spldart@bzflag/contributor/spldart> has quit IRC (Remote host closed the connection) | 02:49 | |
| *** _I_Died_Once <_I_Died_Once!~I_Died_On@unaffiliated/idiedonce/x-1828535> has quit IRC (Ping timeout: 240 seconds) | 03:49 | |
| *** Zehra <Zehra!~Zehra@unaffiliated/zehra> has quit IRC (Quit: Gone for now.) | 04:19 | |
| *** infobot <infobot!ibot@c-174-52-60-165.hsd1.ut.comcast.net> has quit IRC (Ping timeout: 260 seconds) | 05:31 | |
| *** dngor <dngor!abuse@p3m/dngor> has quit IRC (Ping timeout: 240 seconds) | 05:52 | |
| *** dngor <dngor!abuse@107-146-022-246.biz.spectrum.com> has joined #bzflag | 05:58 | |
| *** infobot <infobot!ibot@c-174-52-60-165.hsd1.ut.comcast.net> has joined #bzflag | 06:04 | |
| *** ChanServ sets mode: +v infobot | 06:04 | |
| macsforme | tupone: in your VBO manager branch, the VBO is bound prior to drawing by vbo[type].enableArrays()? | 06:44 |
| macsforme | how do you pick which VBO to bind? is there literally only one VBO per data format? | 06:45 |
| tupone | the vbo is bound only after enableArrays | 07:10 |
| tupone | and there is only one vbo per data format. | 07:11 |
| tupone | I'm experimenting an encapsulation of the vbo index, so that the enableArrays, vertexData ... are methods of the vbo index | 07:13 |
| macsforme | when a VBO needs to grow, do you completely reallocate it and upload all the data again? | 07:26 |
| tupone | not yet. I allocate it at the beginning for a certain size. If not enough, abort :( Need to read size from bzdb maybe. I was thinking about reallocating | 08:14 |
| tupone | you can see the size on VBO_Vertex.cxx last lines. Those value are from playing on various world | 08:18 |
| macsforme | and you did all this to avoid VBO switching? | 09:33 |
| tupone | yes | 09:59 |
| *** _I_Died_Once <_I_Died_Once!~I_Died_On@unaffiliated/idiedonce/x-1828535> has joined #bzflag | 11:01 | |
| *** spldart <spldart!~spldart@bzflag/contributor/spldart> has joined #bzflag | 17:14 | |
| *** ChanServ sets mode: +v spldart | 17:14 | |
| *** The_Noah <The_Noah!~The_Noah@47.185.193.111> has quit IRC (Quit: Leaving) | 19:36 | |
| macsforme | tupone: in my view, it's not a robust design to preallocate (or really, over-allocate to be safe) GPU memory based on some constant in the code, then just abort if we go over | 20:47 |
| macsforme | the constant for the buffer size is a maintenance burden (Flash pointed out a few examples of similar problems recently, I think) because it has to be adjusted if we add other things to draw, and it's a waste of GPU memory to over-allocate memory to avoid risking a client abort... GPU memory should be allocated as it is needed in most cases | 20:50 |
| macsforme | this is all in addition to the fact that the whole design is very complex, based on a theoretical optimization that we don't even know if we need yet | 20:52 |
| *** infobot <infobot!ibot@c-174-52-60-165.hsd1.ut.comcast.net> has quit IRC (Ping timeout: 255 seconds) | 20:53 | |
| macsforme | I believe we need to get away from writing code like this, where it is complex and difficult to maintain... this is not a resource-intensive game, so we can afford to write clear, straightforward code (which we should always start with anyway), with reasonably good efficiency, and then optimize further as needed | 20:55 |
| macsforme | to speak directly, I believe we need to reject that patch and start over... I would start by writing a header with an interface that meets the basic requirements, then the implementation can come after we mostly finalize that | 20:59 |
| spldart | o_O *intrigued at the prospect of upcoming further discussion* | 21:02 |
| macsforme | I can help with this too... I am just working on other things first, like moving the bundle/resource handling under the platform module so we can use SDL 2 for it, fixing some window issues on macOS, and whatever else we need to get the next 2.4 release out | 21:02 |
| tupone | ok. If you don't like I will stop doing it. No problem | 21:06 |
| BZNotify | bzflag: atupone closed pull request #232 "Vertex buffers object handling" (https://git.io/Je9kl) | 21:07 |
| spldart | oh boy | 21:07 |
| spldart | :( | 21:07 |
| macsforme | I will add to my to-do list to write an interface that meets the requirements as I see them, then we can compare ideas on that, then we can probably use large portions of your patch in the implementation itself | 21:12 |
| *** spldart is now known as short_circuit | 21:14 | |
| *** short_circuit is now known as Guest50503 | 21:14 | |
| *** ChanServ sets mode: -v Guest50503 | 21:16 | |
| *** Guest50503 is now known as Foo_man_choo | 21:19 | |
| *** Zehra <Zehra!~Zehra@unaffiliated/zehra> has joined #bzflag | 21:47 | |
| *** Chestal <Chestal!~thilo@vmax.unix-ag.uni-siegen.de> has quit IRC (Ping timeout: 268 seconds) | 23:34 | |
| *** Chestal <Chestal!~thilo@vmax.unix-ag.uni-siegen.de> has joined #bzflag | 23:34 | |
| *** ChanServ sets mode: +o Chestal | 23:34 | |
Generated by irclog2html.py 2.17.3.dev0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!