05:06 < bridge> gm 05:06 < bridge> i wish fish had bash jobs 05:06 < bridge> parallelism was one of the main reasons behind the rust port but they haven't yet added that feature ig 06:04 < bridge> the fish is sadly unemployed 06:12 < bridge> Suprisely to find my mods. 06:13 < bridge> Surprisely to find my mods. 06:13 < bridge> Surprised to find my mods. 06:13 < bridge> :nouis: 06:15 < bridge> Wow 06:15 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1429321950018404493/image.png?ex=68f5b74c&is=68f465cc&hm=9353e4679e7d46951808354b5ebb09e32dee989411c0ec614ed8525c9b923c66& 06:20 < bridge> Quite interesting that 06:21 < bridge> 8 of 10 my most played maps are InfClass map. 06:21 < bridge> :nouis: 06:23 < bridge> Oh wait 06:23 < bridge> That's my current name 06:23 < bridge> I need to check my old name either. 09:38 < bridge> @jupeyy_keks results for only show gametiles (on my windows PC ofc) 09:38 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1429373085491920970/DDNet-benchmark-worst_0.001_FPS.png?ex=68f5e6ec&is=68f4956c&hm=120df4d685fb7bfa819560a357a48fc309e2c086a0928d6c600b8172f254ab3c& 09:38 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1429373086016475189/DDNet-benchmark-AVG_FPS.png?ex=68f5e6ec&is=68f4956c&hm=c3f7efacf4459d251f2185b7945821819e470682c1e3d3ab19a8a2be45045242& 09:38 < bridge> @jupeyy_keks results for only show gametiles (on my windows PC ofc) 09:38 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1429373085491920970/DDNet-benchmark-worst_0.001_FPS.png?ex=68f5e6ec&is=68f4956c&hm=120df4d685fb7bfa819560a357a48fc309e2c086a0928d6c600b8172f254ab3c& 09:38 < bridge> @jupeyy_keks results for only show gametiles (on my windows PC ofc) 09:38 < bridge> @jupeyy_keks results for only show gametiles (on my windows PC ofc) 09:38 < bridge> (Nice just blank data, sec) 09:40 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1429373520197976064/DDNet-benchmark-worst_0.001_FPS.png?ex=68f5e753&is=68f495d3&hm=244dcc08ed5603896f5f5382bd6f3c2882f05112c775ee8639e78209ff1ad5e9& 09:40 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1429373520651091998/DDNet-benchmark-AVG_FPS.png?ex=68f5e753&is=68f495d3&hm=4f5e9466e8ec01d1ac5314bec844c8ec0e481dfbb7884bdbc361b58585b6d162& 09:48 < bridge> hmm here the 18.x version row would be interesting 09:53 < bridge> What happened 18.9 09:57 < bridge> there may be other things beeing measured here, I noticed that the player doesn't have a freeze ninja animation on some older clients for example 09:58 < bridge> also the hud is different, but there was also a map rendering regression in between 10:04 < bridge> do we want to optimize "only show gametiles" ? because we could easily do that already on map load, but this would prevent beeing able to toggle it 10:10 < bridge> I am pretty sure this is the now fixed skin issue visible here, because you see it between 19.4 and 19.5, and not (much) map rendering related was changed between this verions 10:20 < bridge> oh 10:20 < bridge> all my fps gone 10:41 < bridge> That is unbelievable amounts of variability between the versions. Did you control for everything? Did you measure a couple times? Did you average? 10:42 < bridge> it says "worst 0.001" so you'd expect variability 10:42 < bridge> Ah, that's what I missed on the graphs 10:43 < bridge> Honestly the avg also has an unexpected amount of variability. We have been steadily ruining the performance it seems 10:45 < bridge> At 18.9 we started to ruin it even faster. Then it seems we returned to our previous trajectory for 19.5 10:45 < bridge> yeah I'm also curious what happening in 18.9 10:45 < bridge> when was the new HUD update? 10:45 < bridge> I don't remember exactly 10:48 < bridge> damn it was way back in 16.3 10:48 < bridge> damn it was way back in 16.0 10:49 < bridge> not even on the cahrt 10:49 < bridge> not even on the chart 10:51 < bridge> hmm there's like 9 versions between 17.4.2 and 18.9.1 so I'm probably not gonna find it skimming the change logs 10:54 < bridge> i mean we had a few versions with the broken/slow skin loading stuff 10:54 < bridge> but i'd have expected 19.5 to be fast again 11:11 < ws-client1> **** @12944qwerty okay snap parsing kinda works. But there is no dead reckoning yet so the prediction is always behind. Not sure if I will work on that any time soon ._. 11:11 < ws-client1> **** this is what it can do now https://tube.zillyhuhn.com/video.php?t=pygame_tw_07.mp4&u=chiller 11:12 < bridge> <12944qwerty> dead reckoning is calculating the values in between ticks? 11:12 < ws-client1> **** yea 11:12 < ws-client1> **** it needs physics implementation and map data 11:12 < ws-client1> **** to fully predict tee movement that is not sent over the network 11:12 < bridge> <12944qwerty> ah 11:12 < bridge> <12944qwerty> that video does not help and hurts my brain trying to understand which circle is which tee 11:13 < ws-client1> **** haha 11:13 < ws-client1> **** the default tee is the python client and its always in the middle 11:17 < bridge> <12944qwerty> oh dead reckoning is why it's glitching a bit 11:25 < ws-client1> **** yea and it gets worse without aim changes 11:26 < ws-client1> **** time to add python bindings to teeros lib :D hope its ready 11:28 < ws-client1> **** @teero777 i can use your lib for dead reckoning right? is it ready? how i do it 11:31 < bridge> I don't know why it gradually went down, but the lowest is probably 19.2.1 because of dynamic skin loading, which was being updated way too often. 11:38 < bridge> <12944qwerty> xd teero's repos just say don't read or use this code 11:42 < ws-client1> **** well yea :D 11:51 < bridge> its a running gag from jupsti's dd-pg repo 11:56 < ws-client1> **** oh really? i thought its actually not ready yet 11:56 < bridge> tja, insider 12:02 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1429409465001836647/image0.png?ex=68f608cd&is=68f4b74d&hm=686db6c0565b0c411083d817fa45a7713b0b8fd9392d009cf8dc332c9b0222dd& 12:07 < bridge> https://cdn.discordapp.com/attachments/293493549758939136/1429410606112440422/content.png?ex=68f609dd&is=68f4b85d&hm=724629988d152955046b1f8df2435fc012648204308d9bd0a8c08e024762cc5f& 12:15 < ws-client1> **** omg teero help 12:15 < ws-client1> **** https://zillyhuhn.com/cs/.2c54dcae-0349-4982-b670-da43e34bf03f.png 12:16 < ws-client1> **** why the madafak not link it in 12:16 < ws-client1> **** ima ask chartgpt 12:17 < ws-client1> **** omg rare chartgpt W 12:17 < ws-client1> **** `sudo ldconfig` fixed it, how did i never encounter that command in my life? 12:18 < bridge> bcs u rarely self compile and then install apparently 12:18 < ws-client1> **** no i do 12:18 < ws-client1> **** maybe i can blame teero 12:18 < ws-client1> **** for me that usually always worked 12:18 < ws-client1> **** i assume this is cmake responsibility 12:19 < bridge> when i install custom mesa drivers i always had to do it 12:19 < ws-client1> **** why does it not run when i did `sudo make install` 12:19 < ws-client1> **** i did not had to do that with ddnet_protocol and sdl i think 12:19 < bridge> maybe u already had it installed b4 12:19 < bridge> anyway, it works so what 12:19 < ws-client1> **** oh its only once per name? 12:19 < ws-client1> **** yea idk im just surprised 12:57 < bridge> just FYI 18.9 is the only 18.x version I benchmarked, means it could be any 18.x version, there are missing a few in this graph 12:57 < bridge> exactly, map rendering did not change between 19.4 and 19.5, looks to me more like skinloading was the issue 12:59 < bridge> the average is calculated for 30 seconds, on 7 maps, I don't think benchmarking the same map repeatedly will improve the variability by much 13:02 < bridge> I should add a map average to the graph 14:46 < ws-client1> **** @Teero i published u as pyson package axaxaxaxax 14:47 < ws-client1> **** https://github.com/ChillerDragon/ddnet_maploader_py 15:19 < bridge> WAHT XD 15:19 < bridge> nice 15:19 < bridge> I was thinking of making it a single header library 15:19 < bridge> Aren't there better libs for phyton for that? 15:20 < bridge> phytorn 15:21 < bridge> @chillerdragon but why, we have twmap-py ? 16:25 < bridge> idk? More lightweight? my lib only loads the layers necessary for the physics. imo it's really simple. 16:26 < bridge> waiting for your physics as a wrapped python package :p 16:30 < bridge> chiller you are not following me on gh I feel betrayed :tear: :tear: :tear: 16:32 < bridge> I need to step up my GitHub influencer game 16:32 < bridge> 100 followers by the end of the month 16:32 < bridge> how are patches handled btw? I guess you are just merging the patch into the client and not doing a new release on the current master? 16:32 < bridge> step one link your gh to you dc 16:32 < bridge> what? Are you talking about my physics? 16:32 < bridge> already done 16:32 < bridge> no about ddnet patches 17:02 < bridge> Yes twmap is really nice! I just wanted to learn how to build python bindings to be prepared for the physics one. Also competition is always good. Also simple C is cuter than rust :p 17:02 < bridge> @teero777: so are the physics ready to be used? 18:19 < bridge> No xd 18:20 < bridge> Well yes kind of but they can change at any moment 18:20 < bridge> Also I need to write docs 21:31 < bridge> C-chan 21:31 < bridge> .__.