00:07 <+bridge> [ddnet] I was just killed by switch weapon on pickup 00:07 <+bridge> [ddnet] enabled by the update 😄 00:21 <+bridge> [ddnet] ouch 00:58 <+bridge> [ddnet] @deen https://gist.github.com/heinrich5991/cdfc97e03568aa3306ea701f36e219d2 diff two updates 01:34 <+bridge> [ddnet] we should distinguish "default" and "not set" value. As user might explicitly chose default value so on update we should not change his value 01:46 <+bridge> [ddnet] idk how to implement that in a nice way 01:47 <+bridge> [ddnet] we change variables in CConfig all over the place, it's not clear whether it's the default value or the explicitly chosen default value 02:05 <+bridge> [ddnet] can we please change it back from type to server pls it looks so bad 02:05 <+bridge> [ddnet] or make it something like Maps 02:05 <+bridge> [ddnet] Novice Maps 02:05 <+bridge> [ddnet] Brutal Maps 02:07 <+bridge> [ddnet] or "difficulty" 02:08 <+bridge> [ddnet] idk if its good to mix in fng/zcatch/etc with the ddnet tab too u know 02:10 <+bridge> [ddnet] u could have all of that in a "pvp" tab 02:35 <+bridge> [ddnet] @heinrich5991 idk how to implement that in not a nice way without tons of refactoring work 02:35 <+bridge> [ddnet] The only thing I been thinking about is to declare some class X {int Value; bool WasModified;} 02:35 <+bridge> [ddnet] Define implicit cast to/from int for this X type 02:35 <+bridge> [ddnet] Not sure if it is possible and how hard it will impact the perfomance cuse never tried it in c++ 07:51 <+bridge> [ddnet] @holic I renamed it to Maps already 08:24 <+bridge> [ddnet] @Learath2 I guess I should have told you to test without ffmpeg installed locally 😄 09:20 <+bridge> [ddnet] @deen I should have checked with otool... 09:21 <+bridge> [ddnet] Me too 09:21 <+bridge> [ddnet] I totally forgot about that 09:22 <+bridge> [ddnet] @heinrich5991 I guess the best way would be to have a function in CConfig that we always use to set config vars 09:57 <+bridge> [ddnet] I think this was my worst release in a while, broke so many things 09:58 <+bridge> [ddnet] I guess we can't run the DDNet executable on the GitHub CI thingie? 09:58 <+bridge> [ddnet] On which platform? 09:58 <+bridge> [ddnet] All 09:59 <+bridge> [ddnet] My guess is that we don't have a display and SDL2 has no easy way of just telling it to software render to /dev/null 10:00 <+bridge> [ddnet] Oh, I see what you meant, we can run it with a framebuffer probably 10:00 <+bridge> [ddnet] Yeah, that would be nice 10:00 <+bridge> [ddnet] problem is, what do we look for? 😄 10:00 <+bridge> [ddnet] something like `./DDNet quit` would be enough to test for proper linking 10:01 <+bridge> [ddnet] I think just having xvfb on linux would work 10:40 <+bridge> [ddnet] Wh not write a test that tried to connect to a certain game server and the you will see nit only linking issues but it will also check the basic features of the client? 10:40 <+bridge> [ddnet] points query feels more slow :\ 10:40 <+bridge> [ddnet] Why not write a test that tried to connect to a certain game server and the you will see not only linking issues but it will also check the basic features of the client? 10:41 <+bridge> [ddnet] nvm it's ok on empty server 10:41 <+bridge> [ddnet] @Fin sure, that's even more fancy 10:42 <+bridge> [ddnet] Well. But I guess its not too hard? You could even do it via cli if we give a run option or something like they did on ath 10:42 <+bridge> [ddnet] And the just check for the incoming connection on the server? 10:45 <+bridge> [ddnet] Or is that already possible? Like `$ DDNet connect 127.0.0.1:83030`? 10:57 <+bridge> [ddnet] Yes, possible: `DDNet ddnet:127.0.0.1:8303` 10:57 <+bridge> [ddnet] So you can link `ddnet:` links to be opened with DDNet client 10:57 <+bridge> [ddnet] but testing graphical applications this way is rather brittle 10:59 <+ChillerDragon> is it even possible to start the client without a graphical enviroment? To fake some screen or something like that 10:59 <+bridge> [ddnet] We could observe the traffic on both sides and compare it to a known good sample which should make breakages obvious, but we are making this more and more complex while we don't even have the most basic testing 😄 11:00 <+bridge> [ddnet] ChillerDragon: xvfb would work on linux for that, no such thing on mac or windows, if we had our own buildbots maybe we could 11:01 <+ChillerDragon> would be kinda cool to use xvfb in the ci for at least linux then huh? 11:01 <+bridge> [freenode] Sounds better, I like @Learath2 idea with the traffic :o 11:02 <+ChillerDragon> oh i should hav read all messages before writing probably xd 11:18 <+bridge> [ddnet] i mean since we build it anyway we can also test it manually real quick? 11:18 <+bridge> [ddnet] @deen 11:18 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/752820201019539476/Screen_Shot_2020-09-08_at_4.17.58_AM.png 11:18 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/752820209605148692/Screen_Shot_2020-09-08_at_4.18.11_AM.png 11:18 <+bridge> [ddnet] lot of people are having trouble 11:18 <+bridge> [ddnet] seeing entities after the updates? 11:19 <+bridge> [ddnet] rejoining the server sometimes fixes it 11:19 <+bridge> [ddnet] Is FNG right? 11:20 <+bridge> [ddnet] on 0.7? 11:21 <+bridge> [ddnet] its my fng server yes and 11:21 <+bridge> [ddnet] its a 0.6 server that supports 0.7 11:21 <+bridge> [ddnet] they are all using ddnet 11:23 <+bridge> [ddnet] @onby 11:23 <+bridge> [ddnet] https://github.com/ddnet/ddnet/issues/2750 11:23 <+bridge> [ddnet] oh 11:23 <+bridge> [ddnet] so its known issue 11:23 <+bridge> [ddnet] ok 11:24 <+bridge> [ddnet] i'll wait for a response of the network ppl 11:24 <+bridge> [ddnet] else i'll fix it xd 11:24 <+bridge> [ddnet] 👍 11:32 <+bridge> [ddnet] @Jupstar ✪ what is the issue? 11:32 <+bridge> [ddnet] the gameobj gameinfo ext is sent as a snapshot 11:32 <+bridge> [ddnet] but it needs to be sent as a packet after joining the server 11:32 <+bridge> [ddnet] before the map is known 11:32 <+bridge> [ddnet] I don't think that's a feasible change 11:33 <+bridge> [ddnet] entities on fng used to work for people 11:33 <+bridge> [ddnet] Why do you need to know the gametype so early? 11:33 <+bridge> [ddnet] bcs the map is built onmapload 11:33 <+bridge> [ddnet] @onby I'm guessing the new renderer is doing a lot of work in advance 11:33 <+bridge> [ddnet] Yeah 11:33 <+bridge> [ddnet] its not built every frame anymore 11:33 <+bridge> [ddnet] rejoining the server seems to fix it for *most* people 11:33 <+bridge> [ddnet] but not all 11:34 <+bridge> [ddnet] yes bcs then they already got the snapshot 11:34 <+bridge> [ddnet] Let me take a look, but I don't think we can easily move this out of the snap without a lot of compatibility cruft 11:34 <+bridge> [ddnet] ok thx 11:35 <+bridge> [ddnet] @Jupstar ✪ is it maybe possible for you to rebuild it when more info is available? 11:35 <+bridge> [ddnet] thats pretty expensive 11:35 <+bridge> [ddnet] cant they get gametype from server browser 11:35 <+bridge> [ddnet] it was like that before 11:35 <+bridge> [ddnet] i doubt it changes often for most servers 11:35 <+bridge> [ddnet] when i first implemented my code 11:36 <+bridge> [ddnet] @onby relying on silly strings for determining info like this is very brittle 11:36 <+bridge> [ddnet] what does it rely on currently? 11:36 <+bridge> [ddnet] to make only red entities show up 11:36 <+bridge> [ddnet] on uninitialized memory xD 11:37 <+bridge> [ddnet] .-. 11:37 <+bridge> [ddnet] You send a GAMEINFO_EX object, which has the needed information 11:37 <+bridge> [ddnet] wait so 11:37 <+bridge> [ddnet] the red is shown bcs its a ddrace tile 11:37 <+bridge> [ddnet] is this something i can fix serverside 11:37 <+bridge> [ddnet] red spike is ddr tile? 11:37 <+bridge> [ddnet] probably yes 11:37 <+bridge> [ddnet] i mean the index 11:37 <+bridge> [ddnet] Not our fault you relied on information that's not there yet 11:37 <+bridge> [ddnet] not the behaviour 😄 11:37 <+bridge> [ddnet] its not his fault 11:38 <+bridge> [ddnet] me? 11:38 <+bridge> [ddnet] :x 11:38 <+bridge> [ddnet] Definitely not his fault 11:38 <+bridge> [ddnet] his= 11:38 <+bridge> [ddnet] it worked before bcs it used the mod string 11:38 <+bridge> [ddnet] whoever implemented gameinfo just didnt thought about buffering 11:39 <+bridge> [ddnet] so lets pls just be constructive 11:39 <+bridge> [ddnet] Whoever implemented the new renderer didn't check whether the information he used is actually populated at the time of use 11:39 <+bridge> [ddnet] If you want to play blame shifting 11:39 <+bridge> [ddnet] as i said 11:39 <+bridge> [ddnet] It's rather trivial 11:39 <+bridge> [ddnet] when i programmed it it used the mod string 11:40 <+bridge> [ddnet] not any packets 11:40 <+bridge> [ddnet] look the commit in the issue 11:40 <+bridge> [ddnet] ok so to be clear 11:40 <+bridge> [ddnet] i dont have to fix my server right 11:40 <+bridge> [ddnet] no, except u want features from the client, else the client should assume its unknown mod and just display all tiles 11:41 <+bridge> [ddnet] hmm 11:41 <+bridge> [ddnet] can i ask smth, kinda off topic 11:41 <+bridge> [ddnet] which distro has most minimal live iso? 11:41 <+bridge> [ddnet] can i do anything to help the client understand that its fng entities xd 11:41 <+bridge> [ddnet] i wondered if the ddnet codebase of my mod mightve been mesing it up 11:42 <+bridge> [ddnet] i dunno if the client allows snapshots before connected 11:42 <+bridge> [ddnet] i dont think os 11:44 <+bridge> [ddnet] @Jupstar ✪ can't it be updated when the entities info changes? 11:45 <+bridge> [ddnet] I know a mod that actively uses that IIRC 11:45 <+bridge> [ddnet] it can but its expensive and will cause lags 11:45 <+bridge> [ddnet] what are we talking? 1ms? 11:45 <+bridge> [ddnet] depends on hardware 11:45 <+bridge> [ddnet] but building VRAM isnt fast 11:46 <+bridge> [ddnet] hm. I mean it's in the very first snapshot 11:46 <+bridge> [ddnet] perhaps one could delay building the vram until the first snapshot is there? 11:46 <+bridge> [ddnet] (would still break the server that changes the gametype info) 11:47 <+bridge> [ddnet] I'd just rebuild it, if a server changes gameinfo that late we can just tell them that it incurs a cost 11:50 <+bridge> [ddnet] im ok with that 11:51 <+bridge> [ddnet] how can i compare the objects? 11:51 <+bridge> [ddnet] okay, different question 11:51 <+bridge> [ddnet] how does the filtering work? 11:52 <+bridge> [ddnet] or rather, the rendering? 11:52 <+bridge> [ddnet] does it render the whole layer to a texture? 11:52 <+bridge> [ddnet] if m_DontMaskEntities 11:52 <+bridge> [ddnet] is not set, it will look if its a valid ddrace tile 11:52 <+bridge> [ddnet] does it upload the texture indices to the graphics card? 11:52 <+bridge> [ddnet] it uploads vertices 11:52 <+bridge> [ddnet] it doesnt even upload vertices that arent valid 11:52 <+bridge> [ddnet] so like put this part of the entities texture there 11:53 <+bridge> [ddnet] it adds valid tiles to a list 11:53 <+bridge> [ddnet] rest is ignored 11:53 <+bridge> [ddnet] Looking to replace the texture in place? 11:53 <+bridge> [ddnet] yes 11:53 <+bridge> [ddnet] yes, but how does it upload the valid tiles? 11:53 <+bridge> [ddnet] just by position in the entities layer? 11:53 <+bridge> [ddnet] well it uploads it as vertices xD 11:54 <+bridge> [ddnet] I don't know what that means ^^ 11:54 <+bridge> [ddnet] it knows the positon of the entities layer yes 11:54 <+bridge> [ddnet] fill the list with fng tiles by default for quick fix? 🙂 11:54 <+bridge> [ddnet] no 11:54 <+bridge> [ddnet] lol 11:54 <+bridge> [ddnet] @Jupstar ✪ so if we were to upload all tiles and change the entities texture to only include those that we want, it would work fine? 11:54 <+bridge> [ddnet] the problem always was, that you could e.g. spam a front layer into fng 11:55 <+bridge> [ddnet] why is that a problem 11:55 <+bridge> [ddnet] because people did it and it caused entities to be unplayable 11:55 <+bridge> [ddnet] maybe its intentional 11:55 <+bridge> [ddnet] yes 11:55 <+bridge> [ddnet] yes, in fng it works anyway, but in ddrace its annoying 11:55 <+bridge> [ddnet] @heinrich5991 yes that would work 11:56 <+bridge> [ddnet] ah nice, and without delay, too, I guess? 11:56 <+bridge> [ddnet] but it will then create front layers independent of the mod 11:56 <+bridge> [ddnet] yes without any delay 11:57 <+bridge> [ddnet] so if we have a fng front layer 11:57 <+bridge> [ddnet] or generally, a non ddrace front layer 11:57 <+bridge> [ddnet] should be fine 11:57 <+bridge> [ddnet] isnt the "front" layer supposed to be unused for all types of maps except ddrace 11:57 <+bridge> [ddnet] yes but we dont know the mod 😄 11:58 <+bridge> [ddnet] so better built everything 11:58 <+bridge> [ddnet] i c 11:58 <+bridge> [ddnet] the front layer doesn't exist by default I think 11:58 <+bridge> [ddnet] It doesn't indeed 11:58 <+bridge> [ddnet] also the credits of the entities will then work 11:58 <+bridge> [ddnet] why does it show the front layer in fng in that case 11:58 <+bridge> [ddnet] so either remove them or still check some tiles 11:59 <+bridge> [ddnet] well u right @onby 11:59 <+bridge> [ddnet] we can still just not render the front layer 11:59 <+bridge> [ddnet] in any other mod 12:00 <+bridge> [ddnet] only gamelayer, and the mod programmers have to add it to ddnet, if they require other layers 12:01 <+bridge> [ddnet] i suppose this would fix the fake tiles that @ᶰ°Konͧsti makes too 12:01 <+bridge> [ddnet] <ᶰ°Konͧsti> most faketiles are fixed already 12:02 <+bridge> [ddnet] so i guess we need ddnet_front.png then? 12:02 <+bridge> [ddnet] and game seperatelly 12:02 <+bridge> [ddnet] yes 12:02 <+bridge> [ddnet] working on it 12:02 <+bridge> [ddnet] ok thx 😄 12:02 <+bridge> [ddnet] really 12:03 <+bridge> [ddnet] What will you do about the credits? 12:03 <+bridge> [ddnet] still strip them in all gametypes if there's a possible position 12:04 <+bridge> [ddnet] Hm, don't we already clean up the layers from invalid tiles btw? 12:04 <+bridge> [ddnet] that's the problem ^^ 12:04 <+bridge> [ddnet] we do that with incomplete information 12:04 <+bridge> [ddnet] heinrichs solution is really the best 12:04 <+bridge> [ddnet] if the mapper creates invalid tiles and destroys performance in his maps 12:04 <+bridge> [ddnet] its the mappers fault 12:07 <+bridge> [ddnet] So we'll have to strip the credits from the source files as it'd be tricky to strip them in the code 12:07 <+bridge> [ddnet] we can also just create the texture on fly acutally 12:07 <+bridge> [ddnet] we can still stirp them in source 12:08 <+bridge> [ddnet] when loading entities just create front and gamelayer by cheking for valid tiles 12:08 <+bridge> [ddnet] ah 12:08 <+bridge> [ddnet] But won't we need a table of valid tiles for a given mod? 12:08 <+bridge> [ddnet] well, we kinda do this today 12:08 <+bridge> [ddnet] but instead of modifying the texture, we modify the game tiles in the (loaded) map 12:08 <+bridge> [ddnet] Ah, same idea but we use it to build the entities texture instead of cleaning the layers 12:08 <+bridge> [ddnet] Great solution actually 12:09 <+bridge> [ddnet] yes 😄 12:09 <+bridge> [ddnet] okay, may I hand it back to you? @Jupstar ✪ 12:09 <+bridge> [ddnet] ^^ 12:09 <+bridge> [ddnet] yes, i'll fix it 12:09 <+bridge> [ddnet] thanks! 12:09 <+bridge> [ddnet] Not like any of us has any idea how the rendering stack works anymore 😛 12:09 <+bridge> [ddnet] so when map is rendered the gameinfo object is initialized in some way? 12:10 <+bridge> [ddnet] GameInfo object is actually uninitialized when the map is loaded I think 12:10 <+bridge> [ddnet] yes 12:10 <+bridge> [ddnet] It has a default state, but it takes atleast one snap without gameinfoex for it to get there 12:11 <+bridge> [ddnet] it initilaized with null paramters tho 12:11 <+bridge> [ddnet] if not found 12:11 <+bridge> [ddnet] but only at the first snapshot tick 12:11 <+bridge> [ddnet] Maybe initialize it to `GetGameInfo(0, 0, &ServerInfo);` OnConnected? 12:12 <+bridge> [ddnet] well i dunno rn when the client calls render for the first time 12:12 <+bridge> [ddnet] after three snapshots arrived I think 12:12 <+bridge> [ddnet] then its already fine 12:14 <+bridge> [ddnet] I thought you need it to be initialized onmapload 12:14 <+bridge> [ddnet] no we change the texture on fly 12:15 <+bridge> [ddnet] as soon as it knows its fng, it will display fng tiles, but invalid tiles are transparent 12:16 <+bridge> [ddnet] Well when you first render the map on OnMapLoad, it will be uninitialized 12:16 <+bridge> [ddnet] we dont require it here anymore 12:16 <+bridge> [ddnet] But I think it won't display anything until the 3rd snapshot 12:16 <+bridge> [ddnet] @deen you can put it into ddnet_scripts if you haven't already 12:16 <+bridge> [ddnet] it will build all tiles 12:17 <+bridge> [ddnet] @deen just tell me where the new position of the script is so I can put a link there 12:19 <+bridge> [ddnet] We seem to have an issue capturing the mouse 12:20 <+bridge> [ddnet] on macOS mouse clicks won't register until I switch to another app and back 12:21 <+bridge> [ddnet] does ddrace have a front layer too? 12:21 <+bridge> [ddnet] not ddnet 12:21 <+bridge> [ddnet] Yep 12:21 <+bridge> [ddnet] are the tiles different? 12:21 <+bridge> [ddnet] Not sure, though is ddrace still relevant? I don't think anyone hosts it anymore 12:22 <+bridge> [ddnet] ok then i build it up like ddnet 12:29 <+bridge> [ddnet] At least http://ddrace.info/ points to us 😄 12:29 <+bridge> [ddnet] well, I own the domain ^^ 12:29 <+bridge> [ddnet] ah! 12:29 <+bridge> [ddnet] :poggers: 12:30 <+bridge> [ddnet] @Learath2 any idea about getting stable builds for steam_api.dll? 12:30 <+bridge> [ddnet] you could apply diffocsope to two subsequent builds 12:33 <+bridge> [ddnet] @heinrich5991 what's up with that demo gif on diffoscope website? https://diffoscope.org/ I just watched 2 minutes of progress bars until I got bored 😄 12:33 <+bridge> [ddnet] 😄 12:33 <+bridge> [ddnet] idk 12:34 <+bridge> [ddnet] I'll write an email 12:34 <+bridge> [ddnet] @heinrich5991 https://github.com/ddnet/ddnet-scripts/blob/master/update/diff_update.py 13:26 <+bridge> [ddnet] @deen have you put it into diffoscope yet? 13:26 <+bridge> [ddnet] `diffoscope file1 file2` 13:37 <+bridge> [ddnet] @heinrich5991 Some bytes are different, but doesn't tell me much 13:37 <+bridge> [ddnet] can you paste the diff somewhere? 13:38 <+bridge> [ddnet] Seems like all it does is hexdump and then diff 13:38 <+bridge> [ddnet] sometimes it shows more intelligent stuff 13:38 <+bridge> [ddnet] like what headers differ 13:38 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/752855371202166845/message.txt 13:38 <+bridge> [ddnet] whole diff? 13:38 <+bridge> [ddnet] yes 13:39 <+bridge> [ddnet] can you also post the binaries please? 13:40 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/752855826208784394/steam_api.dll 13:40 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/752855876494164090/steam_api.dll 13:41 <+bridge> [ddnet] Oh, are those just timestamps maybe? 13:42 <+bridge> [ddnet] the first one is a checksum 13:42 <+bridge> [ddnet] so fixing the second one probably suffices ^^ 13:44 <+bridge> [ddnet] the second one is in the beginning of the .edata section 13:45 <+bridge> [ddnet] hmm. why does the checksum even differ? does the checksum cover more than the PE header itself? 13:46 <+bridge> [ddnet] > 4 13:46 <+bridge> [ddnet] > 4 13:46 <+bridge> [ddnet] > Time/Date Stamp 13:46 <+bridge> [ddnet] > The time and date that the export data was created. 13:46 <+bridge> [ddnet] I assume it should be over the entire file 13:46 <+bridge> [ddnet] ha, my guess was right! 13:46 <+bridge> [ddnet] yes, the second one is a timestamp 13:46 <+bridge> [ddnet] https://docs.microsoft.com/en-us/windows/win32/debug/pe-format#export-directory-table 13:46 <+bridge> [ddnet] so we just need to tell the linker to use a fixed timestamp and all good I guess 13:46 <+bridge> [ddnet] yes, sth like that 13:46 <+bridge> [ddnet] what compiler is it? mingw? 13:46 <+bridge> [ddnet] https://wiki.debian.org/ReproducibleBuilds/TimestampsInPEBinaries 13:46 <+bridge> [ddnet] yes 13:47 <+bridge> [ddnet] `-Xlinker --no-insert-timestamp` 13:47 <+bridge> [ddnet] from your link 13:47 <+bridge> [ddnet] > This works for executables, but DLLs builts like this cause problems with bound executables: updating a DLL without changing its timestamp causes bound executables to crash. 13:47 <+bridge> [ddnet] hmm 13:48 <+bridge> [ddnet] So we should set the timestamp, but only update it from the old value when something actually changed. That seems hard to do 13:49 <+bridge> [ddnet] let's see what bug it causes and whether it applies to us 13:49 <+bridge> [ddnet] 😄 13:50 <+bridge> [ddnet] @deen seems it's only important for a feature called "binding" 13:50 <+bridge> [ddnet] https://docs.microsoft.com/en-us/archive/msdn-magazine/2002/march/inside-windows-an-in-depth-look-into-the-win32-portable-executable-file-format-part-2 13:50 <+bridge> [ddnet] ctrl-f binding 13:50 <+bridge> [ddnet] we don't do that, so null timestamps should™ be fine 13:51 <+bridge> [ddnet] sounds good 13:51 <+bridge> [ddnet] do you want to make a PR? 13:51 <+bridge> [ddnet] oh, this affects our normal build 13:51 <+bridge> [ddnet] yes 13:54 <+bridge> [ddnet] Mac does the same btw, but we don't have an autoupdater there 13:56 <+bridge> [ddnet] can you give me two builds for that one? 13:59 <+bridge> [ddnet] Apparently in our case there are no timestamps, interesting 13:59 <+bridge> [ddnet] so nothing to fix on mac 13:59 <+bridge> [ddnet] 👍 14:14 <+bridge> [ddnet] @deen fixed 14:29 <+bridge> [ddnet] what the purpose of masking entities ? Its just for graphics? 14:32 <+bridge> [ddnet] in the editor? 14:32 <+bridge> [ddnet] to prevent mappers from adding entities that don't exist yet, but might later be added 14:32 <+bridge> [ddnet] ingame 14:32 <+bridge> [ddnet] probably peformance 14:35 <+bridge> [ddnet] I am missing why we can not ignore the 'mask_entities' flag when loading not DDNet.png 14:36 <+bridge> [ddnet] I am missing why we can not ignore the 'mask_entities' flag when loading not entities\DDNet.png 14:37 <+bridge> [ddnet] or we build like layot for any image for future? Probably I need to read what verticies are to understand it, @Jupstar ✪ 14:37 <+bridge> [ddnet] or we build like layot for any image for future? Probably I need to read what verticies are to understand it, @Jupstar ✪ ? 14:45 <+bridge> [ddnet] why does the editor not remove speedups instead sets their force to zero? 14:46 <+bridge> [ddnet] @BannZay dunno what u mean but the flag is too late 14:46 <+bridge> [ddnet] why we do initialize these indicies for everymap instead of every server type? 14:47 <+bridge> [ddnet] like keep initialized indicies for fng for future 14:47 <+bridge> [ddnet] like keep initialized vertices for fng for future 14:47 <+bridge> [ddnet] @BannZay the purpose of masking entities ingame is to not show entities that don't have an effect 14:47 <+bridge> [ddnet] hmh, to fix DDNet.png? 14:47 <+bridge> [ddnet] im finished except that i need a force restriction for speedups, because else they are saved, 14:47 <+bridge> [ddnet] so no mod compability for speedup layers, until the editor is fixed xD 14:47 <+bridge> [ddnet] why its on the image then? 14:48 <+bridge> [ddnet] who programms the editor here? 14:48 <+bridge> [ddnet] nobody and everyone 🙂 14:49 <+bridge> [ddnet] there were not that many changes to the editor I think 14:49 <+bridge> [ddnet] @Jupstar ✪ but now that you mention it, you can take it over as well. seems like there are no more graphics problems left 😄 14:49 <+bridge> [ddnet] famous last words 14:49 <+bridge> [ddnet] no graphics problems left 14:50 <+bridge> [ddnet] dejavu 14:50 <+bridge> [ddnet] https://cdn.discordapp.com/attachments/293493549758939136/752873500875685938/screenshot_2020-09-08_14-50-02.png 14:52 <+bridge> [ddnet] and tune zone also working forgot it here 14:53 <+bridge> [ddnet] @Jupstar ✪ are vertices will be rebuilded in case when I left one fng server and connect another? Is there any sense to do it? 14:53 <+bridge> [ddnet] it will rebuild the entities 14:53 <+bridge> [ddnet] since only entity rendering is the problem 14:53 <+bridge> [ddnet] not every tile layer 14:54 <+bridge> [ddnet] the entities textures to be precise 14:55 <+bridge> [ddnet] @deen you mean rendering the editor with texture arrays? 14:55 <+bridge> [ddnet] nah, just improving it in general 14:55 <+bridge> [ddnet] xD 14:56 <+bridge> [ddnet] its soo messy, probs easy to reprogram it, but don't really want to, way too much work xD 14:57 <+bridge> [ddnet] anyway, should i restrict speedups with force zero, else they are displayed, but it removes the possibilties for other mods to use speedup layers with speedups with force 0 xD 14:58 <+bridge> [ddnet] no other mod uses them I guess ^^ 14:58 <+bridge> [ddnet] I did a bit with the editor, but instead started spending time on editor2 over in 0.7 14:58 <+bridge> [ddnet] Looks like a much cleaner base to work with 14:59 <+bridge> [ddnet] oh yeah, i saw it too, nice additions 15:01 <+bridge> [ddnet] Native support for undo-redo is nice 15:01 <+bridge> [ddnet] I recently watched a talk by a guy working on photoshop, their approach to undo-redo kinda shocked me 😄 15:02 <+bridge> [ddnet] what do they do? 15:02 <+bridge> [ddnet] Instead of saving the changes, they just say fuck it and save entire snapshots on every save 15:02 <+bridge> [ddnet] every change* 15:02 <+bridge> [ddnet] ez ram 15:02 <+bridge> [ddnet] do they do CoW? 15:02 <+bridge> [ddnet] I thought this was insane, because people edit massive files with photoshop but they have a cute optimization 15:02 <+bridge> [ddnet] Yes CoW 15:03 <+bridge> [ddnet] https://www.youtube.com/watch?v=QGcVXgEVMJg Interesting talk and an interesting approach to polymorphism all in all 15:04 <+bridge> [ddnet] what use case do they have for polymorphism? 15:06 <+bridge> [ddnet] In photoshop you have multiple objects within a document that need to be drawn 15:06 <+bridge> [ddnet] That is the example he was using 15:07 <+bridge> [ddnet] I wonder how well it will age as compilers get better and better at devirtualization 15:08 <+bridge> [ddnet] The main arguments he gives against runtime polymorphism is that calls that go through the vtable aren't cheap 15:10 <+bridge> [ddnet] yea, I thought you'd do ecs or whatever they call it nowadays 15:10 <+bridge> [ddnet] What he creates throughout the talk is quite similar to an ecs 15:11 <+bridge> [ddnet] using existing work might have been nice for the new mapeditor 15:12 <+bridge> [ddnet] I found the same project as last time when I thought about it: https://www.mapeditor.org/ 15:15 <+bridge> [ddnet] i've no plan if i destroyed compability to blockworlds 15:15 <+bridge> [ddnet] or ddrace 15:15 <+bridge> [ddnet] (not ddnet) 15:15 <+bridge> [ddnet] i dunno what layers they have/had 15:16 <+bridge> [ddnet] can we remove "-Winconsistent-missing-override" 15:17 <+bridge> [ddnet] because i like to use override now, since its safer 15:19 <+bridge> [ddnet] it's only safer when you use it consistently, right? 15:20 <+bridge> [ddnet] well i use it for future implemented functions 15:20 <+bridge> [ddnet] so its already safer there 15:20 <+bridge> [ddnet] you cannot have NULL virtual functions for these 15:21 <+bridge> [ddnet] future implemented functions? 15:21 <+bridge> [ddnet] functions i implement now or in future 15:21 <+bridge> [ddnet] like that dont exist yet 15:22 <+bridge> [ddnet] NULL virtual functions? 15:22 <+bridge> [ddnet] from base class 15:22 <+bridge> [ddnet] You mean pure virtual functions or sth else? 15:22 <+bridge> [ddnet] yes 15:26 <+bridge> [ddnet] About editor: couldn't the "editor2" from vanilla be adapted? 15:26 <+bridge> [ddnet] yes, would be nice 15:26 <+bridge> [ddnet] but i fixed the bug anyway 15:26 <+bridge> [ddnet] editor2 has muuuch work to be ready 15:27 <+bridge> [ddnet] It's LordSk's project and he isn't working much on it. I did a handful of PRs that kept him active but I don't think there was any activity after I had to take a break 15:28 <+bridge> [ddnet] If we want to adopt it for ddnet we'll need to adopt the development aswell 15:28 <+bridge> [ddnet] I thought it's finished already :c 15:28 <+bridge> [ddnet] @heinrich5991 how would we do envelopes and stuff if we use a tile editor like the one you linked? 15:30 <+bridge> [ddnet] I thought about extending it, but true, envelopes are a huge features that is missing from it 15:34 <+bridge> [ddnet] Could someone do me a build for Win x64 with https://github.com/ddnet/ddnet/pull/2730 added in? 15:34 <+bridge> [ddnet]