Guild icon
DDraceNetwork
Development / developer
Development discussion. Logged to https://ddnet.tw/irclogs/ Connected with DDNet's IRC channel, Matrix room and GitHub repositories — IRC: #ddnet on Quakenet | Matrix: #ddnet-developer:matrix.org GitHub: https://github.com/ddnet
Between 2021-10-05 00:00:00Z and 2021-10-06 00:00:00Z
Avatar
Avatar
murpi
  • veritasium
gerdoe 2021-10-05 04:10:33Z
justatest
Avatar
louis 2021-10-05 05:28:20Z
score: 2940
Avatar
Avatar
louis
score: 2940
perrodecerro 2021-10-05 05:37:04Z
nouis
Avatar
Avatar
Cipy29
You guys made it so vali can't ddos teeworlds and 1st thing he does is ddos facebook i hope ur happy 😤 😤 😤
Cøke 2021-10-05 07:50:39Z
wait what did u say ? xD
07:57
you seem to be just as funny as björn when you ask such questions haha, he is not even noticed without this ddos shit because no one is interested in him. through the ddos notices at all someone that he exists, hope he continues to do it at big sites and moves for it where he belongs (edited)
Avatar
Cipy29 2021-10-05 08:01:28Z
pepeH
Avatar
Anime.pdf 2021-10-05 08:06:13Z
Whos vali
08:06
Tell him about kog
08:06
troll
Avatar
Cøke 2021-10-05 08:10:26Z
i think he is not interested in such small servers
Avatar
Cipy29 2021-10-05 08:14:20Z
hey guys I have this weird problem -- the moment i press the Editor button on the main menu my desktop computer starts making a weird high pitched noise, the moment I exit the editor it instantly stops (I don't know which part is making the noise lmao I've never heard smth like this before) -- it doesn't happen in other uses of my computer, or in other parts of the ddnet client(nvm as I was typing/testing this I realized a similar noise happens when I open the settings option (quite a bit lower pitched than the editor sound) xd, still never heard it outside of ddnet)
08:14
lmao this sounds so stupid
08:15
also it's not the cpu fan spinning too fast, it's barely spinning
08:16
lmao it only happens in the editor when the UI elements are present (if I press tab to toggle them noise goes away)
Avatar
deen 2021-10-05 08:16:45Z
coil whine
Avatar
heinrich5991 2021-10-05 08:16:49Z
this might be interesting for #developer
08:17
but I don't understand it. might be related to rendering text?
Avatar
deen 2021-10-05 08:17:00Z
can depend on exact power consumption of your gpu
Avatar
Cipy29 2021-10-05 08:17:38Z
how come it's so instant, what happens the moment I press editor that makes it start
Avatar
deen 2021-10-05 08:17:50Z
you can try f1, gfx_refresh_rate and change to another value there, try 0, 480, 120
08:18
Because your GPU instantly switches to a specific power mode I guess
08:18
and rendering all the text might be more/less taxing than the normal menu and ingame
Avatar
Cipy29 2021-10-05 08:21:33Z
it was set to 0, I set it to 480 and it went away
08:23
I also now noticed it happens (to a lesser degree(less noisy), and not at all instant) when in game, a lot harder to notice
08:23
should editor have capped fps by default?
Avatar
Learath2 2021-10-05 08:28:44Z
I have this in the entire game and I get different pitches depending on the state of the ui
08:28
I think it’s bad isolation on the sound output :/
Avatar
Cipy29 2021-10-05 08:30:47Z
sry i don't get it, what does sound output have to do with this (is there sound in editor)?
Avatar
Learath2 2021-10-05 08:31:06Z
Oh, you get noise outside?
08:31
I get it on my speakers and headphone
Avatar
Cipy29 2021-10-05 08:31:44Z
yep, not from the speaker, from the computer itself
Avatar
Learath2 2021-10-05 08:32:43Z
Sounds like coil whine alright. I don’t really know a way to fix it
Avatar
Cipy29 2021-10-05 08:33:29Z
capping gfx_refresh_rate (i capped it to 480) like deen said solved it
Avatar
Learath2 2021-10-05 08:33:45Z
Great \o/
Avatar
deen 2021-10-05 08:35:15Z
Bonus point: You now have lower power consumption
Avatar
Cipy29 2021-10-05 08:36:29Z
poggers time to offset that with mining bitcoin
08:39
feature request: make gfx_refresh_rate 0 placebo, use gpu power that'd be used for 1k fps+ go towards mining ddnet crypto, ez sustainable business model trollet
Avatar
deen 2021-10-05 08:42:15Z
You found the one way of making money that's even worse than advertisements, so no thanks 😄
Avatar
Learath2 2021-10-05 08:42:53Z
https://blog.cloudflare.com/october-2021-facebook-outage/ cloudflares blog posts are so nice to read. They have some people that write so good
Understanding How Facebook Disappeared from the Internet
Today at 1651 UTC, we opened an internal incident entitled "Facebook DNS lookup returning SERVFAIL" because we were worried that something was wrong with our DNS resolver 1.1.1.1. But as we were about to post on our public status page we realized something else more serious was going on.
Thumbnail
Avatar
deen 2021-10-05 08:44:49Z
Because their blog targets potential employees, not potential customers
Avatar
Learath2 2021-10-05 08:46:22Z
Btw apparently the root issue was an invalid configuration value in persistent storage, causing the configuration validator to keep going off when it finds an invalid cached config value. Feedback loop 😛
Avatar
Fän 2021-10-05 08:48:28Z
which caused their bgp sessions to go down 🙂 (edited)
Avatar
Learath2 2021-10-05 08:48:50Z
And took down the database that was hosting the config 😄
08:50
An interesting failure though, I’m sure their blog post was simplified but it sounds like bad engineering practice to do a database query when you find an invalid config value. Shouldn’t you be pushing these changes rather than polling them?
08:54
Oh and the amount of data dns hosters get is insane, I never quite considered how much they can track
08:58
Ah the config thing I read was a previous post-mortem, nvm. So it’s still unknown what actually happened. Someone nuked the routing table? 😛
Avatar
Ryozuki 2021-10-05 09:15:08Z
Fortunately, 1.1.1.1 was built to be Free, Private, Fast (as the independent DNS monitor DNSPerf can attest), and scalable, and we were able to keep servicing our users with minimal impact.
09:15
never miss a oportunity to add self flattery
09:15
bluekitty
Avatar
Learath2 2021-10-05 09:32:03Z
https://youtu.be/2S2v45_NSYs
CNBC
Why It’s Impossible To Put Tracking Microchips In Covid Vaccines
Avatar
Ryozuki 2021-10-05 09:32:35Z
the fact u need a vid is hilarious
Avatar
Learath2 2021-10-05 09:32:35Z
You know we have went wrong somewhere along history if people need to make videos like this
Avatar
Ryozuki 2021-10-05 09:32:43Z
monkalaugh
Avatar
Cipy29 2021-10-05 09:36:18Z
my covid vaccine has coil whine help
09:37
[marked as duplicate]
Avatar
Learath2 2021-10-05 09:48:49Z
I wanted to get one of those biohax chips in my hand actually, but I don't think I can put it there myself without lidocaine
Avatar
Jupstar ✪ BOT 2021-10-05 12:00:27Z
just buy some fancy 8k GPU for 5000 dollars, it will be so underloaded by teeworlds that it won't consume alot of power (@Cipy29)
Author icon
Cipy29
my covid vaccine has coil whine help
2021-10-05 09:36:19Z
12:00
for me the diff between 1k fps and 500fps ingame is under 10 watts
12:01
and my card has a TDP of like 300 Watts
Avatar
Jupstar ✪ BOT 2021-10-05 12:09:35Z
but fps caps for different context sound good, editor & menu vs ingame & demos Maybe we should add it
12:12
the question is always, what is a good default value xD vsync would probably the best for such stuff But we also know the display's refresh rate now
Avatar
Fän 2021-10-05 12:37:37Z
hey @Learath2
12:37
any idea where in the source code it does the serverinfo request?
Avatar
Learath2 2021-10-05 12:39:46Z
CServer::SendServerInfo is what sends it, but UpdateServerInfo -> CacheServerInfo is what generates it. It's only generated when something about the server changes
Avatar
Fän 2021-10-05 12:43:22Z
Trying to cache the packet using a hardcoded request
12:43
but it doesn't seem to respond at all
Avatar
Learath2 2021-10-05 12:50:38Z
Hm, if you show me the request I can take a look at it, but caching at the firewall for too long doesn't sound like a great idea to me
12:51
The new central serverlist is already fairly stale (5s) If it gets cached for longer on the firewall it'll lead to extended periods of stale serverinfo
Avatar
Avatar
Learath2
The new central serverlist is already fairly stale (5s) If it gets cached for longer on the firewall it'll lead to extended periods of stale serverinfo
Fän 2021-10-05 12:52:43Z
30s
12:52
this is just for older clients anyway
12:52
anyway, I just hardcoded
12:52
\x78\x65\x5b\x06\x00\x00\xff\xff\xff\xff\x67\x69\x65\x33\xdb
Avatar
Learath2 2021-10-05 12:53:25Z
the central serverlist also uses the serverinfo request for now, until @heinrich5991 finishes up the server part
Avatar
Fän 2021-10-05 12:53:31Z
seems the first 4 bytes are random
12:53
and the last byte too
Avatar
Avatar
Learath2
the central serverlist also uses the serverinfo request for now, until @heinrich5991 finishes up the server part
Fän 2021-10-05 12:53:42Z
thats whitelisted
12:53
so that won't be cached
12:54
tbh, I could probably make the current protection less "aggressive"
12:55
which means it would allow all traffic, until an attack starts
12:55
players who are already connected will stay connected
12:55
but while the ddos attack is going on, only newer clients can join (edited)
Avatar
Learath2 2021-10-05 12:55:43Z
src/engine/shared/network.h documents the packet header, it shouldn't be random
Avatar
Avatar
Learath2
src/engine/shared/network.h documents the packet header, it shouldn't be random
Fän 2021-10-05 12:55:52Z
its a random "token"
12:56
sha1 apparently
Avatar
Learath2 2021-10-05 12:56:11Z
The last byte is a token in that packet
Avatar
Fän 2021-10-05 12:56:34Z
well
12:56
from 3 samples
12:56
the packet looks like this
Avatar
Learath2 2021-10-05 12:56:58Z
2 bytes of extra token, 1 byte of basic token, those should be the only random things there
Avatar
Fän 2021-10-05 12:57:31Z
"78 65 5b 06 00 00 ff ff ff ff 67 69 65 33 db" "78 65 27 eb 00 00 ff ff ff ff 67 69 65 33 3e"
12:57
so yeah
12:57
buffer[2] and buffer[3] are random
12:57
and last byte
12:57
does it matter though?
12:58
As in, can I just set this to anything?
Avatar
Learath2 2021-10-05 12:58:57Z
random should be fine iirc, but it's been a while since I touched that part of the code
Avatar
Fän 2021-10-05 12:59:07Z
well
12:59
doesn't seem to respond (to me atleast)
Avatar
Avatar
Fän
tbh, I could probably make the current protection less "aggressive"
Fän 2021-10-05 13:00:29Z
thoughts?
Avatar
Avatar
Fän
doesn't seem to respond (to me atleast)
Learath2 2021-10-05 13:04:04Z
Hm, you can try debug 1 on the server to see if any error is triggered, or you can add some dbg_msg along the path. I need to go take care of some personal stuff
Avatar
Avatar
Fän
thoughts?
Learath2 2021-10-05 13:05:33Z
This sounds like a good idea, but how would you extract the list of established clients from the gameserver? Will you just keep the filter monitoring?
Avatar
Fän 2021-10-05 13:05:43Z
Pretty much
13:06
I won't really be able to explain the entire thing, but It'll be able to seperate connected players from other traffic 100%
Avatar
Avatar
Learath2
Hm, you can try debug 1 on the server to see if any error is triggered, or you can add some dbg_msg along the path. I need to go take care of some personal stuff
Fän 2021-10-05 13:08:26Z
Was related to sport
Avatar
heinrich5991 2021-10-05 14:14:46Z
@Fän yes, can be set to anything
Avatar
Fän 2021-10-05 14:17:36Z
figured. tyvm
Avatar
[quakenet] afterfx BOT 2021-10-05 14:34:21Z
test
Avatar
[quakenet] heinrich5991 BOT 2021-10-05 14:35:33Z
bye afterfx
Avatar
Fän 2021-10-05 14:35:46Z
man.
Avatar
heinrich5991 2021-10-05 14:37:00Z
hm?
14:37
afterfx was already gone before I could say anything
Avatar
denkbar 2021-10-05 14:37:54Z
test unsuccessful then
Avatar
heinrich5991 2021-10-05 14:44:34Z
https://skins.tw/how-to/ddnet-integration
Teeskins: DDNet integration
This tutorial will show you how to integrate Teeskins Teeworlds database into DDNet.
14:44
This does not longer work with Steam. DDNet had to remove this feature for the steam version, since it conflicts terms of services. You can still use the downloadable version on DDNet's website.
14:44
we don't support cl_skin_download_url on steam?
Avatar
Ryozuki 2021-10-05 14:45:34Z
we do
Avatar
Learath2 2021-10-05 14:47:09Z
we just don't provide skins that's all
Avatar
Ryozuki 2021-10-05 14:53:38Z
@NeXus ^
Avatar
GitHub BOT 2021-10-05 15:09:40Z
Author icon
Learath2
[ddnet-web:master] 1 new commit
55b7b66 Add HappyFray - Learath2
Avatar
heinrich5991 2021-10-05 15:30:19Z
@Learath2 could you test https://github.com/ddnet/ddnet/pull/4095 on macos?
[Concept] Add new mouse state (ingame mouse) by Jupeyy · Pull Reque...
This would remove ui_mousesense and use the desktop cursor position instead for menus and editor This would also fix the editor cursor to not move faster than in menus. That's probably how ...
Thumbnail
Avatar
Learath2 2021-10-05 15:30:55Z
mh, it's so hard to compile on macOS
15:31
let me give it a quick try
Avatar
Jupstar ✪ BOT 2021-10-05 15:33:49Z
while you at it, can you test the client with HiDPI off?
Avatar
heinrich5991 2021-10-05 15:34:09Z
wörk wörk
Avatar
Jupstar ✪ BOT 2021-10-05 15:34:18Z
to see if SDL fixed the weird bug i send a few weeks ago
15:35
we abuse Learath, bcs no other macos dev that takes the time <.<
Avatar
Fän 2021-10-05 15:35:25Z
wurk wurk (edited)
15:35
I do have macOS
15:35
but I cba
Avatar
heinrich5991 2021-10-05 15:35:50Z
do we even have another developer with macos?
15:35
apparently Fän
Avatar
Learath2 2021-10-05 15:36:16Z
I still have to do sed -i "" 's/11\.0/10.15/g' CMakeCache.txt so damn annoying
Avatar
Fän 2021-10-05 15:36:22Z
only "dev" i do on macOS is sometimes C in vscode
Avatar
Learath2 2021-10-05 15:36:34Z
and I have to delete the bundled sdl so it will stop linking to it
Avatar
Fän 2021-10-05 15:37:10Z
Can whoever is doing this
15:37
stop ddosing the server while I'm still working on it
15:37
<.<
Avatar
Jupstar ✪ BOT 2021-10-05 15:37:18Z
i dont even understand why its so hard to compile under macos :D, we dont really have troublesome libs imo
Avatar
Learath2 2021-10-05 15:37:35Z
It's macOS
Avatar
Fän 2021-10-05 15:37:42Z
everything is hard under macos
Avatar
Jupstar ✪ BOT 2021-10-05 15:37:57Z
ok best argument
15:38
have to admit
Avatar
Learath2 2021-10-05 15:38:02Z
pkg-config is provided by homebrew, which will cause wrong sdk paths to end up in the includes
Avatar
Fän 2021-10-05 15:38:13Z
ok who ever tf is ddosing
15:38
stop it
15:38
get some help
15:38
I'm trying to work here
15:38
<.<
Avatar
Learath2 2021-10-05 15:38:44Z
isn't this like any normal day for you? 😄
Avatar
Fän 2021-10-05 15:38:47Z
It is
15:38
but usually I don't work on it actively (edited)
15:38
while being attacked
Avatar
Learath2 2021-10-05 15:39:52Z
mh, #4095 doesn't work on macOS
Avatar
DDNet BOT 2021-10-05 15:39:53Z
https://github.com/ddnet/ddnet/pull/4095
[Concept] Add new mouse state (ingame mouse) by Jupeyy · Pull Reque...
This would remove ui_mousesense and use the desktop cursor position instead for menus and editor This would also fix the editor cursor to not move faster than in menus. That&#39;s probably how ...
Thumbnail
Avatar
Learath2 2021-10-05 15:39:57Z
not sure how to help you debug it though
Avatar
GitHub BOT 2021-10-05 15:40:05Z
Author icon
heinrich5991
[ddnet/ddnet] Pull request opened: #4197 Properly export antibot DLL symbols on Windows, this time for real
0750ae0fbd79f66aedfbb7ff1092007ae69a6d9c was missing the function prototypes that used ANTIBOTAPI.

Checklist

  • [ ] Tested the change ingame
  • [ ] Provided screenshots if it is a visual change
  • [ ] Tested in combination with possibly related configuration options
  • [ ] Written a unit test if it works standalone, system.c especially
  • [ ] Considered possible null pointers and out of bounds array indexing
  • [ ] Changed no physics that affect existing maps
  • [ ] Tested the ...
Avatar
Learath2 2021-10-05 15:40:52Z
@heinrich5991 https://master2.ddnet.tw/ddnet/15/servers.json failed. libcurl error: SSL certificate problem: certificate has expired um why?
15:42
@Deleted User I think the problem is that the mouse isn't grabbed at all
Avatar
heinrich5991 2021-10-05 15:42:18Z
looks fine from my side
15:42
Expires on: December 23, 2021
Avatar
Learath2 2021-10-05 15:42:59Z
hm looks fine on browser but libcurl doesn't agree
Avatar
Jupstar ✪ BOT 2021-10-05 15:43:20Z
did you use bundled libs and update the git submodule? (@Learath2)
Author icon
Learath2#2281
@Deleted User e4d67636 I think the problem is that the mouse isn&#x27;t grabbed at all
2021-10-05 15:42:13Z
Avatar
Learath2 2021-10-05 15:43:33Z
I can't use bundled libs
15:43
libsdl doesn't include the required patch to even run on macOS without crashing instantly with SIGILL
Avatar
Jupstar ✪ BOT 2021-10-05 15:44:12Z
but ingame should still be grabbed shouldnt it?
15:44
guess thats the most important
Avatar
Learath2 2021-10-05 15:44:48Z
mh, also the cursor is out of sync with the desktop cursor
15:45
so I can only get about halfway down the window before my desktop cursor peeks out the botto
15:45
ingame is still grabbed properly
Avatar
Jupstar ✪ BOT 2021-10-05 15:46:33Z
mhh, guess we need someone with current SDL to test
15:46
that sounds like bug that any SDL dev would instantly catch
Avatar
Learath2 2021-10-05 15:47:07Z
I mean my own version of sdl is 2.0.16
15:47
Well what did we add in our bundled build? I'll roll it up to master and add those too
Avatar
Jupstar ✪ BOT 2021-10-05 15:47:50Z
but this sounds like a bug that would affect anything in existence (@Learath2)
Author icon
Learath2#2281
so I can only get about halfway down the window before my desktop cursor peeks out the botto
2021-10-05 15:45:10Z
15:47
anything that uses SDL
Avatar
heinrich5991 2021-10-05 15:48:44Z
if it's not a new bug, other devs might have worked around it
Avatar
Learath2 2021-10-05 15:48:46Z
I don't know what you do, so I don't know if it would affect everything that uses sdl
15:49
do you just take the position of the desktop cursor?
Avatar
heinrich5991 2021-10-05 15:49:05Z
I mean
libsdl doesn't include the required patch to even run on macOS without crashing instantly with SIGILL
also sounds pretty bad 😛
Avatar
Jupstar ✪ BOT 2021-10-05 15:49:19Z
it just uses the desktop cursor pos relative to the window
Avatar
Avatar
heinrich5991
I mean
libsdl doesn't include the required patch to even run on macOS without crashing instantly with SIGILL
also sounds pretty bad 😛
Learath2 2021-10-05 15:49:41Z
It will work if you compile with an older sdk, it makes appkit disable some asserts
Avatar
Jupstar ✪ BOT 2021-10-05 15:51:20Z
but i also dunno what tsfreddie added with all this m_ScreenHiDPIScale stuff
Avatar
Learath2 2021-10-05 15:51:45Z
disabling highdpi fixes it
Avatar
Jupstar ✪ BOT 2021-10-05 15:51:54Z
ok
Avatar
Learath2 2021-10-05 15:52:05Z
it also makes things look pixellated though 😄
Avatar
Jupstar ✪ BOT 2021-10-05 15:52:11Z
why cant we just fetch the window size and the canvas size in macos
Avatar
Learath2 2021-10-05 15:52:32Z
also it's not the greatest experience since I can sort of escape the "grab"
15:52
it's not really grabbing anything
Avatar
Jupstar ✪ BOT 2021-10-05 15:53:54Z
let me try a patch, and you have to test xd
Avatar
Avatar
Jupstar ✪
why cant we just fetch the window size and the canvas size in macos
Learath2 2021-10-05 15:54:06Z
Well as long as you fetch it at the correct time it used to work just fine, which is why I was very skeptical when tsfreddie was doing the dpi stuff. I just didn't interfere because I didn't have the time to investigate or do it myself
15:55
I have not once in my life seen an airline as janky as brussels airlines
15:56
I submit documents, the website says you'll receive an automated mail acknowledging the submission, I get no automated mail
15:56
How do you fail to implement an automatic mail, it's just a function call
Avatar
Jupstar ✪ BOT 2021-10-05 15:58:28Z
ok pushed it
16:00
but i dunno m_ScreenHiDPIScale is a float
16:00
i dont like to rely on floats
16:00
this really needs an update some day
Avatar
heinrich5991 2021-10-05 16:01:10Z
™
Avatar
Learath2 2021-10-05 16:01:35Z
I don't like the fact that we need to even know about whether the underlying canvas is hidpi or not
Avatar
Jupstar ✪ BOT 2021-10-05 16:02:25Z
yeah
16:02
just split to canvas width/height and windows width/height
Avatar
Learath2 2021-10-05 16:02:31Z
Still doesn't work
Avatar
Jupstar ✪ BOT 2021-10-05 16:02:42Z
i mean for displaying resolutions its ok
16:02
in menus?
Avatar
Learath2 2021-10-05 16:03:15Z
I can only move the cursor on one quarter of the window, so it does line up with the 2x scaling this monitor has
16:03
in all menus
Avatar
Jupstar ✪ BOT 2021-10-05 16:03:34Z
oh yeah, rip, can you test if it works on editor, just so i know the idea atleast works
Avatar
Learath2 2021-10-05 16:03:55Z
editor works
Avatar
Jupstar ✪ BOT 2021-10-05 16:04:13Z
ok, in menus it saves the state in a var, i'll update it
16:04
we should also rename "screenwidth"
16:04
its actually canvas width
16:04
bit confusing
Avatar
Learath2 2021-10-05 16:04:52Z
There is one more problem though
Avatar
heinrich5991 2021-10-05 16:05:15Z
mac CI is broken. /Applications/Xcode_12.5.1.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX11.3.sdk/System/Library/Frameworks/ApplicationServices.framework/Frameworks/QD.framework/Headers/ColorSyncDeprecated.h:1761:3: error: unknown type name 'ptrdiff_t'
Avatar
Learath2 2021-10-05 16:06:05Z
Idk how best to show it but at the edges of the screen both the desktop and the teeworlds cursor are rendered
16:06
s/screen/window/
16:07
@heinrich5991 I can take a look if you do a more verbose build
16:07
I'd like to see CMakeCache.txt and the output of the compiler if you want to give it a go
Avatar
heinrich5991 2021-10-05 16:08:05Z
how do I get a more verbose build? this is the output of the compiler: https://github.com/ddnet/ddnet/runs/3804985058
Avatar
Learath2 2021-10-05 16:09:18Z
I think cmake --build --verbose works, but don't remember
16:09
or make VERBOSE=1 is also usable
Avatar
Fän 2021-10-05 16:10:35Z
GER1 is now joinable on all versions (edited)
16:10
🙂
Avatar
heinrich5991 2021-10-05 16:10:51Z
nice 🙂
Avatar
Jupstar ✪ BOT 2021-10-05 16:10:56Z
ok updated
Avatar
Learath2 2021-10-05 16:10:59Z
Just checked ninja -v is exactly the output I'm looking or
16:11
for*
Avatar
Avatar
Fän
GER1 is now joinable on all versions (edited)
deen 2021-10-05 16:11:05Z
Thanks!
Avatar
heinrich5991 2021-10-05 16:11:37Z
can confirm 0.7 works
Avatar
Jupstar ✪ BOT 2021-10-05 16:12:18Z
@Learath2: i really dont want to annoy you, but could you also check if// SDL_GL_GetDrawableSize reports HiDPI resolution even with SDL_WINDOW_ALLOW_HIGHDPI not set, which is wrong if(SdlFlags & SDL_WINDOW_ALLOW_HIGHDPI) SDL_GL_GetDrawableSize(m_pWindow, pCurrentWidth, pCurrentHeight); else SDL_GetWindowSize(m_pWindow, pCurrentWidth, pCurrentHeight);in backend_sdl.cppis really needed for non HiDPI
16:12
its like the most random if branch we have
16:12
SDL_GL_GetDrawableSize should always give the canvas size
Avatar
Learath2 2021-10-05 16:13:04Z
There is a point in the startup where it will return the wrong size :/
16:13
I'll check with dbg_msg in a second
Avatar
Jupstar ✪ BOT 2021-10-05 16:13:31Z
thats a rather big bug in SDL imo then
16:13
i just cant imagine they didnt fixed it yet
Avatar
Learath2 2021-10-05 16:14:07Z
it's more likely an issue in the way apple does gl
Avatar
Jupstar ✪ BOT 2021-10-05 16:14:29Z
but SDL could just internally check if HiDPI is on
16:14
and if not, give the window size
16:14
having todo it on the user side is very counter intuitive
Avatar
Learath2 2021-10-05 16:14:57Z
Okay, your latest push works fine
16:15
now I'll check th ebranch
16:15
the*
16:16
yeah, this happens right at startup, I think it's the bug I'm thinking of, but I'll print it out and check anyway
Avatar
Jupstar ✪ BOT 2021-10-05 16:16:39Z
time to get SDL dev, so many things that confuse me
Avatar
Learath2 2021-10-05 16:17:20Z
I think it still triggers a window resize when the correct resolution is available btw, atleast on my resizing branch I never had to think about it after I started reacting on all resize events
Avatar
Jupstar ✪ BOT 2021-10-05 16:18:20Z
i think i added support for the correct resize event
16:18
but tsfreddie probably wants it for the HiDPI variable
16:18
maybe we should just update that variable on the resize events
Avatar
Learath2 2021-10-05 16:20:18Z
Um, for me it behaves exactly as expected, @TsFreddie what did you see that made you add this?
16:20
[2021-10-05 18:19:46][dbg]: drawable w=2576 h=1798 [2021-10-05 18:19:46][dbg]: window w=1288 h=899 this is with hidpi on
Avatar
Jupstar ✪ BOT 2021-10-05 16:20:47Z
you have to test it with off
Avatar
Learath2 2021-10-05 16:20:48Z
with hidpi off it returns the same size for both, which is sane
Avatar
Jupstar ✪ BOT 2021-10-05 16:20:57Z
yes
16:21
that sounds correct
16:21
maybe he tried on SDL 2.0.14
Avatar
Learath2 2021-10-05 16:21:30Z
what we really want there is the canvas size, right?
Avatar
Jupstar ✪ BOT 2021-10-05 16:21:32Z
and it got fixe
Avatar
Learath2 2021-10-05 16:21:42Z
I'll try just using that see if anything breaks
Avatar
Jupstar ✪ BOT 2021-10-05 16:21:43Z
yes
Avatar
Learath2 2021-10-05 16:22:11Z
actually, can't possibly change anything since we use the drawable size when hidpi
Avatar
Jupstar ✪ BOT 2021-10-05 16:22:42Z
but isnt that expected?
16:22
just use drawable size for both
16:22
that would be the proper fix
Avatar
Learath2 2021-10-05 16:23:18Z
Yeah, I mean me actually removing the branch is a useless test, since it can't possibly change anything
Avatar
Jupstar ✪ BOT 2021-10-05 16:24:14Z
yeah i want it to be removed, bcs it isnt correct
Avatar
Learath2 2021-10-05 16:24:25Z
If we remove it lets wait for @TsFreddie to be around so he can check it out
Avatar
Jupstar ✪ BOT 2021-10-05 16:24:27Z
on linux the drawable size is smaller, e.g. if you have window decorations
16:24
e.g. close button
Avatar
Learath2 2021-10-05 16:24:52Z
Jupstar: any idea on the desktop and tw cursor both being displayed?
16:25
it means you can interact with the decoration, which is very bad UX. you can resize the window, or move it around
Avatar
Jupstar ✪ BOT 2021-10-05 16:25:33Z
or better: its not clear what the window manager does, when triggering a resize event by ourself
16:25
when does it happen? (@Learath2)
Author icon
Learath2#2281
Jupstar: any idea on the desktop and tw cursor both being displayed?
2021-10-05 16:24:52Z
16:25
always?
Avatar
Learath2 2021-10-05 16:26:17Z
in menus only
Avatar
Jupstar ✪ BOT 2021-10-05 16:26:52Z
in editor it works?
Avatar
Learath2 2021-10-05 16:27:24Z
Ah, no it doesn't work there either
Avatar
Jupstar ✪ BOT 2021-10-05 16:27:59Z
does SDL_ShowCursor(false); fail maybe :o?
Avatar
Learath2 2021-10-05 16:28:06Z
it's very weirdly grabbed, like it's not really grabbed but something is at each tick moving the cursor back in bound
Avatar
Jupstar ✪ BOT 2021-10-05 16:28:56Z
wtf xD
16:29
do they teleport the mouse? XD
Avatar
Learath2 2021-10-05 16:29:37Z
maybe some macOS restriction that they can't properly grab it? idk the apis you use, maybe check the functions see if they document any macOS quirks?
16:29
I know it's hard to get raw input on macOS
Avatar
Jupstar ✪ BOT 2021-10-05 16:29:48Z
well we use SDL_SetWindowGrab which sounds correct
Avatar
Learath2 2021-10-05 16:30:10Z
like if I move my mouse quick enough it'll be outside for a split second
16:30
if I move it constantly around the edges it'll jitter in and out
Avatar
Jupstar ✪ BOT 2021-10-05 16:32:04Z
alternativly dont grab the mouse at all under macos, dunno why this happens
16:32
how does it work in other games?
16:32
ever played anything?
Avatar
Learath2 2021-10-05 16:32:30Z
let me check factorio
16:33
why would anyone add a messenger link to their company website if they don't intend to have anyone actually watching the chat?
16:37
factorio doesn't grab the mouse at all, neither in-game nor in menus
16:37
though their gameplay doesn't require grabbing, ours does
Avatar
Jupstar ✪ BOT 2021-10-05 16:38:16Z
but ingame our grabbing works, doesnt it
16:38
its really a matter of, if non ingame should grab
Avatar
Learath2 2021-10-05 16:38:31Z
yep, so we just don't grab in menus and it's fine I guess
16:38
I think I even prefer not grabbing in menus and editor as I said in the PR
Avatar
Jupstar ✪ BOT 2021-10-05 16:39:01Z
yeah for editors sounds good tbh
16:39
but yeah, i cant judge on it
16:39
in non fullscreen no problem
16:39
but in fullscreen it will minimize your app xD
16:39
and windows + opengl = true love
16:40
so maybe on not enable it on macos
16:40
since fullscreen doesnt work anyway?
Avatar
Learath2 2021-10-05 16:43:03Z
wait what? on fullscreen if you don't grab it will minimize?
Avatar
Jupstar ✪ BOT 2021-10-05 16:46:14Z
no, clicking outside the window
16:46
will minimize your app
Avatar
Learath2 2021-10-05 16:46:32Z
ah on macOS it won't
16:47
ah I see what you mean, I think even on macOS it would, you mean clicking desktop on a second screen, right?
Avatar
Jupstar ✪ BOT 2021-10-05 16:48:53Z
yes
Avatar
Learath2 2021-10-05 16:49:14Z
which is why borderless fullscreen is superior
Avatar
Jupstar ✪ BOT 2021-10-05 16:49:32Z
in many things yes
16:49
but it usually doesnt disable compositing
16:49
so less fps, and fps is everything
16:50
and ofc you couldn't play 5:4, which some ppl do in this community
Avatar
Learath2 2021-10-05 16:50:33Z
I don't have a compositor even 😛
Avatar
Jupstar ✪ BOT 2021-10-05 16:51:05Z
macos might be fine, windows sucks, KDE too xd
16:51
with wayland it worked perfect
16:51
but wayland has forced vsync
16:51
and rn my system crashes with wayland so cant test anyway xD
Avatar
heinrich5991 2021-10-05 16:52:03Z
I thought you need to play 4:3 stretched to be a pro
Avatar
Jupstar ✪ BOT 2021-10-05 16:52:30Z
i bet most of these ppl dont even know that they dont have 144hz using such resolutions anymore
16:52
one of the reasons i added support for it
16:54
did we update macos for the CI?
16:54
or random fail
Avatar
heinrich5991 2021-10-05 16:56:00Z
not random
16:56
Learath2 wanted to help me debug it, but I havne't gotten to it yet
16:56
feel free to try to debug it
Avatar
Jupstar ✪ BOT 2021-10-05 16:56:36Z
looks like a SDK fail
16:57
i just dont understand why it happens rn, what triggered it xd
Avatar
Learath2 2021-10-05 16:57:49Z
looks like an sdk mixup to me, as in two different versions of the sdk getting both included, but I can't know for sure
Avatar
Jupstar ✪ BOT 2021-10-05 16:58:05Z
ok on my local branch it works, i just need to build artifacts
Avatar
Learath2 2021-10-05 16:58:06Z
not without seeing CMakeCache and/or the verbose build
Avatar
GitHub BOT 2021-10-05 17:01:17Z
Author icon
heinrich5991
[ddnet:pr_ddnet_verbose] 1 new commit
5de27dc Verbose - heinrich5991
17:01
Author icon
heinrich5991
[ddnet/ddnet] New branch created: pr_ddnet_verbose
Avatar
heinrich5991 2021-10-05 17:02:02Z
https://github.com/ddnet/ddnet/runs/3805775384 I added output of CMakeCache.txt and verbose build (hopefully)
17:02
I'll eat some dinner now though
Avatar
Jupstar ✪ BOT 2021-10-05 17:07:25Z
@Learath2: for me it works with the cursor not being shown, only when i try to leave the window
17:07
then i also get this jittering
Avatar
Learath2 2021-10-05 17:07:35Z
on macOS?
Avatar
Jupstar ✪ BOT 2021-10-05 17:07:46Z
yes
17:08
mojave
17:08
whatever version that is
17:09
the jittering isnt nice, but i think its ok, the 5 users using macos have to live with it xDD
17:10
nice resizing works fine on macos too
17:11
i really dont have any major bugs, except that i cant use fullscreen at all
17:11
not even desktop fullscreen
17:11
ah wait
17:12
on this build it even works
17:12
only not on the nightly built from ddnet.tw
17:12
is the nightly still using the old libs @deen ?
17:12
just want to be sure its not our bundled libs
17:15
@Learath2: does "real" fullscreen not work for you?
Avatar
Learath2 2021-10-05 17:15:20Z
nope
Avatar
Jupstar ✪ BOT 2021-10-05 17:15:34Z
ok, then i guess its a driver bug
Avatar
Learath2 2021-10-05 17:15:38Z
It doesn't render anything at all
Avatar
Jupstar ✪ BOT 2021-10-05 17:15:39Z
bcs it works with the software renderer
Avatar
Learath2 2021-10-05 17:15:56Z
it does grab all input including the touchpad though, so it's really hard to quit
17:16
you need to blindly open the console and type quit
Avatar
Jupstar ✪ BOT 2021-10-05 17:16:47Z
ok really weird
17:16
wait do we even support real fullscreen
17:17
ah only if you restart?
17:17
17:17
/ Todo SDL: remove this when fixed (game freezes when losing focus in fullscreen) SDL_SetWindowFullscreen(m_pWindow, SDL_WINDOW_FULLSCREEN_DESKTOP);
17:17
oh xd
Avatar
Avatar
Jupstar ✪
is the nightly still using the old libs @deen ?
deen 2021-10-05 17:18:21Z
No, nightly is using new libs, it takes them from ddnet/ddnet-libs master state
Avatar
Jupstar ✪ BOT 2021-10-05 17:19:38Z
i hope SDL doesnt detect that you use old SDKs or smth and gives a broken SDL
17:19
bcs the SDL from brew works without problems
Avatar
Learath2 2021-10-05 17:20:23Z
I don't know if my patch made it into 2.0.16 btw, if it did brew version should now work though
Avatar
Jupstar ✪ BOT 2021-10-05 17:20:23Z
mh, but fullscreen works
17:20
even on nightly now
17:20
i hope it doesnt prefer the brew SDL and i am just confused
Avatar
Learath2 2021-10-05 17:21:02Z
do otool -L DDNet and check what it linked to
17:21
on macOS it's like a dice roll what sdl gets linked to
Avatar
Jupstar ✪ BOT 2021-10-05 17:21:27Z
wait maybe it crashed bcs gfx_resizable was 0 before
17:22
AHH
17:22
indeed
17:22
gfx_resizable 0 is evil
17:22
guess we should default it to 1 now, that it works anywhere anyway
Avatar
deen 2021-10-05 17:22:44Z
or just remove it and always be 1?
Avatar
Jupstar ✪ BOT 2021-10-05 17:23:11Z
with gfx_resizable 1 fullscreen works on macos for me
17:23
also changing it on fly
17:23
yeah
17:23
even better
Avatar
GitHub BOT 2021-10-05 17:23:44Z
Author icon
def-
[ddnet/ddnet] Pull request opened: #4198 Always support resizing, remove gfx_resizable

Checklist

  • [ ] Tested the change ingame
  • [ ] Provided screenshots if it is a visual change
  • [ ] Tested in combination with possibly related configuration options
  • [ ] Written a unit test if it works standalone, system.c especially
  • [ ] Considered possible null pointers and out of bounds array indexing
  • [ ] Changed no physics that affect existing maps
  • [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--u...
Avatar
heinrich5991 2021-10-05 18:36:16Z
@Learath2 did you find anything in the log output?
Avatar
Learath2 2021-10-05 18:36:40Z
I also had dinner
18:36
let me take a look now
18:38
aha, I see it
18:40
PC_CURL_INCLUDEDIR:INTERNAL=/Library/Developer/CommandLineTools/SDKs/MacOSX11.sdk/usr/include
18:42
I think the new system protection stuff isn't letting us delete commandlinetools anymore
18:42
Can you try checking if it does get deleted after the rm -rf?
Avatar
gerdoe 2021-10-05 18:43:46Z
+
18:43
ooooh sorry
18:44
my cat jus ttyped +
Avatar
Learath2 2021-10-05 18:44:14Z
this is another reason it's a massive pain to compile on macOS btw, I have no idea what the correct way to do this is
Avatar
Jupstar ✪ BOT 2021-10-05 18:47:22Z
is there no stub like mingw? so its easy to cross compile and not always broken xd
Avatar
heinrich5991 2021-10-05 18:48:20Z
there's also cross compilation for mac, but we'd like to be able to be compiled on mac, too
Avatar
GitHub BOT 2021-10-05 18:48:36Z
Author icon
heinrich5991
[ddnet:pr_ddnet_verbose] 1 new commit
fbb7141 a - heinrich5991
Avatar
Jupstar ✪ BOT 2021-10-05 18:49:29Z
yeah but you can also use it on macos i guess? Xd
Avatar
Learath2 2021-10-05 18:50:51Z
every time I encounter this issue I remember the temper tantrum the developer of homebrew threw because google didn't hire him
Avatar
heinrich5991 2021-10-05 18:51:17Z
can you elaborate?
Avatar
Learath2 2021-10-05 18:51:35Z
He failed a simple coding question, iirc it was reversing a linked list
18:52
it's very telling by the mess he created
Avatar
Jupstar ✪ BOT 2021-10-05 18:52:52Z
In file included from /Users/runner/work/ddnet/ddnet/src/macos/notifications.mm:3: In file included from /Applications/Xcode_12.5.1.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX11.3.sdk/System/Library/Frameworks/Foundation.framework/Headers/Foundation.h:168: In file included from /Applications/Xcode_12.5.1.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX11.3.sdk/System/Library/Frameworks/Foundation.framework/Headers/NSAppleEventDescriptor.h:9: In file included from /Applications/Xcode_12.5.1.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX11.3.sdk/System/Library/Frameworks/ApplicationServices.framework/Headers/ApplicationServices.h:43: In file included from /Applications/Xcode_12.5.1.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX11.3.sdk/System/Library/Frameworks/ApplicationServices.framework/Frameworks/HIServices.framework/Headers/HIServices.h:26: In file included from /Applications/Xcode_12.5.1.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX11.3.sdk/System/Library/Frameworks/ApplicationServices.framework/Frameworks/HIServices.framework/Headers/HIShape.h:22: In file included from /Applications/Xcode_12.5.1.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX11.3.sdk/System/Library/Frameworks/ApplicationServices.framework/Frameworks/QD.framework/Headers/Quickdraw.h:31: /Applications/Xcode_12.5.1.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX11.3.sdk/System/Library/Frameworks/ApplicationServices.framework/Frameworks/QD.framework/Headers/ColorSyncDeprecated.h:1761:3: error: unknown type name 'ptrdiff_t' ptrdiff_t rowStride;I wonder if the apple devs have nothing like a CI that catches such simple bugs
18:53
i mean a missing typedef... doesnt sound too hard to catch, some header swizzler, that makes sure every header has what it needs
Avatar
Avatar
GitHub
Click to see attachment 🖼️
Learath2 2021-10-05 18:55:18Z
Aha, newer commandlinetools actually makes this a symlink
18:55
so, we should try not deleting this at all now
18:55
Since it should correctly symlink to the proper place now
18:55
though I don't see how it could cause this issue :/
18:56
mh, it would probably symlink to the commandlinetools version of the sdk though, I don't really know what to do there
Avatar
Jupstar ✪ BOT 2021-10-05 20:24:20Z
what even made the ddnet repo use the new macos sdk version? I just tested and my ddnet github repo uses an older version is it controlled by github -> random?
Avatar
Learath2 2021-10-05 20:29:16Z
I'd think random
Avatar
Learath2 2021-10-05 21:04:18Z
I can't decide whether to take my laptop on vacation with me or not
Avatar
Jupstar ✪ BOT 2021-10-05 21:09:22Z
flip a coin
Avatar
Learath2 2021-10-05 22:03:33Z
If I take the laptop I have to take my ugly bag :/
Avatar
SPYRES 2021-10-05 22:20:10Z
well depends where u going and how long
Avatar
Learath2 2021-10-05 22:21:26Z
sweden, a week
Avatar
SPYRES 2021-10-05 22:22:29Z
who u going with
Avatar
murpi 2021-10-05 22:22:48Z
flipping a coin and realizing how much you dislike the outcome is more useful for decision making than the actual result of the flip
Avatar
SPYRES 2021-10-05 22:23:34Z
if u going with ur gf, then no. if u going with mates, then no, if u going to see family, then yes. if u going by urself then go to the club instead
Avatar
Learath2 2021-10-05 22:26:17Z
what is this gf thing you are talking about? 😛
Avatar
SPYRES 2021-10-05 22:26:21Z
unless you do this which i approve @Learath2 https://www.youtube.com/watch?v=4lNaATnghJE
YaitsTac
Riding A Dirt Bike While In Zoom Class With A Green Screen Attached...
Avatar
Learath2 2021-10-05 22:26:26Z
Going to visit a friend
Avatar
Avatar
SPYRES
unless you do this which i approve @Learath2 https://www.youtube.com/watch?v=4lNaATnghJE
Learath2 2021-10-05 22:27:24Z
lmao
Avatar
Avatar
Learath2
lmao
SPYRES 2021-10-05 22:27:42Z
u get work and pleasure in one! easy solution
Avatar
Avatar
murpi
flipping a coin and realizing how much you dislike the outcome is more useful for decision making than the actual result of the flip
Learath2 2021-10-05 22:28:06Z
let me try
22:29
you know what, I'll take it, just incase ddnet decides to blow up while I'm away
Avatar
SPYRES 2021-10-05 22:31:07Z
Sounds like we need a DR scenario when @Learath2 not here
22:31
If you give me the password, I'll keep it safe 😉
Avatar
GitHub BOT 2021-10-05 23:51:58Z
Author icon
heinrich5991
[ddnet/ddnet-web] Pull request opened: #143 Ten million finishes
Avatar
heinrich5991 2021-10-05 23:52:10Z
@murpi still awake and can review? ^
23:57
ah @louis, maybe you're from a better time zone and can proofread this? ^
Exported 433 message(s)