| *** Cobra_Fast is now away: offline | 00:03 | |
| *** Cobra_Fast is back | 00:03 | |
| BZNotify | bzflag: Zehra opened issue #337 "[Bug] flagid in bz_ePlayerDieEvent may be incorrect" (https://github.com/BZFlag-Dev/bzflag/issues/337) | 02:07 |
|---|---|---|
| Zehra | In some moments, the proper method of bug reproduction will be demonstrated with associated plug-in. | 02:39 |
| Agatha | Zehra: you can pick up a flag from a distance away from it, so this isn't workable even in theory. Why not just keep track of the position of each flag? | 02:56 |
| *** I_Died_Once <I_Died_Once!~I_Died_On@c-73-184-170-223.hsd1.ga.comcast.net> has quit IRC (Ping timeout: 258 seconds) | 02:58 | |
| Zehra | Agatha: Thanks for the insights. This seems like a way more viable approach and makes more sense. | 02:59 |
| Zehra | on second thought, now I realize one thing. I mistook the allowedFlagGrab event with the flagGrabEvent. | 03:00 |
| Zehra | Thus I may be able to pull the positional data of the flag before it is grabbed. | 03:01 |
| BZNotify | bzflag: Zehra commented on issue #337 "[Bug] flagid in bz_ePlayerDieEvent may be incorrect" (https://github.com/BZFlag-Dev/bzflag/issues/337#issuecomment-1763657368): The following has been tested and reproduced with `BZFlag 2.4.26`... | 03:09 |
| Zehra | Updated with steps for reproduction and with plug-in for additional ease of bug reproduction/documentation. | 03:10 |
| *** Flash <Flash!~Flash@user/flash> has joined #bzflag | 04:00 | |
| *** Flash_ <Flash_!~Flash@user/flash> has quit IRC (Ping timeout: 258 seconds) | 04:02 | |
| *** FastLizard4 is back | 04:10 | |
| *** FastLizard4 is now away: AWAY from keyboard | 04:12 | |
| *** Zehra <Zehra!~Yukari@user/yukari> has quit IRC (Quit: Quit.) | 04:34 | |
| *** FastLizard4 is back | 05:27 | |
| *** FastLizard4 is now away: AWAY from keyboard | 05:27 | |
| *** FastLizard4 is now away: IN-GAME - Starfield | 05:28 | |
| *** Cobra_Fast is now away: offline | 06:26 | |
| *** Cobra_Fast is back | 06:26 | |
| *** Sgeo <Sgeo!~Sgeo@user/sgeo> has quit IRC (Read error: Connection reset by peer) | 08:02 | |
| *** FastLizard4 is back | 09:06 | |
| *** FastLizard4 is now away: AWAY from keyboard | 09:30 | |
| *** I_Died_Once <I_Died_Once!~I_Died_On@c-73-184-170-223.hsd1.ga.comcast.net> has joined #bzflag | 10:12 | |
| *** FastLizard4 is back | 10:41 | |
| *** FastLizard4 is now away: AWAY from keyboard | 11:22 | |
| *** FastLizard4 is now away: GONE - Screen Detached and Disconnected from IRC (I'm probably asleep, at work, or doing something in real life) | 11:48 | |
| *** FastLizard4 is back | 17:16 | |
| *** FastLizard4 is now away: AWAY from keyboard | 17:25 | |
| *** Flash_ <Flash_!~Flash@user/flash> has joined #bzflag | 18:04 | |
| *** Flash <Flash!~Flash@user/flash> has quit IRC (Ping timeout: 258 seconds) | 18:07 | |
| Juest | any idea why 2.5/2.6 has increased jitter? | 20:23 |
| Juest | on windows | 20:23 |
| SpringTank | bugs | 20:49 |
| SpringTank | Someone was coding in light mode and it attracted bugs. | 20:49 |
| blast007 | Juest: nope, not yet | 21:06 |
| blast007 | does it happen if you host a server locally too? | 21:07 |
| Juest | blast007: dont remember but i think yes probably | 21:10 |
| blast007 | how many FPS are you getting when you have the jitter issue? | 21:14 |
| Juest | didnt check | 21:14 |
| blast007 | I am seeing a higher jitter on master than 2.4, but it's still not *high* | 21:20 |
| blast007 | 2ms on 2.4, 6 to 7 on master | 21:20 |
| blast007 | (to a remote server) | 21:20 |
| Juest | i had a jitter high as 70ms | 21:20 |
| blast007 | that's what I saw previously, but now I don't know how to recreate it :) | 21:21 |
| blast007 | I did just have 4ms even connecting to localhost | 21:22 |
| blast007 | oh wow, in 2.4 on localhost, I'm getting 200+ ms of jitter, but only 1ms of lag :P | 21:24 |
| blast007 | fixed the jitter on 2.4 (by commenting out the QueryPerformanceCounter recalibration code, which is part of why bzfs hosting on Windows is so bad), and now it shows 0 (as it should be, all things considered) | 21:29 |
| blast007 | Juest: have you never run a git bisect before? | 21:30 |
| Juest | blast007: not really something i do but i think i could give it a try | 21:30 |
| blast007 | it might be more annoying if you don't have Visual C++ 2017 though, since 2019 is going to probably upgrade the project files | 21:31 |
| blast007 | hmm, actually a bisect might not be super useful against a localhost server, since at least some of the revisions will have the buggy QPC recalibration... | 21:32 |
| blast007 | I'll see if I can work through a bisect here | 21:34 |
| Juest | so this jitter bug affects 2.4? | 21:34 |
| Juest | blast007: i do have vs 2017, 2019 and 2022 installed | 21:35 |
| Juest | but not old versions | 21:35 |
| blast007 | different bug | 21:35 |
| Juest | oh ok | 21:35 |
| Juest | i had the high jitter on master when joining your server | 21:36 |
| Juest | i can check the jitter on linux to double check | 21:36 |
| blast007 | on 2.4 Windows, the QPC recalibration, at least on the server, makes time run at like 2/3rds speed, so as you might imagine, that screws some stuff up :) | 21:36 |
| blast007 | I think I removed the recalibration code on master | 21:37 |
| blast007 | the basic idea of git bisect is that you find a known good version and a known bad version, and then it will skip to revisions between those and you tell git if that step was good or bad, and it helps you track down the revision that introduced an issue | 21:38 |
| Juest | im seeing 54-66ms of jitter in a local server on master at 180 fps | 21:39 |
| Juest | on a i7 4790k windows 7 | 21:39 |
| Juest | checked in release build | 21:39 |
| blast007 | let me try on my Windows 7 system | 21:42 |
| blast007 | is this on a Wifi connection, or wired? | 21:42 |
| Juest | wired | 21:42 |
| Juest | and localhost | 21:42 |
| Juest | i have more jitter on 2.4 than on master | 21:42 |
| blast007 | oh, right.. localhost.. so.. kinda irrelevant :) | 21:42 |
| blast007 | 2.4 on localhost? | 21:43 |
| Juest | yes on localhost | 21:43 |
| blast007 | yeah, the 2.4 server time is skewed, so calculations like that will be very broken | 21:43 |
| Juest | 260-270-330 | 21:43 |
| Juest | ah bummer | 21:43 |
| Juest | which bzexcess server i can use as control for testing on 2.4? | 21:43 |
| blast007 | any of them. they're all on the same server. | 21:44 |
| Juest | the jitter is good on your server | 21:47 |
| Juest | 138 +- 2ms | 21:48 |
| Juest | a spike of 17-24 ms at one point | 21:48 |
| Juest | another master issue is the libpng warnings about gAMA | 21:49 |
| Juest | the jitter is like 60ms on master | 21:49 |
| Juest | i've grabbed a unknown flag lol | 21:51 |
| Juest | what it does? | 21:51 |
| blast007 | unknown flag in master? | 21:51 |
| Juest | yes | 21:52 |
| blast007 | huh, that's.. weird :P | 21:52 |
| Juest | on your 1.7 good flags map | 21:53 |
| blast007 | when you were holding that flag, I also couldn't right click to identify you | 21:54 |
| Juest | oh huh | 21:54 |
| Juest | any idea when those bugs we found broke? | 22:08 |
| blast007 | nope | 22:08 |
| blast007 | but that's what git bisect can help with :) | 22:08 |
| Juest | :D | 22:08 |
| blast007 | we found 4 new bugs: | 22:09 |
| blast007 | * Can't respawn after a SteamRoller death | 22:09 |
| blast007 | * Can't identify tanks | 22:09 |
| blast007 | * Lock on warning does not show up | 22:09 |
| blast007 | * "Unknown" flags spawning (shows up as "Unknown" when you grab them and has an abbreviation of --) | 22:10 |
| Juest | wait, the gm trajectory works fine? | 22:10 |
| blast007 | yeah, it was still curving towards you on my screen | 22:10 |
| blast007 | I was aiming it past you so it didn't hit | 22:10 |
| Juest | ah i think i saw that | 22:10 |
| Juest | i see, barely noticed | 22:10 |
| Juest | does git bisect take up storage? | 22:10 |
| blast007 | no | 22:11 |
| Juest | ah good | 22:11 |
| Juest | could do it | 22:11 |
| blast007 | it just jumps back and forth through revisions and updates the working directory to that revision | 22:11 |
| Juest | i could use windbg with the pdb files in place | 22:11 |
| Juest | i need the pdbs because im not doing debugging on the same computer as the one that built the game | 22:12 |
| blast007 | for this you wouldn't need the debugger even | 22:12 |
| blast007 | it's more about tracking down which revision introduced a specific issue | 22:12 |
| Juest | figuring out the piece of code that causes the bug is what i need the debugger for | 22:12 |
| Juest | unless you know the piece of code responsible for this | 22:12 |
| blast007 | oh, sure, if you want to go at it from that angle | 22:13 |
| blast007 | I was going to bisect it | 22:13 |
| Juest | sure go ahead | 22:13 |
| blast007 | solo bots might be enough to test 3 of the issues | 22:13 |
| Juest | oh, and the bug with the rogue skin not applying? what code do you think is it? | 22:13 |
| blast007 | I did have a fix for the rogue skin, but to make it easier I also removed that server-side plugin event ;) | 22:14 |
| blast007 | so... I haven't pushed that quite yet | 22:14 |
| Juest | Ah | 22:14 |
| Juest | understood | 22:14 |
| Juest | your server is not exactly vanilla? | 22:14 |
| blast007 | no, it is | 22:15 |
| Juest | well it has plugins | 22:15 |
| Juest | nvm i did not read that properly | 22:15 |
| blast007 | the plugins I have shouldn't really affect anything | 22:16 |
| blast007 | nothing gameplay related at least | 22:16 |
| blast007 | okay, I lied.. I guess I do have the 'genocide' plugin (which I wonder if that was the Unknown flag) and thiefControl (which does affect the behavior of thief) | 22:17 |
| Juest | :D | 22:18 |
| Juest | does the unknown flag go away if you unload the genocide plugin? | 22:18 |
| blast007 | or are they ALL unknown if I just do +f G{30} with no other flags :D | 22:19 |
| Juest | could i ask you to have a server to use as testing ground like this? | 22:19 |
| Juest | or localhost is there for that matter? | 22:20 |
| Juest | i was using the start server functionality | 22:20 |
| blast007 | xs.bzexcess.com:5432 is running with +f G{30} +f ST, so it should be 30 geno and 1 stealth | 22:21 |
| Juest | i noticed, there's no way to change ports and it waits until a existing bzfs instance shuts down before starting a instance | 22:21 |
| Juest | on master? | 22:21 |
| blast007 | yeah | 22:21 |
| Juest | why not use 255432? | 22:21 |
| Juest | oh did you mean 25432? | 22:22 |
| blast007 | beause that's higher than the maximum port of 65535 | 22:22 |
| Juest | the master servers are running on 2nnn compared to 2.4 running on the nnnn | 22:22 |
| blast007 | let me check if it actually shows up on the list. I just hand-write a config quick. | 22:22 |
| Juest | ah | 22:22 |
| Juest | its not showing up | 22:23 |
| Juest | 5432 you sure? not 25432? | 22:23 |
| blast007 | it's just a number. | 22:24 |
| Juest | anyways im getting error connecting to server on that port | 22:24 |
| blast007 | k, let me check | 22:24 |
| Juest | for consistency please host it on 25432 to differenciate master and 2.4 | 22:24 |
| Juest | firewall is the issue perhaps? | 22:27 |
| *** Sgeo <Sgeo!~Sgeo@user/sgeo> has joined #bzflag | 22:28 | |
| blast007 | I had typoed the port number the second time :) | 22:31 |
| blast007 | but.. geno wasn't the problem | 22:31 |
| Juest | the game still accepts input when in the background | 22:40 |
| blast007 | ah, we have a Null flag type now | 22:41 |
| blast007 | so it's adding that :) | 22:41 |
| blast007 | I bet that's part of flag effect decoupling | 22:42 |
| *** Zehra <Zehra!~Yukari@user/yukari> has joined #bzflag | 23:20 | |
| Zehra | I heard of some developments/bug reports for master/2.5/future 2.6 happening. | 23:22 |
| Zehra | I do suggest that the scoring system be changed on quite a few plug-ins, as they lead to sub-optimal scoring. | 23:23 |
| Zehra | Additionally, there is the concern that such plug-ins do corrupt the scoring of statistic sites. (Virtually all custom game modes within the official distribution are known for said issue.) | 23:24 |
| Juest | would be nice to have some kind of qa or unit testing to catch those bug gotchas that accumulate over time | 23:26 |
| Zehra | My suggestion would be that some form of team scoring can be added to KeepAway and KoTH, as both modes do not feature proper team scoring integration, either by flaw or by design. | 23:26 |
| Juest | interesting | 23:27 |
| Zehra | Juest, I agree. I am very interested in the possibility of aggregating the data of bug reports/known issues and summarizing them in some form of an FAQ. | 23:27 |
| blast007 | unit tests for BZ would be like this: https://www.reddit.com/r/ProgrammerHumor/comments/176sggj/allunittestspassing/ | 23:28 |
| Juest | why not actually fix them besides making bulletins of discovered bugs? | 23:28 |
| Zehra | Hence, it would be possible to simply address most bugs or trace them rather easily via search formats. | 23:28 |
| Juest | i see | 23:28 |
| Zehra | well, some bugs are nearly impossible to reproduce, as they are glitches, generally relating to the cache or similar. | 23:28 |
| Zehra | a.k.a. corrupted configs leading to clients not properly starting..etc | 23:29 |
| blast007 | Zehra: no, Github Issues are how we track bugs | 23:29 |
| blast007 | why would we create a summary of github issues? | 23:30 |
| blast007 | just look at the issues | 23:30 |
| Juest | lol, the joke is that the sink works but its almost broken except for the basic functionality? | 23:30 |
| Zehra | blast007: But many "bugs" are actually user errors. | 23:30 |
| Zehra | a.k.a. a configuration option being incorrect or needing to be added. | 23:30 |
| Zehra | Similarly, many of the "options" people may want, already exist, but aren't readily known. | 23:31 |
| blast007 | Zehra: then if they are user errors, why would they be github issues? | 23:31 |
| blast007 | Juest: yeah. It's passing all the unit tests. Water runs, goes down the drain. :) | 23:32 |
| Zehra | blast007: The bugs and known issues would reference the forums bug/known issues sub forum. | 23:33 |
| Juest | ... | 23:33 |
| Zehra | Actual issues are added to GitHub. | 23:33 |
| blast007 | Zehra: I guess I don't understand what your point is then | 23:34 |
| Juest | Zehra: configuration errors are bugs as long its not the user misunderstanding | 23:35 |
| blast007 | yeah, the configuration file being filled with NUL bytes is a bug | 23:35 |
| Juest | only if it was not intentional | 23:35 |
| blast007 | might be some race condition with writing the config from two clients at once | 23:35 |
| blast007 | cuz I did get someone's corrupted config once, and the size was appropriate for a configuration file, but the entire contents was NUL bytes | 23:36 |
| Juest | if you click save configuration on multiple instances at the same time it corrupts? | 23:36 |
| blast007 | I was thinking that one scenario would be if you had multiple clients running, all with unsaved changes, and you log off or shut down | 23:37 |
| Juest | that too | 23:37 |
| Juest | i was having in mind that as well | 23:37 |
| blast007 | all the clients would try to save the config are roughly the same time, so that might trigger the issue | 23:37 |
| Juest | something like a log off is likely to cause that | 23:37 |
| Juest | particularly on windows | 23:37 |
| Zehra | Or abrupt power cutoff. | 23:38 |
| blast007 | power cutoff wouldn't try to save the config | 23:38 |
| Juest | no Zehra, config gets saved when the application does its shutdown routine | 23:38 |
| Juest | in linux its more common to have killed applications instead of a proper log out like on windows | 23:39 |
| Juest | as in, do a logout via killing apps instead of telling them to shut down | 23:39 |
| blast007 | maybe it's something that would be solved by just not overwriting the file to begin with | 23:40 |
| blast007 | use a rename operation and a second file with the new contents | 23:40 |
| Juest | maybe use intermediary file states? | 23:40 |
| blast007 | or maybe something where we need to lock the file | 23:41 |
| Juest | yeah something like config.new that then gets moved to config.cfg | 23:41 |
| Juest | great idea | 23:41 |
| Juest | config.lock | 23:41 |
| blast007 | or we go crazy and switch to using an SQLite database ;) | 23:41 |
| Juest | but it would cause problems with multiple instances | 23:41 |
| Juest | unless you dont care about the config when using multiple instances | 23:41 |
| Juest | about saving it* | 23:41 |
| blast007 | we do | 23:41 |
| Juest | as the user i meant | 23:41 |
| blast007 | multiple clients is relatively common | 23:42 |
| Juest | why not encourage the practice of specifying the config file for each multiple instance then? | 23:42 |
| blast007 | no | 23:42 |
| Juest | huh | 23:42 |
| blast007 | people aren't going to to the command line to run a second client | 23:42 |
| blast007 | they're clicking on the menu shortcut | 23:43 |
| Zehra | @blast007: https://forums.bzflag.org/viewtopic.php?t=19628 https://forums.bzflag.org/viewtopic.php?t=20171 Example of two threads relating to bugs which could be resolved by FAQ. | 23:43 |
| blast007 | Zehra: that's a game bug | 23:43 |
| blast007 | (the first one) | 23:43 |
| * Juest rolls eyes | 23:43 | |
| blast007 | the second one would NOT be a *frequently* asked question | 23:44 |
| blast007 | that was a very specific scenario | 23:44 |
| Zehra | I probably read the forums too much that everything looks like FAQ :p | 23:45 |
| Juest | sounds like your brain is the problem here :) | 23:46 |
| Zehra | I know, I'm only human after all... don't place the blame on me. :P | 23:46 |
| Zehra | Juest, also I highly doubt most players run their clients via command line or some shortcut. | 23:47 |
| blast007 | I'm actually ChatGPT. Gotcha. | 23:47 |
| Zehra | lol | 23:47 |
| blast007 | Zehra: most *do* run it via a shortcut, cuz that's what Windows uses :) | 23:48 |
| Juest | linux users who installed from the repositories too | 23:48 |
| blast007 | on linux it's through a desktop entry (a .desktop file) | 23:48 |
| Juest | and mac of course, has a very involved way of specifying arguments so most people run it without arguments | 23:48 |
| blast007 | mac is special | 23:48 |
| Juest | mhm, binary bundles land | 23:49 |
| blast007 | there's no menu there for apps, really | 23:49 |
| Juest | you invoke the app directly | 23:49 |
| Juest | and its a gui basically | 23:49 |
| blast007 | a .app is actually a directory that is treated special | 23:49 |
| blast007 | you can cd into an .app | 23:49 |
| Juest | i know :) | 23:50 |
| Juest | its called bundles | 23:50 |
| Juest | they're* | 23:51 |
| Juest | is it possible to toggle hideadmin on yourself? | 23:51 |
| Juest | how are people admins and not admins sometimes? | 23:51 |
| Zehra | I got a few plug-ins for that trick. | 23:52 |
| blast007 | it's a group permission | 23:52 |
| Zehra | Also depends on global groups. | 23:52 |
| blast007 | also, some server might use /password instead of groups, so they'd be a normal user until they supply the admin password | 23:52 |
| blast007 | servers* | 23:52 |
| Zehra | ^ | 23:52 |
| Juest | how can you tweak the global group permissions? | 23:52 |
| Zehra | You use a global group file... | 23:53 |
| blast007 | I don't recommend using /password on public servers though, since it's too easy to typo the command and send the admin password to the public chat :) | 23:53 |
| Juest | global groups live on the forum dont they? | 23:53 |
| Zehra | a.k.a. sample.globalgroup: spawn talk...etc | 23:53 |
| blast007 | yeah, global groups are created on the forum and you manage members then | 23:53 |
| Zehra | give me a second... | 23:53 |
| Juest | i asked how do you manage the permissions? | 23:53 |
| blast007 | but the groupdb file that bzfs reads is what assign permissions to groups | 23:53 |
| Juest | oh | 23:53 |
| Juest | so the permissions is handled by the groupdb irrespective of global or local group? | 23:54 |
| blast007 | yeah | 23:54 |
| blast007 | and you don't have to own a group to assign rights to it | 23:54 |
| Juest | long ago i had a setup of a bzfs | 23:54 |
| Juest | but i lost that stuff | 23:54 |
| Zehra | that's how some servers assign admins, like they'll add in one of the more popular group configurations. | 23:55 |
| Zehra | oh, also don't forget, superUser can be used for permissions... no global groups needed. | 23:56 |
| Zehra | But that is pretty rare, if done at all now days. | 23:57 |
| Zehra | Might be good if you only grant like 3-5 people permissions or so. | 23:57 |
| * Juest shrugs | 23:57 | |
| Juest | there's some inconsistencies with the server setup and things like that | 23:58 |
| Juest | what folder does bzflag read when listing worlds in start server? | 23:58 |
| Zehra | the /saveworld folder... known as "worlds" IIRC. | 23:58 |
| Juest | ah | 23:59 |
| Zehra | basically where the config file is and a directory down known as "worlds". | 23:59 |
| Zehra | IIRC | 23:59 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!