| 00:00.38 | blast007 | if you can replicate it in debug build it's probably better than release with debug info |
| 00:01.53 | blast007 | I not really sure what limitations the latter has |
| 00:02.36 | Juesto | none, it's just easier to debug that way |
| 00:02.56 | blast007 | easier than a debug build? |
| 00:03.39 | Juesto | yes because you don't have nt noise and not the slowdowns due to debug checks |
| 00:06.52 | blast007 | I guess the slowdowns aren't that much of a problem if I'm not on an atom :) |
| 00:07.28 | Juesto | kind of |
| 00:07.39 | Juesto | depends |
| 00:07.48 | *** join/#bzflag Zehra (~Zehra@unaffiliated/zehra) |
| 00:08.08 | JeffM | Debugging release is harder because optimizer is on and stepping will get odd |
| 00:08.24 | JeffM | But sometimes you do have to do it |
| 00:09.21 | JeffM | I most often just try debug and if it dupes all is good |
| 00:09.43 | JeffM | But Iâm also not on a toy cpu |
| 00:10.05 | Juesto | I'm more for looking at stacktrace |
| 00:10.31 | Juesto | meh |
| 00:10.36 | JeffM | Then it can work, unless stuff gets inlined |
| 00:11.08 | JeffM | What problem are you trying to debug? |
| 00:12.29 | blast007 | a crash when you hit connect again while in the middle of connecting to a server |
| 00:13.01 | blast007 | easy fix: disable the connect button while in the middle of connecting :P |
| 00:13.54 | JeffM | Yeah |
| 00:14.09 | JeffM | Iâm assuming you canât dupe? |
| 00:14.35 | JeffM | If it was me Iâd try debug so I could get the exception message |
| 00:15.15 | blast007 | nope, can't dupe |
| 00:15.22 | blast007 | tried on three computers, two windows, one linux |
| 00:18.18 | *** join/#bzflag infobot (ibot@rikers.org) |
| 00:18.18 | *** topic/#bzflag is https://BZFlag.org || http://ohloh.net/p/bzflag || https://bzflag.org/help || Channel Logs: http://infobot.rikers.org/%23bzflag/ || 2.4.14 released! https://forums.bzflag.org/viewtopic.php?f=62&t=19953 |
| 00:18.18 | *** mode/#bzflag [+v infobot] by ChanServ |
| 00:18.19 | JeffM | Ahh , connect spawn, reconnect |
| 00:18.26 | JeffM | Sure |
| 00:18.31 | JeffM | Donât do that |
| 00:18.44 | Juesto | spawn and connect happening simultaneously |
| 00:18.55 | JeffM | Yeah |
| 00:18.59 | Juesto | i know, it's still a bug, JeffM |
| 00:19.03 | JeffM | Thatâs joint to jack the state |
| 00:19.15 | Juesto | wdym |
| 00:19.42 | JeffM | Spawn should set a flag to block joininternetgame2 until itâs complete |
| 00:19.47 | Juesto | also it's expected that the client refuses to connect anymore after a lot of attempts? |
| 00:20.26 | Juesto | and if you back out when it's connecting it's possible to use leave game |
| 00:20.41 | JeffM | The network code is super fragile |
| 00:20.49 | JeffM | Iâm surprised it works at all |
| 00:20.54 | Juesto | perhaps that flag is crashing the game? |
| 00:21.02 | Juesto | no sarcasm? |
| 00:21.05 | JeffM | That flag doesnât exist |
| 00:21.10 | Juesto | huh.... |
| 00:21.14 | Juesto | oh i see |
| 00:21.14 | JeffM | Iâm saying it should |
| 00:21.28 | Juesto | right i misinterpreted,sorry |
| 00:21.43 | JeffM | Pffft I can create a server that crashes ever client that connects to it |
| 00:21.49 | Juesto | lol |
| 00:21.50 | JeffM | Itâs super jankey |
| 00:22.10 | JeffM | Bzflag is not an Internet game |
| 00:22.15 | Juesto | why it was made like that |
| 00:22.32 | Juesto | it is a game intended for lan parties? |
| 00:22.34 | JeffM | Itâs a lan game from 1994 |
| 00:22.38 | JeffM | Yeah |
| 00:23.00 | Juesto | well why people even try to online with it |
| 00:23.04 | Juesto | welp |
| 00:23.05 | JeffM | It just happens to work on the internet |
| 00:23.23 | Juesto | of course, isn't that expected? |
| 00:23.24 | JeffM | They like pain |
| 00:23.27 | blast007 | we should switch to IPX/SPX |
| 00:23.34 | Juesto | oh boy |
| 00:23.51 | blast007 | null modem connections |
| 00:23.57 | JeffM | None of the networking or simulation is tailored for internet play |
| 00:23.58 | blast007 | smoke signals |
| 00:24.05 | JeffM | It all needs to be redone |
| 00:24.18 | Juesto | should that happen soon? |
| 00:24.42 | JeffM | Should? It should have happened 10 years ago |
| 00:24.49 | Juesto | why it didn't |
| 00:25.07 | Juesto | development in this game is just too strange |
| 00:25.14 | JeffM | Nobody did it |
| 00:25.14 | blast007 | because nobody did it |
| 00:25.21 | Juesto | lame |
| 00:25.35 | JeffM | Open source projects can be strange |
| 00:25.41 | JeffM | Indeed |
| 00:25.46 | Juesto | not all are |
| 00:25.56 | Juesto | this one's a unique case imo |
| 00:25.58 | JeffM | Itâs a big take, no single person wants to do it alone |
| 00:26.10 | Juesto | and a team? |
| 00:26.16 | Flash | anyone ever looked at ZeroMQ? |
| 00:26.44 | JeffM | nothing in bzflag has ever been done as a team |
| 00:26.52 | JeffM | team implies strong cooperation |
| 00:27.09 | JeffM | that is not something that traditionally happens in bzflag. |
| 00:27.21 | JeffM | the "patches welcome" mentality still holds |
| 00:27.46 | JeffM | but as for your bug, sure makes complete sense, add it to the list of others that are in there. |
| 00:28.33 | Juesto | :) |
| 00:28.37 | JeffM | my guess is the socket isn't flushed before the spawn gets returned. |
| 00:28.43 | JeffM | so the message is processed |
| 00:28.43 | Juesto | where? github? |
| 00:28.59 | Juesto | it's very specific where it crashes |
| 00:29.15 | Zehra | Maybe check the code a bit? |
| 00:29.29 | JeffM | the question is what "should" happen in that race conditon? does the rejoin override any outstanding spawns? or shoudl the rejoin wait for the spawn to complete. |
| 00:29.35 | JeffM | that's a specication problem |
| 00:29.37 | Juesto | sorry for the not-so-good str blast007 |
| 00:29.39 | JeffM | both are valid solutions. |
| 00:29.52 | Juesto | dilemma! |
| 00:30.20 | JeffM | I'd say that all outstanding responces should be ignored once a rejoin happens |
| 00:30.39 | JeffM | this would be another case where the 2 byte security header would help |
| 00:30.44 | Juesto | JeffM: actually the flag you proposed will be enough to fix this bug |
| 00:30.59 | JeffM | each session would have a different header and it be easy to ignore the spawn if it came in. |
| 00:31.10 | JeffM | that would be the minimum dirty way |
| 00:31.18 | JeffM | adding a global is rarely the right solution ;) |
| 00:31.21 | Juesto | so, pretty hacky |
| 00:31.27 | JeffM | yeah |
| 00:31.33 | Juesto | well not really global |
| 00:31.46 | JeffM | it would have to be in the global scope since there isn't a "client" object. |
| 00:31.51 | blast007 | closing the socket and resetting the state is probably the right solution |
| 00:31.54 | JeffM | and it would have to span messages. |
| 00:32.07 | *** join/#bzflag spldart (~spldart@bzflag/contributor/spldart) |
| 00:32.07 | *** mode/#bzflag [+v spldart] by ChanServ |
| 00:32.10 | Juesto | so, cancelling the spawn? |
| 00:32.12 | JeffM | possibly, but my fear is VERY laggy connections |
| 00:32.26 | Juesto | oh boy |
| 00:32.26 | JeffM | you the user choose to connect, that means you the user want to cancel the spawn |
| 00:32.39 | Juesto | well just disconnect the previous |
| 00:32.48 | *** join/#bzflag short_circuit (~spldart@bzflag/contributor/spldart) |
| 00:32.48 | *** mode/#bzflag [+v short_circuit] by ChanServ |
| 00:32.54 | JeffM | the spawn return time is not controllable by the user, the reconnect is |
| 00:32.59 | Juesto | And there we go, yay brainstorming |
| 00:33.16 | JeffM | a null check in the spawn processor is probably the simplest fix |
| 00:33.20 | JeffM | my guess is it's checking mytank |
| 00:33.28 | JeffM | and mytank is null |
| 00:34.09 | Juesto | there's a certain cancellation when the user reconnects, that is not as wide as it should |
| 00:34.36 | JeffM | my guess is the other code makes assumption and doesn't check for validity |
| 00:34.41 | Juesto | we'll see, I'll debug build on fedora |
| 00:35.05 | JeffM | you never know what you'll get with packets, so it's best to make the spawn code more robust |
| 00:35.22 | JeffM | then the client is protected against malicious servers too |
| 00:35.31 | Juesto | better :) |
| 00:35.31 | JeffM | not just the case you happen to be hitting now |
| 00:35.40 | blast007 | if (myTank) leaveGame(); |
| 00:36.07 | Juesto | lol that would not let you spawn at all i think |
| 00:36.19 | Juesto | i mean you would get disconnected |
| 00:36.21 | JeffM | it's the only way to be safe |
| 00:36.33 | Juesto | ye |
| 00:36.51 | Juesto | so only let it connect if there's isn't anything else going on |
| 00:36.59 | blast007 | Juesto: no, that's the current code |
| 00:37.03 | Juesto | oh |
| 00:37.10 | blast007 | when you want to connect, it runs that |
| 00:37.12 | JeffM | normal C++ bzfs shoves a LOT of stuff out the pipe when it shouldn't, or could wait. |
| 00:37.34 | Juesto | so unnecessary precache/preload? |
| 00:37.42 | JeffM | leave game should check to see if an connection is in progress or not |
| 00:37.45 | blast007 | is 'delete myTank' a problem if myTank is null? |
| 00:37.54 | JeffM | I could see calling leaveGame in the middle of a connection being bad. |
| 00:37.59 | Juesto | null exception |
| 00:38.26 | JeffM | delete is supposed to null check |
| 00:38.33 | JeffM | but it's generaly considered bad form |
| 00:38.58 | JeffM | is the ptr set to null after the delete ? |
| 00:39.02 | blast007 | yes |
| 00:39.33 | JeffM | it's MsgAlive? |
| 00:39.37 | blast007 | just thinking if we could just *always* call leaveGame() when connecting and not only when myTank isn't NULL |
| 00:39.49 | JeffM | on spawn? |
| 00:39.55 | blast007 | no |
| 00:39.59 | JeffM | or at the start of join? |
| 00:40.11 | blast007 | when requesting a connect |
| 00:40.23 | JeffM | yeah there is no harm in always parting. |
| 00:40.37 | Juesto | doesn't that already happen when you connect again? |
| 00:40.52 | blast007 | leaveGame() resets a lot of state and deletes the serverLink |
| 00:41.06 | Juesto | ah, yeah that seems better |
| 00:41.12 | blast007 | Juesto: only if you're far enough along so that that myTank is set |
| 00:41.17 | JeffM | well I see one bug already |
| 00:41.24 | Juesto | :D |
| 00:41.28 | JeffM | Player* tank = getPlayerByIndex(playerIndex); |
| 00:41.28 | JeffM | if (tank == myTank) { |
| 00:41.28 | JeffM | <PROTECTED> |
| 00:41.36 | JeffM | getPlayerByIndex CAN return null |
| 00:41.44 | JeffM | myTank can ALSO be null |
| 00:41.47 | JeffM | null == null |
| 00:41.58 | Juesto | ah, i spotted an attempt to load observer tank texture |
| 00:42.27 | Juesto | that shouldn't be happening |
| 00:43.29 | JeffM | I'd add if (tank == nullptr) |
| 00:43.29 | JeffM | break; |
| 00:43.29 | JeffM | after Player* tank = getPlayerByIndex(playerIndex); |
| 00:43.29 | JeffM | in playing.cxx : handleServerMessage : in the case MsgAlive: section |
| 00:47.15 | Juesto | uhm, where do i do configure? |
| 00:47.39 | Juesto | in which dir |
| 00:50.24 | Juesto | I'm confused |
| 00:59.35 | Juesto | blast007: i dont see any configure but configure.ac |
| 01:00.48 | blast007 | ./autogen.sh && ./configure --enable-debug && make |
| 01:01.23 | blast007 | README.Linux has info about dependencies |
| 01:02.11 | Juesto | I've seen, looks outdated |
| 01:02.40 | blast007 | ? |
| 01:03.03 | Juesto | now fedora uses dnf from 22 onwards |
| 01:03.17 | blast007 | so? |
| 01:04.27 | blast007 | (the 2.6 file does show dnf, and splits out CentOS which uses yum) |
| 01:04.55 | blast007 | but it's not like we cover *every* single package manager |
| 01:05.30 | Juesto | should i worry about the no s spewed by autogen.sh? |
| 01:05.52 | blast007 | no s? |
| 01:06.02 | Juesto | "no" results |
| 01:06.14 | blast007 | that's probably configure |
| 01:06.27 | Juesto | ah hm |
| 01:07.44 | blast007 | if you run the configure without the make, you can look at the results of configure that it writes out at the end |
| 01:15.25 | Juesto | still, should i worry? |
| 01:16.42 | blast007 | I think there will be things that say no |
| 01:17.29 | Juesto | there's no tag for 2.6? |
| 01:20.34 | blast007 | no, because it's not released |
| 01:20.47 | blast007 | master is 2.5, which will be released as 2.6 |
| 01:23.49 | Juesto | :/ |
| 01:23.57 | Juesto | âSceneNode::SceneNode()â:SceneNode.cxx:41:40: error: âmemsetâ used with length equal to number of elements without multiplication by element size [-Werror=memset-elt-size] |
| 01:23.57 | Juesto | <PROTECTED> |
| 01:23.57 | Juesto | <PROTECTED> |
| 01:23.57 | Juesto | cc1plus: all warnings being treated as errors |
| 01:24.08 | Juesto | ehhhh |
| 01:24.23 | blast007 | which version of the game? what version of gcc/g++? |
| 01:26.53 | Juesto | HEAD |
| 01:27.11 | Juesto | 8.1.1 |
| 01:27.29 | blast007 | master? |
| 01:27.37 | Flash | where is SceneNode.cxx? |
| 01:27.38 | Juesto | HEAD is 2.4 |
| 01:27.50 | Flash | that's a gcc version :) |
| 01:28.33 | blast007 | src/geometry/ |
| 01:28.37 | Juesto | I'm on fedora 28 |
| 01:29.08 | Flash | well that line was changed 15 years ago :) |
| 01:29.48 | Flash | but it looks like an error to me |
| 01:29.51 | Juesto | ? |
| 01:30.25 | blast007 | should be a * not a &, right? |
| 01:30.53 | Flash | yes |
| 01:31.03 | Flash | sphere is in fact an array of 4 GLFloats |
| 01:31.08 | blast007 | yup |
| 01:31.21 | Flash | yay for compiler warnings! |
| 01:38.57 | Juesto | welp |
| 01:39.20 | Juesto | what version was used to compile |
| 01:39.28 | Juesto | the released version |
| 01:41.05 | Juesto | how do i disable treat warnings as errors |
| 01:43.42 | *** join/#bzflag BZnotify (BZnotify@gateway/service/github.com/x-cythbbnwfhpljygs) |
| 01:43.42 | BZnotify | [13bzflag] 15blast007 pushed 1 new commit to 062.4: 02https://git.io/vppKO |
| 01:43.43 | BZnotify | 13bzflag/062.4 148dcd4c3 15Scott Wichser: Fix a memset. |
| 01:43.43 | *** part/#bzflag BZnotify (BZnotify@gateway/service/github.com/x-cythbbnwfhpljygs) |
| 01:43.50 | blast007 | pull that change |
| 01:45.12 | Flash | Juesto, the debug build is aggressive on purpose (that 15 year old error should have been fixed 15 years ago) |
| 01:45.13 | Juesto | me? k |
| 01:45.26 | Flash | but the production build doesn't have -Wall |
| 01:46.12 | Flash | you can disable any particular warning with specific compiler flags, but if one is going to be developing, one must learn to interact with the compiler/build system |
| 01:46.27 | Juesto | amazing how it got through all those years |
| 01:47.05 | blast007 | there's lots of silly code :) |
| 01:47.13 | Juesto | i might remove -Wall from debug config maybe |
| 01:47.15 | blast007 | 2.4.4 cleaned up a *lot* of crap |
| 01:47.20 | blast007 | or don't |
| 01:47.26 | blast007 | just build and see if it works now |
| 01:48.25 | Juesto | blast007: now it's geometry/SphereSceneNode, 555, 566 |
| 01:48.52 | Juesto | ignored-qualifiers error |
| 01:48.53 | blast007 | 45 seconds left on this Fedora 28 ISO :) |
| 01:49.10 | Juesto | :d |
| 01:49.49 | Juesto | kinda happy but guilty that I'm getting someone onto this |
| 01:50.09 | Juesto | I mean willing to fix it |
| 01:50.13 | Juesto | or something |
| 01:50.13 | blast007 | not the first time a newer compiler has triggered was warnings |
| 01:50.30 | blast007 | I'm using Debian, so I'm on GCC 6.3.0 :) |
| 01:50.39 | Juesto | heh |
| 01:51.00 | Juesto | fedora's as new as may |
| 01:51.16 | Juesto | may 2 to be exact |
| 01:53.24 | JeffM | Zehra: did you find your api function? |
| 01:56.46 | Zehra | JeffM: I read your post and checked 2.4.14 and it was there. :) |
| 01:57.10 | JeffM | Cool |
| 01:57.23 | JeffM | Why were you using year old headers? |
| 01:58.19 | Zehra | I'm probably a bit stupid, but most of my plug-in work was done on 2.4.10 and I basically sort of stuck with it. :/ |
| 01:58.55 | Zehra | I'll have to update to the latest version of Visual Studio too. |
| 01:59.35 | JeffM | Just work build from git |
| 02:00.38 | Juesto | hmm, kinda surprised at the bug I've hit :d |
| 02:00.41 | Juesto | well meh |
| 02:00.42 | Zehra | I'm not exactly sure how to do it. :p |
| 02:00.46 | Juesto | not so surprised noe |
| 02:00.57 | JeffM | Itâs good to learn |
| 02:01.12 | Juesto | Zehra: which part exactly? |
| 02:01.49 | Zehra | Juesto: How to work the build from Git. |
| 02:02.02 | Juesto | work the build???? |
| 02:02.13 | Juesto | I'm not sure what that means |
| 02:02.32 | Juesto | you use git client and use vs |
| 02:02.36 | Juesto | i guess (?) |
| 02:02.42 | JeffM | Get code, get dependencies, set environs, build |
| 02:02.58 | Juesto | it's easier on linux lol |
| 02:03.00 | blast007 | Zehra: latest 2.4 needs Visual Studio 2017 |
| 02:03.19 | blast007 | Juesto: it's really not |
| 02:03.22 | JeffM | Itâs the same on Linux, |
| 02:03.24 | Juesto | vs17 is free |
| 02:04.00 | Juesto | blast007: last time i tried it was just too complicated |
| 02:04.11 | Juesto | and broken |
| 02:04.14 | blast007 | when was that? |
| 02:04.25 | Juesto | before the upgrade to vs2017 |
| 02:04.32 | Juesto | 2.4 something |
| 02:04.40 | JeffM | Itâs pretty easy |
| 02:04.48 | Zehra | blast007: I have the 2015 version, so I basically need to update to 2017. :p |
| 02:04.55 | blast007 | Zehra: yup |
| 02:04.59 | JeffM | The hardest thing is setting the environment variable |
| 02:05.17 | JeffM | And knowing to restart vc after |
| 02:05.18 | blast007 | and that'll go away in 2.5 |
| 02:05.22 | Juesto | bzflag-dependencies repo was kinda broken |
| 02:05.31 | JeffM | Going relative path? |
| 02:05.34 | blast007 | yup |
| 02:05.37 | JeffM | Cool |
| 02:05.44 | Juesto | actually one can build with 2015 |
| 02:05.47 | Juesto | i think |
| 02:05.55 | blast007 | the project files are 2017 |
| 02:05.56 | JeffM | Should see if there are nugget packages for any of them |
| 02:06.04 | Juesto | there's compatibility |
| 02:06.19 | Juesto | no? |
| 02:06.36 | blast007 | I haven't tried, because I have 2017 |
| 02:06.39 | JeffM | 2017 can build 2015, 2015 cant build 2017 |
| 02:06.51 | JeffM | It canât go forward in time |
| 02:07.02 | Zehra | Juesto: I might have something on my blog on how to do the environment variable... I'm checking for it now. |
| 02:07.23 | Juesto | JeffM: possible if 2017 is installed afaik |
| 02:07.48 | blast007 | if you have 2017 installed, why would try to open a 2017 project in 2015? |
| 02:07.49 | JeffM | Itâs super easy |
| 02:08.06 | Juesto | native targeting |
| 02:08.19 | Juesto | and the project format hasn't changed, blast007 |
| 02:08.22 | JeffM | Yeah if the target is there |
| 02:08.50 | JeffM | 2017 is better than 2015 |
| 02:08.50 | blast007 | didn't say iti did, just saying it wouldn't make sense |
| 02:09.11 | Juesto | :P |
| 02:09.27 | blast007 | and it certainly doesn't help Zehra because he doesn't have 2017 ;) |
| 02:09.36 | Juesto | anyways how do i disable -Wall ? |
| 02:09.46 | Juesto | for debug builds |
| 02:11.28 | blast007 | probably have to edit the configure.ac and rebuild |
| 02:11.37 | blast007 | rebuild the configure script, I mean |
| 02:12.32 | Juesto | why is upnp disabled |
| 02:13.00 | blast007 | because you don't have the extra dependency, but it's silly anyway |
| 02:13.18 | Juesto | which |
| 02:13.22 | blast007 | I don't know |
| 02:13.30 | blast007 | but you don't need/want it |
| 02:13.32 | Juesto | it's set to disabled by default |
| 02:13.57 | blast007 | it's just to poke holes in your NAT to run a public game server |
| 02:14.24 | Juesto | how far i have to look in the file to find -Wall |
| 02:14.34 | Juesto | ik what upnp does |
| 02:14.36 | blast007 | editors have search |
| 02:14.39 | Juesto | it's useful |
| 02:15.01 | Juesto | just found lol |
| 02:15.08 | blast007 | if people can't figure out how to forward ports in their router, they have no business running a game server |
| 02:15.47 | blast007 | and it's only useful for people behind NAT. a VPS or dedicated server will have no use for that. |
| 02:16.56 | Juesto | i know |
| 02:17.02 | Juesto | -_- |
| 02:17.14 | Juesto | is it optional functionality |
| 02:17.28 | Juesto | i think it could be compiled in |
| 02:17.39 | Juesto | without much issue |
| 02:17.52 | blast007 | I want to remove it |
| 02:17.53 | Juesto | and I'd put upnp for the listen not bzfs |
| 02:17.57 | Juesto | :/ |
| 02:17.59 | Juesto | oh well |
| 02:18.14 | blast007 | for the listen? |
| 02:18.40 | blast007 | is a rebel and uses yum on Fedora 28 |
| 02:18.44 | Juesto | yea, client's "start server" |
| 02:18.47 | Juesto | LUL |
| 02:19.00 | blast007 | that uses bzfs |
| 02:19.03 | Juesto | yum is a symblink to dnf |
| 02:19.17 | Juesto | and if bzfs isn't installed? |
| 02:19.27 | blast007 | then it doesn't start a server |
| 02:19.41 | Juesto | oh |
| 02:20.19 | Juesto | most of the functions section gives a no |
| 02:20.25 | blast007 | I want to remove that menu too ;) |
| 02:20.26 | Juesto | during configure |
| 02:20.43 | Juesto | lots of cleanup? |
| 02:20.48 | blast007 | yeah |
| 02:21.06 | Juesto | good idea for 2.6 i guess |
| 02:21.28 | Zehra | looks forward to 2.6 |
| 02:21.31 | blast007 | yup |
| 02:21.38 | Juesto | feels like he came at a fortunate time |
| 02:21.52 | blast007 | lots of cleaning planned for 2.6 |
| 02:22.12 | blast007 | replacing our custom audio code with OpenAL, for instance, and protocol fixes |
| 02:23.30 | blast007 | (also, you can use SDL2-devel instead of SDL-devel, as we can use SDL2) |
| 02:23.56 | Juesto | the bug i got fixed too maybe |
| 02:24.01 | blast007 | the configure script will pick SDL2 if it's available (and at least 2.0.3) |
| 02:24.24 | Juesto | I've installed both sdl |
| 02:24.55 | blast007 | k |
| 02:42.03 | JeffM | Building on windows https://usercontent.irccloud-cdn.com/file/4XrAZ1v5/BuildingBZ.mp4 |
| 02:42.52 | *** join/#bzflag DTRemenak|RDP (~DTRemenak@about/essy/CrazyCoder/DTRemenak) |
| 02:42.52 | *** mode/#bzflag [+v DTRemenak|RDP] by ChanServ |
| 03:05.03 | *** join/#bzflag BZnotify (BZnotify@gateway/service/github.com/x-rbahtwlhrufhumyz) |
| 03:05.03 | BZnotify | [13bzflag] 15blast007 pushed 2 new commits to 062.4: 02https://git.io/vppic |
| 03:05.03 | BZnotify | 13bzflag/062.4 14bc3d695 15Scott Wichser: Fix warnings about "type qualifiers ignored on cast result type" |
| 03:05.03 | BZnotify | 13bzflag/062.4 1401bacd4 15Scott Wichser: Fix warning "unnecessary parentheses in declaration of âskyboxNamesâ" |
| 03:05.03 | *** part/#bzflag BZnotify (BZnotify@gateway/service/github.com/x-rbahtwlhrufhumyz) |
| 03:06.13 | Juesto | all fixed? |
| 03:06.51 | blast007 | yup |
| 03:07.10 | blast007 | well, it at least builds now, but I'm sure it will still crash for you ;) |
| 03:07.43 | *** join/#bzflag the_map (the_map@gateway/vpn/privateinternetaccess/themap/x-38220869) |
| 03:08.10 | Juesto | ah |
| 03:08.26 | Juesto | I've ditched -Wall anyway |
| 03:08.54 | Juesto | nvm |
| 03:10.12 | Juesto | apparently not |
| 03:10.21 | Juesto | or whatever |
| 03:17.36 | BulletCatcher | Compiling the test program with GLFW is more "interesting" than with SDL2. First, windowed mode works the way it ought to. |
| 03:17.43 | BulletCatcher | With fullscreen mode 1 I see the window pop up, then the screen changes resolution and I have an 800x600 portal to my 1920x1080 desktop that I can move around. |
| 03:17.48 | BulletCatcher | The window with the colorful triangle is there, and when I press ESC in the window my display goes back to its normal 1920x1080 self. |
| 03:17.50 | BulletCatcher | I haven't been brave enough to try modes 2 or 3. |
| 03:20.36 | blast007 | 2 will start fullscreen (at around 800x600) and 3 will run fullscreen at your desktop resolution |
| 03:22.18 | BulletCatcher | Another fun side effect: my mouse seemed to "remember" the 800x600 viewport and wouldn't move out of it for maybe a minute even though it was back at 1920x1080. |
| 03:22.30 | blast007 | weird.. |
| 03:31.30 | BulletCatcher | Back to the SDL2 version: it is simply catching signals because kill -HUP makes it quit but kill -INT does not. |
| 03:33.07 | BulletCatcher | You may be able to reproduce that behavior in windowed mode. |
| 03:37.47 | BulletCatcher | This is on my on-board GeForce 7025 GPU with the NVIDIA 304.137 drive (OpenGL version 2.1.2). |
| 03:52.19 | Juesto | hm |
| 04:21.42 | *** join/#bzflag BZnotify (BZnotify@gateway/service/github.com/x-tejfhhmamtmobgfp) |
| 04:21.42 | BZnotify | [13bzflag] 15JMakey pushed 1 new commit to 062.4: 02https://git.io/vppXU |
| 04:21.42 | BZnotify | 13bzflag/062.4 14303a753 15Jeff Makey: Remove m4/curses.m4, which is no longer needed with PKG_CHECK_MODULES macro. |
| 04:21.42 | *** part/#bzflag BZnotify (BZnotify@gateway/service/github.com/x-tejfhhmamtmobgfp) |
| 04:42.06 | JeffM | <PROTECTED> |
| 04:43.29 | Flash | is that a choice that can be isolated to Windows? |
| 04:43.38 | Flash | I don't remember nuget in any other context |
| 04:44.06 | JeffM | Yes, itâs part of the vc projects |
| 04:44.29 | Flash | I feel so dirty ... I've been using VC this year |
| 04:46.02 | JeffM | But nuget does totally exist for Clu packages on other systems, Linux and OS X |
| 04:46.21 | JeffM | Clr packages, sorry |
| 04:47.36 | JeffM | Nuget -restore your project.sln |
| 04:48.09 | JeffM | Itâll get all your deps and put em in the GAC |
| 04:48.37 | Juesto | osx can use visual studio 2017 now |
| 04:48.49 | Juesto | it's tied to xamarin |
| 04:49.58 | JeffM | Sort of, it uses mono |
| 04:50.18 | JeffM | But yes itâs a good .net ide on OS X |
| 04:51.31 | JeffM | Monodev/Xamarin studio is not the worst thing though |
| 04:52.40 | Flash | package managers and package manager managers |
| 04:53.32 | Flash | one ring to rule them all... |
| 04:54.08 | allejo | VS for mac is not too bad. not even close to VS on windows, but it's usable. no C++ support though :( |
| 04:55.23 | JeffM | That isnât its goal |
| 04:56.11 | Flash | if we are going to be cross-platform, we have to respect the dominant build environments on each platform |
| 04:56.30 | Flash | meaning VS on windows, Xcode on Mac and ... nothing on Linux |
| 04:56.53 | JeffM | Prayer and chewing gum is traditional on Linux |
| 04:57.22 | Flash | I do have to say, though, that after spending some quality time with CMake, I am more sympathetic to replacing our autotools based config |
| 04:57.23 | allejo | lol |
| 04:57.45 | Flash | it even builds the VS projects under VS 17 |
| 04:58.16 | Flash | @JeffM I forget ... are you a scotch guy? |
| 04:58.56 | JeffM | I canât really drink |
| 04:59.09 | JeffM | Make is not horrible |
| 04:59.22 | Flash | I thought someone in this community was a scotch guy |
| 04:59.29 | JeffM | I prefer using .sln files on all platforms |
| 05:00.17 | Flash | that's a Microsoft solution |
| 05:00.49 | JeffM | Nope |
| 05:01.27 | Flash | https://file.org/extension/sln |
| 05:01.41 | JeffM | msbuild BZFSPro.sln, works on the big 3 |
| 05:02.38 | JeffM | Most commonly associated with, not âexclusively associated withâ |
| 05:03.26 | Flash | honestly, I've never heard of msbuild and I just finished a survey of C++ build environments |
| 05:03.33 | Flash | after failing miserably with gradle |
| 05:03.40 | JeffM | I did not say c++ |
| 05:03.54 | Flash | fair enough |
| 05:04.14 | Flash | I'm just saying, for cross-platform, it's an edge case |
| 05:04.31 | JeffM | C# works fine cross platform |
| 05:04.37 | JeffM | You assume too much |
| 05:05.03 | Flash | I am not assuming. I also know that C# is not widely used on non Microsoft platforms |
| 05:05.57 | JeffM | Itâs used a lot in gaming, anything using unity uses it |
| 05:06.23 | JeffM | And it works just fine |
| 05:06.25 | Flash | and that may be my blind spot. My main experience is not in gaming |
| 05:06.40 | JeffM | Thankfully bzflag is a game |
| 05:07.37 | JeffM | Also in pure c# there is no need to build for each platform, one bin all systems |
| 05:07.49 | Flash | or Java |
| 05:07.52 | Flash | or Python |
| 05:07.57 | Flash | or ... |
| 05:08.05 | JeffM | Similar runtime setup to java |
| 05:08.13 | JeffM | Just with less suck |
| 05:23.20 | Juesto | why connecting to a server just breaks |
| 05:23.34 | Juesto | also im getting spam in gdb about ansi not supported |
| 05:23.48 | Juesto | ANSI code not supported |
| 05:25.05 | Juesto | wow |
| 05:25.12 | Juesto | I'm stuck but no crash |
| 05:25.49 | Flash | 2.4 is stuck supporting older compilers |
| 05:26.24 | Juesto | okay i crashed it again |
| 05:26.49 | Juesto | it gets harder to reproduce as speed increases, blast007 |
| 05:27.44 | Juesto | i still don't have symbols for some reason |
| 05:27.50 | Juesto | I'm running debug build |
| 05:28.34 | Juesto | nvm actually |
| 05:32.30 | Juesto | still getting unknown/unresolved stacks |
| 05:35.00 | Flash | Juesto, what platform? |
| 05:46.36 | Juesto | Fedora, Flash |
| 05:47.17 | Flash | what do you mean by "unknown/unresolved stacks"? |
| 05:52.59 | Juesto | that it's not showing functions but addresses |
| 05:54.34 | Flash | can you pastebin a stack trace? |
| 06:22.08 | Juesto | nop, xterm is making things difficult |
| 06:22.30 | Flash | are you remoting in? |
| 06:22.53 | Flash | xterm is probably the worst possible terminal |
| 06:23.02 | Flash | gnome-terminal is much more friendly |
| 06:24.04 | Juesto | http://paste.fedoraproject.org/paste/rdBf7rxIB61iXwTMQfk6Q |
| 06:25.04 | Juesto | xfce one is nice too |
| 06:25.14 | Juesto | i don't like gtk3 style |
| 06:25.19 | Juesto | is rxvt good? |
| 06:25.26 | Flash | don't know |
| 06:25.29 | Juesto | gnome overall has became heavier |
| 06:25.31 | Flash | that link doesn't work |
| 06:26.46 | Juesto | https://paste.fedoraproject.org/paste/rdBf7rxIBF61iXwTMQfk6Q |
| 06:26.48 | Juesto | now |
| 06:27.23 | Flash | how did you generate that? |
| 06:27.51 | Juesto | gdb |
| 06:27.56 | Juesto | bt command |
| 06:28.09 | Juesto | backtrace |
| 06:28.15 | Flash | it doesn't work because your memory is all corrupted |
| 06:28.25 | Juesto | what do you know |
| 06:28.37 | Flash | #5 should be somewhere in main. 000000 is a bad address |
| 06:28.48 | Juesto | it said i need dnf install debug-info |
| 06:29.02 | Flash | #4 and #3 are way high up; those are unlikely locations |
| 06:29.11 | Juesto | dnf debuginfo-install |
| 06:29.13 | Flash | #2 is too low |
| 06:29.17 | Juesto | :/ |
| 06:29.26 | Juesto | idk that's a debug build |
| 06:29.37 | Juesto | i got something different i don't remember now |
| 06:29.50 | Flash | so in gdb, if you set a break in main, can you get a good trace? |
| 06:29.58 | Juesto | idk |
| 06:30.06 | Juesto | I'm not seeing symbols |
| 06:30.17 | Juesto | I'll build in Windows with pdbs |
| 06:30.33 | Flash | switching platforms is creating multiple unknowns |
| 06:30.52 | Juesto | nah i reproduced the bug well |
| 06:31.05 | Juesto | it's just a matter of hitting it timely |
| 06:31.22 | Flash | if you are going to debug in windows, then you need to focus your development there |
| 06:31.37 | Flash | bouncing back and forth is not a primary debugging technique |
| 12:41.18 | blast007 | JeffM: I have zero interest in NuGet, at least for a C++ application. It's just a bunch of outdated packages built against the wrong version of Visual Studio. |
| 12:46.40 | blast007 | I do prefer the way we have it now where we can update our dependencies at will and choose build options. |
| 12:47.57 | blast007 | the goal is to use the same bzflag-dependencies repo for building some/most macOS dependencies |
| 14:26.07 | Juesto | :p |
| 14:26.16 | Juesto | and eventually linux i assume, yes? |
| 14:26.56 | blast007 | not for normal builds, but maybe for flatpak/snaps? dunno how those work at all |
| 14:30.51 | Juesto | :/ |
| 14:31.32 | Juesto | should i post my bug to bzf github? |
| 14:31.45 | blast007 | sure |
| 14:32.01 | JeffM | did you try changing the code I sugested? |
| 14:32.49 | Juesto | not yet and my computer is not behaving well in Fedora after all, I'll dev in Windows |
| 14:32.59 | Juesto | later I'll do that |
| 14:33.55 | Juesto | I'll post it, please leave me the suggestion noted there, thanks in advance |
| 14:34.27 | Juesto | my post will be a bit vague but it'll do |
| 14:39.41 | Juesto | there it is |
| 14:42.58 | *** join/#bzflag JeffM (sid48618@gateway/web/irccloud.com/x-lxxihsiwngtqsshj) |
| 14:42.58 | *** mode/#bzflag [+v JeffM] by ChanServ |
| 14:58.11 | *** join/#bzflag Juesto (~AndC@186.143.198.248) |
| 14:58.28 | *** join/#bzflag Juesto (~AndC@unaffiliated/juest) |
| 15:43.41 | *** join/#bzflag spldart (~spldart@bzflag/contributor/spldart) |
| 15:43.41 | *** mode/#bzflag [+v spldart] by ChanServ |
| 16:37.03 | *** join/#bzflag Juesto (~AndC@190.194.204.95) |
| 16:37.13 | *** join/#bzflag Juesto (~AndC@unaffiliated/juest) |
| 16:38.28 | Juesto | what's the expected value when the env for dependencies isn't used? |
| 16:49.33 | blast007 | sorry, I'm not understanding the question |
| 16:58.51 | *** join/#bzflag Juesto (~AndC@rev190-105-167-003.pccp.net.ar) |
| 16:58.51 | *** join/#bzflag Juesto (~AndC@unaffiliated/juest) |
| 17:05.31 | JeffM | 4 |
| 17:10.46 | blast007 | Juesto: can you rephrase the question? I'm not understanding what you're asking. |
| 17:11.18 | Juesto | what happens when bz_deps isn't there |
| 17:11.23 | JeffM | it won't build |
| 17:12.08 | JeffM | you'll get errors about a missing envrironment variable, and missing headers, and libs |
| 17:12.37 | Juesto | ah right i remember a bit |
| 17:13.54 | JeffM | the dependenices are easy to setup |
| 17:14.09 | Juesto | i used to get more errors before lol |
| 17:15.57 | JeffM | I posted a video how to set it up, it's very very easy |
| 17:17.52 | *** join/#bzflag Juesto (~AndC@rev190-105-167-003.pccp.net.ar) |
| 17:17.52 | *** join/#bzflag Juesto (~AndC@unaffiliated/juest) |
| 17:31.59 | JeffM | Flash: ZeroMQ is not suitable for gaming, it's too rigid, hides connect/disconnect, and assumes you trust both ends (since it seems to be designed for server to server messaging) |
| 17:32.20 | JeffM | it's an interesting idea and a cool thing for server to server backend stuff. |
| 17:32.32 | JeffM | but a game needs to know when someone disconnects, to you know.. remove the tank |
| 17:32.52 | *** join/#bzflag Juesto (~AndC@rev190-105-167-003.pccp.net.ar) |
| 17:32.52 | *** join/#bzflag Juesto (~AndC@unaffiliated/juest) |
| 17:33.45 | JeffM | it's best suited to atomic transactions |
| 17:35.04 | Juesto | hm? |
| 17:35.12 | JeffM | that was not for you |
| 17:35.20 | JeffM | stay connected ;) |
| 17:35.31 | JeffM | for you, just install the deps and build |
| 17:35.36 | Juesto | sorry, wifi switching too much or something |
| 17:35.44 | Juesto | updating vs lol |
| 17:36.06 | JeffM | umm yeah sure ... lol? |
| 17:37.26 | JeffM | oh sweet, someone forked raknet and kept it going |
| 17:37.53 | blast007 | there were lots of forks. is there a viable one? |
| 17:38.00 | JeffM | https://www.slikesoft.com/?page_id=1224&lang=en |
| 17:38.05 | JeffM | looks decent |
| 17:38.14 | JeffM | last commit was 6 days ago |
| 17:38.15 | Juesto | visual studio gg |
| 17:38.18 | Juesto | meh |
| 17:38.28 | JeffM | it's the fork with the most changes. |
| 17:39.26 | JeffM | they added unicode support |
| 17:39.37 | JeffM | and have been extending the docs too |
| 17:39.57 | JeffM | may be worth investigating |
| 17:40.47 | blast007 | cool |
| 17:41.27 | blast007 | certainly has an actual presense online. Forums, IRC, discord, etc. |
| 17:41.36 | JeffM | yeah |
| 17:41.41 | JeffM | it at least looks the part :) |
| 18:42.30 | *** join/#bzflag Juest- (~AndC@rev190-105-167-003.pccp.net.ar) |
| 18:42.39 | *** join/#bzflag Juest- (~AndC@unaffiliated/juest) |
| 18:47.15 | *** join/#bzflag Juesto (~AndC@rev190-105-167-003.pccp.net.ar) |
| 18:47.15 | *** join/#bzflag Juesto (~AndC@unaffiliated/juest) |
| 18:53.47 | *** join/#bzflag Juesto (~AndC@190.194.204.95) |
| 18:53.47 | *** join/#bzflag Juesto (~AndC@unaffiliated/juest) |
| 19:17.02 | Juesto | before trying to compile should i compile the deps right! |
| 19:17.04 | Juesto | ? |
| 19:17.35 | Juesto | Can I merge the dependencies with the main source? |
| 19:17.38 | blast007 | yeah, or grab the binaries from the releases page on the github repo |
| 19:18.37 | Juesto | shouldn't the batch file for building have vs edition detection? |
| 19:20.51 | blast007 | for the one version I support? |
| 19:21.01 | blast007 | oh, edition |
| 19:21.44 | blast007 | it could, but meh |
| 19:21.57 | blast007 | do you have something other than the Community Edition? |
| 19:23.13 | Juesto | yea |
| 19:23.27 | Juesto | there's scripts to detect the installed edition |
| 19:23.31 | blast007 | edit the file or use the bins |
| 19:23.45 | blast007 | batch scripts? |
| 19:23.54 | Juesto | I'm manually integrating the deps into the main |
| 19:23.58 | Juesto | yeah |
| 19:24.02 | Juesto | via batch |
| 19:24.09 | Juesto | querying registry, etc |
| 19:24.36 | blast007 | why are you making more work for yourself? |
| 19:24.57 | blast007 | https://github.com/BZFlag-Dev/bzflag-dependencies/releases |
| 19:25.07 | Juesto | vsinstaller has vswherw |
| 19:25.09 | blast007 | there's the binaries I made for 2.4.14 |
| 19:25.11 | Juesto | vswhere* |
| 19:25.27 | blast007 | download it, extract it, set BZ_DEPS |
| 19:26.04 | Juesto | i prefer having everything together in ide |
| 19:29.56 | Juesto | somewhat I'm aiming for a better managed thing |
| 19:34.18 | JeffM | you don't see the deps in the IDE |
| 19:34.21 | JeffM | just use the bins |
| 19:34.34 | JeffM | even if you build them they don't show in the IDE at the same time as bz |
| 19:34.43 | JeffM | the bins are WAAAAAAAAAAYYY simpler |
| 19:35.02 | blast007 | I think he's actually adding them to the solution |
| 19:35.10 | JeffM | oh god, don't do that |
| 19:35.21 | JeffM | then you'll have to set different dependency paths |
| 19:35.36 | JeffM | the environment var is build around them all being bins in the dir |
| 19:35.43 | JeffM | they are not project references. |
| 19:36.19 | JeffM | it'll probalby screw up the build order too |
| 19:37.08 | JeffM | get it workign the normal way first, THEN change it after you sure it's not busted ;) |
| 19:37.52 | JeffM | "you need new tires, let me take the hood off and take the engine out, I like to have all the parts in once place" |
| 19:38.28 | blast007 | it's just silly because that's how we *used* to have it and then we fixed it :) |
| 19:38.36 | JeffM | yeah it was a mess |
| 19:38.40 | JeffM | updating was a huge pain |
| 19:38.45 | JeffM | and it was a pain to get it working |
| 19:38.58 | JeffM | I did a lot of fiddling to get it to work the old way |
| 19:39.05 | JeffM | the environ made it much simpler |
| 19:41.10 | JeffM | I really think there should be a null check in msgAlive for the return value from findPlayer |
| 19:41.22 | JeffM | there is no valid spawn case where the spawning player is null |
| 19:42.12 | Juesto | it potentially could be improved |
| 19:42.38 | JeffM | a null check is just basic common sense |
| 19:42.38 | Juesto | by making the projects inside, such as bzf etc run the dep compile before compiling |
| 19:42.49 | JeffM | why? |
| 19:42.56 | blast007 | please god no |
| 19:43.57 | blast007 | I don't want to build the dependencies for every single clone of our repo |
| 19:44.08 | blast007 | I don't want the dependencies cluttering our our source releases |
| 19:44.29 | blast007 | stop wasting your time |
| 19:47.53 | JeffM | and ours |
| 19:54.01 | Juesto | found the current vsvc++ projects/solutions for most of the deps |
| 19:55.44 | JeffM | are you going to edit all the bzflag projects to point to them? |
| 19:59.26 | Flash | JeffM: good point about ZMQ and connect/disconnect |
| 19:59.51 | JeffM | for what it does it's kinda cool |
| 19:59.58 | JeffM | I may use it for some other stuff |
| 20:00.11 | Flash | I'm just starting to play with it at work |
| 20:00.19 | Flash | I need a data bus |
| 20:00.26 | JeffM | yeah that's what it's for |
| 20:00.31 | JeffM | not an end user connection |
| 20:00.31 | Flash | and most of the "new" things are java-centric/only |
| 20:00.51 | JeffM | there is a C# native implementation of it too |
| 20:01.10 | Flash | native? didn't know that |
| 20:01.19 | Flash | I thought most of the bindings sat on the C layer |
| 20:01.24 | JeffM | yeah someone just rewrote it in C# :) |
| 20:01.38 | JeffM | there is one that binds to C, and one where they just redid it in C# |
| 20:01.55 | JeffM | sockets are already pretty easy in C# so it probalby wasn't hard |
| 20:02.10 | JeffM | TCPListener and TCPClient take care of most of the BS |
| 20:02.28 | Flash | I had a typo in my "hello world" program and client and server weren't on the same port |
| 20:02.32 | Flash | but no indication |
| 20:02.56 | Flash | so I fired up wireshark and I'm like "what the heck is going on with the TCP connection protocol?" |
| 20:02.57 | JeffM | BZFSPro uses TCPListener and it works pretty well. |
| 20:03.10 | JeffM | adding IPV6 support was like 6 lines of code. |
| 20:03.41 | Flash | the client just kept trying to connect to the wrong port, with absolutely no complaint |
| 20:03.57 | JeffM | yeah it didn't seem to have a lot of error reporting |
| 20:04.10 | Flash | as you say, designed for trust |
| 20:04.11 | JeffM | little too much "Don't worry about it" for gaming :) |
| 20:04.18 | JeffM | soo much so |
| 20:04.42 | JeffM | since all it's methods are atomic, they are effetivly connectionless |
| 20:04.53 | JeffM | so any connection specific info has to be sent as part of the payload |
| 20:05.01 | JeffM | any anyone can fake a payload |
| 20:05.12 | JeffM | "sure I'm player ID 12... I just died" |
| 20:06.34 | JeffM | it also looked like each socket was bound to a single paradim, request reply, pub etc.. |
| 20:06.46 | Flash | yes |
| 20:06.47 | JeffM | for BZ you need to mix that up |
| 20:06.59 | Flash | but ZMQ sockets are not necessarily TCP/UDP sockets |
| 20:07.04 | JeffM | some stuff is request reply from the client, some is request reply from the server, some is just pub |
| 20:07.13 | Flash | an unfortunate overload of terms |
| 20:07.21 | JeffM | can you bind 3 ZMQ sockets to the same UDP socket? |
| 20:07.32 | Flash | I don't know. Haven't gotten that deep |
| 20:07.41 | Flash | the paradigm is that you don't (mostly) care |
| 20:07.44 | JeffM | also there was no real way to flag some transactions as loosable and some as not. |
| 20:07.50 | Flash | obviously firewall stuff applies |
| 20:07.51 | JeffM | player update pubs can be lost |
| 20:07.55 | JeffM | logins can't |
| 20:08.05 | Flash | login would have to be a request |
| 20:08.10 | Flash | update is a pub |
| 20:08.29 | JeffM | yes but how can I say send both over UDP, and it's ok if the pub fails, but the request can't |
| 20:08.41 | Flash | I haven't thought through the semantics of application, just learning it |
| 20:08.51 | JeffM | I checked into it, you can't really |
| 20:08.53 | Flash | however, UDP by definition must be "loseable" |
| 20:08.55 | JeffM | or at least not in a clean way |
| 20:09.12 | Flash | so it is inappropriate for login |
| 20:09.13 | JeffM | if you are relying SOLELY on UDP for transpo verification |
| 20:09.30 | JeffM | the entire point of a gaming network lib is to give you more services over what the transport layer gives you |
| 20:09.49 | JeffM | raknet works entirely over UDP, but still has message priority and stuff |
| 20:10.00 | JeffM | so you can say "this message must get there" and "this message can fail" |
| 20:10.24 | Flash | so it implements UDP retry as a protocol over the protocol |
| 20:11.04 | JeffM | yes, but only for the message types that need it |
| 20:11.15 | JeffM | and it's not as greedy as TCP |
| 20:11.33 | JeffM | it wont' try AS hard as TCP since it knows that it's a game, and if it fails they are probalby disconnected. |
| 20:11.49 | JeffM | SLikeNet really does look like it would work prefectly well |
| 20:15.48 | JeffM | single port UDP, open source, cross plat, Object Replication, RPC, etc.. |
| 20:16.10 | JeffM | hell it even does voice communications |
| 20:16.21 | JeffM | and matchmaking |
| 20:16.55 | JeffM | I believe it also has lag simulation for testing. |
| 20:23.02 | Juesto | i just optimized the build procedure somewhat |
| 20:23.31 | Juesto | detection to be precise |
| 20:25.46 | JeffM | whut? |
| 20:39.46 | JeffM | fullbuild only takes 1:33 for me, not sure why anything needs to be optimized. |
| 20:40.09 | Juesto | uh |
| 20:40.55 | Juesto | i just fixed the necessity to edit the bat to change vs editions |
| 20:41.25 | JeffM | you are not clear about what you are talking about then |
| 20:42.43 | Juesto | sorry |
| 21:10.08 | *** join/#bzflag Juest- (~AndC@rev190-105-167-003.pccp.net.ar) |
| 21:10.08 | *** join/#bzflag Juest- (~AndC@unaffiliated/juest) |
| 21:42.22 | *** join/#bzflag Juesto (~AndC@190.194.204.95) |
| 21:42.22 | *** join/#bzflag Juesto (~AndC@unaffiliated/juest) |
| 21:49.12 | Juesto | so i just "temp fixed" the warning for observer tank and such textures by copying from rabbit |
| 21:49.46 | Juesto | i had to copy the pdbs manually |
| 21:50.42 | blast007 | copy them where? |
| 22:03.35 | JeffM | whut? |
| 22:03.43 | Juesto | okay there we go, it's crashing at return id in player::getid |
| 22:03.52 | Juesto | player.h 361 |
| 22:04.17 | JeffM | who's the caller of that? |
| 22:04.21 | Juesto | blast007: i copied the pdbs from the respective outputs |
| 22:04.32 | Juesto | handle servermessage |
| 22:04.43 | JeffM | what is the current message? |
| 22:04.49 | Juesto | playing.cxx 2831 |
| 22:05.27 | Juesto | again, it's the collision between spawning and establishing a connection |
| 22:05.37 | JeffM | that's in msgMsg |
| 22:05.43 | JeffM | that's chat |
| 22:06.12 | Juesto | playing.cxx 3165 is the next one in the stacktrace |
| 22:06.24 | Juesto | and yes I'm on Windows, it's easier to debug |
| 22:06.37 | JeffM | sure that's the main message pump |
| 22:06.59 | JeffM | but 2831 is when processing MsgMsg |
| 22:07.02 | JeffM | aka a chat message |
| 22:07.05 | Juesto | the one to blame is player.h 361 |
| 22:07.23 | Juesto | the downloading is parsed as a chat message? |
| 22:07.40 | blast007 | why would it be crashing there? we null check srcPlayer |
| 22:07.44 | Juesto | How do i get a console window of the game in Windows m |
| 22:07.46 | Juesto | ? |
| 22:07.51 | JeffM | yeah the.h isn't at fault |
| 22:07.54 | JeffM | that retursn a number |
| 22:07.58 | JeffM | it's the caller |
| 22:08.14 | blast007 | the Windows client doesn't write to stdout |
| 22:08.15 | JeffM | that's like saying 12/0 crashes and it's 0's fault |
| 22:08.22 | Juesto | playing.cxx 6835 |
| 22:08.24 | blast007 | or at least, it doesn't write it out to a command prompt |
| 22:08.29 | Juesto | Oh well |
| 22:08.46 | Juesto | i would be seeing it on windbg if it did anyway |
| 22:09.04 | Juesto | would it be useful to have stdout? |
| 22:09.06 | JeffM | yeah that callstack doesn't make sense |
| 22:09.14 | blast007 | Juesto: is this a release build with debug symbols? |
| 22:09.34 | Juesto | debug with pdbs |
| 22:09.38 | Juesto | not release |
| 22:09.46 | JeffM | still does't make sense |
| 22:09.48 | Juesto | oh wait |
| 22:09.54 | JeffM | was scrPlayer null? |
| 22:10.00 | Juesto | yeah i did compile in debug |
| 22:10.29 | JeffM | put a breakpoint on 2831 and see |
| 22:10.33 | Juesto | 4th in the callstack is 7404 on playing.cxx |
| 22:10.48 | JeffM | also when doing normal debug you should not have to copy the PDBs, just set your debug target command correctly |
| 22:10.58 | blast007 | don't even think he's using the debugger in visual studio. he mentioned windbg. |
| 22:11.03 | JeffM | callstack doesn't mean anything, it's the data that is important |
| 22:11.03 | Juesto | so just without pdbs? |
| 22:11.17 | JeffM | debug automaticly builds the pdbs |
| 22:11.31 | Juesto | and the exception itself? |
| 22:11.32 | JeffM | just you can't debug the default target becuse it has no data folder |
| 22:11.39 | Juesto | the pdbs don't get copied |
| 22:11.43 | JeffM | no the actual state of the variables in each stack fame |
| 22:11.48 | JeffM | they don't NEED to be copied |
| 22:11.57 | JeffM | what sln did you build? |
| 22:12.04 | JeffM | fullbuild or just bzflag? |
| 22:12.09 | Juesto | i copied the data folder |
| 22:12.21 | blast007 | you don't need to copy the data folder... |
| 22:12.22 | JeffM | oh god |
| 22:12.29 | JeffM | sooo backwards |
| 22:12.34 | JeffM | what sln did you build? |
| 22:13.09 | Juesto | off fullbuild.sln, bzflag project with it's dependency tree |
| 22:13.21 | blast007 | https://github.com/BZFlag-Dev/bzflag/blob/2.4/README.WINDOWS#L32-L37 |
| 22:13.29 | JeffM | right click on the bzflag project and pick set as startup project |
| 22:13.39 | JeffM | then right click on bzflag again and go to properties |
| 22:13.46 | JeffM | pick the debug item from the list on the left |
| 22:14.07 | Juesto | blast007: that's outdated because i have bzf in a bin_Debug_Win32 folder |
| 22:14.10 | JeffM | set your debug command and working direction to the bin dir. |
| 22:14.12 | JeffM | https://usercontent.irccloud-cdn.com/file/7QJrK61k/image.png |
| 22:14.23 | JeffM | we know dude, we friggin wrote it |
| 22:14.36 | JeffM | it's super easy to setup |
| 22:14.52 | JeffM | then just debug right from visual studio, no copying anything |
| 22:15.00 | Juesto | isn't the same as windbg basically? |
| 22:15.06 | blast007 | Juesto: it doesn't make it wrong |
| 22:15.20 | blast007 | it's copied to a folder at the root leve of the source directory |
| 22:15.22 | JeffM | I don't know what windbg is, and I don't care what it is. |
| 22:15.24 | blast007 | level* |
| 22:15.44 | JeffM | set a friggin breakpoint on the call to getId and see what the object is |
| 22:15.46 | Juesto | i just prefer to use windbg over the hanging vs but i don't mind using it |
| 22:16.08 | JeffM | then tell us what the srcPlayerObject is before the call to getId |
| 22:16.11 | Juesto | fine, sorry for pissing ya |
| 22:16.12 | Juesto | ugh |
| 22:17.04 | JeffM | if return iid; crashes your system and srcPlayer is a valid object, your computer is broken |
| 22:17.45 | JeffM | if srcPlayer is null, and somehow made it past the srcPlayer == null test, your computer is broken. |
| 22:18.04 | JeffM | so ether srcPlayer is garbage or something else is going on. |
| 22:18.32 | Juesto | recalling I've mentioned a warning about player ids |
| 22:18.57 | JeffM | it's still anumber |
| 22:19.05 | JeffM | the wrong number won't crash |
| 22:19.18 | JeffM | not at that point |
| 22:20.34 | Juesto | wait hold on |
| 22:20.35 | JeffM | origText will also tell you what the chat message is, that may help deduce the state. |
| 22:23.18 | blast007 | if there's memory corruption the stack trace will be next to useless |
| 22:23.39 | JeffM | yeah |
| 22:23.56 | blast007 | in such a situation, something like valgrind might be more useful |
| 22:24.30 | Juesto | I'm trying to figure out how to get a function state in windbg, *checks manual* |
| 22:24.34 | JeffM | step through the code first. |
| 22:25.09 | Juesto | I'm trying to work in the crashed state |
| 22:25.16 | Juesto | wait, step in or out |
| 22:25.29 | Juesto | ? |
| 22:27.30 | JeffM | in goes in, out goes out |
| 22:27.43 | JeffM | I'd just set a breakpoint and look at the watches |
| 22:36.36 | Juesto | the last message seen on the connection status before gui backs out is connection established |
| 22:36.43 | Juesto | is that helpful? |
| 22:37.39 | blast007 | I compiled a list of devices that support Open GL 2.1 / 3.3 and OpenGL ES 2.0 / 3.0: https://gist.github.com/blast007/917f72b59fac372fb2f09cd2f555b94f |
| 22:38.55 | blast007 | going OpenGL 3.2 we'd lose first and second generation Intel Core (i3/i5/i7) integrated graphics support, at least on Windows (Mesa might fill in the gaps on Linux) |
| 22:39.16 | blast007 | (and any Raspberry Pi running the OpenGL driver) |
| 22:39.44 | blast007 | going with OpenGL ES 3.0 instead of 2.0 would mostly lose us Raspberry Pi support |
| 22:40.16 | blast007 | and maybe various android devices? that's going to be a big mix of support... |
| 22:40.41 | Juesto | definitely Android is going to have some weird touch controls |
| 22:40.50 | Juesto | or anything touch screen |
| 22:41.04 | Juesto | and without keyboard |
| 22:41.51 | blast007 | yeah, it would need some onscreen controls designed |
| 22:43.41 | Juesto | about origText, how's the thing? I've used -echo if it helps |
| 22:45.15 | Juesto | so yeah it's probabl crashing at playing.cxx 2831 |
| 22:45.35 | Juesto | I'll breakpoint that |
| 22:46.29 | Juesto | the comment says message is to me |
| 22:46.49 | Juesto | I'm not sure |
| 22:48.11 | Juesto | I'll give those breakpoints a try |
| 22:52.25 | Juesto | oh boy, windbg seems to require asm and program addresses knowledge |
| 22:52.54 | Juesto | I'm not even sure which thread it is |
| 22:53.20 | Juesto | I'm getting through |
| 22:53.55 | blast007 | use the visual studio debugger then. you just click before a line to set a breakpoint... |
| 22:54.12 | blast007 | you seem to like doing everything the hard way :) |
| 22:58.07 | Juesto | I'm trying to learn |
| 22:58.36 | Juesto | what's the symbol equivalency of srcPlayer and origText? |
| 23:04.17 | Juesto | okay im attaching |
| 23:04.31 | Juesto | with vs |
| 23:05.38 | blast007 | or just do what we said and point the debug paths.. |
| 23:06.38 | JeffM | bzflag is not a good place to learn how to program :) |
| 23:15.19 | blast007 | it's where I learned, and I turned out fi--okay, nevermind |
| 23:15.36 | Juesto | im stepping it and it's all inside xstring..... |
| 23:15.43 | JeffM | there are no strings on that line |
| 23:16.04 | Juesto | xutility |
| 23:16.21 | JeffM | what line in playing.cxx? |
| 23:16.23 | Juesto | i made it break on playing.cxx 2809 |
| 23:16.43 | JeffM | sure that's a std::string |
| 23:16.55 | JeffM | what is srcPlayer |
| 23:18.02 | Juesto | how to tell? nothing in Watchs |
| 23:18.34 | JeffM | you have to add it to watches |
| 23:18.39 | JeffM | or just hover over it |
| 23:18.50 | JeffM | it can't read your mind |
| 23:19.50 | Juesto | right click or what |
| 23:20.14 | JeffM | drag the item into the watches, or double click and type it in |
| 23:20.32 | blast007 | there's also a 'Locals' tab |
| 23:20.40 | JeffM | yeah it may be in there |
| 23:20.41 | JeffM | or autos |
| 23:20.46 | JeffM | but making a watch is simple |
| 23:22.10 | Juesto | which thing I'm looking for, again? |
| 23:22.18 | blast007 | srcPlayer |
| 23:22.25 | JeffM | you couldn't scroll up in chat? |
| 23:22.40 | Juesto | annoying how vs tries to load *all* symbols |
| 23:22.46 | Juesto | when windbg is on demand |
| 23:22.57 | JeffM | never been a problem for me |
| 23:23.05 | Juesto | huh |
| 23:23.08 | JeffM | you've never run it before, so it hasn't cached them |
| 23:23.12 | Juesto | well it's loading useless junk |
| 23:23.17 | JeffM | just chill |
| 23:25.45 | Juesto | okay |
| 23:25.51 | Juesto | it gave me an id |
| 23:25.52 | Juesto | but |
| 23:25.58 | Juesto | its just too weird |
| 23:26.10 | Juesto | what |
| 23:26.14 | JeffM | you should be able to open the ptr up and look at the structure |
| 23:26.15 | Juesto | says im admin? |
| 23:26.17 | JeffM | see if it looks ok |
| 23:26.38 | Juesto | how do i tell |
| 23:27.06 | JeffM | name looks ok, IDs are not big and stuff |
| 23:27.10 | JeffM | just sane data |
| 23:27.16 | Juesto | it seems to go nuts |
| 23:27.28 | JeffM | then it's probably a bad ptr |
| 23:27.30 | Juesto | as soon it hits player.h 361 |
| 23:27.53 | JeffM | does this only duplicate on this machine? |
| 23:28.00 | Juesto | Maybe joininternet game? |
| 23:28.15 | Juesto | i tried on the same machine but in two oses |
| 23:28.24 | Juesto | the str to crash is |
| 23:28.26 | JeffM | any other machines? |
| 23:28.37 | Juesto | try to reconnect as soon you read preparing world |
| 23:28.40 | JeffM | I'm wondering if the machine has bad ram |
| 23:29.06 | Juesto | i can try a fresh system and config |
| 23:29.13 | Juesto | borrowing my mom's laptop |
| 23:29.16 | JeffM | that won't fix your ram |
| 23:29.21 | JeffM | if course you are |
| 23:29.26 | Juesto | trying to reproduce |
| 23:29.42 | Juesto | I'll run a mem test |
| 23:29.53 | Juesto | it got 8gb of ram |
| 23:30.11 | JeffM | what server are you connecting to? |
| 23:30.13 | blast007 | that's how much RAM my router has |
| 23:30.59 | Juesto | ugh |
| 23:31.01 | JeffM | is it a super slow CPU? |
| 23:31.14 | blast007 | he was using an atom-based celeron, iirc |
| 23:31.21 | JeffM | oh wow |
| 23:31.53 | blast007 | though fwiw, I couldn't reproduce it on my Windows 8.1 atom mini-pc |
| 23:32.10 | JeffM | I can spam the connect button and have no issues |
| 23:32.23 | Juesto | exactly how i said? |
| 23:32.27 | JeffM | yep |
| 23:32.33 | blast007 | I've tried spamming it and waiting until various parts of the join |
| 23:33.02 | JeffM | if bzflag was multi-threaded I could see an issue |
| 23:33.02 | spldart | <PROTECTED> |
| 23:33.20 | JeffM | but nothing can change the player ptrs while it's processing a chat message |
| 23:33.24 | blast007 | spldart: yeah, it's using 3% right now, so probably a bit overkill, but it's what I had :) |
| 23:33.43 | JeffM | the join function would have exited by then |
| 23:33.51 | spldart | soab! That's alot for a router XD |
| 23:34.21 | JeffM | it's what video cards have these days, why not routers too? |
| 23:34.26 | blast007 | spldart: little quad-core 64-bit x86 AMD SoC |
| 23:34.55 | blast007 | (AMD's answer to Atom chips, essentially) |
| 23:35.15 | JeffM | I thought those were GEODES |
| 23:35.30 | blast007 | well, AMD's answer to *modern* Atoms |
| 23:35.43 | JeffM | ahh |
| 23:35.53 | blast007 | http://www.fit-pc.com/web/products/specifications/fitlet-models-specifications/?model[]=FITLET-GX-C67-FLAN-W |
| 23:35.58 | spldart | aren't 64bit and x86 mutually exclusive terms? Wait... Am I messing up a joke? |
| 23:36.02 | JeffM | I used to have an APU in a plex server , it was ok |
| 23:36.29 | JeffM | spldart: no there are other 64 bit archetectures, like 64bit ARM |
| 23:36.30 | blast007 | same idea, but that would have targeted a higher TDP likely |
| 23:37.03 | JeffM | AMD64 is the X86 compatible 64 bit extensions |
| 23:37.08 | spldart | I was aware of 64 arm. I just always thought of x86 as 32 bit. My bad :( |
| 23:37.15 | JeffM | there is ARM64 and IA64 |
| 23:37.43 | JeffM | 64 bit PC CPUs also include 32 bit x86 compatiblity |
| 23:37.50 | JeffM | that's why you can run bzflag ;) |
| 23:37.51 | blast007 | IA64 being Itanium |
| 23:37.54 | JeffM | yeah |
| 23:37.58 | spldart | x86 makes me think 286 through pentium 4 |
| 23:38.00 | JeffM | glad that thing died |
| 23:38.04 | Juesto | i have no idea about any memory corruption |
| 23:38.07 | JeffM | reminded me of pentium pro |
| 23:38.19 | JeffM | Juesto: try to dupe on another machine |
| 23:38.24 | Juesto | ok |
| 23:40.29 | spldart | Oh. Don't know if it even matters. I pulled my old install, the next install and the current install of debian. I have no idea why the short lived middle install would span bzflag between both monitors yet the one before and the one after did not. lsmod shows no different in loaded modules or anything. Same version of Debian. Damndest thing :??? |
| 23:40.47 | JeffM | different driver |
| 23:40.58 | JeffM | it means the driver reported your desktop as a single screen |
| 23:41.06 | spldart | see: lsmod shows no difference. |
| 23:41.14 | JeffM | something clearly is |
| 23:41.15 | JeffM | ;) |
| 23:41.17 | blast007 | what GPU? |
| 23:41.23 | spldart | nvidia |
| 23:41.27 | spldart | nouveau |
| 23:41.32 | JeffM | it may just be an option |
| 23:41.34 | spldart | lsmod was identical |
| 23:41.38 | spldart | all were debian 9.4 |
| 23:41.43 | Juesto | I'll try the ram i had before, i changed it with a 8gb stick |
| 23:41.50 | spldart | problem gone now so I stopped fighting with it. |
| 23:42.09 | JeffM | I've never had good luck with the open source nvidia driver |
| 23:42.12 | JeffM | nv all the way |
| 23:42.27 | spldart | I'm really considering that jeff |
| 23:42.32 | spldart | any caveats? |
| 23:42.35 | blast007 | wasn't 'nv' the old 2D only driver? |
| 23:42.46 | JeffM | maybe, I don't know what the current one is called |
| 23:43.01 | JeffM | spldart: richard stalman will think you are a poopy head |
| 23:43.12 | blast007 | https://linux.die.net/man/4/nv |
| 23:43.13 | JeffM | but your system will work, so for me, a fine trade-off ;) |
| 23:43.14 | spldart | giggle |
| 23:43.49 | JeffM | ahh I guess it's called "nvidia" then the driver number |
| 23:44.20 | spldart | details details |
| 23:44.23 | JeffM | on ubuntu, there is a control panel that will just go install them and make things work |
| 23:44.48 | JeffM | https://wiki.debian.org/NvidiaGraphicsDrivers#NVIDIA_Proprietary_Driver |
| 23:44.53 | spldart | lspci says VGA compatible controller: NVIDIA Corporation G92 [GeForce 9800 GT] (rev a2) |
| 23:45.18 | spldart | caveats on switching from free to nvidia proprietary modules? |
| 23:45.22 | blast007 | at least it's not hybrid laptop graphics - those are terrible |
| 23:45.29 | JeffM | are you offended by closed source drivers made by the same people who made the hardware so they know the right thing to do? |
| 23:45.30 | spldart | YES!!!!!!!1 |
| 23:45.48 | spldart | Jeff... No I am not ;) |
| 23:45.52 | JeffM | or would you prefer hippies guessing at a black box? |
| 23:46.11 | blast007 | I have a laptop with hybrid intel/amd GPUs, and that's bad enough on Windows.. can't imagine it's any better in Linux |
| 23:46.11 | JeffM | then there is probalby no reason to not use propritary |
| 23:46.22 | JeffM | yeah linux would NOT like dual GPUs |
| 23:46.38 | blast007 | Windows barely likes it |
| 23:46.44 | JeffM | yeah, mine is like that too |
| 23:46.50 | JeffM | updating is "fun" |
| 23:46.59 | spldart | Blast> recently had my first expererience with that with a eee pc box |
| 23:47.02 | blast007 | it blacklisted Firefox from running on the Radeon, for instance |
| 23:47.09 | spldart | can't remembrder the model exactly |
| 23:47.20 | JeffM | one of these days I want to get ryzen laptop |
| 23:47.28 | spldart | twas an unpleasant bit.....ch....uhm |
| 23:47.32 | blast007 | one of these days I want to find a replacement mobo/headpipe for that laptop that's from the intel only model |
| 23:47.38 | blast007 | heatpipe* |
| 23:48.48 | spldart | linux was flat pissed that part of the gpu was on cpu and part was external on mobo. |
| 23:49.03 | spldart | I hacked as far as I could but it just wasn't worth it |
| 23:49.43 | spldart | Now I'm thinking of tinkering with a Nuc |
| 23:51.22 | spldart | My broh got further than me in Ubuntu but still not really running right |
