| 00:19.15 | *** join/#bzflag infobot (ibot@rikers.org) |
| 00:19.15 | *** 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 |
| 00:19.15 | *** mode/#bzflag [+v infobot] by ChanServ |
| 01:14.36 | allejo | so... I got a bit too eager and forgot how to tell time. kinda started shutting non-active servers. they'll be back after maintenance. for now, enjoy the ones left |
| 01:43.52 | Juesto | oh huh |
| 03:05.11 | allejo | maintenance time! hopefully I don't break anything |
| 03:31.54 | Juesto | ! |
| 04:01.48 | *** join/#bzflag BZnotify (BZnotify@gateway/service/github.com/x-vjjwxvxrpuqephwg) |
| 04:01.49 | BZnotify | [13bzflag] 15JeffM2501 opened pull request #132: Fix for octree crash after allman formatting changes (062.4...06octree_allam_format_fix) 02https://git.io/vhL4w |
| 04:01.49 | *** part/#bzflag BZnotify (BZnotify@gateway/service/github.com/x-vjjwxvxrpuqephwg) |
| 04:02.18 | JeffM | that should fix the crash issue after the formatting change |
| 04:05.58 | JeffM | and prevent further issues if astyle is run again |
| 04:06.09 | JeffM | just say no to macros with multiple lines.... |
| 04:06.22 | JeffM | write a function instead :) |
| 04:09.40 | Juesto | hm |
| 04:28.45 | *** join/#bzflag kierra (~quassel@00.alucard.sujevo.io) |
| 05:10.59 | *** join/#bzflag moriah (~quassel@2605:9880:0:5f2:225:90ff:fed7:86e6) |
| 05:52.10 | *** join/#bzflag kierra (~quassel@00.alucard.sujevo.io) |
| 09:08.21 | *** join/#bzflag kierra (~quassel@2605:9880:0:5f2:225:90ff:fed7:86e6) |
| 09:09.18 | *** join/#bzflag allejo (~allejo@bzflag/developer/allejo) |
| 09:09.18 | *** mode/#bzflag [+v allejo] by ChanServ |
| 09:16.11 | *** join/#bzflag bertman_ (~bert@unaffiliated/bertman) |
| 09:44.29 | allejo | so this all took longer than I wanted. I'll be restoring the rest of my servers later today when I wake up. mofo and league servers have been restored |
| 11:22.24 | *** join/#bzflag BZnotify (BZnotify@gateway/service/github.com/x-bnfjezfuogodioza) |
| 11:22.24 | BZnotify | [13bzflag] 15blast007 pushed 2 new commits to 062.4: 02https://git.io/vhLdE |
| 11:22.24 | BZnotify | 13bzflag/062.4 144b1cf68 15Jeffery Myers: convert inline macros into actual functions, so that astyle won't screw everything over. |
| 11:22.24 | BZnotify | 13bzflag/062.4 14f546f56 15blast007: Merge pull request #132 from JeffM2501/octree_allam_format_fix... |
| 11:22.24 | *** part/#bzflag BZnotify (BZnotify@gateway/service/github.com/x-bnfjezfuogodioza) |
| 11:31.18 | *** part/#bzflag Guest21686 (~quassel@2605:9880:0:5f2:225:90ff:fed7:86e6) |
| 11:32.36 | blast007 | as for that weird character issue we've had in Bundle.cxx. Looking at a hex editor at a pre-allman version of that file, I'm seeing 0xC285, which is not a valid unicode character. In Notepad++ it shows as NEL, which stands for NExt Line. It's some unicode new line character. |
| 11:33.07 | blast007 | there's also some weird thing where closing the client via F12, at least, is hanging some of the time |
| 11:33.50 | blast007 | from pausing in the debugger, it looks like it's getting stuck swapping the GL buffer |
| 11:34.09 | blast007 | (I paused, set a breakpoint after that call, and it never hits the breakpoint when I resume) |
| 11:45.52 | *** join/#bzflag kierra (~quassel@unaffiliated/kierra) |
| 12:04.07 | *** join/#bzflag spldart (~spldart@bzflag/contributor/spldart) |
| 12:04.07 | *** mode/#bzflag [+v spldart] by ChanServ |
| 12:08.45 | blast007 | only tested the hang on Windows so far, haven't tried on Linux. |
| 12:12.31 | spldart | holy smackeroni! |
| 13:06.11 | *** mode/#bzflag [-v Foo_man_choo] by ChanServ |
| 14:16.26 | JeffM | blast007: the thing to look for is macros with multiple lines |
| 14:24.28 | *** join/#bzflag Notify (~notify@104.225.5.10) |
| 14:24.47 | blast007 | will that just be ones that are in the form #define something(n) \ (and then stuff on one or more lines below that) ? |
| 14:25.15 | blast007 | if so, I see several with: git grep -e '#define .*(.*).*\\' |
| 14:33.32 | *** mode/#bzflag [+v spldart] by ChanServ |
| 15:06.24 | JeffM | yes |
| 15:06.29 | JeffM | the issue is with where they are called |
| 15:07.14 | JeffM | if they were called in a single line thing that used to have { } around it, and now that { } got optimized out, when the macro replacement happens, only the first line of the macro goes in the IF or else instead of all of it. |
| 15:07.57 | JeffM | this is why macro use should not be a substitute for inline functions :) |
| 15:08.28 | JeffM | the same error would have happended if a user had not known and removed the { } thinking it was not needed. |
| 15:12.26 | JeffM | blast007: what is the exit issue you see? do you have a callstack from the debugger? |
| 15:14.27 | JeffM | oh yeah it totaly throws an exception in the kenel or driver when it does that |
| 15:15.06 | blast007 | throws an exception when it hangs during F12? |
| 15:15.24 | JeffM | yeah |
| 15:15.38 | JeffM | setting up breakpoints in cmdQuit now |
| 15:16.25 | blast007 | in debug builds we do have a couple macros that replace glBegin and glEnd, defined in include/bzfgl.h |
| 15:17.59 | JeffM | that shoudl be fine nobody is going to do if(x) glBegin(); |
| 15:18.14 | blast007 | ah, ok |
| 15:18.45 | JeffM | it's not all multi line macros, its ones that would be used in a single line if/else/for/while, etc... |
| 15:19.11 | blast007 | Thanks, I understand what to look for now. :) |
| 15:19.43 | JeffM | WOW. that throws the exception right away. it doesn't even get a chance to exit the loop! |
| 15:20.04 | JeffM | so it's not in the cleanup, it's in the command processor or something |
| 15:20.38 | JeffM | I put a breakpoint after playing loop was done, and it never got hit |
| 15:21.22 | blast007 | yeah, it never left the loop. noticed the audio thread was still going, so I put a breakpoint after the loop. never got to the closeSound() call (or closeAudio, whatever it is) |
| 15:21.58 | blast007 | so I put a breakpoint inside the loop (after hitting F12) and it never hit it, so the game loop was stuck somewhere |
| 15:22.06 | Juesto | huh? |
| 15:22.39 | JeffM | I have a feeling it's in the windows keymaping stuff |
| 15:22.47 | JeffM | my guess is it doesn't happen on SDL platforms |
| 15:23.37 | JeffM | going to try to revert WinDisplay.cxx and see if that is the issue |
| 15:25.40 | blast007 | I don't think we use that |
| 15:25.49 | JeffM | you think it's SDL? |
| 15:25.53 | blast007 | should be using SDL2 stuff for everything except input and maybe audio |
| 15:26.00 | JeffM | because I put a breakpoint in the SDL keys and it never got hit |
| 15:26.23 | blast007 | oh, is WinDisplay used for input? ... |
| 15:26.24 | JeffM | yeah wind display is not compiled |
| 15:26.29 | JeffM | I think so |
| 15:26.34 | blast007 | heh |
| 15:26.39 | JeffM | or maybe not, it's not built |
| 15:27.26 | JeffM | no HAVE_SDL is true |
| 15:27.48 | blast007 | so maybe it was just joystick input that was left as non-SDL on Windows |
| 15:28.04 | JeffM | that would make sense, for force feedback |
| 15:28.58 | JeffM | yeah, I get an exception BEFORE the SDL event is triggered on F12 |
| 15:29.18 | JeffM | I'm going to debug the last release and see what that did |
| 15:29.30 | blast007 | is F12 some magic key in visual studio? though I think I still had it happen in a release build.. |
| 15:29.38 | JeffM | it should not be |
| 15:30.13 | blast007 | https://stackoverflow.com/questions/18997754/how-to-disable-f12-to-debug-application-in-visual-studio-2012#41130093 |
| 15:30.23 | JeffM | wow |
| 15:30.26 | JeffM | never knew that |
| 15:33.59 | JeffM | going to exit from the menu should trigger the same command right? |
| 15:36.35 | blast007 | should be very close, at least |
| 15:37.03 | JeffM | I get no issue in release pressing F12 |
| 15:37.22 | blast007 | let me try a release build again |
| 15:37.38 | JeffM | I get no issue in debug, pressing F12 when not attached to the debugger |
| 15:37.42 | blast007 | perhaps I was using a debug build from the start.. dunno |
| 15:38.09 | JeffM | I'm building an older version now and seeing if I get the same debugger F12 issue in that build |
| 15:41.00 | JeffM | This is what I did in release https://usercontent.irccloud-cdn.com/file/MbMmjHiB/BZFlag_exit.mp4 |
| 15:42.44 | JeffM | yeah I get the same F12 issue in the last release, so I don't think that part is related to the code chagnes |
| 15:43.43 | JeffM | outside the debugger everything seems fine for me with F12, both fullscreen and windowed |
| 15:53.05 | JeffM | and with that registry change, and a reboot, I no longer get an issue in the VC debugger. |
| 15:53.09 | JeffM | it quits just fine |
| 15:57.36 | *** mode/#bzflag [-v Foo_man_choo] by ChanServ |
| 16:33.36 | blast007 | maybe the behavior is F12 in a debugging session is just sporatic.. I can't get it to hang at all now |
| 16:33.45 | blast007 | s/is/of/ |
| 16:36.22 | blast007 | (though maybe it's because I'm using remote desktop) |
| 16:41.27 | JeffM | there could be some command latency in that case |
| 16:42.26 | JeffM | I'm glad it's not a general problem though |
| 16:42.36 | blast007 | I also wasn't seeing any exception, so maybe I was having a different issue |
| 16:42.40 | JeffM | maybe |
| 16:42.49 | JeffM | GL and remote desktop is odd :) |
| 16:42.51 | blast007 | I unfortunately also updated VirtualBox between now and then |
| 16:43.26 | blast007 | so maybe the update fixed something, because it *seemed* like it was getting stuck doing a GL buffer swap |
| 16:43.30 | JeffM | I've been doing testing connecting and quitting and I haven't seen issues in like 10-20 runs on my machine. |
| 16:43.51 | JeffM | I would like to see the list of multi-line macros and review them |
| 16:44.06 | JeffM | most of them can probably just be replaced with inline functions |
| 16:44.13 | JeffM | let the optimizer sort it out ;) |
| 16:45.21 | JeffM | once you apply the formatting to master, I'll make a PR that cleans up the begin/end stuff |
| 16:48.47 | blast007 | this should find (most? all?) of the multi-line macros: git grep -e '#define .*(.*).*\\' |
| 16:49.07 | JeffM | yeah I just did that |
| 16:49.13 | JeffM | it finds too much but that should help |
| 16:49.22 | JeffM | JESUS CHRIST!!!! |
| 16:49.26 | JeffM | https://www.irccloud.com/pastebin/SDlKeDYm/ |
| 16:49.48 | JeffM | a 64 line macro!!!! |
| 16:50.02 | blast007 | good god |
| 16:50.09 | JeffM | who hates functions! |
| 16:50.31 | JeffM | yeah I'll fix that in master when you apply the formatting.... |
| 16:51.10 | JeffM | all so some other code can look like this |
| 16:51.13 | blast007 | I will try to get that done tonight. Got VS2017 installed on my actual system so it won't be so horribly slow |
| 16:51.14 | JeffM | https://www.irccloud.com/pastebin/GRcVoA2r/ |
| 16:51.48 | JeffM | OPENMEDIA is another abomination |
| 16:53.00 | JeffM | there are only 9 real multi line macros left |
| 16:53.27 | JeffM | the other 21 that show up in the search are #defs that use a \ in a string |
| 16:53.33 | JeffM | #define ANSI_STR_FG_GREEN "\033[32m" |
| 16:53.43 | JeffM | BZ_CONNECT_HEADER "BZFLAG\r\n\r\n" |
| 16:53.48 | JeffM | those are fine |
| 16:54.35 | blast007 | I only had 12 results from that git grep here. maybe it behaves differently on Windows. |
| 16:54.40 | JeffM | probalby |
| 16:56.25 | blast007 | 14 from git bash on windows |
| 16:56.31 | blast007 | in 2.4 |
| 16:58.10 | JeffM | yeah my windows search just found more, and didn't search linuxDisplay |
| 16:58.47 | JeffM | really just render and openmedia sould be changed. |
| 16:59.56 | JeffM | MD5STEP doesn't even NEED a new line |
| 17:00.35 | JeffM | PNGTAG can be collapsed too |
| 17:00.59 | JeffM | though with libPNG it may not even be needed |
| 17:01.05 | blast007 | PNGTAG is gone in master |
| 17:01.10 | JeffM | good |
| 17:01.27 | JeffM | the RENDER one is just frigtening |
| 17:01.29 | blast007 | and MD5 should get killed anyway :) |
| 17:02.13 | JeffM | now I want to find who put that monstrosity in there |
| 17:05.50 | JeffM | yeah Chris did RENDER, its' been there since commit one |
| 17:11.35 | *** mode/#bzflag [+v spldart] by ChanServ |
| 17:20.21 | JeffM | blast007: and yes the menu does call the same quit command as the keyboard :) |
| 17:20.56 | blast007 | I wasn't 100% sure because I saw somewhere else that there was also a message sent, but maybe that was for handling of /quit |
| 17:21.20 | JeffM | I had a breakpoint set in the command and it got hit when I went to the menu |
| 17:22.30 | JeffM | does master need to still support SGIImage files? |
| 17:25.51 | blast007 | I removed SGIImage files in master |
| 17:29.01 | JeffM | ah cool |
| 17:29.09 | JeffM | then OPENMEDIA probalby isn't there anymore |
| 17:29.20 | JeffM | I have a change for RENDER and I've tested it |
| 17:30.07 | blast007 | OPENMEDIA in master is only used for PNG and WAV files |
| 17:32.05 | blast007 | I have been playing around with OpenAL as well. We'll have to convert some of our sounds to mono if we want them to be positional. And I'm a bit unclear on how their alGenBuffers works. Most of the examples I've seen only allocate a single buffer... |
| 17:32.58 | blast007 | I'll have to throw my code up somewhere when I get home. |
| 17:35.21 | blast007 | I think the general idea is to allocate enough sources (via glGenSources) to cover how many simultaneous sounds you want playing, and then you reuse them as they finish playing. |
| 17:36.20 | JeffM | yeah rolling buffers is a common thing |
| 17:36.31 | JeffM | some games let the user configure how many buffers |
| 17:36.32 | blast007 | maybe I only need a single buffer and I just need to store the audio data in some other variable, then load it into a buffer, assign it to a source, and play the source |
| 17:36.36 | JeffM | 16, 32, etcc. |
| 17:37.00 | JeffM | the issue comes up when you run out of your preset number, because allocating buffers is slow |
| 17:37.16 | JeffM | so you have to be able to kill lower prority sounds |
| 17:37.40 | JeffM | are there any wrapers around openAL |
| 17:37.47 | JeffM | like a "sound engine" that does all that for you? |
| 17:38.38 | JeffM | https://github.com/R4stl1n/cAudio |
| 17:39.25 | blast007 | I also saw this, but it doesn't have any tags/releases yet: https://github.com/kcat/alure |
| 17:39.46 | JeffM | at the very least use the ideas they have |
| 17:39.59 | blast007 | though alure *is* packaged in Debian |
| 17:40.02 | JeffM | https://www.ambiera.com/irrklang/features.html |
| 17:40.49 | JeffM | nevermind that's not open source |
| 17:41.07 | blast007 | alure is in the same user/organization that openal-soft is in |
| 17:41.15 | JeffM | that sounds like a good thing |
| 17:41.32 | JeffM | go as high level as you can |
| 17:41.53 | blast007 | we could add some sick module music with alure :P |
| 17:41.58 | blast007 | beep boop |
| 17:42.11 | JeffM | it worked for unreal |
| 17:42.24 | blast007 | yup |
| 17:56.17 | spldart | Holy crap! I didn't know Psensor could monitor gpu temp,graphics throuput, mem usage, pcie buss usage etc etc 8~O |
| 18:00.25 | *** join/#bzflag The_Noah (~The_Noah@cpe-97-99-253-228.tx.res.rr.com) |
| 18:09.55 | JeffM | server side GM shot tracking is going to require server side player state interpolation too |
| 18:42.18 | JeffM | once SSPs are done, you could have server side autopilot |
| 18:42.43 | JeffM | just add a flag that forces a client to respect player updates for MyTank |
| 20:13.58 | *** join/#bzflag DTRemenak (~DTRemenak@cpe-172-251-230-211.socal.res.rr.com) |
| 20:13.59 | *** join/#bzflag DTRemenak (~DTRemenak@about/essy/CrazyCoder/DTRemenak) |
| 20:13.59 | *** mode/#bzflag [+v DTRemenak] by ChanServ |
| 20:17.48 | *** join/#bzflag bertman (~bert@unaffiliated/bertman) |
| 21:30.33 | *** join/#bzflag Zehra (~Zehra@unaffiliated/zehra) |
| 22:03.20 | JeffM | how can you hide the hud? |
| 22:03.57 | JeffM | may have gotten bored https://usercontent.irccloud-cdn.com/file/7UhZpCiz/image.png |
| 22:04.41 | blast007 | /localset noGUI 1 |
| 22:04.52 | blast007 | (and then turn off the radar and scoreboard) |
| 22:05.20 | blast007 | I'm working on the merge right now |
| 22:07.06 | allejo | ohhhh so *you're* "SuperTestMachine" |
| 22:07.52 | JeffM | https://usercontent.irccloud-cdn.com/file/wnMZhque/image.png |
| 22:07.53 | JeffM | yeah |
| 22:08.01 | JeffM | each of those tanks is on the blue team |
| 22:08.07 | JeffM | with different skins |
| 22:08.17 | allejo | :O |
| 22:08.38 | JeffM | you pick the skin at join time |
| 22:08.40 | JeffM | https://usercontent.irccloud-cdn.com/file/AQGltwJk/image.png |
| 22:09.00 | JeffM | and once team selection is moved post join, I can have the server send a list of extended skins |
| 22:09.14 | allejo | that's awesome |
| 22:09.22 | JeffM | right now each team has 3 skins |
| 22:09.27 | JeffM | standard, hex, and digital |
| 22:09.56 | JeffM | gives people a small bit of customization |
| 22:10.02 | allejo | what about hats? :p |
| 22:10.14 | JeffM | it'd just be another ID |
| 22:10.18 | JeffM | but sure |
| 22:10.29 | JeffM | it'd just be another tank part |
| 22:10.33 | JeffM | eHat |
| 22:10.44 | allejo | hehe eHat |
| 22:10.52 | JeffM | in this, the server can override the skin choice too |
| 22:11.12 | JeffM | it's basicly another field like "team" |
| 22:11.19 | JeffM | it gets set in the same places team does |
| 22:12.21 | JeffM | it would be a very easy change for 2.6 to give the users some more options. |
| 22:12.43 | JeffM | adds 2 bytes to MsgEnter and MsgAddPlayer |
| 22:12.56 | JeffM | and 2 bytes to the future MsgSetTeam |
| 22:13.01 | allejo | 2 bytes? but think of the bandwidth! |
| 22:14.10 | blast007 | when we're done, we'll probably shave off around a hundred bytes from a typical join ;) |
| 22:14.44 | JeffM | https://usercontent.irccloud-cdn.com/file/aYIThyXV/image.png |
| 22:14.53 | blast007 | iirc, right now, we send fixed length strings of 32 and 128 characters for callsign and motto |
| 22:14.54 | JeffM | those are the skinds I currently have |
| 22:15.00 | allejo | those custom skins would make life easier for what i want to do to make things more color blind friendly |
| 22:15.40 | allejo | ooo me like |
| 22:15.48 | allejo | did you make those or find them? |
| 22:15.58 | JeffM | if you can come up with a set of colorbind skins, I can easaly add a local option to force it to always load those for a player that is CB |
| 22:16.03 | JeffM | they are public domain |
| 22:16.10 | JeffM | the digital one is from the us goverment :) |
| 22:16.29 | allejo | oh nice |
| 22:16.29 | JeffM | the hex one I had found a while ago, it's what I used for NuTank |
| 22:16.49 | allejo | yea I have a copy of that one somewhere with your model |
| 22:16.56 | blast007 | public domain is kinda iffy since some countries have no concept of public domain |
| 22:17.07 | blast007 | (though we do use PDCurses) |
| 22:17.13 | JeffM | it allows relicense |
| 22:17.16 | JeffM | so good enough :) |
| 22:17.36 | JeffM | or I can always draw some new ones |
| 22:17.48 | allejo | how does licensing work across countries? e.g. public domain |
| 22:17.59 | JeffM | it's up to the author to sue |
| 22:18.19 | JeffM | so if it's PD, then it was made PD in a country that supported PD, thus the author probably won't sue anyone |
| 22:18.42 | JeffM | I could do a solid color one too I guess |
| 22:18.48 | JeffM | that'd be pretty easy |
| 22:40.41 | JeffM | https://github.com/JeffM2501/bzflag/tree/skin_id |
| 22:40.50 | JeffM | ended up with 5 skins |
| 22:41.11 | JeffM | https://usercontent.irccloud-cdn.com/file/kGSjz1Ms/image.png |
| 22:50.07 | JeffM | allejo: who did you think the SuperTestMachine was? |
| 23:01.49 | JeffM | wow, that ended up looking like a creeper from minecraft https://usercontent.irccloud-cdn.com/file/KIfRV3b2/image.png |
| 23:06.48 | blast007 | *Self-destructing in 5 seconds* |
| 23:08.57 | JeffM | I'm fixing the treads too |
| 23:09.03 | JeffM | where is the max texture size applied? |
| 23:13.29 | JeffM | https://usercontent.irccloud-cdn.com/file/m2pXaRU3/image.png |
| 23:13.35 | JeffM | there, that's better |
| 23:13.56 | JeffM | should probably increase the resolution of that texture too |
| 23:36.57 | Juesto | nice |
| 23:43.01 | JeffM | We can tweak the colors once people play with it |
