06:18.13 | *** join/#bzflag ibot (~ibot@rikers.org) |
06:18.14 | *** topic/#bzflag is http://my.BZFlag.org || http://cia.vc/stats/project/BZFlag || http://my.BZFlag.org/w/Getting_Help || Channel Logs: http://ibot.rikers.org/%23bzflag/ || BZFlag 2.4.2 is in devlopment, please take a task || The PLAN http://my.bzflag.org/w/Development_RoadMap |
06:18.14 | *** mode/#bzflag [+o ibot] by ChanServ |
06:58.40 | *** join/#bzflag brlcad (~sean@BZ.BZFLAG.BZ) |
11:10.41 | *** join/#bzflag survery (~thomas@dslb-188-110-169-245.pools.arcor-ip.net) |
11:10.43 | *** part/#bzflag survery (~thomas@dslb-188-110-169-245.pools.arcor-ip.net) |
12:47.21 | *** join/#bzflag mdskpr_ (~mdskpr@108.25.139.196) |
13:11.05 | *** join/#bzflag kierra (~jolie@unaffiliated/kierra) |
13:35.37 | *** join/#bzflag spldart (~spldart2@c-98-201-137-215.hsd1.tx.comcast.net) |
13:35.37 | *** join/#bzflag spldart (~spldart2@bzflag/contributor/spldart) |
13:35.37 | *** mode/#bzflag [+v spldart] by ChanServ |
13:55.44 | *** join/#bzflag DarkCalff (DC@173.231.40.98) |
14:25.57 | *** join/#bzflag ts (~ts@bzflag/developer/ts) |
16:02.52 | *** join/#bzflag yosef (468a95fa@gateway/web/freenode/ip.70.138.149.250) |
16:11.21 | *** join/#bzflag Gabe_G23 (u557@gateway/web/irccloud.com/x-kbjahdicdvcocnhw) |
16:11.36 | *** join/#bzflag Gabe_G23 (u557@bzflag/player/GabrielG) |
16:41.29 | *** join/#bzflag JeffM (~JeffM@adsl-75-16-90-217.dsl.irvnca.sbcglobal.net) |
16:41.29 | *** join/#bzflag JeffM (~JeffM@unaffiliated/jeffm2501) |
16:41.29 | *** mode/#bzflag [+v JeffM] by ChanServ |
17:16.59 | *** join/#bzflag Erroneous (~DTRemenak@about/essy/CrazyCoder/DTRemenak) |
17:17.00 | *** mode/#bzflag [+v Erroneous] by ChanServ |
17:27.56 | *** join/#bzflag temporalD (~a_temp_di@bzflag/serverop/TemporalDistraction) |
17:29.14 | *** join/#bzflag cygal (~cygal@wtf.awesom.eu) |
17:32.14 | *** join/#bzflag randomparticle (~randompar@about/essy/snick/randomparticle) |
17:57.59 | *** join/#bzflag Anxuiz (~Anxuiz@unaffiliated/anxuiz) |
18:03.17 | *** join/#bzflag jh` (~janne.has@212-149-234-105.bb.dnainternet.fi) |
18:58.36 | *** join/#bzflag temporalD (~a_temp_di@bzflag/serverop/TemporalDistraction) |
19:13.37 | *** join/#bzflag blast007 (~blast007@bzflag/developer/Blast) |
19:13.37 | *** mode/#bzflag [+v blast007] by ChanServ |
19:21.06 | randomparticle | i notice in TimeKeeper.cxx there are some global static definitions such as: |
19:21.07 | randomparticle | static struct timeval lastTime = { 0, 0 }; |
19:21.39 | randomparticle | why is that used rather than a local static since the value is not accessed outside of the function? |
19:22.13 | JeffM | probalby just for memory savings |
19:22.23 | JeffM | since we create multiple timekeepers |
19:22.35 | randomparticle | ah, ok |
19:23.18 | randomparticle | another thing i've noticed: the bz time starts at zero. is that necessary? |
19:23.36 | randomparticle | if you didn't have to start from zero, you wouldn't have to calculate differences |
19:24.01 | JeffM | everyone who uses a timekeeper should be computing deltas ayway so I don't know if it matters |
19:24.14 | JeffM | I don't think we actualy care how long we have been runiing |
19:24.18 | randomparticle | would simplify the code and make it slightly more efficient |
19:24.30 | JeffM | I doubt it is a huge performance issue |
19:24.47 | randomparticle | true. but an accumulation of small performance issues eventually becomes a large performance issue |
19:24.56 | JeffM | not always |
19:25.03 | JeffM | it's not like we are hard up for CPU cycles |
19:25.08 | JeffM | we always have to wait on GL anyway |
19:25.42 | JeffM | have you checked to see if 2.4 has the clock fixes from 2.99? |
19:25.48 | JeffM | cus that may affect your issue |
19:25.57 | randomparticle | doesn't look at first glance as though it does |
19:26.07 | randomparticle | haven't investigated it fully though |
19:26.16 | JeffM | you should backport that then and see if it helps |
19:26.36 | randomparticle | what i've done is take out the time code and put it into a small C program i'm running constantly |
19:26.43 | randomparticle | if i get a big leap, it will print it out |
19:27.14 | randomparticle | but so far runningg 3 hours without anything weird happening |
19:27.15 | JeffM | that may not show you anything, as the issue may be that something else is blocking the timer and putting it into a state |
19:27.36 | JeffM | it could be say GL is hanging for a bit and making timekeeper wait longer then it is comfortable |
19:27.41 | JeffM | your app won't ever get into that state |
19:27.45 | randomparticle | yeah |
19:27.56 | JeffM | tearing things apart is not always the best way to fix them ;) |
19:28.16 | randomparticle | true. but i'm just eliminating another possibility |
19:28.26 | JeffM | and probably elimnating the cause |
19:28.36 | randomparticle | that's possible also |
19:28.58 | randomparticle | whatever it is that's happening, it's not happening very often |
19:29.09 | randomparticle | once every several thousand frames |
19:29.39 | randomparticle | is quite hard to debug because it means waiting around for a long time |
19:30.10 | *** join/#bzflag jh` (~janne.has@212-149-234-105.bb.dnainternet.fi) |
19:30.37 | JeffM | backport the changes, see if that helps |
19:30.48 | JeffM | if DT goes negative it could cause your issue |
19:31.48 | randomparticle | but it's massively positive |
19:31.49 | randomparticle | that's odd |
19:32.05 | JeffM | yeah kinda like an unsiigned wraparound? ;) |
19:32.22 | JeffM | hence why getting a negative in there could be bad |
19:32.22 | randomparticle | was mentioning that possibility to NTH yesterday |
19:32.35 | randomparticle | but i'm not so sure |
19:32.48 | randomparticle | i've had it happen twice within five minutes, or not at all for 2 hours |
19:32.56 | JeffM | DTRemenak also mentioned that to you when you started this, when he asked if the clock stuff got backported |
19:33.01 | randomparticle | the clock isn't going to wrap around twice in 5 mins i think |
19:33.02 | JeffM | it was to prevent just that case |
19:33.09 | JeffM | no but the DT may |
19:33.12 | JeffM | if it's small enough |
19:33.41 | JeffM | subtracting 1 from 0 in an unsigned byte makes it go to 255 |
19:33.59 | JeffM | that kind of wraparound, not the adding kind |
19:34.20 | JeffM | and some of the fixes were to make sure that the deltas never went below 0 |
19:34.40 | JeffM | so again, backport them and see how it goes |
20:01.00 | *** join/#bzflag kierra (~jolie@unaffiliated/kierra) |
21:03.46 | *** join/#bzflag Erroneous (~DTRemenak@about/essy/CrazyCoder/DTRemenak) |
21:03.46 | *** mode/#bzflag [+v Erroneous] by ChanServ |
21:08.21 | *** join/#bzflag I_Died_Once (~I_Died_On@unaffiliated/idiedonce/x-1828535) |
21:10.17 | *** join/#bzflag Propane_Tank (~Propane_T@host-115-98-2-96.midco.net) |
21:15.04 | *** join/#bzflag dcat (~dcat@c-98-249-63-154.hsd1.va.comcast.net) |
22:26.36 | *** join/#bzflag Gilly (~btw@about/essy/ilkimys/gilly) |
23:27.08 | *** join/#bzflag KTL (~KTL@213.219.154.3.adsl.dyn.edpnet.net) |
23:37.12 | *** join/#bzflag meeba (~lamer@c-98-245-155-89.hsd1.co.comcast.net) |
23:46.59 | *** join/#bzflag Anxuiz (~Anxuiz@unaffiliated/anxuiz) |