| 00:11.13 | *** join/#bzflag TimRiker (~timr@bzflag/projectlead/TimRiker) |
| 00:11.13 | *** mode/#bzflag [+o TimRiker] by ChanServ |
| 00:43.05 | *** join/#bzflag spldart (~spldart2@bzflag/contributor/spldart) |
| 00:43.05 | *** mode/#bzflag [+v spldart] by ChanServ |
| 01:05.31 | brlcad | BulletCatcher: okay (re ftgl patch) |
| 01:36.51 | *** join/#bzflag TD-Linux (~wheeeeeee@about/essy/indecisive/TD-Linux) |
| 02:02.06 | *** join/#bzflag JeffM2501 (~JeffM@unaffiliated/jeffm2501) |
| 02:02.06 | *** mode/#bzflag [+v JeffM2501] by ChanServ |
| 02:03.25 | *** join/#bzflag R0b0t1 (~Enigma@unaffiliated/r0b0t1) |
| 02:04.23 | *** join/#bzflag McSpider (~McSpider@xplr-ts-v10-208-114-184-180.barrettxplore.com) |
| 02:12.53 | *** join/#bzflag Yassen (Yassen@unaffiliated/yassen) |
| 03:30.41 | *** join/#bzflag I_Died_Once (~I_Died_On@c-69-136-38-243.hsd1.ga.comcast.net) |
| 04:03.49 | *** join/#bzflag McSpider (~McSpider@xplr-ts-v10-208-114-184-180.barrettxplore.com) |
| 04:25.38 | *** join/#bzflag TimRiker (timr@bzflag/projectlead/TimRiker) |
| 04:25.38 | *** mode/#bzflag [+o TimRiker] by ChanServ |
| 07:18.18 | CIA-10 | BZFlag: 03bullet_catcher * r21193 10/trunk/bzflag/src/other/curl/ (18 files in 6 dirs): Add some missing files from upstream curl, omitting the ones that are generated during the build process as well as the tests and packages subdirectories. |
| 07:21.08 | CIA-10 | BZFlag: 03bullet_catcher * r21194 10/trunk/bzflag/src/other/curl/ (7 files in 5 dirs): |
| 07:21.08 | CIA-10 | BZFlag: Include BZFlag-specific curlbuildWin32.h in distributions. |
| 07:21.08 | CIA-10 | BZFlag: Make Subversion ignore the presence of built HTML and PDF files. |
| 07:21.08 | CIA-10 | BZFlag: Replace the use of the "roffit" program used for generating HTML man pages with "groff -Thtml -man" to eliminate another external dependency. |
| 07:21.08 | CIA-10 | BZFlag: Disable regeneration and cleanup of libcurl.vcproj due to missing dependencies in upstream distribution. |
| 07:21.09 | CIA-10 | BZFlag: Do not try to distribute Makefile.dist, as it is missing from the upstream distribution. |
| 07:21.10 | CIA-10 | BZFlag: Disable superfluous clean-local Makefile rule that causes a maintainer-clean error. |
| 07:45.56 | *** join/#bzflag ibot (~ibot@rikers.org) |
| 07:45.56 | *** topic/#bzflag is http://cia.vc/stats/project/BZFlag || http://my.BZFlag.org/w/Getting_Help || Channel Logs: http://ibot.rikers.org/%23bzflag/ || http://ohloh.net/p/bzflag || 2.0.16 is the current release, use it wisely |
| 07:45.56 | *** mode/#bzflag [+o ibot] by ChanServ |
| 07:46.08 | CIA-10 | BZFlag: 03trepan * r21195 10/trunk/bzflag/src/ (5 files in 2 dirs): |
| 07:46.08 | CIA-10 | BZFlag: * gutted FontSizer |
| 07:46.08 | CIA-10 | BZFlag: - replaced tiny, blah, blah, blah with: fontSizes (ex: "8 12 16 24 32 ...") |
| 07:46.08 | CIA-10 | BZFlag: - based fontSizes on the view height, and nothing else |
| 07:46.08 | CIA-10 | BZFlag: * the FontSizer::setMin() call does nothing |
| 07:46.09 | CIA-10 | BZFlag: * the FontSizer::resize() call just loads the bzdb variable's values |
| 07:46.16 | trepan | and now the screen will refresh before you've been to the bathroom and back |
| 07:49.11 | trepan | forget to mention, it now snaps to the fontSizes values; i think snapping was disabled before |
| 08:05.01 | cygal | ~trepan++ |
| 08:05.15 | cygal | I had noticed this class last summer, but I did not dare touching it |
| 08:05.24 | cygal | (well, did not really know what to do with it..) |
| 08:07.49 | trepan | it was poorly conceived, and poorly implemented |
| 08:08.54 | trepan | using ratio values in setMin() for different contexts really make no sense (ex: look for how "alertFontSize" is used in different parts of the code w.r.t. FontSizer::setMin()) |
| 08:10.15 | trepan | still thinks the whole font setup (ftgl and friends), should be replaced by freetype + a couple of source files :) |
| 08:58.59 | *** join/#bzflag bier|tp (~bier@p54A5431D.dip.t-dialin.net) |
| 09:22.20 | *** join/#bzflag dhouthoo (~dhouthoo@80.169.61.18) |
| 10:06.47 | *** join/#bzflag Gnurdux (~gnurdux@c-68-50-118-143.hsd1.md.comcast.net) |
| 10:11.48 | trepan | cygal: fwiw, i took another look; my changes probably need a rewrite as well (ceil, floor, and round style calls would be better -- but that would touch about 30 files) |
| 11:58.58 | *** join/#bzflag dhouthoo_ (~dhouthoo@80.169.61.18) |
| 12:00.04 | *** join/#bzflag delusional (~delusiona@unaffiliated/delusional) |
| 12:17.59 | *** join/#bzflag tupone (~alfredo@gentoo/developer/tupone) |
| 12:17.59 | *** mode/#bzflag [+v tupone] by ChanServ |
| 12:24.51 | CIA-10 | BZFlag: 03atupone * r21196 10/branches/v2_0branch/bzflag/src/common/PlayerState.cxx: typo, clear velocity[1] too |
| 12:56.42 | *** join/#bzflag delusional (~delusiona@unaffiliated/delusional) |
| 13:45.39 | *** join/#bzflag Upsetter (~Ups@89.246.207.84) |
| 14:13.15 | *** join/#bzflag dcat_ (~dirtycat@ip70-188-17-162.rn.hr.cox.net) |
| 14:25.21 | *** join/#bzflag dhouthoo (~dhouthoo@2002:54c5:7a99:0:9227:e4ff:feea:516d) |
| 14:27.30 | *** join/#bzflag dhouthoo (~dhouthoo@2002:54c5:7a99:0:9227:e4ff:feea:516d) |
| 14:28.38 | *** join/#bzflag dhouthoo (~dhouthoo@2002:54c5:7a99:0:9227:e4ff:feea:516d) |
| 14:29.52 | *** join/#bzflag dhouthoo (~dhouthoo@2002:54c5:7a99:0:9227:e4ff:feea:516d) |
| 14:31.00 | *** join/#bzflag dhouthoo (~dhouthoo@2002:54c5:7a99:0:9227:e4ff:feea:516d) |
| 14:32.20 | *** join/#bzflag dhouthoo (~dhouthoo@2002:54c5:7a99:0:9227:e4ff:feea:516d) |
| 14:33.22 | *** join/#bzflag dhouthoo (~dhouthoo@2002:54c5:7a99:0:9227:e4ff:feea:516d) |
| 14:33.51 | *** part/#bzflag dhouthoo (~dhouthoo@2002:54c5:7a99:0:9227:e4ff:feea:516d) |
| 14:34.33 | *** join/#bzflag dhouthoo (~dhouthoo@2002:54c5:7a99:0:9227:e4ff:feea:516d) |
| 14:41.22 | *** join/#bzflag temporalD (~temporal@bzflag/serverop/TemporalDistraction) |
| 14:50.57 | brlcad | so... it's better because it does less now. hm |
| 14:52.21 | brlcad | the performance on the scaling sucked, but "should" have only been a one-time cost to render glyphs .. unless someone kept wiping out cached glyphs (which would be bad regardless) |
| 14:53.36 | brlcad | horizontal scaling made the menus always show regardless of the screen ratio so you could play bzflag on a 24" display turned sideways, for example, without text running off the display |
| 14:56.28 | brlcad | when originally implemented, it worked for any ratio of window size within the bounds of the minimum and maximum font sizes |
| 14:57.48 | trepan | it's better because it doesn't chew up ridiculous amounts of time and texture memory |
| 14:58.23 | brlcad | it didn't do that originally either, everything was pretty much instantaneous (and heavily profiled) |
| 14:58.50 | trepan | well it sure did it before i changed it |
| 14:59.01 | brlcad | I don't doubt that |
| 14:59.56 | brlcad | it took a lot of work to clear out an absurd number of state resets before, I wouldn't be surprised if they're back |
| 15:02.04 | brlcad | how much texture memory was it using? |
| 15:23.29 | trepan | a number of megapixel count textures |
| 15:28.03 | *** join/#bzflag Wyk3d (~Wyk3d@cl-86-125-185-198.cablelink.mures.rdsnet.ro) |
| 15:32.58 | *** join/#bzflag Scrumpy (~Scrumpy@82.95.193.193) |
| 15:33.39 | trepan | r21197 should help with the <<24" display turned sideways>> use case |
| 15:35.31 | *** join/#bzflag Cobra_Slow (~coprah@static.108.245.40.188.clients.your-server.de) |
| 15:35.53 | *** join/#bzflag fahadsadah (fahad@unaffiliated/fahadsadah) |
| 15:37.30 | *** join/#bzflag essy (~essy@freenode/staff/sportchick) |
| 15:37.30 | *** mode/#bzflag [+v essy] by ChanServ |
| 15:37.38 | *** join/#bzflag TutTankAhmon1 (~Olivier@LAubervilliers-151-11-10-35.w193-251.abo.wanadoo.fr) |
| 16:10.37 | CIA-10 | BZFlag: 03trepan * r21197 10/trunk/bzflag/src/bzflag/FontSizer.cxx: * use the smaller viewport dimension for picking font sizes |
| 16:15.10 | *** join/#bzflag goingberserk (~abc@pD9E86977.dip.t-dialin.net) |
| 16:44.31 | brlcad | trepan: iirc, a similar hack was originally done in fontsizer, but it failed miserably for non-square fonts |
| 16:44.37 | *** join/#bzflag MindstormsKid (~msk@Wikipedia/MindstormsKid) |
| 16:46.52 | brlcad | not saying I care terribly about the feature |
| 16:47.32 | brlcad | but that was the main reason why that class existed in the first place, and it did that well -- the performance issues happened later and were in bz-land and freetype-land |
| 16:49.03 | brlcad | to say it was "poorly conceived, and poorly implemented" is more what I care to correct, mischaracterizing the problem and eliminating the feature as a solution |
| 16:52.20 | *** join/#bzflag blast007_bz (~blast@BZ.BZFLAG.BZ) |
| 16:55.02 | trepan | if the use of the ratiometric bzdb variables in concert with the FontSizer context sizes was the intended use of the class, then i would say it was poorly conceived |
| 16:55.48 | trepan | using 1/64 for 100% view height context vs using 1/64 for a 40% view height context makes no sense |
| 16:56.38 | trepan | as for the ftgl/freetype/whatever slowdown, does anyone know when that started happening? |
| 16:57.11 | trepan | i why ftgl's font atlases are so wasteful (i'm seeing a lot of empty space for certain fonts) |
| 16:57.17 | trepan | s/i why/and why/ |
| 16:57.41 | brlcad | the view height ratios were already there, it hooked into existing bzdb setup |
| 16:58.05 | trepan | using them for varying contexts doesn't really make any sense |
| 16:58.13 | trepan | varying context heights that is |
| 16:59.00 | brlcad | height ratios were to not have tiny fonts on super high-res displays |
| 16:59.12 | *** join/#bzflag goingberzerk (~abc@pD9E86977.dip.t-dialin.net) |
| 16:59.59 | trepan | but using the same ratios for different context heights does not make sense |
| 17:00.52 | trepan | i well understand using ratios to scale the fonts to the view height, the re-written FontSizer does the same thing |
| 17:00.54 | brlcad | what do you mean by different contexts? the bzdb vars? |
| 17:01.05 | trepan | FontSize(sizex, sizey) |
| 17:01.11 | trepan | *FontSizer |
| 17:02.37 | trepan | either way, unless the slowdown in the FTGL path is fixed, creating multiple large sized fonts just to test height/width is not reasonable |
| 17:03.06 | trepan | tweak the ratios down, and that'll be that for now ;) |
| 17:04.53 | brlcad | the use of multiple fontsizers for different context sizes isn't exactly a problem of the fontsizer .. it's job was to size for a given context, ideally one font sizer that matches your display size |
| 17:05.51 | trepan | "but using the same ratios for different context heights does not make sense" |
| 17:06.05 | trepan | not necessarily a FontSizer problem, but that is the way it was being used |
| 17:07.16 | brlcad | so it was adapted faultily, that happens a lot |
| 17:07.26 | brlcad | i'm sure a lot of the performance problems were the same cause |
| 17:07.48 | brlcad | the re-entry of context resets as a quick fix for gl crud is another one |
| 17:08.23 | trepan | who added that? |
| 17:08.40 | brlcad | I've seen that happen at least a half-dozen times by various people |
| 17:08.49 | brlcad | freetype can be absurdly slow at rendering glyphs, fwiw |
| 17:09.11 | brlcad | that's why there was advancement code to pre-render glyphs at start-up so it was at least a one-time cost |
| 17:09.37 | trepan | that's but part of the reason, texture transfers would be another |
| 17:09.54 | brlcad | also part of why it made sense to combine fonts into a fixed number of font sizes since character maps are rendered in sets based on size |
| 17:10.16 | trepan | even if you had pre-render image memory lying around on the CPU, you'd want to transfer it to GPU side textures for rendering, regardless of freetypes speed |
| 17:10.58 | trepan | and fixed sizes are to reduce the amount of memory used on the GPU |
| 17:11.19 | brlcad | memory on the GPU wasn't originally a problem at all |
| 17:11.24 | brlcad | it was render time |
| 17:12.10 | brlcad | I can't imagine that bz's calls into ftgl are using up texture memory unless something is just buggered somewhere in the pipeline |
| 17:12.39 | brlcad | do you have a profile? |
| 17:12.44 | trepan | or ftgl just doesn't handle texture atlases all that well |
| 17:14.03 | trepan | not lately |
| 17:14.44 | brlcad | could be .. but I'd be surprised -- the only time-sink I can think of is it calling through to freetype to precache kerning metrics for visible characters |
| 17:15.03 | brlcad | and that sped things up for bz big-time when it was added |
| 17:15.57 | *** join/#bzflag JeffM2501 (~JeffM@unaffiliated/jeffm2501) |
| 17:15.57 | *** mode/#bzflag [+v JeffM2501] by ChanServ |
| 17:16.19 | *** join/#bzflag JeffM_ (~JeffM@adsl-75-50-184-13.dsl.lsan03.sbcglobal.net) |
| 17:17.10 | brlcad | can set precomputeKerning = false in src/other/ftgl/src/FTFace.h to test if that's a factor |
| 17:29.06 | *** join/#bzflag xaver (~xaver@pD9ED695A.dip.t-dialin.net) |
| 17:42.37 | *** part/#bzflag Upsetter (~Ups@89.246.207.84) |
| 17:55.40 | *** join/#bzflag JeffM2501 (~JeffM@unaffiliated/jeffm2501) |
| 17:55.40 | *** mode/#bzflag [+v JeffM2501] by ChanServ |
| 17:58.18 | *** join/#bzflag AAA_awright_ (~a3@ip24-251-157-63.ph.ph.cox.net) |
| 18:39.21 | *** join/#bzflag Qew (~qew@unaffiliated/qew) |
| 18:55.46 | trepan | brlcad: http://trepan.bzflag.bz/FontSizer.log (from before the change) |
| 18:56.11 | trepan | note the dependence on the fontSize, would that be caused by precomputeKerning? |
| 18:58.01 | *** join/#bzflag Tupone (~Tupone@gentoo/developer/tupone) |
| 18:58.02 | *** mode/#bzflag [+v Tupone] by ChanServ |
| 19:02.46 | brlcad | trepan: hm, that's very interesting .. looking deeper |
| 19:05.15 | *** join/#bzflag goingberserk (~abc@pD9E86977.dip.t-dialin.net) |
| 19:07.11 | brlcad | yeah, I wouldn't think precomputeKerning is at play there .. if it were, subsequent calls for the same face, same size, should be drastically reduced |
| 19:07.45 | brlcad | the few I see repeated are within noise, with some even slower on the second call |
| 19:08.11 | *** join/#bzflag Pimpinella (~frank@gondolin.pimpi.org) |
| 19:20.31 | *** join/#bzflag goingberserk (~abc@pD9E86977.dip.t-dialin.net) |
| 19:36.32 | *** join/#bzflag Erroneous (~DTRemenak@about/essy/CrazyCoder/DTRemenak) |
| 19:36.32 | *** mode/#bzflag [+v Erroneous] by ChanServ |
| 20:03.24 | *** join/#bzflag goingberserk (~abc@pD9E86977.dip.t-dialin.net) |
| 20:04.13 | *** join/#bzflag TimRiker (~timr@bzflag/projectlead/TimRiker) |
| 20:04.13 | *** mode/#bzflag [+o TimRiker] by ChanServ |
| 20:10.33 | *** join/#bzflag nielsle (~nielsle@1503032406.dhcp.dbnet.dk) |
| 21:04.14 | *** join/#bzflag bambino (~bambino@36.255.223.67.in-addr.arpa) |
| 21:16.33 | *** join/#bzflag bier_ (~bier@p54A558D9.dip.t-dialin.net) |
| 21:17.22 | *** join/#bzflag bier|tp (~bier@p54A558D9.dip.t-dialin.net) |
| 21:17.50 | *** join/#bzflag bier (~bier@p54A558D9.dip.t-dialin.net) |
| 21:25.25 | *** join/#bzflag goingberserk (~abc@pD9E86977.dip.t-dialin.net) |
| 21:46.18 | *** join/#bzflag thefirstdude (~bla@dsl092-208-150.chi1.dsl.speakeasy.net) |
| 22:30.20 | *** join/#bzflag blast007_bz (~blast@bzflag/developer/Blast) |
| 22:30.20 | *** mode/#bzflag [+v blast007_bz] by ChanServ |
| 23:10.00 | *** join/#bzflag ab0oo (~johng@c-76-97-205-109.hsd1.ga.comcast.net) |
| 23:10.32 | ab0oo | has anyone had problems with no audio with Kubuntu Lucid and BZflag? |
| 23:10.51 | blast007 | do you have the correct SDL installed for your sound system? |
| 23:11.19 | *** join/#bzflag GvzEvxre (~timr@bzflag/projectlead/TimRiker) |
| 23:11.19 | *** mode/#bzflag [+o GvzEvxre] by ChanServ |
| 23:11.31 | ab0oo | good question. |
| 23:11.50 | blast007 | ex: if you use pulseaudio, there's libsdl1.2debian-pulseaudio |
| 23:12.09 | ab0oo | no, pulseaudio is a nonstarter with kubuntu |
| 23:12.14 | blast007 | k |
| 23:12.20 | blast007 | alsa? |
| 23:12.27 | ab0oo | preferred. |
| 23:12.34 | ab0oo | I'd prefer to use alsa |
| 23:12.57 | ab0oo | ahh, let me see here... |
| 23:13.02 | blast007 | the one up there I gave was just an example |
| 23:13.10 | blast007 | there's a -alsa one too |
| 23:13.35 | ab0oo | hmm, alsa is listed as "c" for conflicting. |
| 23:13.38 | *** join/#bzflag `ls` (~timr@bzflag/projectlead/TimRiker) |
| 23:13.38 | *** mode/#bzflag [+o `ls`] by ChanServ |
| 23:13.51 | ab0oo | yep... pulseaudio is the one configured |
| 23:14.28 | ab0oo | Couldn't open audio: No available audio device |
| 23:14.31 | blast007 | isn't audio on linux fun? ;) |
| 23:14.35 | ab0oo | I feel closer. |
| 23:15.02 | ab0oo | I'd been using linux since 1994, and I've never, not once, found the appropriate person to biatchsmack for the horrible audio support |
| 23:15.10 | blast007 | hehe |
| 23:15.58 | ab0oo | so, according to kde, I'm using gstreamer as my "audio backend" |
| 23:16.43 | blast007 | hmm |
| 23:17.01 | ab0oo | I have this installed: libsdl1.2debian-alsa |
| 23:17.14 | blast007 | might have to find out what gstreamer is piping to |
| 23:17.33 | ab0oo | alsaplayer works (i.e. I can play an audio file with alsaplayer) |
| 23:18.59 | ab0oo | well, when all else fails, delete your .bzf/2.0/config file |
| 23:19.19 | ab0oo | suddenly, audio works. my gameplay will be so much better when I can hear teleporters and missle locks. |
| 23:19.21 | blast007 | heh |
| 23:19.42 | ab0oo | thanks for the pointer to SDL. I'd have never found that. |
| 23:20.07 | ab0oo | replacing the pulseaudio with the alsa SDL driver was key, I believe |
| 23:20.30 | blast007 | that one is a pretty common issue |
| 23:20.39 | *** join/#bzflag remote (~remote@unaffiliated/remote) |
| 23:21.48 | *** join/#bzflag Think_Differentl (~js@bzflag/player/ThinkDifferent) |