Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2017-12-25

00:09:14 Quit JdGordon (Ping timeout: 260 seconds)
00:16:05 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
00:31:47 Quit ZincAlloy (Quit: Leaving.)
00:39:08__builtinhmm, licensing actually might be a slight issue
00:40:27__builtinthe apogee code is all GPLv2 or later, which is fine
00:42:19 Quit JdGordon (Ping timeout: 264 seconds)
00:47:21__builtinbut the Build engine code is under a custom BUILDLIC.txt, which prohibits commerical distribution
00:53:03 Quit dys (Ping timeout: 248 seconds)
00:54:48 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
01:09:33***Saving seen data "./dancer.seen"
01:15:59jhMikeSI don't want dirty hack in the vuprintf
01:17:04__builtinah, yes
01:17:18__builtinI don't think it actually needs it, to be honest
01:18:18__builtinthat was more for quake, actually
01:18:19 Quit michaelni (Ping timeout: 264 seconds)
01:18:36jhMikeSOne secret: if you format a string and it doesn't understand the format, the formatter is output unchanged
01:19:28__builtinby the way, are you fine with everything being enabled in vuprintf?
01:19:37jhMikeSthat means after formatting, the %f s will be left there do with whatever you like. it won't read the parameters out either
01:19:56jhMikeSI don't care. I just didn't want to enable more that was being used.
01:21:06jhMikeSit probably won't add too much because it uses one formatter for all compatible types
01:24:03__builtinas for licensing, the Build engine source can remain under the original BUILDLIC, right?
01:24:47__builtinif someone wants to make a commerical fork, they'll just have to know that part is not GPL-compatible and remove it
01:26:12 Join dys [0] (
01:31:03jhMikeSI can't answer that one
01:31:15 Join michaelni [0] (
01:33:48__builtinfor now I'll leave it as-is until someone complains :)
01:38:51 Quit Moarc (Ping timeout: 256 seconds)
01:47:55 Quit jhMikeS (Ping timeout: 256 seconds)
01:48:05 Join Moarc [0] (
01:55:59 Quit pamaury (Ping timeout: 260 seconds)
03:03:03 Quit Moarc (Ping timeout: 252 seconds)
03:06:02 Join Moarc [0] (
03:09:34***Saving seen data "./dancer.seen"
04:35:45 Join Piece_Maker [0] (
04:38:30 Quit Acou_Bass (Ping timeout: 272 seconds)
04:38:30 Nick Piece_Maker is now known as Acou_Bass (
04:46:47 Join jhMikeS [0] (
04:47:15jhMikeS__builtin: does duke have a lot of floating point?
04:47:29__builtinvery little
04:47:46__builtinI think the engine is completely integer-based
04:48:18__builtinwell, almost
04:48:42__builtinthe sound code uses some floats, though
04:49:14__builtinthere's probably enough in there to slow it down a bit
04:50:54jhMikeS I wonder if it would help if it were compiled with softfp or hard would help for arm v6
04:51:28__builtinhave you tried it on the gigabeast?
04:51:44jhMikeSno, that's why i"m wondering
04:51:50jhMikeSI was going to though
04:52:12jhMikeSthe vfp should be enabled already
04:53:24jhMikeSwhere are the gcc opts set for that?
04:55:41jhMikeSthat will set it for the game itself and the SDL lib?
04:56:05__builtinyes, duke is built with SDLFLAGS as well
04:57:02__builtinour gcc doesn't support hard float, does it?
04:58:36jhMikeSI don't know. I don't know why the vfp was enabled (not by me) if it didn't
04:59:51jhMikeSI tried softfp which puts the arguments in normal registers not straight into fp registers
05:00:32__builtin-mfloat-abi=hard gives, "unimplemented"
05:02:10jhMikeSit got through with softfp
05:02:30__builtinyeah, but will that give any benefit?
05:03:06jhMikeSyeah, it's still hw floating point instructions but arguments are passed with the standard ABI (r0-r3, [sp])
05:03:55__builtininteresting, I'll test
05:04:32jhMikeSwhich output file should I disassemble to verify?
05:06:24__builtinhmm, running the output with softfp gives a freeze
05:07:06__builtinlooks like it is generating FPU instructions, though
05:09:37jhMikeSyep, they're there
05:09:38***Saving seen data "./dancer.seen"
05:10:02jhMikeSmaybe I should double check the setup
05:10:29__builtindoes it run on your gigabeat?
05:10:51jhMikeSI didn't know you had a beast
05:11:09__builtinI don't, only ipod6g
05:11:42jhMikeSdoes that have vfpu?
05:12:04*__builtin should probably check...
05:12:36jhMikeSbeast is working fine
05:13:44__builtinwhere is VFP support enabled for the beast?
05:15:07jhMikeSin it's crt0
05:15:13jhMikeSit's working fine
05:16:50__builtinthe VFP isn't mentioned in the datasheet
05:16:52jhMikeSline 286
05:17:05__builtin(for the ipod6g's CPU)
05:17:44jhMikeSoh, no idea about that one
05:18:18jhMikeSwell, it definitely works if there's an enabled vfp
05:19:06jhMikeSmaybe the same coprocessor instruction works across all?
05:19:17__builtinI imagine any speed benefit would be hard to see on the beast
05:19:31jhMikeSseems to be getting an FPS boost but can't say if it's that doing it. need a control test
05:19:43__builtinyou mean the code to enable the VFP?
05:20:20jhMikeSyeah, I don't know how on that SoC
05:20:50__builtinI don't think it even has any floating-point hardware
05:24:46jhMikeSit can definely be enabled for armv6 anyway
05:25:24__builtinyes, as long as the hardware exists, at least
05:26:10__builtinapparently the "F" in a core name signifies VFP presence
05:26:55__builtinthe beast has an arm1136jf-s core, while my ipod6g has an arm926ej-s
05:30:23__builtinjhMikeS: do you have any other device that can run duke?
05:31:26jhMikeSvfp is worth about 3 fps at the game start (just sit there and wait for FPS to settle)
05:31:51jhMikeSthere's also dips due to sounds that don't happen with the vfp enabled
05:32:28jhMikeSI haven't tried gigabeat-f but that has the same screen size as the beast and isn't nearly as powerful
05:33:04__builtinthe reason I'm interested is that someone has told me that it crashes on a device I don't have
05:33:15jhMikeShmmm actually vfp looks like it's getting 4-5 fps inscrease
05:35:41jhMikeSwhich one?
05:35:51__builtinthe Zen X-Fi
05:36:06jhMikeSI don't have that
05:36:29__builtinanything else that can run duke?
05:38:40jhMikeSthe 2nd most powerful thing I have right now would be gig-f, otherwise it's sandisk players, some PP and coldfire targets
05:39:34__builtinI'd be interested to see how it runs on PP or sandisk players with a small screen
05:39:41__builtincould you test on those?
05:40:14__builtinin the meantime I'll see why it won't build for m68k
05:41:59jhMikeSfor armv6 definely add SDLFLAGS += -mfloat-abi=softfp
05:41:59jhMikeSSDLFLAGS += -mfloat-abi=softfp
05:43:10jhMikeSI'll try gigabeat f then
05:43:47jhMikeSoh, yeah, sure
05:44:05jhMikeSdon't expect much though the screens are much smaller
05:44:12jhMikeSwhich seems to help a lot
05:44:30__builtinhonestly I'm expecting it to crash
05:46:02__builtincould I add that flag in tools/configure instead?
05:46:42jhMikeSmaybe but really it shouldn't be in the core. I think you brought FP libs into the core which bloated the hell out of it
05:49:16jhMikeSh10-20 gb, h10-6gb, e200-v1, clip-v1 (ha), fuze-v2 ?
05:49:52__builtinwhat's the question?
05:50:09jhMikeSany preference?
05:50:14__builtinit won't run on the clip
05:50:18__builtinno greylib support
05:51:20__builtintry either h10, e200v1, and fuzev2 in any order you prefer
05:54:14jhMikeSwhich files do I need again. I still have timidity, that goes in .rockbox, right?
05:55:12__builtinyou need duke3d/* and timidity/* both under .rockbox
05:55:27__builtinyou may want to delete duke3d.cfg when copying
05:56:50jhMikeSgigabeat f gives undefined instruction at 01a4ef70
05:57:11__builtincan you map that?
06:01:36jhMikeSwell, it's on the audiobuffer
06:03:55__builtinhmm, are the addresses in the ELF correct?
06:07:07 Quit TheSeven (Ping timeout: 265 seconds)
06:08:41*__builtin goes to bed
06:09:12jhMikeSgoodnight. I think somehow vfp instructions got made
06:11:15jhMikeSthrere's a vpush in initgroupfile
06:12:53jhMikeSsomehow the vfp thing got set even though I had #if ARM_ARCH >= 6 around it
06:17:09jhMikeSok it runs but audio is going way too fast
06:19:10jhMikeScrap, forgot makefile isn't preprocessed :p no wonder it did that
06:33:25jhMikeSdid a completely clean build to be sure. the sample rate setup is wrong
06:40:17 Quit JdGordon (Ping timeout: 248 seconds)
06:47:00 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
06:48:01 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
06:49:43jhMikeS__builtin (logs): e200 gets stuck on Checking sound inits
06:50:11jhMikeSsoft lockup (backlight still works with hold)
06:55:44 Quit JdGordon (Ping timeout: 252 seconds)
07:08:10 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
07:09:42***Saving seen data "./dancer.seen"
07:13:31 Quit JdGordon (Ping timeout: 268 seconds)
07:34:49 Join go2m [0] (
07:35:56 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
07:36:20 Part go2m
07:48:29 Quit JdGordon (Ping timeout: 272 seconds)
07:55:08 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
08:08:17 Quit JdGordon (Ping timeout: 248 seconds)
08:19:12 Join ender` [0] (
08:20:19 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
08:27:57 Quit JdGordon (Ping timeout: 240 seconds)
08:35:27 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
08:43:32 Quit JdGordon (Ping timeout: 252 seconds)
08:50:34 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
08:53:51 Join johnb6 [0] (
09:02:17 Quit this_is_a_nick (Read error: Connection reset by peer)
09:09:43***Saving seen data "./dancer.seen"
09:12:05 Quit johnb6 (Ping timeout: 240 seconds)
09:25:23 Join johnb6 [0] (
09:55:29 Quit johnb6 (Ping timeout: 248 seconds)
10:09:37 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:bda9:6d24:4436:b425)
10:09:49 Quit ZincAlloy (Client Quit)
10:10:30 Quit aphirst (Ping timeout: 268 seconds)
10:11:52 Join aphirst [0] (~aphirst@unaffiliated/aphirst)
10:44:44 Quit JdGordon (Ping timeout: 264 seconds)
10:51:30 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
11:02:40 Quit JdGordon (Ping timeout: 252 seconds)
11:09:44***Saving seen data "./dancer.seen"
11:09:46 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
11:51:15 Join johnb6 [0] (
11:51:36 Quit johnb6 (Client Quit)
12:10:55 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:b93c:b4eb:426c:e26)
12:11:05 Quit ZincAlloy (Client Quit)
12:20:10 Join b3z [0] (
12:21:21b3zHello I have a Sansa clip zip with Rockbox firm. with a broken screen. Its there any way to still use it remotely or in some useful way or should i treat it as trash?
12:32:14 Quit JdGordon (Ping timeout: 256 seconds)
12:38:20 Join ZincAlloy [0] (
12:38:44 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
12:40:41 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
12:42:46 Quit bzed (Ping timeout: 252 seconds)
12:44:34 Quit ZincAlloy (Quit: Leaving.)
12:44:49 Join bzed [0] (
13:04:31 Quit JdGordon (Ping timeout: 272 seconds)
13:09:45***Saving seen data "./dancer.seen"
13:27:05 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
13:28:32 Quit dys (Ping timeout: 264 seconds)
13:40:37 Quit pamaury (Ping timeout: 272 seconds)
13:41:21 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
13:43:27 Quit jhMikeS (Ping timeout: 240 seconds)
13:46:17 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:84e4:e073:abd2:d32b)
13:56:33 Join dys [0] (
14:01:05 Quit dys (Ping timeout: 260 seconds)
14:01:20 Quit JdGordon (Ping timeout: 248 seconds)
14:08:35 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
14:18:23 Quit JdGordon (Ping timeout: 248 seconds)
14:21:15 Quit ZincAlloy (Quit: Leaving.)
14:27:41 Join ZincAlloy [0] (~Adium@
14:30:44 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
14:36:16 Join johnb6 [0] (
14:36:59johnb6you can activate voice for menu entries and title and use it in "blind mode".
14:38:43 Quit ZincAlloy (Quit: Leaving.)
14:40:46 Quit akaWolf (Quit: leaving)
14:41:25 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
14:42:03 Quit johnb6 (Ping timeout: 272 seconds)
14:48:07 Quit akaWolf (Quit: leaving)
14:48:15 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
14:50:43 Quit akaWolf (Client Quit)
14:57:48 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
15:09:49***Saving seen data "./dancer.seen"
15:10:32 Quit sanchaez (Ping timeout: 264 seconds)
15:23:35 Join sanchaez [0] (~sanchaez@
15:25:50 Quit sanchaez (Read error: Connection reset by peer)
15:27:46 Quit b3z (Read error: Connection reset by peer)
15:27:57 Join b3z [0] (
15:28:31 Join sanchaez [0] (~sanchaez@
15:50:07 Join dys [0] (
15:53:29 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:2833:829d:6a9:8ce6)
15:53:44 Join this_is_a_nick [0] (
15:55:31 Join johnb6 [0] (
16:02:40 Quit ZincAlloy (Quit: Leaving.)
16:17:09 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:2833:829d:6a9:8ce6)
16:20:31 Quit JdGordon (Ping timeout: 248 seconds)
16:21:52__builtinb3z: we have several users who operate it without the screen
16:25:50__builtinjhMikeS (logs): from that it looks like it's crashing in SoundStartup()
16:27:31 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
16:35:24__builtinalso, for the Makefile, you'll have to do some shell ickery to compare values
16:44:53 Quit johnb6 (Quit: Nettalk6 -
16:56:40 Quit ZincAlloy (Quit: Leaving.)
17:09:50***Saving seen data "./dancer.seen"
17:34:27 Join Jinx [0] (~Jinx@unaffiliated/jinx)
17:38:34 Join johnb3 [0] (
17:48:26__builtinit seems that the MV_MixFP* functions in mv_mix.c are responsible for a large amount of the overhead of sound
17:58:19__builtinI'm going to attempt to rewrite them as fixed-point
18:23:20 Quit alexweissman (Remote host closed the connection)
18:24:32 Join alexweissman [0] (
18:31:26 Quit aphirst (Ping timeout: 240 seconds)
18:33:11 Join aphirst [0] (~aphirst@unaffiliated/aphirst)
19:04:34 Quit this_is_a_nick (Ping timeout: 272 seconds)
19:09:51***Saving seen data "./dancer.seen"
19:21:47 Quit JdGordon (Ping timeout: 268 seconds)
19:23:13 Quit alexweissman (Remote host closed the connection)
19:28:21 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
19:39:31 Quit b3z (Remote host closed the connection)
19:39:59 Quit johnb3 (Quit: Nettalk6 -
20:09:27 Join JannF [0] (
20:16:50 Quit JannF (Quit: Leaving.)
20:44:51 Join alexweissman [0] (
21:09:53***Saving seen data "./dancer.seen"
21:34:29 Join this_is_a_nick [0] (~amofiuhr_@
21:44:13 Join Soap [0] (~Soap@rockbox/staff/soap)
21:46:10 Quit Soap_ (Ping timeout: 252 seconds)
21:52:53 Join ZincAlloy [0] (~Adium@
21:58:43 Quit dys (Remote host closed the connection)
22:10:23 Quit TheSeven (Ping timeout: 265 seconds)
22:13:54 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
22:29:44 Quit ZincAlloy (Quit: Leaving.)
22:55:13 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:a0d2:828:e453:413)
23:08:58 Join jhMikeS [0] (
23:09:54***Saving seen data "./dancer.seen"
23:24:14 Join cela [0] (c24b1876@gateway/web/freenode/ip.
23:26:17 Join [Saint] [0] (cb64d42d@rockbox/staff/saint)
23:27:47celaDuke3d works well on Fuze+ nice work, tried on Sony NWZ-E384 stops with message "ATOMIC 1.5 ->CRC You should try to ge" text gets cut off, will test with other arm color devices and report back. :-)
23:29:18 Quit cela (Quit: Page closed)
23:47:40 Quit this_is_a_nick (Ping timeout: 272 seconds)
23:51:00 Join cela [0] (c24b1876@gateway/web/freenode/ip.
23:52:52celaDuke 3d runs well on Zen X-Fi, only issue it seem I cant find a button to get back to the game menu, only to restart the level, so you have to play Duke nuke'em forever!
23:52:57[Saint]OK, so, theme update:
23:53:35[Saint]To extend the range of support, I'm going to have to end up doing something I didn't want to do, and split the theme up into small/medium/large categories.
23:53:39 Quit cela (Client Quit)
23:54:39[Saint]The reason why I need to do this is a somewhat complex one, but I need to do it to bring the minally supported resolution down.
23:54:50[Saint]errr, *minimally.
23:56:29[Saint]Even though the assets are loaded and displayed dynamically, the theme still needs to survive being parsed by the skin engine on any given target.
23:57:00gevaerts__builtin: I seem to have unstuck the build system a few times recently by restarting one of my build clients. It's of course possible that the client is at fault, but my suspicion is that the server sometimes gets stuck but can wake up again if *something* happens, and disconnecting a client is something
23:57:42gevaertsSo maybe next time it's stuck you could try connecting a client, or even just connecting to the server port
23:58:44[Saint]If a theme is loaded with out of bounds values, it will be discarded. To minimise the chances of this happening I'll be splitting the theme up in to three smaller sub-themes of small, medium, and large, each supporting a range of four different font glyph sizes.

Previous day | Next day