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 2012-01-04

00:00:05funmanbluebrother^: i'm getting xcode 3.2.6
00:00:05 Join T44 [0] (
00:00:39thomasjfoxkugel: Don't we already up/down the volume levels in the software mixer?
00:00:59kugelbut depending on the device, touching hte master volume is the right thing (e.g. android)
00:01:03thomasjfoxkugel: At least the volume controls work with the latest build
00:01:19 Quit MethoS- (Remote host closed the connection)
00:01:28kugelthomasjfox: what do you mean?
00:01:37kugelthe latest build of what?
00:01:48thomasjfoxkugel: The current build of trunk
00:01:58thomasjfoxWith the new volume slider popup in cabbie
00:02:06kugelon maemo?
00:02:22kugelyou don't use alsa-lib there dont you?
00:02:36thomasjfoxBut I don't adjust the volume either
00:03:00kugelright, that's volume via dsp, what I mentioned before
00:03:06kugelthat works yes
00:03:08 Quit Topy (Ping timeout: 240 seconds)
00:03:46thomasjfoxanyhow, it's getting late, I want to give your patch a final run :)
00:03:57kugelbtw, I think now that there's a real mixer that should be moved from dsp to the mixer
00:05:19kugelhm, no it's actually HAVE_SW_VOLUME_CONTROL and you don't define that
00:05:33kugelTONE is bass/treble
00:06:48thomasjfoxI thought support was added along with the new software mixer
00:07:42kugelthe software mixer can adjust the volume of each channel, but not all together IIRC
00:08:19kugelbut the dsp also only applies to the music channel
00:09:05kugelit wasn't added the control the channel's volume via buttons
00:09:19thomasjfoxkugel: New version of the patch works out of the box on the N900. Trying a N8xx build now
00:09:36kugeljhMikeS knows more
00:09:46kugelthomasjfox: so, volume control didn't actually work before?
00:10:42thomasjfoxkugel: At least not the one inside rockbox IIRC. Never used it before, I always use the hardware buttons to controll the master volume
00:11:19kugeloh hmm
00:11:23kugelno idea
00:13:52thomasjfoxkugel: N8xx also seems to work
00:13:55 Quit ender1 (Quit: Computers work fine as long as users aren't allowed anywhere near them.)
00:14:31kugelthomasjfox: pcm-gstreamer implements volume control
00:14:32thomasjfoxWith your refactoring it will probably get easier to write a "driver" for libhal
00:15:10kugelsince the first version
00:15:17thomasjfoxkugel: I know. The first version of the maemo patch also made use of it
00:15:47thomasjfoxIt was just very annoying to adjust the volume inside rockbox + adjust the master volume on the phone
00:16:22kugelyou sounded like volume control isn't implemented on maemo
00:17:04thomasjfoxIt's not implemented on the maemo port of rockbox
00:17:10thomasjfoxThough that "changed" with the software mixer
00:17:20kugelit is implemented
00:17:26thomasjfoxAnd only with your recent cabbie changes it become intuitive to use
00:17:47kugel"volume inside rockbox" is what I mean
00:18:18thomasjfoxanyhow, my goal for today was to commit the SDL optimizations patch and test your changes
00:18:27thomasjfoxBoth goals are accomplished :)
00:18:34kugelheh, nice. and thank you again
00:18:44kugelI guess I can commit it then
00:19:11*thomasjfox bets there will be red :o)
00:20:05kugelthomasjfox: you're saying since the software mixer rockbox controls the master volume?
00:20:14kugelI'm still confused
00:20:26thomasjfoxkugel: No, the master volume stays unchanged
00:20:40kugelwhat changed then?
00:20:40thomasjfoxmaster volume = host system volume
00:21:19thomasjfoxIt was my impression that it added support for changing the volume of the music channel inside rockbox
00:21:45thomasjfoxLet me grab and build without software mixer
00:21:46kugelno, that's not added
00:22:15kugelwhile it can do this in theory, you can't control that. it's done automatically when voice is speaking
00:23:37 Join stoffel [0] (
00:24:17thomasjfoxkugel: You're right. Even the old build can adjust the volume of the music with the up/down buttons
00:25:16thomasjfoxargh. Now it gets clear to me
00:25:32thomasjfoxOf course we adjust the volume on our own gstreamer channel
00:25:40thomasjfoxBut we don't touch the master host volume
00:25:55 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
00:26:53 Quit Strife89 (Read error: Operation timed out)
00:27:04thomasjfoxI had experimental support to sync the master host volume to rockbox's gstreamer channel and that was just working like crap
00:27:45kugeloh, that's unfortunate
00:27:50kugelit works nicely on android
00:28:37thomasjfoxI also remember why:
00:28:54thomasjfoxIf you push the master volume over a certain limit, the speaker protection kicks in
00:29:31thomasjfoxThe speaker protection is done in software and chews around ~12% extra CPU, no matter how loud the music is playing
00:31:23 Quit TomColler (Ping timeout: 268 seconds)
00:31:29thomasjfoxkugel: Cu, gotta count some sheep.
00:31:42 Quit thomasjfox (Quit: Konversation terminated!)
00:33:24 Join TomColler [0] (
00:35:02 Quit pamaury (Remote host closed the connection)
00:36:22 Quit nosa-j (Ping timeout: 248 seconds)
00:41:47 Join nosa-j [0] (
00:44:41CIA-88New commit by kugel (r31548): Rework powermgmt to enable code re-use on appliation and sims. ...
00:44:53CIA-88New commit by kugel (r31549): android: Add preliminary eclipse support for building the native code, disabled by default.
00:47:36CIA-88r31548 build result: 53 errors, 4 warnings (kugel committed)
00:50:07CIA-88r31549 build result: 52 errors, 2 warnings (kugel committed)
00:53:47 Quit bertrik (Ping timeout: 240 seconds)
00:58:16***Saving seen data "./dancer.seen"
01:07:30CIA-88New commit by kugel (r31550): Fix reds by implementing a few more stubs, and undefining HAVE_STORAGE_FLUSH in the sim. ...
01:09:45CIA-88r31550 build result: 0 errors, 2 warnings (kugel committed)
01:10:48CIA-88New commit by kugel (r31551): android: move pcm_shutdown() to audiohw_close() now that it's properly called from powermgmt.
01:11:35 Quit TomColler (Remote host closed the connection)
01:12:36CIA-88r31551 build result: 0 errors, 4 warnings (kugel committed)
01:13:10CIA-88New commit by kugel (r31552): Fix beast warning.
01:14:48 Quit domonoky (Ping timeout: 240 seconds)
01:15:05CIA-88r31552 build result: 0 errors, 2 warnings (kugel committed)
01:15:24 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
01:18:49kugelgit ci -va
01:19:25CIA-88New commit by kugel (r31553): Fix intermittent warning on sdl app build.
01:20:48 Join nosa [0] (
01:21:09CIA-88r31553 build result: 285 errors, 0 warnings (kugel committed)
01:21:59funmansizeof(tmp) -> temp?
01:22:15funmanwhy is this code duplicated?
01:22:23 Quit nosa-j (Ping timeout: 252 seconds)
01:22:23 Nick nosa is now known as nosa-j (
01:23:08CIA-88New commit by kugel (r31554): Fix oops. Too late now.
01:24:57funmankugel: can't cpufrequency and scalingfrequency be factored?
01:25:04funmanthey are only used in debug menu afaict
01:25:06CIA-88r31554 build result: All green
01:25:45kugelfunman: hm?
01:25:46funmanfrequency_linux(int cpu_number, bool cpufreq) /* cpu == !scaling */
01:26:07kugelgo ahead if you like
01:26:19funmanwhich target use this?
01:26:35kugelall those which had lots of reds a few minutes ago :P
01:26:45funmantrue :)
01:26:46kugelhosted builds on linux
01:31:30 Join deute [0] (
01:32:16deuteforum down?
01:34:11CIA-88New commit by funman (r31555): *frequency_linux(): factorize
01:34:26funmandeute: yes it's down for me
01:35:57CIA-88r31555 build result: All green
01:37:49 Join KiwiCam_ [0] (
01:40:23 Join sansa-clip-plus [0] (
01:42:10 Quit sansa-clip-plus (Client Quit)
01:48:11 Quit TheLemonMan (Quit: WeeChat 0.3.6)
02:01:24 Quit stoffel (Remote host closed the connection)
02:03:53 Join Axe3 [0] (
02:04:42Axe3Will my FM transmitter work with my ipod 30gig video if i install rockbox?
02:07:29 Join Keripo [0] (
02:08:02 Quit deute (Quit: CGI:IRC (Ping timeout))
02:10:17 Quit lebellium (Quit: ChatZilla 0.9.88 [Firefox 10.0/20111228055358])
02:14:05funmanjust try and tell us
02:14:36funmando you plug it on the jack headphones or on the apple usb port?
02:15:05Axe3apple usb port
02:16:43funmansome accessories work some don't, check on the wiki
02:16:57Axe3thanks ill check it out i browsed the website but didnt see anything
02:17:39Axe3also would i still use iTunes to put songs on my ipod
02:18:03funmanyou can but it's simpler to just copy paste folders
02:18:28Axe3well if i can do that that would just be awesome lol
02:19:04funmanif you can copy paste?
02:21:21Axe3what device do you use rockbox on
02:25:32Axe3wait so i can have the original OS on my iPod AND rockbox?
02:30:50funmansure, check the manual it should have some FAQ
02:35:34 Join perrikwp_ [0] (
02:37:30 Quit jlbiasini (Quit: jlbiasini)
02:37:36 Quit perrikwp (Ping timeout: 240 seconds)
02:40:52 Join alienkid10 [0] (~alienkid@unaffiliated/alienkid10)
02:41:12alienkid10are the forums down for a known reason?
02:41:33alienkid10I'm unable to connect for some reasion
02:42:10JdGordonscorche|sh: ^
02:49:13 Nick perrikwp_ is now known as perrikwp (
02:52:51 Quit factor (Read error: Operation timed out)
02:57:45 Part alienkid10 ("QUIT :bye")
02:58:17***Saving seen data "./dancer.seen"
03:04:14 Part domonoky
03:05:04 Join factor [0] (~factor@
03:07:00 Join Topy44 [0] (
03:07:01 Quit T44 (Read error: Connection reset by peer)
03:07:41 Quit Topy44 (Read error: Connection reset by peer)
03:07:55 Join Topy44 [0] (
03:09:33 Quit Axe3 ()
03:26:01 Quit Topy44 (Read error: Connection reset by peer)
03:26:06 Join T44 [0] (
03:35:21 Quit T44 (Read error: Connection reset by peer)
03:35:22 Join Topy44 [0] (
03:36:11 Quit Topy44 (Read error: Connection reset by peer)
03:40:14 Join Topy44 [0] (
03:42:12 Quit KiwiCam (Quit: Leaving)
03:47:24 Join pocketprotector [0] (
03:47:38pocketprotectoryo guys, whats up with
03:53:39JdGordonthe server it and the forums are on are down
03:53:45JdGordonscorche: ^?
03:55:36pocketprotectoroh. well i have a very nice server if i can some how contribute to this project.
03:55:58pocketprotectori would not mind helping host the sites.
03:56:34pocketprotectorits 10MBps up.. usually i get about 12-13
03:56:36JdGordonI think we are fine for hosting, its just that server was being fiddled with recently and obviously not quite up again yet :)
03:57:00pocketprotectoroh okay.
03:57:19pocketprotectormaybe a mirror isn't such a bad idea
04:18:45 Quit pixelma (Disconnected by services)
04:18:47 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:18:50 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:19:45 Quit amiconn (Disconnected by services)
04:19:46 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:20:08 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:29:29 Join dys` [0] (
04:30:58 Quit dys (Ping timeout: 252 seconds)
04:31:22 Quit [7] (Disconnected by services)
04:31:32 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:32:46 Quit Scromple (Quit: Leaving)
04:37:15 Join Scromple [0] (~Simon@
04:38:32 Join evilwombat [0] (
04:38:52evilwombatFor Recorder devices, does Rockbox require a partition table, or can the entire disk be a VFAT filesystem?
04:39:16evilwombat(I am putting in solid state storage and want to avoid that whole partition-starts-on-a-flash-block-boundary mess)
04:40:30JdGordonI know we support superfloppy devices, not sure about for the main storage though
04:40:44funmanbluebrother^: Undefined symbols:
04:40:44funman "___memcpy_chk", referenced from:
04:40:45funman _swd_initdict in libuclppc.a(n2e_99.o)
04:40:50funmanbluebrother^: with gcc-4.0
04:40:57funmanbluebrother^: building mkamsboot 1.5
04:41:08JdGordonevilwombat: are you talking about a Archos recorder?
04:41:20evilwombatJdGordon, yes. ajbrec v1
04:41:39evilwombatseems to work
04:41:53JdGordonwhy do you care about the speed penalty of the block boundaries? the usb->ata chip is only usb 1.0 anyway
04:42:00evilwombatdude, sweet!
04:42:06JdGordonno where near fast enough to notice the flash write speed issues
04:42:11evilwombatI have the usb2.0 version
04:42:29evilwombatalso, wear penalty of modifying two blocks when one is enough
04:43:11JdGordon99% of the disk will only be written to once
04:43:24evilwombataah, ok
04:43:46evilwombatwell, archos recorder + compactflash = success.
04:44:14evilwombatI found a nifty CF / IDE adapter which is shaped like a regular HDD.
04:44:20evilwombatand fits very nicely
04:44:29evilwombatSandisk 8GB cf card
04:44:52JdGordonwhatever floats your boat :)
04:45:22evilwombatRB seems very pessimistic about the low watermark, though... refilling waaay earlier than necessary. Spinup time, what's that?
04:45:30evilwombatoh well
04:45:49evilwombati expect the battery life on this setup to be vastly increased.
04:58:20***Saving seen data "./dancer.seen"
05:10:40evilwombatAssembled and appears to be working- now to stress-test the ATA bridge :)
05:11:06evilwombatThe key thing is the shape of the CF/IDE adapter. This one has screw holes in all the right places.
05:12:01JdGordonyou're doing it wrong if you cant get a CF card into the space for a 2.5" hard disk!
05:12:30evilwombatWell, the HDD is kind of part of the whole structure. I believe the 4 main screws actually screw into the disk mounting holes
05:13:53evilwombatThis is the adapter I used-
05:14:34evilwombatthe ATA bridge and hte card play mostly nicely
05:14:56evilwombatI've had to power-cycle to get it to fully enumerate once, but that may just be usb-storage being dump again
05:15:28evilwombatI've grown so accustomed to expecting it to vibrate and then lag like hell whenever the disk light blinks
05:18:59JdGordonI put a 32GB CF in my ipod video
05:19:03JdGordonseems a massive waste in a AJBR
05:19:29 Nick dys` is now known as dys (
05:24:05 Quit factor (Read error: Connection reset by peer)
05:28:31evilwombatJdGordon, how so?
05:29:28evilwombatThe only other RB device I have is an H300, and that was randomly given to me by a friend after I had already gotten the CF adapter.
05:29:52evilwombatAnd the only other use I have for CF cards is booting an ancient ARM4, and the bootloader on there hates partitions >1GB, anyway
05:30:44evilwombatNow, putting one into the H300, THAT would be cool. But, need adapter
05:31:06JdGordonbecause its huge and clunky... if i wanted to use a AJBR I'd put in a 1T 2.5" disk and have all my music available
05:31:39evilwombatAh, fair. My music collection is 4.6 GB, so no issues there. I will consider putting one into the H300
05:34:52 Join factor [0] (~factor@
05:36:51 Quit Horscht (Quit: Verlassend)
05:52:57 Quit Keripo (Quit: Leaving.)
05:53:17 Join Rob2222 [0] (
05:57:04 Quit Rob2223 (Ping timeout: 240 seconds)
06:21:33CIA-88New commit by funman (r31556): dbg_partitions: make static
06:21:45CIA-88New commit by funman (r31557): dbg_ports() : move proto to system.h
06:21:53CIA-88New commit by funman (r31558): dbg_ports(): remove duplicate prototype
06:24:03CIA-88r31556 build result: All green
06:25:33 Join guymann [0] (~charles@
06:26:42CIA-88r31558 build result: All green
06:29:25CIA-88New commit by funman (r31559): dbg_hw_info() : add prototype to system.h
06:29:33CIA-88New commit by funman (r31560): dbg_hw_info(): remove duplicate prototypes
06:30:07funmanmost debug-target.h now only have a definition of "DEBUG_CANCEL"
06:30:17funmani guess we could use keymap context for that
06:31:57CIA-88r31559 build result: All green
06:34:20CIA-88r31560 build result: All green
06:36:54JdGordonfunman: there is a action_get_userabort() which is better
06:37:04JdGordonthoguh that's all in firmware/ isnt it?
06:37:56 Nick Jack87|Away is now known as Jack87 (Jack87@nasadmin/admin/jack87)
06:38:21funmanaction_userabort() ?
06:39:04JdGordonthats the one
06:50:20 Join nosa [0] (
06:50:55 Quit nosa-j (Ping timeout: 240 seconds)
06:50:56 Nick nosa is now known as nosa-j (
06:52:29scorcheJdGordon: yeah - i upgraded something that should have been harmless and it stuffed everything up
06:52:32scorcheworking on it
06:53:11scorchei might have to re-do my entire apache config...
06:53:43Topic"Rockbox is 10 years old! | Please read before speaking: | Please direct offtopic/social chat to #rockbox-community | This channel is logged at | Forums and Themesite are temporarily down for maintenance" by ChanServ (ChanServ@services.)
06:55:51JdGordonoh joy
06:58:24***Saving seen data "./dancer.seen"
07:00:21[Saint]gevaerts: will be happy.
07:00:35[Saint]No forum spam.
07:07:21evilwombatSo yeah. This IDE adapter actually takes two CF cards (that is, master and slave). Though they lock me up at a mental institution for this, I'll still ask- what would it take to get RAID-0 working in RB, and on the host PC?
07:08:17evilwombatWouldn't it just be hilarious to get the ATA bridge to export both block devices, apply mdadm to them, and teach rockbox to assemble the array?
07:08:28*evilwombat hides
07:11:52 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
07:13:35kugelpfunman: I would prefer to not move stuff that targets implement into general headers
07:15:11kugelpI'd rather make eg debug-target.h a common header
07:16:04kugelpbut I realize we're not consistent there
07:17:52funmankugelp: well the function is general
07:17:57funmanonly implementation varies
07:18:12funmane.g. pcm driver
07:20:11funmancargo cult++
07:20:39funmanthat button is defined almost everywhere but needed only in a few targets
07:21:17CIA-88New commit by funman (r31561): dbg_pcf(): use action_userabort
07:21:46kugelpI do like the the way it was implemented for ata recently
07:23:25funmanit can still be moved after though
07:23:27kugelpthe functions aren't general, they specifically exist for the debug menu
07:23:37CIA-88r31561 build result: All green
07:23:39funmanyeah but they exist on every target
07:23:58funmanthere could be a debug-target.h in firmware/export if you really want
07:24:23kugelpor debug-driver.h :-)
07:27:38kugelpsuch a file would be nice for lcd, backlight and more
07:33:51CIA-88New commit by funman (r31562): sh/debug-target.h: move in .c
07:33:55CIA-88New commit by funman (r31563): coldfire: move debug-target.h content in .c
07:34:01CIA-88New commit by funman (r31564): imx31: move DEBUG_CANCEL definition to .c
07:34:05CIA-88New commit by funman (r31565): s3c2440: move debug-target.h content to .c
07:34:09CIA-88New commit by funman (r31566): pnx0101: debug-target.h content is not used
07:34:14CIA-88New commit by funman (r31567): s5l8700: move debug-target.h content to .c
07:34:21CIA-88New commit by funman (r31568): s5l8702: move debug-target.h content to .c
07:34:26CIA-88New commit by funman (r31569): imx233: move DEBUG_CANCEL definition to .c
07:34:35CIA-88New commit by funman (r31570): PP: move DEBUG_CANCEL definition to .c
07:34:37CIA-88New commit by funman (r31571): rk27xx: move debug-target.h content to .c
07:34:58CIA-88New commit by funman (r31572): as3525: move debug-target.h content to .c
07:36:14CIA-88r31562 build result: All green
07:37:26CIA-88New commit by funman (r31573): imx233: move dbg function declaration to system-target.h
07:38:38CIA-88r31572 build result: All green
07:39:22funmanthat leaves one target-specific function
07:39:36funmanimx31: __dbg_dvfs_dptc()
07:39:52funmanand 15 empty debug-target.h
07:40:43CIA-88r31573 build result: All green
07:43:35Mode"#rockbox +o scorche" by ChanServ (ChanServ@services.)
07:43:42Topic"Rockbox is 10 years old! | Please read before speaking: | Please direct offtopic/social chat to #rockbox-community | This channel is logged at" by scorche (~scorche@rockbox/administrator/scorche)
07:43:49Mode"#rockbox -o scorche" by ChanServ (ChanServ@services.)
07:46:50 Quit Topy44 (Quit: Leaving)
07:49:56 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
07:56:29 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
07:59:25funmanjhMikeS: can imx31 __dbg_dvfs_dptc be merged in dbg_ports or dbg_hw_info ?
08:00:40 Quit knittl (Read error: Operation timed out)
08:03:00jhMikeSHere! I would say it's not really compatible. It would be better to give to rename it to whatever the cpu frequency screen is.
08:03:29CIA-88New commit by funman (r31574): zenutils: remove unused empty file
08:05:00jhMikeSdbg_cpufreq I guess it is, which is mostly descriptive
08:05:28funmanbut dbg_cpufrea is static to debug_menu.c
08:05:42CIA-88r31574 build result: All green
08:05:46funmanit debugs the cpufreq coutner afaik
08:06:51jhMikeSright. I'd keep the display unique string since it doesn't have a counter
08:07:18 Join Wojtson [0] (
08:07:34funmanperhaps we could move targe tspecific entries to target tree by using weak symbol
08:08:00jhMikeSI guess that's not a problem
08:08:54pamauryfunman: why did you move dbg_hw_target from debug-target.h to system-target.h ?
08:09:11funmanpamaury: i'll remove debug-target.h
08:09:20 Quit Wojtson (Client Quit)
08:09:25 Join Wojtson [0] (
08:09:45jhMikeSfunman: actually it is, since certain variables and functions don't exist on i.MX31
08:09:46funmanpamaury: feel free to move it again to imx233-specific-stuff.h
08:10:03funmanjhMikeS: hm?
08:10:31jhMikeSit doesn't define HAVE_ADJUSTABLE_CPU_FREQ
08:10:55jhMikeSunless set from the debug screen, it's all automatic
08:11:34funmanhm ADJUSTABLE_CPU_FREQ means boost/unboost
08:12:17Wojtsonhello, could anybody know how I can restore original firmware (uninstall rockbox temporarily) in MPIO HD300 ?
08:14:59[Saint]The manual probably knows this.
08:17:46funmani think adding entries with NULL functions debug_menu.c' menuitems isn't going to work
08:21:04 Quit pamaury (Remote host closed the connection)
08:24:58 Quit Scromple (Quit: Leaving)
08:26:14 Join kadath [0] (
08:29:43funmanfast solution: add an extern proto in debug_menu.c
08:29:55funmanfor __dbg_dvfs_dptc
08:29:55 Quit Wojtson (Quit: CGI:IRC)
08:33:39 Join ender` [0] (
08:34:44 Join petur [0] (~petur@rockbox/developer/petur)
08:35:59 Join knittl [0] (
08:35:59 Quit knittl (Changing host)
08:35:59 Join knittl [0] (~knittl@unaffiliated/knittl)
08:42:05CIA-88New commit by funman (r31575): debug_menu.c: cleanup ...
08:43:01funmanif we move functions from debug_menu.c to targe ttree and still declare them in debug-target.h, we'll still hav #ifdef in debug_menu.c
08:44:31CIA-88r31575 build result: 35 errors, 0 warnings (funman committed)
08:45:44 Quit evilwombat (Quit: Leaving)
08:48:53CIA-88New commit by funman (r31576): missing parens
08:50:58funmanjhMikeS: ?
08:50:59CIA-88r31576 build result: All green
08:52:54funmanhm dbg_cpufreq proto neds to be somewhere
08:55:43jhMikeSSure if we don't want a more descriptive string in that case
08:56:00jhMikeSI suppose it really doesn't matter
08:56:18funmanand it'll only be yourself who look into this menu ever, right? :)
08:56:45jhMikeSI don't know. saratoga wanted to be able to set the speed for codec testing iirc
08:56:49*[Saint] wonders how hard/easy it would be to: a - include a link in the Rockbox commit log tweets, and b - string long commit messages into several tweets if needed instead of just the first 160 chars
08:56:54funmanit still doesn't work, the protype needs to be in a public place
08:57:06funmandreamlayers want to enable -Wmissing-prototypes
08:57:32jhMikeShis mentioning it was a big reason I ever wrote that code
08:57:52funmanwhy doesn't it use boost/unboost btw?
08:58:27***Saving seen data "./dancer.seen"
09:00:45jhMikeSbecause the frequency is controlled automatically by hardware, if needs more it gets more all by itself, and it uses three frequencies in normal use
09:01:20jhMikeSit's automatic control for core voltage as well
09:04:15jhMikeSafaik it's unique in that respect from any other native target
09:11:53jhMikeSI suppose "boost" is possible in an indirect way by setting the load bits which add software load hints to the hardware's calculations
09:15:53jhMikeSnothing has behaved such as to really indicate it should be done. it works quite fine as-is with all the heavy-duty codecs and video
09:17:37funmanlooks cool indeed
09:20:44desowindoes C99 actually allow declaring variable inside for initialization?
09:25:49 Join liar [0] (
09:32:43 Join nick_p [0] (
09:33:46 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:36:19 Join y4n [0] (y4n@unaffiliated/y4ndexx)
09:46:22 Join ihst [0] (~ihst@
09:49:18 Quit ihst (Client Quit)
09:52:12 Quit scorche (Disconnected by services)
09:52:16 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
10:19:41 Quit nick_p (Quit: Leaving)
10:23:02 Quit Poodlemastah (Read error: Operation timed out)
10:28:54 Join stoffel [0] (
10:29:01 Join Poodlemastah [0] (
10:45:27 Quit [Saint] (Remote host closed the connection)
10:50:38 Join GodEater_ [0] (93722cd1@rockbox/staff/GodEater)
10:51:19 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
10:56:05 Join MethoS- [0] (~clemens@
10:58:12 Quit MethoS- (Remote host closed the connection)
10:58:29***Saving seen data "./dancer.seen"
11:03:46 Quit [Saint] (Read error: Connection reset by peer)
11:06:26 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
11:06:39 Join Topy44 [0] (
11:07:38 Quit stoffel (Read error: Operation timed out)
11:22:22 Join nosa [0] (
11:25:48 Quit nosa-j (Ping timeout: 244 seconds)
11:25:49 Nick nosa is now known as nosa-j (
11:46:15 Join TheLemonMan [0] (
11:47:38 Join TomColler [0] (
11:50:20 Join stoffel [0] (
11:52:17TheLemonManle-hrm, i'm trying to get usb to run but i get a swi error, i bet the code jumps in some random section
12:03:17CIA-88New commit by jethead71 (r31577): PCM mixer: Get rid of an obsolete macro cruft.
12:05:43CIA-88r31577 build result: All green
12:11:53 Join dfkt [0] (
12:11:58 Quit dfkt (Changing host)
12:11:58 Join dfkt [0] (dfkt@unaffiliated/dfkt)
12:21:22 Quit jhMikeS ()
12:31:10 Quit einhirn (Read error: Connection reset by peer)
12:36:32 Join jhMikeS [0] (
12:36:32 Quit jhMikeS (Changing host)
12:36:32 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
12:36:37 Quit factor (Ping timeout: 240 seconds)
12:46:54 Quit [Saint] (Remote host closed the connection)
12:52:48 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
12:54:55 Quit anewuser ()
12:56:47 Join anewuser [0] (~anewuser@
12:57:00 Quit anewuser (Changing host)
12:57:00 Join anewuser [0] (~anewuser@unaffiliated/anewuser)
12:58:31***Saving seen data "./dancer.seen"
13:03:15 Quit petur (Quit: *plop*)
13:05:48 Quit jhMikeS (Ping timeout: 240 seconds)
13:08:12 Join jhMikeS [0] (
13:08:12 Quit jhMikeS (Changing host)
13:08:12 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
13:09:37pixelmakugel: re FS #8778 - I believe amiconn said that it's not the right fix and can make things worse for some of these Ipods but can't remember details
13:09:38fs-bluebot iPod 1G scrollwheel not enabled (patches, unconfirmed)
13:10:29TorneBagder: when are you/zagor gonna be available to prod at git stuff some more? :)
13:14:51 Quit stoffel (Read error: No route to host)
13:17:03 Join stoffel [0] (
13:20:02TheLemonManyay got an undefined instruction from whitin a data abort one
13:20:29TheLemonMani smell mmu problems
13:21:25 Join Thra11 [0] (~thrall@
13:27:02 Quit liar (Read error: No route to host)
13:27:43 Join liar [0] (
13:32:17 Join konfiot [0] (
13:33:23 Quit konfiot (Client Quit)
13:37:24 Join mortalis [0] (~mortalis@
13:38:10 Join Buschel [0] (
13:40:24 Quit [Saint] (Read error: Connection reset by peer)
13:41:42 Join [Saint] [0] (~Saint]@unaffiliated/saint/x-8516940)
13:42:31 Quit stoffel (Ping timeout: 240 seconds)
13:45:47Misanthroposfirst let me thank you for working on usb support for the clip+ - it seems to work well with resent r31536
13:47:32Misanthroposwhile the #DEFINE HANLDE_USB_BY_OF is a compiler option - I wonder if a setting entry for "play while charging" would be welcomed - i like it, since i sometimes want to continue listening when I come to the office and have it recharged instead of moving files
13:48:21[Saint]This should be possible already.
13:48:31Misanthroposwith the hide internal drive option?
13:48:47Misanthroposor how?
13:48:59Misanthroposbecause that did not work :)
13:50:10gevaertshm, apparently that's not enabled on clip+
13:51:44gevaertsMisanthropos: try holding a button while plugging in
13:52:05Misanthroposomg.. works.. thx :D
13:52:42*[Saint] does a happy dance.
13:53:37[Saint]I thought it should work, but gevaerts' "no" confused me so I backed away, I wasn't sure what it applied to.
13:53:48Misanthroposme neither :D
13:53:50gevaertsthe manual wasn't updated apparently
13:54:13[Saint]Shock horror! :)
13:54:26gevaertsI thought the manual text would be derived from the actual options, but for some reason it's not
13:54:52[Saint]I actually thought it was too...
13:55:24gevaertsWell, it mostly is, but the AMSv2 situation is a bit special
13:57:24CIA-88New commit by gevaerts (r31578): USB chargung now works the normal way on AMSv2, so remove the hack that treated it differently. Also fix the fuze+ workaround for the lack of ...
13:58:09 Nick scorche` is now known as scorche (~scorche@rockbox/administrator/scorche)
13:59:20CIA-88r31578 build result: All green
14:01:15 Join einhirn [0] (
14:13:36 Join lorenzo92 [0] (
14:14:24 Quit nosa-j (Ping timeout: 240 seconds)
14:14:54 Join nosa-j [0] (
14:24:53[Saint]lorenzo92: for what its worth
14:25:44[Saint]For what its worth, a damaged rfs filesystem causes no additional CPU usage on any of my Sammy mobiles.
14:26:24Tornethat doesn't really make any sense
14:26:59lorenzo92ah don't ask me either...I don't really know how that could be possible!
14:27:06TorneThe only thing i can see going crazy if the fs is broken is the initial dircache scan
14:27:30Tornebut, if the fs is broken enough that that scan takes much longer/more cpu, then i would expect the fs is probably incomprehensible and it may never terminate
14:27:35Torneat which point, yeah, tough, your FS is broken
14:27:50Tornei can't imagine anything that would cause it to use loads of cpu and then *not also break completely*
14:28:06Tornehaving said that our FS implementations are not super rugged :/
14:28:18lorenzo92running fsck.vfat on pc, I got something like bad total cluster size + unknow fat number
14:28:23TorneOh, wait, this is a linux device
14:28:28lorenzo92yep :)
14:28:41TorneSo yeah, it's probably not in our code at all that it's using the cpu then :)
14:28:48Torne!blame kernel
14:29:27Tornedo we still do dircache on hosted platforms?
14:29:45[Saint]Afaik yea.
14:29:55Tornei guess that invovles just stat'ing everything on the vfs
14:30:21Tornei dunno what this "rfs" thing is but it sounds like there is some nonstandard filesystem foo going on
14:30:44Torneif that's implemented badly and it grinds to a halt on corrupt metadata then, yeah, stating everything is likely to do that :)
14:31:48[Saint]Hum...maybe not. Unless I've lost the setting somewhere. The menu seems to have changed on me.
14:32:11Torneif we *aren't* doing dircache, then i really don't know :)
14:32:27Tornebut it seems sensible to default to blaming the people who built the kernel if there's a nonstandard fs driver involved
14:32:33Tornebecause people cannot code that stuff for shit
14:32:41Torneespecially consumer hardware vendors :)
14:34:09lorenzo92Torne: exactly :D
14:34:28Tornethere's quite probably nothing we can do about that either
14:34:33Torneother than "fix kernel"
14:34:35lorenzo92there is this rfs, well I found the source codes
14:34:40Tornewhat is it?
14:35:02[Saint]Wow...shit. I'm out of the loop. We can monitor CPU on raaa finally?
14:35:04TorneOh, right
14:35:08Torneit's another journalled fat
14:35:16TorneYeah, I would expect this is totally broken.
14:35:19lorenzo92no time now to explain as I have to go, but it's like a fat32 layer that *should* avoid errors lol
14:35:36Tornecan you actually mount the disk over usb?
14:35:47Tornethen yeah, it's totally broken
14:35:49Torneby definition
14:35:50lorenzo92it's a fat drive after all
14:37:15Tornei vote: replace it with ext4, impleemnt mtp :)
14:37:39lorenzo92not all users are linux guys :D
14:37:46TorneSee under "mtp"
14:37:52GodEater_all those attempts at journalling FAT just need to die in a fire
14:37:54TorneUSB MSC is braindamaged. don't allow it.
14:38:05TorneGodEater_: psh! symbian rugged fat actually works.
14:38:12Torne(because it doesn't use a journal)
14:38:22Tornefor a value of "works" that excludes anything bad that happens while it's mounted over USB MSC
14:38:34GodEater_so it exploits a loophole in my pronouncement above then!
14:38:36Tornebut that's also true of all the journalled fat's
14:39:05GodEater_or presumably any file system at all
14:39:11GodEater_if the FS driver is shoddy
14:39:46TorneWell, it's not true of, say, ext4, because if you mount an ext4 filesystem on the host then the host will also do journalling/barriers/etc
14:39:49Torneand thus your integrity gets preserved
14:39:54Tornebecause the fs actually has the integrity designed into it
14:39:55GodEater_goddamn it - why does Spotify Mobile NEVER REMEMBER what I was listening to last
14:40:07GodEater_I meant over MSC
14:40:11GodEater_rather than mtp
14:40:17TorneEven over msc..
14:40:31TorneIt's not very *portable* to do so, but it's not unreasonable to unmount an ext4 disk, expose it over usb msc, then mount it again
14:40:33lorenzo92I still need to see how much time does it need fsck.vfat on device's bootup....
14:40:36GodEater_I bet someone somewhere could write an EXT4 driver shitty enough :)
14:40:48TorneWell, yeah. but they mostly haven't :)
14:41:14gevaertsDAPs should use those fancy cluster filesystems
14:41:15TorneNot sure if anyone has actually written an IFS that can write to ext4
14:41:31Torneinstallable file system
14:41:31GodEater_save me from acronyms
14:41:39TorneFS plugins for windows
14:41:45GodEater_oh nice
14:41:52Tornethere's an IFS that can *read* ext2/3/4
14:41:59Torneand write ext2 at least.
14:42:05Tornethere's one that can write ext3 but it's *really dodgy*
14:42:08Torneor was last time i checked :)
14:42:54GodEater_people write shitty FS drivers all the time!
14:43:58 Quit lorenzo92 (Remote host closed the connection)
14:44:06TorneThere are not that many, was all i meant. :)
14:44:38Tornethere's only a few, and all but the "real" one are rarely used ;)
14:45:23[Saint]Some digging proves me mistaken...apparently there's no dircache on raaa...I wonder why? Its on other flash-based ports.
14:45:44[Saint]Because we dumbly walk the entire fs?
14:46:10*[Saint] shrugs.
14:46:10GodEater_I don't understand the need for dircache on flash based targets?
14:46:21gevaerts[Saint]: it seems reasonable to assume that the underlying kernel does caching
14:46:32GodEater_ are there some which seem so dog slow we still think it's worth it?
14:46:51[Saint]Not to my knowledge.
14:46:53Tornethe other flash based ports have it because it's faster
14:46:58Tornedon't they?
14:47:11GodEater_faster than what?
14:47:13GodEater_not having it?
14:47:13Tornei mean, the difference is not necessarily interesting
14:47:29Tornei would be stunned if there wasn't a trivially measurable performance improvement
14:47:34Tornei might be surprised if any humans cared
14:47:37Tornebut that's different :)
14:49:07[Saint]If there is indeed a speed increase, its not perceivable by normal humans :)
14:49:25[Saint]I certainly can't tell if its enabled or no.
14:49:55TorneRAM is way, way faster than flash, and dircache answers questions with many less "roundtrips" as well
14:50:02Torneso yeah, it should be measurably faster ;P
14:50:12Tornebut iv'e never had it turned on on my clipv2, so i dunno :)
14:50:20GodEater_all praise to Cacheri
14:52:11[Saint]Debug - database info bothers listing "dircache ram: no", but its probably just not smart enough to not display that if there's no dircache.
14:52:34[Saint]Or, menu changes hid the setting REALLY well.
14:53:54[Saint]Whoops "RAM Cache: no" even. Get it right Hayden...
14:57:02TheLemonMandoes UIE show the physical address in case of a data abort
14:58:26 Quit liar (Read error: Connection timed out)
14:58:32***Saving seen data "./dancer.seen"
14:58:54 Join liar [0] (
14:59:07 Join stoffel [0] (
15:03:21TorneTheLemonMan: it does on the platforms where you can deduce it
15:03:27Tornei.e. the ones with an MMU
15:03:40Tornethere is no way to find out on ARMs without cp15.
15:03:55 Join jlbiasini [0] (
15:04:49Torneit'll say "address 0xblah" if it can figure it out
15:04:59Torneif not then that cpu doesn't know.
15:05:08TheLemonManoh, then the address makes sense
15:05:12jlbiasinipamaury: I've made a new patch to have usable keymaps in plugins games
15:05:29Tornethe "at blah" is the PC address
15:06:33TheLemonMani think can be safely used even on the bootloader map
15:11:27 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
15:11:31 Join factor [0] (~factor@
15:13:36kugelthere's no dircache on raaa
15:15:35[Saint]I wasn't sure if I remembered incorrectly or not...and the debug
15:15:46[Saint]menu further confused
15:17:26[Saint]Seems like it probably wouldn't be too hard for it to not display values for features that aren't actually present.
15:19:49[Saint]"RAM cache: no" means! Not enabled, but could be" on the targets that have it...but seems to mean "not present at all" elsewhere.
15:19:58kugelthat's the database menu
15:20:23kugelRAM cache isn't related to dircache :)
15:20:54kugelif dircache is available there's an extra dircache debug menu
15:21:49[Saint]Ah...I am getting screwed up tonight aren't I...
15:22:03kugelRAM cache is a feature of the db, that's available on RaaA
15:22:28kugelah actually it isnt, but it could be
15:23:14[Saint]Yeah...just noticed its not present there. I was indeed mixing the two.
15:24:47[Saint]So my "debug menu should only display values for things that are present" still applies...I was just wrong about what I thought the string applied to :)
15:27:01[Saint]kugel: when did CPU stats sneak into raaa?
15:27:19kugelit's not only raaa, sim also
15:27:32[Saint]Its nice to not have to use OS Monitor to see when raaa is eating my MHz.
15:28:28kugelOS monitor and rockbox are likely to not report the same numbers, though
15:28:52 Quit liar (Read error: Connection timed out)
15:30:23TheLemonManseems that the problem is in i2c_add_node, if i comment it out it gives a much nicer undefined instruction
15:30:36[Saint]Oh...indeed there is some difference.
15:31:35[Saint]+/- ~2MHz
15:31:37kugelyou know, measuring the cpu usage already uses cpu. let alone displaying the results at the same tim
15:31:45 Quit tchan (Read error: Connection reset by peer)
15:32:01kugelthen, on android, it makes a difference whether raaa is in foreground or not
15:32:30 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
15:32:56kugelfinally there are different methods to do the measurements and to sanitize the results
15:33:02 Quit tchan (Read error: Connection reset by peer)
15:33:45 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
15:37:47[Saint] that's...interesting.
15:38:23[Saint]Holding a finger in one spot on the screen in the CPU stats menu causes up to 100% CPU usage.
15:38:37[Saint]Constant screen updates?
15:39:48[Saint]Yeah...scrolling is causing some huge spikes in usage.
15:41:25[Saint]2%, to 50~100% just from holding a finger in one spot.
15:43:18 Join MethoS- [0] (~clemens@
15:49:38*[Saint] is interested as to whether or not anyone else sees these huge spikes in CPU usage from tricking the screen into updating like this.
15:51:41[Saint]Seems like a flaw in the scrolling. If my finger isn't actually moving, just resting on the screen, I don't think it should need to update a bajillion times a second ans slowly ramp the CPU up to 100% usage.
15:53:33[Saint]I might be wrong, but it seems like a "bad thing".
16:02:47 Join n1s [0] (~n1s@rockbox/developer/n1s)
16:19:57 Quit pamaury (Remote host closed the connection)
16:30:17 Quit Thra11 (Read error: Operation timed out)
16:39:27 Quit [Saint] (Read error: Connection reset by peer)
16:39:30 Join [Saint_] [0] (~Saint]@unaffiliated/saint/x-8516940)
16:42:32 Quit KiwiCam_ (Quit: leaving)
16:42:46 Join KiwiCam [0] (
16:47:04 Nick [Saint_] is now known as [Saint] (~Saint]@unaffiliated/saint/x-8516940)
16:47:31 Quit Zambezi (Read error: Connection reset by peer)
16:48:21 Join Zambezi [0] (
16:53:37 Quit kadath (Ping timeout: 240 seconds)
16:57:00 Join kadath [0] (
16:58:07 Quit mortalis (Quit: KVIrc 4.1.1 Equilibrium
16:58:35***Saving seen data "./dancer.seen"
17:01:44 Join pamaury [0] (
17:01:44 Quit pamaury (Changing host)
17:01:44 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
17:02:05 Quit factor (Read error: Connection reset by peer)
17:13:04GodEater_is it just me that can't see "z" in Magnus' email?
17:13:08GodEater_or am I blind?
17:13:35gevaertsz is y :)
17:13:44GodEater_that explains a lot :D
17:14:04GodEater_I've re-read that email five times thinking "I'm missing something here"
17:18:42 Join factor [0] (~factor@
17:29:51 Join liar [0] (
17:42:11 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 9.0.1/20111220165912])
17:44:46 Quit jlbiasini (Remote host closed the connection)
17:51:25 Join Thra11 [0] (~thrall@
17:53:11 Quit GodEater_ (Quit: Page closed)
17:55:13dfktis there some trick to rockbox apk creating/signing, so that i don't have to uninstall an older version before installing a newer one?
18:00:36bluebrother^if you build the apk on the same machine each time you don't need to uninstall first −− as long as the new apk is using the same certificate you're good.
18:00:38[Saint]You should never have to, provided the builds were signed by the same key.
18:00:58bluebrother^the only problem is that the debug keys expire after 1 year :)
18:01:07bluebrother^but that should be bearable
18:01:35dfktweird, i always build in the same virtual machine, and i always have to uninstall
18:01:50bluebrother^btw, it seems that building with ndk versions below r7 doesn't work anymore. Intentional?
18:02:01[Saint]bluebrother^: yes.
18:02:09bluebrother^I usually update using make install. No problem for me.
18:02:21bluebrother^[Saint]: why that?
18:02:46[Saint]Some change in what was explicitly linked.
18:03:10[Saint]...something to this extent.
18:03:17bluebrother^yes, but is it intentional to require the most recent ndk? Are we always following the latest ndk?
18:03:43[Saint]To target ics I believe of was necessary.
18:03:52[Saint]Usually, I don't think so.
18:04:01[Saint]*it was
18:04:02dfktbluebrother^, i do make > make zip > make apk... i guess i don't need make zip, and when to use make install - as last?
18:04:25[Saint]You DO need make zip
18:04:26bluebrother^dfkt: not completely sure. I usually use make && make zip apk install :)
18:04:45dfktah ok, so make install as last, thanks :)
18:05:01bluebrother^well yes, it installs the apk to the device :)
18:05:05[Saint]dfkt: that wont fix it.
18:05:19[Saint]Something is busted in your setup.
18:05:34dfktwell, i always sideloaded manually, no push or something
18:05:44bluebrother^is your VM killing the certificate store maybe?
18:05:45[Saint]That's not a pro lem.
18:05:57dfktdifferent patches used in builds won't affect the signing for sure?
18:06:00[Saint]I sideload also.
18:07:04dfktmake apk complains about being deprecated in my setup, maybe that's an issue
18:07:08[Saint]Something with your signing is busted I figure..
18:07:15[Saint]And no, that's fine.
18:07:27dfkthmm, ok, thanks
18:07:28[Saint]Make apk *is* depricated.
18:07:45[Saint]But, that's fine.
18:08:20dfktguess i will have to deal with uninstalling and recreating shortcuts/widget, it's not that much of an issue anyways
18:08:44dfktat least less hassle than setting up a new build environment :)
18:09:13[Saint]...if you say so lol
18:13:41 Quit pamaury (Quit: No Ping reply in 180 seconds.)
18:14:01 Join pamaury [0] (
18:14:01 Quit pamaury (Changing host)
18:14:01 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
18:32:02 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
18:42:53 Part TomColler
18:52:40CIA-88New commit by bluebrother (r31579): Extend flashing hint a bit. ...
18:54:14[Saint]Is anyone aware what handles the commit tweets? And, if there's any code I can be nosey at?
18:54:39[Saint]I swear it could be done in a way that makes it...useful.
18:54:42CIA-88r31579 build result: All green
18:56:05[Saint]I'd really like to include a link to the commit, and the string long commit messages across multiple tweets when needed.
18:56:36[Saint]But I can't seem to find what is handling the sending of said tweets.
18:56:59[Saint]*to string
18:58:18bluebrother^we have commit tweets?
18:58:27[Saint]We do indeed.
18:58:37***Saving seen data "./dancer.seen"
18:59:26Bagdertalk to Zagor about the details of that
18:59:57[Saint]Ah...hmmm. I never seem to catch him.
19:00:25[Saint]As they are now, they're not entirely useful.
19:00:42*bluebrother^ uses the RSS feed instead
19:01:16[Saint]Does it include a link to the commit its announcing?
19:01:40[Saint]That's what I want out of twitter...
19:01:50bluebrother^well, people following some of the commit message templates that are common among git users might improve that. Simply take the first line as short description and post only that
19:02:00 Quit saratoga (Ping timeout: 258 seconds)
19:02:15 Join Keripo [0] (
19:03:50[Saint]My "problem" is, twitter notifies me of a commit and I think "oh, I want to look at that", and I've got to go search of out manually.
19:04:11[Saint]I'd really like to just click a link if I want to.
19:05:54bluebrother^works fine with the RSS feed :)
19:06:05bluebrother^well, it includes a link to the full commit message only, though.
19:06:32[Saint]Yeah...I'd like to be taken to the diff.
19:07:05bluebrother^never bothered too much about that −− the diffs are usually still on the front page, so it's not too far away :)
19:07:49[Saint]But then I have rom launch the browser, it smacks of effort! :p
19:08:01[Saint]*to launch
19:09:19bluebrother^my browser is almost always open :)
19:09:44[Saint]On your phone? :)
19:13:26bluebrother^no, but I usually don't read diffs on the phone
19:20:01 Quit nosa-j (Ping timeout: 252 seconds)
19:20:26 Join nosa-j [0] (
19:21:50 Join Strife89 [0] (~Strife89@
19:24:19 Join Horscht [0] (
19:24:19 Quit Horscht (Changing host)
19:24:19 Join Horscht [0] (~Horscht@xbmc/user/horscht)
19:33:14 Quit stoffel (Ping timeout: 240 seconds)
19:33:19 Join krabador [0] (
19:38:31 Join bluebrother [0] (
19:38:31 Quit bluebrother (Changing host)
19:38:31 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
19:38:35 Join lebellium [0] (
19:40:56 Quit bluebrother^ (Read error: Operation timed out)
19:41:08 Quit fs-bluebot (Ping timeout: 240 seconds)
19:42:42 Join fs-bluebot [0] (
19:58:49 Join mars___ [0] (
19:59:34 Join stripwax [0] (
20:00:12 Quit mars___ (Remote host closed the connection)
20:05:06 Quit Strife89 (Ping timeout: 252 seconds)
20:06:47 Join bertrik [0] (
20:07:00 Quit bertrik (Changing host)
20:07:01 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
20:07:09 Join Strife89 [0] (~Strife89@
20:09:08remlapHi is m4b album art supported folder.jpg doesn't seem to work for it
20:09:49 Join lovasoa [0] (~olojkine@
20:11:19gevaertsremlap: folder.jpg should work
20:11:32gevaertsAre your files progressive jpegs maybe?
20:12:45remlapah never mind
20:12:46remlapit was png
20:12:52remlapthanks gevaerts
20:12:56remlapnow a jpg and fine
20:15:02gevaertsAh, good :)
20:15:13remlapwhy it had jpg extension
20:15:15remlapI have no idea
20:19:03CIA-88New commit by bluebrother (r31580): Rockbox Utility: add support for mkimxboot bootloader. ...
20:20:48CIA-88r31580 build result: All green
20:41:54lovasoakugel: latest svn refuses to build for the 0: I get rockbox/samsungYPR0/firmware/libfirmware.a(system-ypr0.o): In function `system_init':
20:41:56lovasoasystem-ypr0.c:(.text+0x20): undefined reference to `gpio_init' (And i made a make veryclean)
20:42:17funmanisn't the ypr0 in the build system yet?
20:43:47 Quit liar (Remote host closed the connection)
20:43:57 Join jlbiasini [0] (
20:44:07 Quit Xerion (Ping timeout: 240 seconds)
20:49:06kugellovasoa: works for me
20:51:44lovasoakugel: So why do I get this error. I checked that system_init() is defined in gpio_ypr0.c, but this file doesn't seem to be compiled during the build
20:52:54 Quit pamaury (Read error: Operation timed out)
20:54:31kugellovasoa: what does make $PWD/firmware/target/hosted/ypr0/gpio_ypr0.o do?
20:55:11 Join toa [0] (
20:56:46lovasoakugel: /firmware/target/hosted/ypr0/gpio_ypr0.o doesn't exist. gpio_ypro.c does not compile!
20:57:29kugellovasoa: the command should create the file
20:57:36kugel'make $PWD/firmware/target/hosted/ypr0/gpio_ypr0.o'
20:58:02lovasoakugel: Aah. OK I try this
20:58:26kugelperhaps you have local changes? gpio_ypr0.c is in svn, and added to sources properly
20:58:38***Saving seen data "./dancer.seen"
20:58:55lovasoaI have the gpio_ypro.c from svn
21:00:01lovasoamake tells me "No rule to make blabla"
21:00:31kugellook if the file is on firmware/SOURCES
21:03:29[Saint]kugel: any thoughts on scrolling in raaa causing insane CPU usage?
21:06:31[Saint]All my raaa devices are showing really high CPU usage caused by scrolling, well, actually not even scrolling...just holding a finger in one spot on the screen.
21:07:32kugelyes because it's updating the screen very oftem
21:08:41kugelthe "doesn't even move by one pixel" case isn't considered because it's so unlikely; you usually don't do this and even then you probably move by a pixel or to without noticing
21:10:37[Saint]I did some playing around to get the statusbar displayed, and holding a finger on the screen uses more CPU than flac playback with peak meters in the wps
21:10:40 Quit y4n (Quit: 6,000,000 ways to die — choose one.)
21:11:00[Saint]Which I had considered to be a "worst case scenario".
21:11:36[Saint](Statusbar was so I could watch os monitor)
21:12:10kugelwhy flac?
21:12:34kugelflac is probably the cheapest codec after wav :)
21:13:08[Saint]It seems to eat more CPU than the mp3s I have.
21:14:24[Saint]That's....interesting. that's my CPU, but, not my experience...
21:14:58kugelanyway, user interaction is the most expensive operation
21:16:06kugelyou want it both smooth and without latency
21:16:30kugelso it's driving cpu mad, especially updating the diplay
21:16:38[Saint]Something just seems really wrong to me that scrolling can use up to 100% CPU...I just can't shake that feeling. I'm probably not looking at it right, it just feels wrong to me.
21:19:04kugelit's doing lcd updates at the max. framerate
21:22:30[Saint]Wouldn't something like ~25 fps still be just as smooth?
21:22:55[Saint]Or, a barely notable difference at best?
21:27:35kugelI don't see a reason to cap it though
21:29:43[Saint]Well...scrolling in a very long list could in theory be enough to eat all available CPU.
21:29:54[Saint]Smells like a reason to me.
21:30:25[Saint](If it could be avoided)
21:33:37CIA-88New commit by learman (r31581): AAC/M4A: Playback time wrapped after 2 hours and 42 minutes. Increase the wrap limit to about 27 hours.
21:35:14CIA-88r31581 build result: All green
21:39:14lovasoakugel: I managed to compile the latest revision. I had a modified firmware/SOURCES. I made a "svn revert". :)
21:43:03 Quit toa (Changing host)
21:43:04 Join toa [0] (~toa@unaffiliated/toa)
21:49:34[Saint]"svn stat" ftw.
21:52:09 Quit Thra11 (Ping timeout: 252 seconds)
22:03:49 Join blind_tester [0] (
22:04:12 Join _jhMikeS_ [0] (
22:04:12 Quit _jhMikeS_ (Changing host)
22:04:12 Join _jhMikeS_ [0] (~jethead71@rockbox/developer/jhMikeS)
22:04:12 Quit jhMikeS (Disconnected by services)
22:04:13 Nick _jhMikeS_ is now known as jhMikeS (~jethead71@rockbox/developer/jhMikeS)
22:06:50blind_testerrockbox is great! i configured a sansa clip+ for a blind friend. talk-files are present for files+dirs. during selection i can hear the talk. but if i start at title 1 and i am now at title 3: how can i determine the actual title using talk? there is an option where i can *see* the title info but thats not enough for him.
22:15:02blind_testerI have developed a ruby script which scans a folder recursive, examines the file- and dirnames, rates the text (italian, english, german) and creates .talk-files using Mary TTS. I don't want to spend much time in registering as a developer etc. - is anybody here, who would take the source, have a look at it and maybe add it to ?
22:19:16kugelblind_tester: you can register on the wiki, with your real name. then I can give you write accesses in a second
22:20:41blind_testerThank you!
22:24:25bertrikCan someone tell me more about the recent USB fixes? Have we actually identified and fixed real problems with AMSv2 USB, for example?
22:33:45funmanbertrik: USB just works
22:34:05funmanthere are still a few problems though, that TheSeven is investigating
22:34:06bertrikcool, but how did it get fixed?
22:34:43funmanin one or several of my numerous commits, dunno which ones exactly
22:38:11bertrikfunman, does your clip zip display get a reddish tint when setting a low brightness?
22:38:20bertrik(or some other tint)
22:44:04 Join Scromple [0] (~Simon@
22:44:18blind_testerI tried to register for the wiki but i am not sure if all is fine. Can i get write access for my my account (MichaelArlt)?
22:46:10TheSevenblind_tester: done
22:46:19blind_testerThank you!
22:47:41blind_testerI have found 2 errors in deutsch.lang: "Einschalten in %d..." should be "Einschalten" and "%d Hz-Band Verstärkung" should be "Hz-Band Verstärkung" (in my opinion).
22:48:21TheSevenso it is shown as "<number> %d Hz-Band ..." on the player?
22:49:13blind_testerAehm in the in voice (for creating the talk files)
22:49:41*TheSeven is not sure how that stuff works
22:50:00TheSeven%d is usually a placeholder for a numeric value
22:50:37blind_tester% is only in <source> and <dest> - normally not in <voice> - only in these 2 translations...
22:51:24TheSevenonly in the german language or in others (especially english) as well?
22:51:47blind_testerthe text is used to generate the voice before %d is known. he would talk "%d" :-)
22:53:07funmanperhaps it could talk the real number
22:53:08blind_testerno - in english is no % −−> swcodec: "hertz band gain"
22:55:00bertrikhm, so maybe the number is added programmatically
22:55:01blind_testerbut it makes no sense to pregenerate a talk-file which talks "%d Hz-Band Verstärkung" and talk "25" before. the result would be: "25 %d Hz-Band Verstärkung"
22:55:10CIA-88New commit by gevaerts (r31582): Reorganise USB initialisation to not depend on a specific enumeration sequence, by Bartosz Fabianowski, with minor tweaks by Michael Sevakis ...
22:55:23funmanblind_tester: talk engine would skip the "%d" of course
22:56:38blind_testerit can't skip it. the rockbux utility generates the talk-files using espeak and it gives espeak "%d Hz-Band Verstärkung"... so what do you think will espeak generate?
22:57:47CIA-88r31582 build result: 0 errors, 78 warnings (gevaerts committed)
22:58:39***Saving seen data "./dancer.seen"
22:58:49blind_testerHave a look at english.lang - no %d - only: "hertz band gain"
22:59:52funmanblind_tester: true
23:02:08funmanwho is our talk expert?
23:02:22funmanthe %d should probably implemented, or removed from all voice strings
23:03:31[Saint]Does of not voice numerals in this instance?
23:03:38[Saint]*it not
23:03:39bertrikthe code for this string first does talk_number(*setting, false); then talk_id(LANG_EQUALIZER_GAIN_ITEM, true);
23:04:11bertrikso the %d in the voice string appears to be redundant
23:05:15blind_testerbertrik: Thank you
23:07:10CIA-88New commit by gevaerts (r31583): Move some variable declarations to inside #ifdef USB_FULL_INIT, to avoid warnings for bootloaders that don't have the full USB state machine
23:08:48 Join megal0maniac [0] (
23:09:23CIA-88r31583 build result: All green
23:09:59megal0maniacHi, all. Could anyone with a little more insight than myself, please tell me whether there is hope for a stable port for the Fuze+?
23:11:06bertrikthere is always hope :)
23:12:02megal0maniacI'm just curious, because the Clip Zip port happened extremely quickly. But am I correct to assume that most of the work was already done with the Clip+?
23:12:44gevaertsThe fuze+ is basically not related to any of the earlier sansa players, while the clip zip is basically a clip+ with a different screen
23:13:27gevaertsThe fuze+ port seems to be advancing nicely these days
23:15:09n1sbertrik: correct, the voice doesn't handle printf style place holders (how could it with pregenerated clips?)
23:15:39n1sthis could perhaps be a feature if we get an integrated tts some time which would simplify a lot of code
23:15:57megal0maniacI see. Despite the bugs, it's running 100x better than the OF, which is embarrassingly slow in my humble opinion. What would "Cannot create buffering thread" mean? MPEG Player throws it when attempting to play any video, even known supported formats.
23:15:58[Saint]n1s: I thought it spoke individual digits in these cases.
23:16:09n1spixelma: is the de facto maintainer of deutsch.lang i think so perhaps she would like to fix them
23:16:16[Saint]Like "one zero zero MHz"
23:16:46n1s[Saint]: it can do that but that is not done like this
23:17:14blind_testernls: but don't drop support for talk-files if you have an embedded tts - only if you support automatic language detection and have several voices :-)
23:17:32n1syou then need to talk_number(number); talk_id(lang_id);
23:17:47blind_testernls: language detection for the file- and foldernames :-)
23:17:56 Quit Zarggg (Quit: Rebooting client...)
23:18:12n1sblind_tester: yeah, it could probably never completely replace off line pre generated files
23:19:18n1s[Saint]: it could probably be done by spelling the whole string a digit/letter at a time after filling in the numbers with sprintf :)
23:23:32 Part jlbiasini
23:28:25adnapanyone else having problems with rockbox force screen locking on android?
23:28:45 Join Zarggg [0] (
23:36:49blind_testerFor all who are interested: The ruby script which recursively creates language specific .talk files is online: .
23:40:51 Quit stripwax (Quit:
23:41:47 Join stoffel [0] (
23:47:31 Quit krabador (Ping timeout: 268 seconds)
23:54:48 Quit n1s (Quit: Ex-Chat)
23:55:16megal0maniacWhat can I do to help speed up development of a port, apart from actual development. I can't do much more than compile a clean build, but I'm keen to assist
23:57:47 Quit remlap (Read error: Connection reset by peer)
23:58:58 Join krabador [0] (

Previous day | Next day