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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2011-02-10

00:00:07 Quit shai (Quit: Leaving)
00:00:17TheSevenbut on the power consumption side i don't see any effect, while in emcore the effect is massive
00:00:26TheSeven(emcore is also ticking at 100Hz, so that shouldn't matter)
00:00:49 Join JdGordon| [0] (
00:00:49 Quit JdGordon| (Changing host)
00:00:49 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
00:00:59*TheSeven likes the realtime "MHz in use" line in the buffering screen :)
00:01:17TheSeven(29-30MHz for MP3 right now)
00:01:47*TheSeven wonders why on earth it's boosting 3% of the time if it's sleeping 74% of the time
00:02:17pamauryis there something special I should now about ldr and mov pc, rx on ARM. For some reason I can't make it work
00:02:29 Quit webguest95 (Quit: CGI:IRC (EOF))
00:02:42*TheSeven looks for a power-hungry codec
00:02:55TheSevenbetween ape 3000 and 4000 would be perfect :)
00:04:22 Quit domonoky (Read error: Connection reset by peer)
00:05:55TheSevenaccording to my buffering screen hack c3000 is using ~56MHz
00:06:25*TheSeven wonders if he should make a better version of that hack and throw it into flyspray or even svn
00:10:03thomasjfoxkugel: Shutdown handling needs some more work, the last.FM file / database play counts still isn't written or read back
00:10:21thomasjfoxkugel: But embedded album art works fine :)
00:12:01TheSevenhm, c4000 plays like 50% realtime, but i don't see *any* difference in current draw compared to MP3
00:12:30kugelthomasjfox: meh/cool
00:12:39*TheSeven wonders if we can get rid of everything that relies on tick accuracy
00:12:47 Join robin0800 [0] (
00:12:57TheSevenit would be very helpful if i could just reduce the tick rate to 50 or even 25Hz while not boosting
00:13:23TheSevenwe'd still have an accutate USEC_TIMER, so i see no reason why that shouldn't work if done properly
00:18:00thomasjfoxkugel: I hope I'll find enough time to fix the shutdown stuff before Sunday. Though this will depend on "external" demands ;)
00:19:28 Quit robin0800 (Read error: Connection reset by peer)
00:19:57JdGordon|TheSeven: PLEASE do it!
00:20:12JdGordon|we had this argument with doing that for android when the app was hidden
00:20:28 Join robin0800 [0] (
00:20:52TheSevenJdGordon|: Seems like that's a can of worms though :/
00:20:53JdGordon|there is absolutly no reason the tick rate needs to stay constant, or heigh when the only important tihng happening is audio playback (i.e no screen)
00:21:36 Quit thomasjfox (Remote host closed the connection)
00:21:53TheSevenJdGordon|: wouldn't the best approach even be dynamically scheduling wakeups, if that's possible on the hardware?
00:22:01TheSeventhat would reduce both load and latency
00:22:54TheSevendo we have a usec timer or something equivalent on all platforms?
00:23:20 Join enthdegree [0] (
00:23:28TheSeveni mean, in the worst case we could just keep the tick timer constant on platforms that don't have one, and fake a usec timer based on that
00:24:03 Quit sideral (Quit: Leaving.)
00:29:04TheSevenwe should be at ~12h battery life on the ipod classic now
00:29:17TheSevenhow does apple manage to do three times as much?
00:30:02TheSeveni'm doing aggressive undervolting, clock gating, clock scaling and proper sleeping now
00:30:50 Quit [Saint] (Disconnected by services)
00:30:51 Join S_a_i_n_t [0] (S_a_i_n_t@
00:31:12LloreanTheSeven: I know on the gigabeat Swhen we tried to do manual clock scaling initially it actually harmed battery life. I think it was theorized that the CPU had automation to handle such things itself if we left it alone after the bootloader initialized it.
00:31:41TheSeveni don't think we have such a thing here
00:31:51LloreanI also remember there was quite a while on the PP iPods where we were asking similar questions along the lines of "we're doing an awful lot, yet they still beat us by 30%, why is this?"
00:31:58LloreanThough... three times is a bit much
00:32:22TheSevenapple seems to be running that thing at ~14mA average, juding from their 36h runtime claims
00:32:32TheSevenwe're at like 40 now
00:32:46LloreanAre the 36h claims realistic?
00:32:53 Quit bertrik (Ping timeout: 250 seconds)
00:32:54TheSeveni think so
00:33:11TheSevenat least their 24h claims on the nano2g were actually like 26h
00:33:30TheSevenand they were running that one at ~16mA (we're at ~13 now)
00:33:52LloreanSo do we need to use vastly less CPU, or is it other things draining power.
00:34:01JdGordon|TheSeven: dunno, the reason we were looking at the tick is because the tick tasks are mostly nop's when the display is off so there is no need to even bother calling them so often
00:34:02TheSevenit seems to be other things
00:34:04LloreanI know we were losing a lot of power to things like accessory power supply, etc.
00:34:08TheSevenCPU load is at ~30MHz
00:34:26JdGordon|it was theorised that the constant wakeups was sucking cpu time from android (Although thats been fixed a different way now)
00:34:48TheSevenif that thing is completely idle and i'm running emcore on it, i get down to ~12mA at very very low clock speeds (48MHz)
00:35:19LloreanDon't we clock under 48mhz on the PP ones?
00:35:26LloreanShouldn't we be able to get even lower on the Classic?
00:35:39LloreanI thought we were able to get as low as about 24mhz on the PP iPods with MP3 decoding happening.
00:35:56TheSevenwe can't get below 108MHz while keeping the tick rate constant and a decent bus speed
00:36:32TheSeven54MHz might be doable at half the tick rate, but everything below that kills USB and the HDD
00:36:34LloreanSo that's what needs to be solved, or is it just impossible to do that for some reason?
00:36:37TheSevenalso the LCD will start lagging
00:36:52LloreanWell, we don't need the HDD except during buffering (which boosts anyway), and doesn't USB also boost?
00:37:00TheSevenyes, they do
00:37:22TheSevenbut nevertheless i'm pretty sure that apple is running it constantly at 216MHz
00:37:57TheSevenalso that whole clock setup thing is not really well understood yet
00:38:07LloreanBut if you can get down to ~12mA at low clock speeds, that suggest at most 12mA is going to all other hardware, right? So if we're at 40 now, that means we could trim to 28, still double what you said Apple's at?
00:38:21TheSeveni think i've figured out the most important parts of it, but there are still questions left, like why we can't clock AHB == APB
00:38:41TheSevenLlorean: that's not true
00:38:48TheSeventhe 12mA is the absolute minimum i can reach
00:39:13TheSeveni.e. clocking at 24MHz or something, USB off, HDD off, backlight off, codec off, everything undervolted, CPU idle
00:39:33LloreanAaah, okay
00:39:50LloreanEven then it suggests you must be missing something if they can hit 14mA while decoding.
00:39:58 Quit mlt- (Quit: CGI:IRC (EOF))
00:39:59kugelTheSeven: can you disable the display?
00:40:01TheSevenyep, but what?
00:40:14S_a_i_n_tLlorean: some can go lower than 24Mhz also...but at a rather large cost to UI functionality.
00:40:16TheSevenkugel: i can, but apple isn't doing it
00:40:22S_a_i_n_t(hence the boost on GUI patch)
00:40:49TheSevenS_a_i_n_t: 54MHz is already lagging gui-wise
00:40:53kugelthe displays usually give huge power savings
00:41:10LloreanTheSeven: I think he meant PP iPods
00:41:23TheSevenhm, maybe they're re-initializing the LCD at some point and putting it into some lower power mode
00:41:33TheSeveni don't have datasheets for the LCDs yet :/
00:42:14LloreanWell, best of luck at least. The progress so far is still pretty exciting.
00:42:34TheSevenkugel: let me try LCD off, backlight off, codec off, USB off, everything off @54MHz
00:43:42TheSevenoh, wait
00:44:10TheSevenemcore, backlight off, USB ON, 216MHz, idling except for a 100Hz tick: 37mA
00:44:21TheSevenso lower than rockbox at 108MHz with USB off
00:45:07TheSeventhe LCD accounts for 3.5mA
00:45:29 Quit kugel (Remote host closed the connection)
00:47:40pamauryhaha, I now have the proof that my WinCE bootloader is mapping the SDRAM at 0x8c000000.
00:49:14 Quit ender` (Quit: A man without religion is like a fish without a bicycle.-- Vique's Law)
00:49:23 Quit DerPapst (Quit: Leaving.)
00:54:51TheSeven8.3mA :D
00:55:15TheSeven54MHz, idle, usb off, hdd off, backlight off, lcd off, codec off, undervolted
00:55:54TheSevennow we need to figure out how to make it actually do something useful at that current :)
00:56:22TheSeven11.8mA with the LCD powered on, that leaves us like 2mA to work with...
00:57:39LloreanHow much does the codec use then?
00:58:43 Quit Jerom (Quit: Leaving.)
00:58:56TheSevencertainly depends on the headphones and volume settings, but usually more than 2mA
00:59:19TheSeventhe electronics use barely anything, but it needs to supply some power to the headphones :)
01:00:11TheSeven~12mA with the LCD active
01:02:02TheSevenalso the hold switch being locked increases current consumption by ~4mA.
01:02:10TheSeventhere must be something wrong with that
01:03:40S_a_i_n_t+~4mA from applying hold!?!
01:03:58 Join krazykit [0] (
01:04:15TheSeventurning on the hold switch physically powers the clickwheel down
01:04:38TheSevenso we might need to change some gpio settings to prevent current from leaking through clamp diodes on the data lines
01:05:55*S_a_i_n_t wonders how different the hardware would be if the manufacturer gave a flying f**k about power consumption.
01:06:21TheSeven~14.5mA at full CPU load with LCD on
01:06:26***Saving seen data "./dancer.seen"
01:06:29TheSevenso CPU load at 54MHz only accounts for 2.5mA
01:06:54S_a_i_n_tthat's a fairly cheap LCD
01:17:14DEBUGEOF from server (Connection reset by peer) (snapshot: netstuff.c line 545)
01:17:14***No seen item changed, no save performed.
01:17:17***Started Dancer V4.16
01:17:17***Connected to on port 6667
01:17:17***Logfile for #rockbox started
01:17:17Mode"logbot :+i" by logbot
01:17:21***Server message 501: 'logbot :Unknown MODE flag'
01:17:21 Join logbot [0] (
01:17:21 Join Zambezi [0] (Zulu@
01:17:21 Join krazykit [0] (
01:17:21 Join S_a_i_n_t [0] (S_a_i_n_t@
01:17:21 Join enthdegree [0] (
01:17:21 Join robin0800 [0] (
01:17:21 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
01:17:21 Join factor [0] (~factor@
01:17:21 Join liar [0] (
01:17:21 Join stripwax [0] (
01:17:21 Join milk [0] (
01:17:21 Join markun_ [0] (~markun@rockbox/developer/markun)
01:17:21 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.
01:17:21 Join Horscht [0] (~Horscht@xbmc/user/horscht)
01:17:21 Join powell14ski_ [0] (
01:17:21 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
01:17:21 Join MaynardWaters [0] (
01:17:21 Join Galois [0] (
01:17:21 Join Dreamxtreme [0] (~Dre@
01:17:21 Join sasquatch [0] (~username@
01:17:21 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
01:17:21 Join toffe82 [0] (
01:17:21 Join eGen_ [0] (
01:17:21 Join MethoS- [0] (~clemens@
01:17:21 Join mudd1 [0] (
01:17:21 Join T44 [0] (
01:17:21 Join panni_ [0] (
01:17:21 Join mystica555 [0] (
01:17:21 Join CIA-70 [0] (~CIA@
01:17:21 Join Rob2222 [0] (
01:17:21 Join Raptors [0] (
01:17:21 Join dfkt [0] (dfkt@unaffiliated/dfkt)
01:17:21 Join einhirn [0] (
01:17:21 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
01:17:21 Join rasher [0] (~rasher@rockbox/developer/rasher)
01:17:21 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
01:17:21 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
01:17:21 Join Barahir [0] (
01:17:21 Join ThomasAH [0] (
01:17:21 Join parafin [0] (
01:17:21 Join ack` [0] (
01:17:21 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
01:17:21 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
01:17:21 Join YPSY [0] (
01:17:21 Join TBCOOL [0] (
01:17:21 Join logiclost [0] (
01:17:21 Join ranmachan [0] (
01:17:21 Join Farthen [0] (
01:17:21 Join @ChanServ [0] (ChanServ@services.)
01:17:21 Join MagusG [0] (
01:17:21 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
01:17:21 Join scorche|sh [0] (~scorche@rockbox/administrator/scorche)
01:17:21 Join pikytcus [0] (
01:17:21 Join amee2k [0] (
01:17:21 Join Elfish [0] (amba@2a01:4f8:100:90a1:abc:abc:abc:abc)
01:17:21 Join Kohlrabi [0] (
01:17:21 Join feisar- [0] (
01:17:21 Join Loto [0] (~nfs@xbmc/user/Loto)
01:17:21 Join crwl [0] (
01:17:21 Join preglow [0] (
01:17:21 Join yosafbridge [0] (
01:17:21 Join dionoea [0] (~dionoea@videolan/developer/dionoea)
01:17:21 Join jepler [0] (~jepler@emc/developer/pdpc.professional.jepler)
01:17:21 Join balintx [0] (
01:17:21 Join bzed [0] (
01:17:21 Join Dhraakellian [0] (
01:17:21 Join kkit|sh [0] (
01:17:21 Join AlexP [0] (~alex@rockbox/staff/AlexP)
01:17:21 Join guymann [0] (~charles@
01:17:21 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
01:17:21 Join ehntoo [0] (
01:17:21 Join n17ikh [0] (
01:17:21 Join Sochiro [0] (~Sochiro@
01:17:21 Join Battousai [0] (~bryan@gentoo/developer/battousai)
01:17:21 Join Quazgaa [0] (
01:17:21 Join jfc [0] (
01:17:21 Join Strife89 [0] (~Strife89@
01:17:21 Join jae [0] (
01:17:21 Join plux [0] (
01:17:21 Join froggyman [0] (~seth@unaffiliated/froggyman)
01:17:21 Join linuxguy3 [0] (
01:17:21 Join amiconn [0] (quassel@rockbox/developer/amiconn)
01:17:21 Join zu [0] (
01:17:21 Join Kitar|st [0] (
01:17:21 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
01:17:21 Join pixelma [0] (quassel@rockbox/staff/pixelma)
01:17:21 Join elcan [0] (
01:17:21 Join mordocai_afk [0] (~mordocai@
01:17:21 Join timccc [0] (~timccc@
01:17:21 Join Guinness [0] (
01:17:21 Join Slasheri [0] (miipekk@rockbox/developer/Slasheri)
01:17:21 Join alexbobP [0] (
01:17:21 Join Xerion [0] (
01:17:21 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
01:17:21 Join [fred] [0] (
01:17:21 Join literal [0] (
01:17:21 Join simonrvn [0] (
01:17:21 Join w0m [0] (~wom@
01:17:21 Join jordan` [0] (
01:17:21 Join scorche [0] (~scorche@rockbox/administrator/scorche)
01:17:21 Join iq [0] (~iq@unaffiliated/iq)
01:17:21 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
01:17:21 Join thegeek_ [0] (
01:17:21 Join Bagder [0] (~daniel@rockbox/developer/bagder)
01:17:21 Join niekie [0] (quasselcor@CAcert/Assurer/niekie)
01:17:21 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
01:17:21 Join blast007 [0] (~blast007@bzflag/developer/Blast)
01:17:21 Join GodEater [0] (~bibble@rockbox/staff/GodEater)
01:17:21 Join sbhsu [0] (~a6530466@
01:17:21 Join vedos [0] (
01:17:21 Join soap [0] (~soap@rockbox/staff/soap)
01:17:21 Join aevin [0] (eivindsy@unaffiliated/aevin)
01:17:21 Join merbanan [0] (
01:17:21 Join tmzt [0] (~tmzt@
01:17:21 Join Torne [0] (~torne@rockbox/developer/Torne)
01:17:21 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
01:17:21 Join Utchy [0] (
01:17:21 Join Rondom [0] (
01:17:21 Join Hadaka [0] (
01:17:21 Join pjm0616 [0] (~user@
01:17:21 Join maraz [0] (
01:17:21 Join ved [0] (
01:17:21 Join miceh [0] (
01:17:21 Join knittl [0] (~knittl@unaffiliated/knittl)
01:17:21 Join simabeis [0] (
01:17:21 Join FOAD [0] (~dok@
01:18:04 Quit robin0800 (Quit: Leaving)
01:19:01TheSevenCPU idling at 216 vs. 108MHz: 5.0mA difference
01:20:00 Quit toffe82 (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
01:20:02TheSevenAHB clock: 108 vs. 54MHz (guessed): 3.6mA difference
01:20:50TheSevenAPB clock: 54 vs. 27MHz (guessed): 0.3mA difference
01:22:27CIA-70r29267 build result: All green
01:22:53TheSevenCPU at full load at 216 vs. 108MHz: 4.2mA difference
01:23:29TheSevenCPU at 108MHz at full load vs. idle: 6.8mA difference
01:24:14S_a_i_n_tthat last ones...not what I would have expected.
01:25:35TheSevenCPU at 216MHz at full load vs. idle: 6.7mA difference
01:26:15TheSevenwhen closing that circle, we have 0.8mA of error somewhere :/
01:26:37 Quit dfkt (Quit: -= SysReset 2.53=- Sic gorgiamus allos subjectatos nunc.)
01:28:04TheSevenso in theory we should get an 8.6mA gain from boosting
01:28:30 Quit factor (Ping timeout: 240 seconds)
01:29:06TheSevener, FAIL!
01:30:24 Nick amee2k is now known as LoR3Z (
01:30:43CIA-70New commit by 03theseven (r29268): iPod Classic: Do boosting the right way round ;-)
01:30:54TheSevenat 33.0-33.5mA idle now
01:31:28 Nick LoR3Z is now known as amee2k (
01:31:36TheSevenplaying mp3 at ~35mA
01:32:23TheSeven~45MHz in use
01:32:39TheSeven(but the buffering debug screen probably accounts for quite a bit of that)
01:32:48 Quit pamaury (Remote host closed the connection)
01:32:56 Join sprotte24 [0] (
01:33:08 Quit timccc (Ping timeout: 250 seconds)
01:33:50TheSevenmain menu vs. WPS makes a ~1mA difference
01:34:06S_a_i_n_tthat's be screen updates?
01:34:33CIA-70r29268 build result: 5 errors, 0 warnings (theseven committed)
01:35:03S_a_i_n_tif you're using iLike, it'll probably be using the most whilst scrolling text, and/or when the "HDD" anim is on ;)
01:35:21S_a_i_n_tbut that's a pretty cheap theme to run.
01:35:48CIA-70New commit by 03theseven (r29269): Fix red (remove accidentally committed debugging code)
01:35:59S_a_i_n_tI've made silly themes that make a marked decrease in battery life.
01:36:43TheSevenS_a_i_n_t: the emCORE boot menu also makes quite a bit of difference
01:37:03TheSeven~6.5mA more current while it's scrolling
01:37:22S_a_i_n_t"'s pretty" ;)
01:37:40TheSeven80% of that not being the actual rendering, but pusing the data to the LCD!
01:38:18S_a_i_n_t...ok, wow.
01:38:41S_a_i_n_tbut, it's so pretty. :D
01:39:35CIA-70r29269 build result: All green
01:41:04TheSevenS_a_i_n_t: but on the other hand, the headphones volume makes like ~13mA difference
01:41:25TheSevenso apple can't even meet their 36h in theory, if you're listening at full volume
01:41:42TheSevenand why on earth is there very noticable clipping at -5dB!?
01:42:03*TheSeven remembers not getting any clipping up to 0dB on the classic
01:42:25TheSevener, wait. that would be really bad...
01:43:59TheSevenapple might have done a really bad cheat here
01:46:29S_a_i_n_tI am pretty sure that the runtime isn't measured at "full volume".
01:47:02S_a_i_n_tI've tried to do the math there on a few targets and it seemed sceptical that it was measured with full draw at the ehadphones.
01:47:50*TheSeven currently wonders where the codec gets its power from, and if it might have been hit by some undervolting
01:49:49 Quit stripwax (Quit:
01:52:09TheSevensomething can't be right
01:52:33TheSevenif i power down that LDO completely, the codec doesn't respond
01:52:43TheSevenat 0.9V, it responds, but there is no sound
01:52:52TheSevenat 1.0V, everything seems to be working
01:53:22TheSevenapple drives it at 1.8V, there there is no noticable difference in audio quality or the clipping limit
01:53:31TheSeven(always seems to be around -6dB)
01:53:59 Quit mudd1 (Ping timeout: 240 seconds)
01:54:00 Quit liar (Ping timeout: 255 seconds)
01:54:16TheSevenso what the heck is going on here?
01:54:29TheSevenis that the logic power supply, and the audio power supply is taken from somewhere else?
01:54:42TheSevenand if yes, from where? how did it get hit by undervolting?
01:54:56TheSevenit isn't taken from the memory or LCD regulator, i've already checked that
01:55:18TheSeventhe only other thing i've undervolted is the CPU core, which is in a voltage range where i really doubt that the codec could be abusing it
01:55:42 Join Keripo [0] (
01:58:05 Join dfkt|x [0] (
01:58:05 Quit dfkt|x (Changing host)
01:58:05 Join dfkt|x [0] (~dfkt@unaffiliated/dfkt)
02:03:24 Quit dfkt|x (Remote host closed the connection)
02:04:45 Quit sprotte24 (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
02:15:53 Quit MethoS- (Remote host closed the connection)
02:15:56 Quit Sochiro (Read error: Connection reset by peer)
02:20:44Unhelpfulkugel: you found very little information regarding embedded album art in ogg vorbis because vorbis tags are not a standardized metadata format, they're a labelling format and anything you do with them is ad-hoc.
02:21:41 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
02:35:14 Join afk [0] (Dre@
02:35:30 Quit afk (Read error: Connection reset by peer)
02:38:12 Quit Dreamxtreme (Ping timeout: 240 seconds)
02:40:25 Join Dreamxtreme [0] (~Dre@
02:42:25 Join qurvel [0] (
02:46:17 Quit Keripo (Quit: Leaving.)
02:48:57 Join Keripo [0] (
02:50:46 Quit Keripo (Client Quit)
02:51:32 Join antil33t [0] (
02:55:28 Quit krazykit (Ping timeout: 250 seconds)
02:56:41 Join Keripo [0] (
02:57:25S_a_i_n_tAny idea what information to display on a .wps when there's no AA present, and no next track present, keeping in mind that track, album, artist, genre, year, bitrate/VBR, codec and sample rate are already displayed?
02:58:37S_a_i_n_tHaving both no AA *and* no next track is reasonably unlikely...I suppose, but it looks barron enough to warrant displaying more info in the event it does happen, and adding another conditional UI viewport to center the info onscreen in the case needlessly complicates the code.
02:59:04S_a_i_n_tit's simpler for me to just add some "filler info", but I'm not sure I still wish it to be of value.
02:59:48*S_a_i_n_t requests Llorean's analytical mind.
03:00:45JdGordon|how about a massive "PLaylist finished"! splash?
03:00:54JdGordon|or there is a good spot for your easter egg
03:01:08S_a_i_n_thaha ;)
03:01:41S_a_i_n_tno, I'd like to base the easter egg around horrible dsp/eq settings.
03:01:44JdGordon|I fail to see the correlation between aa being present and next track being avilable
03:02:12S_a_i_n_tthere isn't one, but, in those instances the screen is left barron.
03:02:15 Join Xerion_ [0] (
03:02:25 Join krazykit [0] (
03:02:38S_a_i_n_tno AA, and no next track info...on a pretty large screen. just looks crap.
03:02:52S_a_i_n_tSo I want to suppliment more info about the present track.
03:02:56LloreanS_a_i_n_t: I think something like "JdGordon" suggested that makes it clear the playlist is about to end might be nice.
03:03:30S_a_i_n_tthere's already a "X of X in Playlist" thing.
03:03:36LloreanBut you'd want a conditional for the "move to next folder" option so that you could conditionally display something like "no more files" vs "folder is about to change"
03:03:42 Quit Xerion (Ping timeout: 240 seconds)
03:03:43 Nick Xerion_ is now known as Xerion (
03:03:54LloreanThe X of X thing isn't always accurate if you've got folder based playback going. :)
03:03:55S_a_i_n_twhich is smart enough to know when there's only one track, so doesn't do 1 of 1, etc.)
03:05:47S_a_i_n_tHmmm...perhaps I could display time and date info.
03:05:57S_a_i_n_tthat doesn't pop up in the .wps anywhere else.
03:06:38S_a_i_n_tI really just need "something" to fill a gap, as that is simpler than re-writing code to make another conditional UI viewport just for this instance to center the info.
03:07:05S_a_i_n_tbut, that would be the "smart" thing to do, but the screen would still look very empty.
03:07:14 Quit Rob2222 (Ping timeout: 240 seconds)
03:09:00LloreanA large font countdown timer to the end of the playlist? :-P
03:10:09 Join Choicefresh [0] (~Liam@unaffiliated/choicefresh)
03:10:11S_a_i_n_tin the case of No AA, it already uses 35pt font with vastly increaded ascent and descent, and it still looks small ;)
03:10:42S_a_i_n_tI'd need to make a larger font by hand just for this case.
03:11:55LloreanI think having a clock / date only show up when the playlist is about to end seems a little bit odd. It's a situation where the user, if they ever see it, isn't going to see it very often in the future and wonder if it's a bug or something.
03:12:07LloreanI actually like the date/time to be visible all the time anyway
03:12:30LloreanMaybe you could alternate date/time and next track info, and then just show date/time if there's no next track?
03:13:07S_a_i_n_tSo do I, but without going to the extent of recreating a larger "built in" (imatation thereof) status bar, there's no place for me to include the clock/date.
03:13:22S_a_i_n_tIf I did use the built in bar, I could not read it on device.
03:13:54LloreanWhat about alternating sublines with date/time alternating with next track info, so that the user's used to seeing it down there. Then the lack of next track info delivers a bit more of a message.
03:14:18Choicefreshhey, i asked this question yesterday but went to sleep before i got an answer −− if i upgrade from the stable release to the current build, will it work with more accessories?
03:14:21S_a_i_n_tHmmm, that's a reasonable idea.
03:14:39LloreanChoicefresh: Probably not, I don't recall seeing any significant changes to accessory support since the last release.
03:14:43S_a_i_n_tChoicefresh: Unlikely.
03:15:03S_a_i_n_tThere's usually no real reason to *not* use the current build, though.
03:15:04LloreanS_a_i_n_t: I haven't seen your theme, so I don't know how bad that'll look. :)
03:15:28S_a_i_n_tI'll post some screenshots shortly.
03:16:20 Quit thegeek_ (Read error: Connection reset by peer)
03:16:31S_a_i_n_tit's kind of an "enhanced cabbie"
03:16:42ChoicefreshIs there anywhere where I can find a list of differences or a changelog between the release and the current build? I think I'm pretty satisfied with the release, except for battery life and accessory compatibility problems.
03:17:19***Saving seen data "./dancer.seen"
03:18:52 Quit qurvel (Ping timeout: 276 seconds)
03:21:00 Quit BHSPitMonkey (Ping timeout: 255 seconds)
03:22:02S_a_i_n_tI see we have embedded AA in SVN now...does anyone know the order of preference over embedded vs. "usual" AA?
03:22:41S_a_i_n_tI mean, will tracks that have embedded art suddenly start displaying that in preference over my "Folder.jpg" etc.?
03:24:13 Join timccc [0] (~timccc@
03:24:24 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
03:24:42S_a_i_n_tI think it would also be nice if searching for/using embedded AA could be disabled by the user.
03:25:44S_a_i_n_tSo that those that maintain their own album art aren't hit with a needless check for embedded art. (I assume these checks would add up over time)
03:26:56Strife89S_a_i_n_t: It prefers the embedded AA.
03:27:16S_a_i_n_taaaaaaw! that sucks!
03:27:48S_a_i_n_twell...I guess it's all good, but, I have to make sure to strip embedded AA in this case.
03:28:15S_a_i_n_tI find per-track AA simply wasteful.
03:30:17S_a_i_n_tfor large libraries, embedded art (especially particularly large images) can add a lot of space back if removed.
03:31:20 Quit Keripo (Quit: Leaving.)
03:34:02S_a_i_n_tI assume the manuals haven't been built yet to reflect the change in supporting embedded AA?
03:39:04Choicefreshso SVN is the changes in the current build?
03:41:08 Join Keripo [0] (
03:49:23S_a_i_n_tChoicefresh: Correct.
03:50:19S_a_i_n_tif you're using the latest release build, then "Major changes since version 3.7 (in SVN)" is what concerns you re: updating
03:51:33LloreanS_a_i_n_t: Did the patch include a manual patch?
03:51:44LloreanDid the *AA* patch...
03:52:20S_a_i_n_tLlorean: JdGordon|: Strife89: froggyman: others: <−− screenshots (random) of my "enhanced-touch-cabbie)
03:52:43S_a_i_n_t*top right clearly shows the "barron no AA/No next track" case.
03:53:04S_a_i_n_tLlorean: yes, it does include a manual patch.
03:53:05JdGordon|those arrows really look horrible
03:53:19LloreanWhat *are* those arrows?
03:53:21S_a_i_n_tcome up with a better solution ;)
03:53:35S_a_i_n_tLlorean: next/Prev/ffwd/rwd
03:53:44LloreanAh, touchscreen.
03:54:03JdGordon|S_a_i_n_t: honestly, the way svn cabbie does it is better than your way
03:54:08JdGordon|those arrows just look hideous
03:54:09LloreanI'd just shift the AA to a side, and have a vertical column of "control" icons on the right of it.
03:54:12LloreanRather than centered AA
03:54:14JdGordon|(nott hat i like the hidden controls)
03:54:18 Join Zarggg [0] (
03:54:29S_a_i_n_tJdGordon|: I really dislike that popup.
03:54:48Strife89S_a_i_n_t: That's NICE.
03:54:49JdGordon|as do i
03:54:50S_a_i_n_tthough, I'm thinking about only displaying these arrows when the screen is touched.
03:54:56LloreanI also don't really like the large, widely spaced font in the menu. That spacing just looks really awkward to me.
03:55:08S_a_i_n_tLlorean: it is necessary.
03:55:14LloreanFor what?
03:55:17S_a_i_n_tfor "touch friendlyness"
03:55:49LloreanYou can't make the letters fill the whole space using a larger font, though?
03:56:05LloreanThe lines can be just as big, just actually "full", right?
03:56:13JdGordon|S_a_i_n_t: also, the QS really looks like ass :p
03:56:21S_a_i_n_tif the font is that large, then menu entries get get off and need to scrool...which looks crap.
03:56:32S_a_i_n_tthe added ascent/descent is really the only option.
03:56:57Llorean"looks crap" vs "looks crap"
03:57:04S_a_i_n_tJdGordon|: you know why? because you need to add quickscreen to %cs ;)
03:57:13JdGordon|i know :)
03:57:25*JdGordon| might just do that now
03:57:27LloreanHow touch unfriendly is it really?
03:57:57LloreanIt sounds like our touch precision is pretty bad then?
03:58:00S_a_i_n_teven with the lists that wide, it's still hard to hit a single option without a stylus
03:58:11S_a_i_n_tand no, the screen is fine in the OF.
03:58:17S_a_i_n_tI believe the failing is in RB
03:58:22LloreanAs I said, *our* precision.
03:58:38S_a_i_n_ttouchscreen is usable, but hardly "precise"/
03:58:42LloreanHow does touch work in a list like that.
03:58:55LloreanCan you hold your finger down, move it around, see it live-highlight what you hover over, then release to select?
03:58:56S_a_i_n_thow do you mean?
03:59:17S_a_i_n_tno, it only highlights the "last touched" list entry.
03:59:24LloreanAh, a little frustrating.
03:59:34LloreanSo you can't see what you're highlighting without actually invoking it?
03:59:40S_a_i_n_tif you're scrolling, it doesn't highlight what's under your finger...unfortunately.
04:00:56S_a_i_n_tholding your finger down and scrolling just moves the list up/ doesn't actually hightlight any selection, no.
04:01:31*JdGordon| thinks the huge line height looks pretty ugly, but agrees that when its the font height its not easy to use
04:01:41LloreanIt'd be nice if your delta was low enough, it treated your finger as a pointer, and only scrolled if you moved quickly, or moved to the "end" of the visible list in either direction.
04:01:48S_a_i_n_tthe list item touched before scrolling is briefly highlighted, but the action of scrolling cancels the hightlight.
04:02:10JdGordon|S_a_i_n_t: why dont you just invisibly split the top half off the wps into a |<< and >>| button?
04:02:22 Quit panni_ (Read error: Connection reset by peer)
04:02:29LloreanWould it be possible to get the font at least a little bit larger. So maybe only long menu items scroll? Basically, get it as large as possible without scrolling "Resume Playback"
04:02:32LloreanOr "Now Playing"
04:02:34Lloreanwhichever is longer.
04:02:48S_a_i_n_tI thought about that...but I'm not a fan of touch areas that aren't obviously marked, or "buttons".
04:03:24S_a_i_n_tLlorean: there's lots of longer menu entries than those ;)
04:03:41LloreanYes, but they're seen less often.
04:03:46LloreanWhereas that screen is seen *constantly*
04:03:48LloreanAs is the files list.
04:03:48S_a_i_n_tthe height the font is now, almost no menu entry needs to scroll...I can use the lists, and it's perfectly readable on device.
04:04:01LloreanSo having them not look like crap vertically is better than not looking like crap on less often seen screens horizontally
04:04:32 Quit Keripo (Quit: Leaving.)
04:05:15 Quit MaynardWaters (Remote host closed the connection)
04:05:24S_a_i_n_tI personally think that the added ascent/descent looks better than a larger font with list items spaced closely together, but, I add "personally" as I realise it's quite subjective.
04:05:55 Quit pixelma (Disconnected by services)
04:05:57 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:05:59 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:06:01S_a_i_n_tlists in the Andoid UI also have a similar sytle.
04:06:19S_a_i_n_t(lots of room between list entries)
04:06:42*Llorean shrugs.
04:06:52 Quit amiconn (Disconnected by services)
04:06:53 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:07:11 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:07:22S_a_i_n_tI did try the same font with no added ascent of descent, and it looked crap...I also tried a slightly larger font, with less ascent/descent added...and (to me) it also looked crap.
04:07:29LloreanAre you considering posting this as an optional theme, or is this proposed as a default theme version for the Android devices?
04:07:36 Quit TheSeven (Ping timeout: 246 seconds)
04:08:00 Join Barahir_ [0] (
04:08:24S_a_i_n_tI'm considering this as an optional theme, but I am hoping to create a version of it that could in theory become the default for touch devices of this resolution.
04:08:32LloreanIf it's your own thing, I don't really care much and was just offering my opinion on how it looks. Style is subjective, and it's yours. If it's a potential update to the default for touch use, other than having a lot of new features that should be ported to other ones, the spacing isn't in line with the current design and changes in the actual aesthetic need discussion.
04:08:52LloreanSo, for your version - fine. For a default - probably worth trying to make it "look" more like the other Rockbox cabbies.
04:09:45S_a_i_n_tI'd rather go to the extent of finding a middle ground...and updating the existing touch cabbies to look more like it ;)
04:10:09S_a_i_n_tand yes, I would like to port some of the newer touch features to existing touch cabbies
04:10:16LloreanAs I said, you'd have to talk about actually changing the chosen default theme design then.
04:11:34S_a_i_n_tStrife89: most convincing response I've had so far, btw, thanks ;)
04:12:05 Quit Barahir (Ping timeout: 276 seconds)
04:13:39 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:17:43 Quit Kitar|st (Ping timeout: 250 seconds)
04:19:34 Join Keripo [0] (
04:23:03 Join Kitar|st [0] (
04:24:19 Quit saratoga (Ping timeout: 245 seconds)
05:03:26 Quit t0rc (Remote host closed the connection)
05:05:02 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
05:16:19 Quit Llorean (Read error: Connection reset by peer)
05:17:22***Saving seen data "./dancer.seen"
05:18:59 Quit t0rc (Quit: Give someone code, help them with one project. Teach someone to code, help them rule the world.)
05:21:38Strife89S_a_i_n_t: Is that for the Android port, or something else? Whatever it's for, I'd get it JUST to use that theme
05:22:16 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
05:22:22RaptorsAre there any devs online?
05:22:33RaptorsIs there a version for the Fuze+ in the works?
05:22:45Strife89Not at the moment, IIRC
05:23:06Raptors:( The hardware looks sick but people are complaining about the fireware
05:23:26Strife89If someone gets interested enough, work may begin one of these days.
05:23:33Strife89Ports take time, though.
05:24:39Strife89Personally, I'm happy with an 8GB Fuze and a big microSD
05:25:06RaptorsI like my fuze too but I want a better screen
05:25:48Strife89There IS an iPod Classic port in the works, but it's considered Unusable right now.
05:26:28RaptorsI am anti-apple so I'm not going to get that
05:26:39Strife89If you own an Android device with a good screen (and are willing to use an Unusable-grade port), there's also that.
05:26:48RaptorsI guess I'll hope someone will make it...
05:37:55 Quit Horscht (Quit: Verlassend)
06:00:29 Quit milk (Ping timeout: 245 seconds)
06:05:43 Join milk [0] (
06:09:28 Quit milk (Client Quit)
06:18:53 Join factor [0] (~factor@
06:34:15 Join sideral [0] (~sideral@rockbox/developer/sideral)
06:38:40 Quit sideral (Client Quit)
06:52:58S_a_i_n_tJdGordon|: jsyk, I have already prepared the "Quickscreen" graphic for the theme ;)
06:53:37JdGordon|oh have you?
06:54:18S_a_i_n_tyes, I did it after:
06:54:19S_a_i_n_t[15:57] <JdGordon|> i know :)
06:54:20S_a_i_n_t[15:57] * JdGordon| might just do that now
06:54:21DBUGEnqueued KICK S_a_i_n_t
06:54:21S_a_i_n_t :P
06:56:20JdGordon|so... guess what!?!
06:56:25CIA-70New commit by 03jethead71 (r29270): buffering: Don't execute move-handle-ony case if handle is of metadata type (atomic) and must be kept fully buffered. Manage handle corruption guard ...
06:57:01S_a_i_n_tI tried looking at a "better" way to do %cs, so things like that guy on the forum was talking about with a tag that goes true for X after entering a specific screen...but I couldn't really understand how to add other screens to the currecnt scheme. :/
07:00:16CIA-70r29270 build result: All green
07:01:06JdGordon|S_a_i_n_t: if a proper screen system happens, and enough screens are fixed for it, then a "just got to this screen" tag would be trivial
07:01:19JdGordon|but i have a feeling it isnt very useful except for fm and wps
07:01:55S_a_i_n_tI imagine it could be used in the .sbs, but...only for "bling".
07:02:05S_a_i_n_tit would add little practical value there.
07:03:57S_a_i_n_twho should I bug about a syntactical error in the manual?
07:04:24JdGordon|AlexP: or pixelma prob
07:05:18JdGordon|syntactical, you mean bad english or latex?
07:06:02S_a_i_n_thaha...nevermind, I only imagined I was looking at the manual. apparently I clicked the wrong link and it's CustomWPS instead.
07:06:08S_a_i_n_tI'll change it myself ;)
07:06:57 Quit JdGordon| (Quit: leaving)
07:17:26***Saving seen data "./dancer.seen"
07:28:38 Join mudd1 [0] (
07:38:52 Join liar [0] (
07:59:05 Quit mudd1 (Ping timeout: 276 seconds)
08:01:06Choicefreshderp. i'm trying to install a theme onto my rockbox via rockboxutility, and it seems to install correctly, but it doesn't show up in the rockbox theme list nor in the rockbox info panel in rockboxutility.
08:01:19 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
08:02:59 Join LinusN [0] (
08:02:59 Quit LinusN (Changing host)
08:02:59 Join LinusN [0] (~linus@rockbox/developer/LinusN)
08:04:13 Join sideral [0] (~sideral@rockbox/developer/sideral)
08:05:14 Join mudd1 [0] (
08:13:26 Join B4gder [0] (~danielx@2a00:1a28:1200:9::2)
08:13:26 Quit B4gder (Changing host)
08:13:26 Join B4gder [0] (~danielx@rockbox/developer/bagder)
08:13:48 Join bluefoxx [0] (
08:13:53 Quit rasher (Read error: Operation timed out)
08:15:16 Quit sideral (Quit: Leaving.)
08:15:24 Join sideral [0] (~sideral@
08:15:25 Quit sideral (Changing host)
08:15:25 Join sideral [0] (~sideral@rockbox/developer/sideral)
08:27:53 Join rasher [0] (~rasher@rockbox/developer/rasher)
08:29:51 Quit BHSPitMonkey (Remote host closed the connection)
08:36:50 Join ender` [0] (
08:39:39Choicefreshany ideas on how to fix the problem i'm having?
08:43:05Choicefreshi guess i'll figure it out tomorrow. good night
08:43:22 Join bertrik [0] (
08:43:22 Quit bertrik (Changing host)
08:43:22 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
08:43:37 Part Choicefresh
08:53:51 Join Guest84580 [0] (
08:53:52 Quit Guest84580 (Changing host)
08:53:52 Join Guest84580 [0] (~bjst@rockbox/developer/Zagor)
08:54:21 Nick Guest84580 is now known as Zagor (~bjst@rockbox/developer/Zagor)
09:01:53 Join pamaury [0] (
09:01:53 Quit pamaury (Changing host)
09:01:53 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:13:34 Join n1s [0] (~n1s@rockbox/developer/n1s)
09:17:27***Saving seen data "./dancer.seen"
09:20:46sideralAre genre tags (as stored in id3->genre_string) guaranteed to be single words, or is it possible to have multiple genres in there (and if so, comma- or space-separated)?
09:21:36Zagorthe genre tag is chaos. expect anything.
09:21:46Zagoror, more common, nothing :-)
09:22:29sideralAlright Zagor, thanks! That's what I was afraid of...
09:23:38 Join efyx [0] (
09:25:07 Join JdGordon1 [0] (
09:27:14Slasherisideral: many genre names contain multiple words anyway
09:28:17Slasherii think there are thousands of different genres, subgenres and so on
09:29:07B4gderand two people won't even agree in which genre a single song would be in
09:29:13Slasheriindeed :)
09:29:15B4gdergenre is just completely broken
09:29:23JdGordon1most songs dont fit in only one genre
09:29:36JdGordon1i mean, they change genre mid track
09:29:59sideralif there are multiple genres, how are they separated typically? Commas?
09:30:02 Quit bertrik (Ping timeout: 240 seconds)
09:30:07B4gderand for people who are into that particular music style, they would still miss the accurate or precise genre they'd want
09:30:11sideralCan there be a space in a genre name?
09:30:36n1ssideral: yes
09:31:43 Join hebz0rl [0] (
09:31:46Zagor"christian rock" or some such
09:31:50Zagor"death metal"
09:31:55Slasherisideral: it depends how those words are separated. Such as Folk Metal, Death Metal, Black Metal, Nu Metal and so on
09:32:18LinusNpetur's h120 finally runs rockbox again!!!
09:32:26ZagorLinusN: yay!
09:33:23sideralMy questions relate to Llorean's idea of matching only whole genre names with autoresume. My idea would be to take the search string, do a strstr on id3->genre_string, and then look for non-alpha characters left and right of the match.
09:34:41sideraldo a search string "death metal" would match "podcast,death metal" but not "death metallic"
09:35:33ZagorI would say using genre for _anything_ is bound to result in dissappointment
09:35:54B4gdernobody uses genre
09:36:03B4gderor everybody sets their own
09:36:07B4gderits just a mess
09:36:12Zagorexcept for a very small section of really anal users who write every tag on every track themselves
09:36:44sideralI believe there are quite a few people who define their collection of podcasts to be the files who have a (manually set) "podcast" genre
09:36:55B4gderI doubt that
09:37:00sideralI think some podcasting software offers doing that automatically
09:37:05B4gderI think most people just download podcasts
09:37:16B4gderand if the genre is set, then sure it works
09:37:21ZagorI think path is generally much safer than genre
09:37:24sideralBut I think you're right, that might be useless complexity
09:37:34Zagorand still rather unsafe
09:37:39sideraland all podcasts end up in the same folder anyway
09:38:51Slasherisideral: if you are planning to add some genre searching to the db, i would instead suggest of adding one 4 byte integer to every track, that could match the track to 32 different "genres" or whatever sections user wants to put the track in
09:39:30Slasherithen user could give names to each of those 32 sections and for every track use a multi selection to specify those sections where the track should be in (from wps directly)
09:39:44Slasherii would like that kind of searching capability
09:39:58sideralSlasheri: nice idea, but that wasn't the goal. I'm trying to improve the resume-on-automatic-track-change filter
09:42:03sideralI think I'll just allow configuring a list of directories (with full path names, as suggested by Llorean) and get rid of the genre-matching stuff
09:42:28sideralI love discussions that end up chopping features :)
09:43:25sideralOn a related note:
09:43:25sideralWhen I change the setting to match complete words only, I'd like to rename it from “Custom path/genre substrings (comma-separated)” to “Custom paths (comma-separated)”. If I manage to do this before the 3.8 feature freeze and before any translator has catched up with the previous name, can I just change the language string or should I properly deprecate it?
09:43:25sideral(I assume that each deprecated string takes up some resources in the final binary; let me know if this is not the case.)
09:44:24siderals/complete words/dirnames/
09:48:42n1sa deprecated string takes 3 bytes IIRC
09:49:38n1ssince the meaning changes i think deprecating and creating a new string is nicer but it's not a big deal
09:49:53 Quit factor (Read error: Connection reset by peer)
09:50:42 Join factor [0] (~factor@
09:50:48Zagorfor such a recent commit I don't think changing the meaning is a big deal
09:51:55n1sdid any translations pick it up?
09:52:02sideralnot yet
09:52:12n1sthen it's fine to just change it
09:54:51 Quit Keripo (Quit: Leaving.)
09:54:57sideralCool. Thanks guys!
09:54:57sideralI'll do the lang string change right away (before I adapt the implementation) to minimize the risk of anyone picking it up, and follow up with the implementation tonight after some testing. (The new string matches the current implementation as well (as it does strictly more then the new implementation), so there will be no inconsistency.
09:55:25 Join wodz [0] (
09:56:45wodzI think I understand how rockchip 'update' firmware format. The code to handle this in SDK is a *BIG* mess.
09:57:33sideralAh, but I'll use "directory" rather than "path" as that seems to be the RB standard name
09:57:37Zagorwodz: nice
09:58:25 Join petur [0] (~petur@rockbox/developer/petur)
09:58:59wodzZagor: Do we care much about 'clean room' reverse engineering aproach?
09:59:56 Join Keripo [0] (
10:00:23 Quit JdGordon| (Ping timeout: 255 seconds)
10:00:25Zagorwodz: we care very much about not copying any code
10:00:37wodzthat is clear
10:00:42 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
10:00:50Zagorhowever reading the sdk code and doing a clean reimplementation is not a problem
10:00:51 Quit Keripo (Client Quit)
10:00:56wodzZagor: ok
10:03:09sideralQuick question to a native speaker: In a menu "Resume on automatic track change", does "Custom directories (comma-separated)" sound acceptable and self-describing enough?
10:03:49CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
10:03:49*S_a_i_n_t doesn't grasp what the "comma seperated" is for.
10:04:17sideralyou can specify multiple directories, separated by commas
10:04:59S_a_i_n_tis this a list of dirs that will always resume?
10:05:40sideralIt's the list of dirs in which a track will be resumed even if it's played as the next track in automatic track change
10:05:59 Quit alexbobP (Read error: Operation timed out)
10:06:08 Join alexbobP [0] (
10:06:38sideralmanual track selection will always attempt to resume all files when autoresume has been enabled
10:06:41S_a_i_n_tthat seems a terribly obscure setting to me ;)
10:08:04S_a_i_n_tif manual track selection will always resume...then, what's the need for this selection of folders?
10:08:23S_a_i_n_tif a resumable track is encountered in a playlist, is it *not* resumed automatically?
10:08:37sideralby default, no
10:08:45S_a_i_n_tall these resume related settings are getting the better of used to be so simple ;)
10:08:56sideralyou can enable it globally or customize the dirs in which this will happen
10:09:20S_a_i_n_twouldn't it be better to have a key file for this, than a setting?
10:09:27S_a_i_n_tIMO, it would.
10:09:45S_a_i_n_tsimilar to database.ignore/unignore
10:09:51sideralthis is to support playlists of podcasts as well as playlists of audiobook chapters
10:09:59S_a_i_n_t"always.resume" or similar.
10:10:19S_a_i_n_tit seems to me that the folders would be relatively "set and forget".
10:10:25sideralSaint: Yes, that has been suggested a few times. The problem is that we'd have to look for this file for each track playback
10:10:40S_a_i_n_tis that a problem?
10:10:56sideralI think so
10:11:21S_a_i_n_twell, it's only going to look for this file if you specifically enable resume, no?
10:11:31S_a_i_n_tso, you're only taking a hit if you've chosen to.
10:11:57S_a_i_n_tiiuc it wouldn't affect those that don;t use it with a random spinup/read.
10:12:15sideralI think it's better to not take that hit, and to configure it in the configuration settings where it belongs IMHO
10:12:53S_a_i_n_tI just don't think it's something that will get set enough to warrant crowding the menu.
10:13:28sideralI agree the performance hit wouldn't be too serious, as noting the existence of this file probably could be cached pretty well
10:13:29S_a_i_n_tI think placing a file manually in the chosen dirs is a rather elegant approach, actually.
10:13:59*S_a_i_n_t shrugs
10:14:18sideralI think users want to find all settings belonging to a feature in one place.
10:14:32S_a_i_n_thow are the dirs entered into this "comma seperated list"?
10:15:00sideralwith the text editor currently. AFAIK we don't have a UI for selecting (multiple) dirs
10:15:04S_a_i_n_tif it involves the vkeyboard, I can't see it getting *too* much use...that thing is a bitch.
10:15:18S_a_i_n_tanother reason why I think a keyfile is more elegant.
10:15:27S_a_i_n_tmanually typing dir names is just...painful.
10:15:34 Join Keripo [0] (
10:15:48S_a_i_n_tusers will probably edit the .cfg manually anyway...which is why I don't see placing a kayfile as a big deal.
10:16:51S_a_i_n_tbasically, not to discredit you...but I just see it as needlessly complicated to do something that could be done "under the hood".
10:16:56sideralI understand you reasoning. But to me the keyfile approach is inferior because it makes the feature less discoverable
10:16:59 Join JdGord [0] (~jonno@
10:17:32S_a_i_n_tI think the setting is aimed more towards those that would be comfortable doing it this way...but, that's subjective I guess.
10:18:16 Quit mc2739 (Read error: Connection reset by peer)
10:18:21S_a_i_n_tIf I had to type dir entries manually to achieve this in a working state, I'm (personally) just going to say "stuff this" and edit the config manually.
10:18:41S_a_i_n_tI'd be wlling to guess I wouldn't be the only one.
10:18:44sideralWhich is cool, at least you have discovered the config option :)
10:20:52sideralAnd to make it more discoverable, I'm looking for a good description. You seem to think "Custom directories (comma-separated)" does not cut it? :)
10:21:26S_a_i_n_tI think that it will be very hard to accurately describe what this does without it reading like a novel ;)
10:22:00S_a_i_n_tto me, "Custom directories (comma-separated)" gives no indication of what it actually does.
10:22:22S_a_i_n_tperhaps it may, in taken with the context of the menu it is located within, but on it's
10:22:33sideralIn a menu "Resume on automatic track change", mind you
10:23:00 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
10:24:42S_a_i_n_tI think it's one of those setting where it's not really possible to give it a name that will clarify it's function without already knowing of it's existence and operation.
10:25:15sideralPerhaps "In custom directories only" would be better, with a default that makes it obvious there can be a comma-separated list, such as "/podcast,/podcasts"
10:25:18S_a_i_n_t"Custom directories (comma-separated)" is as good as anything I guess, but if you've no idea what it does it's incredibly vague.
10:26:03S_a_i_n_thow does "/podcast,/podcasts" make it obvious there's a comma seperated list?
10:26:15sideralthere's a comma in it? :)
10:26:33S_a_i_n_tah...crappy font, couldn't see it ;)
10:27:57S_a_i_n_tIf you do put in an example dir(s), make sure to name them so that they won't be likely to apply resume to a dir that already exists without the users knowledge of the setting.
10:28:10S_a_i_n_t"/foo,/bar,/baz" etc.
10:28:36S_a_i_n_tas opposed to "/music,/podcasts"
10:29:10sideralWell, by default the setting is off, and canceling the text editor does not change it. So I think a useful default would be good. And it does actually make sense to enable this for podcasts only
10:29:40gevaertsLinusN: thanks!
10:30:13S_a_i_n_tsideral: though, you can't really assume to know what dir structure podcasts will be in.
10:30:25S_a_i_n_tthey won't necessarily be in /podcasts.
10:30:44S_a_i_n_tI think an example dir is better to be just that, an example unlikely to exist.
10:31:02sideralRight. That's one of the reasons the setting is off by default, and the user is put to the text editor right away when selecting it
10:31:45gevaertsAlexP, soap: can you ban maxmara1?
10:34:25sideralSaint: Hmm, I'll ponder this some more throughout the day, but I think I tend towards giving a useful default.
10:34:52sideralAnyway, thanks for having this discussion, it was extremely useful!
10:38:02CIA-70New commit by 03sideral (r29271): autoresume: Simplify configuration option for resume on automatic ...
10:38:37sideral(as I said, this is the quick lang string fix only, the implementation will be cleaned up tonight)
10:42:13CIA-70r29271 build result: All green
10:48:07 Quit JdGord (Ping timeout: 240 seconds)
10:48:45 Join DerPapst [0] (
10:52:33n1ssideral: what happens if the path contains a comma?
10:54:08sideraln1s: you won't be able to match such paths with this feature
10:54:37n1sisn't it better to use a char that isn't allowed in filenames?
10:54:49n1suse it as a separator
10:55:55 Quit mc2739 (Ping timeout: 240 seconds)
10:56:13 Join sprotte24 [0] (
10:57:03sideraln1s: that's a good idea, now that I've removed "comma-separated" from the description (which I liked better than "semicolon-separated", for example, because it was shorter and better understandable)
10:57:31sideralwhich character do you reckon?
10:57:43sideralis ":" still illegal on FAT?
10:58:45n1si'd go with : or |
10:59:16n1s static const char invalid_chars[] = "\"*/:<>?\\|";
11:01:05sideralSo that's why copying my .ogg albums over to my MP3 player always chokes up on funny names... FAT sucks ;)
11:01:10sideralOK, ":" it is
11:01:17Zagor: is a pretty standard path separator
11:01:49sideralOn Unix, yes. Not sure what Windows uses these days
11:02:18Zagoroh, they use ;
11:02:38Zagorright, duh, since they use : for drive name.
11:03:34 Join qurvel [0] (
11:04:26 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
11:06:01sprotte24I just bought a Sandisk Sansa Fuze. I am interested in changing to the rockbox firmware.
11:06:02sprotte24My question:
11:06:04sprotte24The original Sandisk firmware uses nice pictures for the main menu options like "Music", " "FM radion", etc.
11:06:05sprotte24I don´t find this in the rockbox manual. Does this mean that the rockbox does not have such niche menues for selection?
11:06:19Zagorrockbox looks very different, yes
11:07:02sprotte24I did not find any main menu picture in the rockbox manual?
11:07:18sprotte24only text screens?
11:07:59Zagorthe rockbox GUI is not icon focused, no. it's built around menus.
11:09:29sprotte24thx. why not, do pictures waste too much memory space?
11:10:00sprotte24pictures are much better to recognice than litte text lines.
11:10:07Zagornot at all
11:10:16Zagoricons always have to be explained by text
11:10:18n1ssprotte24: that is of course you opinion
11:10:31CIA-70New commit by 03theseven (r29272): iPod Nano 2G: Fix current leak through clickwheel GPIOs when clickwheel is powered down
11:10:53Zagorand we have some pretty complex items in our menus that you'd have to scratch your head for a while to come up with an icon for
11:11:07 Quit wodz (Quit: Leaving)
11:11:52Zagorbut the main thing is that we're programmers, not graphics artists. when we add a new feature, adding a text menu entry is infinitely easier and quicker than drawing a new icon.
11:11:59*TheSeven expects r29272 to add another 5% battery runtime :)
11:12:50sprotte24Is there any risk to install rockbox on my Sandisk Sansa Fuze (firmware V01.02.26A)?
11:13:34 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
11:13:38sprotte24Are there experiencesd that a user cannoot remove rockbox if he do not linke rockbox?
11:13:57CIA-70r29272 build result: 10 errors, 0 warnings (theseven committed)
11:14:11Zagorsprotte24: no, it's very safe
11:14:22n1ssprotte24: you can remove it, and even dual boot when rockbox is installed, the manual explains these things
11:15:00TheSevenbut, of course, as usual, no warranty for anything, you do everything at your own risk
11:15:39sprotte24I can sitch between the old firmware and rockbox at any time with dual boot?
11:15:45 Join GeekShad0w [0] (
11:16:07 Quit factor (Read error: Connection reset by peer)
11:16:30CIA-70New commit by 03theseven (r29273): Fix red
11:16:32sprotte24I think, installing rockbox is the only way to really test it and judge the quality
11:17:10TheSevensprotte24: AFAIK yes, but I don't have a Fuze
11:17:28***Saving seen data "./dancer.seen"
11:17:36n1ssprotte24: as i said, it's in the manual
11:17:59sprotte24n1s: thx, I will read tis.
11:19:54CIA-70r29273 build result: 10 errors, 0 warnings (theseven committed)
11:20:21sprotte24On my orig. Sansa Fuze I have no selcetion for deside to stop play after one record, it continues with the next music. Do I have a choice on the rocket box to stop after the end of a record?
11:22:46CIA-70New commit by 03theseven (r29274): Fix more red. Two had been hiding.
11:23:51*TheSeven wonders if sprotte24 is using some T9-like input method that's not very good at auto completing :)
11:24:01 Quit GeekShad0w (Quit: The cake is a lie !)
11:24:54sprotte24TheSeven: I write with chatzilla IRC from firefox on my windows pc
11:25:29*S_a_i_n_t would have just said "yes" and run with it ;)
11:26:12CIA-70New commit by 03jethead71 (r29275): audio_peek_track should copy the struct mp3entry instead of pointing directly into the buffer. Despite the dire warning, caller does in fact ...
11:26:27sprotte24I do not use T9, I do not like writing SMS :-)
11:26:34CIA-70r29274 build result: 2 errors, 0 warnings (theseven committed)
11:27:19sprotte24I live in Germany, so my English is not so good
11:27:41ZagorTheSeven: whack-a-mole! :)
11:30:00TheSevenZagor: no, *that* one wasn't mine!
11:31:31 Join TheSeven|Mobile [0] (~theseven@rockbox/developer/TheSeven)
11:31:35CIA-70r29275 build result: All green
11:31:52TheSevensprotte24: "rocket box" seriously looked like autocompletion going nuts :)
11:32:12TheSevenZagor: proof!
11:32:26Zagor"rocket box" has to be our new years' edition
11:33:14 Join factor [0] (~factor@
11:33:26S_a_i_n_tIt's quite impressive to type like Google Translate, without Google Translate ;)
11:36:26 Quit TheSeven (Ping timeout: 240 seconds)
11:40:36 Quit markun_ (Ping timeout: 250 seconds)
11:57:16 Join markun [0] (~markun@rockbox/developer/markun)
12:01:08 Quit mudd1 (Ping timeout: 240 seconds)
12:05:36 Join PurlingNayuki [0] (~PurlingNa@
12:05:48PurlingNayukiHi everyone.
12:06:07PurlingNayukiWhat should I do to set up the Android SDK?
12:08:13S_a_i_n_tFollow the directions on the Google developer website.
12:08:50PurlingNayukiThen if I want to use the in Rockbox?
12:09:03S_a_i_n_tit's best not to.
12:09:24S_a_i_n_tjust follow the directions here, then you are assured to get the latest available copies.
12:09:25S_a_i_n_tmiawing and celling
12:09:35S_a_i_n_tdammit, ignore that last post.
12:10:02S_a_i_n_targh! why did the link do that!
12:10:03PurlingNayukiOh no.
12:10:54PurlingNayukiBut I execute that in Rockbox.
12:12:13S_a_i_n_tit should be long as it points to SDK r9 and NDK r5b
12:12:15PurlingNayukiAnd how to build
12:13:05S_a_i_n_tI'm not even sure what that is for.
12:13:44PurlingNayukiI ran ../tools/configure in rockbox/android.
12:13:57PurlingNayukiBut which one should I choose?
12:14:21S_a_i_n_twhat do you mean?
12:14:31 Quit sprotte24 (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
12:14:31gevaertsYou should build in a dedicated directory, not in a random source directory
12:14:53S_a_i_n_tyou should not need to run in rockbox/android/
12:15:01S_a_i_n_tyou should create your own build dir.
12:15:07S_a_i_n_tnot build in randon source dirs.
12:15:18PurlingNayukiBuild as usual in the android/ directory of Rockbox trunk (configure with tools/configure and follow the directions) (../tools/configure && make -j3)
12:15:20PurlingNayukiThat is from the offical website.
12:15:26S_a_i_n_tooop, too late, thanks gevaerts.
12:15:58gevaertsPurlingNayuki: android/README is *more* official
12:16:10PurlingNayukiWell thanks.
12:16:19PurlingNayukiI'll read it at once.
12:16:23S_a_i_n_tyes, there is no need to build as an individual step.
12:16:44 Quit Keripo (Quit: Leaving.)
12:17:15S_a_i_n_tMy advise would be to not bother with InstallToolchain either, personally...but, if you want to, go for it.
12:18:08S_a_i_n_tall you need (additional to the rockbox source) is the S/NDK and a JRE
12:24:58 Join dfkt [0] (dfkt@unaffiliated/dfkt)
12:30:04 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
12:30:09 Quit TheSeven|Mobile (Read error: Connection reset by peer)
12:31:01PurlingNayukiNow I'm trying to build my first Android Rockbox/
12:31:18*PurlingNayuki is very happy
12:31:52*PurlingNayuki is also nervous and don't want to wait
12:33:59PurlingNayukiCC apps/bookmark.c
12:34:01PurlingNayukiIn file included from /home/yzflcyq/rockbox/apps/bookmark.c:34:
12:34:03PurlingNayuki/home/yzflcyq/rockbox/apps/recorder/icons.h:33:33: error: bitmaps/rockboxlogo.h: No such file or directory
12:34:08*PurlingNayuki needs help!
12:35:26 Quit JdGordon| (Ping timeout: 240 seconds)
12:35:58 Quit JdGordon1 (Ping timeout: 240 seconds)
12:39:02 Part LinusN
12:47:40S_a_i_n_tit sounds as though your PATH is incorrect, or, your source checkout is imcomplete.
12:52:11n1sIIRC rockboxlogo.h is generated by the build system so something has gone wrong prior to that error
12:59:15 Quit pamaury (Ping timeout: 240 seconds)
12:59:29 Join LinusN [0] (~linus@rockbox/developer/LinusN)
13:00:59 Join francesco_ [0] (
13:01:22PurlingNayukiI remember that I used sudo when running
13:13:20 Quit petur (Quit: *plop*)
13:13:23francesco_hi there.. does anyone know if it's possible to change colors to the clip+ v1 screen?
13:13:36Torneno, it's a monochrome display
13:16:05francesco_Torne: actually it's yellow and blue
13:16:15francesco_i'd like to change the yellow to something else...
13:16:35Zagorfrancesco_: you can't. the pixels are only one color.
13:16:51Tornefrancesco_: It's definately monochrome :)
13:16:53Zagorthe top pixels are yellow, the rest blue.
13:17:00TorneThe top bit is monochrome yellow, the bottom bit is monochrome blue
13:17:17francesco_Zagor and Torne: ok thanks
13:17:32***Saving seen data "./dancer.seen"
13:25:35*linuxstb reads about the .ONESHELL Makefile directive and wonders if it would be useful for Rockbox
13:27:19francesco_is there a way to avoid that everytime clip+ get plugged to usb it switch to OF?
13:29:26 Join pamaury [0] (
13:29:26 Quit pamaury (Changing host)
13:29:26 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
13:31:43 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
13:38:26Zagorfrancesco_: no, we use OF for usb
13:38:52Zagor(on this target)
13:44:16AlexPgevaerts: OK
13:45:06 Quit TheSeven (Ping timeout: 245 seconds)
13:46:59 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
13:50:00 Join petur [0] (d408b802@rockbox/developer/petur)
13:52:33 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
13:55:13 Join mudd1 [0] (~cmertes@2001:638:504:20e0:221:70ff:fe83:655e)
14:04:27 Quit antil33t (Read error: Connection reset by peer)
14:04:37 Join antil33t [0] (
14:06:45pamauryI'm considering to get a fuze+, do you think Sandisk is still willing to give us some hardware (as they pretended a few month ago (don't remeber who they contacted)
14:08:56 Join Rob2222 [0] (
14:09:47 Join MethoS- [0] (~clemens@
14:10:43Zagorpamaury: if you plan to develop on the device, the fund can buy it for you
14:11:15 Quit GeekShadow (Read error: Connection reset by peer)
14:11:22 Join GeekShad0w [0] (
14:11:26pamauryI plan to try. If the fund buys it and I fail, I'll give the device to someone else of course.
14:17:02pamauryZagor: I buy it and the "fund" gives me the money back, or the "fund" really buys it for me ?
14:17:23Zagoryou buy it. sending money is easier than sending goods.
14:19:01pamauryOk, now I need to find a cheap fuze+ if possible
14:21:01 Join kevku [0] (~kevku@2001:7d0:0:f9af:feed:feed:feed:feed)
14:27:15 Quit petur (Quit: reboot)
14:27:57sideralfrancesco_: There is experimental USB support for the Clip+. If you want to help debug it, see FS #11664
14:28:47Tornenote that "helping debug it" might include your clip locking up and having to wait for 36 hours for the battery to run down before you can use it again
14:38:18 Join robin0800 [0] (
14:48:20 Join petur [0] (d408b802@rockbox/developer/petur)
14:51:03 Join komputes [0] (~komputes@ubuntu/member/komputes)
14:54:12 Quit B4gder (Remote host closed the connection)
14:56:51 Join markun_ [0] (5571f93d@rockbox/developer/markun)
15:05:35 Join Topy [0] (
15:06:16 Quit T44 (Read error: Connection reset by peer)
15:17:33***Saving seen data "./dancer.seen"
15:21:21 Join Lol [0] (
15:21:47 Nick Lol is now known as Guest28722 (
15:22:05Guest28722does anyone speak english?
15:23:21 Quit Guest28722 (Client Quit)
15:36:57PurlingNayukiWell English is the most widely used language all over the world.
15:36:58 Quit benedikt93 (Quit: Bye ;))
15:38:04PurlingNayukiWhen I let it make apk it display:
15:38:06PurlingNayukiTHIS TOOL IS DEPRECATED
15:38:32Tornebecause we call a tool that's deprecated, i expect
15:39:23PurlingNayukiYou really have a sense of humor, I expect.
15:40:41Tornewell what do you want?
15:40:52Tornei wasn't joking, that's the answer.
15:41:06Tornesome tool we use while building the apk is deprecated in whatever version of the android sdk you have.
15:41:31PurlingNayukiSo it is.
15:42:05Torneit suggests we should maybe use another tool at some point, but it's not relevant to a user building it
15:43:52gevaertsIt's a bit silly, really
15:44:20 Part PurlingNayuki
15:44:21gevaertsThe recommended replacement is calling some API function in the android tools. The tool we use does exactly that
15:47:02 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
16:01:07 Quit krazykit (Quit: awe yeeeeeee)
16:03:18 Quit n1s (Quit: Lmnar)
16:04:32 Quit benedikt93 (Quit: Bye ;))
16:24:40 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
16:27:22 Part LinusN
16:28:41 Quit petur (Quit: Page closed)
16:38:03 Quit robin0800 (Ping timeout: 240 seconds)
16:40:15 Join toffe82 [0] (
16:45:43 Join {phoenix} [0] (
16:48:03 Quit Topy (Ping timeout: 240 seconds)
16:55:40 Quit bluebrother (Disconnected by services)
16:55:41 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
17:05:32 Join tintamarre [0] (
17:10:03 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
17:14:12 Quit GeekShad0w (Read error: Connection reset by peer)
17:14:45 Join GeekShad0w [0] (
17:17:35***Saving seen data "./dancer.seen"
17:17:45 Quit TheSeven (Ping timeout: 265 seconds)
17:20:20 Part Zagor
17:22:49 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
17:25:50 Join thegeek [0] (
17:35:50 Quit {phoenix} (Read error: Operation timed out)
17:38:06 Join {phoenix} [0] (
17:39:16 Quit sideral (Ping timeout: 240 seconds)
17:40:39 Join T44 [0] (
17:42:14 Quit francesco_ (Remote host closed the connection)
17:44:22 Quit markun_ (Ping timeout: 245 seconds)
17:51:18 Part toffe82
17:54:02 Join sprotte24 [0] (
18:00:26 Quit DerPapst (Quit: Leaving.)
18:00:33 Quit balintx (Remote host closed the connection)
18:03:11 Join n1s [0] (
18:03:11 Quit n1s (Changing host)
18:03:11 Join n1s [0] (~n1s@rockbox/developer/n1s)
18:09:33 Quit jae (Quit: leaving)
18:10:12 Join L-Strife89 [0] (
18:11:18 Join TheLemonMan [0] (
18:22:36 Quit Dreamxtreme (Read error: Connection reset by peer)
18:24:15 Join milk_ [0] (
18:24:17 Quit Utchy (Ping timeout: 276 seconds)
18:26:14 Quit plux (Ping timeout: 240 seconds)
18:28:50 Join Dreamxtreme [0] (~Dre@
18:35:30 Join plux [0] (
18:36:38 Quit qurvel (Ping timeout: 250 seconds)
18:46:23 Quit TheLemonMan (Read error: Connection timed out)
18:47:02 Join TheLemonMan [0] (
18:47:24 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
18:47:56 Join Utchy [0] (
18:53:14 Quit TheLemonMan (Quit: free(me))
18:53:56 Join kugel [0] (~kugel@rockbox/developer/kugel)
18:54:34kugelS_a_i_n_t: building in android/ works just fine
18:54:47 Quit Dreamxtreme (Quit: Going!)
18:55:02kugelbuilding there actually allows you to use eclipse for the java bits
18:55:29kugel(otherwise eclipse doesn't pick up for building the .apk)
19:00:24 Join robin0800 [0] (
19:00:59 Quit hebz0rl (Quit: Leaving)
19:02:27 Join balintx [0] (
19:04:10 Join sideral [0] (~sideral@
19:04:10 Quit sideral (Changing host)
19:04:10 Join sideral [0] (~sideral@rockbox/developer/sideral)
19:08:57 Quit evilnick_B (Ping timeout: 244 seconds)
19:15:33 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
19:17:39***Saving seen data "./dancer.seen"
19:37:11 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
19:38:39 Join bertrik [0] (
19:38:46 Quit bertrik (Changing host)
19:38:46 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:39:33 Join Horscht [0] (
19:39:33 Quit Horscht (Changing host)
19:39:33 Join Horscht [0] (~Horscht@xbmc/user/horscht)
19:40:11 Quit robin0800 (Quit: Leaving)
19:44:01 Join Jerom [0] (~jerome@
19:45:50 Quit {phoenix} (Remote host closed the connection)
19:48:28 Join {phoenix} [0] (
19:54:55 Join TheLemonMan [0] (
20:00:56 Quit Horscht (Quit: Verlassend)
20:05:18 Join panni_ [0] (
20:07:06 Join Seerious [0] (
20:07:09 Join shai [0] (
20:07:28 Part Seerious
20:08:34 Quit shai (Client Quit)
20:12:12 Quit sideral (Quit: Leaving.)
20:13:34CIA-70New commit by 03theseven (r29276): iPod Nano2g/Classic clickwheel: Configure GPIO pins as Hi-Z while hold switch is engaged
20:13:48TheSevenhopefully the mole has been whacked now
20:15:26 Quit GeekShad0w (Read error: Operation timed out)
20:15:35 Join Horscht [0] (
20:15:36 Quit Horscht (Changing host)
20:15:36 Join Horscht [0] (~Horscht@xbmc/user/horscht)
20:18:55CIA-70r29276 build result: All green
20:30:05 Join Buschel [0] (
20:32:40 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
20:34:07*Buschel gets linker errors with r29270
20:34:22Buschel(simulation, cygwin)
20:35:08Buschelseems connected to the embedded AA changes
20:41:02 Quit evilnick_B (Ping timeout: 245 seconds)
20:47:05 Quit L-Strife89 (Quit: Returning to the dorm)
20:48:00 Quit liar (Read error: Connection reset by peer)
20:48:51 Join liar [0] (
20:57:51 Join offense [0] (
20:58:55 Quit offense (Client Quit)
21:00:43Buschelthe linker errors occurs when building the test plugins with r29258 or later
21:01:04kugelBuschel: it's the plugin
21:01:14kugeltest_jpeg plugn*
21:01:18kugelI got it as well, but haven't looked into it yet
21:02:32Buschellooks like there is something going wrong in case JPEG_FROM_MEM is defined
21:07:58 Quit AlexP (Quit: Please insert girder)
21:08:51 Join AlexP [0] (~alex@rockbox/staff/AlexP)
21:11:22Buschelkugel: how does this look? ->
21:12:23Buschelat least it compiles, but I do not have any test files to test with
21:12:56kugellooks good
21:13:04 Join fml [0] (
21:13:12kugelI didn't pay attention when making the wrapper
21:13:39 Join mystica555_ [0] (
21:13:45Buschelno prob
21:14:07fmlS_a_i_n_t: Hello. I've seen in the logs that you wanted to play with the patch for menu reordering. Have you got any results?
21:17:16fmlOr was that [Saint]? Is that the same person?
21:17:41***Saving seen data "./dancer.seen"
21:20:22 Quit liar (Ping timeout: 255 seconds)
21:20:23CIA-70New commit by 03Buschel (r29277): Fix test plugin build.
21:21:34 Join liar [0] (
21:24:24CIA-70r29277 build result: All green
21:27:48 Quit Rob2222 (Read error: Connection reset by peer)
21:27:58 Join Rob2222 [0] (
21:28:45 Join Keripo [0] (
21:38:39*linuxstb looks around for thomasjfox
21:41:57 Quit Jerom (Quit: Leaving.)
21:46:11 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
21:46:53 Quit n1s (Quit: Ex-Chat)
21:49:27 Quit mystica555_ (Read error: Connection reset by peer)
21:50:49 Join mystica555_ [0] (
21:53:16 Quit mystica555 (Ping timeout: 240 seconds)
21:55:41CIA-70New commit by 03alle (r29278): Fix a couple of typos in the comment in english.lang. Whether the comment is correct is another question which is not the subject of this commit.
21:56:47*TheSeven contemplates an evil plan
22:00:04CIA-70r29278 build result: All green
22:04:32 Quit fml (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
22:05:14TheSevensome quick grepping suggests that we might actually be able to fake current_tick based on USEC_TIMER at least on singlecore devices
22:05:28 Join mystica555 [0] (
22:05:31dfktdoes rockbox on android actually write a scrobbler.log somewhere, or is this not implemented yet?
22:05:38TheSevenand the non-hardware-dependent tick tasks don't seem to really care about how often they are called
22:11:10kugeldfkt: it seems not working for raaa
22:11:20kugelthomasjfox is investigating it
22:12:25kugelTheSeven: how many tick tasks are there anyway?
22:12:33kugelbacklight, button, and?
22:12:53TheSevenand dozens of drivers in the target tree
22:13:02TheSevenno idea what those need tick tasks for though
22:13:18TheSevenand that timeout thing of course
22:13:41kugelbut you must not forget that the thread scheduler also depends on the timer
22:13:56TheSevenyes, but it shouldn't depend on the tick rate being constant
22:14:27kugelno, but if you make it too low then the timeouts threads are asking for will be off
22:14:43kugel(although I doubt they need them very accurate)
22:14:48TheSevenwhat i'm contemplating is to scale the tick rate with the CPU speed, and maybe if we're idle even calculate for how long and schedule the next tick at whatever time we need it
22:15:47TheSevenwhile latencies may be critical while boosting, i don't think they are while unboosted
22:16:13TheSevenreducing the tick rate if there is no pressure on the scheduler will even increase efficiency :)
22:16:43kugelit would be sure nice to have that
22:17:05TheSevenso we have 420 occurrences of current_tick and 28 occurrences of tick_add_task...
22:17:23kugelperhaps replace current_tick with a function call, and make it possible to override it in the target tree
22:17:37TheSevenhm, the ultimate hack would probably to increment current_tick by 2 or 4 depending on the current scaling factor :)
22:17:41kugelI find it ugly anyway that plugins/codecs can access it directly (*rb->current_tick)
22:18:14kugelTheSeven: we considered that one actually :) I think it's not that much of a hack
22:18:36TheSeventhis one could probably work out without much effort
22:19:07TheSevenbut where is it being incremented? i can't seem to find that using grep
22:19:14TheSevenasm code involved?
22:19:17kugelin call_tick_tasks()
22:20:05TheSevenhm, in the long term i think we should get rid of HZ completely and work with microseconds instead
22:20:28TheSeven2000 seconds of delay (before it wraps) should be enough for everything, right?
22:21:12TheSevenand if we don't have that much timer accuracy, things shouldn't mind if it increments in steps of 1000 or whatever
22:22:08kugelI don't agree with that
22:22:33kugelif you can't count on the timer to have microsecond resolution, then it shouldn't have microsecond resolution in the first place
22:23:36 Join bluebrother [0] (
22:23:37 Quit bluebrother (Changing host)
22:23:37 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
22:24:45kugelis 10ms insufficient?
22:26:59 Quit bluebroth3r (Ping timeout: 255 seconds)
22:27:25 Quit Stummi (Quit: Bye!)
22:29:20 Quit bluebrother (Ping timeout: 276 seconds)
22:29:34TheSevenkugel: I'm thinking more along the lines of more efficiency (and higher-quality profiling data) if the accuracy is available
22:30:03TheSevenbtw my tick divider proof of concept patch seems to work fine so far
22:30:14kugelwhy would go down to 1us increase efficiency?
22:31:27 Quit Buschel (Ping timeout: 265 seconds)
22:32:10TheSeveni would go to 1us to unify things
22:32:28TheSevenfor scheduling purposes, 1ms is probably sufficient for almost everything, but not for profiling
22:36:38 Join user890104 [0] (
22:37:56 Part sprotte24
22:39:08TheSevenhm, rockbox barely sleeps 60% of the time at 27MHz
22:39:19TheSevenso 10MHz of idle load... where is that coming from?
22:39:31TheSevenand the UI is really sluggish
22:41:28 Quit benedikt93 (Quit: Bye ;))
22:43:48 Quit TheLemonMan (Quit: free(me))
22:48:32TheSevenTorne: do i remember correctly that it was you who implemented that dynamic USB current selection thing?
22:48:47TheSevenis it to be expected that it will only ever draw 100mA in charging mode, and 500mA only in UMS mode?
22:48:59TheSevenbut that's what happens on my classic
22:49:01Torneit should switch to 500mA once it's been configured
22:49:06Torneunless you set it to off
22:49:09TheSevencharging is set to "force"
22:49:13TorneDoesn't matter
22:49:26Tornedo you actually have a working usb stack?
22:49:44TheSevenwell, it does switch to 500mA for UMS, but not if i hold menu while plugging it
22:49:46Tornedoes it use the same detection mode as PP ipods?
22:49:56Torneand do you have HID enabled?
22:50:27TheSevenno, i don't have HID enabled
22:50:35TheSevenand what do you mean with "detection mode"?
22:50:40TheSevenUSB or power detection?
22:51:11Torneis it doing polling or interrupts?
22:51:21Tornetry turning HID on, unless that doesn't work :)
22:51:29Tornemaybe the host isn't configuring our dummy device properly
22:51:32TheSevenint usb_detect(void) { if (charger_inserted()) return USB_INSERTED; return USB_EXTRACTED; }
22:51:32 Quit factor (Read error: Connection reset by peer)
22:51:41TheSevenhost OS is windows 7
22:51:55Tornenot thta kind of detection
22:51:58Tornedetection of an active host
22:52:06Tornehow does it tell there's a host and not just power on Vcc
22:52:19TheSeveni don't think it detects that at all
22:52:30TheSevenhow could it do that?
22:52:42TheSeveni.e. how would one implement it API-wise?
22:52:43Tornethe other drivers do it
22:52:48Tornelet me look
22:52:56TheSevenshouldn't the USB core just check for a bus reset to determine that?
22:53:38Tornesee USB_DELAYED_INSERT code
22:54:07TorneDoes it immediately go into USB mode even if you only plug in a charger, if you don't hold menu?
22:54:29TheSeven(which is annoying me quite a bit)
22:54:30TorneRight, then you haven't done detection "right"
22:54:36Tornedefine USB_DELAYED_INSERT i think
22:54:50TorneIt distinguishes USB_POWERED from USB_INSERTED
22:54:58Tornethe usb core is responsible for handling host detection
22:54:58TheSevenwhere would one define that? config/ipod6g.h?
22:55:08Tornei dunno, look at the PP ipods
22:55:20Tornethe target code tells it when ther'es usb power
22:55:28Tornethen it turns on the usb controller and wiats for an interrupt
22:55:38Tornethe usb core then transitions it to USB_INSERTED when it gets the interupt for a bus reset
22:55:57TorneIt may be that the charger stuff doesn't work properly for targets that don't do delayed insert handling :)
22:56:05 Join lol [0] (
22:56:09Torneor it might be something else
22:56:10TheSevenwell, apparently
22:56:19TornePRetty sure it works fine on PP though
22:56:31 Nick lol is now known as Guest98973 (
22:56:45TheSevenis there any reason to not always enable this though? i mean, it's a USB core matter that doesn't really depend on the driver, right?
22:57:04Guest98973Hey, Is there a supported Lua documentation for rockbox?
22:57:10TheSeven/data/rockbox-trunk/firmware/usb.c:482: error: "USB_UNPOWERED" undeclared
22:57:21Tornemaybe it's a different define :)
22:57:21 Join Sof [0] (
22:57:30Tornelook to see what ipodvideo defines ;)
22:57:35 Quit Keripo (Quit: Leaving.)
22:57:42Torneincluding indirectly via config.h based on its usb type ;)
22:58:25 Join Keripo [0] (
22:58:31SofHo no. I've landed in a geek world... :-) hello all
22:58:50 Quit Guest98973 (Client Quit)
22:58:56TheSevenTorne: there's USB_DETECT_BY_DRV and USB_DETECT_BY_CORE
22:59:02TorneThat's the one
22:59:10*Torne has to idle for now.
22:59:15 Quit domonoky (Read error: Connection reset by peer)
22:59:39SofI've a little question about the sansa clipp+ with rockbox. Is there a way to lock the keyboard while playing?
23:00:12bertrikYes, I think so, but I forgot the key combo for that
23:00:21literalhold the middle button and the home button, I believe
23:00:29TheSevenSof: If there is such a thing, it should be described in the manual :)
23:00:42literaland yeah, it's in the manual
23:02:12Sofhm... OK sory I'll look into the manual. I thought it was quicker to ask here ;-) htanks! :-)
23:02:27 Join bluebrother [0] (
23:02:27 Quit bluebrother (Changing host)
23:02:27 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
23:03:42CIA-70New commit by 03jethead71 (r29279): Always do proper mp3entry copying instead of using bufread directly. Add bufreadid3() helper.
23:04:49 Quit bluebrother (Read error: Operation timed out)
23:05:52*jhMikeS wants to officially deprecate USB_DETECT_BY_DRV and insist on USB_DETECT_BY_CORE instead since it's more robust and actually easier to implement at the driver level
23:07:27 Join bluebrother [0] (
23:07:27 Quit bluebrother (Changing host)
23:07:27 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
23:07:43 Join factor [0] (~factor@
23:08:06Softhanks for your clear answers guys. really helpfull. bye! :-)
23:08:11 Quit Sof ()
23:08:33CIA-70r29279 build result: All green
23:08:58 Quit evilnick_B (Ping timeout: 245 seconds)
23:09:32 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
23:10:31 Quit ender` (Quit: If Klingons had invented Usenet, killfiles really would...)
23:13:37 Quit efyx (Quit: Quitte)
23:17:04 Quit mudd1 (Remote host closed the connection)
23:17:43***Saving seen data "./dancer.seen"
23:34:12 Join robin0800 [0] (
23:35:02*TheSeven needs some kind of a function generator plugin for rockbox
23:35:17TheSevene.g. square/sine at configurable amplitude and frequency
23:35:32 Quit robin0800 (Remote host closed the connection)
23:35:32S_a_i_n_twhat for?
23:35:40 Join robin0800 [0] (
23:35:58TheSevenmainly to test other hardware :)
23:36:17Bagder <= perl script to make sine table with
23:36:32amiconnJust play an appropriately prepared wav file?
23:36:52TheSeventuning on-target was the whole point of it
23:36:58TheSevenespecially the frequency
23:37:37jhMikeSI have a blit synth plugin that I was messing with but never really did much with
23:38:31TheSevenwhat i currently need is rather trivial: produce at least one high sample at a specified interval
23:39:02TheSevenan arbitrary duty cycle square wave at a configurable frequency
23:39:07jhMikeSit does that
23:39:20jhMikeSwithout aliasing
23:40:05jhMikeSthe pure unitegrated impulse train will give the high samples
23:46:38jhMikeSall I have is a stale old test_sampr patch for it :\
23:47:56 Quit {phoenix} (Remote host closed the connection)
23:48:05 Quit bertrik (Quit: :tiuQ)
23:49:43amiconnTheSeven: The coldfire targets also have cpu clock based timers. That doesn't stop us from clock scaling. Clock scaling just needs to adjust the prescalers accordingly
23:50:02 Quit kevku (Quit: KVIrc 4.0.2 Insomnia
23:50:49jhMikeSTheSeven: fwiw">
23:51:54 Join mudd1 [0] (
23:51:56 Quit mudd1 (Client Quit)
23:52:41 Join mudd1 [0] (
23:58:33 Quit komputes (Remote host closed the connection)

Previous day | Next day