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 2009-10-05

00:01:46TheSevendoesn't change anything
00:02:43 Quit gregzx ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
00:02:51linuxstbCan you upload your audiotest code somewhere?
00:05:36TheSeveni ditched the whole wmaudio stuff now and replaced it with my own, still doesn't work
00:06:17linuxstbYou could try some plugins that play audio - e.g. metronome
00:07:14 Quit Thundercloud (Remote closed the connection)
00:07:14TheSevenmetronome fails with a beeping piezo
00:07:27linuxstbAh yes, it will try the user timer (currently TIMERD...)
00:07:41linuxstb(we need to get a second timer working in Rockbox)
00:08:40linuxstbOK, then maybe test_sampr - it's one of the test plugins, so isn't compiled by default. Add the line "test_sampr.c" to apps/plugins/SOURCES to enable building it.
00:09:13TheSeveni think i'll do some splashf'ing now
00:11:04 Quit smacnay ("Leaving")
00:11:49 Quit matsl (
00:16:48TheSevenwhere can i see the contents of the logf buffer?
00:17:54domonoky1TheSeven: you need to build a logf enabled build, and then its in the debug screen somewhere (afaik)
00:18:12TheSevenhow do i enable that?
00:19:19 Quit bluebrother ("leaving")
00:19:48 Quit bertrik ("sleep")
00:20:59DerPapst#include "logf.h" iirc
00:21:08*DerPapst doesn't know for sure :P
00:21:56linuxstbTheSeven: You need to enable logf when you run tools/configure (in the advanced options I think). Then add #include "logf.h" (if it's not already there), but with "#define LOGF_ENABLE" _before_ the #include line
00:22:10 Quit barrywardell ()
00:22:10 Quit webguest67 ("CGI:IRC (EOF)")
00:22:31linuxstbThis is so you can leave calls to logf in your code, and only enable them when you need to debug that part of the code.
00:22:49linuxstb(and so that the logf buffer isn't spammed with messages from all over Rockbox)
00:22:57 Quit domonoky1 (Read error: 104 (Connection reset by peer))
00:23:05 Join barrywardell [0] (
00:23:15gevaertsof course you can compile with -DLOGF_ENABLE for lots of fun :)
00:23:30linuxstbOn targets with usb-serial (hint, hint!), you can logf over usb...
00:24:20TheSeven...or over UART
00:25:06TheSevenin fact, in our current situation, USB could be easier, as we have both sample code and UART requires hardware hacking
00:25:07 Quit stripwax (Read error: 131 (Connection reset by peer))
00:25:54CIA-43New commit by tomers (r22930): Use similiar constants in genlang and language.c. ...
00:26:10CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
00:26:36*pamaury thinks the beautiful logf screen in the debug menu he coded is even easier to use than uart and usb-serial
00:27:01gevaertspamaury: not if you have lots of data :)
00:27:10linuxstbOr want it real-time...
00:27:36pamauryI've noticed it during my intensive debug sessions with MTP :)
00:27:42gevaertsthat's when pamaury's other contribution comes in : make usb-serial actually work :)
00:27:45CIA-43New commit by kkurbjun (r22931): Brickmania: Some code refactoring - no functional changes.
00:28:09 Quit barrywardell ()
00:29:02 Join barrywardell [0] (
00:29:29pamauryThe problem with usb-serial is that it uses two usb endpoints, on some target with few endpoints, it can be a problem when both MSC and HID are activated
00:30:15gevaertsindeed. That would be less of a problem if choosing which drivers to enable would be a bit easier...
00:31:23pamauryThat's why in my svn code, I've changed usb-core: when a special button is pushed(< on e280), only the activated drivers are enabled and I have one entry per driver. easy...
00:33:06liar|netbookTheSeven: on my nano2g ftl_init fails: /* Something went terribly wrong.... */ return 1; <= this is where the ATA error: 1 comes from
00:33:19pamauryperhaps it would be cool to have a option in the debug menu which says: 1)"MSC-HID mode" 2)"MTP+HID" mode 3)... n)"Selective mode"
00:33:50amiconnusb serial has another disadvantage over real rs232 for debugging: it's more complex and (I think) does more buffering, so chances are higher that things go wrong with the transfer
00:33:50TheSevenliar|netbook: this means something before, probably ftl_init or vfl_init failed
00:34:21 Join notlistening [0] (n=tom@
00:34:38 Quit barrywardell ()
00:34:41amiconnTheSeven: Hmm, iiuc rockbox should also handle previous unclean shutdown in its ftl?
00:35:07gevaertspamaury: we'll need something like that at some point, accessible for normal users (i.e. not in debug)
00:35:16*gevaerts is off now
00:35:34 Quit bmbl ("Bye!")
00:35:43TheSevenamiconn: it can't do that currently, and implementing that will be very tricky and error-prone
00:35:57*pamaury goes to bed also
00:36:03 Quit pamaury ("exit(*(int *)0 / 0);")
00:36:06TheSevenand the only case where this usually happens is if you bootstrap rockbox through ibugger, loaded through the notes exploit
00:36:29TheSeveni.e. if the OF has touched the flash
00:37:23amiconnAnd also if we want to get rid of the OF completely (i.e. flash rockbox)
00:38:02TheSevenamiconn: flashing rockbox to nor implies a pretty high bricking risk on these devices
00:38:48TheSeven...WE HAVE SOUND!!!
00:38:49kugel:D awesome!
00:39:17***Saving seen data "./dancer.seen"
00:39:33DerPapstCongratulation TheSeven and linuxstb :D
00:39:45TheSevennow I need to find out WHY this actually works now
00:39:57amiconnTheSeven: Bricking risk when flashing to nor is always quite high. That's why we have fallback mechanisms on all targets where we can do this atm
00:40:00TheSevenenabling logf seems to have fixed it... o.0
00:40:14TheSeven(or rather the complete rebuild caused by that)
00:40:34 Join chandoo [0] (
00:41:14amiconnFull flashing is usually only done once, after that we only update the rockbox image in flash; the fallback and selector are rarely updated (and never automatically - full flashing is a separate process)
00:41:21linuxstbTheSeven: Time to double-check that... Otherwise, maybe an alignment issue?
00:42:02TheSevenamiconn: if you flash something that has a correct checksum but doesn't work, you're bricked.
00:42:10TheSevenso developing that means trashing lots of them.
00:42:32TheSevenand I don't see no reason at all why we should do this (can't see any advantage)
00:42:47linuxstbNormally a much faster boot time.
00:43:02TheSevenabout 2 seconds difference
00:43:21amiconnAlso running 100% our code
00:43:22DerPapstbut the oder ipods didn't require any flashing too...
00:43:25TheSeven(or 10 to 20 seconds after an unclean unmount, as norboot is repairing the flash)
00:43:46linuxstbDerPapst: No, it's not _required_, just nicer.
00:44:07TheSevenDerPapst: did anybody at iPL ever dare to flash a modified AUPD?
00:44:11Galoiswhat do we have sound on?
00:44:13linuxstbBut it's generally a lot of work (and risk), so few targets support it.
00:44:15amiconnDerPapst: No target *requires* flashing rockbox (some require flashing our bootloader though), but it's still desirable
00:44:22DerPapstTheSeven: not that i know of
00:44:28TheSevenit's pretty safe since nano 3g, but the 2g is still high-risk
00:44:31linuxstbTheSeven: I _think_ ipodwizard does that - e.g. to change the boot logo.
00:45:03TheSevenlinuxstb: only flashing the logo is also relatively safe, but flashing code isn't
00:45:09amiconnThe archoses fortunately have an ultimate safety net (unless you've one of the very rare bootrom-less devices)
00:45:13TheSeventhe code in NOR is our only way to recover
00:45:32TheSevenamiconn: that's true for the ipods, too, since 3g
00:45:43linuxstbWe did the equivalent of flashing code to NOR on the Sansa Clip - it just needs a _lot_ of care...
00:45:44DerPapstthe advantage over a flashed ipod compared to the risk of bricking it isn't worth it... for me at least
00:45:51TheSeventhe 2g has a bootrom DFU mode, too, but that can only be triggered by a bad flash checksum, not by the user
00:46:12amiconnThe coldfire irivers don't, and yet flashing was done (well they can be recovered, but only using a bdm, which isn't a cheap piece of hardware)
00:46:31linuxstbBasically adding code a few instructions at a time, and checking the disassembly lots of times before praying...
00:46:32TheSeveni don't think anyone will care about 2 seconds boot time and showing an apple logo during these (even though we could even replace that logo without too much risk)
00:47:07linuxstbTheSeven: I don't think anyone is saying it's important. But it's always a nice thing when it's done.
00:48:01TheSevenI wouldn't even flash it if it was tested hundreds of times before
00:48:14*DerPapst too
00:48:46TheSevenwe may consider that on the newer devices which enter BR DFU after double-resetting, but for the old ones it's just not worth the risk
00:49:19*TheSeven is step by step reverting his changes now until it breaks again
00:50:41*TheSeven just noticed *why* it failed before, when he looked at his shell history
00:50:41linuxstbGalois: 2nd generation iPod Nano.
00:51:51*DerPapst bets TheSeven forgot to actually install the new build :P
00:52:19TheSevenno, i accidentally added a trailing ´ to the destination file name while hitting enter
00:52:22 Join AsaelReiter [0] (
00:52:36TheSeven(while copying over the new build)
00:52:59DerPapstwell. at least the same effect ;)
00:53:21Galoislinuxstb: rock on
00:53:47TheSevenlinuxstb: your wmcodec change didn't show any effect (i.e. still works after reverting)
00:54:07linuxstbTheSeven: OK, good. Because that made no sense...
00:54:58 Quit ender` (" A computer program will always do what you tell it to, and seldom what you want it to.")
00:55:08TheSevenone reason that could cause additional trouble with a clean svn build could be the stack size, just as a reminder
00:56:48TheSevenlinuxstb: I'll send you a really ridiculous patch now
00:56:56TheSevenin fact i'm considering just pasting it into IRC
00:56:59TheSevenit's 4 lines
00:57:03linuxstbGo ahead...
00:57:24 Nick fxb is now known as fxb__ (
00:57:28liar|netbookTheSeven: ftl_open fails because ftlcxtfound == 0, but i have no idea of anything(never had to do with ftl code before)
00:58:23saratogaamiconn: i'm curious if you think buffering the codec files ahead of time makes sense on the clip
00:58:50TheSevenliar|netbook: try commenting the break below /* This will trip if there was an unclean unmount before. */
00:58:55amiconnIt makes sense in that it means less special casing, as I already said
00:59:23amiconnIt also makes sense in that it seems to help finding bug in buffering
01:00:02 Quit AsaelReiter ("CGI:IRC (EOF)")
01:00:22*TheSeven always forgets the correct URL
01:00:52TheSevenI'm pretty sure none of these lines actually fixed the real problem
01:01:01TheSevenif anything, it's the last change
01:01:53TheSevenlinuxstb: I think you can commit that right away, huh?
01:02:23linuxstbTheSeven: Yes, although let me test first ;)
01:02:35TheSevenI didn't check if the upper changes make any difference, but the old code was definitely just wrong.
01:02:36linuxstbBut no need to post to flyspray, if that's what you're asking.
01:02:53 Join coolrun34 [0] (
01:03:03TheSeven(not that removing that GPIO code from audiotest would have prevented it from working...)
01:03:43coolrun34how are you guys managing your song playlists
01:03:53coolrun34i installed rockbox because my ipod would keep reseting
01:03:56coolrun34and itunes would crash
01:04:12coolrun34seems to work fine
01:04:29TheSevenlinuxstb: if it fails for one or the other reason, try a complete rebuild. I think that's what actually fixed the last problems for me.
01:04:34coolrun34but my question is i cant see all the songs that were on there physically before the rockbox isntall
01:05:23saratogamt: ping
01:05:28linuxstbTheSeven: Yes, it doesn't work for me... Let me do a rebuild (although I can't think why that would help...)
01:06:44linuxstbcoolrun34: might help
01:07:32linuxstbTheSeven: I am just running current SVN, plus the stack change and audio.patch
01:08:38TheSevenlinuxstb: that's roughly what I'm running, too, besides from some unrelated FTL changes (i.e. force mount flag, even if unclean)
01:09:39coolrun34thank you
01:09:40coolrun34so much
01:09:43coolrun34it actually helped
01:09:50coolrun34sorry about the lame questions
01:09:53coolrun34but u did help me out
01:10:35*TheSeven is updating to current SVN now and will then do another patch with all of his (unrelated) changes
01:11:15 Join Thundercloud [0] (
01:13:39linuxstbTheSeven: No, still no luck. It just stays at 0:00
01:14:00TheSevenoops, i once again overlooked something
01:14:16TheSeveni in fact disabled the clock setup again
01:16:58 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
01:17:16TheSevenfunny. still works after re-enabling it
01:17:43TheSevenI'll do another make clean && make && make fullzip now to get rid of any old crap that may be remaining
01:18:28*TheSeven needs a faster build machine
01:18:33linuxstbIf it works, I could try your zip.
01:20:27TheSevenbtw, i have the impression that there is some slightly noticable clickering every half a second
01:20:38TheSevenstill need to check whether it's the mp3 that's at fault though
01:24:24 Quit faemir ("Leaving")
01:24:36TheSevenanother thing i consider a bug: when i open a file, rockbox almost hangs for several seconds while filling its buffers
01:24:52TheSevenprobably caused by busy waiting in the nand driver
01:25:19linuxstbYes, I meant to ask you about that - you just start the DMA transfer and busy-wait for it to finish?
01:25:24 Quit Thundercloud (Remote closed the connection)
01:25:48TheSevenyep, I wanted to keep things safe, slow and stupid...
01:25:57linuxstbOf course.
01:26:17saratogadoes the notes exploit still exist on the Nano5G?
01:26:42TheSeventhey fixed it in the 4g firmware that came out some weeks before
01:26:44linuxstbOr the latest OF version on the 4G (but you can downgrade easily)
01:26:55saratogaso they saw what you were doing?
01:27:26saratogahardware wise are the newer nanos similar to the 2G?
01:27:32saratogasame SOC and such?
01:27:42TheSevenno, newer models of that soc
01:28:53TheSeven2g = 8701, 3g/classic = 8702, 4g = 8720 (the ipod touch 2g soc!), 5g = 8730
01:29:09 Quit MethoS- (Remote closed the connection)
01:30:16linuxstbTheSeven: How's that build doing?
01:30:46TheSeventesting it right now
01:31:10TheSevenoh NO!
01:31:16TheSeventhe file browser is broken!
01:31:27TheSevensame bug as with the other browsers earlier
01:32:06TheSevenand audio is broken again
01:32:56liar|netbookTheSeven: now ftl_vfl_read returns 4
01:33:15TheSeven=> you fed it a block number that's out of bounds
01:33:29*TheSeven is confused
01:37:44 Quit mt2 (Remote closed the connection)
01:40:05*kugel too
01:43:41kugelcurse albumart being tied to the playback engine!
01:44:19kugelI can't really think of a way to do multiple albumarts in a clean and safe way
01:45:52*TheSeven curses whatever code is fooling him right now
01:46:19TheSeventhere must be some special voodoo when rebuilding rockbox in a clean directory
01:48:02*TheSeven once again notices some clickering
01:48:33linuxstbTheSeven: So it's working again?
01:49:41TheSevenyes, besides that noise
01:50:37*TheSeven wonders what makes rockbox fail to unmount the flash all the time now
01:50:37 Join Blue_Dude [0] (n=chatzill@rockbox/developer/Blue-Dude)
01:52:00Blue_DudeRegarding threads: at any given time, how many separate threads are running, and what is a good way to track them?
01:53:01saratogaBlue_Dude: the debug menu has a thread viewer
01:53:13saratogai think theres about 10
01:55:00Blue_DudeA read something in the log about how you're not supposed to put up anything on the display outside the main thread, or something like that. May have been a splash screen. I'm trying to get a handle on playback threads. So far I have audio, codec and voice. Anything else I need to know about playback program flow?
01:55:27 Join killan [0] (
01:55:28 Quit killan_ (Read error: 104 (Connection reset by peer))
01:55:50 Join fdinel [0] (
01:55:52Blue_DudeIt seems some operations are limited to specific threads. Perhaps because certain threads "own" certain parts of the hardware?
01:55:56TheSevenbtw, what's the point of prebuffering more than a single track on flash targets?
01:56:23kugelwhat's the point of not doing it?
01:56:27Blue_DudeTheSeven: good question. I'm trying to figure out the buffering logic myself...
01:57:24kugelBlue_Dude: so I am :)
01:57:44linuxstbTheSeven: The Rockbox playback logic is somewhere you don't want to venture into...
01:57:58TheSevenreading lots of stuff ahead that the user may never even want to play may be useful on targets that have storage media that need lots of power to start up, but i can't see the advantage of doing that with flash media
01:58:33kugelagain, what's the advantage of not doing it?
01:58:53TheSeventhis is playing music on my ipod right now
01:59:03Blue_Dudekugel: buffering.c, playback.c and pcmbuf.c are hopelessly entangled. I'm looking at breaking out them out further but I'm having a hard time getting an overall picture of what's supposed to be happening.
01:59:04TheSevenkugel: in fact, it may waste power on flash targets
01:59:15kugelBlue_Dude: add albumart also :)
01:59:50linuxstbBlue_Dude: Just delete it all and start from scratch...
02:00:26Blue_Dudekugel: and crossfade, and metadata, and ridiculously complicated playlist construction options...
02:01:01Blue_Dudelinuxstb: don't tempt me. I've got half a mind to gut it and start over.
02:01:40kugelTheSeven: you may waste power both ways
02:01:51Blue_DudeAnd howls go up around the globe as everyone's favorite features get left behind...
02:02:07TheSevenbut on disk targets it may actually save some, while on flash targets, it can't
02:02:09kugelreading from flash isn't 100% free, so that you still want to have it sleeping most of the time
02:02:16linuxstbBlue_Dude: Just make sure you don't break gapless, and I'll be happy...
02:02:53Blue_Dudelinuxstb: If I have to even touch codec operation, I'll shrivel into a ball and weep softly.
02:03:00Blue_DudeSo gapless is safe.
02:03:06kugelthere's absolutely no point in adding a ifdef hell to this relatively ifdef-free area for questionable reasons
02:03:07TheSevenlinuxstb: first make nano2g play a single track without gaps :-D
02:03:20JdGordonwhat the fuck? buffer is full then the whole buffer gets trashed and track skips!
02:03:31JdGordonafter about 20s of playing the track
02:03:42saratogaTheSeven: on our current flash targets, theres an advantage to doing infrequent large flash transactions (presumably it lets the hardware spend more time unclocked)
02:03:58saratogathis results in slightly better battery life for large verses small buffers
02:04:17JdGordonsaratoga: hey, you have any more luck with the clip?
02:04:18saratogabut it may not be true on future targets (and eventually someone should check on the nano2g)
02:04:18TheSevensaratoga: does it really make a difference if that buffer is 3MB or 15MB?
02:04:35saratogaTheSeven: 30KB vs. 30MB did make a difference
02:04:41saratogai don't know about smaller differences
02:04:43 Quit GeekShad__ (Read error: 104 (Connection reset by peer))
02:04:47kugelTheSeven: you think it's helpful to leave lots of MB just unused?
02:05:05 Join GeekShad__ [0] (
02:05:06*kugel doesn't see the point of having this discussion again and again
02:05:12saratogaJdGordon: my research is taking up nearly all my time these days, but I've been in email contact with Nico about some things
02:05:31kugelpeople should stop to think of flash as it would be as cheap as ram
02:05:39TheSevenlinuxstb: btw, clocking still isn't right, sounds like 44100 played at 48000
02:05:44linuxstbTheSeven: Hmm, that zip doesn't work for me...
02:05:57kugelsaratoga: I'm yet to have a crash on my clip
02:06:21kugelwith your logging build
02:06:21saratogakugel: its very erratic
02:06:28saratogai got one after about 4 hours
02:06:41saratogadfkt reported something like 20 hours without a crash
02:06:50saratogathe logging seems to change buffering somehow
02:07:00saratogai should probably sit down and make it more efficient so it introduces less overhead
02:07:16kugelit basically does a lot more flash access doesn't it?
02:07:16saratogabut for now I have enough data from that build
02:07:29saratogaprobably 2x more then normal
02:07:40saratogaso it slows down buffering a lot
02:08:34 Join ben [0] (
02:08:40kugelJdGordon: any idea how to do multi aa cleanly?
02:08:43 Quit GeekShad__ (Read error: 104 (Connection reset by peer))
02:08:47saratogakugel: i think move_handle has something definately wrong with it, i just don't yet really understand the ring buffer well enough to be sure
02:08:58linuxstbTheSeven: Yes, I think my bootloader test was around 48KHz.
02:09:01 Nick ben is now known as Guest62636 (
02:09:02kugelyea that ring buffer is magic for me too
02:09:11saratogawe should ask Matthias Schneider about it more
02:09:14JdGordonkugel: cant say i've thought about it much.... I dont see why statsbar needs to wait for it...
02:09:18TheSevenlinuxstb: I'll do exact measurements now
02:09:29saratogaboth of the things he spotted look impossibly complicated to me and yet I think they're both right
02:09:40saratogahe may have a better understanding of buffering then we've given him credit for
02:09:41kugelJdGordon: because I don't want it to be forgotton
02:09:50JdGordonit wont be....
02:10:02kugeland you get strange behavior if you even try to use aa in the statusbar
02:10:05linuxstbTheSeven: Did you see my comment saying your zip doesn't work for me? (i.e. stuck at 0:00)
02:10:30JdGordonkugel: what sort of strange behaviour?
02:11:03kugelask gevaerts (or the guys on the tracker), i haven't actually tried it myself
02:11:06TheSevenlinuxstb: yes, i saw it, but i have no idea why it's behaving like that.
02:11:13kugelbut I think gevaerts even had a crash, but I'm not sure
02:11:40kugeljust imagine two skins fighting about code that's designed for a single skin
02:12:02JdGordonthey arnt fighting about the skin.. its the aa....
02:12:14JdGordonim going for breakfast... bbs
02:12:17kugelwhat I said..
02:13:32TheSevenlinuxstb: this *could* in fact be trouble prebuffering one of the files
02:13:47TheSeveni had some strange FTL reactions again with that build
02:13:59 Join AndyI [0] (n=pasha_in@
02:19:09 Quit pixelma (Nick collision from services.)
02:19:09 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
02:19:29 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
02:19:59 Quit amiconn (Nick collision from services.)
02:20:01 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
02:20:15 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
02:20:21 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
02:20:50 Quit DerPapst ("Leaving.")
02:21:50kugelsomeone floods our tracker with patches based of code that's not on the tracker or in svn
02:23:27kugelthat tasks should be closed, especially since the code that they're based off has no chance to get in SVN (I assume?)
02:24:34 Quit kugel (Remote closed the connection)
02:25:07 Quit AndyIL (Read error: 110 (Connection timed out))
02:25:47 Quit benime (Read error: 105 (No buffer space available))
02:27:35*linuxstb wonders what kugel was talking about
02:30:08mc2739FS #10640, 10642-10648 - cowon d2 patches based on someones code which he did'nt / couldn't provide
02:32:37 Quit JdGordon ("Leaving.")
02:33:22 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
02:35:19linuxstbmc2739: Ah, OK...
02:36:35saratogaso the guy who opened the task isn't the author?
02:36:53 Quit liar|netbook (Remote closed the connection)
02:36:55linuxstbSee here for example -
02:39:00saratogawhat should I look for?
02:39:08saratogabah i see
02:39:18***Saving seen data "./dancer.seen"
02:41:18saratogajust close that one then
02:47:53JdGordonanyone wanna try a 1 line patch to not draw AA so often?
02:51:19mc2739JdGordon: sure
02:53:20 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/freenode/x-5d8ddf4e4090fa5d)
02:54:25TheSevenlinuxstb: I have a calibrated sampling rate now
02:54:51linuxstbTheSeven: Nice. Do you still have the clicking?
02:55:12TheSevenno idea, need to check with real audio
02:55:24TheSeven(was using a 441Hz sine to calibrate it)
02:55:34linuxstbHow accurate is it?
02:55:43linuxstbI mean can you clock it at exactly 44.1Khz?
02:55:46mc2739JdGordon: what do I need to look for?
02:56:08JdGordonmc2739: that AA still draws as expected... and gets cleared as expected if you have a funky skin
02:56:50TheSevenI think the accurracy of that freq is probably higher than that of my measuring equipment
02:57:50TheSeven(used a OscIn * 4 * 49 / 8 / 4 / 4 PLL and divider chain)
02:57:57mc2739JdGordon: funky - as in the aa being overlayed
02:58:41TheSevensounds good - i think the clicking went away
02:58:52mc2739I don't think I have any of those
02:59:04TheSevenand i have re-enabled the clock setup, and it still works
02:59:19TheSevenbut what in fact really seems to have an impact is your weird patch
02:59:49TheSevenafter my latest adjustments (and my copying-firmware-to-ipod process fooling me again) this was the key to make it work
03:01:12CIA-43New commit by jdgordon (r22932): redraw the AA much less frequently to save bettery life and also let skins draw over the AA...
03:01:46TheSevenlinuxstb: is there a DS of that codec?
03:02:09TheSevenoh, all over the place...
03:03:21TheSevenfor those keen to try: i uploaded a new zip file, same path
03:06:09linuxstbTheSeven: Yes, success.
03:06:13 Quit Guest62636 (Read error: 110 (Connection timed out))
03:06:27mc2739JdGordon: I'm not seeing any problems with the themes I have
03:06:37JdGordongood to hear :)
03:07:03 Quit lifeless_ (Read error: 60 (Operation timed out))
03:07:12TheSevenlinuxstb: my playback screen freezes at 0:00, even though it plays back fine, even the next few tracks
03:07:27TheSevenclickwheel is unresponsive
03:07:32linuxstbYes, I was about to say that mine has frozen hat 0:07
03:07:44 Join Guest62636 [0] (
03:08:10linuxstbMaybe interrupts are broken somehow?
03:09:00linuxstbScrolling the wheel (to change volume) seems to freeze the UI
03:10:07linuxstbIf I don't scroll the wheel, the UI seems fine...
03:12:56TheSevenmine is too unresponsive to even break it :-)
03:13:15TheSeventhe funny thing is that this didn't happen with the old build
03:13:26linuxstbAre clocks enabled?
03:13:47TheSevenonly the patch i sent and your weird patch
03:14:43linuxstbAnyway, nice work getting this far. I need to sleep now, but I'll read the logs in the morning...
03:14:57*TheSeven needs to sleep, too
03:16:28linuxstbBut it seems volume changing is broken for some reason - changing volume from the sound settings menu also freezes the UI...
03:16:43*linuxstb really goes to sleep now..
03:16:59TheSevenobviously something funny is going on with that codec
03:17:06TheSeventhat sampling rate thing is weird, too
03:22:02 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
03:31:27 Quit Guest62636 (Read error: 110 (Connection timed out))
03:38:18 Join funman [0] (n=fun@rockbox/developer/funman)
03:49:37 Quit chandoo ("Leaving")
04:04:35 Join swima-on-the-lap [0] (
04:06:10 Quit swima-on-the-lap (Client Quit)
04:06:51 Quit TheSeven (Nick collision from services.)
04:07:09 Join The_Seven [0] (
04:07:21 Nick The_Seven is now known as TheSeven (
04:09:02CIA-43New commit by kkurbjun (r22933): Button Driver: Add button_status support with data - allows for reading the /immediate/ value of the touchscreen
04:13:16CIA-43New commit by kkurbjun (r22934): Brickmania: Touchscreen improvements - Take advantage of the immediate button data values. Reduce the wait time if the score needs to count up at the ...
04:14:24 Quit intrados (Read error: 110 (Connection timed out))
04:15:47 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
04:17:29CIA-43New commit by kkurbjun (r22935): Fix Red
04:26:14 Join intrados [0] (
04:34:25 Join Guest62636 [0] (
04:35:42 Nick Guest62636 is now known as benime (
04:36:11 Join damone [0] (
04:36:58damonehi everyone
04:38:20 Quit Rondom (Nick collision from services.)
04:38:31 Join Rondom [0] (
04:38:32damoneI was wondering, would it be possible to make rockbox read data from a usb port and interpret especific commands as button presses?
04:39:22***Saving seen data "./dancer.seen"
04:39:36funmanit should be doable with some hacking of the source code
04:40:50damonei was thinking of connecting a hacked nintendo pad to play rockboy
04:41:53kkurbjundamone: without some significant hardware on the pad's side doing that would not be reasonable, you need a chip that can act as a usb host, and you would need a device that uses our software stack
04:42:06damoneI can make a pad that sends commands noproblem
04:42:37damonebut i dont understand much about the RB stack
04:42:44funmanyou need a usb host controller on the other side
04:43:03kkurbjuncool, well, I would imaging using the UART stuff would probably be the way to go
04:43:20damonesure, np
04:43:42 Quit funman ("free(random());")
04:43:54kkurbjundamone, what are you thinking of using for the pad?
04:44:00kkurbjunin terms of a host interface?
04:44:07damoneif someone helps me with the RB side i would be happy to make a open proyect so everyone can make his own pad
04:44:42damoneI think a cheap AVR Micro
04:44:53damoneand a powersupply
04:45:33damonecode to output uart via usb
04:45:51damonei have done it before
04:46:11kkurbjunnice, I didn't realize those devices had host capability
04:50:50kkurbjunThe USB uart as it stands in rockbox is not currently written for controlling the player, but with some hacking I'm sure it could be done - I'm not familiar with the USB UART as I've never used it, but some players use a standard UART for remote controls.
04:51:49damonewell, if anybody is willing to help me with the Rockbox stuff, i can implement a nice host-usb pad to play rockboy =)
04:51:57damoneor doom
04:52:05damoneor any game for that matter
04:52:39 Join safetydan [0] (n=safetyda@rockbox/developer/safetydan)
04:55:22 Join docjay5191 [0] (
04:55:47kkurbjunIt sounds like a really interesting project, if I ever get the USB UART driver working on the player I've been working on lately I would be interested in playing around with it, but my time is limited, so you might be waiting for a while if you relied on me, the channel is more active in the morning (MST), so you might have some luck if you ask around then.
04:56:14 Part docjay5191
04:56:31kkurbjunis there a good demo board for the AVR micro w/ usb host capability?
04:57:13kkurbjunit looks like the AT90USB is the one with host capability
04:58:09kkurbjunand digikey has a kit for it for 30$
04:59:31damonebut there is an open source library that emulates usb host in almost any atmega
05:00:27damoneit's low speed (1.5Mb/s) but I think it should work just fine
05:01:28kkurbjunthat should be plenty fast, I think the mr500 uses the uart at 9600 baud and it's fast enough for games and the like
05:04:33kkurbjunrockbox supports the apple remote interface to some degree, looks like someone did a similar project with the standard uart on the ipods:
05:04:45kkurbjunthe usb uart would be more universal though
05:05:25damoneyeah, I think uart would be the way to go
05:05:26kkurbjunyou could support all of the portal player devices, the gigabeat S, the mr 500
05:09:47damonemmm, also interesting
05:10:44kkurbjunyeah, thats what I was looking at, atmel's site has a coule of host device demonstrations
05:11:00kkurbjunfor that particular USB variant
05:11:14kkurbjunand the source for the demonstrations seems for the most part GPL compatible
05:12:23kkurbjunsorry, I was looking at that board, not LUFA
05:12:43kkurbjunatmel has their own demonstrations that includes a UART setup
05:14:43 Join lifeless_ [0] (n=lifeless@
05:18:50damoneyeah, the implementation pdfs are well explained
05:19:10damonethat one would be nice to use i think
05:19:43kkurbjunwhat about this board:
05:19:57kkurbjunthat gives you everything
05:20:22damoneyeah, but 30 bucks
05:21:03damonei think I can implement a cheaper one using the 3bucks one
05:22:15kkurbjunyeah, it's certainly possible, depends on how much you need extra
05:23:06kkurbjunwith a pcb, connectors, clock (?), etc, it might not be that much less after you factor in your time putting it all together
05:24:36damoneyeah, but I would be happy to make some kind of cheap kit so other people can use it
05:26:13kkurbjunif your kit could compete with a 30$ pre-made then it might not be a bad option
05:28:28damoneobviously im thinking about one integrated to a gamepad
05:29:17damonemaybe i can make a dual one, so one side classic AB and + controls
05:29:43damoneother side, more like WASD + trackball or something for doom
05:30:02kkurbjunahh, so is this less of a diy and more of a buy it pre-made idea?
05:30:27damoneno, not trackball, a linear potentiometer would work better
05:30:27kkurbjunI was just thinking you connect it up to a snes pad or something
05:31:08damonei was thinking more like open development
05:31:22damone?so anyone can buy a premade, kit, or just make one themselves
05:34:38 Quit intrados (Read error: 145 (Connection timed out))
05:35:04LloreanIdeally it would be made with a USB A female in the controller (in my opinion) so that people could just use their sync cables to attach the controller to any SWUSB target.
05:36:08damoneyep, i was thinking the same thing =)
05:37:54 Join intrados [0] (
05:45:24 Quit Horscht ("Verlassend")
05:47:57damonegotta go, cya all!
05:48:18damoneim looking forward to make that controller
05:48:21 Part damone
05:53:38 Quit mc2739 ("Leaving")
05:54:27 Join mc2739 [0] (n=mc2739@rockbox/developer/mc2739)
05:58:34 Join intrados1 [0] (
06:04:16 Quit intrados (Connection timed out)
06:04:58 Join mc2739_ [0] (n=mc2739@rockbox/developer/mc2739)
06:05:12 Quit mc2739_ (Read error: 104 (Connection reset by peer))
06:08:56 Join mc2739_ [0] (n=mc2739@rockbox/developer/mc2739)
06:10:55 Quit mc2739 (Nick collision from services.)
06:11:04 Nick mc2739_ is now known as mc2739 (n=mc2739@rockbox/developer/mc2739)
06:20:30CIA-43New commit by kkurbjun (r22936): Simulator buttons: Remove a hack and make brickmania work the same as on target.
06:25:29CIA-43New commit by kkurbjun (r22937): Brickmania: Fix the flip_sides powerup on touchscreens.
06:37:03 Join JdGordon1 [0] (
06:37:34 Join Ninjanub [0] (
06:37:36 Part rwong
06:37:41NinjanubHey, I need some help
06:38:08NinjanubMy sansa shoes two copys of the same song.
06:38:16 Quit JdGordon1 (Client Quit)
06:38:22NinjanubAny idea?
06:39:26***Saving seen data "./dancer.seen"
06:39:58NinjanubMy sansa E200 is having issues with rockbox
06:41:18cool_walking_Ninjanub: You just have to wait a bit because people aren't always looking at their screen, or this window.
06:41:44cool_walking_and obviously not everyone can help
06:41:46 Quit Ninjanub (Client Quit)
06:45:25 Quit coolrun34 ()
06:46:09 Join coolrun34 [0] (
06:55:51 Quit intrados1 (Read error: 145 (Connection timed out))
07:00:05 Join matsl [0] (
07:02:54CIA-43New commit by kkurbjun (r22938): Brickmania: Add an extra check in the hit logic to make sure that the brick isn't /hit/ after it has been removed from play.
07:06:48 Part mc2739 ("Leaving")
07:10:35 Join intrados1 [0] (
07:22:51 Join einhirn [0] (
07:23:48 Join stoffel [0] (
07:30:15 Quit intrados1 (Operation timed out)
07:39:02 Part safetydan
07:46:02 Quit coolrun34 (Read error: 145 (Connection timed out))
07:47:18 Quit T44 (Read error: 104 (Connection reset by peer))
07:52:38 Quit hd (Client Quit)
07:52:47 Join HellDragon [0] (
07:55:01 Join Topy44 [0] (
08:01:57amiconntomers: Fyi, "langv2" had nothing to do with the language version byte in the generated .lng file, so "langv4" makes no sense
08:02:13 Join T44 [0] (
08:04:03amiconnIt refers to the format of the .lang. The current, tag-based format which allows to restrict strings to certain targets or features is called langv2 versus the old, simpler format
08:05:32 Quit CaptainKwel (Remote closed the connection)
08:07:49tomersamiconn: Thanks. I'll fix that
08:08:34amiconnIt probably makes sense to just write 'lang' now, since the v2 format is in use for quite some time now
08:08:48 Quit Topy44 (Read error: 104 (Connection reset by peer))
08:09:09 Join mc2739 [0] (n=mc2739@rockbox/developer/mc2739)
08:09:57 Quit einhirn (Read error: 110 (Connection timed out))
08:11:02CIA-43New commit by tomers (r22939): Revert usage string change from r22930. ...
08:18:35 Part nosa-j
08:20:42amiconnkkurbjun: Does your comment about the fixed point precision mean that brickmania can be ported to the Player if the precision is upped to 4 bits? ;)
08:21:45 Quit cool_walking_ ("Page closed")
08:23:30 Join intrados1 [0] (
08:24:15 Quit antil33t ()
08:24:24 Join bertrik [0] (
08:26:24 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
08:26:50 Join Grahack [0] (
08:29:57 Join antil33t [0] (n=Mudkips@
08:33:31 Join flydutch [0] (
08:33:58 Quit Grahack ("Leaving.")
08:35:19 Quit GodEater (Read error: 113 (No route to host))
08:36:37 Quit lifeless_ ("Ex-Chat")
08:39:24 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:39:27***Saving seen data "./dancer.seen"
08:39:31 Join Rob2222 [0] (
08:40:26 Join ender` [0] (
08:42:42CIA-43New commit by tomers (r22940): Add USB HID mouse support for H10 ...
08:48:01 Join esperegu [0] (n=quassel@
08:49:45CIA-43New commit by tomers (r22941): Fix red by r22940 (undefined button used)
08:57:04 Quit Rob2223 (Read error: 110 (Connection timed out))
08:57:06 Quit daurn (Read error: 110 (Connection timed out))
08:58:07 Join daurn [0] (i=daurnima@freenode/staff/daurnimator)
08:59:05 Join Thundercloud [0] (
08:59:34pixelmakkurbjun: are the FIXED3 in front of the bmpheight and bmwidth defines in brickmania correct? Just looked for a cause of the pad looking weird on the c200 (can't test until the evening)
09:01:31 Quit antil33t ()
09:04:11 Quit stoffel (Remote closed the connection)
09:06:42pixelmaI admit, I don't know what they are good for... just asking
09:08:26 Quit BHSPitLappy (Remote closed the connection)
09:20:20 Quit bertrik (Read error: 113 (No route to host))
09:32:54 Quit thegeek (Read error: 60 (Operation timed out))
09:37:08 Quit sig11 (Read error: 60 (Operation timed out))
09:41:15 Join einhirn [0] (
09:43:45 Quit Thundercloud (Remote closed the connection)
09:46:35 Join antil33t [0] (n=Mudkips@
09:49:52 Join antil33t1 [0] (n=Mudkips@
09:49:58 Quit antil33t (Client Quit)
09:49:58 Quit antil33t1 (Remote closed the connection)
09:50:05 Join antil33t [0] (n=Mudkips@
09:50:32 Join kugel [0] (n=kugel@rockbox/developer/kugel)
09:51:30kugelkkurbjun: sleep() does yield also. therefore brickmania was garuanteed to yield each frame, and is now likely to yield twice in a row
09:53:00CIA-43New commit by tomers (r22942): Add H10's mouse buttons in USB HID mode to manual
09:59:32kugelheh, kkurbjun hit the same trap I hit. r22936 broke scrollwheel sims (at least the e200 one which I tested). I should've add a comment about why the hack is needed :/
10:04:33 Join casainho [0] (n=chatzill@
10:05:36casainhokugel: ping
10:09:53*Llorean just realized something
10:10:05LloreanWith the addition of a fourth action to the quick screen, we gave up entirely on restoring the original QS functionality?
10:12:52 Join thegeek [0] (
10:15:43 Join antil33t1 [0] (n=Mudkips@
10:15:52 Quit antil33t (Nick collision from services.)
10:16:03casainhokugel: did you bought that Mini2440 board?
10:20:56 Quit kugel (Read error: 110 (Connection timed out))
10:31:19JdGordonLlorean: no...
10:31:32JdGordonand well... its fairly silly to tihnk it was ever coming back
10:37:22 Join robin0800 [0] (
10:38:05LloreanWell, it can't be brought to several targets at all now.
10:39:28***Saving seen data "./dancer.seen"
10:43:40 Quit casainho ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824085743]")
10:44:02 Join darkham [0] (
10:45:39markundo we want to support itunesnorm if there are no replaygain tags?
10:46:24Zagoris there any reason not to?
10:47:28markunI guess not :)
10:47:57markunI found a formula to convert it here:
10:48:10markun(but the other way around)
10:48:31markunhow is the gain value stored in rockbox? in dB or as a factor?
10:51:20 Quit BHSPitMonkey (Read error: 113 (No route to host))
10:55:02 Quit ender` (Read error: 145 (Connection timed out))
11:02:46 Quit darkham ("Sto andando via")
11:02:57 Join Lynx_ [0] (
11:07:23 Join KBH [0] (
11:07:55 Join petur [0] (n=peter@rockbox/developer/petur)
11:25:38 Quit HBK (Read error: 110 (Connection timed out))
11:32:22 Join barrywardell [0] (n=barrywar@rockbox/developer/barrywardell)
11:32:24LloreanJdGordon: Why not have the quick screen simply launch the lua program on a set-named lua script, then people can extend the lua functionality so that they can have the quickscreen do *anything*
11:32:48JdGordonthats what I was tihnking with being able to run plugins
11:32:56JdGordonbut I dont know if lua can talk?
11:33:42LloreanYeah, but I'm sure it could be expanded to have access to the normal voices at least
11:34:00*JdGordon glares at midgey :p
11:34:14JdGordonnot untill plugins in general can talk
11:35:05JdGordonI wonder if any of the core could be replaced with lua scripts? trim it down a bit maybe
11:36:43*JdGordon wonders how doable playlist "shuffle append" would be to add
11:37:02rasherHeyo, rtl menus
11:38:53JdGordonyep :)
11:40:13*JdGordon tries out the patch
11:41:03 Quit daurn (Read error: 60 (Operation timed out))
11:41:04 Quit JdGordon (Read error: 54 (Connection reset by peer))
11:41:48 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
11:41:52JdGordonapparently lang files arent rebuilt(?) with make if english.lang wasnt changed
11:54:03 Join DerPapst [0] (
11:58:33 Join daurn [0] (i=daurnima@freenode/staff/daurnimator)
11:59:41 Quit robin0800 (Remote closed the connection)
11:59:57 Join robin0800 [0] (
12:01:28 Quit robin0800 (Remote closed the connection)
12:02:31CIA-43New commit by tomers (r22943): Manual: Add instructions to switch USB HID keymap modes
12:03:03tomersrasher, JdGordon: Let me know how it goes for you :-)
12:03:47JdGordonwell, just tried it and it works :)
12:04:10JdGordonexcept it looks like the icons got squshed.. it might just be im not used to the c200 display though
12:04:14tomersIt broke some screens, and I must have been wrong here and there...
12:04:59JdGordoni just tried the lists
12:05:19tomersTry to reboot, and watch
12:06:23JdGordonapart from that comment I added, i dont see any reason to not commit it in the current form and fix the screens as they show up
12:07:57tomersI am a new committer... This introduces instability (only for the Hebrew and Arabic langs). I think I needed to get feedback before committing. Are you sure it is ready (after fixing your remark)?
12:08:04*linuxstb wonders if anyone has any ideas about what kind of bug could cause some screens in Rockbox on the Nano2G to not display the actual list item text (or the selection bar) on some screens. For example, the file browser works, as does most of the menus, but the "browse XXX" menu items don't...
12:08:32*rasher gets nervous
12:09:11 Join Lynx [0] (
12:09:34JdGordontomers: well, i can only say from the code in the list it looks fine... the statusbar handling looks a bit wierd also now that i give it another look :p
12:09:38 Nick Lynx is now known as Guest45437 (
12:10:16tomersJdGordon: But do you think I should commit and fix, or wait until I fix it first?
12:10:27JdGordonlinuxstb: YES!!!! same problem as the mr500 had... 0x0000000 is valid RAM but that is where the translated string pointters are set to...
12:11:21JdGordontomers: small graphical glitches are well worth it IMO... if it causes crashes then no, but it doesnt.. (does it?)
12:11:24tomersDoes someone knows what triggers the daily manual build? Is it being compiled daily at a certain hour?
12:11:43linuxstbJdGordon: Ah, I sort of remember that. What's the fix?
12:11:43ZagorJdGordon: pointers as in plural? surely only one text should be there?
12:11:43tomersJdGordon: It doesn't. I'll commit soon
12:12:07tomersJdGordon: Not using the screen's width
12:12:16JdGordonZagor: be pedantic later :p
12:12:47Zagorheh, I was actually confused rather than pedantic
12:13:41JdGordonlinuxstb: settings.h around line 194
12:14:34tomersJdGordon: In order to fix the graphic glitches, I would like to add rtl parameter to some the lcd's API (which is being shared with the plugins, which are language agnostic). What's your opinion
12:14:39JdGordonZagor: ah sorry :) the fake addresses the translated strings get set to is supposed to be in invalid RAM so functions know if it needs to do the translation or not
12:15:02ZagorJdGordon: aha
12:15:38JdGordontomers: my only opinion is rtl support is great and anything it needs should be done...
12:16:05JdGordonI wonder if rtl should be something linked to viewports?
12:16:54tomersWhat do you mean exactly?
12:18:47JdGordonyou wouldnt need to change the api at all then, the default NULL viewport would be set for ltr, and then lists would be rtl, and all the graphical glitches would disappear
12:20:38JdGordonthat way screens which arnt converted to work will still be usable... just slightly wrong
12:26:01 Quit Lynx_ (Read error: 110 (Connection timed out))
12:26:01 Quit notlistening (Read error: 104 (Connection reset by peer))
12:26:01 Nick Guest45437 is now known as Lynx_ (
12:26:08CIA-43New commit by jdgordon (r22944): Fix FS #10617 (fix spotted by Teruaki Kawashima), something about the %C not working if %?C is the first tokens in the skin
12:30:13*linuxstb doesn't understand that VIRT_PTR fix, and simply adding "|| CONFIG_CPU==S5L8701 doesn't fix the Nano2G
12:30:13*TheSeven is confused by FTL failing to unmount the flash as soon as you've played audio once
12:32:33JdGordonlinuxstb: you need to have VIRT_PTR set to a pointer which is invalid, or unmapped, or whatever
12:33:21JdGordonTheSeven: audio plays, data is written to disk... is that supposed to work?
12:34:25TheSevenno idea, but even after stopping audio playback, the unmount will fail when shutting down
12:34:48TheSeven(and i also once saw an error message that it couldn't write some settings file when I was trying to pause)
12:35:16JdGordon.playlist_control and config.cfg both get written on shutdown
12:35:22JdGordonand during playback
12:36:02linuxstbJdGordon: How much IRAM does the DM320 have?
12:37:15*linuxstb sees "#define IRAMSIZE 0x4000"
12:37:25*tomers Needs Latex guru to resolve "FS #10652 - Manual: Can't use nopt inside btnmap "
12:38:11TheSevenlinuxstb: any idea on how to resolve those lockups when writing to the wmcodec?
12:38:17linuxstbJdGordon: So for the Nano2G (176KB IRAM), how can it be fixed?
12:39:09JdGordonI guess using 0xffff0000 would work, assuming thats more than the amount of RAM it has, and it has nothing mapped to that are
12:39:15linuxstbTheSeven: No... Unless it's some kind of interrupt issue, with i2c and I2s interrupts happening together.
12:39:30***Saving seen data "./dancer.seen"
12:40:27TheSevenJdGordon: I'm currently compiling a build with them mapped to an address space hole
12:41:17JdGordonassuming it is infact the same bug.. sounds like it is thoguh
12:41:27TheSevenJdGordon, linuxstb: Yep, that *fixed* it !
12:41:45JdGordonno ! needed.... thats how its supposed to work
12:41:59JdGordonthat whole virtual pointer stuff shold be moved out of settings.h ....
12:44:46TheSevenlinuxstb: I'll go for I2C polling instead of interrupts now, let's see if that works better
12:45:28TheSevenor just add a timeout to the interrupt driven i2c
12:49:35TheSevenok, polling-based i2c fixed it
12:49:55TheSevenand if i look at system-s5l8700.c, i also see why.
12:52:23TheSevenbtw, system_reboot() has a typo in the ifdef
12:54:27TheSevenand if i fix that it reveals a red.
12:55:13TheSevenand if i fix that, it reveals another bug.
13:00:41 Quit intrados1 (Read error: 105 (No buffer space available))
13:01:10 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:06:49TheSevenlinuxstb: what should I do about those IRQs getting lost?
13:07:16TheSevenjust add a timeout to the wakeup wait and call it in a loop until the completed bit is set?
13:14:25linuxstbWhy would the interrupts get lost?
13:15:48TheSevenno idea, but some I2C ints just never get through
13:16:29 Join evilnick [0] (n=evilnick@rockbox/staff/evilnick)
13:16:37TheSevenbtw, how do plugins and codecs access the firmware functions? don't they need to be re-linked after compiling a new firmware?
13:17:34*TheSeven is getting annoyed by builds taking too long
13:18:04TheSevenvmware really doesn't seem to do a good job with disk I/O
13:24:49linuxstbTheSeven: They use the "plugin API" - basically a struct of pointers (to variables and functions. See apps/plugin.[ch]
13:24:50 Join mt_ [0] (n=MTee@
13:30:59 Quit TheSeven (Remote closed the connection)
13:32:05 Join TheSeven [0] (
13:32:25 Part LinusN
13:34:07 Join Sajber^ [0] (
13:36:25TheSevenlinuxstb: as soon as i change some code anywhere, the FTL starts behaving a little strange, until i do a full rebuild
13:38:12 Join MethoS- [0] (n=clemens@
13:38:27 Join funman [0] (n=fun@rockbox/developer/funman)
13:39:09*linuxstb scratches head
13:39:34TheSeventhis *could* in fact be caused by *copying* that new build, I'll need to check this
13:40:21TheSevenbut i'm still wondering where these I2C ints get stuck
13:40:42TheSevenbtw, why are you feeding the DMA such small chunks of audio and not the whole PCM buffer?
13:42:49 Quit mt (Read error: 101 (Network is unreachable))
13:43:40CIA-43New commit by tomers (r22945): RTL support in menus
13:51:42JdGordonarg... the settings_list.c change was naughty!
13:52:11 Join AsaelReiter [0] (
13:52:41JdGordonnot really a big deal, but it might annoy some people
13:53:33AsaelReitertomers: כשנמצאים בתפריט, הכפתור השמאלי חוזר לרמה הקודמת, והימני נכנס לרמה הבאה. לא כדאי להפוך את זה ב-RTL?
13:54:15JdGordonenglish in this channel only please
13:54:28tomersAsaelReiter: The keys will definitely be flipped in RTL mode in the near future
13:54:43AsaelReitertomer: thanks.
13:56:12markunmaybe we should have an optional UTD mode for chinese and japanese ;)
13:56:22TheSevenlinuxstb: do you think i should commit that i2c hack?
13:56:47JdGordontomers: lots of red to fix :)
13:57:05 Quit barrywardell ()
13:57:40tomersJdGordon: Working on it...
13:58:14*JdGordon wants inverted-upsdiedown-top-left-to-bottom-right!
13:59:57CIA-43New commit by tomers (r22946): Fix some of the reds
14:00:49 Join barrywardell [0] (
14:01:11gevaertsZagor: how long does it take for changes in blockedclients to take effect?
14:01:49Zagor10 minutes iirc
14:02:53gevaertswhen I changed it a few days ago it seemed to take long
14:03:29Zagorit wasn't really tested when I added it so it is perhaps buggy
14:05:03*gevaerts looks for fixed reds :)
14:05:50linuxstbTheSeven: Regarding your DMA question, I'm not sure - that's just how it works on all other targets. Maybe they had a limit on DMA size.
14:06:40 Quit AsaelReiter ("CGI:IRC")
14:07:43linuxstbTheSeven: Regarding committing, my preference is always to wait and try and find a real solution. But I guess that for new ports, it's not as important to do things right first time...
14:08:50TheSevenyep, it's also bugging me why those interrupts get lost in the first place, but we can at least prevent that from locking up the main thread
14:10:17*TheSeven will wait for tomers to fix the reds before committing
14:12:09 Join intrados1 [0] (
14:16:31CIA-43New commit by zagor (r22947): Fixed html
14:17:01 Quit JackWinter2 (Read error: 104 (Connection reset by peer))
14:17:02linuxstbTheSeven: BTW, the "view OS stacks" debug screen now works (with JdGordon's fix), and that shows the main stack is only reaching 41% for me (with a 16KB stack). Even after trying pictureflow (but I have no albumart on my nano2g)
14:17:45TheSeveni didn't have albumart either, but pictureflow stkov'ed at 12K
14:18:58CIA-43New commit by tomers (r22948): Fix red - forgot one #ifdef
14:19:29 Join JackWinter [0] (
14:19:31 Quit AlexP (Read error: 110 (Connection timed out))
14:19:44CIA-43New commit by tomers (r22949): Fix red - bootloaders don't link language support
14:20:54 Join tomers_ [0] (
14:21:14*tomers_ Lost Internet connectivity while trying to commit fixes for reds
14:21:37tmzt01:42 < CIA-43> New commit by tomers (r22949): Fix red - bootloaders don't link language support
14:21:48tmztand 948
14:24:01 Quit mt_ (Read error: 110 (Connection timed out))
14:24:16linuxstbTheSeven: Could you check what your stack usage is? Maybe it depends on the recursion level or something... (my directories are shallow)
14:24:39TheSevenlinuxstb: I just have .rockbox and 2 MP3 files in the root
14:25:56TheSeveni get main 86% after running pictureflow
14:27:05TheSeven22% after boot
14:28:43*tomers_ No more reds :-)
14:28:47linuxstbTheSeven: Strange... Do you have any other files on the disk ?
14:28:55CIA-43New commit by theseven (r22950): Add myself to COMMITTERS
14:28:59*linuxstb doesn't have a clue what pictureflow does...
14:29:03CIA-43New commit by theseven (r22951): Fix displaying strings on stack for iPod Nano 2G
14:29:09CIA-43New commit by theseven (r22952): Fix a #ifdef typo and the reds hidden by it
14:29:16CIA-43New commit by theseven (r22953): Reduce impact of lost interrupts on S5L8700 I2C
14:29:23 Quit tomers (Read error: 113 (No route to host))
14:29:23CIA-43New commit by theseven (r22954): Make iPod Nano 2G audio work
14:29:38linuxstbWelcome - you'll shortly own the front page of ;)
14:29:55 Join froggyman [0] (i=www-data@
14:29:58 Join maruk [0] (
14:30:01mc2739TheSeven: congrats!
14:30:26tomers_Congrats!!! How do you commit several patches at once?
14:30:44TheSeveni just committed them one after another
14:30:54funmantomers_: you can also use git svn dcommit
14:30:59tomers_Quick typing :-)
14:31:36funmanTheSeven: so you now have sound ?
14:31:57TheSevenlet's see if audio actually works with that build - i haven't increased the stack size (which i did when testing)
14:32:42funmanwell, according to you have had sound for some time already o:
14:33:15TheSevennot in an app build
14:34:16 Join robin0800 [0] (
14:35:28TheSevencan anybody try actually building and testing r22954 with a clean tree?
14:36:49TheSevenlinuxstb: in fact, the front page looks really nice :-)
14:37:49*funman quickly finds 5 consecutives things to commit
14:39:24*TheSeven wonders what's causing some targets to change binary size on almost every unrelated commit
14:39:33***Saving seen data "./dancer.seen"
14:40:14*JdGordon has a damn wierd bug he cant figure out
14:40:20funmanTheSeven: random gcc behaviour
14:40:40JdGordonthe stupid peakmeter apparently makes hidden viewports clear themselves and stop the scrolling.. even when they were hidden anyway
14:41:49janne_oksanenthere's a smiley face character in UTF-8?
14:43:10marukTheSeven: I will try r22954 tonight (I'm in France).
14:44:07 Join kachna [0] (
14:44:54 Nick evilnick is now known as evilnick_home (n=evilnick@rockbox/staff/evilnick)
14:45:03TheSevenbtw, just noticed "view partitions" in the debug menu assumes 512-byte sectors
14:45:09 Quit intrados1 (Read error: 60 (Operation timed out))
14:45:55TheSevenand "disk info" is printing some random garbage for vendor, model, ...
14:46:33Zagoryou mean your flash doesn't have a proper IDENTIFY response? :)
14:47:41TheSevenwell, this isn't even implemented in the driver
14:47:50*TheSeven wonders where that screen gets its information from
14:47:58marukTheSeven: 1st, it compiles cleanly.
14:48:06 Join jgarvey [0] (n=jgarvey@
14:48:37TheSevenwell, not really cleanly, there are some yellows
14:49:34 Join intrados1 [0] (
14:54:14TheSevennow let's go for a better NAND driver
14:55:44JdGordonassigning a bool to a bool should never cause any problems right? this is without any type casting at all...
14:56:15linuxstbTheSeven: IMO, #elif is nicer... (in settings.h). You also could have used "defined(CPU_S5L870X)" (and a higher address) to save bertrik a future headache.
14:57:05 Join DaveDave1port [0] (
14:57:07TheSevenlinuxstb: OK, I'll change that in my next commit round
14:57:46*TheSeven wonders if we know the NAND controller's interrupt
15:00:31CIA-43New commit by tomers (r22955): Have scollbar in the right side while in RTL mode
15:08:41 Quit DaveDavenport (Read error: 110 (Connection timed out))
15:13:33 Join faemir [0] (n=faemir@
15:20:09CIA-43New commit by tomers (r22956): Fix yellow
15:24:11*TheSeven wonders why the main thread lockups when changing volume came back with his nand driver improvements
15:29:36 Quit froggyman ("CGI:IRC (EOF)")
15:43:06preglowhas there been any shit with the mailing lists lately?
15:43:18preglowfor some reason, i've stopped getting rockbox related mails
15:43:40janne_oksanenI got a bunch of messages during the night
15:43:55janne_oksanenso it was at least working this morning
15:43:56Zagorpreglow: yes it has. at&t added us to their spam block list temporarily, which created massive bouncing and resulted in some accounts being unsubscribed
15:44:20ZagorI meant to write that on the front page, but it got lost in my head somewhere on the way...
15:44:37preglowguess i'll have to go about subscribing again then
15:45:22kkurbjunpixelma: FIXED3 is needed around the bitmaps - everything should be positioned/calculated in fixed point in the main portion of the program and the fractional dropped when making a call to a function that needs just the integer portion - INT3 drops the fractional.
15:45:42kkurbjunkugel: why is that hack needed?
15:46:17 Join evilnick [0] (i=0c140464@rockbox/staff/evilnick)
15:48:29 Join mt [0] (n=MTee@rockbox/developer/mt)
15:48:54kkurbjunamiconn: I guess it might be possible, isn't the player a charcell though?
15:51:44amiconnIt is, but there are a few user definable characters. See the playergfx lib :)
15:52:11amiconn20x14 pixels on the newplayer lcd (which is the more common type)
15:53:49 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
15:57:11bluebrothertomers_: what's the point of rr22955? You can select the scrollbar to be left or right, and I don't see a reason why it should do something automagically, that's only confusing.
15:57:31bluebrotherplus, the settings value in the menu is "left" and "right" ...
15:57:49TheSevenboth the NAND and ECC driver don't like yielding, any idea why?
15:57:59*bluebrother wonders if that change was discussed first
15:58:52ZagorTheSeven: the driver as such, or the device?
15:59:08Zagorthe device not liking it sounds like strict timing
15:59:16*bluebrother spots an earlier commit and still thinks it's confusing.
15:59:30tomers_bluebrother: If I switch between Hebrew and English, the meaning of having a scroll bar next to the tree's root is kept. I tried to discuss it but no one participated in the discussion. Do you think we should revert?
16:00:10bluebrotherwell, a user has no idea what "show opposite" would be on showing the scrollbar. That's quite confusing.
16:00:37 Join kugel [0] (n=kugel@rockbox/developer/kugel)
16:00:37tomers_It makes the code nastier, but that's doesn't matter. The user's perspective is important. I wanted two have similar meaning for both RTL and LTR users. The default for RTL users is against intuition...
16:00:52bluebrotherhow do other applications (f.e. windows) handle RTL? Do they switch the position of their scrollbars?
16:00:52kugelkkurbjun: try an e200 sim
16:01:10tomers_bluebrother: They switch the position of the scrollbars
16:01:25kkurbjunI don't like the idea of fliping the scrollball automatically, or making the code nastier, if the user prefers it on the other side they can flip it manually
16:01:38kkurbjunkugel, I did, but do you know why it doesn't work?
16:01:53bluebrotherhmm. Then at least the wording in the menu should get improved.
16:02:09LloreanI tend to agree, if there's a manual scrollbar position option, automatically swapping it is bad. Users can set it as they like explicitly.
16:02:16linuxstbtomers_: Nasty code definitely matters... (as a general rule).
16:02:17tomers_I will let you decide either way, and do what you decide
16:03:06kkurbjuntomers, the position of the scrollbar is effectively arbitrary, an English speaker might use it on the left or right side and the same goes for any other language, I don't think it should be a forced flip
16:03:11bluebrotherWindows, f.e., defaults to displaying the scrollbar on the right in LTR languages. Rockbox shows it on the left as default. Just having an option to "show" and "show opposite" the scrollbar is counter-intuitive for users.
16:03:12 Quit JdGordon (Read error: 113 (No route to host))
16:03:35bluebrotherdo the icons get flipped as well?
16:03:46linuxstbTheSeven: What happens when you yield in the NAND/ECC drivers? e.g. an immediate freeze or random freezes?
16:03:49tomers_I think default for LTR is left (as it is), but default to an RTL language is in the right
16:04:16tomers_bluebrother: The icons are shown in the right. The bitmap itself is not being mirrored
16:04:31bluebrotherit could be called as "scrollbar at beginning of text", and "at end of text".
16:04:55bluebrothertomers_: I wouldn't expect the icons to get mirrored, I was referring to their position :)
16:04:55tomers_Should I have such a long string as an option?
16:05:15TheSevenlinuxstb: no freezes, just the transfer failing
16:05:20tomers_Their position is mirrored already (still having some glitches, though)
16:05:26bluebrotherwell, a shorter string would be better, of course. Still I think the current wording is confusing at best.
16:05:30TheSevenso probably something messing with some resources used by it in between
16:05:51bluebrother"show scrollbar" "start of line", "end of line" maybe?
16:05:53tomers_I agree that it is confusing. Can you please think of an alternative wording?
16:06:04tomers_Looks nice to me
16:06:09tomers_Do it
16:06:19kkurbjunI think it should just be left or right side as it was
16:06:30LloreanI think left or right is fine.
16:06:31kkurbjunit made sense and worked intuitively
16:06:39Lloreanespecially if it means the code can be simpler.
16:06:53linuxstbTheSeven: Maybe different threads accessing the disk is causing issues. So some kind of locking would be needed.
16:06:56LloreanScrollbars are more user preference than anything else anyway.
16:07:05bluebrotherLlorean: left or right as in not adjust according to RTL / LTR?
16:07:26Lloreanbluebrother: exactly, yes.
16:07:26kkurbjunbluebrother: that is my vote
16:07:33tomers_kkurbjun, Llorean: The code it nastier with left/right. It's not good for RTL users to have scrollbar defaulted to the left
16:07:47linuxstbTheSeven: Although I'm sure other drivers must handle these waits in some way, so looking at some other low-level NAND/ATA drivers might give some tips.
16:08:01Lloreantomers_: At a LTR user I have scrollbars on the right in 99% of my software anyway. I don't quite understand why it's "not good"
16:08:10kkurbjunwhy is it bad for them to have it left or right
16:08:17bluebrotherI'm in no strong favor if it should be a simple left-right or a start-end of line setting.
16:08:39LloreanScrollbars are a themeing option, not a language option, aren't they?
16:08:50bluebrothertomers_: why it's "not good" to have teh scrollbar on the left for RTL languages?
16:08:58LloreanEspecially with later themes, they'll probably be skinnable. Language shouldn't affect themes if at all possible.
16:09:14LloreanTo me this means scrollbar position should never be changed unless manually by the user, or by a change of theme.
16:09:39tomers_Llorean: mmmm... Interesting points. Are all in favor of reverting to Left/Right?
16:10:03bluebrothertomers_: why is the code nastier with left-right? It's still putting the scrollbar somewhere, and it simply needs to decide where to put it depending on different values
16:10:33tomers_bluebrother: Checked it and it's not that bad. I take my words back.
16:10:48tomers_I'll do it tonight
16:12:11*bluebrother suggests putting such changes on the tracker first so that more people can comment on it. IRC is always dependent on the people currently being online.
16:20:24bluebrothertomers_: btw, noticed that the HID setting description in the manual is somewhat off of the LatexGuidelines: description items should end with a ., not :, and the space afterwards should be outside of the []. See
16:20:53bluebrothernot that it's a big problem, but it would be good to be consistent here. I don't believe we're really consistent yet, but anyway ...
16:24:12funmanSome progress on recording on Sansa AMS: I see an I2SIN interrupt with the fifo full bit set, but no consecutive interrupts
16:25:31funmanAnd i found why headphones were muted
16:28:06 Join liar|netbook [0] (n=liar@
16:29:30TheSevenwhich header do i need for panicf?
16:30:30funmanpanic.h iirc
16:30:52TheSevenah, i was trying panicf.h
16:31:55TheSevenplayback starts almost instantly now, and no more gui lockups
16:32:10gevaertsjust by adding panicf?
16:32:53TheSevenno, that was just to fix a yellow i caught while fixing the nand driver
16:33:58funmansadly using the recording patch on clip deadlocks instantly when entering recording or fm screen
16:34:42TheSevenlinuxstb: you suggested using a different define in settings.h, but I can't see that defined anywhere?
16:35:13tomers_bluebrother: I wasn't aware of this document. I'll fix that. BTW can you help me with "FS #10652 - Manual: Can't use nopt inside btnmap" ?
16:35:58funmanFM screen deadlocks on current build as well :/
16:36:25linuxstbTheSeven: firmware/export/config.h
16:36:25gevaertsisn't it time to add the nanog2 to the build table?
16:36:39linuxstbgevaerts: If you want to fix our yellows first... ;)
16:36:45linuxstbBut yes, it's close.
16:36:56 Join Lss [0] (
16:37:06TheSeventhere's quite a lot of them, and I'll introduce 4 more with my next commit which I won't bother to fix
16:37:41TheSevengevaerts: there will even be a red with current SVN
16:37:55 Quit barrywardell ()
16:38:03TheSevenlinuxstb: found a solution for that one in the meantime?
16:38:16TheSeven(other than dropping vorbis support :-D )
16:38:17gevaertsok, better wait a few more minutes then :)
16:39:08bluebrothertomers_: I'
16:39:36***Saving seen data "./dancer.seen"
16:39:41bluebrotherve spotted the FS entry, though I haven't found time to look into details yet. opt is some kind of beast ...
16:40:35preglowZagor: could you check if i'm still subscribed to the commiters list? i don't know of any way in which to do that
16:40:39preglowaddy starts with thomj@
16:41:20funmanpreglow: did you receive an email yesterday ?
16:41:30Zagoryou were nomail-flagged for bounces. I cleared that flag now.
16:41:46bluebrotherI was also planning to replace opt by a c-like preprocessor. It's already somewhat working, but gpp (the preprocessor I used) doesn't know about variadic macros. And the gcc preprocessor doesn't trim removed lines which breaks LaTeX, unfortunately
16:42:21CIA-43New commit by theseven (r22957): Fix the stack string issue for the Meizus, too
16:42:24 Quit matsl (Read error: 110 (Connection timed out))
16:42:27CIA-43New commit by theseven (r22958): iPod Nano 2G NAND/ECC driver and FTL improvements (still polling)
16:42:35preglowfunman: no
16:43:02preglowZagor: so i'm just flagged on the rest of the lists too, then? i assume i can't just resubscribe and have things work again
16:43:17ZagorI'll take a look
16:43:49 Join toffe82 [0] (n=chatzill@
16:44:52Zagorpreglow: flag cleared in on rockbox, -dev and -cvs. I didn't see you on -sf
16:45:50preglowi monitor flyspray by other means, so no
16:45:56preglowZagor: thanks
16:46:10*TheSeven doesn't like to make ECC interrupt-driven before we worked out what's causing that interrupt loss
16:46:24linuxstbTheSeven: Why is vorbis giving a red? It doesn't for me.
16:46:26*tomers_ going
16:46:36*TheSeven needs to re-check
16:46:40TheSevenin the morning, it still did
16:46:52linuxstbI fixed it over the weekend, I forget exactly when.
16:46:59linuxstb(in SVN)
16:48:37funmancan you listen to the radio while recording it on e200/c200 ?
16:50:08mc2739funman: you can record the radio while listening to it on the e200
16:51:08 Quit antil33t1 (Read error: 104 (Connection reset by peer))
16:51:14 Join antil33t [0] (n=Mudkips@
16:52:04 Join explore [0] (
16:54:25TheSevenwow. is there really no FMC int on the s5l8700?
16:55:11mc2739funman: I don't know think you can record on the mic and listening to the radio
16:55:55funmanmc2739: thanks, i meant record the radio & listen to the radio at the same time
16:58:29funmani'm using e200/c200 code from audio-c200_e200.c (as3514 specific) so I'll assume it's correct
16:58:32 Join barrywardell [0] (
17:00:14 Quit evilnick_home (Read error: 104 (Connection reset by peer))
17:00:33 Join evilnick_home [0] (
17:05:14 Join ender` [0] (
17:06:26 Quit Zagor ("Don't panic")
17:06:33 Quit krazykit ("Connection reset by beer")
17:07:37 Quit ender` (Client Quit)
17:10:11liar|netbookTheSeven: wow, the latest svn works even with my nano 2g
17:12:34TheSevenhuh? I didn't change *anything* regarding the way it mounts...
17:13:40 Join ender` [0] (
17:14:08 Nick funman is now known as a (n=fun@rockbox/developer/funman)
17:14:18 Nick a is now known as a_home (n=fun@rockbox/developer/funman)
17:14:23liar|netbookTheSeven: you changed nothing in ftl_open?
17:14:35liar|netbookTheSeven: or in ftl_vfl_read?
17:14:39 Nick a_home is now known as funman (n=fun@rockbox/developer/funman)
17:15:36TheSevenonly some mutex stuff, and before that patch this just can't have been called multiple times simultaneously
17:17:26kugelkkurbjun: the data in button_read_device() and button_get_data isn't entirely the same
17:17:36kugelI was confused by that also but didn't investigate it much further
17:17:39 Quit Lss (Read error: 104 (Connection reset by peer))
17:22:50 Join krazykit [0] (
17:22:51TheSevenlinuxstb, gevaerts: that red has indeed went away
17:25:53 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
17:26:36linuxstbTheSeven: If pictureflow is definitely giving you that high stack usage (as proven in "view OS stacks"), then I guess we need to increase the stack. Do you want to commit that change?
17:26:59TheSeveni'll try it with *some* albumart on my ipod first
17:27:49linuxstbUnless of course there's some known bug with pictureflow...
17:31:42funmanmc2739: I have posted an update to fs#10371
17:38:20 Quit barrywardell ()
17:41:18 Join barrywardell [0] (
17:43:22 Join panni_ [0] (
17:46:01TheSevenlinuxstb: I found out what's going on with the timers :-/
17:46:15TheSeventhis is really ugly :-(
17:47:21 Quit kugel (Read error: 110 (Connection timed out))
17:52:07linuxstbTheSeven: What?
17:54:36TheSeventhey muxed the ints
17:57:00 Quit funman ("free(random());")
17:58:28 Quit panni_ (Read error: 104 (Connection reset by peer))
17:59:02 Join panni_ [0] (
17:59:37 Quit barrywardell ()
18:00:29linuxstbTheSeven: Just the timer ints?
18:01:29CIA-43New commit by theseven (r22959): Fix the user timer on iPod Nano 2G
18:01:43TheSeventrying doom now
18:03:10TheSevenok, doom works, even though the LCD is a little small
18:04:35*gevaerts expects more enthusiasm for that
18:07:23 Quit faemir ("Leaving")
18:09:23TheSevenlinuxstb: OK, what's next on our list?
18:09:25linuxstbTheSeven: I have a feeling there was a reason bertrik chose TIMERD for the Meizu...
18:09:47linuxstb(but I may be wrong... Although we should check with him.)
18:09:50 Join coolrun34 [0] (
18:09:59TheSevenyep, I'll ask him as soon as he turns up
18:10:21TheSevenif there was one, this'll need a file split
18:10:31linuxstbOr we use TIMERB?
18:10:43linuxstbOr rather, we find a pair of timers that work on both...
18:10:52TheSeventimerb is the tick (on both, i think)
18:12:04linuxstbBTW, have you noticed the "test_*" plugins in apps/plugjns? They could be useful for testing/benchmarking/optimising things. There's one for disk access, and another for LCD.
18:12:37linuxstbI guess a nice power-off would be quite useful at this stage...
18:14:13linuxstbPlus we need a "lcd_blit_yuv()" function to get mpegplayer working.
18:15:54 Quit antil33t (Read error: 54 (Connection reset by peer))
18:16:07 Join antil33t [0] (n=Mudkips@
18:18:31 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:20:52 Quit maruk ("Leaving.")
18:22:42 Join wincent [0] (
18:24:00 Quit krazykit ("Connection reset by beer")
18:25:34TheSevenlinuxstb: can our LCD controllers do YUV to RGB in hardware? if no, we'd just need to copy that from another target, right?
18:26:59 Quit kachna (Read error: 110 (Connection timed out))
18:30:12 Join bertrik [0] (
18:31:47 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:32:31bertrikehm, the s5l8700 timer change for nano2g probably breaks other s5l8700 targets
18:33:18tomers_bluebrother: Please approve "FS #10654 - Fix descriptions in manual to obey Latex guidelines"
18:33:26tomers_so I can commit
18:34:27TheSevenbertrik: that's what I was about to ask you. Which timers can safely be used on Meizu?
18:34:47TheSevenIf this is possible, I'd like to use the same numbers on the nano, but timer D is the piezo there
18:35:35bertrikthe various targets use different hardware timers for pwm, piezo, etc.
18:36:00bertrikso some timers need to be on a specific timer instance, while others can be chosen more freely
18:36:06bluebrothertomers_: nice to see someone actually going through the manual and fixing it :) Looks really good.
18:36:30TheSevenyes, the question is: is timer C used for such a thing on any target?
18:36:30bertrikwe'd have to check the different wiki pages about the use of the timers
18:36:51bertrikTheSeven, that needs to be figured out by searching the s5l8700 target wikis
18:37:43bluebrotherone minor hint: you changed one double-space to a single space (last change in bookmarking.tex). Though it doesn't do any harm it's not necessary −− LaTeX treats multiple whitespace as single whitespace. Putting a bigger gap after a sentence ending stop is common in some languages (though not english)
18:38:00bertrikTheSeven, what timers does the nano2g use?
18:38:11bertrik(for hardware functions like pwm, piezo, I mean)
18:38:35bluebrotherLaTeX has an option for this, it's called "frenchspacing". If we would use that we didn't need to care about as LaTeX does this itself, the only thing we would need to care about is marking stops that don't end a sentence as such.
18:38:39TheSeveni think only timer D = piezo, the others didn't show any obvious hardware reaction
18:39:39***Saving seen data "./dancer.seen"
18:40:29CIA-43New commit by tomers (r22960): FS #10654 - Fix descriptions in manual to obey Latex guidelines
18:41:03 Nick tomers_ is now known as tomers (
18:41:17bertrikok, I'll try to make a list
18:42:52bluebrothertomers: oh, and another thing about whitespace in case you don't know: the only situation where LaTeX can get problems with whitespace is when you insert hard spaces, i.e. a "\ ". Unfortunately opt adds some possible breakage as well (though I haven't understood what exactly is happening in those cases, the result is at least an added hard whitespace)
18:43:28bertrikTheSeven, the nano2g backlight is not connected to a timer output?
18:43:51TheSevenno, that's handled by the PMU
18:44:24bertriktimer A = Samsung YP-S3 PWM for LCD backlight, timer B = unused, timer C = Meizu PWM for LCD backlight, timer D = nano2g piezo
18:45:08bertrikrockbox needs at least two free timers: one for the kernel tick and another for a generic timer
18:46:14tomersbluebrother: When are the daily manuals gets compiled?
18:47:03bluebrothertomers: I don't know (Zagor might know), but I _think_ somewhere between 4am and 6am european time
18:49:59bluebrotherhmm, browsing the folders on the download server comes up with something around 9pm, though that listing doesn't tell about the timezone used. Maybe the time the manuals are built has changed, too
18:51:01 Quit tomers (Read error: 54 (Connection reset by peer))
18:51:29 Join tomers [0] (
18:51:41TheSevenbertrik: so we'll need to do at least 2 versions
18:51:47 Join Grahack [0] (
18:52:10gevaertscan't you share the kernel tick and the piezo? :)
18:53:35polobricolois ther a define which includes all the Portal Player ipods ?
18:54:45gevaertswhat exactly do you need?
18:55:17polobricoloi'm going to work on the ipod accesory protocol
18:55:48 Join GodEater [0] (
18:56:01polobricolobut the actual source code uses the pp uart driver without checking if you are actually building for a pp ipod
18:56:48TheSevenlinuxstb: I'm observing UI freezes if i change volume while album art is still loading
18:57:31gevaertspolobricolo: there seems to be IPOD_ARCH, which is defined to 1 by PP ipod targets, but I'm not sure how safe that one is
18:58:25TheSevenwhat about adding a PORTALPLAYER or IPOD_PORTALPLAYER define?
19:00:47gevaertsYou could check CONFIG_CPU (should be PP5020 or PP5022 for the old ipods, except for the very old ones that are PP5002)
19:00:57mtWas there something changed/broken that affected the sim ? not all the buttons are working for me.
19:01:03gevaertsbut yes, a PP502x define could be useful
19:02:14mttomers: Nice work on RTL .. I want to test, but I'll have to get the buttons working back in my sim build first :/
19:08:36polobricoloTheSeven: the audio doesn't work for me
19:09:27TheSevenit's stuck at 0:00 and doesn't play anything?
19:10:04polobricoloi just built it 5 minutes ago using the lastest svn
19:10:42polobricolowith a m4a file it works
19:10:55polobricolobut it didn't with a mp3 file
19:11:35*TheSeven suspects a bad file
19:11:39TheSevencan you try another mp3?
19:12:12polobricoloTheSeven: it doesan't work neither
19:13:24 Quit DataGhost (Nick collision from services.)
19:13:31 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
19:13:33polobricolobut my m4a file only plays the beginning. rockbox detects it lenght to 1 mn and 18 s
19:13:42polobricolomy song i about 3 mn
19:15:19polobricolostrange if i start playing my m4a file then i go back to main menu (with the song still playing) and i then open another m4a file my ipod does the same than for mp3
19:15:40polobricolobut if when playing my file i press next, it works
19:16:44polobricoloTheSeven: even stranger if i start listening a m4a then i press next and i read a mp3 file it works .
19:17:32CIA-43New commit by tomers (r22961): Invert buttons in RTL mode
19:19:38amiconnpolobricolo, TheSeven: That sounds like a restart bug in the driver
19:19:45polobricololooks like the init of the music layer is buggy on the 2g
19:22:17 Join Zagor [0] (n=bjst@
19:29:19 Join notlistening [0] (
19:29:47 Join funman [0] (
19:29:58notlisteningdomonoky, what stage are you at now on the work? Did ou solve the connection problem?
19:30:53domonokynotlistening: i put a patch into the tracker. it can already generate voice/talk files with opensapi :-)
19:31:54domonokywhats missing: setting for the wavformat. better handling of the answer from the server (the problem with list answers), and improved error handling.
19:32:02 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:32:05 Quit JackWinter (Remote closed the connection)
19:32:06 Join matsl [0] (
19:32:12domonokyalso setPitch doesnt seem to have any effect on the server.
19:33:24 Join JackWinter [0] (
19:34:56notlisteningSet pitch is a strange one the way it works but it might be broken as i have implemented it two ways
19:35:41notlisteningwith my client it set the setting for each bit of speech generation but i guess you want to be able to set once and it be remember right?
19:41:58 Join Thundercloud [0] (
19:42:49domonokyi want to set all settings after the start of the the tts, and then just set the outpufile and voice things repetably
19:42:58*tomers Fixing my red. Found much more elegant solution
19:45:19rasherTheSeven: Welcome aboard!
19:45:27notlisteningokay right just fixed the timeout issue, you said about error handeling was that my end or yours?
19:46:41notlisteningand did you figure out the disconnection issue?
19:47:38domonokyyes, with error handling i meant rbutils side. the disconnection issue happens, when i send you something which doesnt end with "\n"
19:48:07 Nick fxb__ is now known as fxb (
19:49:04domonokyafter you recieved that string without "\n" the server closes the connection and says "Client closed connection".. :-) a little bit misleading
19:50:06polobricolohow do i get the value of PCLK on the nano2g ?
19:50:59polobricolocan i directly use cpu_frequency ?
19:51:41notlisteningah i an not sure why not adding the \n causes it to close the connection but the commincation is set by line to a \n is needed ;)
19:53:20 Join kachna [0] (
19:53:48notlisteningmaybe this is a CTLF and LF issue?
19:53:48CIA-43New commit by tomers (r22962): Fix red: Invert buttons in RTL mode ...
19:57:31domonokynotlistening: could be, as i am testing on windows at moment.
19:58:21notlisteningguess there might be a bit of "i don't care because tcl takes care of it going on" :D
19:58:24 Join daggett [0] (n=denis@
19:59:27notlisteningSo to summaries I need timeouts and a end of list status code correct? Would a description of the possible error codes be helpful and when they occur?
19:59:39 Join kugel [0] (n=kugel@rockbox/developer/kugel)
20:00:44domonokya fix for the timeout, setPitch and some sort of end-oflist marker (or just the list size at the beginning) is all i need, i think :-)
20:01:40domonokydescriptions of possible errors might be nice, but not really needed ( i will just gice the error text to the user if the error-code isnt the good one) :-)
20:02:18kugeltomers: Why/how are the buttons reverted in RTL mode?
20:02:47kugelso that you press LEFT to select an item?
20:03:11tomerskugel: Yes.
20:03:36domonokynotlistening: also in the protokoll description were a few errors (getEngine and setOut do return something)
20:04:55kugelI'm not exactly seeing why RTL means inverting the buttons also. does other software do that?
20:05:09pixelmakugel: I bet that feels more natural to people used to RTL reading
20:06:01kugelpressing left to skip to the next track just because text is right aligned doesn't sound very natural to me
20:06:09pixelmayou could try how it feels to you to use left to go into the directory and right for leaving etc,
20:06:37pixelmait's not only right aligned, you read from right to left...
20:07:02kugelI know...
20:07:07kugelbut that was before already IIRC
20:07:35mtHmm .. Clean sim build, buttons still don't work :/ .. Anyone having such issue ?
20:08:13kugelit just wasn't aligned to the right. I don't understand that changing the user interface which hasn't changed for years so extensively just because of the text alignment is needed
20:08:40kugelsome plugins will probably go crazy as well
20:08:47 Join Utchybann [0] (
20:09:07pixelmathe progress bar should probably drawn from right to left too I guess
20:09:24 Join faemir [0] (n=faemir@
20:09:57amiconnpeakmeter too?
20:10:06kugelI also think that buttons behavior depending on the current language can be really confusing
20:10:41funmankugel: try running some programs with LANG=il on unix to see how the interface changes
20:10:47pixelmait's probably more something like "it's done now because someone else looked into this" not "change the buttons because it's now right aligned"
20:11:00pixelmabut it is a problem for e.g. the manual too...
20:11:07amiconnkugel: Proper RTL support has been on the "wanted features" list for quite some time
20:11:23kugelamiconn: sure, but I wasn't aware that it means mirroring buttons
20:11:43CIA-43New commit by tomers (r22963): Revert renaming of the scrollbar global setting by r22945, following discussion in IRC
20:11:45pixelmaI mean how to describe the two different use cases now
20:12:11 Quit notlistening (Remote closed the connection)
20:12:19ZagorI agree with kugel. we should at least have clean, working text-only RTL before embarking on inverting the entire GUI
20:12:38amiconnI'm not sure what the UI conventions for RTL language are. But if mirrored button behaviour is part of that, it should be done imho
20:12:47 Join stoffel [0] (n=quassel@
20:12:50Zagorand perhaps get input from some RTL users...
20:13:11 Quit tomers (Connection reset by peer)
20:13:39 Join tomers [0] (
20:13:46gevaertsThere are actually even more issues. Which direction does the progress bar on the wps go?
20:14:14funmannormal reading direction
20:14:18*pixelma feels ignored
20:14:19*kugel also sighes about the inclusion of apps/ stuff into firmware/
20:14:52amiconnkugel: Where? I was checking for that, perhaps not closely enough
20:15:28pixelmagevaerts: amiconn already asked "and the peak meter?"
20:15:37gevaertsah yes, that too
20:15:49 Join dfkt_ [0] (
20:15:49*gevaerts apologises to all Arnolds
20:15:53pixelmaquoted freely
20:16:06kugelpixelma: that's a problem yes (just like with other configurable button solutions) but as the manual is english only I consider it not so critical
20:16:35 Join pamaury [0] (
20:16:47gevaertsI think it's essential that the manual gets this right, especially on targets where the buttons aren't fully symmetrical
20:16:48kugelI'm not entirely happy with the button reverting.
20:17:03funmanthere could be a note specifying which buttons/screens can change their order
20:17:29pixelmamaybe it can get a mention somewhere in the beginning (if it stays this way). But as I said I can imagine this to be more logical for native RTL language users
20:17:35tomersSorry all for not participating in the discussion above. I think changing buttons is good for list view. Not for WPS
20:18:02kugelPeople might actually have gotton used to the known interface. Now having their buttons dependant on their language selection without the possibility to go back to the known state
20:18:07gevaertsMaybe we should start with handling text properly, and then do some research (also known as a forum poll) to find out what actual users want
20:18:27pixelmasounds like a plan
20:18:30 Join freenodesucks [0] (
20:18:36Zagoryes, and work on a separate monster gui-invert patch outside of svn
20:18:39 Quit dfkt (Nick collision from services.)
20:18:42kugelI bet that even if it may be more naturual for lists, it will be unaturual for a lot dozens of other places and plugins
20:18:54gevaertsresearch could of course also include people living in rtl regions going out and asking their friends
20:19:28 Nick freenodesucks is now known as dfkt (
20:19:35ZagorI have a feeling hebrew users are pretty used to not getting all their GUIs reversed just because they get proper typesetting
20:20:03funmanwhy not just ask tomers ?
20:20:14gevaertsIt certainly can't be a simple reverse-buttons thing in the driver. I'd expect every single button assignment to have to be looked at individually
20:20:15pixelmaor arabic users
20:20:30tomersI'm here. WPS and progressbar need not change. Just tree view
20:20:45kugeltomers: your change means inverting it for everything I think (I haven't looked very detailed at it, but noticed that you changed it at the button driver level)
20:20:57pixelmafunman: didn't he state a comment by committing it?
20:21:03tomersmt get help us too, as he is the maintainer of the arabic language file
20:21:04Zagorfunman: heh, I didn't know tomers was israeli. apologies, tomers
20:21:45kugeldoom while using RTL language is probably funny (although you can change the button maping there), or rockblox
20:22:13*gevaerts votes to implement RTL in the LCD drivers :)
20:22:21tomersButtons are not inverted in plugins. IT IS CONTEXT SENSITIVE
20:22:22amiconnkugel: Check again - it's *not* switched for everything
20:22:26funmanwhen playing doom in RTL language, do you control the bots and not the player ?
20:22:44kugelok, sorry then, as I said I haven't looked very detailed at it
20:23:19 Quit JackWinter (Read error: 104 (Connection reset by peer))
20:23:46 Join freenodesucks [0] (
20:23:55gevaertstomers: is it not needed for wps and friends because nobody does that and you're used to it, or is LTR handling for next/previous really more natural?
20:24:06kugelah looks like r22962 changed it to be apps logic, I haven't looked at that at all
20:24:07 Join merbanan [0] (
20:24:16 Quit funman ("leaving")
20:24:21 Join JackWinter [0] (
20:24:31tomersgevaerts: It shouldn't change WPS afaik. I'll check it
20:24:58amiconnr22962 is just a better imlementation - it was implemented in a context sensitive way before
20:25:07 Quit dfkt (Nick collision from services.)
20:25:11 Nick freenodesucks is now known as dfkt (
20:25:25 Quit dfkt_ (Nick collision from services.)
20:25:27tomerskugel: I was in a bit of rush to fix things, so I just committed once in a while when I made progress...
20:26:04amiconnkugel: You didn't answer my question though - where does it include apps stuff in firmware/ ?
20:26:40amiconnThis would indeed be something that had to be changed asap
20:26:48tomerskugel, amiconn: Make sure you look at the latest version, as indeed I had stuff in drivers before, and fixed it
20:26:56pixelmatomers: gevaerts' question was about the why it's not needed in the WPS
20:26:57kugelI did, lcd-bitmap-common.c
20:27:25 Join Strife89 [0] (n=Strife89@
20:27:37kugelsee above.. (ok, I typo'd bitmap-common.c)
20:27:38tomerspixelma: As I understand what gevaerts mean, skipping tracks needs to use the same buttons both in RTL and LTR
20:28:15amiconntomers: You had stuff in the button driver, but that wasn't including apps stuff from firmware. Just the other way round, which is the correct way. The new solution is more elegant though
20:29:38pixelmatomers: his question was if it really feels more natural to you this way or if this is just because other devices don't do that
20:29:53TheSevenlinuxstb: around?
20:30:05TheSevenwhat about having a shot at RTC and battery ADC?
20:30:18 Quit JackWinter (Remote closed the connection)
20:30:20rashertomers: I'd expect if the progressbar was reversed as well, you'd want to skip track with the opposite buttons? Which would you prefer?
20:30:32polobricoloTheSeven: i managed to get the battery value
20:30:33markunanyone here who plays Nero encoded AAC files in rockbox?
20:30:41gevaertsand which would random people who are not used to computers prefer?
20:30:44polobricolousing your
20:30:51bertrikTheSeven, there is a basic ADC driver already, can you reuse that?
20:31:02tomerspixelma: When the main menu is displayed RTL, you expect to press LEFT, which is OUT of the tree, to enter deeper, and RIGHT, which is INTO the tree, to go back one level. That's all. It's unusable otherwise
20:31:11TheSevenpolobricolo: yep, I already have the battery reading code in the bootloader, but we need to fit that to the rockbox apis
20:31:26 Join JackWinter [0] (
20:31:29TheSevenbertrik: the battery ADC is on the PMU chip, not the SoC
20:31:32bertrikTheSeven, the only RTCs implemented so far for the s5l8700 targets were external ones (over I2C)
20:31:32gevaertstomers: ah, ok, so if the WPS worked RTL (progressbar and things), you'd also want the buttons to be reversed?
20:31:35tomersgevaerts: I believe it's unusable without swap left and right
20:31:40amiconnkugel: oh
20:31:49TheSevenbertrik: yep, the RTC is in the PMU, too
20:31:51amiconntomers: see kugel's link ^^
20:31:59bertrikTheSeven, ok
20:32:29tomersgevaerts: Yes, indeed. But since the WPS is still LTR oriented, we don't need to change buttons, therefore avoiding some usage mess
20:33:11kugelrtl could possibly be a viewport property, with viewports_set_defaults() set it depending on the rtl status. that would also allow for overriding it in some cases
20:33:28tomersgevaerts, kugel: The alternative is to change the API to pass RTL swap. Or is there something I didn't think of yet?
20:33:56*gevaerts stays out of implementation discussions. He isn't familiar with GUI-related code at all
20:34:00*mt thinks having an RTL WPS should be an option, not something that changes with lang-file selection.
20:34:30tomersmt: That's another feature. Lets leave it for later :-)
20:34:33mtAlthough there could be different default options for different lang files.
20:34:36linuxstbpolobricolo: Is "#ifdef CPU_PP" what you want?
20:34:41amiconntomers: Have a flag in the lcd driver that is set when loading or changing settings
20:34:55mttomers: Yeah, I'm just suggesting. :)
20:35:23linuxstbTheSeven: To answer your earlier question - yes, lcd_blit_yuv is simply a matter of copying another implementation. Although there does seem to be far too much of that going on...
20:35:30CIA-43New commit by bluebrother (r22964): Add some more options to rbutil deploy script. ...
20:35:44polobricololinuxstb: if there aren't any non-ipod portal player supported by rockbox then yes
20:36:06bluebrotherdomonoky: anything left that we could have in the rbutil deploy script apart from osx support?
20:36:13kugelpolobricolo: iap is only compiled for ipod so that shouldn't be an issue
20:36:18amiconnlinuxstb: I would like to see those unified, but they're all slightly different due to the different ways to acces the lcd controller etc
20:36:29 Quit JackWinter (Remote closed the connection)
20:36:39linuxstbamiconn: Yes, me too...
20:36:57domonokybluebrother: osx support :-)
20:37:19 Join JackWinter [0] (
20:37:42tomersamiconn, kugel: Should I have a flag in the lcd driver as you suggested, or having rtl as a viewport property, as kugel suggested? I am a bit confused now :-)
20:37:44bluebrotherdomonoky: send me a mac ... :P
20:38:02domonokybluebrother: i can give access to JdGordons mac :-)
20:38:33bluebrotheris python part of a default osx install btw?
20:38:41kugeltomers: it will be the same, just that you can override rtl per viewport and not per drawing call with the later
20:38:50domonokybut osx-support in the deploy scipt might get interesting, because we want it to be build not-statically :-)
20:38:58kugelI would prefer the viewport one
20:39:12kugelthe latter*
20:39:22bluebrotherdomonoky: you spotted that my last commit adds support for building dynamically?
20:39:34tomerskugel: This option sounds much better
20:39:40*domonoky needs to read more carefully :-)
20:39:43***Saving seen data "./dancer.seen"
20:39:52UtchybannTheSeven, linuxstb: I have just installed r22962 on my nano2g.
20:40:03kugeltomers: it may be a bit more work though :) not everything is viewport'ified
20:40:07bluebrotherall this framework and fat binary thing might be interesting though.
20:40:27domonokybluebrother: at least on jdgordons mac, python is already installed.
20:40:34tomerskugel: It's a good change to move towards a viewported world :-)
20:40:37UtchybannTheSeven: 'Rockbox info' only show 960MB, I have a 4GB nano.
20:41:01kugelbut lists generally are, you shoulds be fine with doing "vp->rtl = lang_is_rtl" at the beginning of list drawing (or even at init)
20:41:14gevaertsUtchybann: probably "just" cosmetic
20:41:24 Quit JackWinter (Remote closed the connection)
20:41:35TheSevenUtchybann: known bug
20:41:42TheSevenit's just calculating it wrong
20:41:44tomersCan you please tell me how puts_style_offset() should know the rtl mode?
20:41:51Utchybanngevaerts: probably, but read or write access seems to freeze...
20:42:00 Join JackWinter [0] (
20:42:23kugeltomers: actually, for future features, implementing it as a bit field may be even better (vp->flags |= VP_RTL_FLAG)
20:42:25TheSevenUtchybann: then it wouldn't boot up at all
20:42:43 Quit JackWinter (Remote closed the connection)
20:42:48 Join notlistening [0] (
20:42:57UtchybannTheSeven: yes, in fact I have freeze when I try to play a file.
20:43:00tomerskugel: I repeat my last question: how puts_style_offset() should know the rtl mode?
20:43:15TheSevenUtchybann: do you have any album art on it?
20:43:37TheSevenas soon as there are jpegs around, mine is a little shaky, without them it's rock solid :-)
20:43:46 Join JackWinter [0] (
20:43:47kugeltomers: if (current_vp->flags & VP_IS_RTL) xpos = current_vp->width - w - xpos;
20:44:00UtchybannTheSeven: you mean as cover.jpg in the directory or as id3tag ?
20:44:18 Quit coolrun34 ()
20:44:27TheSevenbut i'm having trouble with jpgs in the folder
20:45:25UtchybannTheSeven: in fact, I have quickly install rockbox and use file uploaded by gtkpod. I will upload some files in a more standard way.
20:46:21UtchybannTheSeven: I play them via Database not Files menu.
20:46:58tomerskugel: I missed that global variable
20:47:17kugelthat's a pointer of the currently used viewport
20:47:34tomersYeah, I meant to say that I overlooked it :-)
20:48:29TheSevenUtchybann: database didn't even work for me
20:49:00Utchybannlinuxstb: I have also see few strange pixels when running 'cube' demo. A blue pixel that should be black.
20:49:41UtchybannTheSeven: I can browse the database. It seems good for me. But When I try to play a file, rockbox freeze.
20:50:09UtchybannTheSeven: did you init the database ?
20:50:37TheSevenUtchybann: initing the database freezes for me
20:50:51TheSevenand concerning the cube: your lcd having a stuck pixel?
20:50:55TheSevenmine has several.
20:51:06 Join jferguson [0] (
20:51:13TheSevenin fact it's not the LCD itself, but GRAM, that has bad cells
20:53:06 Quit robin0800 (Remote closed the connection)
20:55:07UtchybannTheSeven: the pixel is not stuck. When switching to wireframe rendering, I don't get the blue pixel.
20:55:53UtchybannTheSeven: playing via 'Files' menu works nicely.
20:56:50TheSevenThen blame it on the database, not audio
20:57:07 Quit esperegu (Read error: 104 (Connection reset by peer))
20:57:23 Join esperegu [0] (n=quassel@
20:57:36 Quit notlistening (Remote closed the connection)
20:57:47UtchybannTheSeven: or because my other files are upload for itunesdb with strange filename and probably cover in id3 tag.
20:59:08polobricoloUtchubann: can you read mp3 files ?
20:59:08polobricolodoes anyone know what is Auto Flow Control (for the UART on the s5l8701)
20:59:51 Join esperegu_ [0] (n=quassel@
21:00:18 Join hd [0] (
21:00:28 Quit hd (Read error: 104 (Connection reset by peer))
21:01:03 Part jferguson ("ERC Version 5.2 (IRC client for Emacs)")
21:01:14Utchybannpolobricolo: yes
21:01:48polobricolostarnge, i can't. are you using the lastest svn ?
21:02:27UtchybannI'm running r22962.
21:03:58 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
21:06:39 Join hd [0] (
21:07:23 Join FlynDice [0] (
21:09:24 Quit bluebrother (Read error: 131 (Connection reset by peer))
21:09:32 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
21:09:48 Join esperegu__ [0] (n=quassel@
21:09:52 Quit esperegu_ ("No Ping reply in 90 seconds.")
21:12:27UtchybannTheSeven: I have tried to change the WPS while listening to music and something freeze but I still hear music.
21:13:17domonokypolobricolo: Auto Flow Control is a standard feature of uarts, it controls the sending speed with the RTS/CTS signals. dont know if that is needed for iap.
21:14:03 Quit esperegu (Read error: 110 (Connection timed out))
21:15:39 Quit HellDragon (Connection timed out)
21:17:24 Join _zic [0] (
21:17:53 Join barrywardell [0] (
21:22:26polobricolousing Utchybann's rockbox.bin my audio works !
21:23:10polobricoloi now have to test the iap
21:24:26polobricolooh i get a divide by zero
21:24:54*polobricolo realises he didn't enable the auto baud feature
21:25:06linuxstbTheSeven: Regarding the next things to do, I'm very busy with other things for the next few days, but plan to do work around the installation process, in order for us to add official Nano2G support to Rockbox. i.e. finish my ipodpatcher work, get a Rockbox bootloader working, and release new ipodpatcher binaries with that bootloader embedded. I'll probably also write a Rockbox plugin to do the bootloader encryption (so it's all self-su
21:25:13polobricoloi will check that tommorow
21:25:59TheSevenlinuxstb: I just implemented the ADC, an am trying to figure out the RTC apis
21:27:52polobricoloTheSeven: does the s5l8701 support autobaud for the UART. i can't find it in the datasheet
21:28:08TheSeveni don't think it does
21:28:14TheSevenisn't such a thing usually done in software?
21:28:15polobricoloor this there a way to disable it in rockbox
21:29:42polobricoloi just had a look at the PP driver. it sets the baudrate to 115384bps when rockbox asks for autobaud
21:29:58 Join krazykit [0] (
21:31:01linuxstbTheSeven: Do you mean the Rockbox APIs? Do you know how to read the actual RTC data? Is there one in the S5L that the Apple firmware is using?
21:31:29TheSevenlinuxstb: No, it's using the one in the PMU, and yes, I know how to get to the data
21:31:50TheSeveni'm currently just wondering on how to best organize the code, as there are 3 things sharing the same i2c stuff now
21:32:08linuxstbIs that a PCF, like earlier ipods?
21:32:16TheSevena different one, I think
21:32:22TheSeventhis demands for another drivers/pcfXXXXX.c, I think, but I don't know the exact model number right now
21:33:36*kugel bought that board
21:34:09kugelthe mini2240 one
21:34:17linuxstbTheSeven: So it has a different register layout to the pcf50605/6 ?
21:34:26TheSevenit has a different address, at least
21:34:30TheSevenneed to check the regs
21:35:05polobricoloThe ipod doesn't even pay music through the dock.
21:35:32polobricoloit ooks like you need to activate it manually
21:36:06TheSevenno, reg layout is completely different, too
21:36:10kugeldomonoky: did your board arrive already?
21:37:16kugelI got a friend of mine one also. he's a passionate network coder, maybe I can get him to implement a lan stack for rockbox :p
21:38:06TheSevenkugel: which board?
21:38:28kugela china one with a s3c2440 soc
21:38:56*TheSeven is wondering if he should go for a friendlyarm board
21:39:05 Quit DataGhost (Nick collision from services.)
21:39:13 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
21:39:15 Quit matsl (Remote closed the connection)
21:39:18polobricolothe iap doesn't work :(
21:39:19kugelTheSeven: if you do, consider this one :p
21:39:34polobricoloprobably a problem with the uart autobaud
21:40:07linuxstbkugel: Hey, don't distract him! ;)
21:40:20TheSevenpolobricolo: why should a remote need autobaud?
21:40:45domonokykugel: no my board is still delayed in customs somewhere in china :-)
21:41:16polobricoloi don't know. but rockbox initialize the baudrate to 0 (meaning autobaud mode)
21:41:31*polobricolo wonders if it is possible to diable that
21:41:50domonokykugel: network should be easy with the lwip stack (i already used that somewhere else) so we only need the nic driver and ofcourse apps using it :-)
21:42:14*TheSeven eyed at which is in fact the same board
21:42:31polobricoloTheSeven: and all the docks don't use the same baudrate
21:45:28TheSevenlinuxstb: should I go for a pmu-nano2g.c and pmu-target.h in firmware/target/arm/s5l8700/ipodnano2g?
21:45:28 Quit amiconn (Remote closed the connection)
21:45:28 Quit pixelma (Remote closed the connection)
21:45:29kugelTheSeven: indeed, it's equal except for the price :p
21:45:55TheSevenkugel: but would ship from germany (where i am)
21:46:04TheSeven(at least if it was actually in stock, which it isn't)
21:46:06kugeloh, I am too
21:46:32*kugel values saving 50% over the shipping origin
21:47:07TheSevenit isn't really 50%, especially with additional taxes/customs/whatever may come up when buying it in the US or china
21:47:19 Join amiconn [0] (i=quassel@rockbox/developer/amiconn)
21:47:28 Join pixelma [0] (i=quassel@rockbox/staff/pixelma)
21:47:30domonokykugel: its less then 50% if you add up shipping and taxes. but still cheaper (and actually in stock) :-)
21:47:43TheSevenin stock, but stuck :-P
21:47:54bluebrotherso it's in stuck :)
21:48:12domonokytrue :-)
21:48:15kugeldomonoky: did you pick ems express shipping? they promise 3-7 days shipping time
21:49:00domonokykugel: no i choose dhl (more expensive, but more convienient for me (using Packetstation) )
21:49:25 Quit petur (Remote closed the connection)
21:49:28UtchybannTheSeven: playback works fine even with cover.jpg present in folder.
21:50:00TheSevenUtchybann: I only had glitches when I tried to change the volume while albumart was still loading and such stuff
21:50:52UtchybannTheSeven: ok. I have problem when trying to erase dynamic playlist. => need to reboot.
21:55:08 Part Grahack
21:55:26linuxstbTheSeven: Seems reasonable - I'm assuming it's going be quite specific to the Nano2G anyway...
21:58:02 Quit Strife89 ("Temporary quit.")
21:58:52polobricoloTheSeven: PCLK is 50mhz right ?
21:59:19CIA-43New commit by tomers (r22965): - Use pointers instead of repeating access to array element ...
22:00:23TheSevenpolobricolo: I suspect it's 50MHz in iLoader and 100MHz in rockbox
22:04:00*polobricolo is going to write a UART logger tomorrow
22:04:10polobricolobut i have to go now
22:05:19kugeltomers: I fear a red :)
22:05:56CIA-43New commit by tomers (r22966): Fix red - type
22:06:13*tomers had a typo typing typo
22:06:46domonokyfail ! :-)
22:08:56 Quit merbanan (Read error: 145 (Connection timed out))
22:11:17 Join Stephen_ [0] (n=S@
22:11:29 Quit stoffel (Read error: 113 (No route to host))
22:12:04 Quit _zic ("Ухожу")
22:12:48 Join Strife89 [0] (n=Strife89@
22:14:09 Quit tomers (Read error: 54 (Connection reset by peer))
22:14:38 Join tomers [0] (
22:14:52 Part JackWinter ("Konversation terminated!")
22:16:55 Join stripwax [0] (
22:21:35CIA-43New commit by theseven (r22967): Reworked iPod Nano 2G PMU code, added RTC and battery ADC.
22:21:57 Quit gevaerts (Nick collision from services.)
22:22:04CIA-43New commit by tomers (r22968): Implement RTL as a viewport's bit-field
22:22:05 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
22:25:03 Quit bluebrother (Read error: 104 (Connection reset by peer))
22:25:08 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
22:31:26polobricolodoes the PANIC power off on the nano2g cause a clean unmount of the FTL
22:32:46TheSevendepends on which one you get
22:33:04TheSeven"Unmounting flash failed" or "Poweroff not implemented yet"
22:34:36 Quit bluebrother (Nick collision from services.)
22:34:41 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
22:35:17*TheSeven thinks this was the last commit from him today
22:36:50linuxstbTheSeven: So is the time displaying in Rockbox now?
22:37:01CIA-43New commit by tomers (r22969): Fix yellow
22:37:03TheSevensetting it also works properly
22:37:34linuxstbDid you test that it is the same as when in the OF?
22:37:49 Join tomers_ [0] (
22:37:51TheSevennot with rockbox, but with earlier testing code
22:37:57TheSevenshould be ok.
22:38:18TheSevenbtw, what is the commits-per-day record? We've been pretty busy today.
22:39:14kugelyea, today was extra-ordinary
22:39:22linuxstbTheSeven: Hmm, rtc-nano2g.c has CR/LF line-endings...
22:39:45***Saving seen data "./dancer.seen"
22:40:05linuxstbAnd pmu-target.h is iso-8859-1
22:40:30TheSevenfunny, as I copied both of them from other files and then ditched the contents (but not the header)
22:40:42 Quit barrywardell ()
22:41:13*linuxstb wonders why "file" thinks rtc-nano2g.c is Pascal
22:41:26 Quit bmbl (Read error: 145 (Connection timed out))
22:42:03tomers_kugel: Have you seen my latest commits? Is it ok?
22:42:24 Join drf|laptop [0] (
22:42:25kugelnot sure
22:42:32TheSevenwhat's the default encoding for rockbox? UTF8 with or without BOM?
22:42:33drf|laptophey guys
22:42:36drf|laptopSo I'm just curious
22:42:40linuxstbTheSeven: utf-8 without
22:42:43drf|laptopany chance of getting Rockbox for the Zune?
22:42:44*drf|laptop disappears
22:43:06linuxstbdrf|laptop: Just follow this simple guide -
22:43:30kugeltomers_: I'm thinking you don't need to set it at every drawing
22:43:39drf|laptopBut seriously. Is it because Microsoft isn't open-source or just because I'm one of only few people who likes Zunes?
22:43:48kugelthe RTL should be correctly set for the parent coming in
22:44:10markundrf|laptop: the zune is similar to the gigabeat S, but they fixed a security problem which we use for getting rockbox to run.
22:44:11linuxstbdrf|laptop: It's because there are security protections to prevent people running their own code, and no-one has found a way to bypass them yet.
22:44:32drf|laptopwhat is the gigabeat S?
22:44:35drf|laptopnever even heard of it
22:44:45kugelI thought the only needed thing would be to add rtl flagging to viewport_set_defaults(), and ensure that viewportmanager_theme_changed() is called during font loading
22:44:56drf|laptopisn't there a Zune SDK? or is that only for the software aspect
22:45:05linuxstbdfkt: Basically an early version of the Zune, sold by Toshiba.
22:45:17 Join barrywardell [0] (
22:45:19drf|laptopI've never even heard of that
22:45:33linuxstbdrf|laptop: But we're getting off-topic for #rockbox
22:45:38 Quit tomers (Read error: 113 (No route to host))
22:45:51DerPapston my ipod sim (5G) lists can't be navigated using the scrollwhel anymore. is that related to the ltr changes?
22:45:55drf|laptopah, well, I just wanted to know why Zune wasn't supported
22:46:46bluebrotherthere's another reason why: because nobody did the work.
22:47:01 Nick fxb is now known as fxb__ (
22:47:26drf|laptopI figured as much :/
22:47:34drf|laptopIf I was able to help I would, but I'm not about to tear my Zune apart
22:47:45drf|laptopnor am I any good at writing code...
22:49:55*TheSeven will commit encoding and line break fixes, upload another build to clustur, and then go to bed
22:50:20tomers_kugel: Sounds much better. I'll try
22:50:37DerPapsttomers_: saw my message?
22:52:15tomers_DerPapst: Yes. Can you expand on the issue? Maybe it's related, and maybe not... You can try different builds and pinpoint it, but today there were tons on commits. Are you a developer? Can you dive into the code and try to locate the problem?
22:53:56DerPapstwell, i'm not too familiar with coding for rockbox (mostly did plugin stuff). i can check which revision broke it for me though.
22:54:29 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
22:54:42CIA-43New commit by theseven (r22970): Encoding fixes
22:56:00kugeltomers_: drawing shouldn't fix the viewports more than needed (i.e. not more than setting the rectangles up), proper rtl fixing can be done beforehand and much less often
22:56:08*DerPapst wonders why logbot prepends a 03 before the committer's name...
22:56:37kugelthat's because your IRC client is crap
22:56:48kugeland CIA-43 != logbot
22:56:57mtProbably color code
22:57:35bluebrotherDerPapst: that's a color code, not a simple 03
22:57:40 Quit barrywardell ()
22:58:11Horschtso, I have a windows batch script that uses various programs to sync my ipod/rockbox player to my local media library. It's not yet universal, would the "third party" section of the forums be a good place to ask for input?
22:59:06 Join barrywardell [0] (
22:59:18DerPapstkugel: my client isn't crap because it shows it correctly and i know that logbot istn CIA...
22:59:48linuxstbtomers_: Was your patch at FS #10650 what you committed as r22945 today?
23:00:10DerPapstbluebrother, mt: thanks. i thought colour is only the first control char. ;)
23:00:26 Part drf|laptop
23:01:03tomers_linuxstb: Yes, but the code was changes since
23:01:28tomers_Actually, I don't remember if it is FS #10650 == r22945 exactly
23:02:42linuxstbMany reasons... 1) The commit message didn't refer to a flyspray task; 2) The flyspray task is still open; 3) The flyspray task said "At the moment, this patch breaks some less common screens. This is a preliminary work. It still needs lots of development.", which doesn't sound like something that IMO should have been committed...
23:05:05*LambdaCalculus37 knows he missed a massive load of fun on the nano2g front...
23:06:05TheSeventime to try the latest build, then.
23:06:29tomers_linuxstb: 1. You are correct. 2. I will close it. 3. Screens are broken only for RTL langs, and I discussed it on IRC. People thought I should commit.
23:06:59LambdaCalculus37TheSeven: I haven't had a single chance to do a single thing Rockbox-related now.
23:07:06LambdaCalculus37And I miss it!
23:07:47LambdaCalculus37TheSeven: So what's going on with the nano2g? And congrats on commit access! :)
23:09:19TheSevenR/W NAND and FTL, Audio, RTC, battery ADC and whatever I overlooked (some bugfixes) done, besides some instability issues. (and I won't trust the RW FTL yet)
23:12:10 Join domonoky1 [0] (
23:12:31 Quit TheSeven ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
23:13:58CIA-43New commit by tomers (r22971): Add rtl flagging to viewport_set_defaults(), and ensure that ...
23:13:59 Join TheSeven [0] (
23:15:46LambdaCalculus37TheSeven: Are there any instructions to load a test build on the nano2g?
23:15:48 Quit esperegu__ (Read error: 113 (No route to host))
23:16:02LambdaCalculus37I want to at least help test out the port; it's the best I can offer right now.
23:16:10mtDerPapst: It seems there's something that broke up/down buttons in e200 sim build too (haven't tried on target yet) .. have you found a problematic revision ?
23:16:14TheSevenit's basically installing iLoader and unpacking the rockbox zip to the drive
23:16:32LambdaCalculus37TheSeven: Got a link to iLoader on the wiki?
23:16:43TheSevenlinuxstb: is ipodpatcher already usable?
23:17:00TheSevenif not:
23:17:02linuxstbNot for dual-booting, but the "replace the OF with iloader" feature works fine.
23:17:08DerPapstmt: not yet. cygwin is tooo slow.
23:17:37*DerPapst needs to config his linux vm to compile sims
23:18:25DerPapsti'm checking if r22960 still works. that is one before the inverted button commit
23:19:20 Quit Zagor ("Clint excited")
23:19:47 Quit Strife89 ("If you hold a Linux shell to your ear, you can hear the C.")
23:19:58LambdaCalculus37TheSeven: Thanks, I'll give it a try soon.
23:20:31pixelmadidn't kugel say it was related to some hack for sims being removed earlier today?
23:21:18pixelmathat the scrollwheel in sims doesn't work anymore I mean
23:21:50DerPapsthe mentioned some hack.. but i haven't checked what it was good for.
23:22:00DerPapstthe scrollwheel in the wps still works though
23:22:11DerPapstit's just broken in the lists
23:22:49mtpixelma: Didn't know/read that. :(
23:23:57mtI re-built, re-pulled a clean copy, and re-built once more trying to make sure it wasn't a problem in my local tree !
23:24:19pixelmaI'm not sure what and which commit it related to exactly and hoped to get his attention
23:25:43mtI guess I'll revert to the latest revision as of yesterday and rebuild, again. :/
23:29:29 Quit domonoky (Read error: 110 (Connection timed out))
23:29:37*TheSeven will go to bed now
23:30:08LambdaCalculus37Good night!
23:30:32DerPapstnighty night TheSeven
23:30:41CIA-43New commit by tomers (r22972): Updated Hebrew language (fixed position of question marks)
23:32:12 Quit ender` (" MCSE = Moron Claiming System Expertise, Must Call Someone Else, Minesweeper Champion, Solitaire Exp")
23:32:30CIA-43New commit by tomers (r22973): Fix red: Protect viewport RTL flag manipulation with #ifdef HAVE_LCD_BITMAP
23:33:21DerPapstmt, tomers_: scrolling in list is already broken in r22960.
23:34:36tomers_DerPapst: ipod5g simulator?
23:35:07DerPapste200 sim seems to be broken too
23:35:35*tomers_ Checking scrolling in e200 sim
23:35:44DerPapsti'd guess all the other ipod sims too.
23:36:13CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
23:36:17tomers_DerPapst: Scrolling in LTR (english)?
23:38:14UtchybannTheSeven: rtc seems fine.
23:38:54 Quit Llorean ("Leaving.")
23:40:31 Quit liar|netbook (Read error: 148 (No route to host))
23:42:46markunsaratoga, GodEater: I think you guys are overreacting a bit here:
23:43:34markunor should I post that in #rockbox?
23:43:54markunah way, this *is* #rockbox :)
23:44:02DerPapsthehe :D
23:44:18markunway -> wait
23:45:02gevaertsmarkun: I agree
23:45:46gevaertsrestoring a device to a state the service people are likely to be familiar with is not a bad thing to do
23:46:55CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
23:47:47markungevaerts: even if installing rockbox means that technically you have no warranty anymore, I would want to send it back if for example the buttons would fall of
23:47:54gevaertsof course
23:48:17 Join liar|netbook [0] (n=liar@
23:48:52*gevaerts also strongly doubts that rockbox would mean that technically you have no warranty anymore in such cases
23:49:34mtkugel: Is there something I should do to fix the scroll wheel in the sim ?(e200 sim)
23:49:38DerPapstmt, tomers_: r22944 already seems to be broken. so it's not related to the LTR committs
23:50:05evilnickThere was some noise about this with the iRiver port. IIRC, iRiver were fine repairing hardware issues on Rockboxed players for a time but then changed their tune.
23:50:13kugelDerPapst: as pixelma said, kkurbjun broke that
23:50:39n1sgevaerts: what if you played too much doom so you broke the buttons? :)
23:50:56 Join safetydan [0] (
23:51:13gevaertsn1s: broken buttons are likely to be an edge case anyway :) We're speaking about clear manufacturing defects here
23:51:16DerPapstkugel: thanls
23:51:39*DerPapst reverts r22936 then
23:51:58 Join Llorean [0] (
23:52:25gevaertsof course the OP asks about making sure that no traces remain. That's going a bit further than just putting the OF back
23:52:31 Quit bluebrother ("leaving")
23:55:21*gevaerts thinks that it's also extremely hard to remove all traces of something if an sd card and its manufacturer are involved
23:57:05markun"our sniffer dogs found traces of rockbox, so we're not going to replace the buttons"
23:57:52gevaerts"I was framed!"
23:58:38gevaertscould happen on the mr500...
23:58:49 Quit TopyMobile (Read error: 110 (Connection timed out))

Previous day | Next day