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 2009-07-18

00:00:38shotofaddsgevaerts: i'm not sure that explains it. the call to card_detect() is only wrapped in HAVE_HOTSWAP, and card_detect() is always #defined to card_detect_target()
00:00:52shotofaddsI don't see where multi-driver comes into it
00:01:05gevaertshm, you're right, it shouldn't matter
00:01:46shotofaddsalso when building a sim gcc warns about the apostrophies in the mv.h #error lines
00:02:14shotofaddsbut that one's easier to solve ;)
00:02:25*linuxstb thinks if the Elio gets supported by Rockbox, it will stand head and shoulders above all others in terms of ugliness
00:02:43 Quit CaptainKwel ("Page closed")
00:02:52gevaertsshotofadds: indeed. My local copy doesn't have #error lines in mv.h anymore :)
00:02:56scorchelinuxstb: is it a good brick though?
00:03:16linuxstbscorche: Yes, I think so. So it's has a use.
00:03:21 Quit tarbo (Connection timed out)
00:04:08 Quit Zarggg_ (Read error: 104 (Connection reset by peer))
00:04:21 Join Zarggg [0] (
00:04:24 Quit kugel (Remote closed the connection)
00:04:27*pixelma finds more than one model of Tatung Elios through google and wonders which one it is
00:05:02 Quit Torne (Read error: 60 (Operation timed out))
00:05:05 Join Torne [0] (
00:05:23LambdaCalculus37amiconn: Still a strange device for boasting a hard drive and an SD slot.
00:05:52amiconnIt looks rather similar to an H300 apart from the buttons
00:06:23pixelmabutton pad looks interesting too... 9 buttoms
00:06:27pixelmabuttons too
00:06:37amiconnIt also seems to have USB host. That's make triple storage...
00:07:10CIA-69New commit by roolku (r21932): make ELIO_TPJ1022 bootloader compile again
00:07:29pixelmaor is the center one not a button?
00:08:10*gevaerts starts to collect clues
00:08:40LambdaCalculus37We need more devices like the TPJ1022 to give us a challenge. :)
00:09:27 Quit LambdaCalculus37 ("Fwump")
00:09:29shotofaddsgevaerts: that hack in root_menu.c is blatantly wrong anyway with multi-driver, but that's not the point...
00:09:36pixelmaLambdaCalculus37: or the DellDJ?...
00:09:43pixelmagah he left
00:10:09roolkupixelma: the centre is just a button light (multi-colour though)
00:10:48pixelmahmm, misleading. But it seems to have side buttons?
00:11:31amiconnHmm, looks like the Elio uses an 1.8" drive with side connector (2.5" hdd type)?
00:11:36roolkupixelma: vol+/- on the right, record/ab-repeat hold on the left
00:12:00JdGordon|shotofadds: i cant tlook at the code right now... what is root_menu.c doing there?
00:12:10roolkuamiconn: yes unfortunately. a 2.5" hd that i tried drew to much current
00:12:31amiconnHmm, but that should allow for pretty easy CF modding
00:12:33pixelmaroolku: sounds like an "interesting" keypad... at least there are already 8 buttons on the front
00:12:47shotofaddsJdGordon|: "quick hack to stop crashing if you try entering the browser from the menu when you were in the card and it was removed"
00:12:58amiconn10 buttons, and a multi-colour light
00:13:12*amiconn really needs to get that multicolour indicator on the M3 going
00:13:16JdGordon|shotofadds: ah ok.. so a legit bandaid :)
00:13:23JdGordon|that very uch sounds like my text :)
00:14:01roolkupixelma: I think the unit is quite interesting as a device, but very ugly and awkward :)
00:14:10shotofaddsJdGordon|: yeah, it's sane atm. but it checks for pathnames beginning '<' and assume they're all removable
00:14:35linuxstbroolku: Have you investigated hard drive replacements for it? i.e. how big do they go?
00:15:14amiconnI think CF is the better way here. There won't be that many drives in that form factor.
00:15:38roolkulinuxstb: when I got the unit I tried to get a replacement drive, but they were incredibly expensive on ebay - I believe laptops use them
00:15:41amiconnWhat's interesting though is that the 20GB model is already 7mm thick (the 40GB model is .95mm)
00:15:45pixelma10 buttons, I thought it was 12 (well 12 functions, maybe 10 physical buttons, just like Volume Up/Down looks like one on the c200 too)
00:15:49amiconnErr, 9.5mm
00:21:13B4gderZagor: the score added per round should probably somehow be adjusted by the number of clients participating
00:21:22*gevaerts now has a three-drive sansa :)
00:21:27B4gderlike perhaps "percentage of total score this client worked on"
00:21:41Zagorpercent is good
00:22:08B4gderhm, but on second thoughts that too will vary by the amount of participators
00:22:31shotofaddsgevaerts: does it build in the sim? ;)
00:22:33*JdGordon| one ups gevaerts and sets up 2 ramdisks on his e200...
00:22:35Zagorheh, right. you want me to think too?
00:22:41gevaertsnothing like spending hours to see why this drive doesn't work and then finding that it's just because it was formatted as FAT12
00:22:46ZagorI'm just here to agree :)
00:22:47JdGordon|and partitions the microSD to have 4
00:22:55shotofaddsha! I had that problem with the first SD card I tried
00:23:48 Quit aaron424 ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060308]")
00:24:15gevaertsshotofadds: so what now? Should I live dangerously and commit, or wait a bit? :)
00:24:17linuxstbAnyone know if any existing targets use i2c for backlight control? (the Nano 2G does...)
00:26:14ZagorB4gder: I think the rolling average will fix the score issue good enough
00:26:24B4gderyes, I think you're right
00:26:25shotofaddsgevaerts: you know I'd rather see this committed sooner rather than later. once thing to maybe do first is make the existing hotswap drivers take note of the first_drive parameter (although this won't make a difference in practise)
00:26:48gevaertsshotofadds: indeed it won't. Let's fix that later
00:27:12shotofaddslinuxstb: D2 uses I2C for backlight brightness and power (it's controlled by a PCF50606 PWM output)
00:27:35shotofaddsgevaerts: go for it then!
00:28:20shotofaddslinuxstb: although it does also have a GPIO toggle for quick on/off
00:28:45linuxstbshotofadds: Ah, OK. I was just curious...
00:28:52 Part toffe82
00:28:52CIA-69New commit by gevaerts (r21933): Commit FS #9545, storage cleanup and multi-driver support
00:29:36*gevaerts gets ready to count reds
00:31:57gevaertsit takes reasonably long, so it can't be horrible
00:32:31B4gderwe need to get more toplists running, for users and clients etc
00:32:59*amiconn would appreaciate a progress indicator/ auto-reloader
00:33:42B4gder4 3 ... 2...
00:33:56B4gderBuild round completed after 294 seconds.
00:34:46gevaerts100. *very* reasonable!
00:36:23JdGordon|you should be building before commit!
00:36:26JdGordon|naughty naughty!
00:37:04*gevaerts is sure that JdGordon| has only ever had green builds :)
00:37:41JdGordon|any chance that the build server can be hoocked up to logbot so we can get a message in here with the score when its finished?
00:38:09ZagorJdGordon|: actually the cia bot is easier to talk to. and yes, that is planned.
00:38:24shotofaddsgevaerts: and a negative delta too, well done :)
00:39:06gevaertsshotofadds: that could be by accident :)
00:39:44CIA-69New commit by gevaerts (r21934): don't try to call nand_soft_reset() or nand_disk_is_active(), they don't exist
00:40:20gevaertsOK. Now the tree is ready for those storage stub cleanups :)
00:40:27shotofaddsgevaerts: cunning fix :/
00:40:40CIA-69New commit by dave (r21935): Basic backlight driver for Nano 2G
00:40:51shotofaddsmore to the point, the tree is ready for the D2 SD driver
00:41:11CIA-69New commit by gevaerts (r21936): svn properties, late as usual
00:41:23 Quit dmb (Read error: 60 (Operation timed out))
00:42:00 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
00:44:16 Join dmb [0] (n=Dmb@unaffiliated/dmb)
00:44:17JdGordon|Zagor: I dont know if its valuable or not.. but could you add the heigh and low scores for each client also?
00:44:36Zagorhigh and low?
00:44:46JdGordon|round score
00:44:53JdGordon|although its probably meaningless...
00:45:10Zagoryou mean the highest and lowest build they did this round?
00:45:26gevaertsZagor: does it skip updating the web page if build rounds are close to each other?
00:45:37JdGordon|no, the clients best and wrost scores in the average window..
00:45:47Zagorgevaerts: no, it's new a bug I just fixed...
00:45:58gevaertsah, a *new* one! :)
00:46:08Zagorthe best kind
00:46:18B4gdernew bugs are much better than old ones
00:46:28*gevaerts fully agrees!
00:46:46ZagorJdGordon|: ah. well with a simple -old/6 + new/6 scheme we don't save the individual scorse
00:47:28Zagorthe rolling average is not yet implemented though
00:48:25Zagorit's currently a total average of all builds
00:48:27gevaertsWhy is the delta table ordering so "interesting"?
00:48:52B4gderit is completely separate
00:49:11 Quit jordan` (Read error: 110 (Connection timed out))
00:49:13B4gderand builds who are not listed in the script will end up leftmost
00:49:26Zagorgevaerts: c3 beat monster with 3 points! :)
00:49:59gevaertsZagor: I told you it would be faster!
00:50:31ZagorB4gder: was there a reason for this special sorting?
00:50:42 Join jordan` [0] (i=gromit@
00:50:56B4gderwell, just as the script is separate it has no real idea of the "other" sorting
00:51:07B4gderso I just made it possible to custom the order
00:51:35B4gderthe original sorting is based on the name, which the second script doesn't know
00:51:39*shotofadds will wait til sunday to commit the D2 stuff. there's too much accompanying stuff to write (installation changes, wiki notes etc) plus I won't be here tomorrow to cope with the fallout when it all goes titsup...
00:52:15 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
00:52:18 Join Galois [0] (
00:52:24ZagorB4gder: the green table is sorted by build name nowadays. I'll sort the size table the same way.
00:52:41 Quit scorche (" rawr...that is all...rawr")
00:53:06B4gderah right, that should work
00:53:39amiconnZagor: Still disconnects...
00:53:52amiconn(at end of round)
00:53:52Zagoramiconn: that was me restarting the server manually
00:53:54saratogado we use make -j or do builds in parallel now?
00:54:03Zagorsaratoga: make -j
00:55:06 Quit w1ll14m ("CGI:IRC")
00:55:08ZagorI thought about doing the bootloaders single-core but that would just mean giving the slow clients even less to work with (since fast newcomers would gobble up bootloaders even quicker)
00:56:00Zagori.e. the system supports both methods. it's just that we currently specify all builds as multicore builds.
00:56:09gevaertsIt seems the file that rasher's binsize scripts picked up is gone as well
00:56:47gevaertswhich is why hex-gevaerts was slow these last few rounds
00:57:22 Quit B4gder ("It is time to say moo")
00:57:58Zagorgevaerts: aha, didn't know about that. sizes can be fetched from /data/$rev.sizes for now, and from the database
00:58:49gevaertsrasher: ping ^^
00:59:24***Saving seen data "./dancer.seen"
00:59:44 Join wincent [0] (
00:59:46shotofaddsgevaerts: any idea how to get around this multi-driver sim problem? should card_detect() even be available in the sim? or should it be stubbed somewhere?
01:00:31*gevaerts looks
01:01:11shotofaddsgevaerts: I think you only got a low score because there are no Onda Sim builds in the table...
01:01:41gevaertsthat's not my fault! ;)
01:02:18 Join w1ll14m [0] (
01:02:24wincentEvil memory allocator bug got hunted away!
01:03:03JdGordon|shotofadds: it would be nice if card_detect() and removable storage in geneeral could be made to work in the sim..
01:03:11JdGordon|stub it is probably the easiest
01:03:57w1ll14mwould it be possible to edit config file for gigebeats, so ab repeat is enabled ?
01:04:40gevaertsshotofadds: at least e200.h defines HAVE_HOTSWAP in #ifndef SIMULATOR
01:05:02shotofaddswell that's one solution...
01:05:48 Join CaptainKwel [0] (
01:06:00linuxstbroolku: Have you tried USB on the elio?
01:06:13 Quit robin0800 ("Leaving")
01:06:58roolkulinuxstb: I can't remember to be honest - I got the unit a while ago and just dug it out again
01:08:38 Quit shotofadds ("Leaving")
01:09:47linuxstbgevaerts: The Nano2G doesn't build any more...
01:09:56linuxstb(and I'm guessing the Meizu might not either)
01:10:11linuxstbfirmware/target/arm/s5l8700/ata-nand-s5l8700.c:32: error: ‘nand_spinup_time’ redeclared as different kind of symbol
01:12:01CIA-69New commit by gevaerts (r21937): remove obsolete nand_spinup_time() declaration
01:12:33linuxstbgevaerts: Thanks!
01:12:57gevaertsthis is a bit that does need cleanup. At least now it's possible to get started :)
01:15:39 Quit PaulJam (Nick collision from services.)
01:15:46 Join PaulJam_ [0] (
01:17:39 Join mt [0] (n=mt@
01:19:26CIA-69New commit by gevaerts (r21938): fix ata-nand-s5l8700 compilation properly this time
01:21:26 Quit petur ("Zzzz")
01:23:19linuxstbgevaerts: Can you make it work next? ;)
01:23:36gevaertslinuxstb: I thought that was your job!
01:24:02 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
01:24:17 Quit PaulJam_ (".")
01:25:35 Join stripwax [0] (
01:25:52 Quit Ubuntuxer ("Leaving.")
01:26:16 Join Strath [0] (
01:26:49 Quit DarkDefender ("Leaving")
01:28:50gevaertsrasher: did you see Zagor's line about /data/$rev.sizes?
01:34:54 Quit Zagor ("Leaving")
01:36:02 Join JdGordon| [0] (i=ad752425@gateway/web/freenode/x-8b17dc484c09461f)
01:36:22JdGordon|anyone know anything about bufgetdata()?
01:36:28JdGordon|specifically.. how safe it is to use?
01:37:51 Quit m67_l3 (Client Quit)
01:37:52GreatBeaverhow come h120 sounds so good?
01:38:03GreatBeaverit sounds almost as good as my $200 sound card
01:38:08 Quit stripwax ("")
01:38:09GreatBeavermaybe 75%
01:38:26GreatBeaverwrong forum
01:39:29 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
01:40:11 Quit martian67 (SendQ exceeded)
01:41:26 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
01:42:16 Quit martian67 (SendQ exceeded)
01:43:26 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
01:43:50JdGordon|... I know its possible for the handle to move in the buffer... but does tht actually happen? or was it changed so they dont move?
01:43:57 Join stripwax [0] (
01:44:13JdGordon|I want to grab the handle pointer and keep it for the songs duration
01:44:20 Quit martian67 (SendQ exceeded)
01:44:36amiconngevaerts: That is a huge delta for the Ondas....
01:45:22 Quit stripwax (Client Quit)
01:45:26gevaertsamiconn: it is, yes. It may be because some functions were just never linked in up to now
01:46:11gevaertsthe commit actually activated the NAND driver (even though it's not functional yet)
01:46:16Unhelpfulwhat does bloat-o-meter say?
01:47:44 Join martian67 [0] (
01:48:08 Quit tchan ("WeeChat 0.3.0-rc1")
01:48:21 Quit martian67 (SendQ exceeded)
01:48:38JdGordon|hmm...maybe there is no point trying to do what I want.... if the hwcodecs cant grab a buffer like swcodec can then the option will have to stay anyway
01:50:23 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
01:51:00 Quit roolku ()
01:51:07saratogalets move around the mdct library a bit and see what happens
01:51:35amiconngevaerts: Oh btw, USB mode on the beast works on my hub when the charger is connected. That supports the overcurrent theory, right?
01:52:59ej0rgeovercurrent theory?
01:53:29 Quit linuxguy3 (Read error: 104 (Connection reset by peer))
01:53:40 Join linuxguy3 [0] (
01:53:46CIA-69New commit by saratoga (r21939): Rearrange the MDCT library lookup tables so that codecs can access them. Access them in cook. Remove old cook trig tables. Replace old 32x16 bit ...
01:56:47gevaertsamiconn: it does, yes
01:57:11amiconnHmm, I wonder whether it is a good idea that new clients have an initial score of zero. New fast clients will grab all the easy builds and leave nothing for the actually slow clients
01:57:42*amiconn observes this in his other vm, which is now building for the first time
01:58:12gevaertsthat would only be a problem if we have new fast clients all the time. I don't really expect that
02:00:17saratogais that red my fault ?
02:00:32amiconnHmm, also my new client scored 28000 in this round. If the limit for fast client is 30000, that might cause this client to never enter the 'fast' category if there aren't enough builds
02:01:15saratogahmm warnings in files I didn't touch so probably not
02:01:16amiconnIt *should* be as fast as the other VM was before (same host, same number of cores; old VM is shut down and retired from the build system)
02:01:19gevaertsamiconn: it's a little over 26000
02:01:56amiconnAlso, does the server reconsider categries every round, or only if the client reconnects?
02:02:58gevaertsevery round
02:03:29gevaertsthe cutoff score is actually the score of the heaviest build
02:05:35 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
02:10:00Unhelpfulamiconn: i was looking at the data reordering you did for mpeg idct... i was thinking that a similar thing might me a good idea for jpeg, but that supports multiple transform sizes, and the optimal ordering would vary for each...
02:10:47 Quit martian67 (Read error: 60 (Operation timed out))
02:11:55Unhelpfuldoing it for *one* pass could be fairly simple, and could be done much as the transposed vs non-transposed cases are handled now
02:15:42 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
02:16:18saratogaUnhelpful: ping
02:17:30saratogais the code here for the negative power of 2 case optimal?
02:21:00Unhelpfulsaratoga: is the purpose of the extra shift and add to round the result?
02:21:38 Join mobyone [0] (n=50d4240b@gateway/web/cgi-irc/
02:22:22saratogaUnhelpful: I'm not sure, it seemed odd to me
02:22:58amiconnmingw32-sdl cause sred & yellow
02:23:07saratogai suppose the negative case is really too rare to make much difference anyway
02:24:00 Quit JdGordon| (Ping timeout: 180 seconds)
02:24:20*amiconn installs plain sdl
02:24:33Unhelpfulit seems pretty odd to me, too... if it's supposed to be returning x*(2**i) then i'm not really sure what the -i case is doing.
02:24:54mobyoneanyone of you know about rockbox dissapering from the ipod, and the ipod just turning back to normal?
02:24:57CIA-69New commit by saratoga (r21940): Optimize overlap_math by only doing shifting if theres gain, and moving the check for sign outside of the for loop. 3% speedup on PP5024.
02:26:01Unhelpfulif the purpose is to round the result because some of the bits have been shifted out, i'd think this would be more correct:
02:26:38Unhelpful...actually, i think that might be the same :)
02:26:51 Join mc2739 [0] (
02:26:54martian67any oldschool c64 fans here?
02:27:10martian67filters and clean samples have been figured out
02:27:15martian67thats a real SID :)
02:27:42mobyonei remember jumpman junior lol
02:29:22CIA-69New commit by mt (r21941): Generic codec-extradata parsing, in preparation for addition of AAC/RM.
02:33:46 Quit mobyone ("CGI:IRC (EOF)")
02:34:47 Quit bertrik (Read error: 60 (Operation timed out))
02:37:44amiconnSorry for being too slow so that another round was slightly broken - the VM is fixed now
02:39:08*amiconn thinks that crosscompiled sims should be a build option, not just selected by the first available sdl-config
02:45:49 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
02:53:46 Quit martian67 (Read error: 110 (Connection timed out))
02:56:29saratogais it possible to optimize the standard sequential mono channel buffer to interleaved stereo conversion any beyond just doing out[2*i] = in[i] in a for loop?
02:58:12Unhelpfulsaratoga: surely an unrolled loop using ldmia/stmia would be an optimization on ARM?
02:58:51 Join ReKleSS [0] (n=ReKleSS@
02:59:25***Saving seen data "./dancer.seen"
02:59:26saratogaUnhelpful: I assumed gcc would do that for me if I'm sequentially loading, but maybe not . ..
02:59:44Unhelpfulsaratoga: where does the code in question reside?
03:00:18Unhelpfulhrm, actually, unrolling and ldmia can work on the input side.. the output side is trickier, eh?
03:00:32 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
03:00:33saratogaUnhelpful: libcook/cook_fixedpoint.h
03:00:46saratogaoutput_math, but I'm about to change it
03:00:49saratogalet me pastebin the new version
03:01:54 Quit Thundercloud (Remote closed the connection)
03:02:27saratogahuh prefetch abort on ARM but the sim works fine and I didn't touch any ASM
03:03:11JdGordonmislaigned data
03:03:55saratogaJdGordon: int32_t outbuf[2048] __attribute__((aligned(32)));
03:04:09saratogaall I did was change that to 32 from 16 bit
03:04:37*JdGordon is out of ideas then :p
03:04:49saratogai'll try it without explicit alignment
03:07:35 Quit w1ll14m (" HydraIRC -> <- Po-ta-to, boil em, mash em, stick em in a stew.")
03:07:38Unhelpfulsaratoga: for the 1->2 case i suspect you could do some very clever tricks on armv6... on lower arm versions i'm less sure?
03:08:12saratogastill no luck
03:08:39 Quit martian67_ (Read error: 110 (Connection timed out))
03:09:54 Nick YpsyZNC is now known as Ypsy (
03:10:04saratogaUnhelpful: ok if its hard I don't want to deal with it just yet, we have much easier optimizations available right now
03:10:44Unhelpfulie, you could use ldmia to load in, say, 8 16-bit values in 4 registers, use the halfword-packing instructions to expand that to 8 registers that each contain the same value in top and bottom half, and then store those with stmia
03:11:18saratogathey're actually all 32 bit values
03:11:45saratogahuh prefetch abort at 0x7c327c32
03:11:50saratogai wonder how the hell it got there
03:12:11Unhelpfuli thought our PCM data was 16-bit, except for while passing through DSP?
03:12:38 Quit dmb (Read error: 104 (Connection reset by peer))
03:13:50saratogaUnhelpful: codecs operate at 32 bit precison mostly internally, and then tell the DSP engine what precision they're giving it
03:14:10saratogai'm trying to switch it to output 32 bit rather then 16 bit now, but its giving me aborts on ARM for whatever reason
03:15:28 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
03:15:32Unhelpfulsaratoga: ah. hrm, well, if it's already 32-bit, i'm not really sure that you can do anything special with stores while going from mono to stereo, but you could still load four or more values with ldmia, and then store them with str
03:15:49saratogahaha what the hell, putting it in IRAM instead of DRAM fixes the abort
03:16:05Unhelpfulsaratoga: i suspect you have an alignment issue ;)
03:16:16 Join scorche [50] (n=scorche@rockbox/administrator/scorche)
03:16:34saratogayes probably some other struct is being moved around
03:17:25saratogahow do I best go about trouble shooting this?
03:19:43saratogaUnhelpful: so this likely occurs because a word is being accessed thats not word aligned right?
03:20:07saratogaso I should be looking at structs containing char or int16_t? or something else?
03:20:35 Join z35_ [0] (
03:20:44 Quit z35_ (Read error: 104 (Connection reset by peer))
03:23:03 Quit mt (Read error: 113 (No route to host))
03:23:16 Join z35_ [0] (
03:23:27 Quit z35_ (Read error: 104 (Connection reset by peer))
03:23:36 Quit martian67 (Read error: 110 (Connection timed out))
03:27:25Unhelpfulsaratoga: i don't really know... any global or local variable should already have the proper alignment. the only time i've managed to mess up alignment is when doing dynamic allocations.
03:27:44FlynDicenotlistening: Sorry to take so long but not much time for irc for me lately... Yes please test the patch mc2739 posted in the forum with as many sd cards as you can. I can't remember if you were one of the ones with troublesome cards before. That patch basically reactivates voltage scaling at 1.10 v for the unboosted state but it boosts on disk accesses. My card worked just fine without the boost at 1.10 but I could make it
03:27:46FlynDicefail at 1.05 v. With the boost during disk accesses my sd works just fine at the 1.05v setting. We would be planning to use a 1.10v setting for the final product though.
03:28:26 Quit scorche (Read error: 104 (Connection reset by peer))
03:29:26saratogaok i'll leave this as an excercise for MT
03:30:26 Join martian67 [0] (
03:31:00 Quit martian67 (SendQ exceeded)
03:31:35 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
03:38:20CIA-69New commit by saratoga (r21942): Commit FS #10413 by Michael Chicoine. Fixes upside down menu in mpegplayer when watching videos with the screen inverted.
03:38:42 Quit martian67_ (Read error: 110 (Connection timed out))
03:42:04 Quit martian67 (Read error: 60 (Operation timed out))
03:42:30CIA-69New commit by kkurbjun (r21943): Brickmania: Make the defines dependent on screen size generic calculations
03:45:37 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
03:49:09saratogaare there any stats for commits/day for rockbox over time?
03:49:45JdGordonon ohloh and
03:50:40 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
03:52:19 Join toffe82 [0] (
03:53:26 Join dmb [0] (n=Dmb@unaffiliated/dmb)
03:55:32saratogaI don't see historical data on either though
03:57:09CIA-69New commit by kkurbjun (r21944): Brickmania: Correct a bug in the ball/brick collision (large screens (640x480) did not collide with the bottom row of bricks)
04:00:06 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
04:01:00 Quit martian67_ (SendQ exceeded)
04:01:35 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
04:07:16 Quit TheSeven (Nick collision from services.)
04:07:31 Join The_Seven [0] (
04:07:35 Nick The_Seven is now known as TheSeven (
04:08:44 Quit martian67 (Read error: 110 (Connection timed out))
04:09:30 Join scorche [50] (n=scorche@rockbox/administrator/scorche)
04:14:22 Join m67_l3 [0] (n=martian6@about/linux/regular/martian67)
04:16:24 Part wincent ("Kopete 0.12.7 :")
04:18:52 Join dys` [0] (
04:21:03 Quit mc2739 ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
04:23:13 Quit martian67_ (Read error: 110 (Connection timed out))
04:30:09 Quit dys (No route to host)
04:31:34 Quit GreatBeaver ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
04:36:47 Quit stettberger (Read error: 60 (Operation timed out))
04:36:47 Quit webmind (Read error: 60 (Operation timed out))
04:36:48 Join stettberger_ [0] (
04:37:35 Join webmind [0] (
04:59:26***Saving seen data "./dancer.seen"
05:03:45 Nick Ypsy is now known as YpsyZNC (
05:06:42 Quit r0b- (Read error: 110 (Connection timed out))
05:07:03 Join r0b- [0] (
05:16:47 Quit m67_l3 (Remote closed the connection)
05:20:06 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
05:20:39 Quit martian67 (SendQ exceeded)
05:21:34 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
05:22:28 Quit martian67 (SendQ exceeded)
05:23:04 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
05:24:04 Quit martian67 (SendQ exceeded)
05:30:11 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
05:31:06 Quit martian67 (SendQ exceeded)
05:31:52 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
05:41:13 Quit martian67 (Read error: 60 (Operation timed out))
05:42:04 Join gartral [0] (n=Gartral@unaffiliated/gartral)
05:42:17 Join blackbir [0] (
05:42:34gartralhey all.. i cant build today... i keep getting make: *** No rule to make target `/home/gareth/rockbox/sansa/apps/bitmaps/native/rockboxlogo.176x220x16.o', needed by `/home/gareth/rockbox/sansa/bitmaps/rockboxlogo.h'.
05:44:36Unhelpfuldid you try again after a make clean?
05:45:22gartrali have a script too build, it runs make veryclean each time
05:45:47 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
05:46:28gartralalthough, looking through, it seems my SOURCES file in that dir got currupted somehow, so at least i know what too fix.. sorry
05:47:05CIA-69New commit by kkurbjun (r21945): Bubbles: Make new screen support a bit more generic and add support for 640x480 and 480x640 screen sizes
05:49:43gartralok, no.. half my sources are awol
05:50:14Unhelpfulgartral: then it sounds like you need a fresh checkout.
05:51:47 Quit scorche (Read error: 104 (Connection reset by peer))
05:53:39 Join scorche [50] (n=scorche@rockbox/administrator/scorche)
05:56:46gartralyep, pebkac error
05:57:32 Quit gartral ("Why must we be chicken or sheep when we have the blood of the Fox, Wolf?")
05:58:29 Part blackbir
06:00:48 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
06:08:26 Quit r0b- (Read error: 110 (Connection timed out))
06:08:50 Quit martian67 (Read error: 110 (Connection timed out))
06:15:21 Join m67_l3 [0] (n=martian6@about/linux/regular/martian67)
06:16:13 Quit m67_l3 (SendQ exceeded)
06:16:52CIA-69New commit by kkurbjun (r21946): Brickmania, Chessbox, Blackjack, and Solitaire: Add support for 640x480 and 480x640 screens
06:24:02 Quit martian67_ (Read error: 110 (Connection timed out))
06:27:39 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
06:28:35 Quit martian67 (SendQ exceeded)
06:29:29 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
06:30:12 Quit saratoga ("Page closed")
06:30:19 Quit martian67 (SendQ exceeded)
06:30:53 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
06:38:36CIA-69New commit by kkurbjun (r21947): Clix: Add support for screen widths larger than 480
06:45:07 Join funman [0] (
06:47:15 Join feisar--_ [0] (
06:47:15 Quit feisar-- (Read error: 104 (Connection reset by peer))
06:53:59 Quit martian67 (Read error: 110 (Connection timed out))
06:55:25 Join petur [0] (
06:56:23 Join homielowe [0] (n=homielow@
06:58:39 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
06:59:27***Saving seen data "./dancer.seen"
06:59:51 Quit martian67 (SendQ exceeded)
07:00:26 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
07:10:04 Part toffe82
07:15:31 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
07:16:27 Quit dmb (Read error: 113 (No route to host))
07:19:14 Quit CaptainKwel (Remote closed the connection)
07:23:47 Quit martian67 (Read error: 110 (Connection timed out))
07:27:26 Quit martian67_ (Read error: 60 (Operation timed out))
07:30:31 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
07:46:29 Join martian67 [0] (
07:53:28 Quit martian67_ (Success)
07:57:29 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
07:58:04funmanI think I can safely stop expecting an answer from the product marketing manager at austriamicrosystems
07:59:27Unhelpfulwhen did you write them? :)
08:03:32 Join r0b- [0] (
08:04:17 Join n1s [0] (n=n1s@rockbox/developer/n1s)
08:08:42 Join _lifeless [0] (n=lifeless@
08:14:17 Quit Horscht (Read error: 110 (Connection timed out))
08:22:57 Quit __lifeless (Read error: 110 (Connection timed out))
08:24:59 Join robin0800 [0] (
08:28:02 Quit martian67 (Remote closed the connection)
08:28:23 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
08:29:21 Join stoffel [0] (
08:40:23 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
08:40:23 Quit pixelma (Nick collision from services.)
08:40:31 Quit amiconn (Nick collision from services.)
08:40:32 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
08:40:40 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
08:40:50 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
08:44:05 Join Rob2222 [0] (
08:45:34 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
08:53:28 Quit martian67 (Success)
08:57:15 Quit petur (Remote closed the connection)
08:57:54 Join petur [0] (
08:59:28***Saving seen data "./dancer.seen"
09:00:45 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
09:01:29 Quit Rob2223 (Read error: 110 (Connection timed out))
09:08:21 Quit martian67_ (Success)
09:14:43*Unhelpful has managed to move a good bit of the lcd bitmap driver code into an lcd-bitmap-common.c... also consolidated some similar-ish functions.
09:15:23Unhelpfulit knocks out rather a *lot* of code duplication... and it's a binsize win on *most* targets, though i not that coldfire b/w is a bit up :/
09:15:45Unhelpfulthere's a FS for a similar attempt already, iirc... i'll put a patch up there soon, i think :)
09:15:55 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
09:23:43 Quit martian67 (Read error: 110 (Connection timed out))
09:37:10 Join DarkDefender [0] (
09:38:04markunamiconn: wasn't there a reason why you thought the LCD code shouldn't be shared?
09:38:34 Quit martian67_ (Remote closed the connection)
09:39:36pixelmakkurbjun: maybe it's because I'm still not fully woken up and am not used to this sort of abbreviations but I had to read a few times over b/w to understand that it meant "between" (my first thought was "black and white"), and I already saw yesterday that you used w/ for with. All that in comments where I don't see the reason to not spell it out...
09:40:17 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
09:41:21 Quit z35 (Read error: 110 (Connection timed out))
09:44:46Unhelpfulmarkun: for the most part the shared code did not differ at all between targets or LCD types, or differed only for color vs b/w
09:46:04Unhelpfulall of the puts_* variants that call puts_style_offset, for the purpose of allowing you to omit a few arguments, for example
09:54:22 Join merbanan [0] (
09:55:41 Quit martian67 (Read error: 60 (Operation timed out))
09:59:45 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
10:00:25 Join mt [0] (n=mt@
10:00:45 Quit martian67 (SendQ exceeded)
10:01:09 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
10:08:51 Join bertrik [0] (
10:10:42 Quit martian67 (Read error: 60 (Operation timed out))
10:15:55 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
10:21:32*pixelma wonders if MMCs would work with the D2's SD slot (in Rockbox)
10:25:39markunUnhelpful: I know, I was never against it
10:27:18funmanpixelma: says MMC work in the OF so there is no technical limitation
10:28:30Unhelpfuli started thinking it should be done when i was working on pictureflow text scrolling... if you want to make a change to scrolled text drawing you have to touch so many files :/
10:30:16pixelmafunman: well, I thought it should work in the OF since I've been told that the MMC and the SD protocol are very similar and most other devices that support SDs support MMC as well. But I'm not sure if it works in Rockbox already (though there is an MMC driver for the Ondios I don't know if the SD one is now totally seperate and what it can do)
10:30:37 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
10:30:48 Quit at0m ("new elecs for teh studio!")
10:31:35funmanpixelma: since the protocols are similar the current SD driver could be modified to support MMC if it doesn't already (the original ingenic code did that, but MMC was removed since the onda have a µSD slot not compatible with MMC)
10:31:36 Quit martian67_ (Remote closed the connection)
10:32:10pixelmawhich is why I asked about the D2
10:32:14funmanif another MMC driver is written, perhaps the ondio-specific ata_mmc.c could be splitted into 2 parts, one of them would be reused by the new driver
10:32:36 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
10:33:07 Quit stoffel (Read error: 104 (Connection reset by peer))
10:34:07 Join stoffel [0] (
10:34:51funmanis a SD (and MMC) driver for cowond2 in the works? I can't see one in SOURCES
10:35:10 Quit martian67_ (SendQ exceeded)
10:35:34 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
10:36:27 Quit robin0800 (Read error: 110 (Connection timed out))
10:37:51pixelmathis is what I read from the mails in the dev-ml - the plan is to switch storing settings etc. on the SD card since the inbuilt memory is still read only. I think a working SD driver must be existent for this first
10:38:08 Quit martian67 (Read error: 110 (Connection timed out))
10:38:18 Join robin0800 [0] (
10:44:48bertriklinuxstb, on most sansas, the backlight is controlled over i2c (the analog parts, including the codec, is connected through i2c)
10:49:16funmanall but c200v2 and m200v4
10:58:11 Quit n1s (Read error: 110 (Connection timed out))
10:59:29***Saving seen data "./dancer.seen"
11:00:22 Join m67_l3 [0] (n=martian6@about/linux/regular/martian67)
11:02:21 Quit petur ("real life....")
11:05:44FlynDicefunman: RE: SD & pl180 I think we're assuming PCLK and MCLK are the same inata_sd_as3525 and I don't think that is correct. See p2-2 &p2-7 of pl180 docs
11:06:09FlynDiceer incorrect of course...
11:08:08FlynDiceIt says PCLK >= 3/8 MCLK and both are less than 100 MHz
11:08:18 Quit martian67_ (Success)
11:11:09 Join PaulJam [0] (i=Paule@
11:11:37 Join graey [0] (
11:12:47 Quit PaulJam (Nick collision from services.)
11:12:56 Join PaulJam_ [0] (i=Paule@
11:13:18graeyhello all, I'm finishing the guide for setting up a rockbox development debian/ubuntu VM (vbox) today, should I merge it with the existing page in the wiki or create a new one for it?
11:15:26 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
11:16:01 Quit goffa (Remote closed the connection)
11:16:05 Join goffa [0] (n=goffa@
11:16:33funmanFlynDice: hum why isn't there an equality there ..
11:17:01funmani believe this makes timeouts calculation and frequency for identification mode incorrects, do you see something else?
11:18:56FlynDicefunman: Still trying to make sense of it, I can't find out how to determine MCLK yet but we base everything off of PCLK right now assuming PCLK=MCLK
11:19:14 Quit stoffel (Read error: 113 (No route to host))
11:23:23FlynDicefunman: No I don't see anything else but I haven't looked very hard yet at that, still trying to determine how MCLK is figured out
11:23:24 Quit m67_l3 (Read error: 110 (Connection timed out))
11:23:49funmanso far we only know mclk <= 165.3MHz and mclk <= 100MHz
11:24:12funmanI remember looking at this information some months ago but don't remember the output ..
11:27:46FlynDiceThere are other MCLK's for the as3525 with separate dividers(issi, i2so) but there's no connection to the pl180 MCLK that I can determine
11:28:03FlynDiceer i2si
11:29:22funmani think they are not related
11:30:27 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
11:30:41funmanthe diagram on page 101 of the pdf doesn't show SD module
11:31:03funmanbtw .. the ide clock is needed for SD to operate properly
11:31:42funman(and more than that the ide peripheral can't be reset for SD controller to work at all)
11:34:19 Join n1s [0] (n=n1s@rockbox/developer/n1s)
11:38:38 Quit martian67 (Read error: 110 (Connection timed out))
11:39:55FlynDicefunman: Im googling for info now with fingers crossed ;)
11:44:39 Quit r0b- (Read error: 104 (Connection reset by peer))
11:45:43 Quit funman ("free(random());")
11:45:47 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
11:53:14 Quit martian67_ (Read error: 110 (Connection timed out))
11:56:29linuxstbbertrik: Does the timer interrupt (and hence things like sleep() and current_tick) work on the Meizu?
11:59:03 Nick dys` is now known as dys (
12:00:22 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
12:03:41bertriklinuxstb, yes the timer interrupt works, I think sleep works (at least it doesn't block) and I haven't tried current_tick
12:04:27linuxstbInterrupts in general are working (the i2c interrupt does for example), but for some reason the timer doesn't.
12:04:44linuxstbAlthough it could just be related to the clock settings - I haven't touched those...
12:05:29bertrikthe s5l8700 datasheet is a bit weird, at some points it claims to have 4 timers and at other it claims to have 2, the timer output pins are partially incorrect too
12:05:41linuxstbIn SVN, most of crt0.S is commented out for the Nano, as it caused a freeze. I've been slowly re-enabling bits, and it's just the setting of CLKCON which makes things freeze.
12:05:47bertrikthe meizu does indeed have 4 timers, but maybe the s5l8701 has only 2
12:06:15linuxstbAlthough you're using timerb?
12:07:09bertrikcurrently, TIMER A is unused (thinking of making it a usec timer), TIMER B is the kernel timer, TIMER C is used for backlight PWM and timer D is used as the general timer
12:07:15linuxstbI tried displaying the value of that timer (in a while(1) loop) and it was incremenenting up to about 0x3000 and then looping back to 0 without running the rockbox timer init, and then it changed to incrementing up to about 0x100 before looping back to 0.
12:08:07*linuxstb wishes the crt0.S had some comments...
12:08:22 Quit martian67 (Read error: 110 (Connection timed out))
12:09:25bertrikCLKCON controls basically all clocks (FCLK, PCLK, HCLK, MCLK)
12:10:13bertrikdisplaying the default PLLPMS would give a lot of info how the clocks are configured by default
12:10:36 Join AndyI [0] (i=AndyI@
12:13:07linuxstbPLL0PMS = 0x00021200, PLL1PMS=0x00010103
12:13:14bertriklinuxstb, weird to see the timer running up to different values
12:13:51 Join Exiled [0] (n=4b8e8c7a@gateway/web/cgi-irc/
12:15:21bertriklinuxstb, I also found that PCLK actually seems to use twice as fast as expected (i.e. running at HCLK frequency instead of HCLK/2), but I could be mistaken this early in the port
12:15:27 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
12:16:16bertrikthe datasheet seems to suggest to use PLL0 for the system clocks and PLL1 for the audio clocks (to set the sample rate)
12:16:24 Quit gevaerts (Read error: 60 (Operation timed out))
12:16:40 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
12:23:37 Quit martian67_ (Read error: 110 (Connection timed out))
12:23:45 Quit AndyIL (Read error: 110 (Connection timed out))
12:26:36 Join stripwax [0] (
12:28:12linuxstbbertrik: Is there anything else from the Meizu port I can test on the Nano? I guess your audio driver... ;)
12:28:46Unhelpfulok, the merge of scrolling and styled text drawing for all bitmap targets is at FS #4817. yay for tasks that won't die. :)
12:29:21bertriklinuxstb, the ADC would be nice, you can probably recognise a battery voltage on one of the channels
12:29:29stripwaxHm, is the twiki user "AsdFgh", real name "Zxc Vbn", supposed to be there?
12:30:05bertrikor PCM+DMA maybe, you can dump the DMA current address and DMA count registers to see if it is actually running
12:30:32linuxstbbertrik: Yes, I was thinking about adc...
12:31:16linuxstbI guess the next big thing is the clickwheel driver - hopefully the hardware is the same as the 1st gen nano, and I just need to change the low-level code (i2c and gpio IIRC)
12:31:24bertrikthose PLL settings are quite low, the meizu bases everything off a 32768 Hz crystal, maybe the nano uses a much faster crystal
12:31:57 Quit PaulJam_ (".")
12:33:09linuxstbbertrik: Yes, it does. TheSeven has just told me that the Meizu code would result in my nano running at 11GHz... ;)
12:35:01bertriklinuxstb, I think we need to know the default setting of CLKCON
12:35:26 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
12:35:57bertrikBTW, I'm not completely sure how the ADC interrupt is supposed to work, there is no way to clear it in the peripheral itself as far as I can tell from the datasheet
12:36:02pixelmastripwax: probably not, but only the Swedes are able to delete pages
12:36:45bertrikin another datasheet of a samsung device, the register map was the same, except it did have an additional explicit interrupt clear register
12:38:10 Quit martian67 (Success)
12:41:32 Quit stripwax ("")
12:43:45 Join stripwax [0] (
12:43:56 Join _zic [0] (
12:45:25 Nick stettberger_ is now known as stettberger (
12:45:58linuxstbstripwax: I think the Swedes go through the wiki occasionallly and clean up invalid registrations. No idea how frequently they manage to do it though...
12:48:11TheSevenbertrik: battery ADC seems to be on the power management asic, accessible via i2c
12:52:39 Join mirak [0] (
12:59:32***Saving seen data "./dancer.seen"
13:10:04 Join wincent [0] (n=wincent@
13:12:15 Quit Exiled ("CGI:IRC (EOF)")
13:15:30 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
13:20:32 Join r0b- [0] (
13:23:25 Quit martian67_ (Read error: 110 (Connection timed out))
13:23:57 Join mcuelenaere [0] (
13:30:53 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
13:31:25CIA-69New commit by dave (r21948): Various Nano2g changes - enable the rest of crt0.S (including clock setup) and make the nano2g bootloader do something useful (displaying gpio ports). ...
13:32:29bertriklinuxstb, I'll have a try right away to see if meizu m3 bootloader still works
13:32:40linuxstbbertrik: Thanks. This shouldn't affect you at all.
13:33:08stripwaxThe clix keymap on ipod either doesn't make much sense or doesn't work properly, I think. What is the physical activity associated with BUTTON_SCROLL_BACK / BUTTON_SCROLL_FWD on ipod? Is that the same action as increasing volume from wps?
13:33:12linuxstbbertrik: I would be curious to know the values of TBCNT if you can print them in a loop in the bootloader.
13:33:25linuxstbOn the nano it appears to loop from about 0x00 to 0x50
13:34:26bertrikthe meizu m3 bootloader still works as before
13:36:39CIA-69New commit by dave (r21949): Nano2g button driver - currently only implementing the hold switch. Should have been part of previous commit.
13:37:54bertriklinuxstb, I would expect it to run up to 99 (0x63)
13:38:05bertrikkernel-s5l8700 assumes a PCLK of 100 MHz
13:38:24linuxstbOK, let me keep track of the max value of it...
13:39:01 Quit martian67 (Read error: 110 (Connection timed out))
13:39:06bertrikIt would probably be a good idea to make #defines or functions somewhere to get various clock speeds (PCLK etc)
13:39:43bertrikif the meizu clock setting is different from the nano2g
13:40:26stripwaxah, clix doesn't check BUTTON_REPEAT. I guess ipod only ever really sends BUTTON_SCROLL_FWD|BUTTON_REPEAT and not BUTTON_SCROLL_FWD on its own
13:40:40 Quit robin0800 (Read error: 60 (Operation timed out))
13:41:57linuxstbbertrik: It seems to up to 98
13:42:10linuxstb(unless I'm just missing 99...)
13:43:14bertrikah, there seems to be an off-by-one error in the driver
13:43:54bertrikthe timer prescaler register should be filled with (value-1) but the timer count register can be filled directly with the interval value (without subtracting 1)
13:44:00 Join Thundercloud [0] (
13:44:30linuxstbbertrik: But for some reason the timer interrupt is not firing for me...
13:44:47bertrikso if TxDATA0 = 100, I expect it to run up to 99 with a 100-cycle interval
13:45:45bertrikmaybe the interrupt bit mapping is different
13:45:55 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
13:46:32bertrikyou could enable more interrupts in INTMSK and see which one triggers when the count is reached
13:48:47linuxstbYes, that seems likely... I think TheSeven mentioned some interrupts were different (I2C was the same though)
13:49:07 Quit r0b- (Read error: 110 (Connection timed out))
13:49:28 Join r0b- [0] (
13:51:31bertriksystem-s5l8700 has default handlers for all interrupts that are "weakly linked" (IIRC) so you can override them in other code. Anyway uncaught interrupts are indicated by a panic this way.
13:53:47 Quit martian67_ (Read error: 110 (Connection timed out))
13:54:04CIA-69New commit by bertrik (r21950): S5L8700: fix off-by-one error in kernel timer and fix incorrect interrupt mask in pcm driver.
14:00:20 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
14:00:35 Join robin0800 [0] (
14:01:35 Quit aditya (Read error: 110 (Connection timed out))
14:02:27 Join aditya [0] (n=aditya@
14:09:09 Quit martian67 (Read error: 110 (Connection timed out))
14:15:28 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
14:21:03gevaertsBagder, Zagor: did you switch to this last-6 average? It looks like you did, but without clearing the averages first, so e.g. kkurbjun-monk-kkurbjun has a totally wrong average now
14:23:50 Quit martian67_ (Read error: 110 (Connection timed out))
14:25:00mcuelenaereBagder: could you add at least one sim for the Onda's? Currently the VX747 sim is broken and it would be usefull to see that in the build table
14:25:36 Quit evilnick_home (Read error: 104 (Connection reset by peer))
14:30:29 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
14:33:15CIA-69New commit by stripwax (r21951): Make clix playable on ipod by fixing keymap (so up doesn't quit), button repeat logic (so holding down select doesn't repeatedly fire select events), ...
14:38:06 Quit stripwax ("")
14:38:20CIA-69New commit by mcuelenaere (r21952): Fix PDbox compilation
14:38:41 Quit martian67 (Read error: 110 (Connection timed out))
14:45:28 Join m67_l3 [0] (n=martian6@about/linux/regular/martian67)
14:47:22 Join Sajber^ [0] (
14:48:58 Join evilnick_home [0] (n=evilnick@
14:52:43CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
14:52:43CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
14:52:43 Quit gevaerts ("Reconnecting")
14:52:53 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
14:54:11 Quit martian67_ (Read error: 110 (Connection timed out))
14:54:52 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
14:54:54mcuelenaeregevaerts: is the VX747 the only target using MULTIDRIVE currently in SVN?
14:55:05gevaertsmcuelenaere: also VX777 IIRC
14:55:16gevaertsshotofadds plans to add one tomorrow though
14:55:44mcuelenaerethe VX747 sim seems to be broken
14:56:11mcuelenaereapps/bookmark.c uses strip_volume(), which is defined in firmware/common/dir_uncached.c; but this isn't compiled in sim-builds
14:58:32CIA-69New commit by gevaerts (r21953): Flyspray: FS #10326 ...
14:59:35***Saving seen data "./dancer.seen"
14:59:40mcuelenaeremoving the function to dircache.c fixes this, but I'm not sure if this is the right thing to do
14:59:45*gevaerts has a look at this onda sim
15:00:30 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
15:01:11 Quit J-23 (Nick collision from services.)
15:01:50 Join teru [0] (
15:04:47 Join J-23_ [0] (
15:05:30 Nick J-23_ is now known as J-23 (
15:05:34gevaertsmcuelenaere: this seems to be a general incompatibility between sim builds and multivolume. Other multidrive builds (sansa,ondio) declare the NUM_DRIVES bit in an #ifndef SIMULATOR block
15:05:43pixelmamcuelenaere: wouldn't that break the function on targets without dircache?
15:06:29mcuelenaerepixelma: I don't know; like I said I wasn't sure if that's the right thing to do
15:06:46mcuelenaeregevaerts: so the Onda's need to #define the multidrive part only #ifndef SIMULATOR?
15:06:55 Join spacetraveler [0] (n=spacetra@
15:07:07gevaertsmcuelenaere: that would at least make the problem go away.
15:07:39gevaertsThe proper solution would of course be to fix the simulator to handle multidrive/multivolume, but that will be a lot more work
15:08:01gevaertsAnother solution would be to handle this in config.h
15:08:26 Join z35 [0] (n=z35@
15:08:37 Quit m67_l3 (Read error: 110 (Connection timed out))
15:09:23gevaertsI don't think #undef-ing things in config.h depending on another define (SIMULATOR) is very clean though
15:09:41 Quit mirak ("Ex-Chat")
15:10:28mcuelenaeregevaerts: I saw that MULTIDRIVE gets defined in config.h depending on CONFIG_STORAGE, shouldn't the MULTIDRIVE defines in config-*.h get removed then?
15:10:43mcuelenaereor do you want them to be explicitly in there?
15:10:56 Quit spacetraveler (Client Quit)
15:11:38gevaertsmcuelenaere: yes and no. It gets defined on CONFIG_STORAGE_MULTI, but not for e.g. the sansas that have two drives but only a single driver
15:12:14gevaertsso yes, you could remove it for the ondas
15:12:35mcuelenaerebut don't those define MULTIVOLUME?
15:13:06 Join Stephen_ [0] (n=S@
15:13:10mcuelenaereah no they don't
15:13:12gevaertsnot anymore
15:13:29gevaertsalso, it doesn't mean the same thing anymore
15:13:41 Quit DarkDefender (Remote closed the connection)
15:14:05gevaertswell, they do get MULTIVOLUME from config.h, but it's a slightly different sort of multivolume :)
15:14:26Mikachui have two partitions on my nano, so i've had multivolume enabled, do i need to change anything?
15:15:02 Nick fxb__ is now known as fxb (
15:15:25 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
15:15:38gevaertsMikachu: if you just added #define HAVE_MULTIVOLUME, that should still work the same
15:15:51Mikachuokay, we'll see :)
15:15:59 Quit z35 (Read error: 113 (No route to host))
15:16:11Mikachui'm also going to check if hid has started working for me yet
15:17:13gevaertsinternally things changed. There now is HAVE_MULTIDRIVE, HAVE_MULTIVOLUME and NUM_VOLUMES_PER_DRIVE. If you set HAVE_MULTIVOLUME yourself, you'll get NUM_VOLUMES_PER_DRIVE==4. If you get it because of HAVE_MULTIDRIVE, you'll get NUM_VOLUMES_PER_DRIVE==1
15:18:36CIA-69New commit by mcuelenaere (r21954): Fix Onda VX747(+)/VX777 simulator build
15:18:50CIA-69New commit by gevaerts (r21955): adapt c200v2 and view defines to the new drive/volume handling scheme
15:21:16mcuelenaeregevaerts: was there a reason for moving the USB_ENABLE_* defines from usb_core.h to config.h?
15:23:13gevaertsmcuelenaere: I think they are really configuration things, and therefore belong in config.h. I've been unconfortable with them being in usb_core.h since about the time I put them there...
15:23:18CIA-69New commit by mcuelenaere (r21956): Fix warning when compiling with logf() enabled
15:23:35 Quit martian67 (Read error: 110 (Connection timed out))
15:27:01 Quit martian67_ (Read error: 60 (Operation timed out))
15:27:44CIA-69New commit by gevaerts (r21957): re-add the usb_core.h include. r21953 was a bit too aggressive
15:30:06Mikachugevaerts: well, i tried hid again, and it's worse now, it hangs rockbox when i press a button, when i unplug the cable it wakes up
15:30:22gevaertsdid you update your bootloader since last time?
15:30:29 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
15:30:35Mikachui'll try that
15:33:15CIA-69New commit by mcuelenaere (r21958): Make the CPU boost log scroll if lines are too big.
15:35:26 Quit daurnimator (Read error: 113 (No route to host))
15:35:52*gevaerts thinks that we're committing too fast. The build system is lagging!
15:36:54Mikachugevaerts: didn't help
15:38:44gevaertsMikachu: what exactly do you mean by "it hangs rockbox when i press a button"?
15:39:13Mikachui mean what i say, it stops responding to input when i'm in remote_control, and stops sending data to the computer when in disk mode
15:39:58*gevaerts hasn't seen anything like this at all :\
15:40:19Mikachuwell, as long as it works for everyone else, don't worry about it, but if you want me to test something let me know
15:40:22gevaertsIs this an official build?
15:40:30Mikachuno, but i tried that too
15:40:42Mikachui haven't tried an official build with the new bootloader yet, i guess it's worth a try
15:41:34Mikachuheh, it's a bit tricky to download a current build when the build system is compiling.. it removes the old ones before the new ones are finished
15:41:52 Join CaptainKwel [0] (
15:41:53*bertrik sees that the battery charger in the Meizu turned itself off after a few hours
15:42:44bertrik(that's a good thing, it's supposed to do that)
15:43:34Mikachugevaerts: no luck
15:44:35 Join dmb [0] (n=Dmb@unaffiliated/dmb)
15:45:17Mikachumultivolume works fine though
15:51:10CIA-69New commit by teru (r21959): set YOFS to 0 for portrait LCDs. ...
16:00:21 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
16:04:13 Nick YpsyZNC is now known as Ypsy (
16:08:09 Quit martian67 (Connection timed out)
16:14:39 Join Tomers [0] (
16:15:24 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
16:23:36 Join mirak [0] (
16:23:45 Quit martian67_ (Read error: 110 (Connection timed out))
16:27:04 Quit timc (Read error: 110 (Connection timed out))
16:29:15 Join Ubuntuxer [0] (
16:30:20 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
16:32:54 Quit robin0800 ("Leaving")
16:33:01 Quit ReKleSS ("Leaving")
16:37:02 Join robin0800 [0] (
16:37:16 Join faemir [0] (n=faemir@
16:39:07 Quit martian67 (Connection timed out)
16:45:27 Join m67_l3 [0] (n=martian6@about/linux/regular/martian67)
16:47:25teruDose anyone know how to make background of bubbles?
16:53:16 Quit martian67_ (Connection timed out)
16:55:35Tomersgevaerts: Implementing dynamic USB class drivers as a plugin would not allow multiple drivers (HID + Mass)
16:56:11kkurbjunpixelma: sorry for the confusion, I'll try to be conscious of that in the future.
16:56:29kkurbjunteru: are you trying to make a new bubbles background?
16:57:05gevaertsTomers: why not?
16:57:25Tomersgevaerts: Is it possible to load more than one plugin at once?
16:58:26gevaertswith the current infrastructure, no. There are several possibilities though
16:59:39***Saving seen data "./dancer.seen"
16:59:55terukkurbjun: no. background of bubbles for gigabeat dosen't match other parts.
17:00:37TomersAlso, USB drivers are not organized properly, but scattered around...
17:01:10gevaertsThe reasons why this doesn't work that I can see are : (a) only one plugin can have the UI. That's only a problem for UI plugins of course. (b) Plugins are linked at fixed addresses. This is the serious one, but we could link (some) plugins as position independent code
17:01:20kkurbjunteru, yeah, I think that is because bubbles was designed around a 4/3 screen while the gigabeat is a 3/4.
17:01:28pixelmabubbles should make better use of portrait screens, for that you will need to place the preview better
17:02:05kkurbjunI do think that empty space at the bottom of the screen looks bad though, it might be better to place all that information atht ebottom of the screen
17:02:20gevaertsTomers: also, we could conceivably have a single plugin that has all class drivers in it
17:02:37pixelmaand similar stuff. I once made a mockup of how I'd imagine that but couldn't get it to work properly (gave up quite early I admit though)
17:02:53Tomersgevaerts: Some sort of usb class manager
17:03:00*pixelma wonders if that mockup is still available
17:03:46Tomersgevaerts: OK. I'll start reading... and implementing...
17:04:51gevaertsTomers: one more point that I think is important : I think we should keep storage in core, although it could have the UI bit (i.e. the USB screen, with possibly speed and activity indicators) in a plugin.
17:05:00 Quit dash32 ("Verlassend")
17:05:51gevaertsAnyway, I think you should start by exploring a bit. There are a lot of possibilities, each with their own advantages and disadvantages
17:06:21gevaertsdon't worry if you don't have much committable code for a while :)
17:06:54 Join mitk [0] (
17:10:42mcuelenaeregevaerts, Tomers: I'm wondering whether this shouldn't be part of a plugin rework? Adding another 'plugin' space doesn't seem right to me (we already have codec & plugin)
17:11:28gevaertsmcuelenaere: exactly. I strongly prefer to see this in the normal plugin buffer
17:11:31Tomersgevaerts: Why should we keep the storage in core? What are the tradeoffs?
17:12:09pixelmaMikachu: re. the HID build problems on your nano - what OS are you using exactly? I guess it's some linux distro but how you describe it reminds me of that OSX problem (Rockbox USB hung in the USB screen for me, sometimes turning back to normal after disconnecting, sometimes needing a hard power off on my c250)
17:12:37mcuelenaeregevaerts: that would mean we couldn't load plugins when USB is plugged in though
17:12:57Mikachupixelma: the thing is it works perfectly fine with my friend's ipod video
17:12:58gevaertsTomers: it's a bit more resilient. It's not too hard to mess up an installation enough to make plugins not work. Of course, if we have bootloader USB on all involved targets it's not too important anymore
17:13:08Mikachupixelma: and same thing with the two ipods on his computer
17:13:21gevaertsmcuelenaere: maybe. That depends on how exactly it
17:13:24gevaerts's done
17:14:24gevaertsAlso, even if this happens, I'm not sure if it's a real problem.
17:15:26pixelmaMikachu: some USB ports, or cables?
17:15:29mcuelenaereit depends, if the USB 'plugin' has a GUI or not
17:15:37pixelmaerr... same
17:15:37 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
17:16:06Mikachupixelma: different cables too, yeah
17:16:06 Quit mirak ("Ex-Chat")
17:16:27CIA-69New commit by Ubuntuxer (r21960): Commit FS #10350, prevents to save an unchanged highscore and move the function show_highscore to the lib
17:16:33Mikachupixelma: i mean different cable on his computer and mine, but i tried the his and my ipod with the same cable/port on my computer
17:16:49gevaertsmcuelenaere: there's a reason I recommended to start by exploring a bit :)
17:17:24*mcuelenaere remembers an IRC talk about making plugins use the buffer space
17:18:38gevaertsmcuelenaere: that would require this long talked about possibility to grab a bit of the buffer without stopping playback. I don't think it would be fair to ask Tomers to start digging in the buffering code now :)
17:18:57mcuelenaeregevaerts: yes that's true :)
17:19:03gevaertsunless of course he wants to :)
17:19:19mcuelenaerewasn't there another SoC student working on the buffering code?
17:19:53gevaertsyes. BryanJacobs has to work in there because of the nonstandard requirements of his codec
17:19:57Tomersgevaerts: That would be for GSoC 2010, but I already graduated :-)
17:20:09 Join mirak [0] (
17:20:31teruthis is screen dump of bubbles from gigabeatf sim. it doesn't looks good and I hope this to be fixed.
17:21:30mcuelenaereteru: does it have some kind of garbage at the bottom on-target?
17:21:35mcuelenaeremy Onda VX747 has that..
17:21:37gevaertsTomers: ah, that's a bit unfortunate ;)
17:22:47Tomersbut the whole point of GSoC is to get you in the community, and I'll of course continue working on Rockbox (hopefully) after the project ends
17:23:16 Quit m67_l3 (Connection timed out)
17:25:29gevaertsTomers: so you'll do that bit next year anyway? ;)
17:25:39gevaertsteru: it looks exactly the same on target
17:26:47 Quit martian67 (Read error: 60 (Operation timed out))
17:29:01Tomersgevaerts: I'll try to convince Software Engineering students to apply for Rockbox in 2010 :-)
17:29:38gevaertsgood :)
17:30:28 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
17:32:04 Join dash32 [0] (n=dash32@
17:37:02 Quit Ubuntuxer ("Leaving.")
17:37:42pixelmaat least the text looks a bit off in that bubbles screenshot
17:39:45 Quit MrDuck (Read error: 113 (No route to host))
17:40:06pixelmaI wonder whether the changes since yesterday aren't oversimplifying things
17:44:16 Quit robin0800 (Connection timed out)
17:44:39pixelmateru: why did you change the NUM_SCORES to 5 on all targets? It was LCD_HEIGHT/10 before and none of the targets have an LCD_HEIGHT of 50
17:45:07 Join robin0800 [0] (
17:45:25pixelmaI haven't looked what it is used for but I guess for placement of the score number on the screen
17:51:54teruI thought 32 is too many and LCD_HEIGHT is defined to 5 for some targets. also, LCD_HEIGHT/10 is not likely to fit current score displaying.
17:53:30 Quit martian67 (Connection timed out)
17:55:39pixelmakkurbjun: with all your changes, are you sure that the plugins look the same (or at least reasonable) on all target)? Quite a lot of changes (thanks for the abbreviation reply by the way :) )
17:58:34 Quit Stephen_ ("Leaving")
18:02:38kkurbjunpixelma: I checked it on the archos devices, the nano, the H10, the gigabeat F, the ipod vieo, and the mrobe 500 to get a sampling of screen sizes
18:03:15kkurbjunI don't think it it 100% exact, but it looked fine from the testing I did and it makes the plugin easier to maintain
18:05:56pixelmasmall H10 or big H10?
18:06:36kkurbjunI think that one has the same height/width so I wanted to check that it looked ok on that too
18:08:16pixelmayes it has, the big one would be a good test case as well as it has the common 160x128 screen size. I would have tested all different sizes but then I know I'm always very carefull with these things, maybe a bit too much...
18:09:09kkurbjunyeah, larger screen sizes are less trouble as they have more slack for placement
18:10:40 Join flydutch [0] (
18:11:02mitkHi. There is problem between r21932 and r21933. uSDHC card has disappeared in Fuze at all with r21933. With 21932 Fuze still can see it.
18:11:26mcuelenaeregevaerts: when I define CONFIG_STORAGE as (STORAGE_NAND|STORAGE_SD), Rockbox doesn't seem to know that the SD card is removed when I hotswap it (with only STORAGE_SD it shows an empty root). Any ideas?
18:11:42gevaertsmcuelenaere: ah yes...
18:15:03gevaertsmcuelenaere: it really worked before? other drivers seem to call disk_unmount() when detecting a remove
18:15:26 Quit robin0800 (Success)
18:15:47gevaertsThat bit still needs to be corrected for most drivers. They now call disk_unmount(card_number), while it should be disk_unmount(card_number+first_drive)
18:15:53mcuelenaeregevaerts: there's no real 'before'; I'm just now trying to get hotswap to work
18:16:03gevaertsah, ok :)
18:16:07 Join robin0800 [0] (
18:16:17mcuelenaerebut yes, it worked before :)
18:16:42mcuelenaerethe driver currently does disk_unmount(0) & disk_mount(0)
18:17:21mcuelenaeregevaerts: where does card_number come from?
18:17:21 Join webguest67 [0] (n=5a0d85e1@gateway/web/cgi-irc/
18:17:25 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
18:18:00gevaertsmcuelenaere: in your case that's the 0. You need to correct that with int first_drive that you get from sd_num_drives()
18:18:16shotofaddsgevaerts, mcuelenaere : hotswapping seems to work on the D2 so you can follow that example (fs#10415)
18:18:35 Quit webguest67 (Client Quit)
18:18:45mcuelenaereshotofadds: thanks :)
18:18:56shotofaddsthe only problem is that the file broswer won't show any files at all if drive 0 isn't present...
18:19:06shotofaddsI assume there's a dodgy assumption somewhere
18:20:07shotofaddsalso every extra drive is named "<microSDx" but I guess that's easy enough to correct eventually
18:20:20gevaertsI wouldn't be surprised. Code that uses HAVE_HOTSWAP assumes that this is only for 1 and higher
18:20:22mcuelenaereshotofadds: why do you check for HAVE_INTERNAL_SD? The D2 doesn't have any internal SD's, does it?
18:20:30mcuelenaereSD cards*
18:20:34 Quit _zic (Remote closed the connection)
18:20:50shotofaddsthe driver is intended to work with other targets that do have internal SD, eg DAX, M200 v1
18:21:04mcuelenaereah ok
18:21:04shotofadds(ignoring the fact it's currently TCC78x specific and those are TCC77x targets)
18:21:27linuxstbbertrik: Have you looked at NAND or USB on the meizu?
18:21:31 Quit Tomers (Read error: 113 (No route to host))
18:22:01mcuelenaereshotofadds: drive 0 is present in the Onda's, but it doesn't contain any useful data (no FTL yet); does this cause problems?
18:22:01shotofaddsI got that wrong, the DAX doesn't have internal SD. just m200/c100 v1
18:22:07bertriklinuxstb, no, not seriously
18:22:17shotofaddsmcuelenaere: drive 0 is the SD card
18:22:20 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
18:22:43 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
18:22:43mcuelenaereah, ideally that should be NAND (when the FTL is figured out)
18:23:00shotofaddsright now it's ideal as-is
18:23:03pixelmateru: here was my idea for bubbles on portrait screens (just faked, though I have the background picture done already)
18:23:25gevaertsMaybe we need a way to make this order target dependent
18:23:48shotofaddssounds like a good idea
18:24:15kkurbjunpixelma: that looks really good, I like the setup
18:25:07gevaertsunfortunately I can't really think of a nice way to do it. Maybe something like #define STORAGE_PRIMARY STORAGE_NAND?
18:26:20shotofaddsI'm not sure how we'd handle that in storage.c. Maybe MrSomeone can sort it out for us?
18:26:34pixelmakkurbjun: if you like, I could send you the background picture. I had it working halfway with the vertical offset but somehow couldn't find out how to do that with the shot start position - it was either in the old place or basically offscreen. At that point I gave up
18:26:45gevaertsbasically only storage_init() needs to change
18:27:54terupixelma: looks great.
18:28:53gevaertshm, it seems that 640x480 needs a lot of new bitmaps
18:29:11 Quit pabs (Remote closed the connection)
18:29:22 Join pabs [0] (
18:32:20pixelmaI try to remember if I also used bigger bubbles for that. I remember sending the graphics to one of the Gigabeat guys who were there in the beginning until the extra gigabeat channel was turned down and a short time after that
18:32:38 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
18:33:00 Join DarkDefender [0] (
18:34:53 Quit martian67 (SendQ exceeded)
18:35:46 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
18:41:05 Quit _lifeless (Remote closed the connection)
18:43:46CIA-69New commit by Toni (r21961): correct bottom viewport height calculation in the recording screen
18:47:50 Quit shotofadds ("Leaving")
18:48:25 Quit teru ("Quit")
18:50:29 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
18:51:14 Join Toni1 [0] (n=91fec306@gateway/web/cgi-irc/
18:51:52Toni1strange, my commit did not trigger the build system
18:52:49bertrikoh yes it did
18:53:04Toni1yes, I saw it right now :)
18:56:02 Quit Toni1 (Client Quit)
18:58:52 Quit martian67 (Read error: 110 (Connection timed out))
18:59:41***Saving seen data "./dancer.seen"
19:00:33 Join Blue_Dude [0] (n=chatzill@
19:03:41Blue_DudeA proof-of-concept patch for dynamic range compression is at FS #10449. It reduces gain on the fly to prevent clipping of pre-amplified signals.
19:04:07 Join m67_l3 [0] (n=martian6@about/linux/regular/martian67)
19:12:57 Quit bmbl ("Bye!")
19:13:01 Quit faemir ("Leaving")
19:13:28 Quit martian67_ (Connection timed out)
19:14:48 Quit robin0800 ("Leaving")
19:20:17 Quit dmb ("Leaving")
19:24:28 Quit merbanan (Remote closed the connection)
19:26:50mcuelenaeregevaerts: I do this, but still the filesystem doesn't want to unmount:
19:27:49 Quit Trista689 (Remote closed the connection)
19:28:56*mcuelenaere thinks the SD thread should be unified
19:31:02 Join stripwax [0] (
19:31:26 Join Tristan [0] (
19:35:57mcuelenaerehuh? somehow the SD drive has drive number 1?
19:41:57UnhelpfullImbusMacBookPro-lImbus seems to have some problems
19:43:09 Join dmb [0] (n=Dmb@unaffiliated/dmb)
19:46:26 Join DarkSpectrum [0] (
19:49:03 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.5.1/20090715094852]")
19:49:38 Join aaron424 [0] (
19:51:33 Quit dmb (Remote closed the connection)
19:52:51aaron424How do you boot to the original firmware of the m200 v4s?
19:54:11bertrikboot by inserting the USB cable, or hold left while booting IIRC
19:54:15 Join BryanJacobs [0] (
19:54:34 Join _lifeless [0] (n=lifeless@
19:54:40 Quit BryanJacobs (Client Quit)
19:55:14aaron424that does not work. rockbox boots when I plug in the usb cable
19:56:53bertrikand the left button?
19:57:41aaron424nope. rokcbox boots pretty fast, but ive held down the left button and it does not even show text. (eg, it doesn't say it cant find the OF
19:58:43bertrikthen I don't know,maybe Domonoky or funman knows
20:00:18aaron424he's not here.. I'll see later
20:07:32aaron424do you hold down the left button before you turn it on or just after you hit power?
20:07:50 Join yziquel [0] (i=55da623a@gateway/web/freenode/x-c6ecad043a796d4f)
20:08:45 Quit mitk ("Leaving")
20:08:47 Join dys` [0] (
20:09:48yziquelhi. as i have a weird corrupted vfat filesystem, I wish to format the usb drive. how should i do this so that it remains clean / compatible with rockbox?
20:10:08linuxstbOn what device?
20:10:28bertrikaaron424, I don't have a m200v4 so I don't know. It could also be that setting the hold button (if it has one) works
20:11:23yziquellinuxstb: Olympus M:ROBE
20:11:54yziquellinuxstb: /dev/sdb
20:11:57linuxstbyziquel: Then I have no idea
20:12:10yziquellinuxstb: where can i find out?
20:12:28yziquellinuxstb: whether documentation or source code...
20:12:34linuxstbIf you wait, maybe someone here can help.
20:12:48yziquellinuxstb: i'll wait. thanks.
20:13:34linuxstbRockbox itself just needs a FAT32 filesystem. However, I don't know what the original firmware on m:robe requires, and you should keep that happy as well.
20:13:39 Quit TheSeven (Read error: 104 (Connection reset by peer))
20:14:29Unhelpfulsee if the OF has a recovery/format option, perhaps?
20:14:40 Quit dys (Connection timed out)
20:15:01yziquellinuxstb: I think i do not care that much about the original firmware anymore. But isn't VFAT better?
20:15:07 Join TheSeven [0] (
20:15:25yziquelUnhelpful: cannot start the OF anymore, anyway...
20:16:03yziquellinuxstb: or is it possible to use some ext filesystem?
20:16:50linuxstbyziquel: "vfat" just refers to the long filenames support. And no, Rockbox only supports FAT
20:17:29Mikachuyou can have vfat on any of fat12/16/32 afaik
20:18:20 Join stoffel [0] (
20:19:01 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
20:21:24yziquelso i just make one unique vfat partition?
20:22:07 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
20:22:48saratogaUbuntu seems to have some trouble with HID support on my Sansa, if I scroll the wheel the volume maxes out and then continues to flash on teh screen for a good 20-30 seconds before it stops getting the "vol+" signal
20:24:47yziquelhmmm. just checked the partitions on /dev/sdb: there's 4 partitions and they overlap (ugh!).
20:25:22saratogahuh and my e200v1 never mounts in MSC mode but my e200v2 does
20:25:26saratogai wonder whats going on
20:25:35linuxstbIt's called Ubuntu...
20:25:50 Join roolku [0] (
20:26:12linuxstbThe e200v1 not mounting sounds like the libgphoto issue - we have a wiki page about that.
20:26:26roolkuyziquel: if you are talking about the mrobe:100 - it is formatted as a superfloppy - no partitions
20:26:40saratogaI patched that bug a while ago and have not had issues since, but I will check to see if one of the newer ubuntu updates broke it again
20:27:39yziquelroolku: yes i am talking about mrobe:100. Dunno what a superfloppy is, and i therefore do wonder why i have overlapping partitions.
20:28:12yziquelroolku: so how should i format the beast?
20:28:14roolkuyziquel: your partition tool might misinterpret random data as a partition table
20:28:39yziquelroolku: my partition tool is fdisk /dev/sdb. ok.
20:29:21linuxstbroolku: Does the OF require anything on the disk to boot?
20:29:30roolkuyziquel: putting a fat16 partition on /dev/sdb should work
20:30:05yziquelroolku: mkfs -t vfat /dev/sdb ?
20:30:17saratogaI stand corrected, Ubuntu seems to have unpatched my libgphoto2
20:30:20roolkulinuxstb: IIRC only the directory structure
20:30:49linuxstbsaratoga: How very helpful ;)
20:31:13roolkuyziquel: yes
20:31:24yziquelroolku: thanks. doing it.
20:33:24yziquelroolku: mkfs.vfat: Device partition expected, not making filesystem on entire device '/dev/sdb' (use -I to override)
20:33:39yziquelroolku: so i'll override this.
20:34:31yziquelroolku: done.
20:35:16roolkuyziquel: then you need to create a hidden "System" Directory with the rockbox bootloader on it
20:35:24 Quit m67_l3 ("out")
20:35:29yziquelroolku: hidden?
20:35:36 Join m67_l3 [0] (n=martian6@about/linux/regular/martian67)
20:35:55roolkuyziquel: attribute hidden - it may work without
20:35:58yziquelroolku: are things documented from this point on?
20:37:06roolkuyziquel: I don't know it rbutil does any checks, but I think it would at least expect a System directory containing a pp5020.mi4
20:37:59 Quit aaron424 (Read error: 104 (Connection reset by peer))
20:38:53yziquelroolku: i do not understand what a hidden attribute is. it doesn't seem to be chmod dialect.
20:38:57 Quit flydutch ("/* empty */")
20:39:03saratogai'd forgotten how fast the e200v1 boots if you replace the OF bootloader
20:39:27saratogai have to hold the power button longer to shut it off then the entire boot process takes
20:39:48roolkuyziquel: no, it is a DOS/FAT attribute - there is a good chance you won't need it
20:40:32roolkuyziquel: have you tried running rbutil?
20:41:43linuxstbyziquel: I think the only way to set it on linux is via mtools (the mattrib command)
20:42:02linuxstbIt doesn't map to Unix permissions/modes
20:42:09 Nick Ypsy is now known as YpsyZNC (
20:42:24yziquelroolku: no. do not know what rbutil is.
20:42:33yziquellinuxstb: i'll have a look.
20:43:05roolkuyziquel: rockbox utility:
20:45:01yziquelroolku: i like command line installation procedures...
20:45:30yziquelroolku: but i'll have a look at packaging it for debian.
20:45:50AlexPyziquel: It is built as a static binary
20:46:16AlexPyziquel: If you want to know why people think it is pointless to package it, have a look at the dev mailing list archive this month
20:46:27roolkuyziquel: you can look at the manual install chapter in the manual
20:47:30roolkuyziquel: copy the bootloader to the System directory and to the root
20:47:50roolku... and extract to the root
20:48:12yziquelroolku: yes, i've already done this 4 months ago. i know the music.
20:48:39roolkuyziquel: and? Is it working? :)
20:48:57 Quit stripwax (Read error: 104 (Connection reset by peer))
20:49:23 Join _zic [0] (
20:50:01yziquelroolku: my girlfrind is forcing me to leave my computer and to start eating. i'll be back.
20:51:10gevaertsmcuelenaere: that does look OK to me...
20:51:35Unhelpfullinuxstb: there's a mount option that appears to be for mapping between windows and unix hidden-file conventions? it's hard to say what it does, man page just has "dots, nodots, dotsOK=[yes|no]: Various misguided attempts to force Unix or DOS conventions onto a FAT file system."
20:59:42***Saving seen data "./dancer.seen"
21:05:22 Quit antil33t ()
21:06:27 Join mc2739 [0] (
21:13:37 Join petur [50] (n=petur@rockbox/developer/petur)
21:15:14 Join stripwax [0] (
21:19:51*bertrik tries to think of a keymap for the meizu M3
21:20:05 Join at0m [0] (
21:21:03bertrikhow are touchpads currently handled? just by position of touch? or is there also some concept of sliding?
21:23:50pixelmaI don't think there is the concept of sliding... although I seem to remember that markun started some work on this for the Gigabeat F touchpad (the cross)...
21:27:33bertrikI don't even know exactly how scrollwheels are handles, I can imagine that a sliding touchpad can be made to look like a scroll wheel.
21:29:38mc2739there seems to be a problem with the sd card on the sansa e200v1 and e200v2 - with the card inserted on boot, it is not visible under files. If the card is removed and inserted, it will show up.
21:30:42AlexPbertrik: no sliding as of yet AFAIK
21:31:22AlexPbertrik: I guess we should start with a list of the essential buttons :)
21:31:56pixelmamc2739: which revision are you running?
21:32:48mc2739pixelma: r21961 from the current builds page
21:32:57mc2739(on the e200v1)
21:32:58saratogamc2739: I just confirmed on my e200v1
21:33:04bertrikI'm quite close to compiling a regular rockbox image without unresolved linker symbols, I'd love to just browse through the rockbox menus, but that needs some sort of working keymap :)
21:33:20linuxstbJust up, down, left and right...
21:33:36linuxstbThere's nothing similar you can copy?
21:33:51AlexPbertrik: I think for the menus, up and down could be touch up and down, select is middle touch, and back is <>
21:34:02bertriklinuxstb, left and right is a luxury that is not present on the meizu m3 :P
21:34:04AlexPlinuxstb: Have you seen a Mieze M3? :)
21:34:26linuxstbAh yes, I remember - the first target without left/right...
21:34:31AlexPbertrik, then we have click M and click play left to play with
21:34:39linuxstbBut no, I haven't seen one...
21:34:42saratogagevaerts: did sd work on boot up with r21933 for you?
21:34:58mc2739also, 18.11.02 # <mitk> Hi. There is problem between r21932 and r21933. uSDHC card has disappeared in Fuze
21:35:08saratogaor FS #9545 as you may remember it
21:35:11*linuxstb finds and looks for the buttons...
21:35:20AlexPlinuxstb: It has a button marked <> that is both click and touch sensitive, and a vertical pad that is position sensitive and has a click at the top and bottom as well
21:35:27 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
21:36:07AlexPlinuxstb: That is it...
21:36:07AlexPplus a hold switch
21:36:10linuxstbThen I guess the pad needs to be up/down - i.e. slide your find up for up, slide it down for down - similar to a wheel, but without the ability to carry on forever...
21:36:24*linuxstb wonders how that typo happened
21:36:52AlexPAnd I'd have touch in the middle of the pad as select, and <> as back
21:36:53linuxstbSo you have "touch", "short press" and "long press" for that <> button...
21:36:59AlexPlinuxstb: yep
21:37:21linuxstbThat's not going to be fun...
21:37:26AlexPand then long press and short press for the top and bottom of the pas
21:37:33*linuxstb also realises why there's no sim yet...
21:37:41bertrikthe OF uses M (upper button) as "back" and the play button as "OK"
21:37:44AlexPlinuxstb: And then whatever touch gestures you want to do on the pad
21:37:49pixelmaAlexP: and the center of the pad?
21:38:09AlexPpixelma: no click, just that you have touched the centre and not moved your finger
21:38:14pixelmaalthough if you do gestures that will be hard to stop at the center
21:38:21AlexPbertrik: I think M would be better as menu
21:38:34bertrikAlexP, yes
21:39:18linuxstbWhat information does the driver give you regarding touching the vertical pad? The ipod wheel driver gives a position in the range 0-95
21:39:41bertrikwe get a position 0 (bottom) to 255 (top)
21:39:46linuxstbWhich is then translated into "SCROLL_FWD" and "SCROLL_BACK" events.
21:41:30linuxstbThis is going to make games requiring 4 dimensions of real-time movement (e.g. pacbox) impossible...
21:42:01 Quit stoffel (Remote closed the connection)
21:42:24AlexPSo we have in total: touch, short click and long click on <>, short click and long click on M, short click and long click on >||, and then whatever scroll or buttons we want to make on the pad
21:42:34AlexPlinuxstb: yes, some plugins are just a no can do
21:42:45pixelmawould it be possible to distinguish between the gesture and a touch of a certain region, so you could have slides and then top, center, bottom "button"?
21:42:47Mikachuyou could map four positions to directions on the pad, like hjkl in vi :)
21:43:08AlexPpixelma: I think so, yes
21:43:18AlexP(IIUC) :)
21:43:27pixelmaplus the <>
21:43:51 Quit m67_l3 (Connection timed out)
21:43:51AlexPyeah, and then the three actual physical click regions
21:44:45pixelmahmm, doesn't sound too bad by the number, "just" positions then
21:45:12AlexPyeah, the lack of left and right are the major issue
21:46:41mc2739I confirmed e200v1 sd card missing on boot started with r21933 - works fine on r21932
21:47:24 Join SirFunk [0] (
21:48:57gevaertssaratoga: hm, no it doesn't
21:50:16*gevaerts tries to claim that r21933 could not possibly have caused that
21:50:30 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
21:55:22 Quit Horschti ("Verlassend")
21:58:12 Quit martian67_ (Connection timed out)
22:00:32gevaertshm, the big difference is that disk.c now uses storage_present() and not card_detect() anymore
22:00:53 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
22:02:48TorneHorscht: I've been experimenting with the ATA patch for ipodvideo and believe i have made it stable :)
22:04:29 Join aaron424 [0] (
22:07:09Torneer, DMA patch
22:08:21mc2739aaron424: the m200v4 booting to rockbox on usb insert is correct - the usb detect signal has not been found yet. The left key is the proper button to boot to the OF. Try pressing it before you press the power button.
22:09:58mc2739aaron424: if you hold the left button while inserting the usb cable, it should boot the OF and connect the usb
22:25:12 Join mc2739_ [0] (
22:25:55 Quit mc2739 (Nick collision from services.)
22:25:57 Nick mc2739_ is now known as mc2739 (
22:26:18 Quit aaron424 ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060308]")
22:26:30 Join aaron424 [0] (
22:28:30 Quit n1s (Read error: 110 (Connection timed out))
22:28:56 Quit aaron424 (Client Quit)
22:30:48 Join n1s [0] (n=n1s@rockbox/developer/n1s)
22:32:25 Join Strife89 [0] (n=4892d0bb@gateway/web/cgi-irc/
22:32:31 Join mc2739_ [0] (
22:33:34 Quit mc2739 (Nick collision from services.)
22:33:36 Nick mc2739_ is now known as mc2739 (
22:33:44 Quit Strife89 (Client Quit)
22:35:20 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
22:38:33 Quit saratoga ("CGI:IRC")
22:38:52 Quit martian67 (Read error: 110 (Connection timed out))
22:41:06 Join Strife89 [0] (i=4892d0bb@gateway/web/freenode/x-c5aa9d4891cd3935)
22:41:48Strife89Oy, vey ....
22:42:48Strife89So the iPod Color finally arrived. I turned it on and browsed the files before connecting it to a PC with Ubuntu 8.04.2.
22:43:14Strife89I started doing several file operations, having as many as three at once.
22:43:55Strife89Smack in the middle of the transfers (after about 20 minutes), the iPod simply stopped all disk actvity and seemed to disconnect from the computer.
22:44:21Strife89I waited, then physically unplugged it, then plugged it back in.
22:44:48Strife89I was able to browse the files, so I assumed all was well. I unmounted the iPod....
22:45:03Strife89... and was greeted with a PANIC: mount(0).
22:45:45Strife89I can no longer do anything with the iPod. No matter how I turn it on, I get the dead iPod screen and an Apple URL.
22:46:13Strife89I don't have a lot of time, but ...
22:46:19Strife89Any ideas?
22:46:48Mikachuwhat about emergency disk mode?
22:47:13Mikachu(select+play _just_ after the display powers up on my ipod)
22:47:46Strife89It doesn't seem to have any effect.
22:47:53Strife89It's an iPod Color.
22:48:02TheSevenhold reset (menu+select) until it reboots, and then immediately switch to select+play
22:48:15 Part Strife89
22:48:19*AlexP thinks that whoever's plan it was to trick BryanJacobs into doing buffering, it is working :)
22:48:24 Join Strife89 [0] (i=4892d0bb@gateway/web/freenode/x-c5aa9d4891cd3935)
22:48:38Strife89Crap, please repeat that...
22:48:51TheSeven[22:46]<TheSeven>hold reset (menu+select) until it reboots, and then immediately switch to select+play
22:49:58Strife89Doesn't appear to work.
22:50:13TheSevendoes the reset work?
22:50:31Strife89The reset works just enough.
22:50:36 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
22:50:48Strife89That is, I see the Apple logo, then the error message.
22:50:52Strife89That's it.
22:51:08TheSeventry pressing and holding the other combo as soon as the screen turns off before the apple logo comes on
22:51:43Strife89Ugh, still no dice.
22:52:29TheSeventhat just must work, it always does
22:52:46TheSeventry holding menu + left instead, that should go to diagmode
22:52:54TheSevenah i mean left + center
22:54:10Strife89THAT isn't working either. >.<
22:54:18Strife89Perhaps the HDD died?.....
22:54:37Strife89Wait, never mind.
22:54:46Strife89It works if it's the first thing I do.
22:55:26TheSevenwell the file transfers crashing could be an indication for a slowly dying HDD
22:55:34Strife89.... But selecting "DiskMode" simply brings up the error message after a moment.
22:56:04TheSeventhat looks like real HDD trouble then
22:56:47Strife89It also locks up trying to get HDD info in the Diag menus.
22:57:13gevaertsmaybe just a bad connection?
22:57:17Strife89Before displaying empty information, at least.
22:57:25 Quit mc2739 (Read error: 110 (Connection timed out))
22:57:39Strife89gevaerts: Got a page for opening the thing?
22:57:56gevaertsI'm sure there's one on
22:58:16TheSevenok, that's either the connection, or an entirely dead HDD
22:58:28 Quit martian67_ (Connection timed out)
22:59:46***Saving seen data "./dancer.seen"
23:02:13Strife89Google is not my friend today.
23:06:31Strife89TheSeven: Thanks. :)
23:08:44CIA-69New commit by stripwax (r21962): Simplify my previous change, don't need separate defines for ipod
23:08:49 Join faemir [0] (n=faemir@
23:13:54Strife89Thanks for the help, hopefully I'll get this thing fixed. ...
23:14:57 Quit _zic (Remote closed the connection)
23:15:13 Quit dash32 (Read error: 110 (Connection timed out))
23:16:00 Join dash32 [0] (
23:16:33 Quit Strife89 ("Gotta go.")
23:35:39 Join martian67_ [0] (n=martian6@about/linux/regular/martian67)
23:38:57 Quit martian67 (Connection timed out)
23:41:07yziquelroolku: everything works fine. g'night.
23:46:56 Quit martian67_ (Read error: 60 (Operation timed out))
23:50:28 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
23:51:42 Join aidy [0] (
23:57:59AlexPwhat ho
23:58:54 Join saratoga [0] (i=9803c6dd@rockbox/developer/saratoga)

Previous day | Next day