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

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

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

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

#rockbox log for 2008-12-11

00:02:01 Quit ibseco ()
00:02:07 Quit bluebrother ("leaving")
00:03:58maffek bai
00:04:07 Quit maffe ("IRC ist obsolet!")
00:10:38 Join Dvnty|CL [0] (
00:10:56Dvnty|CLanyone here ? :D
00:11:18Hillshumyou can ask a question
00:11:44alexbobpDvnty|CL: /who
00:11:51Dvnty|CLi have got the H10 6GB IRiver
00:12:27Dvnty|CLits just strange and it says there are system files missing
00:12:36Dvnty|CLis here someone who got the H10 system files ?
00:13:54alexbobpwe have rockbox system files, does that count?
00:14:06BigBambiDvnty|CL: Is this a rockbox related question?
00:14:27 Join neddy [0] (n=john@nat/sun/x-2f71553f6a5d5ac3)
00:14:48Dvnty|CLi had rockbox on my H10 but now i wanted to change it back to the normal system
00:15:05Dvnty|CLbut it doesnt start normaly - it says system files are missing
00:15:12Dvnty|CLso maybe someone also got the H10 6 GB
00:15:19Dvnty|CLand can give me those files somehow
00:15:32 Quit DataGhost (Nick collision from services.)
00:15:40 Join DataGhost [0] (n=dataghos@unaffiliated/dataghost)
00:15:55 Quit ender` (" Two things are infinite: the universe and human stupidity, even though I'm not yet sure about the universe. -- A. Einstein")
00:18:28 Join ajonat [0] (n=ajonat@
00:27:35Dvnty|CLi reinstalled rockbox but it doesnt work oO
00:29:10Dvnty|CL"System Files Missing" To restore the system files, connect H10 to recovery tool. -.-
00:29:40*Dvnty|CL drives over BigBambi
00:30:22LloreanDvnty|CL: Please, don't pester specific people about it if your question isn't answered.
00:30:44Dvnty|CLhe seemed to be active ^^
00:31:27 Quit mcuelenaere ()
00:31:29gevaertsDvnty|CL: that doesn't mean he knows anything about your issue
00:32:38BigBambiDvnty|CL: I don't have an H10, I have never used one, and I don't have the system files
00:33:09 Quit itcheg (" ajax IRC Client")
00:33:38LloreanDvnty|CL: The appropriate wiki page suggests just using the iRiver recovery tool.
00:34:04 Join soap [50] (n=soap@rockbox/staff/soap)
00:34:18 Join itcheg [0] (i=62db4767@gateway/web/ajax/
00:35:58 Quit gromit` (Read error: 110 (Connection timed out))
00:36:15***Saving seen data "./dancer.seen"
00:37:24 Quit Thundercloud (Remote closed the connection)
00:46:33 Quit faemir (Remote closed the connection)
00:47:03 Join gromit` [0] (
00:49:10 Quit itcheg (" ajax IRC Client")
00:50:41 Join MethoS- [0] (
00:51:02 Quit Dvnty|CL (Read error: 110 (Connection timed out))
00:53:03 Quit Nico_P (Remote closed the connection)
00:53:45 Quit herrwaldo ("Konversation terminated!")
00:57:07 Quit lasser (Read error: 110 (Connection timed out))
01:05:58 Quit MethoS_ (Read error: 110 (Connection timed out))
01:06:46 Join itcheg [0] (i=62db4767@gateway/web/ajax/
01:07:03 Quit n1s ()
01:10:10 Join itcheg1 [0] (i=62db4767@gateway/web/ajax/
01:10:31 Quit itcheg (Nick collision from services.)
01:10:41 Nick itcheg1 is now known as itcheg (i=62db4767@gateway/web/ajax/
01:12:25 Quit bmbl ("Woah!")
01:18:20 Quit massiveH ("Leaving")
01:30:17 Part toffe82
01:31:18 Quit gromit` (Read error: 110 (Connection timed out))
01:37:15 Quit itcheg (" ajax IRC Client")
01:38:31 Quit culture (Read error: 110 (Connection timed out))
01:40:19 Join thegeek [0] (
01:54:50 Join itcheg1 [0] (i=62db4767@gateway/web/ajax/
02:09:01 Join ebil|work [0] (n=andy@
02:11:53 Quit parafin (Read error: 60 (Operation timed out))
02:14:06 Join parafin [0] (
02:15:56 Quit HellDragon (Remote closed the connection)
02:17:49 Join HellDragon [0] (
02:27:57 Join Davide-NYC [0] (
02:28:05 Join massiveH [0] (
02:32:30Davide-NYCUI question / discussion: How does one get back to the FM Radio from the WRS on an H100?
02:35:21 Nick fxb is now known as fxb__ (
02:36:06 Quit parafin (Remote closed the connection)
02:36:09 Quit XavierGr (Nick collision from services.)
02:36:18***Saving seen data "./dancer.seen"
02:36:22 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
02:36:23Davide-NYCI got it but it's not elegant. STOP (takes you to the FM Radio Menu) and then LEFT (takes you out of the FM Radio Menu back to the FM Radio)
02:37:49UnhelpfulFS #7528 seems to already have a fix, but is still open and no fix in svn?
02:40:20 Join parafin [0] (
02:43:37rasherSeems like a prime candidate for committing
02:46:22 Part XavierGr
02:46:24 Quit ebil|work (Read error: 54 (Connection reset by peer))
02:46:51 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
02:47:19 Quit Llorean (Read error: 131 (Connection reset by peer))
02:47:45 Join Llorean [0] (
02:48:37Unhelpfulso that should be ok during freeze, then?
02:49:57LloreanIt's a bug fix, isn't it?
02:51:30Unhelpfulso it would seem. there's a lack of feedback on whether it actually solved the stated problem, but it's pretty clear that it can't make anything worse.
02:52:19 Join ebil|work [0] (n=andy@
02:52:39Unhelpfuli've got it ready to commit here, i just want to make sure i'm following the freeze rules properly
02:53:05LloreanI'd say, go ahead with it.
02:56:25 Nick pabs_ is now known as pabs (
02:56:47*rasher is surprised to see his own comment on FS #7528
03:06:23 Quit thegeek (Read error: 104 (Connection reset by peer))
03:13:44 Quit krazykit` (Remote closed the connection)
03:15:01 Quit massiveH ("Leaving")
03:15:09 Quit Davide-NYC (Read error: 104 (Connection reset by peer))
03:15:10 Join krazykit [0] (
03:22:56 Quit jhulst (Read error: 60 (Operation timed out))
03:24:27 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
03:25:35 Join massiveH [0] (
03:29:40 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
03:30:49 Quit advcomp2019 (Nick collision from services.)
03:30:52 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
03:31:56 Quit pixelma (Read error: 60 (Operation timed out))
03:40:39 Quit DataGhost (Nick collision from services.)
03:40:47 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
03:43:35 Quit amiconn (Read error: 110 (Connection timed out))
03:49:33 Join ajonat_ [0] (n=ajonat@
03:56:06 Join ebil [0] (n=andy@
04:01:37 Quit ajonat (Read error: 104 (Connection reset by peer))
04:08:03 Join gregorovius [0] (
04:09:37gregoroviusis it possible to modify the button maps in a 5th gen ipod to match the original behaviour?
04:10:32LloreanYou can compile a build with whatever button map you like.
04:11:21 Join AndyIL [0] (i=AndyI@
04:12:12gregoroviusI was hoping for a hidden option somewhere, I think compiling rockbox is out of my league right now
04:13:11 Join blkhawk- [0] (
04:13:34LloreanThere is no option to change keymappings.
04:14:09 Quit ebil|work (Read error: 110 (Connection timed out))
04:15:08soapgregorovius: considering the significant differences in the iPod stock menu structure and Rockbox's - I think you would find such a key map nearly worthless.
04:17:17gregoroviussoap: I don't see why... I'm hardly ever using the menu button now
04:20:26*soap is trying to remember the stock actions well enough to defend his claim ;)
04:21:03gregoroviusheh, well, i've only been using rockbox for a couple days, but
04:22:13gregoroviuswhen navigating the menus, <prev> does what menu does in the original firmware, <next> duplicates the center press, play takes you to the wps and menu toggles between where you are and the root menu, but I find I never use that
04:22:53gregoroviusI'd rather have menu go up one level and <prev> go to the root menu
04:22:53 Quit AndyI (Read error: 110 (Connection timed out))
04:24:16 Join __lifeless [0] (n=lifeless@
04:24:36 Quit Aurix_Lexico (Read error: 110 (Connection timed out))
04:24:45 Quit _lifeless (Remote closed the connection)
04:26:29Unhelpfulit's not *that* hard to build-your-own, really
04:28:21 Quit blkhawk (Read error: 110 (Connection timed out))
04:28:59gregoroviusit doesn't bother me *that* much either
04:29:08 Nick blkhawk- is now known as blkhawk (
04:29:17gregoroviusI was using a nokia n800 as my music player
04:30:15gregoroviusputs a lot of things in perspective
04:31:19Lloreangregorovius: Other than the original firmware, is there some reason you'd expect a button labelled "Menu" to go "back" while the button designating the back direction to take you to the main menu?
04:32:43gregoroviusyou've got a point there, Llorean... i'd say basically because <forward> doesn't act like <ok> I don't expect <back> to act as <cancel>
04:33:15LloreanForward does act like 'ok'.
04:33:29LloreanYou can enter folders with it, etc.
04:33:47gregoroviuscolor me suprised =)
04:34:11gregoroviusok, i guess i'll shut up now
04:36:20***Saving seen data "./dancer.seen"
04:38:12 Quit ebil (Read error: 110 (Connection timed out))
04:38:33n17ikh|Lappydoes the resizer preserve aspect ratio?
04:40:00n17ikh|Lappynever mind, the wiki says yes
04:40:18Unhelpful...i was answering "how does" :/
04:40:23Unhelpfulyes, it does.
04:40:57 Join tyamada [0] (
04:41:12 Quit tyamada (Client Quit)
04:41:25 Join Minthe [0] (
04:41:40n17ikh|Lappyand I can have like, cover.130x130.bmp, cover.170x170.bmp, etc, right?
04:41:47n17ikh|Lappyfor multiple WPSs
04:42:03soapyou _can_ - but you don't need that anymore.
04:42:13n17ikh|Lappywell, yeah
04:42:23n17ikh|Lappybut the resizer can only do bitmaps, right?
04:42:59n17ikh|Lappyand probably resizing a bitmap from 170x170 to something like 130x130 using what, nearest neighbor? probably makes it a little blurrier than it could be
04:43:19Unhelpfulare you using a color player?
04:43:41n17ikh|LappyI mean, I haven't tested to see how it looks yet
04:43:45 Quit MethoS- (Remote closed the connection)
04:44:19n17ikh|Lappywhat's standard procedure, resize to the largest possible square for the player and let the resizer do the work from there?
04:44:52 Quit massiveH ("Leaving")
04:46:01Unhelpfulthat should work pretty well. the downscaler is area average, and there's a linear upscaler as well, on color. they both look pretty good. the greyscale targets have the nearest-neighbor scaler, and it looks good enough, as far as having album art look good on 2-bit greyscale goes
04:46:16n17ikh|LappyI'll give it a shot
04:46:39n17ikh|Lappymake some covers with the presized one for the wps, and some just with 176x176
04:46:43n17ikh|Lappysee how they look
04:49:49n17ikh|Lappythe scaling doesn't adversely affect battery life or gapless playback or anything like that?
04:51:11Unhelpfulthere will be some cost for scaling, i'd expect it's pretty small, but nobody's run a battery bench just yet. the scaler checks quite frequently if it needs to yield to other processes, i can't get it to make playback skip, so far.
04:51:57n17ikh|Lappywell, battery life isn't a huge concern to me, as the time I use my player the most is in the car, on a car adapter
04:53:29Unhelpfulreally, it only takes a few hundredths of a second for reasonable size images, so i'm not expecting a huge drain on the battery for album art
04:57:03 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
04:57:22saratogaUnhelpful: out of curiosity, did you count how many multiplies are needed per pixel in the scalar?
04:57:27saratogajust wondering how fast it is
04:59:21 Join aliask [0] (n=chatzill@rockbox/developer/aliask)
04:59:56Unhelpfuli'd done a quick estimate in my head... it's probably wrong, though. ;)
05:04:49Unhelpfulactually counting, for RGB color, i get 12*dw*sh + 15*dw*dh
05:05:05 Quit markun (Read error: 60 (Operation timed out))
05:05:42Unhelpfuli think i know a way to knock the 12 down to 9, at the cost of taking maxmimum input size down from 2^24 to 2^12
05:06:04 Join markun [50] (n=markun@rockbox/developer/markun)
05:06:43saratogais it really 12*dw*sh and not 12*sw*sh?
05:07:31Unhelpfuli'm fairly sure. the first term is the line scaler, and its multiplications are per output pixel, and it's called per input line.
05:08:28saratogaah wouldn't have expected it to not depend on the source width
05:09:32 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
05:09:53LloreanUnhelpful: Maximum input size would be decreased to 2^12 pixels wide, or what?
05:10:49UnhelpfulLlorean: to be precise, what would really be happening is a change from a limit at sw<=2^24 && sh<=2^24 to (sw * sh) <= 2^24
05:11:26LloreanSo, 4096x4096 or dimensions that result in an equal or lesser product?
05:11:37 Quit itcheg1 (" ajax IRC Client")
05:11:45LloreanI personally think that's quite reasonable at least.
05:13:08Unhelpfulbasically, the line scaler right now returns output that's been brought back down to a 0..255 range. it wouldn't be hard to drop the repiprocal multiply from it that does that, and divide by (sw*sh) in the vertical scaler instead of by sw in the line scaler and sh in the vertical
05:14:13Unhelpfulthe setup functions would just store the divisor they need in the scaler context, instead of the reciprocal, and then the vertical scaler would multiply its divisor by that before calculating its scaling reciprocal
05:21:07 Quit ajonat_ ()
05:25:34Unhelpfulsaratoga: it's not entirely independent of the source width, only the multiply count is. adds are about three per input pixel, a few more *dw*sh, and a few per output pixel
05:28:18 Join itcheg1 [0] (i=62db4767@gateway/web/ajax/
05:38:40 Quit itcheg1 (" ajax IRC Client")
05:40:12 Quit Horscht ("IRC is just multiplayer notepad")
05:45:26 Quit perrikwp (" ajax IRC Client")
05:47:43 Quit obo (Read error: 110 (Connection timed out))
05:54:57 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
06:00:05 Quit saratoga ("CGI:IRC (EOF)")
06:08:48gregorovius2 questions: does rockbox utility always fetch the latest version? and, does the upgrade process consist on installing rockbox again, or do I have to take any precautions?
06:09:44 Join Darksair [0] (n=user@
06:10:44Lloreangregorovius: The version you get depends entirely on which option you choose.
06:11:16LloreanSome of them let you choose to get the release version or the current development version, others just get the current dev version.
06:11:36gregoroviusLlorean: I have r19368 installed, and I want a version >r19374
06:11:38LloreanAs for upgrading, you should pick one of the options that doesn't say it installs a bootloader. You don't need to reinstall that to upgrade.
06:12:02gregoroviusok, so none of the quick start options
06:12:32LloreanThe current development version is always the newest SVN revision.
06:12:49Minthequick start... maybe you can remove some rocks.
06:13:04Mintheand codecs.
06:13:11LloreanMinthe: What?
06:13:24gregoroviusok... besides that, my settings should be ok if I choose to install SVN
06:13:56Mintheless plugins faster start, not right?
06:14:01krazykitno, not right.
06:14:16krazykitthe plugins only run when you explicitly load them
06:14:23LloreanSame with codecs.
06:14:39LloreanNeither of them will affect the startup speed, and that's not the 'quick start' he was talking about anyway.
06:14:49MintheOh... Completely dynamic load.
06:15:08 Join fanta2 [0] (n=fanta2@
06:15:29Minthe#iPodLinux-podzilla2 loads all modules when startup so I thought it is the same for Rockbox sorry.
06:15:40LloreanRockbox isn't based on Linux at all
06:16:56MintheI knew. So only how to implement dynamic loading matters.
06:17:50MintheSorry time for left bye
06:17:54 Quit Minthe ("Leaving...")
06:21:36 Join gromit` [0] (
06:22:11 Part sadmac
06:36:21***Saving seen data "./dancer.seen"
06:39:17 Join Nibbler [0] (
06:52:44 Quit perrikwp (" ajax IRC Client")
06:56:06 Quit Darksair ("Emacs = ESC-Meta-Alt-Ctrl-Shift")
06:59:33 Join Darksair [0] (n=user@
07:00:31 Quit Nibbler (Read error: 110 (Connection timed out))
07:02:41fanta2Anybody able to answer a question about the USB stack?
07:03:44 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
07:03:58saratogayou probably want gevarets unless its a simple question
07:05:48fanta2yeah, probably; i've got a strange but reproducable problem with sending data over usb where the entire thread halts until I disconnect, but I don't know if it's a bug or my bad programming skills :P
07:06:02saratogawhich player?
07:07:54 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
07:08:13fanta2It's the Sansa e200 (with e200R firmware)
07:08:36fanta2I'm sending a large amount of data over the usb stack; split into seperate packets
07:09:03saratogaUSB on the sansas has various issues
07:09:09fanta2then I send a response packet to say I'm done, and that last usb_drv_send halts
07:09:10saratogabut you probably know that
07:09:34 Nick flx is now known as flux (
07:12:36fanta2yeah, but it's been surprisingly stable for me up until now
07:13:32 Quit saratoga ("CGI:IRC (EOF)")
07:18:24 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
07:18:48 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
07:20:02 Join bertrik [0] (
07:20:41 Quit JdGordon (Remote closed the connection)
07:22:55 Join JdGordon [0] (
07:25:43 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
07:31:39 Quit gregorovius ()
07:35:01 Join DA-Laptop [0] (
07:36:11DA-LaptopiPod Video should work with Rockbox, right? I'm looking to buy a new player
07:36:27aliaskNot the new ones
07:36:39scorcheall devices rockbox works with are on the front page
07:36:57DA-LaptopThat's the one I'm looking at
07:38:47DA-LaptopAlternatively, I'm also looking at Cowon D2
07:38:51DA-LaptopNot sure which is better
07:40:55scorchewell, if you are looking for a rockbox target, you dont have much choice between the 2 since only one of those is fully supported...
07:43:10DA-LaptopI want something that has the potential for a lot of storage, a large screen and preferably, good battery life
07:43:46DA-LaptopMy music collection is in FLAC, hence the need for a lot of storage. I like music videos, plus watching TV shows, hence the screen. and battery life..well, that speaks for itself.
07:44:23scorchevideo on the ipod 5G in rockbox isnt really what i would call "nice"
07:44:33DA-Laptopoh.. what's wrong?
07:44:35scorcheif i were you, i would look at something liek the gigabeat F
07:44:40DA-LaptopI have an F20
07:44:42Unhelpfulno, but video int the OF is not bad
07:44:46scorcheit has a slow processor and a large screen to update
07:45:29DA-Laptopscorche: My F20 is barelly still alive. Tends to shut down on its own randomly, the audio connector is so bad that I constantly have to move the earphone plug around just to get sound on both phones, and the battery life is down to about 10hrs
07:45:38 Quit fanta2 (Read error: 110 (Connection timed out))
07:45:40DA-LaptopI want something different this time around
07:46:02scorchei would much prefer a gigabeat F to an ipod 5th gen...
07:46:37LloreanIn terms of playback performance the Gigabeat F far outclasses the iPod Video.
07:46:52LloreanThe only real advantage the iPod Video has is being slightly smaller.
07:48:13DA-LaptopNot much of a choice, eh..
07:56:15 Join Nibbler [0] (
07:56:42 Quit bertrik (Remote closed the connection)
07:59:43 Quit amiconn (Remote closed the connection)
08:01:37 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
08:01:37 Join pixelma [0] (n=pixelma@rockbox/staff/pixelma)
08:02:41DA-Laptophow large can SD, SDHC or MMC cards be?
08:03:12scorchethat isnt really rockbox related...
08:03:32B4gderI'm sure wikipedia has all those answers
08:03:36LloreanYeah, their limits really aren't something under our control.
08:05:26 Quit markun (Remote closed the connection)
08:10:54 Join markun [50] (n=markun@rockbox/developer/markun)
08:29:48 Quit BigBambi (Read error: 148 (No route to host))
08:34:39 Join petur [50] (n=petur@rockbox/developer/petur)
08:35:46 Join Rob2223 [0] (
08:36:24***Saving seen data "./dancer.seen"
08:37:08 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:48:10 Quit petur (Read error: 145 (Connection timed out))
08:53:47 Quit Rob2222 (Read error: 110 (Connection timed out))
08:59:27 Join Darksair` [0] (n=user@
09:00:25 Quit JdGordon (Read error: 104 (Connection reset by peer))
09:05:03 Quit Darksair (Read error: 110 (Connection timed out))
09:08:43 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
09:09:08 Join ender` [0] (
09:15:52 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
09:18:11 Join ebil|work [0] (
09:20:04 Join petur [50] (n=petur@rockbox/developer/petur)
09:28:21 Quit jhulst (Read error: 148 (No route to host))
09:34:20 Join lasser [0] (
09:35:38 Quit Darksair` (Connection timed out)
09:45:03 Join Thundercloud [0] (
10:00:04 Quit Nibbler (Read error: 110 (Connection timed out))
10:03:07 Quit Seed ("cu, Andre")
10:06:48 Quit ebil|work (Read error: 145 (Connection timed out))
10:08:42 Quit lasser (Read error: 60 (Operation timed out))
10:17:08*GodEater posts for IRC log posterity : Rockbox git repo is now available over http too : "git clone"
10:17:27GodEatermany thanks to Zagor for sacrificing a few more head hairs to making it possible
10:20:57 Join tkooda [0] (n=tkooda@
10:23:12tkoodaanyone got any comments on running rb-v3.0 vs. rb-r19389 on a Toshiba Gigabeat F40? -I get an almost garunteed freeze when removing+restoring usb power in my car
10:23:19tkooda(using v3.0)
10:26:59LloreanDo you hold down "Menu" when attaching USB power?
10:27:03 Quit Thundercloud (Remote closed the connection)
10:27:30LloreanI don't believe it can distinguish a USB charger from an actual USB connection.
10:36:27***Saving seen data "./dancer.seen"
10:40:08 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
10:42:30tkoodait works fine and is charging while I'm playing. with an older (~1+ year ago?) pre-3.0-revision of rb it worked fine (restoration of usb power while in "car mode" would do the ~10 second rewind (as configured) and resume playing just fine)
10:43:02tkooda..but now, restoring usb power freezes it to the point of needing to flip the battery switch
10:43:11tkooda(can't even power down with power button)
10:43:29Lloreanhave you tried it with a normal charger?
10:43:39LloreanSpecifically, are you sure it's USB related, and not related to the rewind?
10:44:16tkoodathe "normal" (non-usb) charger has (as expected) zero effect on the usb remove+restore lockup
10:44:35tkooda(didn't mean to sound like a jerk there, heh)
10:45:03LloreanAnd it's rewinding properly?
10:45:27tkoodaremoving+restoring "normal" power doesn't perform the pause that the "car mode"+usb_power does though
10:45:39LloreanDo you have car mode enabled?
10:45:42tkoodanormal power doesn't incur the rewind, IIRC
10:45:47LloreanNormal power should trigger car mode exactly like USB power.
10:45:48tkoodawill test again now
10:45:57tkoodahrmm, I guess that does make sense
10:46:03LloreanThat's why I asked you if it relates to the rewind...
10:47:00 Quit rvvs89 (Read error: 60 (Operation timed out))
10:49:07 Join AndyI [0] (i=AndyI@
10:49:19 Quit DA-Laptop (Remote closed the connection)
10:56:40 Join robin0800 [0] (
10:57:18tkoodaunplugging normal power successfully pauses, but doesn't rewind+resume on restore, which is fine with me as long as it doesn't freeze. -usb remove+restore freezes in my car. (was digging trying to find a usb power source that won't attempt to mount it as a mass storage device; think I've got one somewhere)
10:58:13 Join lasser [0] (
10:58:17tkoodathink the next step is to confirm crash using AC->usb power source (not laptop + usb mass storage mount), to re-create situation when I'm in my car with car->usb power adapter
11:00:31 Quit AndyIL (Read error: 110 (Connection timed out))
11:00:50 Join AndyIL [0] (i=AndyI@
11:02:54Lloreantkooda: I thought you said "yes" to holding "menu"
11:03:01LloreanIf you were holding menu, even your PC won't try to mount
11:05:30LloreanBut car mode should work with the mains power.
11:05:41LloreanI don't see a rewind option though. Where is that in the menus?
11:10:05 Quit AndyI (Read error: 110 (Connection timed out))
11:10:28Lloreantkooda: You there?
11:10:48tkoodain my car when I experience the freeze when I turn my car back on after filling up with gas: 1) power on, 2) hold menu and plugin car adapter -> usb power, 3) play audio, 4) turn off car, cutting usb power and pausing play, 5) turn car on, restoring usb power.
11:11:01tkooda..and that's when the device becomes unresponsive
11:11:12LloreanSo you didn't hold Menu before/during #5
11:11:25tkoodanope. I didn't have to before v3.0?
11:11:25LloreanYou need to then too. Every time it receives USB power you should be holding down Menu when it happens.
11:11:34LloreanYou were lucky before 3.0 then.
11:11:44tkoodaI liked that "feature"
11:11:45LloreanThe player is not able to distinguish between USB power and true USB.
11:11:52LloreanIt's not a feature, this is a hardware limitation.
11:12:08LloreanYou need to "tell" it it's not got a computer at the other end of the cable by holding down Menu.
11:12:21LloreanAnd, where do you configure the rewind time?
11:12:24tkoodathe rb software detects the menu button and does something differently.. can't I just have rb do that* action all the time by default?
11:12:43LloreanIf you compile a build with USB set to default to power, and requiring you to hold menu to connect to a PC, sure.
11:12:58tkoodaheh. no menu option?
11:12:59LloreanAgain: Where do you configure the rewind time? I don't see such an option alongside the car adapter mode option.
11:13:15LloreanYou said "as configured" so I assume you configured a time.
11:13:31LloreanAre you sure you weren't possibly using a patched build previously?
11:13:36tkoodaI recall it being a selection of a list of times (like the display sleep times). IIRC
11:14:03tkoodano. zip from, then 3.0 zip
11:15:05 Join ender1 [0] (
11:15:19Lloreantkooda: Well, can you find the option you said you were using then?
11:15:22LloreanI'd like to test it myself.
11:15:38tkooda[04:13] <tkooda> looking
11:16:43LloreanThere's a patch specifically to always us USB power when in Car Adapter mode.
11:16:58tkoodadon't see it now. I actually only (thought I??) remembered using it with my previous F40 using pre-3.0..
11:17:09LloreanWhich would explain the behaviour you describe rather well.
11:17:21LloreanIt's in the tracker, task 6654
11:17:26tkoodak, thanks
11:18:10 Quit ender` (Read error: 60 (Operation timed out))
11:20:01tkoodahrmm.. I think it'd be nicer if that patch simply inverted* the default action, rather than requiring you to disable car mode entirely in order to mount as mass storage
11:21:20LloreanComments about the patch should go there.
11:26:32tkooda solves what I'm looking for
11:28:09 Join Darksair [0] (n=user@
11:28:38tkoodaone byte change:
11:28:52 Join DerDome [0] (n=DerDome@
11:39:17 Quit tvelocity ("Αποχώρησε")
11:44:42 Join Llorean1 [0] (
11:45:28 Part _Auron_
11:48:59 Quit Darksair (Connection timed out)
11:50:56 Quit __lifeless (Read error: 145 (Connection timed out))
11:51:34 Join n1s [0] (n=nils@rockbox/developer/n1s)
11:55:22 Join |MegafEee| [0] (n=Linux@
11:55:29 Quit Llorean (Nick collision from services.)
11:55:33 Nick Llorean1 is now known as Llorean (
11:56:27 Quit |MegafEee| (Remote closed the connection)
11:58:33 Join PaulJam [0] (
11:58:42 Join _lifeless [0] (n=lifeless@
11:59:20 Quit robin0800 (Read error: 54 (Connection reset by peer))
12:09:16 Join funman [0] (n=fun@rockbox/developer/funman)
12:15:27 Join moos [0] (
12:24:25 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
12:28:39 Join Darksair [0] (n=user@
12:29:46 Join TheSphinX^ [0] (
12:31:12soapAm I the only one who, for some hard to define reason, has problems with the "Repairing broken players" sub forum and wish it wasn't nested?
12:31:22 Quit linuxstb (Read error: 145 (Connection timed out))
12:31:28soapI am constantly passing over it unintentionally.
12:36:28***Saving seen data "./dancer.seen"
12:38:53 Quit AndyIL ()
12:40:10 Join culture [0] (
12:47:47 Join AndyI [0] (i=AndyI@
12:49:55 Join Aurix_Lexico [0] (n=comrade@
12:50:54 Join Seed [0] (
13:02:13 Join MethoS- [0] (
13:07:43 Join countrymonkey [0] (n=4b05639a@gateway/web/cgi-irc/
13:08:25 Quit countrymonkey (Client Quit)
13:09:00 Join webguest82 [0] (n=4b05639a@gateway/web/cgi-irc/
13:09:15webguest82Where can I download French voices?
13:13:38dionoeayou'll have to generate them yourself
13:13:49dionoeaas far as I know you can do the generation in the rockboxutility program
13:14:51webguest82I don't have a French voice.
13:15:37webguest82The closest I have to a french voice is what comes with my screen reader, which is proprietary and not sapi complient.
13:16:52 Quit lasser (Remote closed the connection)
13:17:13webguest82Has anyone made any downloadable ones?
13:18:41funmanany one could forward me the as3525 datasheet? I want to check the JTAG header on the Clip
13:21:09funmanB4gder: thanks, i got it
13:27:40pixelmawebguest82: I could install espeak+mbrola tools in windows which someohow integrates with SAPI, espeak provides free voices. Found all needed info there but don't remember much as it's quite a while ago
13:28:50webguest82Thanks, but no thanks. I had bad experiences with espeak-zh, and am never touching it again; I couldn't understand what it was saying, even though I am a native speaker!
13:30:07gevaertswebguest82: mbrola is very much better than the default espeak engine
13:31:56pixelmait also depends on the voice itself. I could understand the swedish voice quite well although I am not a native speaker and the voice is a little bit noisy but ok
13:31:56webguest82Never could figure out how to get it to work.
13:32:24 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
13:32:37 Join J-23_ [0] (
13:33:04webguest82Could someone help me with mbrola?
13:33:05 Quit J-23 (Nick collision from services.)
13:33:07 Nick J-23_ is now known as J-23 (
13:33:50LambdaCalculus37webguest82: This channel is about Rockbox. What does mbrola have to do with Rockbox?
13:34:11 Quit webguest82 ("CGI:IRC")
13:34:18 Join webguest74 [0] (n=4b05639a@gateway/web/cgi-irc/
13:34:21 Quit webguest74 (Client Quit)
13:34:34 Quit Darksair (Read error: 104 (Connection reset by peer))
13:34:54 Join webguest82 [0] (n=4b05639a@gateway/web/cgi-irc/
13:35:04webguest82I was asking about french voices.
13:35:24webguest82My browser closed the window on me, did anyone say anything?
13:35:47scorchewebguest82: as we have told you many times in the past, you can look at for the logs
13:36:35 Quit TheSphinX^ ("XChat@Linux")
13:37:43webguest82Could someone help me with Mbrola? I tried to do it once, never got it to work.
13:42:47 Quit webguest82 ("CGI:IRC (EOF)")
13:44:22 Quit Aurix_Lexico ("Leaving.")
13:50:35 Quit LambdaCalculus37 ("Ka-chunka")
13:55:20 Join itcheg1 [0] (i=62db4767@gateway/web/ajax/
13:55:53 Quit DerDome ("Leaving.")
14:04:09 Join lasser [0] (
14:04:52 Quit itcheg1 (" ajax IRC Client")
14:04:54 Join mofux [0] (
14:06:17funmando we have an idea of the position of the AS3525 BGA on the Clip ?
14:06:29funmani.e. where is the pin A1 on the PCB
14:06:37 Join minikillan [0] (
14:06:46minikillanhi guys
14:06:55funmani can't make the connections to the JTAG header 'match'
14:08:00minikillananyone who knows anything about the rockbox for iAudio 7 project?
14:08:15funmanthe top square (when looking at the front of the pcb, hold/power button at the top) is linked to this jtag header, and if the top pin is A1, that means JTAG_TMS, JTAG_TDI, JTAG_TCK, and gpio C7
14:08:27B4gderminikillan: I don't think there's a lot to know
14:08:28funmanminikillan: just look at the status wiki page (linked on the website)
14:08:47scorcheas well as the forum thread on the subject
14:10:35minikillan(thanks) :)
14:13:55 Join DerDome [0] (
14:14:30 Join massiveH [0] (
14:16:04 Nick funman is now known as Unhelpf (n=fun@rockbox/developer/funman)
14:16:10 Nick Unhelpf is now known as funman (n=fun@rockbox/developer/funman)
14:16:40 Join fredddy [0] (
14:20:55 Quit culture (Read error: 110 (Connection timed out))
14:22:53pixelmahmm, is the new bmp scaler supposed to ba able to scale up?
14:23:48 Join Darksair [0] (n=user@
14:31:00 Nick Bensawsome is now known as NinJew (n=Bensawso@unaffiliated/bensawsome)
14:31:04funmanfredddy: hi, you have a fuze don't you? would you test some code on it?
14:31:29 Nick NinJew is now known as Bensawsome (n=Bensawso@unaffiliated/bensawsome)
14:36:31***Saving seen data "./dancer.seen"
14:38:01fredddyfunman: yes
14:39:58funmando you already have the bootloader installed? I'm testing if the patch builds and i'll paste it
14:40:18fredddyand could you take a look at FS9617? With this fixes all plugins would compile and work fine in the sim
14:41:20fredddyI have a bootloader but its a bit older I think
14:41:23 Join lazka [0] (
14:43:41lazkasuggestion: make versioning of rbutilqt like rockbox -> 3.1. sync releases, so every new port has install support in every release.
14:44:21 Quit minikillan ()
14:44:49lazkaa friend of mine was confused by the seperation of the installer / build.. i think that this might help.
14:45:13funmanfredddy: just build as normal firmware and keep your old bootloader. I added another comment on FS #9617 (by the way if you use 'FS#' a direct link to the task is made on the irc log)
14:45:50 Part lazka ("cya")
14:46:03Zagorlazka: but that would mean we cannot release a new installer until the next rockbox release. because surely a 3.2 or 3.1.1 installer that installs 3.1 would be just as confusing?
14:47:17funmanfredddy: it tests 2 things: 1/ if the tuner is detected (if it's not you'll have to remove the panicf() from firmware/drivers/tuner/si4700.c) and 2/ reads data from a mysterious device (which might be the buttons)
14:47:37 Join lazka [0] (
14:48:05funmanfredddy: oh I forgot something .. please add "lcd_udpate();" in apps/main.c just after the lcd_puts(); (search for "fuze" to get the exact location)
14:48:10 Part B4gder
14:48:24Zagorlazka: but that would mean we cannot release a new installer until the next rockbox release. because surely a 3.2 or 3.1.1 installer that installs 3.1 would be just as confusing?
14:49:05lazkaZagor, I don't see a problem with 3.1.1 ... maybe
14:49:23funmanor '3.1.a' ?
14:49:29lazkajust a suggestion.. I don't know the rb development process exactly
14:49:46fredddyhmm svn dosen`t compile a fuze build (without the patch from funman)
14:50:16funmanfredddy: i know i just noticed (the fuze config defines si4700 radio present but wasn't tested, so now that the radio driver is present it doesn't build)
14:50:18scorchelazka: the installer is just the can choose which build you want to install from it...
14:50:55 Join LambdaCalculus37 [0] (i=44a04303@rockbox/staff/LambdaCalculus37)
14:51:01funmanoh? I thought it only used downloaded builds (i never used rbutil)
14:51:18 Join itcheg1 [0] (i=41d59de2@gateway/web/ajax/
14:51:32scorchefunman: well, i mean from release, or latest, etc
14:52:30funmanfredddy: if the tuner is detected i will commit the missing part for FM
14:54:11funmanJdGordon: do you have time to test this patch on your e200v2 as well?
14:54:17 Quit reacocard (Read error: 104 (Connection reset by peer))
14:54:26 Quit Bensawsome ("The awsome is gone :(")
14:55:34fredddyfunman: It says no si4700 tuner
14:56:03 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
14:56:06LambdaCalculus37funman: The build you sent me yesterday is starting to give me errors.
14:56:08 Quit itcheg1 (Client Quit)
14:56:46LambdaCalculus37I was getting *PANIC* SD TIMEOUT errors from time to time on my Clip.
14:57:01funmanfredddy: alright, just remove the else panicf(...) in firmware/drivers/tuner/si4700.c then, and add the missing "lcd_update()" to apps/main.c
14:57:20*LambdaCalculus37 just got that error message again
14:57:29LambdaCalculus37*PANIC* SD : DATA TIMEOUT,
14:57:30funmanLambdaCalculus37: I met this problem as well, we need to add proper error handling to the SD driver
14:57:55LambdaCalculus37funman: Strange, though... I was using my Clip for about two hours last night without a single problem.
14:58:09LambdaCalculus37I was able to get through three albums and some FM radio listening without a hitch.
14:58:19funmanit's still unstable, and errors in SD controller are unpredictable
14:59:17LambdaCalculus37I know this is still unstable code, so I know there will be errors and panics from time to time. It's just a little funny how it works perfectly for a bit then starts doing this.
14:59:52funmanI feel the same, just estimate ourselves lucky when it works ;)
15:00:18funmanI think the proper way is to mark transfered sectors as bad when an error happens and transfer them again
15:00:56funmanBut i'm not sure of the execution flow between the sd thread, the dma interrupt which marks the end of a transfer, and the sd interrupt which says a problem happened in a transfer
15:01:53funmanWe should interrupt on data_block_end to be sure that a block has been transfered safely
15:01:56fredddyfunman: where does the lcd_update go ?
15:02:42funmanfredddy: after lcd_puts() in apps/main.c (line 145, between "fuze hack" comments)
15:03:07 Join Nic0_P [0] (
15:05:38 Quit J-23 ("THIS IS ZNC")
15:06:06 Join J-23 [0] (n=zelazko@
15:06:35 Join bapdog [0] (
15:06:36 Quit Nico_P (Nick collision from services.)
15:06:43 Nick Nic0_P is now known as Nico_P (
15:07:48fredddyfunman: I get a grey screen now on startup
15:08:07funmannothing printing ? :/
15:08:22funmanwhat is the last revision you had built before?
15:08:52Nico_PZagor: here?
15:09:16fredddyno I see the rockbox logo (the one from the bin file) and then I get a grey screen
15:09:24Nico_PWhat would you think of installing the PluginTree for TWiki?
15:10:31fredddyI got to the menu with an svn version without the tuner
15:11:31funmanother changes have happened before the tuner was committed, don't you remember the exact revision / or the approximate date?
15:12:05ZagorNico_P: I think it looks like a good tool. possibly a little taxing for the server though.
15:12:17 Quit bmbl (Connection timed out)
15:12:36Nico_PI think it would help us get more organized
15:12:39Zagorwith some caching it could help make the wiki more fathomable (?)
15:13:14ZagorI plan to await the release of FosWiki before doing anything to the twiki installation though,
15:13:31Nico_Phow long will that be?
15:13:59Zagorleast I heard they were aiming for a release this year.
15:15:21fredddyI don`t now wich version it was sorry
15:16:05Nico_PZagor: well, fair enough
15:16:23funmanfredddy: ok, can you svn up and see if the builds display the menu ?
15:16:40funmani committed r19390 and r19391 to fix the build
15:17:09fredddyfunman: If i don`t call the button_i2c_init(); Funktion its okay
15:17:46 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
15:17:50fredddyit prints FFFFFFF
15:17:59 Join ebil|work [0] (n=andy@
15:19:23fredddyno its FFFFFFFF backlight turns off after a few seconds
15:19:51 Quit Darksair ("Emacs = ESC-Meta-Alt-Ctrl-Shift")
15:20:24funmanhum we need to disable backlight turning off
15:21:02funmanwell, this FFFFFFFF means that reading failed, so no need to go further :/
15:21:15 Quit Horscht (Client Quit)
15:21:49 Join Schmogel [0] (
15:22:33funmanfredddy: let me re read the init function perhaps i did something wrong here
15:22:43fredddyIt has to fail since i turned off the init function
15:22:57 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
15:23:46 Quit DerDome ("Leaving.")
15:24:22funmanright, but the problem is not in the init function (which doesn't use hardware functions) but in the other code (which return -1 if init hasn't been called, or display a grey screen)
15:24:45fredddyit works turning off the backlight in config
15:25:09 Part lazka ("cya")
15:25:14 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
15:25:23fredddyshould I put an panicf in every function and tell you wich one doesn`t work ?
15:26:03funmannope, the functions are too simple and have no checks
15:27:29funmancan you try this new patch : ? it uses another GPIO pin (perhaps the one in the first patch relates to LCD)
15:27:52funmanthis one applies on a fresh svn cechkout
15:32:17 Join aneqrs [0] (
15:34:13fredddyhm it gives some HUNK #1 failed can I ignore them ?
15:35:39funmanfirst use "svn revert -R . ; svn up ; rm firmware/target/arm/as3525/button-e200v2-fuze.c" and apply the patch again, it shouldn't have failed. if it does again please paste the patch output
15:35:49 Join mc2739 [0] (
15:36:11funmanhello mc2739
15:37:43funmandid you try to access the FM chip using i2c on your e200v2 ?
15:38:14mc2739yes, but no sound yet
15:38:50 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
15:38:51funmanbut could you at least get access to i2c registers ? (verify the chip ID for example)
15:39:07funmanthe FM chip could be on line1 (like e200v1) or line2 (like clip & m200v4)
15:39:37mc2739I'm not sure about i2c registers since there is no working buttons yet
15:40:10 Join faemir [0] (
15:40:14funmani'd just put panicf() / lcd_puts() in firmware/drivers/tuner/si4700.c
15:40:20mc2739but I used an nvram.bin from my e200v1 so that it was on the correct frequency for one of the local stations
15:40:40mc2739ok, I'll try that
15:40:43funmanby the way we found i2c communication on a mysterious device which might be button or wheel on fuze
15:40:58funmansince the gpio mappings are exactly the same so far on e200v2 and fuze, it might be the same
15:41:20mc2739nice - is that in the patch I saw in the logs?
15:41:40funmanyes, but i didn't verify if the PIN mapping was the same for e200v2
15:42:49fredddy hier the output it doesn`t patch
15:42:55bapdogi noticed that quite a few of the supported rockbox targets don't have USB. can that be fixed?
15:43:12n1sbapdog: yes it can but it's not easy
15:43:13funmanbapdog: sure it can, if someone does fix it.
15:43:31bapdogso we have to dual boot them to load files?
15:43:48funmanfredddy: can you paste apps/main.c.rej please?
15:43:59 Join MethoS-- [0] (
15:44:43n1sbapdog: yes, or use memory cards with players who has slots for that
15:44:46bapdogcos i see that the e280 v2 doesn't have USB working yet, can we assume that it isn't going to work
15:45:08gevaertse280 v2 isn't a supported rockbox target
15:45:17n1sbapdog: no one can say if USB on that will ever work
15:45:23bapdogi thought it's in vevelopment
15:45:38funmanfredddy: and apps/main.c now please?
15:45:42gevaertsin development is a *very* different thing than supported
15:45:45pixelmahmm... looks like some change since last Monday broke scrolling in conditional viewports (my first impression, need to invesigate further)
15:45:58n1sbapdog: it is, but that doesn't mean that it necessarily will be finished
15:46:16bapdogi know that, but i'd hope you wouldn't be attempting such a task without *some* hope
15:46:40J-23what about >1GB?
15:47:15bapdogi'm not a hardware person, but i'd like to help out if I can. i did a few of my own mods around version 2.2
15:47:22n1sbapdog: hope doesn't finish ports, hard working devs do, they are however still pretty bad at predicting the future
15:48:05bapdogi have a couple of archos recorders that are getting a bit shaky
15:48:27LambdaCalculus37bapdog: There were ports started that never got finished. The iriver iFP series has a partial port that's currently stalled, and another was for the Archos Gmini devices, which was finally stopped and pulled from the source tree when the only developer working on it was forced to stop.
15:48:31pixelmaI have a WPS which worked fine wrt scrolling last week, updated build today and scrolling only works in some of them and if I see correctly, the ones where scrolling works are fixed ones and in preloaded/conditional viewports it doesn't
15:48:32bapdogi know things have moved on a lot
15:49:03funmanLambdaCalculus37: 'forced to stop' as in legally, or as in he choose to stop his work ?
15:49:22bapdogand i know if i read the mailing lists and the archives i'd know all this, so i hope you don't mind me asking
15:49:41pixelmato add: in 2 sims, should try on target
15:49:55 Quit tchan ("WeeChat 0.2.7-dev")
15:49:59funmanfredddy: everything looks alright, did you check if the patch itself had correct unix line endings ? (you can convert it with fromdos or gedit for example)
15:50:09 Join tchan [0] (
15:50:44bapdogi want an e280, and I saw some v1 units for sale on ebay, I guess it's not worth the gamble buying a v2 and crossing fingers
15:53:09LambdaCalculus37funman: "Forced to stop" as in he had a bunch of real-life problems that got in the way, and required his focus much more.
15:56:28funmanmc2739: at a first sight e200v2 seems to use the same pins than fuze once again (B1 for clock, A0 or A5 for data)
15:56:36fredddyfunman converting to unix worked
15:58:03 Quit mofux (Remote closed the connection)
15:58:09 Quit MethoS- (Read error: 113 (No route to host))
15:58:17mc2739funman: ok - building now
15:59:02funmanmaybe i need to properly analyse the delay and use a correct function instead of reusing the fmradio delay
16:01:08mc2739funman: is this a problem? si4700.c:54: implicit declaration of function 'panicf'
16:02:26funmanmc2739: nope, but i disabled tuner on e200v2/fuze so it won't be called anyway (detection apparently failed on fuze)
16:04:19 Join obo [0] (n=obo@rockbox/developer/obo)
16:05:25 Join mofux [0] (
16:05:54 Join Zoko007 [0] (n=zoko007@
16:07:58 Part Zoko007
16:10:26mc2739funman: how about this one?
16:11:04 Quit mofux (Remote closed the connection)
16:11:13fredddyfunman, READ stays 0 doesn`t matter wich button I press
16:11:19funmanmc2739: button should be unsigned char, but that's not a real problem for now
16:11:33funmanfredddy: and what about wheel scrolling?
16:11:52funmanthe OF doesn't actively poll the value, but only does so when a change on pin B0 is detected
16:12:16funmanyou might want to add a panicf() if the read value is != 0 to detect a change
16:13:37fredddyso every button sends an interrupt to B0 ?
16:14:14funmanwe're not even sure it's buttons or wheel (but probably one of them)
16:17:03fredddycould we use kugels button patch an check all pins in INT_GPIOB?
16:17:13funmanfredddy: if you replace #define SDA 0 by #define SDA 5 in firmware/target/arm/as3525/button-xx.c : do you get again a grey screen?
16:18:11funmanfredddy: afaiu the OF only use this i2c communication when a change in B0 happens, kugel might help here because i don't know how interrupts on GPIO work (register settings etc)
16:20:34fredddyyes I get a grey screen again
16:21:54funmanok, if you replace the while(1) loop by a int x;for(x=0; x< 20; x++) loop, does it show a normal screen after 2 seconds?
16:23:48pixelmanot scrolling on target either :\
16:24:49fredddyit goes to main screen
16:25:33funmani understand than pin A5 is related to LCD but i'm not sure
16:25:38 Join kugel [0] (n=chatzill@unaffiliated/kugel)
16:25:42 Join mofux [0] (
16:26:25 Quit perrikwp (" ajax IRC Client")
16:27:45funmankugel: why do you think button_i2c_read() is weird?
16:27:47kugelteleporting is nice
16:28:11kugelbutton is 0, and you return button (i.e. 0) if ret is 0
16:28:34kugel"return ret ? ret : button;"
16:28:59kugelah, I overlooked the &button
16:30:09 Part LinusN
16:31:45kugelfunman: nevermind then. I'm gonna put up a some modifications of the lcd_driver. a) remove lcd_delay b) add a write another lcd_write_cmd i found in disassembly, c) modify lcd_enable a bit, and d) little fixes in lcd_update_rect
16:32:08kugelWho would be the person who would take a look at it before committing?
16:33:31kugelfunman: btw: is the diff the button driver you (think you) found in the disassembly?
16:35:28kugelanother question: I'm reading "detection of the fm tuner has failed", but I apparently overlooked what made you think so. The best way to tell if a tuner is detected is to look at the main menu. The radio item will only be there if the radio is detected at runtime
16:36:32***Saving seen data "./dancer.seen"
16:38:57funmankugel: i can commit if you say it's ok (also if you could have a look at a lcd fix on e200v2 on flyspray that would be great)
16:39:09funmanmc2739: I think you are the author of this patch?
16:39:19funmankugel: yes it's (partial) disassembly
16:39:33mc2739yes, I did that one
16:39:59funmankugel: just read the recent backlog, i have put a panicf() is si4700 init failed and fredddy confirmed the panic
16:40:37mc2739I also saw the same panic - "no si4700 tuner"
16:40:49kugelfunman: without buttons it's a bit hard to say if it's "ok". But the display behaves definitely saner with my changes
16:43:49 Join jfc^3 [0] (
16:43:57kugelbecause I cannot really force lcd updates
16:47:57mc2739kugel: I used autostart.rock to test lcd_update_rect - I had to modify apps/main.c to activate the autostart feature
16:50:21 Quit aneqrs ()
16:55:13 Quit faemir (Remote closed the connection)
16:59:02 Quit Zagor ("Client exiting")
17:00:00 Quit jfc (Read error: 110 (Connection timed out))
17:04:57 Join faemir [0] (
17:09:46 Quit moos ("Rockbox rules the DAP world")
17:11:50funmankugel: btw when i say partial disassembly, i overlooked some functions which seemed related to OF operating system, and not directly related to i2c over GPIO, so it *should* be complete
17:12:12 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
17:13:01 Join robin0800 [0] (
17:13:12 Quit Hillshum ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
17:13:49 Quit faemir (Remote closed the connection)
17:15:36 Quit aliask ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111318]")
17:20:31 Quit perrikwp (" ajax IRC Client")
17:21:15 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
17:23:27kugelfunman: but it doesn
17:23:30kugel't work?
17:23:57funmanwhat doesn't work ?
17:24:32funmani don't know, you are responsible for testing and interpreting ^^
17:24:52kugelI just thought because someone already tested today
17:24:54funmanperhaps you missed that : this code is only ran when a change on B0 is detected (it's run in the isr)
17:25:27funmanand the code use A5 or A0 depending on some boolean in the OF (I read that this boolean was initialized to 0, but I may have missed a part)
17:26:33kugelI think I can provide the isr for gpiob
17:28:40 Join faemir [0] (
17:36:39 Nick fxb__ is now known as fxb (
17:39:13kugelwriting this one register changed the colors on my fuze
17:40:43funmanfredddy mentioned using A5 showed a grey display
17:40:56 Join toffe82 [0] (
17:41:20 Join {phoenix} [0] (
17:42:18funmanthe delay used is 4 timer cycles = 2 micro secodns
17:42:25funman2.66 micro seconds*
17:44:35funmanvoid i2c_delay(void) { int old = TIMER1_VALUE; while(TIMER1_VALUE + 4 < old) ; } would do the job
17:47:20kugelfunman: the lcd driver sets A5
17:48:32 Quit linuxstb ("Leaving")
17:48:37funmanoh then it makes sense to use A0
17:49:14 Quit jhMikeS (Nick collision from services.)
17:49:20 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
17:49:59kugelfunman: wait a bit before committing
17:50:58funmankugel: no argument to lcd_write_cmd(0x22) ? (cmd 0x21 has an argument, just like cmd 0x47 & 0x48
17:51:21kugelfunman: write_reg != write_cmd
17:51:33funmanoh sorry
17:51:40kugelno problem :)
17:51:52 Join Jaykay [0] (
17:52:45 Quit massiveH ("Leaving")
17:53:06kugelfunman: "GPIOA_PIN(5) = (1<<5);" is what the lcd driver does. is that really setting the pin?
17:54:02 Nick jfc^3 is now known as jfc (
17:54:02funmanyes, setting bit5 in the register will drive pin5 high
17:54:12kugelif I do "GPIOA_PIN(5) = 1;" I get a white screen as well
17:54:57kugeler? I thought I only access the specific bit for that pin with GPIOx_PIN
17:55:26funmanGPIOA_PIN(5) = 1 /* << 0 */; has unspecified results as far as i know
17:55:45funmanwhen reading register for pin X you only get pin X set or not (all other are unset)
17:55:55funman.. you only get *bit* X set
17:56:19kugelso a pin is 8 bits or what? I thought it's 1bit per pin
17:57:21kugel(I mean one significant bit)
17:57:21funmanthe register itself is 8 bit, but only 1 bit has a meaning, you shouldn't modify/read the others
17:57:41 Quit petur ("home->...")
17:57:52funmanyep, the (1 byte wide) register for PIN X only have one significative bit : the BIT X
17:58:04kugelyea, that's how I thought it is. And I thought GPIOx_PIN only reads/writes that significant but
18:00:05funmannope it reads one byte
18:00:37funmanthe definition of 'byte' is 'the smallest area of memory accessible' (usually 8 bits, but it can be more on exotic hardware)
18:00:41 Join herrwaldo [0] (
18:00:48kugelhm, then the button driver does it wrong I suppose
18:01:04kugelfunman: yea, makes sense
18:01:55funmanfor example on ARM there is ldrb (load byte = 8 bits), ldrh (load half-word = 16 bits) and ldr (load word = 32 bits), you can't go below 8 bits
18:02:10kugelfunman: btw: I updated the patch once more. I tested it by mapping the working power button to button down and play a bit in quickscreen
18:02:20kugelfunman: I know what a byte is :)
18:02:36 Quit perrikwp (" ajax IRC Client")
18:02:36Jaykay@the developers: as i did the day before yesterday, i suggest closing of
18:02:47 Quit mc2739 (Read error: 110 (Connection timed out))
18:02:50kugelI was just under the impression the _PIN macros only access that significant bit
18:02:52Jaykayit does something maybe unwanted automatically
18:03:04 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
18:03:24funmankugel: byte != octet !
18:03:26kugelI haven't looked to much at the macros, I admit
18:04:13kugelanyway, it's cleared up now and I can use my new knowlegde now :)
18:04:15funmanthe macros just access the byte without using unnecessary "& (1<<x)", but you should use it when writing to the pin
18:04:19toffe82I finally have some time to install rockbox on my ippod 5.5 with 80gb hd, I use the last version of rbutil, the ipod was detected as ipo and I choose the option with the 80gb hd, everything download and install
18:04:31rasherJaykay: I'd like to leave that to the dev who opened the patch
18:04:47toffe82but when I restart the ipod I had a panic error , cluster size not supported 4096
18:04:56toffe82do I miss something ?
18:05:03Jaykayrasher: ok...
18:06:10 Quit MethoS-- (Remote closed the connection)
18:07:05 Quit Schmogel (Read error: 110 (Connection timed out))
18:07:51funmantoffe82: did you try restoring the ipod with itunes and reinstalling rockbox again?
18:08:22toffe82I restore it but didn't have time to install again, I will do it tonight
18:08:40Jaykaya question concerning is this still needed? because it was opened in march 06 and last synced in aug 07....
18:08:48toffe82the firmware version of the ipod is 1.3 and was 1.3
18:09:01toffe82before the restore
18:09:53kugelfunman: so what needs to be done now with the buttons and isr on gpiob?
18:11:22 Join Nibbler [0] (
18:12:07funmankugel: use a correct delay, run the code and see what meaning we can give to the value we read
18:12:34funmanit could be 1 button per bit, or arbitrary (position or steps increment) value from the wheel
18:13:22kugelcan I grab your latest patch somewhere?
18:14:19funmanJaykay: if it's synced, yes it'd be wanted
18:14:24kugelfunman: btw: i tested my lcd stuff, by mapping the working power button to BUTTON_DOWN and playing in quickscreen a bit
18:15:09funmankugel: , but you may want to use the delay function I pasted here
18:15:39Jaykayfunman: not the answer i hoped for but thanks^^
18:16:28kugelfunman: btw: I remember that I had some problems with lcd_update(), I always used lcd_update_rect for such stuff
18:16:49funmankugel: so shall i commit it ?
18:18:02kugelif noone else wants to look at it (maybe linuxstb?)
18:18:59funmanhum if you have problems with lcd_udpate() i'd look at lcd_window_*
18:19:48toffe82funman: why restoring the ipod would change something ?
18:19:57toffe82different formatting ?
18:25:09funmanyes, i don't know if the cluster size is hardware/partition table/or filesystem specific (I'd go for filesystem but i'm not sure)
18:26:00funmankugel: then i'll let linuxstb do it :
18:27:05kugelfunman: I agree
18:29:22kugelfunman: read doesn't change
18:29:33kugelREAD: 0, no matter what I press
18:29:33funmanstays at 0 ?
18:30:07kugelI haven't changed the delay loop yet. Where should that what you posted go into?
18:30:14 Quit n1s ()
18:30:23funmannow try to put a panicf() if it's != 0, or to only run that in the GPIOB isr (where you unmask only pin 0)
18:31:03funmankugel: button_delay()
18:31:11kugel"i2c_delay(void) { int old = TIMER1_VALUE; while(TIMER1_VALUE + 4 < old) ; }" looks weird. Doesn't the timer value increase?
18:31:34funmanit decreases until it reaches 0 and wrap
18:31:53funmanhum there it is what the other check was : check for wrapping
18:32:04*funman double hums
18:32:39funmanyou should use TIMER2 (which is the always active kernel tick) and not TIMER1 (off by default plugin timer I guess)
18:33:09 Join karashata [0] (n=karashat@
18:33:21funmanat worse it'll spend 10ms instead of 2us, but feel free to correct the function (i.e. check if the newly read value is > at the old one)
18:33:21kugelfunman: &button is incompatible pointer type
18:33:34 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:33:34funmanjust use unsigned char
18:34:13funmani.e. unsigned char button = 0; (i2c functions read 8 bits at a time)
18:36:35***Saving seen data "./dancer.seen"
18:39:05kugelfunman: it shows FC now, but still doesn't change
18:39:19kugelI made sure the lcd update works by showing a counter
18:39:58funmankugel: how do you know the value doesn't change but is hidden by the high refresh rate ?
18:41:13kugelret seems to be 0xfc, panicf on if(button) doesn't kick in
18:42:37funmanhum perhaps I'm wrong, i thought -4 would be 0xfffffffc
18:43:16funmanmaybe modify the function to give a unsigned char pointer so you can check the return value separately
18:43:55 Quit andrew__ (Read error: 60 (Operation timed out))
18:44:50kugelfunman: which function? button doesn't seem to be changed at all
18:45:24funmankugel: make int button_i2c_read(void) become int button_i2c_read(unsigned char *button)
18:45:38 Join dabujo [0] (
18:48:25 Quit Jaykay ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
18:49:20funmankugel: -4 means writing of the address failed (see generic_i2c.c) so perhaps the delay function is too fast, try playing with that
18:49:25 Join Lynx_ [0] (
18:49:29 Join miepchen^schlaf [0] (
18:50:31Lynx_Hi guys! I just had a friend install rockbox on her sansa (using rockbox utility), and she sais now her computer (win) does not recognize the player anymore when plugged in via usb...
18:50:49kugelfunman: button_i2c_read(unsigned char *button); and printing button shows that button increases the same way as my counter
18:50:55LambdaCalculus37Lynx_: Tell her to reboot back into the OF (hold |<< while turning on the Sansa).
18:51:21Lynx_LambdaCalculus37: usb support in rockbox does not work yet at all?
18:53:55LambdaCalculus37Lynx_: We still have some nagging issues to sort out before we can activate it for normal builds.
18:54:14LambdaCalculus37So she'll have to boot back into the OF every time she wants to add files or update her build.
18:54:32Lynx_LambdaCalculus37: right, thanks
18:55:37funmankugel: can i see your diff please?
18:55:44LambdaCalculus37Lynx_: Also, let her know that the manual covers a lot of information about using her Sansa with Rockbox, so have her download and read it.
18:55:46Lynx_did the images from just disappear?
18:56:27Lynx_LambdaCalculus37: yeah, did that, the response was 'but it's so much' ;) But I'll help her with stuff when I see her tomorrow
18:57:54kugelfunman: I need to head off now, I'll be back later and read the logs
18:57:56 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
18:58:58funmankugel: hm the delay should be while(TIMER2_VALUE + 4 > old)
18:59:49funmanand check the return value of i2c_read_data() because it can be important to know where it failed
19:00:16funman+ did you know you can use format specifiers in panicf() ? like panicf("button detected : %x", button)
19:00:20 Quit funman ("leaving")
19:01:17 Join Jaykay [0] (
19:01:39 Join massiveH [0] (
19:02:35 Join kugel [0] (n=chatzill@unaffiliated/kugel)
19:03:12kugelerr, funman already gone?
19:03:28LambdaCalculus37toffe82: Try this out... in firmware/drivers/ata.c, change #define SECTOR_SIZE from (512) to (4096), build, and see what happens.
19:04:04*LambdaCalculus37 smacks himself
19:04:13toffe82LambdaCalculus37:only this ?
19:04:28LambdaCalculus37toffe82: Actually, scratch that for a minute.
19:04:44*LambdaCalculus37 reads ata.c again to make sure he didn't miss anything
19:05:50*LambdaCalculus37 requests the opinion of an iPod dev
19:06:23toffe82looks like everybody is sleeping today :)
19:06:35LambdaCalculus37toffe82: Indeed. :)
19:06:47LambdaCalculus37Don't change anything yet, because I want to make sure that this will be correct.
19:07:11kugelfunman: (for the logs) with the diff I gave you button doesn't change at all, also with the corrected delay (so it's "> old"). ret is still -4
19:07:23 Quit kugel (Client Quit)
19:07:45toffe82LambdaCalculus37: I have to reinstall a machine to compile, my virtual machine died after an update
19:08:18LambdaCalculus37toffe82: Let me try it over here on my end.
19:08:26toffe82LambdaCalculus37: ok
19:08:42*LambdaCalculus37 gets out his PowerBook and decides to try out that change
19:10:24LambdaCalculus37toffe82: Just to be sure... this iPod video was originally a 30GB, but with an 80GB hard drive upgrade, correct?
19:12:08 Join fml [0] (n=4fd3ee1e@gateway/web/cgi-irc/
19:12:36toffe82LambdaCalculus37: it is what they say on the site
19:13:49LambdaCalculus37toffe82: If they say so. :)
19:13:51*LambdaCalculus37 builds
19:15:32 Join MethoS- [0] (
19:15:57fmlHello. When I go to the FM radio screen, then go to the main menu and there select "Resume playback" −− shouldn't I get back to the radio? Now, mp3 playback is started.
19:16:16 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
19:16:46 Join tvelocity [0] (
19:16:49fmlSansa e200, r19360, if that matters
19:19:06 Quit Jaykay ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
19:24:55 Quit miepchen^schlaf ()
19:26:37 Join miepchen^schlaf [0] (
19:31:48 Join gregzx [0] (
19:32:16domonokyfml: no that resumes the current playlist. fm-radio is a bit special .. :-)
19:35:58domonokyso fm-radio is not considered as "playback". you also can not "resume" fm-radio. So its not a bug, its a feature :-)
19:38:15 Quit Nibbler (Read error: 148 (No route to host))
19:40:26 Join webguest92 [0] (n=568db29d@gateway/web/cgi-irc/
19:42:32 Quit webguest92 (Client Quit)
19:44:23 Join Thundercloud [0] (
19:48:22fmldomonoky: but I remember there was a discussion about that and it was decided the other way round. And it also worked that way. I.e. resume playback should go to the last "sound producing" screen. I'm almost 100% sure that it worked that way. That's why I'm asking.
19:58:12 Join bertrik [0] (
20:03:24 Quit _lifeless (Remote closed the connection)
20:12:34 Quit Thundercloud (Remote closed the connection)
20:13:01 Join Thundercloud [0] (
20:13:22 Quit Thundercloud (Remote closed the connection)
20:14:55fredddycan someone have a look at FS #9617 ?
20:21:53 Quit {phoenix} (Remote closed the connection)
20:24:53 Join {phoenix} [0] (
20:27:27 Join mib_vkht51si [0] (i=d8ef4426@gateway/web/ajax/
20:27:35 Nick mib_vkht51si is now known as MarcGuay (i=d8ef4426@gateway/web/ajax/
20:28:41MarcGuayfml: I know that if you're in the Radio screen, go to the main menu, and press play (the button equivalent of Resume Playback), it'll return you to the radio. Seems a little inconsistent, but I always just hit Play, anyway.
20:29:57 Join sajes [0] (n=sajes@unaffiliated/sajes)
20:30:12 Join jgarvey [0] (
20:31:20*LambdaCalculus37 waves to MarcGuay
20:31:28 Part sajes ("Off to -community. (Wrong channel :)")
20:32:28*MarcGuay shakes a weary limb
20:33:20 Join Zagor [242] (
20:33:31LambdaCalculus37MarcGuay: I took a look at FS #9524 and had a go at cleaning it up a little bit so it sounds more comprehensive. I'll try and post another diff later today if you'd like to take a look.
20:33:44 Quit krazykit (Read error: 60 (Operation timed out))
20:34:22 Join krazykit [0] (
20:34:33 Join sajes [0] (n=sajes@unaffiliated/sajes)
20:34:44 Quit blahrus (Remote closed the connection)
20:35:29sajesIs there a specific reason Rockbox doesn't work with Sansa E200 V2's? Like copyright/patent issues?
20:35:34MarcGuayLambdaCalculus37: Sure.
20:35:40 Join petur [50] (n=petur@rockbox/developer/petur)
20:36:10 Quit petur (Client Quit)
20:36:11krazykitsajes, completely different hardware
20:36:20 Join petur [50] (n=petur@rockbox/developer/petur)
20:36:29domonokysajes: new hardware, but people are working on a port.. :-)
20:36:37***Saving seen data "./dancer.seen"
20:37:56sajesThanks guys. I'll have to keep checking back on the stable builds status.
20:38:34 Join reacocard [0] (
20:39:03 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
20:39:41 Part sajes
20:40:30fmlMarcGuay: ah, yes, that's it! Ok, so it works as it did for months now. I confused pressing PLAY and selecting "resume playback"
20:41:14MarcGuayJdGordon, anyone: You/we/it should probably decide whether it's called "Quick Screen" or "Quickscreen" for the sake of searching and whatnot.
20:41:31rasherSelecting "Resume playback" should do the same as pressing PLAY, if you ask me
20:42:05MarcGuayI agree.
20:42:11fmlrasher: or the other way around :-)
20:44:31fmlBut seriously: iirc this also was discussed, and many didn't want that "resume playback" does anything but starting music playback (not just producing sound)
20:46:11MarcGuayI suppose it's nice to be able to take a break from your playlist, listen to the radio, and return to your paused spot in the playlist without having to use bookmarks.
20:52:43fmlMarcGuay: ah, I did again (press Enter) Yes. And, actually, it behaves like the "FM radio" menu entry in the sense that the FM entry unconditionally gets you to FM. And ResumePlayback always gets you to mp3 playback. So it's ok if you ask me.
20:55:55MarcGuay"Playback" infers that the sound is recorded as well, and all of the Playback Settings refer to how sound file playback behaves, not FM radio. It might be wise to keep the terms as is.
21:00:52fmlMarcGuay: I'm not argueing with that. Everything is fine!
21:02:21MarcGuayI know. Just thinking out loud... Err... Typing out loud... You know..
21:03:09fmlMarcGuay: hammering at the keyboard (so that it gets loud :-)
21:03:11 Quit jhulst (Remote closed the connection)
21:04:56 Quit gevaerts (Nick collision from services.)
21:05:05 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
21:08:49 Quit massiveH (Read error: 104 (Connection reset by peer))
21:09:01 Join massiveH [0] (
21:09:30 Join freddy_ [0] (
21:10:15 Quit fredddy (Read error: 110 (Connection timed out))
21:13:02 Join akur [0] (
21:17:26 Quit fml ("CGI:IRC")
21:19:07 Quit PaulJam (".")
21:19:19 Part akur
21:21:57 Quit perrikwp (" ajax IRC Client")
21:22:57 Join PaulJam [0] (
21:25:21 Join fml [0] (n=4fd3ee1e@gateway/web/cgi-irc/
21:26:19fmlI use Ubuntu 8.04. What latex packages should I install to be able to build the manual? I get the error about the missing marvosym.sty
21:26:46 Join Aurix_Lexico [0] (n=comrade@
21:27:36 Join akur1 [0] (
21:27:50crwl tells me that marvosym.sty is contained in the package texlive-fonts-recommended
21:30:24 Part akur1
21:30:24 Join akur1 [0] (
21:30:28 Join n1s [0] (n=nils@rockbox/developer/n1s)
21:30:56 Quit {phoenix} (Remote closed the connection)
21:31:02fmlcrwl: thanks!
21:31:15 Quit Lynx_ ("Konversation terminated!")
21:32:00 Part akur1
21:33:43 Quit itcheg (" ajax IRC Client")
21:34:42 Join {phoenix} [0] (
21:39:39 Quit fml ("CGI:IRC")
21:45:38 Quit n17ikh|Lappy ()
21:47:52Zagorisn't mp3 supposed to be working on the clip?
21:48:36LambdaCalculus37Zagor: It sort of works. MP3 playback tends to skip and stutter on the Clip for some odd reason. Vorbis playback is perfect, however.
21:48:47 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
21:49:10bertriklast nights' build didn't play oggs for me
21:49:18ZagorI don't get any sound whatsoever, nor does it count the time. perhaps due to these files being 320kbps?
21:49:27 Join kugel [0] (n=chatzill@unaffiliated/kugel)
21:49:28*Zagor tries some other files
21:49:42bertrikI think it broke recently
21:50:11soapIs it safe to say FS #4753 applies only to H3xx? Or was that just how the initial bug was filed? I see no mention of it happening on the H100, but I want to be sure.
21:50:46soapI was under the impression they were more alike at that level.
21:52:07Zagorno go with 128kbit mp3 either
21:52:34pixelmaUnhelpful: around?
21:52:58gevaertssoap: I'll have a look at it later, once I find out what exactly the bug is about
21:53:54soapI asked because I wanted to clarify affected targets in the wiki page ReleaseNotes31.
21:53:58kugelZagor: try the flash-buffering, funman said it's increases playback reliability
21:54:20kugelBut I have experienced the same on my fuze, which has enough mem to buffer the classic way
21:54:54LambdaCalculus37Zagor: I'm running a build funman gave me yesterday, which (mostly) works.
21:55:34*kugel "Backlight fading is now available on most targets. " in the release notes :)
21:56:34domonokymaybe r19383 broke it ? (perhaps not enough buffer left).
21:57:09LambdaCalculus37domonoky: I had gotten a *PANIC* SD: DATA TIMEOUT, error message.
21:57:12 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
21:57:14 Join Nibbler [0] (
21:57:19kugeldomonoky: does plain svn work for you?
21:57:38*domonoky will test plain svn again.
21:57:45*LambdaCalculus37 will also test plain SVN again
21:57:59 Quit LambdaCalculus37 (" ajax IRC Client")
21:58:29 Join ajonat [0] (n=ajonat@
22:00:13ZagorI get 592KB buffer
22:00:17Zagor593 even
22:00:39domonokyZagor: could be too less, pcm buffer alone needs ~512k ..
22:00:48amiconnkugel: The backlight fading is rather flickery on my c200
22:01:10Zagordomonoky: on clip too?
22:01:51domonokyyes, we use the normal unmodified playback engine.
22:02:07domonokymy working m200v4 has ~880k buffer
22:03:38Zagorand they both have 2MB ram?
22:04:50domonokybut quickscreen/tagcache/pitchscreen is disabled on m200v4, r19382 enabled it for the clip.
22:10:47soapthe release notes look very nice, rasher.
22:10:56kugelamiconn: is it? it's fine on my e200 (for me), same backlight driver
22:11:42 Quit faemir (Remote closed the connection)
22:12:30pixelmaflickery was my impression too (also c200)
22:13:17 Join Aurix_Lexico1 [0] (n=comrade@
22:13:45 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
22:14:00 Join _lifeless [0] (n=lifeless@
22:14:16kugelmaybe there's some hardware differences between e200 and c200 which makes it flickery. You can try adjusting the step delay
22:14:30gevaertsDo we disable keyclicks again for PP in 3.1?
22:15:13kugelgevaerts: isn't in supposed to be fixed?
22:15:36gevaertskugel: I didn't see it fixed, but I may have missed it
22:15:39 Join moos [0] (
22:15:46*gevaerts doesn't use keyclicks himself
22:16:00kugelgevaerts: r19388
22:16:20 Quit Aurix_Lexico (Read error: 60 (Operation timed out))
22:16:47*gevaerts claims to have been asleep at the time
22:16:59pixelmathere was a commit by jhMikeS this week and the comment says it was fixed - didn't try myself though. I also don't know if it also fixes the issue that is mentioned in some forum threads that the recording screen freezes when keyclick is enabled
22:17:35pixelmait's still on the frontpage... ;)
22:17:47 Join faemir [0] (
22:18:16gevaertsThat would be FS #8660. According to the comments in there I don't think it's very related
22:18:32MarcGuayKeyclicks work but recording still locks up.
22:19:18gevaertsMaybe recording can disable keyclicks for the duration?
22:20:31n1sgevaerts: i guess it has to, but as the suggestion in the bug's comment from pondlife i think a generic pcm playback disabling while recording would be nice, to take care of both keyclick and voice in one place
22:21:03 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
22:21:15*pondlife is the devil
22:21:43 Join funman [0] (n=5a241c06@gateway/web/cgi-irc/
22:21:58pondlifeI meant that the current voice-disable should be extended to any other PCM output
22:22:23n1si think it sounds like a sound plan ;)
22:23:35funman320kbps play fine for me on the Clip, I have a checkout from this afternoon but didnt try it yet
22:23:42*linuxstb reads the log and wonders what funman & kugel wants him to commit...
22:23:49 Quit {phoenix} (Remote closed the connection)
22:24:35Zagorfunman: do you have tagcache enabled on the build you are running?
22:24:39funmanlinuxstb: fuze lcd patch added recently on flyspray (must be on the last page)
22:25:19 Join akur [0] (
22:25:36 Part akur
22:25:38 Quit faemir ("Leaving")
22:25:43 Quit pondlife (Client Quit)
22:25:59gevaertsIs FS #7220 really enough of a bug to warrant a release notes mention?
22:27:15*domonoky detects that this suddon shutdowns on m200v4 have something todo with volume, if i turn it too high it will shutdown. if i lower the volume it never happens. strange...
22:27:32gevaertssafety feature :)
22:27:47Zagorgevaerts: I'd say either it is a bug, and then it should be on the list. or it's not a bug and then it should be closed.
22:28:17domonokygevaerts: built-in ear-protection :-)
22:28:18 Join aarcane [0] (
22:28:37Zagorgevaerts: oh right, the text says "the most important". hmm...
22:28:40funmanI have svn plus FS #9332 plus FS #9611 (i think? for the FM) and playback is fine if i ignore some skips
22:28:59Zagorfunman: how much buffer do you have?
22:29:05n1sdomonoky: some dacs have overheat protection, maybe that's kicking in (and since the cpu and dac is in the same chip it takes down the whole thing) ? :)
22:29:40gevaertsZagor: Actually I think it should be either closed or fixed. JdGordon's fix looks easy, and "we think the keymap could be better" sounds like a feature request to me, and we closed that tracker
22:29:44Zagoroh, FS #9332 "read directly from flash" is clearly a big difference from my build
22:29:49oboon the 3.1 release page, shouldn't most of the fixed bugs under plugins really be under core?
22:30:33n1sobo: i got that impression too
22:30:52Zagorobo: definitely all codec related anyway. or moved to a "codecs" section.
22:31:03 Join namnuf [0] (i=5a241c06@gateway/web/ajax/
22:31:17n1si think codecs can be under core
22:31:21MarcGuaygevaerts, Zagor: The other part of the bug is that it's missing from the manual. And yeah, simply removing the key as suggested sounds easy enough.
22:31:26Zagorlooks like someone didn't finish that section
22:31:34 Quit funman ("CGI:IRC")
22:31:47 Nick namnuf is now known as funman (i=5a241c06@rockbox/developer/funman)
22:31:48pixelmaI believe being able to delete a bookmark with a button when browsing the list is long existing feature, not sure as I don't really use bookmarks...
22:32:12n1s8901 is a little bad to have in the list too, It was fixed in the ipod minis with the release of the 3.0 bootloader but it isn't fixed in _all_ targets that can be cf modded...
22:32:14gevaertsobo: oops. My fault I guess (I did the splitting)
22:32:26funmanzagor if zou are familiar with buffering.c you can have a look at why it causes problems (it or another file which uses/is used by it)
22:32:37 Quit J-23 ("ZNC -")
22:32:45kugellinuxstb: a few fuze lcd driver fixes & cleanups
22:32:54 Join tessarakt [0] (
22:33:24Zagorfunman: you mean why it skips?
22:33:27 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
22:33:30linuxstbkugel: Have you tested it in both the bootloader and main Rockbox build?
22:33:42linuxstbAnd has anyone tested it on a different fuze?
22:33:49funmanzagor, i dont know if it skips, but i know it corrupts the memory and cause crashes
22:34:27gevaertsobo: should be better now
22:34:30Zagorfunman: you wrote "I begin testing on the Clip and notice a lot of skips in mp3" in a comment to the patch?
22:34:45funmanfreddy_: FS #9617 looks ok to me, but i cant commit it now, perhaps another developer can
22:35:00kugellinuxstb: not in the bootloader. But in the main build, as far as I could test with 1 button
22:35:13obogevaerts: yup, looks good to me
22:35:22linuxstbkugel: It should be tested in the bootloader - because the bootloader already inits the lcd.
22:35:36funmanZagor: FS #9332 causes skips because it doesnt refill automatically when reaching a low level. I meant checking the original buffering.c (I tried but didnt see anything obvious)
22:36:15*gevaerts leans towards closing FS #7220
22:36:38***Saving seen data "./dancer.seen"
22:36:40Zagorwell, since I can't play audio I can't really test buffering.c... :-) I'll try with tagcache disabled and see if that works better.
22:36:43gevaertsDo people really leave the player in the Recent Bookmarks list that often?
22:37:05 Quit kugel (Nick collision from services.)
22:37:08 Join kugel_ [0] (
22:37:14 Nick kugel_ is now known as kugel (
22:37:35funmanZagor: you have no sound but can you still power off or navigate ?
22:37:45Zagorfunman: correct
22:37:51n1sgevaerts: I think that the argument that we shouldn't use the rec button for anything since people use it to turn on the backlight is not very good :)
22:38:00gevaertsn1s: exactly
22:38:04kugellinuxstb: the main build does init the lcd as well (there's no #ifdef's for the bootloader).
22:38:41linuxstbkugel: I know, but my point is that your patched version of the lcd init is different - so it needs testing in the bootloader
22:38:58n1sgevaerts: and we have the "first keypress enables backlight" setting...
22:39:00pixelmafunman: the Fuze has a screen resolution of 220x176x16, right? If so the bitmaps in the plugins won't be a problem as the H300 and the 4G Ipod Color have this screen size
22:39:12linuxstbpixelma: Correct.
22:39:17funmankugel: but then a mandatory bit could be still made by the old bootloader and not the new build
22:39:59gevaertsn1s: I'll close it
22:40:26*n1s wishes we could release new bootloaders for the iriver h100/h300 sometime...
22:40:30n1sgevaerts: ok :)
22:41:14gevaertsn1s: indeed. That would close a lot of bugs
22:41:37gevaertsn1s: indeed. That would close a lot of bugs
22:41:53linuxstbn1s: indeed. That would close a lot of bugs
22:41:57Zagordo we know why the clip takes so long to shutdown?
22:42:40kugellinuxstb: Ok, I'll do a test
22:42:41n1sgevaerts: btw, iirc you reported some problem with the new sansa bootloader in FS #9369? or was i dreaming?
22:42:51*petur isn't really convinced about closing 7222
22:42:59petur7220 I mean
22:43:00kugelfunman: see why I proposed to wait for another voice :)
22:43:58funmankugel: you deserve my respect for that ;)
22:44:26*domonoky finds it strange, that his m200v4 reliably shutsoff if the headphone vol (hph in as3514.c) goes to 0x19..below its fine. :-/
22:44:58gevaertsn1s: I had some USB detection issues while testing, but I found that I had usb detection issues on my c200 around that time with everything, including a clean OF, so I blame the cable
22:45:37Zagorremoving tagcache bumps the buffer from 593 to 836 KB
22:45:41funmandomonoky: did you measure if there was differences between the OF level and rockbox level ?
22:45:46 Quit aarcane (Client Quit)
22:45:53n1sok, i'll volunteer lambdacalculus' c200 as a guinea pig and if all goes well as Bagder to upload the new versions then :)
22:46:05Zagorand hence working music
22:46:16bertrikdomonoky, the ams sansa battery code assumes a lithium battery, that's probably part of the problem
22:46:25gevaertspetur: I won't object to a fix for the issue, but I don't think we need this sort of debatable things on the tracker for years
22:46:25Zagor...unless I touch the volume keys, that is :-)
22:46:28funmanZagor: hum I had measured much less RAM usage when i enabled it
22:46:33 Quit massiveH ("Leaving")
22:46:50gevaertsn1s: I can test too, since I now also have an e200 with cable :)
22:47:04domonokybertrik: i already declared NO_LOWBATTERY_SHUTDOWN, and its clearly linked to volume..
22:47:15*gevaerts didn't realise that the new sansa bootloaders were still pending
22:47:27kugellinuxstb: it works
22:47:45n1sgevaerts: would be great, seems to work just fine here
22:47:53Aurix_Lexico1what does OF stand for?
22:48:01 Nick Aurix_Lexico1 is now known as Aurix_Lexico (n=comrade@
22:48:08domonokyOF = original firmware
22:48:25Aurix_Lexicoah, ty
22:49:24 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.4/2008111318]")
22:49:38Zagorskipping happens too often to be buffer refill related
22:49:44 Join kugel [0] (n=chatzill@unaffiliated/kugel)
22:50:01funmanZagor: 32kb doesnt last long
22:50:31Zagorthis is svn, not #9332
22:51:07funmanoh, i dont know about SVN, like I said i want to understand why it crashes before looking why it skips
22:51:38Zagorpossibly the same reason
22:53:02*n1s still think some assumption leads to overwriting memory or reading too much
22:53:18n1ssize assumption, that is
22:54:01 Quit XavierGr (Nick collision from services.)
22:54:14 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
22:54:42 Quit MethoS- (Remote closed the connection)
22:55:00funmanZagor: possibly the CPU freq required could be too low (25, 31, and 248MHz), did you see if wav files skip ?
22:55:03 Join MethoS- [0] (
22:55:14 Quit MethoS- (Remote closed the connection)
22:55:17 Quit leachim6 (Read error: 104 (Connection reset by peer))
22:55:18ZagorI'll test that now
22:55:33 Join MethoS- [0] (
22:57:09amiconnkugel: Yes it is. I disabled fade-in of course, because of its irritating delay, but kept fade-out enabled.
22:58:01 Join faemir [0] (
22:58:23kugelamiconn: I find the short flash (of lcd_enable) much more irritating. But I guess that's what the settings are good for
22:58:46 Quit petur ("*plop*")
22:59:05Zagorwav does not skip
22:59:24n1sdo they sound right?
22:59:27 Join n17ikh|Lappy [0] (
22:59:39 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
23:00:19funmanperhaps 25 MHz isnt enough then
23:01:10funmannote: when you change frequencies you should define integer factors of PLLA_FREQ in clock-target.h for clarity
23:01:18 Join culture [0] (
23:01:25gevaertsn1s: forget all my complaints. Seems that my c200 was set to autodetect, which means "try to connect as MTP, and if that fails behave strangely". If I set it to MSC it's fine
23:01:43n1sgevaerts: good to hear :)
23:01:55gevaertsSo I probably have two good cables :)
23:02:29n1sBagder: the sansa bootloaders and sansapatchers in FS9369 are good to go (last 7 attachments)
23:03:18*gevaerts ended up installing the new c200 bootloader using rockbox usb :)
23:03:25Bagdern1s: do they have a version number?
23:04:10 Quit domonoky (Read error: 104 (Connection reset by peer))
23:04:15n1syes, bootloaders are now 5.0 and sansapatcher is 0.7 ;)
23:04:27n1syay, for consistant versioning
23:04:45Bagderdid the older ones have any?
23:04:55Bagderi mean those on the download server atm
23:05:02gevaerts0.6 and 4.0?
23:05:05Zagorwav also does not crash.
23:05:13n1sthe old e200 booltoader was o.4 at least, not sure about c200
23:05:44BagderI'll call them all 0.4 I think
23:05:49n1sZagor: could it have something to do with different codecs requesting different sized chunks of data?
23:05:49Bagderjust to keep them around for a while
23:05:54gevaertsWas it 0.4 or 4.0?
23:06:07n1s4.0 sorry
23:06:24gevaertsBagder: if you just run the old sansapatcher it will tell you
23:06:40Bagderah right
23:06:42Zagorn1s: I don't know. I don't have any theories yet.
23:07:00Bagder"sansapatcher v0.6 with v4.0 bootloaders"
23:07:17 Quit karashata ("G'bye everyone!")
23:07:51 Quit jhMikeS (Read error: 104 (Connection reset by peer))
23:09:13Zagorafter mp3 fails I always get "SD: RX Overrun, RXR Fifo Full"
23:09:21 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
23:10:45funmanZagor: I think it happens when playback stops automatically
23:11:02funmanlike when I disable change folder
23:11:16Zagorwell my playback never gets a chance to stop automatically. it fails long before the track is finished.
23:11:47funmanThis is also a reminder that the SD code should mark transferred data as successfull on the data block end status flag, and retransfer if they failed
23:12:04funmanZagor: whats 'fail' ?
23:12:14 Quit MethoS- (Remote closed the connection)
23:13:21Zagorfunman: playback (both sound and counter) halts and cannot be restarted. after mp3 fails this way, wav no longer works either
23:13:37Zagorand usually pretty soon after that I get SD panic
23:14:16 Join MethoS- [0] (
23:14:19funmandoes the icon displays stopped?
23:14:55Zagorno, it stays at "play"
23:14:56ZagorI can pause and resume but only the icon changes
23:15:12funmanthen perhaps failed SD transfer cause the stopping
23:15:31funmanI'll try to give some love to SD code tomorrow
23:15:42Bagdern1s: 7 new files on the dl server
23:15:58 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
23:16:20Zagorfunman: I doubt that. why would mp3 cause sd failure and not wav?
23:16:47ZagorI rather think it is a memory bug. buffer overrun or wild pointer.
23:17:17n1sthat leaves only all irivers all iaudios and the mrobe
23:17:19funmanif you re using buffering.c and it causes corruption, the number of sectors asked could be largely too much
23:18:01funmanthe fact that wav cause no failures could just be that this hazardous and hardly reproducable bug hasnt happened yet or hasnt touched important memory yet
23:18:58 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
23:19:03funmanThe same file could very well play 10 times and fail on the 11th when i used buffering.c , and I noticed that no codecs were actually better than another (at the exception of wav which gave me longer playback before crashing)
23:19:14 Quit ebil|work (Read error: 110 (Connection timed out))
23:19:21 Join killan [0] (
23:19:27funmanIm not sure if I had playback stopping however
23:20:32 Quit faemir ("Leaving")
23:20:44ZagorI've played the 6-minute wav file through completely several times without problems but never get past 20 seconds in any mp3 file.
23:20:53ZagorI'm going to create some other formats and try those
23:22:06gevaertsfunman: how large reads can the SD driver handle?
23:22:48funmannot more than the actual storage size or than 0x7fffffff sectors I think
23:23:37 Quit jgarvey ("Leaving")
23:23:58gevaertsfunman: I mean, you said "the number of sectors asked could be largely too much"
23:24:41funmanThere is no check, so if it is over the SDRAM size for example it will cause problems
23:25:15gevaertsah ok. The buffering code shouldn't cause that sort of problems then...
23:25:23funmanthough I would expect a panic in this case if invalid memory is accessed
23:26:15Zagoris the playback & buffering code completely untouched for the clip, or is something changed?
23:27:26funmanuntouched, I just added a #ifdef DEBUG sanity check in playback.c I think (it wouldnt meet pcmbuf requirements when using a very low audio buffer)
23:28:00Zagorogg works perfectly
23:28:38kugellinuxstb, funman: I tested the bootloader (in case you didn't get my previous message)
23:28:55funmanZagor: I bet that you will think otherwise after using it for some days, and agree to my conclusion that the failures arent codec specific
23:29:15 Join massiveH [0] (
23:29:35funmangevaerts: we transfer data by batches of 127 sectors, but i believe a wrong starting address could cause problems
23:30:09 Join aneqrs [0] (
23:30:24shotofaddsgevaerts: I don't think I've mentioned it here yet, but I've been able to dump the 'spare' bytes from every page of the D2's NAND over USB (128MB worth) and I'm in the process of writing a PC-based TCC NAND simulator to try and work out where the remaining FTL issues lie. There is light at the end of the tunnel but lack of time is holding me back somewhat :(
23:30:31Zagorfunman: the fact that mp3 triggers the bug so much easier is still useful information
23:30:38gevaertsfunman: sectors as in 512 bytes?
23:31:09gevaertsshotofadds: I guess lack of time is better than lack of ideas :)
23:31:34funmanZagor: if the bug happens at the same location of the same file everytime sure, else I think it doesnt add useful information
23:31:35shotofaddsindeed, but lack of time is kinda frustrating!
23:31:51 Join faemir [0] (
23:32:17funmangevaerts: yes that s it (you define a starting address, say start from sector 32, and then they are transfered 127 by 127 (or less, 127 is a limit))
23:33:04gevaertsOK. So 63.5KB a time
23:34:07funmanonly the bootloader requests 256 sectors = 128kB, else all the requests are < 127 sectors
23:34:19gevaertsFrom my reading of buffering.c it reads 32K at a time, so that should have a very low chance of being the problem
23:34:40funmangevaerts: i dont think buffering causes direct problems to SD
23:34:40n1sfunman, Zagor: perhaps testing on a better supported target with memorysize artificially limited could yield some clues?
23:35:14funmanrather than it corrupts the memory and then random stuff happen (which IMO shouldnt be examined until we find where the corruption happens)
23:35:37funmann1s: good idea, i was also wondering if the simulator builds can be limited in memory
23:35:49shotofaddsgevaerts: In theory, if I write known values to the first 16 bytes of each logical sector using the OF's MSC mode, and then get a dump of the first 16 bytes of each physical sector, that combined with the spares dump "should" be enough information to nail this once and for all.
23:35:56shotofaddsbut as I say, in theory...
23:36:10n1sfunman: i think the sims are pretty far from targets when it comes to memory management
23:36:48 Quit mofux (Remote closed the connection)
23:37:35funmani must add i didnt test the latest revisions of buffering.c
23:38:26 Quit aneqrs ()
23:39:38funmanwhat I have seen is that some memory_handle struct would get corrupted. I remember an index (~= pointer) over 2MB (the whole SDRAM available) which makes me think of corruption, perhaps because of an overflow on the small memory target
23:41:45Zagorn1s: indeed. cramming a c200 build into 2MB could be an interesting test.
23:43:34 Quit krazykit (Read error: 110 (Connection timed out))
23:43:40Zagorotoh didn't domonoky say his m200v4 works fine?
23:45:21 Quit BigBambi (Remote closed the connection)
23:46:13funmanfine equals not better than the clip
23:46:40funmanthat s what I understood at least :/
23:48:48 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
23:49:29Bagderrasher: that rbutil image you gave me, that's a replacement of the existing one on the dl server right?
23:51:13 Quit n1s ()
23:53:59 Quit dabujo (Read error: 54 (Connection reset by peer))
23:58:19 Join daurnimator [0] (n=fake@unaffiliated/daurnimator)
23:58:23PaulJami was just wondering, with the inclusion of the WPS tag for arbitrary settings, is it planned to remove the old tags which became redundant?

Previous day | Next day