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 2009-01-08

00:00:14Zagorgartral1: 1,2,4,8-bit
00:00:15 Join solexx_ [0] (
00:00:30 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
00:00:31 Quit nibbler (Read error: 113 (No route to host))
00:01:09gartral1no 16?
00:01:55Zagornot among the original files, no
00:02:10gartral1well now, were gonna have a real problem..
00:02:31Zagortry running bmp2rb separately and see what happens
00:02:58 Join nibbler [0] (
00:02:59gartral1command not found
00:04:13Zagorlook at the command line. the program is located in tools/
00:04:44gartral1ohh.. well i certainly wont find it in build, will i, lol
00:04:53 Join Zom [0] (
00:05:15gartral1still says command not found
00:05:34BigBambiand is it there?
00:06:12BigBambiyou are doing ./bmp2rb right?
00:06:25gevaertsOr ../tools/bmp2rb :)
00:06:49gartral1oops, no i wasnt
00:07:23gartral1displays usage
00:07:58BigBambiyou need to tell it where the file is - if you type "something" it looks for something in the directories in your PATH, if you do e.g. "./something" it looks in the current directory
00:08:14 Part kerwood
00:09:36gartral1error - Unsupported compression 3
00:09:43toffe82does senfirm works from cygwin ?
00:10:35 Quit Nico_P (Remote closed the connection)
00:11:14Zagorgartral1: use uncompressed bmp
00:11:18moostoffe82: you can use sendfirm as .exe under windows
00:12:31 Quit solexx (Read error: 113 (No route to host))
00:12:37gartral1Zagor: gimp doesnt do that
00:12:56toffe82moos: I wanted to know if the linux version is working if you run it from cygwin and if yes somebody can use perhaps the cygwin.dll to run it so no more problem of licence
00:13:26 Quit itcheg (" ajax IRC Client")
00:13:41moosah ok sorry, don't know then, I don't use cygwin anymore
00:15:20gartral1haha... windows bitmap works and gimps doesnt
00:15:50 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
00:16:10saratogacouldn't rbutil just call sendfirm as a command line app?
00:16:18saratogasort of like closed source programs call lame.exe
00:16:23saratogathats certainly GPL clean
00:16:36gartral1Zagor: THANK YOU
00:16:50Zagorgartral1: you're welcome
00:17:01gevaertsIt would also make really sure that rbutil doesn't inherit dependencies from it
00:18:29 Quit Bagder ("Lämnar")
00:21:43 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
00:23:30 Quit jgarvey ("Leaving")
00:27:23 Join _Auron_ [0] (
00:27:58 Quit Schmogel (Read error: 110 (Connection timed out))
00:28:02kugelsaratoga: yes, but who distributes sendfirm?
00:28:49Zagorwe can still distribute sendfirm, it just cannot be licensed with gpl
00:28:51kugeljhMikeS, Zagor: Is anything holding back the backlight code police patch?
00:29:27kugelZagor: is lgpl possible?
00:29:27 Nick fxb is now known as fxb__ (
00:29:44gevaertsIt could be gpl+exception
00:29:51kugel(iirc yes)
00:31:20jhMikeSkugel: I'm not holding anything back.
00:31:22 Quit bertrik ("Leaving")
00:31:30Zagorkugel: well, I want to strip out the time settings. and if we do that there's no point in first code policing a lot of code that is removed. unfortunately I haven't seen a response from amiconn about this. (I'm asking him since he added the fade time settings)
00:31:52kugeljhMikeS: I asked if something holds it back, not if you do so :)
00:32:41bluebrotherLGPL allows you to link the code (i.e. a library) with non-OSS code. We have the other way round, and I don't think LGPL handles this
00:33:20*bluebrother should go and read it
00:33:29gevaertsbluebrother: how is that different?
00:33:30jhMikeSkugel: I'm speak about my own stance
00:33:32kugelZagor: well, besides that imo it should be done in a separate patch regardless of that half of it is "potentially" removed again, I'd like to get in so I can implement backlight fading for the ams sansas
00:34:43bluebrothergevaerts: as in which part incorporates which. We have OSS - links -> non-OSS, while (IIRC) LGPL deals with non-OSS program -> LGPL libs (glibc and such)
00:34:49saratogasendfirm could simply be BSD licensed, and then we could distribute it
00:35:02kugelbluebrother: we have the other way around? sendfirm is open source, and links windows' closed-source mtp stuff (IIUC)
00:35:05Bagdergevaerts: I think bluebrother is right
00:35:27Bagderbut why use lgpl at all?
00:35:40Bagderwhy not bsd at once ?
00:35:50bluebrotherkugel: yep. And LGPL is linking closed-source apps with LGPL libs.
00:36:04gevaertsHow? GPL and LGPL deal with code and linking, they don't have library/program distinctions
00:36:05kugelah ok
00:36:19bluebrothergood question. I'm not aware of the terms of BSD.
00:36:20Bagdergevaerts: LGPL does afair
00:36:27*gevaerts goes to read
00:36:36Bagderthe limits of LGPL is limited by what?
00:36:50bluebrothergevaerts: then split it in your code <-> other code :)
00:36:54Bagder"the library" originally, but now I think its phrased somewhat differently
00:37:14saratogaBSD basically lets you do anything you want, which is useful here since we want to link to closed source code
00:37:53saratogaLGPL makes little sense since we don't want to include the app in a closed source program, but rather want to link to a closed source DLL
00:37:59*gevaerts stands corrected
00:38:41gevaertsI would use either BSD or GPL+exception in this case
00:39:17kugelbut rbutil can't ship sendfirm in this case, can it?
00:39:27saratogayou can ship it
00:39:31saratogayou just can't link it
00:39:45saratogawhich is fine since we don't need to link it, we can just call it as a seperate exe
00:39:48kugelah yea
00:39:54saratogathink dbpoweramp calling lame.exe
00:40:04Zagorkugel: the macros in backlight.h should be in config.h
00:40:14Bagderyes, executing a separate app is fine
00:40:21bluebrotherBSD sounds like a nice option
00:41:01Zagorbut I really don't want to commit a big patch for code I don't even want us to even have :-/
00:41:01Bagderthe authors of the sendfirm code in use need to agree on that
00:41:32gartral1what does sendfirm do?
00:41:40kugelZagor: so it's rotting, and basically blocking the ams work, until amiconn answered?
00:42:13Zagorhow is it blocking ams work? it's only cosmetics.
00:42:18kugelwell, if yes, I'm just gonna do the ams backlight fading using the old scheme
00:42:35kugeland do duplicated work
00:42:58Zagoryes, I suggest you use the old scheme for now
00:44:26 Quit saratoga ("CGI:IRC (EOF)")
00:45:59*gevaerts can't find a license for the code on which MTP_DLL/sendfirm_win.cpp is based
00:46:37 Quit ender` (" On the other hand, you have different fingers.")
00:50:56***Saving seen data "./dancer.seen"
00:54:16Zagoris anyone here opposed to removing the backlight fade time settings?
00:55:11ZagorI like to compare it to "fade on stop". if we don't need a time setting for that, why do we need it for backlight?
00:55:32*linuxstb predicts a long debate over what the hard-coded values should be....
00:56:00Zagorlinuxstb: well we have rather sensible defaults now, don't we?
00:56:22bluebrotherI quite like to have a rather longish fade-out time on my Ipod ...
00:57:42linuxstbZagor: I dare you to send an email to the dev list asking that question... ;)
00:59:22 Quit DerDome ("Leaving.")
01:00:15 Quit MethoS-- (Remote closed the connection)
01:02:59XavierGrwell I for one object on removing it. The setting is there for a reason. Different usage habbits. It would be pity now that it's there to drop it.
01:03:45XavierGrZagor: I am sure you will find many many people that would not like to sucrifice their usage patterns for a "sane" default value.
01:04:49ZagorXavierGr: I would claim the setting is there because it could be done, not because of user habits. you're not seriously suggesting that not being able to configure your fade-out to 5 seconds instead of 2 is "changing user habits"?
01:06:39XavierGrconfigurability is subjective, I wouldn't mind having a sane defualt value, but I understand that many people like such a feature.
01:07:02Zagorthe feature stays. only the micro-management goes.
01:07:24XavierGralso the same applies for crossfade timings which I think you suggested to drop too
01:07:27Zagoralong with a truckload of #ifdefs and messy code
01:07:38Zagorno I didn't
01:07:54XavierGrin the crossfade issue I think it would be more than just pitty to drop
01:08:08ZagorI consider crossfade a lot more "core" to rockbox than backlight fading
01:08:22XavierGroh sorry if it wasn't you I don't remember exactly but there was a discussion about it
01:08:35Zagoryes, there was
01:09:29XavierGris it buggy at the moment? (backlight fading with timings)
01:09:37gartral1i agree, crossfadeing is awsome, backlight fadeing, i have turned off anyway
01:10:28ZagorXavierGr: I don't know
01:11:46XavierGrI think that for targets that hardware-support it, it works good
01:11:58Zagorgartral1: I think the feature is nice, but the time settings add too little value to justify the cost they incur in the code
01:13:09 Join itcheg [0] (i=62db4767@gateway/web/ajax/
01:13:21XavierGrZagor: but if you think that it will make a huge difference in code complexity then maybe a discussion on the mailing list would be more definite
01:14:15Zagorunfortunately, a discussion on the mailing list is never definite...
01:14:18XavierGrBut as linuxstb suggested it might bring many objections. I guess you never know until you try :P
01:15:17XavierGrZagor: yup you are right, sometimes it doesn't conclude to something solid.
01:15:27Lloreanlinuxstb: I don't think the logo swapper page should be around. The only reason I edited it was to make the warning bright red and a bit clearer.
01:16:01gartral1yea, logoswar doesnt even work anymore
01:17:14XavierGrI remember some discussion about different prefferences regarding fade-in vs fade-out. E.g some liked fade-out but never wanted to encounter fade-in to be able to have quick access to the screen, the solution was to set the fade-in timeout to 0 ms
01:17:58XavierGrif these settings are lost then maybe it will be a dillema for those poeple that liked only the fade-out effect
01:18:28 Join Bensawsof [0] (
01:18:51UnhelpfulZagor: if you don't like the core_ prefix, that's fine... would having the same name as the wrapped core file be OK, and then perhaps wrappers.h or api_wrappers.h for the rb-> defines?
01:18:51pixelmaZagor: but people sometimes don't follow IRC, so a discussion on the mailing list gives more chance for people who care to notice. And it is often encouraged (here) and I think this feature has been around for a long time if only for some (one?) target(s), would be fair IMO
01:19:15Zagorpixelma: yes, you are right
01:19:31 Quit itcheg (" ajax IRC Client")
01:19:32ZagorUnhelpful: I think the same name as core is good
01:20:43Unhelpfulok... there's another bmp.c in pluginlib already, but there are no function name conflicts, so i can include the core bmp.c there on platforms that will use it... and the name still says which functions it provides.
01:20:58 Quit tessarakt (Read error: 110 (Connection timed out))
01:21:09 Join tessarakt [0] (
01:21:29kugelUnhelpful: just to let you know, lucent has the same corruptions on his fuze
01:22:01gartral1what does the installer flag do at configureation?
01:22:28Unhelpfulif everything that needs scaling can reasonably be adapted to either use the bmp loader frontend, or act as a frontend to the core scalers, there wouldn't be anything left in bmp.c except for save_bmp, i believe.
01:22:58Unhelpfulkugel: good to know. do we have any real idea what's going on yet? have you figured out why the test_disk plugin fails?
01:23:27kugelnot really
01:23:49kugelI looked at the driver, but I didn't find anything obviously wrong
01:23:53Unhelpfulhrm, does test_disk get errors for lucent as well?
01:24:10kugelbut funman is apparently back, and he said he's going to work in it
01:24:13Unhelpfuland did you ever try smaller reads, to see if they "fix" test_disk just as they do the loader?
01:25:32kugelUnhelpful: I don't think it's storage related: a) 2 daps with the same corruption, b) cover art is all fine
01:25:39 Quit Bensawsome (Connection timed out)
01:25:39 Nick Bensawsof is now known as Bensawsome (
01:26:23Unhelpfulvs a pile of other models with similar LCDs using the same code without problems... plus test_disk failing.
01:27:21 Quit BrianRowe ("Ex-Chat")
01:28:35kugelUnhelpful: I still doubt it's storage
01:29:06*Unhelpful is rather unsure what it would be, then
01:29:08kugelUnhelpful: maybe something with RAM is wrong, that would also kinda explain some/most/all of the other issues
01:29:56Unhelpfulit's possible, i suppose...
01:31:20 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.5/2008121623]")
01:32:58UnhelpfulZagor: should just wrappers.h be ok for the plugin API wrapper defines?
01:33:23ZagorUnhelpful: sure
01:33:39 Quit Bensawsome (Read error: 60 (Operation timed out))
01:33:44Unhelpfulgood deal, i can knock out a cleanup for this pretty quickly, then.
01:35:10 Join roebeet_ [0] (
01:36:23 Nick roebeet_ is now known as roebeet1 (
01:36:39 Nick roebeet1 is now known as roe (
01:37:02 Part roe ("Ex-Chat")
01:39:02 Join roebeet_ [0] (
01:39:07 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
01:39:08 Quit tessarakt (Connection timed out)
01:39:31 Join tessarakt [0] (
01:39:52 Join MarcGuay [0] (
01:40:49 Quit Chronon ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
01:41:06 Join chrippa [0] (
01:41:27 Quit flydutch ("/* empty */")
01:41:37MarcGuaylinuxstb: Do you know if defining TCCBOOT requires compiling a dual-bootloader? I get a blank screen if I simply set it for the c100.
01:42:13 Quit bluebrother (Read error: 113 (No route to host))
01:43:04 Join CaptainKewl [0] (
01:43:20 Join markun [50] (n=markun@rockbox/developer/markun)
01:43:49 Part toffe82
01:46:01Unhelpfulis there a way to "erase" a prior declaration of a not-yet-defined function, such that you can write the definition with different qualifiers?
01:47:29 Quit Zagor ("Clint excited")
01:47:30Unhelpfulthat's what i figured. i've been trying to come up with a way for the mem* wrappers in plugin.h to be static inline, so that 1) they can be turned into direct calls to the plugin_api functions 2) they don't increase the size of the plugins that need them
01:47:51 Join Calcipher [0] (
01:48:50 Quit tyfoo ("Carpe diem")
01:49:06Unhelpfuli don't *think* define memset rb->memset will help, since the wrappers are for compiler-generated calls that don't appear in the source code
01:49:07 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
01:49:32 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
01:49:43Calcipherwell, I just bought a Cowon Iaudio D2
01:50:14 Join itcheg [0] (i=62db4767@gateway/web/ajax/
01:50:24Calcipherafter looking at the interface, and already owning a rockboxed Sansa, I want to try rb on it
01:51:55Calcipheris this the only place to get instructions on installation?
01:53:00CalcipherI'm not completely sure where to begin, I'm guessing determining what fw version i'm running, and then...
01:53:40UnhelpfulCalcipher: anything not in the table on the front page is an incomplete / unsupported port. you're probably looking at the right place, already, although it would likely be worth checking the forums for a new ports thread for D2, as well.
01:55:23Calcipherthanks Unhelpful, I'll take a look
01:55:56MarcGuayCalcipher: I don't think the instructions will get any easier than that.
01:57:27Unhelpfulthe "known problems" suggest to me that this is a good deal less usable than pretty much anybody would want for real use
01:57:48Unhelpfulif you're not developing, or testing for a developer, i would suggest you not use rockbox
01:58:33 Quit itcheg (" ajax IRC Client")
02:01:42 Quit moos ("Rockbox rules the DAP world")
02:02:01 Join fdinel [0] (
02:02:11Calcipherwell, I may end up returning the player in that case, I really use rockbox for the voice feedback features, since I have very low vision
02:03:07Calcipherthere are details that are not in the steps, such as anything regarding tcctool
02:03:42Calcipherand how to identify your current fw
02:04:12roebeet_Calcipher: I am a D2 owner as well. I've been watching the RB port status, but there's still work to be done.
02:04:17MarcGuayCalcipher: It's also a touchscreen player, which might be more difficult to use than something with tactile buttons.
02:04:26CalcipherSee I would like to atleast test the in development version before returning the device
02:06:06CalcipherI know touch is not my thing, but from reading the info on the page it sounds like rb allows for navigating with the buttons to a degree
02:06:06roebeet_Calcipher: Check the wiki for details,
02:06:20Calciphermay be just enough for normal function
02:07:11Calcipherroebeet_, thanks, I was looking at that before popping in and asking the same old questions
02:07:57Calcipherroebeet_, do you know where I can find what fw version my player has?
02:08:25roebeet_should show up on the boot screen, at the bottom
02:08:44roebeet_if you are in the US, it's probably 2.57
02:08:48Calciphernormal boot?
02:09:06MarcGuayCalcipher: tcctool is here . The zip contains instructions.
02:09:15Calcipheroh ok, lets see, yes I am in the U.S. by the way
02:09:41Calciphergreat, thanks MarcGuay
02:10:33roebeet_Also the fw version for the D2 is in settings > information
02:11:01Unhelpfulsvn mv just records an add and a delete? i thought that svn had "real" support for history-preserving file renames?
02:11:53Calcipherroebeet_, I see, I have 2.55
02:13:48 Join itcheg [0] (i=62db4767@gateway/web/ajax/
02:14:11MarcGuayCalcipher: Do you have the rockbox source code checked out?
02:14:13 Quit jhMikeS (Nick collision from services.)
02:14:19 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
02:15:18MarcGuayCalcipher: You'll need to do that.
02:15:45MarcGuayThere doesn't appear to be a pre-compiled bootloader.
02:16:09domonokyUnhelpful: svn mv looks like a delete and add, but it really preserves the history.
02:16:52Unhelpfulok, then i think i did the right thing prepping this cleanup :)
02:17:04CalcipherOh, so there is only a pre compiled rb, but not the loader
02:18:18Calcipheroh, so that means I'm going to have to break out my old cigwin setup, thats how I was working with some builds a few months ago
02:18:37 Join akur [0] (
02:19:18Unhelpfulit looks like a "working" bootloader is made by patching an OF file. it's the same way on gigabeat S, and that makes it difficult/impossible to legally distribute a "complete" loader.
02:20:03Calcipheri see, so i still need the 2.55 of, to patch
02:20:19 Quit Thundercloud (Remote closed the connection)
02:21:25CalcipherI'm going to dig around the port thread and see if I can find the stuff I'll need aside from tcctool
02:22:57MarcGuayCalcipher: All you'll need is the OF .bin and the rockbox source. mktccboot is in the tools directory. tcctool is optional but very easy to use in windows.
02:24:39Calcipherwell, I'm not completely sure how to go about this if I want to use the prebuilt RB version that was linked on the steps to install page
02:25:11CalcipherI'll need the source for that specific build then
02:27:07MarcGuayCalcipher: You'll just need to build the bootloader, which should load that prebuilt RB fine.
02:27:46 Quit SUSaiyan ()
02:28:24 Part Bensawsome ("The awsome is gone :(")
02:28:30CalcipherIt will need to be the most up to date svn data to build the bootloader right?
02:29:22CalcipherI'll take a look at the instructions in the zip you linked to
02:29:28MarcGuayCalcipher: Umm. It might help, I'm not sure when the last update to the D2 bootloader was.
02:30:36 Quit Aurix_Lexico (Read error: 110 (Connection timed out))
02:35:48 Join Divadmils [0] (
02:35:56Divadmilshey guys
02:36:24 Part Divadmils
02:36:42 Join Dizzle [0] (n=4c78fda2@gateway/web/cgi-irc/
02:37:32Dizzlehey everyone
02:37:56Dizzlewas just wondering if anyone is working on a build for the sansa fuze?
02:39:15 Quit Dizzle (Client Quit)
02:39:18 Join Dizzle [0] (n=4c78fda2@gateway/web/cgi-irc/
02:39:29 Quit Dizzle (Client Quit)
02:39:37 Join Divadmils [0] (n=4c78fda2@gateway/web/cgi-irc/
02:40:30 Quit Divadmils (Client Quit)
02:41:05 Join webguest66 [0] (n=4c78fda2@gateway/web/cgi-irc/
02:41:49webguest66hey is there going to be a build for sansa fuze? since there is already sansa builds.
02:42:08 Join gstory [0] (
02:42:22Unhelpfulfuze support is still pretty far from being complete, in many ways.
02:42:46webguest66does it use different firmware or something?
02:43:14kadobanwebguest66: it's completely different hardware than the supported sansas, iiuc
02:43:26webguest66okay thanks man
02:43:45webguest66and thanks for the link MarcGuay
02:43:58 Quit robin0800 (Remote closed the connection)
02:46:21 Quit webguest66 ("CGI:IRC (EOF)")
02:50:59***Saving seen data "./dancer.seen"
02:52:37 Quit domonoky (Read error: 104 (Connection reset by peer))
02:52:48 Part gartral1
02:53:46 Nick gstory is now known as GuyInDentonTx (
02:53:48CalcipherI got a bootloader.elf error when trying to make the loader
02:53:59Calciphersomething maybe wrong with my set up
02:55:52Calciphersince it seems I always had to use a command I didn't fully understand everytime I started using it, I'm guessing something was off somehwere so I always had to use the Export part=$path:/opt/arm/bin
02:57:21Calcipherlooks like my build environment isn't right
03:01:13 Join kerwood [0] (
03:04:18 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
03:05:11 Quit MarcGuay ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
03:06:17 Quit tessarakt ("Client exiting")
03:09:20Unhelpfulany comments on this cleanup for resized load in pluginlib before it goes in? i think i've addressed most of what zagor didn't like about it, especially the API stuff.
03:17:27 Quit tvelocity (Connection reset by peer)
03:22:37 Join reacocard [0] (
03:27:30 Join Darksair [0] (n=user@
03:29:57 Quit GuyInDentonTx (Remote closed the connection)
03:44:17 Quit XavierGr ()
03:45:36 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
03:52:10 Quit roebeet_ ("Ex-Chat")
04:03:53 Join blkhawk- [0] (
04:12:30 Join plutonian [0] (
04:12:51plutonianis there a plugin I can use to turn down the cpu frequency on my ipod under rockbox?
04:13:02plutonianor another way I can save some battery life?
04:14:24krazykitas it is, the cpu is only boosted when necessary
04:14:33 Join saratoga [0] (n=41becb3b@gateway/web/cgi-irc/
04:14:44 Join Barahir_ [0] (
04:15:04plutoniankrazykit: so it never runs faster than it needs to?
04:15:07krazykitplutonian, for more than that, i'm sure others know better. it may take some time for a better reply
04:15:20krazykitplutonian, right.
04:19:58 Quit blkhawk (Read error: 110 (Connection timed out))
04:21:53 Quit Zom ("leaving")
04:23:01 Quit BHSPitLappy (Remote closed the connection)
04:24:19plutonianis the ipod a dual core platform?
04:24:44 Join Willwolfe [0] (
04:26:10 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
04:29:32 Quit BHSPitLappy (Remote closed the connection)
04:31:40 Quit Barahir (Read error: 110 (Connection timed out))
04:33:40Unhelpfulhrm... global plugin_api for plugins may have been a good deal easier than i thought :)
04:33:48 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
04:36:19 Quit miepchen^schla (Read error: 110 (Connection timed out))
04:36:19Unhelpfulthe "real" work will be adapting plugins to it, i think... assuming that it works on targets as well as it has on the sim, so far.
04:41:58plutonianthe plugins are written in C, right?
04:42:14kadobanplutonian: yes
04:43:44 Quit itcheg (" ajax IRC Client")
04:44:05Unhelpfulamiconn: i believe you were the one who had said that a global plugin_api provided by the loader would be troublesome for the simulator. what i've done is add a const struct plugin_api rb; to the PLUGIN_HEADER, and add a pointer to it to the plugin_header struct. plugin.c is almost completely unchanged, except for a *(hdr->api) = &rockbox_api; added before calling plugin_start.
04:44:29Unhelpfulit works in sim for bounce.c and bubbles.c so far, i'll try building with those two for a target (e200) next.
04:45:17Unhelpfulit will need every plugin modified to work for it, but it will remove the need for any init functions for things in pluginlib, unless they do more than copy the plugin_api.
04:45:41 Join itcheg [0] (i=62db4767@gateway/web/ajax/
04:48:17Unhelpfulit would also mean that MEM_FUNCTION_WRAPPERS can be removed, and a set of wrapper functions put in the pluginlib, to take care of gcc optimizations that insert calls to mem*
04:51:04***Saving seen data "./dancer.seen"
04:55:48Unhelpfulassuming this idea pans out... is there any reason to continue passing the plugin_api to plugin_start? it seems that if, for some reason, a plugin needs to do something special with the API, it can use the magic global API pointer?
04:59:26Unhelpfullooks like both work on e200 as well... i'll start converting some more.
05:07:46 Quit itcheg (" ajax IRC Client")
05:09:17 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
05:15:59 Quit Willwolfe ("ChatZilla 0.9.84 [Firefox 3.0.1/2008070208]")
05:16:36 Quit PaulJam (".")
05:24:51 Quit Horscht ("")
05:32:33 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
05:37:33 Quit Darksair (Read error: 54 (Connection reset by peer))
05:41:43 Join roebeet_ [0] (
05:48:23 Quit saratoga ("CGI:IRC (EOF)")
05:48:41 Quit nibbler (Read error: 113 (No route to host))
05:55:38 Quit roebeet_ ("Ex-Chat")
05:57:35 Join pixelma_ [0] (n=pixelma@rockbox/staff/pixelma)
05:57:48 Quit amiconn (Nick collision from services.)
05:57:50 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
05:59:41 Quit pixelma (Read error: 60 (Operation timed out))
06:11:27 Quit plutonian (Read error: 110 (Connection timed out))
06:22:59 Join TheSphinX^ [0] (
06:23:03 Quit TheSphinX^ (Read error: 104 (Connection reset by peer))
06:51:08***Saving seen data "./dancer.seen"
06:54:29 Part ajb`` ("ERC Version 5.3 (IRC client for Emacs)")
06:58:11 Join Darksair [0] (n=user@
07:22:11 Join Bagderr [0] (n=daniel@gateway/web/cgi-irc/
07:24:01 Nick Bagderr is now known as B4gder (n=daniel@gateway/web/cgi-irc/
07:36:23 Quit agaffney (Read error: 110 (Connection timed out))
07:41:43Calcipheranyone out there have a cowon d2 and fw 2.57? I don't know why I can't build my bootloader, I get a bootloader.elf error
07:52:43lucentCalcipher: are you linking against an OS build?
07:52:45lucentjust guessing
07:53:17*lucent checks in
07:53:25CalcipherI am not very knowlegdable on building in the first place
07:54:05CalcipherI only know how to do the basics with builds, basically just enough to work with voice building, when it wasn't done through rbutil
07:55:16CalcipherI have a pre built rb build from the d2 section of the wiki, and basically only need the patched bl for fw 2.57
07:55:42CalcipherIf I knew what was giving me the build error in this case I would fix it
07:56:14lucentCalcipher: when you do tools/configure you should pick the number for your player model and then "B" letter for the bootloader
07:56:35lucentit works like this for my Sansa AMS unit anyways
07:56:42 Join funman [0] (n=fun@rockbox/developer/funman)
07:57:04Calcipheryeah I did exactly that
07:57:27lucentokay I'm out of suggestions, but I don't know much to suggest anyways
07:58:01Calcipherwhen I ran the make command it processed most of the files but then gave me a long list or errors related to arm sections
07:58:46Calcipherthanks lucent
08:00:04 Quit BHSPitLappy (Remote closed the connection)
08:00:18 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
08:04:52 Quit Bensawsome ("The awsome is gone :(")
08:05:22amiconn_Unhelpful: Hmm, interesting. Somehow I forgot that the loader needs to access the plugin header anyway, even though I added that header a long time ago. :\
08:05:35 Quit Seed ("cu, Andre")
08:05:42amiconn_The header isn't actually a header in case of sim plugins/codecs, but an exported symbol, as sim plugins are shared objects resp. dlls
08:06:00Unhelpfulthe overlay loader is fixed as well, patch up in a few... i don't think i can actually *test* the overlay loader with the right hardware?
08:07:06Unhelpfulright, but since the sim/target split is all before the actual call, and initializes a pointer to the header structure, whether it be an exported symbol, or just some data at the head of the file, there's no difference, as far as i can tell, in how to init the thing.
08:07:15amiconn_So this method should work... and probably the plugin start function should have the api parameter removed?
08:07:30 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
08:07:54amiconn(doing that requires changing all plugins though)
08:08:24Unhelpfulthat's what i thought as well
08:08:45Unhelpfuli *think* i've fixed all plugins, and all pluginlib functions, removing the _init ones that only copy the API
08:09:14amiconnThe greylib copied the rb parameter into its own struct. This could be removed as well
08:09:26Unhelpfuli think i got that
08:10:47Unhelpfulapparently not quite, i fixed the init but left the pointer in the structure...
08:11:10amiconnTesting the overlay stuff should be possible on any target, with some build system adjustment and a proper linker script
08:12:25amiconnUnhelpful: Obviously you also need to fix any function that uses the pointer
08:13:27amiconn(the linker scripts for the plugins loaded via overlay loader currently only support archos)
08:13:48Unhelpfuli think i have... it looks like the only thing i missed was the pointer in the struct, i'll edit and replace the patch in a moment
08:15:02Unhelpfuli'm actually inclined to think that once the initial bugs are shaken out, this will probably be less bug-prone, since you can't forget to call an init function for many pluginlib components now
08:15:17kadobanout of curiosity, will the rb parameter just because global, or are the plugin functions going to become wrapped or something like that?
08:15:49Unhelpfulit's a global, initialized by the plugin loader
08:15:58kadobanah, nice
08:17:13Unhelpfulwrapping the core functions is problematic, because they're already declared, so we can't make the wrappers inline static, which means there's a chance they won't be inlined, and they will also end up having code in the linked plugin, i believe even if all calls to them *do* get inlined.
08:17:52Unhelpfulwe could use #define func rb->func, since the name of the API pointer is guaranteed, but i'm not sure what sort of problems that might end up causing
08:18:15kadobanyeah, it's probably not necessary anyway...i think it'd be more confusing than just using them regularly
08:19:03Unhelpfuli think that macro-wrapping core functions is probably going to remain limited to the use of importing core files into pluginlib
08:20:22funmanI tried FS #9703 on my Clip, it surely cleans up watermark handling but doesn't bring stable playback yet. The truth is somewhere else..
08:21:17*lucent suggests "THE TRUTH IS OUT THERE"
08:21:44B4gderthe truth can of course be in multiple places
08:21:49Unhelpfuland then there's the weird bitmap load issue... which goes away if you just change the read chunk size.
08:21:56B4gderI doubt you can get fine playback without the watermark mess sorted
08:22:26funmanSure, but I had hoped from zagor's description that it was the only work needed :/
08:22:43B4gderI bet he hoped that as well!
08:22:44lucentUnhelpful: does your device display weird pixel values ? I was reading backlog and kugel mentioned something about me by nickname
08:22:55Unhelpfulamiconn: any pointers on where to start if i want to build plugins for overlay on, say, e200? i only have that and gigabeat to test on as well.
08:23:19Unhelpfullucent: i only have e200v1 and beast, and i've never seen any issues like the ones in his screenshots.
08:23:40lucentUnhelpful: I have a different issue that does not show in screenshots
08:24:03lucenttwo dotted and diagnoal lines of pixels with wrong color values but only on the background image
08:24:04Unhelpfulif it's invisible in screenshots, probably an LCD driver thing?
08:24:14Unhelpful.... that's what's showing in his screenshot.
08:24:22lucentwhat that is?
08:24:25Unhelpfulamong other things, that is
08:24:25lucentI didn't see it
08:24:28lucentoh okay
08:24:30Unhelpfullet me find them...
08:25:19Unhelpfulthere's what i suspect to be pixel shift, visible in the rockbox logo, and also two diagonal lines of bad pixels that start somewhere around the logo
08:25:33Unhelpfuland then there's this:
08:25:34lucentholy shit
08:25:40lucentthat's exactly what I'm seeing
08:25:47lucentI could not make a screenshot that captured it though
08:25:52lucentwhich hardware is that?
08:25:55lucent4gb fuze?
08:26:01Unhelpfulwhichever fuze he has
08:26:22lucentthanks for confirming that for me Unhelpful
08:26:31Unhelpfulif i can find the patch...
08:26:34lucentI mean, about the wrong pixel values
08:26:38funmanI remember kugel has a 4GB fuze yes
08:27:30Unhelpfulif this is not an earlier, broken version of it... the patch reverts *only* the read chunk size change in the revision where he identified the problem as starting
08:27:37Unhelpfuland it seems to fix things for him
08:28:09Unhelpfulhe was also able to get read errors in test_disk, he never got to changing the read/write sizes there to see if the same "fix" worked
08:28:14amiconnUnhelpful: The respective *.make files for those subdir plugins have a check for the affected targets, which makes them use in the respective folder instead of the generic This check would need to be removed and and .ovl used for all targets
08:28:43amiconnThen you'd need to change plugins/SOURCES to build the loaders for all targets instead of just archoses
08:29:12Unhelpfulok... do the mentioned scripts need to be edited themselves? that's the part i'd really have little idea how to do.
08:29:14amiconnAnd finally, the various files would need to be extended to support the target(s) you want to test on as well
08:29:15lucentUnhelpful: I'm trying to suss what you're saying, it's difficult for me to know more than "Lucent, test FOO and tell me if it works or not?" :/
08:29:43*amiconn thinks those files should probably be renamed to
08:29:44 Quit Calcipher ("I-n-v-i-s-i-o-n 2.0 Build 3515 with A Pack Fix By")
08:29:55Unhelpfullucent: well, you can try that patch. all it does is read unscaled bitmaps in 8-pixel chunks, instead of whole lines, at a time.
08:30:52Unhelpfulchessbox/rockboy/zxbox are the only ones using it, right now?
08:30:52JdGordondoes leaving the WPS cancel the a-b mode marks?
08:31:05 Join Rob2223 [0] (
08:32:01amiconnUnhelpful: yes
08:34:07amiconnI think that extending the overlay loader for potential use on arm would be a good thing anyway, for targets like the clip
08:35:07Unhelpfulhrm, trying to link the .ovl files gives me this:
08:35:13Unhelpful/home/chshrcat/crossdev/arm-elf/lib/gcc/arm-elf/4.0.3/../../../../arm-elf/bin/ld: target elf32-sh not found
08:35:45 Quit jhulst (Read error: 113 (No route to host))
08:35:48amiconnYeah, since the files are only suitable for sh1 right now, as I already said
08:36:12amiconnThey need to be extended to support our other architectures as well (and should be renamed)
08:36:17Unhelpfuli see that at around 2:29 :/
08:37:07Unhelpfulwhat needs to change, beside the output format?
08:38:15 Join SUSaiyan [0] (
08:39:06amiconnSeveral addresses need adjustment as well
08:39:07 Join daurnimator [0] (
08:39:44 Quit BigBambi (Read error: 113 (No route to host))
08:39:45amiconnI'm thinking about a solution that reuses (via #include) for most things.
08:40:09amiconnIt would need a little extension there, but not much
08:41:56lucentUnhelpful: may I ask you, what is test_disk ?
08:42:17lucent(I might remember the wrong name, it was asked if I passed some kind of disk test)
08:43:09Unhelpfulit's a plugin, i believe it's not built by default, and you'll need to edit apps/plugins/SOURCES
08:44:53Unhelpfulhrm... what's the right output format for arm, i wonder? the obvious elf32-arm or elf-arm do not seem to get me anywhere, and i'm sure there should be a way to get ld to tell me about supported output formats...
08:45:30 Join daurn [0] (n=daurnima@unaffiliated/daurnimator)
08:46:04Unhelpfulthere we go... elf32-{little,big}arm, no idea what i should be using
08:46:46funmansee xx/ ?
08:47:37Unhelpfulhrm, possibly.
08:48:19*B4gder reads the users list is probably the place to look, but of course it has cases for all of our targets.
08:49:20 Quit Rob2222 (Read error: 110 (Connection timed out))
08:49:43*Unhelpful is beginning to think that "test overlays on e200" is not a very simple quick hack... :/
08:50:14 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:50:56 Join ender` [0] (
08:51:06daurnhey all
08:51:10daurnhows the fuze port now
08:51:13***Saving seen data "./dancer.seen"
08:51:16daurnmy fuze just got delivered!!!
08:51:48*lucent swears about the fuze
08:51:50funmanit's being worked on
08:51:58lucentdaurn: which capacity and hardware revision?
08:52:40daurnfuzes have different revisions??
08:52:41Unhelpfulhrm, if i want to have a generic pixel aspect ratio defined as 1:1, and override it for some targets... i'm assuming i'd want the per-target values in config-<target>.h?
08:52:49 Join Willwolfe [0] (
08:52:52 Join bertrik [0] (
08:52:53lucentdaurn: OF version is either 1.x.x.x or 2.x.x.x
08:52:55funmandaurn: yes, see the sandisk forums
08:53:07lucent1.x... is v1 Fuze
08:54:01daurni have .01.01.11A
08:54:29lucentI suggest upgrading the OF right away, 'cause you're in for a bumpy ride making rockbox work on that thing for daily use
08:54:44daurnwhat to?
08:55:05lucentdaurn: check sandisk sansa forums for the fuze model, and get the latest version (.22?)
08:55:13lucentdaurn: do you own any microSD cards?
08:55:16B4gder has 1.1.22
08:55:22ZagorUnhelpful: yes
08:55:23daurnjust one, and its in my phone
08:55:40lucentwhat capacity and type? SD? SDHC? which class
08:55:45Unhelpfuland as far as setting 1:1 values for default?
08:55:58daurnuh, 1gb, sd I assume
08:56:11lucentinteresting, that might work
08:56:23funmanlucent: what's the status of SD cards for fuze (and e200v2?) ?
08:56:24daurnhow to update?
08:56:26lucent8gb sdhc class 6 here and it magically started to kind of work
08:56:36daurncan only find windows updater
08:56:39funmanI thought all should work flawlessly
08:56:40daurnon linux >.<
08:56:52lucentfunman: kind of unknown really, mine used to crash the firmware, now it mostly works
08:57:10funmandaurn: drop the .bin file on your player, then umount and eject it. easy no ?
08:57:17lucentfunman: I could tell you the revision that it crashed the firmware, and the revision as of yesterday it is working
08:57:47funmanlucent: if it works now, then just cross our fingers for no regression ;)
08:58:03lucentfunman: do we care to know why it started working on 8gb fuze?
08:58:14funmani'm looking at the SD code now but I only have internal storage to test with (Clip has no SD slot)
08:58:23lucentfunman: I never asked if anyone wanted me to chortle through and find which exact rev it was that the change happens
08:58:25Unhelpfulthe way i'm thinking of doing things is by premultiplying the input sizes in recalc_dimension... so i should have a multiplier for each dimension, perhaps PAR_H and PAR_W, both defined to 1 by default
08:58:27ZagorUnhelpful: add something like PIXEL_ASPECT_RATIO to config-xxx.h for the targets that don't have 1:1 and then #ifndef PIXEL_ASPECT_RATIO #define PIXEL_ASPECT_RATIO nnn in config.h, below the platform-specific includes
08:58:56funmanlucent: I would say random effects, how would it crash exactly : SD panic RX OVERRUN or something ?
08:59:01Unhelpfulif you'd rather i spell it out instead of using PAR, that's fine, too :)
08:59:08lucentfunman: the CID request failed
08:59:19lucentfunman: do you know what I am saying about?
08:59:53funmanlucent: oh then yes that might be interesting to see when this was fixed
08:59:59lucentUnhelpful: please spell it out, I don't know what the hell PAR means
09:00:07lucentif I don't know, a new coder won't know
09:00:12daurnfunman: its an exe >.<
09:00:16funmanlucent: hum, you are talking about fuze and SD cards, right ?
09:00:29lucentfunman: Fuze 8gb and an 8gb sdhc microSD card class 6
09:00:35lucentI have no other cards to test with
09:00:38funmandaurn: 1/ get firmware on 2/ unzip 3/ copy .bin file
09:00:41ZagorUnhelpful: how about CONFIG_ASPECT_HEIGHT and CONFIG_ASPECT_WIDTH ?
09:01:02 Join flydutch [0] (
09:01:06lucentfunman: the kugel's 4GB Fuze and a smaller card was working for him, while the same thing was crashing with my 8gb fuze and 8gb microSD
09:01:22lucentso... that's why I'm asking if you want a regression test
09:01:23UnhelpfulZagor: those should be fine. i can add a comment explaining *exactly* what they mean.
09:01:31ZagorUnhelpful: yeah
09:01:49lucentit's kind of like flinging noodles at the wall and hope they stick
09:02:40funmanlucent: some changes were made to the code in my absence, mostly related to SD cards
09:03:04lucentdaurn: in brief, the Fuze has an in-built firmware updater which looks for fuzeN.bin where N is some kind of SanDisk code for geographic region, but all these files are identical it's just a different filename to activate the special regional features
09:03:12lucentfunman: okay
09:03:15daurnah k, thx
09:03:31lucentfunman: any suggestion on which revision to start cracking on?
09:03:40funmanlucent: I think r19447 / FS #9560 is what fixed your problem ? (15 december)
09:04:10*lucent reads
09:04:26lucentuh Walloon language file?
09:04:45lucentone of us is doing it wrong.
09:04:58funman9650* sorry
09:04:58ZagorUnhelpful: I think a comment in lib/resize.h would be nice. "why is the declaration conditioned on LCD_DEPTH?"
09:05:57daurnhmmm, my fuze won't connect to my comp >.<
09:05:59daurn[ 2012.607008] usb 7-1: new high speed USB device using ehci_hcd and address 6
09:05:59daurn[ 2013.243705] ehci_hcd 0000:00:1d.7: port 1 reset error -110
09:05:59daurn[ 2013.243713] hub 7-0:1.0: hub_port_status failed (err = -32)
09:06:12lucentfunman: maybe, r19453 I still had panics
09:06:25Lloreandaurn: As always, problems with the original firmware are really outside our scope.
09:06:26lucentfunman: I will be some minutes testing and then tell you more
09:06:38daurndaurn: ya, true
09:06:54funmandaurn: google for "mtp msc fuze" or similar, you might need to force msc mode
09:06:56Lloreandaurn: But you might want to check to make sure it's in MSC and not Auto, assuming it works like the clip
09:07:06UnhelpfulZagor: better just to remove the condition, really... the function itself will never even be compiled for other depths, inside the lib.
09:07:26funmandaurn: writing instructions into the wiki for these issues is very welcome ;)
09:07:32Unhelpfulmaybe a comment in SOURCES, and in bmp.c, explaining why LCD_DEPTH matters?
09:07:42 Join ze [0] (i=ze@
09:08:27lucentdaurn: Fuze OF "Auto" USB mode is likely to attempt MTP connection, you will need to change your OF settings on the Fuze in the System > USB Mode menu to force MSC
09:08:34daurni have
09:08:38ZagorUnhelpful: won't that cause warnings? "declared but never defined"
09:08:47lucentyou could try rmmod'ing ehci-hcd
09:09:09lucentI tell you this stuff works in Ubuntu Jaunty Alpha 2 live cd
09:09:16lucentno rmmod'ery
09:09:34Unhelpfulwill it? i'd imagine it will just always be assumed to reside in some other file, and when it's never referenced on other targets, it will be forgotten about.
09:11:07ZagorUnhelpful: perhaps, I don't remember. try it :-)
09:11:14Unhelpfuland it won't matter if FS #9770 goes in, at some future point.
09:14:24 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
09:14:29Unhelpfulcompiles and links just fine, for e200
09:15:25daurnok, working now
09:15:35daurnis there a way to get rockbox going easily?
09:15:43daurnor do I need to do bootloadery stuff?
09:16:33 Quit funman ("leaving")
09:16:45 Join funman [0] (n=fun@rockbox/developer/funman)
09:17:19lucentoh and check out rbutil/mkamsboot
09:17:22daurnv2 sansa info is sparse
09:17:50lucentyou'll need an OF bin and tools/configure with 58 and B to build the bootloader
09:17:58lucentthen make for the mkamsboot tool
09:18:14lucentuse it to payload the rockbox bootloader onto the OF bin
09:18:16amiconnUnhelpful: I didn't say it's simple, but its certainly doable, and probably even desirable, to port the overlay mechanism to other targets
09:18:26amiconnHence my idea of reusing
09:19:10Unhelpfuli think as far as testing the changes to the overlay loader for FS #9770, though, it might be a good deal easier to bother somebody who has an archos ;)
09:19:11amiconnZagor: Why would you want to remove the fade time settings? I think they are a nice feature if one likes fades in general
09:19:28*amiconn would rather add them for the other fading method(s)
09:21:08Zagoramiconn: because a) I think they are too micro-management, and b) I would like to clean up the backlight fading code
09:21:23 Part akur
09:21:33amiconnI disagree about (a)
09:22:12 Nick blkhawk- is now known as blkhawk (
09:22:37amiconn*If* I enable fading, I want the fade-in *very* short, because otherwise it slows down access and hence annoys me. Fade-out could be quite long
09:22:48amiconnOthers very obviously want longer fade-in times
09:23:08Zagor"very obviously"?
09:23:35amiconnWhen Slasheri added backlight fading, he chose fade-in == fade-out == 2 seconds
09:24:21amiconnThen I added the time settings, which allowed to go down to 500ms minimum. Short enough for me, but preglow wanted it still shorter...
09:24:26 Join petur [50] (n=petur@rockbox/developer/petur)
09:25:07UnhelpfulZagor: does this look good for comments?
09:25:21Unhelpfulamiconn: 2s fade-in sounds horribly annoying
09:25:41Zagoramiconn: are you sure he actually ever used it with 2s fade-in?
09:25:41amiconnAnd if we offer control about backlight time under various condition (on battery, on charger, hold enabled) I fail to see why we shouldn't offer control about fade times as well
09:27:07ZagorUnhelpful: I'm being really picky now, but the first comment (in SOURCES) fails the "document why, not what" test
09:27:11Willwolfedaurn: I found to be a big help. And of course the devs in here.
09:27:30 Quit bertrik (Remote closed the connection)
09:28:05Zagoramiconn: well, there seems to be enough differing opinions about this that I'll drop it for now
09:28:30UnhelpfulZagor: fair enough, i'll fix that and commit
09:28:46ZagorUnhelpful: excellent
09:30:19funmanZagor: I committed r19714 which should recover from problems in SD transfers. I tried FS #9703 but still notice the exact same playback problems, any specific location I could look at?
09:30:56Zagorfunman: what are the problems?
09:31:25funmanskips, crashes
09:32:29B4gderZagor: AMS replied in a friendly fashion but they can't help out on the synposys end, they "have only contact to Synopsys local sales and support"
09:32:49B4gder(imagine that said without the typos)
09:33:07Zagorfunman: yeah, that happens sometimes for me too with 320kbit mp3, but not other codecs or 128 kbit mp3
09:33:48Zagorfunman: I would suspect buffering if it was consistent for codecs. but as it is now I don't really know where to look. blaming the mp3 codec feels rather unlikely.
09:34:01ZagorB4gder: ok
09:34:02 Quit Tristan (Client Quit)
09:34:48Zagormaybe it's some dma thing
09:35:44funmani somehow doubt it since rocks / rockbox / wav files are well transferred
09:36:03 Nick Barahir_ is now known as Barahir (
09:36:05 Join Tristan [0] (
09:36:12daurnum hmm...
09:36:18daurnI just installed the bootloader
09:36:31Zagorfunman: I know. it's puzzling
09:36:33daurnbut it just says File not found
09:36:58daurnoh, thats just the rockbox files not found
09:37:18*daurn needs to read the manual: but, how do you choose between OF and rockbox?
09:37:20Zagorfunman: oh I mean pcm dma, not disk
09:37:23lucentdaurn: it can't read files over 1GB physical location on FAT
09:37:35funmanZagor: couldnt't you reproduce the same problem with a c200 (simulating low memory)
09:37:41lucentdaurn: either USB presence or "left" button held on poweron
09:37:50funmandaurn: there is no manual for fuze yet
09:37:55Zagorfunman: not with FS #9703 applied
09:38:41Zagorthose crashes were watermark related. but obviously that's not the only problem.
09:39:34*Unhelpful wonders if the bmp load corruption issue might be related... would the differing bitrate change the amount of the MP3 file read at once, or not?
09:40:36ZagorUnhelpful: no. the buffering is completely codec agnostic
09:41:24Unhelpfulhrm, right... i forgot that they have a generic buffer in front of them.
09:42:07Zagorbut I agree the two issues might be related
09:43:17Unhelpfulthe bmp loader issue goes away if you make the read size small enough, though, and doesn't seem to happen at all outside of wps/background - kugel loaded a bunch of covers into pictureflow with scaling both enabled and disabled, and didn't see any corrupted.
09:43:50LloreanUnhelpful: Isn't the corruption very consistent, and very late in the image?
09:44:13LloreanWith scaling disabled it might be "offscreen" or out of the visible area, as it were. And with it enabled, scaling could help cover it up?
09:47:15Unhelpfulif you're scaling down far enough, yes, it will cover small pixel shifts, and possibly even a line or two miscolored. but his wps screenshot at has corruption in the first few lines of the backdrop.
09:48:13Unhelpfulalso, bitmaps are actually stored "backwards" vertically, so the bottom of it when displayed is actually the first part of the file
09:48:55Unhelpfulthe file can be stored top-to-bottom, by giving the height a negative sign in the header. none of the files i've looked at do this, but i'm not dumping the row order in the loader.
09:49:09lucentI see the same purple-ish lines on my Fuze 8gb
09:49:36 Join Thundercloud [0] (
09:49:49lucentam willing to test patches if you want to put me in the rat maze :)
09:49:58Unhelpfullucent: could you maybe try reloading the WPS while playing, and see if the corrupted areas change?
09:50:19Unhelpfulthe only patch i have is the one i linked earlier, and it's a coverup, not a real fix
09:50:54 Quit Willwolfe ("ChatZilla 0.9.84 [Firefox 3.0.1/2008070208]")
09:51:14lucentsilly question from me, how to re-load WPS as you want me to try?
09:51:29lucentbrowse and then select a theme again?
09:52:04Unhelpfulyes, though i suspect you might need to load a different one first, to get it to actually reload
09:53:03Unhelpfulif timing affects the corruption, i would think it would be less consistent if you load a new WPS later than when it's loaded during startup... but, i could be wrong.
09:53:06lucentloading iCatcher, then fighting with bug where a "ghost" button is triggered that cancels the menu level I'm at and stops the song
09:53:40pixelmayou shouldn't need to load a different one first
09:54:24pixelmabut who knows what the Fuze does...
09:54:25funmanlucent: by the way, I suppose you can only access 1GB from the storage inside your fuze 8GB ?
09:54:39lucentloading cabbiev2 interrupts my song
09:54:44lucentis that supposed to happen?
09:55:45lucentthere's something sinister with the way this device is accessing data
09:56:05lucentfunman: something like that, I resized my FAT partition to be less than 1GB
09:56:16lucentfunman: I wouldn't know if it does or does not anymore
09:56:26lucentfunman: want to debug that problem with me?
09:56:46funmanlucent: no sorry, solving this problem requires reading the OF code
09:58:19lucentfunman: do you need a disassembly? are you not interested to read disassembly code or is that a statement of "not worth the time"
09:58:58funmana bit of both, and also that I only have a 1GB "working" Clip, so I couldn't debug it
09:59:05lucentoh okay
09:59:07lucentwant a Fuze?
09:59:49funmanyes, i'm thinking about it; but not now
09:59:57lucentif you have the time and motivation, I might contribute
10:00:10lucentthe problem for me is people are too busy to accept a donation
10:00:24funmanI miss both now, and if I decide to get one I can invest and also look for a v2 fuze
10:00:44Unhelpfullucent: interrupts as in stops, or as in pauses and resumes?
10:00:47lucentah yeah I can't find one, I don't know if that is the mysterious red 8gb fuze
10:00:59lucentUnhelpful: stops, then the song skips ahead
10:01:05funmanv2 Clip might help also since there is 4GB / 8GB models (and 4GB v1 Clip)
10:01:15Unhelpfulskips ahead? that's weird
10:01:33lucentUnhelpful: that's "normal" for what I am experiencing with rockbox and my player hardware, just FYI
10:01:40lucentit's totally buggy
10:02:06lucentI've confirmed the same behavior on other 8gb fuze, per some other people volunteering to try rockbox on their Fuze devices
10:02:17Unhelpfuloh... well, a WPS load that changes the album art size, or enables album art when the previous WPS had none, will cause playback to stop and resume, since a rebuffer is currently required to reload AA
10:03:11lucentI think SD access is triggering a button press on my setup
10:03:39lucentkugel once mentioned something about SD sharing I/O pin configuration with the scrollwheel or button light
10:04:19lucentit's a clusterfuck. I don't feel confident saying that WPS loading or graphics or LCD have an issue, to answer your question anyways about the WPS reload
10:04:25lucentthe bad pixel values remain
10:04:45Lloreanlucent: Please, try not to use such language in here.
10:05:17lucentLlorean: pick on somebody else, I'm not in the mood for it tonight :)
10:05:34LloreanIf you're not in the mood for it, behave. Simple solution.
10:05:57*Anges is in a good mood this *morning* :)
10:06:29Mode"#rockbox +o Llorean " by ChanServ (ChanServ@services.)
10:06:56Kick(#rockbox lucent :I'm not joking, and PMing me to tell me you're not interested in the rules is not the way to avoid them.) by Llorean!n=DarkkOne@rockbox/administrator/Llorean
10:07:23Mode"#rockbox -o Llorean " by ChanServ (ChanServ@services.)
10:16:31 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
10:16:45ZagorB4gder: rbclient works with any ssh port, doesn't it? with bygg in my garage it will be on the same address as lillebror.
10:16:57B4gderyes it does
10:17:15B4gderseveral servers are on non-standard ports already
10:27:43 Quit funman (Read error: 54 (Connection reset by peer))
10:30:12JdGordonanyone got a hwcodec that is having problems resuming a shuffled playlist?
10:30:58JdGordonFS #9531
10:32:07 Part Zagor ("Client exiting")
10:32:17 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
10:34:13 Quit Thundercloud (Remote closed the connection)
10:36:31kadobanis there any way at all to profile a plugin? (even on the sim would be helpful)
10:37:18*Unhelpful swears he just saw such a thing
10:37:20B4gderI figure the standard tools should work in the sim
10:38:01kadobanoh yeah...that makes sense :)
10:38:05Unhelpfullook in apps/plugins/lib/profile_plugin.h?
10:38:36Unhelpfulor maybe that's not what it seems to be
10:41:17pixelmaJdGordon: I experience problems resuming shuffled playlists or at least ones that were started be using "insert shuffled" on a folder on different targets, regardless of hwcodec/swcodec. IIRC amiconn noticed first
10:41:22 Join einhirn [0] (
10:41:37jhMikeSTo any concerned, does this look nice enough? (maybe a better function name than "system_except" ??)">
10:46:18ZagorjhMikeS: system_button_wait()? also you might want to change some of the copyright lines. alan korr definitely didn't write that :)
10:47:02pixelmaJdGordon: best to try with a short album or so... start a playlist with "insert shuffled", if the first played track is not the first in the album - note order, stop, resumse -> right track is resumed but not in the right playlist position. It looks like all the tracks in the playlist are shifted as much as it needs to get the first album track to the first position
10:48:09ZagorjhMikeS: otherwise, it looks great. panic.c was rather messy.
10:48:25jhMikeSZagor: Perhaps, but I guess I'll have to revision search to find out who did. :) It really can perform more tasks to shut things off than a simple button wait. It's really to handle unrecoverable stuff.
10:50:34pixelmaJdGordon: just tried to reproduce on my c200 with a folder containing 6 tracks -> insert shuffled gave me: 4, 3, 6, 2, 1, 5 (WPS correctly showed playlist position as 1/6), stop + resume in the first track gives me: track 4 is resumed correctly, playlist position is now 3/6, view current playlist shows 1, 5, 4, 3, 6, 2
10:50:41ZagorjhMikeS: as far as I see, it waits for a button on the targets where that is simple and just halts/loops on other targets. so either it's "wait for button" or just "wait". either way I'd say it's more descriptive than "except".
10:51:03Zagorbut I'm not going to make a fuss about it :)
10:51:17***Saving seen data "./dancer.seen"
10:51:37 Part Anges
10:53:11jhMikeSsystem_exeception_wait ?
10:53:33Zagorsounds good
10:55:11jhMikeSI'm just trying to do these sorts of bits before the target list is so huge it's really, really hard :)
10:56:14 Quit Lss (Read error: 60 (Operation timed out))
10:57:37ZagorI should add a wait for clip
10:58:29jhMikeSI tried to add while (1); to any file not implementing anything so far
10:58:39Zagorok, good
11:16:11 Join lucent [0] (
11:16:34lucentokay, I found the revision that makes SD kind of work on Fuze 8gb, and it confuses me, it is r19617
11:16:51lucentimplement system_reboot
11:17:18lucentwhat happens is I insert microSD card, the Fuze reboots, and the card works
11:17:24lucentthat's totally weird isn't it?
11:17:48Zagorsounds like some init that we don't do properly
11:17:50lucentbefore r19617, inserting an SD card results in *PANIC* card not initialized
11:18:18Zagorafter jhMikeS commit, you'll get that again :)
11:18:28lucentlol maybe
11:18:56lucentis there any useful information I can get from my device in its after-reboot state where SD is working?
11:19:15Zagorthe reason is that when I added system_reboot() I forgot to add a button wait. so it reboots immediately on panic.
11:19:53Lloreanlucent: Does it work if you insert it while powered off, then?
11:20:10LloreanIt would then just be hotswap failing, right?
11:20:51ZagorLlorean: sounds like it, yeah
11:35:42*amiconn hands jhMikeS his 140 points
11:36:46jhMikeSme will spends those soon
11:46:17bluebrotherhmm, zezayer wps is broken. It flashes the status bar over the wps
11:46:44 Quit midkay (Read error: 104 (Connection reset by peer))
11:47:47 Quit Darksair ("Do you hear that? This is the sound of inevitability. This is the sound of your death, Mr. Anderson.")
11:50:00jhMikeS-258 on gigabeat F seems a bit extreme. :\
11:52:12 Join midkay [0] (n=midkay@rockbox/developer/midkay)
11:52:31 Join goffa_ [0] (n=goffa@
12:05:23 Quit goffa (Read error: 110 (Connection timed out))
12:05:27 Join funman [0] (
12:09:06 Join troy_ [0] (n=toppy@
12:12:18 Join DerDome [0] (
12:12:47 Quit BHSPitLappy ("Ex-Chat")
12:17:18Unhelpful...ok, so absolutely all archos bitmap devices have the same tall pixels? i'm about to start mucking about with this aspect ratio stuff...
12:22:09 Join Jaykay [0] (
12:22:16pixelmaAFAIK all bitmap Archoses, yes. I believe that even the charcell pixels of the Player's display aren't square either but shouldn't matter here ;)
12:23:26Unhelpfuldepends, is there a greylib implementation for it via custom characters? ;)
12:23:51Jaykaywhat does the "user:" field in the lang files do?
12:26:04 Quit GodEater_ (" ajax IRC Client")
12:26:14 Join GodEater_ [0] (i=c2cbc962@rockbox/staff/GodEater)
12:26:22 Quit troy___ (Read error: 110 (Connection timed out))
12:27:37Unhelpfuland the av300 is color and has nice, square pixels?
12:28:19*Unhelpful isn't sure about names... i'm tempted to go with LCD_ instead of Zagor's suggested CONFIG_, but that might seem to be the aspect ratio of the display, rather than the pixels :/
12:29:08funmanjhMikeS: r19713 : setting the volume of line input is dependant on HAVE_RECORDING , but we can have FM on line input, and no recording available. shouldn't it be dependant on INPUT_SRC_CAPS ?
12:29:22 Join moos [0] (i=Mustapha@rockbox/staff/moos)
12:29:33LloreanUnhelpful: Do we have any PIXEL_ anything yet?
12:29:48Lloreanor LCD_PIXEL_?
12:29:51 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
12:30:08ZagorI'm fine with LCD_. LCD_PIXEL_ASPECT_RATIO_HEIGHT too. I guess it won't be used often enough to be a nuisance :)
12:30:11pixelmaUnhelpful: IIRC, the AV300 has a hybrid display with a mono and a colour layer but the port is well...
12:30:28bluebrotherpixelma: do you have an idea why the zezayer wps shows the statusbar over the wps itself? It's easy to spot with h100 sim
12:30:34jhMikeSfunman: quite technically, yes but it matches the other stuff that's there for the moment.
12:30:42UnhelpfulLlorean: nope... i think LCD_PIXEL_ASPECT_HEIGHT might be getting a little long, but i don't think it likely to be used anywhere other than inside recalc_dimension
12:31:23JdGordonpixelma: hmm... ok.. can you pastebin the .playlist_contol file?
12:31:31Unhelpfuland maybe a couple of places where things that resize want to define an area of a specific aspect ratio that fits on the screen - pictureflow comes to mind, for one
12:31:32funmanjhMikeS: oh, thanks for the detail
12:31:36*jhMikeS wonders about supressing the other warnings from endian.h in thread-sdl.c
12:32:05pixelmaUnhelpful: CONFIG_LCD already exists but refers to the lcd controller, I believe. At least I remember domonoky having trouble with this in the m200 port as that one has the same controller but different display (?)
12:32:22funmanjhMikeS: did you see FS #9717 ? - about SDL
12:32:38 Quit n1s ()
12:32:43pixelmabluebrother: maybe it's caused by the recent statusbar changes?
12:33:17Unhelpfulgiven the number of targets actually using this, i was going to just define pixels as 4:5 in the configs for archos targets, and 1:1 in config.h if not defined
12:33:39Unhelpfultrying to base it on some other define might be getting too clever
12:33:44pixelmaJdGordon: did you try to reproduce? Might be easier to compare yourself than waiting for someone else to try and paste everything somewhere
12:33:46bluebrotherpossibly, though I have no idea. Cabbie and boxes work fine
12:34:03JdGordonpixelma: no.. i literally just got back to my comp...
12:34:29Unhelpfulalso, looks like dax and m200 share that controller
12:35:20pixelmabluebrother: doesn't the zezeyer WPS use the standard statusbar? Would need to have a dig around first
12:35:39pixelmacabbiev2 and boxes don't
12:36:03jhMikeSfunman: kind of what I had in mind
12:36:08bluebrotherpixelma: no, zezayer does disable the statusbar
12:36:24bluebrothermaybe it's caused by zezayer not defining a viewport?
12:36:54*JdGordon must be doing something wrong
12:37:50 Quit kachna (Read error: 60 (Operation timed out))
12:38:12pixelmaJdGordon: ok, I could paste that one for you but I think you should be able to reproduce
12:38:24pixelmabecause it happens reliably for me
12:38:28JdGordonno.. its ok.. did a dirplay instead of insert shuffled
12:38:43*JdGordon has a good idea whats the problem
12:38:54JdGordonbut no idea how to fix it
12:39:47JdGordonalthough.... stop and resume is working as expected...
12:41:21pixelmabluebrother: hmm, then I have no idea just now why all the others would work but not zezeyer, can't see a big difference in the .wps files at first glance
12:41:36pixelmaor better: something that could explain it
12:42:42JdGordonpixelma: "start a playlist with "insert shuffled", if the first played track is not the first in the album" <- what does that mean? insert shuffled on a folder will always start the first track after the shuffle
12:43:53pixelmabut the first track in the playlist doesn't have to be the first track of the album. In my example it was track 4
12:45:41funmanlucent: could you dump the CSD register reported by your Fuze ?
12:45:54 Quit einhirn (Read error: 110 (Connection timed out))
12:46:16 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
12:46:49JdGordonAAAHHH... no i get it,,,
12:49:31lucentfunman: CSD? how?
12:50:11funmanadd some snprintf/lcd_puts/lcd_update in ata_sd_as3525.c
12:50:33funmanline 306, just after it has been read from the card
12:50:49lucentokay, yeah I can do that
12:51:21***Saving seen data "./dancer.seen"
12:52:25 Quit Jaykay (Read error: 110 (Connection timed out))
12:53:04lucentfunman: I actually don't know how to output to the LCD though
12:54:11lucentfunman: sorry to trouble you, could you write a quick patch? I am not familiar enough to write the code myself :/
12:55:22lucentfunman: also, just to be clear, the SD starts working after a panic triggers a reboot due to a bug in Zagor's patch
12:56:05lucentwithout the panic-induced reboot that code doesn't get farther than CID it fails trying to get the CID
12:56:25lucentsorry, fails trying to send the CID command
12:56:40pixelmabluebrother: did you try other WPSs?
12:57:21Zagorlucent: did you ever answer Lloreans question? 11:19<Llorean> lucent: Does it work if you insert it while powered off, then?
12:57:24funmanlucent: perhaps some delay mising
12:58:25funmansee (I've no clue if the delay is too low or too high though)
12:58:53lucentoh Zagor, if I boot with your buggy patch then the reboot happens on boot and it works, same as inserting card (reboot and then it works) only at boot
12:59:04*lucent looks
12:59:39Zagorjust to be clear: it isn't the patch that is buggy. my commit of system_reboot() simply didn't include a button wait before rebooting.
13:00:26funmanlucent: here i'm talking about internal storage, not about SD cards (try it with no card in the slot)
13:01:03bluebrotherpixelma: yes, others seem to work fine
13:01:30ZagorjhMikeS' commit now added a while(1) after panic so it won't reboot anymore
13:01:38bluebrotherhmm, Rockboxed also has this issue.
13:01:55lucentfunman: I will try soon, in 1-2 minutes
13:02:32bluebrotherso it looks like %wd not working as expected in some cases
13:02:48lucentZagor I don't mean any insult, it is just easy to say buggy vs. fixed for me
13:03:37lucentit's totally by chance that the untimely reboot makes SD card work for me :)
13:05:52lucentfunman: Data abort at 3004DA88 ? hmm
13:06:16Zagorlucent: no harm done, I just wanted to clarify since "zagor's patch" could be interpreted as FS #9703
13:08:42funmanlucent: oops should be snprintf(str, sizeof(str), ..)
13:09:38lucentlol @ that
13:09:54lucentit's also mad about declaring lcd_puts
13:10:00lucentbut if it links it should work Iguess
13:12:10 Quit linuxstb (Read error: 113 (No route to host))
13:14:35lucentfunman the winning lotto numbers ARE: 92404018, BEFBCFFF, 5F5983D3, 260032 (yeah this one is shorter)
13:15:15 Quit DataGhost (Nick collision from services.)
13:15:23 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
13:15:27lucentFYI the SD internal, and I wrote a smaller FAT FS on it incase that makes any difference
13:15:46funmanit is 0x00260032 then
13:16:55funmanthanks, now that'd be interesting to comapre it to a different capacity model (fuze or e200v2)
13:17:42*JdGordon decides the logic for insert shuffled must be broken
13:18:20JdGordonI can only assume this is a really really old bug
13:18:44pixelmaI'd be not surprised
13:19:05JdGordonI first thought it was the shuffel resume handling i changed a while ago.. but its definetly not
13:20:17lucentfunman: I know other people have contacted me who want to help, they could try
13:22:52pixelmaJdGordon: do you remember how long your changes are ago - at least somewhat, a hint?
13:23:11JdGordonno, but should be fairly easy to find...
13:23:20JdGordonsep 08 maybe?
13:24:00pixelmathat one:;revision=18561 ?
13:25:20Zagorlooks like the 8GB Clip is v2 only
13:26:10funmanZagor: yes that's the theory so far, (4GB is v1 or v2, lower is v1)
13:26:37 Join Aurix_Lexico [0] (
13:28:04 Join AndyIL [0] (i=AndyI@
13:29:56advcomp2019funman, there is someone that says there are 1gb and 2gb v2s too but i have not seen them yet
13:30:59 Join Darksair [0] (n=user@
13:31:09 Quit Darksair (Remote closed the connection)
13:31:30pixelmahmm, compiling sims in cygwin seems to be broken somewhat - it fails when compailing Doom with a lot of "undefined reference to `_my_open'" / "undefined reference to `_my_close'". H100 sim, just updated to rev19719, made clean, rerun configure. (and I believe PaulJam reported something similar a few days ago, didn't try back then)
13:31:36 Join Darksair [0] (n=user@
13:31:50 Quit Darksair (Remote closed the connection)
13:32:24funmanpixelma: i'd look into cygwin headers to see if they redefine open()
13:32:52 Join Darksair [0] (n=user@
13:34:08 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
13:38:28 Join Lss [0] (
13:39:05 Quit AndyI (Read error: 110 (Connection timed out))
13:43:15JdGordonpixelma: was the revision so you could test before that?
13:43:48 Join itcheg [0] (i=62db4767@gateway/web/ajax/
13:44:26pixelmapardon? This was when trying to compile a fresh H100 sim to see what the WPSs do (as bluebrother reported)
13:44:52JdGordonno... 20min ago...
13:46:19JdGordonMORE BUILDS!?
13:47:08JdGordonwe really need someone to fix that build page... it doesnt fit on 1680 wide any more :p
13:47:16ZagorJdGordon: zoom!
13:47:57pixelma asked about the revision to test one before (still not a 100% sure if I understand your question ;) )
13:48:01Zagorat 50% it fits in less than 1000px
13:48:24JdGordonpixelma: yes... i was subltly asking if you had tested yet :p
13:48:55*JdGordon is compiling a old build now to test
13:49:31pixelmaI do so too but don't want to test on target, especially on the c200 as I have Rockbox USB enabled
13:49:57pixelmacompiling a sim now (hopefully that works better than the current sim) :)
13:51:05JdGordongrr... it does work
13:51:22pixelmawhich revision did you try?
13:52:03 Join PaulJam [0] (
13:52:41pixelmaI try with r18560 as the description of r18561 sounded related to me
13:53:02 Quit LambdaCalculus37 ("Ka-chunka")
13:53:05JdGordoni went further back than needed... just chose a ncie round number
13:53:31pixelmaproper method would be bisecting then
13:53:56JdGordonno, i'm 99% sure that if it works pre that commit then thats what did it
13:54:01JdGordonalthough i dont know why
13:54:11 Join pipsqeek [0] (
13:56:11JdGordonyep... 18570 doesnt work
13:57:20JdGordonthe problem is... insert shuffled doesnt actually use the random seed which is what that change fiddled with
13:57:38JdGordonin the .playlist_control an insert shuffled file is the same as a regular insert one
13:57:46*JdGordon is confusled
13:59:09pixelmathe playlist is still random for me (even the same order). It's just that everything is shifted in a way that the album's first track is first in the playlist
13:59:47JdGordonwith which r?
14:00:23 Join Jaykay [0] (
14:00:59pixelmalately, r18560 still compiling ;)
14:01:27 Quit itcheg (" ajax IRC Client")
14:01:34JdGordonoh.. hang on.. yeah... its not shuffling at all.. i'm getting 2,3,4,5,...,1
14:02:49 Join nplus [0] (n=nplus@141.25.Globcom.Net)
14:02:59JdGordonsame with 18550.. seems back then insert shuffle only worked if shuffle was actually on
14:04:00pixelmait always worked for me without having shuffle enabled
14:05:19pixelmaI use it a lot since I sometimes want to hear various tracks shuffled but albums in order and turning shuffle on and off is a bit of a hassle (especially without quickscreen :P )
14:05:49JdGordonyeah, I'm sure it used to work as expected also... but this e200 sim isnt shuffling unless shuffle is on
14:06:06JdGordonbut, it does resume correctly regardless of shuffle
14:07:16 Quit Aurix_Lexico (Read error: 110 (Connection timed out))
14:08:58*JdGordon is stumped
14:10:12pixelmalooks like it is still working correctly in r18560...
14:10:45*pixelma "ups" one revision
14:11:00JdGordondoing the actual insert shuffled also?
14:11:33pixelmayes, my playlist looked like this (same folder of music): 2, 6, 3, 1, 4, 5
14:13:44JdGordonI _may_ have found the offending line
14:13:53JdGordonI dunno why it works for oyu and not me though :/
14:14:01*JdGordon does admit to sucking at testing
14:14:21pixelmapractice ;)
14:14:40JdGordonor my comp is being stupid and rand() isnt very rand...
14:14:50JdGordonjust got 2,3,4,5,1 again :/
14:15:57JdGordonpixelma: can you try adding "playlist->first_index = 0;" before the break on line 2348 in playlist.c and see if that fixes things?
14:16:33Jaykayrasher: now around?
14:17:20pixelmaJdGordon: I'll do that next
14:17:47pixelmayou mean in current SVN, yes?
14:18:52JdGordonor even with 18561
14:19:02JdGordonthe line number might be different though
14:19:05JdGordonbut in that general area
14:22:14pixelmaok, first - it's really r18561 that breaks it and with this revision I also saw the unshuffeling (H100 sim)
14:23:50 Quit AndyIL (Read error: 104 (Connection reset by peer))
14:25:00 Join mikeones [0] (
14:26:24mikeonesanyone know what it is you reseat on a e280 v1.1? Someone said it was under the front cover but I don't see anything removable there.
14:27:10 Join AndyI [0] (i=AndyI@
14:27:19JdGordonthe battery maybe?
14:27:38B4gderI don't understand the question
14:27:41mikeonesI think it was advcomp2019 how was telling me this
14:28:16JdGordonmikeones: what made you open it up?
14:28:20 Join kugel [0] (n=chatzill@unaffiliated/kugel)
14:28:25mikeonesI thought he said reseat but it could have been reset
14:28:27 Join fredddy [0] (
14:28:41B4gderwhat exactly do you want to accomplish?
14:28:45mikeonesJdGordon: its bricked an recovery mode does not work
14:28:46pixelmaJdGordon: just to make sure I don't do anything stupid... I see a break on line 2346 (r19719) in playlist.c here, you mean that?
14:29:07advcomp2019it is the daughter card.. you need to remove the main board but did you try the e200tool first tho?
14:29:08B4gdermikeones: then opening it up won't help at all
14:29:22JdGordonpixelma: yeah, inside the block starting if (nread <= 0)
14:30:04mikeonesB4gder: what do you suggest?
14:30:23B4gdersee the unbrick wiki page
14:30:30mikeonesdid all that
14:30:43mikeonestried two diffrent linux computer
14:30:49advcomp2019B4gder, i have seen the seen people staying that reseating the daughter card works.. somehow it gets unplugged, but that is only if they have dropped it
14:31:00mikeonesit is not seen in any of them
14:31:28mikeonesadvcomp2019: nothing else os working so I thoght I would try that
14:31:35B4gdermikeones: you tried e200tool in both of them?
14:31:51mikeonesbut it is not seen in dmesg nor lsusb
14:32:23B4gderyeah, then I'd lean towards hw problems too so reseating the daughter board is worth a shot
14:36:10advcomp2019mikeones, here is the back side of the main board: and the daughter card is the board near the top near the headphone jack
14:37:05 Quit petur ("gone to visit more indians :/")
14:38:49 Part LinusN
14:39:48 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
14:41:15mikeonesadvcomp2019: the one with the barcode that starts 644?
14:41:27 Join tyfoo [0] (
14:41:52advcomp2019yep that is right
14:42:47 Part B4gder
14:45:41pixelmaJdGordon: looks good (first test in the edited r18561 H100 sim), I'm about to put an edited current revision build on my c200
14:47:08JdGordonpixelma: ok, great... im going to bed so when you're ready commit it
14:47:54pixelmaI shall commit something heavy playlist code related? *getting scared*
14:48:00mikeonesadvcomp2019: show how do you get that off?
14:48:02kugelfunman: hey :) Nice to have you back
14:48:19 Join LambdaCalculus37 [0] (i=44a04303@rockbox/staff/LambdaCalculus37)
14:48:36pixelmaJdGordon: works on the c200 too
14:49:28 Nick fxb__ is now known as fxb (
14:49:34JdGordonok, you can have the credit for the fix :)
14:49:42advcomp2019i do not have a photo but you move it back and forth a little bit while lifting it up
14:50:01 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
14:50:48 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
14:51:25***Saving seen data "./dancer.seen"
14:54:48pixelmaI don't even know what it actually does to explain in the commit message...
14:57:48 Join AndyIL [0] (n=pasha_in@
14:58:53 Quit AndyI (Read error: 104 (Connection reset by peer))
15:02:10funmanhey kugel ! have some time for testing?
15:02:40kugelI read the backlogs a bit, you want to know the CID of my msd?
15:02:55funmancsd. you have a 4GB right?
15:03:12kugelfunman: btw: I did the microsd work, and everyone reported success, except lucent, so I assume it's limited to his card
15:03:46ZagorClip v2 BI0810BIIK-8GB. Version V02.01.16F
15:03:54funmanyep, by the way good work for that !
15:04:12funmanZagor: is that something you own ?
15:04:39kugeloh thanks :) I needed some help of gevaerts though, so I don't want to take all credits :P
15:04:40Zagoryep, went out and bought it just now
15:04:59ZagorI hate the mirror front though :-(
15:05:04kugelfunman: oh, and I have a clip (2GB) now
15:05:17*kugel loves this little cute thing
15:05:54LambdaCalculus37kugel: Isn't it adorable? :)
15:06:04funmanZagor: cool, you self designed to write clipv2 support in mkamsboot ;)
15:06:10LambdaCalculus37Powerful sound from such a tiny little player.
15:06:31Zagorwhatever you do, don't buy the silver one
15:06:46kugelLambdaCalculus37: indeed. Impressive that it's basically on par with the fuze w.r.t. to sound
15:07:02LambdaCalculus37Zagor: I have no plans to get the silver one. I like my blue 2GB Clip just fine. :)
15:07:39ZagorI'm glad I don't expect to actually use it much :)
15:07:51*LambdaCalculus37 wonders just how many Clips are going to populate the Tower of Rockbox at DevCon2009
15:08:15Zagordoes anyone know a good way to open it without too much damage?
15:08:21kugelwell, it's going to have it's own (little and cute) tower ;)
15:08:51funmanZagor: start by top or bottom
15:09:19funmanyou should only damage 1 or 2 clips and leave the other safe (it's enough to close it again though)
15:09:31funmanZagor: but we have photos don't we ? no need to open it
15:09:49funmanI'd just write mkamsboot support and see until which points it boots
15:09:59Zagordo we?
15:10:10Zagoropening is more fun! ;-)
15:10:39funmancheck out the forum thread
15:13:37 Join mc2739 [0] (
15:13:38 Join Schmogel [0] (
15:14:33 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
15:15:19mc2739funman: csd for e200v2 = 92404018, BEFBCFFF, 5F5983D3, 00260032 (same as fuze)
15:15:37funmanmc2739: 8GB ?
15:15:55 Quit jhMikeS (Read error: 60 (Operation timed out))
15:16:03Zagorfunman: I see c200v2 photos but no clipv2
15:16:15funmanalright, so I suppose several SD cards make the full storage
15:16:15 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
15:17:11*liiwi ponders epoxying the external 4gb microsd card into his e280v2
15:17:30 Quit amigan_ ("leaving")
15:17:34funmanttp:// the person didn't upload the pictures
15:19:57 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
15:22:48 Join miepchen^schlaf [0] (
15:23:01advcomp2019funman, do you need more c200v2s? i was given one but the headphone jack is messed up tho
15:23:17 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
15:23:18funmani need more developers with a c200v2 ^^
15:23:21kugeladvcomp2019: sure
15:23:42kugelif you donate it, it would be great
15:24:09funmanit only needs finishing lcd support and looking at the buttons if i'm correct
15:24:29kugelpretty much, yea
15:26:43advcomp2019i might be able to ship to a developer that is in the usa, but i will see.. i am thinking of trying to fix that headphone jack tho
15:28:37 Join robin0800 [0] (
15:30:58Zagoraha, it changes usb id depending on usb mode. 0x7434 in auto and MTP, 0x7435 in MSC mode.
15:31:14robin0800patch charge_c200_e200 not sure its working on c200 it shows discharing in debug screen but battery is at 10%
15:31:22funmanZagor: auto is mtp in your case?
15:31:57Zagorauto gives the same id as mtp, but the descriptors are different
15:32:29Zagorthe SansaClip lsusb dump is what it looks like in auto mode. mtp and msc look slightly different.
15:33:03 Join roebeet_ [0] (
15:34:24kugelI should get a panic, but I don't get one
15:34:35Zagorkugel: when?
15:34:56 Join amigan [0] (
15:35:07kugelanytime the microsd is initialized
15:35:12funmanZagor: by the way when you look at mkamsboot don't forget the clipv2 firmware is slightly bigger (300kB+) so check the nrv2e block size, and test your firmware in a debugger (like skyeye) prior to target
15:35:34funmankugel: i replaced the panics by a retry this morning
15:36:02funmanhm ok (by the way I need the internal storage CSD)
15:36:14kugelfunman: oh
15:36:31funmanand panicf() can take format specifiers (no need to use snprintf)
15:36:53kugelfunman: panicf() does nothing anymroe?
15:38:52 Quit roebeet_ ("Ex-Chat")
15:38:57funmani meant i remove the panicf() from SD driver
15:40:23kugelah lol, call me stupid
15:40:40 Join Barahir_ [0] (
15:41:25kugelsame csd as mc2739 reported, indeed
15:41:29funmankugel: did you try faking the internal storage size ? (by hijacking the calculations using the csd register)
15:41:41*kugel mv'd rockbox.bin instead of rockbox.sansa to the player
15:41:49kugelfunman: yea, doesn't work
15:42:01 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
15:42:35funmanthere is a (quite short) arithmetic function which i'd like to understand
15:43:08kugelI mean, ~4GB is reported, but copying >3GB still stopped after ~800MB
15:43:21 Join einhirn [0] (
15:43:21funman43 instructions, take the bits 95:64 from the CSD in argument
15:44:49Zagorthe clip even says v2 in the scsi disk name: scsi 4:0:0:0: Direct-Access SanDisk Sansa Clip 8GB v02. PQ: 0 ANSI: 0
15:45:10ZagorI never noticed what the v1 says there
15:49:48ZagorI don't recall that the v1 has a PODCASTS dir, does it?
15:50:10advcomp2019Zagor, yes it does in the newer firmware
15:50:20Zagorand a version.sdk?
15:50:46advcomp2019yea that is on the v1s too
15:52:52 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
15:53:17 Quit Jaykay (Read error: 110 (Connection timed out))
15:55:36 Quit jhulst (Read error: 104 (Connection reset by peer))
15:55:46 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
15:56:19 Quit mc2739 ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
15:56:59 Quit Barahir (Read error: 110 (Connection timed out))
16:00:35 Nick Barahir_ is now known as Barahir (
16:04:28 Nick fxb is now known as fxb__ (
16:09:40Zagorfunman: slightly bigger? it's 15 frickin megs!
16:09:44 Join kachna [0] (
16:10:19*linuxstb hands Zagor arm-elf-objdump
16:11:07 Quit robin0800 (Remote closed the connection)
16:11:15funmanZagor: i meant the part we compress
16:11:33funmani.e. 'firmware' as opposed to 'library block' or 'header block'
16:14:02 Join TheSphinX^ [0] (
16:25:13 Join BdN3504 [0] (n=55b2348e@gateway/web/cgi-irc/
16:25:42Zagorfunman: does the v1 have any sort of recovery if I flash a botched bootloader?
16:26:10BdN3504sorry to bother you guys again, but i have a semi bricked sansa e200. can you recommend anyone i could turn to with my problems?
16:26:19linuxstbZagor: afaik, the e200v2 is the only AMS Sansa with a known recovery mode.
16:26:26Zagorlinuxstb: ok
16:26:29funmanZagor: yes, built in mkamsboot
16:26:46Zagorfunman: ?
16:27:37funmanif by 'bootloader' you mean bootloader-clip.sansa ,you could use /dev/random with a correct mkamsboot binary and still recover
16:27:40linuxstbBdN3504: is about all anyone knows...
16:27:49funmanif by 'bootloader' you mean the m300.bin file produced by mkamsboot, you're doomed
16:28:33linuxstbZagor: mkamsboot contains a small "pre-bootloader" which is designed to be very small/reliable. That does the dual-booting, and loads the Rockbox bootloader or OF. The source code is in utils/mkamsboot/dualboot.S
16:28:38funmanlet me look if i can a clipv2 diff
16:29:01funmanINSERT find in sentence
16:29:03 Quit kachna (Read error: 110 (Connection timed out))
16:29:43Zagorlinuxstb: it still requires that we get that right the first time
16:30:50 Join toffe82 [0] (
16:31:04funmanZagor: note that in the Clipv2 code is executed from 0x30000000
16:31:20funmansince the firmware doesn't fit in the 320kB
16:31:39funmanit's written in SDRAM by the internal AMS bootloader
16:32:40linuxstbMaybe the Clipv2 has a recovery mode though - or have people tried?
16:33:18Zagorfunman: where is this information?
16:33:23funmanZagor: sorry i can't find it, the format is simple however : an arbitrary 32 bits word have been inserted and there is a simple checksum in each block
16:33:42funmanZagor: in my brain (from reading the clipv2 firmware)
16:34:05Zagorah. you wouldn't mind jotting down some highlights in twiki?
16:34:05funman*in each of the 2 first 0x200 bytes block of the 15MB firmware file
16:34:29funmani would, i already put the interesting parts in the forum thread
16:34:54funmani can find the link to it though
16:35:01Zagorthe forum is a horrible place to store&search information
16:35:28funmani use it to exchange with other people
16:35:42funmanhoping that someones step up to do the work ;)
16:35:47BdN3504<linuxstb> thanks, i got it fixed. was a bit weird, i put an old sansa firmware on the 16mb partition hoping it would remove the database rebuild. next thing i know is i when i reconnect the device, the rockbox bootloader shows the rb splashscreen and the boot up sequence then begins to flicker with a white light. i couldn't access the sansa as a drive anymore... but i could still access it via recovery mode so
16:38:39 Join MethoS- [0] (
16:39:08 Quit BdN3504 ("CGI:IRC (EOF)")
16:39:56funmanperhaps looking if word at 0x20 is not 0xffffffff is enough to identify clipv2 (or fuzev2 which uses the same format!)
16:40:31linuxstbfunman: mkamsboot now uses the md5sum of the firmware to identify it....
16:41:46 Join tvelocity [0] (
16:41:55 Join Jaykay [0] (
16:44:12 Join jgarvey [0] (
16:44:55funmanlinuxstb: oh right, that's easier
16:45:09funmanbut the md5sum can be ignored can't it ?
16:46:01 Join evilnick [0] (i=0c140464@gateway/web/ajax/
16:51:05Jaykayis there any reson why isnt closed? it was committed in r19673...
16:51:26***Saving seen data "./dancer.seen"
16:51:37funmanZagor: hum mkamsboot already has support for v2 firmware format ..
16:52:02Zagorbut not the clipv2 firmware md5sum
16:52:18funmansure since nobody tested it ^^
16:52:55Zagorand what do you mean v2 runs at 0x30000000. so does v1.
16:53:40funmansee line 0x606 for instance
16:53:52funmanwell no
16:54:25funmanon v1: the internal bootloader loads the firmware block (which is always < 128kB on clipv1, e200v2, c200v2, fuze ..) in IRAM at offset 0 and executes it
16:54:43funmanon Clipv2 (and probably Fuzev2) it loads the > 320kB firmware block in SDRAM at offset 0x30000000
16:54:58Zagorah, ok. that makes more sense.
16:55:30funmanhum Zagor we forgot something
16:55:53funmanyou need first to use an older mkamsboot with a simple test.S in order to find a button / or usb connection for dualboot
16:56:01funmangpioa pin3 is a good first candidate
16:56:56funmandualboot.S should be built as position independant code, but extra details can't harm
16:57:11*funman scratches head
16:57:39 Quit fredddy (Remote closed the connection)
16:57:39funmanin fact the firmware block is loaded in SDRAM, right, but the SDRAM is mapped at 0x0 so both excecute from the same place
16:58:29kugelram is mapped to 0x0? didn't the datasheet say it must be 0x30000000?
16:58:30funmanit will cause problems because clipv1 linker scripts assume 0x0 is the IRAM
16:58:56funmanthere is several aliases, 0x30000000 is always the SDRAM and 0x0 can be the ROM the IRAM or the SDRAM
16:59:59Zagoras3525 boot.lsd only uses IRAM so does it really matter if that is in fact iram or sdram?
17:00:20Zagorapart from the size, of course
17:02:03funmanyes it does, we should use an alias for IRAM, and copy the vectors; to not overwrite ourselves when we load the main rockbox into sdram
17:02:26funmanbootloader/sansa_as3525.c needs to be checked as well
17:03:01funmane.g. use 0x30000020 to be sure to not overwrite vectors,
17:03:07 Quit TheSphinX^ ("XChat@Linux")
17:03:11*pixelma doesn't know what to do with JdGordon's fix
17:03:12funman(as the rockbox loading address)
17:03:28Zagorfair enough
17:03:29 Join TheSphinX^ [0] (
17:04:16pixelmaI'd like to commit it because it looks very simple and fixes a quite annoying bug I couldn't explain what it actually does in the commit message. He just didn't want to commit himself because he was about to leave
17:04:39 Quit Zagor ("Client exiting")
17:06:33 Quit tvelocity (Read error: 110 (Connection timed out))
17:06:40linuxstbfunman: dualboot.S could just be used to extract the images to the correct location - IRAM for the RB bootloader, or DRAM for the OF. So as far as the main Rockbox build is concerned, there would be no difference.
17:07:42Jaykaywhat about it was committed in 19673 and isnt closed
17:07:53 Quit einhirn (Read error: 54 (Connection reset by peer))
17:08:14funmanlinuxstb: but the bootloader/sansa_as3525.c assumes we run from IRAM and DRAM is entirely at our disposition
17:08:20 Join krazykit` [0] (
17:08:56funmanright, use 0x0 for the OF and 0xIRAMALIAS for RB should be plain enough
17:10:08 Join tyfoo2 [0] (
17:10:41 Join tvelocity [0] (
17:11:14moospixelma: simply mention the bug that fixed and the FS# maybe, or of course what JdGordon wake up...
17:12:13 Quit m0f0x (Read error: 110 (Connection timed out))
17:12:23linuxstbfunman: What is "0xIRAMALIAS" ? I thought the RB bootloader ran from IRAM (i.e. 0x0), and the main RB build from DRAM?
17:12:53funmanlinuxstb: you're right
17:13:03funmanin the Clipv2, address 0x0 is the DRAM however
17:13:19pixelmamoos: he said I could commit it but didn't have anything to do with this type of code yet (I tested for him though)
17:13:29linuxstbfunman: Ah, I misunderstood that part....
17:13:33kugelfunman: isn't 0x30000000 still valid?
17:13:40funmansorry to be confusing
17:13:47linuxstbfunman: So we would need to remap IRAM back to 0x0 before using it?
17:13:50funmankugel yes it is, it's a fixed alias
17:13:57funmanlinuxstb: i'm not sure if we can
17:14:07*pixelma wonders why some words seem to be stolen from her sentences
17:14:11 Quit TheSphinX^ ("XChat@Linux")
17:15:00moospixelma: at least you tested it, and seems to fix a bug, right? Isn't it a few lines patch too?
17:15:01linuxstbfunman: Is it the same AMS SoC in the Clipv2
17:15:08 Quit tyfoo (Read error: 145 (Connection timed out))
17:15:08 Nick tyfoo2 is now known as tyfoo (
17:15:15pixelmamoos: I one liner
17:15:41funmanlinuxstb: i think it's an upgrade to it (armv5)
17:15:49 Quit MethoS- (Remote closed the connection)
17:16:17linuxstbfunman: That's what I thought... So we need Bagder to send another nice email to AMS?
17:16:57moospiexelma: then go go go, mention the FS# and JdGordon and you are done
17:17:13funmanlinuxstb: from my side it looks compatible
17:17:23 Join TheSphinX^ [0] (
17:18:51funmanperhaps the difference between clipv1 & v2 is in CCU_MEMMAP register
17:20:04moospixelma: exemple: One liner patch by JdGordon to fix FS#XXX ?
17:20:41funmanby the way the codecs i couldn't link on the clip are aac ape and spc (I could add mpc wma sid nsf and asap)
17:20:50pixelmamoos: thanks for encouraging me, I'll do that then. Hope for no red... :)
17:21:46moospixelma: hehe :) At worst blame JdGordon (/me is joking)
17:22:18moosnice to see you around funman btw
17:22:35 Quit krazykit (Read error: 110 (Connection timed out))
17:22:57funmanthanks ^^ i just wish i could fix more bugs while i'm here
17:23:13 Join gartral1 [0] (n=Gartral@
17:23:21mooshehe, sure you will :)
17:27:23 Join m0f0x [0] (n=m0f0x@
17:27:36 Quit funman ("leaving")
17:28:16*linuxstb thinks pixelma is optimistic ;)
17:29:21pixelmawhy that?
17:29:42linuxstbA patch by JdGordon that results in a green build table first-time? ;)
17:30:27pixelmawell, it was tested by me :P (only test built one sim and one target though)
17:30:34gartral1building the rockbox sim for an e200 v1 fails... for lack of a better word, epically
17:30:38*linuxstb takes that back ;)
17:31:58*pixelma relieved
17:32:27Lloreangartral1: Try "make clean" or "make veryclean"?
17:32:27mooshehe :)
17:33:01moospixelma: let's hope it doesn't add bug somewhere else, hihi ;)
17:33:25pixelmaspeaking of compiling sims - had a problem compiling a sim in cygwin earlier today (after make clean and rerun configure, see logs
17:33:41bluebrothergrrr. Another rbutil bootloader install bug fixed. Why can't people state exact information in the first place?
17:33:56pixelmamoos: I'll let JdGordon|zzz handle that then ;)
17:37:49domonokybluebrother: time for rbutil 1.0.10 ? :-)
17:38:06bluebrotherI want to have a check for an updated version from within rbutil. Is it a good idea to check against a file in svn via viewvc?
17:38:15bluebrotherdomonoky: unfortunately I'd say yes :(
17:38:36bluebrotherbut maybe I can get some small improvements finished first. Let's see ...
17:38:50Jaykaybluebrother: does rbutil need any testing a normal user can do?
17:39:14domonokyyes, if there are other improvements in preparations, we probably should wait a bit with the new release.
17:39:17bluebrotherJaykay: well, you can "just" use it and check if something breaks :)
17:39:49bluebrotherunfortunately testing is a bit nasty due to the number of supported targets (and the fact that I only own a fraction of those)
17:40:15bluebrotherI was thinking about providing svn builds every now and then in the future to have more testers
17:40:28domonokyJaykay: especially testing the bootloader installs would be good
17:40:46Jaykaybluebrother: didnt get it, sorry
17:41:11bluebrotherI _hope_ that was the last bug in bootloader installs. Let's see :/
17:41:15Jaykaydomonoky: so i should deinstall rockbox incl the bootloader and install it again
17:41:19domonokybluebrother: maybe it would be better to just check a file on the download server, which we update when ever we put a new release there.
17:41:56bluebrotherdomonoky: that was the alternative I was thinking about. OTOH if we check against version.h we could scrap that additional file.
17:42:21domonokybluebrother: also remember that we dont always get new binarys for all plattforms at the same time.
17:43:03bluebrothertrue. But that would require updating the versions file every time a new binary gets uploaded. I somewhat fear it getting forgotten every now and then
17:43:33domonokyhm, checking the filelisting of the server would be good :-)
17:43:44bluebrotherand afaik we can't put a php script helping with that on the download server
17:43:55bluebrotherchecking the file listing is a bit more work ;-)
17:43:57Jaykaydoes anybody want to close its committed already
17:44:01bluebrotherbut I considered that too.
17:44:22linuxstbbluebrother: What are you talking about, making rbutil check for new versions of itself?
17:44:42bluebrotherlinuxstb: yes
17:45:32bluebrotherat least having a menu item to do that, and if there's a new version present the user a download link.
17:46:10bluebrothermaybe we can extend that later to download by itself. Though I don't think that makes much of a difference ...
17:46:15linuxstbHaving a "current.txt" on the download server seems most reliable - i.e. the worst that can happen is that it will not be updated. It will never contain a newer version than is actually available on that particular mirror.
17:47:08*Jaykay wonders why everybody is ignoring him...
17:47:42linuxstbJaykay: Your nick is too short - it gets missed between "bluebrother" lines...
17:48:14Jaykayreally.... i'd say his nick is too long
17:48:23 Nick bluebrother is now known as blueb (n=dom@rockbox/developer/bluebrother)
17:48:29 Nick linuxstb is now known as stb (n=linuxstb@rockbox/developer/linuxstb)
17:48:46 Nick LambdaCalculus37 is now known as LC37 (i=44a04303@rockbox/staff/LambdaCalculus37)
17:48:53LC37Jaykay: Satisfied? ;)
17:48:58 Nick LC37 is now known as LambdaCalculus37 (i=44a04303@rockbox/staff/LambdaCalculus37)
17:48:58 Nick Llorean is now known as Llor (n=DarkkOne@rockbox/administrator/Llorean)
17:49:32 Nick domonoky is now known as dom (n=Domonoky@rockbox/developer/domonoky)
17:49:35bluebhmm. We got a patch accepted from someone without real name?
17:49:42 Nick dom is now known as domonoky (n=Domonoky@rockbox/developer/domonoky)
17:49:48 Nick stb is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
17:49:52linuxstbblueb: We did?
17:50:15bluebat least the initial patch author doesn't have a full name in the tracker
17:50:17 Nick Llor is now known as Llorean (n=DarkkOne@rockbox/administrator/Llorean)
17:50:19 Nick blueb is now known as bluebrother (n=dom@rockbox/developer/bluebrother)
17:50:30*bluebrother would like to see a real name policy in the tracker
17:51:15*linuxstb too - I think we can live without the benefit of anonymous bug reports
17:51:23*Llorean agrees.
17:51:29 Join {phoenix} [0] (
17:51:45*Jaykay doesnt know if his vote counts but agrees too.
17:53:00 Quit linuxstb ("Leaving")
18:01:26 Join robin0800 [0] (
18:05:24robin0800jhMikeS: c200 charging patch dosen't kick in on my c250 it appears to have 4.765v even at 10% is your 4.16v correct?
18:06:12*LambdaCalculus37 also agrees on the real name policy
18:13:07 Join m0f0x_ [0] (n=m0f0x@
18:16:18 Quit m0f0x_ (Client Quit)
18:21:17 Quit tchan (Remote closed the connection)
18:21:54 Join tchan [0] (
18:23:55 Join fredddy [0] (
18:27:24 Join karashata [0] (n=karashat@
18:28:21 Quit m0f0x (Read error: 110 (Connection timed out))
18:30:02 Quit Nico_P (Remote closed the connection)
18:30:31moospixelma: you should close the related task to your last commit
18:31:38 Quit jhulst (Read error: 104 (Connection reset by peer))
18:33:14robin0800jhMikeS: 3.7v unplugged 4.7v as soon as plugged in might be seeing the charger not the battery?
18:36:02 Nick fxb__ is now known as fxb (
18:37:20 Join bertrik [0] (
18:38:07 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
18:39:51 Quit Jaykay (Read error: 110 (Connection timed out))
18:39:57 Nick krazykit` is now known as krazykit (
18:44:00 Quit robin0800 (Remote closed the connection)
18:45:19gartral1is this patch for e200s?
18:46:45 Join robin0800 [0] (
18:46:57 Join MethoS- [0] (
18:47:58gartral1i see a functional charge..
18:50:10 Join BigBambi [0] (
18:51:30***Saving seen data "./dancer.seen"
18:52:16 Quit TheSphinX^ ("XChat@Linux")
18:55:42 Join BXCracer [0] (
18:56:14 Join tessarakt [0] (
18:56:29 Quit tessarakt (Read error: 104 (Connection reset by peer))
18:59:04 Join kachna [0] (
19:02:47 Quit jhulst (Read error: 104 (Connection reset by peer))
19:04:37 Join plutonian [0] (n=plutonia@
19:06:32 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
19:09:06 Quit miepchen^schlaf ()
19:10:57 Join miepchen^schlaf [0] (
19:15:37 Quit Darksair ("Emacs = ESC-Meta-Alt-Ctrl-Shift")
19:29:03 Quit MethoS- (Remote closed the connection)
19:33:35gartral1rockdoom wont let me press the wall bottons..
19:34:07 Join djtremolo [0] (
19:34:43 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:34:46djtremolohello guys.. one question :)
19:35:17gartral1please dont ask to ask, just ask
19:35:23djtremoloI have an old archos recorder v1 and I've done RomBox flashing far ago
19:36:07djtremoloNow, when I try to install 3.1 version, this old version just keeps starting even I've removed the .rockbox folder and *.ajz from the root
19:36:30djtremoloHow can I "clean" the bootloader to forget the rombox?
19:36:58domonokyif flashed rockbox, you ofcourse also need to update the flash...
19:37:37 Join bertrik_ [0] (
19:37:43djtremoloI used the graphical tool for updating
19:38:15djtremoloIt works when I manually start the /ajbrec.ajz, but not automatically at boot-up. What do I miss?
19:38:25domonokythe flash update :-)
19:38:38djtremolook... How to do that?
19:38:47 Join funman [0] (n=fun@rockbox/developer/funman)
19:38:47*domonoky searches the wiki for instructions...
19:38:49 Quit bertrik (Read error: 113 (No route to host))
19:39:14 Join Thundercloud [0] (
19:39:33gartral1why wont rockdoom let me hit buttons....
19:40:07Lloreangartral1: Do you mean "why, when I hit buttons in doom, won't they do anything"?
19:40:35gartral1the wall pads that your supposed to punch
19:40:35LloreanOr do you mean "while I'm playing Doom, why can't I hit the buttons in the levels"?
19:40:42LloreanYou're not supposed to punch them.
19:40:52LloreanYou use the "use" button to activate them, just like doors...
19:42:09domonokydjtremolo: contains instructions on how to update the flash. But i dont know if you will get additional problems, because you may have a very old rombox in flash.
19:42:34*gartral1 can effectively be labeled an imbecile for not assigning a key to use
19:43:24gartral1ohh.. use isnt listed
19:43:26djtremolodomonoky: thanks. I've thought the bootbox and rombox is "yesterday" stuff.. that's why I was not brave enough to try that one. I'll see what happens
19:43:56gartral1its game open, right?
19:44:24domonokyas far as i know, bootbox still work. But i dont know if there is still a "rombox" available. But the normal flash image should work.
19:45:20 Join webguest [0] (n=c38975bb@gateway/web/cgi-irc/
19:45:32djtremolodomonoky: I read from somewhere that you can still build the sw by yourself to get a rombox version, but you have to disable the database feature...
19:46:25domonokydjtremolo: yes, rockbox has grown too big to support rombox officially (at least for archos recorders)
19:47:14 Join elinenbe [0] (n=elinenbe@
19:51:44 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
19:52:43djtremolodomonoky: Cool! The instructions asked to run "rombox.ucl", I run "rockbox.ucl", otherwise the instructions worked OK. Thank you.
19:52:55 Quit Xerion (" ")
19:54:10djtremoloI've been following the Rockbox development since it's start but haven't updated for about two years as the sw was good enough for me. Now I wanted to check out what has happened here. The Rockbox team has done good work.
19:54:19 Quit webguest ("CGI:IRC (EOF)")
19:55:22LambdaCalculus37Thanks. :)
20:04:34 Quit plutonian (Read error: 104 (Connection reset by peer))
20:09:52 Nick bertrik_ is now known as bertrik (
20:12:10 Join Bagder_ [0] (
20:21:45 Quit ender` (Read error: 104 (Connection reset by peer))
20:22:27 Quit parafin (Read error: 60 (Operation timed out))
20:23:06 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
20:28:57 Part djtremolo
20:28:57 Quit mikeones ("leaving")
20:28:57 Join Zagor [242] (
20:29:13 Join parafin [0] (
20:29:29 Quit linuxstb (Read error: 110 (Connection timed out))
20:30:06 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
20:33:09 Quit fredddy (Remote closed the connection)
20:33:49 Quit linuxstb_ ("Leaving")
20:34:02 Quit Bagder (Read error: 110 (Connection timed out))
20:34:09 Quit jhulst (Remote closed the connection)
20:36:11Zagorcowon s9: 55 hours battery life!
20:40:30 Quit robin0800 (Read error: 60 (Operation timed out))
20:42:39 Join robin0800 [0] (
20:44:43gartral1e250 25
20:44:49gartral1hours batt life*
20:44:51robin0800two comments on FS #8363 now say still not working on c200
20:48:08 Quit Acksaw (Read error: 104 (Connection reset by peer))
20:50:35 Join Acksaw [0] (
20:51:31***Saving seen data "./dancer.seen"
20:53:33*amiconn blames Zagor's r19704 for the cygwin sim problems
20:55:08PaulJamamiconn: have you seen FS #9771 ?
20:56:38*amiconn goes checking
20:56:49amiconnI don't check the tracker regularly 'cause it's so slow
20:57:16*gartral1 has never had problums with tracker
20:57:56ZagorPaulJam: thanks, I had missed that too
20:58:08*Zagor fixes
20:58:25amiconnCalling a single tracker task takes >30 seconds here
20:58:43Zagoramiconn: but the site isn't slow?
20:59:15Zagorhave you tested with another browser? maybe it's some weird css/javascript issue?
20:59:44amiconnNot yet. I'm using FF 3.0.5 fwiw
21:02:56amiconnSame problem in IE7 and SRware Iron
21:03:14bluebrothermaybe a virus killer issue?
21:03:54 Join Acky [0] (
21:06:05 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
21:06:08amiconnUnlikely. On my linux box it's a bit faster (Iceweasel 3.0.5) but still *a lot* slower than the main page
21:06:41Zagorreally strange
21:08:04PaulJami had noticed very long loading times too, with my sisters laptop during the holidays.
21:08:14amiconnYeah, especially since the tracker only fetches files from (as opposed to the main page, which has some external logos)
21:08:25 Quit funman ("Lost terminal")
21:10:22 Join Xerion [0] (
21:10:41amiconnUnfortunately firebug still wasn't updated for FF3. It would be very useful for finding the cause of the slowness
21:12:10 Join gleesond [0] (n=gleesond@
21:12:13 Part gleesond
21:13:24 Quit karashata ("G'bye everyone!")
21:13:27PaulJamif the problem isn'r restricted to FF3, couldnt you just use an older version of the browser?
21:14:30Zagordoes anyone know why doom doesn't use its' my_open in the simulator?
21:16:54BigBambiZagor: You don't need the sim, just run doom natively :)
21:20:07 Quit Acksaw (No route to host)
21:20:17 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
21:21:00 Join ender` [0] (
21:21:08amiconnZagor: Several .js files take ages for fetching, and additionally many files are fetched in sequence rather than in parallel
21:21:46 Join Anges [0] (
21:22:30*amiconn retries on linux to rule out av delays
21:23:21jhMikeSrobin0800: It will settle at 4.16-4.18V after charging for me. 4.765V would be a wrecked battery so that doesn't sound right.
21:24:20 Join Strife89 [0] (n=michael@
21:25:03 Quit elinenbe (" HydraIRC -> <- IRC for those that like to be different")
21:25:17jhMikeSAlso, I thought reading RTCSUPP was decided on as being correct already for both e200 and c200, not just e200 (which reads properly).
21:25:24bertrik4.765 sounds like the voltage that the c200 shows when the USB cable is connected
21:25:45pipsqeekg'day everyone
21:26:01amiconnSame problem. Highest delay is due to prototype.js. That one takes ~15sec on linux, and ~25sec on windows
21:26:02jhMikeSso does e200 unless you actually read the battery (RTCSUPP)
21:26:29amiconnIt's a 124KB file, which should load in less than a second with a 3Mbit/s line...
21:27:44pipsqeekhow's everyone going?
21:28:46Zagoramiconn: yikes, that is a LONG time! does it take the same time if you fetch it with, say, curl?
21:29:03BigBambipipsqeek: Fine thanks, but this is an on topic support and development channel :)
21:29:05jhMikeSbertrik: I thought you had that part worked out already
21:29:14 Join Aurix_Lexico [0] (
21:30:29*amiconn tries to understand curl argument syntax
21:30:33jhMikeSbertrik: or was that only for e200?
21:30:51*bluebrother suggests wget and runs
21:31:26pipsqeekBigBambi: ah. fair enough. I just installed Rockbox on a dead as nails 3rd Gen iPod and it works perfectly, apart from a battery that only lasts 3 hours instead of the advertised 8. Unlike the 4G Colour Clickwheel model that seems to last longer than advertised. was looking around the website and thought I'd jump in the #chan to say hello and see what's happening. I guess I'll only pop in when I have an issue.
21:31:37 Quit timc`` (Operation timed out)
21:31:42Zagoramiconn: curl -O <url>
21:31:42bertrikjhMikeS, IIRC, on e200 you can get the correct voltage from RTCSUP even when the charger is connected, c200 has no way to see the correct voltage with the charger connected
21:31:46BigBambipipsqeek: There is #rockbox-community :)
21:32:19*amiconn tries with wget on linux
21:32:23pipsqeekah, maybe that should be put on the site :) I only see a mention of this channel
21:32:28BigBambiI suspect your battery issue is just that the battery is old, Rockbox on the 3G should get longer runtime than the Apple firmware
21:32:32amiconnSame slowness. A whopping 12KB/s download speed
21:32:37BigBambipipsqeek: It is in the topic for this channel
21:33:04pipsqeekah, I'm using irssi and have turned off topic view.
21:33:11 Join karashata [0] (n=karashat@
21:33:17jhMikeSbertrik: aha, ok. I suppose you just start it when plugged and handle things a bit differently. I prefer to be cautious and not over recharging but I guess its unavoidable.
21:33:32 Quit robin0800 (Read error: 60 (Operation timed out))
21:33:57jhMikeSIt's not a big deal to treat them both a bit differently I guess.
21:34:09pipsqeekwell, the site has bad wording anyway. "We often hang out on #rockbox....." if it's only for topic support and dev then it should say so. and then specify general chat can be on #rockbox-community instead.
21:34:21 Join robin0800 [0] (
21:34:30pipsqeekyou can see my misunderstanding.
21:34:42pipsqeekanyway. I'm off, thanks for recommending #rockbox-community
21:34:52 Part pipsqeek
21:35:42Zagoramiconn: that is definitely something on your end. I get 700-1000 Kbyte/s on that file, from several different network locations
21:38:24 Join fml [0] (n=4fd3c25a@gateway/web/cgi-irc/
21:39:21fmlHello to all of you. Any comments on FS #9618 and FS #9769 ?
21:39:21jhMikeSbertrik: Do you recall if RTCSUPP reads the same as BVDD on c200 or does it need to really specify them differently there?
21:39:32 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:41:20bertrikjhMikeS, I put some measurements on the SansaCharging wiki some time ago, BVDD and RTCSUP are both too high, but not equal
21:42:03bertrikBTW, the information on that page about reading the charging current for Sansas is incorrect, I'll fix that.
21:42:50 Quit karashata ("G'bye everyone!")
21:45:04jhMikeSRTCSUPP seems closer at least
21:45:32 Quit jhulst (Remote closed the connection)
21:45:52 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:46:38 Join timc`` [0] (n=aoeu@
21:49:32 Quit Strife89 ("Study time. :)")
21:59:27 Join funman [0] (n=fun@rockbox/developer/funman)
22:01:49 Join martin___ [0] (
22:02:28*amiconn wonders about that my_open()
22:03:10 Nick Bagder_ is now known as Bagder (
22:03:45Zagoramiconn: yeah me too.
22:03:46funmanamiconn: did you look in /usr/include/* ?
22:04:02funmanautoconf uses strange prefixes as well, but it can't harm here
22:04:04*Zagor notices rashers sim commit. thanks!
22:04:53 Quit jhulst (Remote closed the connection)
22:06:52amiconnI mean what my_open()/my_close() is good for, and why it's only needed/used on target
22:07:59 Quit nplus (Read error: 110 (Connection timed out))
22:08:19ZagorI don't think I have ever actually ran doom in the sim...
22:09:24*amiconn did, and it worked
22:09:44 Join jhulst_ [0] (n=jhulst@unaffiliated/jhulst)
22:10:08amiconnIt looks like the my_* functions do some magic regarding the file descriptor limit in rockbox, but I don't understand it completely
22:12:21 Join z35 [0] (
22:12:50Horschtis there a windows application that will build my rockbox tagcache?
22:13:04 Quit gevaerts ("leaving")
22:13:20 Join tessarakt [0] (
22:13:51Horschtcurrently, whenever I add new songs to my ipod (5.5G 80GB) I have to rescan/rebuild my tagcache after the next boot
22:14:05Horschtfor so many files, this takes up to 30 minutes
22:14:21bluebrotherHorscht: do you have dircache enabled? That should speed it up
22:14:45Horschtno, how would dircache speed up building my tagcache?
22:15:18bluebrotherwell, it does somehow −− I guess as in checking if a file has changed at all
22:15:42bluebrotheralso, the database can update itself and that should be quite fast
22:15:54Horschtit's not, realy.
22:16:11bluebrotherat least it was for me, but I don't have a player of that capacity.
22:16:21bluebrotherwell, afaik you need to enable dircache for that
22:16:36bluebrotherso I'd strongly suggest enabling that and trying again first
22:16:37 Nick Anges is now known as Anges_kikou (
22:17:19Horschtwill give that a shot tomorrow
22:17:30 Nick Anges_kikou is now known as Anges (
22:17:38PaulJamamiconn: when you compile the uisim on cygwin, do you also get lots of warnings? (warning: -ffunction-sections may affect debugging on some targets)
22:17:58 Quit bertrik (Read error: 113 (No route to host))
22:18:22linuxstbZagor: Do you know that the tools/database/ tool no longer compiles?
22:18:23amiconnLast time I compiled a sim I didn't. I rarely use the sim though
22:20:06Zagorlinuxstb: yes, now :-)
22:20:11*linuxstb blames mcuelenaere's commit of FS #9733
22:20:26linuxstbZagor: OK, just thought I would mention it ;)
22:21:08*linuxstb wonders why logf.c is compiled at all for the database tool
22:21:08 Quit fml ("CGI:IRC (EOF)")
22:21:57 Join miepchen^schla [0] (
22:22:12*linuxstb removed logf.o and -DROCKBOX_HAS_LOGF and it compiles cleanly...
22:22:32pixelmaPaulJam: I compiled a sim this noon and don't remember seeing warnings like this, "just" failing while compiling doom
22:23:02PaulJampixelma: ok, thanks for the info.
22:23:08linuxstbHorscht: There is a tool that might build the database, but I'm not sure how well it's been tested, or how to compile a Windows .exe...
22:23:14 Quit Xerion (Read error: 54 (Connection reset by peer))
22:23:45amiconnActually I have some sims compiled 2008-12-28, and I don't think there were any warnings
22:24:26*amiconn wonders how many files in general and how many music files Horscht has on his ipod
22:24:49Horschtamiconn, ~ 10.000 Music files
22:25:19Horscht+ ~1000 Coverart
22:26:09linuxstbZagor: Hmm, I get an immediate segfault running tools/database/database
22:26:26 Join bertrik [0] (
22:27:14bertrikwho is currently developing for the c200v2?
22:27:46Zagorlinuxstb: Yoshihisa's patch FS #9371 is the way to go, I think
22:28:34 Quit tvelocity (Remote closed the connection)
22:28:58amiconnHorscht: Hmm, weird. I have fewer files on my H180 (~4000 music, ~6500 total) but a full database update takes less than 2 minutes for me
22:29:19amiconnThis is with dircache enabled - with dircache disabled it would take several times longer
22:29:36Horschti will try with dircache enabled tomorrow.
22:29:43 Join tvelocity [0] (
22:29:54HorschtI thought dircache was only usefull if you use the file explorer often
22:30:13amiconnDircache is useful for everything that opens lots of files
22:30:39Horschtalso, would it not "lower" my free memory, i.e. shrinking my audio buffer?
22:31:18Horschtwould it be usefull or make any sense at all to turn dircache on, build the tagcache then disable dircache again?
22:31:21amiconnIt does. You can check how much after having it enabled
22:31:31linuxstbZagor: Have you looked at his most recent patch, or do you plan to?
22:31:32amiconn(in the debug menu)
22:32:25amiconnBut then the benefits outweigh the smaller audio buffer on large-mem targets. Personally I have decided to use dircache on all targets with >=32MB RAM and a harddisk
22:33:16HorschtI think I am missunderstanding the dircache then
22:33:36Zagorlinuxstb: I had planned to, but I'm trying to focus on my buffering work now to avoid it hanging in the air for too long.
22:33:57*amiconn is running a test to see how much slower a db update without dircache is on hos H180
22:34:56HorschtI will give dircache a shot tomorrow after charging and synching my ipod
22:35:59Slasherii should optimize the db reverse scanning sometimes, that is fairly easy to do and would speed up update a lot without dircache
22:36:14 Quit martin___ (Read error: 110 (Connection timed out))
22:36:17kugel\o/ Slasheri
22:36:22Slasherikugel o/
22:36:57Slasherisorry, i haven't been here much lately :)
22:37:00HorschtI always use rebuild instead of update. I tried update once and it drained half my battery and wasn't finished after 2h.
22:37:01pixelmacould some people test one thing for me? I wonder if it's just me seeing this... so if someone has a minute please try the following: play some music, enter file browser from WPS (music still playing), return to WPS by resuming, then try to stop playback - with a recent revision, maybe not the latest
22:37:02linuxstbZagor: I'm just looking now, and it seems to still touch lots of files, and make unobvious changes (e.g. renaming things)
22:37:17amiconnHorscht: Almost 6 minutes for the scan, plus ~30 sec for the db build after reboot
22:37:37amiconnOh, btw, with dircache enabled you won't have to reboot when updating the db
22:37:45Horschtrebuild or update?
22:37:45kugelSlasheri: once you have some time, could you please have a look at , I've been scratching my head on how to implement it in the wanted way (i.e. show non-sort tags, but use sorting tags for sorting internally)
22:37:48pixelmaSlasheri: there was some talk in the forums about the track number guessing
22:37:57bluebrotherrebuild the database will start from scratch, thus taking longer than update ...
22:37:58lucentkugel: I have confirmed that 8gb fuze v1 + microsd = panic
22:38:11kugellucent: what panic?
22:38:26amiconnbluebrother: Of course I compared full rebuilds. I never use update because I don't trust it
22:38:35funmanZagor: linuxstb suggested that we use 0x0 to decompress the OF (be it IRAM or DRAM) and an IRAM alias to decompress rockbox (and execute it from the alias rather than from 0x0, after copying the vectors)
22:38:47 Join petur [50] (n=petur@rockbox/developer/petur)
22:39:07 Join dkerschner [0] (
22:39:20Slasheripixelma: hmm, need to check that :) all tracks are marked if the track number have been guessed -> it should be easy to implement a feature to make it possible to discard guessed tags when browsing
22:39:35 Quit miepchen^schlaf (Success)
22:39:38dkerschnerwhat's a recommended player for rockbox? Looking for something on the not so expensive side heh
22:40:07Bagderdkerschner: see
22:40:20amiconnSlasheri: Imo it would be better to do away with guessing
22:40:45Slasherikugel: i will check that
22:41:00lucentkugel: the one where sending CID command fails
22:41:01*amiconn has a whole bunch of tracks with a (wrongly) guessed track number
22:41:05 Part dkerschner ("ERC Version 5.3 (IRC client for Emacs)")
22:41:07pixelmaSlasheri: IMO it could be removed completely - half of the time it's wrong anyway and in my case a great deal of my tracks don't have track numbers on purpose (because I don't know them or else)
22:41:12Slasheriamiconn: so that remove the guessing completely?
22:41:22kugellucent: that what seems to be fixed magically now?
22:41:49lucentkugel: when I reported that microSD external was working, it is because of Zagor's patch which reboots... so what I saw was that I had microSD inserted on boot, and it rebooted immediately, which left it working
22:41:52Slasheriwell, i have no opinion on that :) maybe that should be voted if it should be completely removed?
22:42:16kugellucent: well, as funman said, might be just lacking some delay
22:42:24pixelmayes, maybe a mail to the dev list?
22:43:17kugelwhy remove that, if it's possible to just ignore the guessed numbers using some parameter in the tagnavi?
22:43:35lucentkugel: it is confirmed on another user's 8gb fuze, with 64mb microSD card size
22:43:58kugelbut on the other hand, the user should be responsible for correct tags, not the app (especially not on an embedded system)
22:44:08kugellucent: where?
22:44:24lucentkugel: "Agnes" user tested for me
22:44:54pixelmakugel: because it complicates things for no real gain?
22:45:00amiconnkugel: It adds code (binsize) and does more harm than it helps
22:45:20funmanI started looking (again) at the OF for this problem but saw nothing obvious
22:45:23kugellucent: I can just say play around a little, such as putting some delays, or loop until csd is correct, or something else. I cannot really do much
22:45:45funmanadvcomp2019: does your 4GB clip works fine (as 4GB) with rockbox ?
22:45:59amiconnI also have several tracks with the tracjnumber filed intentionally left blank, but since the artist name consists of digits (only) the db treats them as having a (bogus) track number of 66
22:46:09lucentkugel: okay! I just wanted to give you correct information, I did the regression testing against SVN revisions and that implement reboot on panick is the revision when microSD sort of works for me
22:46:31pixelmakugel: I also dislike software that wants to know better than I do in general ;)
22:46:37bertrikfunman, I think I have some CSDs saved somewhere
22:46:43kugelpixelma: :)
22:46:45 Join MethoS- [0] (
22:47:02lucentkugel: also I confirm that your hack / workaround for pixel drawing on the LCD makes my incorrect pixel values go away
22:47:12kugellucent: without hardware that fails, I cannot do much, sorry
22:47:27kugelit's not my hack/workaround, Unhelpful provided it
22:47:34lucentmy mistake
22:47:49pixelmawants to know better what I want, I mean
22:48:02amiconnZagor: You forgot to fix the test plugins (at least some of them)
22:48:03funmanbertrik: lucent & kugel tested today, the CSD were the same on 4GB e200v2 4GB Fuze and 8GB Fuze - so I suspect several SD cards are used and we need to select one or the other
22:48:18kugelamiconn: yea, some others might find it useful. I, for myself, didn't even notice the db has such a feature, since my tags are correct
22:48:33Zagoramiconn: ah :-(
22:48:36 Nick jhulst_ is now known as jhulst (n=jhulst@unaffiliated/jhulst)
22:48:39 Quit {phoenix} (Remote closed the connection)
22:48:56amiconnZagor: test_codec.c emulates the codec api, so I now get "error: structure has no member named `PREFIX'"
22:48:57kugelbut there's no need for wasteful features, that don't work as expected, that's true
22:49:47amiconnkugel: My tags are also correct. As I said, the tracknumber field is left blank on purpose.
22:50:10kugelI didn't claim yours are wrong
22:50:16 Quit Nico_P (Remote closed the connection)
22:50:59Slasherikugel: one thing i am worried about adding all kinds of tags to the tagcache is that every new tag increases the db size by 4*number of entries (and memory consumption if db is loaded to ram)
22:51:08amiconnPaulJam: Hmm, I do indeed get that warning for everything in the pluginlib...
22:51:10kugelalthoug you once said some have replaygain tags, which you don't use at all
22:51:35***Saving seen data "./dancer.seen"
22:51:43kugelSlasheri: I was worried about that too
22:51:53kugelI think I commented on that in the tracker
22:51:57jeffroniusCan anyone answer: Is the latest svn e200v2 code functional??? I checked it out compiled and installed the mkamsboot patch on a Sansa e260v2 but after the rockbox bootloader spashsceen I get a blank LCD. The sansa boot still works.
22:52:20bertrikfunman, hmm, nasty... the naf_ceX_n bit in the CCU_IO register could be a possible candidate for selecting a specific flash chip I think
22:52:29kugelwhich is why I didn't really wanted to have both artist and sort artist loaded at the same time if you only browse artist
22:52:36 Quit XavierGr (Nick collision from services.)
22:52:47 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
22:53:08funmanbertrik: let me look at CCU_IO usage
22:53:12Bagderjeffronius: isn't that just about the same like saying it isn't functional?
22:53:41bertrikwe should be able to figure out from the OF how any bank switching is done I think
22:53:42jeffroniusbagder: I am not excluding "operator error"
22:54:15Bagderthe ams sansas are all "shaky" ports still so it's not a big surprise
22:55:20funmanI only see xpd_func_sel being modified
22:55:27*Horscht needs to install Ubuntu in a Virtual machine and create a Rockbox Build environment
22:55:48Slasherikugel: i have thought that db size issue many times but couldn't find out a really better way to do that
22:56:42jeffroniusbagder: i am inbagder: thank you for the reply! I will read up on the wiki
22:56:55Slasherikugel: one way could be to make some numeric tags shorter but that could be extremely hard to implement
22:57:24*linuxstb would like to conclude the sendfirm/MTP_DLL.dll license debate and wishes mcuelenaere was around...
22:57:53 Quit LambdaCalculus37 (" ajax IRC Client")
22:58:17kugelSlasheri: how long are they?
22:58:18bertrikfunman, ok thanks for looking
22:58:27 Quit tvelocity (Remote closed the connection)
22:58:49Slasherikugel: all main tags in the track entry are 4 bytes
22:59:03 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
22:59:36Slasheribut some tags, like disc number, track number, rating, year, etc. could be much shorter
23:00:34 Part gartral1
23:00:45 Join tvelocity [0] (
23:01:02bertrikradio on the sansa clip got a lot quieter in the past few days
23:01:12kugelSlasheri: 4bytes because of the padding? how much speed is gained due to padding? I suppose the Xes take quite a few bytes
23:01:35funmanbertrik: doesn't it depend from the reception?
23:01:43Slasherikugel: btw, one feature i would like to implement sometime would be to make it possible to put a track into different categories. So that there would be a 4 byte category tag that would have a bitmask about in which categories the songs belongs to
23:02:24Slasherikugel: well, that padding is a entirely different story :) that 4 byte thing is only to numeric tags
23:02:27funmani'll try to look in the function i identified as "read_sectors" tomorrow (at 0x5340 in Clip v29 firmware)
23:02:31kugelhehe, sounds like a lot of work though
23:02:35bertrikfunman, no I don't think so. Radio on a slightly older build was fine, now it's very quiet all of a sudden. I suspect svn r19713 has something to do with it
23:02:58 Quit Zagor ("Clint excited")
23:03:01linuxstbSlasheri: What kind of categories?
23:04:22 Quit petur ("ZZzzzz *bonk*")
23:04:29Slasherilinuxstb: user definable categories, something like a playlists. But those would be accesed from the db browser and there could be a quick access from wps that would open a multiple checkbox list to categories a song :)
23:04:33Slasherijust some thoughts
23:04:56Slasheri*to categorize
23:05:08kugelI must admit that this sounds like overkill to me :/
23:05:23Slasherihehe, i would find that useful :)
23:05:40kugelI wouldn't use it at all
23:05:59Slasherii never know exactly all the songs i have on the player and i would like to categorize the good ones
23:06:44Slasherisomething like that, but a more advanced rating system :)
23:06:45*amiconn wonders who would put songs on his dap which he doesn't like
23:07:02Slasherii already use ratings for that but find it a little bit of limiting
23:07:13Slasheri(there can be only a one rating of a song)
23:07:24*linuxstb is still very old-fashioned, and just listens to albums...
23:07:50*linuxstb misses having to get up half-way through and turn it over, and considers such a patch ;)
23:08:39 Quit tessarakt (Read error: 110 (Connection timed out))
23:08:57 Join tessarakt [0] (
23:09:36bertrikfunman, it's the #ifdef HAVE_RECORDING in audiohw_set_master_vol that prevents the line input volume getting set
23:09:57*linuxstb spots more complication with sendfirm - the Linux version is based on an LGPL'd example program from libMTP, so it's not trivial to change its license...
23:10:27Bagderis the example code itself LGPL?
23:11:24funmanbertrik: oh good catch, this check sounded weird but is in accordance with the rest of the file
23:12:11linuxstbBut it's pretty trivial...
23:12:38Bagderif that's the only LGPL code, I figure a rewrite should be fairly easy if that turns necessary
23:13:05*linuxstb propses Rockbox drops support for Windows ;)
23:14:05 Join Jaykay [0] (
23:17:27pixelmaof course, doing it wrong here, starting a current sim with −−background... still like that change though :)
23:17:49*bluebrother is with linuxstb :)
23:19:26linuxstbBagder: Do you have a view on the MTP_DLL.dll license?
23:20:23BagderI haven't followed that discussion at all. What about that license?
23:20:40 Quit ze (Read error: 113 (No route to host))
23:21:44linuxstbBagder: Look in utils/MTP/ at sendfirm_win.c. It links with MTP_DLL.dll, which IIUC contains a lib from a MS SDK
23:22:05linuxstbMTP_DLL.dll is (again, IIUC), a wrapper around that lib
23:22:52amiconnBleh, test_codec even fails to compile for target
23:24:11Bagderlinuxstb: then I guess it depends on the license of the libs that the .dll gets linked with
23:24:23 Join Jaykay_ [0] (
23:24:47linuxstbBagder: Well, I'm assuming (as they're from MS), that it's not going to be GPL compatible...
23:25:20Bagderwell we can't use plain GPL that's for sure
23:26:43amiconnHmm, a single-line fix makes it compile
23:26:57*amiconn is really curious whether it will work though
23:27:01 Quit tessarakt (Connection timed out)
23:27:31 Join tessarakt [0] (
23:27:53 Join Bensawsome [0] (
23:31:52pixelmalinuxstb: what are Rockbox drops and why should Windows get them? :P
23:35:29 Quit Jaykay (Read error: 110 (Connection timed out))
23:37:20 Quit jhulst (Remote closed the connection)
23:38:52 Join Xerion [0] (
23:39:06 Join neddy [0] (n=john@nat/sun/x-64664d65982a0a22)
23:39:47 Quit bmbl ("Woah!")
23:40:29 Part neddy
23:41:40 Quit jhMikeS (Nick collision from services.)
23:41:46 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
23:41:51 Join Strife89 [0] (n=michael@
23:44:01 Quit funman ("leaving")
23:45:04 Quit Jaykay_ (Read error: 110 (Connection timed out))
23:52:20 Join MarcGuay [0] (
23:54:06MarcGuaylinuxstb: Alright. So I've defined TCCBOOT in config-c100 and I'm running "mktccboot OF.bin player.rom dualbootloader.rom".
23:54:36MarcGuayAnd when I upload the dualbootloader with tcctool, I get the LCD lighting up and that's all, hold switch in either position.
23:55:34MarcGuayI added the button_init to crt0.S with no change (I thought maybe it was required to detect the hodl switch). But it's my understanding that even if my rockbox build is crunked, the OF should still load...?
23:55:46MarcGuay*hold switch
23:56:04 Quit tvelocity (Remote closed the connection)
23:57:30linuxstbMarcGuay: Can you post a patch of your changes to crt0.S?
23:58:23linuxstbMarcGuay: Also, how much RAM does the C100 have?

Previous day | Next day