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

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

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

#rockbox log for 2014-01-19

00:01:48 Quit wodz (Quit: Leaving)
00:02:08lebelliumfrom my understanding, there are no more Archos users among the active devs
00:03:38lebelliumor they just keep a very old build because they think it's not worth upgrading anymore
00:04:25saratoga_well it'd be nice to fork at the next release then
00:04:29lebelliumbut we still add some useful features on old players sometimes. Quite recently the Ondio FM got soft key lock
00:05:45[Saint]There's no reason to not go back and add features that won;t increase the binsize by any considerable amount.
00:05:57[Saint]...but, there will also be no obligation for anyone to do so.
00:06:27[Saint]I'm all for forking out HW codec/charcell and applying only bugfixes and trivial features to that branch.
00:07:13[Saint]Leaving us the other targets to play with, without fear of breaking ancient, highly restrictive ports, would be very nice.
00:07:55 Join n1s [0] (~n1s@rockbox/developer/n1s)
00:08:25[Saint]I *don't*, however, want to see them get forked out and then left to rot for all eternity...that would be a huge shame.
00:09:16[Saint]But in saying that, they've been kept on the brink of brokenness, only being massaged back into line occasionally for an, I'm not sure if anything is really lost.
00:14:54[Saint]...being perfectly honest, I think its something we all want, and don't want, simultaneously. No one wants to be the one to push the button, no one wants to be the one that is effectively killing ~13 years of Rockbox history.
00:15:29[Saint]But, in reality, perhaps euthanasia is better than dying of old age and neglect.
00:15:42[Saint]Gah...its a sad affair.
00:17:09lebelliumhard decision to take so that means we're probably still be discussing that next year :)
00:18:05[Saint]I think I may have nailed it with "No one wants to be the one to push the button, no one wants to be the one that is effectively killing ~13 years of Rockbox history."
00:18:34[Saint]...that's how I feel about it, at least. I imagine at least a few others feel similarly.
00:19:29[Saint]There would be a certain beauty, and cleanliness about it, if one of "The Originals" did it.
00:19:54n1sCurrent builds of these ports don't see much use so they are already bit-rotting
00:20:06[Saint]Do we *know* that?
00:20:12[Saint]Or do we just have no reports?
00:20:19lebelliumthat's what I was saying above
00:20:25lebelliumthere are no download stats AFAIK
00:20:31saratoga_bagder can probably get the download stats
00:20:34saratoga_hes posted them before
00:20:41[Saint]That would be interesting.
00:20:59[Saint](I wonder if it can also filter unique downloads)
00:21:10n1s[Saint]: last time o saw stats the archos builds were pretty low
00:21:47[Saint]Right. I wasn't necessarily saying you were wrong. I just didn't know if we know it for a fact, or if we just all felt that way.
00:22:39saratoga_for reference
00:22:52kugelthat's almost six years ago :)
00:22:56*[Saint] doesn't see thos enumbers increasing.
00:23:09[Saint]It actually built then too. :)
00:23:48n1syeah the downloads are sure to have dropped since the recorder stopped building...
00:24:07*[Saint] decides we should make the RSB do it
00:24:25[Saint](do we actually have one, technically, or is it still the last lot since we didn;t renominate?)
00:24:28saratoga_could still check releases
00:24:53AlexPI'd probably drop them at this point
00:24:59n1sbut seriously, the low numbers + few reports from users + the fact that the majority of added features are not relevant/available on hwcodec
00:25:11saratoga_can the build system handle two branches by the way?
00:25:16AlexP[Saint]: I guess it is still the last lot until the next election, whenever that might be
00:25:21n1sum, make the case as imo
00:25:29[Saint]AlexP: right, thanks.
00:25:49AlexPWhich I think might include me now I think about it
00:26:16[Saint]...sad as it may be, it does seem we've reached a consensus. We're all thinking the same way, well, almost.
00:26:23AlexPUnless there has been an election I didn't notice :)
00:26:33saratoga_in terms of developers, who is actually dealing with hwcodec the most? personally none of this has much impact on me since the codecs dont' even build for hwcodec
00:26:34[Saint]But I think it would be best to go to the ML to let users have a say, and inform them.
00:26:49saratoga_i guess kugel, jdgordon because of the theme engine and buffering
00:26:51AlexPsaratoga_: I'm not sure anyone is
00:26:55lebelliumI need to get a FM recorder and get g#497 merged before any support drop
00:26:57fs-bluebotGerrit review #497 at : by Jean-Louis Biasini (changes/97/497/3)
00:27:09[Saint](2~3 years ago...)
00:27:10AlexPAnd those that are it isreally just if it inadvertently breaks
00:27:18AlexPAnd by breaks I mean doesn't build
00:27:22AlexPNot works :)
00:27:39saratoga_ g#497 can probably go in
00:27:40fs-bluebotGerrit review #497 at : Add FM softlock to the fm recorder. by Jean-Louis Biasini (changes/97/497/3)
00:27:49saratoga_unless theres some problem i'm not aware of?
00:27:59[Saint]How would we approach this, we we to handle it in the most transparent way possible?
00:28:06[Saint]*were we
00:28:53lebelliumsaratoga: I tested it for Ondio FM and it got merged. I guess Jean-Louis was just waiting for someone to test
00:29:01[Saint]I think once the descision has been made that that's the course of action that would inevitable follow, we would need to give the users and less active developers a chance to have a say?
00:29:08[Saint]Via the forum/ML?
00:29:26AlexPI don't think users et much of a say tbh
00:29:27saratoga_lebellium: assuming it works fine, i can commit it now
00:29:42[Saint]AlexP: neither do I...but, we need to make them think they do.
00:29:44n1s[Saint]: i'd say bring it to the dev ml let it take enough time for people to think and reply and then perhaps let the rsb decide if no consensus is reached on the ml
00:29:52AlexPBut if you would like to try and reach a decision, the dev mailing list would be a good start
00:30:01lebelliumsaratoga: it works fine for Ondio FM but I can't say for FM recorder, although I don't see any reason why it would be different
00:30:28AlexP[Saint]: heh :) I don't see the point though, the odd person who uses it will say no, nobody else will care, and it won't influence any decision
00:30:34AlexPwrt the user ml that is
00:30:59[Saint]AlexP: n1s: I think the decision has been made already, and has been for some time, its just finalizing it really.
00:31:04saratoga_actually can someone else push 497? i'm not near a machine with git in case it breaks something
00:31:09[Saint]I honestly don;t think any active dev is going to oppose this.
00:31:58n1s[Saint]: then reaching a consensus on the ml should be easy
00:32:41[Saint]AlexP: n1s: The main reason I thought it would at least be nice to include users in this decision is the vain hope I have that it might inspire someone to take up maintaining the branch once its forked.
00:32:54[Saint]...I mean, you never know, and I feel I'd have to try before I put this to bed.
00:33:02AlexPBe my guest :)
00:33:04[Saint]However unlikely it is.
00:33:32*[Saint] is /sometimes/ an optimist, when it suits. :)
00:34:04n1s[Saint]: _optimist_
00:35:00n1sbut sure, i wouldn't expect any constructive discussion but if you're willing...
00:35:44 Quit ender` (Quit: I will not use any plan in which the final step is horribly complicated, e.g. "Align the 12 Stones of Power on the sacred altar then activate the medallion at the moment of total eclipse." Instead it will be more along the lines of "Push the button." --)
00:36:26[Saint]Wow...there's a quit message and a half.
00:39:30 Quit kugel (Ping timeout: 260 seconds)
00:49:33 Quit n1s (Quit: Ex-Chat)
00:56:25***Saving seen data "./dancer.seen"
01:15:40 Quit lebellium (Quit: ChatZilla [Firefox 27.0/20140116125114])
01:23:18toehserWhat is in the way of making Archos FM Recorder work? I have one I can either troubleshoot with or loan out. Not quite sure I want to permanently part with it, but I'm glad to ship it away for an extended period if it is really useful for development (and anyone cares...). Or I can use it to try to help fix the issues, if someone wants to help get me up to speed. Why would the HWCODEC be an issue, wouldn't the code for that "just keep working"?
01:24:34gevaertstoehser: the FM recorder is fine, it's the recorder that's troublesome
01:25:09gevaertsIt's easy to get building again for a while, just find another feature to drop, but that's not a proper solution
01:25:16toehserI thought the FM Recorder was just the Recorder with the Tuner enabled? What was the FM Recorder minus FM called then?
01:26:17gevaertsThe issue is that it's too big for the ROM bootloader, which can be solved by moving over to a bootloader / main firmware model like basically every other target
01:26:39gevaertsThat shouldn't be *too* hard to do, but someone has to be motivated enough to do the work
01:27:15toehserI see, I'm thinking of the "Recorder 20".
01:28:07[Saint]Is there a bricking risk involved, gevaerts? If so, is there a recovery method?
01:28:25gevaertsNo. It's about loading from disk
01:28:39gevaertsThe other issue is that HWCODEC has a rather different playback engine which can cause issues sometimes, and which annoys people
01:28:44toehserSo, it means writing a new bootloader for something that no-one uses... I see the problem.
01:29:50toehserIf you fork it, it is dead, might as well just declare a version that is the 'last archos version', nobody is going to be fixing and working on it if you fork it... I agree with that.
01:29:52gevaertsAnd then of course there are things like a possible future new plugin loader that would allow neat stuff, which does include CPU architecture specific bits, where again Archos is different
01:30:10gevaertsAnd of course there's the Player with its charcell screen :)
01:33:35gevaertstoehser: I very much doubt if shipping devices to people will help here. People who care about the old Archoses usually have a few of them lying around
02:01:58 Join nosa-j [0] (~m00k@
02:06:24saratogashould i keep closing fs tasks for AMS usb bugs or just let people keep making them
02:06:31saratogai guess in theory they may be different issues
02:26:23[Saint]IS it known to be resolved now?
02:26:42[Saint]"the main one", that is. Themes randomly totally fucking USB.
02:27:55*JdGordon wouldnt shed a tear if we dropped charcell
02:30:45[Saint]You've been pretty impassioned on the subject in the past, perhaps you want to be the one to bring it up formally on the ML then? :)
02:35:07saratoga_i don't think usb is fixed, but we already have a bunch of tasks open for it
02:35:18[Saint]ah, right.
02:35:49[Saint]we actually need to be thinking about freezing, RC builds, and a release.
02:36:03[Saint]We need AMS testers for git HEAD.
02:36:21[Saint]Specifically people that saw the theme/USB shennanigans.
02:56:26***Saving seen data "./dancer.seen"
02:58:49 Quit saratoga_ (Quit: Page closed)
03:06:06 Quit Rower (Quit: Hmmm...)
03:10:31 Join Strife89 [0] (
03:47:35PurlingNayukiAnyway git HEAD works fine on my AMSs.
03:50:08[Saint]PurlingNayuki: Don;t you use a unix host OS, though?
03:50:35[Saint](It seems to be Windows specifically that has issues)
03:54:16PurlingNayukiI have Windows for games, Debian for working and study, and had OSX for music stuffs.
03:55:51 Join draba [0] (
03:57:47toehserUSB to the internal storage seems to work fine for me now, but I've seen what seems like more issues to the card, I'll try that some...
04:03:04[Saint]I guess we just need to hear from lebellium, then.
04:04:01[Saint]His "samsunglike" theme is apparently an almost guaranteed way to scre up USB on AMS.
04:04:10[Saint]*screw up
04:13:55PurlingNayukiWhy there's no lang APIs for plugins?
04:42:57 Quit amiconn (Disconnected by services)
04:42:57 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:42:57 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:42:57 Quit pixelma (Disconnected by services)
04:42:58 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:43:01 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:50:38 Quit draba (Ping timeout: 264 seconds)
04:54:34 Quit Strife89 (Ping timeout: 246 seconds)
04:54:35 Join bitz [0] (
04:54:51 Quit bitz (Client Quit)
04:56:27***Saving seen data "./dancer.seen"
04:58:15 Join X2000 [0] (
05:04:58X2000Question: I'm running Liniux and am trying to build Rockbox form source. I can generate a makefile but every time I try to build it I get an error saying that a file called lang.h cannot be found. Google hasn't turned up any helpful results and I can't find the file anywhere in the Rockbox soure direcotry. Where do I get lang.h from?
05:07:54[Saint]Are you sure you meet the build requirements? That may cause odd issues I imagine.
05:07:58[Saint]one sec.
05:08:14[Saint]Apart from a base system, all we need should be "automake bison build-essential ccache flex libc6-dev libgmp3-dev libmpfr-dev libsdl1.2-dev libtool texinfo"
05:09:02[Saint]that's for binaries/simulator, manuals need slightly more.
05:09:14X2000Yeah I already went through and installed all those while trying to run
05:09:49[Saint]WHat target are you trying to build for?
05:10:04X2000ipod video
05:13:38[Saint]working fine here.
05:14:35[Saint]WHat do you mean "while /trying/ to run"?
05:15:00[Saint]That part is kinda critical, you need the armeabi toolchain, there's no avoiding it. It needs to suceed.
05:15:57[Saint]Also, is this a clean working checkout?
05:16:22[Saint](if there are patches applied, which? can you build without them, have you checked?)
05:16:36X2000every time I tried to run it, it would say cnnot run without (package), I installed the package it asked for and repeated until it ran sucesfully
05:17:51[Saint]massaging a base system until a toolchain builds doesn;t necessarily mean you have all required components to build Rockbox binaries proper.
05:18:12X2000I'm trying to start work on a plugin I have in mind. right now I'm just trying to build an unmodified version of the source form git
05:18:34[Saint]If you haven't done so expressly, I would recommend "sudo apt-get install automake bison build-essential ccache flex libc6-dev libgmp3-dev libmpfr-dev libsdl1.2-dev libtool texinfo"
05:19:00[Saint]Or pacman, or yum, or aptitude...whatever floats your boat.
05:19:28X2000thank you apperently build essential wasn't installed. I'll try it now.
05:20:26X2000nope same error
05:21:09[Saint]make veryclean in the build directory, then reconfigure and try again.
05:21:41[Saint]that is the command "make veryclean", and then rerunning configure for the target.
05:22:15X2000seems to be working. Thank you so much
05:22:22[Saint]Not a problem.
05:25:20PurlingNayukiIs FS #12074 fine to commit? Though we can't detect lineout, it's still possible to let users to choose.
05:25:20fs-bluebot Add setting to enable/disable lineout on Sansa Fuze V2 (patches, unconfirmed)
05:25:44 Part X2000 ("Konversation terminated!")
05:28:04[Saint] Comment by MichaelGiacomelli (saratoga) - Monday, 03 December 2012, 15:49 GMT
05:28:04[Saint]Hi Rawl,
05:28:04[Saint]We won't be including this in the main build until someone reverse engineers the mechanism for detecting if a line out dock is connected.
05:28:42[Saint]The setting hijacked here also isn't appropriate.
05:28:59[Saint]That setting is used to *power off* the lineout feature, not to switch to or from it.
05:38:30PurlingNayukiNew settings?
05:38:42PurlingNayukiShould a new setting be added?
05:39:19[Saint]Argh. Its hard to tell, but the last two lines there are from myself, no from the comment.
05:39:22PurlingNayukiWhy auto detect is needed to add this to main build?
05:39:51toehserI haven't gotten a USB related failure for a long time... maybe it is fixed.
05:40:04[Saint]Anyway, I'm not sure that would help matters. As using a new setting doesn't seem to help the fact that there was opposition to this until such time as the line out state could be detected.
05:41:07[Saint]Autodetecting of the lineout state is the way its handled on every other target we can use the lineout on, as far as I know.
05:41:45[Saint]ANd the only lineout related setting is for disabling lineout completely (potentially saving battery).
05:42:35[Saint]It would seem weird to handle it different for one particular target. And it would inevitably lead to situations where the user forgets which output is enabled and wonders why one or the other won't work.
05:42:36PurlingNayukiCurrently we set the AMS lineout to powered off IIUC.
05:43:16PurlingNayukiNope. They set this themselves so they will remember.
05:43:22PurlingNayukiJust like volume cap.
05:43:29[Saint]Unfortunately, that's not how users work.
05:43:58[Saint]At least the volume cap provides some form of visual hint due to the reduced maximum volume.
05:44:50toehserWhy doesn't "backdrop: -" in the theme cfg file always clear the backdrop from a previous theme? Is there some other way to clear it?
05:44:50[Saint]You could certainly try to get this patch included (with a new setting), but I think it would just hit the same wall as it did initially.
05:45:22[Saint]toehser: it _should_.
05:45:28[Saint]Not doing so is a bug.
05:45:31toehserdoesn't though
05:45:53[Saint]You know what to do. :)
05:47:37 Join danielf [0] (
05:50:18 Quit danielf (Client Quit)
05:50:38 Join danielf [0] (
05:50:56 Nick danielf is now known as qsylalyn (
05:53:54 Quit qsylalyn (Client Quit)
05:54:59 Quit TheSeven (Disconnected by services)
05:55:13 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
06:00:44toehserk. I filed a bug. Weird thing: Load cabbiev2 theme, then tomsway2: backdrop cleared. Load 3_eyed_jack theme, then tomsway2: backdrop not cleared.
06:14:52toehserFor that matter, load 3_eyed_jack, then load cabbiev2, backdrop stays 3_eyed_jack.
06:39:43 Join leftright [0] (c5e01404@gateway/web/freenode/ip.
06:40:49leftrightH140: Bug; latest dev build still exhibits the following bug, FS #12769 - rb 3.12
06:40:50fs-bluebot rb 3.12 - iriver H140 bootloader crash: 1st boot fails, Reset pin, 2nd boot Successful. Repeatable. (bugs, unconfirmed)
06:54:55PurlingNayuki[Saint]: Reset lineout after reboot should be fine to users.
06:55:23PurlingNayukiIf they find no phone output most of they will try rebooting first.
06:55:39[Saint]I'm not sure we have any user settings that are gleefully ignored on a reboot.
06:55:48[Saint]That seems..rather disgusting.
06:56:14[Saint]Well, not just ignored, *reset*.
06:56:28***Saving seen data "./dancer.seen"
06:56:56[Saint]Nothing should ever do that in our current system, and I don't think anything does, because its kinda hideous.
06:58:25[Saint]This is why it kinda needs to detect the lineout state, so it "Just Works".
07:13:06 Quit leftright (Quit: Page closed)
07:24:34PurlingNayukiActually there is.
07:24:59PurlingNayukiSleep timer does that in fact, they will be reset on every boot.
07:54:13 Nick SuperBrainAK is now known as DormantBrain (~andy@2001:470:8:a61::5f92:59a1)
07:55:23 Join Gallomimia_ [0] (~gallomimi@
08:05:36 Join draba [0] (
08:24:34 Join Unhelpful_ [0] (~quassel@rockbox/developer/Unhelpful)
08:24:40 Join KiwiCAM_ [0] (~quassel@
08:24:50 Join Mir_ [0] (
08:25:53 Quit zoktar (Ping timeout: 264 seconds)
08:25:53 Quit albb0920 (Ping timeout: 264 seconds)
08:25:53 Quit ladyblink (Ping timeout: 264 seconds)
08:25:53 Quit gelraen (Ping timeout: 264 seconds)
08:25:54 Quit Mir (Ping timeout: 264 seconds)
08:25:54 Quit pystar89 (Ping timeout: 264 seconds)
08:25:54 Quit ranmachan (Ping timeout: 264 seconds)
08:25:59 Join gelraen [0] (
08:26:01 Join zoktar [0] (
08:26:02 Quit kiwicam (Ping timeout: 264 seconds)
08:26:03 Quit Unhelpful (Ping timeout: 264 seconds)
08:26:03 Quit froggyman (Ping timeout: 264 seconds)
08:26:03 Quit zoktar (Changing host)
08:26:03 Join zoktar [0] (~zoktar@unaffiliated/zoktar)
08:26:10 Join ranmachan [0] (
08:26:13 Nick Mir_ is now known as Mir (
08:26:46 Join ladyblink [0] (
08:32:15 Join froggyman [0] (~frogs@unaffiliated/froggyman)
08:45:03 Join albb0920 [0] (~albb0920@
08:45:39 Join n1s [0] (~n1s@rockbox/developer/n1s)
08:48:41 Join danielf [0] (
08:48:58amiconngevaerts: While moving over to a bootloader model as such isn't difficult, it's a bit more involved to make it work for both flashed and non-flashed install, and to make sure the flash image transition is safe
08:50:36amiconnThat means each model should be thoroughly tested. I do have access to Player, Recorder, Ondio SP and Ondio FM, but not FM Recorder and Recorder v2. The last two are identical apart from the model flag though, so having access to either one should be sufficient
08:51:43amiconnIn theory, one should also test with bootrom and non-bootrom variants, but I never came across one of the latter
08:51:58amiconn[IDC]Dragon did, but he's not been around for years
08:52:23 Quit danielf (Client Quit)
08:52:43 Join danielf [0] (
08:53:08 Nick danielf is now known as qsylalyn (
08:55:35 Quit [Saint] (Remote host closed the connection)
08:56:30***Saving seen data "./dancer.seen"
08:57:15 Quit qsylalyn (Client Quit)
08:57:33 Join qsylalyn [0] (
08:57:42 Join [Saint] [0] (~saint@rockbox/staff/saint)
09:08:58 Quit qsylalyn ()
09:09:11 Join pretty_function [0] (~sigBART@
09:09:13 Join qsylalyn [0] (
09:10:15 Quit qsylalyn (Remote host closed the connection)
09:10:26PurlingNayukiAnyway to upgrade Clip zip's firmware from 01.20.21 to 01.20.21?
09:10:49 Join qsylalyn [0] (
09:11:23[Saint]PurlingNayuki: Why would you?
09:11:56PurlingNayukiI modified the bootloader a little to adapt my using habit
09:12:41PurlingNayukiI was on 01.20.21 already before and I use the same version to patch bootloader, rename it to clzpa.bin and put it in the root
09:13:09PurlingNayukiBut after disconnect the USB it just goes to update database, no firmware upgrading.
09:14:34 Join rela [0] (~x@pdpc/supporter/active/rela)
09:15:23*PurlingNayuki is trying to flash the OF back and rockbox it again.
09:16:45[Saint]As far as I am aware, the firmware shouldn;t care if it is updating to the same, or lesser, version.
09:17:09PurlingNayukiOops I know the reason why
09:17:41PurlingNayukiStupid me, the file name should be clpza rather than clzpa
09:30:25 Quit qsylalyn ()
09:39:51 Join sprout_n1w [0] (~Rob@
09:43:50 Quit Gallomimia_ (Read error: Connection reset by peer)
09:44:37 Join Gallomimia_ [0] (~gallomimi@
10:01:10 Join ender` [0] (
10:06:01 Quit sprout_n1w (Quit: leaving)
10:19:47 Quit [Saint] (Remote host closed the connection)
10:21:49 Join [Saint] [0] (~saint@rockbox/staff/saint)
10:29:08 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
10:32:45 Join pystar89 [0] (
10:44:07Barankoupdated my usb ticket
10:44:23Barankoconnecting clip+ while turned off - sansa connects as usb1.0 )
10:44:38Barankolinux case.
10:44:47Barankowin7 all fine with it.
10:53:31 Quit rela (Read error: Connection reset by peer)
10:56:34***Saving seen data "./dancer.seen"
10:58:37 Quit draba (Ping timeout: 252 seconds)
11:04:48 Quit pretty_function (Remote host closed the connection)
11:26:47 Join Rower [0] (
12:13:24 Quit Gallomimia_ (Quit: Gallomimia_)
12:19:31gevaertsBaranko: for the record, full speed is still usb 2.0, and there are *very* few 1.0 devices around (1.1 is much more common) :)
12:21:24[Saint]I don't think I remember the last time I saw 1.0 in the wild.
12:21:32[Saint](in use)
12:21:45copper[Saint]: mice and keyboards
12:23:53[Saint]Not mine, that's certain.
12:24:04gevaertscopper: 1.*0*?
12:24:06gevaerts1.1, sure
12:24:25[Saint]even that's fading out.
12:24:58gevaerts1.1 was basically a bugfix release. There's not been any good reason to ever use 1.0 since 1.1 was released
12:26:19gevaertsAnd the step from 1.1 to 2.0 for low and full speed devices is also reasonably small. You need to change the version number, and I think you may need an extra descriptor, but that's more or less it
12:26:43 Join rela [0] (~x@pdpc/supporter/active/rela)
12:27:02[Saint]Its quite confusing at first.
12:27:08gevaertsOf course, for a mouse or something like that, that might mean having to update firmware that's been fine since 1998 and isn't actually broken, so why bother? :)
12:28:48gevaerts[Saint]: it's fairly straightforward now, but once we have superspeed devices we'll have to explain that while full speed *is* usb 2.0, high speed is not USB 3.0 :)
12:30:21[Saint]"full speed" was rather short sighted, I guess. :)
12:30:49gevaertsso was high speed, for that matter :)
12:30:57bertrikthere's plenty of room beyond full speed, like super, hyper, ultra, etc. :)
12:31:09gevaerts"Fuller speed" and "Fullest speed" would have been better
12:31:56coppersince we're currently at "super speed", I assume USB 4.0 will be "mega speed"
12:32:15[Saint]we could shorten it to ÜSB
12:32:18gevaertscopper: we can't know!
12:32:29copper[Saint]: haha, nice
12:32:33gevaertsIt may be mega speed, or it may be super speed *or* mega speed!
12:32:36copperI would totally vote for that
12:32:58gevaertsWe don't know if they'll refer to the 3.0/2.0 specs for lower speeds or include them in the 4.0 spec
12:33:39[Saint]"fuller than full speed but not quite the fullest speed"
12:33:56[Saint]...has a ring to it.
12:50:09 Join leftright [0] (c5e01404@gateway/web/freenode/ip.
12:51:47leftrighthi. Is mr. Someone perhaps reviewing this bug FS #12769. if not please update the falsh web page stating that flash only works up to v3.8
12:53:13leftrightthats on the H1xx series
12:53:53 Quit leftright (Client Quit)
12:56:38***Saving seen data "./dancer.seen"
13:03:07 Join mortalis [0] (~mortalis@
13:08:43[Saint]Oh. That was cute.
13:09:12 Quit rela (Read error: Connection reset by peer)
13:09:47[Saint]I think it might take a /wee/ bit longer than expected.
13:21:30 Join arcolight [0] (5c6414a3@gateway/web/freenode/ip.
13:27:00arcolightHello. I have a question about rockbox utility and packaging for Debian. I've read on official site, that there are no official packages for different linux distributions. I'm using Debian stable and I want to install rbutil as package. If i want to creaqte deb-package, should I create build scripts and send them to developers or they don't interested in this and I should only created package and post them on official site?
13:27:23arcolightSorry for my English, it's not my native language.
13:30:17 Join juda [0] (~brakha@2a01:e34:ee9a:fc30:5ceb:3c55:afaf:953f)
13:30:50judaj'aurais voulu avoir plus d'information sur rockbox et mon lecteur mp3
13:31:33judaet si c'est possible d'effectuer le changement, comment on fait sur linux?
13:34:06copperjuda: do you speak english?
13:34:39copperwhat is your MP3 player?
13:35:29 Join ikeboy [0] (
13:35:30copperit's not supported by Rockbox, sorry
13:38:13 Part arcolight
13:38:50 Join arcolight [0] (
13:41:07[Saint]arcolight: as long as you conform to the license requirements, you can do whatever you want.
13:55:08 Join eyfour [0] (
13:58:20 Quit eyfour (Client Quit)
13:58:26 Quit arcolight (Quit: Konversation terminated!)
13:59:33 Join arcolight [0] (
14:02:51 Join eyfour [0] (
14:03:11 Part eyfour
14:13:21 Quit ikeboy (Quit: ikeboy)
14:13:53 Part arcolight ("Konversation terminated!")
14:24:03 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
14:26:16Barankogevaerts: at all, i don't remember which of them right
14:26:37Barankoi see that speed musch differs and dmesg shows that now it's only full speed
14:26:49Barankoand want's to be connected to high speed
14:29:01 Join ikeboy [0] (
14:33:07 Join lebellium [0] (
14:36:05 Quit ikeboy (Quit: ikeboy)
14:45:57 Join Strife89 [0] (
14:49:41[Saint]I can create the same effect by connecting an iPod Video while the disk is spinning.
14:56:41***Saving seen data "./dancer.seen"
15:17:18 Quit lebellium (Ping timeout: 252 seconds)
15:21:24 Join lebellium [0] (
15:41:51 Quit Strife89 (Ping timeout: 252 seconds)
15:52:29 Part juda ("Quitte")
16:05:11 Join kugel [0] (
16:05:11 Quit kugel (Changing host)
16:05:11 Join kugel [0] (~kugel@rockbox/developer/kugel)
16:14:50kugeldoes anyone know speex?
16:19:05copperIt's a speech oriented audio codec, but that's not important right now.
16:31:46kugelcopper: right
16:31:56kugelI've got a question on how to use it
16:56:44***Saving seen data "./dancer.seen"
17:10:08 Quit lebellium (Quit: ChatZilla [Firefox 27.0/20140116125114])
17:11:01 Join lebellium [0] (
17:25:43BarankoLooks like i found some dependencies. But as for me it looks very strange,
17:25:43BarankoUsing linux box.
17:25:44BarankoIf i insert card transcend 4gb 6 class and copy 3gb data into internal memory of player (sd memory mounted too, but not used) all goes fine.
17:25:44DBUGEnqueued KICK Baranko
17:25:44BarankoIf i insert card transcend 32gb 10 class and copy 3gb data into internal memory of player (sd memory mounted too, but not used) all goes bad, sansa disconnects, hangs etc.
17:26:10Barankodon't know how it could be, but i reproduce it right here right now.
17:35:54 Join krabador [0] (~krabador@unaffiliated/krabador)
18:04:29 Join krabador_ [0] (
18:04:46 Quit krabador (Ping timeout: 260 seconds)
18:08:23 Quit krabador_ (Client Quit)
18:10:24 Join krabador [0] (~krabador@unaffiliated/krabador)
18:11:21 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
18:16:39 Quit krabador (Quit: Sto andando via)
18:17:46 Quit Baranko (Quit: Ухожу я от вас)
18:23:34 Join krabador [0] (~krabador@unaffiliated/krabador)
18:30:06 Quit ParkerR (Ping timeout: 240 seconds)
18:30:06 Quit habys (Ping timeout: 240 seconds)
18:30:07 Quit Karrde (Remote host closed the connection)
18:30:08 Join ParkerR [0] (
18:30:51 Quit ParkerR (Changing host)
18:30:51 Join ParkerR [0] (ParkerR@unaffiliated/parkerr)
18:35:06 Quit krnlyng (Ping timeout: 260 seconds)
18:40:50 Join habys [0] (~luke@
18:50:19 Join krabador_ [0] (
18:50:58 Quit mortalis (Quit: KVIrc 4.3.1 Aria
18:51:26 Quit krabador (Ping timeout: 260 seconds)
18:56:48***Saving seen data "./dancer.seen"
18:59:52 Quit uwe_ (Ping timeout: 252 seconds)
19:00:45copperHEAD no longer reads APEv2 tags (example file: )
19:00:51 Quit krabador_ (Quit: Sto andando via)
19:01:12copperwell, sim doesn't read them
19:01:16 Join krabador [0] (~krabador@unaffiliated/krabador)
19:01:56copperand my actual Fuze+ craps out when trying to play some MPC files (playback stops)
19:02:04copperI've checked the integrity of my files
19:10:52copperthe sims definitely don't read APEv2 tags
19:10:56 Join uwe_ [0] (
19:10:56coppermy Fuze+ does
19:11:04copperI'm trying to reproduce the playback issue
19:13:44 Quit Rower (Ping timeout: 272 seconds)
19:13:50copperdamn it
19:13:57copperworks fine from the internal storage
19:20:37coppercan't reproduce it either from the microsdxc card
19:20:44copperdatabase fuck up maybe
19:31:33 Quit krabador (Quit: Sto andando via)
19:36:37 Join W00fer [0] (
19:37:01W00ferHello, someone here working on RK27XX port? I have some information I can share
19:42:11copperyup, database fuckup again
19:42:28copperand when the database is fucked up, "initializing" it from Rockbox doesn't help
19:42:45copperI know something's wrong when initializing the database takes a lot longer than usual
19:42:51 Nick DormantBrain is now known as SuperBrainAK (~andy@2001:470:8:a61::5f92:59a1)
19:42:57copperthe fix is to delete it completely and THEN initialize it
19:43:05copperdelete the database files I mean
19:47:05 Join krabador [0] (~krabador@unaffiliated/krabador)
19:57:24 Join Strife89 [0] (
19:58:59 Join krnlyng [0] (~liar@
20:19:19 Quit APLU (Ping timeout: 252 seconds)
20:20:36 Quit uwe_ (Ping timeout: 272 seconds)
20:25:40 Join Karrde [0] (
20:32:10 Join APLU [0] (~mulx@2a01:e34:ee29:12b0::10)
20:33:20bertrikn1s: looking forward to your opus improvements :)
20:33:49 Quit Strife89 (Ping timeout: 264 seconds)
20:34:26n1sbertrik: i haven't really done anything this time around, just imported jmspeex' optimizations, so far
20:35:02n1si think saratoga is planning to write some arm asm for the fft which should speed it up quite a bit i think
20:36:58saratogayeah i have some asm working nicely for the pre-1.1 code
20:37:08saratogaits broken on the current git though, and i haven't had time to figure out why
20:37:28saratogajust simple stuff to help me remember armv4 :)
20:37:42chrisjjIs theer a way to erase the databse from the RB menus?
20:38:19saratogayeah, the initialize database option erases and then rebuilds it from scratch i believe
20:41:28saratogaon a side note, i wonder if parts of the mdct could be shared between opus and our core mdct lib
20:41:45saratoganormally we can't use the mdct lib because opus uses non-power-of-2 transforms
20:42:01saratogabut i guesss the rotation code, which is actually fairly slow, could be made to work with both
20:42:03 Join uwe_ [0] (
20:42:08coppersaratoga: like I said, "initialize" doesn't always work as expected
20:42:22copperno idea why
20:42:52copperI've had that database re-initialization bug for a long time
20:43:39copperit's hard to reproduce, because one first has to cause the DAP to crap out when connected via USB
20:43:53copperand then I don't know what kind of "crapping out" causes the database bug
20:44:27copperdoes "initialize" really, physically delete all database files before rebuilding it?
20:54:51saratogai'm not familiar with the code so i don't know for sure
20:56:43copperit calls remove() on database_idx.tcd and database_N.tcd
20:56:51***Saving seen data "./dancer.seen"
20:57:00copperit doesn't seem to do anything with database.log, not sure that's relevant
20:57:17copperdoesn't touch database.ignore either
20:57:21saratogathe database.log is only made if you ahve that debug option enabled
20:57:30saratogadatabase.ignore just tells the database not to scan that folder
20:57:58CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
20:57:58*gevaerts nods
20:58:13gevaertsdatabase.ignore is not at all part of the database
20:58:34pamauryW00fer: I did some work on rk27xx but the main dev is wodz
20:58:35gevaertsRemoving it on database cleanup would be a rather serious bug
20:58:51gevaertsIt's similar to .nomedia on android and the like
20:58:53pamauryW00fer: he usually reads the logs
20:59:00copperthen I don't know what's wrong
21:03:03saratogaperhaps a rebuild while the device is running triggers some bug that doesn't happen when its first booted?
21:03:11saratogamemory corruption perhaps?
21:04:02copperhmmm, I'm pretty sure I do reboot first, then connect via USB, then erase the files, unplug, reboot, initialize
21:04:08copper(so yeah)
21:04:50copperI mean, I do all that, to make sure I don't run into the bug
21:05:20coppernext time I'll try to think of making a copy of the screwy database files
21:05:38copperand upload them for one of you guys to analyze
21:06:54coppersaratoga: does the main symptom ring any bells? "Initialize" that adds files a lot slower than normal.
21:07:07copperlike 10 times slower
21:07:21saratogano idea
21:07:26saratogai don't know anything about the database
21:11:46 Quit y4n (Quit: only amiga makes it possible)
21:20:29 Quit APLU (Ping timeout: 272 seconds)
21:24:40 Join APLU [0] (~mulx@2a01:e34:ee29:12b0::10)
21:42:09kugelsaratoga, n1s: do you know a bit about speex?
21:53:28*[Saint] kinda expects saratoga to, for some reason.
21:54:16 Join wodz [0] (
21:54:28wodzW00fer: yes?
21:58:51saratogakugel: not really, preglow did most of the work on it when i first joined the project
21:58:54saratogawhat did you want to know
21:59:19kugelsaratoga: I want to feed talk clips in chunks to the speex decoder
21:59:57saratogasorry i don't know much about that
22:00:04saratogayou want me to ask preglow ?
22:01:59kugelI guess I need to align the chunks to frame/packet size but I don't know how to do that
22:02:14saratogawhat are you ultimately hoping to do?
22:02:58kugelhave the talk clips on buflib and copy the to-be-decoded data to a small static memory area
22:03:33saratogawhere are tehy now?
22:03:53kugelon the audio buffer together with buffered audio data
22:05:30kugela) I would like to decouple this and b) the clip cache scheme for lowmem is very inefficient
22:06:13 Quit lebellium (Ping timeout: 264 seconds)
22:06:48kugelwell, actually I pushed the part that decouples it
22:07:37kugelbut to improve b) I would like to allocate clips individually with buflib (now there is one large alloc for all clips & thumbnails)
22:10:03wodzgevaerts: The guy is right about hd300 recovery after flashing bad bootloader. There are some catches but generally if the code doesn't crash too early it is possible to recover this way.
22:10:48[Saint]If its not a guaranteed method, though, gevaerts was right to say what he did.
22:11:10 Join lebellium [0] (
22:11:42wodz[Saint]: bootloader must be screwed in very sever way for this to not work
22:12:16[Saint]Right. But as long as its not a guaranteed method, its just as dangerous as the AMS recovery wiki page in my eyes.
22:12:23wodz[Saint]: I recovered like dozen times bootloader on MPIOs that way
22:13:06wodz[Saint]: Basically you have nothing to lose if you flash bad bootloader. The only alternative is to reflash with BDM
22:14:44wodzwell, you can do hardcore things like swaping nor flash also :-)
22:14:46[Saint]Maybe I parsed it differently, but I read it as a Buzfeed post: "Find one weird HW recovery trick the MP3 fatcats DON'T WANT YOU TO KNOW!"
22:15:05[Saint]It just...I dunno, it made it seem like recovery was always trivial, and we just didn;t know it yet.
22:15:18[Saint]I think gevaerts was right to add some caution to it.
22:17:04wodzYes and no. It *should* be documented as it allows to easily recover in most situations. No because I didn't know it is possible until I f*** up the bootloader on BDM enabled device and observed what is actually happening
22:18:10[Saint]I would have absolutely no problem with the information being added to the wiki if it came with a similar warning as the AMS unbrick page.
22:18:17wodzand it works ONLY because of checks in rockbox bootloader
22:18:49W00ferwodz hi
22:18:51[Saint]ie. "Not an exact science, may not work at all, the best advice is to flash known good code"
22:19:01wodzW00fer: hi
22:19:03[Saint]At your own risk, etc.
22:19:12W00feri have bought a rk2706 player
22:19:19W00ferand extracted the firmware partition
22:19:29W00ferso if you are interested
22:19:47W00feri'm doing now some inquiry about the characteristics of the firmware
22:20:16W00ferbut i need some help as well to progress further
22:20:33wodz[Saint]: Do we encourage people to try random bootloaders or we STATE LOUD TO STICK WITH TESTED BUIKD==
22:21:05W00ferpamaury thanks
22:21:11wodzW00fer: What you want to know?
22:21:23W00feri have seen your work on the tracker and read the logs where you stated rk27xx questions
22:21:47W00fermaybe it's better over pm wodz
22:21:55 Join BobJonkman1 [0] (
22:22:14wodzW00fer: we usually keep discussion public
22:22:15[Saint]wodz: ...I, I'm not even sure why you're asking me that.
22:22:38[Saint]I thought for sure my intentions there are clear.
22:23:36W00ferokay then i'll try to ask it
22:23:51W00ferwould you like to have my firmware?
22:23:55W00ferfor development
22:24:15wodzW00fer: I am not particularly interested.
22:24:49W00fer1: how to update the firmware, as i have read that just overwriting the firmware partition and rebooting boots back to old 'stock' firmware
22:25:07 Join Soukyuu [0] (
22:25:09W00fer2: how to test rockbox safe, without messing with official firmware (it has SD slot)
22:25:47W00fer3: is there anything i can do to post some info about my device so touchscreen controllers can be investigated? I have seen on the wiki that no info was present on chipsets of rk27xx players
22:25:55W00ferwithout opening that is
22:26:27W00fer4: are there any safe scsi commands that do not brick my player for testing. I have some limited linux experience
22:27:32W00feranyway: i tried some investigating with rk27 device manager and it identifies properly. Because i don't know C, nor make i cannot decompile/decompress rk27boot.bin as I have seen on various sites
22:27:47W00ferit's version 16.24, which is newer than on most information available
22:27:51wodzW00fer: The easiest way to test rockbox (or anything else) is to force the player into RKDFU mode and use rk27load to send arbitrary binary to the device. Unfortunately usually this involves opening the device. There is custom scsi command which in theory should force dfu mode but I it doesn't work on my device.
22:28:17W00ferI have external RESET button
22:28:29W00ferif i press that with connecting
22:28:34W00ferdoes it set DFU mode?
22:28:48W00feri have now the following:
22:29:48wodzW00fer: Don't know. You need to check. It should present itself as VID:PID 0x71b:0x3201
22:30:00W00feri have 3203
22:30:22W00ferthat is in normal mode, so not when service partition is opened.
22:30:30W00fersee screenshot
22:31:11wodzin rkdfu mode you don't get any drive
22:31:38W00ferany idea how to send the scsi command from windows?
22:31:56W00ferand how to proceed further?
22:32:12W00ferand does the binary overwrite the original firmware?
22:32:33wodzYou can force this mode by shorting data pins of nand chip with some metal object and turning on in this state. This causes flash read fault and brings you rom rk dfu mode.
22:32:56wodzW00fer: sorry, I have 0 experience with windows
22:33:23W00ferthe scsi command sounds better
22:33:38W00ferthan opening a glued device
22:34:29wodzIf scsi command works than good for you. The procedure with nand I outlined is bullet proof.
22:34:40W00ferwhere is it?
22:34:45W00ferthe command?
22:34:59W00feri can try it from ubuntu
22:35:42W00ferthat is a .c file
22:35:53W00ferhow am i supposed to run that?
22:35:53 Quit bluebrother (Read error: Connection reset by peer)
22:36:17wodzyou need to compile
22:36:27 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
22:36:36W00feryeah without knowledge of the whole toolchain
22:36:41W00fernever compiled software myself
22:38:44 Quit fs-bluebot (Ping timeout: 248 seconds)
22:39:51 Join thomj [0] (~thomj@2001:840:4243:3::100)
22:39:54BobJonkman1Hi all: I have a Sansa Fuze+ that used to run Rockbox fine. But now it won't boot, either Rockbox or the original firmware (VolDown+Power). Just plugging in USB shows "Boot version: 1.0" on the screen and I can access storage just fine, and I can still save files which remain after unplugging-replugging (so it's not the dead internal storage problem). Reset (hold Power for 20sec) doesn't do anything. Recovery mode (VolUp+USB plugin
22:39:57 Nick thomj is now known as preglow (~thomj@2001:840:4243:3::100)
22:40:05 Join fs-bluebot [0] (
22:41:08[Saint]BobJonkman1: have you tried simply copying an unmodified firmware image to the root of starage and then ejecting?
22:41:15[Saint]*storage, even
22:41:36BobJonkman1[Saint]: "starage" sound like it would be a cool thing to have
22:42:10BobJonkman1[Saint]: Haven't tried copying manually, but there's a file in the root now. Hang on, will try again.
22:44:26[Saint]Using a unix-like OS, it is possible to force the decive into a low level SoC based recovery mode and pass a firmware image to it to execute.
22:44:27preglowkugel: i don't quite understand what you're trying to do wrt speex
22:44:54wodzW00fer: The upgrade process, outlined in manual you emailed me, uses special type of .rkw file. It is not the same as the BASE.RKW you have in SYSTEM folder on hidden drive.
22:44:59[Saint]From there, you /may/ be able to restore the device using its own repair/format capabilities in the OF.
22:45:25[Saint]But someone will almost certainly have to walk you through it.
22:45:27pamauryBobJonkman1: is the battery completely discharged ? Maybe that why you can't boot
22:45:31BobJonkman1[Saint]: Nope, the Fuze+ still won't boot with a fresh in root
22:45:42[Saint]BobJonkman1: pamaury is the man to talk to here.
22:45:49kugellet's say I have a voice clip of 1.5K. I want to copy the first 1K to a spare area and decode from that. Then I want to copy the rest (0.5K) to the same spare area and deocde the remaining bytes
22:45:49pamaurytry to plug the device in OF USB mode (hold volume down when plugging usb)
22:45:52[Saint]You;re very lucky he's here. :)
22:45:53kugelpreglow: ^
22:46:07BobJonkman1pamaury: Possible. It's been plugged in for hours, so possibly the battery isn't holding a charge
22:46:09W00ferah you are the same guy of the e-mail
22:46:14W00ferdidn't knew that
22:46:25pamauryBobJonkman1: bootloader 1.0 of Rockbox doesn't charge, only version 2.0 does
22:46:28preglowkugel: sure, sounds sane :>
22:46:42W00feryup that .rkw file is the one I don't have
22:46:43pamauryyou should probably update to the newest bootloader
22:47:12kugelpreglow: I don't know how to do that properly :)
22:47:29W00ferany way I can assemble a .rkw from that hidden partition myself?
22:47:41kugelwith the naive implementation the decoder errors out on the second chunk
22:47:42preglowkugel: well, as long as your buffer size is always larger than the minimum speex frame size, it should just be a matter of copying and decoding
22:47:56preglowif i remember correctly, we do variable size frames by default, tho
22:48:07BobJonkman1pamaury: Hmmm... How to charge it then? Use an external power adapter instead of the computer's USB?
22:48:31[Saint][10:45:51] <pamaury> try to plug the device in OF USB mode (hold volume down when plugging usb)
22:48:55wodzW00fer: In theory yes. I didn't try however. I am not aware of any tool to do this so you would need to write one first.
22:48:57gevaertswodz: hmmm, I somehow read that as h300, not hd300. Feel free to delete my comment (I can't right now)
22:49:17[Saint]Oh...holy crap.
22:49:22[Saint]I did the same thing.
22:49:27[Saint]wodz: gevaerts: ^
22:49:35[Saint](sorry wodz :'( )
22:49:40BobJonkman1pamaury, [Saint]: Plugged in with VolDown, nothing happens.
22:49:53wodzgevaerts: you read it right BUT the mode works the same.
22:50:10wodzgevaerts: Its just how coldfire bootloaders work
22:50:25W00ferwodz do you have any experience with the rk27 sdk?
22:50:27kugelpreglow: what I'm seeing is that the decoder attempts to deocde the whole buffer regardless of frame/packet size (1024 bytes) and sets the overflow bit
22:50:31W00fersince it is available
22:50:33BobJonkman1pamaury [Saint]: Ah, a few seconds later it shows the "Boot version: 1.0" screen, and connects as storage device
22:50:57pamauryBobJonkman1: are you running Linux or Windows ?
22:51:02kugelreading some docs I think that's because we don't insert terminators into the talk clips so it must decode the clip at once?
22:51:07BobJonkman1pamaury: Linux
22:51:13[Saint]I totally forgot about the 1.0 bootloader not charging
22:51:23preglowkugel: maybe it does :> i think mike sevakis had to deal with most of that side of things
22:51:25*[Saint] should probably stick to the devices he knows well
22:51:26wodzW00fer: I know it quite well but never attempted to actually build the code. I use it as a reference only (as a last resort)
22:52:07pamauryBobJonkman1: 32 or 64-bit ?
22:52:09n1skugel: sorry i don't really know anything relevant about it
22:52:18W00ferdoesn't bring me much further after all
22:52:25W00ferstill my question holds
22:52:30W00ferif i can set it in DFU moe
22:52:43W00feram I able to safely run the rockbox code?
22:53:09W00ferthere are reports that rk27 platform is "virtually unbrickable"
22:53:17BobJonkman1pamaury: 64bit (Linux Mint 16, Petra, uname -a is "Linux zelazny 3.11.0-12-generic #19-Ubuntu SMP Wed Oct 9 16:20:46 UTC 2013 x86_64 x86_64 x86_64 GNU/Linux"
22:53:23W00ferbut I don't want to take the chance
22:54:05wodzW00fer: You can try with the tool I pointed you out. You can safely run code in dfu mode. The platform is unbrickable in the sense that the worst you can end up is rom dfu mode :-)
22:54:29pamauryBobJonkman1: download and
22:55:17W00ferbut how to get out of dfu mode?
22:55:22pamaurywhen you are done, plug the device in recovery mode (by holding volume up), then (probably as root) run ./sbloader_64 <path to>
22:55:43pamauryif it works correctly, it should show rockbox USB bootloader mode, but showing "2.0" instead and it should charge
22:55:45kugelpreglow: :'(
22:55:54wodzW00fer: But the truth is that THERE IS NO ROCKBOX for your device. You would need to write drivers.
22:56:24wodzW00fer: If you are lucky this will be tiny part probably but still.
22:56:32W00feri suspected otherwise since the changelog
22:56:39W00ferand the status page
22:56:45W00ferwith 100% greens
22:56:49wodzwhich status page?
22:56:53***Saving seen data "./dancer.seen"
22:57:18preglowkugel: where does the code decoding out of the audio buffer live these days? :)
22:57:29wodzthis is for reference design (RM970 and clones)
22:58:11W00fermaybe mine is a ramos clone
22:58:12W00ferafter all
22:58:24wodzrm970 is not touchscreen
22:58:25kugelpreglow: in talk.c it calls mp3_play_data(). this passes the (entire) voice clip to voice_thread.c
22:59:00W00fertouchscreen is just an add on
22:59:05W00ferextra chip
22:59:10wodzW00fer: You could reuse most of the code but I bet there are device specific part.
22:59:14kugelin voice_thread.c it initializes the deocder and sets the speex buffer to the one from mp3_play_data()
22:59:32W00feris there any active rk27 development?
23:00:11BobJonkman1pamaury: OK, downloaded.
23:00:25wodzW00fer: That depends what you mean by active. I play with the hardware from time to time but my spare time is limited.
23:00:45pamauryBobJonkman1: then run it as I explained
23:00:54W00ferare you the only developer for that platform?
23:01:02 Join saratoga_ [0] (123e1c65@gateway/web/freenode/ip.
23:01:09wodzW00fer: basically yes.
23:01:16BobJonkman1pamaury: OK, here goes. Thanx in advance!
23:01:34preglowkugel: code seems to assume all the data is contained in the bit buffer
23:01:51W00fertoo bad, i would like to contribute as much as I can
23:01:52saratoga_W00fer: if you're not into c programming, basically none of this matters to you since you won't be able to port anything to your device
23:01:58preglowkugel: which is assumed to be one contiguous area of mem
23:02:16kugeland what can I do to change this assumption?
23:02:24W00fersaratoga_ thanks for the comment
23:02:32preglowkugel: havin a look now
23:02:39W00feri'm just more than dumb end user
23:03:15W00feri would like to contribute, sure I'll understand that you can't learn me do C in a day. But I know people that can, but don't have any rockbox experience
23:04:01saratoga_i'd probably start with something much simpler if you want to learn c
23:04:16W00ferstarts with, what can i do to help.
23:04:18saratoga_most people learn to walk before they run
23:04:33W00feri'm not talking coding, thank you
23:04:41wodzW00fer: The port for your device is doable but YOU need to take care of porting. I can share my knowledge about the platform but I don't have time to hack yet another random player.
23:05:15W00ferthere is no "generic" rk27 build that i can execute to test and report results?
23:05:39wodzwell, there is but I doubt it will work
23:05:55W00ferbecause of touchscreen okay
23:05:59saratoga_W00fer: i think you're a little confused about how porting works
23:06:00kugelpreglow: cool
23:06:09saratoga_you have to reverse engineer your device, then write device drivers for it
23:06:15W00feri know that
23:06:21saratoga_theres nothing to test until you've done that
23:06:29 Quit toehser (Read error: Connection reset by peer)
23:06:43W00ferbut since the platform is the same, and drivers are there (although from rm970)
23:06:48kugelpreglow: as I said, I'm assuming we must include more terminators at encoding to make it not having to decode the whole buffer at once
23:06:57kugelbut I may be wrong, I know nothing about speex
23:07:18 Quit n1s (Quit: Ex-Chat)
23:07:23preglownaw, shouldn't need to do anything at encode
23:07:30BobJonkman1pamaury: OK, result: "Cannot probe transfer size". I've read the sbloader −−help; is transfer size the filesize of the .sb file?
23:07:59pamauryno, did you run it as root ?
23:08:03W00ferbut yeah, it might not work, but we didn't test it. Let's just say i think that rk27xxport on rockbox wiki then needs a name change to rm970port
23:08:13W00fernot to confuse people with rk27xx platform players
23:08:13BobJonkman1pamaury: Yes, from a root shell
23:08:18kugelpreglow: when I pass only part of the buffer the decoder returns with 0 at some point with the overflow bit set
23:08:28kugeland bits.charPtr is set to the last byte
23:08:32pamauryhum, then try to specify the transfer size: it should be 1024
23:08:34W00feras it only works on that specific player
23:08:42pamauryiirc the option is "-x 1024"
23:08:54wodzW00fer: The rk27xx part is universal (more or less) but then there are 1000000 ways to hook up additional hardware. Different ways to connect buttons, different lcd controllers, etc
23:08:59BobJonkman1pamaury: OK, will do.
23:09:43W00ferso why is it being reported then as universal platform, as only 1 player will work and most probably none of the other players
23:10:07W00ferand never will
23:10:11BobJonkman1pamaury: Aha! Response: "Status passed", and the device rebooted. Still says "Boot version: 1.0" tho
23:12:36wodzW00fer: Sorry but you don't understand the basics of electronic design then. You start from generic (reference) design and add your stuff. This port is for this generic part. Some manufacturers do pure generics, some change software only and some change something in design.
23:13:06wodzeven rk SDK has many conditionals
23:13:18pamauryBobJonkman1: hum, ok if you don't mind, I will prepare a special firmware to recover from very low battery for tomorrow and we can retry tomorrow ?
23:13:23W00feryeah and i'm enthusiastic to see if from that reference design i can generate some information for you to develop further :)
23:13:32BobJonkman1pamaury: That would be wonderful.
23:13:40W00ferand i do understand that you don't want to develop for some random player
23:13:49wodzW00fer: No you can't.
23:14:18saratoga_W00fer: like i said before, if you're not going to do the work yourself, don't worry about this and just go buy a supported player
23:14:26W00feri already have
23:14:57W00feri just expected a bit more support for someone who is trying to help
23:15:02wodzW00fer: Don't take it personal, maybe your player is nice and dandy but I don't have interest in spending my time on this.
23:15:26saratoga_W00fer: if you're not working on it, then you should not expect anything at all
23:15:27wodzW00fer: I can share the knowledge which will save you something like few hundred hours of work that is
23:16:34W00ferif you are not actively developing the platform, and i will have to do device specific porting on my own its a waste of time (even for my friends who do know C).
23:16:58W00feri was just curious about the active state of development
23:17:06saratoga_talking to you is a waste of time, send whoever you know that wants to work on this in
23:17:16saratoga_then let them discuss what they need to know
23:17:40W00ferwhy are you even participating sara, didn't knew i asked
23:18:08W00ferthanks wodz
23:18:10wodzW00fer: you are rude, to speak mildly
23:19:30W00ferhmm i asked quite politely, just some simple questions.
23:19:57wodzW00fer: No, you insist to do the work for you. That is quite different
23:20:38W00feri just wanted to know how to execute rockbox and send the reports, i never asked for a full player support.
23:20:54W00feror end user ready version
23:21:06saratoga_W00fer: didn't we say that you can't ? what is the problem
23:21:25wodzThere is NO enduser version for your player. I stated this already
23:21:44W00ferit sounds weird to me that nobody is interested in debugging information for their own benefit
23:21:48W00ferbut that's my mistake then
23:22:02W00fersorry bout that
23:22:02saratoga_there is no information you can provide that would be beneficial
23:22:07wodzW00fer: And I said how to run custom code.
23:23:20W00fergood luck then
23:23:36saratoga_anyway, this argument is pointless
23:23:42saratoga_lets get back on topic
23:24:08W00ferthanks for the help saratoga
23:24:12W00ferappreciate it
23:25:13 Part W00fer
23:26:10[Saint]Some people...
23:27:11[Saint]Do we somehow come off to the public as being professional or something?
23:27:37 Quit krabador (Quit: Sto andando via)
23:27:51[Saint]Errr...I mean, do you think somehow people are getting the impression that this is anything more than a community hobby project?
23:27:55BobJonkman1Wow... And here I'm swept off my feet by the level of support *I'm* getting.
23:28:35[Saint]BobJonkman1: We do try and make it painless where possible.
23:28:52[Saint]Its a LOT easier when those seeking support are calm and rational, though.
23:28:53wodz[Saint]: The guy was simply clueless how embedded developments works
23:29:00saratoga_its that awkward problem of needing to know a lot before you can even comprehend how little you really know about software
23:29:18BobJonkman1[Saint]: I've always been exceptionally impressed by the support provided on FAIF projects
23:29:36saratoga_people think that since they can use a computer that they're most of the way to being an experienced embedded engineer
23:29:46saratoga_and get upset when told otherwise
23:29:49 Join krabador [0] (~krabador@unaffiliated/krabador)
23:30:55[Saint]The thing that seemed to push that user over the edge was the suggestion that no one is actually interested in doing this work for him/her and that if s/he wanted it to happen, they were the best candidate for the job.
23:31:06wodzsaratoga_: well they *are* on the way to be experienced embedded engineer. It just something like 1000 hours of learning
23:31:41BobJonkman1A few months ago I was on the #ownCloud channel asking for support, preparing to make a SoftwareFreedomDay presentation on ownCloud. I got to speak to a developer in that channel too, who was actually just a few km from where I was. He actually came to my SFD presentation and helped present! Try to get that level of involvement from commercial projects
23:32:19 Join W00fer [0] (
23:32:20[Saint]BobJonkman1: A huge part of it revolves around people doing what they love.
23:32:29W00fertoo bad u guys talk about me when i was away
23:32:32[Saint]It isn't hard to be passionate about a topic you enjoy.
23:32:33*BobJonkman1 wishes he had better coding chops to make contributions
23:32:35W00ferbut i can read logs
23:32:45SoukyuuShort question; the bootloader linked for creative zen on the wiki page is betav2, i see a link to betav3 on the forum thread, should i be updating or was that only something to test a specific issue?
23:32:50wodzI didn't know much about C or electronics not speaking about asm and reverse engineering when I started
23:32:51[Saint]W00fer: Its not a secret club
23:33:36W00fernope it isn't
23:33:42[Saint]Soukyuu: Dvelopment thread trumps wiki.
23:33:57Soukyuu[Saint], thanks
23:34:23saratoga_i started out knowing how to design processors in verilog and built a few MIPS ones, but still read assembly at a rate of like 3 opcodes a minute after like 8 years on this project
23:34:38Soukyuuis the hierarchy irc > forum thread > wiki?
23:35:17saratoga_usually whatever is newer
23:35:24[Saint]IRC is where we would prefer you to seek support/advice, but the forums are no less accurate or less likely to get a response.
23:35:42 Quit pamaury (Ping timeout: 260 seconds)
23:36:01[Saint]The wiki can be potentially misleading, as there's no obligation for anyone to keep the information contained within current.
23:36:14W00feryeah i experienced that
23:36:26[Saint]So there can be ancient contradictory documentation mixed amongst the latest current documentation.
23:36:51W00ferif it was accurate, i would have never asked this question as i thought that it was accurate for the platform.
23:36:58W00ferbut fine with that
23:37:19preglowkugel: yeah, the decoder pretty much requires the entire frame to be present in the bit buffer
23:37:40[Saint]There's nothing misleading about the Rockchip page(s) I'm aware of.
23:37:55preglowkugel: i can't really tell you too much about this, and i doubt mike could either
23:38:15kugeli found the decoder returns quite a few times before the clip is fully decoded, so it does actually decode in chunks
23:38:19preglowkugel: you basically have to keep feeding the decoder a bit buffer of more than the biggest frame speex does, then keep track of how much it eats
23:38:29wodzW00fer: It is accurate for the platform. There are drivers for timers, lcdif, pll, cache and whatsever. That doesn't mean you can take arbitrary player build around rk27xx SoC and expect everything to be supported
23:38:29preglowkugel: oh sure
23:38:46preglowkugel: what we do now is pass it all the data at once, and feed it the same bit buffer until speex itself says it's done
23:38:55preglowkugel: speex itself knows perfectly well how big its frames are
23:39:00W00ferthen it should display that it is a platform, different than a port
23:39:10W00ferfor end user hardware
23:39:13W00ferit is in the same list
23:39:37preglowkugel: the main speex.c codec we have does this, but it has some ogg framing that hides the raw details
23:39:44W00ferin my opinion
23:40:07preglowkugel: but it too has chunks of multiple speex frames delimited at the proper boundaries
23:40:16saratoga_W00fer: no offense, but you're not qualified to have an opinion here
23:40:23kugelyea I looked at that, the main codec seems to have more information that the voice clips offer
23:40:39W00feri do take it as an offense, sorry bout that
23:40:51preglowkugel: ah, there's one main gotcha here
23:40:59preglowkugel: we never flush the bit buffer to byte align it
23:41:12[Saint]The Rockchip wiki is very clear about it being for the SoC itself.
23:41:31[Saint]It doesn't state it will "Just WOrk" on any arbitrary rk device.
23:41:58W00fersaint, where is that it only is for that soc?
23:42:18W00fermust have misread then
23:42:52[Saint]That page is only documenting the SoC variants.
23:43:02[Saint]No other hardware is mentioned.
23:43:29[Saint]It is possible to run arbirary code on these devices, but that is very far from a fully supported port.
23:44:25W00ferand i asked thus, how i can run that arbitrary code to experiment :)
23:44:39[Saint]And wodz told you.
23:45:20kugelpreglow: how can you do that?
23:45:23preglowkugel: that can be worked around without breaking the files we already have, but it's going to mean you have to manually bit align every time you reorder the buffer of data you give speex
23:46:05[Saint]Ah, slight apology, I thought you were referring to the Rockchip SoC wiki page.
23:46:08kugelcouldnt you stash bits.bitPtr and re-use the last byte?
23:46:14[Saint]...but that doesn;t really change much.
23:46:31preglowkugel: probably, yeah
23:46:35W00ferit states players
23:46:37W00ferpandora, onda
23:46:45preglowkugel: yeah, almost certainly
23:47:24[Saint]iPo...huh? There's no Rockchip iPods.
23:47:35W00fertarget list
23:48:41W00ferbut okay, back on topic
23:49:01preglowkugel: so, basically either learn how to read speex frame sizes to ensure you have a whole one, or base yourself on the max frame size, so you know speex will never run out of data
23:49:20preglowkugel: and fixup the bit buffer when you notice you haven't got enough anymore
23:49:37preglowcopying from pcmbuf
23:52:31kugelwell, the maximum frame size seems easy: I read it's 20ms of audio, so as a simplification the max. frame size is 20ms of pcm data
23:52:50preglowyeah, but drastically less in speex form
23:53:36preglowafaik, speex frames are fixed size, speex in vbr mode will choose from a list of frame sizes it can make for a given frame of pcm audio
23:53:40kugelright but it's still small enough to not be a problem. I'm considering a 1K buffer which should be more than sufficient
23:53:54preglow1k is almost certainly more than you'll need, yeah
23:54:01 Quit foolsh (Read error: Connection reset by peer)
23:54:10kugelI'm off now, thanks for your help
23:54:16preglowno worries
23:58:25saratoga_preglow: any interest in working on codecs these days?
23:58:38 Quit kugel (Ping timeout: 252 seconds)
23:58:42preglowsaratoga_: always, depends on the codec, but no time atm or for the next months
23:58:47 Join toehser [0] (~tom@Connqueror.Toms.NET)
23:58:48saratoga_yeah same here
23:58:54saratoga_any interest in Opus?

Previous day | Next day