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 2011-08-08

00:00:21bertrikfunman, it seems that people with a clip+ with .16 firmware have the problem. I don't know which commit in particular fixed it, but at least a recent bootloader fixes it.
00:02:07bertrika bit odd really, that we apparently fixed something without first running into the problem in the first place
00:03:14 Nick GeekShadow is now known as testtest (~Antoine@reactos/tester/GeekShadow)
00:03:40 Nick testtest is now known as GeekShad0w (~Antoine@reactos/tester/GeekShadow)
00:04:19bertrikBagder, saratoga has a binary somewhere, but I don't know where exactly myself, so I guess we'll have to wait for him
00:06:18 Join [Saint] [0] (
00:07:01jhMikeSseems to be some strange distorion in this version of the vgm codec :\
00:07:33jhMikeSsounds like it's a phonograph record at times
00:08:35jhMikeSor is that the new wax cylinder DSP effect ? :)
00:09:24 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 3.6.18/20110614230723])
00:09:25bertrikBagder, see here for the binary clip+ bootloader :
00:17:18 Quit BHSPitMonkey_ (Remote host closed the connection)
00:18:38 Quit ruler5 (Remote host closed the connection)
00:19:05 Join ruler5 [0] (
00:21:27 Join GeekShadow [0] (
00:23:34 Join T44 [0] (
00:24:38 Join Scromple [0] (
00:27:21 Quit Topy44 (Ping timeout: 250 seconds)
00:27:42 Join nick-p [0] (
00:27:53nick-pHi, I wrote a patch that adds a persistent sleep timer a while ago and have been maintaining it since:
00:28:03 Quit GeekSh4dow (Quit: The cake is a lie !)
00:28:05nick-pThe subject came up recently in the forums, I've been told that the style generally looks good, but to bug people on IRC about unforeseen problems/ getting it committed:
00:29:20 Join stripwax [0] (
00:30:32 Quit pamaury (Remote host closed the connection)
00:42:16 Quit funman (Quit: leaving)
01:00:36 Quit ender` (Quit: Just because I have a short attention span doesn't mean I)
01:00:40 Quit mudd1 (Ping timeout: 264 seconds)
01:00:46 Quit stripwax (Quit:
01:02:19 Quit GeekShad0w (Quit: The cake is a lie !)
01:14:48 Join saratoga2 [0] (9803ec71@gateway/web/freenode/ip.
01:16:00 Nick ruler5 is now known as ruler501 (
01:16:01 Quit bertrik (Quit: :tiuQ)
01:17:26 Nick ruler501 is now known as ruler (
01:17:30 Nick ruler is now known as ruler5 (
01:17:50 Nick ruler5 is now known as ruler501 (
01:19:04 Quit saratoga2 (Ping timeout: 252 seconds)
01:29:11 Nick ruler501 is now known as ruler5 (
01:38:15 Join wtachi [0] (
01:50:56 Quit sideral (Ping timeout: 250 seconds)
01:54:15***Saving seen data "./dancer.seen"
02:10:32 Join Keripo [0] (
02:13:17 Join ntkm [0] (
02:14:33 Quit ntkm (Client Quit)
02:14:39 Quit liar (Quit: hallowed are the ori!)
02:42:15 Quit domonoky (Read error: Connection reset by peer)
02:46:45 Quit Thra11 (Ping timeout: 255 seconds)
02:59:51JdGordonAlexP: can you change wps_tags.tex to show a numbered list for the %cs tag like i've done here?
03:00:03 Nick ruler5 is now known as OriginBot (
03:00:32 Nick OriginBot is now known as ruler5 (
03:09:34 Quit wtachi (Quit: WeeChat 0.3.5)
03:13:45 Join HaimN [0] (~quassel@
03:18:52 Nick ruler5 is now known as ruler501 (
03:19:34 Nick ruler501 is now known as ruler5 (
03:20:00 Nick ruler5 is now known as ruler501 (
03:20:54 Join Judas_PhD [0] (
03:21:10 Quit saratoga (Ping timeout: 252 seconds)
03:23:39 Quit mgue (Ping timeout: 260 seconds)
03:25:16 Join mgue [0] (
03:28:49 Quit Judas_PhD (Quit: This is a quitting message)
03:54:19***Saving seen data "./dancer.seen"
03:55:30 Quit nick-p (Quit: CGI:IRC (EOF))
04:09:12 Join funman [0] (~fun@rockbox/developer/funman)
04:15:17 Join wtachi [0] (
04:16:55 Quit HaimN (Remote host closed the connection)
04:19:20 Quit [Saint] (Disconnected by services)
04:19:21 Join S_a_i_n_t [0] (
04:20:38 Quit amiconn (Disconnected by services)
04:20:38 Quit pixelma (Disconnected by services)
04:20:38 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:20:40 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:20:42 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:20:56 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:38:16 Quit [7] (Disconnected by services)
04:38:28 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
05:11:56 Quit Keripo (Quit: Leaving.)
05:14:31 Nick S_a_i_n_t is now known as [Saint] (
05:26:07 Quit funman (Quit: leaving)
05:43:34 Quit Horschti (Quit: Verlassend)
05:53:19 Join Rob2222 [0] (
05:54:22***Saving seen data "./dancer.seen"
05:57:00 Quit Rob2223 (Ping timeout: 260 seconds)
06:16:29JdGordonmoving here as it makes more sense
06:16:52JdGordonI'd happily scrap in favour of just maintaining cfg files
06:17:06JdGordonwe'd probably only need 3 or 4 different ones
06:17:37[Saint]I think it makes sense to do so.
06:18:06[Saint]I was thinking a .cfg for each target...but you're right. It would be possible to just have a few for all.
06:18:22[Saint]probably more than 4, but less than ~25 odd
06:18:33JdGordoni doubt that
06:19:12JdGordoncharcell, mono, grey, colour, colour+touchscreen, raaa
06:19:34JdGordonactually no
06:19:50JdGordonthe .cfg should still be generated but from a much smaller listing, similar to WPSLIST
06:19:59[Saint]You'd need one for each userfont size at least.
06:21:33[Saint]convbdf and font versioning would be a rather good thing to come of this. I'm not sure where to start there though. Building the fonts each time has *always* seemed unnecessary to me.
06:21:53[Saint](unless the sources have changed, of course, but that's a rarity)
06:22:06JdGordonthey get rebuilt?
06:22:53[Saint]I'm pretty sure they do, but I might be wrong there. I may be seeing it a lot due to my use of git stash/pop
06:25:23[Saint]No, on second thought...I don't think they do. I'm mixing it up with my use git gat stash and pop and another conversation I had the other day where I thought it was wasteful for the buildservers to start clean each time.
06:29:43*[Saint] sees the %cs changes and facepalms... ;)
06:29:49[Saint](in a good way, though ;))
06:31:07[Saint]I need to make a buttload more bitmaps for the titlebar ;)
06:31:32JdGordonI only added 2 more?
06:31:56[Saint]3. but so far that means 9 bitmpas.
06:33:26JdGordonhopefully none more need to be added for a while
06:33:40JdGordonmost screens should be very uniquly identifiable now
06:33:54[Saint]That's good though, I can now present some useful title for all the major screens.
06:33:59*JdGordon is mildly annoyed he didnt add the push/pop_current_activity() thing months ago
06:36:28[Saint]I'm struggling for wording for the "Setting Chooser" list.
06:36:41JdGordonas did I apparently :p
06:36:48[Saint]At the moment, I think I have "Make A Selection"...but that seems very robotic.
06:37:24JdGordonyou really shoudnt be doing the titles though
06:37:32JdGordonit is going to look shit on non english users
06:38:31[Saint]The problem is it looks ass drawing the title on the titlebar with the font engine, even with antialised fonts.
06:40:24[Saint]I'll probably end up dropping it though, you're right. Looking pretty isn't nearly as important as translation.
06:40:54[Saint]If I was insane, I could translate by hand and make a bitmapstrip for each language and load it conditionally checking the language settings ;)
06:41:19JdGordondoesnt work that way
06:42:46[Saint]I'm sure it could...."if lang == foo set var bar" then check the variable in the titlebar in the .sbs/wpb and display the right bitmapstrip set.
06:42:54[Saint]You'd need them all preloaded, though.
06:43:22JdGordonyes, if you wanted to have 25MB of unused bitmpas in ram
06:44:07[Saint]It's insane, though. And doesn't need that much thought. I just though I *could* do it... that doesn't mean I think its practical by any means, though :D
06:44:55[Saint]I need to check now if all the strings I want are in english.lang
06:45:23[Saint]I _think_ that was one of my deciding factors in the end for going with the bitmapstrip also. Wanting to word things slightly differently.
06:46:01[Saint]the difference between "Plugin/Plugins" and a few other strings that didn't seem to quite fit the part.
06:46:47 Quit [Saint] (Quit: Imagination is for turbo-nerds who can't handle how kick-butt reality is. I'm a kick-butt reality master! I would rather die, than be imaginative. I mean that.)
06:48:56 Join [Saint] [0] (~Saint]
07:05:09[Saint]Right...its not an absolute need, but something (that makes sense, isn't robotic sounding) in the egnlish.lang for "Setting Chooser" would be nice.
07:05:31[Saint]If I'm to translate it, the only choice really is to re-use "Settings" for this screen.
07:06:01[Saint]...which isn't so bad, I guess.
07:07:20[Saint]Also, "Plugin" seems to be missing from english.lang
07:07:54[Saint]there's "Plugins" but that doesn't fit the use case of a singular plugin that keeps the .sbs while running.
07:08:05JdGordonyou should never see "Plugin" though
07:08:25[Saint]the use case I outlined above.
07:08:33[Saint]a plugin that keeps the .sbs
07:09:40[Saint]I either have to have it say "Plugins" for now, or have a blank yellow titlebar (not really an option, IMO, it looks awful)
07:11:31[Saint]there's no "Playlist Viewer" either.
07:14:44[Saint]so...ideally, 3 more english.lang strings would be nice. One each for "Setting Chooser", "Plugin" (Perhaps "Current Plugin"?) and "Playlist Viewer"
07:16:23[Saint]I can make substitutions for now, though. at the very least I'd like to get "Playlist Viewer" in there as that's the one case where I substitution really isn't possible. Neither "Playlist", Playlists" or "Current Playlist" really fits.
07:24:50 Quit [Saint] (Remote host closed the connection)
07:27:34 Join [Saint] [0] (
07:30:53 Join Thra11 [0] (~thrall@
07:54:23***Saving seen data "./dancer.seen"
07:57:25 Join mudd1 [0] (
07:59:51 Join pacovila [0] (~pacovila@
08:00:19 Quit pacovila (Client Quit)
08:11:28 Join einhirn [0] (
08:20:40 Join pacovila [0] (~pacovila@
08:21:16 Part pacovila
08:27:53 Join B4gder [0] (
08:27:53 Quit B4gder (Changing host)
08:27:53 Join B4gder [241] (~daniel@rockbox/developer/bagder)
08:32:01 Join Zagor [0] (
08:32:01 Quit Zagor (Changing host)
08:32:01 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
08:41:46 Quit powell14ski_ (Quit: powell14ski_)
08:44:01 Join ender` [0] (
08:44:19 Join Thra11_ [0] (~thrall@
08:44:23 Quit factor (Read error: Connection reset by peer)
08:47:17 Quit Thra11 (Ping timeout: 255 seconds)
08:47:26 Join Keripo [0] (
08:47:55 Quit pixelma (Disconnected by services)
08:47:56 Quit amiconn (Disconnected by services)
08:47:57 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
08:47:58 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
08:48:00 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
08:48:15 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
08:51:26 Join sideral [0] (~sideral@rockbox/developer/sideral)
08:53:00 Join factor [0] (~factor@
08:58:31 Quit Keripo (Quit: Leaving.)
08:59:09 Join n1s [0] (~quassel@rockbox/developer/n1s)
09:13:37 Join antil33t [0] (
09:14:40 Quit Scromple (Quit: Leaving)
09:19:04 Join pamaury [0] (
09:19:05 Quit pamaury (Changing host)
09:19:05 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:19:24 Part Zagor
09:22:38 Join Keripo [0] (
09:46:11 Quit wtachi (Quit: &)
09:48:49[Saint]the Clip(s) are 128x64x1...yes?
09:54:26***Saving seen data "./dancer.seen"
10:08:38 Quit mudd1 (Ping timeout: 252 seconds)
10:13:46 Quit sideral (Quit: Leaving.)
10:25:08 Join liar [0] (
10:44:54 Join evilnick [0] (
10:44:54 Quit evilnick (Changing host)
10:44:54 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
10:45:04 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
11:01:22 Join bertrik [0] (
11:01:22 Quit bertrik (Changing host)
11:01:22 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
11:13:30 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
11:13:47amiconn[Saint]: Yes, but keep the colour split in mind as well
11:14:12[Saint]Oh, I am.
11:14:42[Saint]There's a few annoying things I just cleaned up on the clip port. With luck, it'll get some testing on the forums.
11:17:21[Saint]I fixed up the progressbar, the progressbar background (commented out for some reason, I suspect draw order prior to the inclusing of the "backdrop" bar tag), and made RTL langs a lot more functional by reversing the viewport orientation and alignment when an RTL lang is selected.
11:18:36[Saint]I also added time elapsed/remaining to an alternating subline which alternates between track time and next track info...but that was just for the specific user.
11:20:18evilnickHey [Saint], not sure if you saw this suggestion last time, but would it be possible for you to provide a changelog that relates to the theme? It'd make it a lot easier to test any new features
11:21:43[Saint]Um....I should be able to do so, yes.
11:23:00[Saint]With some luck (I just sidetracked myself with tweaking up the Clip(s) cabbiev2 port) I'll have 240x320 and 480x800 on the tracker tonight, so that should make it easier. I'll outline the features in the initial post, then list changes in the subsequent ones.
11:24:31[Saint]evilnick: ^
11:25:15evilnickOn the RB tracker?
11:25:56[Saint]In truth I should have had it there a long time ago.
11:27:09 Join sideral [0] (~sideral@rockbox/developer/sideral)
11:27:55[Saint]I'll run everything through the SDL app when I get home and make sure no gremlins have snuck in, then I'll put it up there. I'll also put up pre-made theme packages and builds with the theme inclusive on my forum thread regarding the theme ports.
11:28:10[Saint]It's been a long time coming. Stupid sickness and real life
11:28:14*[Saint] shakes a fist.
11:29:37 Quit sideral (Client Quit)
11:34:14 Quit [Saint] (Quit: Imagination is for turbo-nerds who can't handle how kick-butt reality is. I'm a kick-butt reality master! I would rather die, than be imaginative. I mean that.)
11:54:29***Saving seen data "./dancer.seen"
11:55:31 Join sideral [0] (~sideral@rockbox/developer/sideral)
11:56:12 Join user890104_ [0] (~Venci@
11:56:28 Quit user890104 (Disconnected by services)
11:56:33 Nick user890104_ is now known as user890104 (~Venci@
12:06:53kugelgevaerts: for simple yielding functions like I/O we can use mutexes instead of making allocations fail. how does that sound?
12:07:55gevaertsWhat would the effect be? Make other allocations wait?
12:09:08kugel(or alternatively to mutex, "while(locked) yield();"
12:09:37kugelperhaps something to consider later, will continue the fail route
12:11:05gevaertsI'm a bit worried that that would force waiting in cases that's not actually needed, i.e. when another allocation could be moved to make a block that's big enough. I know that a first implementation might not be able to do that anyway, but this would fix that limitation in the API
12:12:25kugelwell, allocations won't fail (if I get things right) since in the end the audiobuffer will be shrinked to make room
12:15:59gevaertsAs long as that's possible, of course
12:18:41 Join [Saint] [0] (~Saint]
12:18:56gevaertsMy intuition says not to go for mutexes or similar, but I can see the advantages of not taking chunks out of the audio buffer if waiting a short time would avoid that, at least for cases that aren't time critical (which no allocation should be, really)
12:25:42 Join A [0] (
12:26:08 Nick A is now known as Guest88537 (
12:26:08 Quit Guest88537 (Client Quit)
12:28:44gevaertsThere are of course also cases where there's an equally "good" solution without waiting, provided you skip the first one
12:39:01 Quit antil33t ()
12:44:28 Join MethoS- [0] (~clemens@
12:51:09 Quit Thra11_ (Ping timeout: 246 seconds)
13:04:09 Quit mgue (Ping timeout: 240 seconds)
13:11:53 Quit liar (Quit: hallowed are the ori!)
13:20:29 Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:26:04 Join chu [0] (~ma@
13:26:38chuhow can I format the partition type to 0xFF
13:27:25Tornea hex editor
13:27:28Torneor sfdisk
13:27:41TorneA hex editor is the most reliable way, probably
13:28:46Tornewhat OS are you using?
13:29:42chuWindows 7
13:31:44Tornewhat have you been using to partition the disk?
13:33:35chuwith some partition manager progame they have the feture to change the partiotion type id
13:33:47Torneyeah; you can try that
13:33:55chubut I don't think the home version of acronis have that feature
13:33:55Tornebut it may not let you set it to FF
13:34:46chuI'm trying the free version of Partition Wizard Home Edition
13:35:07 Join PaulJam_GPB [0] (
13:35:45Torne <- this supports editing the MBR of disks in a reasonable way
13:38:07Torneanyway, when you have done this then hopefully the firmware wont' reformat itself any more
13:38:14Torneyou are not goingto be able to mount the partition from windows though :/
13:38:19Tornesince it won't think the partition exists *either*
13:38:35Torneand a normal rockbox build for the beast will neither recognise that partition type, nor try to mount more than one partition :)
13:38:53Torneso it won't be particualrly obvious that it has worked
13:39:09gevaertsYou could patch rockbox to show a different partition type though, after which windows will see it
13:39:23TorneProbably best to make sure rockbox can read it at all first :)
13:40:20chufirst thing first
13:40:28chui'll try to change the mbr
13:40:36chuand tell you about the result
13:40:48Torneformat the partition and put some files on it first
13:40:57chuwe will try the other thing later :)
13:41:03Tornethat way it will be easier to tell if you have mounted it successfully in rockbox later :)
13:41:35chui'll have to reboot my pc
13:42:01chubecause windows 7 don't allow me to mess with the hdd mount from rockbox
13:46:32 Quit chu ()
13:47:33 Quit PaulJam_GPB (Quit: CGI:IRC)
13:49:27 Join user8293851 [0] (~aoeu@
13:49:44 Quit user829385 (Ping timeout: 252 seconds)
13:49:58 Quit user8293851 (Max SendQ exceeded)
13:50:21 Join PaulJam_GPB [0] (
13:51:00 Join user829385 [0] (~aoeu@
13:54:33***Saving seen data "./dancer.seen"
13:57:15 Join nick-p [0] (
13:57:39 Quit user890104 (Ping timeout: 258 seconds)
13:57:44 Join user890104_ [0] (~Venci@
13:58:23 Nick user890104_ is now known as user890104 (~Venci@
14:02:12 Quit nick-p (Quit: CGI:IRC (Ping timeout))
14:02:41 Join nick^p [0] (
14:03:21nick^pSorry if you've read this before, I tried earlier, but picked a time when no one was around. I wrote a patch that adds a persistent sleep timer a while ago and have been maintaining it since: The subject came up recently in the forums, I've been told that the style generally looks good, but to bug people on IRC about unforeseen problems/ getting it committed:
14:06:10 Join lebellium [0] (
14:06:26 Quit lebellium (Client Quit)
14:08:21kugeljhMikeS: ping
14:08:34*[Saint] discovers that "Channe; Configuration" isn't automagically set to "Custom" if "Stereo Width" is adjusted to a value other than 100%
14:09:16jhMikeSkugel: yessir?
14:09:39[Saint]...making it quite possible that a user would set a custom stereo width, but not actually have it applied from not knowing that the channel configuration need be set to custom.
14:09:52[Saint]making it one of those settings that "doesn't do anything"
14:09:53kugeljhMikeS: nevermind, svn history answered my question :)
14:10:18jhMikeSthat was easy :)
14:12:38sideralnick^p: I think the patch looks fine, but it's a bit unclear to me what it actually changes, even after reading the FS task and the proposed manual changes. Maybe this would be an easier sell if you'd reword the introductory text in FS #10849 to be more self-contained and descriptive?
14:12:39fs-bluebot Dedicated sleep timer menu with persistent duration and "on power up" option (patches, unconfirmed)
14:14:36[Saint]Its perfectly understandable if you wade through the many comments, but I can understand people not wanting to do so ;)
14:17:51amiconnThis is on purpose
14:18:07kugelgevaerts: I _think_ I finished compaction enablement for the most part, now to teach the playback engine to give something up if needed
14:18:19nick^pChanServ: sideral: It replaces the existing volatile Sleep Timer with a Timer whose values a persistent. It also add the option that the Timer can be automatically started on power up. I'll have a go at rewording the patch description, is there an option to edit this?
14:18:49sideralI don't even seem to get the basics... What is persistent about the timer?
14:19:34sideralnick^p: I can edit the original message for you in case you don't have the necessary permission
14:19:40nick^pThe duration of the sleep timer.
14:20:16sideralpersistent across reboot? so I don't have to select the duration each time I start the timer?
14:20:33sideralamiconn: What's on purpose?
14:20:52amiconnStereo width being independent of channel configuration
14:20:53nick^pYes, it's now an option that's stored in the config file. The startup option is also stored there (if set).
14:21:18amiconnThis allows to quickly switch between a custom width and plain stereo (or plain mono)
14:21:24[Saint]amiconn: what benefit does that have?
14:22:11[Saint]If you set it to anything other than 100% it needs to be on "Custom" to why not do that automatically?
14:22:14kugelthere was talk to make custom open the stereo width list when selected
14:22:22kugelbut that doesn't unfortunately work with queickscreen
14:22:41gevaertskugel: finally the playback engine! Good luck ;)
14:22:42kugel[Saint]: switch back to stereo without losing the custom value
14:22:54sideralnick^p: Looks like the timer duration is configured directly on the sleep timer screen, not in the config menu as most other settings. Is there a rationale why that's a good idea?
14:23:16kugelwhich is nice when toggling between headphones and sound systems, for example
14:23:52nick^psideral: Looking at the current description, I can't think of a way to describe it more concisely. Are you not a sleep timer user? ;)
14:24:03kugelgevaerts: my plan was to re-order so that buffering is in the front, and only space from buffering can be taken
14:24:32kugelothers deal with dma or are otherwise not easily movable/shrinkable so I don't want to touch those if possible :)
14:24:49sideralnick^p: I do use the sleep timer occasionally, but I may not be familiar with the sleep-timer experts' terminology :)
14:24:54kugel[Saint]: but it's indeed very confusing
14:25:06[Saint]kugel: It is, yes.
14:25:21gevaertssounds sensible
14:25:23[Saint]Well, its not confusing if you *know* how to use it.
14:25:46nick^psideral: Yes, there was a posting in here that mentioned the Settings menu was reserved for persisten settings. After reading that I move the sleep timer menu to Settings>General>System - I'll try and find it...
14:27:41[Saint]If that is so...then I wonder why Time& Date is in System
14:27:56[Saint]there's several persistent settings there.
14:28:31[Saint](I also happen to think its a rather silly place to keep it)
14:28:45sideralnick^p: If you move the options there, you can use the default way of accessing them in the main menu: short-select starts sleep timer, long-select opens the config submenu
14:29:05gevaerts[Saint]: time is extremely non-persistent
14:29:37sideralI like making the duration a regular config setting. Then I can put it in my quick screen and change it from there.
14:30:20nick^pI can move them back to time & Date, was never really sure why they were there in the first place.
14:30:21[Saint]gevaerts: Alarm wakeup screen & Time format are, though.
14:30:42gevaertsYes, if you're going to make the duration persistent, it should be a setting
14:30:56nick^pI hadn't realised there were long short select advantages though...
14:30:56gevaerts[Saint]: right. That sounds wrong
14:31:17[Saint]IMO its another batch of settings, and should be under settings...but, I'm positive this has been discussed before. I have just never seen it discussed
14:31:25[Saint]it == Time & Date settings
14:32:30sideralnick^p: If you want to change the main menu, this needs a dev mailing-list heads up. To make it less likely that a flamewar will ensue over this, we'll have to sell this right. And for this the FS description should be massaged to have all the needed info on the first screen.
14:32:34[Saint]That would leave the System menu entirely for "system-esque" things ;)
14:33:22 Quit Keripo (Quit: Leaving.)
14:33:39nick^p"sell this right" is why I've muddled along just applying the patch for myself until now :) I've got a sinking feeling...
14:34:05sideralnick^p: I suggest to do that step (move sleep timer to main menu) last. As a first step, add the config settings, but leave "start sleep timer" in the system menu. Once that's done, we can start selling the move to the main menu
14:35:25[Saint]Apart from people just being used to it...I really see no reason for sleep timer to be in the System menu personally.
14:35:43sideralnick^p: Yeah, the politics can be gruesome, but you have to keep in mind that most developers only *sound* gruesome; they're actually just nice people with little spare time some invenstment in the project
14:35:58nick^pSo I replace the current System>Time & Date> Sleep Timer with the sub menu that I've currently placed in the Settings tree?
14:36:09siderals/some/and some/
14:36:40JdGordonI think "System > time&date>sleep timer" is still the correct place for it
14:37:06JdGordonI also odnt understand the point of a persistant setting so will keep his mouth shut (for a change!)
14:37:13[Saint]But, the thing is...I don;t really think Tine & Date belongs in System ;)
14:37:29[Saint]*Time too
14:37:48nick^pThe point is if ever night you want a time of 1 hour, you currently have to set it every night.
14:38:32nick^pI'll pop off and move the menu back into Time & Date
14:38:33[Saint]Time & Date is the only "setting-ish" group in System...I see the System menu as more of a place for review, not for the setting of things.
14:39:12[Saint]If you're unfamiliar with the UI, its probably one of the last places you'd look for time and date settings.
14:39:25sideralnick^p: I'd also prefer to start the sleep timer with a single button press. "Selling" means telling people that this is the intended benefit. The FS task as is doesn't talk about benefits yet
14:39:52 Join chu [0] (~ma@
14:40:19chuI still can't create another partition for rockbox
14:40:20sideralnick^p: Rather than removing stuff, I'd suggest to split the patch into two separate ones.
14:40:45chuacronis say that it can't write the MBR to the disk
14:40:48sideralSaint: I fully agree
14:40:50*[Saint] wonders if "Set Sleep Timer" could be added as a WPS Hotkey function.
14:40:58nick^psideral: by single button press, do you mean turning the device on?
14:41:01chueven after I restar the computer
14:41:08nick^pOr some quick screen option?
14:41:56sideralnick^p: No, I rather meant by having it easily accessible in the main menu. It will only be a single button press for touchscreen users, I guess :)
14:42:14 Quit user890104 (Remote host closed the connection)
14:42:27nick^psideral: I'm not sure what a logical split would be - nothing is removed currently, just moved...
14:43:20 Join user890104 [0] (~Venci@
14:43:33sideralOr a hotkey option, why not. (Although in the long term I'd like to have the hotkey reworked to allow it to start any option from the WPS or browser context menu, into which the sleep timer probably won't fit.)
14:43:55[Saint]The sleep time is persistent, yes? (hence the name, I guess)...have you considered adding "Start Sleep Timer" to WPS Hotkey nick^p
14:44:24[Saint]I'd find that really useful myself, I hate digging through menus when I don't have to :P
14:44:26sideralnick^p: My suggestion would be: patch 1 −− make sleep time option persistent and move it to settings menu. patch 2: move sleep timer to main menu
14:45:04gevaertsDo we really want *more* in the main menu?
14:45:12[Saint]Ewwwww.....function settings divided in different menus is...yuck.
14:45:16nick^pSaint: there's an option to start the timer on startup. I'm not sure what WPS Hotkeys are, I'll look into it.
14:45:16sideralthere could be another split out to make the "on power up" option a separate patch
14:46:25sideralnick^p: I'd suggest to postpone any work on the hotkey until the basic feature (persistent timer value) is in
14:46:35[Saint]I'm not sure sideral meant to say "main menu"
14:46:38[Saint]...I hope not.
14:46:38nick^psideral: ah it currently isn't in main menu, it's Settings>General>System
14:47:30sideralAh, maybe I misunderstood
14:47:39[Saint]adding functions to the Hotkey(s) is really very minor.
14:47:47[Saint]a one line change pretty much.
14:48:30nick^psideral: I'm not keen to split out on startup, it kind of ties the whole patch together...
14:48:52nick^p[Saint]: Cool, I'll have a look
14:48:55[Saint]nick^p: To be honest, I'm not sure how many people would use it other than yourself.
14:49:05sideralSaint: If it was in the main menu, then there would be a natural way to access its options: Long-Select always opens the config settings of the highlighted item
14:49:15[Saint]You say in your task description/coments that its almost solely for your use case.
14:49:25sideralnick^p: OK
14:49:36[Saint]sideral: No way this would get in the main menu.
14:49:50[Saint]No amount of convincing or pleading in MLs could do that I think.
14:50:00nick^p[Saint]: Me and some other insomniacs I know
14:50:13[Saint]Its a tiny feature to warrant a place on the already crowded main menu
14:50:21nick^pI don't think it belongs in the main menu
14:50:59sideralThis indeed raises the question then whether the function and its config should be in the same menu.
14:51:41nick^p[Saint] don't all features start from the developers use case?
14:52:35[Saint]nick^p: I wasn't saying don't include it, but I certainly wouldn;t bother splitting it out into another patch.
14:53:15sideralThen probably the start-timer function should be in System, and its options in Settings -> General -> System. Which one of you probably already said earlier on? :)
14:53:51*[Saint] thinks that's gross, and is yet another reason to move Time & Date to Settings ;)
14:54:40[Saint]I'd rather it all "be under the same roof" whether its System or Settings.
14:55:06sideralSaint: you mean we should put the start-timer function into Settings as well? that's what I'd find a bit strange
14:55:48[Saint]Its not too it?
14:55:56[Saint]Wouldn;t it be on/off?
14:56:05[Saint]that's a setting, no?
14:56:47sideralNo, starting the timer is a function.
14:56:48sideralIf that is a point of contention, I think the way the patch does it now (keep everything under System -> Time&Date) is the way to go forward at first.
14:56:48kugeljhMikeS: what happens if queue_send() from the thread that handles the queue_event ?
14:57:26sideralkugel: the event is queued, and the thread itself later finds it
14:57:44kugelso it's turned into a queue_post() ?
14:57:51[Saint]Well, maybe instead of "Start Sleep Timer" or whatever it should be "Sleep Timer - On/Off"?
14:57:58sideralSaint: er, just System
14:58:08sideralnot System -> Time&Date
14:58:39kugelsideral: queue_send() is supposed to block until the waiting thread did queue_reply()
14:58:58kugelwhich deadlocks
14:59:20[Saint]That way I would then expect to be able to go back to the sleep timer settings and cancel the timer by going to "Sleep Timer - Off if I wanted to go to cancell the timer for whatever reason.
14:59:25sideralkugel: ah, I mixed up the two. then I really have no answer. sorry for chiming in ;)
14:59:30[Saint]Is this currently possible nick^p?
15:00:05 Quit PaulJam_GPB (Quit: CGI:IRC (Ping timeout))
15:00:08sideralSaint: Yeah, I think the patch can cancel the timeout already
15:00:18nick^p[Saint] Yes, you can switch it on/ off
15:01:51[Saint]Ideally, what I'd like is "Sleep Timer - On/Off, Value (the timer duration), Set on Boot - On/Off" for the menu.
15:02:27nick^pThat's exactly what the patch does \o/
15:02:36sideralSaint: That's what the patch does, but it raises precisely the "shouldn't functions and setting be separated into their respective menus" question
15:03:04sideralSaint: for which you have to have an answer, and I don't like yours :)
15:03:38sideralMaybe there's a better answer? You should come up with one; this question *will* be raised
15:04:01[Saint]What exactly do you mean? I'm sure I could name instances where a function and its settings are in the same place.
15:04:13[Saint](If I understand you correctly)
15:04:15nick^pFor the sake of getting it checked in, how about I leave the current new sub menu where it as and add an on/off in the existing Time & Date menu?
15:04:24 Quit chu ()
15:05:01sideralSaint: I meant don't like the answer that arming/disarming the timer is a setting rather than a function
15:05:47[Saint]What (seemingly arbitrary, to me) magic are you using to define the two?
15:05:53sideralSaint: Yeah, there's precedence, but people seem to hate it where we have it.
15:06:33[Saint]Yet another reason (oh, wait...the same one) to move Time & Date ;)
15:06:47sideralfully agree :)
15:06:58[Saint]It all seems to point to this, then we're all happy (I think)
15:07:10sideralSaint: Re my magic to separate the two: Hmm, good question. Not sure I have a good answer
15:07:27sideralLet me try
15:08:21sideralSaint: A setting is a variable whose value is made persistent across reboot, and which is only changed by the user. it's used to configure what a function does.
15:08:49sideralA function is a way to change runtime state, that is, to do something with a side effect
15:09:17sideralA decreasing timeout is not a setting. It's initialization value is.
15:09:32sideralThat's my expectation at least
15:09:37 Join chu [0] (
15:09:52chuI have a gigabeat S60
15:10:08chuI installed a 160Gb HDD on it
15:10:24chucurrently it can only recognize 128Gb
15:10:25Tornechu: where did you get to?
15:10:36[Saint]Right, but the way I see it you're only setting it to on/off really after its set up. Making it as setting in my eyes.
15:10:37chuI still can create the new partition
15:10:45[Saint]even if it is a contentious one ;)
15:10:53chuI'm currently trying Gparted
15:11:29chucan't create the new partition
15:11:38Torneyeah, rockbox disables the ability to write to the MBR :)
15:11:44sideralSaint: With this line of argument, anything is a setting. I'm only turning playback on and off when I select "resume playback" or stop playback, right?
15:11:49Torneto avoid anything messing up the MBR "fix" we apply
15:11:55Tornei forgot about that
15:11:59 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
15:12:09chuso what should I do now
15:12:19TorneYou'll need to make a rockbox build with the MBR fix compiled out
15:12:47chuhow can I add that
15:13:01Torneremove USBSTOR_READ_SECTORS_FILTER() and the equivalent WRITE one from firmware/target/arm/imx31/gigabeat-s/usb-target.h
15:13:16nick^psideral: there's quite a few functions currently in settings "browse .cfg files" "save sound settings" for example...
15:13:40Tornewindows will probably mount the wrong partitions after that :/
15:13:52Torneand linux is likely to refuse to consider the partition table valid at all :)
15:13:59Tornethis part will be entertaining :)
15:14:07sideralnick^p: that's more of a settings meta menu with settings-related functions, really. the actual settings are in the submenus
15:14:55Torneit's *probably* possible to edit the partition table that way, though
15:15:15Tornedo you have a development environment set up to compile rockbox?
15:15:42chuI'm running ubuntu
15:15:59Tornehave you compiled rockbox before?
15:16:22 Quit [Saint] (Ping timeout: 260 seconds)
15:16:27nick^psideral: I'll move the function back and release a new patch in around 10-11 hours, need to get back to work now...
15:16:42TorneRight. You are going to need to comile it a bunch of times, probably, to experiment with this
15:16:46Tornestart with
15:16:49sideralnick^p: No need to haste :)
15:16:56TorneStart off just by making sure you can compile it unmodified
15:17:05Tornethen try the above change to remove the MBR fix
15:17:18gevaertsTorne: doesn't bootloader usb allow MBR access?
15:17:29Tornegevaerts: Don't think so
15:17:34chui think I'll try with this
15:17:36Tornegevaerts: the fix is not conditional on anything afaict
15:17:43chuthe bootloader
15:17:48chuonly recognize 128GB
15:18:10Tornegevaerts: yeah, it's there in the bootloader too
15:18:21gevaertschu: different bootloader
15:18:50chuI only use the one supply with beastpatcher
15:18:58chuare there other one?
15:19:00TorneAnyway, you shouldn't need to touch the rockbox bootloader really
15:19:16chuok I'll try to compile it first
15:19:27chuto see if I can chane the MBR
15:19:56kugelgevaerts: \o/
15:20:09kugelit seems to work :)
15:20:49kugelI forgot to set a side space for new handles so my shrink callback actually checks shrinking from both sides
15:20:53gevaertschu: the bootloader that dislikes large disks is the one in flash, which we don't touch.
15:21:07gevaertskugel: nice!
15:22:50 Quit chu (Quit: CGI:IRC (EOF))
15:26:03 Quit nick^p (Quit: CGI:IRC)
15:37:08 Join God_Eater [0] (93722cd0@rockbox/staff/GodEater)
15:38:51kugelgevaerts: so yea, I can now succesfully make an allocation while music is playing after boot .)
15:39:14gevaertskugel: does it work twice in a row? ;)
15:39:49kugelthe second time music doesn't stop because the space from the first time is still free :)
15:42:17 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
15:54:37***Saving seen data "./dancer.seen"
16:02:04dfktis there a simulator/emulator that lets me create WPS's for RaaA somewhat comfortably? specifically looking for 800x480
16:06:39 Join domonoky1 [0] (
16:08:32 Join chu [0] (
16:09:36 Quit domonoky (Ping timeout: 255 seconds)
16:19:20chuI have successful compile rockbox with the modified that you recommend
16:19:44chubut now when I boot the rockbox stop at scan sdick
16:19:58kugelare the sims and targets voice files incompatible?
16:20:08chuand a white sceen with the line
16:20:40chuStkov main
16:20:45chuwhat should I do
16:21:15Tornedid you build it unmodified first?
16:21:26Torneand did that work?
16:21:27chuit boot successful
16:21:42jhMikeSkugel: The thread associated with sending the message is moved off the sender list and is woken when queue_reply is called or another queue_wait is executed.
16:21:53Torneright, well there's no reason why changing the usb stack should cause it to fail to boot..
16:22:04Tornei can see it maybe crashing when you connect USB
16:22:42Tornethe usb storage code is not executed at startup
16:22:45jhMikeSkugel: the thread owning the queue will release the waiting thread when it next calls queue_wait(_w_tmo) or queue_reply(), whichever is first.
16:23:05chuat usb-target.h
16:23:14chuI remove every thing after
16:23:21chuvoid usb_fix_mbr(unsigned char *mbr);
16:23:23jhMikeSblah, how'd I post both of those (thought I deleted the first) :)
16:23:25chuis it correct
16:23:35Torneyou removed the #endif at the end as well?
16:23:50kugelhm, a not-in-sync voice file can actually crash rockbox
16:24:07kugeljhMikeS: what if the thread is the same?
16:24:08chushould I remove it
16:24:18Tornejust the two macro defintiions
16:24:43jhMikeSkugel: don't queue_send to the same thread that does the waiting :)
16:24:58kugelthat's what I figured :)
16:28:32chushould I remove void usb_fix_mbr(unsigned char *mbr); ?
16:30:09Tornepastebin what "svn diff" says
16:30:34chuafter I remove void usb_fix_mbr(unsigned char *mbr);
16:30:39chuand recompile
16:30:42chuI boot normally
16:30:57TorneNo, there is no reason for that to make any difference :)
16:31:04TorneYou must've been doing something else wrong.
16:31:15TorneThat's just a prototype, removing it makes no difference
16:31:44chuI don't understand eithher @@
16:31:54chuanyway since it can boot to rockbox
16:32:04chuI';; try to make the partion
16:35:59chunow gpartd can't no longer recognize the disk
16:36:18Tornequite possibly not
16:36:34Tornethe fix code is in there because the original firmware likes to write weird stuff to the flags which is not, technically, valid :)
16:36:45Torneyou may need to use sfdisk or similar horrors
16:36:49Torneor just a hex editor.
16:37:57chuso what should I do with a hex editor
16:38:17sinthetekhex edit!
16:38:36Torne :)
16:38:54TorneOK, actually creating a partition that way would be a pain
16:39:09Torneok, so I suggest:
16:39:15chuI can see the horror @@
16:39:23Tornedd if=/dev/sdb count=1 | xxd
16:39:29Tornewhere for sdb use, er, whatever the device is
16:39:32Tornepastebin that
16:39:38Tornejust to check that it really is the MBR
16:39:54Torneand see if it is indeed just suffering from the issue with the active flag being set to some nonsense
16:41:07chuubuntu@ubuntu:~$ sudo dd if=/dev/sdb count=1 | xxde 1+0 records in 1+0 records out 512 bytes (512 B) copied, 3.2675e-05 s, 15.7 MB/s No command 'xxde' found, did you mean: Command 'xxd' from package 'vim-common' (main) xxde: command not found
16:41:28Torneyes, you meant command xxd from package vim-common :)
16:41:37Tornetry again :)
16:41:41God_Eaternot xxde
16:42:03chumy bad ^^
16:42:05God_Eaterand also - Torne *REALLY* meant pastebin - not just paste into the channel
16:42:11Torneif you really don't have xxd, use hexdump
16:42:14chuubuntu@ubuntu:~$ sudo dd if=/dev/sdb count=1 | xxde 1+0 records in 1+0 records out 512 bytes (512 B) copied, 3.2675e-05 s, 15.7 MB/s No command 'xxde' found, did you mean: Command 'xxd' from package 'vim-common' (main) xxde: command not found
16:42:28Tornealso you just typed the same thing again
16:42:45God_Eateror at least, pasted the same thing again
16:43:37 Quit markun (Remote host closed the connection)
16:44:01chuI pste the wrong thing
16:44:11God_Eaterthat's a fucking broken MBR
16:44:12TorneThat doesn't look like a beast partition table.
16:44:20TorneGod_Eater: not really
16:44:36TorneIs sdb actually the right device?
16:44:40God_Eaterreally? see line 7 ;)
16:44:47chuit sbe
16:44:50TorneGod_Eater: that's the stderr from dd getting mixed in
16:44:57Tornethen, er, use sde?
16:45:13chuthat what it give when I use sde
16:45:40Torneyou pasted sdb above
16:45:52TorneThat only has one partition, so that can't be the beast
16:46:59Torneright, that looks ok
16:47:09TorneBut yeah, the beast has ruined the active flags, which seems to confuse linux
16:47:43Torneit's supposed to be 00 or 80 but the beast uses 01 for no reason
16:48:12Torneso yeah. try a differnet partitioning tool
16:48:14God_Eateryou've looked at these too often to pick that out so fast Torne ;)
16:48:16gevaertsFixing those has been safe for people in the past
16:48:22Tornegevaerts: Yeah, it shouldn't choke on it
16:48:35Torneit just writes it wrong, it doesn't read them
16:48:46TorneGod_Eater: hey just because i can read a hexdump of a partition table
16:48:55gevaerts says how
16:49:06God_EaterTorne: so can I - just not that fast!
16:49:27Torneyeah, do what gevaerts links to
16:49:33Tornethen tools should be able to work with it okay
16:49:36*God_Eater decides it's time for a cuppa
16:49:49gevaertsOf course use sde instead of sdz in your case
16:50:07God_Eaterwhoever put sdz in that doc is a hero
16:50:12TorneGod_Eater: damn right
16:50:25gevaertsGod_Eater: it just ruined my 26th disk!
16:50:38God_Eaterunless we find someone with 26 disks and a Beast.
16:50:42God_Eaterwhich I guess gevaerts is ;)
16:51:06God_Eaters/someone/an idiot/
16:51:31chulinux can see my partiotions again
16:52:17Torneok. so create the third partition and try setting it to type ff
16:52:22God_Eaterright. Cuppa.
16:52:26Torneif you find a tool that lets you do that, you win
16:52:33Tornedump the mbr again and pastebin it just to check :)
16:52:40TorneWell, er
16:52:47TorneActually you probably want to create it as type b first
16:52:53Tornethen format it as FAT32 and copy some files onto it
16:53:02Tornethen go back in and change the type afterward
16:53:17Tornejust so we can test if it mounted successfully later more easily
16:56:10gevaerts /sbin/fdisk seems to be able to set ff for me
16:56:33*gevaerts does not save. This is his laptop's disk :)
16:56:58kugelah yes
16:57:01kugelnow the fun part starts
16:57:07kugelmysterious crashes :(
16:57:38Tornegevaerts: well it's a matter of whether it saves it without clearing the actualf ields
16:57:43kugelthis one is extra mysterious because I don't think it relates the moving, though
16:58:03chufdisk can change it
16:59:16gevaertsSeems to work at least on a loopback device
16:59:54Torneanyway, yeah. linux is happy to believe that type ff is a real partition, it seems
17:00:02Tornemaybe windows will be too, which would save time
17:00:25Tornethe internet suggests ff is not in fact meant to mean it doesn't exist, maybe it's a xenix bad block relocation table?
17:00:41gevaertsI expect it will. IIRC making windows not see the firmware partition was rather tricky
17:00:43Tornebut in any case i'm pretty sure the beast ignores it
17:00:49Tornegevaerts: why did we do that?
17:01:13Torneafaik windows doesn't repsect the hidden bit any more
17:01:29Tornein fact current windows versions seem to basically ignore the partition type entirely and rely on superblock detection
17:02:49 Quit B4gder (Quit: Konversation terminated!)
17:03:14gevaertsIIRC there were two reasons: the firmware partition is a bit dangerous to expose (too easy to ruin things), especially since it's the first one, and you can't have a removable device with two filesystem-partitions in windows, so we had to mark it as non-removable, which has some (at least UI) disadvantages
17:03:57Tornegevaerts: You can
17:04:16Torneat least, i thought so
17:04:37Tornei'm pretty sure i remember having the beast plugged in and seeing both partitions
17:04:44Torneeven with the mbr patching
17:05:15gevaertsWith tb.inquiry->DeviceTypeModifier = DEVICE_REMOVABLE in usb_storage.c?
17:05:27gevaertsThere used to be an ifdef there for the beast
17:05:33Tornei dunno
17:05:49Tornewhich sense of removable is that?
17:06:08gevaertsThe SCSI INQUIRY one
17:06:13Tornewhat is the device type there for a normal flash drive or whatever?
17:07:10gevaertsI know I've seen the same (i.e. show only the first filesystem) behaviour with sd cards and some flash drives
17:07:30 Part Zagor
17:08:34Torneright, but card readers are removable in the sense of "the drive can be empty"
17:08:37Tornewhich the beast is not
17:08:45Tornesince it's a fixed disk..
17:09:05*Torne will fiddle at home later.
17:09:14gevaertsYes, but *many* devices treat USB-unpluggable the same way
17:09:49Tornesure, i'm not disputing that many devices do it wrong :)
17:09:54Tornei don't see why ti matters for us
17:10:19gevaertsI don't disagree, but I seemed to be losing the argument back then :)
17:10:26Tornein any case, i am pretty sure that windows doesn't respect the hidden bit any more
17:10:40Torneand i don't remember it respecting it for a long time, in fact
17:10:55TorneNT has pretty much always just probed for superblock formats it knows how to mount, and doesn't care if partition types are wrong
17:11:08Tornemy machien happily mounts NTFS on partitions of type b, or type 82, or whatever
17:11:29amiconnTorne: There are specific partition types for the hidden version of the various fat partition types
17:11:45Torneyes, but i don't think it cares
17:12:05amiconnRockbox doesn't, but maybe the host will?
17:12:05 Quit chu (Quit: CGI:IRC (EOF))
17:12:15TorneI mean windows
17:12:29amiconnOh hmm, the probelm is different here - host will work fine (since rockbox usb does lba48)
17:12:41amiconnSo why not just use the ordinary fat32 type?
17:12:48Tornethe beast will format itself if you do
17:12:55TorneYou can only have two partitions
17:13:21amiconnAh, yes
17:13:35Tornefor this purpose the third partition has to be type ff
17:13:41Tornebecuase the firmware assumes that means it doesn't exist
17:13:51Tornebut I believe hosts will ignore it and mount it anyway
17:14:01amiconnSo use a type internally that the beast loader doesn't care about, and let our fancy mbr translator translate it to type fa32?
17:14:19*God_Eater believes this is what torne just said
17:14:20TorneWell, if it works without translating it at all i don't see why we'd bother
17:14:31Tornethe translator basically makes it impossible to have the MBR be writable
17:14:44Tornewhich makes it very hard to actually set this up in the first place, and to repair it if the flash bootloader does decide to reformat :)
17:14:57Torneso if it's possible to do without it that's probably preferable
17:15:04God_Eaterwhat we need is our own flash bootloader
17:15:25God_Eatera la gigabeat F
17:15:26amiconnWe could have a two-way translator
17:15:39Torneyick :)
17:15:49Torneanyway this still proof of concept here
17:15:59Tornehe hasn't got far enough to prove it actually works at all
17:22:50 Quit TheSeven (Remote host closed the connection)
17:24:05 Join petur [0] (~petur@rockbox/developer/petur)
17:27:46 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
17:30:37gevaertsTorne: does the flash code care about the contents of the data partition at all?
17:32:02God_Eatergevaerts: evidence would suggest yes, on account of the occasional "I think I'll reformat everything again" behaviour some of us Beast owners have seen
17:32:26God_Eaterlord alone knows why mind you
17:33:16Tornegevaerts: it validates a bunch of fields in the BPB, but it doesn't read any files afaik
17:33:21Torneit might mount it maybe?
17:33:45gevaertsWith some extra magic, and knowledge of how far you can go without making the boot code blow up, we might be able to have the third "hidden" partition (mostly) overlap the second one, and hide the second one over USB.
17:33:58gevaertsOr maybe that's taking things too far
17:34:06God_Eaterthat's just crazy talk
17:34:12God_Eatergo and have a lie down
17:34:25TorneIf we were going to do that we could just not have a third partition
17:34:34Torneand just patch the mbr to enlarge the second partition to the full drive size
17:34:47TorneNot sure if that will work though
17:35:20gevaertsBut that assumes the things it reads from the FAT make enough sense in the 128GB world to make it not blow up
17:35:49 Join chu [0] (
17:36:03chusfdisk −−print-id /dev/hdb 5
17:36:20chuubuntu@ubuntu:~$ sudo sfdisk −−print-id /dev/sde 3 ff
17:36:33chusfdisk how that the partition 3 is FF
17:36:48chubut gparted still recognized it as fat32
17:37:01chudid I do something wrong
17:37:45gevaertsgparted probably looks at the filesystem inside
17:38:41chuok then what should I do next
17:38:57chuto make rockbox recognize both partition
17:41:37gevaertsI suspect the patch on is enough
17:41:56 Join y4n [0] (y4n@unaffiliated/y4ndexx)
17:47:18 Quit chu (Quit: CGI:IRC (EOF))
17:51:44 Quit petur (Quit: *plop*)
17:54:41***Saving seen data "./dancer.seen"
18:03:58 Join mudd1 [0] (
18:07:32n1sTorne:, gevaerts: did we release a new bootloader after the mbr filter stuff was added?
18:08:45n1skugel: target and sim voices are probably out of sync for most targets because of the features stuff depending on config defines and people like to make stuff #ifndef SIMULATOR even if not strictly necessary
18:09:10kugelthat sucks
18:09:22n1sa sim voice can be built by using the advanced config and selecting both voice and sim
18:09:24 Quit sideral (Quit: Leaving.)
18:09:26n1syes, it sucks
18:09:27kugelcould build a sim-voice thoughh
18:10:14n1si tried cleaning it up a long time ago but it's just one of those things that it's really hard to get people to not do it seems
18:10:30 Join sideral [0] (~sideral@rockbox/developer/sideral)
18:12:35 Join lebellium [0] (
18:12:51kugelyay, first "Please reboot to enable splash" removed
18:35:56 Quit ReimuHakurei (Quit: If I use this, I will disappear, and Shana-tan will remain...)
18:36:20 Join ReimuHakurei [0] (
18:41:01 Quit ReimuHakurei (Read error: Connection reset by peer)
18:41:39 Join ReimuHakurei [0] (
18:42:10 Join Transformer [0] (
18:42:46 Part Transformer
18:51:11 Join dean [0] (
18:51:33deanHey all I was thinking of trying rockbox on my ipod what does it actually do?
18:52:06 Join wtachi [0] ( :)
18:53:32Lloreandean: There's a link on the front page called "Why should you run Rockbox?"
18:54:21deanLlorean: I have read different things does it work on ipod 5th generation?
18:56:01LloreanIt still sounds like you aren't visiting the front page.
18:56:16LloreanThere's a very explicit list of what it supports there.
18:56:40deanit says 5.5g so not sure what that means lol
18:56:48 Quit wtachi (Client Quit)
18:58:52Lloreandean: it says 1st through 5.5th generation.
18:59:00Llorean5 is between 1 and 5.5, right?
18:59:37deanLlorean: True but when you go into manuals there isn't one for 5th generation
19:00:06LloreanWhat do you think the "iPod Video" is?
19:00:13LloreanThere is some minimal expectation that you know what you own.
19:01:52deanMine doesn't look like that though mine looks like the nano but its the 5th generation cos its got a camera on it
19:02:13LloreanNo, that's the iPod NANO 5G.
19:04:00 Join wodz [0] (
19:04:48deanNone of them say nano 5th generation on them tho it shows one that show ipod video and then one that shows photo. There isn't a 5th generation nano on there it only goes up to 2nd gen nano
19:05:13TorneAnd what does that tell you? :)
19:05:14LloreanYes, and the 5th generation Nano isn't listed as supported on the front page of the site either.
19:05:22LloreanThese should be clues. ;)
19:06:06deanLlorean: Thats what I said at the beginning it said it supports upto 5.5g but I was querying if nano was supported thats all lol
19:06:43Torneas it says, the nano 1g and nano 2g are supported
19:06:46Tornethis implies the other nanos are not
19:06:52Torneotherwise we wouldn't list which ones were
19:07:30Lloreandean: You didn't mention "nano" at all when you asked.
19:07:51deanTorne: Ok cool thats all I needed to know I don't suppose you know whether they will update it to support nano 5th generation
19:07:59LloreanYou asked if it supported the iPod 5th generation, which it does. The 5th generation iPod Nano is an entirely different device.
19:08:17deanLlorean: I assumed 5th generation were nano I didn't realise they had older ones which were 5th gen
19:11:47 Join wtachi [0] (
19:12:01 Join Llorean1 [0] (
19:12:19 Quit Llorean (Disconnected by services)
19:12:22 Nick Llorean1 is now known as Llorean (
19:12:26 Quit Llorean (Changing host)
19:12:26 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
19:13:05 Quit benedikt93 (Quit: Bye ;))
19:13:12 Quit dean (Quit: Leaving)
19:24:37 Quit Strife89 (Read error: Operation timed out)
19:28:33 Quit ruler501 (Quit: g2g cya)
19:31:11 Quit mystica555 (Ping timeout: 240 seconds)
19:31:36 Join saratoga [0] (9803ec71@gateway/web/freenode/ip.
19:32:19saratogaBagder: could you upload this to the download server:
19:32:52Bagderand just replace the existing one?
19:33:49saratogamight be a good idea to move the old one to a 3.0 folder
19:34:46Bagder6559ea0f4185dd8e8ba59059b3e08653 bootloader-clipplus.sansa
19:34:50Bagderis now in place
19:35:58 Join mystica555 [0] (
19:36:35 Quit fs-bluebot (Ping timeout: 255 seconds)
19:36:35 Quit bluebrother (Ping timeout: 255 seconds)
19:36:47 Join Jerom1 [0] (~jerome@
19:37:46 Join fs-bluebot [0] (
19:38:04 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
19:45:04 Join ruler5 [0] (
19:47:45 Join Horscht [0] (~Horscht@xbmc/user/horscht)
19:54:45***Saving seen data "./dancer.seen"
19:55:52 Join excalibas [0] (
19:58:50 Quit ruler5 (Remote host closed the connection)
19:59:34 Join ruler5 [0] (
20:02:24 Nick ruler5 is now known as ruler501 (
20:04:34 Join Thra11_ [0] (
20:04:55kugeljhMikeS: do you know a bit about buffering by chance? :)
20:05:23 Quit Thra11_ (Read error: Connection reset by peer)
20:06:05CIA-14New commit by buschel (r30268): Remove a floating point calculation from ay codec.
20:06:24 Join Buschel [0] (
20:06:36 Quit saratoga (Ping timeout: 252 seconds)
20:09:17CIA-14r30268 build result: All green
20:11:29wodzWhen I try to run rockbox binary on rk27xx it barfs on disk_mount_all(). It reads MBR correctly but first sector of the fat is returned as all zeros. When I read the same sector from bootloader with sd_read_sectors() into static buffer data are ok. Any hints?
20:12:09ukleinekwodz: wow, you're already that far, nice
20:12:47wodzfar? I wouldn't say that
20:13:26ukleinekwodz: compared to me, it's far, yes
20:13:56*ukleinek had a question he wanted to ask wodz, but he doesn't remember now
20:15:51sideralBagder: Do you have admin access to the Flyspray server?
20:16:00BagderI do
20:16:26sideralBagder: ukleinek has repeatedly tried to register for FS, but doesn't receive the activation email
20:16:38ukleineksideral: thanks
20:17:53Bagderusing 'ukleinek' as user name?
20:18:26sideralguess so
20:19:29ukleinekBagder: yes, ukleinek at
20:21:10wodzok it is a bug in sd driver apparently - first transfer is successful the second is fucked
20:21:24jhMikeSkugel: I think so
20:21:59kugeljhMikeS: where would I start if I wanted add the ability to free buffer space (from the front)?
20:22:23kugelit's not fatal if it isn't possible at all times
20:22:25jhMikeSin what way?
20:22:26Bagderukleinek: I added you manually now, but I think it should send you a mail with the password, possibly this fails too
20:22:46jhMikeSkugel: basically, you can free from the front or the end, but you must ask the playback engine
20:23:00ukleinekBagder: maybe it generates a log on your side? Unfortunately I cannot access the logs on my end.
20:23:17BagderI bet it does, I just couldn't find it just now
20:23:35 Join Thra11_ [0] (
20:23:42Bagderits my brother, Zagor, who fiddles with this installation mostly
20:24:08jhMikeSkugel: playback does it every time rebuffering has to happen for low buffer or a playlist change
20:25:02Bagderukleinek: if no mail gets through, I can set a password manually in a while
20:25:15jhMikeSkugel: also, atomic audio handles are locked when the codec is running
20:25:43kugeljhMikeS: I'm on the shrink callback in playback.c (exists in my local tree). instead of stopping playback I want to ask buffering if it can let some free at its front
20:26:21 Join lorenzo92 [0] (~chatzilla@
20:26:23ukleinekBagder: there is no hurry, I can collect patches here
20:26:27 Quit lorenzo92 (Client Quit)
20:26:39kugeljhMikeS: not front of the ring buffer (read position) but the real front of the underlying buffer you know
20:27:02 Join lorenzo92 [0] (~chatzilla@
20:27:07jhMikeSkugel: erm... heh
20:27:36kugeltaking away so that it's not available for buffering anymore
20:27:41kugelsafely, that is
20:31:43 Quit lorenzo92 (Client Quit)
20:31:43kugelmy idea is to go the linked list backwards and close handles until the requested space is freed?
20:31:43jhMikeSyou'd have to update buffer indexes
20:31:43 Join lorenzo92 [0] (~chatzilla@
20:31:43jhMikeSto put the buffer start at some other place in ram?
20:31:43 Quit lorenzo92 (Client Quit)
20:31:43jhMikeSyou can't close handles without the engines acknowledgement since they're stored in the track array
20:31:43 Join lorenzo92 [0] (~chatzilla@
20:31:43 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
20:31:43DBUGEnqueued KICK jhMikeS
20:31:43jhMikeSbuffering also stores array indexes
20:31:51sideralMP3 encoding question: what kind of header / padding is a long row of 0x55 bytes, sometimes interspersed with a "LAME3.98." string, sometimes with other junk?
20:32:00jhMikeSthose are also stored in the handles, which would need resync to their new offsets
20:32:47 Quit dfkt (Read error: Connection reset by peer)
20:33:38jhMikeSI'd say how it works is almost completely unsuitable for resizing the buffer like that
20:34:26kugeljhMikeS: the playback engine has knowledege that's no problem
20:34:46kugelshouldn't be anyway :)
20:35:11jhMikeSchanging the buffer start point would still muck up the ridx/widx in the handles
20:35:18kugelI mean the callback is in playback.c, but not usually run from the audio thread
20:35:23jhMikeSplayback doesn't handle those
20:36:33jhMikeSI'm not certain how it would help. playback isn't aware of the physical buffer start, just the logical "first thing in the ring buffer"
20:39:05kugeljhMikeS: playback is aware since it gives the buffer to buffering.c with buffering_reset
20:39:47kugelanyway, I would want to add this as an buffering API and fix up track handles as necessary
20:39:52jhMikeSonly there when giving it memory, but beyond that, not in its operation
20:42:12jhMikeStaking from the end of the buffer would be far simpler
20:42:21kugelit knows the underlying buffer and it knows when it might want to give some of that back to buflib
20:42:54kugeljust because of widx/ridx of each handle or is there another reason?
20:43:24jhMikeSpretty much that
20:43:40kugelthose are pretty easy to fix, no?
20:44:36jhMikeSor use something different
20:47:29 Quit Jerom1 (Quit: Leaving.)
20:47:38 Quit ruler501 (Quit: g2g cya)
20:47:54kugeljhMikeS: what do you mean by that?
20:49:29kugeljhMikeS: could perhaps take from the end and then memmove the entire thing?
20:50:17jhMikeSthat might cause codecs to fail
20:51:01 Join ruler5 [0] (~ruler_501@
20:53:53 Join Keripo [0] (
20:57:45jhMikeSmoving data would violate fixed-location guarantee for the codecs, both atomic and packet types
20:59:06 Join dunkaist [0] (
21:04:06kugelyou say playback can't handle if handle_ids aren't valid anymore? doesn't buffering return some error for invalid handles?
21:05:25jhMikeSit doesn't but skipping into what it expects is already buffered would fail
21:06:20jhMikeSwhatever is in the track list had already been vetted
21:07:03kugeljhMikeS: how would it fail?
21:08:30jhMikeSall the metadata and everything that it expects that it loaded and verified before adding more entries to the list would be suddenly invalid, wreaking havoc with the logic. making it paranoid would make it a hell of alot more complex.
21:09:52wodzDamn, now two reads in row works but multisector read fail. This sd IP is weird
21:10:46jhMikeSkugel: the only thing it expects beyond that point for a track is for the codec to possibly fail to decode and report it, which basically amounts to an auto skip
21:11:56jhMikeSif suddenly things can disappear at any time, we're in for one wild night :)
21:12:16jhMikeS*playback is sparta :p
21:12:50kugelit's certainly possible to do any smarts needed for playback in that callback
21:13:44CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
21:13:44*jhMikeS sure as hell will run away screaming from that one
21:14:01ukleinekwodz: in encode_page, why does key2 exist at all. You could just do key[i&0xf] instead of key2[i]
21:14:07kugelso you would prefer stopping playback every time?
21:15:25kugelwhat would also be possible is to unload all handles except those belonging to the currently playing track and then force-rebuffer
21:15:27jhMikeSkugel: probably not :) I'd recommend to address buffering's shortcomings more than playbacks
21:15:37kugelthat's perhaps easier from the playback pov
21:16:06jhMikeSthat is what it does for the playlist, but it won't guarantee freeing things at the beginning of the memory block
21:16:40kugelthat guarantee isn't needed
21:17:15wodzukleinek: this is straight translation asm->C from dissasm IIRC
21:17:22 Join robin0800 [0] (
21:17:25 Quit Keripo (Quit: Leaving.)
21:17:55jhMikeSas it is now, it expects to rebuffer from its current position
21:18:40kugelthat's ok. it'd just need to handle that the buffer boundaries possibly change on rebufferung
21:26:46 Quit sideral (Ping timeout: 276 seconds)
21:26:51 Quit ReimuHakurei (Read error: Connection reset by peer)
21:32:13 Quit robin0800 (Ping timeout: 255 seconds)
21:33:57*wodz booted rockbox binary on his rk27xx device
21:34:17wodzIts appealing to see rb main menu :-)
21:36:02ukleinekwodz: \o/
21:37:23 Join ChickeNES [0] (
21:38:30 Quit y4n (Quit: PÆNTS ØLF!)
21:41:09 Join robin0800 [0] (~robin0800@
21:44:38 Quit dunkaist (Quit: leaving)
21:54:47***Saving seen data "./dancer.seen"
21:57:34 Quit robin0800 (Read error: Connection reset by peer)
21:57:37 Join Thra11__ [0] (~thrall@
21:59:37 Join ReimuHakurei [0] (
22:00:41 Quit Thra11_ (Ping timeout: 264 seconds)
22:00:58 Join Thra11_ [0] (
22:04:51 Quit Thra11__ (Ping timeout: 260 seconds)
22:08:24 Join robin0800 [0] (~robin0800@
22:09:28CIA-14New commit by wodz (r30269): rk27xx sd driver fixes
22:12:23CIA-14r30269 build result: All green
22:13:15ukleinekwodz: there is a (copied) typo in your commit: s/tranfer/transfer/
22:13:38ukleinektwice even
22:14:37*ukleinek now looks after his rk27xx target
22:21:32CIA-14New commit by wodz (r30270): Fix typos in comments. Thanks ukleinek for pointing out.
22:24:26CIA-14r30270 build result: All green
22:37:09ukleinekwodz: you want some more? :-)
22:37:36wodzsure, go ahead
22:38:00 Quit Llorean (Quit: Leaving.)
22:38:12ukleinekwodz: you sent me a dump of your nand. I'm unable to decrypt it with the encode_page routine from rkusbtool. Isn't that the right procedure?
22:39:06ukleinekwodz: utils/rk27utils/README ~= s/whith/with/
22:42:10wodzukleinek: The routine is correct. This is raw dump and you need to decrypt in 512bytes blocks
22:42:37ukleinekwodz: hmm, that's what I did
22:42:50wodzin raw dump after 512bytes of data there are 16bytes of metadata which you need to skip
22:43:02*ukleinek nods
22:43:37ukleinekthe function is just copyandpasted
22:44:06wodzlooks correct
22:44:09bertrikhow did you figure out the key?
22:44:41ukleinekbertrik: a guy disassembled to ROM
22:45:04bertrikcool, I guess they stored the key along with the code, then? :)
22:45:42bertrikIIRC, that's how it was done with the sansa c200 and e200 bootloaders
22:46:30*ukleinek wonders if it's PEBKAC that the things that work for wody don't work for him
22:46:51 Quit Thra11_ (Ping timeout: 260 seconds)
22:47:19ukleinekargh, broken keyboard layout
22:48:18wodzbertrik: the key and scrambling routine sits in rom of the SoC - it happen to be quite easy to dump :-)
22:48:40ukleinekbertrik: at least for wodz, not for me :(
22:49:26 Join liar [0] (
22:49:57bertrikWe've seen quite a few players now that use encryption but in a totally inadequate manner.
22:50:50bertrikLike the c200 and e200 bootloaders that contain both the algorithm and the key in plain text, and the fuze+ that uses encryption with an all-zero key.
22:51:33wodznanos seem to be exception
22:52:08bertrikI don't really know how TheSeven did the nano 2g
22:52:36Bagderpatience and persistance were involved =)
22:53:19 Join stripwax [0] (
22:57:12 Join mystica555_ [0] (
22:59:19*TheSeven extends his ears
22:59:24 Quit robin0800 (Read error: Connection timed out)
22:59:29*ukleinek retried on i386, but without luck
23:00:16 Join robin0800 [0] (~robin0800@
23:00:51 Quit mystica555 (Ping timeout: 260 seconds)
23:01:23TheSeventhe nano2g was all aes (with an unknown hardware key, used by a hardware crypto accelerator, can't be revealed), and hmac-sha1 (similar hardware accelerator)
23:01:37 Join powell14ski_ [0] (
23:01:49ukleinekwodz: could you provide me with a dump of a rk27load session, optimally with the input files?
23:01:50TheSevenbut as this is all symmetric, it allows us to sign our own firmware if we get into the device by means of a userspace exploit
23:02:09TheSevenso we just needed a userspace exploit once to boostrap our own signing process :)
23:02:47TheSevenon the newer generations they used public key cryptography, but luckily hat a bootrom-level vulnerability that allows us to boot through an exploit
23:03:12 Join Thra11_ [0] (~thrall@
23:03:14TheSeventhe added complexity of proper firmware signing actually made it easier for us to get our code running, we don't even need to do on-device signing any more :)
23:06:04wodzukleinek: I'll try to record rk27load session for you tomorrow
23:06:16ukleinekwodz: great
23:08:08wodz <- which ld script do I need to tweak?
23:09:50wodzmy current
23:11:17 Quit stripwax (Quit:
23:14:08 Quit TheSeven (Remote host closed the connection)
23:21:38 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
23:23:53TheSevenwodz: you need to fix apps/plugins/
23:24:46 Join Thra11__ [0] (~thrall@
23:24:56wodzTheSeven: thanks for the hint
23:27:30 Quit Thra11_ (Ping timeout: 258 seconds)
23:31:23 Quit Thra11__ (Ping timeout: 240 seconds)
23:34:01 Part domonoky1
23:34:07 Join ptrkmj [0] (~chatzilla@
23:37:23 Join Thra11__ [0] (~thrall@
23:37:24 Join Strife89 [0] (
23:41:16 Quit n1s (Remote host closed the connection)
23:46:19 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 3.6.18/20110614230723])
23:48:55 Nick ruler5 is now known as ruler501 (~ruler_501@
23:54:48***Saving seen data "./dancer.seen"

Previous day | Next day