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 2011-10-11

00:00:10 Join Horschti [0] (
00:00:11 Quit Horschti (Changing host)
00:00:11 Join Horschti [0] (~Horscht@xbmc/user/horscht)
00:01:35 Part Zagor
00:02:42 Quit semitones (Remote host closed the connection)
00:02:52 Join semitones [0] (
00:07:09 Quit domonoky (Read error: Connection reset by peer)
00:13:39 Quit robin0800 (Quit: Leaving)
00:14:10 Join robin0800 [0] (
00:15:09 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
00:22:15 Quit Jerom (Quit: Leaving.)
00:24:32 Join Scromple [0] (
00:31:18 Quit AlexP (Read error: Connection reset by peer)
00:37:25 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
00:48:59 Quit zchs (Ping timeout: 252 seconds)
00:49:28 Join zchs [0] (
00:51:07 Quit keyb_gr (Ping timeout: 260 seconds)
00:51:46 Quit simonlnu (Quit: bbl)
00:53:09 Quit TheSeven (Disconnected by services)
00:53:20 Join [7] [0] (~TheSeven@rockbox/developer/TheSeven)
01:00:56 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
01:06:58 Quit mudd1 (Ping timeout: 258 seconds)
01:11:42 Join mudd1 [0] (
01:13:36 Quit stripwax (Quit:
01:15:51 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
01:18:38 Quit ender` (Quit: Female software engineers become sexually irresistible at the age of consent, and remain that way until about thirty minutes after clinical death. Longer if it's a warm day. -- Scott Adams)
01:20:31 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
01:26:23 Quit hilbert (Quit: Textual IRC Client:
01:32:12 Quit robin0800 (Quit: Leaving)
01:43:03JdGordonZagor: I think swapping database and files is a bad idea because current users will get confused, and it gives no real advantage. (it seems to be change for the sake of change)
01:43:38JdGordonand "Rockbox info" and the t&d screen are the only submenus in system which users would want to go in frequently which is why I put them at the top of that list.
01:50:11 Quit froggyman_ (Ping timeout: 240 seconds)
01:51:03 Quit mudd1 (Ping timeout: 258 seconds)
01:57:12***Saving seen data "./dancer.seen"
01:59:34 Quit MethoS- (Read error: Connection reset by peer)
02:05:20 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
02:15:46 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
02:27:38 Quit Keripo (Ping timeout: 256 seconds)
02:44:49 Join esalazioneHg [0] (
02:46:59esalazioneHghi guys, i've found from what version rockbox stopped working on my iPod nano 2G. it's the 3.7.1. how can i put a newer version?
02:48:58esalazioneHg(more infos: from 3.8 it hangs after some seconds, in 3.9.1 it plays songs for ~8 secs, then hangs with a noise and sometimes saying "undefined instruction at ...")
02:49:07 Quit [fred] (Ping timeout: 260 seconds)
02:49:07 Join [Saint] [0] (~Saint]@
02:50:01esalazioneHgi'd like to help you find out what's the problem so you can fix it ;)
02:50:48 Join [fred] [0] (
02:57:07JdGordon[Saint]: the nano2g is your main dap no?
02:59:49esalazioneHg(just discovered the 3.9 works a little better then the 3.9.1...but still freeze after a little...)
03:01:46[Saint]JdGordon: kinda...why?
03:02:04[Saint]I use my Color mostly
03:03:30[Saint]For n2g stuff, I think user890104 is your man. But, ask away.
03:03:48 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
03:04:04[Saint]JdGordon: ....?
03:04:42esalazioneHg[Saint] on my nano 2G the last working rockbox version is the 3.7.1
03:05:38[Saint]That's....weird. my version isn't current, but its a LOT newer than that.
03:05:38esalazioneHg[Saint]all the newer version hangs after a while (from 8 secs to ~30 secs), sometimes saying "undefined instruction at ..."
03:05:43 Join ChickeNES [0] (~ChickeNES@
03:05:49[Saint]What are the symptoms?
03:06:12[Saint]Oh, sorry. Hmmmm.
03:06:56[Saint]Do you use the Rockbox bootloader? Or a freemyipod launcher?
03:07:09esalazioneHgi read on the forum that another user had this problem, but no solution was given and the theard is a bit old
03:07:30[Saint]Like EMCORE, EMBIOS or iLoader.
03:07:47[Saint]Or do you just use the Rockbox bpotloader?
03:07:48esalazioneHgi'm using the rockbox one, i installed it when the nano 2G was still unstable :P
03:08:43esalazioneHgi tested the apple's firmware and it's fully working
03:08:46[Saint]Google freemyipod, or head over to #freemyipod-support and install EMCORE
03:09:20esalazioneHgwhen the playback hangs on rockbox sometimes i hear a noise
03:10:29[Saint]Yeah, right...I've no idea what's going on. But I can tell you virtually no one involved in the nano2g's development is using the Rockbox bootloader.
03:10:58esalazioneHgok, i'm looking for emCORE
03:10:58[Saint]Even if you weren't having problems, I'd still advise you use emCORE
03:11:43esalazioneHguhm, ok, i'll try it right now...maybe it's better to add this advice on the nano2g's manual
03:13:02user890104esalazioneHg: here's a similar case to yours:
03:13:11[Saint]No. Its not anything to do with the Rockbox project.
03:13:27esalazioneHguhm...."Your device is fully supported by emCORE, but there are some bugs left, so it's not yet recommended for end users. If you would like to dualboot Rockbox and Apple's firmware, it would be safer to use Rockbox's bootloader or iLoader"
03:13:37[Saint]user890104: can you guide him through noteboot please?
03:13:41esalazioneHgthey suggest to use rockbox's bootloader :D
03:13:53[Saint]Old docs, don't worry.
03:15:30user890104esalazioneHg: you basicly need to download then put it in your ipod's Notes folder, then eject it
03:16:07 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
03:16:22user890104if you have any problems with the bootloader then, please ask in #freemyipod-support
03:16:54esalazioneHg[user890104]ok, i'm trying, thanks
03:17:39[Saint]Thanks user890104 , I'm on my phone, trying to do grocery shopping ;)
03:17:44[Saint]I appreciate it.
03:17:53user890104my ipod is with the most recent (development) emcore build and rockbox 3.9.1, and it works fine
03:18:08user890104[Saint]: you're welcome :)
03:25:38user890104but i'm going to sleep now, it's 04:25 here
03:26:31esalazioneHguhm, we're near :P i'm trying with the ipodpatcher because the bootnote didn't work
03:26:46esalazioneHgthanks for your help, i hope it'll work ;)
03:29:03[Saint]Hmmmmm...that's no good.
03:29:23[Saint]Noteboot should *always* work.
03:30:40esalazioneHgi simply have to put the noteboot in the Notes folder then eject and reboot with the HOLD on (orig. firm.) or off (rockbox)?
03:32:21[Saint]Safely eject, with hold off.
03:34:51 Join Keripo [0] (~Keripo@
03:38:59esalazioneHg3.9.1 still hangs
03:39:55esalazioneHgis there a way to see if i'm using emCORE or the rockbox bootloader? i think it's still using the rb's one
03:41:28esalazioneHg(anyway, ipodpatcher fails with "[ERR] *** ipodpatcher does not support the 2nd Generation Nano! ***")
03:41:54 Quit Keripo (Ping timeout: 260 seconds)
03:42:18esalazioneHg(i remember i had the same error the first time i used it and someone give me a new version of the ipodpatcher, in this IRC channel)
03:48:22esalazioneHgipodpatcher v4.0 works, ipodpatcher 5.0 no
03:57:13***Saving seen data "./dancer.seen"
03:58:13[Saint]If you don't get a boot menu, you messed up the install.
03:58:50[Saint]Its very obvious when emCORE is installed.
04:01:13[Saint]The emCORE installer should have launched and guided you through the installation.
04:01:15 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
04:03:20esalazioneHgusing the old ipodpatcher (v 4.0) i installed emCORE, but IT hangs!!! o.0
04:04:06esalazioneHgi can boot the original firm., but if i scroll the wheel in emCORE, it freeze...and if i lunch rockbox i've got the same error as before
04:04:22esalazioneHgthe orig.firm is working well
04:10:56esalazioneHgi really have to go' 4 am here
04:11:13esalazioneHgi'll be there tomorrow again ;) thanks for your help
04:14:07 Join [Saint_] [0] (~Saint]@
04:15:59 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
04:17:16 Part esalazioneHg
04:18:04 Quit [Saint] (Ping timeout: 258 seconds)
04:23:32 Quit amiconn (Disconnected by services)
04:23:33 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:23:48 Quit pixelma (Read error: Operation timed out)
04:24:00 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:24:33 Join pixelma [0] (quassel@rockbox/staff/pixelma)
04:33:36 Quit fiesh (Ping timeout: 258 seconds)
04:34:34 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
04:36:18 Join fiesh [0] (
04:37:26 Quit [7] (Disconnected by services)
04:37:36 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:44:14 Join FritoDude [0] (
04:45:02FritoDudeI'd like the Clip V2's recording screen to show more clearly whether I'm recording or not.
04:45:14FritoDudeIt has a small icon on the top line that is a square, circle, two triangles, which show that is it stopped, recording, and paused.
04:45:29FritoDudeMy guess of the simplest implementation of this is to invert the screen (black becomes white, white becomes black) while recording.
04:45:47FritoDudeChanging the draw mode to "SOLID+INVERSEID" for the various viewports would seem to do the trick. The idea would be to find where the recording state icon is changed, & insert some code there to change the drawmode.
04:46:48FritoDudeI've look through the code and would like some clues about where the code is that actually changes the recording status is.
04:46:53FritoDudeAny ideas?
04:49:18[Saint_] can do it all with a theme (.sbs specifically, most of the rec screen is hard-coded) and display whatever icon(s) you like.
04:50:34 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
04:51:02 Quit [Saint_] (Quit: AndChat - Android IRC Client -
04:51:02Llorean[Saint_]: He's talking about inverting the display when recording is happening, not really about icons.
04:52:18 Join [Saint] [0] (~Saint]@
04:52:21FritoDudeIf I could use much larger icons, that might do the trick. Right now the recording status icons are in the top line and are pretty tiny. Can I change the size with a theme? I tried a bunch of the canned themes without luck.
04:53:05[Saint]Yes, you can.
04:53:09FritoDudeOf course if I use a theme then I won't get to take advantage of the two hours I've already spent understanding the code! ;-)
04:53:31[Saint]Check out the manual re: theme syntax
04:54:04FritoDudeI'll have a look now.
04:55:05JdGordonFritoDude: the recording screen is the only one which doesnt provide full skin support yet, so if you want a custom coded fix have a look at apps/recorder/recorder.c
04:55:05 Join froggyman [0] (~seth@
04:55:07 Quit froggyman (Changing host)
04:55:07 Join froggyman [0] (~seth@unaffiliated/froggyman)
04:57:48FritoDudeI've worked on understanding recording.c and haven't yet figured out where the recording status icon is changed.
04:58:06FritoDudeHere's my understanding so far: There is a giant while loop in the function apps/recording/recording.c::recording_screen() that seems to intercept button presses.
04:58:06FritoDudeThe function apps/recording/recording.c::recording_screen() seems to set up the gui. It contains a big while loop that runs the whole time the user is on the recording screen. It has a switch on which button the user presses. If, for example, the user presses the pause button ("ACTION_REC_PAUSE"), it invokes the function rec_command(). rec_command() invokes pcm_record.c::audio_pause_recording().
04:58:06FritoDudeaudio_pause_recording() puts a "PCMREC_PAUSE" event on the the pcmrec_queue. pcmrec_thread() handles the event by calling pcmrec_pause(). pcmrec_pause just seems to set static variable "is_paused" to true.
04:58:29JdGordonoh, you want it more visible outside of the rec screen?
04:58:59FritoDudeI just want recording status to be more visible when I'm on the rec screen.
05:00:32FritoDudeA perfect solution might be a "basic" recording screen that just has the elapsed time and a huge recording status icon. I don't need the gain or peak meter. I'm just using it to record notes to myself when I walk or bicycle. But, I thought just inverting the screen might be a lot easier to implement.
05:03:38[Saint]An .sbs is the way to do this.
05:03:41 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
05:03:45[Saint]Not hacking in the core.
05:05:37FritoDudeI'd sure like it if I could do this in the .sbs file. I'm in the middle of looking at the theme syntax in the clip v2 manual. I'm trying to figure out how I tell it to do something to the recording screen but not to the others.
05:06:31JdGordonlike I said, the rec screen doesnt handle the sbs very well.... the ui viewport will be used to draw all the current display which doesnt give you much room to do anything else
05:07:12 Quit ChickeNES (Quit: Computer has gone to sleep.)
05:10:14FritoDudeJdGordon, are you saying that if I use a line like "ui viewport: X,Y,[width],[height],[font]" in the sbs that it'll be ignored by the recording screen?
05:11:32 Join jhMikeS [0] (
05:11:32 Quit jhMikeS (Changing host)
05:11:32 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
05:11:34JdGordonthe rec screen will try to fit into whatever you give it
05:11:47JdGordonbut thats the whole text area and list (though both will shrink if needed)
05:15:55 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
05:16:32FritoDudeI'm still kind of hazy about the theme approach, but what if I gave the top half of the screen to the ui viewport? I think you are saying that then I'd have to scroll around the ui viewport to see all the info. I'd be ok with that because I don't use that information much. Then my issues would be to figure out how to display the recoding status more prominently in the bottom half of the screen,
05:16:32FritoDude& also to worry about whether changing the sbs is going to mess up some of the other screens.
05:16:46 Join Rob2223 [0] (
05:17:05JdGordonuse %?cs to check if you are in the recording screen and set a differen ui viewport accordingly
05:18:07 Quit jhMikeS (Ping timeout: 248 seconds)
05:20:29 Quit Rob2222 (Ping timeout: 256 seconds)
05:21:25FritoDudeThanks JdGordon, I've just found that %?cs syntax in the manual.
05:24:42FritoDudeNow my issue will be how to display the recording status. Is Playback Mode the way to do it? I see values for "Stop", "Recording", & "Recording Paused" for the %?mp tag.
05:25:32 Join froggyman_ [0] (~seth@
05:26:44JdGordonsounds about right
05:26:46 Quit froggyman_ (Read error: Connection reset by peer)
05:26:50 Quit froggyman (Quit: Ex-Chat)
05:27:05 Join froggyman [0] (~seth@
05:27:05 Quit froggyman (Changing host)
05:27:05 Join froggyman [0] (~seth@unaffiliated/froggyman)
05:29:22 Join froggyman_ [0] (~seth@
05:29:34 Quit froggyman_ (Read error: Connection reset by peer)
05:30:31FritoDudeSo, the idea would be to use %cs to shrink the ui viewport on the recording screen. Then I'd add another viewport to that screen, and use the %mp tag to conditionally display some bmp images in that viewport. Does that sound right?
05:31:32 Join [Saint_] [0] (~Saint]@
05:32:12[Saint_]FritoDude: yep. You've got it.
05:32:30[Saint_]Not many people grasp the skin syntax so quickly.
05:32:48[Saint_]Perhaps I've found my padawan.... ;)
05:33:42 Quit [Saint] (Ping timeout: 245 seconds)
05:34:56 Join [Saint] [0] (~Saint]@
05:35:18FritoDudeWe'll see if I can actually implement the darn thing in addition to pretending like I understand it. ;-)
05:37:30 Quit [Saint_] (Ping timeout: 255 seconds)
05:38:37FritoDudeThanks for the help everyone.
05:39:18 Quit [Saint] (Ping timeout: 255 seconds)
05:40:13 Quit FritoDude ()
05:57:16***Saving seen data "./dancer.seen"
05:59:18 Part fiesh
06:04:59 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
06:16:15 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
06:44:07 Join FritoDude [0] (
07:02:51 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
07:05:48 Quit fatline (Quit: This computer has gone to sleep)
07:13:05 Quit gevaerts (*.net *.split)
07:13:06 Quit tmzt_ (*.net *.split)
07:13:06 Quit feisar- (*.net *.split)
07:13:06 Quit avacore (*.net *.split)
07:14:08 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
07:14:08 Join tmzt_ [0] (
07:14:08 Join feisar- [0] (
07:14:08 Join avacore [0] (
07:16:48 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
07:17:27 Join Zambezi_ [0] (Zulu@
07:17:56 Quit niekie (Quit: No Ping reply in 180 seconds.)
07:18:05 Join niekie [0] (~niek@CAcert/Assurer/niekie)
07:18:07 Quit Zambezi (Ping timeout: 248 seconds)
07:22:36 Quit niekie (Excess Flood)
07:23:48 Join niekie [0] (~niek@CAcert/Assurer/niekie)
07:37:34 Join [Saint] [0] (~Saint]@
07:43:43 Quit Xerion (Ping timeout: 248 seconds)
07:57:19***Saving seen data "./dancer.seen"
07:59:43 Quit [Saint] (Read error: Connection reset by peer)
08:03:42 Join Keripo [0] (
08:06:14 Join Buschel [0] (
08:07:54 Quit Keripo (Ping timeout: 252 seconds)
08:20:26 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20110928134238])
08:27:44 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
08:47:50 Join Keripo [0] (~Keripo@
08:47:53 Quit FritoDude ()
08:52:14 Quit Keripo (Ping timeout: 260 seconds)
08:54:28 Join ender` [0] (
08:55:51 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
09:06:44 Quit ender` (Read error: Connection reset by peer)
09:06:50 Join ender` [0] (
09:07:10 Join hilbert [0] (
09:10:12 Join LinusN [0] (
09:13:42 Quit Scromple (Quit: Leaving)
09:18:33 Join einhirn [0] (
09:26:22 Quit Zambezi_ (Changing host)
09:26:22 Join Zambezi_ [0] (Zulu@unaffiliated/zambezi)
09:26:36 Nick Zambezi_ is now known as Zambezi (Zulu@unaffiliated/zambezi)
09:38:16ZagorLinusN: I'm going with "percent of full screen width", even within smaller viewports. otherwise everything gets a lot more complex.
09:38:24 Join Keripo1 [0] (
09:39:16LinusNZagor: back to IE HTML hell then ;-)
09:40:08Zagorluckily we don't do html :)
09:41:26Zagoror rather, we don't have legacy data assuming relative viewport dimensions (only positions)
09:41:54LinusNi'm sure you thought this through.... :-)
09:42:35 Quit Keripo1 (Ping timeout: 240 seconds)
09:43:00 Join Keripo [0] (
09:47:42 Quit Keripo (Ping timeout: 252 seconds)
09:49:41 Join mortalis [0] (
09:57:23***Saving seen data "./dancer.seen"
09:57:32 Join Keripo [0] (
10:01:37JdGordonZagor: oh, you're fiddling with the lcd drivers?
10:01:43JdGordoncare to take over a patch i was working on?
10:02:27JdGordonremove a bunch of cruft, make everything pixel based instead of char based
10:02:57Zagorwhat happens to charcell?
10:03:01JdGordonjdgordon/rockbox/tree/lcd_puts_rework"> is the git branch, i dont tinhk i got around to putting it on fs
10:03:12JdGordoni was hoping it owuld magically work
10:04:31Zagorwhat parts are char based today?
10:05:36ZagorI've got my plate pretty full with this percent test right now, but it sounds like something we probably want.
10:06:25JdGordon is the diff
10:07:15JdGordonbut yes, i didnt deal with charcell at all
10:07:33 Join mudd1 [0] (
10:07:59 Quit mortalis (Quit: CGI:IRC)
10:08:59Zagoryeah I need to deal with that before my percent test can work. thanks.
10:12:22 Join dfkt [0] (~dfkt@unaffiliated/dfkt)
10:51:05 Quit einhirn (Ping timeout: 255 seconds)
10:51:17 Join ntkm [0] (
10:55:58 Quit ntkm (Client Quit)
10:57:43 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
10:59:15kugelpZagor: we agreed on dropping archos/charcell if it gets a too heavy maintenance burden didn't we?
11:00:03Zagorsomething like that
11:00:31Zagorbut I don't think this qualifies as too heavy.
11:01:28kugelpI guess its up to you as you do the work :-)
11:12:15 Quit sinthetek (Read error: Connection reset by peer)
11:13:48Zagoraugh, I'm colliding with the remote scrolling code
11:14:07Zagorbah, screw that for now
11:14:11 Join sinthetek [0] (~sinthetek@unaffiliated/sinthetek)
11:18:53gevaertsZagor: it's a rather remote issue anyway :)
11:20:10Zagorwhee, it compiles
11:24:20 Join wodz [0] (
11:28:23 Join GodEater_ [0] (93722cc8@rockbox/staff/GodEater)
11:33:47 Join Horscht [0] (~Horscht@xbmc/user/horscht)
11:35:50wodzkugelp, pamaury: any thoughts about crt0.S factoring?
11:36:11 Quit Horschti (Ping timeout: 240 seconds)
11:36:18pamaurywodz: I wrote some thoughts during the week-end I think
11:36:22 Quit mystica555_ (Remote host closed the connection)
11:39:27pamauryI think to refactoring I would write a generic "crt0-lib.S" file which contains the common parts. And then, either you compile it and call it from crt0.S, either you write it using as macros, include it in crt0.S and use it
11:40:05 Part LinusN
11:41:06 Join bluebrother [0] (
11:41:06 Quit bluebrother (Changing host)
11:41:06 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
11:41:16wodzpamaury: regarding imx233 and calling C code in crt0.S wouldn't it be better to rewrite this into asm? You are calling functions from mmu-arm.S anyway
11:41:20 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
11:41:53wodzC code only serves as wrapper as I understand
11:43:19pamauryit's just a bit more readable, I could rewrite it in assembly code but that's not what will prevent crt0 refactoring :)
11:44:14 Quit fs-bluebot (Ping timeout: 248 seconds)
11:44:16 Quit bluebrother^ (Ping timeout: 258 seconds)
11:45:03wodzbut it will converge imx233 crt to one of the standard layout :-)
11:45:22 Join fs-bluebot [0] (
11:45:44pamauryI don't see the point, it's already quite standard and not all platforms share the same memory layout, there will always be a target specific part
11:46:13wodzI am talking about the need for additional SP setup in your case
11:46:49pamaurythat's really a detail
11:47:57pamaurythe two differences between imx233 crt0 and standard are: 1) self-moving code 2) mmu setup
11:55:13pamaurywodz: what are your thoughts for crt0 refactoring ?
11:57:26***Saving seen data "./dancer.seen"
12:09:38wodzpamaury: I am biased to main crt0.S and from there calling target specific parts
12:11:50pamaurywodz: here is an example of a crt0-as-lib refactoring (just a proposal):
12:14:26 Join mortalis [0] (
12:16:53wodzpamaury: for me this macros look really cryptic. But I must admit this is very compact solution
12:18:15pamaurythe problem of the main crt0 with callbacks is that you will put callbacks everywhere *and* you will have to standardize all the names w.r.t linker *and
12:18:37pamauryyou will have to have a way to choose between branch and ldr for vectors
12:19:31mortaliswodz: sound is working now, but paly speed is little bit accelerated
12:19:51pamauryand I don't see how you will implement self copying code (except if you implement it for all targets)
12:19:57wodzmortalis: that means you misconfigured I2S clock
12:22:58wodzpamaury: what you mean by everywhere? You probably need one call to setup target specific parts - after vectors and before zeroing bss and stack setups
12:35:45 Join MethoS- [0] (~clemens@
12:49:58ukleinekmortalis: I looked at the usb dump you posted on the forum
12:50:29ukleinekmortalis: for me it looks as if the first four blocks of nand repeat several times. Is that possible?
12:51:23ukleinekmortalis: and btw my player behaves as yours (though I didn't try to shortcut NAND pins)
12:52:36 Quit kadoban (Ping timeout: 244 seconds)
12:58:12 Join TomColler [0] (
12:59:10mortalisukleineik: yes, looks like they are repeating
13:01:43mortalisukleineik: I was mistaken when sayed that in mask ROM mode rk27load works fine. It loads only stage1 and stage2 but fails to upload usercode.
13:05:40wodzmortalis: so how do you test rb code?
13:06:11mortalisRK27 upgrade image in virtualbox
13:07:16wodzmortalis: but where you inject code?
13:07:46wodzukleinek: on my device first 5 blocks where identical
13:08:25wodzukleinek: I guess it is some (poor) sort of protection considering raw nand chip is not very reliable device
13:09:11mortalisi used rkwpatch from alemaxx to create BASE.RKW, then I replaced BASE.RKW in official firmware image.
13:13:16wodzmortalis: Is stage2 executed at all? (you can craft stage2 image which for example blink backlight to see)
13:14:24wodzsince windows DFU driver for all rk27xx devices is the same I wouldn't expect differences in rom mode
13:17:07wodzmortalis: and about volume - I think this may be harder than I first thought - skin engine needs current volume to show this in WPS, settings code expects this info unconditionally AFAIK and there may be other parts I am not aware of.
13:18:10mortaliswodz: ok, I'll test stage2 may be today or tomorrow
13:19:14wodzmortalis: What variant of rk27xx is your player based on?
13:19:23mortaliswodz: 2706b
13:20:49wodzmortalis: I suspect you set wrong scramble parameter for stage2 and it isn't executed at all.
13:28:06mortaliswodz: I also sniffed usb traffic in rom mode. It uploads stages 1,2,4 from rk27boot.bin. Fourth stage separated to blocks of 4096 bytes. After that it acts like in 'other' not rom DFU mode.
13:30:00wodzmortalis: cool! what version of rk27boot.bin you use? There are a few slightly different floating around
13:31:34mortaliswodz: your can extract it from this image
13:32:16wodzah ok you used the one dedicated to your player
13:32:38wodzI'll look at it when find some time (unlikely heh)
13:40:42LalufuIs it possible that disabling interrupts for too long on a PP based board makes the serial driver/chipset wonky?
13:42:40Torneshouldn't do, but you might drop data on the floor
13:43:27wodzTorne: as arm guy could you comment discussion about factoring crt0.S?
13:43:46kugelwodz, pamaury: think I favor crt0-lib.S but in the end both approaches work
13:43:55Lalufukugel recently (this weekend) committed a patch that disabled interrupts during buflib moves.
13:44:24Lalufuthe iap driver I'm using dynamically allocates it's RX/TX buffers from buflib when it receives the first packet of what might be an IAP message
13:44:56Lalufuif the player is playing that usually causes a buflib move, which disables interrupts for quite some time
13:45:12kugelthe interrupts are disabled only shortly
13:45:23kugel1 callback+1 memmove
13:45:50LalufuI noticed that the player stops/resumes playing if IAP allocates from buflib.
13:45:59LalufuI assume that's because some buffer needs to be moved?
13:46:32kugelno, because the audio buffer needs to be shrinked to make room
13:47:04Lalufuah, ok.
13:47:24kugelyou cannot reasonably play audio during that action
13:48:22kugelwell, you probably can with more clever code :)
13:49:03LalufuYeah. The effect I see is that I somehow lose the serial link to some devices after the allocation. I can see packets on the serial interface before the allocation, but none after, until the player is reconnected.
13:49:48LalufuI'll have to see if reverting the "no interrupts" patch fixes that, if so I have to come up with my own locking to handle buffer moves.
13:54:25 Quit mortalis (Quit: CGI:IRC (EOF))
13:56:08*ukleinek will have a deeper look at his player's nand now
13:56:33kugelLalufu: do you call core_get_data() properly?
13:57:16 Join T44 [0] (
13:57:26Lalufucould you elaborate on how to call it properly?
13:57:28 Quit Keripo (Quit: Leaving.)
13:57:30***Saving seen data "./dancer.seen"
13:59:08LalufuI call it once, right after core_alloc_ex().
13:59:16Lalufuand adjust my cached pointers in a move callback
13:59:16JdGordonLalufu: how big are the tx/rx buffers?
13:59:42Lalufu64k for RX, 512 for TX, plus a handful of bytes for headers
14:00:04Lalufuwe have to use 64k for RX, since the device could send a data block that long
14:00:12JdGordonis the alloc call disalbing interrupts for too long?
14:00:22Lalufuwe could use 64k for TX, too, but the use of that is optional for the things IAP does so far
14:00:44LalufuI don't know the actual cause of the problem, I have to investigate that further.
14:00:47JdGordonon the first packet you could use the plugin buffer and then when it is safe you can alloc for the next packets
14:01:18 Quit Topy44 (Ping timeout: 248 seconds)
14:01:30Lalufuwhat plugin buffer?
14:02:10kugelLalufu: ah okay, sounds alright. remember that you still must not pass it to functions that yield
14:02:18JdGordonplugins are loaded into a static 512K (or smaller on some targets) buffer, they almost never use the full size so you can use plugin_get_buffer() to grab the remaining buffer space
14:02:43Lalufuas far as I can tell I don't do that, the buffer is used by iap.c only
14:03:03Lalufuis iap.c a plugin? I don't think it is.
14:04:06JdGordonthat call is for core to use to get that otherwise wasted buffer
14:04:15JdGordongood for very temp storage
14:04:17Lalufuaaah, I see
14:11:47kugelLalufu: though, for new code I recommend calling core_get_data()
14:12:44JdGordonnot if it means playback would stop
14:12:54JdGordonunless it is a permenant buffer
14:13:08LalufuThat's a bit tricky, as I call other functions from my code, and those might yield, I don't know.
14:13:08JdGordonor at least not completly temporary
14:13:21Lalufuso i'd have to call core_get_data() after every call to an outside function
14:13:30JdGordonLalufu: if you dont know for sure, lock your handle
14:13:53kugelLalufu: yea, is that a problem?
14:14:10JdGordonwhich is something buflib should be providing :/
14:14:13kugelthe lock is only necessary if you pass the buffer to functions
14:15:34 Join mortalis [0] (
14:17:48LalufuCurrently I only access the buffer via a set of pointers into the buffer. Those are updated in the move callback.
14:19:15JdGordonare they void* or char*'s?
14:19:29JdGordonanything else and you need to be careful doing pointer arithmatic :)
14:19:54Lalufuall char*
14:28:00Tornewodz: er, i dunno really, i haven't looked at anything other than crt0-pp.S in great detail
14:28:40Tornewodz: for pycorn i wrote most of the bootstrap code in C and only used asm for the bare minimum of stack setup, but that's not much less than what our crt0's do anyway, i expect?
14:28:45Torneat least, except the imx31
14:32:36 Quit hilbert (Quit: Textual IRC Client:
14:34:31JdGordonyou know what the solution to the constant UI arguments is?
14:34:55JdGordonremove the ui from the core codebase and have it completly loadable instead :)
14:35:39ZagorI know, what if we make something like... librockbox. that'd be neat!
14:35:49JdGordonsure would!
14:38:43Lalufuand have the capability for multiple, concurrent control schemes?
14:39:24 Join WalkGood [0] (~4@unaffiliated/walkgood)
14:41:32JdGordoni've got "handle SIGUSR1 nostop pass" in my .gdbinit file, and it seems to work mostly, but when starting rockbox it still stops after priting "Program received signal SIGUSR1, User defined signal 1." the last time, anyelse getting that also?
14:41:37JdGordonor know how to stop it
14:42:36kugelJdGordon: I think that's a gdb bug. I have this also
14:43:06JdGordonits really annoying!
14:49:50ukleinekit seems I don't have repetitions
14:57:05 Join mystica555_ [0] (
14:59:09ukleinekmortalis: if you're interested I have a Python tool to dump the nand flash
15:01:07 Join freddyb [0] (
15:04:25 Quit antil33t (Read error: Connection reset by peer)
15:05:50mortalisukleinek: nice. Could you upload it?
15:06:49 Join antil33t [0] (
15:09:39Zagoronce we go all-pixels, we should remove all the xy-ing in the lcd function names
15:10:48 Quit mortalis (Quit: CGI:IRC)
15:14:42 Quit wodz (Quit: Leaving)
15:17:51 Join hilbert [0] (
15:22:00 Quit mystica555_ (Remote host closed the connection)
15:23:56 Join mystica555_ [0] (
15:29:29 Quit echelon (Remote host closed the connection)
15:29:59 Join Thra11 [0] (~thrall@
15:32:09 Quit freddyb (Quit: Leaving)
15:32:21 Join echelon [0] (~echelon@gateway/tor-sasl/harel)
15:34:39Zagorcan anyone remind me of the purpose of having offsets in addition to position?
15:37:28JdGordoni think it was a pixel offset to start drawing the first char from
15:38:22Zagoryeah but what does it do that can't be done by increasing x the same amount?
15:39:26JdGordonsee my patch, I think I solved this
15:39:33JdGordonx was char based
15:40:37Zagorin that case your patch should be able to go much further, basically eliminating all the _offset calls
15:41:24ZagorI am looking at your patch, since it is required for my percent experiment
15:42:26Zagorthe player sim boots up fine with it btw, but crashes on navigation
15:43:22JdGordonplease close fs#12237 when you're done with this :)
15:43:23fs-bluebot WPS/FMS string space bug introduced by r30302 (bugs, unconfirmed)
15:46:57 Quit WalkGood (Ping timeout: 252 seconds)
15:47:09Zagorwe have 14 different lcd_puts* calls!
15:50:42JdGordoncalls or functions?
15:51:08JdGordonmy diff removes a few :)
15:51:21JdGordonand yes, that number is why i started working on it
15:51:39JdGordonisnt it nice how most are wrappers?
15:51:48JdGordonwrapper->wrapper->wrapper->function :/
15:51:55Zagorit makes me reach for my machete
15:54:26ZagorI want one: lcd_puts(x, y, style, scrolled, fmt, ...)
15:54:27kugelZagor: offset is for skipping first N chars of the string
15:54:49Zagorno it's not
15:54:58kugelit's used for some strange feature where you move the whole list to the left, but also for scrolling IIRC
15:55:48Zagorwe even have xyoffset functions
15:56:29kugelyes, I added that for pixel scrolling because everything is line based
15:56:42kugelI added the yoffset part
15:57:09 Join WalkGood [0] (
15:57:09 Quit WalkGood (Changing host)
15:57:09 Join WalkGood [0] (~4@unaffiliated/walkgood)
15:57:10kugelI would have used the pixel function if they supported scrolling
15:57:31***Saving seen data "./dancer.seen"
15:57:37kugelZagor: what is the xoffset for, then? I
15:59:49Zagorah, you mean first N pixels of the string
16:00:06kugelcould be pixels yes
16:01:26Zagorright, this is needed for smooth scrolling
16:01:53JdGordonyoffset is different again isnt it? doesnt it cut the first yoffset pixels from the top of the line?
16:02:03Zagorso lcd_puts(x, y, xofs, yofs, style, scrolled, fmt, ...) then :-)
16:02:27ZagorJdGordon: how is that not the same as xoffset?
16:02:32Zagoronly on y axis
16:02:56JdGordonyeah, that was more in response (and badly worded) to kugel
16:03:03JdGordonnever mind
16:04:02Zagorxcrop ycrop might be more obvious
16:04:46JdGordonand I think lcd_puts(x,y,str), lcd_putsf(x,y,fmt,...) and lcd_puts_ex(x,y,xcrop,ycrop,fmt,...) are the minimum you'd want to be nice api-wise
16:04:59JdGordoneven if the first two are #defines
16:05:28Zagorwhy puts and putsf ?
16:06:09JdGordonmost of the time (i.e everything in firmware/ ) uses them
16:06:19JdGordonand im lazy
16:06:45JdGordongoing through 5 layers of wrappers is insane, but 1 is fine
16:07:16JdGordonpresumably the lcd api is the oldest bit of code, which is why it has evolved so painfully? :)
16:07:48Zagoryeah, it's pretty old
16:10:53Zagor415 calls in firmware/ and 934 in apps/ :-)
16:11:17Zagorto lcd_puts*
16:11:46Zagorsome of those in firmware/ are the implementations of course
16:12:05Zagoranyway, this is post-3.10 material
16:12:23JdGordonwhy so man in apps? all plugins?
16:12:55Zagoryeah, 804 in apps/plugins/
16:16:21kugelyoffset is different unfortunately
16:18:59 Quit gevaerts (*.net *.split)
16:18:59 Quit tmzt_ (*.net *.split)
16:18:59 Quit feisar- (*.net *.split)
16:19:00 Quit avacore (*.net *.split)
16:19:18Zagorkugel: how?
16:21:17 Join feisar- [0] (
16:24:26 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
16:24:26 Join tmzt_ [0] (
16:24:26 Join avacore [0] (
16:32:21Zagoryoffset looks like just ypos adding to me
16:33:09 Join keyb_gr [0] (
16:33:45Zagorcompensating for ypos being line oriented
16:41:41kugelZagor: that's about it yes
16:42:03Zagorso we don't need it in a pixel-oriented api
16:44:27Zagorback to lcd_puts(x, y, xcrop, style, scrolled, fmt, ...)
16:44:27 Join toffe82 [0] (
16:46:59Zagordid anyone (Llorean?) write down a list of the known problems we need to solve before 3.10?
16:52:27 Part Zagor
17:02:58 Join tmzt [0] (
17:05:33 Quit tmzt_ (Ping timeout: 256 seconds)
17:05:34 Quit avacore (Quit: Leaving)
17:05:43 Join avacore^ [0] (
17:21:56 Join matze` [0] (
17:24:11 Join ChickeNES [0] (~ChickeNES@
17:25:32 Join TheLemonMan [0] (
17:31:05 Quit zchs (Ping timeout: 252 seconds)
17:31:30 Join GermanMushroom [0] (
17:32:49 Join n1s [0] (~quassel@rockbox/developer/n1s)
17:44:32 Quit hilbert (Ping timeout: 260 seconds)
17:45:28 Join powell14ski_ [0] (~powell14s@
17:49:19 Join AlexP [0] (~alex@rockbox/staff/AlexP)
17:56:33 Join kadoban [0] (
17:57:32***Saving seen data "./dancer.seen"
18:05:36 Join mortalis [0] (
18:05:51 Join wodz [0] (
18:06:04CIA-14New commit by wodz (r30741): Switch to SYS mode on arm FS #12322 by me
18:06:34 Join fatline [0] (
18:07:21 Join hilbert [0] (
18:08:20CIA-14r30741 build result: 100 errors, 9 warnings (wodz committed)
18:12:56wodzcan someone tell me whats going on?
18:13:17Torneoh dear.
18:13:21*Torne has a look :)
18:14:04n1slooks like a full disk or some other disk problem
18:14:10AlexPwodz: Yeah, that build client is buggered :)
18:14:14Torneon every client?
18:14:24Torneoh, wait
18:14:26AlexPTorne: Same client for all those errors
18:14:26n1sseems to be allAsuka-w1ll14m
18:14:42n1syeah, erroriing out instantly is really fast :)
18:14:47AlexPIt needs a banning :)
18:14:49Tornen1s: True!
18:15:46wodzI would love to see real build log for this commit
18:16:16AlexPwodz: I think Bagder or Zagor can trigger a round manually
18:16:19TorneSome builds did happen
18:16:24n1swodz: ask zagor to ban the clent and rerun the build
18:16:36AlexPanyone can ban it
18:17:13Tornethe list is in svn, right?
18:17:29AlexPI can't remember which file though
18:17:37*AlexP prods gevaerts
18:17:46*gevaerts jumps
18:17:51AlexPAs he has done a fair amount of banning
18:17:52n1shmm, maybe not full disk but rather permission problems
18:18:03 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:18:33CIA-14New commit by gevaerts (r30742): Block Asuka-w1ll14m for brokenness
18:18:50AlexPta :)
18:23:32 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:30:33 Quit n1s (Remote host closed the connection)
18:35:45 Join freddyb [0] (
18:37:00 Join Strife89 [0] (~Strife89@
18:48:14 Join Jerom [0] (~jerome@
18:50:27 Quit ChickeNES (Quit: Computer has gone to sleep.)
19:11:05LalufuGood news: it's not the disabled interrupts in the buflib code.
19:11:16LalufuBad news: I don't know what it is.
19:11:36 Quit GermanMushroom (Ping timeout: 245 seconds)
19:13:13 Join simonlnu [0] (R41zQiWaA1@unaffiliated/simonrvn)
19:16:44 Join y4n [0] (y4n@unaffiliated/y4ndexx)
19:21:42 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
19:25:59 Quit toffe82 (Remote host closed the connection)
19:27:40 Quit Thra11 (Read error: Operation timed out)
19:29:19 Quit hilbert (Quit: Computer has gone to sleep.)
19:39:11 Join hilbert [0] (
19:43:25 Quit freddyb (Read error: Connection reset by peer)
19:44:46 Quit mortalis (Quit: CGI:IRC)
19:46:45 Quit simonlnu (Ping timeout: 252 seconds)
19:47:23 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
19:48:16 Join Thra11 [0] (~thrall@
19:49:07 Join simonlnu [0] (u44lL3jjKK@unaffiliated/simonrvn)
19:54:26 Quit TomColler (Remote host closed the connection)
19:54:50 Nick kugel is now known as kugelp (~kugel@rockbox/developer/kugel)
19:55:50 Join zchs [0] (
19:57:33***Saving seen data "./dancer.seen"
19:57:48 Quit fatline (Quit: This computer has gone to sleep)
19:58:49 Quit WalkGood (Quit: CaminaBien ♪ ♫ ♪ ♫)
20:08:56 Join fatline [0] (
20:16:05 Quit fatline (Quit: This computer has gone to sleep)
20:19:17 Nick kugelp is now known as kugel (~kugel@rockbox/developer/kugel)
20:20:04 Quit Jerom (Quit: Leaving.)
20:20:29 Join Xerion [0] (
20:25:44CIA-14r30741 build result: All green
20:26:29 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
20:26:46Zagorta-daa :)
20:27:36 Quit TheLemonMan (Quit: WeeChat 0.3.5)
20:32:24 Join freddyb [0] (
20:33:24 Join lovasoa [0] (~lovasoa@
20:34:15freddybkugelp: Do you (or anyone else) have any objection to the current version of FS #12299? (Assuming I can get Andree's player sorted out.)
20:34:16fs-bluebot Font Glyph cache is no longer getting saved (bugs, new)
20:40:14 Join mortalis [0] (
20:41:46 Quit lovasoa (Read error: Connection timed out)
20:41:57 Join lovasoa [0] (~lovasoa@
20:45:45 Quit lovasoa (Client Quit)
20:51:42 Join fatline [0] (
20:53:34 Join GermanMushroom [0] (
21:08:31 Quit keyb_gr (Ping timeout: 252 seconds)
21:17:25 Quit pamaury (Remote host closed the connection)
21:17:59 Quit GermanMushroom (Read error: Connection reset by peer)
21:21:25kugelfreddyb: havent looked at the patch yet
21:38:04 Join Buschel [0] (
21:38:09freddybok, you're not opposed to the concept of a glyph file per font, correct?
21:39:19freddybBuschel, I'm reading your post now.
21:39:38Buschelfreddyb: sorry, that my setup makes such problems :/
21:41:41freddybNo worries. :)
21:41:56 Quit bluebrother (Quit: reboot)
21:42:07 Quit fs-bluebot (Remote host closed the connection)
21:42:17BuschelI hope you get this sorted out... You do not experience any of my issues?
21:46:35 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
21:47:24 Join fs-bluebot [0] (
21:47:33freddybI have usb screen issues, especially with HID on, but I think it's a different issue. I got zero sized cache files at one point... Maybe I have to switch powermgmt to unload the fonts instead of dumping the caches because I don't see any other place the font handles get closed.
21:53:29ZagorJdGordon: didn't you take a shot at "turn LCD_WIDTH/HEIGHT into variables"? I can't find that patch.
21:54:44Zagorah, fs#11615
21:54:45fs-bluebot Dynamic screen size (patches, new)
21:57:37***Saving seen data "./dancer.seen"
22:06:28 Quit Horscht (Read error: Connection reset by peer)
22:06:56 Join Horscht [0] (~Horscht@xbmc/user/horscht)
22:10:10 Quit mystica555_ (Quit: Ekkusu Chatto)
22:10:51 Join mystica555_ [0] (
22:11:30 Quit mystica555_ (Remote host closed the connection)
22:14:38 Quit matze` (Remote host closed the connection)
22:14:38 Quit Buschel (Quit: ChatZilla 0.9.87 [Firefox 7.0.1/20110928134238])
22:21:43 Join Thra11_ [0] (
22:24:29 Quit Thra11 (Ping timeout: 260 seconds)
22:25:58 Quit kadoban (Quit: bye)
22:26:04 Join kadoban_ [0] (
22:29:29 Join mystica555_ [0] (
22:39:33 Quit mystica555_ (Read error: error:1408F10B:SSL routines:SSL3_GET_RECORD:wrong version number)
22:41:41 Join mystica555_ [0] (
22:43:49 Quit mystica555_ (Read error: error:1408F10B:SSL routines:SSL3_GET_RECORD:wrong version number)
22:45:42 Join liar [0] (
22:50:27 Quit mortalis (Quit: CGI:IRC)
22:57:49 Join mortalis [0] (
23:02:56 Quit mortalis (Quit: CGI:IRC)
23:09:18 Join mystica555_ [0] (
23:23:57Zagorsurprisingly, lcd_clear_display() gets angry when lcd_frambuffer = NULL
23:27:18bertrikwow, there are a lot of button contexts
23:30:49bertrikperhaps a silly question, but is it possible to go to the wps when there is no audio playing?
23:31:06Zagoronly when resuming
23:31:27Zagoror if you are paused
23:31:57bertrikso it's basically impossible to be in the wps with audio stopped?
23:32:04Zagoriirc yes
23:39:10 Quit freddyb (Ping timeout: 258 seconds)
23:46:30 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)
23:57:40***Saving seen data "./dancer.seen"

Previous day | Next day