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

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

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

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

#rockbox log for 2009-07-23

00:00:53Skip2for Sansa Fuze?
00:02:01kugelsaratoga: maybe, I'll see if I can pull from his repo
00:02:21kugelhmm, not quite up-to-date
00:04:01kugelour touchscreen guys really have no talent for porting cabbiev2 :(
00:05:19aidyshotofadds: i'll test
00:09:52 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
00:10:02 Join safetydan [0] (n=deverton@rockbox/developer/safetydan)
00:10:28CIA-8New commit by 03kugel (r22003): Commit part of FS #9873 - Utilise buttons for playlisting ...
00:10:48aidyshotofadds: works :)
00:16:54 Quit phoboss ("Leaving.")
00:18:52 Join Thundercloud [0] (
00:19:36kugelPaulJam_: fixed :)
00:20:57kugelJdGordon|: any idea why the lists have a pointer array (a pointer to a vp for each screen), instead of just a single pointer (expecting that the vp for the remote is right after)?
00:21:46JdGordon|umm... thats a bad assumption
00:21:56JdGordon|i tihnk...
00:22:11kugellists have struct *viewport[NB_SCREENS], all other code just has viewport[NB_SCREENS]
00:22:24 Part wincent ("Kopete 0.12.7 :")
00:22:31linuxstbkugel: Does SELECT+PLAY works on ipods? Long select and long play are already mapped to things.
00:22:34kugelI don't think it's a bad assumption, unless we start mallocing viewport structs
00:22:36 Join BCM43 [0] (
00:22:55kugellinuxstb: I expect so
00:23:00JdGordon|hmm... yeah fix it
00:24:06pixelmaI don't like "short Rec" being used on the c200 and I believe "go to recording screen" on "long Rec" would need "Rec" as a pre condition then too
00:24:15 Join wincent [0] (
00:24:35 Part wincent ("Kopete 0.12.7 :")
00:25:13CIA-8New commit by 03alex (r22004): Fix manual button tables for disktidy, mpegplayer, text editor and ...
00:25:39 Join wincent [0] (
00:26:02*AlexP must be half asleep tonight
00:26:30 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
00:27:45kugelpixelma: why would it need that?
00:28:53 Quit Neovanglist (Read error: 60 (Operation timed out))
00:32:54 Quit bertrik (Read error: 113 (No route to host))
00:33:26 Part wincent ("Kopete 0.12.7 :")
00:37:13CIA-8New commit by 03kugel (r22005): Fix reds (disable for m200, take one of the two pitchscreen combos for mrobe100)
00:39:30JdGordon|arg, that commit got in when I was afk...
00:39:46*JdGordon| still tihnks that not giving the option of which playlist action to use is stupid
00:40:09JdGordon|oh, misread
00:40:12JdGordon|ignore me...
00:45:19 Join mcuelenaere [0] (
00:46:09kugelJdGordon|: I will :)
00:47:54JdGordon|I dont suppose anyone has looked at the ipod button driver recently have they? the buttons not working after hold is still there and driving me bonkers :p
00:49:39 Quit BCM43 ("Leaving")
00:50:18 Quit ender` (" To get as fewest unhappy people as possible, always bully the same ones.")
00:51:36*kugel slaps his forehead
00:52:37 Quit B4gder ("It is time to say moo")
00:53:10linuxstbJdGordon|: I don't think anyone has looked at it in about three years...
00:53:51JdGordon|yeah :/ thats what I thought.... I looked at it ages ago but didnt see anything obvious that would cause it
00:54:55linuxstbI have a faint recollection of someone from IPL saying the clickwheel hardware needed re-initialising after hold. I'm not sure if I tried it though...
00:55:28linuxstbI think it was the ipodloader2 dev - so looking at that code might give a clue.
00:55:48CIA-8New commit by 03kugel (r22006): Fix red caused by failure to close comments properly.
00:56:19JdGordon|and actually using the wheel re-inits it? whereas pressing it doesnt?
00:56:52linuxstbIt possibly gets confused...
00:57:47linuxstbIIUC, the clickwheel is controlled by a little microcontroller, which sends messages regarding wheel/button events.
00:59:11CIA-8New commit by 03rob (r22007): D2: Notify the backlight driver when the HOLD button is toggled. Fixes FS #10451.
01:00:14JdGordon|ok, ill try maybe forcing an init or something..
01:02:08***Saving seen data "./dancer.seen"
01:05:04mcuelenaerekugel: RE r21998: doesn't the sizeof(..)/sizeof(..) get reduced to a constant by GCC? (as sizeof() doesn't support dynamic sizes AFAIK)
01:05:05 Quit LambdaCalculus37 ("off to get food")
01:05:27linuxstbJdGordon|: There seems to be code in loader2 to handle this -"> (search for "aaaa")
01:06:29*mcuelenaere didn't know of backlight_hold_changed()
01:06:49linuxstbi2c_reset_status and i2c_reset_count just look like debugging, so you can ignore those.
01:07:36 Nick WarpCycle is now known as DarkSpectrum (
01:08:26linuxstbJdGordon|: Seems straightforward to copy into button-clickwheel.c - I hope it works...
01:09:13 Quit DarkSpectrum ()
01:16:00 Join DarkSpectrum [0] (
01:17:01 Quit shotofadds ("Leaving")
01:19:19 Quit PaulJam_ (Nick collision from services.)
01:19:24 Join PaulJam [0] (
01:20:14Unhelpfulmcuelenaere: sizeof() should be compile-time constant, but i'm not sure what happens with, for example, a dynamic array
01:21:05mcuelenaereUnhelpful: I think GCC would bail out saying it can't determine the size of the variable
01:22:28Unhelpfulmcuelenaere: really? since it can allocate the array dynamically, i'd expect it to generate an expression along the lines of len * sizeof(elem) for sizeof(dynarray)
01:23:28pixelmakugel: I thought it's needed for things where you have button release and repeat with the same button (looks "cleaneer" to me). Also, I don't understand "I expect it to" on linuxstb's question - you can try out these action things in a sim - unlike electrical or mechanical restrictment that some targets have. On that note I wanted to tell you that the X5 pad mapping won't work because you can't have combos there except ones with "Power" but that's
01:23:28pixelma also not really usable since there is a hard power off if you hold "Power" for too long
01:23:39 Quit DarkDefender (Remote closed the connection)
01:23:59 Quit fdinel (Read error: 110 (Connection timed out))
01:24:40kugelI don't think it's needed
01:25:19kugeland no, I won't test some 40 sims which do not reflect hardware restrictions on my slowmo laptop
01:26:01pixelmathen ask for testers, explicitly
01:27:01Unhelpfulmcuelenaere: it compiles with our 4.0.3 arm toolchain, though the test function i gave it isn't exactly optimal. int test(int l) { int arr[l]; return sizeof(arr); } works, and it calculates the size just by shifting r0, but it goes on to save sp to another register, subtract the array length from it, and then restore it...
01:27:03kugelIIRC I mailed to the -dev list early enough
01:28:48kugelfeel free to fix, as I said, they could possibly be improved on some targets
01:29:11pixelmaalso, I would have tried some sims even when not all, I think one of the 4g pad Ipods would have been one of the first
01:29:23 Nick fxb is now known as fxb__ (
01:29:30pixelmavery kind of you
01:29:35mcuelenaereUnhelpful: ah ok, I found but it doesn't help much
01:29:55Unhelpfulmcuelenaere: it seems like gcc docs never do help much :/
01:29:59mcuelenaereanyway, the sizeof() is probably constant'd in this case (touchactions seems to be a static array with a static size)
01:30:18kugelwhat's the problem? Since when are we required to commit perfect code? And what's the problem with putting stuff into SVN for wider testing?
01:30:37mcuelenaerekugel: it isn't a problem, just nit-picking :)
01:30:51kugelmcuelenaere: it seems to be a problem...
01:31:08kugelfor pixelma, at least
01:31:35mcuelenaereoh sorry, wrong subject
01:31:39*mcuelenaere meant the sizeof() thingy
01:32:16kugelah, hehe :)
01:32:26 Nick advcomp2019__ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
01:34:52linuxstbkugel: Unless it's a major change being actively worked on, I would say we should always aim to commit perfect code - there's never any urgency....
01:35:18kugelwell, I aim for it, of course.
01:36:07linuxstbBy testing things?
01:36:40kugellinuxstb: btw, select+down works well on the e200 (both long select and long down are mapped), so I don't see a problem with the ipods
01:37:03 Quit Thundercloud (Remote closed the connection)
01:37:09kugelif people don't care about what I write to the -dev list, i can't help it
01:37:32linuxstbThen it's obviously not a feature people are interested in... Which says something else.
01:38:30mcuelenaereI was observing the stack usage a bit earlier, and saw the codec thread switching a lot between running and sleeping state; after looking in the code it seems every codec loops while(wait_for_takeoff()) sleep(1); Shouldn't this get replaced with a wakeup or similar?
01:39:20Unhelpfulthat certainly looks like a case for an event queue...
01:39:43mcuelenaerean event queue seems a bit overkill to me
01:40:01kugelwhy overkill?
01:40:12mcuelenaerea simple wakeup_wait() (and wakeup_signal()) is enough
01:40:27mcuelenaere(although wakeup objects aren't enabled in all targets by default)
01:40:49mcuelenaeredoesn't a event queue struct require more binsize than an wakeup one?
01:41:54Unhelpfulthat probably depends on whether we'd want to queue *other* events to codecs?
01:42:27 Join funman [0] (n=fun@rockbox/developer/funman)
01:42:52 Quit BradC (Read error: 113 (No route to host))
01:43:11mcuelenaereUnhelpful: like what?
01:44:10Unhelpfulmcuelenaere: i'm not sure... i'm assuming we handle track skips by just resetting the codec or loading a new one, so we don't need events for that?
01:45:25mcuelenaereyes, I think something like that; IIUC there's a thread which handles this and sets a flags when the currently loaded codec needs to decode another file; otherwise it loads another codec
01:46:00mcuelenaere(the code I mean is while (!*ci->taginfo_ready && !ci->stop_codec) ci->sleep(1); , it's duplicated in all codec_main() functions)
01:48:36kugelfunman: the fuzev2 arrived already
01:49:13funmankugel: nice!
01:49:19funmandid you start looking at LCD code?
01:50:08kugelI just came back from a 5-day rockbox-break and saw it in the mailbox
01:50:59funmanfuze OF 2.1.17 : init_lcd at 0x8768 , lcd_write_reg at 0x8894
01:51:21kugelmcuelenaere: that seems like 2 events
01:51:40mcuelenaerekugel: sure, but they can both be handled by 1 wakeup
01:52:00mcuelenaerebut you could of course use an event queue if you really wanted
01:52:10kugelyes, you can handle infinte events with a wakeup+flag combo
01:52:46kugelI don't really care, but what you pasted looks "improvable" :)
01:53:06*mcuelenaere just wants all threads to sleep when nothing is being done :)
01:53:10kugelfunman: can I download that version somewhere (/me goes looking on Bagder's page)
01:53:40pixelmakugel: it's just that I try my best and test as much as I can if I make some changes of things that can be tested (e.g. bitmap changes in plugins are tested in sims of all affected screen size) since I know it's easier for me to try and get things right in the beginning than if someone else has to clean up later, maybe bit by bit. Even if it's more effort (basically just time though) for me in the beginning it would be more effort later. If
01:53:40pixelma someone else would have told me "this button combo doesn't work" in code I just commtted and I am familiar with, I wouldn't have told him "go fix it yourself"...
01:53:41kugelindeed there is it
01:54:00funmankugel: i think Bagder hasn't fuzev2/clipv2 firmwares
01:54:47kugelpixelma: 9 days on the -dev ml, I have nothing more to add
01:57:50funmanthe memory regions at 0x0 and 0x30000000 seem different : for example 0x9ADC branch to 0x30018A69, and 0x18A69 is not a function entry point
02:03:03 Quit PaulJam (Nick collision from services.)
02:03:11 Join PaulJam_ [0] (
02:05:26 Quit r0b- (Read error: 110 (Connection timed out))
02:05:36 Quit Lynx_ (" HydraIRC -> <- s0 d4Mn l33t |t'z 5c4rY!")
02:06:40 Join Hillshum [0] (n=hillshum@unaffiliated/hillshum)
02:07:49 Join r0b- [0] (
02:10:57 Join DarkSpectrum- [0] (
02:11:00funmanfuzev2 seems to have the same SD code than clipv2 (which is good for us!)
02:11:16*DarkSpectrum- is working on WPS for tuner, this may be a dumb question but how much ram does each preset use?
02:11:53 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
02:12:05 Nick DarkSpectrum- is now known as DarkSpectrum (
02:12:35linuxstbkugel: BTW, yes, SELECT+PLAY works as a combo on the ipod - if you press and hold both buttons at the same time. Pressing either first, then the other will obviously not work, and that's how I've always used combos. Looking at the keymaps, the e200 seems about the only target to do these kind of combos.
02:12:52linuxstb(I'm not saying it's bad, just not how combos have been in the past)
02:13:27kugelI can press select then down, on my e200 to get to the id3screen
02:13:33funmanDarkSpectrum: memory for FM preset is statically allocated anyway, why the question?
02:13:47kugelso it's not so obvious to me
02:13:53DarkSpectrumjust trying to figure out how much ram i might have to play with
02:14:37funmanDarkSpectrum: you don't need to think about it
02:14:38DarkSpectrumi'm new to all this and i'm just trying to plan everything i can think of out
02:15:01 Join JdGordon|_ [0] (i=ad75c29c@gateway/web/freenode/x-9a2daefeb7f4755e)
02:15:14linuxstbkugel: Doesn't that depend on how long you hold select before pressing down? Long select is mapped to the context menu.
02:15:33JdGordon|_linuxstb: thanks
02:16:02linuxstbJdGordon|_: "Thanks, it works", or "Thanks, I'll try it" ? I can't recreate that bug on my ipod video...
02:16:07funmanthere is a hardcoded limit of 64 presets which consist of 4 bytes for the frequency (int) and 28 bytes for the preset name (27 chars + '\0')
02:16:12linuxstb(so I can't try to fix it...)
02:16:13 Quit mcuelenaere ()
02:16:14kugellinuxstb: down, then select doesn't work
02:16:17JdGordon|_linuxstb: thanks ill try it
02:16:48DarkSpectrumalso is there a reason that the current is "/recorder/tuner.h" etc... or can i start a "/tuner" ?
02:17:04 Join fdinel [0] (
02:17:06DarkSpectrumerr /apps/*
02:17:26kugelthe repeat kicks in after 0.5s or so here, so if I begin to hold select I have this 0.5s to press down
02:17:30funmanDarkSpectrum: i think the original reason is that radio code was for archos recorder, so moving it into /apps/tuner would be good IMO
02:17:53JdGordon|_DarkSpectrum: yes, moving all radio/preset stuf into apps/radio or tuner would be great
02:18:01kugellong select is also mapped to the context menu on the e200
02:18:16JdGordon|_*cough* semi customizable buttons
02:18:45AlexPfunman: I don't think DarkSpectrum was talking about presets, but rather about coding a WPS for the FM screen - A WFMS if you like
02:19:18AlexPer, me misread
02:19:22AlexPIgnore me :)
02:20:18JdGordon|_hang on a sec... DarkSpectrum are you trying to do WPS for the tuner?
02:20:38linuxstbDarkSpectrum: You're asking about creating an apps/tuner/ directory, and moving all the radio-related code there? If so, then yes, I think that's a good idea.
02:20:57DarkSpectrumyep, thats what i'm talking about
02:21:24DarkSpectrumi'll do that first then before getting into the bulk of the FMPS
02:21:30kugellinuxstb: do you have an ipod to test the combo?
02:21:51DarkSpectrumerr better TPS, TunerPS
02:21:58linuxstbkugel: Yes, I did just test.
02:21:58AlexPWFMS :)
02:22:12AlexPor WTS perhaps
02:22:18linuxstbkugel: (before I said "BTW, yes, SELECT+PLAY works as a combo on the ipod")
02:22:21funmanWhile Tuning Screen?
02:22:30AlexPor While Tuner Screen
02:22:36AlexPOr While FM Screen
02:23:16AlexPanyway, I don't think the extension is the critical bit
02:23:28 Quit Skip2 ("Beatings will continue until morale improves. | I have things to do and schemes to plan.")
02:23:29AlexPAlthough knowing us it'll cause the most arguments :)
02:23:30DarkSpectrumwhere should users store their .tps's :) ?
02:23:39kugellinuxstb: the "at the same time" thing makes me curious, though. It doesn't have to be so simultanously on the e200
02:24:00linuxstbMaybe I'm just slower than you...
02:24:06DarkSpectrum.rockbox/TPS ?
02:24:32kugelat the very same time actually doesn't work.
02:24:34AlexPDarkSpectrum: I'd assume it'd follow the same model as the WPS
02:24:35kugelfor me that i
02:25:01AlexPDarkSpectrum: But that can easily be moved when it works
02:25:01JdGordon|_DarkSpectrum: the best firs step to getting a WPS-like thing going for the tuner is to clean up radio.c and move all its drawing code out so its sperate...
02:25:05DarkSpectrumor put them in the current .rockbox/WPS with somehow having the current WPS files having a TPS subsection?
02:25:18AlexPDarkSpectrum: I don't think this is your major issue
02:25:51linuxstbkugel: It could also be hardware differences. I've never used an e200, and I'm guessing you've never used an ipod?
02:26:00kugelthat's true
02:26:22JdGordon|_as for the name/format.. there are 2 ideas how to do it.. the best way to do it is probably add some new radio tags and a "mode" tag to the ezxsting wps stuff, and then just have a monolithic .wps for both screens
02:26:32kugelbut BUTTON_REL and BUTTON_REPEAT should be the same for all targets (it's handled in button.c, not in hardware specific code)
02:27:09AlexPJdGordon|_: Are there any targets that can share WPSs, but one doesn't have a radio?
02:27:28JdGordon|_I dunno
02:27:32PaulJam_H300 and ipod video (i think)
02:27:34AlexPIf so, would all radio stuff get ignored if there isn't a radio present
02:27:42kugelDarkSpectrum: we had some talk about it during the devcon, bertrik wanted to give it a shot (but was probably interrupted by the (damn :p) meizus)
02:27:42linuxstbAlexP: All the targets with "optional" radios...
02:27:53JdGordon|_hmm... no, we should support seperate .wps files also
02:27:59AlexPOr would you need to stuff all radio tags inside have radio cnditionals?
02:28:01DarkSpectrumso should i find something else to work on then?
02:28:17*DarkSpectrum just wants a pretty tuner :P
02:28:22JdGordon|_go for it
02:28:24AlexPDarkSpectrum: No, I don't think bertrik is working on it at all
02:28:29AlexPSo go ahead :)
02:28:30*JdGordon|_ is happy to help you there
02:28:39AlexPI'd like a nice FM screen too
02:28:50AlexPWith e.g. station logos as "album art"
02:29:25kugelyea that would be neat
02:30:02DarkSpectrumand i want to try to make it work with the current wps system, hopefully by making it a section in current .wps files it would make it easier/encourage people to theme both at the same time
02:30:17AlexPDarkSpectrum: That is potentially difficult though
02:30:33kugelI think we agreed on seperate files (?)
02:30:41JdGordon|_DarkSpectrum: the only way we would accept it is if it sed the current system to do the drawing.. we dont want two seperate thme systems
02:30:48AlexPDarkSpectrum: As PaulJam_ says, h300 has radio, ipod photo doesn't but share WPS. Some e200 have radio, some don't
02:30:52AlexPas linuxstb said
02:30:58AlexPkugel: I agree
02:31:10DarkSpectrumi see your point
02:31:17DarkSpectrumok seperate it is
02:31:23AlexPI'm purely talking about the actual files
02:31:23linuxstbCouldn't it be done with conditional viewports? Just adding a "if radio screen" tag?
02:31:55AlexPlinuxstb: I guess so, but I can see it making WPS files quite (even more) confusing
02:31:55*linuxstb isn't convinced about separate files - one file for everything could be simpler
02:32:13AlexPIs there a technical reason to prefer a single file?
02:32:17JdGordon|_we should support that radio screen tag so both can share, AND the ability to use two completly seperate .wps files
02:32:19linuxstbAlexP: Maybe more confusing for a WPS author, but not for users...
02:32:25JdGordon|_AlexP: much simpler codewise
02:32:29AlexPlinuxstb: hmmm, true
02:32:35DarkSpectrumlinuxstb: thats what i was originaly thinking and it would make a "system wide" theme easier for people to make
02:32:46linuxstbAnd as JdGordon|_ said, much simpler codewise as well.
02:32:50JdGordon|_people might very well want two screens that look nothing alike
02:33:01AlexPI was thinking thought that seperate ones let you mix and match
02:33:16AlexPi.e. WPS from one theme, WFMS from another
02:33:22JdGordon|_both would *have* to be supported :)
02:33:32DarkSpectrumforum poll? see what people want?
02:33:33*JdGordon|_ keeps repeating self with stronger language :D
02:34:07 Quit PaulJam_ (Nick collision from services.)
02:34:12 Join PaulJam [0] (
02:34:33DarkSpectrumok seperate it is
02:34:33kugelI think if we allow seperate wps and rwps, we should also allow seperate wfms
02:34:53AlexPwps and rwps are physically different screens though
02:34:56*linuxstb doesn't see a need to support two different themes at the same time...
02:35:01kugelDarkSpectrum: wps for remotes
02:35:01AlexPDarkSpectrum: For the LCD remotes
02:35:03DarkSpectrumoh n/m
02:35:15JdGordon|_DarkSpectrum: first step... rip the current drawing code out of radio.c and stick it into a seperate function....
02:35:38AlexPyes, if they were seperate it would also necessitate a RWFMS as well
02:35:43DarkSpectrumfirst step = clean up radio.c and move everything ;)
02:36:03JdGordon|_yep :)
02:36:15AlexPWe can argue after that :)
02:36:20JdGordon|_I have an old patrch on flyspray which is a start you might like to look at
02:36:22AlexPer, s/argue/discuss/ :)
02:36:29DarkSpectrumhave the # ?
02:37:26JdGordon|_but really. if you arnt interested in how it works and just want to get themeing going then it can be skipped
02:37:48DarkSpectrumno, i've allready been reading through it
02:38:03JdGordon|_its nasty isnt it? :D
02:38:13*DarkSpectrum refuses to say anything ;P
02:38:45JdGordon|_my aim with that was to move the lower lever radio stuff back into firmware where it belongs
02:39:26kugelfunman: any idea how to sync ecrits work properly (without bricking my fuze)?
02:39:34DarkSpectrumi'm sure i can hack something out
02:39:49DarkSpectrummay not be pretty at first but....
02:40:06JdGordon|_well.. you cant make it any worse :)
02:40:08funmankugel: what's wrong with his diff ?
02:40:27kugelwhich diff? I only saw the git repo which is pretty old
02:40:42funmanthe diff between svn and the head of his repo
02:40:55kugelit's out of sync
02:41:10funmanwith mkamsboot?
02:41:23kugelhis repo hasn't been rebased recently, it doesn't have the dualboot/ separation for example
02:41:42DarkSpectrumJdGordon: a start? crap thats changing alot allready
02:42:25JdGordon|_yeah, then i lost interest when it got confusing
02:42:52DarkSpectrumTBH i don't know if i will concern myself with firmware, that may be a little to advanced for me
02:43:10JdGordon|_probably a wise choice :)
02:43:33DarkSpectrumwith what i'm thinking all i need to do is edit and pretty much cut and paste code from other places
02:43:39*linuxstb wonders who ecrits is, and what he's been working on
02:44:14JdGordon|_DarkSpectrum: there shouldnt be much pasting.... lots of cut and add though
02:44:22funmankugel: can't you merge the diff manually ? it's not very long
02:44:26kugellinuxstb: he did mkamsboot support & dualboot for the fuzev2
02:44:41kugelfunman: I don't want to break things..
02:44:55DarkSpectrumyeah but alot of what i need to do allready exists, it's only putting it together
02:45:00linuxstbkugel: Ah, nice. But he hasn't submitted a patch, just a git repo?
02:45:07kugelyou're the mkamsboot expert so I thought it might be a good idea to ask you
02:45:29kugellinuxstb: yea, he also was a bit active in our forums
02:45:53*JdGordon|_ back in 15
02:46:08linuxstbIt should be easy (and useful) to set up regression testing - just get a set of original firmware files, and write a script to patch them with mkamsboot. Then make your changes, then run the script again...
02:46:53kugelmkamsboot doesn't handle fuzev2 without his patch
02:47:40DarkSpectrumok off readin through this stuff, if anyone has any sugestions page me please, i can use all i can get
02:47:57linuxstbkugel: Yes, I know. What are you worried about breaking then?
02:48:24kugeldualboot if I fail to sync properly
02:48:54 Quit JdGordon|_ (Ping timeout: 180 seconds)
02:48:57kugelbut I guess I could compare the output of his old mkamsboot with my sync'd mkamsboot
02:50:58funmankugel: (you need to modify scramble as well, and also add a target to tools/configure)
02:51:48funmanand add the missing trailing ',' line 188 of mkamsboot.c ^^
02:54:42CIA-8New commit by 03mcuelenaere (r22008): Lua: return nil when kbd_input() gets aborted and make it possible to preset the input string (based on FS #10455 by Christophe Gragnic)
02:55:35 Part JdGordon
02:56:14 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
03:02:11***Saving seen data "./dancer.seen"
03:03:13*JdGordon cant seem to find where the wps loads the .wps file :/
03:04:35 Quit timc (Remote closed the connection)
03:05:09 Nick Ypsy is now known as YpsyZNC (
03:05:18funmankugel try ascodec_write(0x1b, 0x80) to turn the fuzev2 backlight on (0x85E4)
03:07:08kugelfunman: you addresses differ from mine, did you strip something?
03:07:44funmanaf6eceaa0358295886837e2c7d66fb98 fuze-2-1-17.bin
03:07:57funmanthat's the firmware block (136844 bytes)
03:08:20kugelI have write_reg at 0x8488
03:09:29kugelI think I found ams_dbop_init
03:09:48JdGordonlinuxstb: woohoo... worked first go :)
03:10:02funmanstrangely enough I find 2 different fuzev2 IDs in the OF (0x6d for 1.17 and 0x70 for 2.26)
03:11:18funmankugel: you forgot to remove the OF header (0x400 + 0x8088 = 0x8488)
03:11:34kugelso you stripped something
03:11:49funmansure, i stripped what ever is not code
03:11:59kugelI asked you and you said no, ffs
03:12:19funmansounded like an evidence
03:12:54funmani always disassemble the arm code, with the reset vector being at 0, so i make sure every absolute offsets are correct
03:13:52funmanutils/AMS/hacking/extract_fw.c should be able to extract the firmware for you (perhaps it needs adaptation for fuzev2)
03:14:50 Quit funman ("free(random());")
03:20:16kugelDBOP_BASE is apparently the same
03:24:19 Quit Hillshum ("Leaving")
03:27:17JdGordonDarkSpectrum: looking into it.. it shouldnt actually be too much work to get a WPS going in the fm screen... (well quick+dirtily anyway)
03:29:12DarkSpectrumsame thinking here
03:32:27kugelapparently no afsel setting
03:57:56 Quit gevaerts (Nick collision from services.)
03:58:07 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
04:06:14 Quit TheSeven (Nick collision from services.)
04:06:32 Join The_Seven [0] (
04:06:36 Nick The_Seven is now known as TheSeven (
04:10:52 Join midgey [0] (
04:15:45 Join DarkSpectrum- [0] (
04:15:46 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
04:20:59 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
04:29:05 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
04:40:36 Quit togetic (Read error: 104 (Connection reset by peer))
04:40:54 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
04:42:48 Quit DarkSpectrum- ()
04:47:13 Join DarkSpectrum [0] (
04:59:03 Quit togetic (Remote closed the connection)
04:59:42 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
05:02:03DarkSpectrumJdGordon: where does radio.c first get called from, i can't find it
05:02:16***Saving seen data "./dancer.seen"
05:02:46 Quit PaulJam (Nick collision from services.)
05:02:53 Join PaulJam_ [0] (
05:08:28 Quit togetic (Remote closed the connection)
05:08:49 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
05:12:40 Quit JdGordon ("Leaving.")
05:14:01 Quit PaulJam_ (Nick collision from services.)
05:14:07 Join PaulJam__ [0] (
05:15:22 Quit togetic (Remote closed the connection)
05:15:43 Join JdGordon|_ [0] (i=1816d253@gateway/web/freenode/x-27994fbcd8da88ce)
05:15:48 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
05:15:58 Nick JdGordon|_ is now known as JdGordon (i=1816d253@gateway/web/freenode/x-27994fbcd8da88ce)
05:17:09 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
05:17:35 Quit togetic (Remote closed the connection)
05:17:58 Join togetic [0] (n=togetic@unaffiliated/ibuffy)
05:28:27 Join midgey_ [0] (
05:28:27 Quit midgey (Read error: 104 (Connection reset by peer))
05:32:09 Join SirFunk [0] (
05:35:21 Quit Sajber^ (Read error: 54 (Connection reset by peer))
05:45:28 Quit r0b- (Read error: 104 (Connection reset by peer))
05:45:43 Join r0b- [0] (
05:46:10 Quit PaulJam__ (".")
05:49:44 Quit Horscht ("Verlassend")
05:56:01 Join BradC [0] (
05:56:45 Quit CaptainKwel (Remote closed the connection)
05:58:47 Quit kugel (Read error: 110 (Connection timed out))
06:21:04 Quit SirFunk (Read error: 104 (Connection reset by peer))
06:33:14CIA-8New commit by 03kkurbjun (r22009): Superdom: Add support for 640x480 and 480x640 screens, simplify screen defines.
06:33:35 Join AndyI [0] (i=AndyI@
06:38:36CIA-8New commit by 03kkurbjun (r22010): Forgot to update sources for last revision (r22009)
06:42:53 Quit DarkSpectrum ()
06:45:48 Join DarkSpectrum [0] (
06:46:15CIA-8New commit by 03kkurbjun (r22011): Fix Onda
06:50:01 Join JdGordon1 [0] (
06:50:25 Quit JdGordon ("Page closed")
06:51:29 Quit AndyI ()
06:53:39 Quit AndyIL (Read error: 113 (No route to host))
06:55:48 Join AndyI [0] (i=AndyI@
07:02:19***Saving seen data "./dancer.seen"
07:10:02 Quit n00b2Hack ("Saliendo")
07:10:07*JdGordon1 just loves when things work in the sim but not on target :/
07:23:57CIA-8New commit by 03jdgordon (r22012): Get cuesheets working again (hopefully fixes FS #10460), on swcodec only read the cuesheet from the buffer once
07:24:45CIA-8New commit by 03jdgordon (r22013): this file also...
07:36:30midgey_JdGordon1: what was the point of r22013?
07:40:05 Quit Lss (Read error: 104 (Connection reset by peer))
07:40:21 Join Lss [0] (
07:43:22pixelmalinuxstb, kugel: the new button combo on the Ipod is Select+Play without pre condition, the ID3 screen shortcut on e200 is Select+Down with Select as pre condition and that makes the combos work with holding one button first and then the other, even when short and long presses of one of the buttons are mapped too...
07:51:12JdGordon1midgey_: cuesheet_subtrack_changed() has some side effects which need to happen even if do_full_update is true
07:52:18midgey_couldn't you have switched the ordering without introducing a new bool?
07:52:30midgey_although i guess, gcc should produce the same code
07:52:51JdGordon1yes, probably
07:53:51midgey_i didn't realize that the ordering changed, now it makes sense
08:05:02 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
08:25:00 Quit HellDragon (Read error: 104 (Connection reset by peer))
08:25:22 Join HellDragon [0] (
08:28:04 Join flydutch [0] (
08:31:11 Join ender` [0] (
08:33:43 Join Grahack [0] (
08:36:41 Join n8ature [0] (n=namick@
08:38:03 Part safetydan ("Leaving.")
08:38:35 Join Rob2223 [0] (
08:43:47 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
08:43:47 Quit pixelma (Nick collision from services.)
08:43:53 Quit amiconn (Nick collision from services.)
08:43:57 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
08:44:03 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
08:44:04 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
08:47:29 Quit TheSeven (Read error: 113 (No route to host))
08:47:59 Join bertrik [0] (
08:56:50 Quit Rob2222 (Read error: 110 (Connection timed out))
09:00:31 Join cloudnine [0] (
09:02:23***Saving seen data "./dancer.seen"
09:03:01 Join petur [50] (n=petur@rockbox/developer/petur)
09:05:18 Join webguest71 [0] (n=c1355313@gateway/web/cgi-irc/
09:05:27 Quit webguest71 (Client Quit)
09:12:09 Quit midgey_ ()
09:19:37 Quit BHSPitMonkey (Remote closed the connection)
09:19:48 Quit Stephen_ ("Leaving")
09:28:50 Quit cloudnine ("ChatZilla 0.9.85 [Firefox 3.5.1/20090715094852]")
09:32:14 Join Thundercloud [0] (
09:36:58 Quit bertrik (Read error: 113 (No route to host))
09:50:55 Quit Thundercloud (Remote closed the connection)
09:59:46 Join TheSeven [0] (
10:00:01 Quit kachna (Read error: 113 (No route to host))
10:03:27 Quit lyngaas (Read error: 110 (Connection timed out))
10:07:25 Join webguest73 [0] (n=436ed320@gateway/web/cgi-irc/
10:07:33 Quit webguest73 (Client Quit)
10:08:22 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
10:13:47 Quit linuxstb (Read error: 113 (No route to host))
10:43:24 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
11:02:27***Saving seen data "./dancer.seen"
11:08:28 Join _zic [0] (
11:39:13 Quit DarkSpectrum (Read error: 60 (Operation timed out))
11:40:32 Join DarkDefender [0] (
11:42:29 Join barrywardell [0] (
11:44:37 Join DarkSpectrum [0] (
11:53:20 Join Lynx_ [0] (
12:07:33 Join Sajber^ [0] (
12:21:14 Join zu [0] (
12:22:09 Quit SUSaiyan (Success)
12:26:26 Join bertrik [0] (
12:29:43 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
12:36:02 Quit DarkDefender (Remote closed the connection)
12:39:11 Join DarkDefender [0] (
12:42:42 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
12:43:36 Join desowin_ [0] (n=desowin@atheme/member/desowin)
12:47:39 Join dash32 [0] (
13:02:28***Saving seen data "./dancer.seen"
13:11:42 Join wincent [0] (
13:12:30 Quit bertrik (Read error: 113 (No route to host))
13:13:44jhMikeSHellooo again. :) Since I've lost sync of late, is there anything going on that blocks FS #10458 from being committed?
13:14:24AlexPhello :)
13:14:56*jhMikeS waves at AlexP
13:15:17AlexPHow is it going?
13:16:28 Join kugel [0] (n=kugel@rockbox/developer/kugel)
13:17:50jhMikeSAlexP: Pretty well all in all, thank you.\
13:20:04AlexPgood stuff
13:21:06kugelpixelma: is that a problem?
13:22:08kugeljhMikeS: I think it lacks a committer :)
13:23:15jhMikeSkugel: Guess it's my charge then ... in it goes :)
13:26:10 Join Lear [0] (i=chatzill@rockbox/developer/lear)
13:27:14CIA-8New commit by 03jethead71 (r22014): Commit FS #10458 by Christopher Williams - Fix incorrect AIFF and WAV encoding while recording in mono.
13:32:21 Quit daurn (Read error: 110 (Connection timed out))
13:32:32 Join SUSaiyan [0] (
13:38:34 Quit Lear ("ChatZilla 0.9.85 [Firefox 3.5.1/20090715094852]")
13:38:49 Quit mt (Read error: 113 (No route to host))
13:40:27 Join daurn [0] (n=daurnima@unaffiliated/daurnimator)
13:52:51 Quit r0b- (Read error: 110 (Connection timed out))
13:53:10 Join r0b- [0] (
14:06:29 Nick fxb__ is now known as fxb (
14:07:37 Quit bmbl ("Bye!")
14:16:52 Join DarkSpectrum- [0] (
14:17:11 Quit TheSeven (Read error: 60 (Operation timed out))
14:18:09 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
14:28:27 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
14:30:01 Quit DarkSpectrum- (Read error: 104 (Connection reset by peer))
14:30:07 Join DarkSpectrum [0] (
14:40:26 Join funman [0] (n=fun@rockbox/developer/funman)
14:44:46funman : answer from AMS about AS3531/Clipv2 datasheet.
14:45:33funmanThey "would like" me to sign a NDA, but this removes the possibility of distributing the datasheet to other developers
14:45:49funmanI am not sure how to explain this problem to them
14:46:58Tornehave they shown you the actual NDA terms?
14:47:51scorcheof course you would have to be permitted to release GPOL code based on it as well...
14:48:05funmanI had received a copy of the terms some time ago
14:48:09 Join courtc_ [0] (n=court@unaffiliated/courtc)
14:48:09Tornethat would be what i was getting at, yah
14:48:23funmanI don't remember seeing anything specific about the GPL
14:48:44Tornewell no, but some people's NDAs consider "implementing all the stuff in the datasheet and publishing it as open source" to be "disclosure"
14:50:12funmanThe parties agree to keep all disclosed confidential information strictly confidential.
14:50:36Tornethat could well be interpreted to mean "don't put all the constants in a header file and put them on the internet"
14:50:41scorcheallowing other developers to see it would be nice, of course, but the most crucial part is to be able to release GPLed code from info on the datasheet, otherwise you could find yourself out of development pretty quickly..
14:50:53funman.. doesn't apply to information developed indenpendently of any such confidential information
14:51:11Tornebut reading the datasheet then writing rockbox drivers based on the datasheet is not independent. :)
14:51:15Torneit depends how arsey they would want to be
14:51:16scorchesure, but legal stuff can be complex...
14:51:25funmanso a header of such constants/registers would be dependent on the datasheet
14:51:42AlexPfunman: I think you really need to have a specific agreement that you can release (GPL) code based on it
14:51:51AlexPfunman: Otherwise legally it is just too murky
14:51:52Torneyah. it gets done for linux sometimes
14:52:05Tornepeople get NDAs with specific terms that do allow code to be released
14:52:10Tornebut still not to share the actual docs
14:52:14AlexPWe already have that from AMS
14:52:18Tornebut that's normally very specific agreements
14:52:18AlexPFor other chips
14:53:01AlexPThe datasheet can't be released, but other devs can see it privately, and the code written based on it can be released under the GPL (otherwise it obviosuly wouldn't be in Rockbox)
14:53:02funmanAlexP: do you mean the as3525 ?
14:53:07AlexPfunman: possibly
14:53:11AlexPI can't remember
14:53:28funmanAlexP: i don't know under which (formal) terms Bagder got the as3525 datasheet
14:53:47AlexPBut either way, I'd say that unless it is specifically agreed that you can release GPL code based on information from the datasheet, then Rockbox can't use it
14:54:00AlexPIt is just too unknown otherwise
14:54:25funmanI wonder if I should use top-posting (like him)
14:54:37scorchei believe it was the AS3514 that we had an NDA for
14:54:54 Quit courtc (Read error: 111 (Connection refused))
14:55:05scorcheas i understand it, we were given the AS3525 free on an NDA but still keep it a bit private to respect AMS
14:55:16scorches/free on/free of
14:55:29AlexPWhich is fair enough
14:55:56linuxstbThat's my understanding - that it was just an informal agreement with Bagder that we could release code based on it, but not distribute publically (just to Rockbox devs who needed it)
14:56:12scorchebut yes...directing them to the NDA which i assume we had for the AS3514 would be a gopod start in procuring another one
14:56:28scorchefunman: is this person aware of our past communications/arrangements?
14:56:43funmanI had mentioned them in the past but I'm going to do it again
14:56:55linuxstbBut I would assume any NDA, unless clauses are negotiated specifically, wouldn't allow code based on that information to be published in source form. Otherwise, what's the point of the NDA?
14:57:11AlexPfunman: Seeing as we are all guessing to an extent, it might be wise to get Bagder involved :)
14:57:15AlexPlinuxstb: yes, I agree
14:57:27scorchelinuxstb: right...without that clause it is too much of a risk
14:57:53AlexPIt just isn't worth it to Rockbox to get any where near that sort of potential mess
14:58:10AlexPEspecially as we have a resonable relationship with AMS
14:59:25funmanperhaps it just make their lawyers happy
14:59:49AlexPWhich is a good thing
15:02:32***Saving seen data "./dancer.seen"
15:03:18funmanscorche: do you remember who had signed the NDA for AS3514 docs?
15:04:29AlexPBagder dealt with it I'm pretty sure
15:06:15kugelfunman: are you guessing that this is the ascodec function or are you sure? What you told me to try yesterday didn't work
15:06:39funmankugel: i am sure this is the ascodec function, but other things might be needed to enable LCD backlight
15:07:58kugelfunman: it looks like it is indeed, but there are so many other calls to it, and I've not managed to find something corresponding in the v1 firmware
15:09:50funmankugel: 0x6CC is ascodec_read_write(reg, set, clear, unknown)
15:10:06funmanand it does : reg = (reg & ~clear) | set;
15:10:16 Quit daurn (Read error: 110 (Connection timed out))
15:10:36kugel0x6cc in the v2
15:10:55funmanthe same function exists in v1 firmware (it is used at least in early init, after clocks init, and before other hardware init -lcd, sd, ..-
15:11:01kugelset/clear as parameters?
15:11:19 Quit DarkSpectrum (Read error: 110 (Connection timed out))
15:11:32kugelI thought r1 was the value, and r2 read/write
15:12:05funmanthe i2c saddr is in r0
15:12:14funmanhm ok
15:12:47funmantrue I always see r2 set to 0
15:13:12 Part n8ature
15:14:07 Join daurn [0] (n=daurnima@unaffiliated/daurnimator)
15:14:19kugelwhere is the function in the v1 firmware? I found something similar, but it doesn't do the cmp r0, #64 check
15:15:02funmankugel: well i'm not sure for the function in v2 firmware : i just called it i2c_write(saddr, byte, unknown1, unknown2)
15:15:53kugelnot sure anymore?
15:15:56funmanin clipv1 of (v29) : at 0x428 , i2c_read_write(saddr, clear, set, unknown)
15:16:30kugelI don't have a clip disassembly here
15:16:34funmanIn fact I remembered this function in v1 OF (for which I am sure)
15:16:50 Join TheSeven [0] (
15:18:39funmanI work on this Clip OF version, and then compare it to the OF for other models, so it became my most disassembled OF
15:18:44 Part LinusN
15:21:06 Join arohtar [0] (n=faemir@
15:21:22 Join darkless [0] (
15:21:43 Join DarkSpectrum [0] (
15:22:35kugelfunman: ah yea, that's the one I suspected
15:22:51kugelbut it doesn't do the r0 > 64 check, any ideas?
15:24:25funmanno, but a saddr > 0x40 is allowed anyway; the registers repeat themselves after 0x40 (0x40 == 0, 0x41 == 1 ..)
15:25:15kugelthat could be why
15:26:19kugelthe v2 firmware just returns 0 for >64 it seems
15:27:01 Quit evilnick ("Page closed")
15:29:42kugelfunman: I can't seem to find where the v1 firmware sets the backlight brightness
15:31:31 Join evilnick [0] (i=0c140464@gateway/web/freenode/x-41b88816c4c921bd)
15:36:33funmankugel: 44A4 in Fuze OF v01.01.11
15:37:38funmanthe last function before returning is i2c_write___(saddr, value, unknown) here
15:37:41 Join PaulJam [0] (
15:38:49 Join Lear [0] (i=chatzill@rockbox/developer/lear)
15:39:55kugelfunman: that calls a different ascodec function though
15:40:14kugel(which I also identified as a candidate), is there indeed more than 1?
15:40:25funmanthere are a lot
15:44:34funmankugel: here I see i2c_read__ i2c_read_write i2c_read___ i2c_write___ i2c_mod_reg i2c_read i2c_write
15:44:47funmani'm not sure anymore what is the difference between all of them ;)
15:50:07 Join AsaelReiter [0] (n=d44c7de0@gateway/web/cgi-irc/
15:50:50 Join carl [0] (n=5950bebb@gateway/web/cgi-irc/
15:52:38carlI'm a rockbox user, that use daily builds. I wanted to pick up the last version aka 22014 for my gigabeat S, but in the server the last built is 21994
15:52:46 Join lu6cifer [0] (
15:52:47 Part lu6cifer
15:53:33carldoes anyone know why it isn't the last here? And if I could fin somewhere else the last build for gigabeat S
15:53:42 Join fdinel [0] (
15:54:42carlI saw lot of commits about the new server system, then maybe that a bug somwhere ???
15:55:52funmanwhere did you see 21994 was the last build?
15:56:04carlthe info.txt
15:56:27carlthe system seems to have stoped to build for those targets since
15:56:47carland of course the date of the file
15:56:54carlthe 21st ;)
15:57:43carlwho manage the build server? maybe there is a bug somwhere
15:58:50carlfunman: any thought?
15:59:41 Quit parafin (Read error: 60 (Operation timed out))
16:00:23carlany command for commiters or admins to build again for all targets?
16:00:47carlsince I saw that the build table seems to be green...
16:00:51funmanon the file I just linked, rockbox-info.txt says r22014
16:00:52n1scarl: the zip from that link is "Version: r22014-090723"
16:01:32funmancarl: perhaps you got a gigabeats build from another page?
16:01:37carloh, then I'll retry, thanks to check with me
16:01:37kugelfunman: no success :)
16:01:56 Join parafin [0] (
16:02:00funmankugel: well it can be a long process, just keep trying!
16:02:32funmanto find how to control clipv2 backlight, I injected some code to return to rockbox from a certain point in the OF
16:02:50funmanI used dichotomy to find where backlight control happened
16:02:59carlfunman: I just substitued the "f" in the gigabeat F download path to "S", then no risk to be wrong :)
16:03:08*carl is retring
16:03:34kugelfunman: what did you do?
16:04:00kugeljust moving your "bl rockbox" until it happened after the backlight?
16:04:09carlit seems to be a cache issue, but I don't know how to proced :(
16:04:30funmanafter and before
16:04:50funmancarl: perhaps there is a proxy between you and rockbox website
16:04:55kugeland it was the function you wanted me to try?
16:05:05funmankugel: just giving you suggestions
16:05:46carlfunman: wee!!! that worked with IE
16:05:47kugelyou had success on the clipv2, which function (call) was it?
16:06:18carlthanks funman and n1s
16:06:49carlhave a good day, and keep working this amazing project
16:06:54 Quit carl ("CGI:IRC")
16:07:06 Quit arohtar (Client Quit)
16:07:23funmankugel: a mix of various functions, including 0x7284 in v02.01.16
16:09:33funmani still have the diff i had used to modify OF with a branch to rockbox, but you'll have to be careful with it.
16:10:02funmanecrips didn't try using more than 320kB of SDRAM on Fuzev2, while on Clipv2 we can use safely at least up to 1MB
16:12:11kugelI think backlight_set_brightness is at 0x18468
16:12:28funmanit's not "dichotomy" but "binary search algorithm" :/ (french is "dichotomie"
16:13:14 Join barrywardell_ [0] (n=barrywar@
16:13:29funmankugel: fuzev2 02.01.17 ?
16:15:21 Nick courtc_ is now known as courtc (n=court@unaffiliated/courtc)
16:16:55kugelI might be lacking some init calls
16:17:06funmankugel: what if you use ascodec_write(0x1b, 0x80|0x1f) ?
16:17:52funmanhaving the list of those i2c registers would definitely help! (i hope we get the doc from AMS!)
16:18:01kugelfunman: I have my stuff directly after system_init(), is that too early?
16:18:38funmanyou must remove bits of system_init
16:18:54kugelwhich bits?
16:19:01funmanlet me attach my synced diff to fs#10047
16:19:19funmanpll/cpufreq setting for example
16:20:55kugelmorpion.c !?
16:20:59funmansdram init / peripheral reset / PLL, CGU_PERI, CGU_PROC settings
16:21:18funmankugel: i wanted something to play with ;) but i removed this game from the last patch
16:21:50kugelI'm right now using the stock fuzev1 code
16:22:21funmancheck what i changed for teh clipv2 and add a FUZEV2 case
16:22:45funmani am still not sure which files we will be able to share with as3525 code
16:23:06 Quit funman ("free(random());")
16:29:29 Quit barrywardell (Read error: 110 (Connection timed out))
16:33:07 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
16:33:49*LinusN needs a temporary user account on a US linux box
16:34:58 Quit Lynx_ (Read error: 104 (Connection reset by peer))
16:35:26 Join Lynx_ [0] (
16:36:57 Join slickrik [0] (n=4dde4054@gateway/web/cgi-irc/
16:37:19 Quit wincent (Read error: 110 (Connection timed out))
16:37:58kugelif I at least would know my code is actually executed :(
16:38:00LearHm, AAC decodes differently on arm target and on x86 sim...
16:39:33linuxstbLear: Different output pcm data?
16:40:25 Quit slickrik (Client Quit)
16:40:28LearYes, I compared test_codec output.
16:41:02 Quit parafin (Read error: 54 (Connection reset by peer))
16:41:30 Join parafin [0] (i=parafin@2001:470:1f0b:81:0:0:0:1)
16:42:27kugelbacklight on fuzev2
16:46:14 Join PaulPosition [0] (
16:48:58 Quit PaulJam (Nick collision from services.)
16:49:04 Join PaulJam_ [0] (
16:51:18 Quit kkurbjun ("Leaving.")
16:52:54PaulPositiona few weeks ago, cabbiev2 wps was slightly changed on the (clip?) target to show numbers (db?) when changing up/down. Is that desired behavior for all targets, an accidental commit or a case of "..and then I won't have to patch every time, who cares anyway?" ?
16:54:53Learlinuxstb: Out to be a rounding thing; the arm code doesn't appear to do the "+ (1 << (SCALEBITS - 1))" thing the C implementation of the MUL macros does.
16:55:58linuxstbLear: Ah, is that our optimisation, or part of libfaad?
16:57:45Learlinuxstb: I think that's our code; bits of it was taken from libmad.
16:58:38Learlinuxstb: Or maybe not; it was part of the initial checking, said to be "unmodified libfaad".
16:58:51linuxstbWould you say it needs fixing? I guess it means we're losing one bit of accuracy?
16:58:56 Join Xerion [0] (
16:59:31linuxstbWould be interesting to see if current libfaad has changed it...
17:02:35***Saving seen data "./dancer.seen"
17:03:31 Quit parafin (Read error: 60 (Operation timed out))
17:03:58LearNot important. Could be nice for testing/verification perhaps. But I just tried removing the adds from the C code, which is what the sim ought to use, but there are still differences, and they aren't one bit differences...
17:05:30 Quit AsaelReiter ("CGI:IRC (EOF)")
17:07:09 Quit gevaerts (Nick collision from services.)
17:07:19 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
17:08:34 Quit petur (
17:08:34 Quit einhirn (
17:08:34 Quit mortti (
17:08:34 Quit goffa (
17:08:34 Quit bubsy (
17:08:34 Quit Erant (
17:08:44NJoinpetur [50] (n=petur@rockbox/developer/petur)
17:09:32 Join goffa_ [0] (n=goffa@
17:09:34 Nick YpsyZNC is now known as Ypsy (
17:09:42NJoinmortti [0] (
17:11:07 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
17:13:25 Join DarkSpectrum [0] (
17:15:08 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
17:16:31 Join DarkSpectrum [0] (
17:16:58 Quit dash32 (Remote closed the connection)
17:19:40 Quit PaulJam_ (Nick collision from services.)
17:19:46 Join PaulJam [0] (
17:20:23CIA-8New commit by 03learman (r22015): Fix possible null pointer dereference in cuesheet handling. Causes problems in the simulator.
17:20:39 Quit Sajber^ (Read error: 54 (Connection reset by peer))
17:20:42JdGordon1PaulPosition: bit of 2 and 3
17:20:57JdGordon1i mean, 1 and 3
17:23:31 Quit TheSeven (Nick collision from services.)
17:23:46 Join The_Seven [0] (
17:23:50 Nick The_Seven is now known as TheSeven (
17:24:46 Join toffe82 [0] (n=chatzill@
17:26:19 Join Erant [0] (i=erant@
17:27:17 Join jgarvey [0] (
17:32:01saratogaLear: theres a lot of differences between the c and arm versions beyond just the rounding, the IMDCT uses different instruction ordering, the multiply by a trig value macro is probably less accurate, etc
17:32:33 Join bubsy [0] (i=Bubsy@
17:32:33 Join CaptainKwel [0] (i=2669ecc2@gateway/web/freenode/x-9539ef3d508dd745)
17:32:44saratogawhen trouble shooting I usually subtract the two wav files and plot the difference
17:32:59saratogaif theres a real problem it will be obvious
17:33:58Learsaratoga: did you see my message about the "ticking" I got with an AAC file?
17:34:27 Join dash32 [0] (
17:39:46 Join domvamp [0] (n=55250502@gateway/web/cgi-irc/
17:40:07domvampdoes anyone know something about rockbox for ipod 6th gen ?
17:41:12linuxstbThere's no such thing...
17:41:23kugelLear: IMO cuesheet_subtrack_changed() doesn't behave correctly if it can't handle NULL
17:42:15domvampthanks...but someone is working on it or it's "impossible"
17:43:05JdGordon1kugel: ? it does work correctly
17:43:34saratogaLear: did you see my reply to it
17:43:44Learkugel: Yes, perhaps. But I did the smaller change. :)
17:43:45kugelJdGordon1: it does id3->cuesheet without checking if id3 is NULL
17:44:00JdGordon1id3 should never be null seems no one..I'll wait and check periodically on the site....thanks again to all....regards
17:44:45LearJdGordon1: Can happen in the sim when starting playback.
17:44:45 Quit domvamp ("CGI:IRC (EOF)")
17:44:49kugelshould sounds a bit loose
17:45:04kugelcan also happen on hwcodec IIRC
17:45:05JdGordon1ok, it must never be null at that point
17:45:16JdGordon1Lear's commit is good
17:46:05kugelok, so it must not, but it is, where's the real bug then?
17:46:40JdGordon1thats what I was thinking
17:46:54JdGordon1gui_wps_update is getting called while id3 is null which it shouldnt
17:47:56JdGordon1back in 15
17:52:09 Part LinusN
17:52:55saratogaLear: do you have example output with and without the clicking?
17:53:21 Join BryanJacobs [0] (
17:57:00 Quit darkless ("Leaving")
17:58:34 Quit Grahack ("Leaving.")
17:59:24 Quit linuxstb (Connection timed out)
18:01:29 Join lyngaas [0] (
18:02:00 Join JdGordon| [0] (i=ad75d7aa@gateway/web/freenode/x-7d20054cbdcfd5a2)
18:02:07Learsaratoga: Yes, 30 sec clips in flac, a few megs each. One is recorded output, the other is test_codec output.
18:02:52LearIt is subtle though; I've only heard it with headphones.
18:04:07TorneJdGordon|: is your patch for FS5230 (hold bug on ipod) supposed to be instead of the one dreamlayers proposed?
18:04:32JdGordon|umm, either or i tihnk
18:04:43JdGordon|I didnt realise he had a patch for it
18:04:50JdGordon|im not sure which is better
18:05:09Torneyours *looks* nicer
18:05:28Tornehis just has some outl() with mystery hex numbers :)
18:05:31Tornebut that doesn't mean a lot :)
18:05:33JdGordon|hardly mine, linuxstb told me what to add and I added it :p
18:05:49Tornewell, yah, i'v ebeen running with th epatch already on there ipod_4g_button_int.patch
18:05:57Torneand it fixes the problem on my 5.5g
18:06:21Tornei'll try the patch you posted, though
18:08:58 Join n00b2Hack [0] (n=Administ@
18:10:27 Quit _zic (Remote closed the connection)
18:16:12JdGordon|kugel: I didnt see an updated custom vp patch?
18:17:26JdGordon|arg, if someone is bored.. care to try and see why audio_current_Track() is being called twice per second rom the WPS? it shold have clicked last night that that was bad
18:17:46 Quit petur ("work->home")
18:17:59 Quit PaulPosition ()
18:19:25 Quit PaulJam (Nick collision from services.)
18:19:30 Join PaulJam_ [0] (
18:23:14LearJdGordon: Is it? Doesn't look like it (looking at the code only)...
18:23:18kugelJdGordon|: want to see one?
18:25:47JdGordon|Lear: debbuging the cuesheet last nightsomething was triggering it every .5s or so.. just had a quick look and yeah i cant see it either
18:26:00JdGordon|kugel: if we want to get it commited soonish.. yeah :)
18:28:10LearJdGordon: Well, I don't get it in the sim.
18:29:06JdGordon|maybe it only happens with a .cue loaded
18:29:19JdGordon|in whcihc case its not soo bad
18:29:21 Join bertrik [0] (
18:30:23LearJdGordon: Could be via cue_draw_markers, when the progressbar is drawn.
18:30:26JdGordon|ah yes! twice in cuesheet.c
18:31:49 Quit PaulJam_ (Nick collision from services.)
18:31:52JdGordon|its not really an expensive call anymore so maybe not worth worying about
18:31:57 Join PaulJam__ [0] (
18:34:53 Join _zic [0] (
18:35:10 Quit n00b2Hack ("Saliendo")
18:38:04 Quit _zic (Remote closed the connection)
18:38:36 Quit JdGordon| (Ping timeout: 180 seconds)
18:39:35 Join benutzer [0] (
18:41:40 Join _zic [0] (
18:43:09 Join Hillshum [0] (n=hillshum@unaffiliated/hillshum)
18:44:27 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-ef614237ff50ae8d)
18:46:04 Quit DarkSpectrum (Read error: 104 (Connection reset by peer))
18:51:47 Quit desowin_ ("KVIrc Insomnia 4.0.0, revision: 3332, sources date: 20090703, built on: 2009/07/12 16:20:08 UTC 3332")
18:53:35 Part toffe82
18:58:40 Quit jhMikeS (Read error: 104 (Connection reset by peer))
18:58:45 Quit Lear ("ChatZilla 0.9.85 [Firefox 3.5.1/20090715094852]")
19:02:38***Saving seen data "./dancer.seen"
19:08:37 Quit DarkDefender ("Leaving")
19:21:49 Join parafin [0] (
19:23:50 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
19:25:24kugelJdGordon|: there it is
19:26:27 Nick fxb is now known as fxb__ (
19:29:15 Quit agaffney (Read error: 110 (Connection timed out))
19:30:42 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:31:41 Join drakonik [0] (
19:31:49 Quit barrywardell_ ()
19:33:35drakonikSo, is it normal for Rockbox to not acknowledge a keypress for 5 to 10 seconds?
19:34:01Hillshumdrakonik: I don
19:34:08Hillshum't think so
19:34:23*Hillshum kicks his enter key for moving to the left
19:35:05HillshumAll I have are AMS Sansas but they are quite snappy
19:35:13drakonikIt varies.
19:35:19Hillshumdrakonik: what target do you have?
19:35:26drakonikIpod video 30gig
19:38:18 Join Ubuntuxer [0] (
19:38:42GodEater5 to 10 seconds is not normal for any rockbox target
19:38:50drakonikI figured.
19:39:47ej0rgedrakonik: What's it doing? if it has to wait for the drive to spin up and then read a directory, 3 or 4 seconds may be normal in some conditions - if you don't have the directory cache enabled for example
19:40:21ej0rgeor if the drive is failing
19:40:22 Join timc [0] (
19:40:30ej0rgestill assuming a hard drive rather than flash
19:40:33drakonikej0rge: As far as I can tell, it's not waiting for the drive to spin up. It's highly inconsistent, so I can't really test it for certain
19:40:37drakonikRight, it's a hard drive
19:42:14PaulJam__drakonik: try if it gets better when you enable directory cache.
19:42:25drakonikHm. I'll try that.
19:43:42ej0rgeor database cache if you're using the database
19:43:50ej0rge("load to RAM" option)
19:43:58drakonikDon't get me started on what's broken in the database
19:44:05drakonikBut yeah, I'm using the database and it is loaded to RAM
19:44:08ej0rgehey, i don't even use it
19:44:27drakonikOh, the database in general works fine. But on my specific ipod, it's broken.
19:45:20 Quit PaulJam__ (".")
19:45:38CIA-8New commit by 03Ubuntuxer (r22016): Fix wrong text colour in help, adapt pegbox manual to r21867
19:49:34drakonikAnyway. The DB is already cached to RAM when the keypresses are lagging so badly.
19:49:47 Join PaulJam [0] (
19:49:49 Quit Ubuntuxer ("Leaving.")
19:49:52ej0rgedrakonik: have you tried checking the filesystem?
19:50:00drakonikChecking how? like chkdsk?
19:50:51 Join efyx_ [0] (
19:53:01drakonikEh, might as well.
19:54:29drakonikOh. Well.
19:54:40drakonikSomething about 'lost chains'.
19:56:13drakonikWhy does that keep coming up? No matter how many times I reformat the thing, it keeps having these 'lost chains'.
19:56:51HillshumPossibly a disk issue then
19:59:18 Quit kugel (Remote closed the connection)
19:59:59drakonikOh. That's awesome.
20:00:23drakonikAnd the battery guesser is broke to hell too
20:00:59gevaertsdrakonik: do you always cleanly unmount the device before unplugging?
20:02:51 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
20:07:14 Join webguest27 [0] (n=59cccaa5@gateway/web/cgi-irc/
20:08:00drakonikInteresting. Whenever it reboots, the lost chains are back.
20:08:34drakonikNow to find out whether the ipod firmware is causing it, or if Rockbox is
20:10:15webguest27hiya guys and girls is anyone around that can help me?
20:11:09drakonikJust ask.
20:11:12 Quit webguest27 (Client Quit)
20:11:17drakonikDon't ask if you can ask, just ask the question.
20:11:29 Join webguest39 [0] (n=59cccaa5@gateway/web/cgi-irc/
20:13:14drakonikRockbox seems to think that a full battery instantly drains down to zero after a reboot
20:13:25webguest39hello girls and guys
20:13:30bluebrotherhello webguest39, I'm your friendly answering bot. How may I help you?
20:13:35 Join Zagor [242] (n=bjst@rockbox/developer/Zagor)
20:14:55bluebrotherZagor: the new build system doesn't update the build-info file rbutil uses anymore
20:15:09bluebrotheror at least I assume it's the build systems fault :)
20:15:12Zagorbluebrother: ah, there you are. which file is that? where is it located?
20:15:44Hillshumwebguest39: Just ask. Anyone who knows the answer will answer for you
20:17:02Zagorbluebrother: do you use the timestamp or just the revision?
20:17:38webguest39yes its one of those questions)...has anybody else managed to get rockbox installed on on?
20:18:02 Join Thundercloud [0] (
20:18:46Hillshumwebguest39: most of us here have installed rockbox on at least one DAP
20:18:49bluebrotherZagor: both is used.
20:19:13 Join dys [0] (
20:19:28 Quit r0b- ("( :: NoNameScript 4.02 :: )")
20:19:37Hillshumwebguest39: I personally have it on two
20:20:05webguest39GoogleChrome and irc dont mix i think..i meant to say a tevion mp4 player
20:20:23LambdaCalculus37No, only players listed on the front page work with Rockbox.
20:21:41 Join r0b- [0] (
20:22:24HillshumPlus a few more that almost work
20:22:35 Quit flydutch ("/* empty */")
20:23:09GodEaterHillshum: please don't complicate it :)
20:24:54webguest39i suppose if i could find out how to get at the bootloader i could compile rockbox myself maybe? i have a chainsaw?
20:25:05 Join darkless [0] (
20:25:25bluebrotherwebguest39: no. You need to port it to the hardware, which means you need to write drivers for it.
20:25:46bluebrotherunless you can reuse existing drivers. Plus, you need the Rockbox bootloader running on it
20:27:11LambdaCalculus37And also finding out how firmware updates are performed on the platform, so you can get an idea of how to inject custom code onto it.
20:28:05 Quit drakonik ("My landmass erupts with kittens.")
20:28:18GodEaterwebguest39: most rockbox ports take at LEAST a year to get anywhere near ready, and most take longer
20:28:29HillshumLambdaCalculus37: I think that was meant by "get at the bootloader"
20:30:51bluebrotherwebguest39: there's a wiki page with the most important questions about porting. Look for the NewPort wiki page
20:31:44 Join kachna [0] (
20:31:44 Quit webguest39 ("CGI:IRC (Ping timeout)")
20:32:10Zagorbluebrother: done
20:32:29 Join aaron424 [0] (
20:32:58bluebrotherZagor: thanks :)
20:33:59 Quit aaron424 ("ChatZilla 0.9.85 [Firefox 3.0.12/2009070811]")
20:35:45bluebrotherZagor: can you change "revision" back to "rev"? ;-)
20:36:58bluebrotherthanks again, now rbutil shows the correct values again :)
20:37:35 Join roolku [0] (
20:37:51 Quit dash32 ("Verlassend")
20:38:32roolkuZagor: could we get timestamps back in the build table (in addition or instead of the revision)?
20:39:00Zagorin addition, yes. not instead. revision is the "real" data, timestamp merely chrome
20:39:57Zagorroolku: you run an impressive number of clients btw :)
20:40:18roolkuZagor: they don't seem to contribute much to the overall time though :(
20:41:16Zagorroolku: oh they do contribute. and with my new algorithm they will even more.
20:42:05Zagorwe can't all run 300+ monster servers. 5x60 p/s clients is nearly as good.
20:43:05roolkuZagor: distribution is probably a real problem - I noticed one client had all of its builds canceled once
20:43:21 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
20:43:37Zagoryes, it happens to one or two clients almost every round.
20:44:50Zagorthe best-fit fixes that, and many other inefficiencies. once it works, that is :-)
20:46:41Zagorthe basic distribution is done. I'm currently working with dealing properly with clients that perform worse than expected. (trigger a recalculation of build allocations)
20:48:36Zagorone interesting thing is that I cannot even wait for clients to finish their first build before determining if they are running slower than expected.
20:48:43roolkuZagor: I suspect it will trigger quite often? Depending on load and network congestion?
20:50:01roolkuZagor: maybe a quick benchmark just before a build round? Or does it waste too much time?
20:50:56Zagorroolku: we already have benchmarks in the form of past performances. variations are due to momentary variations in system load, which a benchmark won't show either.
20:51:15Zagorwe simply have to deal with it
20:51:49roolkuZagor: that's why I suggested benchmarks immediatly before the run, but I guess there are still fluctuations
20:52:34Zagoryes, it will always fluctuate. we simply have to adjust to it.
20:53:05roolkuZagor: Zagor: I considered experimenting with distcc since the machines are on a reasonably fast network, but the repository only has an old(ish) version and I am not sure if it worth the bother to manually install it
20:53:17roolkuZagor: also it would make the tree dirty
20:54:08Zagoras I understand it the machines generally aren't on fast enough networks to make distcc efficient.
20:54:34Zagorwe currently upload a fraction of what distcc does, and it still takes up considerable time
20:55:40roolkuZagor: the b?? machines should be. I would have one rbclient running that distributes the compiles to the other local nodes
20:55:53Zagorah, that could be interesting!
20:56:26JdGordon|is the code parallel enough for the many clients to do one build simultanously?
20:56:38JdGordon|.. or is gcc that good?
20:56:57roolkuJdGordon|: I guess only a test would show
20:58:48roolkuJdGordon1: on my home pc (core i7) I get the fastest buildtimes with -j 16 - not sure if this is comparable though - it seems the longest time is spent generating dependencies which is sequential :(
20:59:11Zagoryes the depgen is a bit of a tar pit
21:00:32Zagorit's a bit tricky to parallellize too
21:00:48HillshumCould it be cached
21:02:03gevaertsnot easily
21:02:39***Saving seen data "./dancer.seen"
21:11:48 Quit bluebrother (Nick collision from services.)
21:11:52 Join bluebroth3r [0] (n=dom@rockbox/developer/bluebrother)
21:13:54 Nick bluebroth3r is now known as bluebrother (n=dom@rockbox/developer/bluebrother)
21:16:07ej0rgeroolku: I haven't fiddled with -j building rockbox on my quad-core, but i observed that i can build linux in 2.5 minutes with -j6, while building on a single core takes almost 18 minutes.
21:16:40*gevaerts claims that this is impossible
21:16:55ej0rgegevaerts: just reporting what i observed
21:17:18ej0rgeroolku: Not to my knowledge
21:17:20gevaertsej0rge: ccache, VFS cache, or magic are involved
21:17:30ej0rgegevaerts: Could be
21:17:57 Quit PaulJam (Nick collision from services.)
21:17:57ej0rgethis was "time make -j6" vs "time make" - both after "make clean"
21:18:02 Join PaulJam_ [0] (
21:18:12ej0rgethere were slight differences in the configuration of the kernel, but not huge differences.
21:18:42gevaertsej0rge: well, my bet is VFS cache, with -j6 last
21:20:15ej0rgegevaerts: I did -j6 first
21:20:28gevaertsin that case, magic
21:20:32ej0rgeI'll have to try some more controlled tests
21:21:52 Quit bluebrother ("leaving")
21:23:47ej0rgegevaerts: I've recompiled the kernel several times on my new rig, mostly due to having picked some wrong options due to having made the wrong assumptions about what chips are really under the board-maker labels on the motherboard. I timed several of them and got the ~150 second result on most builds, with -j6 being a few seconds faster than -j4. Often rebooting between builds, but not always. I did a 'make clean' for every timed build because i als
21:24:16ej0rgegevaerts: one time, I neglected to include the -j argument, and it took a very long time to compile. That's pretty much what i know right now.
21:24:36saratogadisk access time seems to play a big part, its possible rebuilding it gave you more file cache hits
21:25:09ej0rgesaratoga: possibly. And I'm not sure how to assure that the files aren't cached.
21:27:49ej0rgeoutside of rebooting fresh, i mean, which is out of the question right now since i'm actually using it
21:30:35 Join faemir [0] (n=faemir@
21:35:25 Join To [0] (
21:36:07 Join Asa_Thor [0] (
21:37:23 Quit Asa_Thor ("ChatZilla 0.9.85 [Firefox 3.0.12/2009070609]")
21:43:50 Quit PaulJam_ (Nick collision from services.)
21:43:54 Join PaulJam [0] (
21:45:51 Join Froopie [0] (n=thetechz@
21:46:14Froopieis rockbox coming to the ipod nano 3 gen any time soon?
21:46:35GodEaterFroopie: nope
21:46:40Froopieoh well
21:46:51Froopielater guys.
21:46:53 Part Froopie
21:47:46bertrikgevaerts, the USB DFU spec doesn't really say what file exactly is to be sent to the device, right?
21:48:58bertrikgevaerts, the meizu_dfu tool sends some commands to determine CPU type, is that part of the DFU spec?
21:49:17gevaertsI'm not sure, but I think so
21:49:21bertrikis the meizu_dfu tool a true USB DFU compliant tool at all?
21:50:00gevaertsit was written to clone the recovery tool that's mentioned on somewhere, so probably not
21:51:43bertrikI could temporarily remove the actual upload part from meizu_dfu and see what happens when talking to the yp-s3
21:52:38bertrikas far as I understand, the tools first sends a small flash-writer program (the SST... file), which is then executed and runs the rest of the firmware downloading and flashing process, right?
21:53:43gevaertsyes, although with some tweaking of the .lds file, you can build directly runnable code as well
21:54:09bertrikhave you actually sent and run code like that?
21:55:14bertrikcool, that would be a nice and probably safe way to start hacking
21:55:36gevaertshave a look at the history for s5l8700/
21:55:40bertrikto actually write my own firmware to the device, I guess I need the appropriate "SST" file for this player
21:57:17TheSevenhey, audio working on nano2!
21:57:21gevaertswell, you can probably also flash it using OF upgrade procedures. Probably only once though
21:57:25bertrikTheSeven, nice!
21:57:28 Join wincent [0] (
21:57:43gevaertsTheSeven: meizu is still further along though! You don't have radio yet!
21:58:01TheSevengevaerts: well, there is no such thing as a radio :-P
21:58:07gevaertsnot my fault :)
21:58:28bertrikTheSeven, was the rockbox code any use in getting audio to work?
21:59:09 Quit LambdaCalculus37 ()
21:59:24TheSevennope, i haven't even tried to understand that after i saw that lot of changes linuxstb had done to make it output anything at all, thereby damaging all the other targets
21:59:31TheSeveni've written that from scratch
21:59:47gevaertsbertrik: I suspect the bit about interrupts not working when using DFU is a bit of a red herring. Copying the interrupt vectors to the right place after booting shouldn't be too much work
22:01:29 Quit coney (Read error: 60 (Operation timed out))
22:01:32 Join coney [0] (n=james@2001:470:8:8b:203:47ff:fedf:1267)
22:02:26bertrikgevaerts, I'll try the get_cpu first
22:03:51bertrikgevaerts, the USB VID/PID is the same as the meizu m6sl
22:04:31gevaertsthat sounds promising
22:06:29bertrikit hangs on the GET CPU step
22:07:42bertrikdmesg says usbfs: USBDEVFS_CONTROL failed cmd meizu_dfu rqt 161 rq 255 len 63 ret -110
22:09:28bertrikI commented out the usb_mimic_windows part though, I'll try to enable that
22:18:08TheSevenbertrik: i guess linuxstb will be happy, though, to have a look at my code tomorrow and check what he is doing differently
22:20:36 Quit PaulJam (Nick collision from services.)
22:20:37 Join PaulJam_ [0] (
22:22:34pixelmakugel: is what a problem?
22:24:46 Quit BryanJacobs ("Java user signed off")
22:24:55 Quit PaulJam_ (Nick collision from services.)
22:24:55 Join PaulJam [0] (
22:25:07 Quit n1s ("Lämnar")
22:25:13 Join IuDeX [0] (n=52a0f8f7@gateway/web/cgi-irc/
22:26:10IuDeXIS there something to test on Clip v1? ( I want to help ;p)
22:27:15 Quit PaulJam (Nick collision from services.)
22:27:22 Join PaulJam_ [0] (
22:27:29IuDeXAny patch or something like that ;)\
22:28:45 Quit kachna (Remote closed the connection)
22:32:49 Join petur [0] (
22:33:22IuDeXwhoops, so I dont see any ideas or something WHY playbacks stop on clip ;/
22:34:30GodEaterI imagine if anyone *knew* why, they'd have fixed it already
22:35:12 Quit jgarvey ("Leaving")
22:36:42 Join TheEvilPrince [0] (
22:37:01*Hillshum had a file that kept stopping predictably on his clip. Now the thing won't play anything
22:37:15*Hillshum should really post details online
22:38:22TheEvilPrinceHi there. Is anyone able to help me regarding a problem with my iAudio X5 please?
22:38:38IuDeXHillsum do do you have fix?
22:38:49 Quit _zic (Read error: 60 (Operation timed out))
22:39:05HillshumIuDeX: for the "won't play anything part" or the general crashes?
22:40:29IuDeXHillsum: My Clip crashes at 30sec on all songs. (is it answer to your question? my English isnt very well)
22:41:43gevaertsTheEvilPrince: it's usually a good idea to actually state your problem. People don't know if they will be able to help if they don't know the problem...
22:41:57TheEvilPrinceI have just installed the lastest version of Rockbox for the first time on my X5. I added a .flac and a .mp3 to the MUSIC directory. Both do not play. They have on occassions played a minute or so after I pressed play...
22:42:19TheEvilPrinceIt basically freezes without playing anything
22:43:02HillshumIuDeX: I was asking if you meant "Do you have a fix for the general problem of frequent crashing?" or "Is your Clip now able to play songs again?"
22:44:17Unhelpfulyay, asteroids:
22:44:19IuDeXHillshum: No, my Clip cant resume songs again. I have to restart.
22:45:02HillshumIuDeX: I really don't know what's causing the problem.
22:45:17TheEvilPrinceI deleled the .tcd files as recommeneded in a forum thread, but to no avail
22:46:12gevaertsTheEvilPrince: have you checked the filesystem?
22:46:32TheEvilPrinceWhat is that?
22:47:12 Quit IuDeX ("CGI:IRC")
22:50:09bertrikTheSeven, I hope so, we tried to do things too much according to the datasheet perhaps instead of looking in the original firmware. The s5l8700 has quite a few errors unfortunately
22:50:44evilnickTheEvilPrince: What OS are you running? Windows, linux, Mac, etc.
22:54:29evilnickOK, so connect the DAP to your PC
22:54:44TheEvilPrinceThe .flac and .mp3 work fine with the cowon firmware
22:55:42evilnickWhat drive letter is it?
22:57:05evilnickCool. Click on the Start Menu, then Run...
22:57:18evilnickType in cmd and click Okay
22:57:49evilnickThen in the command prompt (black old-skool looking window)
22:57:53evilnickType in:
22:58:16dz"black old-skool looking window"... you mean what I'm typing in right now? :P
22:59:13evilnickchkdsk H: /F
22:59:50 Join aaron424 [0] (
23:01:08TheEvilPrinceMicrosoft Windows XP [Version 5.1.2600]
23:01:09TheEvilPrince(C) Copyright 1985-2001 Microsoft Corp.
23:01:09TheEvilPrinceC:\Documents and Settings\User>chkdsk H: /F
23:01:09DBUGSent KICK TheEvilPrince to server
23:01:09TheEvilPrinceThe type of the file system is FAT32.
23:01:09TheEvilPrinceChkdsk cannot run because the volume is in use by another
23:01:09***Alert Mode level 1
23:01:09DBUGsent MODE #rockbox +b *!*n=theevi*@*
23:01:09TheEvilPrinceprocess. Chkdsk may run if this volume is dismounted first.
23:01:09Kick(#rockbox TheEvilPrince :No flooding!) by logbot!n=bjst@rockbox/bot/logbot
23:01:09***Alert Mode level 2
23:01:09Mode"#rockbox +b *!*n=theevi*@* " by logbot (n=bjst@rockbox/bot/logbot)
23:01:27GodEaterlogbot BANS for that ?
23:01:36dzI should hope it's a knockout
23:01:39Mode"#rockbox +o scorche " by ChanServ (ChanServ@services.)
23:01:49Mode"#rockbox -b *!*n=theevi*@* " by scorche (n=scorche@rockbox/administrator/scorche)
23:01:57 Join TheEvilPrince [0] (
23:01:58HillshumDidn't Bagder say logbot is hardened from old battles?
23:02:25GodEaterTheEvilPrince: next time use a pastebin :)
23:02:29Mode"#rockbox -o scorche " by ChanServ (ChanServ@services.)
23:02:42***Saving seen data "./dancer.seen"
23:03:01dzTheEvilPrince: eject the drive in explorer
23:04:54TheEvilPrinceNo problems detected
23:06:23 Quit jfc (Read error: 104 (Connection reset by peer))
23:07:26CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
23:10:01TheEvilPrinceSo any idea why playback is freezing up?
23:10:29krazykitbecause it's not a completed port and there are still bugs to be worked out?
23:11:00ej0rgekrazykit: the X5 has been stable for a while now
23:11:10***Alert Mode OFF
23:11:18ej0rgeTheEvilPrince: you're using a current build of rockbox or a release?
23:11:19krazykitoh, i thought he was on about the fuze. wrong person i guess
23:11:41 Join fdinel [0] (
23:11:47TheEvilPrinceI've tried the release
23:12:24TheEvilPrinceI also tried the current build. I assume I was correct in just copying the current build files over and replacing the release files?
23:12:28ej0rgeTheEvilPrince: I've been using r21947 on my X5L for a few days
23:12:42ej0rgeTheEvilPrince: Correct
23:13:04TheEvilPrinceSo replace the .rockbox folder in its entirety
23:13:19ej0rgeTheEvilPrince: So, it works normally with the OF, but not with rockbox?
23:14:34ej0rgeTheEvilPrince: have you tried playing tracks that were playable in the OF? (I realize this is a dumb question)
23:14:56 Join jfc [0] (
23:15:07TheEvilPrinceYes :D
23:15:16TheEvilPrince1 .flac and 1 .mp3
23:15:16 Quit darkless ("Leaving")
23:15:24TheEvilPrinceBoth play in OF but not Rockbox
23:16:11ej0rgeand both 3.3 and the current build have the same problem?
23:16:25ej0rge(just making sure we understand)
23:17:29ej0rgeTheEvilPrince: out of curiosity, maybe something in the rockbox config is messed up. Delete the entire .rockbox folder and unpack the current build there again?
23:17:34TheEvilPrinceWith the new build. I only copy the .rockbox directory. Not the rockbox-iaudiox5 ?
23:17:58TheEvilPrinceOk I will try that
23:18:40ej0rgerockbox.iaudio should be in the .rockbox folder iirc
23:19:57TheEvilPrinceIt is
23:20:30TheEvilPrinceIs there supposed to be anything in the folders FIRMWARE and _SYS ?
23:20:44ej0rgerockbox doesn't care about those folders
23:21:19 Part wincent ("Kopete 0.12.7 :")
23:22:57 Quit evilnick ("Page closed")
23:24:46TheEvilPrinceJust to add all settings are factory settings
23:25:11HillshumRockbox didn't come from a factory!
23:25:32TheEvilPrinceLol, you know what I mean ;)
23:25:44TheEvilPrince"Standard" settings
23:25:48gevaertsTheEvilPrince: do other files play correctly?
23:26:14TheEvilPrinceOn the OF yes, but nothing will play in RB
23:29:14 Join soap-work [0] (n=42c07542@gateway/web/cgi-irc/
23:29:27soap-workWhat style tags do these files have?
23:29:35soap-work(the X5 problem)
23:29:48bertrikgevaerts, linux thinks the descriptor of the yp-s3 in DFU mode is bad, does it say the same for the M6SL?
23:29:58bertrikI get ** UNRECOGNIZED: 09 21 03 0a 00 00 08 00 01
23:30:05bertrikwhen doing an lsusb -v
23:30:12gevaertsI think it says that for lots of devices
23:31:29 Quit Rob2223 ()
23:31:50 Join Rob2222 [0] (
23:34:34TheEvilPrinceWhat tags should they have?
23:35:43TheEvilPrinceAnd is there a program which can tag this way?
23:36:05soap-workID3 tags on the MP3 (many people have ID3v1 and APE tags on MP3) and vorbis comments on FLAC (again, APE is not uncommon, nor is ID3)
23:36:52 Join jae [0] (
23:37:08soap-workThe only reason I mention that is if they play in the OF and not Rockbox I'm suspecting something Rockbox is less tolerant of than the OF, like tagging to standards.
23:37:41CIA-8New commit by 03peter (r22017): Another pdbox patch by Wincent Balin (FS #10416): switch to using TLSF as memory allocator. Probably the last patch I commit for him, next changes are ...
23:39:04TheEvilPrinceOk soap-work. Is there a program which can tag vorbis exactly how Rockbox wants?
23:40:14gevaertswait. Is it vorbis or flac?
23:40:56TheEvilPrinceFlac files
23:43:27soap-workIt would be a non-standards compliant tagger which /wouldn't/
23:43:49soap-workI would like to think any decent tagger would tell you what style tags the files have currently.
23:44:22soap-workbut I'm not home so I can't say which program I use which I know shows said info.
23:44:47TheEvilPrinceI used Exact Audio Copy when ripping some CDs to .flac
23:46:00HillshumTheEvilPrince: mp3tag will let you use multiple tag types
23:46:12HillshumYou're on XP right?
23:46:53*gevaerts finds
23:47:05 Quit petur ("Zzzzz")
23:47:15gevaertsthat sounds like your problem
23:47:39TheEvilPrincei am
23:49:19 Quit efyx_ (Remote closed the connection)
23:51:09soap-workdoes not explain the mp3 issue.
23:52:06soap-workand I rip to FLAC with EAC, add replaygain with FB2K, ahhhhh but I play the MP3s I transcode, not the FLAC. I can test that EAC thing when I get home.
23:53:02TheEvilPrinceOk :)
23:53:50 Quit ender` (" There are two major products that come out of Berkeley: LSD and UNIX. We don't believe this to be a coincidence. -- Jeremy")
23:54:36TheEvilPrinceBut if I remove the tags and just have a filename the .flac should work, right?
23:55:41TheEvilPrinceOk, just got an MP3 working but it started to play a minute or so after I hit play
23:57:07 Quit Lynx_ (" HydraIRC -> <- Like it? Visit #hydrairc on EFNet")

Previous day | Next day