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 2008-05-04

00:00:23Bagderand yes I agree it sounds like the AMS mode
00:00:25linuxstbAre you sure that's a v2 for that price? There are a lot of refurbished m2x0s around, and at least the one I bought was a v1.
00:00:41linuxstbBut for that price, you could risk it anyway...
00:01:00 Join fyre^OS [0] (
00:01:18toffe82BigBambi: it was done via I pod first and via caddy after
00:01:27saratoga22the V2s apparently came out 18 months ago for the M series, so I doubt theres too many refurb V1s still around
00:02:01 Join CyBergRind|w [0] (n=cbr@
00:02:23linuxstbMaybe I was just lucky then... (I wanted a v1).
00:02:55saratoga22hmm the new ones are only $20, so that'd make more sense
00:03:48linuxstbWhat capacity are the ones for that price? I paid 20GBP (almost $40) for my refurbished m260 (4GB) in the UK.
00:03:57saratoga221 GB
00:04:06 Join Synergy6 [0] (
00:04:12linuxstbOK, I feel better then ;)
00:04:12 Quit Xerion (Read error: 104 (Connection reset by peer))
00:04:15saratoga22perfect for bricking
00:04:40 Join moos [0] (n=c40cd998@gateway/web/cgi-irc/
00:04:46 Join Xerion [0] (
00:05:12moosjhMikeS: ping
00:05:19toffe82saratoga22: will you bid on the S30 , the one finishing in 2H ?
00:06:23 Quit DerDome ("Leaving.")
00:07:44saratoga22i already did :)
00:07:51 Quit bluebrother ("Verlassend")
00:09:26toffe82ok, just to don't bid on this one ;)
00:10:21saratoga22were you looking to get another one?
00:14:56 Quit barrywardell ()
00:16:27toffe82not really, just for spare parts and who knows, the lcd on this one is perhaps working
00:18:41 Quit fyrestorm (Read error: 113 (No route to host))
00:19:08 Quit [CBR]Unspoken|w (Read error: 110 (Connection timed out))
00:20:07saratoga22i think probably half the gigabeats on ebay are broken
00:21:01 Quit fyre^OS (Read error: 113 (No route to host))
00:22:44 Quit dabujo ("( :: NoNameScript 4.2 :: )")
00:23:29 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
00:24:23 Quit robin0800 (Read error: 104 (Connection reset by peer))
00:25:47mcuelenaerewhy does gcc assembly "x &= ~y;" to "LDR R0, =[~y]\nAND x, x, R0" instead of "BIC x,x,#y"? is there an optimisation flag for this?
00:26:05 Quit Horscht ("electromagnetic radiation from satellite debris")
00:29:50linuxstbIs y a constant?
00:30:42linuxstbI'm sure I've seen gcc use bic in those cases.
00:30:56mcuelenaereyes, y is a constant
00:31:25mcuelenaereif gcc would do this all the time, wouldn't be this a total loss of binary size (and RAM usage)?
00:32:48linuxstbWhat's the exact statement it's not using bic for?
00:33:38mcuelenaerefor example "IO_CLK_MOD1 &= ~(CLK_MOD1_VENC | CLK_MOD1_OSD);"
00:33:46mcuelenaereCLK_MOD1_x are macro's
00:33:49amiconnmcuelenaere: Depends on your y. In order to use bic, y needs to be an allowed arm constant
00:33:58 Quit Synergy6 ("Adios")
00:34:08mcuelenaereit should be possible, as the OF uses BIC
00:34:17mcuelenaereand the code is identical
00:34:26mcuelenaerewell, maybe not in C; but in ASM it should be
00:36:14*amiconn isn't surprised though that gcc doesn't pick up all possible optimisations
00:36:48mcuelenaeremaybe a stronger optimisation has to be set in the Makefile?
00:37:17amiconnIf you set higher -O levels, all you achieve most of the time is bigger code which runs slower
00:37:27linuxstbgcc definitely uses bic quite a lot though - to implement &= ~x
00:37:41mcuelenaeremm strange
00:38:01amiconnlinuxstb: Yes, but I'm not surprised it doesn't pick up all opportunities to do this
00:38:37bertrikmaybe it has something to do with IO_CLK_MOD1 being volatile or not
00:38:41amiconngcc isn't very good at optimising (perhaps it is on x86 - that's something I don't know at all)
00:39:22amiconnThat's why the most time critical code parts in rockbox are often written in asm
00:39:24mcuelenaerethis is a pastebin
00:39:38mcuelenaereabove is rockbox, below OF
00:40:26linuxstbbertrik: I've just tried "REG &= ~0x18" (which IIUC is your statement), and gcc used bic...
00:40:37linuxstbThis is compiling for the m200
00:41:01mcuelenaereIO_CLK_MOD1 is a volatile unsigned short*, would that be it?
00:41:27linuxstbIn my code regs are longs (but also volatile)
00:41:59mcuelenaerethey have to be 16-bit accesses
00:42:10 Quit gevaerts ("sleeping")
00:42:39linuxstbAh, that's the difference - gcc uses bic for longs, but not shorts...
00:42:46amiconnmcuelenaere: That means that gcc cannot use bic without a separate hint
00:43:04mcuelenaereamiconn: and that hint is?
00:43:10bertrikhmm, does it *really* have to be 16-bit?
00:43:22mcuelenaerehmm, as far as I know yes
00:44:26amiconnA short only fills half of a reg, and hence it cannot use bic for &=. You need to experiment a little - maybe some casting will do the trick
00:44:42mcuelenaeredatasheet says: "Address 0x0003:xxxx - This area should be accessed by half word (16-bit)"
00:45:12amiconnA somewhat nasty expression that should do the trick though:
00:45:54amiconnIO_CLK_MOD1 = (unsigned int)IO_CLK_MOD1 & ~(CLK_MOD1_VENC | CLK_MOD1_OSD);
00:46:51mcuelenaereand if I would do #define IO_CLK_MOD1 (unsigned int)(*(volatile unsigned short*)(0x***)); would that work too?
00:47:45amiconnI wouldn't care too much about this though, unless this code is executed several thousands of times per second
00:48:12mcuelenaereah ok, thanks
00:48:13linuxstbamiconn: That doesn't seem to work...
00:49:07mcuelenaereand why does GCC insists on doing "LDR R3, =function\nBLX R3" ? I've not seen "BL function " very much
00:49:18bertrikall calls on arm being long (as amiconn explained to me) probably also helps a lot in increasing the binsize
00:49:33amiconnNot all calls are long
00:49:52bertrikok, I misunderstood then
00:49:59amiconnCalls within the same section are short, and calls to static functions are always short
00:50:39amiconnCalls to iram functions are always long though (because it's a separate section), except on iFP
00:51:21mcuelenaereah, you're right now I see
00:51:49mcuelenaereI just thought it would always do long calls (which would double binsize for these calls)
00:53:31bertrikI was confused by the -mlong-calls compiler option thinking it forced *all* calls to be long
00:53:43amiconnThe only other reason where this bic vs. ldr/and issue matters would be if you're short on memory
00:54:02amiconnNot sure how much ram the zvm has
00:54:08mcuelenaere64MB :)
00:55:28amiconn...or for some reason the loadable firmware file size is limited
00:56:15 Quit mf0102 (Remote closed the connection)
00:56:15mcuelenaerenaah, not that I know of :)
00:56:19amiconnBoth reasons apply to the archoses, and the first one applies to the iFP (but that port seems to be almost dead, unfortunately)
00:56:57amiconn(archos isn't arm based of course)
00:58:01 Quit faemir (Remote closed the connection)
01:04:25linuxstbMy two telechips devices (M260 and DAX) only have 2MB RAM... But at least they're flash-based.
01:05:08amiconnYeah, but nevertheless you should try to keep rockbox' memory footprint low on them
01:05:34linuxstbOf course.
01:05:52linuxstbThe hardware's very simple though - and the LCD is just 128x64 mono
01:06:10amiconnThere are several cases where more ram gives better experience - higher jpeg size limit
01:06:15amiconn...for instance
01:06:23amiconn*more free ram
01:07:33Nico_PI'd be curious to see how the buffering code behaves on such targets
01:07:55linuxstbNico_P: I was meaning to ask if you've tested buffering with a tiny buffer...
01:07:59Nico_Pit ought to be fine of course, but still
01:08:02Nico_Plinuxstb: no
01:08:19 Quit bepe86 (Read error: 104 (Connection reset by peer))
01:08:20 Join Bepe_ [0] (n=Bepe@
01:08:38Nico_Pbut I guess it's the same as a 32 MB buffer and even bigger tracks
01:09:15amiconnlinuxstb: Maybe you event want -Os on those targets
01:09:22linuxstbDo we still have that 512KB codec malloc buffer?
01:09:30n1slinuxstb: the voice file for most targets is around 800k so it could get cramped in there...
01:09:35n1slinuxstb: i think so
01:10:08Nico_Pisn't vorbis the only codec that really needs it?
01:10:29linuxstbYes, I think so. But others use it a little as well - just because it's there...
01:10:47n1smaybe it doesn't need to be so frigging huge
01:11:14amiconnAnd the codec and plugin ram should probably be reduced. Not sure whether 32KB of plugin ram (as on archos) are sufficient though - arm code tends to be bigger than sh1 code
01:11:15linuxstbAnd I guess I can get away with an almost Archos-sized plugin buffer.
01:11:39amiconn48KB should definitely do though
01:11:46linuxstbYes, that's the figure I was thinking.
01:11:52amiconniFP used 64KB iirc
01:12:03amiconnAnd that one only has 1MB ram
01:13:11 Quit csc` ("If you can't do anything smart, do somethin right ~Serenity")
01:13:40 Quit lee-qid (Read error: 110 (Connection timed out))
01:14:22linuxstbIt's not even complete yet, and rockbox.bin is about 390KB for the m200...
01:14:31*amiconn ponders whether he should pick up the iFP port...
01:14:40amiconnTry -Os ...
01:14:48linuxstbJust about to...
01:15:25linuxstbIs changing it in the Makefile in the build directory enough?
01:15:43n1sshould be + make clean of course
01:16:16 Join petur [50] (n=petur@rockbox/developer/petur)
01:16:24linuxstbHmm, down to 366064 (was about 392000)
01:16:54Nico_Pamiconn: considering the small size of the potential audience, I'd say it's not a top priority port
01:17:16Nico_Pbut it could be fun of course
01:18:07amiconnIt's somewhat challenging
01:18:20amiconnJust 1MB RAM and 60MHz single core arm7tdmi
01:19:21amiconnNot sure which other DAPs exist that use the PNX0101
01:20:31*linuxstb feels spoilt - 2MB RAM and 120MHz
01:21:27amiconnlinuxstb: The size gain sounds reasonable. ARM code is bigger than SH1 code, and swcodec has somewhat more core code than hwcodec
01:21:57 Quit styleism (Read error: 110 (Connection timed out))
01:21:59linuxstbYes, including speex...
01:22:25amiconnMaybe someone should have a look at thumb again, for non time critical stuff
01:23:14amiconnlinuxstb: The spoiling goes one step further - iirc telechips is arm9
01:24:05*linuxstb promises not to clock it faster that 50MHz ;)
01:24:32*amiconn is curious whether the logikdax will be able to decode ape realtime, and which compression level(s)
01:24:53linuxstbThere's a list of PNX0101 players here -
01:24:59linuxstbSeems quite a few.
01:25:28n1samiconn: will the ifp be able to? ;)
01:25:47amiconnn1s: Haha, even the PP targets aren't...
01:26:12linuxstbamiconn: IIRC, armv5 has got some DSP instructions which should help with APE, so it should be reasonable.
01:27:26*amiconn wonders why he remembered 60MHz for the PNX0101
01:29:27*Nico_P is pleased to see users don't seem to be complaining about playback too much ATM
01:29:30toffe82saratoga22: still winner ?
01:30:46amiconnlinuxstb: Did you take a look at that libavcodec range decoder?
01:31:08moosNico_P: playback seems just fine, just encountered problems with playlist operations
01:31:34*mcuelenaere wonders himself why linux doesn't have a logical target tree like Rockbox..
01:31:49Nico_Pmoos: it's not quite as fine as I'd like it to be, but it seems good enough that I can rest for a bit before going back to it :)
01:32:46*amiconn would in fact like to do that just for the challenge, but can't afford to do so atm :/
01:32:54mooshehe :)
01:32:56amiconnToo many other things on my list already
01:33:03linuxstbamiconn: Only superficially, and I couldn't see how it could replace what APE is using. I also don't think the ffmpeg have done anything with it either.
01:33:37linuxstbBut I'm sure there must be room for optimisation there somewhere (or elsewhere in that part of the code).
01:34:55amiconnI think the predictor can be optimised a bit more on coldfire, by carefully checking which register is "reserved" for how long, and then converting some stack vars into register vars
01:35:19***Saving seen data "./dancer.seen"
01:35:23amiconnBut more than half of the decoding time is spent in the range decoder at -c1000
01:38:10amiconnOptimising the predictor might be enough to make -c3000 usable on coldfire
01:39:05 Quit ompaul (Client Quit)
01:42:14saratoga22toffe82: yes still winning
01:42:46saratoga22doesn't FAAC use the malloc buffer too?
01:43:19linuxstbThe mp4 parser does - for the seektable (a list of frame sizes)
01:43:27linuxstbI'm not sure about FAAD itself
01:43:33linuxstbBut I expect so.
01:43:42saratoga22mp4 frames are variable sized?
01:44:23saratoga22or do you mean AAC frames
01:45:00 Quit ender` (" Trying to establish voice contact ... please yell into keyboard.")
01:45:02linuxstbI mean mp4 packets
01:45:31linuxstbIIRC, they can vary both in duration (number of samples) and size in bytes.
01:45:44saratoga22how annoying
01:46:02saratoga22how is one supposed to deal with this normally?
01:46:24saratoga22i can't image the format expects lots of space to buffer seek tables
01:46:37 Quit bertrik ("bye")
01:48:34*linuxstb shrugs
01:49:19n1s4.45 runtime on my S30 with q4 oggs
01:50:37n1sa whooping 38% of the qouted 12.30 runtime for the OF, room for improvement :)
01:52:26BigBambin1s: hmmm
01:52:34n1snow let's hope batter_bench worked ok
01:54:51*linuxstb wonders how the beast would do with insane APE files
01:55:23*n1s has 0 ape files and is not aware of any linux encoders
01:55:43Nico_Pn1s: only 4.45? that's really not very good
01:55:46linuxstbThe command-line encoder works fine with Wine - do you want it?
01:56:10n1slinuxstb: sure, can't test now though, needs to recharge
01:56:22linuxstbAh, in fact I have a Linux binary...
01:57:03n1sNico_P: we don't power down the HD, scale frequency (or adjust core voltage)
01:57:18Nico_Phmm, true
01:58:10 Quit mcuelenaere ()
01:58:19n1sand maybe some hardware inside the soc as well as outside is left on?
01:59:13Nico_PI have trouble imagining WinCE as very battery efficient though :p
01:59:40linuxstbAs shown by it's tendency to keep the hard disk spinning...
02:00:08Nico_Pindeed. something tells me that's not all
02:01:19linuxstbI guess we'll find out soon enough.
02:01:54 Quit petur ("killed by baileys")
02:03:18 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
02:03:33toffe82saratoga22: winner :)
02:05:27Nico_Panother beast owner? :)
02:06:27BigBambiLong live the beast hegemony!
02:06:42XavierGrIndeed, I will start using it as my main player
02:07:05XavierGrI need to battery_bench the beast to make sure that it lives to its legend.
02:08:53 Quit Nico_P ("bed time")
02:09:07 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
02:09:42saratoga22i've still got 10 minutes to go!
02:10:05saratoga22i think theres going to be more beast developers then beast users
02:11:08toffe82saratoga22: I thought you were bidding on the broken one , it finished at 15$
02:11:28saratoga22no i'm bidding on a working one
02:13:38scorchetoffe82: what was broken about it?
02:13:58toffe82lcd ? they didn't say it
02:14:07saratoga22the guy didn't say, other then that it wouldn't turn on
02:14:28scorchethats a shame...for $15..
02:14:49saratoga22unless people say the screen works, i tend to assume its broken, since half of ebay gigabeats are people selling ones with broken screens
02:15:18krazykiti need an s60 with a broken screen if saratoga22 wins the working one ;)
02:15:43 Join webtaz [0] (
02:15:48saratoga22those seem hard to find
02:15:53saratoga22i looked for a bit
02:16:15toffe82I misunderstood, I wanted to bid on the broken one :(
02:16:18krazykiti'm not finding a lot of s60s on ebay, unless somehow i just fail at searching.
02:16:52saratoga22toffe82: sorry, didn't realize there'd be two auctions ending at almost the exact same time
02:17:07webtazI'm trying to sync a patch, and could need a bit help with the playback_control (or the changes wich viewport did to it)
02:17:19toffe82that's ok, there are some more ;)
02:19:36linuxstb_webtaz: What's the problem? (or what's the patch?)
02:19:47webtazdict2 patch
02:19:48 Quit linuxstb (Nick collision from services.)
02:19:51 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
02:20:42webtazi managed to fix some changes wich are related to viewports, by taking the changes from a actual build which has this patch build in
02:21:03webtazbut this build uses an old version of the playback_control.h i think
02:21:19webtazso i cant reuse the term of it
02:21:25 Nick JdGordon|zzz is now known as JdGordon (i=jonno@rockbox/developer/JdGordon)
02:21:28linuxstbHave you changed the call to playback_control(rb) to playback_control(rb,NULL) ?
02:21:59 Quit XavierGr (Nick collision from services.)
02:22:10 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
02:22:10webtazno that's it
02:22:21webtazi just cant see what the arguments have to be
02:22:34webtazi just saw there has to be one more
02:22:35linuxstbI just looked at all the other plugins that use playback control - they all call it with (rb,NULL).
02:23:08webtazwell i started my player, but i just couldnt find a game that uses this
02:23:21webtazi think i should have tried some others
02:23:26linuxstb'grep playback_control *.c' (in the plugins directory)
02:23:55webtazi'm working with windows
02:24:10webtazjust applying patches and compiling with linux
02:24:19saratoga22alright won a gigabeast
02:25:10BigBambigigabeat S aka gigabeast aka the beast
02:25:10 Quit saratoga22 ("CGI:IRC")
02:25:26linuxstbA typo that stuck...
02:25:31BigBambias it is monstrously powerful, hence a beast
02:25:31webtazi got a F
02:25:39BigBambiF is good too
02:25:50webtazi think S was too expensive when i bought it
02:25:52 Join saratoga [0] (n=9803c50e@gateway/web/cgi-irc/
02:25:59webtazalso not to get in Europe
02:26:08webtazor difficult
02:26:39webtazi was lucky with taxes
02:27:00webtazcustoms -> taxes on goods which go over borders?
02:27:53saratogaor at least the taxmen who check the goods
02:28:08BigBambicustoms duty = the tax
02:32:15orang1In Canada, there is no duty on US computer electronics coming into the country but there is still the provincial tax.
02:33:19webtazcoming into ^^ i just could want sth coming out to me
02:33:21BigBambiSame thing really. In Europe you have to pay the VAT of the country you are importing it into. However, this is off topic...
02:34:02webtazbut interresting
02:35:03BigBambioff topic nonetheless
02:36:46webtazcome on, at the moment here is no topic, i think, and if you think wide it is related to rockbox :)
02:37:00webtazbut I'll soon bring on a new topic
02:37:12webtazwhen, during compilation, the next error occurs
02:37:49BigBambiwebtaz: Please read the topic and the guidelines
02:38:06linuxstbThis channel is logged though - and people read the logs to see what developments they missed...
02:38:10BigBambiThis is a logged development and support channel for Rockbox only
02:39:06BigBambifor off topic ramblings please see #rockbox-community
02:39:24webtazalready found my way in that channel ;)
02:39:36jhMikeSmoos: pong
02:42:32n1sjhMikeS: i did a battery_bench and got 4.45 on my S30, will try to make a first stab at a discharge curve tomorrow
02:44:34jhMikeS4.45? I think we need to start turning unused devices off. :)
02:44:51saratogaor at least spinning down the hard disk
02:45:16BigBambior both :)
02:45:17n1ssaratoga: it spinns down but is not powered down, but i think i know the gpio for that
02:45:27saratogamisread then
02:45:42jhMikeSthere's two GPIOs for the HD power
02:46:23 Quit Zarggg (Read error: 104 (Connection reset by peer))
02:46:36n1sthen I think I know one of them :)
02:46:42 Join Zarggg [0] (
02:47:18n1sif I unset bit 5 in gpio3dr rockbox hangs when trying to access the drive
02:47:50n1sjhMikeS: do you know which ones they are?
02:47:58jhMikeSone is to enable the ata interface and the other is the actual power
02:48:12jhMikeSI'm looking for my notes on that :)
02:49:44jhMikeSI thought maybe MCU2_16 was the ATA interface and MCU3_5 was the HD power (3_5 shut it down)
02:51:47jhMikeSI'll try to confirm on the tracing board for the other
02:52:49n1sok, it's bedtime for me now :) goodnight
02:52:52 Quit n1s ()
02:55:59 Quit spiorf (Remote closed the connection)
02:56:19 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
02:57:34 Quit Horscht (Client Quit)
02:57:37 Join webguest46 [0] (n=443be67b@gateway/web/cgi-irc/
02:57:43 Quit webguest46 (Client Quit)
03:00:27 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
03:14:32 Join Monkeytamer [0] (
03:21:32 Join Hillshum [0] (
03:32:11 Join Monkeytamer_ [0] (
03:35:23***Saving seen data "./dancer.seen"
03:37:24 Quit webtaz (Read error: 110 (Connection timed out))
03:38:45JdGordonanyone have any ideas how the quickscreen should be voiced?
03:40:40jhMikeSaudibly? :p
03:41:36JdGordonjust when the option changes?
03:43:10jhMikeSI suppose the only other option is to start voicing the settings upon entry and stop that if a button is pushed and then voice the change
03:48:23 Quit Monkeytamer (Read error: 110 (Connection timed out))
03:50:08 Join sdoyon [0] (
03:50:47JdGordonspeak of the devil...
03:51:01JdGordonsdoyon: looking forward to a talking quickscreen patch :)
03:51:05jhMikeSJdGordon: pong
03:51:26sdoyonJdGordon: huh?
03:51:36*jhMikeS responded to above statement :)
03:51:54 Quit tessarakt ("Client exiting")
03:52:11JdGordonsdoyon: the quickscreen has been udpated so it should be voicable... not sure the best way to do it though so I havnt
03:52:31jhMikeS...about the "devil" thing :p
03:52:32sdoyonHmm... FS #6171 has been sitting in the tracker forever...
03:52:41*scorche waves at sdoyon
03:53:53*Hillshum wants to write in the wiki to help with the v2 port
03:54:21JdGordonsdoyon: ah ok.. so you want it all voiced on eneter then on changes?
03:55:10 Join homielowe [0] (n=eric_j_l@
03:55:12sdoyonYes, seems right, since you can interrupt it anytime if you know what you want.
03:55:44sdoyonGuess I need to catch up on this. Are the options the same?
03:56:07JdGordonif I get my way they wont have to be though :)
03:56:20sdoyonAh. I see.
03:58:08sdoyonI never committed that patch because I got side-tracked speculating on quick access to functions useful for voice users... you probably vaguely remember a thread about that ages ago. Mainly my conclusion is the two really useful things we're missing is speaking the time and battery level, without having to go all the way to the info menu.
03:58:45sdoyonIn my copious spare time (cough) I meant to see if we could just hack this in on top of the current quickscreen, triggering on SELECT.
03:59:04sdoyonOutrageous, or conceivable?
03:59:44sdoyonSELECT for time and LONG SELECT for battery. Seem that key is unused for most targets in the quickscreen?
04:04:16Hillshumcan i get wiki write access to help with sansav2?
04:11:33JdGordonsdoyon: done
04:11:40JdGordonnot the clock thing though
04:11:56JdGordonHillshum: HiltonShumway??
04:12:38sdoyonOh! OK I'll try it.
04:12:50HillshumJDgordon: yeah, sorry
04:16:50saratogaHillshum: which sansa do you have?
04:17:29Hillshumto be specific
04:17:59saratogaHillshum: can you verify that the "special mode" here does not work on the e200?
04:18:47advcomp2019saratoga, i looked at that and that is just msc mode
04:19:04saratogais it the same MSC mode as you get in the OF?
04:19:25Hillshumbut it still could act as recovery mode
04:19:33saratogaso the player boots into the OF normally before connecting?
04:20:07advcomp2019saratoga, yea.. i have a clip and the folder are the same too
04:21:02saratogawhat happens when you disconnect? does it need to reboot or does it stay in the OF?
04:21:15Hillshumalso, when it Auto-Dectect or the first OF MSC will load if OS does not have MTP
04:21:42Hillshumit refreshes database and stays in of (i think)
04:21:55*Hillshum gets his v2
04:22:37 Quit Monkeytamer_ ("ChatZilla 0.9.81 [Firefox 3.0b5/2008032620]")
04:23:54 Quit daurn ("Cyas later...")
04:23:56Hillshumif the flash was written
04:24:12Hillshumboth in MSC and MTP
04:24:20advcomp2019saratoga, it just depends.. if something writes to it, it will do "database refresh" other then that it will boot normal
04:24:44saratogawell if its not rebooting afterwards, its not the AMS UMS mode
04:24:50saratogaso scratch that idea
04:24:50Hillshumit doesnt give the splash screen
04:25:53*Hillshum will try after holding down |<<
04:27:12Hillshumif it will MSC(mebby i need auto detect)
04:27:50 Quit Bepe_ (Read error: 104 (Connection reset by peer))
04:28:06 Join bepe86 [0] (n=Bepe@
04:29:37Hillshumhmm... 5+ times won't load MSC While |<< is down. Maybe that got scratched in v x.16
04:29:59saratogahow are you doing it?
04:33:28Hillshumpush |<<, plug in, see WMP pop up with MTP
04:34:06saratogathe hold thing doesn't make any difference?
04:37:14 Quit homielowe ()
04:38:42saratogai guess that feature isn't on the e series
04:38:46saratogaodd that they would do that
04:38:58Hillshumit did, will switch to .14
04:40:19advcomp2019i remember the clips trick and it does not work any more too
04:40:38 Join LambdaCalculus37 [0] (n=1800dac6@gateway/web/cgi-irc/
04:41:07*LambdaCalculus37 just read the news... sound on the Gigabeast! :D
04:41:22LambdaCalculus37jhMikeS: Excellent work! Congratulations! :)
04:42:17jhMikeSLambdaCalculus37: thanks
04:42:35Hillshumsaratoga advcomp2019: not .14 MTP
04:43:49LambdaCalculus37jhMikeS: I've started some initial work on the manual. I need to get a keymap so I can add the key options to the manual (right now most of it is based off the F manual).
04:44:28jhMikeSLambdaCalculus37: Nico_P just updated the keymap
04:44:46sdoyonJdGordon: uh the new quickscreen seems to just lockup my x5...?
04:44:55Hillshumno firmware .14 MTP or Auto-detect special mode
04:45:08JdGordonsdoyon: ?
04:45:24LambdaCalculus37jhMikeS: Okay, I'll take a look later on.
04:46:13sdoyonJdGordon: music stops, it doesn't talk, and I can't seem to get out of it. Sighted help has gone to bed so ... ;-)
04:46:48LambdaCalculus37toffe82: Here?
04:46:50JdGordonumm... crap
04:47:03JdGordontalking was working fine in the sim..
04:47:26JdGordonI dont know the x5 keymap so cant help there
04:47:56sdoyonJdGordon: and I need a long POWER to reboot. I know the keymap.
04:48:14sdoyonJdGordon: well, I'll back out your talking rev and see.
04:52:01JdGordonsdoyon: works fine on my sansa..
04:52:14sdoyonJdGordon: r17344 doesn't crash it.
04:53:24 Quit BHSPitLappy (Remote closed the connection)
04:53:26JdGordonok, i dont have time to fiddle atm so ill revert that commit
04:53:39JdGordonactually.. no time, can you back out that commit?
04:54:10sdoyonJdGordon: Uh OK, will do. I'll look at it later, but it's getting late for me too.
04:56:02*LambdaCalculus37 pokes toffe82
04:58:54 Join daurnimator [0] (n=daurn@unaffiliated/daurnimator)
05:03:10 Quit sdoyon ("ircII EPIC4-2.6 -- Are we there yet?")
05:03:30 Join miepchen^schlaf_ [0] (
05:05:01 Quit saratoga ("CGI:IRC (EOF)")
05:06:55 Quit LambdaCalculus37 ("CGI:IRC 0.5.9 (2006/06/06)")
05:08:33 Join csc` [0] (n=csc@archlinux/user/csc)
05:11:28 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
05:12:31 Quit csc` (Client Quit)
05:12:53 Join csc` [0] (n=csc@archlinux/user/csc)
05:16:47 Join blistov [0] (n=blistov@
05:16:54 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
05:17:12blistovi just upgraded my e200 to the current build, and now it crashes when i plug it into usb.
05:17:16blistovrather, hangs.
05:17:24blistovwhat happened to the archives?
05:21:20HillshumRB USB or sandisk usb?
05:24:10 Nick fxb is now known as fxb__ (
05:24:33advcomp2019blistov, you plugging it in while in Rockbox?
05:25:15blistovhaven't updated rockbox in about 6 month.s
05:25:30advcomp2019can you update it?
05:29:08advcomp2019can you please then try it again? another thing is rockbox does not have usb support for the PP based targets yet
05:30:22blistovright, but it used to be able to do usb power while in rockbox.
05:30:30blistovhas that option just been taken away again?
05:32:54advcomp2019yea that is still there but if i remember right you need to hold the select button while plugging it in(at least that works for me)
05:33:29 Quit moos ("CGI:IRC (EOF)")
05:35:05 Join moos [0] (n=c40cd998@gateway/web/cgi-irc/
05:35:25***Saving seen data "./dancer.seen"
05:49:12 Join Rogue84_ [0] (
05:49:52blistovdoesn't work for me with the current build.
05:49:57blistovjust locks hard.
05:50:02blistovhave to pull the battery.
05:51:27advcomp2019hold on.. i will test it with my sansa as soon as i get back in a few minutes
05:56:31 Quit Rogue84 ("BitchX: faster than a speeding bullet, more powerful than a locomotive")
05:58:35 Quit jhulst (Remote closed the connection)
05:58:50 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
05:59:24 Quit brent113_ (Remote closed the connection)
05:59:57scorcheyou dont have to pull out the battery...just hold the power button for 15 seconds...what you are experiencing is something that i think is happening to a small subset of devices that for some strange reason arent rebooting as they should...just reboot into the original firmware when you need to use USB instead of relying on Rockbox to reboot your device
06:00:10 Join brent113_ [0] (n=brent113@
06:03:43 Part orang1
06:04:37Rogue84_Hiya need help install rockbox on ipod 80gig video 5th gen
06:05:17scorchewhat issue are you having?
06:05:34Hillshumarnt all the 80gbs 5.5g?
06:05:44Rogue84_ipod not found
06:05:52Rogue84_with ipod patcher
06:05:55scorcheHillshum: well, there is the classic
06:06:21scorcheRogue84_: are you sure it is an ipod video and not an ipod classic?
06:06:27Hillshumoh yeah, but 5gs are 30 and 60?
06:06:50Rogue84_not sure
06:07:23scorcheRogue84_: what color is it?
06:07:42scorchelike this?
06:08:09scorchethen you do not have an ipod have the ipod classic which is the 6th gen
06:08:34Rogue84_then irockbox won't work
06:08:42Rogue84_thank you
06:09:54advcomp2019blistov, i can use rockbox while the usb cable plug in when holding the select button with the latest build just fine
06:13:07 Quit jhulst (Remote closed the connection)
06:20:55 Join DerDome [0] (
06:22:47LloreanjhMikeS: Around?
06:24:19LloreanWith the new bootloader I get "File Not Found"
06:26:19jhMikeSexactly when did that start? just on this revision?
06:26:54 Part blistov ("Kopete 0.12.4 :")
06:27:10LloreanNo, a couple revisions back didn't work either.
06:27:18LloreanI haven't narrowed it down to exactly when it started yet.
06:27:34LloreanBut I tried earlier today to update from one that still defaulted to the OF to a current one, and it didn't work.
06:27:43jhMikeSI don't seem to be having the trouble so if you could find it I'd appreciate it
06:27:59LloreanLet me just clarify that I'm not doing something stupid.
06:28:04 Quit Rogue84_ ("changing servers")
06:28:08jhMikeSmoos is having the same error too
06:28:15LloreanIt should be safe for me to simply drag an nk.bin that i compile onto the hidden partition, overwriting the old one?
06:29:13*moos will try with last svn build
06:31:45 Join alphanumeric [0] (
06:32:42alphanumericdoes rockbox for sansa support morse code input? i've checked the forums and the manual (it is in the manual for Ipod video)
06:33:06LloreanjhMikeS: I'm too tired to go through SVN revisions tonight, but I'll check tomorrow. Best bet is to just follow changes in /bootloader/gigabeat-s.c, I assume?
06:35:54jhMikeSLlorean: or in crt0.S
06:37:17jhMikeSLlorean: you had a functioning bootloader USB mode?
06:38:01jhMikeS+ booting rockbox.gigabeat?
06:38:20LloreanI had bootloader v08 working dual booting.
06:38:36LloreanBut since it was OF first, I decided to switch it out while updating to a current build.
06:38:45LloreanI attempted to create a dual boot nk.bin, it booted the OF but not Rockbox.
06:38:53LloreanThis was bootloader v10 btw
06:39:20LloreanThen I saw it was more verbose now so updated to v11, but didn't bother creating a dual boot one. USB mode works fine, but I get a "File Not Found" when it goes for rockbox.ipod
06:39:24LloreanEr .gigabeat
06:39:55 Join AndyIL [0] (n=pasha_in@
06:39:58 Join miepchen^schlaf [0] (
06:40:04 Quit miepchen^schlaf_ (Read error: 104 (Connection reset by peer))
06:40:26moosLlorean: exactly same situation here
06:41:26jhMikeSso perhaps v9 works but not v10?
06:42:46LloreanI have not tested v9.
06:43:20jhMikeSv8 to v9 was just a reversal, v10 was using the program flow prediction
06:44:09 Quit HellDragon (Client Quit)
06:44:43LloreanWell when I'm on a computer I can compile on in the morning, I'll start there.
06:48:29 Join HellDragon [0] (i=jd@unaffiliated/helldragon)
06:49:43alphanumericsorry for asking again, but does rockbox for sansa support morse code input? i've checked the forums and the manual (it is in the manual for Ipod video)
06:55:05 Join steve0 [0] (n=3ab26fe4@gateway/web/cgi-irc/
06:55:42steve0just wondering if rockbox can play .rtf files? (i have rockbox on my gigabeatf40)
06:56:38alphanumericsteve0: it will view it as plain text, so you'll see all of the hidden formatting. It will look really ugly.
06:57:21alphanumericsteve0: if you want to demo what it will look like on your pc, just right click on the file and open with notepad (assuming you're a windows user)
06:57:45steve0ok thats not to bad
06:58:13steve0i clicked on the file and nothing happened in my gigabeat if i use "open with" which program should it be using?
06:58:57alphanumericsteve0: I think it will say viewer
06:59:17alphanumericsteve0: hold on, let me find my sansa
07:00:19 Quit AndyI (Read error: 113 (No route to host))
07:01:57alphanumericsteve0: can't find it right now, probably out in the car. it's called like text viewer or just viewer. You can format the rtf in wordpad or word, just file -> save as -> text document
07:02:54steve0ok ill try that it worked in .rtf just has alot of crap like /par non stop all the way through it
07:04:08alphanumericsteve0: yeah, there's no rich text file support. plain text works just fine anyway.
07:06:34steve0ok thx for the help
07:06:54alphanumericsteve0: sure no problem
07:08:47 Join daurnimas [0] (
07:09:33 Quit steve0 ("CGI:IRC (EOF)")
07:11:56 Quit daurnimas (Client Quit)
07:14:46 Join daurnimas [0] (
07:16:46 Quit moos ("CGI:IRC")
07:19:24 Quit daurnimator (Connection timed out)
07:19:24 Nick daurnimas is now known as daurnimator (
07:29:40 Join fyrestorm [0] (
07:32:55alphanumericleave rockbox
07:32:58 Part alphanumeric
07:35:29***Saving seen data "./dancer.seen"
07:38:35 Join rdvon [0] (
07:39:09rdvonI hope you don't get this question too much, but I was wondering if the latest zen will ever be supported?
07:44:19scorche|shrdvon: well, it isnt going to happen with no one working on it
07:44:59rdvonOh. :(
07:45:28rdvonNo doom?
07:46:51rdvonHas rockbox development stopped altogether?
07:47:37scorche|shnot in the can see the commit log on the front page
07:47:53scorche|shno one is working on porting "the latest zen"
07:49:18rdvonWell.. I'm happy with my zen, but I don't see why it has to transfer videos to wmv.
07:49:24rdvontakes forever.
07:53:40 Quit rdvon ("ChatZilla 0.9.81 [Firefox]")
07:54:27Mouser_XWMV = :(
07:54:55Mouser_XThe 'beast has a 500 mhz CPU, and it does a poor job of playing WMVs.
08:03:43 Join Rob2222 [0] (
08:08:29 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
08:19:31 Quit jhulst (Remote closed the connection)
08:19:46 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
08:21:46 Quit Rob2223 (Read error: 110 (Connection timed out))
08:28:51 Quit jhulst (Remote closed the connection)
08:29:06 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
08:36:40 Part toffe82
08:39:14 Quit jhulst (Remote closed the connection)
08:39:29 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
08:52:13 Quit jhulst ("Konversation terminated!")
09:16:22 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
09:16:51 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
09:21:35 Join blistov [0] (n=blistov@
09:21:48blistovare there plans to finish pictureflow?
09:26:05 Quit kushal_12_27_200 ("This computer has gone to sleep")
09:35:33***Saving seen data "./dancer.seen"
09:35:56 Join spiorf [0] (
09:41:20 Join barrywardell [0] (n=barrywar@
09:42:24 Quit csc` (Remote closed the connection)
09:44:35JdGordondoes anyone know if there is a limit to how much talking can be queued up before wierd things start happening?
09:45:30amiconnWeird things should never happen
09:45:38amiconnThe talk queue is 64 entries deep
09:45:55amiconnIf you try to queue more, all that happens is that the extra entries are dropped
09:46:58JdGordonok, so in the sim that talking quickscreen worked fine.. on my e200 it says "shuffle on repeat show files"... any idea why the repeeat and show files values are skipped?
09:47:36*JdGordon rebuilds the voice file to make sure...
09:48:02amiconnYou can hear the effect if you enable filename spelling and then hover over a loooong filename
09:49:00*amiconn wonders what kind of crash sdoyon experienced
09:50:01*JdGordon also.. apart from the skipping the voice, it works fine on the sansa
09:50:18JdGordonunless its a coldfire thing
09:51:00amiconnWell if you get different behaviour on sim and PP already, there must be a problem
09:51:04JdGordontagcache needs to be made to skip <microsd> files if the card i missing
09:51:11 Join davina [0] (
09:51:45amiconnI guess it has to do with pointer types given how much pointer fiddling the settings code is doing
09:52:54JdGordondidnt we enable the shadow variable name warning in the builds?
09:52:59JdGordonI tihnk that may be the problem
09:53:08JdGordonint value = (int)value;
09:53:50amiconnEspecially casting int* <-> bool* is asking for trouble if you're not *very* careful
09:54:23JdGordonyeah, although I tihnk thats all ok here..
09:54:47JdGordonpretty sure that line is the problem.. i changed the temp_var vairbale name to value and didnt check carefully enough :p
09:56:00amiconnIf you cast a boolean 'true' to int, you'll end up with -1 on most platforms
09:56:36amiconnBut that's not guaranteed. Could also be 1, or 255, or probably something else
09:57:35amiconnSo if you need to expand a bool to 0/1, you need to use ?: order to make it portable
09:58:30JdGordonreturn *(bool*)setting->setting==true?1:0;
09:58:40amiconnYes. Usually a bool has all bits set. If sizeof(bool) == 4, it's obvious that you end up with -1
09:59:35amiconnOn out targets though, sizeof(bool) == 1 (iirc). Then it depends on whether the compiler treats bool as signed, and whether it performs the extension to 32 bit before or after making it signed if not
10:00:02amiconnJdGordon: I was referring to your (int)value
10:02:06JdGordon:) its working correclty on the sansa now
10:04:52*amiconn also thinks that it's a very bad idea to name a new variable with a different type the same as the existing one
10:05:00amiconnVery confusing when reading the code
10:07:27 Join bertrik [0] (
10:09:12JdGordonthe point was that is wasnt a new variable.. it was a rename which was missed...
10:11:13amiconnint value = (int)value;
10:11:26amiconn'value' being a bool outside that block...
10:11:48JdGordonno.. its void* outside that block.,..
10:12:12JdGordonbut the line was int value = (int)temp_var;
10:12:27JdGordonthat line got missed
10:12:30JdGordonall fixed now
10:18:12 Join n1s [0] (n=nils@rockbox/developer/n1s)
10:25:21amiconnJdGordon: Here are some screenshots regarding the somewhat strange item placement:"> (higher than I'd expect) and (lower than I'd expect)
10:26:31 Join Mathiasdm [0] (
10:27:32 Join ZincAlloy [0] (n=d9eee40f@gateway/web/cgi-irc/
10:27:49*linuxstb wonders why the Gigabeat S bootloader has about 300KB of zeros in the middle...
10:28:24 Join webguest12 [0] (n=4e6a42e5@gateway/web/cgi-irc/
10:29:18linuxstbSeems to be the interrupt vectors - the load address is after the bss.
10:30:44 Quit webguest12 (Client Quit)
10:31:49 Join miepchen^schlaf_ [0] (
10:31:55 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
10:40:15n1sjhMikeS: is the voltage reading on the S known to be right?
10:40:25 Nick fxb__ is now known as fxb (
10:48:59 Join ReKleSS [0] (
10:49:43ReKleSSerr... if I press the power button on my H120 and nothing happens, does this mean my bootloader install went bad?
10:51:10n1sis is charged?
10:51:49ReKleSSI got a battery low warning earlier, but it's on AC power now
10:52:08n1sand nothing is on the display?
10:52:15ReKleSSno, it just sits there
10:52:34n1sand the firmware upgrade seemed to go ok?
10:52:37bluebrotherwas it powered when you flashed the bootloader?
10:52:53ReKleSSn1s: it normally shuts down afterwards, right?
10:52:54ReKleSSbluebrother: yes
10:53:32 Join ender` [0] (
10:53:35n1sReKleSS: yes, it does, try resetting by pushing a straightened out paperclip into the reset hole on the bottom of the player
10:54:30ReKleSSstill nothing
10:55:42n1shmm, then I don't know if there are more things to try, doesn't look good :/
10:58:07Horschtpower outage during flash process is never a good thing
11:04:33ReKleSSbah, that's my second H120 in a few months
11:06:41ReKleSSis there a known way to recover the flash?
11:06:48bluebrotherif it was on AC power during flashing there shouldn't be a possibility of failing. Except if you flash file is corrupt
11:07:01bluebrotheryes: using external hardware (a BDM)
11:07:22ReKleSSbluebrother: maybe, I was trying to upgrade to the 7_pre4 bootloader
11:07:28 Join aliask [0] (n=chatzill@rockbox/developer/aliask)
11:07:56bluebrotherhmm −− have you patched the hex file and flashed it using the OF flashing functionality?
11:08:07bluebrotheryou can flash the v7 bootloader using the iriver_flash plugin
11:08:54bluebrotherbut recovering using a BDM means you need to have special hardware, need to open the player and solder some wires in it
11:08:55ReKleSSis there any real difference?
11:08:59 Quit Rob2222 ()
11:09:12aliaskReKleSS: Hey
11:09:15 Join Rob2222 [0] (
11:09:28bluebrotherthere shouldn't be a functional difference. But I guess the plugin has been more used ;-)
11:09:50*n1s benches again...
11:10:25bluebrothersorry, no idea what could've happend. Maybe ask Slasheri
11:11:01n1sthe log from yesterday ended on 3.6V about an hour before the player died...
11:11:34n1skind of inexact to extrapolate a shutdown voltage from that...
11:11:48 Quit aliask (Client Quit)
11:16:38 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
11:16:54SlasheriReKleSS: you patched the hex file with the 7_pre4 bootloader?
11:17:12 Join pixelma [0] (n=Marianne@rockbox/staff/pixelma)
11:17:12Slasheriyou absolutely shouldn't do that.. that ruins your player, sorry
11:17:30Slasherithat .iriver must be flashed inside rockbox
11:17:35Slasherionly .bin file can be patched
11:17:52ReKleSSoh wait, like that
11:17:54ReKleSSno, I patched the bin file
11:18:21Slasheriwhich one? could you tell me exactly what you did
11:18:22ReKleSSdescramble -> mkboot -> scramble
11:18:46ReKleSSwith from
11:19:21Slasheribut that archive contains bootloader.iriver, not bootloader.bin
11:19:37Slasheribootloader.iriver is not in compatible format to be used with mkboot
11:19:45 Quit JdGordon ("Konversation terminated!")
11:20:02 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
11:20:27ReKleSSthen, it's bricked?
11:20:29SlasheriBDM is the only way to recover your unit now, please contact LinusN. Hmm, need to clarify those flashing instructions
11:21:16ReKleSSactually, the instructions seem pretty clear, I just didn't read closely enough
11:21:24ReKleSSbut a big warning would have been nice
11:21:39 Join JdGordon [0] (
11:21:56 Quit linuxstb (Read error: 110 (Connection timed out))
11:22:48Slasherii will add it, sorry to hear that. never thought somebody would try to use mkboot for that
11:23:40Slasherithe iriver_flash plugin contains many safety measurements, but mkboot contains none
11:23:50ReKleSSwell, back to a H340 for now...
11:24:20bluebrotheroh, there is a pre4 binary around? Nice.
11:24:35SlasheriReKleSS: but the unit can be fixed, please ask LinusN if he's willing to
11:24:36scorcheReKleSS: might i ask what you are doing with the device now?
11:24:52Slasheribluebrother: have been for a sometime :)
11:25:06bluebrotherSlasheri: just noticed it :)
11:25:13bluebrothera brief changelog would be nice.
11:25:22ReKleSSSlasheri: will do, I haven't found any other DAP I actually like
11:25:33Slasherisame here :)
11:26:37Slasheribluebrother: good idea. that release contains mainly bugfixes
11:26:49Slasheriand support for H110 also
11:28:31ReKleSSgah, sweden
11:28:32ReKleSSexpensive shipping
11:29:06linuxstb_Does it matter what I set the boot flags for the two partitions to on the S?
11:29:28 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
11:33:26Mouser_XI think 0 works, but I'm not the one to ask.
11:33:50Mouser_XMost people set it to non-bootable, IIRC.
11:35:38***Saving seen data "./dancer.seen"
11:36:38linuxstbHas anyone tried wiping the OF's files from the main partition? I'm _assuming_ the OF bootloader won't care about them...
11:37:32Mouser_XIt hasn't for me yet.
11:37:44Mouser_XAnd yes, someone did try it out before I got rid of them.
11:39:23 Quit amiconn (" HydraIRC -> <- IRC for those that like to be different")
11:40:16pixelmais it worth it to commit a "fix" for CREDITS which only removes a duplicate again?
11:41:36linuxstbpixelma: Yes, I think so.
11:41:43linuxstbNo commit is too small ;)
11:42:16*pixelma hits enter key
11:43:30 Join lee-qid [0] (
11:47:14 Quit barrywardell ()
11:48:05 Join barrywardell [0] (n=barrywar@
11:49:43*linuxstb does an insane APE test on the beast
11:51:25linuxstbMouser_X: So you've deleted everything from the main FAT32 partition?
11:51:49Mouser_XI put it all in a RAR archive, and then deleted it, so basically yes.
11:52:02Mouser_X(I put it in the archive, just in case I needed it later.)
11:52:06linuxstbAh yes, a backup could be sensible ;)
11:53:19Mouser_XI don't dare touch the firmware partition.
11:53:58linuxstbFrom the sounds of it, the S's bootloader is happy with a brand new (unpartitioned) disk, so it _should_ be able to recover...
11:54:30linuxstbIt would be useful to know what's needed and what's not, and even if the device can be repartitioned...
11:54:32Mouser_XI'm waiting for someone else to experiment on that. I wouldn't mind shrinking it though, if it doesn't get mad at me for doing so.
11:54:45linuxstbI would even try just using a single partition...
11:54:52 Quit hnakibfdi (Read error: 110 (Connection timed out))
11:56:55 Nick fxb is now known as fxb__ (
11:59:00 Quit barrywardell ()
12:00:26 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
12:02:47 Quit XavierGr (Nick collision from services.)
12:02:58 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
12:07:45 Join Gekz [0] (n=brendan@unaffiliated/gekz)
12:09:08 Join mmnk [0] (i=cz@faeroes.freeshell.ORG)
12:09:31 Quit bepe86 ()
12:10:25mmnkhas anybody tried to compile the sources with mingw or from cygwin ?
12:10:51n1smmnk: cygwin is used by quite a few people
12:12:35bluebrotherthere is also some information about cygwin in the wiki, but not MinGW
12:13:27 Quit linuxstb (Read error: 110 (Connection timed out))
12:14:20Gekzhow about MSYS?
12:14:24Gekzwhich is MinGW
12:15:28 Quit JdGordon (Read error: 104 (Connection reset by peer))
12:16:46bluebrotherI guess you'll run into quite some problems due to dependencies of the build system.
12:17:07bluebrotherplus, you'll need to build the compilers yourself −− at least "our" versions aren't around
12:17:17bluebrother(unless I missed something about this ;-)
12:17:30 Join JdGordon [0] (
12:17:49 Quit ZincAlloy ("CGI:IRC (Ping timeout)")
12:18:29*bertrik uses cygwin with the gcc packages from the rockbox package repository and it works fine, although a little slow
12:19:14 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
12:19:41JdGordonso just an idea... not really sure how feasible it is.. but how would everyone react to having inline settings in the menus, but only the selected item would show its value?
12:20:32bluebrotherJdGordon: remember the patch that inlined the settings using two lines? I really liked that.
12:21:03JdGordonyeah, I tihnk 1 moving double height item would look better
12:21:04bluebrotherand if you make that like "two lines for selected, one line for others" I think I'd like that too
12:21:32Gekzis rockboy still developed?
12:21:33JdGordonthe quickscreen changes means there isnt much new code which needs to be added for it to work... just fiddling with the current menu code
12:21:41bluebrotherpreferrably one could turn that off to have all items showing their values in(new)line
12:21:58 Join Thundercloud [0] (
12:22:30bluebrotheri.e. a switch to make everything two lined. And maybe also give an option to get the old behaviour ;-)
12:24:13JdGordoni imagine the current behaviour would be default.. but yes, having all 3 would be good too
12:24:39JdGordonalthough.. you complained about the quickscreen delta.... this wont be free you know.,..
12:24:56bluebrotheryep, sure.
12:25:16*JdGordon would like to kill off the last of the gui_textarea functions also
12:25:35bluebrotherbut this is new functionality −− the quickscreen was kinda surprising me as it doesn't add really new stuff
12:25:38JdGordonyes/no screen seems to be the last user
12:26:10JdGordonit adds alot actually... user font, voicing, remove lots of SYSFONT LANG ids...
12:26:18JdGordonpotential to customize it
12:27:49JdGordon:'( I was searching in the wrong directory... heaps still uses gui_textarea :(
12:28:00bluebrothervoicing was added in the same commit? Ah.
12:28:20JdGordonno, shortly after
12:30:38bluebrotherhmm? /me starts getting confused about that stuff
12:31:16bluebrotherbut anyway, as there were no real other complaints forget mine ;-)
12:33:21 Join leftright [0] (n=d9e1c061@gateway/web/cgi-irc/
12:33:45bertrikJdGordon: interesting idea! I think I like it.
12:34:06*linuxstb_ is puzzled by the Gigabeast's performance with APE -
12:34:37leftrightSlasheri; you aound ?
12:35:34Nico_Plinuxstb_: not as fast as you expected?
12:35:45*JdGordon is late, but sends his congrats to the beast guys for sound :D
12:36:57Gekzwhat's the link to view the svn?
12:37:54Gekzyes just about to say found it
12:38:06JdGordondoes the splash screen handle \n?
12:38:20leftrightSlasheri; could you look over the H1xx bootloader flash instructions ?,please, as I have reorganised the bootloader instructions to make them easier to follow and consistant with the other flash instructions,
12:38:46Gekzwheres the rockboy source
12:39:05bertrikJdGordon: IIRC you asked for volunteers to work on the battery bench plugin to make it use the ata_idle callback instead of polling the ata_disk_is_active. I may have a go at it.
12:39:17Slasherileftright: ok, i will check that. sounds good :)
12:39:23linuxstb_Nico_P: It's only about 10%-20% faster than the F...
12:39:46JdGordonbertrik: yay :)
12:39:46Nico_Plinuxstb_: AFAIK it's not running at full speed
12:40:31jhMikeSn1s: it should be since it's a direct hookup to the voltage sense
12:40:46bertrikJdGordon: however I only have a flash based player
12:41:03JdGordonthats ok
12:41:49Slasherileftright: looks good
12:42:19bertrikmy plan is to used time triggered measurements only (say once a minute) and put them in the buffer, then flush the buffer on the ata idle callback.
12:42:55Nico_Plinuxstb_: would APE gain from an FPU?
12:43:04jhMikeSNico_P: the core clock divider is set to /2 by the loader and we're not touching that yet.
12:43:18bertrikJdGordon: the buffer should be able to hold at least something like an hour, if it gets too full it spins up the disk
12:43:25linuxstb_jhMikeS: So in effect it's running at 50% of potential speed?
12:43:48linuxstb_Nico_P: It would benefit from integer vector math...
12:44:01Nico_Pdoes the beast have that?
12:44:31linuxstb_I don't know - I need to investigate. But at least the armv5 DSP instructions will help.
12:44:36jhMikeSlinuxstb_: a bit less than 50% potential since the PLL is running at 528MHz, not 532MHz
12:44:42bertrikJdGordon: I'm a bit concerned about what happens toward the end of the battery capacity, it would be a pity if we lost the contents of the battery measurement buffer because the players shuts down. Or is the shutdown done in an orderly fashion?
12:45:03 Join bluebroth3r [0] (n=dom@rockbox/staff/bluebrother)
12:45:26 Join styleism [0] (
12:45:29 Quit bluebrother (Nick collision from services.)
12:45:33 Nick bluebroth3r is now known as bluebrother (n=dom@rockbox/staff/bluebrother)
12:45:35JdGordonbertrik: well, it needs to flush its cache either when the disk is spun up, or when its getting full, or if the batt is getting too low
12:46:11jhMikeSView HW Info gives a rundown of many of the clock settings
12:46:19bertrikJdGordon: anyway, I won't commit anything battery bench related before consulting with you or amiconn
12:47:50 Join Synergy6 [0] (
12:47:51 Quit soap (Read error: 110 (Connection timed out))
12:49:07n1sjhMikeS: great, my battery bench log from yesterday looked a little strange as the last entry (about an hour before the player died) was at 3,65V down from starting at 3,96
12:49:38n1sI'm rerunning now with a half an hour timeout for battery_bench to hopefully get more values at the end...
12:49:39 Quit leftright ("CGI:IRC (EOF)")
12:50:41 Nick rvvs89_ is now known as rvvs89 (
12:53:52JdGordonXavierGr: you around?
12:54:32linuxstb_jhMikeS: That makes sense then - my APE tests showed the S about 10-20% faster than the F at the current clock. So still lots of optimisation to do to get the insane ape files playing...
12:54:58jhMikeSn1s: was the shutdown attempted by rockbox at that voltage or the player just died on it's own?
12:55:44jhMikeSWell, figure the beast is running 264MHz and the F at ~295 so it's definitely more clock-efficient
12:55:56linuxstb_Yes, I would have expected that.
12:56:11*Nico_P suggests a test at full speed
12:56:57jhMikeShave at it but I'm not sure the core voltage is set properly to run that fast so it could crash under full load
12:58:27jhMikeSthough I think an LCD speed test went ok at full speed
12:59:16linuxstb_Have you thought about how cpu scaling will work?
13:00:03jhMikeSsomewhat. if we want to use DVFS then CPUFREQ_NORMAL will be variable and scaling hardware controlled.
13:00:59linuxstb_What does "scaling hardware controlled" mean?
13:01:14jhMikeSFreescale was kind enough to have a DVFS table for 27MHz reference frequency in the linux code
13:02:09jhMikeSThe hardware algorthm requests frequency and voltage changes through interrupts but decides by how much and when base upon the loads in various systems
13:02:30jhMikeScache and bus activity and such
13:03:04jhMikeSthe mc13783 has dedicated lines for DVFS
13:04:52n1sjhMikeS: that build was with NO_LOW_VOLTAGE_SHUTDOWN (or whatever it's called defined so rockbox should'nt have shut down, speaking of that the build I'm using now will try to shut down, oh well, we'll see how it goes
13:06:12 Nick mmnk is now known as tamal (i=cz@faeroes.freeshell.ORG)
13:08:42JdGordongreen across the board :D
13:08:57JdGordonalmost reclaimed the voice delta
13:09:12 Quit tamal ("tamal has no reason")
13:09:26 Quit JdGordon ("Konversation terminated!")
13:09:40jhMikeSn1s: I'm considering checking it against a meter reading...should be easy to do. maybe I'll see if can do it now.
13:11:03 Join JdGordon [0] (
13:11:42JdGordonbtw.. if anyone wants a green delta... with a tiny bit of effort I think add_dir in tree.c can be removed
13:12:11JdGordonI tihnk its only used to the root.m3u8 playlist creation which can be done through the playlist catalog anyway
13:12:52 Join gnakinklp [0] (i=0@
13:14:50 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
13:15:42 Quit Mathiasdm ("Yuuw!")
13:17:16 Join Mathiasdm [0] (
13:19:49 Quit merbanan (Remote closed the connection)
13:20:24 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
13:22:10*gevaerts swears that there was no duplicate in CREDITS when he committed that file. The svn server must have added it
13:24:18*pixelma thanks gevaerts for the opportunity to increase her commit count ;)
13:24:18bluebrotheralways this damn complicated technical stuff ... ;-)
13:24:22 Quit lee-qid ("aufwiederbyebientotsayonara")
13:26:01gevaertshmm, two 'g's, my grep only had one. That could explain it
13:26:41 Join webtaz [0] (
13:26:46 Quit styleism (Read error: 110 (Connection timed out))
13:29:23n1sjhMikeS: looking at powermgmt-imx31.c why is the adc value multiplied with 2303 and not 2300? (not that it would make a big difference)
13:29:57jhMikeSn1s: battery reading is right on the money
13:30:17jhMikeSn1s: because it rescales to divide by 512 instead of 511
13:34:48 Join desowin [0] (n=desowin@atheme/member/desowin)
13:35:40***Saving seen data "./dancer.seen"
13:38:50jhMikeSpardon me, 1024 instead of 1023. -512 to 511 is the current senses.
13:39:23amiconnjhMikeS: Is there a reason why thread.c:core_sleep() is separate for CPU_TCC780X, IMX31L and DM320?
13:39:31 Join gregzx [0] (
13:39:34amiconnIt's exactly the same code...
13:40:12*jhMikeS 's code was copied twice?
13:41:05jhMikeSthough I don't know if the IMX31L will remain as it is since there's some errata that will need to be dealt with regarding that
13:53:08amiconnjhMikeS: I'm trying to track down that dropout problem on disk spinup on my mini
13:53:44amiconnI found that somehow interrupts are disabled for a fraction of a second, but I don't have an idea where that happens...
13:56:42 Join w1ll14m [0] (
13:57:14jhMikeSamiconn: during a call to sleep or yield it happens?
13:57:30amiconn...but I don't have an idea where that happens...
13:58:00jhMikeSI thought you meant more like where outside ata.c
13:58:05amiconnI just put a GPIOB_OUTPUT_VAL ^= 0x08; into kernel.c:TIMER1(), and disabled the normal backlight handling
13:58:19w1ll14mJhMikeS: i was wondering... on the GigabeatSInfo page, it says that the mes60v only has a tuner...
13:58:28amiconnThis gives me half-bright backlight whil ethe tick is running continuously
13:58:41*jhMikeS 's S30 has a tuner
13:58:43w1ll14mbut my mes30vw has a tuner to.
13:58:54 Join faemir [0] (
13:59:01amiconnBut on spinup (from power-off), I get a short period of either full backlight or no backlight
13:59:14 Quit faemir (Remote closed the connection)
13:59:39 Join robin0800 [0] (
14:00:00 Join faemir [0] (
14:00:32 Quit faemir (Remote closed the connection)
14:00:55n1sw1ll14m: my S30 has a tuner too so I don't know where that info comes from...
14:01:01amiconnIf I extend the sleep(HZ/50) in ata.c to HZ/5, the problem goes away, so it has to do with retries somehow. The question is why ata retries obviously disable interrupts...
14:01:11 Join faemir [0] (
14:01:14 Nick fxb__ is now known as fxb (
14:01:32 Quit robin0800 (Client Quit)
14:01:42gevaertsDoes the beast have USB host in the OF ?
14:01:51 Join robin0800 [0] (
14:02:16jhMikeSit has the hardware for it I know that much
14:02:25amiconnI suspect this to be the real cause for the G5.5/80 problem
14:02:44jhMikeSamiconn: I was thinking along similar lines about that
14:02:46w1ll14mn1s: Silicon Labs Si4700 (Only on 60GB model, MES60V) i guess that says that the Si4700 is only in the 60GB version?
14:03:11amiconnThe question is how this could be tracked down...
14:03:28jhMikeSw1ll14m: then toffe82 made me up a strange 'beast or it's just a software thing
14:03:50w1ll14mhmmm i'll have a look on my gigabeat
14:04:29 Quit robin0800 (Client Quit)
14:04:44 Join robin0800 [0] (
14:06:23n1sjhMikeS: your don't have fm?
14:06:30 Quit robin0800 (Client Quit)
14:06:47 Join robin0800 [0] (
14:07:48jhMikeSn1s: yes it does
14:08:32BigBambin1s: Does the radio appear in the OF on the 30?
14:08:44n1syes, I eve tried it
14:08:49jhMikeSamiconn: I'm trying to think how it's even possible to have them disabled across a task switch
14:09:48amiconnjhMikeS: I'm not sure whether the tasks are switched. Monitoring task switches the same way (xor'ing the backlight) doesn't work, unfortunately. It gives lots of flicker
14:10:24 Join mf0102 [0] (n=michi@
14:11:10jhMikeSpossible that an interrupt handler somwhere has unbalanced disable/enable?
14:11:11amiconnIt also can't be a new problem, as the very first revision after my ata poweroff commit shows it. I wonder why I didn't notice it back then though
14:11:23jhMikeSnvm, wrong mode
14:12:01 Quit robin0800 (" HydraIRC -> <- Would you like to know more?")
14:12:16 Join robin0800 [0] (
14:12:38jhMikeScould another interrupt be firing repeatedly and it's actually stuck in interrupt mode for some time?
14:14:11*amiconn has no idea how to track this down either
14:14:40robin0800whats different about sansa c200 firmware 1.01.07 from earlier ones? why sanpatcher dosen't work with it?
14:16:16*amiconn now has an idea how to check for excessive irq disabled time
14:16:29jhMikeSsome kind of checking if the interrupt handler fires more than expected or less than expected (would be a good start anyway)
14:18:16 Join waldo [0] (
14:18:19pixelmarobin0800: I don't know of an 1.01.07 version - does it have an MSC mode setting? If so, what is it set to?
14:18:26 Quit XavierGr (Nick collision from services.)
14:18:39 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
14:19:23pixelmayou need your Sansa in MSC mode and I know that Sandisk removed the option for it in firmware versions 1.01.05 and 1.01.06...
14:27:00LloreanjhMikeS: Just checking in the log, I didn't see anything more about my problem, so I'm beginning testing for which revision it began.
14:28:33jhMikeSLlorean: moos tried v8 but still had the problem but you status v8 worked for you
14:29:06LloreanjhMikeS: v10 (r17211) just worked fine for me too
14:30:31LloreanAnd 17289 just worked.
14:30:39*Llorean suspects a current one will now be working, for no apparent reason.
14:31:45*amiconn added some counters in syste,-pp502x.c:irq() which can be displayed in the debug menu
14:31:54LloreanjhMikeS: Yes.. SVN current bootloader works for me now. :(
14:31:58LloreanMeaning I'm no help any more
14:32:08jhMikeSdid you charge the battery overnight or anything?
14:32:14 Join ZincAlloy [0] (n=d9eee40f@gateway/web/cgi-irc/
14:32:58robin0800pixelma no 1.01.05 &1.01.06 have no msc mode 1.01.07 does again its available as an update note wiki page also says 1.01.07 wont work
14:33:20amiconnJust use an older one then...
14:33:51LloreanjhMikeS: Nothing at all. The S has been sitting on a window sill
14:34:33 Quit JdGordon (Read error: 104 (Connection reset by peer))
14:34:58robin0800pixelma 1.01.06 can go in msc mode with a key combination hold on and rewind held as you plug in usb!
14:35:08bertrikhmm, the quickscreen is voiced even when I turned off all voice options in voice settings menu
14:35:59pixelmarobin0800: heard about it and that's why I talked about a setting...
14:36:05amiconnjhMikeS: No unusual count of interrupts
14:36:31 Join JdGordon [0] (
14:37:54robin0800pixelma I have 1.01.06 working like this just curious about 1.01.07
14:39:21pixelmarobin0800: i have no idea about that version, sorry. By the way - you're statements are hard to read without any punctuation... ;)
14:39:44*amiconn wonders why robin0800 cares about OF version
14:40:32amiconnYou just need one that allows installing rockbox, and that's it
14:41:12robin0800amiconn it seems to improve usb transfer, faster!
14:41:18jhMikeSamiconn: is there any piece of code the mini and 5g share that others don't?
14:41:29 Join hannesd [0] (
14:43:05w1ll14mis it just me or is beast not good at charging the battery in rockbox?
14:43:26jhMikeSw1ll14m: it has to be tuned on and off by software and we're not doing that atm
14:46:44w1ll14mohh ok
14:47:39 Quit linuxstb_ (Read error: 101 (Network is unreachable))
14:47:46 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
14:50:44 Quit ZincAlloy ("CGI:IRC (Ping timeout)")
14:52:48 Quit robin0800 (" HydraIRC -> <- s0 d4Mn l33t |t'z 5c4rY!")
15:00:06 Quit Ave (Read error: 110 (Connection timed out))
15:02:04 Quit faemir (Remote closed the connection)
15:02:44 Join faemir [0] (
15:02:44 Quit faemir (Remote closed the connection)
15:03:33 Join faemir [0] (
15:05:15 Quit faemir (Remote closed the connection)
15:05:42 Join faemir [0] (
15:11:18 Join Ave [0] (
15:12:31*JdGordon should have kept that commit back for some contentious feature commit later :p
15:15:05 Join merbanan [0] (n=banan@
15:16:15*amiconn tries a single core build
15:22:47 Join soap [50] (n=soap@rockbox/staff/soap)
15:23:23w1ll14mit's a beast for real :| it's almost scary
15:23:39Nico_Pw1ll14m: how so?
15:24:05w1ll14mit plays a movie very smooth, even with software eq enaled:|
15:24:25BigBambiOnly running at half speed to, without other stuff
15:24:44w1ll14mit's really a nice hardware platform
15:24:51 Quit gregzx ("ChatZilla 0.9.81 [Firefox]")
15:24:54w1ll14mtoshiba did a good yob
15:26:15w1ll14mhmm when i enable eq, my ogg's will not allow me to change volume.. strange
15:28:41 Join sdoyon [0] (
15:30:40*jhMikeS notices choppiness during buffering that seems to indicate some ata trouble there (320x240 29.97 fps)
15:33:20 Join BitTorment [0] (
15:33:33bertrikJdGordon: I just noticed that the quickscreen is voiced even though voice is turned off in my settings
15:34:08amiconnjhMikeS: Problem also exists when building single core - very weird
15:34:32JdGordonbertrik: ah crap.. which of those silly talk functions doesnt check the setting :'(
15:34:34*JdGordon fixes
15:35:07*JdGordon thought they did the check
15:35:20jhMikeSit might be that the ata driver doesn't yield as much as others because it's ready for transfer almost immediately (it fills 62MB very, very fast)
15:35:43***Saving seen data "./dancer.seen"
15:39:01amiconnThe ata driver itself doesn't yield if transferring data is possible
15:39:13amiconnIt only yields when it has to wait for the disk
15:39:42amiconnHowever, the high-level code that does the buffering has to make sure it doesn't read too much at once
15:40:10JdGordonbertrik: thanks, fixed
15:40:19jhMikeSit seems to almost 100% of the time you call it. recording doesn't even do it's own yield for writing data.
15:40:36amiconnAta transfers are limited to 256 sectors anyway (unless it uses lba48, then up to 65536 sectors are possible, but rockbox doesn't do that). fat.c ensures this
15:46:50jhMikeSwhere's the info about ata_idenfify_info contents?
15:47:00jhMikeSata spec?
15:48:08amiconnWhat are you looking for?
15:49:15jhMikeSstuff about M/UDMA modes
15:50:48amiconnWell there's a number of capability and status flags for dma modes. I don't know how one is supposed to use these modes though
15:51:31jhMikeSbyte 63 has the flags
15:52:41jhMikeSthat's MDMA, 88 lists the supported UDMA modes
15:53:30amiconnThe flags aren't the problem. What I don't understand is ho wthose modes are supposed to work. How could the disk get dma access to the memory?
15:57:54jhMikeSthe controllers on both gigabeats state that they support direct transfers to and from the drive
15:59:27 Part Gekz
15:59:46 Part webtaz
16:00:36jhMikeSso I suppose you give the controller a buffer to send or request a read and the ata controller handles it, sort of like the USB controller does it.
16:01:25 Quit w1ll14m (Read error: 104 (Connection reset by peer))
16:02:59 Join ali_as_ [0] (
16:05:42 Quit ReKleSS ("Leaving")
16:06:06 Quit Ragnor (Nick collision from services.)
16:06:16 Join Ragnor [0] (
16:10:54JdGordonsdoyon: oy! mits off that delta ive reclaimed!! it was MINE!
16:13:06n1sjhMikeS: latest battery bench went down to 3,61V in the last log entry, still far from the 3,4 defined as shutoff, what do you think of going with 3,6 as low and 3,5 as shutoff for now?
16:13:38JdGordonsdoyon: also.. the plugin min api version should have been bumped if you added new items into the middle of the global_settings struct
16:14:04jhMikeSn1s: and the hardware is shutting it down?
16:14:37jhMikeSI suppose it would have to be
16:14:46n1sthis time I'm no certain but the first bench had low batt shutdown disabled
16:15:09sdoyonJdGordon: settings? Ah darn.
16:15:21*jhMikeS checks the PMIC low batt threshold value (if that exists)
16:17:28JdGordonsdoyon: if you do bump it... remember to sort any functions hanging on the end of the struct
16:17:29 Join barrywardell [0] (n=barrywar@
16:18:08sdoyonJdGordon: I suppose I should just move my fields to the end right?
16:18:53JdGordonguess that would work also
16:20:55n1sjhMikeS: it seems like the HD makes the voltage drop enough to shut it down because both bench logs end about 1 timeout before the player died
16:21:58 Quit ali_as (Read error: 110 (Connection timed out))
16:22:06n1sthinking that the pmic seems to be mostly designed for phones (no HD) maybe we need to configure something to allow such drops
16:22:35 Join fml [0] (n=4fd3d74e@gateway/web/cgi-irc/
16:22:56fmlNico_P: ping
16:23:44fmlSeems you have to introduce a new WPS tag for the study mode (just committed)
16:24:04*jhMikeS sees the gigabeat S disk supports UDMA modes 1,2,3,4 which mean max transfer rates of 66.7 MB/s.
16:25:50 Join Lear [0] (i=chatzill@rockbox/developer/lear)
16:25:57Nico_PjhMikeS: so we can fill the buffer in 1 second? :D
16:26:20 Quit faemir (Remote closed the connection)
16:26:29jhMikeSif it would ever actualize nearly that :)
16:26:46 Quit XavierGr (Nick collision from services.)
16:26:57 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
16:27:10*Nico_P didn't understand
16:27:13 Join faemir [0] (
16:29:03n1sjhMikeS: it seems like the hardwired poweroff is under 2.6V for 4ms and doesn't seem configurable
16:29:22n1sbut how can it drop that low from 3.6... ?
16:29:50*bluebrother notices another setting missing from the manual
16:29:56jhMikeSNico_P: I think alot of factors like head seeks and such would prevent that from actually happening but otoh PIO modes only support up to 16.7 MB/s but such delays would be a larger part of the time spent
16:30:41fmlbluebrother: you mean the study mode?
16:31:08Nico_Pright. I wasn't hoping to reach that figure of course, but it's still very promising. what's the max disk rate on typical current targets?
16:31:21 Quit faemir (Remote closed the connection)
16:31:25Nico_P16.7 MB/s?
16:31:37 Join faemir [0] (
16:31:41fmlWasn't it the rule that a feature is only committed with the accompanying piece of the manual?
16:32:13n1swe would never get any new feature then ;)
16:32:43 Join kushal_12_27_200 [0] (n=kushal@
16:33:23*jhMikeS proposes a rules manual and appointing someone to keep it in sync with the rules
16:33:38fmln1s: he-he! BTW: have you noticed FS #8953? It's about moving the cursor in the calculator (we talked about it a bit)
16:34:01n1sfml: yes, I'll try to take a look at it
16:34:23bluebrotherfml: yes.
16:34:53*bluebrother too slow today
16:34:59 Quit faemir (Remote closed the connection)
16:35:26bluebrotherfml: no, it's not a rule. It has been suggested to be a rule though
16:35:58fmlok, not a hard rule but something to strive for
16:36:39bluebrotheryes. But as far as I can see it's not widely known. Or mostly ignored ;-)
16:37:19sdoyonFrankly, I thought there was a 50% chance someone would start yelling because of the new context menu entry. If no one does, then I'll be nice and see how to add stuff to the manual :-).
16:37:35 Join faemir [0] (
16:37:38fml(self irony on): even in FS #8953!
16:38:34bluebrothersdoyon: if I read correctly n1s already volunteered.
16:38:50sdoyonAh good :-).
16:38:59 Quit advcomp2019 (Read error: 113 (No route to host))
16:39:16bluebrotherthe only concern with not adding desrciptions to the manual in the first place is that it mostly gets forgotten then −− and we direct users to the manual quite frequently ...
16:39:18n1si did?
16:40:03fmlbluebrother: he meant FS #8953
16:40:18bluebrothern1s: like 5 minutes ago when you hilighted fml? But it looks like you were referring to something else.
16:40:39n1sok, never mid :)
16:41:18n1sI do wonder what this talk of added contex menu ites is about though?
16:41:29sdoyonThought so... ;-)
16:44:15n1sah, in the wps context menu, well IMHO that menu is pretty much useless anyway so i don't mind :)
16:49:21sdoyonGot to run, bye
16:49:23 Quit sdoyon ("ircII EPIC4-2.6 -- Are we there yet?")
16:50:24*fml follows sdoyon
16:50:47 Quit Lear ("ChatZilla 0.9.81 [Firefox 3.0pre/2008050306]")
16:52:14 Quit fml ("CGI:IRC (EOF)")
16:53:49*amiconn wonders what the heck study mode is
16:55:40amiconn...and why it adds so much code :(
16:58:01 Quit Synergy6 ("Adios")
16:59:29 Quit faemir (Remote closed the connection)
17:01:15 Join fdinel [0] (
17:04:33 Join tessarakt [0] (
17:08:45 Join faemir [0] (
17:11:00 Join sdoyon [0] (
17:14:56 Quit Hillshum ("ChatZilla 0.9.81 [Firefox]")
17:16:57 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
17:18:25*shotofadds see that almost-green build table and gets the feeling he really should implement the D2's user timer one of these days...
17:21:17 Join moos [0] (n=c40cd9b4@gateway/web/cgi-irc/
17:23:09 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
17:24:18 Join csc` [0] (n=csc@archlinux/user/csc)
17:24:57moosLlorean: hi, what did you do to solve your beast boot problem?
17:25:19 Join ibycus [0] (
17:25:42 Quit barrywardell ()
17:25:56ibycushi everyone, is there an option in rockbox to display a generated song title taken from the database instead of the song filename when viewing the dynamic playlist?
17:26:33bluebrotheribycus: no. The playlist viewer only displays the filenames
17:26:50ibycusaah ok
17:27:19ibycustrying to sort of do a dualboot thing with rockbox and apple's firmware, but apple's thing insists on using ridiculously obscure filenames
17:27:36ibycusso my playlist is full of songs with totally unhelpful names
17:27:41ibycusthanks anyway
17:27:59bluebrotherwell, scrap the apple firmware ;-)
17:28:44bluebrotherusing tags from the database has been suggested before (there was even a patch but that was rather slow, so not suited)
17:29:50ibycuswell i never used to use the apple firmware at all
17:30:03ibycusbut i've got an itrip thing that only works with it
17:30:34ibycusso i dont have much of a choice
17:30:54bluebrotheryou might want to check the wiki −− it has a page with information about accessories.
17:31:16bluebrotheriirc the itrip only requires power, and this should be switchable now (using the debug menu)
17:31:25amiconnjhMikeS: The too long interrupt disabling happens in switch_thread(), but I don't know why
17:31:26ibycusoooooh really?
17:31:34ibycusthat's really interesting
17:32:02bluebrotherbest check for Buschel −− he did quite some work with power stuff on the Ipods.
17:32:05jhMikeSamiconn: perhaps not enough cycles between disable/enable somewhere?
17:33:06bluebrotherah, no −− it's in the system menu. You can enable accessory power there.
17:33:28amiconnIs there a reason why thread switches have to disable interrupts?
17:33:30ibycusnow that is really worth knowing
17:33:39ibycusthank god i might not have to fight with gtkpod anymore
17:34:06jhMikeSamiconn: because interrupt interact with the lists and would corrupt them in short order otherwise
17:34:27amiconnBut why?
17:34:42amiconnI'd expect interrupts to be independent from threading...
17:35:10jhMikeSnot to mention the latency issue between being destined to sleep, having the tick fire, and then having to continue to wait out a tick while a thread is on the run list
17:35:46***Saving seen data "./dancer.seen"
17:35:55jhMikeSthey're very intertwined with the scheduler
17:36:25ibycusbluebrother: where in the system menu, i can't see anything like that?
17:36:25jhMikeSI haven't seen a kernel where they're not
17:37:02amiconnThe old rockbox kernel wasn't iirc. It only used the interrupts for wakeup, and that was it
17:37:05bluebrotheribycus: Settings / General / System / Accessory Power Supply
17:37:15ibycusbluebrother: yeah just got it, thanks :-)
17:37:43ibycushow long has that been included in rockbox for?
17:37:48ibycusand is it always set to yes by default?
17:37:58jhMikeSadd/remove from a linked list isn't atomic of course and the scheduler routinely accesses locks and lists in both modes
17:38:22bluebrotherdon't remember exactly −− maybe a few weeks. And no, iirc it defaulted to no a while.
17:38:33ibycusexcellent, that's really good news
17:38:40ibycusill have to test that later on
17:38:59amiconnjhMikeS: Imo interrupts should never access the thread lists, and then there is no problem with non atomic accesses
17:39:25ibycusdoes it use any power when turned on with nothing plugged in, im guessing no?
17:39:47jhMikeSamiconn: interrupts are only allowed access to wakeup which does access way around that
17:42:37*amiconn isn't convinced at all
17:42:49amiconnThe interrupt just needs to wake up the cpu from sleeping
17:42:57amiconnEverything else can happen in normal mode
17:44:04amiconnCheck whether a thread is ready to run, and if not, go to sleep again
17:45:00jhMikeSit does...but if a thread becomes ready to run via an interrupt while it's headed for a sleep, you'll end up waiting out a whole tick with a thread on the run list which isn't good at all
17:45:00 Join toffe82 [0] (
17:45:14bluebrotheribycus: well, I guess it's taking a little power −− otherwise it wouldn't make much sense adding an option for turning it off.
17:45:34bluebrotherbut I don't own any accessories so I'm not too familiar with that stuff.
17:47:29jhMikeSsleeping the core with a thread on the run list is a state that's never wanted so it's made to be atomic with interrupts and both cores
17:47:50*amiconn thinks it would be better to *occasionally* sleep for one tick that to fiddle excessively with interrupt enabling & disabling
17:48:22 Part ibycus
17:48:43 Quit sdoyon ("ircII EPIC4-2.6 -- Are we there yet?")
17:49:29 Join saratoga [0] (n=41becb3b@gateway/web/cgi-irc/
17:53:13*jhMikeS would prefer to guarantee a few uS latency at most esp. with rapid-fire interrupt wakeups like OTG transfers
17:54:27amiconnIf it needs to be really fast you will react to the interrupt in the isr, not in a thread
17:55:42amiconnBtw, is there a reason why interrupts are enabled *after* core_sleep?
17:55:42 Join Synergy6 [0] (
17:56:10amiconnThis seems to cause the observed problem, although I'm not sure yet
17:59:19amiconnWhile ata waits for the disk there seems to be no runnable thread for quite a number of iterations within the for(;;) loop in switch_thread
17:59:56jhMikeSbecause the are masked at core level during the sleep and is fine on the majority of targets since they wake the core anway. this makes it atomic to have it pass through the sleep unimpeded if the interrupt fired while detined for it.
18:00:19 Quit BitTorment ("Do directly to guantanamo bay. Do not pass go. Do not collect your human rights.")
18:00:31 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
18:00:54jhMikeStagets that don't allow wakeup with core level masking were carefully crafted to have it work
18:01:34amiconnMoving the enable_irq() before the actual core sleep reduces the problem, but doesn't fix it
18:02:10jhMikeSspinning in a yield loop should never reach the sleep code since that thread is never taken off the run list
18:03:15 Quit Ave (Read error: 110 (Connection timed out))
18:03:31jhMikeSit mostly just ends up selecting the same thread again since that's likely the only one awake
18:04:54amiconnWell, for some reason the 'impossible' seems to happen...
18:04:54 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
18:05:10jhMikeSbut how far back did you say this problem went?
18:06:03amiconnr15444 made it show up
18:06:19amiconnBefore that we had no disk poweroff on ipods
18:12:03 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
18:12:28jhMikeSat that time and switch_thread didn't routinely disable interrupts at the revision. there was a queue in between interrupt wakes and the run list but this was actually slower and more complicated to implement so most calls in that loop wouldn't disable them except for certain occasional ops.
18:13:46jhMikeSwait, this happens in the yield loop? /me isn't quite with it today
18:14:09amiconnI think so, but I'm not sure yet
18:15:37amiconnThis is really weird. What happens is that rockbox seems to talk to the disk too early after wake-up. There are no ata errors in the init sequence though, but there must be a few retries in the following access, which disables interrupts for too long
18:15:44amiconn(more than 5 ticks)
18:16:19amiconnNow the question is how ata retries manage to disable interrupts for so long.
18:17:28amiconnThe relevant places are in switch_thread(), as I found by instrumenting disable_interrupt(), enable_interrupt(), disable_interrupt_save(), and restore_interrupt()
18:18:01amiconnHmm, actually I can imporve that instrumentation....
18:21:52mcuelenaeredoes someone know how I could test if I have access to a TLV320AIC23? (without IS/DSP, only IC)
18:22:35*preglow wonders why the study mode setting is called justy that
18:22:58preglowseems like a really poor name for what it does
18:23:16preglownot that i like the setting much at all anyway
18:24:03jhMikeSwhat does it do?
18:24:23preglowit seems to change the functionality of right/left to skip some seconds back/forward instead of skip track
18:24:59jhMikeSit's one of those things like "Caption Backlight" which makes one go "wtf is does that mean" :)
18:25:14preglowi don't see the need for anything like this anyway
18:25:20preglowif you don't want to skip tracks: use the hold switch
18:25:24bluebrotherI remember my old SonicBlue player having a mode called study. But that was something completely different: it skipped back a couple of seconds when resuming
18:26:14amiconnjhMikeS: The irq disable that holds for too long definitely happens at thread.c line 1970, and the enable that catches it (with my instrumentation) happens at line 1981
18:28:42jhMikeShow is that code being reached in a yield loop?
18:30:21 Join Buschel_away [0] (
18:30:27jhMikeS...or it doesn't get reached until something sleeps and then the enable occurs
18:30:54jhMikeSmore accurately until everything sleeps at least once
18:32:10 Quit rp- ("leaving")
18:33:22 Join Av3 [0] (
18:34:56 Quit kushal_12_27_200 ("Leaving")
18:35:11 Nick Buschel_away is now known as Buschel (
18:36:45 Join simonrvn_ [0] (i=simon@unaffiliated/simonrvn)
18:36:45 Quit Buschel (Read error: 104 (Connection reset by peer))
18:36:53 Join Buschel [0] (
18:37:19 Quit simonrvn (Read error: 113 (No route to host))
18:37:24 Nick simonrvn_ is now known as simonrvn (i=simon@unaffiliated/simonrvn)
18:39:32 Join Buschel_ [0] (
18:39:32 Quit Buschel (Read error: 104 (Connection reset by peer))
18:39:32 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
18:40:09 Quit moos ("CGI:IRC (Ping timeout)")
18:41:08Nico_PjhMikeS: what's the typical disk transfer rate for out current targets?
18:41:54jhMikeSamiconn: you're saying there's not an interrupt enable for at least five ticks after line 1970?
18:41:59 Join Xerion_ [0] (
18:42:19jhMikeSNico_P: There's Wiki page about that with extensive data on that
18:42:24 Quit Xerion (Read error: 104 (Connection reset by peer))
18:42:27 Nick Xerion_ is now known as Xerion (
18:43:03*Nico_P goes looking for the wiki page
18:43:06Nico_Pah, thanks
18:46:03 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
18:46:36amiconnjhMikeS: During normal operation, all is well. The problem happens only when the disk needs to spin up again after a power-off.
18:48:02amiconn...and then only because we're talking to the disk too early, so it needs to perform retries. The retries itself succeed though
18:48:43amiconnAll this causes a short dropout n the music, even if the spinup is totally unrelated to buffering (e.g. loading a plugin, or browsing the disk with dircache disabled)
18:48:56jhMikeSI not certain if you mean by "holds for to long" means that irq_enable isn't called again for a long time
18:49:13amiconnThat's what I mean
18:49:46amiconnMy instrumentation stores the location and USEC_TIMER value for the last disable_irq() or disable_irq_save()
18:50:28amiconnOn enable, it checks whether more than 50000 usecs have passed since the last disable, and panics if so, showing the location of the disable and the location of the enable
18:51:03jhMikeSthe only place that should be possible is that the core wake is somehow not happening for > 5 ticks in which case the core isn't registering the tick interrupt for > 5 ticks. the timer interrupt pending should wake it on the next tick and then it should enable them and service the tick.
18:53:08jhMikeSthat sequence definitely happens normally but the power on seems to have a side effect on it
18:53:47amiconnIt's not the power-on itself
18:54:25amiconnIf I extend the sleep() delay after ata power-on, the effect doesn't happen
18:54:30 Quit mf0102 (Remote closed the connection)
18:54:55jhMikeStrue, but it appears that attempt to access the disk happens too early then something messes up registering the tick
18:56:56*amiconn tries with simple round-robin scheduling
18:57:15amiconnUnfortunately disabling HAVE_RIORITY_SCHEDULING in config.h doesn't work
18:57:36amiconnThe build breaks in thread.c when doing that
18:58:32 Quit XavierGr (Nick collision from services.)
18:58:43 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
18:58:59jhMikeSarchos builds without it so I guess something isn't in sync for swcodec
18:59:24XavierGrJdGordon: you called?
19:02:03jhMikeSamiconn: the priority system did change pretty radically since 15444 though - mostly new code from then
19:02:07XavierGrbertrik: as JdGordon said, if low voltage triggers a safe shutdown, disk activity should cause battery bench to flush all data to the text file.
19:02:33XavierGrbertrik: if not, (player shutsdown because of no more juice) the last measurement will be lost
19:02:40 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
19:03:01amiconnXavierGr: Problem is here that low-voltage shutdown prohibits disk writes
19:03:05XavierGrbertrik: I don't think there is a way to save those last measurements on the second case, but if you have an idea please speak
19:03:34XavierGramiconn: you mean that the safe shutdown feature won't try to save settings etc?
19:03:53XavierGrsafe shutdown = the one that rockbox triggers even if the player has some juice left
19:04:24jhMikeSamiconn: change thread.h line 328 to #if NUM_CORES > 1 which it should be
19:05:18jhMikeSactually, the #ifdef isn't even needed around that at all
19:07:18bertrikXavierGr: yes, but does a safe low voltage shutdown trigger disk activity?
19:07:41bertrikXavierGr: for flash targets we should flush the cache often as there is no penalty of spinup
19:08:24*jhMikeS also wonders if TIMER1 just somehow fails to run for too long
19:08:24amiconnbertrik: Too much special casing....
19:10:08 Part blistov ("Kopete 0.12.4 :")
19:10:16 Quit fyrestorm (Read error: 113 (No route to host))
19:11:30XavierGrbertrik: I don't think we need to flush often even on flash targets maybe an exception on safe shutdown for them
19:12:06XavierGrbertrik: keep in mind that we need battery_bench to be as transparent as it can be in order to avoid altering the results cause of the plugin itslef
19:12:14 Join moos [0] (n=c40cd9b4@gateway/web/cgi-irc/
19:12:14amiconnShutdown should call the ata callback anyway unless it's critical batt level
19:12:37XavierGramiconn: afaik it is (normal shutdown) but I am not sure about the safe shutdown
19:13:07 Nick Buschel_ is now known as Buschel (
19:13:38bertrikXavierGr: I think there's not much of a penalty for flash targets to flush often
19:14:25jhMikeSamiconn: fixed
19:15:08amiconnbertrik: That's not the point. There is no reason to make battery_bench behave different on flash targets than on ata targets.
19:16:43pixelmabertrik: I don't think (and wouldn't like for some reason) that the plugin behaves differently just because it runs on a flash target. That's also one of the reasons why I haven't used battery_bench with the hack on my c200 yet
19:16:48amiconnjhMikeS: This is really puzzling. I almost suspect something is calling yield() in isr context...
19:18:24bertrikhmm, not the point? the goal of the battery bench plugin is to collect battery data. Missing data for some reason directly interferes with its main purpose.
19:18:34jhMikeSverify the mode bits are supervisor then
19:18:42XavierGrbertrik: also we are talking of a very small error margin, 30 minutes max with 128kbps mp3s on a 32mb ram target
19:18:48*bertrik checks whether flash targets even support the ata notify calback
19:18:52XavierGragain in the worst case
19:19:31amiconnbertrik: Yes they do, and they have to
19:19:47XavierGrbertirk: forcefully writing data is also interfering with its main purpose (although I agree that it is lesser for flash targets still...)
19:20:16 Join fml [0] (n=4fd3c831@gateway/web/cgi-irc/
19:20:58amiconnIt also costs a little bit battery
19:21:02bertrikwe were talking about flash targets :P
19:21:35amiconnWhile there is no spinup, writing every measurement mains that sectors are rewritten many more times
19:21:57Buschelamiconn/XavierGr: from my understanding the write-protection is done if voltage is below battery_level_dangerous[] whereas the safe-shutdown is triggered by battery_level_shutoff[]. a solution can be to set the write-protection limit below the safe-shutdown limit
19:22:03fmlWhy would putting some vars in the middle of the settings structure make plugin API bump necessary (see one of the recent commit comments)?
19:22:18 Join simonrvn_ [0] (i=simon@unaffiliated/simonrvn)
19:22:29amiconnfml: The settings struct is part of the plugin api
19:23:23 Join styleism [0] (
19:23:53fmlamiconn: the struct itself or its address?
19:24:01Buschelamiconn/XavierGr: these voltage levels can be defined for specific targets
19:24:17amiconnIts address of course
19:24:45bertrikamiconn: I think that writing sectors multiple times is not a big issue. There's built-in wear leveling AFAIK and it's not _that_ often (say once a minute)
19:25:02bertrikbut OK, I won't mind keeping it little simpler
19:25:15amiconnIt's not about wear
19:25:27amiconnIt's about extra power consumption
19:25:39fmlamiconn: but then adding something in the middle should do no harm?
19:26:03amiconnfml: Of course it does. Plugins rely on the api versioning to ensure it's compatible
19:26:28amiconnAnd if you change a struct that's exposed to the api, it does have an effect on compatibility
19:27:18 Quit simonrvn (Nick collision from services.)
19:27:19 Nick simonrvn_ is now known as simonrvn (i=simon@unaffiliated/simonrvn)
19:27:37*bertrik highly doubts the power consumption arguments but has no measurements to back it up
19:27:37amiconnjhMikeS: Simple scheduling doesn't solve the problem btw
19:27:47fmlamiconn: ok, I git it!
19:28:26*Buschel just saw he added a var in the middle of this structure with r17192 :/
19:28:29 Join crzyboyster [0] (n=4b596e38@gateway/web/cgi-irc/
19:29:00pixelmabertrik: what would be simpler in having a code path for hd based targets and another one for flash targets?
19:30:09bertrikI don't know until I try it out, but I think it would not be simpler
19:30:21crzyboysterQuick question: What exactly is the quickscreen?
19:30:57bertrikQuick answer: something that's probably described in the manual :P
19:31:37 Quit crzyboyster (Client Quit)
19:32:49pixelmaquick leave...
19:33:03*bertrik was just typing the slightly longer answer
19:33:27jhMikeSthe build table is starting to look unusually green
19:34:05 Quit saratoga ("CGI:IRC (EOF)")
19:35:49***Saving seen data "./dancer.seen"
19:36:18 Quit moos ("CGI:IRC (EOF)")
19:37:10Buschelif it is needed to correct my faulty submitted with r17193, someone must take over −− got no svn-access today :(
19:37:44Buscheli added a new var within (and not at the end of) the user settings structure
19:38:46*Buschel creates lots of typos if his son is crying...
19:43:22 Join K4rP4D [0] (n=KrPD@unaffiliated/krpd)
19:47:31soapIs it not true that the (relatively) large voltage reading between a LiIon @ low (no drive spinning) load and at higher (drive spinning) load is inevitably going to cause a problem writing the last battery-bench results to a file? When the drive isn't spinning the battery appears fine, but as soon as you spin the disk you suddenly go below the critical level. It appears to me that my batteries' (on all my players) voltage reading become even more load-
19:47:31soapsensitive as they age.
19:48:16*bluebrother just needed a couple of edits only because the preview doesn't show the website menu to get the layout right on RockboxUtility :(
19:49:58 Join |AhIoRoS| [0] (n=ahioros@
19:49:59soapBoth my Gigabeat and Video now show a near 10% point drop of battery level on sustained drive accesses, only to slowly creep back up after the drive stops.
19:50:22bluebrotherat least that page is now nicer.
19:50:43Nico_Pbluebrother: yes, it looks good :)
19:51:25bluebrothernow the only thing I'm still disliking is the bad scaling of the images done by my browser. We'll see ...
19:52:32bluebrotherI tried the old wx version again a couple of days ago. Fascinating to see the difference ;-)
19:53:08*Nico_P doesn't understand why in the HTML manual, the navigations links are ordered [next] [prev]...
19:56:20fmlamiconn, Buschel: I think adding a var in the middle of settings struct (and bumping plugins API version) is ok since it's better to have well organized code. And plugins are compiled all together anyway!
20:01:09n1ssoap: yup, voltages tend to drop a bit when a disk starts sucking a lot of current (especially when spinning up)
20:01:33 Nick Av3 is now known as Ave (
20:02:02 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
20:03:02mcuelenaereI'm confused about copy_read_sectors(buf, wordcount) (called in ata_read_sectors) -> wordcount means the amount of bytes*4 to be transferred, right?
20:03:34 Join KrPD [0] (n=KrPD@unaffiliated/krpd)
20:03:52 Quit hannesd (Read error: 110 (Connection timed out))
20:03:57 Quit KrPD (Connection reset by peer)
20:04:09Buschelfml: i didn't bump the api version
20:04:10 Quit fml ("CGI:IRC (EOF)")
20:04:15 Quit Mathiasdm ("Yuuw!")
20:16:16 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
20:17:20 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
20:19:28linuxstb_mcuelenaere: ATA words are 16-bit - I think that's what copy_read_sectors expects - i.e. bytes/2
20:19:43linuxstb_mcuelenaere: See the C implementation of that function in ata.c
20:20:14 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
20:21:42 Quit K4rP4D (Connection timed out)
20:22:00amiconnmcuelenaere: No, it means ata words, which are 16 bits
20:22:22mcuelenaerelinuxstb was first :)
20:24:12 Join Av3 [0] (
20:25:49 Join K4rP4D [0] (n=KrPD@unaffiliated/krpd)
20:26:02 Quit amiconn (" bbl")
20:28:32 Quit druidbartek (Read error: 110 (Connection timed out))
20:28:36 Quit Horscht ("IRC is just multiplayer notepad")
20:28:40 Join gromozekin [0] (n=gtfo@
20:28:51gromozekinhi there. i have weird idea man
20:29:27gromozekinis anybody there
20:29:31gromozekinhow can listen 2 me?
20:29:53bluebrotherwell, looking at the channel I see 135 people around right now.
20:29:56pixelmabluebrother: will you be around later too. I could need a tester with an m:robe100 but won't be around for a while...
20:30:17pixelmashould have been a question ;)
20:30:47gromozekinso, is it real to port rockbox on creative zen xtra player with big big hdd)
20:30:49bluebrotherpixelma: thought that ;-) I'll be around for a while now (well, unless something unexperienced happens ;-)
20:31:01 Quit Ave (Read error: 110 (Connection timed out))
20:31:01 Nick Av3 is now known as Ave (
20:31:15pixelmanice, until later then
20:31:28 Part pixelma
20:31:37 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
20:32:50bluebrothergromozekin: Rockbox is currently being ported to the Zen Vision:M.
20:35:02bluebrotherbut as far as I know this doesn't imply the xtra in any way.
20:35:30bluebrotherthe NewPorts forums has all information about new ports.
20:36:05 Quit BlakeJohnson86 ("Leaving.")
20:36:07 Quit Synergy6 ("Adios")
20:37:23 Join BlakeJohnson86 [0] (
20:37:58 Join robin0800 [0] (
20:40:30gromozekinzen vision m
20:40:34gromozekinis coloured?
20:40:54 Quit tedr0ck (Read error: 104 (Connection reset by peer))
20:42:17 Join BlakeJohnson87 [0] (
20:43:28gromozekinbut my player is b/w
20:43:32 Join dabujo [0] (
20:43:41gromozekinand it connects only through special prog
20:44:27bluebrotherwell, the Zen Vision:M is obviously not a Zen Xtra.
20:45:03gromozekini know)
20:46:24 Join saratoga [0] (n=9803c50e@gateway/web/cgi-irc/
20:51:01 Join tedrock [0] (
20:55:08 Quit BlakeJohnson86 (Read error: 110 (Connection timed out))
20:58:47 Join lee-qid [0] (
21:02:10 Quit gromozekin (Remote closed the connection)
21:05:59 Quit XavierGr (Nick collision from services.)
21:06:12 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
21:10:32 Quit spiorf (Remote closed the connection)
21:16:31 Quit XavierGr (Nick collision from services.)
21:16:31 Quit BlakeJohnson87 (Read error: 104 (Connection reset by peer))
21:16:39 Quit SUSaiyan ()
21:16:42 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
21:17:41 Join BlakeJohnson86 [0] (
21:19:10 Quit mcuelenaere ()
21:19:46 Join Horschti [0] (
21:20:32 Quit Horscht (Nick collision from services.)
21:20:40 Nick Horschti is now known as Horscht (n=Horscht@xbmc/user/horscht)
21:35:50***Saving seen data "./dancer.seen"
21:36:57 Join DataGhost [0] (n=dataghos@unaffiliated/dataghost)
21:49:11 Join bughunter2 [0] (
21:49:32 Join barrywardell [0] (n=barrywar@
21:49:49 Join Mathiasdm [0] (
21:50:38 Quit XavierGr (Nick collision from services.)
21:50:49 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
21:55:02 Join Bright-SUSaiyan [0] (
21:55:22 Quit Xerion (Read error: 113 (No route to host))
21:57:11 Quit desowin ("KVIrc 4.0.0 Insomnia")
21:57:53 Quit Horscht ("IRC is just multiplayer notepad")
21:58:19 Join m0f0x [0] (
22:05:49 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
22:07:16 Join chrisjs169 [0] (n=chrisjs1@unaffiliated/chrisjs169)
22:08:49chrisjs169question - I'm finally about to add the anti-aliased fonts patch to the tracker - should I have both the patch and the convttf file in one task, or should I keep them separate?
22:12:29 Quit barrywardell ()
22:12:56linuxstbWhat's the convttf file, and how is it related to the patch?
22:15:15 Join gregzx [0] (
22:15:52 Join OlivierBorowski [0] (
22:16:43 Quit bertrik ("bye")
22:17:32n1shmm, inserting the headphone connector while the gigabeast is "off" turns it on
22:18:54*n1s has working wake up alarm :)
22:19:46 Part Buschel
22:20:50 Quit lee-qid ("aufwiederbyebientotsayonara")
22:20:53Nico_Pchrisjs169: you should probably have them under the same task if one requires the other
22:22:16 Quit OlivierBorowski (Remote closed the connection)
22:22:49 Join SirFunk [0] (
22:23:00LloreanI assume convttf is the tool to create anti-aliased fonts from ttf files?
22:25:53 Join MegaRain [0] (n=chatzill@
22:26:11 Quit MegaRain (Client Quit)
22:27:05 Join Megarain [0] (n=Rainer_P@
22:27:53chrisjs169Nico_P, ok
22:28:29Nico_Pthanks for posting it btw
22:29:21 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
22:29:47 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
22:30:28MegarainHey, I'm having trouble restoring my iPod(4gen, grayscale, 20gig) so I can put rockbox on it, yet every time I use iTunes, its just corrupts before I can even connect it to the machine. Ideas?
22:30:51n1sbroken drive maybe?
22:32:08MegarainIt was working the other day, and the device still boots up, etc. And it will go into device mode. Is there an easy way to diagnose this?
22:35:43 Quit Horscht (Remote closed the connection)
22:35:54n1swhat do you mean when you say it corrupts then?
22:37:51MegarainStep1: Restore with iTunes
22:37:51MegarainStep2: Plug into wall, select language, looks like everything is Fine.
22:37:51MegarainStep3: Plug back into PC, get error saying that iPod may be corrupted, ether unplug or restore.
22:38:33n1swhat happens if you disable the itunes thingy and run a regular disk scan on it?
22:39:03 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
22:39:21n1sis the step 3 error coming from itunes?
22:39:23MegarainNotes: OS is vista SP1
22:39:59n1scan you access the partition in vista (like a regular harddrive) ?
22:40:22 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
22:40:32MegarainYes. Says its apple iPod, drive letter D
22:40:56n1srun a disk scan on that and check for bad sectors
22:42:02MegarainHold on, going to force it into disk mode and then try again.
22:42:04bluebrotherwasn't there an issue with Itunes on Vista?
22:42:24bluebrotherdoing stuff like that: telling you you'd need to restore the Ipod?
22:42:27MegarainThat's what I was wondering, one reason I was going to use Rockbox.
22:42:34*n1s has never used an ipod nor vista and only itunes while drunk :)
22:43:07amiconnjhMikeS: Still around?
22:44:46bluebrothergoogle needed like a fraction of a second ...
22:44:51n1swhy do we only allow alarm wake up to be set in hours and minutes and not days too?
22:45:06 Join Xerion [0] (
22:45:10pixelmabluebrother: I drew a new cardset for the m:robe100 (used in solitaire and blackjack) - would be cool if someone with the real device could try it out and give an opinion...
22:45:23bluebrotherpixelma: sure.
22:46:23MegarainHey Blue, I've got itunes 7.6, and that patch doesn't do anything in SP1.
22:46:37 Quit styleism (Read error: 110 (Connection timed out))
22:49:39 Quit dabujo ("( :: NoNameScript 4.2 :: )")
22:50:19 Quit tessarakt ("Client exiting")
22:53:13 Quit n1s ()
22:53:17 Quit Horscht ("We don't make mistakes, we just have happy little accidents")
22:53:54MegarainHerm.......though I would double check to make sure I had that update anyways, and I do.
22:58:50MegarainOk, now I only get error 69. Good.
22:59:03 Quit Zarggg (Read error: 104 (Connection reset by peer))
22:59:03 Join countrymonkey [0] (n=4b05639a@gateway/web/cgi-irc/
22:59:49chrisjs169linuxstb, back, sorry. convttf is needed to convert ttf fonts into the anti-aliased version of the fnt file
22:59:51countrymonkeyWhy are new lang strings popping up so frequently these days? Yesterday, I made a chinese update (8957) and today I made another one (8960). Why?
23:00:20countrymonkeyIt seems weird that no new changes were made in feb or mar, but now in apr and may there are so many.
23:00:25 Join Zarggg [0] (
23:00:39krazykitcountrymonkey, perhaps because new features are being added that need to be voiced?
23:00:52countrymonkeyYes, 8960 is another update that needs to be committed.
23:01:02*gevaerts thinks that it's probably spring bringing new things
23:01:17pixelmawell, development isn't predictable...
23:01:30pixelmaor not much ;)
23:01:48 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
23:02:14Lloreancountrymonkey: It looks like you've removed the unknown string in your patch. is there a reason for this?
23:02:18countrymonkeyIt just strikes me as odd that we took such a break, and then all of a sudden things are popping up like crasy. And, I had a question. I cannot type chinese on my computer so can I keep translating the phrases with google translate and verify they are correct by reading the caricters?
23:02:57LloreanPeople develop in there spare time. There's absolutely no reason to expect that all months should be equal. Especially since many new features don't require new strings, so it's more or less random when new ones are needed...
23:03:13countrymonkeyThat was from yesterday's commit.
23:04:47LloreanWhich commit removes the unknown sting?
23:04:55countrymonkeyHold on, it is in my source tree.
23:05:02 Part Megarain
23:05:28countrymonkeyNever mind. It is in my source tree.
23:05:28LloreanNone of the recent changes to english.lang seem to according to the official changelog. They all just add strings, as far as I can see
23:05:53LloreanYou really need to be more careful with your patches.
23:06:19countrymonkeyHmmmmm. That's weird. I'll make a new patch as to what I have now in my source tree.
23:07:04 Quit Horscht ("I am root. If you see me laughing, you better have a backup")
23:07:10LloreanI must admit that basic things like that make me somewhat uncomfortable about trusting your translations. Since I can't verify them myself anyway, the fact that you have shown signs of being somewhat lax in the past in terms of getting strings right, etc, makes me worried about them.
23:07:26LloreanEspecially if you're just using google translate, then reading over them and saying "that's good enough"
23:09:29amiconnLlorean: Didn't JdGordon deprecate the SYSFONT strings for the quickscreen?
23:09:46 Quit domonoky (Read error: 104 (Connection reset by peer))
23:11:13amiconnjhMikeS: I never expected that switch_thread() is called with interrupts disabled that often... Something is extremely fishy here
23:11:27countrymonkeyI won't put one in unless it is exactly what I wanted. Take "enable study mode" for an example. Google translate gave me "da kai shre san de shren ze" (just sounding it out) but what I wanted was "kai shre san de shren ze". What I did was simple strip the "da" off "da kai" and I got what I wanted and put it in.
23:12:15 Join Megarain [0] (n=Rainer_P@
23:12:32MegarainI keep getting an error that I can't corrrectly directory. Ideas?
23:12:47countrymonkeyI don't understand.
23:13:20MegarainHold on, I'll get the exact wording.
23:13:28Lloreanamiconn: According the comment, unknown is also supposed to be used, for example, when time is unknown.
23:13:38 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
23:14:34amiconnLlorean: I was just referring to your remark that the recent changes were only adding strings
23:14:42countrymonkeyIs that it's only purpose?
23:15:20countrymonkeywhen time is unknown? If so, I have a bad translation and need to fix it.
23:15:45Lloreancountrymonkey: Did you read the description?
23:15:47Lloreanamiconn: Ah, okya
23:16:13Llorean"generic string for unknown states, such as an unset clock"
23:16:17countrymonkeyI must have missed something. Hold on, I will go back...
23:16:35bluebrothercountrymonkey: if you know what you want, why are you using google translator anyway? This doesn't sound trustworthy to me ...
23:17:10LloreanI'm also now *more* worried if you're not reading the descriptions of the strings, and just trying to translate them without context.
23:18:24BigBambiIf you know the language well enough to translate, why do you need google translate at all?
23:18:28 Join crope`` [0] (
23:18:57countrymonkeySo, reject the patch and wait for a new translator?
23:19:17countrymonkeyI cannot "type" the characters on my western keyboard.
23:19:46BigBambiSo you are just using it to get the characters? If so, OK, I misunderstood
23:20:04countrymonkeyYes, I am using it to get the charicters
23:21:54pixelmabluebrother: what I wanted to ask you about rbutil... does it give any guidance for those who use it to install the bootloader on an coldfire Iriver or an Iaudio which need some more steps than just copy a file somewhere (simplified a lot)?
23:22:36countrymonkeyTrust me, if I could type the characters, I would do it. Using google translate is difficult.
23:22:57countrymonkeyespecially when you want "exactly" and not "good enough" like I do.
23:23:44countrymonkeyI want "exactly".
23:25:01pixelmareally? For all languages you try to translate?
23:25:58countrymonkeyno! Just chinese because I know it. The only reason I could do bulgarian yesterday was because it was all string deprication
23:26:59 Quit crope` (Read error: 110 (Connection timed out))
23:27:17pixelmayes, noticed that :)
23:28:07MegarainI'm a little confused, how do I add music once rockbox is set up?
23:28:45bluebrotherpixelma: yes, it adds the needed stuff to the progress list. But maybe I should recheck that −− I bet there is room for improvement ;-)
23:29:05countrymonkeyThat "unknown" thing was just accident and looking at the description plus the comments that were made, it is a good translation. Looking at your comments alone, it looked bad, but reading the description cleared things up.
23:29:19 Quit K4rP4D ("Leaving")
23:29:28countrymonkeyYou put the music on the player like you do a usb flash stick.
23:29:51pixelmabluebrother: I should probably use it once to try out on my M5, one of these days...
23:30:15bluebrotherMegarain: you _can_ use itunes if you want to. That will scramble your filenames so you need to use the database.
23:30:23bluebrotherbut most people prefer drag and drop ;-)
23:30:53bluebrotherpixelma: reminds me that I don't know of any way distinguishing between X5 and X5V.
23:31:01*bluebrother puts not on todo list
23:31:16countrymonkeyItunes only works on ipods.
23:31:17 Join bertrik [0] (
23:31:31bluebrotherwhy are those lists only getting longer?
23:31:53pixelmayou said you were "not" adding it ;)
23:32:09bluebrothercountrymonkey: he was talking about Itunes earlier ...
23:32:24bluebrotherso I think it's safe to assume his player is an Ipod.
23:32:39bluebrotherpixelma: right ... should do that instead ;-)
23:33:44*amiconn needs a thread.c wizard
23:33:45BagderSendBootme()::Error on SendUDP() call
23:33:53Bagderthat's an interesting string (from a Zune)
23:33:55amiconnObviously I'm unable to understand how it works :(
23:34:20pixelmabluebrother: part of the problem is that it seems no X5V user is available. If I remember correctly someone wanted to check whether radio detection (or something like that) works on the X5s and didn't find someone with a V model without radio...
23:34:38 Quit Horscht (Nick collision from services.)
23:34:55 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
23:34:58 Quit ender` (Read error: 104 (Connection reset by peer))
23:35:47faemircan anyone here recommend me a good ipod replacement online store?
23:35:54***Saving seen data "./dancer.seen"
23:35:56faemiras in, battery/harddrive parts.
23:36:18countrymonkeyan ipod itself or accessories? Rockbox compatible or don't care?
23:36:37faemiroh sorry, for 5.5gen
23:36:53faemirso presumably new batteries and harddrives then do work with rockbox? or do you have to get the right pones?
23:36:54faemirones *
23:37:04faemirbut yes rockbox compatible :P
23:38:09countrymonkeyYou have to find the exact part. I would say, ipodjuice for batteries, and ebay for everything.
23:38:59faemirwhat happens if you don't have the exact part? I mean, i've seen 'bigger' batteries for sale with longer life
23:39:03faemirbigger capacity *
23:39:52MegarainHey, I have a question,if I formated my iPod's disc, will that brick it?
23:40:00 Join ender` [0] (
23:40:09countrymonkeyThat's ok. But I meant exact type and brand. A little bigger is okay, I have heard. I never tried it myself.
23:40:24countrymonkeyYou need to format with itunes.
23:40:33MegarainFigured as much.
23:40:42Lloreancountrymonkey: Brand is irrelevant.
23:41:09gevaertscountrymonkey: why ? mkdosfs works fine
23:41:12LloreanAnd you don't need to format with iTunes, there are ways listed in the wiki of preparing an iPod drive's partition layout without the use of itunes.
23:41:30countrymonkeyIt is? Okay. Anyway, the HarddriveReplacement wikipage has all you need.
23:42:19countrymonkeyI know. But that is the only appley official way.
23:42:55*gevaerts didn't know that we cared about what apple thinks
23:43:20countrymonkeyI don't either, but...
23:43:43LloreanThe "official apple" way is more or less irrelevant since this is #Rockbox, and what's relevant is "thinks *we* are okay with"
23:45:31countrymonkeyWe are okay with every hack and crack you can think of (unless it will cause harmful effects later).
23:46:49bluebrotherMegarain: Ipods are (almost) unbrickable
23:47:26 Quit simonrvn ("bbl")
23:47:42MegarainWell bluebrother, I need a way to completely clear the hard drive on my ipod. I have a feeling the reason it keeps corrupting is that there is junk data on the device.
23:47:50 Join webguest45 [0] (n=51d10a99@gateway/web/cgi-irc/
23:48:05 Quit webguest45 (Client Quit)
23:48:26 Quit DerDome ("Leaving.")
23:48:33bluebrotherMegarain: you can safely put it into emergency disc mode and do whatever you like with the data on the disk. Just be aware that to boot up the partition layout and the apple bootloader (i.e. the apple firmware) has to be present.
23:48:44countrymonkeyYou can format it, both ways will do, although I'll admit I use the automatic itunesian way.
23:49:16bluebrotherin fact, I managed to trick the Ipod accept a CF card by "restoring" to that card first (manually)
23:49:24Lloreancountrymonkey: No, we aren't okay with "everything ever." We respect copyright, and plenty of "hacks and cracks" as you put it, don't. Please, stop just saying everything that comes to mind.
23:49:33Megarainbluebrother: So, if I put it into disc mode, I can then format the disc....?
23:49:35Bagderthe recovery.bin file within the zune fat fs is interesting
23:49:38LloreanWe are also especially not okay with unofficial install procedures, bug reports on unofficial builds, etc.
23:49:42MegarainHow do you manually restore?
23:50:06 Quit chrisjs169 ("Leaving")
23:50:24bluebrotherMegarain: yes. We have the information in the wiki −− IpodManualRestore
23:50:34 Quit countrymonkey ("CGI:IRC (EOF)")
23:50:42bluebrotherbut those are based on linux.
23:50:43MegarainI'll go take a look
23:51:55 Quit faemir (Remote closed the connection)
23:52:16 Quit amiconn (Nick collision from services.)
23:52:22 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
23:52:59 Join faemir [0] (
23:54:30Megarainbluebrother: Any non-*nix instructions?
23:55:31bluebrothernone that I know of. There's a dd port to windows though
23:55:32 Quit XavierGr (Read error: 104 (Connection reset by peer))
23:55:46bluebrotherbut I guess it's much easier if you just use a linux live CD for restoring.

Previous day | Next day