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 2008-01-30

00:00:11einhirnamiconn: (or some other audible volume difference)
00:00:22amiconnHmm. Are you using analog volume or the digital stereo matrix?
00:01:04einhirnamiconn: digital stero matrix - the effect on SPDIF was mission critical for me ;)
00:04:38amiconnI guess the glitches happen when the track change happens during rebuffering
00:05:04JdGordon|wis the forums still acting up for anyone? is am i haaving wierd proxy issues or something? new posts not being displayed even after refreshing
00:05:08amiconnWhile rebuffering, the mpeg thread delays event processing a bit, due to the bitswap
00:05:22 Quit DefineByte ("Bye all")
00:05:32einhirnamiconn: that would explain it...
00:06:05amiconnAnother reason to try a playback engine unification... if only the swcodec engine would be more stable... :/
00:06:24einhirnamiconn: thats why I thought about a "priority message" to the queue.
00:06:57amiconnA priority message wouldn't help - the delay happens even with a single message in the queue
00:07:13amiconnThe mpeg thread simply doesn't process *any* message for a short while
00:07:22einhirnamiconn: yes, that's what I figured...
00:07:32 Quit davina (Remote closed the connection)
00:07:53einhirnMaybe the replaygain change could be posted to another queue that isn't busy with bitswapping?
00:08:32amiconnThere is no other queue - the mpeg thread is the only thread handling hwcodec playback
00:08:47JdGordon|wamiconn: you have it backwards.... merge swcodec into hwcodec is hw is the better oine....
00:09:51einhirnamiconn: and I assume that posting in a queue that doesn't handle hwcodec playback is a no-no?
00:10:50amiconnSeee line 1234 (svn) - the call to swap_one_chunk() can take up to ~8ms if the buffer is above low watermark, and even up to ~65ms if it is below
00:11:17amiconnAnother queue wouldn't help either, as the bitswap call is monolithic, i.e. doesn't yield
00:11:40einhirnamiconn: I see...
00:11:45amiconnJdGordon: That wouldn't be possible - the hwcodec engine couldn't handle swcodec due to its design
00:12:25amiconneinhirn: The only solution would be to swap in smaller amounts - but then we need more than one thread, i.e. a better engine
00:14:00amiconnThe assembler optimised bitswap manages to swap ~1MB/second
00:14:15 Quit julle- ()
00:14:18amiconn11 clock cycles per byte
00:14:20einhirnamiconn: I assume that swapping couldn't be done in the 0x2000 byte block sizes that the DMA transfers use?
00:14:44einhirnwould probably take to long in the isr.
00:15:01 Quit roolku ()
00:15:23amiconnswapping in the isr is a non-no, unless the chunks become a lot smaller
00:15:59einhirnI see - 0x2000 is 8kb...
00:16:08amiconnyup, and hence ~8ms
00:17:35amiconnBtw, the extra 6dB can be gained by changing the codec register 7
00:18:09amiconnIts default setting is 0x40, which equals 0dB. 0x7f equals (almost) +6dB
00:18:22 Quit jhulst (Remote closed the connection)
00:18:51 Quit BigBambi (Remote closed the connection)
00:19:29 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
00:25:55 Quit obo ("bye")
00:29:37 Quit defukt (Read error: 113 (No route to host))
00:30:01 Quit BigBambi (Remote closed the connection)
00:31:42 Join cool_walking_ [0] (n=cb3b81c3@gateway/web/cgi-irc/
00:37:45 Quit lee-qid (Success)
00:38:17***Saving seen data "./dancer.seen"
00:38:42einhirnamiconn: re +6db will look into that, thanks
00:39:09amiconnOf course that only helps analog out
00:39:16 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
00:39:29 Quit JdGordon|w (" ajax IRC Client")
00:39:59 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
00:42:00 Quit Arathis ("Bye, bye")
00:42:10 Join hsvsunshyn [0] (n=95a662ca@gateway/web/cgi-irc/
00:42:37 Join HSV [0] (n=95a662ca@gateway/web/cgi-irc/
00:42:37 Quit hsvsunshyn (Client Quit)
00:45:16 Join _stink__ [0] (
00:46:12 Quit spiorf (Remote closed the connection)
00:46:12 Quit HSV (Client Quit)
00:46:52 Join HSV [0] (
00:46:52JdGordon|wtalk of the vm build server makes /me prod badger to adds my server again :p
00:48:03*scorche|w thought JdGordon|w said he didnt have a server
00:48:56JdGordon|wwell... my main box
00:49:28 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
00:49:32 Quit n1s ()
00:50:07scorche|wwell, i would hate to have a build server that is absent every time you reboot into vista to play games =P
00:50:11 Quit _stink__ (Client Quit)
00:50:49JdGordon|wwhich is almost never...
00:52:58DerPapstawesome games like solitaire...
00:53:20JdGordon|wkpatience ftw!
00:54:33rasherscorche|w: It's still better than not having it in the pool, surely
00:54:47HSVIs there an appropriate place to indicate a type of dock that is compatable with Rockbox? (in this case, an add-on ipod interface for a Kenwood car stereo) Just in case someone else with a Rockboxed iPod has a Kenwood stereo and wanted to know if the adapter (which works with many/most Kenwoods) works with Rockbox...
00:55:40 Quit aggelidis ("Konversation terminated!")
00:56:02 Quit scorche|w ("CGI:IRC")
00:58:18 Quit hannesd ("Client suicide")
00:59:06 Quit _stink_ (Read error: 110 (Connection timed out))
00:59:34HSVexcellent. 1000x thanks, linuxstb
01:03:46 Quit ender` (" Documentation is like sex: when it's good, it's very good, and when it's bad it's still better than nothing.")
01:06:42 Quit JdGordon|w (" ajax IRC Client")
01:07:21 Quit FMA1394 (".<UPP>.")
01:09:40 Quit ompaul (Read error: 113 (No route to host))
01:15:30 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
01:16:25 Quit n17ikh|Lappy ()
01:18:58 Part sandberg
01:19:02 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
01:20:14HSVas per the main TWiki page, could someone please grant write permission to me? I would like to include above mentioned details in above mentioned TWiki page. My TWiki name is HaroldVan.
01:20:54 Quit kugel (Read error: 104 (Connection reset by peer))
01:21:27 Join Orac2 [0] (
01:21:40 Nick Orac2 is now known as Troy (
01:22:06 Nick Troy is now known as pwncakes (
01:22:25linuxstbHSV: Done
01:23:54 Part pixelma
01:24:04HSVagain, thank you, kind linuxstb
01:24:57einhirnamiconn: I think that I'll leave the DAC_IN_DSP setting out for now, since I don't want to risk clipping -> "7Fhex 200% (+6 dB gain) If the sum of both mixing inputs exceeds 100%, clipping may occur in the successive
01:24:57einhirnaudio processing."
01:26:10einhirnBut I think I'll definately try it out.
01:26:26einhirnJust want to upload that patch so I can go to bed ;9
01:26:31pwncakeslinuxstb: Where can i get the source for tcc tool? Ive searched rockbox/google etc.
01:26:48rasherpwncakes: in svn?
01:26:54rasherunder utils/
01:27:16 Quit DerPapst (Read error: 104 (Connection reset by peer))
01:27:17pwncakescheers, thanks.
01:27:33rasherThe tarball should include it as well
01:32:11 Quit JdGordon|w (" ajax IRC Client")
01:34:47*rasher sighs
01:35:07rashernot exactly instant gratification, this
01:35:36rasherecho all|tools/ <−− takes a while
01:41:27einhirnI'll go to bed now
01:41:29*einhirn waves
01:41:36 Quit pwncakes ("Do fish get thirsty?")
01:41:44 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
01:42:36HSVthank you again for your help, linuxstb. Changes made, and I do not think I even broke anything.
01:45:32 Join jac0b [0] (
01:45:51jac0bdoes anyone know of any good cans to use for the treadmill
01:46:10soaplook at that - nice links and everything HSV. Thanks.
01:46:51soapOnly point - and it would be hard for you to know - when I started that wiki page all the entries were alphabetical. But I see the Car Audio section had already drifted from that original plan - so how were you to know?
01:47:32soapAhh - all the sections have drifted from alphabetical. I can fix that. Thanks again for the entry.
01:47:46HSVindeed, soap. I am planning on testing the interface with my old 4G grayscale. I can alphabetize them when I update it...
01:48:03HSVI am happy to contribute
01:48:33soapwell, we can promise you won't be able to control the iPod from your kenwood - as the apple accessory protocol, not the serial interface on which it runs, are implemented.
01:49:07rashersoap: don't you mean that neither are implemented?
01:49:21soapthe big question on accessories like that, IMHO, is if audio in/out works - as some accessories won't even pay attention to the line-out signal unless they hear the iPod talking to them,
01:49:41HSVright... I did not think remote or display would work. I was afraid that even the audio out would not work
01:49:43soapthank you rasher - "nor" turned into "not"
01:50:07HSVaudio and charge both work perfectly
01:50:56 Quit MethoS_mobile (Remote closed the connection)
01:51:05 Join MethoS_mobile [0] (
01:51:10HSV(the deck actually says "E60", which presumably means that the ip500 is connected, but it is not getting any display info from it)
01:54:34soapError 60?
01:54:40soapShit man - you better watch it.
01:55:20soapLast time I saw that code the headunit caught on fire and lead to a crash which killed a school bus full of orphans.
01:56:45 Quit DerDome ("Leaving.")
02:01:32 Part jac0b ("Ex-Chat")
02:02:42 Quit waldo ("Konversation terminated!")
02:03:17HSVreally? I thought E70 was the "Kill nearby orphans" error code...
02:04:21HSVI looked it up in the manual, and it says "Communication Failure between KCA-iP500 and iPod."
02:05:26HSV(reminds me of an old UNIX joke: '/earth full. Please delete anyone you can.'
02:18:41 Quit soap (Read error: 110 (Connection timed out))
02:20:08 Quit weezerle ("...und tschüss!")
02:21:44 Quit MethoS-- (Read error: 104 (Connection reset by peer))
02:23:46 Quit rocko (Read error: 110 (Connection timed out))
02:24:01 Join Imper [0] (
02:24:35 Quit MethoS_mobile (Read error: 110 (Connection timed out))
02:26:18ImperHey all, quick question I hope I can find an answer to. Currently using a Sansa e200 player with Roxkbox r16172. Is it normal for it to take 2-3 minutes to choose a new EQ Preset?
02:26:49ImperWell, more specifically, 'load' a EQ preset
02:27:57LloreanjhMikeS: Any idea if a recent change might've caused this:
02:28:15LloreanImper: Do you mean literally 120-180 seconds?
02:29:14LloreanIt's not normal for it to take a long time, but I'm curious if you mean that as a realistic measure, or just mean "take a longer than reasonable amount of time"
02:30:08ImperYes, if I move from Classical to Jazz (for example), it sits there with "Loading" for 2-3 real minutes... 120-180 seconds. Actually clicked 'Jazz' when I started this message and it is still loading.
02:31:25 Join jcollie [0] (
02:32:06LloreanThat's not normal, and you're the first person to report something like that.
02:32:06 Quit axionix (Read error: 104 (Connection reset by peer))
02:32:45 Join axionix [0] (
02:34:54ImperLlorean, if I'm the first person to report it, it is probably something I'm doing wrong on my end. I'm only about three days into using Rockbox and still figuring out the intricates. And I know that when I was converting files over the original firmware was experiencing some issues. Maybe it is time I fix up my MP3 collection. Glad to hear it isn't normal (to a point), thank you.
02:37:10ImperOh, and guess while I'm here. Is there any way to disable (or require holding) the Record button? For anyone who has a Sansa they will know the troubles with the location of that button, it is better off disabled.
02:37:52LloreanI own a e200, and haven't run into any trouble with it, so I'm not sure what you mean.
02:38:20***Saving seen data "./dancer.seen"
02:41:26ImperOh, maybe I should explain. I have a protective cover for my Sansa with a belt-clip on the back. Which I commonly swivel to the side. The button is then on the bottom, ready to be hit any time I sit down. Probably hit it about 15-20 times a day in one way or another. Could be that the cover is just too tight on the button.
02:42:12LloreanI haven't used the recording feature on mine, but I thought you had to hold down the button for a second for it to activate already
02:42:32LloreanIf you want to avoid accidental button presses, you should probably use the Hold switch.
02:43:52ImperYes, would be nice, but my cover unconveniently covers the Hold switch. All truth be told, it does seem to be better with Rockbox, guess the original firmware would activate on button press. Is it possible to extend that 1second time?
02:46:58 Part HSV
02:48:48LloreanNot without changes to the code.
02:53:19 Quit JdGordon ("Konversation terminated!")
02:58:16 Join JdGordon [0] (
03:10:19 Quit JdGordon (Read error: 104 (Connection reset by peer))
03:12:28 Join JdGordon [0] (
03:18:43 Quit Imper ("ChatZilla 0.9.80 [Firefox]")
03:33:28 Quit tvelocity ("Αποχώρησε")
03:48:32 Quit jott ("Reconnecting")
03:48:36 Join jott [0] (n=j@unaffiliated/jott)
03:59:11 Join goffa [0] (n=goffa@
04:04:32 Join gtkspert [0] (
04:05:51 Join soap [50] (n=soap@rockbox/staff/soap)
04:09:38 Join perrikwp [0] (n=chatzill@
04:11:29 Join brent0n [0] (n=470bd10b@gateway/web/cgi-irc/
04:11:53brent0nhello :)
04:12:34brent0nis there a plugin that reads PDFs?
04:14:20cool_walking_You can convert PDFs to plain text with numerous utilities, such as pdf2text, but the layout may be screwed up.
04:14:49brent0ndamn :(
04:15:13LloreanWell, nobody actually interested in it has written a plugin to do it.
04:15:37brent0nis there a good way to read books on the rockbox?
04:15:47cool_walking_Plain text.
04:15:48LloreanBuy audiobooks.
04:16:00LloreanIt's an *audio player* firmware.
04:16:03cool_walking_or that.
04:16:06brent0ni'm using the web client, so my responses are a bit delayed
04:17:54 Quit gtkspert_ (Read error: 101 (Network is unreachable))
04:19:13brent0nrockbox is good for doing things audio players don't normally do :P
04:19:21brent0nWAYYY delayed, btw
04:19:32brent0ni said that stuff 2 minutes ago
04:24:47 Quit brent0n ("CGI:IRC")
04:25:41 Nick fxb is now known as fxb__ (
04:28:07 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:28:21 Join miepchen^schlaf [0] (
04:32:41 Join webguest12 [0] (n=3a065c15@gateway/web/cgi-irc/
04:32:46 Quit webguest12 (Client Quit)
04:32:59 Join Tsukasa [0] (n=3a065c15@gateway/web/cgi-irc/
04:33:16 Join aneka [0] (n=kvirc@
04:34:15 Quit Tsukasa (Client Quit)
04:36:10 Quit soap (Remote closed the connection)
04:36:44 Join soap [50] (n=soap@rockbox/staff/soap)
04:37:09 Join Llorea1 [0] (
04:37:25 Quit Llorean (Nick collision from services.)
04:37:27 Nick Llorea1 is now known as Llorean (
04:37:55 Quit soap (Remote closed the connection)
04:37:55 Quit w0rd54 (Client Quit)
04:38:23 Join soap [50] (n=soap@rockbox/staff/soap)
04:38:24***Saving seen data "./dancer.seen"
04:42:35 Join w0rd54 [0] (
04:44:50 Quit Thundercloud (Remote closed the connection)
04:48:49 Quit w0rd54 (Read error: 104 (Connection reset by peer))
04:49:15 Join w0rd54 [0] (
05:03:13 Quit ol_schoola_ (Connection reset by peer)
05:03:35 Join ol_schoola_ [0] (
05:05:59 Quit cool_walking_ ("CGI:IRC (EOF)")
05:09:39 Quit csc` ("Powering Off")
05:15:21 Join cool_walking_ [0] (n=cb3b81c3@gateway/web/cgi-irc/
05:15:49 Join DaCapn [0] (
05:26:55 Quit DrDnar (Read error: 110 (Connection timed out))
05:28:44 Join mf0102 [0] (n=michi@
05:30:45 Join holt_kessler [0] (
05:36:58 Join fyrestorm [0] (
05:40:31 Quit Horscht ("")
05:47:42 Quit holt_kessler ("Leaving")
05:53:45 Join toffe82 [0] (
06:06:39 Join TheLastEmotion [0] (
06:07:26TheLastEmotionI think I might've bricked my Gigabeat F40 but I am not sure
06:07:53toffe82What happened ?
06:08:19TheLastEmotionWell, I had rockbox on it and it worked really well for a few months, then I went to load it up and it keeps giving me either or of two errors.
06:08:55TheLastEmotionEither system error 00000010 or 00000020
06:10:59TheLastEmotionIs the disassembling talking about unplugging the harddrive then plugging it back in?
06:11:03TheLastEmotionBecause I did that one already lol
06:11:57toffe82Your computer doens't recognize the gigabeat ?
06:12:17cool_walking_Yes most likely talking about taking it apart, unless there is some Gigabeat specific thing I don't know about (Which I wouldn't, since I don't own a Gigabeat)
06:12:55 Quit MattAndrew ("CGI:IRC (Ping timeout)")
06:16:02TheLastEmotionI am downloading the root folder that someone put up to see if that works seeing that my GB has no files whatsoever on it.
06:18:21qwmscorche|sh: you around?
06:18:28qwmscorche: or you.
06:21:52 Join BenniBoya [0] (n=3a065c15@gateway/web/cgi-irc/
06:22:28BenniBoyahey can someone help me
06:22:38BenniBoyai cant seem to build my boot loader
06:24:11 Quit DaCapn ()
06:24:35BenniBoyais anyone here?
06:26:13TheLastEmotionAnd now it seems that the harddrive isn't loading in windows.
06:26:17TheLastEmotionHi, Benni :]
06:27:45krazykitBenniBoya, it's generally not recommended to build your own bootloader unless you know what you're doing.
06:27:46BenniBoyawhat did u mean by the hard drive thing?
06:28:09TheLastEmotionThe harddrive of my Gigabeat F40 is now not showing up in windows explorer at all.
06:28:49BenniBoyawhen i run ../tools/configure it says
06:28:58krazykitBenniBoya, please don't paste in here
06:29:04BenniBoyaoh ok
06:29:09krazykituse or something
06:29:16BenniBoyawhich is?
06:29:21BenniBoyaim new to linux
06:29:21krazykita website?
06:29:34BenniBoyaits just text
06:29:38BenniBoyacan i paste that?
06:29:51krazykitpaste that text into and then put that link here
06:30:01krazykitthis is to prevent cluttering the logs and flooding the channel
06:33:55BenniBoyacan someone look at that and see if anything is wrong
06:34:29 Quit desowin ()
06:35:16cool_walking_Looks like you don't have the build environment set up correctly.
06:35:52 Quit cool_walking_ ("CGI:IRC (EOF)")
06:36:06 Join cool_walking_ [0] (n=cb3b81c3@gateway/web/cgi-irc/
06:36:10BenniBoyai have run the rockboxdev script and downloaded the arm stuff
06:36:10cool_walking_Gah. Screw you, web client.
06:36:30krazykitBenniBoya, you haven't changed your $PATH
06:36:46BenniBoyahow do i do that?
06:36:57cool_walking_Not you BenniBoya, the web IRC client :)
06:37:04cool_walking_open /etc/profile in a text editor
06:37:11krazykitread that.
06:37:55BenniBoyaok ty
06:38:08TheLastEmotionSo yeah, my gigabeat is not being recognized at all by my computer...
06:38:27***Saving seen data "./dancer.seen"
06:38:29krazykitTheLastEmotion, right, you've said that. have you tried the recovery procedure?
06:38:58TheLastEmotionWell it was being recognized before I disconnected and then reconnected the hdd.
06:39:34cool_walking_Okay, but your other symptoms say you should still do the recovery procedure
06:39:34TheLastEmotionThe GB shows the USB logo but the computer isn't showing it at all in win explorer.
06:39:53TheLastEmotionHow exactly would I go about doing the recovery procedure?
06:39:59krazykityou would read the wiki
06:42:53BenniBoyaum it didnt work
06:43:25krazykitBenniBoya, you need to log out and back in
06:44:23BenniBoyalogout and in or restart?
06:44:41krazykitthe former.
06:44:48BenniBoyaok ill brb
06:44:58TheLastEmotionMy problem is now with the fact that I can not see my gigabeat within windows explorer to actually restore the GB system files.
06:45:13 Quit BenniBoya ("CGI:IRC (EOF)")
06:46:10TheLastEmotionI did the disconnect/reconnect and now explorer won't even show the GB as a harddrive or anything, there's no letter, no nothing.
06:46:52 Join BenniBoya [0] (n=3a065c15@gateway/web/cgi-irc/
06:47:29 Quit tedrock (Read error: 113 (No route to host))
06:47:46BenniBoyait didnt work
06:48:00 Join tedrock [0] (
06:49:00TheLastEmotionAnyone around? :o
06:49:23krazykitTheLastEmotion, when someone's around, they'll answer. please don't repeat the question over and over
06:49:38 Join toffe82_ [0] (
06:49:52krazykitBenniBoya, in a terminal, do "echo $PATH". does it give you /usr/local/arm-elf/bin in the output anywhere?
06:49:57TheLastEmotionSorry about that..
06:50:17BenniBoyadoesnt give me anything
06:50:44BenniBoyawhere am I ment to run it from
06:51:07krazykitBenniBoya, a terminal. xterm, konsole, gnome-terminal, whatever.
06:51:57BenniBoyait didnt display anything running it as user or root
06:52:37krazykiti've gotta go to sleep now. good luck with your problem
06:52:48BenniBoyaok thanks
06:53:44cool_walking_BenniBoya: It's case-sensitive, are you sure you did "echo $PATH" ?
06:55:18BenniBoyait doesnt display anything about arm
06:58:05cool_walking_I think the tools/configure script should have told you where it put the compilers, do you remember where?
06:58:21BenniBoyadont worry, ive found a fix
06:58:38 Quit SacredTerror (Read error: 110 (Connection timed out))
06:58:41BenniBoyaill just run this in my build folder: export PATH=/usr/local/arm-elf/bin:$PATH
06:59:03cool_walking_Yes, but that won't persist across reboots
06:59:25cool_walking_You need to add that line to the bottom of the /etc/profile file
06:59:57BenniBoyato the bottom?
07:00:32cool_walking_Or edit the line that's already there
07:00:39 Quit midgey ()
07:01:03BenniBoyaso I would put export "PATH=/usr/local/arm-elf/bin:$PATH" in my etc/profile
07:01:11BenniBoyabut where in wetc/profile
07:01:30cool_walking_Just chuck it at the very bottom, so you know it's not inside any "if" blocks
07:02:28BenniBoyaunder "export PATHumask 022"
07:02:49cool_walking_It doesn't matter what's already there.
07:02:52BenniBoyaok ty
07:03:51BenniBoyathis isnt entirely related, but if I had a set of commands to mount my ntfs harddrive
07:04:17BenniBoyawould i add the line to /etc/fstab
07:05:11 Quit toffe82 (Read error: 110 (Connection timed out))
07:05:15cool_walking_Just follow the syntax of the entries already in there, and run "man fstab".
07:05:15BenniBoyawhere would I add it
07:05:33BenniBoyai mean i want it to run at boot
07:05:52cool_walking_Just make sure it has the "auto" option.
07:05:59BenniBoyai dont under stand
07:08:06cool_walking_The syntax is something like "<device> <mount point> <filesystems> <options>"
07:08:34BenniBoyaif I wanted to to run while it booted up, where would i paste it?
07:08:34BenniBoyai have the command but i just need somwhere where it will run automaticly at boot
07:09:19LloreanBenniBoya: Please take non-Rockbox chatter elsewhere, as per the channel guidelines
07:09:28Lloreancool_walking_: Please remember to encourage people to follow the guidelines
07:09:39LloreanYou've been around here enough to know about them.
07:09:50BenniBoyaok sorry
07:10:03cool_walking_#rockbox-community okay?
07:10:19BenniBoyaumm.. how do i get there from the web client
07:10:31cool_walking_type "/join rockbox-community"
07:13:51 Part toffe82_
07:16:57 Quit cool_walking_ ("CGI:IRC")
07:17:11 Join cool_walking_ [0] (n=cb3b81c3@gateway/web/cgi-irc/
07:21:43 Quit linuxstb ("Leaving")
07:25:09 Quit BenniBoya ("CGI:IRC")
07:25:27 Join BenniBoy1 [0] (
07:29:17 Join BadTVDream [0] (
07:30:34 Join cool_walking_200 [0] (
07:32:00 Quit perrikwp ("ChatZilla 0.9.80 [Firefox 3.0b3pre/2008012704]")
07:34:57 Quit TheLastEmotion (Read error: 110 (Connection timed out))
07:48:32 Join lee-qid [0] (
07:52:19 Join Ranger- [0] (n=792d8fa2@gateway/web/cgi-irc/
08:01:03 Quit Ranger- ("CGI:IRC (EOF)")
08:04:11 Join karashata [0] (
08:12:06 Quit z35 (Remote closed the connection)
08:12:22 Join rangerirc [0] (
08:13:30 Nick rangerirc is now known as deranger (
08:14:09derangerhey is there anyone around?
08:14:22 Join kimi-sharamin [0] (
08:14:39derangerI just formatted+re-installed rockbox and lost my awesome equaliser settings
08:14:42 Quit karashata (Nick collision from services.)
08:14:47 Nick kimi-sharamin is now known as karashata (
08:14:59derangerany audio experts who can give me a hand setting up the equaliser?
08:20:22 Quit BigBambi (Remote closed the connection)
08:23:41 Part BenniBoy1
08:24:37 Quit cool_walking_ ("CGI:IRC")
08:24:47 Nick cool_walking_200 is now known as cool_walking_ (
08:25:31 Join spiorf [0] (
08:29:00 Join Febs__ [0] (
08:30:37 Quit midkay (Read error: 110 (Connection timed out))
08:35:04 Join midkay [0] (n=midkay@rockbox/developer/midkay)
08:37:00 Join Rob222241 [0] (
08:38:28***Saving seen data "./dancer.seen"
08:42:54derangerthanks :)
08:42:56 Quit deranger ("<[eV]Rygrass|zzzzzzzz> i like waking up to pms")
08:43:48 Quit spiorf (Remote closed the connection)
08:47:21 Quit Febs (Read error: 110 (Connection timed out))
08:48:48 Quit jhulst (Remote closed the connection)
08:55:26 Quit Rob2222 (Read error: 110 (Connection timed out))
08:56:46 Join ender` [0] (
08:59:04 Quit cool_walking_ ("Miranda IM! Smaller, Faster, Easier.")
08:59:52 Join Ziroday [0] (n=nick@unaffiliated/ziroday)
09:01:29ZirodayHi, I have the iPod Nano 1st Gen, after trying to use rbutil to install rockbox it complained about not being able to install the bootloader, I am aware there is another program that is meant to be able to install the rockbox bootloader on iPods however the name escapes me, anybody know what its called?
09:02:03LloreanIt too is in the manual..
09:02:28ZirodayLlorean: ah, apologies just spotted it - ipodpatcher
09:04:47 Join BenniBoya [0] (
09:05:10BenniBoyahey, im following the instructions on building the bootloader, but it cant seem to load ipod_fw
09:05:26BenniBoyai dont think it is compiled, as there is only ipod_fw.c
09:06:11BenniBoyais anyone here?>
09:07:15 Join petur2 [0] (
09:08:06GodEaterBenniBoya: ipod_fw isn't used anymore
09:08:25GodEaterwhich instructions are you following, and why ?
09:08:30 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
09:08:56 Join OlivierBorowski [0] (
09:09:05BenniBoyaim following the "How to install the Rockbox bootloader on your iPod" instructions from the wiki
09:09:15BenniBoyaand because im changing a bit of code in it
09:09:27GodEaterBenniBoya: which bit of code would that be ?
09:09:38BenniBoyathe bit at the end
09:09:44BenniBoyaim swapping where the ipod fw boots
09:09:47BenniBoyawith the rockbox
09:10:18BenniBoyaso rockbox boots when the holdm switch is on
09:10:21GodEaterthose instructions are ancient
09:10:29BenniBoyaoh lol
09:10:31 Quit aneka ("KVIrc 3.2.4 Anomalies")
09:10:36BenniBoyawould they work though?
09:11:05GodEatercan I ask why this bootloader change is so important to you ?
09:11:14GodEaterthe apple firmware stays loaded once you've booted it once
09:11:21BenniBoyai reset alot
09:11:22GodEaterunless you leave your ipod alone for about a week
09:11:32BenniBoyai dont like playing music in rockbox
09:11:39BenniBoyabut like the games and movies
09:11:58BenniBoyaive changed toe code and compiled toe bootloader right i think
09:12:11BenniBoyabut now am just patching my ipods
09:12:11GodEaterwell, no - those instructions won't work currently
09:12:17GodEaterthey need re-writing really
09:12:23GodEaterbut we've not ever bothered because we don't see the point
09:12:38BenniBoyaim just doing this for personal use
09:13:02BenniBoyaif i could build ipod_fw it would work
09:13:06GodEaterso you have a bootloader-ipodvideo.ipod file ?
09:13:43GodEaterforget about ipod_fw
09:13:43GodEaterit's not used
09:13:53GodEaterand if you somehow getting it working, it'll likely break something on your ipod
09:14:12BenniBoyai changed the ipod.c file
09:14:14BenniBoyabuilt it
09:14:23GodEaterI don't care
09:14:29GodEaterjust answer the question
09:14:42GodEaterdo you, or do you not have an bootloader-ipodvideo.ipod in your build directory
09:14:48BenniBoyai dont
09:14:56amiconnipod_fw is most probably buggy, so I wouldn't use it. Iirc linuxstb fixed several bugs regarding firmware partition handling in ipodpatcher, which is based on the ipod_fw code
09:15:05GodEaterthen you've done something else fundamentally wrong
09:15:20BenniBoyai built the bootloader with ipod nano sttings
09:15:23amiconnThese fixes were never backported afaik, as ipod_fw is outdated, afaik even the ipl people use ipodpatcher nowadays
09:15:24GodEateramiconn: correct
09:15:44GodEaterwe should delete this page from the wiki
09:15:46BenniBoyathe builder made a sub folder of build called bootlaoder
09:15:49GodEateror at least mark it as dangerous
09:16:08markunBenniBoya: what's the name of the bootloader file you've built?
09:16:10GodEaterjust look in the build directory
09:16:10GodEaterignore the bootloader subdirectory
09:16:19GodEaterdo you have a bootloader-ipodnano.ipod in there ?
09:16:24BenniBoyait built to a boorloader sub dir
09:16:53GodEatercan you take a directory listing of your build directory, and put it in a pastebin ?
09:16:57GodEaterbecause I don't believe you
09:17:03BenniBoyahow do i do that?
09:17:06markunBenniBoya: I'll try to build the bootloader, let's see what happens.
09:17:43GodEatermarkun: I just built one for the video - I get a .ipod file
09:17:54BenniBoyatry build for nano
09:17:59markunBenniBoya: I got this file:
09:18:06BenniBoyaooooh now i see it
09:18:08BenniBoyawasnt looking soz
09:18:22GodEaterwhy do we bother ?
09:18:22BenniBoyasoz everyone
09:18:35BenniBoyato teach people lol?
09:18:52 Quit JdGordon ("Konversation terminated!")
09:19:01GodEaterjust put it on your nano with "ipodpatcher -a bootloader-ipodnano.ipod"
09:19:21BenniBoyawill that patch my fw or over write it
09:19:39BenniBoyabecaue the bootloader is 51kb
09:19:47GodEaterit patches it
09:19:54GodEaterthe clues in the name of the binary "ipodpatcher"
09:20:01BenniBoyaso ru ipod patcher from my build dir
09:20:33BenniBoyaonly oooone thing
09:20:39BenniBoyamy fwpatcher isnt compiled
09:20:52GodEateryou don't need to compile ipodpatcher
09:20:55GodEaterit's available for download
09:21:05BenniBoyaoh ok
09:21:14BenniBoyai thought u ment fwpatcher
09:21:21GodEaterdid I type fwpatcher ?
09:21:39BenniBoyai thought they same ting
09:21:50*GodEater gives up
09:22:01*BenniBoya doesnt blame him
09:22:23markunBenniBoya: anyway, just download it here:
09:22:46BenniBoyashould i chuck it in my build dir?
09:23:01markunyou don't have to, but you could I guess
09:23:18BenniBoyai may as well...
09:23:27BenniBoyaive already got the boot loader on my ipod
09:23:31BenniBoyais it ok to install over it?
09:24:01markunI really don't know anything about ipods, sorry
09:24:13markunmaybe GodEater can help? :)
09:25:41BenniBoyaGodEater: do u recon its ok to install my custom boot loader over my already installed stock one?
09:26:24BenniBoyai guess to be on the safe side ill remove it woth ipod patcher
09:27:12 Join Axio_ [0] (
09:27:51LinusNit should be ok to install over it
09:28:57BenniBoyaomg its not running...
09:29:51BenniBoyait says ./ipodpatcher acess denied
09:30:09BenniBoyapermission denied
09:33:07BenniBoyadoes anyone knoe why?
09:34:20 Join Febs___ [0] (
09:34:21GodEaterchmod +x ipodpatcher
09:34:23 Nick Febs___ is now known as Febs (
09:34:38GodEaterand you need to run it as root
09:35:17 Join cool_walking_ [0] (
09:35:24BenniBoyahey cool_walking
09:36:45GodEaterI've updated that wiki page now. ipod_fw is no longer mentioned anywhere.
09:37:40 Join CaptainSquid [0] (
09:38:47BenniBoyaaccept on my computer, mwahahahaha!
09:40:04 Part LinusN
09:40:20cool_walking_BenniBoya: run it as root
09:40:34BenniBoyait worked
09:40:47BenniBoyacan some1 do me a favout
09:40:55BenniBoyamy new bootloader doesnt boot into rockbox...
09:41:06BenniBoyacould some1 look at my ipod.c file?
09:42:05cool_walking_Ah, grabbed the log before GodEater replied, but joined after...
09:42:11BenniBoyaoops i see where i went wrong
09:45:06 Join gtkspert_ [0] (
09:48:29GodEateranyone else seen that only the absolute latest ipods will work with the iTunes video rentals ?
09:49:23GodEaternot that we care of course ;)
09:49:46 Part BenniBoya
09:50:31 Quit Febs__ (Read error: 110 (Connection timed out))
09:54:32 Join xushi [0] (n=xushi@unaffiliated/xushi)
09:58:47 Quit gtkspert (Read error: 101 (Network is unreachable))
10:00:02 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
10:04:15GodEater I'd really love to be helping out with the battery benches on the ipods at the moment
10:04:39GodEaterbut I'm not sure how useful they'd be given mine is nigh on unusable at the moment because of the all the disk problems
10:06:11LinusNit would surely be nice to find the reason for the recent buffering issues
10:06:34LinusNin the meantime, meybe we should revert the spinlock changes
10:07:36LinusNi'm pretty sure that the spinlocks work fine, but the buffering/playback surely suffers
10:11:10GodEaterso does everything else
10:11:19GodEaterthe UI is unusable while the disk is going mad
10:12:04LinusNthe first thing to find out is where the disk accesses are coming from
10:12:29GodEaterwell on mine currently it's the dircache update which is causing the thing to go mad when it first starts up
10:12:39GodEaterI've stopped using the database, and turned off the updates for that
10:12:43GodEaterso that's the only thing left
10:13:46GodEaterthe update is a lot slower than I'd expect too
10:14:29 Quit BadTVDream (Read error: 110 (Connection timed out))
10:14:31GodEaterand playback is taking over a minute to start up from poweroff
10:15:06 Quit amiconn (Nick collision from services.)
10:15:12 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
10:20:14GodEater8860 entries took 355 seconds to scan
10:20:31GodEateronce that's done, playback and buffering become ok
10:21:31GodEaterwhile it's going though, the buffering thread can't load enough off disk to keep up with the codec's demand for data
10:28:10GodEaterturning off dircache completely eliminates the dropouts I'm getting
10:28:40GodEateralthough the real and usefl buffers are taking a LONG time to fill up all the way
10:31:07 Quit nicktastic (Remote closed the connection)
10:31:20 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
10:31:37 Join xushi_ [0] (n=xushi@unaffiliated/xushi)
10:33:43markunGodEater: maybe we should set the scanning thread to a lower priority?
10:34:19GodEaterit isn't already ?
10:38:04markunno idea what the priorities are set to
10:38:31***Saving seen data "./dancer.seen"
10:40:22GodEaterseems to be PRIORITY_BACKGROUND currently
10:40:41GodEaterbut I'm not sure what the IF_PRIO() macro does quite
10:40:54 Quit xushi (Read error: 110 (Connection timed out))
10:41:05amiconnIt just masks out the parameter for targets without priority scheduling
10:41:10 Join MethoS_mobile [0] (
10:41:26GodEateris there a lower priority than background ?
10:44:11GodEateronly LOWEST_PRIORITY it would seem
10:44:35LinusNis this a problem on all targets?
10:44:51GodEatermy gigabeat F is fine
10:45:08GodEaterI have no other portalplayer targets to try it on
10:45:14GodEaterand the H140 is at home today
10:45:39LinusNok, so which target is problematic for you?
10:45:55GodEatermy 5.5G 80GB
10:46:15amiconnIt might be all PP targets, but my quick tests didn't show any of the reported problems on any PP target I have available for testing
10:46:18 Join MethoS- [0] (
10:46:21GodEaterwhich, as we know, has a somewhat spastic disk controller in it anyway
10:46:41amiconn(which include G5.5, mini G2, H10 small, 2nd Gen)
10:47:00 Quit MethoS- (Read error: 104 (Connection reset by peer))
10:47:11GodEateryou're not a dircache user though are you amiconn ?
10:47:33 Join crazy_bus [0] (n=philip@
10:55:57LinusNi'm not sure i understand his corelock changes
10:56:09GodEaterwho is "he" ? :)
10:56:10LinusNthe corelock is a pure busy wait
10:56:19LinusN"he" is jhMikeS
10:59:06 Quit nicktastic (Read error: 113 (No route to host))
11:01:27 Join waldo [0] (
11:02:24LinusNGodEater: can you do a test for me?
11:03:23LinusNin firmware/export/config.h
11:03:51LinusNline 461
11:04:31GodEaterlooks empty to me
11:04:35GodEaterperhaps I need a svn up
11:04:56 Join midkay_ [0] (
11:04:56LinusNadd #define CONFIG_CORELOCK SW_CORELOCK
11:06:40GodEaterthere's a lot of warnings on build now :)
11:07:27amiconnGodEater: I *do* use dircache on these targets except mini G2
11:07:43GodEaterah sorry amiconn
11:08:36amiconnI use dircache if there's a slow disk (that rules out mini G2 - spinup time <= 500ms), and at least 32MB ram (that rules out the X5)
11:09:46*GodEater hopes all these warnings about redefining CONFIG_CORELOCK are ok
11:09:48amiconnLinusN: The corelock must be busy wait afaiu...
11:09:57LinusNamiconn: i guess so too
11:10:27amiconnAnd SW_CORELOCK is just for those PPs with a broken swp(b)
11:10:36amiconn(actually broken bus controller)
11:12:24GodEaterok - compiled
11:12:29GodEaterand installed to ipod
11:12:53GodEaterplayback is still glitchy
11:13:26GodEaterat least he UI is responsive now though
11:16:33LinusNok, he changed the ATA thread from using spinlocks (which yield when waiting) to use mutexes (which busy waits for the corelock)
11:17:07LinusNso i guess that has an impact if two threads are reading the disks at the same time
11:17:08 Join Casainho [0] (
11:17:30GodEateras dircache and buffering would be :)
11:17:44amiconn> ping jhMikeS
11:17:59GodEatersummon jhMikes
11:18:13LinusNi wish i knew why he changed that
11:18:29GodEaterI suspect there was a good reason =/
11:19:00LinusNGodEater: you could try to patch -R this one:
11:19:32GodEateris it worth combining that with the CORELOCK you just made me change?
11:19:34amiconnLinusN: The busywait for corelock should never be longer than a few microseconds
11:19:39GodEateror go back to the same as it was previously ?
11:20:07amiconnIt's just there to serialize the mutex access itself, not the mutex operation as a whole
11:20:14LinusNamiconn: i see
11:20:22LinusNGodEater: the same
11:22:08 Quit midkay (Read error: 110 (Connection timed out))
11:22:26GodEaterpatch seems to think that patch is already applied
11:22:32GodEateron a clean svn tree
11:22:37LinusNpatch -R
11:22:43GodEaterah of course
11:22:45GodEateryou even said
11:22:49GodEater<−− dumbass
11:23:01LinusNGodEater: hmm, reverting the ata patch might not help, since he changed the spinlock code too
11:23:21GodEateryeah - it doesn't build
11:23:48GodEaterwhich revision did it go in in ?
11:23:52GodEaterI can revert to that
11:24:12LinusNit came in 16105
11:24:27LinusNso revert to 16104
11:24:47 Quit MethoS_mobile (Read error: 110 (Connection timed out))
11:25:20LinusNhmm, is it only 505G that suffers from this?
11:26:38LinusNmy guess is that all targets suffer, but 5.5G suffers the most
11:27:03LinusNsince it does a lot more calls to ata_read_sectors()
11:27:06GodEaterthe forums would seem to indicate there are other targets
11:27:17GodEaterbut most of the complaints are against the 5.5G
11:28:10 Join homielowe [0] (
11:29:13amiconnThe strange thing is that I don't see that effect on your G5.5 ...
11:29:25LinusNsince ata_read_sectors() uses the mutex, and the FAT driver performs lots of 1-sector reads
11:29:27GodEater16104 must have been a red build
11:29:33GodEaterI've gone back to 16103
11:31:43 Join rocko [0] (i=rocko@
11:33:37GodEaterheh - how ironic - 16103 is a patch by me :)
11:35:38GodEaterthat's a vast improvement LinusN
11:36:03GodEateralthough the real and usefl buffers are still filling very slowly
11:37:33GodEateralthough, as before when I looked at this - it seems that's a function of "resume at power on"
11:37:45GodEaterif I start playing a new folder, all the buffers fill up straight away
11:38:28LinusNdo we use both cores nowadays?
11:38:37amiconnWell, with the dircache scan running in parallel, it will fill the buffers slowly
11:38:46amiconnThat is to be expected
11:39:15amiconnLinusN: Only in mpegplayer and the spc codec afaik
11:39:27 Join pixelma [0] (n=d5494d0c@rockbox/staff/pixelma)
11:40:16LinusNok, so what do we do? i vote for reverting the mutex/spinlock changes until we find out what is causing this
11:40:37pixelmahmm... still no /query allowed in the webclient
11:41:23markunLinusN: maybe wait for jhMikeS?
11:41:39GodEaterI'm puzzled why amiconn can't see this behaviour on his 5.5G
11:41:44pixelmaGodEater: I could build r16104 fine when I tested the day before yesterday
11:41:47LinusNwell, he must be aware of this by now, and still hasn't done anything
11:41:56GodEateramiconn: would it be worth me sending my config.cfg to you to try with?
11:42:15GodEaterpixelma: odd - wouldn't work here for me
11:42:22GodEaterI got a load of errors in settings_list
11:43:45LinusNamiconn: how many files are there on your/mine 5.5G?
11:44:49pixelmaGodEater, amiconn: I noticed that the problems don't always show the same. On my c200 I had problems with resume, when I changed my WPS around a bit it went away (but now I can't "insert" one single track into an empty playlist ("disk" activity shows in the status bar for a second on my c200 but it doesn't play)
11:46:02 Join MethoS_mobile [0] (
11:46:34pixelmaI also have the impression that first buffering after boot takes longer and the UI is not responsive during the time. It's probably not as noticable as on a hd based player
11:46:57Casainhohello :-)
11:47:30GodEaterpixelma: I assume dircache is not a factor then ?
11:47:32Casainhoanyone knows if mobile sdram is just available in BGA packages?
11:48:23LinusNCasainho: "mobile"?
11:48:41pixelmaGodEater: probably not (there's no dircache on the Sansas)
11:48:47Bagder"MDDR is an acronym that some enterprises use for Mobile DDR SDRAM, a type of memory used in some portable electronic devices"
11:48:49GodEaterpixelma: that's really what I was asking :)
11:48:58BagderLinusN: I'd never heard of it either...
11:49:08LinusNCasainho: many rockbox targets have SDRAM in non-bga packaging
11:49:47Bagder(quote from wikipedia)
11:50:14CasainhoI ask help for this:
11:50:21amiconnLinusN: A few hundred files... all .flac btw
11:50:55LinusNamiconn: a few hundred, maybe that's not enough to make dircache stall the player?
11:50:56GodEateramiconn: any chance you could load it up with a few thousand mp3s ?
11:51:30CasainhoI guy said to me that low power version of SDRAM are the mobile versions, but we wants to know if they are available in other tahn BGA....
11:51:37Casainhoto solder by hand...
11:52:09LinusNok, then i guess the iaudios and irivers use high power versions
11:52:35 Quit crazy_bus (Remote closed the connection)
11:52:41CasainhoLinusN: :-)
11:53:01Bagderquite possibly they don't use DDR at all
11:53:04pixelmaweird, the inserting works for me currently - but jhMikeS told me there was a race condition somewhere so maybe it's not that weird after alle
11:57:31pixelmaBagder: something else - not sure if it's worth it but the H10 manual currently stops in the main menu (fm radio) chapter, I fixed it but it won't be available until the next daily build round...
11:58:13 Join PaulJam [0] (
11:58:29amiconnGodEater: I can put my whole collection on it tonight and test. But I doubt I will see that effect, given that I don't experience many of the reported bug, probably due to the many features I never use....
12:00:31Bagder"Sorry, SMF was unable to connect to the database." ...
12:00:55Bagderit came back with a reload...
12:01:49GodEateramiconn: hence my "shall I send you my config.cfg" as well ?
12:02:07GodEaterok seriously, we need to move the forums :)
12:02:35 Quit xushi_ ()
12:02:51pixelmaBagder, LinusN: is one of you also able to look at the webclient configuration or is that Zagor's field only?
12:03:16BagderI prefer to leave that for Zagor
12:03:21*LinusN hides and points to Zagor
12:05:00 Join Mathiasdm [0] (
12:05:05pixelmaok, I should leave again for now, bbl
12:06:02 Quit pixelma ("CGI:IRC 0.5.9 (2006/06/06)")
12:10:13 Join Febs___ [0] (
12:21:03 Quit MethoS_mobile (Read error: 113 (No route to host))
12:27:11 Quit Febs (Read error: 110 (Connection timed out))
12:31:46 Nick fxb__ is now known as fxb (
12:38:34***Saving seen data "./dancer.seen"
12:39:52GodEaterforums have died again
12:41:09 Part Ziroday
12:43:05markunWe might be able to put ZIF-40 drives in our Gigabeat F's. What do you guys think?
12:44:19GodEaterthat would be nice :)
12:44:51GodEaterdoes he mean "ZIF to ZIF" ?
12:48:39GodEaterdo ZIF drives not come with those ?
12:50:34GodEaterdo you have an S yet ? :)
12:50:39markunno :)
12:51:09GodEaterso I suppose it's my job to connect the S's drive into my F and see what happens ?
12:51:31 Join Arathis [0] (
12:55:12 Quit PaulJam (".")
12:56:11 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
12:57:08 Nick fxb is now known as fxb__ (
13:02:58 Join BigBambi [0] (n=86ceaf37@rockbox/staff/BigBambi)
13:04:16BigBambiGodEater: Nico_P already tried the S drive in the F - it didn't work
13:07:06 Quit Arathis ("Bye, bye")
13:07:23*LinusN is confused by the new locking code
13:08:31preglowthen ask! :)
13:08:54LinusNjhMikeS: ping
13:10:06markunGodEater: doesn't look like it's the same pinout, but I hope I'm wrong
13:11:47GodEaternot if what BigBambi just said is true it seems
13:14:20jhMikeSLinusN: hey (good timing btw)
13:17:25LinusNjhMikeS: i'm trying to understand the mutex code and the core locking
13:17:41jhMikeSanything is particular?
13:17:49LinusNand i feel very blonde :-)
13:18:08LinusNin mutex_lock()
13:18:46LinusNlet's begin with the CORELOCK_SWAP case
13:19:20 Nick Febs___ is now known as Febs (
13:19:21LinusNi guess you use the xchg8 stuff to make sure you have exclusive access to m->locked, right?
13:19:45jhMikeSm->locked doubles as the busy wait + the locked flag
13:20:23LinusNso after the while(), it doesn
13:20:38LinusN't matter if you check locked, or m->locked, right?
13:21:15jhMikeSif STATE_BUSY is returned, the variable is being examined and it must loops. If something else is returned, that is the current value.
13:21:50LinusNbecause it says if(locked), which confused me a little
13:22:08LinusNwell, if(locked == 0)
13:22:09jhMikeSthat's the temp value
13:22:42jhMikeSit's an atomic swap, so it exchanged the busy value into the variable and gets the current value back.
13:23:39 Quit BigBambi ("CGI:IRC (EOF)")
13:23:50jhMikeSlocked will be the value that was there, and m->locked will have busy which causes the other core to bust wait until the busy value is replaced with something else
13:24:10*jhMikeS isn't typing well but hopes the point is being made :)
13:24:22LinusNwhich can potentially take quite some time, right?
13:24:26 Join Arathis [0] (
13:24:39jhMikeSnot long...a couple microseconds
13:25:20 Quit SirFunk (Read error: 110 (Connection timed out))
13:25:29jhMikeSan entire queue send turnaround from COP->CPU->COP reply takes about 15 microseconds
13:25:42LinusNwell, you would have to write to m->locked to unlock it, wouldn't you?
13:26:04jhMikeSyes, that's what it does
13:27:17jhMikeSthis lock will only be seen if the COP were to access the mutex too. something being used on a single core will never block there.
13:27:32 Quit Jon-Kha (Read error: 145 (Connection timed out))
13:27:59 Join SirFunk [0] (n=Sir@
13:29:30*jhMikeS should say "spin" not "block" which involved moving a thread to a wait queue
13:31:18LinusNi still don't get it
13:32:10LinusNif a thread calls mutex_lock() on a locked mutex, the other core would spin like hell on the lock until m->locked is updated
13:32:20jhMikeSthe corelock (spin or peterson's algorthm) is to guard agains concurrent access from two cores when one operation must complete
13:32:29LinusNwhich it won't be until the lock is unlocked
13:32:51LinusNbut this is not the software corelock
13:33:03jhMikeSno, until the core is done modifying the object itself
13:33:27LinusNbut it won't modify it if it finds out that it is already locked
13:33:29jhMikeSthey both have the same precise purpose
13:33:46jhMikeSthe corelock != the mutex
13:34:04LinusNit is in the corelock_swap case, isn't it?
13:34:32jhMikeSin both cases it's the same thing exactly, just implemented more effeciencly if swp works
13:35:46LinusNok, let's move on to the #else case
13:35:54jhMikeSthe mutex's corelock is unlocked when the thread that will block yields. the corelock is serialization for the mutex object itself
13:37:30LinusNso m->cl is unlocked when it yields?
13:37:38jhMikeSfunctinally identical. on single core, corelock_lock is just blank
13:38:05jhMikeSit unlocks the (effective) spinlock or m->cl inside switch_thread
13:38:41LinusNok, so how can we explain that disk access slows down to a crawl on some targets
13:38:57jhMikeSwhich targets and which circumstances?
13:39:10GodEater5.5G, during dircache update / buffering
13:39:15LinusNdircache update upon boot, most notably 5.5G
13:39:34markunonly when buffering and updating happen at the same time?
13:39:47GodEatermost noticeable theb
13:39:55GodEaterbut it still sucks
13:40:09GodEaterI'm still getting playback dropouts even after the dircache update has finished
13:40:12GodEaterjust not as often
13:40:13jhMikeSbut it slows down both?
13:40:51GodEatermy dircache update this morning took 355 seconds to finish
13:40:56GodEater~8000 items
13:42:37jhMikeSdircache will block any thread trying to modify it for the entire dircache update (if modifiying it via a file api). this became apparent after making it the default call path.
13:42:59jhMikeSthen I looked at the code which confirmed that
13:43:31GodEaterbut playback isn't modifying it though is it ?
13:44:32jhMikeSI don't think so. I'm just mentioning something that can cause things to freeze up a :)
13:44:43jhMikeSso only 5.5G is doing this?
13:44:59GodEaterit's only massively noticeable on the 5.5G
13:45:07GodEaterI suspect it's happening everywhere though
13:45:39LinusNjust check all the "insane disk activity" reports
13:46:15LinusNplayback will try to write to .playlist_control
13:46:19jhMikeS"insane meaning" ?
13:46:30LinusNas in accesses the disk for minutes and minutes
13:46:44LinusNand unresponsive while it does it
13:47:03GodEateryeah, UI is dead to all intents and purposes
13:48:15jhMikeSwhat else did you need enabled? I just tried H120 without difficulty.
13:48:42LinusNhmm, 8397 might be another issue though
13:49:11 Quit jcollie ("Ex-Chat")
13:49:18jhMikeSI tried inititing database while updating dircache, that will freezing until dircache is finished
13:49:46GodEaterjhMikeS: I was also using a theme with AA - but to be honest, on the 5.5G, it doesn't matter
13:50:34LinusNjhMikeS: and reverting your mutex changes solves it
13:50:45GodEaterI think what really ground my 5.5G's disk to dust was having dircache + an already built database doing auto-upda
13:50:50GodEater+ load to ram
13:50:59GodEaterAND resuming playback on start up
13:51:09GodEaterI had to hard reset that to stop it melting
13:52:13GodEaterLinusN: is it fair to say the 80GB 5.5G does double the number of ata_read_sectors() of any other target ?
13:52:21GodEateror just "more" ?
13:52:27jhMikeSthis is odd since mutexes are what was originally there.
13:52:38LinusNmore than double, methinks
13:53:00*GodEater goes to get more drink
13:55:09LinusNjhMikeS: i have a question regarding the xchg8 version of the mutex
13:55:44LinusNi'm not sure how it is supposed to work
13:56:11LinusNassume this case:
13:56:34LinusNo the mutex is locked by a thread, i.e m->locked is 1
13:57:27LinusNthen another thread attempts to read m->locked with xchg8
13:57:47jhMikeSthe while loop will let the core pass since STATE_BUSYu8 is equal to 0xff
13:57:47LinusNafter the xchg8() call, m->locked is set to BUSY, right?
13:58:20jhMikeSbusy = 0xff, locked = 1, unlocked = 0
13:58:41LinusNand since locked is then 1, m->locked is not written to again
13:59:08LinusNand if some other thread tries to read it, it will be stuck on BUSY
13:59:35jhMikeSno it won't, when the thread is granted the lock, m->locked is set to 1, not 0xff
13:59:49LinusNhmmmmm, is that what the TBOP_SET_VARu8 is for?
14:00:02LinusNwhen is it set to 1?
14:00:15jhMikeSyes, after the threads context is saved
14:00:53LinusNmaybe the comment /* Block until the lock is open... */ confused me
14:01:16*preglow hugs multi-threaded programming
14:01:20jhMikeSthat's part of the blocking operation.
14:01:34LinusNjhMikeS: but does that set it to 1?
14:01:51jhMikeSsee var_u8_p, var_u8_v
14:02:20 Quit Arathis ("Bye, bye")
14:02:33jhMikeSwhen it is safe to do so, the 8-bit variable at that location is set to the specified value
14:02:42LinusNi see it
14:02:48LinusNnow i get it
14:03:12LinusNi must say that the comment doesn't explain what the code does
14:03:18 Join LambdaCalculus37 [0] (n=44a0430f@gateway/web/cgi-irc/
14:03:34LambdaCalculus37Morning, everyone!
14:03:50markunmoning :)
14:04:03LinusNi mean the comment on line 1046
14:04:10 Join Arathis [0] (
14:04:34 Join kugel [0] (i=kugel@unaffiliated/kugel)
14:04:36LinusNwhen does it yield by the way?
14:04:48LinusNon the block_thread_no_listlock() call?
14:05:18 Quit karashata (" Try HydraIRC -> <-")
14:05:21LambdaCalculus37Who's the NSF codec maestro?
14:05:46preglowLambdaCalculus37: hcs, afaik
14:06:26amiconnWhy would ata_read_sector() on the G5.5 be called more often than on other targets??
14:06:43LinusNamiconn: because of the larger logical sectors
14:06:54LinusNthey are read 1 by 1
14:07:00LambdaCalculus37preglow: Doesn't seem to be here right now... I need to ask him something about the NSF codec's behavior as of late.
14:07:08jhMikeSLinusN: the thread is put on the mutexes wait queue and a new thread is switch in or the core goes to sleep if no threads are on the run queue
14:07:29amiconnLinusN: Sure, but that doesn't mean ata_read_sectors() is called more often
14:07:40preglowLambdaCalculus37: then you need to wait for him :)
14:08:18amiconnRockbox always reads in 512-byte amounts anyway. The ata driver splits and joins as needed
14:08:37LambdaCalculus37preglow: And wait I shall. :)
14:08:42LinusNamiconn: i don't really remember, but i believe we sacrificed some performance on the 5.5g, resulting in more ata_read_sectors() calls...
14:09:02LinusNi might be wrong
14:09:10 Join tvelocity [0] (n=tony@
14:09:54amiconnThe only situation where we're sacrificing performance is when writing in single 512-byte sector portions
14:10:12 Join MethoS_mobile [0] (
14:10:22 Join DerDome [0] (
14:10:44amiconnThat always does a read of the 1024-byte disk sector, replace the 512 bytes in question, and write the 1024-byte sector again
14:10:52LinusNamiconn: ah, yes
14:11:05jhMikeSLinusN: though putting a thread on the wait queue puts in on the end of the run queue where it may have to wait for other threads to execute first whereas the struct spinlock would keep it in place farther up the list. this difference is likely the reason as I see it now.
14:11:25amiconnAnd since the disk firmware doesn't expect the same sector being read again which was just written, it's not even in the disk's cache, causing a lot of physical seeking
14:11:27LinusNjhMikeS: sounds reasonable
14:11:43 Quit Arathis ("Bye, bye")
14:12:13amiconnThis only applies to directory contents and the fat. The actual data is larger, at least one cluster at a time
14:12:29 Join Arathis [0] (
14:14:50 Join MethoS- [0] (
14:15:26 Quit Arathis (Client Quit)
14:18:23 Join Arathis [0] (
14:21:24LinusNjhMikeS: so what do we do?
14:24:11 Join Thundercloud [0] (
14:24:14jhMikeSLinusN: We could go back to struct spinlock for the time being. Threads being woken are guaranteed CPU time after a wake whenever they are reached on the run list but there question is about where freshly woken threads go.
14:25:41 Join midgey [0] (
14:25:42LinusNbut the current incarnation of spinlocks are only using corelock(), right?
14:26:35LinusNwhich doesn't yield at all
14:27:02Casainhoanyone can tell me if a system with ARM9, using a data flash, should be with 32 bits?
14:27:25jhMikeSit was changed for that purpose on dual-core, yes. mutexes will ultimately work I think but perhaps fresh threads should move up the list to "insert next".
14:27:25preglowwhat do you mean? data bus width?
14:27:45preglowmost flash is 16 bit width, afaik
14:27:47Casainholike, a 16M bit memory...
14:28:10preglowyou're asking what capacity the flash should be?
14:28:44Casainhofor store a bootloader + rockbox, what capacity It shoudl have? in Mbit?
14:28:48jhMikeShmmm...I could put a simple patch together to try that
14:28:49 Join Zom_ [0] (
14:28:57LinusNjhMikeS: please do
14:29:32preglow32mbit/4mb should be enough
14:29:44preglowrockbox core is muuuuch smaller than that
14:29:50LinusNjhMikeS: but the question is: should unblocked threads always be inserted on top?
14:29:52preglowwould even have room for some plugins
14:30:46 Quit Zom (Nick collision from services.)
14:31:08 Nick Zom_ is now known as Zom (
14:31:12Casainhopreglow: thank you :-)
14:31:26 Join jcollie [0] (n=jcollie@
14:31:36jhMikeSLinusN: not sure. this would be explicit wakeup, not timeout which seems like it's a "something must run now" situation.
14:32:04LinusNjhMikeS: you have a point there
14:32:27jhMikeSit's likely been starved for awhile already...hmmm
14:34:32 Quit mf0102 ("Verlassend")
14:35:06 Quit jcollie (Client Quit)
14:37:46jhMikeSif it's been blocked because something else was accessing the disk
14:38:37***Saving seen data "./dancer.seen"
14:38:50Casainhopreglow: how much is the core of RB?
14:39:08Casainhowhat could be the minimum flash memory size?
14:39:59 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
14:40:11 Quit midgey ()
14:41:59jhMikeSthere's only one situation that should change this and that's in core_perform_wakeup and the running list isn't NULL and one or more threads are on the list after the current one
14:45:22LinusNi guess at boot, there could be many threads reading a lot from disk, and that's dircache, the database, and the main thread (loading the config+wps), and
14:45:37LinusNeven the playback thread if it resumes
14:46:20LinusNand if they have to yield after each ata_read-sectors() it would take ages to finish
14:47:14 Quit cool_walking_ (Read error: 110 (Connection timed out))
14:47:35 Quit Thundercloud (Remote closed the connection)
14:48:24 Join Thundercloud [0] (
14:49:29 Quit CaptainSquid (Remote closed the connection)
14:49:58 Join Stummvoll [0] (
14:50:47StummvollHi.. Is there Anywhere an german Support Chan for Rockbox?
14:51:58jhMikeSyield itself is very fast. anything accessing the disk gets high priority so it will finish with the shared resource. I'd almost go so far as to force a task switch if a thread was blocked on an object (giving sort of a preemption effect). I'll just try the simple rearrangement first.
14:52:03GodEaterStummvoll: no - english only
14:53:22 Join CaptainSquid [0] (
14:53:28Stummvollhm ok
14:54:28StummvollHow can I play the Tracks in random sequence?
14:54:31LinusNjhMikeS: i trust you with this :-)
14:54:49GodEaterStummvoll: use the quick menu to turn on shuffle
14:54:54LinusNStummvoll: enable the Shuffle setting
14:55:05Stummvollwhere ist that menu?
14:55:20LinusNin the playback settings
14:55:40GodEaterStummvoll: which player do you have ?
14:56:01StummvollSansa e260
14:56:09markunStummvoll: here you can find the manual:
14:57:19GodEaterI believe a long hold on the "menu" button whilst you're playing music back should summon the quick menu on the sansa
14:57:22 Quit ol_schoola_ (Read error: 104 (Connection reset by peer))
14:57:46 Join ol_schoola_ [0] (
14:57:49jhMikeSLinusN: Thanks :) Fairness is tricky and First-Come-First-Server (called "round robin" around here even though RR is a preemtive algorthm) isn't always the fairest thing.
14:58:12StummvollGodEater: Yes, that's it.. Thanks ;)
14:58:18GodEaterStummvoll: no problem
15:00:25 Quit BitTorment (Read error: 110 (Connection timed out))
15:03:31 Join BitTorment [0] (
15:06:58 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
15:09:32 Part Stummvoll ("Pinguin Timeout")
15:11:40*preglow has a hard time seeing it ever being the fairest thing
15:11:45preglowsince it doesn't even take fairness into account
15:11:49 Join jcollie [0] (n=jcollie@
15:13:25jhMikeSpreglow: there's other algorthms and I think Shortest-Job-First had minimal average latency in cooperative threading but then you need to keep track of run time statistics.
15:14:30 Join Jon-Kha [0] (
15:14:37 Join DefineByte [0] (
15:14:47 Quit Febs (Read error: 110 (Connection timed out))
15:14:48preglowwell, if you want a scheduler that has a consept of fairness, then you need to anyway
15:14:52preglowconcept too
15:15:08preglownot saying we really need that
15:15:31GodEaterwe don't *want* fairness do we ?
15:16:58preglowno, i don't think so
15:17:16jhMikeShow do you define the fairness that we don't want?
15:17:33GodEaterwhy bother if we don't want it ? :)
15:18:03jhMikeSpriority isn't equal to unfair
15:18:38GodEaterspeaking personally, I want my audio to work
15:18:49GodEaterhowever that needs to be achieved in terms of threading priority
15:29:02jhMikeSmaybe this works or has a positive effect? hopefully I did the list linking right but it seems to be running.">
15:29:32 Quit pondlife (Read error: 104 (Connection reset by peer))
15:33:30 Join MethoS-- [0] (
15:37:49 Quit BitTorment (Read error: 104 (Connection reset by peer))
15:40:26 Join BitTorment [0] (
15:48:02 Quit MethoS- (Read error: 110 (Connection timed out))
15:50:38 Join n1s [0] (
15:51:48 Join Febs___ [0] (n=chatzill@
15:51:55 Nick Febs___ is now known as Febs (n=chatzill@
15:57:24 Join jgarvey [0] (
15:59:52 Quit DefineByte ("Bye all")
16:02:27 Quit qwm (Read error: 104 (Connection reset by peer))
16:03:35 Quit Mathiasdm ("Yuuw!")
16:04:25 Join Mathiasdm [0] (
16:05:13 Join scorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
16:08:21LambdaCalculus37iPod chipset... who knows it well?
16:12:30*LambdaCalculus37 just realized he set himself up for a zinger
16:14:14jhMikeSGodEater/LinusN/anyone: Check the patch? I definitely don't want to make anything worse but it appears ok and I have no target with said problems.
16:17:11 Quit CaptainSquid ("Miranda IM!")
16:18:59 Join ingrediant [0] (
16:21:46 Quit ingrediant (Client Quit)
16:22:12LambdaCalculus37Ummm... right. =\
16:22:26jhMikeShit and run pr0n spammers? :p
16:22:43LambdaCalculus37jhMikeS: No, she's hot but clothed.
16:23:16jhMikeSsome have a fetish for clothed women ;)
16:24:19 Quit Arathis ("Bye, bye")
16:24:23LambdaCalculus37But I still think that was kind of... distracting. :)
16:24:49 Quit LambdaCalculus37 ("CGI:IRC")
16:24:55 Join LambdaCalculus37 [0] (n=44a0430f@gateway/web/cgi-irc/
16:25:21 Join Arathis [0] (
16:25:56 Join japc [0] (n=japc@
16:27:00LinusNjhMikeS: where is the patch?
16:28:09jhMikeShere again:">
16:29:24LinusNjhMikeS: thanks, will try to test it this evening
16:29:46LinusNGodEater might be a good tester, btw
16:31:54*LinusN has to go
16:31:59 Part LinusN
16:38:40***Saving seen data "./dancer.seen"
16:48:43GodEaterjhMikeS: just got back from the gym - will test now :)
16:48:55 Join mokkurkalve [0] (
16:49:24jhMikeSI have another touch that could improve things too. Only problem is playback does weird things when changing scheduling patterns.
16:49:53GodEaterone patch at a time :)
16:52:00jhMikeSit also drops CPU boost about 7% on a test file that uses goes at about 41%
16:52:19 Join pixelma [0] (n=d5494dd0@rockbox/staff/pixelma)
16:53:50*jhMikeS will take that one back...the tone controls were disables and they're usually on. :p
16:55:24 Nick fxb__ is now known as fxb (
16:56:05GodEaterso far not much difference
16:56:11GodEaterplayback is taking an age to restart
16:56:35GodEaterdircache is building MUCH faster though
16:57:04GodEaterand now pb has started - not so glitchy
16:57:18GodEateralthough still not perfect
16:57:29GodEaterhave had two dropouts so far
16:57:32jhMikeSI'll supply a patch with the other little trick
17:00:20GodEater271 seconds this time
17:00:28GodEaterand 8 drop outs in audio so far
17:00:44jhMikeSwhen do those occur btw?
17:01:10jhMikeSis it the codec getting behind or the encoded audio buffer going empty?
17:01:21 Join Lonnie_ [0] (
17:01:38GodEaterhow do I tell
17:01:51jhMikeSfrom the buffer screen
17:02:33GodEaterpcmbuf starts dropping really quick
17:02:38GodEaterand real and usefl are both empty
17:02:55GodEateralloc remains full pretty much all the time
17:03:01jhMikeSif real and usefl and empty, I suppose PCM would too
17:03:34GodEateryour second patch claims to want to reverse an earlier one
17:03:43GodEaterdoes that mean I should apply it indepently of the other one ?
17:03:44 Join aneka [0] (n=kvirc@
17:03:44jhMikeSyes, remove the old first
17:04:04GodEaterok done
17:04:24 Join desowin [0] (n=desowin@atheme/developer/desowin)
17:08:58GodEaterthe ui is dreadfully unresponsive at first power up
17:09:28GodEaterok - pb has started
17:09:36GodEater1 dropout already
17:10:19GodEaterreal goes from "having stuff in it" to "having nothing at all" in about a microsecond, then pcmbuf drops off
17:10:30GodEaterbut then real goes back to having stuff again
17:11:55jhMikeSThis do_timed_yield in tagcache.c is woefully infrequent too
17:12:13GodEaterthere's no database involved here currently
17:12:39GodEaterI don't understand this
17:12:46GodEaterhow can the real bar be about half full
17:12:48GodEaterthen go to nothing
17:12:52GodEaterthen go back to hald full again
17:13:00GodEaterall in a fraction of a second ?
17:13:18GodEaterwhere did the data go, and then where did it come back from ?
17:13:45jhMikeSI don't know for sure but I think some of this stuff is just race conditions
17:14:42jhMikeSIf I change scheduling orders, "bugs" come and go with the variation and that just should never happen
17:15:59*jhMikeS has seen negative sizes in the buffering screen if watching it while database is updating
17:16:14*scorche|w wondered if GodEater switched keyboards
17:17:30GodEaterto what ?
17:18:00GodEaterjhMikeS: well - this second patch is performing MUCH better than the first one
17:18:12GodEateronly a couple of drop outs while the dircache was building
17:18:27scorche|wGodEater: one with a much bigger "enter" key :)
17:19:30GodEateroh screw off
17:21:53jhMikeSGodEater: I thought you were saying it was dying badly :\
17:23:38GodEaterI'm saying I still get audio drop outs
17:23:45GodEaterbut the disk thrashing appears to have gone
17:23:56GodEaterand the drop outs are less frequent now
17:23:58jhMikeSprior to the mutex use you got none at all?
17:24:09GodEaterno hang on
17:24:11GodEaterthat's a lie
17:24:24GodEaterI got them after 15444 - the ipod ATA power off stuff
17:24:38GodEaterbut we never established what was to blame for that
17:24:51GodEaterwhether it was the ATA sleep stuff itself, or some changes in Nico's buffering code
17:25:23GodEaterok - just go another drop out, pcmbuf empty - this time real and usefl both still had stuff in them
17:25:27jhMikeSATA poweroff and the buffering changes were in a single revision?
17:25:30GodEater(having rebooted again)
17:25:47GodEaterno - I just noticed the dropouts started after that revision
17:26:07GodEaternegative value in usefl for a second now
17:26:15 Join Rob2222 [0] (
17:26:17GodEaterand again a drop out
17:27:03GodEaterand another one - all positive values in everything then
17:27:08GodEater(except the empty pcmbuf)
17:27:41jhMikeSI image you might get more dropouts from that screen. Is the LCD slowish on that?
17:27:43GodEaterI'd say this still isn't useable as a DAP currently - but it's better than it was
17:27:54GodEaternot that I've noticed
17:27:58GodEaterthink amiconn has mended it
17:29:30GodEaterI say commit your diff at least
17:30:55jhMikeSprep_bufgetdata does sleep(1)...this isn't so good either
17:31:07GodEateronce dircache has finished building it's markedly improved
17:31:19GodEaterit just takes nearly 3 minutes for that to happen :(
17:31:35*GodEater shouldn't own so much music
17:35:36amiconnWhat does the stats plugin say?
17:35:40 Join mf0102 [0] (n=michi@
17:35:50 Join JdGordon [0] (
17:37:18GodEaterFiles: 8084, Music: 5640, Dirs: 766, Max files in Dir: 439
17:38:02 Quit pixelma ("CGI:IRC 0.5.9 (2006/06/06)")
17:38:07GodEateris that a single dir that "Max files in" ?
17:38:22 Quit MethoS-- (Remote closed the connection)
17:38:34GodEaterthere's no way I have a directory that big
17:39:47GodEaternot unless it's one of the ones under .rockbox I've never looked in
17:49:50jhMikeShow long does a disk read take on 5.5g where it isn't yielding?
17:50:11GodEateryou tell me how to measure it, and I'll let you know
17:53:38 Join xnyhps [0] (n=sphynx@
17:53:44 Quit Rob222241 (Read error: 110 (Connection timed out))
17:57:17jhMikeSI'm not even sure atm exactly what to measure and where
17:58:45GodEaterwell in that case I'm going to the pub ;)
17:59:20jhMikeSif cache_sector never goes through wait_for_* procedure if the data is already there, that could starve other things
18:07:09nicktasticGodEater, haha
18:07:14 Quit Arathis (Read error: 110 (Connection timed out))
18:08:10 Quit petur2 ("work->home")
18:09:11 Join j0tt [0] (n=j@unaffiliated/jott)
18:10:03amiconnjhMikeS: On G5.5/80, cache_sector will have a miss on every other sector
18:10:25amiconn(physical sector size of the MK8009GAH is 1024 bytes)
18:14:24 Join dberg918 [0] (
18:16:07dberg918does anyone know when WPS uploading capabilities will be back up again?
18:17:01jhMikeSamiconn: so it reads 1K from disk at a time? I wonder about forcing yield between two 512 bytes reads. If cached, the CPU burst should be very much shortened.
18:19:43 Quit jott (Read error: 104 (Connection reset by peer))
18:20:09scorche|wdberg918: well, we dont like to give out time estimates...Real Life is awfully unpredictable
18:20:51markunwhy were they taken down?
18:22:51dberg918I was just curious, I'm just about finished making a WPS for the Gigabeat that uses viewports, and I'm eager to share
18:23:15scorche|wbandwidth constraints...they will be enabled again once i change to my server which would happen once we switch to the new site which would happen when viewports is committed/stable, but as it is taking longer than anticipated, i should probably just switch it now...
18:25:20dberg918the new site?
18:28:24scorche| is currently unofficial..."the new site" will be the official version replacing both the WpsGallery in the wiki and the current
18:30:37dberg918I didn't realize there was a
18:31:02dberg918but apparently I navigated to it without knowing through the normal website
18:31:52 Quit Zom (Remote closed the connection)
18:31:57scorche|wi assumed you meant that because that is the only theme-related site that has uploading disabled...
18:32:03 Join Zom [0] (
18:33:29 Join Domonoky [0] (n=Domonoky@
18:34:29DomonokyCasainho: have you read my pm in the forum, about the analog gnd and vdd on the extension board of rockboxPlayer v1 ?
18:34:54 Quit lee-qid ("aufwiederbyebientotsayonara")
18:35:36LloreanSlasheri: you around?
18:36:53LloreanSlasheri: Well, if/when you might have some time, would you mind commenting on FS# 7598? It looks like something that ought to find its way to SVN, but there's been no direct developer comment on it, and I think you'd probably be most suited. Thanks.
18:38:27CasainhoDomonoky: yes, I read - thank you :-)
18:38:43 Join BigBambi [0] (n=alex@
18:38:44***Saving seen data "./dancer.seen"
18:39:00Casainhobut I will not continue with prototype - now with prototypeB
18:39:16Casainhowaht do you think about prototypeB by Matt?
18:39:34 Join xushi [0] (n=xushi@unaffiliated/xushi)
18:40:06markunCasainho: wow, didn't know that oled screens were cheap
18:40:10 Quit xnyhps ("...")
18:40:24markunwell, relatively
18:40:27 Quit aneka ("KVIrc 3.2.4 Anomalies")
18:41:41 Quit HellDragon (Read error: 113 (No route to host))
18:41:48markunCasainho: nice clock project he did :)
18:42:33Casainhomarkun: yes, and he must do the board in the next 6 weeks :-)
18:42:57Casainhoso, hardware for we play on the next 6 weeks :-) :-)
18:44:06Domonokymaybe the oled display would also be interesting for the other prototype.. it doesnt cost much, and seems faster (as its connected paralell)
18:45:35markunand energy efficient
18:46:01Casainhoyes, energy :-)
18:46:37CasainhoDomonoky: there is no need to other prototype, If this will be good - I know, I changed my ideas....
18:46:41SlasheriLlorean: hi, i already plan to commit that. The latest patch looks ok
18:47:57DomonokyCasainho: its seems this guy has some experience, so maybe it would be good to switch over and help him :-)
18:48:04markunSlasheri: I wrote a patch for the database scanner and wanted your opinion:
18:49:31Slasherimarkun: so now there could be database.unignore deeper inside a folder marked with database.ignore?
18:49:44markunit's usefull for me
18:49:52 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
18:49:54markunand the scanning is still very fast with dircache on
18:50:00Slasheriok, doesn't sound that bad. looks ok to me
18:50:19CasainhoDomonoky: I am doing that right now ;-) :-)
18:50:53 Join petur2 [0] (
18:51:05markunSlasheri: I noticed a problem with a database.ignore file (but maybe only with my patch, I didn't compare)
18:51:09CasainhoDomonoky: also I think that information on TWiki help him, at least a lot of work doned can help in future :-)
18:51:26markunwhen everything is ignored the "initialise database" screen doesn't exit
18:52:03 Join bertrik [0] (
18:53:07Slasherimarkun: hmm.. you mean there is then no option in the menu to initialise the db?
18:53:18 Join Arathis [0] (
18:53:20Slasherithat doesn't sound possible :)
18:54:24Casainhobye bye :-)
18:54:29 Quit Casainho ("ChatZilla 0.9.80 [Firefox]")
18:54:38Slasherimarkun: but currently there is a known problem that empty db doesn't initialise because commit file doesn't exist at all
18:54:56markunaha, maybe that was the problem
19:05:04 Join perrikwp [0] (n=chatzill@
19:13:51SlasheriLlorean: hmm, i am afraid i can't commit that yet.. there is some mysterious/bad code causing null-pointer reference
19:14:26SlasheriLlorean: and i already fixed a compile issue on non-multivolume targets
19:14:42 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:16:02 Quit JdGordon|w (" ajax IRC Client")
19:20:16 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
19:21:08 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
19:21:09*jhMikeS wonders what's up with VMWare's inability to set its own clock correctly even with the timezone right :\
19:21:26LinusNjhMikeS: laptop?
19:21:27 Part triconda
19:21:38jhMikeSno, desktop
19:22:47 Join PhilLight [0] (n=51aef816@gateway/web/cgi-irc/
19:23:09 Quit dberg918 ("Ex-Chat")
19:23:43jhMikeSIf I have it set right, it's messed up at the next restart
19:24:25LinusNnot fun
19:24:44PhilLightSlasheri/Llorean: Re #7598, I haven't tested it at all on any real target other than e200
19:24:45LinusNjhMikeS: i read the logs, looks like your patch made a difference at least
19:25:03SlasheriPhilLight: i really wonder if it doesn't crash at the beginning
19:25:26Slasheriit really should shouln't work.. i am just writing a comment how to fix it
19:26:47jhMikeSLinusN: The one thing that makes the most difference is the one that I think may be unsafe (forcing yield if the mutex was locked). Playback hates badly it in any case.
19:27:49LinusNlooks like we need a good debugging tool for this
19:28:40 Quit perrikwp ("ChatZilla 0.9.80 [Firefox 3.0b3pre/2008012904]")
19:29:01 Quit JdGordon|w (" ajax IRC Client")
19:29:07PhilLightSlasheri: Okay. I've been using it constantly since I wrote it without any issues, but at the time there were a couple of things that I thought might be a problem. Unfortunately I lost the bit of paper I wrote them down on, and haven't really checked the patch since.
19:29:48jhMikeSLinusN: In your case, is it just the PCM buffer that goes empty or real and usefl too then PCM follows?
19:30:03 Join d34df00d [0] (n=d34df00d@
19:30:18 Part d34df00d ("Konversation terminated!")
19:30:38LinusNjhMikeS: i haven't used my ipod, nor have i seen the issue on any other player
19:30:54LinusNi am merely responding to all the user reactions
19:30:58 Join lee-qid [0] (
19:31:30jhMikeSnone of mine have it but the picking one or both changes in the patches actually seems to help them
19:32:10 Quit homielowe (Read error: 110 (Connection timed out))
19:33:37SlasheriPhilLight: now i added a comment, i hope it helps :)
19:33:40jhMikeSproblem is playback hates them both or frankly any at all at this point so I don't know what to do with it
19:34:12LinusNjhMikeS: i guess we need to invent some debugging support to investigate the threading
19:34:50jhMikeSin playback?
19:35:51LinusNsome kind of profiler
19:36:10 Quit japc (Remote closed the connection)
19:36:19 Join FOAD_ [0] (
19:37:28PhilLightSlasheri: Thanks for that. I'll try to have a look at it in the next couple of days.
19:38:20amiconnLinusN: The problem with any kind of debugging code here is that it changes the timing, so that it might be impossible to catch race conditions this way
19:39:12preglowwell, with cooperative threading that shouldn't be a problem, should it?
19:39:18jhMikeSI already see many of the problems in playback. Basically the only safe approach is consider it concurrent no matter what kind of threading and then timing changes don't affect it.
19:39:28 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:39:48jhMikeSpreglow: I've found cooperative threading to be a bug in itself.
19:39:58preglowyeah, and i agree
19:40:13preglowand the mindset it promotes
19:40:15 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
19:40:32preglowwhich is "let's do threading, but be pretty sloppy about it"
19:41:17*amiconn hates preemptive threading
19:42:14preglowhard to hate anything so convenient
19:42:18*amiconn is syncing the music dirs from H180 to G5.5
19:42:21preglowi dislike all threading, but that's because i suck at it
19:42:41amiconnPreemptive threading gives much more headaches than its worth
19:42:51preglowso does cooperative threading...
19:42:56preglowjust look at the current nightmare
19:43:12 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
19:43:15preglowremove one yield() and add another one, and bang, new bugs!
19:43:15amiconnSo why doesn't it hurt single core targets?
19:43:44preglowit does...
19:43:46jhMikeSit does. playback suffers the same problem with a change there
19:44:12 Quit JdGordon|w (Client Quit)
19:44:43amiconnI never observed the reported crazy disk activity so far. That's what I'm syncing the music for - testing with the full set present on G5.5
19:44:45jhMikeSI wrote mpegplayer code just as if it were all concurrent and no change underneath breaks it whether single or multi
19:44:51amiconnWill take quite a while....
19:45:37 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
19:46:27LinusNthe rockbox code was quite neat and simple back in the archos days, and i'd say that the simple threading contributed to it
19:47:01LinusNbut now the demands for more advanced threading has snuck on to us
19:47:13preglowcode isn't as simple as in the archos days anymore, i'm afraid
19:47:21preglowwe have databases, dircaches, sw codecs, sw playback, dsp
19:47:25LinusNthat's what i'm trying to say
19:47:34preglowit's inevitable unless we want a dumbed down firmware, and we don't do we? :)
19:47:34jhMikeSmuch mpegplayer code is accessed on both cores and it survives and the code is pretty much identical
19:48:47*LinusN goes to put the kids to bed
19:48:50 Part LinusN
19:48:50 Join _ke [0] (
19:49:03 Quit JdGordon|w (Client Quit)
19:49:35 Join spiorf [0] (
19:49:43preglowbut that being a fact, i'd rather want proper threading than the halfway thing we have now
19:50:52_kei got a usb-ac-adapter, which can power my usb things, like my ipod. unfortunately it doesnt send any usb host signal, so rockbox simply doesnt start in usb mode, while the ipod firmware does. any hint what i could do?
19:52:07preglowjhMikeS: cooperative threading in a dual core environment kinda qualifies as a halfway solution to me
19:52:28jhMikeStrue. it scales poorly to multiple processors.
19:53:03 Quit FOAD (Read error: 110 (Connection timed out))
19:53:03 Nick FOAD_ is now known as FOAD (
19:53:13jhMikeSand lets slow code hog CPU from time-critical code
19:53:35preglowbtw, how are things in gigabeast land?
19:54:14jhMikeSI got windows to send firmware to the device with a windows sendfirm.exe
19:54:36preglowstill no actual target coding?
19:55:59jhMikeSnot yet. I'm preparing myself mentally to start that stuff (monster datasheet and all). I'm just tying up mpegplayer loose ends like using the core DSP.
19:56:27 Join styleism [0] (
19:56:46preglowi don't think i would have been able to keep my hands off it
19:56:57pixelmajhMikeS: speaking of which there were 2 report of mpegplayer looking odd on H10 (one here and one in the forums)
19:57:11*amiconn kept his hands off the M3 for a full month already
19:57:24jhMikeSpixelma: after that ASM yuv thing perhaps?
19:57:31pixelmaI guess
19:58:10amiconnSimply because I also have a few loose ends to fix for other targets, and not that much motivation to actually do this :/
19:58:22 Join FOAD_ [0] (
19:58:29jhMikeSMy H10 likes sending two pixels with one instuction...perhaps not all do :\
19:59:06pixelmainstuctions can't work :P
19:59:32preglowamiconn: geh, is it just me or doesn't that thing have a display?
19:59:49amiconnIt does not have a main lcd, only the remote
19:59:53jhMikeS:) /me types too rapidly before he forgets what he's thinking and makes all sorts of mistakes
19:59:57preglowwell, that'll be new for rockbox
20:00:03amiconnNot really
20:00:13amiconnThe remote display will be handled as the 'main' lcd
20:00:42amiconnAnd I know how to avoid too much code duplication - my idea will even reduce duplication of code that's already in svn
20:01:37amiconnIf I wouldn't handle the remote lcd as the main one, adapting stuff would be *a lot* of work.
20:02:14preglowi can imagine
20:02:24preglowand handling a disconnected lcd will be easy, i suppose
20:02:37preglowamiconn: same old coldfire stuff?
20:02:48*preglow likes coldfire
20:02:58amiconnWith radio (unlike the M5)
20:03:18amiconnThe most interesting detail will be the LED status display
20:03:48*jhMikeS wants a coldfire that runs a 500MHz and has associative caches...what a beast that would be
20:04:24amiconnThe 500MHz aren't necessary... a data cache would be a very helpful detail though
20:04:36amiconnAnd a faster sdram interface
20:04:57jhMikeSI wonder how fast you'd have to run such a coldfire to equate to say a Gigabeat F
20:05:12pixelmajhMikeS: here's the other report btw.
20:05:16 Quit xushi (Remote closed the connection)
20:06:04amiconnI'm still impressed that PP502x doesn't loose much speed if it accesses uncached (!) sdram, while coldfire struggles like hell if memory intensive code isn't handcrafted to use burst mode wherever possible
20:07:03amiconnjhMikeS: Depends very much on the actual code... when comparing the ape filter performance, you can clearly see that coldfire has higher clock efficiency than arm9 (gigabeat f)
20:07:05jhMikeSthe only thing that brings ARM to a crawl is lack of code caching.
20:08:22amiconnThis is due to the emac
20:08:23 Quit styleism ()
20:09:08jhMikeSemac goes without saying re: multipying
20:10:23 Quit Mouser_X (Read error: 110 (Connection timed out))
20:10:40amiconn Mind the footnote as well
20:10:40preglowyeah, arm hates multiplies
20:10:44preglowcraptacular cpu it is
20:10:55 Join Mouser_X [0] (
20:14:21 Quit cagnulein (Remote closed the connection)
20:14:38 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
20:15:21 Quit FOAD (Read error: 110 (Connection timed out))
20:15:21 Nick FOAD_ is now known as FOAD (
20:15:44*jhMikeS wonders if he should just patch the H10 YUV in SVN or post a patch since the changes make no visible difference on his unit
20:18:03 Join cagnulein [0] (n=Cagnulei@
20:18:38*pixelma tries to remember if there is someone around ATM with a 20GB H10
20:19:00pixelmaArathis: aren't you?
20:19:23Arathispixelma: yes
20:20:01pixelmadid you try mpegplayer and noticed any glitches recently?
20:20:29Arathisdidn't use mpegplayer the last month
20:22:07jhMikeSeven the surest patch with the most instuctions that masks everything hardly has any slowdown
20:22:07pixelmacould you try with a recent build and if you get something like help jhMikeS testing?
20:23:09Arathispixelma: sorry, not at the moment
20:23:44pixelmanevermind then :)
20:24:25*jhMikeS will throw the patches in FS and see which work with preference for the one that has fewer instuctions :)
20:25:43*pixelma wonders if that is on purpose now... :)
20:32:26 Quit miepchen^schlaf (Read error: 113 (No route to host))
20:32:30 Join miepchen^schlaf [0] (
20:35:58 Quit bertrik ("bye")
20:38:45***Saving seen data "./dancer.seen"
20:40:45jhMikeSamiconn: I thought about that re: the uncached ram and some effect is may be from large uncached ram buffers not evicting the working data in the cache since writing so much data basically guarantees flushing of lines. Perhaps there's also a write buffer that always bursts?
20:41:18 Quit MethoS_mobile (Remote closed the connection)
20:45:42 Quit JdGordon|w (" ajax IRC Client")
20:45:45LloreanSlasheri: I just wanted to find out if it was being looked at. :)
20:46:22 Join qwedsa_ [0] (
20:46:26 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
20:51:54 Quit cagnulein (Remote closed the connection)
20:55:04 Join cagnulein [0] (n=Cagnulei@
20:55:56 Quit PhilLight ("CGI:IRC")
20:58:51 Quit JdGordon|w (" ajax IRC Client")
20:59:34 Quit TMM (Remote closed the connection)
21:02:30 Quit qwedsa (Read error: 110 (Connection timed out))
21:03:16 Join TMM [0] (
21:04:13 Join MethoS- [0] (
21:04:55jhMikeSshould mpegplayer be able to play at different speeds? (possible using the pitch setting but needs some special clock handling).
21:07:58amiconnI don't think it should
21:09:40jhMikeSyeah, that's going if crossfeed + video isn't almost as well but none of that needs any special code written beyond using the core DSP.
21:10:34 Quit MethoS- (Remote closed the connection)
21:10:58 Join MethoS- [0] (
21:13:08 Quit MethoS- (Remote closed the connection)
21:13:21 Join stripwax [0] (
21:14:17 Join MethoS- [0] (
21:14:47 Quit MethoS- (Remote closed the connection)
21:24:02 Quit Lear ("ChatZilla 0.9.80 [Firefox 3.0b3pre/2008012804]")
21:31:56 Join sakuramboo [0] (n=sakuramb@
21:32:21 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
21:34:32 Quit Domonoky (Read error: 110 (Connection timed out))
21:41:30sakurambooi just installed rockbox on a 4th gen iPod Photo, everything works beautifully, but when i change the themes, it makes the text in the menus black, is there a way to change that to white text but to also keep the text in the play window black with the iCatcher theme?
21:42:54_kei got a usb-ac-adapter, which can power my usb things, like my ipod. unfortunately it doesnt send any usb host signal, so rockbox simply doesnt start in usb mode, while the ipod firmware does. any hint what i could do?
21:46:31 Quit Mathiasdm ("Yuuw!")
21:48:33 Join goffa_ [0] (n=goffa@
21:52:23 Join perrikwp [0] (n=chatzill@
21:53:18preglow_ke: why would you want it to start in usb mode? rockbox can charge without entering usb mode, afaik
21:53:25 Join sandberg [0] (
21:53:31_kepreglow, yeah? it doesnt seem so here
21:53:41preglownever tried myself
21:54:27 Quit goffa (Read error: 110 (Connection timed out))
21:54:33sandbergWhat's new with the Iaudio port? Any progress?
21:56:03scorche|wiaudio is a brand...not a device...which one?
21:56:14LambdaCalculus37Going home now... good night!
21:56:15sandbergWell, the Iaudio 7 to be specific.
21:56:46sandbergStarted looking at the source code the other day, was hoping to make myself useful...
21:57:05 Quit _ke ("umount /mnt/me")
21:57:06 Quit LambdaCalculus37 ("CGI:IRC (EOF)")
21:58:43 Join Buschel [0] (
22:00:57pixelmasandberg: I think TMM started working on it - not sure if he's really around and if I remember correctly...
22:01:51 Quit jhulst (Read error: 113 (No route to host))
22:04:02 Quit nicktastic ("Leaving")
22:05:25Buschelhi, just read the today's log. did i understand correctly that there was a change in svn (about 16105) which may degrade the battery runtime of ipod videos?
22:05:49Buschelthe 60/80GB ones
22:06:55Buschelthis would explain some strangeness in the last battery benchs which were made with my patch versions...
22:07:28 Join MethoS_mobile [0] (
22:08:35Horschtyou mean 16105 or 16150 ?
22:09:16 Join bertrik [0] (
22:09:36desowinBuschel: redid test on my sansa (same svn revision, same files, same settings, just without patch) - v13 gives 3 hours more runtime; I've put both battery benchamrks on wiki
22:10:19Buscheldesowin: yes i noticed it −− +20% :)
22:11:02Horschtwow, impresive
22:11:51desowinHorscht: from 15h, to 18h
22:12:35Horschtdamn, still 4h behind stock FW for my model :)
22:13:39desowinI dunno if it already beat stock FW at my sansa
22:13:52desowintoo bad OF doesn't play .ogg
22:13:53Horschtyeah, i got an ipod, so...
22:14:50 Quit mf0102 ("Verlassend")
22:14:54Horschtstill impressive, though
22:16:47 Join midgey [0] (
22:16:53preglowhow much does sansa get with retailos, then?
22:18:11Buschelretailos should reach 20h on sansa.
22:18:17desowinit's marketed as 20h
22:18:22preglowthat much
22:18:31Buschelit has a 750mAh battery
22:18:55 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
22:18:57amiconnBuffering is a little slow on G5.5, but I still don't see any "mad" disk activity while buffering
22:19:43amiconnNo skipping either - and that is with default wps, i.e. with peakmeters
22:20:09Buschelthe strange thing is that most of the current measurements on 5.5G 60/80 show battery runtime which are similar to runtimes before HDD poweroff −− when using svn
22:21:12 Quit desowin ()
22:23:25 Quit JdGordon|w (" ajax IRC Client")
22:24:30 Join crzyboyster [0] (n=4b596e5e@gateway/web/cgi-irc/
22:25:23 Quit pixelma (Read error: 104 (Connection reset by peer))
22:25:24crzyboysterI'm having an odd problem with the "new" sliding puzzle plugin
22:27:02crzyboysterI load it up and it looks like this >
22:27:06Buschelthis strangeness in power consumption seems to only happen to the 5.5 60/80GB, my 30GB behaves as expected
22:27:59crzyboysterAnybody know why?
22:28:05 Join JdGordon|w [0] (i=836b004b@gateway/web/ajax/
22:29:34 Quit MethoS_mobile (Read error: 110 (Connection timed out))
22:29:37*amiconn really wonders what GodEater does that he observes this behaviour
22:30:09amiconnThe only strangeness I had was database related
22:30:42 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
22:31:05amiconnAfter putting all the additional music on, I started a rebuild. After finishing the scan, it started to commit right away (using the dircache buffer, I know)
22:31:43amiconnBut then I pressed a button to leave the database browser. Instead of leaving the browser, it stopped committing....
22:32:27amiconnThe only thing I could do was shut down. Everything else hung, but the shutdwon worked
22:32:56pixelmacrzyboyster: did you build rockbox yourself
22:33:01amiconnOn next boot dircache had to do a foreground scan, and db commit restarted. After that, everything was back to normal
22:34:17crzyboysterpixelma: No, it is the current build (from yesterday)
22:35:43krazykitthen it's not very current, now is it?
22:35:46TMMsandberg: I have been working on it, slowly :)
22:36:01crzyboysterWill try with todays build
22:36:48BigBambiwhich one - there are usually many per day
22:36:51sandbergTMM: What's the current status?
22:37:11TMMsandberg: working on an LCD driver, but I'm very strapped for time
22:37:29*JdGordon|w prods linuxstb incase he is watching the log...
22:37:35pixelmacrzyboyster: that looks like something I get if bmps in a plugin change size but keep the name (building my own). It then needs a "make clean" before the bild but that is something the build servers always do (at least they should)
22:37:45TMMsandberg: we can run code, know how to flash the thing, got a way to run code without flashing
22:37:48pixelmanot only in a plugin
22:38:12sandbergTMM: Saw the tcctool, it uploads a new image to RAM and not flash, right?
22:38:14 Join Axio__ [0] (
22:38:17*pixelma starts ignoring her typos ;)
22:38:24TMMsandberg: yes
22:38:37sandbergTMM: So it's safe to play with?
22:38:47***Saving seen data "./dancer.seen"
22:38:54TMMsandberg: totally
22:38:55 Quit Axio_ (Read error: 113 (No route to host))
22:39:10TMMsandberg: well, unless you do something wonky like 'flash the thing' from your own code
22:39:26sandbergTMM: No likely to happen... :)
22:39:36 Join SirFunk_ [0] (
22:39:45sandbergTMM: Was thinking about trying to get some basic stuff working, like the LCD... :)
22:39:47TMMsandberg: truth be told, I haven't worked on it for well over a month
22:45:43 Quit petur2 ("brb")
22:46:02 Quit SirFunk (Connection timed out)
22:50:08 Join p3tur [0] (
22:50:53crzyboysterpixelma: The problem still happens with the latest current build...
22:52:36 Quit Bagder ("*plopp*")
22:53:33 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
22:55:21 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
22:55:29 Quit sakuramboo ("Leaving")
22:55:41 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
22:58:17 Join Domonoky [0] (
23:01:06 Join aneka [0] (n=kvirc@
23:02:49 Join HellDragon [0] (i=jd@unaffiliated/helldragon)
23:03:30preglowpixelma: what do you think about ?
23:03:32 Quit HellDragon (Success)
23:03:47 Quit Buschel ()
23:04:53 Quit perrikwp ("ChatZilla 0.9.80 [Firefox 3.0b3pre/2008012904]")
23:05:12 Quit bertrik ("bye")
23:06:54 Join karashata [0] (
23:07:31pixelmapreglow: hmm... didn't we already have this discussion some time ago? I admit that I don't feel strongly one or the other way
23:07:49 Join HellDragon [0] (i=jd@unaffiliated/helldragon)
23:08:02 Join PimpIntheFlesh [0] (
23:08:16rasherI'm just going to go ahead and be in favour, even though I'm not really a manual writer
23:08:20PimpIntheFleshi have a quick question
23:08:50 Quit Domonoky (Read error: 104 (Connection reset by peer))
23:09:00PimpIntheFlesheverytime i plug my sansa e250r into the charger it makes the Music folder hidden and add several folders how do i stop it
23:09:38preglowrasher: i tilt slightly in favour as well
23:09:43BagderPimpIntheFlesh: you can't, unless you ask sandisk
23:09:49rasherPimpIntheFlesh: That's the OF doing that. You can charge (but not transfer files) if you hold down select while plugging in the cable. Rockbox needs to be running for this to work.
23:10:00preglowpixelma: well, if you don't feel strongly either way, have read enough english manuals to have an opinion on how it's usually done? :)
23:10:28PimpIntheFleshso i have to change those everytime?
23:10:42PimpIntheFleshnot a big deal, but annoying
23:11:09scorche|wPimpIntheFlesh: or you can just change your view setting or use a differently named folder
23:11:09 Quit axionix (Read error: 104 (Connection reset by peer))
23:11:15 Quit crzyboyster ("CGI:IRC")
23:11:27PimpIntheFleshalso what folders can i delete with out causing problems
23:11:40 Join axionix [0] (
23:11:43PimpIntheFleshor i should say what files and folders can i delete
23:12:10Lloreanpreglow: I'm in favour of *mostly* removing contractions, but there are sentences (I don't know if such exist in the manual) that read more awkwardly without contractions, and I think "appropriate use" of them lends readability and makes it a bit more "friendly" feeling to less technical readers.
23:12:23scorche|weverything but /system/ and /.rockbox/ should be fine
23:12:31PimpIntheFleshthank you
23:12:34LloreanBut I'll admit, so long as nobody's going to go through on a case by case basis and pick out some to keep, it's probably better just to be rid of them all.
23:12:47PimpIntheFlesha let me say this thing is freaking awsome
23:12:53PimpIntheFleshkudos to the developers
23:13:22pixelmapreglow: no, but I'm sure it sounds more professional without contractions. The question is if we want that? :)
23:14:00PimpIntheFleshthank you guys for your help
23:14:34HorschtI think a manual should refrain from using contractions, tbh
23:14:38Lloreanpreglow: As a "technical" document, it shouldn't have contractions. As something nontechnical humans read, and need to feel comfortable with (and strange things make people less comfortable) I think it could be debated without end.
23:15:07 Quit PimpIntheFlesh ("Leaving")
23:15:07Horschtmakes it easier for us non-native english speakers
23:15:43rasherI think it should tend towards "no contractions", and problematic passages can be fixed afterwards
23:16:35pixelmaah, the wrong bmp is used for the Nano sliding puzzle - wonder how that was meant to work...
23:17:08pixelmaI understand what it should do but that doesn't work for the Nano...
23:17:10Horschti even go as far as saying to use the negation as little as possible. i.e. instead of saying "thou shalt not kill", say "you should let others allive" (bad example, but it does count as one!)
23:17:58Horschtyou know, be more positive about it
23:18:06preglowLlorean: problem is formulating that as a nice manual writer rule...
23:18:27Lloreanpreglow: Yeah. I think the best bet is just removing contractions, and keeping it that way.
23:18:35LloreanAt least it guarantees consistency.
23:18:36rasher"Avoid contractions, unless it sounds awful"
23:18:39 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
23:18:47preglowrasher: yeah, i guess that'll do
23:19:28pixelmapreglow: had a look through the patch and in most places I'd also use the long form only in a very few not
23:19:52Bagder"in case of doubt, we play rock, paper and scissors about it on IRC"
23:21:04Lloreanscissors. I mean paper, paper!
23:21:24Horschttoo bad, you said scissors first
23:21:39pixelmathat would be rock, box, and scissors in correct rb terms ;)
23:21:42*scorche|w is reminded of,-ever!.aspx again...
23:21:47LloreanI think the safest bet is to just remove all contractions, and if anyone complains about a sentence being hard to understand, we can fix it later.
23:22:02 Quit ompaul (Client Quit)
23:22:05rasherLlorean: sounds good
23:22:07Horschtbox would always win, no, pixelma?
23:22:25pixelmadepends on what it is made of
23:22:42Horschthm.... software
23:22:43scorche|wBagder: that link may remind you of many early sansa R adopters :)
23:22:55Horschtanyways. nn
23:24:23Bagderearly Rockbox adopters have proven many times to blindly jump before asking any questions or thinking about what they actually are doing
23:24:50Bagderthat's part of the Rockbox charm ;-)
23:24:59 Quit p3tur ("gonne")
23:25:14*preglow finds a track called "rock box" and vows to start a port so he can play it as first track
23:25:33Horschti had my ipod unpacked for 5 whole minutes until i first installed rockbox bootloader
23:25:34Bagderpreglow: zune! ;-P
23:25:39rasherpreglow: nano 3g
23:25:40Lloreanpreglow: There's still time to get the "First song" slot on the Gigabeast
23:26:09*preglow scrambles
23:26:55 Quit JdGordon|w (" ajax IRC Client")
23:29:06 Join Daemo1 [0] (
23:29:31Daemo1How long does it normally take for the rockbox database to create?
23:30:02preglowdepends on target and amount of songs/song format
23:30:11preglowit might easily take a good wehile
23:30:13preglowwhile too
23:30:45Daemo1~500 songs mostly in .mp3 and .m4a. Its taken at least 10mins to get through 72 songs
23:31:21Daemo1sorry I dont follow?
23:31:26preglowwhat kind ofp layer?
23:31:46Daemo1just the inbuilt rockbox one...
23:31:59preglowi mean do you have an ipod, iriver, sansa, etc
23:32:03Daemo1oh sorry
23:32:08Daemo1ipod nano 1st geb
23:32:30preglowahh, uyes[Bahh, yes
23:32:30preglowme too
23:32:39preglowdatabase building is very slow on it
23:33:02preglowargh, shell lags
23:37:51preglowdon't really know why it's so slow either, we're probably doing something wrong
23:38:05preglowpixelma: but ok, if no one else beats me to it, i'll just tune that patch up a bit and apply it sooner or later, then
23:41:11 Join m0f0x [0] (
23:43:55 Quit BigBambi (Remote closed the connection)
23:44:45 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
23:45:55 Quit Thundercloud (Read error: 104 (Connection reset by peer))
23:46:25 Join Thundercloud [0] (
23:48:20 Quit goffa_ (Remote closed the connection)
23:48:48 Join goffa [0] (n=goffa@
23:55:14pixelmahmmmh, the new sliding_puzzle probably doesn't look good on the small H10 either
23:56:13 Quit jcollie ("Ex-Chat")
23:57:44 Join csc` [0] (n=csc@archlinux/user/csc)
23:57:52markunjhMikeS: will there be separate DSP settings for mplayer, or will they use the system ones?

Previous day | Next day