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 2007-08-16

00:00:04amiconn1st/2nd, 3rd, 4th, mini, mini g2, color, nano, video - imho
00:00:16***Saving seen data "./dancer.seen"
00:00:20linuxstbOr that...
00:00:38amiconnHmm, or even put the nanos after the color
00:00:42amiconnThe minis I mean
00:00:58amiconnNah, linuxstb's order is better
00:01:23pixelmaI think I would rather see the Minis and the Nanos seperated from the other Ipods
00:01:43amiconnJust because of the disk size?
00:01:57DerPapstthe date where they are announced by apple is ok i thinl
00:02:06DerPapstwhich would be linuxstb's order
00:02:11pixelmaeven if it's not chronologically right, but I wouldn't expect them in the row - who knows the timeline that well?
00:02:12amiconnIn fact linuxstb's is very close to what seems to be the internal order of development with the exception of the mini G2
00:02:47pixelmaI mean the Ondios could be also somewhere within the Recorders, or not?
00:02:55 Quit BigMac (Read error: 104 (Connection reset by peer))
00:03:02amiconnNope, they came later. Way later.
00:03:14amiconnFrom the lcd parameter table in the OF it looks like there was a mini with PP5002 in development
00:03:32pixelmaanyways I just stated what I would expect, I mean they are different models
00:03:34Bagdertaste this
00:03:44DerPapstpixelma: 1st/2nd, 3rd, 4th, color, video, 1st mini, 2nd mini, nano?
00:03:48amiconnThe archos table is chronologically correct
00:04:04pixelmaDerPapst: yes, something like that
00:04:05amiconnpixelma: Well, they're all different models
00:04:09rasherDerPapst: that's how I'd order them. First by "size", then chronological
00:04:21amiconnBy size?
00:04:23pixelmaok... different production lines or whatever
00:04:35rasheramiconn: "full-size ipod" "medium-size" "tiny"
00:04:37amiconnA mini G2 can be 6GB while a 1st/2nd gen can be 5GB
00:04:44rasheramiconn: physical size
00:04:47 Join sneakums [0] (
00:05:05rasherProduct line is probably a better word
00:05:15DerPapstminis and nanos aren't *real* ipods :P
00:05:16rasherIt's two words, actually.
00:05:48DerPapstjust some striped down ones ;)
00:05:58rasherBagder: why are the irivers separated?
00:06:04*linuxstb was about to ask that
00:06:11Bagderno reason...
00:06:32 Quit robin0800 (Read error: 110 (Connection timed out))
00:06:39linuxstbBagder: 5GB first? ;)
00:07:00Bagdernow then
00:07:04 Join BigMac [0] (
00:07:04BagderI swapped the iaudios too
00:07:18rasherI still vote for ordering the ipods by product line, then age
00:07:29rasherNot that it strictly matters, of course.
00:07:33BagderI think I favour that too
00:07:57Bagderlike now
00:08:11linuxstbrasher: What would your order be?
00:08:16rasherlinuxstb: like it is now
00:08:30*linuxstb can't keep up...
00:08:42Bagderhehe, I moved the minis/nano to the end of the ipods
00:08:50*DerPapst can live with both
00:09:05linuxstbI like it as it is now (or was 5 seconds ago...)
00:09:12 Join Zagor [0] (
00:09:14n1sBagder: nice, now how about that size table, tha inconsistent order always anoyyed me... :-D
00:09:39Bagderoh well, perhaps I'll poke on it
00:09:52BagderI think I'll first get this index2 thing more official look and use
00:09:59rasherWithout these issues solved, Rockbxo development will grind to a halt.
00:10:03amiconnBagder: Eh, where?
00:10:24bluebrotherhow about starting a new row on each manufacturer?
00:11:16Bagderamiconn: on the site, to make it easier for non-devs to get a current build
00:11:35amiconnThe table still looks the same to me...
00:11:36Bagderbluebrother: that would be many lines
00:12:39bluebrotherlooks nice :)
00:14:07pixelmabluebrother: I don't think that's necessary - and I guess that would make it look a bit chaotic
00:14:07amiconn_maybe_ put the iaudios in front of the irivers?
00:14:07Bagdernot a bad idea, as that will align the ipods better
00:14:07amiconnThat would group cf targets together, and would make ipods start on a new row (but not end anymore)
00:14:43 Quit webguest57 ("CGI:IRC (Ping timeout)")
00:14:49DerPapstmhm :)
00:14:52amiconnnice :)
00:14:55pixelmanow the 2 Mini builds are divided...
00:15:26amiconnBut I *might* break that nice little order soon :)
00:15:38n1sV2 rec seems a bit out of place....
00:15:41Bagderyeah, let's stop this now as things will change over time anyway
00:16:01amiconnn1s: eh, why?
00:16:20DerPapstthe manual page should use this order then too
00:16:45n1samiconn: it appears between gigabeats and iaudios here, maybe my page is strange?
00:16:49BagderDerPapst: yes, I'll just work on this a bit more first and when it is settled I'll use the same order on all these tables
00:17:05 Part agm3nt
00:17:11amiconnNot here...
00:17:19amiconnMaybe you're looking at a different page
00:18:03DerPapstBagder: ok. take your time :)
00:19:32 Quit perrikwp ("ChatZilla [Firefox]")
00:19:32*amiconn puts a millivolted rockbox build on his recv1
00:19:34n1samiconn: probably a cahche issue, I'll shut up now
00:19:58amiconnLet's see what I've messed up. The charging algorithm is somewhat.... complicated....
00:20:10pixelmaare the 2 Ipod Video "builds" necessary on the manual page? Does it need to be that consistent there?
00:20:44DerPapstprobably not
00:20:52amiconnWell, *someone* should start working on the ram layout
00:20:52DerPapstsince they are exactly the same
00:20:55 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
00:21:01pixelmaeh, was looking at the daily builds page, sorry
00:21:11 Quit stripwax (Read error: 104 (Connection reset by peer))
00:21:12*DerPapst grabs the frist lds file
00:21:36 Join elinenbe [0] (
00:21:36*n1s hands DerPapst a pot of coffe
00:22:10*pixelma runs away from typo classes ;)
00:23:28*Bagder goes to sleep
00:24:37 Join Llorean [0] (
00:25:00DerPapstifdef hell is indeed not an exaggeration
00:25:14*DerPapst is scared
00:25:29amiconnwhere? The lds files?
00:25:34linuxstbBagder: I've just noticed that the Source Archive link in the current build table doesn't have a revision number - is that possible to add?
00:25:53amiconnThere are quite a few ifdefs there. Not the biggest concentration of them though
00:26:06Bagderlinuxstb: sure is... I'll work on it
00:26:28linuxstbBagder: OK. (and sorry to wake you)
00:26:39pixelmabtw... now the Ipod Video manual is only under the heasline "Ipod Video 30GB" - some of them might think that there is a different (or none) for the other Ipods...
00:26:58petur[00:22] * pixelma runs away from typo classes ;)
00:27:10*pixelma wasn't fast enough
00:27:11LloreanShouldn't the headline by "iPod Video Any" for the 30gb, since it's compatible with all of them?
00:27:13linuxstbpetur: She's obviously graduated...
00:27:36amiconnLlorean: 'by'?
00:27:47LloreanApparently I should enroll.
00:27:49linuxstbLlorean: Or just "iPod Video" ?
00:28:12 Quit ender` (" First things first, but not necessarily in that order.")
00:28:13Lloreanlinuxstb: I'd be happy with that too, but I thought at least the 30gb build used to be called that anyway?
00:28:23bluebrotheranyone wants to practise typing? :D
00:28:39*DerPapst notices that his typo clases are a huge success :D
00:28:58*bluebrother just uploaded a screenshot to the wiki
00:29:24Bagderpixelma: fixed
00:31:00amiconnHmm, just one wrong voltage display in debug :)
00:35:20 Join webguest57 [0] (i=1824134c@gateway/web/cgi-irc/
00:37:34webguest57i extracted a theme into my .rockbox folder and it's not showing up... how do i make it show up?
00:38:59bluebrotherwell, I just answered your lol ...
00:39:11bluebrotherthat was the first that popped into my mind.
00:39:21bluebrotherlmao would have been the next.
00:39:28pixelmaor rtfm
00:39:38*bluebrother is nice today :)
00:39:52bluebrotherseriously, "it doesn't work" is _really_ a bad description.
00:40:25bluebrotherit's lacking a bunch of details ...
00:41:10 Join Llorea1 [0] (
00:41:29*amiconn found an oold bug in the battery voltage debug delta display
00:41:38pixelmawebguest57: maybe you'll find something in the "Browse Themes" paragraph of our nice manual, don't know exactly if that's what you are looking for
00:41:54webguest57ok well i just installed rockbox fresh and i went to the themes page, downloaded a theme and extracted the folders into the .rockbox folder
00:41:56 Quit rotator ()
00:42:13bluebrotherand what does the description of that theme say? What build did you install?
00:42:16 Quit Llorean (Nick collision from services.)
00:42:19 Nick Llorea1 is now known as Llorean (
00:43:25webguest57the current build
00:43:58 Quit Zagor ("Client exiting")
00:44:08 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
00:44:25Febswebguest57: which theme are you trying to use?
00:44:51*bluebrother already asked that and waited patiently for the answer
00:46:16*amiconn wonders whether he should create battery-<target>.c files, containing both the adc->voltage conversion and the various tables for percentage calculation
00:46:22BigBambiAnd have you selected the theme from the menu?
00:46:24bluebrotherwebguest57: what does the "Min Version" line tell you?
00:46:46 Quit ompaul ("night all")
00:46:48webguest57senab build?
00:46:51amiconnDoes that make sense?
00:47:03bluebrotheris senab build the same as current build?
00:47:09JdGordonamiconn: where are they now?
00:47:13webguest57i dont know?
00:47:26BigBambiThe survey says... No
00:47:27Lloreanwebguest57: No, they aren't, that's why they're named different things...
00:47:29bluebrotherwell, usually someone would assume it's not
00:47:48amiconnJdGordon: In powermgmt.c, and the current adc->voltage conversion is too inflexible to adapt to the combined 1st/2nd gen build
00:47:59Febswebguest57: The theme you want to use requires that you use a custom build. That is why it won't work with the standard current build.
00:48:00JdGordonthen go for it...
00:48:16amiconnLots of ifdefed tables there, DerPapst would like them ;)
00:48:44bluebrotherclicking that "Senab build" link would even have had information about that ...
00:49:30n1sJdGordon: thanks for fixing the STRINGCHOICE translation :-)
00:50:00Febswebguest57: so you have two choices. You can either choose a different theme (one that works with a standard build), or you can install the custom, unsupported build.
00:50:07JdGordonn1s: :)
00:50:08n1sI have one other menu related bug if you're interested?
00:50:18JdGordonput it on fs.. need to gget to uni soon
00:50:25n1sJdGordon: ok
00:50:34webguest57where else can i download the senab build?
00:50:38webguest57the link is down
00:51:13bluebrotherall builds except the official one are unsupported.
00:51:48bluebrotherwhich means we don't care. Btw, that build is heavily outdated anyway.
00:52:01*bluebrother suggest using a theme that works with an official build
00:52:19Febswebguest: you can find unsupported builds here: But as bluebrother mentioned, we don't provide support here for them.
00:53:24 Quit Siku ()
00:54:51DerPapstGood night everyone
00:54:55 Quit DerPapst ("So Long And Thanks For All The Fish!")
00:55:39webguest57can i have more then one theme on my ipod?
00:56:45bluebrotheryou already have. There are some preinstalled ...
01:00:09bluebrothernight all.
01:00:21 Quit bluebrother ("*zzz*")
01:18:24 Quit lee-qid (Read error: 110 (Connection timed out))
01:18:28 Quit JdGordon ("Konversation terminated!")
01:31:48*n1s -> sleep();
01:31:51 Part n1s
01:32:45 Quit Febs (Read error: 110 (Connection timed out))
01:36:17 Quit hannesd_ (Read error: 104 (Connection reset by peer))
01:37:06 Join hannesd_ [0] (
01:37:32 Join hannesd__ [0] (
01:42:06 Join DWSR [0] (
01:43:09DWSRhey everyone. I'm having a problem with my Rockbox install. I have the database set to auto-update, and it worked the first time I did so, but I changed some of the songs that are on my iPod (Nano 1G) and now it does not find the songs that I've added.
01:50:57 Join cmvjk [0] (
01:51:19 Part pixelma
01:51:23DWSRhey everyone. I'm having a problem with my Rockbox install. I have the database set to auto-update, and it worked the first time I did so, but I changed some of the songs that are on my iPod (Nano 1G) and now it does not find the songs that I've added.
01:52:09krazykitpatience, man. the "i" in irc doesn't stand for instant, you know.
01:52:29krazykitdoes updating manually fix the problem?
01:52:37krazykiter, add the songs
01:52:57DWSRManually updating what songs are on my iPod?
01:53:07 Quit hannesd (Read error: 110 (Connection timed out))
01:53:07 Nick hannesd__ is now known as hannesd (
01:53:07 Quit hannesd (Killed by (Nick collision))
01:53:07 Nick hannesd_ is now known as hannesd (
01:53:45krazykitno, manually updating the database
01:54:01DWSRThat actually reverses the situation. It only finds the stuff I added, but nothing else.
01:54:14 Quit bdgraue (Remote closed the connection)
01:54:22 Join hannesd_ [0] (
01:54:39krazykitwell, i can't help you much. i don't use the database at all
01:55:20 Join perrikwp [0] (n=chatzill@
01:55:37*petur thinks the MCIMX31 reference manual is crap (as far as USB is concerned)
01:58:04*amiconn hopes for green
02:00:19***Saving seen data "./dancer.seen"
02:01:03 Join aliask_uni [0] (i=82c20d65@gateway/web/cgi-irc/
02:01:15*petur spots yet another TMS320 'port' in the wiki
02:01:47amiconnHow can there be another? We don't even have one...
02:01:57*amiconn looks at linuxstb
02:02:10*amiconn whispers AV300
02:02:37peturI meant as new, starting or would be port
02:02:59aliask_uniPity that wiki is down though
02:03:22aliask_uniFor the m:robe M500
02:03:54amiconnArgh, red, but it wasn't me. One build server is short on diskspace...
02:04:15amiconnWho runs ?
02:05:06peturI thought Bagder did a cleanup each time?
02:05:46amiconnThat doesn't clean ccache of course
02:05:48homieloweI have been having trouble trying to find info since the initial wiki by the guy who put linux on it seems to have dissapeared+ google cache seems to hate me
02:09:03 Quit DWSR ("Leaving")
02:09:11 Join Febs [0] (
02:10:04 Quit Arathis ("Verlassend")
02:13:05aliask_unihomielowe: Have you tried emailing him?
02:13:17peturhmmm all those freescale USB controllers look the same... the datasheet of the MFC5329 contains a slight bit more info, still not much :/
02:13:34*homielowe slaps his forehead
02:17:52aliask_unihomielowe: Do you have one? (at)
02:18:28aliask_uniAs in, do you own a mr500i
02:20:29homielowealiask_uni: I would dissasemble one if I had one, there is one member at another forum that is willing to mail a MR-500 to a rockbox developer for free as long as he will get it back(eventually), this member currently owns two
02:24:29aliask_uniIt sure looks interesting, but the whole touchscreen thing would be difficult
02:29:06homielowealiask_uni: shirour(the porter) had a demo running with a touchscreen driver implemented
02:29:26 Part homielowe
02:33:48 Quit petur ("Zzzz")
02:36:33webguest57is there any other emus for rockbox besides rockboy and the other one that i can't recall?
02:37:04Lloreanwebguest57: Everything official is built-in and included in the manual.
02:40:46aliask_uniWhat does rockbox need from the .rockbox folder to boot?
02:41:47LloreanJust to boot, all it needs is (.ipod, .iriver, .mi4, whatever)
02:41:56LloreanBut it won't be able to play music or do *too* terribly much without other files
02:43:02aliask_uniI'm thinking about tacking the rockbox.gigabeat file on to the end of the bootloader file and transferring it across, but I can't put the .rockbox directory on the hdd using this method
02:44:28aliask_uniDoes that sound feasible?
02:44:47LloreanI don't know much about the player: You can only copy certain file formats to it?
02:46:07aliask_uniWell, any time you upgrade the firwmare using the official software, it formats the main partition - so even if I could copy files across to it, they'd be wiped.
02:46:22LloreanI mean, on the iPods, you don't necessarily need a bootloader, you can put the main rockbox binary (rockbox.ipod) in the right place instead.
02:46:33LloreanBut you wouldn't be able to play music until you can get the codecs across somehow
02:47:00aliask_uniYes, it'd just be for testing the main binary while we're waiting for the USB driver
02:47:15LloreanThen that'd probably work, yes.
02:49:20aliask_uniAlso interestingly, the original firmware has usb msc dll's, so we might be able to get into MSC mode. No idea how though.
02:50:17LloreanThat'd of course make things easier.
02:51:06aliask_uniIndeed, but I'm not much good with a disassembler so I'm a bit lost as to how to access it.
02:52:27LloreanWell it's somewhat possible that it's just part of WinCE distro, and use of it isn't implemented at all, really
02:52:46aliask_uniQuite right
02:53:10LloreanIs anything/much known about the USB hardware for the purpose of a separate USB driver for in the bootloader?
02:53:54aliask_uniWell, it's all in the datasheet
02:54:08aliask_uniThe USB hardware is all on the CPU
02:54:21LloreanWell, I'm not familiar with the datasheet, and I know those can be of varying quality. ;)
02:54:50aliask_uniI haven't really looked into it, but I believe austriancoder is using the very same datasheet for his work
02:55:15LloreanAh, then hopefully you'll soon have something to work with, right?
02:55:43LloreanI mean, once you've got a bootloader USB mode, the question about the MSC dll doesn't matter too much, I think
02:55:53aliask_uniAnyway, I've got a lecture
02:56:09 Quit aliask_uni ("CGI:IRC")
03:02:21 Quit Thundercloud_ (Remote closed the connection)
03:05:19 Quit datasleep (Read error: 104 (Connection reset by peer))
03:11:07 Quit thegeek (Read error: 104 (Connection reset by peer))
03:19:53 Quit kubiixaka (Read error: 110 (Connection timed out))
03:26:33 Join EbErT [0] (
03:34:49 Join perrikwp_ [0] (n=chatzill@
03:38:03 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
03:48:51 Join HellDragon [0] (n=Nocebo@unaffiliated/helldragon)
03:53:01 Quit perrikwp (Read error: 110 (Connection timed out))
03:54:32 Quit perrikwp_ ("ChatZilla [Firefox]")
03:54:44 Join perrikwp_ [0] (n=chatzill@
03:54:46 Nick perrikwp_ is now known as perrikwp (n=chatzill@
03:56:42 Quit Nico_P (Remote closed the connection)
03:57:59 Join jhulst [0] (
04:00:21***Saving seen data "./dancer.seen"
04:00:26 Nick fxb is now known as fxb__ (
04:08:29 Quit BigMac ("Leaving")
04:27:40 Nick fxb__ is now known as fxb (
04:28:27 Join BigMac [0] (
04:33:50 Quit BigMac ("Leaving")
04:34:07 Join sarixe [0] (
04:36:58webguest57where is a good place to find doom wads? i tried that reccomened site but all the ones i've tried don't work
04:37:20LloreanTry googling.
04:37:27webguest57i just did
04:37:32LloreanThey're normal Doom wads, and don't really have anything to do with Rockbox other than "we play them"
04:37:45LloreanJust like you shouldn't ask "where's the best place to find MP3 files" or "where's the best place to find jpegs" in here.
04:38:19 Join ender` [0] (
05:06:59 Join BigMac [0] (
05:28:06 Join mexicankiller [0] (n=chatzill@
05:28:50mexicankillerHi,guys what a bad day :(
05:30:47 Join saratoga [0] (i=98039a34@gateway/web/cgi-irc/
05:41:36saratogaanyone who can take a look at the build server?
05:43:08mexicankillerwhat do you need?
05:43:38saratogasomeone to make it work again
05:43:59mexicankillersorry I'm the wrong person
05:44:07saratogayeah i know that
05:44:12mexicankillerok xD
05:44:27saratogai guess i need to wait a couple hours for the europeans to wake up
05:44:53mexicankillerwhat time is there?
05:45:14krazykit6 or 7 am
05:45:28krazykitpretty early.
05:45:31mexicankillerhere is 10:45 pm
05:47:08Lloreansaratoga: I wish they'd make some sort of... not publicly accessible, but available to committers, method of kicking it.
05:47:37 Join saratoga2 [0] (i=98039a34@gateway/web/cgi-irc/
05:47:37 Quit saratoga ("CGI:IRC (EOF)")
05:47:46saratoga2yeah that would be nice
05:47:49 Join JdGordon [0] (i=82c20d66@gateway/web/cgi-irc/
05:48:06saratoga2on the upside, once that build goes through, i'm pretty sure i'll finally have the wma codec faster then ogg on ARM
05:48:26saratoga2coldfire is another story
05:48:28JdGordonhey all
05:48:52saratoga2JdGordon: been meaning to ask you if the sansa LCD gets switched off when the backlight goes off?
05:49:22JdGordonim not sure,I tihnk it does
05:49:44saratoga2yeah i think Toni made a patch for it, just wondering if it ever got in
05:50:23JdGordonI was playing in the backlight code a while ago and I thinki remember it turning off when the BL goes off
05:50:50JdGordontarget/arm/sansa/sansa-e200/backlight-e200.c if you want to check....
05:51:13jhMikeSit turns of the driver chip too now and has staby mode
05:52:09JdGordonLlorean: you dont have access to the webclient config, do you?
05:52:36LloreanI would assume not, since I don't know what you're talking about.
05:52:59JdGordonI dont think#rockbox-community should be blocked
05:53:12JdGordongrr, this keybpard is worse than the one at home...
05:53:34 Quit sarixe ("Peace")
05:53:50mexicankillerJdGordon:what happens with your keyboard?
05:55:20JdGordonthe space is broken, i have to use too much force on it
05:55:28JdGordon... it feelswrong
05:56:12mexicankillerthrow your keyboard about 1 inch
05:56:37JdGordonuni might not appreciate me bhreaking their equipoment
05:56:41mexicankillerin the same place as it is
05:57:01mexicankillerit doesn't going to break
05:57:16mexicankillerI did that with mine and things work better
05:57:57JdGordonand fixed
05:58:01*JdGordon rules!
05:58:35mexicankillerso did you do that?
05:59:30JdGordonno, pulled it out and reseated it
05:59:34JdGordonthe space bar
06:00:24***Saving seen data "./dancer.seen"
06:00:26 Join killdeas [0] (i=bd9cea8a@gateway/web/cgi-irc/
06:00:47JdGordonanyone knoow what timezone nls is in?
06:01:27scorcheJdGordon: IrcNicks?
06:01:27 Quit killdeas (Client Quit)
06:01:39JdGordonTZ not listed :(]
06:02:05 Quit JdGordon ("CGI:IRC (EOF)")
06:02:06 Join killdeas [0] (i=bd9cea8a@gateway/web/cgi-irc/
06:03:24 Quit killdeas (Client Quit)
06:06:23 Join killdeas [0] (i=bd9cea8a@gateway/web/cgi-irc/
06:07:25 Quit killdeas (Client Quit)
06:08:12 Join killdeas [0] (i=bd9cea8a@gateway/web/cgi-irc/
06:08:24 Quit killdeas (Client Quit)
06:08:55 Join killdeas [0] (i=bd9cea8a@gateway/web/cgi-irc/
06:09:41 Quit killdeas (Client Quit)
06:10:02 Join killdeas [0] (i=bd9cea8a@gateway/web/cgi-irc/
06:10:05 Quit killdeas (Client Quit)
06:14:36 Join JdGordon [0] (i=82c20d68@gateway/web/cgi-irc/
06:17:11 Join kubiix [0] (
06:19:22*jhMikeS just found something very, very wrong with timer handling on pp targets
06:21:34JdGordonhow very very wrong?
06:22:07jhMikeSextremely badly super wrong :P
06:22:37JdGordonhaha, dangerous wrong? or just a stupid mistake?
06:22:39jhMikeSreading TIMER1_VAL from both cores
06:22:50jhMikeSno, dangerous
06:24:50jhMikeSwhen coordinating cores for thread wakeup, it would slow to a crawl
06:25:34JdGordoncould this be related to our cop problems?
06:26:17jhMikeSvery well could be since reading it from the COP will cancel the interrupt for the CPU for some time
06:28:49jhMikeSheh, the games run so much smoother too :)
06:29:49 Quit Rick (Read error: 104 (Connection reset by peer))
06:30:51 Join Gibbed [0] (
06:30:52 Join gotthardt [0] (
06:30:58JdGordonjhMikeS: you seen 7598?
06:31:04JdGordondircache for sansa :)
06:31:12LloreanjhMikeS: Games run smoother? Are you running things on the COP?
06:31:15JdGordonnot that its really needed
06:31:43jhMikeSLlorean: just running timer ticks on the CPU only and waking the COP during the tick
06:31:51 Part gotthardt
06:31:54saratoga2is gcc smart enough to use the store multiple and load multiple ops on ARM?
06:32:04 Join gotthardt [0] (
06:32:10jhMikeSsarataoga2: it does quite often
06:32:24jhMikeSJdGordon: I agree...not much point
06:32:43LloreanjhMikeS: Database auto-update/removal
06:32:55LloreanIf anything, Database should be locked on rather than off.
06:32:55JdGordonno, but it would be useful for future targets which may need dircache and have hotswap
06:33:02LloreanErr, dircache, not database
06:33:14JdGordonyou mean default on?
06:33:27 Join ToHellWithGA [0] (
06:33:28LloreanJust remove the setting entirely, but with it on instead of off.
06:33:35LloreanAssuming there's no harm to "on" by that point
06:33:44JdGordonamiconn will KILL you if that happened!
06:34:32 Part gotthardt
06:34:37JdGordonthe dircache dependacy in database shold be removed though
06:34:38jhMikeShmmm...I get a higher boost % but it could be because it sleeps more
06:35:33jhMikeSoh, no...I know why :P
06:35:33JdGordonshuoldnt more sleeping lower the %?
06:35:42LloreanJdGordon: He'd kill me if it was locked on just for Sansa and maybe Nano?
06:35:45ToHellWithGAi'm probably just missing it, but is there a "clear playlist" button?
06:35:46LloreanTo remove any confusion?
06:35:47jhMikeStimers change speed with CPU frequency I think
06:36:06ToHellWithGAs/button/menu option/
06:36:22JdGordonLlorean: oh, I thought you meant for all targets. yeah removing the option and enableing it for flash disks wouold be OK
06:36:29JdGordonslightly longer load time thuogh
06:36:36scorcheToHellWithGA: just stop playback...
06:36:54jhMikeSthe counter during normal frequency will count off fewer ticks then
06:37:25saratoga2does this COP issue affect codecs in any way?
06:37:36JdGordonLlorean: actually, I tihnk just changing the defaut to on would be better, (on flash targets)
06:38:14jhMikeSsaratoga2: I hope so so they can run on it soon
06:38:33saratoga2also, how many registers am I allowed to use on ARM?
06:38:48saratoga2can i change the frame pointer for instance?
06:39:38jhMikeSsaratoga2: r0-r3 = scratch, the rest are non-volatile (r12 seems rather ambiguous though). r14 can be scratch with the routine
06:40:41webguest57is there a way to lock the volume at a certain deceble?
06:40:44jhMikeSsaratoga2: you can allocate stack space with sub sp, sp, <shifter>
06:41:04saratoga2so i shouldn't be using r12 or 15 in my asm functions?
06:41:06JdGordonwebguest57: there is a patch for that
06:41:25webguest57thank god lol
06:41:30jhMikeSr14 = lr so you can stack that return address and then use r14
06:41:55jhMikeSr15 = pc so you can do add pc, pc, <shifter> to do jumps
06:42:26saratoga2if i tell gcc i'm going to write to r14 it will be smart enough to push the stack pointer for me, right?
06:42:55jhMikeSyes, if it's in the clobber list
06:43:18webguest57where can i find the patch? under patches? lol
06:43:18saratoga2thanks, good luck with the COP
06:44:11JdGordonjhMikeS: does the sansa charge in rockbox with usb? I seem to remember it only charges at like 100mA or something?
06:44:12jhMikeShopefully that badness takes care of problems on pp5020
06:44:32JdGordon... as apposed to the full 500mA
06:44:33LloreanJdGordon: Same as iPods when you charge them from normal (non-wall-charger) USB
06:44:37jhMikeSJdGordon: I don't know the current but it seemed to charge when I left it there
06:45:05 Join toffe82 [0] (n=chatzill@
06:45:08JdGordonok cool
06:45:54webguest57JdGordon: Where can I find that patch? What's it called?
06:46:23JdGordonon the patch tracker... i dont know what its called, it likeoly it was closed so you have to search closed tasks also
06:49:49webguest57would you know what patch catagory it would fall under?
06:50:11JdGordoni just remember there is one :)
06:50:17*JdGordon off to class
06:50:47 Quit JdGordon ("CGI:IRC (EOF)")
06:56:05 Quit ToHellWithGA (Read error: 110 (Connection timed out))
06:56:50webguest57lol I just found that too
06:57:31webguest57How do I install it?
06:57:57saratoga2webguest57: see the working with patches wiki page
06:58:07saratoga2i believe its listed under the index link on the front page
07:02:57webguest57ok that's too confusing for me
07:03:33webguest57You guys should incorperate something like that in your next build or somethin'.
07:05:24saratoga2webguest57: patches that should be incorporated into builds are
07:05:37saratoga2we choose not to commit many of them
07:06:16LloreanThat patch wasn't implemented in the right way, iiuc.
07:07:29LloreanAlthough since my target has discrete volume adjustment, I just keep track of the level in my head.
07:09:24 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
07:09:34webguest57I made a folder in the root of the .rockbox filder and i put some gameboy roms in it... How do I get to it?
07:10:12alienbiker99move it to the root of your player instead, it will be easier
07:10:38alienbiker99but you have to select show all in file view to see .rockbox
07:12:07webguest57ok ty
07:13:19saratoga2is it legal to do this in ARM: "sub r1, r1, #16;" ?
07:14:05jbitwhy wouldnt it be?
07:14:14 Quit mexicankiller ("ChatZilla [Firefox]")
07:14:27saratoga2i'm having trouble with the assembler
07:15:43webguest57is it possible to make rockboy a little faster/less choppy?
07:15:46jbityou can even do short-hand "sub r1, #16" in most arm assemblers (which will result in what you put)
07:16:05saratoga2i'm trying to do this without luck
07:16:07saratoga2"sub p, p, #32;"
07:16:17saratoga2where p is a variable passed in and listed in the clobber list
07:17:04jbitin gcc inline asm?
07:17:31saratoga2" Error: register expected, not 'p,p,#32' −− `sub p,p,#32'"
07:17:32jbitunless soemthing big changed.. that isn't how it works... the boost back down to normal :\
07:17:39jbityou have to use %0,%1,%2,etc
07:17:52jbitrefering to first, second, third, (etc) arguments
07:18:45saratoga2oh forgot the brackets
07:19:48saratoga2 "sub %[p], %[p], #32;" works as expected
07:20:51 Quit elinenbe (Read error: 110 (Connection timed out))
07:21:06jhMikeSamiconn: (for the logs) if you thought the scheduler was complicated before ... ;D
07:22:40saratoga2jhMikeS: so no luck with these fixes?
07:23:30webguest57How can I speed up RockBoy?
07:25:39EbErTdrop frames
07:26:21 Quit jhulst (Remote closed the connection)
07:26:31jhMikeSsaratoga2: it's working fine
07:26:55webguest57its set at 6 max
07:27:14 Part EbErT
07:27:48jhMikeSit runs much better than svn now
07:29:13saratoga2jhMikeS: you mean COP is working?
07:29:54webguest57well i cant wait when this becomes 100%
07:30:02webguest57Pokemon rules :P
07:30:53jhMikeSsaratoga2: it has been working but things are smoother without that timer problem
07:31:15saratoga2jhMikeS: well i mean working and getting the anticipated speedup
07:31:27saratoga2i assume this is a pretty big step towards that?
07:31:29jhMikeSI'm not using it for decoding yet
07:32:02saratoga2well good luck, i'm getting some sleep
07:32:14jhMikeSIf things will run as on the dual targets with the proper scheduler mods then that can be done next
07:33:34jhMikeSjust fixing the timer thing seems to have dropped a 1-2% boost though...hmm
07:37:11 Quit saratoga2 ("CGI:IRC (Ping timeout)")
07:37:41 Quit alienbiker99 (Read error: 104 (Connection reset by peer))
07:37:46 Join Doomed_ [0] (
07:39:43 Join alienbiker99 [0] (
07:40:53 Part toffe82
07:42:10 Join desowin [0] (n=desowin@
07:56:01 Quit Doomed_ (Read error: 110 (Connection timed out))
07:58:03 Quit krazykit (Read error: 110 (Connection timed out))
08:00:25***Saving seen data "./dancer.seen"
08:01:33 Quit alienbiker99 (Read error: 110 (Connection timed out))
08:02:02 Join perrikwp_ [0] (n=chatzill@
08:02:38 Join RaRe` [0] (n=Laffin_B@
08:10:41markunwebguest57: which player do you have?
08:11:30 Quit perrikwp_ ("ChatZilla [Firefox]")
08:11:44 Join perrikwp_ [0] (n=chatzill@
08:17:00 Quit webguest57 ("CGI:IRC (EOF)")
08:18:12amiconnjhMikeS: Eh, what-where?
08:19:30jhMikeSyou mean abou the timer? in TIMER1()
08:20:14 Quit perrikwp (Read error: 110 (Connection timed out))
08:21:09 Quit RaRe (Read error: 110 (Connection timed out))
08:21:20amiconnI mean your remark regarding the scheduler
08:22:44amiconnRegarding TIMER1_VAL, why would it cancel the interrupt for the other core? Does the interrupt controller not latch the interrupts?
08:24:59amiconnBtw, the PP interrupt controller has an interesting property regarding core wakeup
08:25:25jhMikeSwhen syncing the core wakeups (having to disable interrupts first)
08:26:09jhMikeSI need two blocks of code that take equal clocks to execute to provide a lock-free sync
08:26:36amiconnIf a core is put to sleep (via CPU_CTRL/COP_CTRL) with interrupts disabled in the CPSR, but some interrupts enabled in CPU_INT_EN(2)/COP_INT_END(2), the respective core will still wake up when one of the enabled interrupts arrives
08:27:03jhMikeSyes, that's where the trouble showed up
08:27:06amiconnIt won't take the interrupt vector of course, but it will wake up
08:27:19jhMikeSit will after reenabling them
08:27:39amiconnyes of course, if they're not cleared before
08:28:13jhMikeSI changed it to sync pulse the COP from the CPU timer to provide the tick there
08:28:19amiconnThis behaviour is perfectly understandable because both core sleep and the interrupt controller are not part of the standard arm core
08:28:51amiconnAnd in fact, without that, wakeup from suspend wouldn't be possible
08:29:15jhMikeSit's mandatory too because you can't have an interrupt during the two code blocks that must sync by cycle
08:29:51amiconn...because in suspend, DRAM is inaccessible, hence enabled interrupts would make the system crash, but the cpu must still wake up on GPIO interrupts (button presses)
08:30:11jhMikeSso you put the handler in IRAM?
08:30:39 Join lost|X40 [0] (
08:30:42amiconnIt had to be there
08:34:55 Quit hfliezfh (Read error: 110 (Connection timed out))
08:36:07 Quit lostnihilist (Read error: 110 (Connection timed out))
08:38:33amiconnI'm not sure I understand...
08:39:08amiconnIs the TIMER1 interrupt enabled for both cores?
08:40:02jhMikeSnot in my implementation and that fixed up the timer problem. I'm just calling core_wake(COP) from TIMER1.
08:41:01amiconnHow does the timer problem show up? I would expect to see problems in svn builds as well
08:41:51jhMikeSit shows up as anything tick based being almost completely unresponsive but scrollwheel and SD insert functioning normally
08:42:22amiconnBut the tick seems to count correctly. 100 ticks per second....
08:43:03 Join alienbiker99 [0] (
08:43:03jhMikeSsomething with coordinating things causes it to go nearly dead
08:43:19amiconnWhat about svn?
08:43:48jhMikeSthe ticks are smoother than SVN so it does cause some glitching as one core races to read TIMER1_VAL
08:44:22amiconnHmm, I can't follow
08:44:23jhMikeSusing the wakeup call much like OF does is better
08:45:01 Join Wofl [0] (
08:45:02amiconnIn svn the timer definitely works correctly, otherwise we wouldn't observe 100 tick per second, correct?
08:45:09jhMikeSit may be 100 ticks/second but they're not as evenly spaced. it shows in the plugins running
08:45:19 Join ddalton [0] (
08:46:16amiconnI never observed such a problem....
08:46:26 Join Zagor [0] (
08:46:42jhMikeSif you clear the interrupt on the COP then the CPU disables it's own and sleeps it won't get the wakeup, right?
08:47:18 Quit ddalton (Client Quit)
08:47:50amiconnIf it disabled it afterwards, sure. But I would expect both cores to take the interrupt if they both have them enabled in the intc and in CPSR
08:47:52jhMikeSthe COP seems to beat it to the clearing since it's got nothing else to do
08:48:27amiconnThere's another thing in your core_sleep that I don't understand. If the core now goes to sleep again until it gets signalled it has to wake up, how does music playback work?
08:48:34 Join n1s [0] (
08:48:40jhMikeSbut COP leaves the sleep and beats the CPU to the clear, the CPU then never takes it especially if the FIQ is running
08:48:44amiconn(i..e how does the fiq wake up the core?)
08:49:08jhMikeSit wakes it like any other interrupt
08:49:36jhMikeSin fact, I'd had music playing fine and no ability to use buttons
08:50:18 Join aliask [0] (
08:50:19jhMikeSso CPU took FIQ, COP took TIMER1 and left CPU with no timer
08:51:39amiconnHmm, seems I read something wrong. The core only goes to sleep unless there is *already* a signal in the mailbox that it shouldn't?
08:51:52 Join webguest74 [0] (i=c023110a@gateway/web/cgi-irc/
08:51:57amiconnBut how does that solve the TIMER1 problem?
08:52:05jhMikeSmailboxes don't wake cores without and interrupt enabled on them
08:52:09 Join miepchen^schlaf [0] (
08:52:48jhMikeSoh, wait you're talking about my setup
08:53:14jhMikeSIf the mailbox is signaled, it skips the sleep, yes
08:53:58jhMikeSthat's basically used by a kernel object to wake the other core to handle waking a thread on it
08:55:27 Join B4gder [0] (
08:55:53 Join Rob2222 [0] (
08:56:37jhMikeSof course it can be used to just wake the core for any reason. it won't check threads early since there's no core tasks queued.
08:56:37 Quit webguest74 (Client Quit)
08:58:13amiconnDo we need interrupts on the cop enabled at all (in CPSR)?
08:59:08amiconnI would think we could forbid that, just enabling the TIMER1 bit in COP_INT_EN so the cop will wake up from time to time if it wenr to sleep
08:59:12jhMikeSnot for this. I have none enabled atm. for PCM playback on the COP of course. I don't plan on leaving FIQ disabled all the time but having the I2S source masked.
09:00:27amiconnWy do we want pcm on cop?
09:00:37jhMikeSI want pcm on the core that starts it
09:00:44 Join petur [0] (n=petur@rockbox/developer/petur)
09:01:43amiconnWe could also solve the TIMER1 problem without a different core sleep, and without disabling interrupts on the cop
09:01:51 Join atsea- [0] (i=ariel@gateway/tor/x-f9bb9498f790ef3d)
09:02:01amiconnJust stop our interrupt dispatcher dispatch TIMER1 interrupts to the cop
09:03:18*jhMikeS can't parse that one :)
09:03:42jhMikeSI need a different core sleep/wake for the scheduler however so that a message has no latency (well, a couple uS)
09:04:14 Join chrisjs169_ [0] (
09:04:34amiconnsystem-pp502x.c, just remove lines 65 and 66, and the 'else' in line 67
09:05:29 Join Matt-W [0] (i=maw@growl/Matt-W)
09:05:40jhMikeSoops, forgot to remove that one
09:06:25jhMikeSsure, guess that would work. the COP shouldn't mess with clearing it, that's about it really.
09:06:34 Quit BobShield (Read error: 104 (Connection reset by peer))
09:07:25amiconnYes, and the cop will still wake up every tick if the appropriate bit in COP_INT_EN is enabled
09:07:26jhMikeSI'm not sure what the point with TIMER2 is there
09:07:47amiconnTIMER2 might be used on either core. It's the user timer
09:08:08amiconnBut that's currently just an arbitrary decision; nothing uses it on cop
09:08:14jhMikeSit'll need more sophistication than that to work and only one at a time
09:08:42amiconnFor TIMER2, only one core is allowed to have it enabled in its *_INT_EN
09:09:07 Join BobShield [0] (
09:09:11jhMikeSthe HW enforces that or something?
09:09:42 Quit HellDragon (Client Quit)
09:10:28 Join webguest74 [0] (i=c023110a@gateway/web/cgi-irc/
09:12:11*jhMikeS doesn't know if that was a suggestion or something re'd :\
09:12:31amiconnNo, our convention enforces that
09:13:36amiconnTIMER2 is the user timer, and only used for (maximum) 2 things which are mutually exclusive
09:13:40 Quit Rob222241 (Read error: 110 (Connection timed out))
09:14:07amiconn(1) Backlight PWM fading on PP targets with backlight on GPIO, i.e. no hardware PWM
09:14:35amiconn(2) the grayscale library (and some related stuff, like the test_framerate plugin)
09:15:07 Join tedr0ck [0] (
09:16:01jhMikeSah, greyscale runs on COP in mpegplayer
09:16:20amiconnI don't think so
09:16:44amiconn(unless that changed recently)
09:16:46jhMikeSso no greyscale targets are dual-core enabled?
09:16:58 Quit alienbiker99 (Read error: 104 (Connection reset by peer))
09:17:00amiconnAfaik, video decoding runs on cop and audio decoding on cpu in mpegplayer
09:17:21jhMikeSand rendering
09:17:26amiconnAnd on greyscale targets, it's audio that suffers from the greyscale lib sucking cpu power in interrupts
09:18:02amiconnThe cop may call the rendering functions, but the isr is running on cpu
09:18:42jhMikeSvo_draw_frame is called on COP but I've not looked much at the hookin with greylib
09:20:01jhMikeSjust leaving TIMER1 dispatch out for COP seem to be working fine
09:20:34 Quit chrisjs169 (Read error: 110 (Connection timed out))
09:21:26 Quit qweru ("KVIrc 3.2.0 'Realia'")
09:22:39 Join JdGordon [0] (
09:22:45amiconnIn fact the user timer cannot run on cop in its current implementation
09:22:59amiconnfirmware/timer.c line 265
09:23:08amiconnIt always sets CPU_INT_EN
09:23:10GodEaterLlorean, you here ?
09:25:13jhMikeSwouldn't be hard to flip it to the registering core and cancel on the other if one's enabled there
09:25:37amiconnIt can never be registered twice anyway
09:26:15 Join mikeyisback [0] (
09:27:39jhMikeStiming in the games/other plugins is much more regular. a very visible difference.
09:28:25 Join ie [0] (i=c27f0811@gateway/web/cgi-irc/
09:30:18 Quit tedrock (Read error: 110 (Connection timed out))
09:30:22ieAn OT question (it's quiet here :-): does anybody own an iAudio U3? What's your experience with it?
09:30:49B4gderno rockbox, no fun ;-)
09:31:20B4gderguys, try the new current build page
09:31:41ieB4gder: if it has good HW it could become the next RB target ;-)
09:32:02B4gderie: well... good hw is generally not what makes the next target...
09:32:12*jhMikeS notices something missing but isn't sure :\
09:32:24GodEaterie: a new target is born when someone writes rockbox for it...
09:32:34*petur wonders what the difference is with M3, which amiconn claims he will port rb to...
09:32:47B4gderM3 has a coldfire I believe
09:32:54*GodEater oohs and aahs at the new builds page
09:32:56mikeyisbackHello, ive been looking at the installation instructions for about the past hour and i cannot find out how to fix my toshiba gigabeat to run rockbox, can anyone help me?
09:33:32GodEatermikeyisback: which bit of the instructions are you stumped on ?
09:33:33B4gderthe new target order on the build page will be used on the other target tables too soonish
09:33:36jhMikeSthat TCC datasheet seems a bit PortalPlayer-ish somehow...can't explain it
09:33:57B4gderalthough that we have that leaked tcc760 data sheet...
09:34:07mikeyisbackive loaded the firmware and the loader, but when i turn on my gigabeat; it says rockbox loaded and thats it
09:34:11B4gderpity none of the cool targets use the 760 ;-)
09:34:25ieB4gder: wouldn't it be better to sort targets alphabetically?
09:34:30peturB4gder: current build page is indeed better for the users, I'll have to create a shortcut for the build table - two clicks away :/
09:34:32 Quit lost|X40 (Read error: 104 (Connection reset by peer))
09:34:34jhMikeShehe ... if it's like other things, one is like the other
09:34:46B4gderpetur: yeah, I share that feeling
09:34:55scorcheie: they are
09:35:05B4gderjhMikeS: indeed, I'm sure it'll be partly helpful at least
09:35:44B4gderalphabetically doesn't make much sense imho
09:35:55scorchewell, they are alphabetical
09:36:07B4gderthey are, on manufacturer at least
09:36:23scorcheexcept for ipod/iriver actually...
09:36:31GodEatermikeyisback: did you remember to switch off the battery?
09:36:31B4gderah yes
09:36:35iescorche: ah, sorry, I looked at the delta page at that moment
09:36:52B4gderthe delta table has a somewhat funny order ;-)
09:37:01GodEatermikeyisback: so you get the rockbox logo on the screen ?
09:37:01B4gderI'll tweak that one day
09:37:58mikeyisbackno, i get the firmware loading screen, and it says loading rock box........ Rockbox loaded
09:38:03mikeyisbackand then it just sits there
09:39:13*jhMikeS thinks both on-disk and in-RAM size deltas should be included
09:40:44 Quit tedr0ck (Client Quit)
09:40:53 Join davina [0] (
09:40:59B4gderwhy on disk?
09:41:02*jhMikeS won't be able to cheat anymore by hiding things in BSS sections :P
09:41:30jhMikeSbinary vs. loaded?
09:42:19B4gderI just don't see the point
09:42:59mikeyisbacksweet i got it to work
09:43:22mikeyisbacki got a bad fwimg01 somehow
09:43:35 Join lost|X40 [0] (
09:44:25 Join tedrock [0] (
09:45:44 Join chrisjs169__ [0] (
09:47:07jhMikeSjust an idea to show a truer size increase like when stacks sizes are increased or whatever
09:47:47B4gderwe'd need to poke the build script to be able to extract that info
09:47:57 Quit ptw419 ()
09:48:07B4gderbut you're right, it would make sense
09:49:21B4gderadjusting the delta table script is easy if the info would just be provided! ;-)
09:49:26 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
09:50:10 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
09:50:14jhMikeSit's hard to make the build script send the real size info back with the zip?
09:50:50B4gderif it would store that info in the rockbox-info.txt file, it would work nicely
09:51:03B4gderthat's where the current info is stored
09:51:22 Part mikeyisback
09:51:40 Quit Soap (Read error: 110 (Connection timed out))
09:51:51B4gdertools/ is what creates it
09:52:00 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
09:52:16B4gderI figure we'd need to scan the map file or something
09:54:00B4gderhm, is it "loadaddress" - "end" ?
09:54:15jhMikeSI thought the BSS info was public in the header or does that vary too much per target?
09:54:29 Join ddalton [0] (
09:54:45B4gderin the header of what?
09:55:19 Quit ie ("CGI:IRC")
09:55:20jhMikeSthe binary, but then that's encrypted on mi4 anyway. another aspect I haven't explored too much.
09:55:34B4gderyes, and compressed on archos etc
09:55:43ddaltonOk a question for everyone: would you prefer the time to be voiced before the free disk space under the info screen?
09:56:13B4gderjhMikeS: so scanning the map file seems like the better choice really
09:56:38B4gderI'll give it a shot and see
09:57:36ddaltonhow do I check out revision 14324?
09:57:53B4gder"svn up -r14324"
09:57:58pixelmaB4gder: is the buildserver issue already solved?
09:58:09B4gderwhich issue?
09:58:13B4gderwe have plenty...
09:58:38pixelma"deepthought" has no disk space left
09:58:47B4gderoh, I didn't see that
09:58:52n1sany talk clip users around?
09:58:59B4gderI'll disable it
09:59:04GodEaterhehe - I didn't mention the diskspace thing because I assumed someone has already seen it!
09:59:09pixelmaI think it's ender`s
09:59:17B4gderit is
09:59:21ddaltonn1s y?
09:59:32B4gderender`: ping!
10:00:09n1sddalton: I was looking at FS #6237 and since I don't use them personally I would like some input, do you think it's a good change?
10:00:24B4gderso funnily enough, right now the dailies are newer than the current ;-)
10:00:29***Saving seen data "./dancer.seen"
10:00:35B4gderat least for a couple of targets
10:01:11ddaltonn1s let me take a look.
10:01:41n1sddalton: all it does is shortening the hover delay from 1/2 a second to 1/4
10:01:47 Quit chrisjs169_ (Read error: 110 (Connection timed out))
10:01:56n1shover delay is the delay before the talk clip is loaded
10:02:38 Quit webguest74 ("CGI:IRC (Ping timeout)")
10:03:03ddaltonsounds good to me maybe even shorter if possible. I will try it out on a build I am about to get. But yes I would think it would speed file browsing up.
10:03:06 Join chrisjs169_ [0] (
10:03:34ddaltonI will play with it and post an update if I think it should be shorter.
10:03:38ddaltonwhat do you think?
10:03:41n1sddalton: the delay is there to prevent loading of clips if you browse past a lot of files
10:04:28n1sddalton: well, I don' tuse the feature myself so I will listen to what others think :-) to people for and no-one against is a commit :-)
10:04:45ddaltonn1s Well I haven't tried it but surely if you are clicking through a dir quickly it would move to the next clip. (Are the clips interupted?)
10:05:05pixelman1s: amiconn uses them (in car) ;)
10:05:24ddaltonwell I think it sounds good let me try it out and I will give you feedback.
10:05:39n1sddalton: thanks
10:05:46n1samiconn: any comments?
10:06:19n1spixelma: how is voice working? mostly ok?
10:07:22 Join mikeyisback [0] (
10:07:35 Quit perrikwp_ (Read error: 110 (Connection timed out))
10:08:40pixelmaI didn't have problems with builds from the 14th - I just find the downloadable voice a bit slow but that's another thing
10:09:15mikeyisbackOk, i got the rockbox working now, but it wont play mp3s, any idea where i can find codecs for it?
10:09:38n1smikeyisback: they should be included if you installed correctly
10:10:49pixelmamikeyisback: you don't try the .sat encrypted ones?
10:11:04B4gderjhMikeS: there, I've added it rockbox-info.txt now
10:11:23B4gderI'll see how it works and then add to the delta table soonish if it works fine
10:11:30ddaltonn1s If you want to commit it it is a very small patch. And it is very easy to change back if it isn't wanted.
10:11:54mikeyisbackno, it didnt have any sat files on it
10:11:58n1sddalton: yes but IMO it's better to ask first :-)
10:12:18mikeyisbackits my wifes old mp3 player, i had to format it and reload its orginal firmware to get this far
10:12:28n1sddalton: so it works ok for you, is it an improvement?
10:12:29B4gderRAM usage: 1299948
10:12:31ddaltonfor the more advance stuff in rockbox what do I need to know in c? Does one c book cover all of this?
10:12:33B4gder(on e200)
10:13:00B4gderddalton: you never learn the hard real-world stuff by reading books
10:13:11pixelmamikeyisback: then it should work fine if you extracted the complete rockbox build directly to the root of your device
10:13:13n1sddalton: well, knowing c is only one side, you need to know rockbox too
10:13:16jhMikeSB4gder: that was fast :)
10:13:27mikeyisbacki did, yet the mp3s wont play
10:13:32markunmikeyisback: do you get some kind of error when you try to play?
10:13:41ddaltonn1s Well I tried it last year but I can try it again I am just checking out rev 14324 because I have a couple of patches that improve voice feedback. And I really can't be bothered updating anything.
10:13:50ddaltonat the moment.
10:14:23n1sddalton: ok, I'll see if amiconn has something agains it and commit if he doesn't object strongly
10:14:27ddaltonso once I read my c book from front to back what should I do next?
10:14:36n1sddalton: hack!
10:14:42markunmikeyisback: the codec should be .rockbox/codecs/mpa.codec
10:15:10ddaltonso how do I find out how rockbox works. Like settings buttons, voice displaying stuff and all the other stuff.
10:15:52n1sddalton: browsing the source, changing things to see what happends and asking about specifics in here
10:16:35mikeyisbackit was in the root directory.. strange
10:16:55ddaltonn1s how long should it take me to learn the language before coding on rockbox?
10:17:16 Quit Zagor ("Client exiting")
10:17:29markunmikeyisback: maybe try a different unzip program?
10:18:19n1sddalton: I don't know, I started out by doing simpler things and learning as I went... and I'm no wizard in c yet :-)
10:19:26 Quit chrisjs169__ (Read error: 110 (Connection timed out))
10:21:23B4gderperhaps it would be interesting to also store "maximum audio buffer size" etc...
10:22:05mikeyisbackgot it
10:22:10mikeyisbackgot my themes working too
10:22:13mikeyisbackthis thing rocks
10:22:41jhMikeSwhy not? that's just ram_size - loaded_size right?
10:22:54B4gderyeah i guess
10:23:39B4gderbut I'll make use of this first
10:24:11B4gderit'll be interesting to squeeze in this in the delta table...
10:25:09B4gderat next commit the used ram size will get logged for each build
10:26:42jhMikeSthis should be interesting to see :)
10:26:51B4gderI agree
10:27:29*n1s wants to do a one-char commit but waits for amiconn's feedback...
10:27:32jhMikeSIt accounts for anything copied to IRAM then discarded where it matters?
10:27:58B4gderit simply stores the "end - loadaddress" size
10:28:22 Part mikeyisback
10:29:27*jhMikeS wonders why if PP timers don't vary with CPU frequency, the time the COP spends asleep changes proportionally with the inverse of the CPU frequency.
10:30:37 Join DerPapst [0] (
10:31:41jhMikeSactually it's not a proportion, it sleep less time at 24 and 80 than it does at 30 :|
10:33:54amiconnpetur: M3 is coldfire based like the M5 and X5, but a bit more different from them than these two from each other
10:34:25amiconnIt has a different button layout, *no* main LCD, and no PCF50605 power management chip
10:34:44n1samiconn: what do you think about shortening the delay before a talk clip is loaded from 1/2 to 1/4 second?
10:37:39jhMikeShmmm...must just be a timing thing. backlight duration is always correct.
10:40:00 Quit chrisjs169_ (Read error: 110 (Connection timed out))
10:40:04 Join ie [0] (i=c27f0814@gateway/web/cgi-irc/
10:40:15ieamiconn: does it have a HD inside?
10:40:55*jhMikeS now needs some SH ASM learnin' since thread functions will be able to return normally
10:41:03aliaskie: The M3 has a hdd in it
10:44:33amiconnn1s: I think the hover delay is perfect as it is
10:44:42amiconnI wouldn't want it to be shorter
10:45:06 Join webguest20 [0] (i=c023110a@gateway/web/cgi-irc/
10:45:18n1samiconn: I'll reject the patch then
10:45:26amiconnB4gder: Btw, the delta table on the current build page has a bug
10:45:41amiconnProbably introduced with keeping the .zips until the new ones are ready
10:46:06 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
10:46:07amiconnIf a build fails, it shows a delta of 0 instead of -
10:46:23B4gderah, yes
10:47:34ddaltonn1s what things did you start out with?
10:47:37ddaltonin c?
10:48:30B4gdernow, enjoy a sorted delta table
10:48:56B4gdersorted in the same order as the other tables that is
10:48:59 Quit hcs ("Leaving.")
10:49:37amiconnB4gder: I wonder how much use the occupied ram info is, as it doesn't (can't) take buffer_alloc'ed memory into account
10:50:15B4gderright, but otoh the occupied ram can never be less than what is reported
10:50:29jhMikeSamiconn: SH uses regparms?
10:50:36B4gderit's just a hint if it grows or shrinks
10:51:02n1sddalton: since I was interested in rockbox I started out with some small tings and posted patches, this is my first patch that got committed
10:51:26ieB4gder: why is iriver before ipod?
10:51:48B4gderto make the ipods somewhat line aligned in the table
10:51:50amiconnjhMikeS: Yes. r0..r7 are scratch, with r4..r7 being regparms
10:52:02B4gderit is not alphabetical
10:52:03amiconnr8..r15 are preserved (r15 is stack pointer)
10:52:44ddaltonI know we don't like more settings in rockbox. (Why is this?) but couldn't we make a setting under the voice menu that said "hover delay?" and then options like 1 2 3 4 and so on? That way everyone can have it how they like. Just a suggestion.
10:52:59ieB4gder: ah, I thought you wanted to make it alph...
10:52:59jhMikeSis there some cost to offet addressing or it doesn't have that (referring to @%0+ )
10:53:02amiconnB4gder: Yeah, but now some clever dev can make ram usage appear to shrink, by moving stuff from static allocation to buffer_alloc
10:53:29amiconnjhMikeS: @%0+ is post-increment...
10:53:30*jhMikeS hasn't been through the manual yet enough yet
10:53:31Aveany news on the nano bug?
10:53:41B4gderamiconn: right, but the delta always need some further info to make real sense
10:53:48n1sddalton: we have some 180+ settings already, they make the user interface and code complicated. that's the reason we want to keep them to a minimum
10:53:50ddaltonn1s: So you wrote the whole playlist thing? or not. (I haven't checked out the link yet.
10:53:52AveI think llorean has not seen the bug but he has it..
10:54:01 Join freqmod [0] (
10:54:09Avemp3's play back pretty ok but vorbis fails miserably, he should try vorbis just to be sure
10:54:12n1sddalton: no, just what was changed in that diff, like 8 lines or so
10:54:20amiconnjhMikeS: You can only use post-increment on loads, and pre-decrement only on stores
10:54:54n1sddalton: however we are not against adding settings if there is a good reason to do so
10:55:20ddaltonn1s: Well now I know. But I really think two settings that we could have should be: rec button config. So everyone can choose what it does. And then most people are happy And for this one the hover delay. What do you think?
10:55:21jhMikeSah, ok. totally full-descending stack-centric.
10:55:22ieB4gder: if you'd place ipod before iriver it would still align, just the other row
10:55:56ieB4gder: just an idea: start each family in a new row
10:56:06B4gderthat'd make too many rows
10:56:29B4gderactually about everyone has their own idea of the order
10:56:40n1sddalton: the rec-button is a sort-of infected issue and I'd reather not get into it, I'm fine with a hover delay setting if it is felt that it is needed.
10:56:52jhMikeSamiconn: but there's no equivalent of "move.l d0, 8(%a0)" or "mov r0, [r1, #8]"?
10:57:06ddaltonn1s so your saying for example if I added a "Voice time as" setting to p7561 this would be bad. Yes I know noone wants it. And I am guessing it isn't coded very well.
10:57:23ddaltonthe patch I am talking about when I say it isn't coded very well.
10:58:07jhMikeSSH7034.pdf looks a bit sketchy as an instruction ref
10:58:55ddaltonn1s: I really think the rec button could be useful. For example in my case I would want it to voice some info on a long press like time battery level and free disk space. Where as sighted people might want it to go to the radio. Anyway just suggesting this.
10:59:30ddaltonn1s do you know how to do buttons?
11:00:00n1sddalton: I haven't done anything to them, what do you want to do though?
11:01:02 Join bluebrother [0] (i=yiyxPutb@rockbox/staff/bluebrother)
11:01:15jhMikeSah, it does have a measly 4-bit displacement for mov.l :P
11:01:27 Join agm3nt [0] (
11:01:42ddaltonn1s: I want to make a long press of rec perform a function I will write. It will just say the time as well as battery level and maybe free disk space. (If I can be bothered) Because I don't have a lot of time.
11:01:45 Part agm3nt
11:02:05ddaltonn1s any ideas?
11:05:17amiconnjhMikeS: There is, but constant offset only supports small offset
11:06:13amiconnThere are 4 bits defining the offset. It can only be positive. Depending on whether you access char, short or int, the offset can be 0..15, 0..30 (even), or 0..60 (dividable by 4)
11:06:48*amiconn wonders what jhMikeS is doing
11:06:59jhMikeSfinally spotted that. I wanted to make the function clobber-risk-free. GCC used a reg that was part of the non-volatile context on cf and that gave me some major headaches when it did.
11:07:46amiconnWhich function?
11:07:48n1sddalton: I'm not sure, this is the commit that made long-press on rec go to the rec-screen maybe it's a starting point, asking JdGordon is probably a good idea
11:08:02*jhMikeS is adding a small feature to allow thread functions to user return which basically involves calling remove_thread(NULL) from the core after it returns.
11:08:03amiconnSome registers in the context switch are clobbered on purpose
11:09:08JdGordonn1s: i fixed the title in the plugin browser... still working on voice though
11:09:14jhMikeSsome are but there's problems if it does it with save_context which can blast data to be loaded for a newly born thread running for the first time.
11:09:22 Join spiorf [0] (
11:09:30n1sJdGordon: ah, nice :-)
11:10:00amiconnWhy do you want to allow thread functions to return, btw?
11:10:01jhMikeSit started to use a2 on coldfire for %0 which made nice crashes.
11:11:21jhMikeSconvenience. allow certain cleanups. have thread back in core code sooner. I don't want an interface version for each processor (to avoid usesless #ifdef-ing).
11:12:40ddaltonok JdGordon can you help me with this?
11:13:37jhMikeSdoing that itself adds a few instructions for a startup stub
11:13:41JdGordonddalton: busy atm
11:14:01peturddalton: look at the changes of that commit and learn
11:14:20jhMikeSbut actually gets some out of load_context
11:15:25ddaltonpetur: I will do that once I read this c book. Like a lot of people on here have said to me.
11:16:35peturddalton: once you read that, also read the code and try to understand it
11:16:46markunddalton: have you ever programmed in another language?
11:17:28ddaltonyeah but I want to make sure I study everything properly this time. I think I was just trying to get into rockbox to soon. But it hasn't really worked.
11:18:28ddaltonhey under the plugins menu are they folders or menu items that don't speak?
11:18:53peturddalton: yes, and that bug is already in flyspray
11:19:19ddaltonso there menu options?
11:19:29 Quit ie ("CGI:IRC")
11:19:48peturno, categories..
11:20:51ddaltonactually aren't they folders because under .rockbox/rocks aren't there folders. (I haven't checked but that is what I remember)
11:21:23ddaltonso they aren't standed menu options or dirs?
11:22:01ddaltonblind people don't even use plugins anyway so I don't think it is ergent
11:22:34peturthey are folders at the moment but really ought to be translatable names like in a menu
11:22:36jhMikeSamiconn: there also tends to be alot of uncertainty on how to stop terminate a thread and alot programmers do it wrong. I'll just make it easy and less bug-prone.
11:24:16 Quit bluebrother ("bbl")
11:25:35n1spetur: they are translatable, the problem I reported in the tracker was that the actual dirname was shown as title when inside a category and JdGordonsaid he has a fix for that
11:26:04peturoh, I misread it then....
11:29:23amiconnjhMikeS: Don't forget that SH1 works similar to arm when calling subroutines
11:29:44jhMikeSis the return addess linked or stacked?
11:30:10amiconnIt does *not* put the return address on stack by default, but has a 'procedure return' register for that (pr), the equivalend of arm's lr (r14)
11:30:21amiconnJust that pr is a separate register, not an alias name
11:32:13amiconnAnd if you want to store/fetch pr, you cannot use mov, you use sts/lds
11:32:29jhMikeSbsr seems to handle pr. param1 goes in r0, then bsr. bsr just takes a label if near enough?
11:33:27amiconnYes, bsr can be used to call +/-2048 bytes (iirc)
11:34:09jhMikeSwhat about a pointer call? I see a faked br and rts to do the jump for that.
11:36:38amiconnIt's +/- 4096 bytes#
11:36:47amiconnThis is a trick to save a register
11:38:33jhMikeSthere's JMP @Rm, Rm->PC
11:38:58amiconnBut then you'd need another register to clear the start address from the context
11:39:33amiconnThe trick used is that the thread start address is first copied from r0 to pr, then r0 is zeroed, and then this zero is written back into the context
11:39:54amiconnThe rts then jumps to the destination (because it was moved to pr)
11:40:21amiconnTake care - bra, bsr, jmp have a delay slot on SH
11:40:43amiconnSo the mov.l r0,@%0 after the rts *is* executed
11:41:22n1swow, seems like voice for almost all sound settings is broken, it's interrupted after about one second and does not continue untill you exit out of a setting screen...
11:41:44jhMikeSah, a trick that would've frustrated the heck out of me. might actually be able to exploit that.
11:42:09 Quit gtkspert (Remote closed the connection)
11:42:23 Join gtkspert [0] (
11:42:33amiconnIf we want to use jmp, clearing the start address entry requires another register
11:42:42*n1s bets a cookie on pcmbuf_set_low_latency();
11:43:04jhMikeSI can load r0 with the context pointer but after the jmp I suppose so the stub has the context pointer in r0
11:43:25DerPapstthere is another voice issue too. dunno if it's only on the sim or on the target too. if you have started playback on an iPod sim then pause playback and then browse the menu nothing is spoken. then when you return to the wps and unpause playback the last voice is spoken.
11:44:06amiconnjhMikeS: *Never* try to load a register from a pc-relative memory address in the delay slot on sh1.
11:44:48amiconnThat guarantees serious headaches - the pc relative addressing is executed *relative to the jump destination*
11:45:10n1sDerPapst: no voice while paused is a known issue, and one that can't be fixed without some re-architecting of the playback code... IIUC
11:45:17amiconnThere's comments on that behaviour in rockbox' crt0.S
11:45:18jhMikeSit would be mov.l @%0,r0 ... so just a register copy
11:45:30DerPapstn1s: ah.. didn't know that
11:45:38amiconnRegister does work of course
11:46:04*DerPapst hopes for MoB and the rework of playback.c then :P
11:46:29jhMikeSthe initial thread start would be the stub, the thread stub takes the context in r0 and does the dirty work
11:48:02preglowjhMikeS: how's multicore stuff going?
11:48:51 Quit freqmod ("Lost terminal")
11:49:23jhMikeSpreglow: pretty well actually. very fine-grained locking for this. no core blocking unless they want the same resource unlike previous attempts.
11:50:36 Join bluebrother [0] (i=oztNiqrA@rockbox/staff/bluebrother)
11:51:22ddaltonn1s what needs to be done for the voice to work when playback is paused?
11:52:36n1sddalton: I'm not sure really but i think it is something about making the sates of playback and pcmbuf stuff independent so that the buffer can send stuff to the dac when playback is paused or something like that
11:53:38preglowjhMikeS: sounds good
11:54:08preglow'bout time we got serious with dual core
11:54:10jhMikeSthe cache coherency stuff is whole other problem. it does strange stuff. I've got a test_cache plugin that writes identical random data to a cached and uncached buffer on the CPU and has the cop compare the two and return an error if they don't match. odd thing is, if it makes it through the first test cycle, it seems to keep working or otherwise both cores have the wrong data.
11:54:25 Join JdGordon_ [0] (
11:54:31 Quit ender` (Read error: 104 (Connection reset by peer))
11:54:33n1sjhMikeS: it seems like setting low latency when entering the sound menu items cause voice to break, should i pop it on flyspray?
11:55:32jhMikeSn1s: that is a known problem. I think the low latency causes the codec thread to remain boosted which starves the voice thread. I haven't tried to boost them both at the same time yet but that may work.
11:56:15jhMikeSboosted = thread priority boost of course
11:56:50n1sjhMikeS: where should I look to try to change that?
11:56:59 Quit webguest20 ("CGI:IRC (Ping timeout)")
11:57:07preglowjhMikeS: i hate caches that don't manage themselves...
11:57:17jhMikeSn1s: in pcm_buffer.c where it's done
11:57:51n1sjhMikeS: thanks
11:58:09jhMikeSpreglow: they never do. even gigabeat needs management but at least it's documented.
11:58:18 Join ender` [0] (
11:58:36jhMikeScf needs management but only for an instruction cache
11:58:53pixelmasomeone already looked into why voice with enabled spelling of filenames breaks mpegplayer?
12:00:07preglowjhMikeS: they do if they snoop the bus
12:00:18jhMikeShmmm...voice commands keep getting sent after loading it?
12:00:33***Saving seen data "./dancer.seen"
12:01:38jhMikeSpreglow: our caches mind their own business :)
12:05:17 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
12:07:10 Quit JdGordon (Read error: 110 (Connection timed out))
12:09:18aliaskWe can include copyrighted code in rockbox as long as it's released under the GPL, right?
12:10:09dionoeaas long as it's released under a GPL compatible license
12:10:41dionoea(all rockbox code is copyrighted by it's author. AFAIK no one made their code public domain)
12:11:05*aliask doesn't know much about how copyright works
12:11:19dionoeathe FSF website has a nice page about that
12:12:25 Join webguest20 [0] (i=c023110a@gateway/web/cgi-irc/
12:12:47ddaltonI type "nano" with out the quotes into cygwin and it says command not found. does this mean I don't have nano installed?
12:17:30ddaltoncan I install linux apps on to cygwin. I want to install a patch thing that apparently is better for lots of patches.
12:21:17 Quit tedrock (Read error: 110 (Connection timed out))
12:22:03 Join chrisjs169_ [0] (
12:24:36DerPapstrerun cygwinsetup.exe
12:24:45 Join pondlife [0] (
12:24:45DerPapsti bet you can install it there then
12:25:00*bluebrother wonders what this patch thing might be ...
12:25:50ddaltonBut I downloaded the zip file from another website.
12:26:25 Quit webguest20 ("CGI:IRC (Ping timeout)")
12:26:49 Join datachild [0] (
12:26:52ddaltonthe patch tool is called "quilt "
12:27:29pondlifeB4gder: 10000 commits, congratulations
12:28:04pondlifeHeading up to the top 40 too
12:28:23 Join Nick_Brackley [0] (i=cb57494b@gateway/web/cgi-irc/
12:28:41ddaltonwow there are a lot of editors in cygwin. My conection is going at 250 kb per second and it has been going for about 10 minutes.
12:29:06bluebrothersounds a bit like the patch and merge thingy git can do ...
12:29:15ddaltononly on 70%
12:29:33ddaltonso can i install it on cygwin?
12:29:56*jhMikeS wonders how to get a subroutine address in a register if too far for bsr
12:31:29 Quit Nick_Brackley (Client Quit)
12:33:43 Quit BigBambi (Read error: 104 (Connection reset by peer))
12:34:08 Join BigBambi [0] (n=Alex@
12:34:49 Join lee-qid [0] (
12:35:53n1sjhMikeS: I had no luck with the priority stuff, not that I knew what I was doing but I think I set the priority for the voice thread to the same as the codec thread and it made no difference...
12:36:07n1sposted it to fs so it's not forgotten
12:37:10 Join FOAD_ [0] (
12:38:42jhMikeSn1s: ok. was worth a check anyway.
12:39:19n1sjhMikeS: i'm not in anyway sure that I actuallly set the thread pririty correct...
12:41:20 Join tedrock [0] (
12:45:52 Join elinenbe [0] (
12:46:04amiconnjhMikeS: Jost load the subroutine address into a register and use jsr
12:46:20amiconnVery much like long calls work on arm
12:46:41 Join agm3nt [0] (
12:47:57 Part agm3nt
12:48:53 Quit FOAD (Read error: 110 (Connection timed out))
12:48:54 Nick FOAD_ is now known as FOAD (
12:48:58aliaskanyone familiar with button_get versus button_read_device?
12:49:04 Join peppo [0] (
12:49:14peppothere's a converter from avi container to rockbox video format?
12:51:40peppowhat's a reasonable bitrate?
12:51:44jhMikeSn1s: setting it identical should work
12:51:47peppo(for a dvd rip source)
12:52:08aliaskpeppo: What mp3 player have you got?
12:52:16peppoipod video 80gb 5.5g
12:52:25aliaskGive 600kbit/s a shot
12:52:35peppo:) that's what I had entered
12:52:53n1sjhMikeS: I can show you a diff of my changes if you'd like
12:52:57peppooh mama this is slow :)
12:53:00 Join Damme [0] (
12:53:22peppo3fps/s :/
12:53:47peppoI'll try vlc instead
12:53:54GodEaterimagine how much slower it'd be on the ipod itself....
12:54:11*aliask was thinking the same thing
12:54:29jhMikeSn1s: sure
12:54:48ddaltonhow do I create a blank file in emacs?
12:55:11Dammehi guys, anyone got any tip for my ipod? I tried to upgrade it's firmware and after 6h I got boored waiting any longer, I disconnected the USB and it rebooted, and then it stuck like 3pixels into the upgrade screen (Black with logo and a bar) and after that I resettet it (nothing happend for 1h) and now it's compleatly dead, nothing happens, no loadingsymbol no applelogo no backlight, windows does not recognize it.. any
12:55:31 Quit elinenbe (Read error: 104 (Connection reset by peer))
12:55:43Dammestupid itunes wanted to upgrade firmware and my sister sat by computer and pressed ok (I've used rockbox on it..)
12:55:51markunpeppo: with a ipod video you are much better off watching videos in apple's firmware
12:55:59jhMikeSamiconn: mov.l sub_routine, Rn?
12:56:11GodEaterDamme: resetting it in the middle of an apple firmware upgrade is the only known way to brick an ipod =/
12:56:12peppomarkun, well I'd still have to convert my sources which are in xvid
12:56:27markunpeppo: sure, but playback will be much smoother
12:56:31DammeGodEater it was already dead.. =( Nothing happend for 1h
12:56:35peppookay, then I'm giving up :)
12:56:40peppobetter get a laptop instead :)
12:56:45GodEaterDamme: then I feat it is not fixable by you
12:56:48markunpeppo: did you try one of the sameple movies?
12:57:07peppomarkun, no.
12:57:24DammeGodEater itn't it possible to connect some interface straight to the motherboard or something? (I havn't checked if it has any special connectors though..)
12:57:51GodEaterDamme: 1) Maybe 2) Do you own such an interface? 3) Do you know what to do with it once connected ?
12:58:12peppomarkun, cool, thanks, I'll try
12:58:48DammeGodEater no no and no but I am an electronic freak and I use to make alot of special interfaces now and then and I was hoping someone else done this before :)
12:58:53 Join gpd [0] (
12:58:56jhMikeSn1s: if that were to work, then that's how to do it
12:59:08GodEaterDamme: well if you can get LinusN's attention he *may* have more of an idea on that.
12:59:09amiconnjhMikeS: Not that simple. SH can only have small literal constants embedded in the instruction. Larger values need to be loaded from memory, like on arm
12:59:30DammeGodEater thanks, I'll ask him later on
12:59:42jhMikeSamiconn: that's what I'm not clear about how re: the assembler syntax for that
13:00:27amiconnThis is because SH has fixed instruction size, like arm (only that it has 16bit instead of 32bit, so more like thumb)
13:01:46amiconnI think the = syntax should work in gcc embedded asm
13:02:12jhMikeSon ARM, better not forget the .ltorg directive on naked functions or the literal pool will quickly be too far away. :P
13:02:37ddaltonso it isn't possible to create a new file in emacs?
13:06:18B4gderddalton: yes it is
13:06:35B4gderjust "open" a file that doesn't exist
13:06:42B4gderedit text, save
13:07:10 Join tedr0ck [0] (
13:11:41 Quit tedrock (Read error: 110 (Connection timed out))
13:14:11jhMikeSamiconn: is "jsr @Rm" on SH like "jsr (%a2)" on coldfire? i.e. not a memory lookup.
13:14:37gpdI just started using rockbox on my 30G iPod Video 5th - so far it has crashed 3 times... is this to be expected?
13:14:57pondlifegpd: I'd think not... what sort of crash?
13:15:09gpdjust seems to hang - last time was with the fire demo
13:15:15amiconnjhMikeS: Yes, no memory lookup involved
13:15:16gpdprobably a buggy plugin?
13:15:38pondlifeMaybe, but it should be fine.
13:15:40amiconnIn fact you can also write jsr %a2@ on cf (that's gnu asm syntax)
13:16:09 Quit Seed (Nick collision from services.)
13:16:15amiconnYou'll see that syntax when disassembling with m68k-elf-objdump
13:16:17 Join Seed [0] (
13:16:39ddaltonhow do I get emacs to do the indentation?
13:17:05jhMikeSodd that I have to write "_remove_thread" and not "remove_thread" like on the others.
13:18:55Dammebah, why didn't apple put in a failsafe-boot-mode ... my ipod seams bricked for real =(
13:22:48preglowDamme: it's not...
13:23:06preglowthat's a rare thing to see
13:23:15DammeIt failed during firmwareupdate..
13:23:25preglowoh, then it just might be, heh
13:24:02Dammein worst case I'll desolder the memorychip and reprogram it or something :P
13:25:19*B4gder demands a commit!
13:26:45*petur commits some code in the repository at work :p
13:26:55peturdid that help?
13:26:58B4gdergood work!
13:27:56gpdok - crashed again - this time "Searching 5 found(play/pause to abort) - play/pause does nothing :(
13:30:18pixelmagpd: did you install the Ipod Video 60/80GB build by any cnahce?
13:30:58gpdno - 30GB Video
13:31:44gpdcrashed again - this time playing a track...
13:31:47*gpd gives up
13:31:57jbitpetur: so you're mentoring for usb otg? how far is it?
13:32:28 Join thegeek [0] (
13:32:33peturjbit: I have no idea... austriancoder is gone again :(
13:32:47jbitwell somebody mentioned you're stuck
13:32:47DerPapstDamme: you surely can't desoilder that chip. it's glues on the PCB
13:32:54jbitor he's stuck
13:32:55jbitrather :)
13:33:02DammeDerPapst too bad :)
13:33:07*DerPapst is the someone
13:33:07peturjbit: yes, transmitting seems to fail
13:33:49jbitpetur: i wrote an i.Mx USB device controller driver a while ago, I heard the pp chips usb stuff is similar?
13:34:27 Quit Damme (Read error: 104 (Connection reset by peer))
13:34:36ddaltonso it isn't possible to get emacs to do the indentation automatically?
13:34:38B4gderjbit: yes, very similar
13:34:51 Part pixelma
13:34:55B4gderddalton: it is certainly possible
13:35:18jbitpetur: was thinking it might be fun to have a look at it :)
13:35:19petursimilar yes, but not exactly the same. Yesterday I found 3 bits the i.MX datasheet said to be fixed at 0 or 1, with the register dump showing something else (have that stuff at home, not here)
13:35:27peturjbit: please do
13:35:40preglowthe recording side of dsp doesn't ever convert to 32 bit samples, no?
13:35:55jhMikeSfor e200: Binary size: 481280, Actual Size: 480168?
13:36:01peturjbit: this is the last testversion I think:
13:36:02jbitpetur: is there a patch or a svn/cvs branch or something I can check out?
13:36:05jbitah okay :)
13:36:10jhMikeSpreglow: yes indeedy for wavpack
13:36:14 Quit GodEater (Remote closed the connection)
13:36:47preglowjhMikeS: just thinking about maybe adding the riaa deemphasis filter as a recording option... and then i'd like to reuse the eq functions
13:37:21jhMikeSeq on the recording side?
13:37:31preglowbut that won't work unless the samples are 32 bit non-interlaved
13:37:36jbitpetur: so receiving data works then?
13:37:45jhMikeSoh, just for the deemph
13:37:47preglowjhMikeS: not eq as such, i'd just want to use the eq_filter function for filtering
13:37:49jbitin device mode or host? :)
13:37:54*jbit is guessing device
13:37:59 Join mahound [0] (n=mahound@
13:38:14preglowjhMikeS: but i guess that's perhaps one of those "just do it on a computer" things
13:38:22mahoundi ahev rockbox on my ipod
13:38:36preglowmahound: congratulations
13:38:37peturjbit: the setup frame is received, but when sending the response, the controller doesn't signal it is done, it just pretend it was never asked to send
13:38:37mahoundwhen i plug it, gnome-volume-manager doesn't recognize it
13:38:44mahound"no sensible filesystem"
13:38:48mahoundany clue?
13:39:02jhMikeSthe only format where filtering in advance really makes sense is mp3 or _perhaps_ wavpack. wav and aiff, I'm not so sure.
13:39:19 Join nerochiaro [0] (
13:39:25jbitpetur: weird, waht have you guys been testing on? ipod or...?
13:39:36*jbit only has a 30gb 5g ipod here ;-/
13:39:42peturbut the code also runs on ipod
13:39:43preglowjhMikeS: it might make sense if you're recording straight to lossy, but does anyone serious really do that anyway? :>
13:40:00peturjbit: feel free to test and fix
13:40:02jbitand last(ish) question: does the patch stop the ipod rebooting into disk mode? :P
13:40:04jhMikeSI think for voice, lectures, meetings and such, yes
13:40:16preglowyou don't need riaa deemphasis there, heh
13:40:25preglowso i'll just forget about this and stick to using it as a plugin
13:40:27DerPapstjbit: no
13:40:37DerPapstthat is independent from rockbox
13:40:40jhMikeSnot really. I think some actually do radio with mp3
13:40:50jbitDerPapst: i mean
13:41:01jbitat the moment, when rockbox is running and you plug a usb cable in
13:41:07jbitit'll reboot into diskmode :)
13:41:17DerPapsthold down menu ~20seconds
13:41:22mahoundok this started happening at the same time as i turn my ipod on and rockbox says "Database is not ready. Initialize now?"
13:41:27preglowamiconn: is there any reason why vinyl (mm cart) straight to h120 line in shouldn't work? i still have treble problems
13:41:31DerPapstwhile plugging usb in
13:42:06jhMikeSThough I don't know why you'd do WavPack if you want to compress again later...maybe disk space. That's so CPU intense trying anything serious besides mono mixdown is certainly too much.
13:42:30mahoundany clue?
13:43:28 Join Damme [0] (
13:43:37Dammestupid internet
13:43:52jhMikeSpreglow: There is one thing that may make real sense - a DC blocker but that's really cheap to do. Some of these DAPS seem to have quite an offest.
13:44:24Dammehehe I had one day of warrenty left :) the only thing I need to do now is to restore my hdd in my ipod so apple cant see I used rockbox :)
13:44:27preglowthat's dirt cheap, simple one pole highpasses are pretty common
13:45:14jhMikeSI used a two-pole in the blit synth, even that hardly phased it (no pun :)
13:45:18preglowmahound: if rockbox itself can read the disk fine, then there's no reason gnome shouldn't be able to read it
13:45:20jbitpetur: have you guys tried reversing the DM software or similar to see if it's doing something differently?
13:45:31DerPapstDamme: i suppose you have a zif to ide connector?
13:45:44DammeDerPapst no but I'll make one somehow
13:45:59Dammeor buy :)
13:46:04preglowjhMikeS: that too should work fine, yeah, it just needs to have no overshoot to make our lives simpler
13:46:10peturjbit: all is based on the work of MrH who made it work in manufacturing mode
13:46:14 Quit ender` (" On the contrary, if you never procreate, neither will your kids.")
13:46:19DerPapsti think there is a pdf that has some good explanations... but it's in german
13:46:38DammeI think I'll understand the technical terms
13:46:41jbitpetur: hrm, okay
13:46:50 Quit ddalton ("I was using BOFHNet IRC version 1.2 by fmillion - get your copy today from !")
13:47:02peturjbit: there are register dumps of the controller (in manufacturing mode) in the wiki
13:47:57DammeDerPapst thanks!
13:47:58DerPapstthe zif to ide adapter (pdf) thing
13:49:31jbitpetur: thanks :)
13:49:34 Join Arathis [0] (
13:51:18 Join austriancoder [0] (n=austrian@rockbox/developer/austriancoder)
13:51:46austriancoderpetur: - lastest patch with some register dumps...
13:51:48*jhMikeS has noob SH asm waiting for testing and an uncharged player ...
13:51:59peturnow look who's there
13:52:00jbitaustriancoder: good timing.. i think :)
13:52:10austriancoderjhMikeS: do you have time to do some usb stuff?
13:52:33peturaustriancoder: 4 days left untill end of GSOC, where have you been?
13:53:06jhMikeSwhilst charging? I suppose. Next thing I'd like to see is restoring the code behavior (longjmp)...or was that tried?
13:53:44 Join ender` [0] (
13:53:55B4gderhey ender`
13:54:02austriancoderpetur: I cleaned up my code... as you maybe know I dont hang in irq everytime I am in front of my computer
13:54:16B4gderender`: deepthought has a disk space shortage
13:54:30austriancoderjhMikeS: lastes patch can be found here:
13:54:43*jhMikeS hangs in irC even when nowhere near the computer
13:54:55petursame here
13:55:21peturaustriancoder: no need to dump the url several times
13:55:51JdGordon_austriancoder: cool, did you have any luck with the reg dumps?
13:56:03 Join spiorf_ [0] (
13:56:05JdGordon_B4gder: can you allow #rockbox-community in the webclient?
13:56:23*aliask would like this too
13:56:30peturaustriancoder: manufacturing mode reg dumps are in the wiki
13:56:30ender`B4gder: i think that was caused by RAM shortage caused by a runaway program yesterday
13:56:38austriancoderpetur: I know
13:56:47ender`disk space is one thing my server won't run out of :)
13:56:50B4gderender`: ah, ok then I'll enable builds on it again
13:57:33 Join webguest62 [0] (i=c023110a@gateway/web/cgi-irc/
13:57:33ender`/dev/evms/rockbox 5,3G 1,3G 4,1G 24% /home/rbclient
13:57:51JdGordon_B4gder: oh, i got a new beast so once i get the cross compiler going you can add it... but my upload still sux, so only the sims are usefull really
13:58:07*ender` 's still waiting for fiber
14:00:35***Saving seen data "./dancer.seen"
14:01:24*jhMikeS has to make a source tree clear to check any usb stuff first
14:03:09 Quit spiorf_ ("Read error: 110 (Connection timed out)")
14:03:15B4gderJdGordon_: we have a pretty solid range of servers right now so there's really no pressure for more
14:03:27 Join spiorf_ [0] (
14:03:46*jhMikeS wonders why he's getting previously applied patch errors
14:04:10 Join Thundercloud [0] (
14:05:03jhMikeSaustriancoder: I'm doing this from a clean tree and the 16th patch. what's up?
14:05:26 Quit spiorf (Read error: 110 (Connection timed out))
14:05:46*austriancoder makes a clean svn checkout
14:06:45preglowgym time
14:06:54*JdGordon_ wants to put his new quad core to good use :p
14:10:10B4gderwe'll make it chew simulator code all day long
14:10:13JdGordon_what happend to the build page??
14:10:31DerPapstJdGordon_: it's dead
14:10:35B4gderit transformed
14:10:40JdGordon_looks good
14:10:41DerPapstbecause they thought it's not needed anymore
14:10:48B4gderwe're done now
14:10:50JdGordon_the voice download link shuold go there also
14:10:53B4gderyou can go home
14:11:02austriancoderjhMikeS: I have no idea :(
14:11:06B4gderJdGordon_: good idea
14:11:13 Nick JdGordon_ is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
14:11:22JdGordonmanual also maybe? just for completness
14:13:40austriancoderjhMikeS: will generate a new patch... one moment
14:15:55 Quit webguest62 ("CGI:IRC")
14:18:15jhMikeSaustriancoder: the lines removed by the patch from arcotg_udc.c don't match what svn has
14:19:32 Quit Febs (Read error: 110 (Connection timed out))
14:21:03*jhMikeS supposes he can just remove the file reference from the patch
14:22:22jhMikeSpatched fine
14:27:08 Join chrisjs169__ [0] (
14:30:10austriancoderfixed patch:
14:30:58JdGordon6 min to make all 3 cross compilers :D 3 of them was downloadign gcc :p
14:31:24 Quit mahound (Remote closed the connection)
14:31:30 Join krazykit [0] (
14:31:55B4gderthat's a mighty speed
14:32:15peturmy dual-core laptop needed 20 minutes....
14:32:32*DerPapst jealous
14:34:29daurnimatorJdGordon: how is the new pc?
14:34:35daurnimatoroh, #rockbox-community :P
14:34:48JdGordonnice n fast :)
14:35:35jhMikeSaustriancoder: I only need logf, or DEBUG too?
14:35:43jhMikeSerrors in core.c
14:35:59austriancoderjhMikeS: only logf
14:37:32jhMikeSusb_ch9.h is missing too
14:37:43austriancoderjhMikeS: one moment
14:42:54 Quit chrisjs169_ (Read error: 110 (Connection timed out))
14:43:01 Join Agm3nt [0] (
14:43:14*aliask is having some strange issues with queues and doesn't know how to debug
14:44:45jhMikeSaustriancoder: just looking at the patch I spotted some errors with the timeouts. besides, they really should be changed back to udelays like MrH had
14:45:39jhMikeSUSEC_TIMER usage even
14:47:05aliaskIt's a hardware timer on the portalplayer chips which increments once every microsecond.
14:48:05bluebrotherB4gder: how about adding a "don't forget to fetch the manual" to the current build page? ;-)
14:48:23B4gderyeah I'll add pointers to manuals and voices
14:48:27amiconnpreglow: I have no idea. If the input impedance is not too low, it should work
14:48:36austriancoderjhMikeS: If you give me an example how to use it, I can change to USEC_TIMER
14:48:37 Join Febs [0] (n=chatzill@
14:48:40bluebrotherdidn't think of voices, yeah :)
14:48:46austriancoderaliask: thanks for the info
14:49:01bluebrotherbut the new page is nice. Much cleaner for users
14:49:02amiconnDid you try recording without deemphasis, and then applying it on a pc with a known-good program?
14:49:49jhMikeSaustriancoder: it's right in MrH's code. the names are a bit different but you can just put the timer_expired stuff back in.
14:50:08austriancoderjhMikeS: okay..
14:50:33amiconnpreglow: Regarding eq - the arm eq (and treble/bass) implementation seems to be kinda inefficient. Just treble/bass makes ogg playback skip (sometimes) on PP5002, even though with linear output, there's quite some headroom left...
14:50:53 Quit BigMac ("Leaving")
14:51:08amiconnPerhaps it's just on PP5002 because of the broken cache, and some function should be in iram?
14:51:59jhMikeSe200 doesn't have trouble with tone controls at all
14:53:02jhMikeSaustriancoder: this one's compiling ok :)
14:53:48*n1s spots an additional slew of voice bugs :-(
14:58:20 Join smail [0] (i=52e0d0a5@gateway/web/cgi-irc/
15:03:18jhMikeSaustriancoder: if this works I take it I should be seeing a clean connect and a recognized device?
15:05:26*aliask wants jhMikeS to say "Well, it looks like it works"
15:05:36austriancoderjhMikeS: You should see in logf a line like "result from driver: 0"
15:06:50jhMikeSgetting - values
15:08:19smailhello i'm trying to compile rockbox for elio but when executing the makefile i got an error with debug_menu.c
15:08:20 Quit Damme (Read error: 110 (Connection timed out))
15:08:31austriancoderjhMikeS: -116 == -ETIMEDOUT | -95 == -EOPNOTSUPP (unhandled usb reqeust)
15:08:44B4gdersmail: you mean the full rockbox, not just the bootloader?
15:09:03B4gderas I doubt it builds for elio
15:09:15smailwhat is the difference ?
15:09:24B4gderthe bootloader is tiny and is made to load rockbox
15:09:29jhMikeSbasically -116
15:09:38smaili see
15:10:17B4gdersmail: elio is still an early port for rockbox
15:10:25smaili now
15:10:25B4gderand nobody works very much on that front
15:10:36jhMikeSaustriancoder: all I see is -116 for a result. no -95
15:10:56smaili wanted to see if i could help
15:10:59austriancoderjhMikeS: I should have the new timing code ready in some minutes
15:11:02B4gdersmail: but a first step is to get the bootloader to load and display something for example
15:11:22smaildirectly on the elio
15:11:22 Join linuxstb_ [0] (i=d556da1b@rockbox/developer/linuxstb)
15:11:37B4gderyes, there is no other way
15:11:53linuxstb_smail: Does the Elio bootloader still compile and run?
15:12:18smailok let me explain what i've done till now
15:12:28linuxstb_I don't think Rockbox ever compiled for the Elio - the port never reached that stage. The next step is getting the LCD working...
15:13:01smaili've managed to make the to work
15:13:05smailso i have
15:13:19smailthe compiler
15:13:20jhMikeSaliask: btw, I'll sure make that known!
15:13:43*aliask is really looking forward to USB
15:13:45smailthen i downloaded the svn
15:14:48smailbut when runing the makefile
15:15:01smaili got some errors
15:15:26smailis there anything i could start with ?
15:16:22B4gdersmail: did you run configure and select to build a bootloader?
15:17:20aliaskAnyone familiar with the queue code, I'm getting stack overflows after initialising the button queue, but I don't know how to trace the cause. Any ideas?
15:17:54smaili've been trying to compile all the rockbox
15:18:03B4gdersmail: don't
15:18:09jhMikeSaliask: since when does that ever happen?
15:18:12B4gderor if you do, you're up to some editing
15:18:29B4gdersmail: aren't you reading what we type?
15:18:37smailit's ok
15:18:43B4gderboth me and linuxstb already said this
15:18:59austriancoderjhMikeS: is usleep (mrh) == udelay (rb) ?
15:19:00linuxstb_smail: It doesn't work. All that works is the bootloader, which is a stripped-down version of Rockbox (it uses the same drivers). So my approach to the port was to write enough drivers for the bootloader to work, and then move on to Rockbox. But the first step wasn't finished - I got stuck with being unable to get the LCD working.
15:19:08aliaskjhMikeS: Well, since the first time I ever tried doing a button_init() - gigabeat S stuff
15:19:27jhMikeSaustriancoder: I believe so...I'll peek again at that
15:20:30linuxstb_smail: If you look at bootloader/elio.c (I think, maybe it's tpj1022.c) then you'll see what little it does, and what debugging info it's creating. It can write to the disk and detect some button presses, but that's about all...
15:20:39jhMikeSaliask: queues really don't use much stack at all
15:21:44smaili know it might sound crazy but where do i start
15:22:08jhMikeSone sure stack overflow source is debugging with splash screens. they take 512 bytes minimum for their buffer.
15:22:39linuxstb_smail: With building the bootloader and installing it... Then you need to a) Become familiar with the Rockbox source code (i.e. do a lot of reading, but feel free to ask specific questions here); b) Start reverse-engineering the original firmware (i.e. read the disassembly...)
15:23:39aliaskjhMikeS: I've been using debugf for my debugging, and commenting out queue_init(&button_queue, true); stops the stkov
15:24:12smailany links to start with ?
15:24:37linuxstb_For what?
15:25:24smaildisassembly ?
15:26:38jhMikeSaliask: that's called on the main thread, right? stkov shouldn't be a factor. queue_init really takes no stack to speak of.
15:27:24linuxstb_smail: This page (by Bagder) has some useful info about MI4 files (the format used for the Elio firmware) -
15:27:51jhMikeSaustriancoder: usleep/udelay are basically equivalent it seems.
15:28:15linuxstb_smail: The "arm-elf-objdump" command listed on that page will give you a very basic disassembly. IDA Pro is also a useful tool.
15:28:17 Join BigBambi_ [0] (n=Alex@
15:28:40LinusNaliask: maybe the button event array is too small, and button_init() overwrites the stack, which happens to be after the button event array in memory?
15:28:41aliaskjhMikeS: Yep, called from main thread iiuc. Any ideas as to how to track down the source of the stkov?
15:28:51B4gdersmail: and if you didn't do disassembly on ARM before, you probably want to read up on the ARM instruction set etc
15:29:11aliaskLinusN: Sounds promising
15:31:40austriancoderjhMikeS: - new error code -5 == -EHWCRITICAL (sending failed)
15:31:51*jhMikeS wonders if gcc has any nice feature to zero-init every stack allocation to catch otherwise undetectable stkov (large buffer mostly unused kind)
15:31:59 Part Agm3nt
15:32:55B4gderthere is a -fstack-check ...
15:32:56jhMikeSaustriancoder: does that count as an improvement in any way like moving to another level of difficulty?
15:33:45 Join agm3nt [0] (
15:33:45austriancoderjhMikeS: makes it easier - i hope
15:34:04austriancoderJdGordon: do you have some time for some a settings menu question?
15:35:59 Join bluebrot1er [0] (
15:36:58 Quit bluebrother (Read error: 110 (Connection timed out))
15:37:13austriancoderJdGordon: I have an linkes list of struct and in the struct is a member called name. Now I want an option to select one of the items from the linked list by choosing an item descriped by the const char* member name.
15:38:29JdGordonis this for debugging?
15:38:35jhMikeSaustriancoder: yup, -5
15:38:54JdGordonim not really sure what you want, but whipping up a custom list handler sounds like what you want
15:39:05austriancoderJdGordon: nope.. i have an linked list of all aviable usb device drivers. and via menu i want to choose one
15:39:21aliaskLinusN: Where's the button event array declared?
15:39:49linuxstb_austriancoder: I think the new shortcuts plugin may do something very similar.
15:40:15austriancoderJdGordon: with the function usb_device_driver_bind("serial"); I can tell the usb stack to use the serial driver.
15:40:28austriancoderlinuxstb_: ahh.. will have a look at it, thanks
15:41:27LinusNaliask: in the queue code
15:41:42LinusNthe button event array is in fact a queue
15:41:53JdGordonaustriancoder: the settings code wasnt really designed for that...
15:42:22 Nick bluebrot1er is now known as bluebrother (
15:42:57*jhMikeS wonders what file implements the ->request call
15:43:07aliaskLinusN: Aren't all queues equal in size? How can one be too small?
15:43:08austriancoderJdGordon: I see..
15:43:15LinusNaliask: beats me :-)
15:43:56austriancoderjhMikeS: firmware/usbstack/driver/usb_serial.c
15:44:17austriancoderJdGordon: what options do i have?
15:45:12*amiconn sends his daily reminder for elio and av300 port to linuxstb
15:45:32austriancoderJdGordon: 1) hardcoded 2) can i give an array of char* to the settings code?
15:45:43LinusNaliask: care to share the code with me?
15:46:04aliaskLinusN: I'm about to commit what I have to the gigabeat-s branch
15:46:36 Quit BigBambi (Read error: 110 (Connection timed out))
15:48:21 Join carlos1234 [0] (
15:49:04JdGordonaustriancoder: sorry, im a bit too tired to be usefull tonight
15:49:17amiconnaliask: Queues cannot be too small. All standard queues are 16 entries deep iirc. If more events queue up, the first ones will just get overwritten
15:49:42amiconnThe only exception is the talk module, which has a special queue that is 64 entries deep
15:49:47austriancoderJdGordon: no problem.. maybe i will find a way, and if not i will ask you again ;)
15:50:59carlos1234miepchen^schlaf: is on this channel only engish the languege?
15:51:33peturcarlos1234: english only, yes
15:52:41carlos1234maybe thisone is not the rightone for me ;-)
15:53:33 Join BigMac [0] (
15:55:37carlos1234petur: from where do you come?
15:55:57peturcarlos1234: .be
15:56:35austriancoderJdGordon: I think we could extend settings api as FILENAME_SETTING shows a list of files in a defined dir - or? If so, then I only need to have a ARRAY_SETTING where I can choose an element from a defined array.
15:56:41aliaskLinusN: It's committed, go nuts :)
15:57:27JdGordonaustriancoder: if you can hard code it it will be the easiest...
15:57:38markuncarlos1234: maybe
15:58:12carlos1234petur: I am from blackforrest
15:58:22austriancoderJdGordon: i could, but a "dynamic" solution would be nicer...
15:58:47JdGordonyou want to save it as a string?
15:58:57JdGordon... in the settings struct
15:59:14*DerPapst doesn't know if it is really necessary to translate "schwarzwald"
15:59:26B4gderah, finally a commit
15:59:33 Join webguest95 [0] (i=c023110a@gateway/web/cgi-irc/
15:59:35austriancoderJdGordon: yup..
15:59:37jhMikeSaustriancoder: this still isn't setup to use the GPIO USB interrupt?
16:00:11austriancoderjhMikeS: nope.. but feel free to add it :)
16:00:37***Saving seen data "./dancer.seen"
16:01:07JdGordonaustriancoder: wouldnt you rather save it as an int?
16:01:17JdGordonand then have a conversion table?
16:01:37austriancoderjhMikeS: I had an problem that irq_src was often 0x0 with irq stuff
16:02:03LinusNaliask: i can't see anything obvious
16:02:14B4gderhah, the complete failure of mine isn't too visible...
16:02:36DerPapstthe zeros?
16:03:10aliaskLinusN: Thanks for looking
16:03:20B4gderadjusted for next commit
16:03:29jhMikeSaustriancoder: it looks like it should plug in now. I wonder if things just aren't timely enough.
16:03:57austriancoderJdGordon: no.. as every driver is identified with an const char* and not with an int - see struct usb_device_driver
16:05:04*DerPapst looks at "broken?" below the jiggabeat F and sansa e2XX on
16:05:18JdGordonaustriancoder: can you add an int identifier for the drivers?
16:05:33JdGordonstoring an int means no exxtra hacks to get it woorking
16:05:54 Nick lost|X40 is now known as lostnihilist (
16:06:01 Quit webguest95 ("CGI:IRC")
16:06:12austriancoderJdGordon: okay
16:06:23B4gderDerPapst: still?
16:06:24carlos1234markun: thx
16:06:43JdGordonaustriancoder: so how far have you got with it? what works?
16:06:54DerPapstB4gder: nope
16:06:59DerPapstit's fine now :)
16:07:01B4gderDerPapst: that can happen if you fetch the page exactly when those zip files are being downloaded, over the old name
16:07:06austriancoderjhMikeS: in the datasheet there is nothing about timings
16:07:09carlos1234i´ll try this
16:07:33 Part LinusN
16:07:37DerPapstB4gder: aha. ok
16:07:58jhMikeSaustriancoder: well, I do recall something about 3ms
16:08:50 Join pixelma [0] (i=d5494d2b@rockbox/staff/pixelma)
16:09:24amiconnB4gder: Do you think it's a good idea to split the battery voltage tables and put them into target tree?
16:09:41B4gderamiconn: yes, that sounds clever to me
16:09:46amiconnAnd if so, should I call the files battery-<target>.c or powermgmt-<target>.c?
16:10:13B4gderdoesn't matter to me
16:11:05 Part gpd
16:11:14austriancoderJdGordon: how should a convertig table look like? any examlpe?
16:11:33amiconnI'll put the adc->voltage conversion there too, since the simple scale factor method doesn't work for the 1st/2nd gen build
16:11:54amiconn(factor is different for 1st and 2nd gen)
16:12:50JdGordonaustriancoder: hmm... actually thats no good... I think for the mean time you shuold do your own handling for it and use the FILENAE_SETTING item
16:12:51amiconnSomeone is eating up my green deltas :|
16:13:29jhMikeSaustiancoder: spinlock_lock is not an irq safe function. functions that serialize with the interrupts should use int oldlevel = set_irq_level(HIGHEST_IRQ_LEVEL);/set_irq_level(oldlevel);
16:13:52austriancoderJdGordon: own handling... hmmm.. where to start
16:14:12JdGordonaustriancoder: im not really sure how to nicely display the list names... how important is this bit?
16:14:38austriancoderjhMikeS: okay... me learned something new
16:15:36austriancoderJdGordon: In the moment there is only one "driver" - a test driver, but in the next days there will be one or two device drivers
16:16:42 Join JavaMan22 [0] (
16:16:46jhMikeSaustriancoder: interrupts don't need to call those of course since it's already in interrupt mode. just regular code to keep interrupts from happening when it can't have them.
16:16:53JdGordoncan you use the filename macro untill sunday? ill whip up something for you then
16:17:11austriancoderJdGordon: jup.. sounds fine
16:17:13JdGordon.... just remind me or ill forget (send me an email on saturday)
16:17:58austriancoderjhMikeS: so i need to do the set_irq_level in usb_arcotg_dcd_irq and it should be okay?
16:18:03JdGordonbed time, ttyl
16:18:07 Quit JdGordon ("Konversation terminated!")
16:18:56 Quit carlos1234 ("Konversation terminated!")
16:20:45 Quit aliask ("ChatZilla [Firefox]")
16:20:57 Join kubiix [0] (
16:21:22jhMikeSaustriancoder: you shouldn't need to there but in normal code that shares data with the interrupt where having an interrupt would corrupt the data if it accessed it at the same time
16:21:52 Quit smail ("CGI:IRC (Ping timeout)")
16:23:10LloreanAve: I've tried Vorbis, MP3, Flac, and SPC without experience the bug.
16:24:28 Join Jon-Kha_ [0] (
16:26:23 Quit idnar (Nick collision from services.)
16:26:26 Join idnar_ [0] (n=mithrand@unaffiliated/idnar)
16:27:21 Nick idnar_ is now known as idnar (n=mithrand@unaffiliated/idnar)
16:30:27 Part agm3nt
16:31:16 Quit atsea- (Remote closed the connection)
16:32:05austriancoderjhMikeS: Here is my try with interrups: - But i get lot of BAD IRQ logf entries
16:35:11 Join atsea- [0] (i=ariel@gateway/tor/x-ca03b8ea93120347)
16:37:22 Quit XavierGr ()
16:42:20*Llorean just noticed B4gder / Bagder's ohloh tag says 10000 commits.
16:42:25 Join XavierGr [0] (
16:42:49B4gderI need to stop committing now with such a fine number there
16:42:54DerPapst1/3 of them are "Oops"-commits :P
16:43:34LloreanOr get another 15k in or so, for an even prettier number. :-P
16:44:10jhMikeSaustriancoder: The GPIOB_INT_LEV must be flipped on every interrupt
16:45:17austriancoderjhMikeS: flipped... and how do i this?
16:46:30jhMikeSyou were returning before executing the GPIOB_INT_LEVEL = ... code
16:51:52austriancoderjhMikeS: now i only get "BAD IRQ"
16:52:08jhMikeSnot me
16:52:33jhMikeSbut something seemed like it was acting better before
16:53:08 Quit Llorean ("Leaving.")
16:53:33austriancoderjhMikeS: can you show me a diff?
16:53:42austriancodersvn diff
16:55:11 Quit pixelma ("bbl")
16:56:59jhMikeSsure, one sec. weren't the interrupts coming correctly right after the initial use of them?>
16:59:26austriancoderjhMikeS: I have no idea...
17:04:02jhMikeS <= some was just tweaking but enabling all USB irqs
17:11:24 Quit kubiix (Read error: 110 (Connection timed out))
17:23:20austriancoderjhMikeS: do you have figured out something about the tx problem?
17:23:28JavaMan22Add theme installation. This relies on a modified server script so the themes site need to incorporate the changes.
17:23:42JavaMan22does this mean themes will work now
17:24:01 Join ie [0] (i=c27f0814@gateway/web/cgi-irc/
17:24:27Nico_PjhMikeS: are you interested in my beginning-of-a-merge of MoB into playback.c ?
17:24:46Nico_Pif you are, I can push a few changes to my plublic git repo
17:24:55ieWill RB work on Sansa e 250? It's not listet on the RB front page but...
17:25:23rasherie: That's a "E200" series
17:26:06ierasher: ah, I missed the word "series"! So I assume the answer is "yes", no?
17:26:15rasherie: Yes
17:26:43rasherI don't believe an E200 even exists. They're all E2x0 afaik.
17:28:42ierasher: ok. Do all of them have the FM radio?
17:29:01jhMikeSaustriancoder: not yet. it's kind of early for me on this one.
17:29:17rasherie: I don't know. You can check the DeviceChart wiki-page for details. (or Sandisk's own pages of course)
17:29:27jhMikeSNico_P: is it actually running now?
17:29:38ieOn one page describing E250, FM wasn't mentioned. This doesn't mean though that it doesn't exist
17:29:39 Join Siku [0] (
17:29:44Nico_PjhMikeS: hell no ! I started changin playback.c today
17:30:04jhMikeSI thought there was a tester plugin (never tried it)
17:30:28Nico_Pthe testing plugin works fine, including real MoB
17:30:48Nico_Pbut now I'm merging all this into playback.c
17:32:02ierasher: from the SanDisk site: "No FM-tuner (and no FM-recording) is available in Europe" This should also be mentioned on the wiki page (it isn't atm)
17:32:49 Quit petur ("*plop*")
17:34:03 Quit Zagor ("Client exiting")
17:34:12rasherie: Actually, I'm unsure whether this is a hardware difference or simply disabled in software. A Sansa guy should probably weigh in right about here.
17:34:19*rasher pings Bagder
17:36:00 Join nut [0] (n=nilay_@
17:36:51 Part ie ("Will read logs later")
17:36:56*JavaMan22 pings rasher
17:41:11 Quit linuxstb_ ("CGI:IRC (Ping timeout)")
17:42:05 Part nut
17:44:14rasherJavaMan22: ?
17:45:22 Join ptw419 [0] (
17:45:57*austriancoder pings austriancoder
17:46:17*JavaMan22 wishes rockboy would be faster
17:50:42indrohi all
17:52:32 Join perrikwp_ [0] (n=chatzill@
17:52:34 Nick perrikwp_ is now known as perrikwp (n=chatzill@
17:52:56indrohow can I boot in disc mode?
17:53:47indroany idea?
17:54:27JavaMan22then hold play and center
17:55:33indrothx, it booted automatically in disc mode when I plugged in usb cable.
17:55:52indrobut I played tomorrow with a few settings, now it doesnt boot in disc mode.
17:56:01indroany idea what It worked anymore?
17:56:35indrois that a setting, which?
17:57:53indroeww, the usb cable was not plugged in right..
17:59:42indroI unpacked the puckman zip into /.rockbox/pacman
17:59:52indrobut when I want play pacman it says, no roms found.
17:59:55indroany idea whats wrong?
18:00:39***Saving seen data "./dancer.seen"
18:00:43indroI have from here:
18:00:48indrois that a valid rom?
18:01:04indroJavaMan22: ping. ;)
18:03:50 Join divo [0] (
18:04:28indrohi divo
18:05:26 Join My_Sic [0] (
18:05:56divoYour site states that rockbox works on the SanDisk Sansa 200-series... Doest that include the "E280"?
18:07:11JavaMan22you reset your ipod by holding menu and center
18:07:23JavaMan22when it is about to start up again hold center and play
18:07:46divoI've got a good offer on the 8gb model (25% off) but I'd like to know wether it supports rockbox.
18:08:04JavaMan22you can check the checksum with cygwin :o
18:08:18indroJavaMan22: I got that. what is with the packbox? any idea why is not working?
18:09:01JavaMan22"" and ""
18:09:04JavaMan22you need both
18:09:23indroohh, okay..
18:11:18indroThis rom is not available for download. It is protected by the ESA.
18:11:27indrocan any one point me to a mirror or send me the file?
18:12:47JavaMan22just keep searching on google youll find eventually
18:15:30indrofound this:
18:15:34indrobut still not working. :/
18:16:30 Quit Arathis ("Verlassend")
18:16:46 Join HellDragon [0] (n=Nocebo@unaffiliated/helldragon)
18:17:22indrotrying next one.. ;)
18:18:13indrostill no luck..
18:18:38JavaMan22ok ill upload a copy for you
18:18:52indrowould be nice.
18:19:32 Quit HellDragon (Read error: 104 (Connection reset by peer))
18:20:58DerPapstGood bye
18:21:00 Quit DerPapst ("So Long And Thanks For All The Fish!")
18:21:25 Join HellDragon [0] (n=Nocebo@unaffiliated/helldragon)
18:21:39 Join Tak [0] (
18:21:57indroworks, thanks! :)
18:23:59indroany idea where are the keybindings for rockboy saved?
18:24:10indroI want them reset to default values.
18:25:09austriancoderis it possible to put a thread in suspend mode?
18:26:09 Join elinenbe [0] (
18:28:42 Join petur [0] (n=petur@rockbox/developer/petur)
18:31:49peturindro, JavaMan22: please do not conduct illegal activities in here (sharing of copyrighted material)
18:33:01JavaMan22you just got here... what are you a spy?
18:33:17Nico_Ppondlife: anything against "bufgetid3" in the buffering API ?
18:33:17peturfyi, the ip address of indro plus the illegal link are logged so that wasn't very clever either
18:33:29pondlifeNico_P: Yes, a bit
18:33:29peturJavaMan22: this channel is logged
18:33:42JavaMan22it wont happen again petur
18:33:43Nico_Ppondlife: why ?
18:34:10Nico_Pmaybe not in the plugin API then, but I think this kind of helper will become handy
18:34:18pondlifeNico_P: I'd prefer it if it was entirely generic, so bufgetid3 would sit in a layer above it. A very thin layer, certainly
18:34:54Nico_Ppondlife: it does use bufgetdata... it's just a wrapper
18:35:05pondlifeThat's exactly why it shouldn't be included.
18:35:15JavaMan22lol i feel like a criminal now :(
18:35:28Nico_Ppondlife: makes sense... but where should I put it then ?
18:35:35pondlifeNo idea :)
18:35:57pondlifeBut it's more a metadata function than a buffering one, no?
18:36:11pondlifeKeep each API as small as possible.
18:36:38pondlife"Less doors and less windows makes for a more secure house."
18:36:40Takhow optimistic should I be about the cowon d2 porting effort? ( )
18:36:56Nico_Ppondlife: maybe in metadata.c then ?
18:37:12Nico_PI'll keep it in playback.c for now
18:37:29pondlifeThat's the way of playback.c.... it doesn't give up without a fight.
18:37:36Nico_Ppondlife: interested in my start of a merge ?
18:37:57pondlifeAlways. Not that I have much time for code reading at the moment.
18:38:03pondlifeBut happy to try running stuff
18:38:23Nico_Pit surely does not run... or even compile for that matter
18:39:42 Join Domonoky [0] (
18:41:07austriancoderjhMikeS: if you get tx working, please send me an mail with the working code...
18:43:16 Join merbanan [0] (n=banan@
18:47:39 Join linuxstb_ [0] (i=d556da1b@rockbox/developer/linuxstb)
18:48:12 Join alienbiker99 [0] (
18:55:29 Join mexicankiller [0] (n=chatzill@
18:55:43 Quit Jon-Kha ("leaving")
18:57:14 Quit Jon-Kha_ ("leaving")
18:57:25 Join Jon-Kha [0] (
18:58:34 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:58:57bluebrotherDomonoky: did you find any new bugs? ;-)
18:59:26Domonokybluebrother: no i was working, didnt have time for rbutil :-)
18:59:58bluebrotherok ... I'll be busy the next couple of days, so I hope everything is fine.
19:00:42*Domonoky will also be away over the weekend.. so lets hope everything is ok, or noboy notices it, if it isnt :-)
19:05:26bluebrotherbtw, I built a static binary with a threads-disabled qt. Should remove the dependency on mingwm10.dll and creates a 3MiB smaller binary
19:05:43bluebrother(which gives a 2.5MiB binary once upx'ed)
19:06:17 Quit alienbiker99 (Read error: 104 (Connection reset by peer))
19:06:28Domonokynice.. and it works without problems without thread support ??
19:06:47bluebrotherwell, I haven't found issues in my quick test.
19:07:04bluebrotherI put a binary online:
19:07:55Domonokyone issue, but not a problem: you resize the theme images, and because of this, they dont look really nice anymore..
19:08:02 Join alienbiker99 [0] (
19:08:12bluebrotherI thought about adding a "show original size" button
19:08:23bluebrotherbut I really don't like resizing the windows.
19:09:31Domonokyyes the resizing window isnt nice.. but perhaps only scale down.. ( no upscaleing) and a show Fullsize button for the downsized images..
19:09:53Domonokyshowing the image fullsize in a small dialog
19:10:18alienbiker99whoever changed the dailybuild page, it looks a lot better now that they are all sorted once agai
19:10:42bluebrotherhmm, might be an idea.
19:10:50bluebrotherwell, gotta run now.
19:10:53bluebrothercu later
19:10:59 Quit bluebrother ("running ...")
19:11:20 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
19:13:38 Quit Febs (Read error: 104 (Connection reset by peer))
19:21:00 Quit pondlife ("disconnected has pondlife")
19:33:04 Join Sliv [0] (i=daba0b01@gateway/web/cgi-irc/
19:33:09 Quit linuxstb_ ("CGI:IRC (Ping timeout)")
19:43:36 Part JavaMan22
19:49:02 Nick chrisjs169__ is now known as chrisjs169 (
19:56:22 Join parafin [0] (
19:57:10 Quit Sliv ("CGI:IRC (EOF)")
19:58:06parafini'm getting no luck compiling rockbox on gentoo :( i used gcc 4.0.3 ,binutils 2.17 and 2.16.1
19:58:24parafinwith 2.17 i got error about missing .got section
19:58:33parafinwith 2.16.1:
19:58:38 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
19:58:39parafin/usr/libexec/gcc/arm-unknown-linux-gnu/ld: section .got [0000000000000000 -> 0000000000000003] overlaps section .text [0000000000000000 -> 0000000000060883]
19:58:57parafinwhat is this section?
19:59:54 Quit jhMikeS (Nick collision from services.)
20:00:00 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
20:00:42***Saving seen data "./dancer.seen"
20:00:47Domonokylooks like your compiler is not correctly configured.. it should be gcc-arm-elf not arm-unkown-linux-gnu i think..
20:00:57 Join kubiix [0] (
20:01:26Domonokyyou can use to build the correct crosscompilers and binutils..
20:01:47parafini don't thing it matters, it's just crossdev script makes binaries with this names, i just fixed Makefile to use them
20:02:17desowinparafin: do you used crossdev?
20:02:22ender`crossdev generated arm-elf-gcc for me
20:02:27Domonokybut its not working, so it could matter.. :-)
20:02:52parafinender`, what command exactly did you use?
20:02:54desowincrossdev -s1 −−gcc 4.0.3 −−target arm-elf
20:03:01desowinpabs: that way?
20:03:10ender`hm, it's been a while
20:03:18parafini used −−target arm
20:03:23 Quit Nico_P (Remote closed the connection)
20:03:24 Quit chrisjs169 (Nick collision from services.)
20:03:24parafini'll try with arm-elf
20:04:03ender`i don't remember what i used, but it did create cross-arm-elf directory in /usr/local/portage
20:04:11*Domonoky bets with arm-elf you dont get this strange section .-)
20:04:26parafini hope so :)
20:05:27ender`btw, my arm-elf-gcc is 4.0.4
20:06:30parafini think rockbox can be built even on 4.2.0 - i get no errors prior that one with gcc-4.2.0
20:06:59ender`since i run a build server, i followed the recommendations
20:08:13Domonokygoogle tells .got section is the Global Offset Table, sounds like relocateable loading.. so something not needed for rockbox :-) so thats definitly from the -gnu gcc/binutils version..
20:11:35 Quit mexicankiller ("ChatZilla [Firefox]")
20:11:48peturXavierGr: fire troubles?
20:12:33XavierGrpetur: somewhat, but there is no forest in crete to burn :P
20:13:02peturyes there is....
20:13:32XavierGrwell no "real" forest like the mainland of Greece
20:15:08 Quit perrikwp (Read error: 110 (Connection timed out))
20:15:44parafini have question about "anti-skip buffer size" preference. I don't quiet understand what it does, 'cause if i set it to 5 sec, my ipod don't spin up every 5 sec...
20:16:28peturparafin: it's the minimum length of music to keep in buffer
20:16:47peturrockbox always tries to buffer as much as possible
20:17:01parafinthat's good =)
20:18:02 Quit ptw419 ()
20:20:08Takif I ever visit greece, I will wear this shirt:
20:31:14parafinhooray, now it compiles fine, thanx
20:32:45 Quit cmvjk (Read error: 110 (Connection timed out))
20:35:30parafinbtw, ipod make click sounds without headphones with original firmware, there is some kind of speaker in there?
20:36:16Domonokyparafin: yes, a small tweeter..
20:36:27 Join Arathis [0] (
20:39:44parafinhow can i shut down ipod so it would resume playing when i switch it on? it stops first and than shut downs, so resume doesn't work...
20:40:30Domonokythere is a resume on startup setting.. please have a look at our very good manual :-)
20:41:58parafinyeah, i enabled it, i found the problem
20:42:12parafini was testinf without headphones
20:42:29parafinand i enabled "don't resume when no headphones" =)
20:47:10linuxstbparafin: Or just pressing PLAY when you boot will resume the previous playlist.
20:48:48parafinit's coincidence, that there is exactly 200 pathes in tracker or not? :)
20:50:00peturit's been worse...
20:51:43 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
20:52:38 Join kubiix [0] (
20:53:42parafinbtw, there is bug in mp3_encoder plugin - in menu where you select bitrate, scrolling doesn't woek correctly
20:53:50 Join ctaylorr_ [0] (
20:53:58 Part divo
20:54:56 Quit ctaylorr (Read error: 110 (Connection timed out))
20:55:44peturparafin: if not already in the tracker, feel free to post a bugreport
20:56:10parafinehhh, i'll have to register
20:57:35peturdo give a valid mail address so that you can be contacted in case of questions
20:58:03peturand if you ever plan to write patches, we have a real name policy...
21:25:38 Join just_asking [0] (i=54c33d1c@gateway/web/cgi-irc/
21:26:04just_askinghiya guys, reading the rockbox page, i still do not know if the Archos 404 is supported. Can anyone help me with this?
21:26:18scorchedoes it list that under supported units?
21:26:18 Quit just_asking (Client Quit)
21:26:43scorche14 seconds with that one
21:27:10*scorche goes about finally getting to the themes site
21:28:12 Join iarwain_ [0] (
21:28:38iarwain_hiya, can anyone help me? I can't seem to find if the Archos 404 is supported or not..
21:28:46scorchedoes it list that under supported units?
21:29:15iarwain_i don't think it is
21:29:25iarwain_lost the page where it says
21:29:26scorchethen what gives yout he impression that it might be?
21:29:42 Part peppo ("part")
21:30:43iarwain_sorry for bothering ^^
21:31:37 Join mexicankiller [0] (n=chatzill@
21:31:37 Quit iarwain_ (Read error: 104 (Connection reset by peer))
21:31:47 Join iarwain_ [0] (
21:31:47 Quit iarwain_ (Remote closed the connection)
21:32:44peturoh, that was the same provider as me
21:34:30amiconnA rather large provider in .be I presume?
21:34:48peturone of the two biggest
21:34:52Takthey should upgrade it to ;-)
21:35:27peturthey have cable, the other big one is ADSL
21:36:57*amiconn notices that most ipods don't have their proper battery type defined
21:37:33amiconnOnly video and nano do. The others use BATT_LIPOL1300 "inherited" from the irivers
21:39:18amiconnHmm, and that creates a problem for my powermgmt target tree split
21:39:52amiconnAll ipods starting from 3g can use the same code except the battery tables, because they all use pcf50605
21:40:07amiconnSo my "one file per target" approach needs a bit of modification...
21:43:47mexicankillerHi guys,any advance on the sansa R?
21:44:54scorcheprogress is posted in the appropriate thread on the forum
21:47:15*parafin is going to have a very nicely patched rockbox :)
21:49:50mexicankillerI'm going to check things about the sansa
21:51:13mexicankilleris there a way to reformat and re-partitionate the sansa memory?
21:52:30scorcheare you referring to the actual storage of the device?
21:52:53amiconnBagder: around?
21:53:15 Nick spiorf_ is now known as spiorf (
21:53:45amiconnA question regarding (C) of source files - when I am going to split out the various pattery tables from powermgmt.c - are all the new files (C) the author(s) of powermgmt.c?
21:54:09amiconnThere will be 14 additional files
21:54:57mexicankillerscorche:I mean to reformat it and then after that make it look as it was
21:55:24amiconnWell, it feels a little wrong. Most of those tables didn't exist back when the original author(s) wrote the code
21:55:46Bagderamiconn: if they're wrong, then correct them...
21:56:07amiconnEh? The tables aren't wrong
21:56:14Bagderthe (c) texts
21:56:32amiconnpowermgmt.c is Copyright (C) 2002 by Heikki Hannikainen, Uwe Freese
21:57:41amiconnI want to split out the various voltage tables. Most targets (and hence most tables) didn't even exist back when powermgmt.c was written
21:58:10rasherShouldn't it be relatively simple to see who wrote which tables, and write correct (c) lines in the new files?
21:58:16rashersvn blame powermgmt.c
21:58:27amiconnBut I also don't know which author(s) provided the various tables
21:58:48BagderI don't think you need to make it complicated
21:59:07Bagderthose files that have clear inheritance from the existing file get the original (c)
21:59:10amiconnrasher: If you do that on current svn, all of them will show me, due to my centivolt->millivolt change
21:59:14Bagderif the files are more or less new, add a new (c)
21:59:51rasheramiconn: ah, true enough.. you could do it on an old revision I guess (of course the same could be true of that)
22:00:43amiconnThey are clearly inherited from the existing file, but they rather clearly inherit contents the original authors never wrote
22:00:45***Saving seen data "./dancer.seen"
22:00:52amiconnThat's my problem...
22:01:38Bagderthere's "ram usage" logged now for the delta table...
22:02:07Bagderit'll an interesting dilemma to add that ;-)
22:03:04amiconnAnd they will be much smaller in the beginning (just the voltage tables for that target plus voltage reading (a one-line function for most targets), but I hope to be able to split other target specific stuff from powermgmt.c later
22:03:40 Join agm3nt [0] (
22:03:53amiconn...for example the charging algoritm: It is recorder v1 *only* (and also only very little of the original code left - most was replaced by gvb)
22:04:16amiconn(Gerald Van Baren)
22:04:59amiconnruncurrent() is another candidate
22:05:04 Quit Domonoky (Read error: 104 (Connection reset by peer))
22:06:21 Quit hannesd_ ("Client suicide")
22:06:22 Quit hannesd ("Client suicide")
22:07:57 Join mexicankiller_ [0] (n=chatzill@
22:09:38 Join mexicankiller__ [0] (n=chatzill@
22:10:14 Quit mexicankiller__ (Client Quit)
22:15:27 Quit Arathis ("Verlassend")
22:16:31 Quit Siku ()
22:16:49 Join Shiny_Mew [0] (
22:21:17 Quit Lear ("ChatZilla [Firefox]")
22:22:32 Quit mexicankiller (Read error: 110 (Connection timed out))
22:26:13 Nick Shiny_Mew is now known as Sivart0 (
22:28:24 Quit mexicankiller_ (Read error: 110 (Connection timed out))
22:30:42 Join cmvjk [0] (
22:30:48 Quit Sivart0 ()
22:36:38 Quit elinenbe (Read error: 104 (Connection reset by peer))
22:39:07parafingcc-4.2.0 doesn't work...
22:39:39parafinwell, will use suggested 4.0.3
22:41:19scorchethere are reasons we recommend versions...
22:41:54pixelman1s: you around?
22:42:26n1spixelma: yes
22:44:27pixelmado you know why the VOICE_EXT_UCL string was removed (for the flash files on archos)? Is there now a more generic string, seems I cannot find one...
22:45:57 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
22:46:22n1spixelma: It wasn't used anywhere?
22:47:43 Quit amiconn (Nick collision from services.)
22:47:44 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
22:48:07amiconnBah, cleaning up ifdef hell is revealing bugs
22:48:29pixelman1s: are you sure? I can't imagine it wasn't but cannot tell for sure as I almost always used spelling of filenames...
22:49:00amiconnE.g. on archos fmrecorder/recorderv2, the NiMH charge curve was used to estimate battery percentage during charging. No BATT_LIION2200 charge curve exists yet
22:51:09pixelman1s: but you are right, in the langV2 patch I still have lying around (version 24) - the only occurence of it was in the old english language file
22:51:24pixelmawonder what happened to it
22:52:47n1spixelma: It can of course be added back in if it's wanted, have you looked at what strings are good to exclude from the hwcodec targets?
22:53:35amiconnThe VOICE_EXT clips are used with file number voicing to tell the type
22:54:07amiconnVOICE_EXT_UCL should be in use on hwcodec, for rockbox flash files (.ucl)
22:55:08pixelmayes, sure. Where has it gone - the langV2 patch "only" removed it from the english.lang - nowhere else...
22:56:16amiconnLooks like it was lost from filetypes.c at some point. Should be readded imo
22:56:27*amiconn hugs baregrep
22:56:53pixelman1s: I only had a quick look, nothing obvious. At the moment I'm looking at the german language file, thought that was also important
22:57:14*n1s has correct voiceing of dB settings :-)
22:57:33 Part Tak ("Ex-Chat")
23:00:56pixelman1s: the only thing I noticed is one of the low battery warnings (I always confuse them but for the non Li-* batteries only one is used) that would exclude it for most Archos (except RecorderV2/FM-Rec but I don't even know how it works there)
23:02:06 Join DWSR [0] (
23:03:33DWSRWhat's everyone's favourite media player software for Windows?
23:03:46scorchewhat does that have to do with rockbox?
23:04:06DWSRRockbox is a media player software.
23:04:16DWSRI'm looking for alternatives to iTunes.
23:04:36rasherAgain, what does that have to do with rockbox?
23:05:01scorchewell, obviously we are missing the connection
23:05:20scorcheso do you mind explaining that connection?
23:05:49DWSRYes. I'm thouroughly lazy.
23:05:51 Part DWSR
23:06:38scorchei feel enlightened
23:06:39 Join mpeccorini [0] (
23:07:06 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
23:09:46parafinit's so cool to have opensourced firmware, but it's so tiring to update all those great patches to last snapshot...
23:10:37scorchewell, someone has to...
23:10:48 Quit lee-qid ("aufwiederbyebientotsayonara")
23:11:20parafini must disclose my real name even to post updated patches or only new one?
23:11:32 Quit spiorf (Remote closed the connection)
23:16:20parafinwhy .bmp files isn't opened by default with rockpaint?
23:17:59 Quit desowin ("use linux")
23:20:51peturparafin: if you do more than just sync to current svn you should, as we add your name to the credits when the patch is accepted
23:21:14parafini see
23:22:22parafindict2 patch doesn't work :( data abort at ...
23:22:53amiconnVoice association for .ucl seems to be gone for a really long time.
23:23:39amiconnHmm, and I know why - it got removed with plugin association
23:24:16amiconnThe .ucl extension is associated with a plugin. Hence it does have an icon, but the extension isn't voiced anymore
23:24:32amiconnGone in r4677, Fri May 21 20:08:24 2004 UTC
23:24:43amiconn3 years+
23:25:09amiconnSo removing VOICE_EXT_UCL from english.lang was correct
23:26:37 Join My_Sic [0] (
23:34:49 Join jhulst [0] (
23:37:52amiconnBah, another 2 targets with the charge curve bug
23:40:15 Quit Bagder (Remote closed the connection)
23:41:04 Join barrywardell [0] (n=barrywar@
23:41:48 Part agm3nt
23:42:23 Join robin0800 [0] (
23:44:25 Join Bagder [0] (n=daniel@rockbox/developer/bagder)
23:51:29 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
23:51:34 Quit midkay_ ("Leaving")
23:53:27 Join malsyned [0] (
23:54:32malsynedIs it possible to adjust the scroll rate on the file browser or get it to have acceleration, like the iPod scrolling does?
23:55:13parafinyou mean not on ipod or what?
23:56:05amiconnmarkun (or other gigabeat coder): The S3C2440 ADC is 10bit ?
23:56:45 Quit ompaul ("Leaving")
23:57:04 Quit jhMikeS (Nick collision from services.)
23:57:10 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
23:57:52parafinrockbox behave strange when i connect usb cable - it show that usb connection screen but sometimes doesn't reboot
23:58:14peturI think that's a known issue

Previous day | Next day