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

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

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

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

#rockbox log for 2007-10-02

00:00:06Bagderruinedtwice: ok, I see
00:00:08ruinedtwiceI checked at the time and noted that the CRC "found" matched precisely a CRC value "advertised as expected" earlier in the scroll.
00:00:08chandlercBadger: if you can point me an some instructions of how to post real firmware dumps, i'm more than happy to
00:00:29chandlercI only have a classic, but i'd be willing to do any comparisons useful
00:00:33 Quit ompaul (Client Quit)
00:00:49chandlercsadly, while i do a great deal of programming in linux, i'm unfamiliar with reverse engineering, especially at the hardware level
00:00:52Bagderruinedtwice: the text output is meant to be displayed on error so in the second case the bootloader clearly thought something was wrong
00:01:25chandlerc(willing to learn if it helps, just pointing out)
00:01:52Bagderchandlerc: I can't tell since I don't know much about that player
00:03:29*pixelma notices that there is no "c200" category in the tracker yet
00:03:44maxkelleyI've noticed such as well.
00:03:46Bagderme fix
00:04:20ruinedtwiceI think I'll take a chance and just toss in a terse precis of the recovery sequences I discovered for each of the above described problems (even though I know that there's a long thread on the forum about the first problem).
00:04:23maxkelleyme fix! (takes out caveman club)
00:04:31markunBagder: was there a firmware dump for the nano 3g?
00:04:35ruinedtwice(shout me down quick, or I'll go ahead)
00:04:40Bagdernot that I've seen
00:04:46 Join jurrie [0] (
00:04:52Bagderin fact, there is very little documented about all of these encrypted ipods
00:05:07Bagderc200 player added to flyspray
00:05:09 Quit jgarvey ("Leaving")
00:05:19pixelmame thanks :)
00:05:36*maxkelley reward Bagder with extra caribou thigh.
00:05:43preglowwe should get an ipod classic for someone with skill
00:05:54maxkelleymad skillz
00:05:58 Quit midkay ("Leaving")
00:06:17BagderMrH's pretty much done with the sansa now, isn't he? ;-)
00:06:23chandlerci would be willing to send one around January in all likelihood
00:06:31peturthere was this crypto guy some time ago asking for a bootrom dump...
00:06:42Bagderah yes
00:06:43 Quit Rob222241 ()
00:06:53chandlercand i'll provide any dumps people want if they can give me some pointers on what they want
00:06:53markunBagder: you can at least ask him if he's interested
00:07:10ruinedtwiceFor the first problem, first I d removed the back plate on the unit and momentarily disconnected the battery, confirmed that Rockbox booted as expected, and then re-attempted to boot OF seeking duplication of the anomally (which occured I'm guessing to no-one's surprise).
00:07:13Bagderbut seriously, I assume public key crypto and that means hard as hell to crack
00:07:25 Join chrisjs169|brb [0] (
00:07:43chrisjs169|brbNice - FS7487 is working
00:07:52Bagder... or rather brute-force factoring
00:07:59chandlercBagder: have to be done by making (guesses sadly) at both the source and the decryption chip, and then brute force from there
00:08:25maxkelleychandlerc: do you have bodyguards to prevent being eaten by steve jobs?
00:08:29pSXAuthorthere is no _way_ you are going to brute force it if its done correctly (ie. a digest hash)
00:08:34ruinedtwiceNext I established a recovery mode connection following the standard protocol and ported the mi4 file for 1.02.18a, followed by disconnection and reboot to OF mode without rockbox (again, I'm guessing to no one's surprise).
00:08:35chandlercmaxkelley: no, but i'm not too worried
00:08:40preglowBagder: i'd expect that too
00:08:55preglowBagder: i'm really quite sure we'll need to hack the hardware on this one to find out anything at all
00:09:00chandlercmaxkelley: if need be, i've got some friends i could mail my ipod to and have them do it... ;] they have much less to be worried about
00:09:01scorche|workwe could always try and contact those quantum computing guys...
00:09:08preglowlogic analyzer action
00:09:11pSXAuthoreven quantum computing won't help
00:09:13Bagderpreglow: quite likely yes
00:09:14ruinedtwiceok so much for obvious reduncancy, now for the second not yet known to me to be commonplace failure.
00:09:19pSXAuthorit only gives a quadratic speedup
00:09:31BagderI hear there are russian and chinese companies doing that stuff ;-)
00:09:31pSXAuthor(and that only in theory ;)
00:09:40chandlercpreglow: anyone who has the know-how to do that given the hardware?
00:09:55maxkelleywe'll create a RockboxIpodEncryptionCracking@Home association.
00:10:13preglowchandlerc: well, i would, as would other people in here, but it's still lots of work
00:10:16Bagderyeah, and in two bazillion trillion years... we've reached 0.02%
00:10:17pSXAuthorso, for example for the 360 signed exes it might reduce the time from 100000000years to 10000years ;p
00:10:19pSXAuthorbig deal ;)
00:10:20sin613so long as i don't have to buy an ipod to pacticipate
00:10:32preglowchandlerc: even doing that seems to be hard, apple have glued the chips to the board with epoxy
00:10:35 Quit animeloe ("This computer has gone to sleep")
00:10:43chandlercpreglow: bastards...
00:10:46maxkelleydon't we all <3 epoxy :)
00:11:07preglowyeah, they really seem to have jumped through hoops to lock us out on this one
00:11:08pSXAuthoryou might be able to read the boot rom with jtag if it exists
00:11:10chandlercpreglow: well, i have a (nearly) free ipod classic, and in 3 months or so, i would be willing to purchase hardware for people to experiment with if that helps...
00:11:15 Quit petur ("Zzzzz")
00:11:23preglowchandlerc: a logic analyzer isn't cheap :P
00:11:30scorche|workpreglow: well, there still is the 2nd gen nano too..
00:11:37chandlercpreglow: hardware being the ipods... ;]
00:11:43chandlercpreglow: and how not-cheap are we talking about?
00:11:46pSXAuthorpreglow: actually - you can buy very good cheap logic analysers these days
00:12:01pSXAuthorfor about $200 you can get really nice ones that attach via usb to a pc
00:12:05preglowchandlerc: easily 5000++ euro
00:12:13preglowthat's for good stuff, though
00:12:20chandlercsee, $200 is a possibility
00:12:38ruinedtwiceIn the second case, porting 1.02.18a via recovery mode failed, as the device "ignored" the attemp, both with and without the rom file, neither "taking" the installation, nor doing anything new and worse.
00:12:39chandlercI just want the very nice hardware of the ipod w/o the very annoying software...
00:12:54pSXAuthorcheap ones don't have massive buffers... but in all seriousness - if you need a large buffer its probably going to be too difficult anyway
00:13:17preglowyeah, perhaps it has jtag
00:13:20chandlercpSXAuthor: link? just for referencing
00:13:21pSXAuthoranyway: i would try to attack it via jtag first... its possible you will be able to read the boot rom
00:13:22scorche|workchandlerc: then you shouldnt have bought a classic ;)
00:13:23chrisjs169|brbruinedtwice, what? i don't feel like looking back up ;)
00:13:30preglowif they went to all this trouble then included jtag, though...
00:13:38chandlercscorche|work: i ordered a video... tehy "upgraded" me for free... bastards
00:13:52maxkelleyI have a grand idea for a new port: Those musical greeting cards! I have a pcb sitting in front of me right now! We can play up to 15 seconds of tin-can quality music! and with flac support, it could be decreased to as far as 0.000001 second!
00:13:53pSXAuthorpretty much ALL consumer electronic devices have jtag... it is used for testing during manufacturing
00:14:02scorche|worksell the classic, then buy a cheaper made + rockbox
00:14:03ruinedtwiceAt this point I noodled around the periphery of the e200tool realm of the damned, partially downloading the linux ISO and printing out some of the half-a-hope dream commentaries of lost souls.
00:14:25chandlercscorche|work: thats actually what i'm looking at doing if there isn't hope for breaking it
00:14:33maxkelleyruinedtwice: I don't think anyone's keeping track of your testimonials.
00:14:38pSXAuthorif its based on a common chipset (eg pp) then the jtag interface may even be documented ;p
00:14:44chandlercscorche|work: i'd love to keep the classic because its thinner / better battery, but .... no rockbox
00:15:25ruinedtwicethat's ok maxkelley, I'm just tossing the descriptions into the log so they can be found via text-search later. I'm confident that they can be of use to someone (even if not here, now, today).
00:15:50preglowit annoys me no one has tried to use the e200 dev boards we have for jtagging yet
00:15:54Bagderruinedtwice: but isn't all this already covered in the unbrick wiki page?
00:15:56maxkelleyruinedtwice: Ah, ok.
00:16:31ruinedtwiceBadger I'm confident that the recovery I "discovered" in the second case is not described in the unbrick wiki page.
00:16:33rasherpreglow: why is that, again?
00:16:48preglowrasher: i have no idea
00:16:53rasherYou'd think it'd be like keeping kids from opening their christmas presents
00:16:57preglowno one has asked for them
00:17:02sin613would something like be of much use for cracking firmware encryption?
00:17:13ruinedtwiceI might be wrong, but I'll just go ahead and toss it in here...
00:17:43scorche|worki saw a "jtag brute forcer" around somewhere...
00:17:58 Join barrywardell [0] (
00:18:24Bagderruinedtwice: it would make about a whole lot more sense to update the wiki page instead
00:18:34Bagderthere's where people look
00:18:44Bagder(because we point them to it)
00:20:28maxkelleypixelma: I can confirm the fm patch. I'm happy to see this port coming together nicely :)
00:20:54ruinedtwiceok, I guess I can update the wiki. (incidentally, I'm guessing that not hosting e200tool for win32 is merciful on your part?
00:21:27Bagderit doesn't work on win32, so yes
00:21:35Bagderit prevents people from trying and failing and then whining
00:21:41Bagdernow they just whine
00:22:34 Join iamben [0] (
00:22:40Bagderbut the source is there so anyone can build it and try
00:23:16*n1s curses integer division...
00:23:38ruinedtwiceright, I was planning to get and mount the linux ISO and give it a whirl, but I stumbled upon a way to get back to base making it a moot point.
00:24:07krazykityou can't just mount the linux iso. you actually have to burn it to a cd and boot from that cd
00:24:23pixelmamaxkelley (and low_light for the logs):
00:24:36maxkelleythanks :)
00:25:01bluebrotheranyone against closing FS #7835?
00:25:04maxkelleypixelma: can you edit fs tasks?
00:25:35maxkelleyMark Arigo :)
00:25:48rasherbluebrother: no. He can always request re-opening if he thinks of a good reason
00:26:04pixelmaoops, thanks for the heads up :)
00:26:17ruinedtwicekrazykit, the heady atmosphere of informality inspired me to imply "burn onto cd and boot from that" by the phrase "get and mount". Good for you, though, to toss in that snippet of specificity. It's bound to be useful to someone, at least in principle.
00:26:26maxkelleyooh, player type is finally the c200!
00:26:53 Nick fxb is now known as fxb__ (
00:27:01krazykitruinedtwice, yeah, well, people have tried doing something like that in vmware and such and started whining when it didn't work. just wanted to make sure you'd do it right :D
00:27:08pixelmaBagder was fix ;)
00:28:05bluebrother=6 :P
00:29:23ruinedtwicewell, it looks like I've successfully repressed my impulse to describe my second recovery. Apart from that, thanks linuxstb_ and Badger, and krazykit for your feedback, guidance, and directions.
00:29:24 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
00:30:08maxkelleyhehe. good to hear you're on your way to getting it resolved.
00:30:31 Quit spiorf (Remote closed the connection)
00:30:37ruinedtwiceI'm going to go ahead now and look up that blue wave of death thread at the forum.
00:30:44 Part ruinedtwice
00:31:10 Join spiorf [0] (
00:32:37maxkelleyI'll be back later.
00:35:46 Join eigma [0] (
00:35:52 Quit Bagder ("*plopp*")
00:35:54eigmalinuxstb_: ping
00:36:21 Join daurn|laptop_ [0] (n=fake@
00:37:56pixelmascorche|work: how are the chances for the official "rockbox-themes" being available soonish? Just wondering if it's worth it to start a new wiki page for c200 themes/WPSs...
00:37:58*Zagor feels the need for dual displays
00:38:43bluebrotherbuild table growing too wide?
00:39:12pSXAuthorpeople still use single monitors?
00:39:19linuxstbeigma: here
00:39:28 Quit scorche|work ("CGI:IRC (EOF)")
00:39:42eigmawhat do you think :S
00:40:11eigmaI can show you a diff of a make at r14918 and then a make at r14919, but it's not particularly useful
00:40:29 Join scorche|work [0] (n=8dc5049d@rockbox/administrator/scorche)
00:40:44*scorche|work kicks cgi:irc for the third time today
00:41:04scorche|workpixelma: i will have a look at the most recent version when i get home and let you know
00:41:19 Quit barrywardell ()
00:41:21pixelmaok, no hurry
00:41:41IsolinearDarn, I was hoping to see 7738 committed when I woke up... :P
00:41:47scorche|worksure there is depressing how long it has been sitting there =/
00:41:51 Quit obo ("bye")
00:42:00 Quit bluebrother ("out for sleep now.")
00:42:18 Join nukem2525 [0] (
00:43:44 Quit ender` (" This space intentionally has nothing but text explaining why this space has nothing but text explaining that this space woul)
00:46:08 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
00:49:12 Quit nukem2525 ()
00:50:56 Quit roolku ()
00:53:02 Quit daurn|laptop (Read error: 110 (Connection timed out))
00:54:12 Join midkay [0] (n=midkay@rockbox/developer/midkay)
00:56:15 Join Mouser_X [0] (
00:56:52Mouser_XI can't remember my login for the flyspray... How do I find out? (I've never actaully used it. Maybe it got deleted?)
00:58:10 Quit kugel ("ChatZilla [Firefox]")
00:58:40ZagorMouser_X: do you remember the email address you registered with?
00:59:01 Join kugel [0] (
00:59:05Mouser_XWell, there's only 2 to choose from.
00:59:19 Quit scorche|work ("CGI:IRC")
00:59:20Mouser_X(It would have been either my yahoo account, or my hotmail acount.
00:59:35Mouser_X) (Both are mouser_x@)
01:00:55 Join Lars_G_ [0] (n=Lars@unaffiliated/lars-g/x-000001)
01:01:06 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
01:01:17Lars_G_How's the midi comming? I forgot who it was fighting it.
01:01:21Lars_G_bah gtg, brb
01:02:42ZagorMouser_X: well your login it is "mouser x". I can reset your password if you like?
01:02:49Zagorlogin id
01:03:29***Saving seen data "./dancer.seen"
01:03:30Mouser_XAh. Let me try again.
01:04:25Mouser_XThanks. I got it.
01:04:34Mouser_X(I guess I mistyped or something...)
01:05:47 Join Falco98 [0] (
01:05:52n1sLars_G_: I've been playing around with the midi plugin the last couple of days and have achieved a quite nice speedup for coldfire targets which was used to improve the playback quality, on pp targets the performance still stinks though
01:06:03Falco98what should i see if i open nvram.bin in a hex editor?
01:06:17Falco98should it be human-readable?
01:06:44n1sFalco98: it will be hex numbers, idk if you can read them
01:07:01Falco98ok.. cuz at the moment, mine is human-readable
01:07:18Falco98it says "# .cfg file created by rockbox r13767-070703"
01:07:49Falco98my build is currently experiencing the bug (which only I seem to ever seE) where if i power-down manually, it has lost all settings upon restart
01:07:49n1sthat's clearly wrong
01:08:02Falco98and if i let it idle poweroff, i retain my settings but lose my dynamic playlist
01:08:31n1sFalco98: are you using voice?
01:08:49Falco98not at all
01:09:01Nico_Psin613: nice link (NSA)
01:09:14Nico_Psin613: has it been used to crack anything ?
01:09:16Falco98when i put a new build on it works fine for weeks, but sometimes will very suddenly start acting this way
01:09:33Falco98and there seems to be no way to fix it other than overwriting the build.. i haven't quite been able to trace it down to one file yet
01:11:00n1sFalco98: maybe file system corruption, have you chaecked it?
01:11:38Falco98this is the third or forth time this specific bug has happened.. i've experienced no other unexplained problems, nor have i had any file corruption or loss
01:11:49Falco98why would this be happening so consistently and nothing else?
01:12:38 Quit mister_pink ("ChatZilla [Firefox]")
01:12:49n1sFalco98: which player do you use
01:12:55Falco98iriver h140
01:12:57RaZorbacKcould someone please tell me what is the file in the source which is generating the config. cfg? i'd like to create a voice-enabled build
01:13:21RaZorbacKa french voice-enabled build
01:13:29 Quit Zagor ("Client exiting")
01:22:44 Quit spiorf (Remote closed the connection)
01:23:28Falco98any luck?
01:24:31n1sFalco98: no ideas, really, have you checked your filesystem?
01:24:44Falco98how do you recommend i best do that?
01:24:53 Join ruinedtwice [0] (
01:25:08n1schkdsk or fsck.vfat dependign on your os
01:25:08Falco98whatever i've done so far hasn't revealed anything special
01:25:22n1schkdsk then
01:25:30Falco98k, running
01:26:07 Quit chandlerc (Remote closed the connection)
01:26:15Falco98hm, it reports that config.cfg was cross-linked
01:27:12ruinedtwicealright I've successfully enabled dual boot on my e200 following the link to the .02 generation sansapatcher. Thanks linuxstb for the referal.
01:27:17Falco98would you recommend a chkdsk /f ?
01:27:44 Join chandlerc [0] (n=chandler@
01:27:59n1sFalco98: if that's the "fix" switch, yes but you should always have a bacup of your data :-)
01:28:23 Quit chrisjs169|brb (Read error: 110 (Connection timed out))
01:28:28Falco98k, well of course the music is all secondary copies, the main ones are on my HD and 100% backed-up on mozy :-P
01:28:37ruinedtwicenow I've googled and searched the forums for this next one, and haven't gotten any hits, namely:
01:29:38 Join RudMan [0] (
01:30:19ruinedtwiceMy H340 intermittently succeeds in hosting my e280. On most attempts the H340 hangs while occasionally the hosting will work (H340 in OF browser mode providing local host USB). The H340 and e200 hardware are known to be in unremarkable good condition. Is this a known issue?
01:31:26Falco98is "config.cfg" a special system file?
01:31:39ruinedtwice(only tangentially related to Rockbox, in as much as both devices have dual boot Rockbox / OF modalities.
01:31:45n1sFalco98: what do you mean?
01:32:08Falco98is it something generated by rockbox itself? or is it a user-saved config file?
01:32:17n1sit's generated
01:32:30Falco98i.e., what does rockbox refer to when it starts, to load settings from?
01:32:54n1sif the config.cfg file exits it is used otherwise defaults are used
01:33:21Falco98so if i were losing my configuration settings upon a manual restart, it would have something to do with config.cfg
01:33:29Falco98(or *could*, anyway)
01:34:00 Join chrisjs169|brb [0] (
01:34:36 Quit eigma ()
01:35:31n1sFalco98: yes
01:35:54Falco98actually, i think nvram.bin and config.cfg have been getting cross-linked somehow
01:35:57n1sprobably that wasn't written correctly on shutdown
01:36:20Falco98i saved the files in an old build when it went bad, and the contents of config.cfg were the binary jumble you'd expect in nvram.bin
01:36:30Lars_G_wont rbox remember the last file you saved conf to and use it? maybe at some point you wrote the config on nvram.bin and it stuck there
01:36:43 Join jac0b [0] (
01:36:45 Join eigma [0] (
01:36:52n1sLars_G_: nope, config.cfg is hardcoded
01:37:03jac0bcan I delete or replace images in the wiki?
01:38:26 Quit chrisjs169 (Connection timed out)
01:38:33pixelmaconfig.cfg only saves settings that are different from the default (built-in) ones...
01:39:21n1sI mean hardcoded as the name of the cfg file which is read on boot
01:39:28 Quit BigMac (Read error: 110 (Connection timed out))
01:39:43pixelmaso at boot it should choose default settings +/- what's in config.cfg
01:39:45 Join barrywardell [0] (n=barrywar@
01:40:28pixelmaah, ok - misread then
01:41:49 Quit ruinedtwice ()
01:42:12pixelmaI shouldn't join in the middle of a discussion and rather get some sleep... nn :)
01:45:13 Part pixelma
01:45:40 Join Soader03 [0] (i=18caf221@gateway/web/cgi-irc/
01:46:38 Quit Nico_P (Remote closed the connection)
01:48:54 Part lamule01
01:49:36 Quit Soader03 (Client Quit)
01:50:27 Quit eigma ()
01:54:58 Quit RaZorbacK (Read error: 110 (Connection timed out))
01:55:51 Join BigMac [0] (
01:56:17 Quit jac0b ()
01:58:20 Join RaZorbacK [0] (
01:58:51 Join carter [0] (i=d8c27c24@gateway/web/cgi-irc/
01:59:16 Join JdGordon [0] (
01:59:34 Join perrikwp [0] (i=98214cbd@gateway/web/cgi-irc/
02:01:46carterhi everyone i am using ipodpatcher.exe to install ipodlinux and i need to know why i am getting this error message: couldnt open input file loader- loader.bin
02:02:18 Quit barrywardell ()
02:02:38carteranyone there?
02:02:41linuxstbcarter: Because you don't have that file in the same directory as ipodpatcher
02:03:02carterplease explain. :)
02:03:14linuxstbExactly what command are you typing?
02:03:45carteripodpatcher 1 -ab loader.bin
02:04:11linuxstbAnd do you have a file called "loader.bin" in the same folder as ipodpatcher.exe?
02:04:28carterno... that would probably work... lol
02:05:57carteromg let me unplug my ipod and see
02:06:01carterthe moment of truth...
02:07:00cartercant load rockbox.ipod: file not found
02:07:15carterim soo a newbie
02:07:33linuxstbSounds like you installed the Rockbox bootloader...
02:07:50linuxstbWhich given that you're in #rockbox makes it a success.
02:12:58Lars_G_carter: what the fuck are you doing in here?
02:13:39Lars_G_carter: now you need to install rockbox.. itself you know
02:13:59Lars_G_unzip for your device in the root mount dir (root dir of the device)
02:14:23Lars_G_btw if this is your first install get the and the fonts zip
02:15:12Lars_G_btw welcome to a world of fun and rockbox.
02:15:18Lars_G_it's been my main FW for a while now
02:15:48Lars_G_in fact I was compiling the lattest svn pull. let me see if it's done
02:17:35 Quit carter ("CGI:IRC")
02:17:57 Quit kugel ("ChatZilla [Firefox]")
02:21:04 Quit BigMac (Read error: 110 (Connection timed out))
02:25:19 Quit hannesd_ (Read error: 110 (Connection timed out))
02:27:47 Nick chrisjs169|brb is now known as chrisjs169 (
02:36:45 Quit XavierGr ()
02:36:58 Join XavierGr [0] (
02:39:42 Quit perrikwp (
02:39:42 Quit animeloe (
02:39:42 Quit sin613 (
02:39:42 Quit Lars_G_ (
02:39:42 Quit qweru (
02:39:42 Quit Mouser_X (
02:39:42 Quit ENd___ (
02:39:42 Quit jhMikeS (
02:39:42 Quit Bger (
02:39:42 Quit ze (
02:39:42 Quit chris__ (
02:39:42 Quit tchan (
02:39:42 Quit Caliban (
02:39:42 Quit solatis (
02:39:42 Quit Galois (
02:39:42 Quit Presence (
02:39:42 Quit Zom (
02:39:42 Quit luckz (
02:39:57 Quit My_Sic (Read error: 104 (Connection reset by peer))
02:40:44 Join blargit [0] (
02:40:44NJoinperrikwp [0] (i=98214cbd@gateway/web/cgi-irc/
02:40:44NJoinLars_G_ [0] (n=Lars@unaffiliated/lars-g/x-000001)
02:40:44NJoinMouser_X [0] (
02:40:44NJoinanimeloe [0] (n=animeloe@unaffiliated/animeloe)
02:40:44NJoinsin613 [0] (
02:40:44NJoinqweru [0] (
02:40:44NJoinENd___ [0] (
02:40:44NJoinjhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
02:40:44NJoinBger [0] (n=Bager@
02:40:44NJoinze [0] (
02:40:44NJointchan [0] (n=tchan@lunar-linux/developer/tchan)
02:40:44NJoinCaliban [0] (n=ianmacd@
02:40:44NJoinsolatis [0] (
02:40:44NJoinPresence [0] (
02:40:44NJoinZom [0] (
02:40:44NJoinGalois [0] (
02:40:44NJoinluckz [0] (
02:40:58 Join theGrue [0] (
02:42:21 Quit luckz (Remote closed the connection)
02:42:54 Join luckz [0] (
02:43:14 Quit Falco98 ("ChatZilla [Firefox]")
02:45:05 Quit Zom (Remote closed the connection)
02:45:12 Join Zom [0] (
02:47:46theGruehi, i've got some questions about rockbox on an iriver h320, is anyone around thats familiar with it?
02:54:26 Quit hcs (Read error: 113 (No route to host))
02:58:55 Quit n1s ()
03:01:56 Quit Thundercloud (Read error: 104 (Connection reset by peer))
03:03:30***Saving seen data "./dancer.seen"
03:03:32 Join BigMac [0] (
03:06:01 Join velixzeen [0] (n=velixzee@
03:15:26XavierGrtheGrue: tell us then and maybe someone will answer
03:19:55 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
03:25:49 Join Ebert [0] (
03:28:51 Quit Rick (Read error: 104 (Connection reset by peer))
03:29:23 Join Rick [0] (
03:31:01 Part velixzeen
03:37:19 Join Gibbed [0] (
03:37:23 Quit Rick (Nick collision from services.)
03:37:37 Nick Gibbed is now known as Rick (
03:41:07 Quit Rick (Read error: 104 (Connection reset by peer))
03:41:29 Join Rick [0] (
03:41:55 Join _BigMac [0] (
03:47:34 Join elinenbe_ [0] (
03:51:46 Join Soader03 [0] (i=18caf221@gateway/web/cgi-irc/
03:58:11 Quit Soader03 ("CGI:IRC")
03:58:17 Quit sin613 ("Leaving.")
03:59:14 Quit BigMac (Read error: 110 (Connection timed out))
04:01:23 Join sin613 [0] (
04:05:41 Quit elinenbe (Read error: 110 (Connection timed out))
04:05:41 Nick elinenbe_ is now known as elinenbe (
04:07:34 Quit Lars_G_ ("leaving")
04:08:56theGruehaha okay
04:09:10theGruei bought a player for cheap that the owner thought needed a new battery
04:09:34theGruei noticed rockbox gives a battery reading when the bootloader starts up, does that have anything to do with capacity?
04:11:51theGrueand my second question was about usb... i seem to need to boot into the original firmware to get it to work. what could the problem there be?
04:12:16chrisjs169theGrue, what player are you referring to?
04:12:45chrisjs169theGrue, there's no USB support in Rockbox.
04:12:57scorchesure there is
04:13:04scorchejust not on PP devices
04:13:16theGruethere's not?
04:13:24theGruei plug it in and it says its going into some usb mode
04:13:28theGruebut doesn't show up
04:13:35scorchewell, what device do you have
04:13:46theGrueand all the iriver faqs say the bootloader is great because it can go into the disk if it's damaged
04:13:54theGrueiriver h320
04:14:00JdGordonsomething is wrong then...
04:14:05JdGordonh300 does have usb in rockbox
04:14:20theGruesorry, i think i have join/parts off, i guess i cant tell if you saw my original question
04:14:24theGruei should fix that
04:14:40sin613he saw it
04:15:30chrisjs169scorche, sorry - i'm just used to PP targets
04:15:33*chrisjs169 will shut up now
04:16:42alienbiker99chrisjs169 is does your build have all the patches once again?
04:16:43 Join low_light [0] (i=4cbdfa51@gateway/web/cgi-irc/
04:17:07chrisjs169alienbiker99, still adding them - a build from a few days ago should be fine though
04:17:31low_lightwho can tell me how I messed up the $Id line in my last commit?
04:17:45sin613pp targets don't have working usb?
04:17:51JdGordonlow_light: in which file?
04:17:55chrisjs169sin613, no
04:17:59 Quit RudMan (Remote closed the connection)
04:18:24low_lightthe last commit
04:18:29JdGordonlooks like the keyword isnt set properly
04:19:18low_lightIs that something I do in my local tree? because I didn't change that line.
04:19:37JdGordonno, its a svn thing, its not set right on the server
04:19:49JdGordonheaps of files arnt set correctly :(
04:19:55JdGordontrying to find the command to fix them
04:20:04sin613chrisjs169: with pp OF are they usually non-HID devices?
04:20:12sin613rather, non-mass storage
04:21:27chrisjs169sin613, the OF has UMS/MSC support, it's just not coded in Rockbox (IIUC)
04:21:31low_lightJdGordon: but in the previous commit amiconn "added missing keywords"...I'm not sure how I would have removed them
04:21:51JdGordonhmm.. no, keywords are set on that file corrrectly
04:21:53sin613ahh. glad i don't have a pp target
04:21:55*JdGordon doesnt know :p
04:21:58 Quit sin613 ("Leaving.")
04:24:00low_lightdo any targets have real-time detection for the radio? (for when there are versions with & without)
04:24:12JdGordonrecorder and sansa
04:24:20JdGordone200 i mean
04:26:13low_lightwhere's the detection for the e200? (hopefully its the same/similar to the c200)
04:28:24JdGordonfigure out what radio_hardware_present() does
04:29:23JdGordondoes the c200 have the same tuer as the e200?
04:31:18JdGordonso you might be lucky and not have to do anything special
04:31:44chrisjs169IIRC the radio is initialized differently, isn't it?
04:35:01low_lightyes, there's a different register bit
04:37:12 Join My_Sic [0] (
04:40:01 Quit perrikwp ("CGI:IRC (Ping timeout)")
04:41:04 Join yapper [0] (i=18ec7c6f@gateway/web/cgi-irc/
04:42:33yapperGreetings ... can someone sort out Wiki write permissions for me?
04:45:10JdGordon404.. please not found
04:49:07XavierGrtheGrue: are you reffering to bootloader usb mode or normal usb mode when rockbox is fully booted?
04:49:09 Part low_light
04:49:43XavierGrtheGrue: if you have problems with the bootloader usb mode, then yes it is a known problem for some H300s
04:50:06XavierGrSVN bootloader doesn't have that problem but I would suggest to you that you wait until a new one is released on the wiki
04:50:56 Quit yapper ("CGI:IRC")
04:51:36 Join yapper [0] (i=18ec7c6f@gateway/web/cgi-irc/
04:55:22 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:55:31 Join miepchen^schlaf [0] (
04:58:51 Quit ENd___ (Read error: 110 (Connection timed out))
04:59:37 Quit yapper ("CGI:IRC (EOF)")
05:00:42 Join tedrock [0] (
05:01:45 Quit tumu (Read error: 104 (Connection reset by peer))
05:01:56 Join tumu [0] (
05:03:32***Saving seen data "./dancer.seen"
05:13:27XavierGrOMFG! I just flashed my H300 and when I first turned on I got an exception error. Fortunately on the second boot it worked okay
05:13:31XavierGrI was scared to death
05:16:00 Join bb__ [0] (
05:17:14tumulittle jolts are nice once in a while :)
05:17:31*JdGordon also enever liked flashing the h300
05:20:18XavierGrfor some seconds I was like "Shit I finally bricked the damn thing!"
05:21:51XavierGranyway good news is that Linus' patch works like a charm
05:22:02XavierGrfresh SVN bootloader works :)
05:22:12XavierGrLinus: Please commit the fix.
05:25:47alienbiker99there is no flashable bootloader like the H1x0 for the H3x0 yet is there?
05:27:39 Quit bb (Read error: 110 (Connection timed out))
05:28:01LloreanXavierGr: Does an exception error on first boot really qualify as "working"?
05:28:07LloreanWe don't want to panic all our other users...
05:30:42XavierGrLlorean: I got this only at the first boot
05:30:48XavierGrcan't replicate it
05:31:18XavierGrit was the first keypress after the flashing procedure, maybe it was pure software/hardware "magic"
05:31:32XavierGralso pondlife didn't get that (tested it before me)
05:31:41LloreanIt would still be nice not to cause everyone to have an exception on first boot after a bootloader update.
05:31:45LloreanAh, okay
05:31:47XavierGralienbiker99: no there isn't
05:31:48LloreanStill, curious why.
05:32:12 Join darksaboteur [0] (
05:33:08XavierGrLlorean: yeah I am quite curious too, but I don't plan to investigate it any further after my initial shock :P
05:34:11JdGordonill try it on my h300...
05:34:18JdGordonjust have to remember how to do it :p
05:34:32*scorche wonders if JdGordon is tempting fate ;)
05:34:48JdGordonalienbiker99: the h100 almost has rombox i tihnk (flash loader)
05:34:49XavierGrJdGordon: see the IriverBoot wiki page
05:35:04JdGordonscorche: well.. its not beign yused atm anyway so briking it is no big deal
05:35:17XavierGrJdGordon: I always forget the process even after 20 times flashing my H300/H100
05:35:54 Join kubiix [0] (
05:36:22XavierGrJdGordon: yes H100 has rombox and rambox, alienbiker99 was saying if H300 has that too, which is not
05:36:58alienbiker99yeah. i forget how to flash the bootloader too, but i'm not going to attempt to try the new svn as my battery doesnt work.
05:37:00JdGordonmisread :p
05:37:47JdGordonXavierGr: which OF version should i get?
05:38:07XavierGrJdGordon: I tried with 1.31k
05:38:10*Llorean swears he's going to ban Dunno one of these days just for being so bloody stubborn
05:38:51LloreanHe seems 100% immune to logic.
05:39:07JdGordonXavierGr: can you dcc me the OF .hex? the zip has a bloody exe in it
05:39:29XavierGrJdGordon: that is 1.31k with revision 14811 and Linus' patch (don't forget the last one otherwise it just wont boot rockbox)
05:39:45XavierGrJdGordon: ok one moment
05:39:46JdGordonwithout the patch..
05:40:01 Quit My_Sic ("Miranda IM! Smaller, Faster, Easier.")
05:41:00XavierGrJdGordon: you mean to send you the OF without the bootlaoder (the original .hex)
05:53:23 Quit kubiix (Read error: 104 (Connection reset by peer))
05:55:30 Join elinenbe_ [0] (
05:56:15 Quit Mouser_X (Read error: 110 (Connection timed out))
05:59:11 Join toffe82 [0] (n=chatzill@
06:02:52 Join webguest80 [0] (i=46f58ac0@gateway/web/cgi-irc/
06:03:29 Join toffe82_ [0] (n=chatzill@
06:05:50 Join eigma [0] (
06:06:16 Quit webguest80 (Client Quit)
06:11:48 Quit miepchen^schlaf ("Verlassend")
06:15:32 Quit elinenbe (Read error: 110 (Connection timed out))
06:15:32 Nick elinenbe_ is now known as elinenbe (
06:20:19 Quit toffe82 (Read error: 110 (Connection timed out))
06:28:12 Quit Rick (Read error: 104 (Connection reset by peer))
06:29:07 Join Rick [0] (
06:33:10 Quit lazka (Remote closed the connection)
06:33:44 Join DC1 [0] (
06:34:02 Quit DC1 (Client Quit)
06:53:03 Quit idnar (Nick collision from services.)
06:53:06 Join idnar_ [0] (n=mithrand@unaffiliated/idnar)
06:56:29 Join n17ikh [0] (
06:58:58 Quit n17ikh|Lappy (Nick collision from services.)
06:59:03 Nick n17ikh is now known as n17ikh|Lappy (
07:03:33***Saving seen data "./dancer.seen"
07:17:54JdGordonam i still here?
07:18:10scorchementally or physically?
07:18:10JdGordonoh well
07:18:19JdGordonnetowkr link-ally
07:18:40scorchebroken keyboard-ally?
07:18:59*scorche really isnt one to talk
07:19:01JdGordonNEVER buy a keyboard/mouse combo for $60... even if its logitech branded
07:20:18 Join sin613 [0] (
07:28:46 Join davina [0] (
07:33:23 Nick fxb__ is now known as fxb (
07:41:04toffe82_I have a problem with this : rb->lcd_clearrect(0, 0, LCD_WIDTH, REC_HEIGHT); I get an error"In function printresult" structure has no member named 'lcd_clearrect'
07:41:39JdGordonlcd_clearrect isnt in the plugin api probably
07:41:46toffe82_it is
07:42:10toffe82_ void lcd_clearrect(int x, int y, int width, int height);
07:42:11toffe82_ Clear a rectangle area.
07:42:42JdGordonwhat file is that in?
07:42:48JdGordonthe plugin api is plugin.h
07:42:59amiconnThat function is gone years ago...
07:43:12toffe82_it is used in a lot of place
07:43:33amiconnIt's definitely not
07:43:33 Nick fxb is now known as fxb__ (
07:43:37toffe82_perhaps not
07:43:48amiconnYou are probably looking at a years old tree
07:44:12toffe82_I just made a svn update
07:45:07toffe82_what can I use in place of this ?
07:47:08amiconnlcd_clearrect was removed on 24 Jun 2005, in r6856, by me
07:47:53amiconnIf you want to clear the whole display, lcd_clear_display(); is best
07:48:15amiconnFor partial clearing, use lcd_fillrect() with a suitable draw mode
07:49:53toffe82_thank you
08:11:02 Quit Seed (Nick collision from services.)
08:11:09 Join Seed [0] (
08:15:02 Quit midkay ("Leaving")
08:18:54 Quit TMM (Read error: 110 (Connection timed out))
08:22:45 Join norbusan [0] (
08:22:50 Part norbusan
08:23:59 Part toffe82_
08:27:42 Quit BigBambi ("Leaving")
08:32:02 Join kubiix [0] (
08:36:18 Join hannesd_ [0] (
08:36:37 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
08:38:26 Quit JdGordon ("Konversation terminated!")
08:41:37 Join JdGordon [0] (
08:48:30 Quit darksaboteur ("ChatZilla [Firefox]")
08:49:54 Join ender` [0] (
08:54:29 Quit FOAD ("I'll be back")
08:59:32 Quit Toki (Read error: 104 (Connection reset by peer))
08:59:44 Join webguest16 [0] (i=cb3b81c3@gateway/web/cgi-irc/
09:01:28webguest16Sorry, SMF was unable to connect to the database. This may be caused by the server being busy. Please try again later.
09:01:44webguest16Is it overloaded, or is something wrong?
09:02:01LloreanA bit of both, this is common at this time of night/day
09:02:09webguest16aah ok
09:02:50webguest16i mean "ahh", not "aah"
09:02:57webguest16the a to h ratio was off
09:03:36***Saving seen data "./dancer.seen"
09:04:31 Join Toki [0] (
09:06:40 Join petur [0] (n=petur@rockbox/developer/petur)
09:11:30 Quit sin613 ("Leaving.")
09:11:36 Quit RaZorbacK ("bye all")
09:13:25 Part webguest16
09:15:49peturhmmmm zune2 is coming and still no zunelinix on version 1?
09:16:20 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
09:17:31 Quit eigma (Read error: 113 (No route to host))
09:26:14 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
09:29:37 Join JdGordon_ [0] (
09:30:30 Join Bagder [0] (n=daniel@rockbox/developer/bagder)
09:31:07 Quit qweru ("moo")
09:31:10 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
09:32:25 Quit markun (Remote closed the connection)
09:34:19 Nick bb__ is now known as bb (
09:35:14 Quit JdGordon (Read error: 110 (Connection timed out))
09:36:40 Join markun [0] (n=markun@rockbox/developer/markun)
09:38:18 Join spiorf [0] (
09:38:36 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
09:42:42 Quit Ebert ()
09:43:32 Quit JdGordon_ (Read error: 104 (Connection reset by peer))
09:43:34 Join JdGordon [0] (
09:49:56pondlifeLinusN: any objections if I commit your Coldfire bootloader fix?
09:50:29 Join barrywardell [0] (n=barrywar@
09:55:39pondlifeHmm, hung build (again)?
09:57:24pondlifeFalse alarm
09:57:56 Join bluebrother [0] (i=t0Y9eXYY@rockbox/staff/bluebrother)
09:58:07*GodEater_ pokes pondlife and points to PM
10:01:31 Quit chrisjs169 (Remote closed the connection)
10:04:49*petur pokes jhMikeS with a red stick
10:04:53pondlifejhMikeS: Your red, not mine!
10:05:03jhMikeSheh. I hate when I forget to do svn add
10:05:12pondlife(or rather, not LinusN's)
10:05:31 Join n1s [0] ('s defintely me
10:06:02peturor rather the missing wm8721.h
10:06:31*pondlife thinks about keywords too
10:10:39pondlifepetur: Is your last comment on now out-of-date?
10:11:12peturpondlife: I forgot to test!
10:11:22pondlifeIt'd be good if you could...
10:11:49pondlifeDon't know if it's related though - my H300 would never boot (not 50/50).
10:11:52peturI could test it around noon if somebody made one (tested)
10:12:09pondlifeI have such a binary, where would you like me to send it?
10:12:34peturzip it and mail to my gmail?
10:14:03*pixelma would like to point Zagor to FS #7881
10:14:52Zagori'll take a look soon. a bit busy right now.
10:15:01 Join ivan`` [0] (n=ivan`
10:15:40pondlifepetur: I don't see your address - try grabbing it from">
10:16:12pixelmaZagor: sure, as I think low_light should commit it and he's not around at the moment...
10:16:51peturpondlife: thanks, received
10:20:37 Join obo [0] (n=obo@rockbox/developer/obo)
10:22:06 Quit spiorf (Read error: 104 (Connection reset by peer))
10:22:21 Join spiorf [0] (n=spiorf@
10:23:32 Quit donsdw (Read error: 110 (Connection timed out))
10:25:10jhMikeSok, I have to stop being retarded here
10:25:52pondlifeI've tried, but it's not easy.
10:26:25peturrofl... /me hands jhMikeS 1543 points :p
10:26:49pondlife+264... ;p
10:27:06jhMikeSracking up the points ;)
10:29:05 Quit JdGordon ("Konversation terminated!")
10:30:41 Join ivan``` [0] (n=ivan`
10:30:53 Quit ivan` (Read error: 110 (Connection timed out))
10:30:56jhMikeSok, this one should take care of it since those build for me
10:31:58 Quit ivan`` (Read error: 104 (Connection reset by peer))
10:32:04 Nick ivan``` is now known as ivan` (n=ivan`
10:32:52 Join roolku [0] (
10:33:48amiconnjhMikeS: Why do we need an extra WM8721 driver?
10:33:50jhMikeShmm...the colored diff doesn't show the full change (ignores spacing changes always?)
10:34:13amiconnThe 8721 is register compatible with the 8731 except for the missing recording
10:34:15jhMikeSbecause it's not compatible with the wm8731. it lacks most bits and registers
10:34:42jhMikeSAlso, because it doesn't support recording, it can be a much simpler implmenentation
10:34:49roolkujhMikeS: morning, I was wondering if you had read the logs from last night ( from around 23.13)
10:35:07amiconn(and in fact some of the WM8721 targets have a WM8711, or even a WM8731 mounted)
10:35:25peturjhMikeS: yes, the diff with ViewVC ignores whitespace
10:35:33roolkujhMikeS: I was considering committing FS #7487 (after some tidy up), but linuxstb said you were planning to implement seeking?
10:35:37 Nick idnar_ is now known as idnar (n=mithrand@unaffiliated/idnar)
10:35:41 Quit jhMikeS (Read error: 104 (Connection reset by peer))
10:37:04 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
10:37:42 Quit barrywardell ()
10:39:15roolkuwhoops, will try again ;)
10:39:17roolkujhMikeS: morning, I was wondering if you had read the logs from last night ( from around 23.13)
10:39:20roolkujhMikeS: I was considering committing FS #7487 (after some tidy up), but linuxstb said you were planning to implement seeking?
10:39:32jhMikeSroolku: not yet
10:40:48jhMikeScombining the ADC codecs with the non-ADC ones is just messy and gets in my way
10:41:22linuxstbjhMikeS: The people working on that seeking patch seem to be working very hard, and I think we should tell them whether their work will be accepted into SVN or not. So I wanted to know if it conflicted with your plans for mpegplayer.
10:41:53jhMikeSlinuxstb: I have to get acquanted with it.
10:43:13jhMikeSthis actually seeks using timestamps?
10:44:42roolkuyou basically set a resume time using a kind of preview mode before you start playing the video - it seeks in constant time steps (0.5min), yes
10:45:00roolkuthere is no actual seeking while you are playing
10:45:42jhMikeSI wonder how close it is to being able to do that anyway then.
10:47:12roolkuif with "that" you mean "seeking while playing" I think there are performance issues
10:47:26jhMikeSroolku: this is synced with the allocator split and minds data alignment on COP?
10:47:54jhMikeSyeah, you need to be able to buffer sparsely and not load all data in between
10:48:58roolkujhMikeS: yes it is synched, as to whether is minds all the data alignment - I don't know. What is it one would have to look out for?
10:49:34roolkujhMikeS: if you wouldn't mind giving it a quick spin?
10:49:45jhMikeSCOP data shouldn't be in the same 16-byte address segment
10:49:55jhMikeSI was going to do that
10:50:17linuxstbjhMikeS: You mean COP data shouldn't be in the same 16-byte segments as CPU data?
10:50:49*amiconn would expect the same convenience from mpegplayer seeking that video.rock offers
10:50:57jhMikeSlinuxstb: yeah, so that cache lines don't cache the same segment on each processor if sharing isn't intended
10:51:06amiconnvideo.rock auto-saves the resume position of a video when you stop it.
10:51:12linuxstbamiconn: But isn't video.rock using CBR files?
10:51:35amiconnWhen playing it again, video.rock asks whether it should resume, or play from the start
10:52:07amiconnThe video part is cbr (as it's uncompressed lcd frames), but the audio part is plain mp2/mp3, cbr or vbr
10:53:54 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
10:54:45 Join TMM [0] (n=hp@
10:54:45 Join JdGordon [0] (
10:54:46jhMikeSroolku: lots of trailing CRs :) maybe it was IE though.
10:55:16jhMikeSroolku: any particular target(s) you'd like checked?
10:56:25roolkujhMikeS: I think gigabeat has been used extensively (but works best)
10:56:38n1samiconn: it seems your theory about symbol clashing causing the track change crash on 64 bit sims was correct :-)
10:56:39roolkujhMikeS: maybe a PP target ?
10:56:40*jhMikeS was going to torture-test on e200
10:57:02jhMikeSok, building
10:58:02roolkucheers (I think the CRs are mine - forgot to strip)
10:58:17 Join Variable [0] (
10:59:19roolkunope I take that back - no CRs :)
11:00:46roolkuback in 20min
11:03:40***Saving seen data "./dancer.seen"
11:07:38 Quit Isolinear (Read error: 110 (Connection timed out))
11:08:55 Quit JdGordon ("Konversation terminated!")
11:10:37 Join JdGordon [0] (
11:12:39 Quit TMM ("Ex-Chat")
11:12:44jhMikeSroolku: so far so good. If you pull thing off you can definitely seek since a seek is just a pause then this.
11:16:07roolkujhMikeS: I don't think it is that simple, but that would certainly be the next step
11:17:09 Join TMM [0] (n=hp@
11:18:45jhMikeSMy gut says it pretty close to that simple since that's pretty much what's done in other players. Perhaps better thread control is needed but that isn't a problem.
11:20:05jhMikeSIt sometimes fails to seek. It has locked a couple times. Somethimes it doesn't find the seek point but I can bail out.
11:20:34TMMquestion: is it theoretically possible to physically damage NAND from software?
11:20:52roolkusolid lock?
11:21:11jhMikeSIncidentally, the files in question: 1) has timestamps that don't match audio frame length. 2) starts with a bad audio frame
11:22:04jhMikeSIt wasn't a hard lock since I could hear activity in the headphones (a bit of really quiet noise)
11:22:18 Join Arathis [0] (
11:22:41pondlifeAnyone understand what's causing the remaining yellow? cfft.c seems to be making a reasonable cast...
11:22:48roolkushall I commit it and it can be improved while in svn?
11:23:11markunroolku: why don't the seeking guys come over here to talk about it?
11:23:39roolkumarkun: you'll have to ask them
11:23:58jhMikeSroolku: lemme just check that it at least starts those files from the beginning since they do work in SVN
11:24:18markunroolku: is it 7487 or 7276?
11:24:50jhMikeSroolku: could we make the options available as well from the resume menu?
11:25:18roolkujhMikeS: yes, that should be possible
11:25:24jhMikeSthose start fine if I don't seek them
11:25:25n1spondlife: it seems to be a gcc bug
11:26:22 Quit atsea-34 (Read error: 104 (Connection reset by peer))
11:26:35jhMikeSI guess go for it. It works pretty well and things that work now continue to work.
11:27:14 Quit spiorf (Remote closed the connection)
11:27:25markunseeking didn't work very well on my files last time I tried
11:27:55roolkujhMikeS: great. I'll probably do this tonight after having sorted out the options
11:28:14roolkumarkun: there is always scope for improvement
11:28:21jhMikeSmarkun: it only failed on some which were glitchy files
11:32:03linuxstbroolku: How does the preview work? Does it use the lcd_yuv_blit() function, or just draws to the framebuffer?
11:33:50*jhMikeS wants to see if anything is different for gigabeat vs. e200 with re: to those glitchy clips
11:34:30 Join webguest58 [0] (i=cba165d1@gateway/web/cgi-irc/
11:35:04 Part webguest58
11:35:13 Join spiorf [0] (
11:35:22 Join Entasis [0] (
11:35:26 Quit spiorf (Read error: 104 (Connection reset by peer))
11:35:59roolkulinuxstb: lcd_yuv_blit() as far as I can tell
11:36:44roolkuat least for colour targets
11:37:32roolkugray_ub_gray_bitmap_part for gray
11:37:44 Join helloworld1 [0] (n=pocketir@
11:39:44 Quit obo ("bye")
11:43:44jhMikeSit acts basically the same for those videos
11:45:24 Join spiorf [0] (
11:46:24jhMikeSfound a video where is fails to find the packet start code but the video is a good one
11:49:52roolkuif you have a link, I can forward it as a test case
11:49:58 Quit Siku ()
11:50:24jhMikeSroolku: see the PM?
11:51:33 Join Siku [0] (
11:57:12 Quit helloworld1 (Remote closed the connection)
12:00:06 Join barrywardell [0] (
12:01:43 Nick fxb__ is now known as fxb (
12:07:39 Quit tictoc (Remote closed the connection)
12:07:42peturpondlife: going to test now. that h300.hex was already tested by you I hope?
12:08:39pondlifeAh, not that particular one, but one built from the same source...
12:08:41 Join tictoc [0] (i=tabac@gateway/gpg-tor/key-0xB9002659)
12:09:03pondlifeI have a shell script to do the mucking about though, so no chance of typos
12:09:22peturboot 1 - ok
12:09:23pondlifeUses the Iriver EU 1.30 firmware FWIW.
12:09:49peturboot 2 too, I'll try a couple of times
12:10:04pondlifeSee if you can beat 50%
12:10:11Nico_Ppondlife: nice commit on playback.[ch] :)
12:10:15peturboot 3 hangs :(
12:10:34pondlifeSo, you still have an issue... maybe put some printfs in?
12:10:42pondlifeTiming I guess.
12:11:01*pondlife has a vested interest here, as he wants to get a bigger HD for his H300...
12:11:11 Join seablue [0] (
12:11:26pondlifeNico_P: Hope that didn't get in your way; I spotted it and thought it might be good...
12:11:47pondlifeYou were probably about to do the same anyway I expect.
12:12:00Nico_Pno, you did well... I hadn't seen that
12:12:21Nico_Pand git helped me fix the conflicts in no time :)
12:13:11Nico_PI got playback.c to compile on my mob branch, but haven't dared running it yet
12:13:24pondlifeScary stuff!
12:13:29Nico_Pit *will* break, the question is how bad
12:13:52*pondlife guesses "Codec failure"
12:13:57jhMikeSNico_P: I know the feeling about doing stuff that you can't even compile for a month that you know will then need meg-debugging. :P
12:14:11pondlifeSometimes it comes out nicely though...
12:14:18Nico_Pyeah, that's exactly it... and I feel like blind
12:14:26Nico_Ppondlife: not this time :p
12:14:40peturpondlife: is there also a very short pause between printing batt voltage and 'loading firmware'?
12:14:59pondlifeIs that new?
12:15:12pondlifeI thought it had always done that
12:15:16peturno idea, just wondering
12:15:39 Join ivan`` [0] (n=ivan`
12:15:53peturanyway, it's not 50% but 30-40% hangs so no change
12:16:28pondlifeThis is a disk problem, right? How well does USB bootloader mode work (when the disk spins up, I mean)?
12:16:30jhMikeSit sounds like the coldfire's are getting as touchy as pp5020 these days
12:16:32peturbah, now it is hanging every time :(
12:16:51pondlifejhMikeS: Nope, this is Petur's H300 only...
12:17:05pondlife(As far as we know.)
12:17:10jhMikeSwhat's so different about his?
12:17:15peturjhMikeS: I do have an 80GB disk in there
12:17:24pondlifeIt's an H380
12:17:25Nico_PjhMikeS: the sim is a good debugging target for playback stuff now that you've given it the same threading as on target, isn't it ?
12:17:25jhMikeSah, then it's the H380
12:17:43jhMikeSNico_P: I think it should be
12:17:50pondlifeNico_P: The sim is very useful for playback
12:18:56Nico_Psegfault !
12:19:10 Part pondlife ("disconnected has pondlife")
12:19:22peturpondlife: bootloader usb just hung too (message on display but not mounted) - but that is on this crappy dell monitor hub
12:21:29*petur reverts to v5 bootloader...
12:26:06*preglow comforts his own nice dell monitor hub
12:29:39*maxkelley pets his.. oh wait. I don't have a hub.
12:30:22 Quit ivan` (Read error: 110 (Connection timed out))
12:39:14amiconnI doubt that the 80GB hdd is a problem
12:39:43amiconnI have the very same hdd in my H1x0, and I don't have any boot problems or bootloader usb mode problems
12:39:55maxkelleywho makes the coldfire platform?
12:40:26 Quit roolku ()
12:41:06n1samiconn: does my latest commit fix the sim crashing for you?
12:41:45amiconnCan't test atm
12:44:09n1sok, it works nicely here now anyway :-)
12:45:36 Join kubiix [0] (
12:46:24 Quit bluebrother ("leaving")
12:48:12 Quit JdGordon (Read error: 104 (Connection reset by peer))
12:48:48 Quit barrywardell (Read error: 110 (Connection timed out))
12:50:20Lloreann1s: Having worked on the midi plugin, any thoughts on whether it'll be able to be made into a codec, or whether someone would be better off starting from scratch for that?
12:50:29 Join JdGordon [0] (
12:51:23linuxstbLlorean: I'm assuming it's memory requirements are the problem with regard to making it a codec. i.e. it has the same problem as other "non-streaming" formats, like MODs.
12:52:27n1sLlorean: I haven't looked at the patch loading code at all but I assume it loads all the needed ones into memory which as linuxstb says requires a lot more memory than the half-a-meg the codecs get
12:52:30Nico_Pmetadata displays in the WPS ! :)
12:52:58preglowthese codecs need special treatment from playback.c and codec.c
12:53:40Lloreann1s, linuxstb: So it would need to only load the necessary patches, and preferably keep them around for all buffered midi files?
12:53:44preglowbest solution (if you ask me) is to have one loader per codec that parses files and loads data as the codec wants to use it
12:53:57 Join barrywardell [0] (
12:54:03jhMikeSwhy is the darn patchfile so huge?
12:54:04preglowLlorean: i don't know how easy it'll be to keep them around
12:54:07Nico_PjhMikeS: I'm still far from getting audio to play though
12:54:19preglowNico_P: w00t!!!
12:54:21n1sLlorean: I think it only does load the needed ones but some individual patches are larger than 2 megs
12:54:37jhMikeSI'm just glad to hear that something's working
12:54:58 Quit GodEater_ (Read error: 110 (Connection timed out))
12:55:47Lloreanpreglow: So as an alternative, your idea is to keep the disk spinning and grab what's needed as it's needed if we can't fit it all?
12:56:19preglowLlorean: no, if we can't fit it all, the file is skipped, won't work
12:56:36n1sLlorean, preglow, maybe patches can be loaded as "metadata" to the buffer when mob gets int :-)
12:56:38preglowwe need to keep everything in memory, doing anything else is just too complex
12:56:43jhMikeSNico_P: another bad thing with huge code besides working on it for month and not being able to compile is when you do finally compile it, it has 1000s of compiler errors. :P
12:57:02preglown1s: but that would pretty much be a hackish way of doing what i'm suggesting anyway
12:57:10 Join Someguy08_idle [0] (
12:57:13Nico_PjhMikeS: that's what I was having... I did compile it nwo and then to know what to fix first
12:57:15n1spreglow: true
12:57:23 Nick Someguy08_idle is now known as Someguy (
12:57:33Nico_PjhMikeS: actually I focused on making it compiling at all costs to be able to follow the code flow
12:58:13 Nick Someguy is now known as ThatGuyAgain (
12:58:14n1sNico_P: is metadata guaranteed to stay in the buffer for the currently playing track with mob?
12:58:17markunfound a nice polish word: zrockboxowania
12:58:30Nico_Pnow I want to get things working little by little
12:58:31jhMikeSI do that if possible. With the recording stuff, it just wasn't at many times.
12:58:35Nico_Pn1s: yes
12:58:54n1sthen that idea could actually work :-)
12:59:17Nico_Pn1s: it needs to be moved sometimes, but it's never supposed to be accessed through a fixed pointer... you request it
12:59:19preglowbut please don't do it :/ the amount of work would be the same as for my idea, and my idea would facilitate other codecs doing stuff that way too
12:59:44preglowmod/xm needs samples too
12:59:47jhMikeSpreglow: codecs owning the buffer seems a reasonable way to go
13:00:00preglowjhMikeS: for everything but actual file buffering, yes
13:00:06n1spreglow: 'kay, would you care to explain your idea a bit more?
13:00:14*Nico_P needs to go make lunch
13:00:15preglowjhMikeS: having the file in the buffer as on disk helps you nothing if you want to play an xm
13:01:17jhMikeSpreglow: I think you had mentioned that before since it's compressed but primitively.
13:01:20preglown1s: basically this: when it comes time to load a file to the buffer, you call a codec-specific loader that parses and loads all data into the buffer the way the codec itself will later want to see it. for midi, you'd probably do little to the main playback sequence data, but you'd load some patches as well and stuff that in the buffer
13:01:25preglowjhMikeS: yes, delta coding
13:02:06preglown1s: for xm, you would need to decompress both pattern data and sample data into the buffer in structs that the codec itself can later use
13:02:25jhMikeSpreglow: are compressed xm very large?
13:02:40preglowi've seen xms above ten megs
13:02:46n1spreglow: sounds good, do you mean "codec-specific loader" as an external sort-of plugin?
13:02:49jhMikeSthat's workable here
13:03:04preglown1s: doesn't really matter, we should just put them in the core for starters
13:03:06linuxstbpreglow: Is the decompression process complicated? We wouldn't want to starve the currently playing codec of CPU time
13:03:12preglowlinuxstb: not at all
13:03:17jhMikeSthis will be a fun little threading excercise no doubt
13:03:24preglowlinuxstb: besides, loading should be a lower-priority process than playback
13:03:42***Saving seen data "./dancer.seen"
13:03:42preglowi'm really starting to see jhMikeS arguments on preemption here...
13:03:48linuxstbpreglow: Yes, but also loading time = disk spinning time, so we want to minimise it.
13:03:51*preglow envisioned careful placing of yield()s ...
13:04:34preglowlinuxstb: sure, but we have no choice anyway
13:04:50preglowif you want to load something, you have to load it while doing other stuff, no matter how long it takes
13:04:55preglowwe have decided to support the stuff
13:05:24preglowif we want codecs to access data as stored in files, we'll need to write them from scratch
13:05:32n1spreglow: we would also need to tell the buffering code what is where in the buffer somehow so we can access it later
13:05:46jhMikeSone issue with preemption on cf is saving those darn accumulators. save macsr, turn off rounding, save accums and accexts. when reloading in reverse order.
13:06:00preglowjhMikeS: how is that an issue?
13:06:16preglown1s: how and why?
13:06:23jhMikeSbecause now we nothing yields until it does movclr
13:06:49preglowjhMikeS: yes, sure, but why is it an issue? it should be straightforward :)
13:06:55jhMikeSARM is just a baby for that stuff though
13:07:05n1spreglow: how would the midi codec for example know where in the buffer a patch were?
13:07:13jhMikeSI guess there's so few calls to switch_thread anyway, you'd never notice it
13:07:35preglown1s: well, i assume the midi codec has a struct or array somewhere with patch addresses
13:07:49preglown1s: that just ends up in the file buffer
13:08:04preglowlots of codec internal state will just actually be placed in the file buffer
13:08:11jhMikeSI guess there would be more attempt to run another task, but if the same task is selected again, then nothing gets saved
13:08:41preglowjhMikeS: i sincerely doubt we'll notice it, no, none of that accumulator handling is overly slow
13:08:46 Join obo [0] (n=obo@rockbox/developer/obo)
13:09:08 Join freqmod [0] (
13:09:11preglowwe could also instate a rockbox policy on never using acc2 and acc3, saves three fetches/saves :>
13:09:19n1spreglow: I think I don't understand your idea fully...
13:09:22amiconnmidi as a codec causes a very special problem
13:09:40jhMikeSThere still would be no more actual context switching than now overall since most threads sleep nearly 100% of the time an take no/minimal CPU
13:10:25amiconnThe midi files are small, the patches are not - and the patches should probably not be loaded twice when playing several midis in sequence
13:10:25preglown1s: all arrays and stuff containing patches and note data in the midi codec for example, just gets loaded by the loader and put in the file buffer, the codec assigns its pointers to the places in the buffer that contains the data
13:10:41preglowamiconn: that'll be a special case i don't really know if it's worth handling
13:11:02preglowperhaps it would be worth making each codec be able of handling several files without unloading, that would handle it
13:11:33amiconnAfaik the codecs already work that way
13:11:39preglowyeah, that's very true
13:11:41preglowforgot that
13:11:56preglowthen we can handle that particular problem in a limited way
13:12:11preglowextending it to handle all cases of patch reuse would be very, very far from kiss
13:12:12amiconnBut midi is still special - it would still need access to old _data_, but _not only_ old data
13:12:13n1spreglow: ah, I thing I get it now, but as Nico_P said we can't rely on pointers being correct, but I guess we could just request the start adress of the data and calculate offsets
13:12:32preglown1s: i'm not talking about metadata
13:12:43preglown1s: for this kind of data, we _would_ be able to assume pointers are correct
13:12:52preglowthere'll be no relocating of this
13:12:55amiconn(unless we want to buffer all patches for a midi file together with that file)
13:13:05n1spreglow: all the simpler then :-)
13:13:47preglown1s: this is the simplest and best way i can see of doing this, but please think up alternatives :)
13:14:06n1spreglow: I'll think about it ;-)
13:14:15Lloreanamiconn: Or we could have codec metadata too, perhaps?
13:14:20preglowjust remember the fact that the file as stored on disk is of limited use to many codecs, they almost always need to be parsed into internal structures, that's the most important bits
13:14:33 Quit freqmod (Remote closed the connection)
13:14:40LloreanStore the patches as "metadata" or whatever you want to call it, for the codec, so they're released when the codec is?
13:15:55linuxstbLlorean: It's not quite the same - the codec is copied into the codec buffer. Leaving the patches in the main buffer will create fragmentation.
13:16:35preglowwhich we really don't want to handle
13:16:50preglowwhich is why i think we should only handle this in a limited fashion
13:16:59jhMikeSI thought about speculatively buffering files regardless of format. If it's skipped, no decompression takes place. If it plays and needs x amount of buffer to play, that amount is flushed including later files the were buffer on a "what if" and it's allowed to take it.
13:17:49preglowjhMikeS: how much time do you think we'd need to fix all the lazy thread coding in rockbox to work with preemption?
13:18:20preglowi don't really have a full grasp on how bad it is
13:19:05jhMikeShmmm...I've been pondering that...not sure. it's mostly a concern for low-level code but of course playback and a few other bits come in to that. If might not be that bad.
13:20:54jhMikeSI think factoring playback will be a huge help in determining that.
13:21:10preglowhaha, i believe you
13:22:30jhMikeSIt doesn't mean throwing mutexes around absolutely everything either. just come careful interface tweaking to some things.
13:24:38 Quit hcs ("Leaving.")
13:25:04 Quit Zagor ("Client exiting")
13:25:14jhMikeSAnd while we're at that, should we divide things into privileged and non-privileged code? I suppose that's optional.
13:25:21preglowwhy would we?
13:25:52preglowi get why we would ultimately do that, yes :) but i think that can wait for sure
13:26:06preglowbeing able to touch ports in all parts of the code with no bloody syscalls is nice
13:26:37jhMikeSswi is such a mess to get into supervisor modes?
13:26:48preglowi'd rather just avoid it
13:27:02jhMikeSwhat high-level code touches ports?
13:27:10preglowno idea
13:27:16preglowbut in what way would this help us far anyway?
13:27:25preglowbuggy code would still be buggy
13:28:12jhMikeSWell, it might not crash from user code as easily. Might be more trouble than it's worth. I haven't been itching to even do that.
13:28:28preglowi think it is more trouble than it's worth, we want buggy code to crash
13:28:44JdGordonthe debug screens are really he only apps layer code which shuold touch ports
13:28:49preglowpreemption might very well make us write better code, but code privileges would not
13:29:02jhMikeSpreglow: i guess you made my point about the threading back at me ;)
13:29:41preglowcoffee time
13:29:53jhMikeSa good thing too is that correct preemtive code is also mutiprocessor safe
13:30:30JdGordonor... we could fix the current code so its multi-proc safe and still have it less complex than preemtive?
13:31:20preglowthat would mean it's more complex than preemptive
13:31:42preglowsince we'd be relying on really fluid stuff. something that would be ok would very probably not be ok in two months time
13:32:12preglowthe only way of doing it properly so that we don't end up in the same old disappearing-maintainer-means-dead-code hell is doing it properly
13:32:30preglowanyway, codecs these days really, really shouldn't be afraid of writing correct multi-threaded code, it's more important than ever
13:32:45 Join atsea-34 [0] (i=atsea-@gateway/tor/x-2677151825aaa682)
13:33:11jhMikeSJdGordon: preemptive != complex...just careful
13:33:30JdGordonmore easy to break preemtive...
13:33:38jhMikeSif you make it multi-proc's also preemtive-safe
13:33:47amiconnpreemption will cause a lot more overhead
13:34:00preglowonly in thread communication
13:34:02preglownowhere else
13:34:13amiconnIn the scheduler as well
13:34:17preglowyielding will happen at more or less the same rate as before
13:34:31jhMikeSJdGordon: and so you won't have suprise race condition that pop-up when new yields appear where you were expecting them
13:34:32preglowthe thread context will become bigger, though
13:34:42amiconnPreemptive threading requires to save *all* registers.
13:34:46jhMikeSpreglow: not much
13:34:55amiconnThat's easily twice as many registers as cooperative
13:34:58preglowbut the current code is becoming too complex for the yield() maze
13:35:01JdGordonjhMikeS: ok, didnt tihnk of that, I guess that makes sense
13:35:45jhMikeSIt depends on _how_ the task switch occured. If it was forced through a function call, then only non-volatile context need be saved anyway.
13:36:15JdGordonso every function call checks if it should context switch?
13:36:26preglowshould we bother about that, though? i'd rather just keep it simple and always store everything
13:36:32amiconnThat sounds even more complicated...
13:36:33jhMikeScall into the kernel that would block a thread would force a switch
13:37:07jhMikeSin this setting, I might just stack the registers
13:37:21JdGordon.. which would eb the equiv of yield() anyway... we dont have that many kernel functions which would work for that
13:37:37jhMikeSJdGordon: no, it's not. yield doesn't block
13:38:19jhMikeSqueues, mutexes, sleep, etc. do/can block a thread.
13:38:22JdGordonoh, misread, you meant do the switch check on kernel threads which are expected to block?
13:38:46jhMikeSthere'd be no check. we'd stack a complete context only if the time quantum ran out
13:38:59jhMikeS...and if an different task was selected
13:39:09JdGordonstill... apart from disk access.. how much of the apps layer calls blocking functions?
13:39:14jhMikeSthe latter almost never happens
13:40:00jhMikeSapps layer would just call queue_wait or whatever the same as always. queues have sync built into them...they are a sync primitive.
13:41:31preglowyes, preemption is definitely starting to sound more and more right
13:42:27amiconnNot at all imo...
13:42:30JdGordoneven using the queue funcs to switch, still is easy to make a function which is going to eat the cpu and not call any kernel blocking functions, so you would need a yield anyway
13:42:45preglowjhMikeS: but how well do you think realtime threads like rtdsp would play with the user interface?
13:42:54 Join GodEater_ [0] (n=bryan@rockbox/staff/GodEater)
13:42:58preglowamiconn: what good arguments against are there?
13:43:25jhMikeSJdGordon: it wouldn't because it would be preempted and something else would run. if nothing else is ready to run, and it's not it eats CPU.
13:43:29preglowthe current deal isn't kiss. stuff crashes and hangs for reasons no one knows because of yield() placement
13:44:08*jhMikeS must be unusal in that threaded programming doesn't give him much trouble. ?
13:44:16preglowthe only excuse to not go preemptive is so we can continue writing bad thread communications code
13:44:25JdGordonjhMikeS: but how could it be preempted if the switching is done in the blockng kernel functions like you said?
13:44:27amiconnRight now you can easily make sure that a certain sequence is run without interruption by other threads. With preemption you can't
13:44:57*JdGordon has an idea!
13:45:01amiconnThat means a lot of code *will* break when introducing preemption
13:45:04preglowdoesn't critical sections do that?
13:45:17JdGordonamiconn: will you agree to code up viewports if jhMikeS agrees to drop this silly preemtion idea? :D
13:45:26jhMikeSJdGordon: when the time quantum runs out, the thread is interrupted and a new task selected. no yield needed. however, yield could still be used just to tell the kernel to try running something else.
13:46:30jhMikeSsome code will break, some will only need a bit of tweaking. nothing to cook up in a weekend for sure.
13:46:35JdGordoncan a thread be run on either cpu? or can it only run on the cpu its created on?
13:47:12jhMikeSJdGordon: well, if I get pp5020 sorted out, they can be switched in my patch.
13:47:29 Join Rondom [0] (
13:47:37preglownow, that i do not see a reason to introduce
13:47:52preglowwe will want to keep certain threads on certain cores since we'll be prioritising certain things
13:48:00preglowlike on ipods, it's pretty important the ui is responsive
13:48:02jhMikeSonly thing is, the thread just call switch_core(new_core)
13:48:21JdGordonyeah, but why have the COP sleeping when the CPU is switching between 3 busy threads?
13:48:52preglowthe rockbox kernel should be a bit better, but there's no reason for it to become a full-fledged server kernel either
13:49:14jhMikeSThe scheduler isn't as busy as some might think. During playback, expect about 60 switches/second. The audio thread sleeps then the buffer is full.
13:49:44preglowmost threads don't do shit
13:50:21jhMikeSnope. idle is about 10-12 switches per second to service all the threads on timeout. mostly that's the scroll thread.
13:51:51 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
13:51:56 Quit iamben (Read error: 104 (Connection reset by peer))
13:52:29jhMikeSthe single biggest cycle burner we do is busy yield loops and that can rack up millions of switches (actually calls to switch_thread that just select the same task again) during a disk read with a fast processor.
13:53:14jhMikeSEven SH runs them by the 100 thousand during playback buffer
13:54:18jhMikeSI must admit I got these measurement with a patched kernel where switch_thread is only about 50-60 instructions.
13:56:03 Quit barrywardell ()
13:58:57 Join donsdw [0] (
13:59:54jhMikeSThe same situation in any case
14:01:11preglowjhMikeS: but how well do you think realtime threads like rtdsp would play with the user interface?
14:01:38preglowrt threads are always troublesome, in my experience
14:01:41pregloweven when simple
14:03:11preglowrealtime dsp
14:03:36jhMikeSah, stupid me. "realtime threads" is really a priority consideration.
14:04:00jhMikeSIt wouldn't need to starve away others to be timely
14:04:38preglowyeah, i guess it wouldn't need to be real realtime
14:04:46preglowwouldn't be the end of the world if it skipped a buffer during intense load
14:05:05preglowbut ui not being interrupted too much is really the only important one
14:05:07jhMikeSI think the right setup could actually improve the situation since priority could be finely adjusted
14:05:10pregloweverything else can be starved to hell
14:06:21 Join Benoitb [0] (
14:06:51jhMikeSI think what's more important is latency where a time-critical task won't run becuase some slow thing doesn't yield
14:06:52JdGordonthe ui shuoldnt have a problem as long as it gets enough time when moving in the lists
14:07:09GodEater_how would you do the prioritisation? Something similar to a "nice" implementation ?
14:07:18BenoitbHi ! I've got a problem on a Gigabeat F40, I compiled the latest SVN and my gigabeat doesn't boot after I turn it off unless I switch the battery off and on
14:08:10preglowGodEater_: still no waking?
14:08:13jhMikeSGodEater_: I guess there's alot of ways. The distribution needs to have maximally high frequency's sort of a dithering thing.
14:08:29BenoitbI had this problem with a recent custom build so I wanted to check with the official SVN without any patch and it's the same problem, so it is in the core of rockbox
14:09:22GodEater_preglow: nope - still no waking
14:09:34LloreanBenoitb: Did the build you were using have the quick startup patch?
14:09:54LloreanBenoitb: There have been a few patches in the past for the gigabeat that cause problems with booting, even after you've removed the build.
14:10:02 Join CaptainSquid [0] (
14:10:14 Join barrywardell [0] (
14:10:14Benoitboh, I see, how can I get it clean and new ?
14:10:20LloreanNo clue.
14:10:24preglowGodEater_: good, sure i solved it, then
14:10:29BenoitbI'll check if those builds had that patch
14:10:29GodEater_yep :)
14:10:32jhMikeSI guess it's similar in goal to what's attemped when framedropping in mpegplayer or so many other applications.
14:11:48 Join Thundercloud [0] (
14:12:11 Quit Arathis ("Bye, bye")
14:13:19BenoitbLlorean : I tried the EvilG's build and it sas it uses the FS #7113 - Gigabeat F really fast restart
14:13:34Benoitbso maybe it is the cause of the mess
14:13:43BenoitbI'll check the page of this patch
14:14:13*jhMikeS needs to find out if any cache service interrupts can be generated on PP targets since that could help automate coherency.
14:16:38BenoitbI see you advise to do a real power cycle, which I think I did. Could it be that I have to wait for a certain amount of time when the battery switch is off before I switch it to on again ?
14:20:21 Join eigma [0] (
14:20:29JdGordonhey eigma
14:20:35 Quit elinenbe (Read error: 104 (Connection reset by peer))
14:20:41 Join elinenbe [0] (
14:21:06JdGordonfigure out what the 3rd spi target was?
14:21:12eigmaI'm pretty sure it's the BL
14:21:44eigmaI haven't been able to find a datasheet though
14:22:06eigmaI can probably fish out the commands from the OF, but it would still be nice to have the DS
14:23:29eigmaI gotta go to school.. tty
14:23:37 Quit eigma (Client Quit)
14:28:47 Join ThUnD3r|Gr33n|Aw [0] (
14:28:51ThUnD3r|Gr33n|Awservus ...
14:28:59BenoitbI tried to leave my Gigabeat unpowered for 10 minutes but it didn't solve the problem, it doesn't want to boot now that I've powered it down
14:29:07ThUnD3r|Gr33n|Awcan somebody give me some help?? i dont unterstand how to install a patch :P
14:30:02Benoitbdid you try the wiki page "simple guide to compiling"
14:30:17Benoitbit seems very understandable
14:31:04 Join norbusan [0] (
14:31:07 Part norbusan
14:33:08 Quit Rondom ("Ex-Chat")
14:33:09markunBenoitb: did you try the battery switch?
14:33:32markunah, that's what you mean by "unpowered" of course..
14:33:44Benoitbmarkun: yes I did, it's the small off/onn switch at the bottom, right ?
14:33:59markuncan you still charge your player?
14:34:04Benoitbyes I can
14:34:11LloreanBenoitb: And the power adapter and USB cable were not attached when you did this?
14:34:16Benoitband it powers up if I switch it back and forth
14:34:28BenoitbLlorean: no
14:34:55Benoitbthe problem is that I do have to switch it each time I want to power it up
14:35:12Benoitbinstead of just pressing the standard power button
14:35:26*amiconn wonders whether markun actually tested the new codepage on charcell (sim) ...
14:35:31Benoitbas if it was left in a bad state when it is shut down
14:35:37markunamiconn: I didn't
14:36:06markunamiconn: does it work after your commit?
14:36:33amiconnBut I tend to think that this combined iso.cp / isomini.cp is somewhat messy
14:37:04amiconnExtending it can easily go wrong, as it relies on the order
14:38:04amiconnI think it would be better to have one file per codepage, even though that means to have a few more files around
14:42:49 Join kugel [0] (
14:43:12jhMikeSsomething is very, very wrong on pp5020 since what basically amounts to a rearrangment (file shuffling) an no difference in what particular things are executed on startup, causes hanging
14:49:40 Quit BjoernErik (Connection timed out)
14:50:13ThUnD3r|Gr33n|Aw/!\ Benoitb /!\
14:50:24ThUnD3r|Gr33n|Awsorry i was away from keyboard ...
14:50:34ThUnD3r|Gr33n|Awi should compile the files ??
14:51:04Benoitbif you want to apply the patch, I'm affraid you have to ;-)
14:51:07 Nick parafin|away is now known as parafin (i=parafin@
14:53:06ThUnD3r|Gr33n|Awyes .. i have to "install" a patch for mpeg player
14:53:13 Join FOAD [0] (
14:53:25 Join jgarvey [0] (
14:53:46 Join CaptainSquid83 [0] (
14:53:47jhMikeSman, pp5020 loves to freeze it you happen to touch certain registers before a device enable
14:56:35ThUnD3r|Gr33n|Awo thanks .. mom .. i'll red a little bit ;)
14:59:16ThUnD3r|Gr33n|Awok .. i dont understand one word _P
15:00:07*jhMikeS hereby christens pp5020 the playback.c of hardware
15:00:17 Quit CaptainSquid (Remote closed the connection)
15:00:27JdGordonoh dear!
15:00:41 Join moos [0] (
15:01:54ThUnD3r|Gr33n|Awoke .. i try to make my way ;)
15:02:27jhMikeSNo other chip I know of is so intolerant. The others tend to kindly ignore certain things.
15:03:42 Join Wiwie [0] (
15:03:46amiconnEver tried reading from nonexisting addresses on cf? Much fun to debug...
15:03:48***Saving seen data "./dancer.seen"
15:04:18 Quit Wiwie (Client Quit)
15:04:24ThUnD3r|Gr33n|Awhow to download this patch ??
15:04:25jhMikeSthose have happened for sure many times but I was able to get to it faster than that one
15:05:16BenoitbThUnD3r|Gr33n|Aw : if you don't know how to compile and don't want to read the docs, you can try an unsupported build that includes the features you want
15:05:25Benoitbyou cann find it in the forums
15:05:48 Part Benoitb ("Kopete 0.12.5 :")
15:05:57ThUnD3r|Gr33n|Awwhere can i find it ?? sorry for these f*** questions .. im new here .. but i want to learn more about this all ;)
15:06:03jhMikeSmaybe a careful going-over needs doing to make sure it's not happening elsewhere
15:06:37amiconnBagder/ LinusN: around?
15:07:11amiconnIt looks like the build system doesn't build sims on 64 bit hosts (??)
15:07:30amiconnThere is a cast warning in font.c that didn't happen to show up since...
15:08:03amiconn... almost 2 years ago
15:08:23LinusNdon't you have a 64-bit build server?
15:08:35amiconnI do
15:08:52LinusNdoes it build sims?
15:09:06 Quit CaptainSquid83 ("Miranda IM!")
15:09:38 Join petur2 [0] (n=petur@rockbox/developer/petur)
15:09:41JdGordonamiconn: there is no warnings in font.c on my amd64 comp...
15:09:41 Quit petur (Nick collision from services.)
15:09:43 Nick petur2 is now known as petur (n=petur@rockbox/developer/petur)
15:10:20JdGordonunless its target dependant?
15:10:32LinusNamiconn: your server is set up to only build sh, arm and 68k target builds
15:10:41amiconnIn fact any bitmap sim should trigger it
15:10:41JdGordonamiconn: any sim?
15:10:50JdGordone200-sim doesnt trigger it
15:10:53amiconn(i.e. anything but Player)
15:10:59amiconnfont.c line 309
15:11:33amiconnLinusN: Why is that? Sim would be even better on slow-upstream connections
15:11:58LinusNamiconn: because you haven't told us that your server can build sims?
15:12:11JdGordonI thought you coundlt configure it to only build sims?
15:12:13amiconnMy box used to build sims
15:12:28amiconnThere are several sim build logs from may
15:13:04LinusNok, i'll add sim to your server's capabilities
15:13:47amiconnMaybe *someone* didn't want to see all those 64 bit warnings?
15:14:08LinusNback when 64-bit was still pretty uncommon
15:15:06JdGordon*couph* add me to only do sims...
15:15:34 Join Bjoern-Erik [0] (
15:15:38JdGordonactually... dont.. cbf setting it up atm :p
15:25:17 Quit Llorean ("Leaving.")
15:25:48 Join Llorean [0] (
15:28:40 Join rvvs89_ [0] (n=rvvs89@unaffiliated/rvvs89)
15:31:31 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
15:39:01 Part LinusN
15:43:54TMMwhat's 'sim'?
15:44:15TMMI never would have guessed :P
15:44:22TMMa sim for players? what is it, why does it need servers? :)
15:44:36 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
15:45:16n1sTMM: the servers build both regular builds and simulator builds that simulate rockbox on a computer
15:45:25 Join webguest83 [0] (i=43a8cc97@gateway/web/cgi-irc/
15:45:41TMMwell, I can do some number crunching soon, my new server should arrive in about 2 weeks
15:45:57 Join Thundercloud_ [0] (
15:46:02TMMhow much load will such a thing generate?
15:46:49webguest83hola nerds
15:47:01n1sthere are usually about 5-20 builds in a day, they take about 5-10 minutes
15:47:04pondlifeTMM: Depends on how often people change the code
15:47:20webguest83my i ask a question
15:47:22Lloreanpondlife: And how much code changes, right?
15:47:28TMMit'll be a dual core AMD 5600+ 4Gb of mem, and about 5Tb of harddisk space
15:47:30Lloreanwebguest83: See the channel guidelines. :-P
15:47:45webguest83Ok basics ipod 5g 30g
15:47:53webguest83 how to i get themes on it because i know you have to unzip them onto
15:48:02webguest83my ipod but i dont know how
15:48:05Lloreanwebguest83: Please, please read the channel guidelines.
15:48:56n1sTMM: That'l do nicely :-) talk to Bagder about setting it up
15:49:19 Quit kubiix ("Miranda IM! Smaller, Faster, Easier.")
15:49:28webguest83ok i read the guidlines'
15:49:35webguest83but can anyone help me
15:50:40Lloreanwebguest83: Well, you might get more help if you follow them. For example, the one about not posting multiple short lines in a row...
15:50:48 Join agm3nt [0] (
15:50:53LloreanA properly formatted theme zip just needs to be extracted straight to the iPod, just like a build.
15:52:23 Quit webguest83 ("CGI:IRC (EOF)")
15:52:37 Join webguest84 [0] (i=43a8cc97@gateway/web/cgi-irc/
15:53:24webguest84can anyone help me with rockbox theams
15:54:07TMMn1s: will do
15:54:19webguest84ok thank you soo much
15:54:25Lloreanwebguest84: Please attempt to use proper spelling, as is asked in the guidelines. And as I said, a properly formatted theme can simply be extracted straight to the player, just like a build.
15:54:34*petur wonders if webguest84 likes spam
15:54:57|Rain|everyone likes spam! that's why the spammers pass out enough for everyone!
15:55:08webguest84so i can go to the Ipod harddrive and find a folder where the theams are?
15:55:36Lloreanwebguest84: "Themes", you demonstrated that you knew how to spell this earlier, and I'm sure you've seen the word a dozen times while reading the manual.
15:55:44 Quit animeloe ("This computer has gone to sleep")
15:56:08webguest84 how to i get Themes on it because i know you have to unzip them onto
15:56:10Lloreanwebguest84: You shouldn't need to find a folder, just unzip straight to the drive and if the theme .zip file is formatted properly it should end up with the files in the right folders.
15:56:45 Quit Thundercloud (Read error: 110 (Connection timed out))
15:56:45webguest84ohh well i got it off the web site and its in zip format?
15:56:54webguest84by the way i have a Mac
15:57:17kugelI love those webguests
15:57:29Lloreanwebguest84: What part of "unzip it straight to the drive" is confusing?
15:57:36LloreanI know it's a zip, that's why I said to unzip it...
15:58:08webguest84ok got ya and when i do that i get a folder. Wait what program should i use to unzip
15:58:49|Rain|my inetrnet masheen said saved file but rockbox theme not in list???/
15:59:08Lloreanwebguest84: Did you unzip straight to the iPod?
15:59:08 Quit webguest84 ("CGI:IRC (EOF)")
15:59:12petursomebody should sue apple for making available computers that are easy to use :/
15:59:15 Join webguest85 [0] (i=43a8cc97@gateway/web/cgi-irc/
15:59:18 Quit Zagor ("Client exiting")
15:59:19webguest85my bad
15:59:26Lloreanwebguest85: Did you unzip straight to the iPod?
15:59:37webguest85well i dont konw how i would do that
15:59:47LloreanWhen picking where to unzip to, pick your iPod...
15:59:51webguest85when i downlode it i get a folder and thats where im stuck
16:00:01kugelHe doesn't know how to unzip
16:00:06webguest85ok and what program will let me do that for the mac'
16:00:12|Rain|who installed rockbox on your ipod for you?
16:00:12webguest85yea im a nub
16:00:14Lloreanwebguest85: How did you unzip the Rockbox build?
16:00:18*GodEater_ wonders when this became a mac support channel
16:00:24|Rain|how did you do that? it came in a zip file...
16:00:27|Rain|unless there's a mac installer as well
16:00:37webguest85there is
16:00:40 Join desowin [0] (
16:00:44Lloreanwebguest85: It's really not our job to explain basic computer operations... I don't use a mac, so you'll need to find an unzip program.
16:00:52peturcurse the creators of rbutil ;)
16:01:01webguest85yep thats it
16:01:07*GodEater_ thought rbutil did theme install too
16:01:08webguest85well i could do it anyways
16:01:15peturrbutil can also install themes
16:01:17 Join midgey|web [0] (i=8dd3a84d@gateway/web/cgi-irc/
16:01:21webguest85but you can get better ones
16:01:37peturwell, there's a good reason to learn your computer better ;)
16:01:41*Llorean bets these "better" ones require a custom build... which he also can't install.
16:01:58peturlearn how to unzip and the reward is a nicer theme
16:02:07GodEater_safe bet I'd say Llorean
16:02:28webguest85i do know how but i wont let me select my Ipod
16:02:43midgey|webwebguest85: open up a terminal, (/Applications/Utilities/
16:03:03midgey|webnavigate to the folder containing the theme zip file
16:03:18GodEater_lost him at "navigate" I'd imagnie
16:03:26midgey|webfor example, your desktop (cd ~/Desktop)
16:03:27kugelI doubt he knows where to type that
16:04:12kugelWho said Windows is for happy users?
16:05:16webguest85sry only 15
16:05:29rvvs89_How does that make any difference?..
16:05:40GodEater_<−− programming since the age of 7
16:05:44webguest85 confused
16:07:10|Rain|you will get slightly more respect if you take the time to learn how to set your nickname, more still if you figure out how to use a proper IRC client, and even more still if you figure out how to follow directions from the people trying to help you :P
16:07:35webguest85lol ok
16:08:27 Quit hannesd_ (Read error: 104 (Connection reset by peer))
16:09:19midgey|webmac os x will automatically unzip a zip to the folder you're currently in, since the files extracted start with a dot (e.g. /.rockbox) they won't be shown
16:09:29webguest85well i think i deleated the root folder and i have to go to school but im gona reinstal rockbox and try to unzip it to my ipod...
16:09:47midgey|webim pretty sure the only way to specify an output location is to use unzip through the terminal
16:09:52webguest85ok so thats wahat the blank files are
16:10:04webguest85 yea but i dont know how to do that
16:10:09webguest85ill just google it
16:10:31 Join bluebrother [0] (i=UiLDVLcJ@rockbox/staff/bluebrother)
16:10:41webguest85well ok friends.. i gtg il get a actual name next time like *Nubprogramer*
16:10:46|Rain|open the finder, go to applications, go to utilities, and go to terminal, just like midget suggested earlier
16:10:59webguest85i know how to open it duh
16:11:12webguest85gtg school......
16:11:19webguest85but im not loging out
16:12:03kugelHe is not logging out
16:12:15 Join scorche|work [0] (n=8dc5049d@rockbox/administrator/scorche)
16:12:31n1sthat won't stop the webirc from committing suicide anyway...
16:12:35kugelImagine when no webguest would log out
16:13:02bluebrotherjust wait for the webserver to get restarted then ;-)
16:15:54 Join donsdx [0] (
16:16:21*GodEater_ gives up looking for the "infinity" symbol in his charmap to run his webguest joke
16:16:39|Rain|221E ∞ INFINITY
16:16:48*|Rain| whistles
16:16:51 Quit webguest85 ("CGI:IRC (Ping timeout)")
16:16:52GodEater_which section is it in ?
16:16:58GodEater_ah - there he goes anyway
16:17:00|Rain|not sure, I just grepped it out of utflist.long
16:17:23*petur hugs ping
16:18:18GodEater_so anyway, our cgi:irc client goes up to "webguest∞"
16:18:22GodEater_lame late gag
16:18:32GodEater_|Rain|: it's in common anyway
16:30:42 Join midkay [0] (n=midkay@rockbox/developer/midkay)
16:32:11TMMGodEater_: what's so funny?
16:32:19TMMGodEater_: I don't get it :)
16:34:25bluebrotherTMM: just increase the number after webguest more and more ... ;-)
16:34:45TMMeuh... you can't
16:35:06TMMwhat's that wooshing sound?
16:35:10TMMjoke going over my head perhaps?
16:36:26GodEater_TMM: The rockbox webserver has ram going up to 11
16:36:50GodEater_bit like Spinal Tap's amps
16:36:52GodEater_it can do anything
16:37:44 Quit Entasis (Read error: 104 (Connection reset by peer))
16:37:50|Rain|...why don't you just make 10 a little bit louder?
16:38:14TMMGodEater_: I don't get THAT either
16:38:25GodEater_TMM: do you even know who Spinal Tap are ?
16:38:53TMMGodEater_: euh... no
16:38:56*bluebrother just found a wikipedia article about Spinal Tap
16:39:28|Rain|copyrightviolationtube will help you out:
16:40:07GodEater_TMM: best read up or watch the video - otherwise it won't make any sense
16:40:37|Rain|hopefully that's the right one
16:40:38 Quit midgey|web ("CGI:IRC (EOF)")
16:40:44|Rain|ah, yep.
16:42:55|Rain|you should probably read up AND watch the video (or watch the entire spinal tap movie), 'cause it still won't make MUCH sense (not that it'll make much more if you watch the movie)
16:43:34GodEater_not unless you're smoking something ;)
16:44:20GodEater_heh -it's been so long since I watched that clip
16:44:32GodEater_"it's one louder than ten"
16:46:22delYsidwhat does "red" stand for?
16:46:57GodEater_where ?
16:47:00GodEater_in what context ?
16:47:14delYsidsvn changelogs
16:47:42GodEater_fixed a red build
16:47:46GodEater_i.e. one that didn't work
16:48:13delYsidoh, this refers to some visual build-type overview?
16:48:36GodEater_yes, the build table
16:48:44delYsidi see, thanks.
16:48:49GodEater_which is desired to be green
16:49:01GodEater_sometimes features yellow (a build with warnings)
16:49:07GodEater_and occasionally red (didn't compile)
16:49:09LloreandelYsid: "Yellow" means compiler warnings, and "Red" means that the build failed either due to compiler or linking errors.
16:49:15*Llorean was too slow
16:54:39pondlifeOften features yellow, sadly...
16:55:22 Join miepchen^schlaf [0] (
16:59:24 Nick sbeh is now known as gentoo (
16:59:34 Nick gentoo is now known as sbeh (
17:00:04amiconnpondlife: Only due to a gcc bug atm
17:03:50***Saving seen data "./dancer.seen"
17:09:56 Join kubiix [0] (
17:09:57delYsidThe substitution of "plugins" to "plug-ins" breaks the pronounciation for the swift engine.
17:10:05delYsidWithout this change, it worked right.
17:13:52*JdGordon looks for volanteers to fix up the manual for FS #7864 so it can be commited
17:14:03 Quit JdGordon ("Konversation terminated!")
17:16:23 Quit agm3nt ()
17:16:44 Quit donsdx ()
17:17:41 Quit donsdw ()
17:18:32Nico_Pis there a way I can add newlines to all logf calls withouth having to change all the calls ? (ie in the macro)
17:19:15bluebrotherchange the logf macro?
17:19:33Nico_Pyeah, but how ? I want to add the newline in the string it gets
17:20:38bluebrotherthe easy way would be to add a printf call after the actual output
17:20:53 Join donsdw [0] (
17:22:10Nico_Pbluebrother: thing is the logf macro is just a substitution
17:22:30Nico_P#define logf _logf, where _logf is void _logf(const char *format, ...) ATTRIBUTE_PRINTF(1, 2);
17:22:52Nico_Pamiconn: any ideas ?
17:23:20pondlifeNico_P: Why not just put the \n in the logf call, like it's meant to be?
17:23:37Nico_Ppondlife: logf calls aren't supposed to have \n
17:23:59pondlifeAh, I'm probably thinking of DEBUGF
17:24:01Nico_Pand don't... what I have is #define logf DEBUGF to print all logf messages to the console
17:24:13Nico_Pbut they print without a newline, which isn't easy to read
17:24:28pondlifeI have a patch to put logfs to the console
17:24:33 Join NonExistenT [0] (n=non3x@
17:24:45 Part NonExistenT ("leaving")
17:24:46Nico_Ppondlife: what does it do ?
17:24:53dionoeaNico_P: #undef logf
17:25:01pondlifeCan't remember! Will look
17:25:16dionoea#define logf(f,args...) _logf(f"\n",##args)
17:26:38pondlifeNico_P: might help
17:27:03Nico_Pdionoea: I replaced _logf with DEBUGF (because that's what I want), but it doesn't work
17:27:45dionoeawhat doesn't work ?
17:28:04Nico_Pdionoea: #define logf(f,args...) DEBUGF(f"\n",##args) doesn't compile
17:28:25Nico_Ppondlife: yeah that would work but requires a logf build...
17:28:28dionoeadid you undef the logf first ?
17:28:42Nico_Pdionoea: it's not defined because I'm not in a logf build
17:28:49 Quit TMM (Read error: 110 (Connection timed out))
17:28:52dionoeawell what error does it print ?
17:28:55pondlifeWhy not use a logf+debug build?
17:29:04Nico_Pactually originally it's #define logf(...)
17:29:27Nico_Pdionoea: I got the following:
17:29:30Nico_Pcommon/dircache.c: In function ‘dircache_new_entry’:
17:29:30Nico_Pcommon/dircache.c:870: error: expected ‘)’ before string constant
17:30:02Nico_Pobvisouly dircache.c:870 has a logf call
17:30:28dionoeacan you paste that line ?
17:30:41Nico_Pah, but in that call there's no string: logf(basedir);
17:30:53dionoeaah, that would be problematic
17:31:10dionoeachange it to logf("%s",basedir)
17:31:22pondlifeIt should be changed anyway
17:31:37pondlifeIn case there's a dir named %d !
17:31:57Nico_Pdionoea: now it works beautifully, thanks :)
17:32:03dionoeayou're welcome
17:32:04 Quit jhMikeS (Nick collision from services.)
17:32:10 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
17:32:32jhMikeSargh, some twisted-ass illegal thing in pcmbuf.c just cost me hours of hell
17:32:54bluebrotherNico_P: sorry, was distracted a couple of minutes
17:33:00pondlifejhMikeS: Time well spent now you know what it was, no?
17:33:13Nico_Pbluebrother: no problem, dionoea gave me the solution :)
17:33:31jhMikeSthe call to pcmbuf_flush_fillpos in the interrupt handler
17:33:48jhMikeSnow my H10 is playing fine with revamped pcm drivers
17:33:57pondlifeLow latency?
17:34:02bluebrotherI would have changed the _logf function itself ;-)
17:34:23Nico_Pbluebrother: I don't want to have to enable logf
17:34:25jhMikeSno, the api revamp that simplifies and refactors it
17:34:32Nico_Pbut that was the simple way ;)
17:34:33 Join Poro [0] (i=551d4ff8@gateway/web/cgi-irc/
17:34:42pondlifejhMikeS: Sounds good to me...
17:34:49PoroRockbox sucks
17:34:53PoroRockbox sucks
17:34:53DBUGEnqueued KICK Poro
17:34:53PoroRockbox sucks
17:34:53PoroRockbox sucks
17:34:55***Alert Mode level 1
17:34:55PoroRockbox sucks
17:34:56***Alert Mode level 2
17:34:56PoroRockbox sucks
17:34:56***Alert Mode level 3
17:34:56PoroRockbox sucks
17:34:57***Alert Mode level 4
17:34:57***Alert Mode level 5
17:34:57PoroRockbox sucks
17:34:59bluebrotherPoro: we know.
17:35:12pondlifeSometimes it blows instead
17:35:14 Join hannesd_ [0] (
17:35:15jhMikeSPoro:'s a bag of shit
17:35:15 Quit Poro (Client Quit)
17:35:15bluebrotherand that's exactly what we like abou tit.
17:35:15LloreanIt's actually got a dual mode function that allows it to blow if you so wish.
17:35:26GodEater_oh he left
17:35:30pondlifejhMikeS: Wronger, it SOUNDS like a bag of shit
17:35:42Nico_P..compared to OF
17:35:44jhMikeSah, right :)
17:35:44bluebrotherwait, where is that blow function? Too bad he left.
17:35:52pondlifeIt's not a bag of shit, just an auditory simile of one.
17:36:03CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
17:36:03*jhMikeS wants to be able to +o and /kickban
17:36:35bluebrotherthe ability to devoice would be nice ;-)
17:36:49jhMikeSas opposed to a metaphorical one?
17:36:51 Join low_light [0] (i=c730180b@gateway/web/cgi-irc/
17:37:07*jhMikeS inclined to use force preemptively :)
17:38:04low_lightjhMikeS: does the e200 use the COP_INT's for anything?
17:38:37scorche|workbluebrother: +b %*!*@* is a mute
17:38:38jhMikeSlow_light: in OF?
17:38:56bluebrotherscorche|work: I guess only ops can do this?
17:38:58scorche|workand sorry...i was out working...wasnt here to kick
17:39:34 Join tikakan [0] (
17:39:57scorche|workyes....onloy ops can
17:40:52 Join Nick_Brackley [0] (
17:41:15tikakanjust a quick question: is it possible to upgrade a 1. gen ipod with a bigger HD? (I tried it 2 years ago with the original Apple FW, but it didnt work.)
17:43:15jhMikeSlow_light: ?? in OF it doesn't at least not up to the menu menu
17:44:09low_lightjhMikeS: the c200 OF uses some COP_INT's and I want to see if they are related to the microsd insertion
17:44:50jhMikeSCOP_INT's can be enabled on CPU. Don't you know what line goes low already?
17:44:59***Alert Mode OFF
17:45:48GodEater_tikakan: do you know the drive interface is /
17:46:40low_lightjhMikeS: I know which GPIOL_INPUT_VAL but switches on insertion, but I can't get the interrupts to work
17:46:45tikakanGodEater: its IDE, isnt it? Should work at least to 32 GB if its one of the very old controllers...
17:47:10pixelmamaybe this page helps...
17:47:28GodEater_tikakan: I've no idea - that's why I'm asking you :)
17:48:34tikakanGodeater: I know its IDE, I got the replacement HD in an external USB box working with an IDe interface.
17:48:39jhMikeSlow_light: does the sim enable that GPIO interrupt too? it won't set the correct bit CPU_HI_INT_EN thinks it running on something else and only uses bit 32
17:48:41tikakanThx, pixelma, gonna check ^^
17:49:45 Quit petur ("sporting time")
17:50:43low_lightjhMikeS: the sim doesn't get very far in the OF before sleeping
17:51:18jhMikeSlow_light: what do you mean "sleeping"?
17:51:51low_lightthe cpu is put to sleep
17:52:13low_lightCPU_CTL = 0x80000000
17:52:38jhMikeShmmm...perhaps the emu is in error and doesn't wake the cpu if interrupts are disabled?
17:53:15low_lightI can only run the emu with the cpu...if I remember, the cop gets stuck on a SWI instruction (are those emulated)?
17:54:04jhMikeSI think the whole ARM core is emulated but I can't say if everything is handled correctly.
17:56:07 Join kugel_ [0] (
17:57:23 Quit kugel (Nick collision from services.)
17:57:33 Nick kugel_ is now known as kugel (
17:58:10low_lightthere are several places where there's a loop around "SWI 0x123456" and it won't break. I don't think the emu was branching to the interrupt handler
17:58:53jhMikeSwhat is that? on h10 it doesn't seem to point to thumb code
18:00:02jhMikeSI'm not even clear on whether SWI actually uses that number or not
18:02:45kugelwhat means rwps?
18:02:47*jhMikeS ought to just write some code to find out and forget the ambiguous arm doc
18:03:53bluebrotherremote wps
18:06:50low_lightjhMikeS: arm doc says "This value is ignored by the processor, but can be used by an operating system's SWI exception handler to determine which operating system service is being requested"
18:07:34jhMikeSso I suppose you have to read the instruction at pc and check the 24-bit value?
18:08:37jhMikeSat lr - #n
18:09:52 Join Arathis [0] (
18:09:55low_lightpage 200 of the arm ref manual
18:11:02Lloreanbluebrother: Actually, as he's using a Nano, it may be that it's Rockbox's fault for the filesystem errors.
18:11:19bluebrotherhow that?
18:11:21LloreanLooking at the various reports on "the nano bug" it looks like one of the symptoms may be fs corruption.
18:11:45LloreanData read into memory, like buffered audio, is read corrupted, so it's not too much a stretch to assume writing may be corrupted as well.
18:12:12bluebrotherhmm. I'm not quite aware of that bug −− is it in FS?
18:12:23rasherIs it ever..
18:12:23LloreanYeah, search for "Nano 1.3.1" I think.
18:12:33rasherFS #7510
18:12:40bluebrotherah, this "playback broken since ..." bug?
18:13:09|Rain|are there any known issues wrt tracks ending up in the database multiple times? I just reinitialized last night, but before I did I had 4 entries for every track
18:13:33 Quit rvvs89_ (Read error: 110 (Connection timed out))
18:13:37bluebrotherremoving tracks from the database only works with dircache enabled.
18:13:58bluebrotherand you might need to load the database to RAM, though I'm not 100% sure about that.
18:14:32 Quit Seed ("cu, Andre")
18:14:51jhMikeSlow_light: something seemed odd about some later notes though. I guess #0x123456 is just filler. perhaps emu not calling the handler properly. I don't think the emu knows about the remapped vectors either.
18:14:58 Join Seed [0] (
18:15:44jhMikeSlow_light: does the the ouput show if bit 4 is set by the OF in CACHE_CTL?
18:17:31jhMikeSand I know the bl sets it and it's cleared initially but if it's set later
18:19:49low_lightno I don't see that it's set
18:20:22 Quit tikakan ("Leaving")
18:20:39 Join MarkFawcus [0] (i=18ec7c6f@gateway/web/cgi-irc/
18:21:08low_lightI don't see that the bl sets it either
18:21:28MarkFawcusWho do I need to talk to about getting Wiki write permission?
18:21:58jhMikeShmmm...this one's different then. e200 bl does and sets remapped vectors for the switchover.
18:22:58jhMikeSe200 and H10 fw are almost identical in the basic routines
18:23:15rasherMarkFawcus: done
18:23:17low_lightmy bit 4 you mean the 5th bit ;) gets set
18:23:25rasherMarkFawcus: assuming your wiki name is MarkFawcus
18:23:37MarkFawcusrasher: thanks.. it is
18:24:38 Quit amiconn (Nick collision from services.)
18:24:44 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
18:24:48jhMikeSlow_light: but e200 and H10 clear it as the first register access and never seem to set it again anytime soon.
18:25:19 Quit MarkFawcus (Client Quit)
18:26:47low_lightjhMikeS: here's right before the switch
18:29:39*Nico_P doesn't understand why the codec doesn't start decoding
18:29:57Nico_Pit loads from disk then disappears
18:31:16jhMikeSinteresting value for CPU_CTL ... and the MBOX8 thing. I also see it clears that first.
18:31:28delYsidrasher: r14758 breaks swift's pronounciation of "plguins". Can you conditionalize your change on $$tts_object{"name"} != "swift" please?
18:32:00rasherdelYsid: I think I'll just make it specific to festival for now
18:32:10jhMikeSplguins? link penguins?
18:32:25 Quit obo ("bye")
18:32:26delYsidrasher: thats fine too, thanks.
18:33:05delYsidswift now says "plug-ines" which sounds very strange :-)
18:33:50jhMikeSI suppose this was to fix "ploogins"?
18:34:47jhMikeSmade it sound scottish
18:36:35markunrasher: does espeak still pronounce it correctly?
18:37:28rashermarkun: I honestly don't know. I think I'll leave it to someone who does. The only thing I know for sure is that festival messes it up, so that's all I should have fixed in the first place, really
18:38:28donsdwSansa e200: Does anyone know if the lag between turning wheel, and the volume inc/decrease is normal?
18:38:37kugelwhat das patch return when a patch succeded/failed to apply
18:38:43kugeltrue and false?
18:39:07kugelor doesnt it return anything?
18:39:47jhMikeSdonsdw: I think that little bump in volume might be due to pcm delay and using tone controls. Do you have them on?
18:39:55donsdwYes lag. I turn the wheel to increase volume and it takes about 2 sec before the volume increases.
18:40:20jhMikeSdonsdw: that doesn't seem normal at all. mine does not do that.
18:40:32donsdwI'll check, jhMikeS. Thanks
18:41:23jhMikeSdonsdw: unless you happen to be running extremely demanding codecs which could make the UI lag
18:42:37donsdwjhMikeS: So far I have not found those settings.
18:42:45*Nico_P has sound !!
18:42:58Nico_Pdon't know exactly why, but still
18:43:26jhMikeSdonsdw: Sound Settings|Bass/Treble
18:43:36pondlifeNico_P: Go on, try rewinding - I dare you!
18:44:50jhMikeSdonsdw: something tell me if you don't know where the setting is, you probably don't have it on since it's off by default
18:45:52donsdwjhMikeS: I set bass, treble, and volume all to zero. No more lag. Thank you! Now I'll play around and see which one caused it.
18:46:04 Join My_Sic [0] (
18:46:04delYsidIs there a reason why on my sansa the scrollwheel does not change volume in the MIDI plugin?
18:46:16 Join Frazz [0] (
18:46:29delYsidIt would make sense that the scrollwheel should behave like in other playing modes I think
18:46:37Nico_Ppondlife: it was very brief and I can't get it to play again... :(
18:47:05pondlifeThat's good - it means you have something to debug.
18:47:05jhMikeSdelYsid: noone added the support?
18:47:20donsdwjhMikeS: No, I'm just not familiar with menus. I was looking in General Settings. Duh.
18:47:52 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:48:57delYsidjhMikeS: I was guessing :-)
18:54:26 Quit Nick_Brackley ("ChatZilla [Firefox]")
18:54:29 Quit pondlife ("disconnected has pondlife")
18:54:41 Join daurn|laptop__ [0] (n=fake@
18:54:50 Nick daurn|laptop__ is now known as daurn|laptop (n=fake@
19:01:09 Quit amiconn (Nick collision from services.)
19:01:15 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
19:01:36amiconnIt seems that really no 64 bit host was building sims for a few months
19:01:48amiconnNow that mine is enabled again, the warning appears (X5 sim this time)
19:02:29*amiconn fix
19:03:53***Saving seen data "./dancer.seen"
19:06:19 Quit daurn|laptop_ (Read error: 110 (Connection timed out))
19:07:17 Join eigma [0] (n=cat@
19:07:54 Join MournBlade [0] (
19:12:56 Quit atsea-34 (Read error: 104 (Connection reset by peer))
19:13:19Nico_PjhMikeS: could you give me a quick briefing on how to make code blocking ?
19:13:53 Join daurn|laptop__ [0] (n=fake@
19:14:51 Join dmccarthy [0] (
19:15:50 Join Redbreva [0] (
19:18:00jhMikeSNico_P: queues are blocking if no messages are in them. mutexes block a thread if another thread has been granted it. the waiting thread will be woken and given the lock when the thead that owns it releases the lock.
19:22:44Nico_PjhMikeS: thanks, I'll try the mutexes
19:23:27 Join Wiwie [0] (
19:23:55 Quit BrianHV (Remote closed the connection)
19:24:15Nico_PjhMikeS: what do you think of having a mutex in buffering.h and using it in buffering.c and playback.c ?
19:24:22Nico_Pto sync buffer accesses
19:24:29 Quit |Rain| (Read error: 110 (Connection timed out))
19:25:20amiconnWhy would buffering need a mutex?
19:25:22jhMikeSNico_P: it all depends. that's explained in the paste.
19:26:26amiconnn1s: Track change now works without crash on 64 bit sim. Had to reconfigure and rebuild fully though
19:26:30Nico_PjhMikeS: thanks
19:27:09Nico_Pamiconn: I'm not rally sure, but some calls may need some
19:27:25amiconnrasher: "Thanks" for breaking "plugins" for sapi.... :/
19:27:27jhMikeSwe're not going down the "cooperative is safe" road on this round :)
19:27:44Nico_PjhMikeS: what do you mean ?
19:28:04amiconnThat has nothing to do with cooperative. Unless I am missing something, buffering is inherently safe
19:28:06rasheramiconn: oh, well I only knew about it being wrong for festival, so that's all I could really put there
19:28:19amiconnEach pointer is only ever modified by one thread
19:28:19 Quit scorche|work ("CGI:IRC (EOF)")
19:28:31 Join scorche|work [0] (n=8dc5049d@rockbox/administrator/scorche)
19:28:54amiconnrasher: The last change (moving the correction from sapi to all-english) was done by you as well...
19:29:22 Join BrianHV [0] (
19:29:23jhMikeSI think this version allows multithread access
19:29:27rasheramiconn: I guess I forgot that it was previously done for sapi, since I didn't add that.
19:30:03 Quit daurn|laptop (Read error: 110 (Connection timed out))
19:30:18 Join ompaul [0] (n=ompaul@freenode/staff/gnewsense.ompaul)
19:31:30jhMikeSamiconn: you mean what I call - "chasing pointers"? head/tail. only one thread has access to modifying each?
19:33:48 Join belarm [0] (
19:33:50jhMikeSyeah, that's safe as long as the rule never, ever has to vary for any reason.
19:33:59 Quit donsdw (Read error: 110 (Connection timed out))
19:34:06amiconnCurrent hwcodec buffering works this way, and it's perfectly stable, even though the consumption (tail) happens in isr context
19:35:03belarmAnyone know the keybinding to stop the "Recording" function on an 80GB iPod?
19:35:21belarmnot had much luck finding it in the docs
19:36:34Nico_PjhMikeS: "only one thread has access to modifying each" > does this mean like playback modifies tail and buffering modifies head ?
19:37:06jhMikeSamiconn: I setup mpegplayer as such too.
19:37:23amiconnNico_P: yes
19:37:45amiconnThe other only ever monitors, never modifies
19:37:51jhMikeSNico_P: but it's only safe if whatever is at the head becomes visible atomically
19:37:51Nico_PI'm pretty sure my code respects that
19:38:25amiconnjhMikeS: It is safe as long as the (global) pointer is only increased *after* everything is actually in place
19:38:28Nico_PjhMikeS: can you ellaborate ? I'm not sure I understand what you mean
19:38:45rasherbelarm: it really should be in the manual
19:39:08amiconnThat way the consumer can never assume data is already available when it actually isn't yet
19:39:29jhMikeSIn mpegplayer, the audio data is added at the tail and then the write_pos is updated. The interrupt reads the data and pulls it off the buffer by modifying read_pos.
19:39:56 Join funky_ [0] (
19:40:12rasherbelarm: But it isn't, for some reason.. hang on.
19:40:18amiconnYup. Exactly like what the hwcodec playback engine does
19:40:34bluebrotherrasher: I guess this is due to the fact that recording is still experimental on Ipods
19:40:36amiconn(and also the even more simplified one in wavplay.rock)
19:40:39 Join Domonoky [0] (
19:40:49rasherbluebrother: It should still be mentioned in the manual
19:40:49jhMikeSThe interrupt never need disabling to sync things
19:41:14bluebrotherrasher: yes, it should. But the manual got outdated quite a bit lately
19:41:28bluebrotherDomonoky: what do you think about an "update" button in the Quick Start tab?
19:42:02Domonokysounds nice..
19:42:02rasherbelarm: It's the global "cancel" button.. whatevr that is on ipod
19:42:11jhMikeSof course I used 64-bit variables which are efficient to add/subtract and good for about 3,000,000 years at 44kHz
19:43:24rasherbelarm: Probably the prev button, it seems
19:46:08jhMikeSI just wonder about other operations like MoB and codec loading and such. playback.c need sync primitives in some select places for that stuff.
19:48:31 Quit dmccarthy ("Leaving")
19:49:23*jhMikeS realizes a nasty thing re: PP irq vectors and caches.
19:51:15jhMikeSheh...funny. shoulda seen that sooner. fiq is changed by playback but the COP reads those vectors too and they're not in IRAM
19:54:06amiconnWhy is that vector changed at all?
19:54:25amiconnI would expect that it is set up by crt0, and then never changed
19:54:39 Quit barrywardell ()
19:54:44amiconn(well, on SH some plugins do nasty vector fiddling....)
19:55:53 Quit bluebrother ("bbl")
19:56:41Nico_PjhMikeS: what exactly is codec_request_buffer_callback(size_t *realsize, size_t reqsize) supposed to do ? advance the tail of reqsize and return the size available in realsize ?
19:56:42 Join cothrige [0] (
19:56:55jhMikeSamiconn: when changing the handler from playback to recording and back
19:56:59cothrigehowdy all
19:57:18amiconnjhMikeS: Why not use one handler, and handle the mode within that?
19:57:25*Nico_P forgot about codecs.h... looking
19:57:26jhMikeSI think it should
19:57:32amiconnOn archos, the MAS isr works that way
19:58:43jhMikeSNico_P: the codec wants some data, it tries to get reqsize bytes and return what was granted in *realsize
19:59:28Nico_Pthat's what I initially understood but I obvioulsy got confused
19:59:38jhMikeSamiconn: I think the concern must be speed since fiqs happen alot but a test and b instruction's not much
20:00:05 Quit ThUnD3r|Gr33n|Aw (" IRcap [8.23] ")
20:00:45amiconnYeah, I know that FIQs happen rather often. MAS interrupts happen much less often (although, for recording there is one interrupt each 30 recorded bytes)
20:00:45 Quit kubiix (Read error: 104 (Connection reset by peer))
20:01:12amiconnThat's already mp3 compressed though, so not *that* often
20:01:16 Quit cothrige (Remote closed the connection)
20:01:18jhMikeSif MAS can handle that, ARM certainly can
20:01:50amiconnHow many sample pairs are transferred per FIQ? 12?
20:02:21amiconnSo ~4000 ints/sec
20:02:43jhMikeS12 pairs for packed. 6 pairs for unpack 32-bit.
20:03:13amiconnOkay, quite a few more than on SH1 for recording (around 700 ints/sec)
20:04:57jhMikeSI don't think a few more instructions will be noticed. The asm I did has 1/2-2/3 the instructions of the current asm fiq (didn't really count)
20:05:41amiconnOh, forgot that SH1 also handles more... in wavrecord.rock
20:06:06amiconnFor wav recording, it's one int per 36 recorded bytes - but that's uncompressed...
20:06:45jhMikeSand wav is 44100?
20:06:46belarmrasher: actually, it seems to be the menu button; but when i chose the Recording option by mistake, i guess the iPod locked up; as choosing it now, i get an interface I didn't see last time.
20:06:58belarmthanks for the help, in any event
20:07:03amiconnCan be any standard rate from 8000 to 48000 Hz
20:07:21rasherbelarm: That's a known bug, I think (leaving the recording screen puts the ipod in a bad state)
20:07:29amiconn(all rates which are allowed for mp3)
20:07:57belarmentering it seemed to do it this time
20:07:59 Quit spiorf (Remote closed the connection)
20:08:17jhMikeSso for 44,100Hz, 16-bit, stereo thats 176,400/36 = 4,900 ints/sec.
20:08:20belarm(rather, the time in question, when i was stuck)
20:08:24amiconn8 or 16 bit (but wavrecord only uses 16), mono or stereo
20:08:31 Join spiorf [0] (
20:09:24jhMikeSI think we'll be ok with a dispatcher on pp then :)
20:16:14amiconnI think so too
20:16:37amiconnBtw, the wavrecord isr is running from dram (!), because archos plugins can't use iram
20:17:25amiconnAnd it still leaves enough time to write the data to disk (but on Ondio it depends on the individual MMC whether it can keep up)
20:23:09jhMikeSI dunno, this might overload it:
20:30:41 Join kugel_ [0] (
20:30:52 Join Zagor [0] (
20:30:55 Quit kugel (Nick collision from services.)
20:30:57 Nick kugel_ is now known as kugel (
20:31:46 Join Ebert [0] (
20:34:13kugelis there a calender in rockbox?
20:34:21 Quit eigma (Read error: 110 (Connection timed out))
20:35:28 Join linuxstb_ [0] (
20:36:12jhMikeShmmm...I cut an instruction from that so I guess there's hope
20:37:38amiconnkugel: Yes but for archos recorders only so far
20:37:56kugelim a little bit confused
20:38:16amiconnThere's calendar patch in the tracker, but that one is unfortunately not suitable for low display resolutions
20:38:18kugelthere is a calendar plugin on the tracker, as well as a calenar button map for sansa
20:38:40kugeldo they belong together? because they are not marked as related
20:38:57amiconnSo someone needs to go ahead and make a "best of both"
20:39:43kugelso they belong together, yes?
20:40:03 Quit My_Sic (Read error: 104 (Connection reset by peer))
20:40:41 Join My_Sic [0] (
20:41:22amiconnI don't know
20:41:25 Quit Wiwie (Read error: 110 (Connection timed out))
20:42:04kugelok, i think im gonna wait with adding the calendar to my patch
20:42:12kugeluntil this questions is answered
20:44:15 Join Wiwie [0] (
20:48:56Nico_Psound \o/
20:49:26markunNico_P: nice!!
20:49:44markunare you developing in the sim?
20:50:05Nico_PI'm still quite far from anything functional, but at least playing one track works
20:54:45Nico_Pcodec_advance_buffer_loc_callback is a bit of a PITA
20:58:24jhMikeSNico_P: hurry up! mpegplayer is way more stable on H10 than playback.
20:58:29 Join eigma [0] (i=eigma@
20:58:38Nico_Phehe :)
20:58:48jhMikeSIn fact, I can't crash mpegplayer at all anymore
20:59:04Nico_PjhMikeS: time for FF/RW then :)
20:59:28amiconnjhMikeS: Do you have some uncommitted changes?
20:59:32pixelmajhMikeS: and how does mpegplay look with voice ui enabled? ;)
21:00:05jhMikeSamiconn: to what? mpegplayer? no.
21:00:18 Quit jgarvey ("Leaving")
21:00:19amiconnThen try starting a video with voice enabled
21:00:28jhMikeSpixelma: it gets a bit green but I like that ;)
21:00:31amiconnThis is bound to crash...
21:00:43amiconn...or become greenish, yes
21:01:13jhMikeSother than that which is so simple to do I'm suprised it's not dealt with already
21:02:50 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
21:03:03jhMikeSamiconn: you said doom disables voice? I know there's some pluginlib thing for that now.
21:03:20hcsHas anyone else noticed that OGGs are crashing?
21:03:25Nico_PjhMikeS: what are you suprised it's not dealt with already ?
21:03:36Nico_Phcs: on what player ?
21:03:55***Saving seen data "./dancer.seen"
21:04:19hcsNico_P: Gigabeat
21:04:29jhMikeSNico_P: It's should be just a matter of disabling talk menus
21:04:34Nico_Phcs: I think it's been fixed
21:04:46linuxstb_Nico_P: You mean avoided?
21:04:49hcsNico_P: ok, I'm just now updating, but I thought I'd check
21:04:59Nico_Plinuxstb_: yeah :)
21:05:50jhMikeShmmm...I guess it was added for backlight, not for voice
21:05:51amiconnjhMikeS: Looks like its PLUGIN_IRAM_INIT(rb) - your own work iirc
21:06:51amiconnIt's the very first thing rockdoom.c:plugin_start() calls
21:08:41amiconnHmm, but mpegplayer calls it too...
21:08:54*pixelma does the unpopular job of nagging Zagor about c200 radio detection ;)
21:09:21Zagorah good, I'll test it right away
21:10:33jhMikeSamiconn: not only those
21:11:01Zagorpixelma: where was the patch?
21:11:08amiconnSo what could be wrong with mpegplayer and voice...
21:11:24 Quit hcs ("Leaving.")
21:11:53jhMikeSI'm gonna try the menu disabling. It has no trouble as long as the talk menu option is off.
21:13:48amiconnHmm, but doom also uses standard menus
21:13:54*amiconn has a suspicion
21:14:07 Quit Wiwie (Read error: 110 (Connection timed out))
21:15:18low_lightZagor, pixelma: I'm not really sure what the radio patch will do when there's no radio...besides not freeze
21:16:43Zagorhehe, my radio-less c200 ... has a radio!
21:16:57jhMikeSthat's impossible!
21:17:07 Join TMM [0] (
21:17:09pixelmaa working one or just in the menu?
21:17:51jhMikeSwhat chip does that use? the same sanyo as e200?
21:18:09low_lightjhMikeS: yes
21:18:09*pixelma gets greedy - one with radio for the price of one without ;)
21:18:21Zagordoes it say anything about radio on the cover of the radio-c200s?
21:18:51Zagoror on the box?
21:18:54pixelmanope, there was only a sign on the box
21:19:26Zagormy box says nothing
21:19:37advcomp2019Zagor, it could be like the e200... there is the radio chip but the firmware was stopping it from being in use
21:19:58amiconnZagor: It's no Rock-box hence no voice UI ;)
21:20:06Zagoradvcomp2019: yeah, seems like it. I just want to be sure I didn't buy the radio version by mistake.
21:20:21Zagoramiconn: :)
21:20:25pixelmayeah, I remember that there were reports of it for some "radio-less" e200s too
21:20:52TMMpixelma: those were not just broken? ;)
21:21:25Zagorbut since the OF didn't have radio enabled, I'm pretty sure this one is not supposed to
21:22:33pixelmahehe, nice :)
21:24:00amiconnMuch like the early v2 recorders...
21:24:40Zagoryeah. seems like everyone is doing.
21:26:02TMMlinuxstb_: linuxstb: now it keeps resetting itself in USB rescue mode as well, there is really nothing else to do but send it back I'm afraid
21:26:27pixelmaso we still don't know what happens if someone has a c200 that really has no radio (or if there are such c200s)
21:26:29linuxstb_TMM: Oh well, it was educational at least...
21:26:31TMMlinuxstb_: linuxstb: did yours show the same behavior without the NAND attached?
21:26:38amiconnjhMikeS: Found a bug in doom instead :/
21:26:53linuxstb_TMM: No, it doesn't reboot.
21:27:13TMMlinuxstb_: well, I guess I should give my chip credit for trying then
21:27:36jhMikeSamiconn: no more green blocks in mpegplayer now either
21:27:51rasherpixelma: can't you just commit the radio-enabling patch and see if any actual radio-less c200 appears. It's not like you could prove they *don't* exist anyway
21:28:00TMMlinuxstb_: did you try reconnecting the NAND while it is plugged in yet?
21:28:05amiconnYeah, but I can make doom crash with menu voices enabled when setting boolean options
21:28:25amiconn(and it says 'yes' for both yes and no)
21:28:35 Quit eigma (Read error: 110 (Connection timed out))
21:28:50amiconnFamiliar bug - nobody seems to check what struct opt_items consists of :(
21:28:58pixelmarasher: yeah, I think so too - but it is low_light's patch...
21:29:12amiconnThe voice id is an id, not a pointer, and "no voice" is -1, not NULL
21:29:29amiconn0 is a valid voice id - happens to be "yes"
21:30:08rasherpixelma: yeah, I meant a more general you. Zagor's c200 having a radio chip is a good indication that all c200s might have one (but not proof, of course)
21:30:13jhMikeSsomethings odd though. buzzing about the menus it seems when the disk spins down from the menu, voice gets trashed and no longer plays
21:30:24jhMikeSthis is from the root menu btw
21:31:17amiconnHrrrm, and rockdoom.c messes up menu_items and opt_items :(
21:31:50 Nick fxb is now known as fxb__ (
21:31:58jhMikeSI don't have to do anything. Just wait for the disk to spin down on H10 and voice is totally clobbered.
21:33:25 Join agm3nt [0] (
21:33:39delYsidjhMikeS: I saw similar behaviour when I voice-enabled metronome
21:33:57delYsidthat uses pcm_play stuff (just as mpegplayer does, doesnt it?)
21:34:50pixelmaagm3nt: does your c200 have a radio (officially)
21:34:55delYsidI think there is a race when using pcm_play and voice.
21:35:32delYsidI can consistantly hang voice output until I reboot if I send some talk_id while pcm is being used
21:35:33jhMikeSwouldn't suprise me
21:36:48 Join eigma [0] (i=eigma@
21:36:59delYsidthats because pcm output reuses the voice channel I guess.
21:37:30jhMikeSThere's only one pcm output for everything
21:37:31agm3ntpixelma: I'm buy it without radio but later I install american firmware with radio support and higher volume
21:38:03 Quit My_Sic (Read error: 104 (Connection reset by peer))
21:38:19pixelmaah, seems it's common then
21:39:39jhMikeSagm3nt: lower volume than american firmware? does some Ministry of Aural Heath mandate that?
21:40:15jhMikeSjust sounds pun
21:40:25rasherjhMikeS: I believe there are some EU regulations on volume... but it's all very weird.
21:40:30delYsidjhMikeS: the EU has max volume laws since a short time now.
21:40:43jhMikeSthough so.
21:40:50delYsidsomething about standard ear protection...
21:41:04jhMikeS*sigh* socialism
21:41:08low_lightpixelma: feel free to commit the patch now...I can't do so for about 8 hours
21:41:34delYsidAs someone who knows a bit about ear damage I actually think thats a fine default.
21:41:55 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
21:42:32hcsyep, ogg works now, thanks
21:42:43agm3ntjhMikeS: yes, eu firmware is with lower volume
21:43:46 Join qweru [0] (
21:46:24low_lightpixelma: and if you are feeling daring, you can enable HAVE_RECORDING :)
21:46:40*jhMikeS finds the corruptions that occur because of frequency scaling on h10 very bothersome
21:47:08 Join Arathis_ [0] (
21:47:18 Quit Arathis_ (Read error: 104 (Connection reset by peer))
21:47:22pixelmalow_light: won't do the latter, want to leave that to the pros ;)
21:48:35pixelmabut radio... if you wish :)
21:50:23low_lightrecording *should* work as-is, but there might be layout issues with the recording screen and keymap problem
21:50:28low_lightI've never used visited the recording screen, not even on my h140
21:50:48low_lightused OR visited
21:51:17pixelmawouldn't wonder if there are keymap problems as well...
21:51:43pixelmaprobably will take a look, but won't commit anything
21:53:25 Quit courtc (Read error: 110 (Connection timed out))
21:54:20 Quit homielowe (Read error: 110 (Connection timed out))
21:54:28jhMikeShrmm...can't pin in on scaling. wtf.
21:55:01low_lightThe EQ screen is messed up on the c200 too, the hs filter isn't displayed...perhaps that screen needs to scroll.
21:55:42amiconnjhMikeS: Frequency scaling should be fine now
21:56:43jhMikeSit seems to affect starting i2s playback somehow though
21:57:11amiconn(now == since more than 2 months)
22:00:36 Join freqmod [0] (
22:00:58jhMikeSok, force enabling the TX FIFO seems to fix it, even though it's not disabled anymore because disabling it kills i2s playback after recording
22:02:14jhMikeSyes, now everything is solid
22:03:03 Join Wiwie [0] (
22:03:24 Quit Arathis (Read error: 110 (Connection timed out))
22:05:22pixelmajhMikeS: would that be something that affects all swcodec targets? Just asking because I saw similar problems with mpegplayer and voice on my M5 (just with whitish blocks not green)?
22:05:36jhMikeSpixelma: yes
22:05:53jhMikeSI can go into the menu just fine now
22:06:24 Join donsdw [0] (
22:06:32 Quit desowin ("use linux")
22:08:03donsdwMy Sansa e200 is now a brick. LCD is lit with message, "Searching... 13 found (PREV to abort)
22:08:20donsdwBut I cannot turn off the player.
22:08:27alienbiker99um.... hold down the power
22:08:27 Quit eigma ()
22:08:38donsdwDoesn't work
22:08:49donsdwDoes not shut down.
22:08:50alienbiker99try holding it hard for 15 sec
22:08:55n1sdonsdw: please don't incorrectly use the term 'brick'
22:09:05alienbiker99its just frozen
22:09:32donsdwn1s: okay
22:09:37advcomp2019remove the battery is another way
22:10:07scorche|workholding down power for a while should work..
22:10:09donsdwalienbiker99: I think the 15 seconds worked. Thanks!
22:10:36 Join mpeccorini [0] (
22:10:42 Join ubercooldave [0] (i=586d0c45@gateway/web/cgi-irc/
22:10:43jhMikeSit's a "soft brick" vs. a "hard brick" :P
22:10:53scorche|workBagder: around?
22:11:06ubercooldavehey guys
22:11:07scorche|workwell, i wouldnt even call that a soft brick
22:11:12*n1s wonders what kind of houses you would build with soft bricks ;-)
22:11:26krazykitn1s, squishy ones :D
22:11:30jhMikeSit's just fresh clay
22:12:00donsdwIt's no brick at all, just a donkey at the non-connector end of the earbuds.
22:12:27amiconnn1s: padded cells ;)
22:12:28n1sso it's a brick that just crashed and didn't burn :-P
22:13:59scorche|workanyone have any more thoughts on the themes site? (newer version is up at hope to get it out this week
22:14:32scorche|workpixelma: and i added the c200 ;)
22:15:00pixelmanot much to see there yet ;)
22:15:29scorche|workwell, there wont be till it goes live
22:15:50scorche|workbut any other changes that anyone thinks that are needed before it debuts?
22:15:51n1sscorche|work: we now support 1G and 2G ipods too
22:15:54*jhMikeS better commit this little bitty before the FS reports roll in by the droves :P
22:16:27ubercooldavei have a few problems with some themes... i got a 30gb ipod 5g and on some themes like the cassette one the GUI when playing a song is not displayed properly, its just the basic one with a different background, and im sure i downloaded a theme for my specific ipod so it should have worked? help appreciated
22:16:35scorche|workyeah...i plan on adding that later tonight, n1s
22:16:56low_lighthmmm...some more c200 magic: DEV_EN &= ~0xc0000137; DEV_EN |= 0xc0000137;
22:16:59pixelmajhMikeS: voice and mpegplayer? There already is a bugreport for it - not really a new one ;)
22:17:02*jhMikeS introduced a little i2s startup bug but has the fix
22:17:05scorche|workubercooldave: does the theme require patches?
22:17:11mpeccorinihi everyone, I have a problem with my ipod nano 1st gen. It was running rockbox just fine, I stopped using it for over a month and now Windows says it's not formatted when I connect it, so I can't add/remove song or anything. Any ideas?
22:17:31scorche|workRedbreva: ping
22:18:17mpeccoriniEven rbutilqt can't detect it
22:18:32ubercooldaveim not sure what you mean, all i did was download the theme and unzip it to the right folders, and it seemed to install quite a few files in different places... the AquaPod theme works very well tho and looks awsome so well done to whoever made that lol
22:18:38jhMikeSok, so 1) fix the i2s 2) fix the mpegplayer
22:19:30low_lightpixelma: you should replace the c200 sim bitmap with your version
22:19:57pixelmayou mean the svg?
22:20:23scorche|workubercooldave: well, what you describe means that the theme you are trying to load is either broken, or requires patches
22:20:57parafinor both :)
22:21:17ubercooldavethanks for the help
22:21:19low_lightpixelma: yeah...scaled of course. I had commited a scan of my c200, not the best image
22:21:27ubercooldavewhat do you mean by patches though?
22:22:23ubercooldaveif i just downloaded a theme like Aurora Black, would it work simply by me installing all the files from the theme into the right place or would i require something else?
22:23:17 Join nico__ [0] (
22:24:14pixelmalow_light: I don't know, every now and then the suggestion comes up to replace all uisim bitmaps with png made of the svgs. Nothing happened so far...
22:25:50krazykitubercooldave, theme installation is covered in the manual
22:26:26krazykitubercooldave, but if the theme requires patches (modifications of the source code, unofficial) you need to either compile yourself or find a custom build
22:26:46ubercooldaveokay lol
22:26:58ubercooldavethank you very much for your help, everyone
22:27:06rasherCan't SDL load pngs anyway?
22:27:12jhMikeSdarn, I wanted to commit two changes really fast to get them in in one build but no luck. :\
22:27:21 Join ilgufo [0] (
22:27:29ilgufogood evening
22:27:39amiconnjhMikeS: You have 60 seconds afaik
22:27:55pixelmarasher: sorry, used that wrong but that was not the point
22:27:56 Join BigMac [0] (
22:27:58nico__i'm trying to do something but can't figure of a way... can you tell me if it's do-able? (...)
22:28:32jhMikeSIt didn't take me more than 30 to do them both it seemed
22:29:01n1snico__: not unless you tell us what you'r etrying to do
22:29:09rasherpixelma: I know what you meant. It just made me wonder if there's a good reason to use BMP rather than PNG. But then, I don't know if SDL comes with png loading - I just thought it might
22:29:26nico__let's say i want rockbox to find all my music of a particular genre (via tags) and play it by album (one album after the other, with an album's songs in the track number order of course)
22:30:14nico__(and if possible, as a bonus, play the albums from a same artist together, say by release date)
22:30:16rasherSeems it doesn't. I guess that's reason enough. No need to add another dependency I suppose.
22:30:33amiconnrasher: The windows sim used bmp because windows has built-in support for bmp
22:30:41n1snico__: should be possible with the database, possibly you will need a custom tagnavi.conf setup, search in the forums
22:30:48amiconnI guess the BMPs were just copied to the sdl sim
22:31:05nico__nls: thanks
22:31:46pixelmanico__: the DataBase page in the wiki also explains the custom_tagnavi
22:32:36amiconnjhMikeS: ear and earier ;)
22:33:03pixelmahe wanted to save some time ;)
22:33:16 Quit ompaul ("#freenode-levelone cheers in a week or so")
22:33:45ubercooldavei think i see now... on the themes website it says custom build see description, but then the discription is very short and dosent mention it :(
22:34:04rasheramiconn: it doesn't look like SDL supports anything but BMP by default, so that's a good reason as well
22:34:08*jhMikeS has an extra ear?
22:34:24 Quit ilgufo ("So Long, and Thanks For All the Fish -")
22:34:53nico__pixelma, you mean this wiki? ->
22:35:07 Join petur [0] (n=petur@rockbox/developer/petur)
22:35:10amiconnjhMikeS: Hmm, that disabling should probably be done in doom as well
22:35:13jhMikeShow eerie :P
22:35:43amiconnI thought the iram init is enough, but it seems it isn't, even if no menu item is actually voiced
22:35:47pixelmanico__: yes... just type DataBase in the "go" box
22:35:49jhMikeSprobably since it hasn't been done it seems
22:36:10jhMikeSthat won't stop a message posted to the voice thread from waking up the codec
22:36:18 Quit MournBlade ("Her lips were wet with venom")
22:36:25 Quit belarm ("Leaving")
22:37:12nico__pixelma: ah, my bad, didn't notice the search was case sensitive :-/
22:37:15amiconnIdeally voice should be made to cooperate with plugins using iram, but that would probably end up rather complex
22:37:26jhMikeSFrankly Doom should just leave music running until a "Play Game" is selected
22:37:38amiconnnico__: The "go" box is not a search box...
22:38:15amiconnjhMikeS: But then it would have to re-init iram.... unless it would be possible to hook plugins into the iram swap mechanism
22:38:16nico__amiconn, my bad again :)
22:38:23jhMikeSI have thought about some sort of memory arbitration where buffers are locked and released
22:39:29amiconnnico__: It *is* confusing...
22:39:32jhMikeSattempting playback at that time would just make it refuse to start which would be a simple request for that buffer
22:40:15jhMikeSAre you saying we should swap iram in such a way that voice/playback and iram using plugins can all run?
22:40:26 Join clowner [0] (i=97520855@gateway/web/cgi-irc/
22:40:54amiconnjhMikeS: At least either voice+playback *or* voice+plugin
22:41:12amiconnAll 3 together would be a bonus
22:41:14 Nick clowner is now known as nameless (i=97520855@gateway/web/cgi-irc/
22:41:27 Quit _BigMac (Read error: 110 (Connection timed out))
22:41:31jhMikeSwell, #1 happens already at least.
22:41:38amiconnThe plugin would have to provide the extra swap buffer in that case... but that's probably too complicated
22:41:58namelessnothing can support ipod nano 2nd for the videos ?
22:42:44amiconnWe definitely do want voiced plugins, the question is whether that applies to any of the plugins which currently use iram (they're all visual)
22:42:48low_lighthmmm...I had to reset the FM region settings in the OF to get the tuning in Rockbox right
22:42:53jhMikeSit would have to happen in time with thread switching and that's quite a hunk to copy on a context switch :)
22:42:56nico__i guess there is no way to sort a playlist according to some criteria? that could be a way to do what i'm looking for...
22:43:27nico__can it be customized too, let's say include a "sort by ..." item along with the "schuffle" one in playlists
22:43:44amiconnjhMikeS: Huh? The core doesn't do that on every thread switch, o if those threads actually need to run, afaik
22:44:16 Quit hcs ("Leaving.")
22:45:02namelessnone can help me ?
22:45:05ubercooldavei just got some patches to install... i am unfortunatly clueless as to how to install them... i checked the manual and it didnt make much sense im very sorry
22:45:10 Quit agm3nt ()
22:46:13low_lightamiconn: should lcd_enable put the c200 lcd in standby mode?
22:46:22Bagderubercooldave: check the wiki instead
22:46:34amiconnI don't know. This mess needs cleaning up too
22:46:40pixelmalow_light: still want me to commit the radio patch? And should I mention something about the c200s that are labelled to not have a radio?
22:46:48amiconnWe don't need 2 different functions to put the lcd to sleep
22:48:22jhMikeSamiconn: with the codec swap, I suppose that's basically what happens anyway
22:48:23low_lightpixelma: sure, go ahead :)
22:48:28 Join toffe82 [0] (
22:48:41 Join advcomp2019_ [0] (n=advcomp2@
22:49:19jhMikeSlow_light: is the lcd slow to start back up? that's the only reason a separate lcd_sleep exists with Renesas-based targets.
22:49:27 Quit nameless ("CGI:IRC (EOF)")
22:49:55ubercooldaveis it easy to install a patch or not?
22:50:08 Join obo [0] (
22:50:14Wiwierockbox rulez!
22:51:05 Join courtc [0] (
22:51:14amiconnjhMikeS: I never saw a difference between sleep and standby...
22:51:16peturpondlife: around?
22:51:34low_lightjhMikeS: I have not tested
22:51:37amiconnThat's in spite of me usually being rather delay sensitive
22:51:58 Quit obo (Client Quit)
22:52:12 Join jac0b [0] (
22:52:27jhMikeSI see it and others commented on it. Initially I did have just full standby mode always.
22:52:46jac0bhow do I load a default cover if one is not found
22:52:52 Quit nicktastic ("Leaving")
22:53:08low_lightso lcd_enable is for display on/off and lcd_sleep is standby on/off?
22:53:16jac0bI put this in the wps but nothing happens %?C<%C|cover.bmp|>
22:53:53jhMikeSlow_light: yes, when implemented as such. the renesas needs many longer delays when coming back from power off compared to just visible display.
22:54:24ubercooldaveis the blogger / theme creator julius here?
22:55:06 Join fm2 [0] (n=chatzill@
22:55:30 Part Domonoky
22:55:46fm2Bagder: hello. Could you please answer some questions about apps/Makefile?
22:55:48*jhMikeS gets nostalgic talking about one of his first rb patches (committed by RaeNye)
22:56:07 Join roolku [0] (
22:56:10Bagderfm2: perhaps!
22:56:35fm2 Bagder: what is the difference between '=' and ':='?
22:57:07fm2 Bagder: and I can't seem to find the slots for compiling files, i.e. cc call.
22:57:29 Quit mpeccorini (" HydraIRC -> <- IRC with a difference")
22:57:32linuxstb_fm2: Look at tools/ (IIRC) - that's included in every Makefile
22:57:40ubercooldavedoes anyone know if julius' jThemes all work with the same custom build or not?
22:57:51 Quit Frazz (Read error: 110 (Connection timed out))
22:58:14amiconnjhMikeS: You notice a difference of ~60 ms ??
22:58:19 Join midgey [0] (
22:58:27Bagderfm2: question #1 is found in the gnu make docs
22:58:39Bagderfm2: what "slot" ?
22:58:50fm2linuxstb: ah! That's it!
22:58:53linuxstb_I assume s/slot/rule/
22:59:06fm2linuxstb: yes
22:59:24jhMikeSdisplay on itself takes 50
23:00:38*jhMikeS tired and doing bad maths
23:00:39 Quit jac0b ()
23:01:10amiconnjhMikeS: display on takes around 60ms, assuming a frame frequency of ~60Hz
23:01:12ubercooldavefor the ipod 5g... does rockbox drain the battery faster than the apple firmware or is it just that rockbox cannot guess the battery remaining acutratly?
23:01:38amiconnResume from sleep is nearly instant, resume from standby takes ~10ms
23:01:38jhMikeSdisplay_on has 80ms of delays
23:01:57 Quit lostnihilist ("Leaving")
23:02:52amiconnEven 100ms difference shouldn't be noticeable...
23:02:56jhMikeSpower_on takes 200. So it's 80 ms vs 280 ms. yeah I see that.
23:03:08Ebertanyone know a person/place to get a small Gigabeat dock-lineout connect??
23:03:13 Quit advcomp2019 (Read error: 110 (Connection timed out))
23:03:17amiconnpower_on needs to be *always* executed
23:03:44jhMikeSno, not if you only do the visible display procedure
23:03:57***Saving seen data "./dancer.seen"
23:04:21toffe82Ebert: if you find a connector, it is easy to do
23:04:35toffe82Ebert: you have the pinout on the wiki
23:04:43 Quit ubercooldave ("CGI:IRC")
23:04:53 Join Klevi [0] (
23:05:03Ebertwhere would i find a connector then?
23:05:34 Join My_Sic [0] (
23:07:01toffe82Ebert: you cannot find the connector by itself but you can buy some cable with the connector look for the toshiba pda e800 syncrho cable
23:07:16amiconnAha, hmm
23:07:16 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@
23:07:25 Quit seablue ("life, death, life, death")
23:07:28*amiconn wonders whether that method saves power at all
23:07:36Ebertah thx toffe82
23:07:45amiconnCompared to letting the display running without backlight I mean
23:10:08jhMikeSprobably not measurably. perhaps it's would be better to leave it on and then just do the whole thing/or not if the user wants to.
23:10:32jhMikeSthough it does affect boost measureably
23:11:15 Join jac0b [0] (
23:11:33jhMikeSbut I mean, if the screen will be left on, then it seems silly not to update it since it's visible in bright light
23:11:34jac0bis there a conditional tag for albumart
23:12:13amiconnjhMikeS: Hmm, comparing that sequence with the HD66789R, it seems the 180ms sequence near the end of the power supply setting could be shortened to 100ms
23:12:50amiconnI wonder whether it's a real design difference, or just a mistake in the docs
23:13:42 Quit jac0b (Client Quit)
23:15:54jhMikeSwhich one are you looking at? x5, e200?
23:16:09 Join Davide-NYC [0] (
23:16:49amiconnX5 is HD66773R (the one where the times are additive
23:16:58amiconnH300 is HD66789R
23:17:12 Join AceNik_ [0] (n=AceNik@
23:17:29AceNik_uys what does i2c mean ?
23:17:58*amiconn recommends wikipedia
23:18:19AceNik_the h10 [20GB] is hanging like at every point possible now, just after 1 min of startup or sometimes at startup
23:18:30 Nick parafin is now known as parafin|away (i=parafin@
23:18:59AceNik_it was working fine , perfect till 14941
23:19:21AceNik_after that i tried 14962, so somethin has gone wrong in between
23:19:49scorche|workmsg Bagder has anyone in the US said anything about the google money?
23:19:54jhMikeSthere's an extra delay on the e200 that isn't there on H300 (both HD66789R). it was part of the OF sequence.
23:20:02 Join lostnihilist [0] (
23:20:15AceNik_oh btw does the h10 with rolo bootin umpteen number of times by executin rockbox.mi4 cause i remember earlier it only was able to do this once
23:20:18Bagderhaha, the answe is no anyway ;-)
23:20:28Bagder(add 'r' somewhere)
23:20:36 Quit nico__ (Remote closed the connection)
23:20:45 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
23:21:20low_lightlater all
23:21:22 Part low_light
23:21:38jhMikeSAceNik_: say what? H10 should rolo
23:22:05 Quit Ebert ()
23:22:06 Join sin613 [0] (
23:22:47AceNik_jhMikeS: earlier while using rolo h10 used to be able to restart with the rockbox.mi4 only once, now it does it as many times as you want to
23:23:08AceNik_as in i execute is first then , after the restart i can execute it again
23:23:35jhMikeSwell...heh. I just did some threading work.
23:23:58AceNik_mike did u just commit the stuff for the12c
23:24:43AceNik_cause my h10 jus hangs liek hell, im not pointing at you patch being wrong, but just if you could hav a look at the 20GB haning
23:24:45jhMikeSrolo isn't so luck with voice enabled though
23:24:54jhMikeSI have a 20GB
23:25:18jhMikeSyou just updated?
23:25:20amiconnjhMikeS: Hmm, both sequence look somewhat different from the datasheet afaics
23:25:25AceNik_well my player just hangs at any instant now, i updated from 14941
23:25:29 Part Redbreva ("User is away.")
23:27:23VariableAny reason #2954 hasn't been committed?
23:27:24 Quit fm2 (Read error: 104 (Connection reset by peer))
23:27:35 Nick Variable is now known as Isolinear (
23:27:49AceNik_guys there is another minor thing if you could sort out the clock settins are not saved in the apps folder, they get saved in the games folder
23:28:28 Join webguest86 [0] (i=0cd834f9@gateway/web/cgi-irc/
23:29:03 Quit webguest86 (Client Quit)
23:29:31 Nick funky_ is now known as kubrick (
23:30:16jhMikeSAceNik_: you can rolo forever but it hangs like hell?
23:30:45AceNik_ok but whats causing the hang even in a normal restart can you figure that out
23:31:51jhMikeSAceNik_: If mine were then I would still be working on it but it's not hanging on anything at all.
23:32:07AceNik_k ill recheck in a while
23:32:18AceNik_probably ill do a clean build
23:32:25AceNik_thanks anyways c ya
23:32:31jhMikeSUmmm...I would recommend that
23:32:33 Part AceNik_
23:34:46 Quit tictoc (Remote closed the connection)
23:34:54 Join haemmy [0] (n=stefan@
23:36:36 Join tictoc [0] (i=tabac@gateway/gpg-tor/key-0xB9002659)
23:37:29 Quit haemmy (Client Quit)
23:37:37*jhMikeS wonders how something's called SVN rXXXXXX if it's not a clean build
23:38:02IsolinearMaybe he hasn't showered yet?
23:38:20jhMikeS:-) Maybe he should shower with his H10?
23:38:24 Quit hcs ("Leaving.")
23:38:42IsolinearHey, your opinion on 2954?
23:38:50 Quit Strath (Read error: 104 (Connection reset by peer))
23:39:46jhMikeSFS #2954?
23:39:56IsolinearYeah, scrolling margins.
23:40:00linuxstb_Isolinear: "viewports"
23:40:12Davide-NYCGuys, quick question: how functional is rockbox recording on the sansa c200?
23:40:20jhMikeSCan't say I've even thought about it at all.
23:40:47IsolinearMakes for snazzy WPSs.
23:41:04pixelmaDavide-NYC: it's not enabled yet and it has the same restrictions as the e200s
23:41:05jhMikeSHey Davide. That was about FS #2954.
23:41:10Isolinearlinuxstb: Viewports?
23:41:35*jhMikeS hopes to get at least 32KHz out of it but that means running the chip at 64KHz
23:41:57 Join Strath [0] (
23:42:13markunI just enabled the ARM optimized mdct in Tremor on my Gigabeat and it goes from 614% realtime to 613%!
23:42:56jhMikeShmmm...gigabeat really hates memory stalls
23:43:11jhMikeSis that in SVN or a patch?
23:43:42linuxstb_Isolinear: A planned future feature in Rockbox -
23:43:50markunwith .icode replaced by .text
23:45:06IsolinearAhhh.. Very nice.
23:46:03amiconnWhat? Gigabeat is only around 600% realtime for vorbis? I would have expected much more given the clock frequency...
23:46:07jhMikeSmarkun: it could also be alignment changes which trigger more cache line fills. data on arm should be aligned on arm as well so that a 16-byte segment takes only one cache fill instead of two.
23:46:19preglowbeen any juicy talk on preemptive multitasking while i was gone? :>
23:46:33 Join atsea-34 [0] (i=atsea-@gateway/tor/x-ceaff0631690e479)
23:46:34markunamiconn: yes, dissapointing :)
23:46:35jhMikeSnothing juicy
23:46:52preglowjhMikeS: still no idea why test_codec fails now?
23:47:13amiconnAnyway, 500 or 600%, doesn't really matter
23:47:30amiconn...especially since the gigabeat doesn't even do frequency scaling
23:48:04markunamiconn: but still we get a better runtime with lower CPU load
23:48:44jhMikeSpreglow: I thought it failed on PP. Obviously markun could test on gigabeat. But I have no idea right now. No other plugin shows any symptoms.
23:48:45amiconnReally? I thought that even frequency scaling doesn't change that, but causes stability problems, and was disabled for those reasons
23:49:09alienbiker99600% realtime is bad?
23:49:27amiconnSo I somehow expected that changing the cpu load also doesn't change battery runtime
23:49:42jhMikeSif you don't do frequency scaling, the player will at least sleep the core alot
23:49:56markunamiconn: I expected that as well, but it turned out to be wrong. Let me look up some measurements
23:50:03amiconnvery odd
23:50:36 Quit linuxstb_ ("ChatZilla [Firefox]")
23:50:43jhMikeSIt didn't seem everything was done that should be though
23:51:18preglowjhMikeS: don't you have a gigabeat?
23:51:26jhMikeSpreglow: yes
23:51:51jhMikeSmarkun: try aligning all the data it accesses to 16-bits...I'm really curious
23:51:58jhMikeS16 bytes :)
23:52:22preglowjhMikeS: then why does markun have to test on gigabeat? :)
23:52:51jhMikeSbecause he came here reporting on the asm mdct test_codec result
23:53:04jhMikeSfor tremor
23:53:56markunjhMikeS: how do I do that?
23:54:48jhMikeSmarkun: in any .asm file .balign 16
23:55:02n1sah, right in asm...
23:55:30jhMikeSthere's some tables in C code too I think
23:57:47jhMikeSThere four possible memory stalls in there as well but that's bugger to mess with on short notice
23:58:18 Quit Wiwie (Read error: 110 (Connection timed out))
23:58:52markunjhMikeS: should I also change the .align to .align 16 in the asm file?

Previous day | Next day