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

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

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

#rockbox log for 2008-01-23

00:01:20 Quit OlivierBorowski (Remote closed the connection)
00:01:57 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
00:03:01 Quit Rob222241 ()
00:04:27 Quit n1s ()
00:04:32 Part LinusN
00:05:51Bagderwe need a commit!
00:07:40 Quit jhulst (Read error: 104 (Connection reset by peer))
00:07:50JdGordon|wbuild 71?
00:09:42stripwaxQuick question - if I make a change to a single codec .c file, why does make take so long ?
00:11:36BagderJdGordon: as in the build after 70 which is what the table contains right now
00:11:56Bagderstripwax: due to makefile flaws
00:12:18stripwaxthanks tho
00:13:02JdGordon|wBagder: that i fugred :p which target?
00:13:05 Quit Redbreva ("Miranda IM! Smaller, Faster, Easier.")
00:13:08Bagdercowon d2 sim
00:18:28 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
00:18:48saratogai was just reading the logs
00:18:59saratogahas it been verified that the apple firmware doesn't vary the voltage?
00:19:22shotofaddsBuild table should update soon...
00:21:21stripwaxhm, recently been getting this a few times - "*PANIC* Updating size on empty dir entry 83" - something fishy in recent builds or do I need a fsck?
00:23:59Nico_Phaha I love the mrobe500i sim with the default theme
00:25:28Nico_PJdGordon|w: what are the button bindings?
00:26:03Nico_PI can't seem to be able to do anything in the WPS
00:26:56preglowsaratoga: i don't believe anyone verified anything like that, no
00:27:03preglowsaratoga: but it is kind of unlikely
00:27:27saratogapreglow: why not? they included a power management chip with software controlled voltage, while sandisk did not
00:27:34saratogapresumably there was a reason for that
00:27:36JdGordon|wNico_P: there are no buttons... you need to use the mouse
00:27:38Bagderoh, my host got the honor of building the first d2 sim build
00:27:40stripwaxhm, fsck is clean. it just seems to 'happen' from time to time and lock up that way.
00:27:40JdGordon|w16135 NE 85th ST
00:27:59preglowsaratoga: the chip does tons of other stuff than just that
00:28:03Nico_PJdGordon|w: and in the WPS?
00:28:06JdGordon|wnow you all know my secret hideout :'(
00:28:29*Bagder sends out "the team"
00:28:32JdGordon|wNico_P: umm.. iirc you need to have a wps which supports the regions... cant remember if that was coomed or not though
00:28:42saratogapreglow: what else does it do?
00:28:42shotofaddsBagder: pride of place on the left hand side :) what on earth is the sort order?!
00:28:51preglowsaratoga: but anyway, it shouldn't be too hard to find out, it wasn't too hard to find the pcf communication function in the bl, i don't think retailos should be any harder, than it should be easy to find out if retailos writes to those regs
00:29:01saratogaBuschel made it sound like rockbox barely touched the device
00:29:02JdGordon|w /dev/random :p
00:29:08preglowrockbox barely does
00:29:13preglowit sets a couple of voltages and uses it for alarms
00:29:15preglowthat's that, afaik
00:29:29saratogai don't think the bootloader would need to scale the voltage
00:29:37preglowno, but retailos would
00:29:38Bagdershotofadds: the sort order is very hard to see in the table itself, partly because the full title it uses are a bit weird, but also because the full titles aren't shown in the pics used as titles for each column
00:29:41saratogaprobably just inits it and hands off to the OS
00:29:57Bagderbut it is plain alphabetical order
00:30:18saratogapreglow: sorry misread
00:30:29 Join cool_walking_ [0] (n=cb3b81c3@gateway/web/cgi-irc/
00:30:36saratogai don't think looking at the bootloader is necessary, we have the data sheet for the device in the wiki
00:31:05saratogaunless you mean to find the ic2 function or whatever?
00:32:16 Quit webguest26 ("CGI:IRC")
00:32:18 Quit jhMikeS (Read error: 104 (Connection reset by peer))
00:32:43preglowsaratoga: yes
00:33:12preglowsaratoga: the bl has one function equivalent to a pcf_write which should be easy to identify calls to
00:33:13 Quit Arathis ("Bye, bye")
00:33:23preglowsaratoga: the dest is one of the parameters, unless i remember wrongly
00:33:48 Join Aevum [0] (
00:34:17preglowi'd love to look into it, but not right now
00:34:26***Saving seen data "./dancer.seen"
00:34:34 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
00:35:34stripwaxHa. Does profiling interfere with backlight on ipod video? (seems like it). Guess the timer is needed
00:35:56preglowi did make an overview over what pcf regs the bl touched, but i lost that info in a stupid "rm -rf" incident
00:36:41*shotofadds intended to get some flash reading code in SVN tonight instead of playing with the sim; it'll have to wait...
00:38:46shotofaddsI can read raw pages from the NAND flash, I "just" need to figure out the logical->physical address translation. I think that may take some time :/
00:38:47saratogapreglow: the datasheet says the package isn't surface mount, so in theory you (or someone with an Ipod) could open the player and stick a DMM lead to the VDD trace and see what the voltage is
00:38:56oboBagder: that's a lot of spam you get... you don't reject anything straight at the MTA level?
00:39:06saratogasorry "VDD pin"
00:39:27Bagderobo: nope
00:39:55*stripwax sighs - profiling has created *two* files both named profile.out :(
00:41:09saratogawow you got profiling to work?
00:41:12preglowsaratoga: i've tried opening my nano, and all i managed to do was warp the entire case...
00:41:12saratogawhat player?
00:41:45*Bagder fades away
00:41:51*preglow burns out
00:42:27hcsmy, my
00:42:52*Nico_P melts down
00:43:01preglowit's better to melt down than fade away!
00:43:03stripwaxsaratoga - ipod video (5g)
00:43:51saratogayou feel like opening it?
00:44:21stripwaxhad to hack some things to prevent IRAM-out-of-space, and make some changes to profile.c to actually build, and fix a #define redefinition in sudoko ..
00:44:50stripwaxfeel like opening it .. ?? not right now - unless it's completely and transparently reversible and has no chance of causing damage
00:45:22saratogai've only opened the 3G and it wasn't particularly hard but i can't say about the 5G
00:45:27preglow5g is worse...
00:45:33saratogathough people do upgrade the batteries so i assume its possible
00:45:34stripwaxneed more than a plectrum
00:45:40preglowit's the same thing as my nano, and i couldn't even bloody budge that open
00:45:50preglowi broke a plectrum trying to open it...
00:46:03 Quit shotofadds ("CGI:IRC")
00:49:22cool_walking_who the hell says "plectrum"?
00:49:30 Quit ender (" Documentation is like sex: when it's good, it's very good, and when it's bad it's still better than nothing.")
00:49:48preglowi just did
00:49:51stripwaxcool_walking_ - "guitar pick" is two words and involves more typing
00:50:12cool_walking_but it sounds cooler. plectrum is too technically
00:50:33preglowplus, it's "plekter" in norwegian, "plectrum" is the first thing i think of, heh
00:50:35krazykitwell this is a technical support and development channel.
00:51:58cool_walking_I broke one too, on my Video :) Butter knife got it open in the end.
00:52:53preglowi've only got a jazz iii left, doubt i'll break that, but it's too thick to work...
00:53:04preglowdon't dare use a knife, i'll just end up hacking it to pieces
00:53:11saratogai just use a razor sharp screw driver
00:53:27PaulPositionlol... Still is better than those big, felt-covered bass guitar picks.. :p
00:53:56 Quit scorche|w ("CGI:IRC (Ping timeout)")
00:57:14 Quit robin0800 (" Want to be different? Try HydraIRC -> <-")
00:58:20*stripwax googles for 5g dismantling tips..
00:58:29saratogatheres videos on youtube
01:00:40 Quit nanok (Remote closed the connection)
01:01:29 Quit z35 (Read error: 110 (Connection timed out))
01:06:00 Quit obo ("bye")
01:09:51 Part pixelma
01:11:30stripwaxok - here's a 5g profile output for vorbis.
01:12:52saratogastripewax: I tried for ages to do mp3 a couple months ago
01:13:00saratogadid you have to make any changes to get profiling working?
01:13:10 Quit JdGordon|w ("CGI:IRC")
01:13:23 Join mokkurkalve [0] (
01:13:35stripwaxyes - as i mentioned earlier
01:14:08stripwaxalac and mpegplayer wouldn't compile due to lack of iram (so I just disabled building them)
01:14:31saratogai vaugely remember doing that, but i still couldn't get it to work
01:14:39saratogai just ended up with blank output files
01:14:44saratogamaybe i did something wrong
01:14:49preglowstripwax: ahh, isn't there a .pl tool to process that into something readable?
01:15:02stripwaxpreglow - someone mentioned it earlier. it doesn't work for non-coldfire apparently
01:15:07preglowwell, shit
01:15:13preglowwould rule if it did
01:15:14saratogai'll take a look
01:15:18stripwaxdoes the above look .. reasonable .. or garbage?
01:15:27preglowstripwax: really, really hard to say, which is why i ask
01:15:30preglowand it'd be very interesting
01:15:35preglowi think lostlogic did this stuff
01:15:39 Join PaulJam_ [0] (
01:15:43stripwaxi'll post a patch for the changes I needed to make to get the rest to build
01:16:10preglowstripwax: could you post the tremor map file as well?
01:16:23stripwaxum, pastebin?
01:17:45preglowmail, ftp, pastebin, news, i can read them all :)
01:18:09stripwax(minimal) patch:
01:18:36stripwaxnot including the fact that I commented out alac, mpegplayer and doom from apps/plugins/SUBDIRS
01:19:11stripwaxnor the ci->profile_thread() / ci->profstop() calls in vorbis.c
01:19:19stripwaxsaratoga - oh, is that the bit you missed out, do you think?
01:19:31amiconnBagder: The svn table script needs info about 'rob'
01:19:54preglowstripwax: what's up with the sudoku change?
01:20:02stripwaxpreglow - vorbis map file:
01:20:13stripwaxpreglow - #define being redefined (as I mentioned earlier)
01:20:42stripwaxINDEX_MASK is defined when doing a profile build (didn't track down by whom/where)
01:20:51stripwax(didn't look, probably obvious if you know the code)
01:21:48 Quit BigBambi (Remote closed the connection)
01:21:58 Quit lee-qid (Read error: 110 (Connection timed out))
01:22:03preglowstripwax: sure, this looks like plausible data
01:22:56stripwaxwasn't sure about some of the insane counts/ticks for e.g. mdct_backward
01:23:29preglowi don't see why that shouldn't have a high tick count, heh
01:23:59stripwaxbut output was skipping pretty badly every few seconds (guess profiling itself needs time to housekeep)
01:24:23stripwaxnot that my last two statements are connected (they are not)
01:24:40preglowyeah, probably
01:24:44preglowdon't really know how it works
01:24:46stripwaxyeah, mdct_backward should have a high tick count, but seemed like a huge call count as well
01:25:11stripwaxbut then I don't know the format of that profile.out so maybe that's just how it is
01:25:22amiconnProfiling slows everything down to a crawl, that's expected
01:25:49amiconn(that's about all I know about it...)
01:26:04stripwaxokie doke
01:26:46saratogafor reference, wma spends ~55% of its time doing mdcts, so i would expect vorbis to do the same
01:27:10stripwaxsaratoga - yeah, I read a paper on Tremor this evening, seems about 50% of the time in mdct
01:27:19stripwaxwho knows if that's true on rockbox though
01:27:37preglowwell, we use the stock mdct...
01:27:38stripwaxwould be fun trying to do mdcts in parallel on both cores..
01:27:41preglowon a crappy compiler.....
01:28:15saratogadidn't a lot of the mdct get converted to ASM anyway?
01:28:17stripwaxpreglow - oh, no-one tried to hand-tune it yet? (I say that like it's trivial..)
01:28:23*stripwax thought so
01:28:56preglowsaratoga: eh, not that i know of
01:29:08preglowsomething did happen for libmad, but i don't think so for tremor
01:29:12*stripwax checks tracker..
01:29:22preglowthat's wrong...
01:29:23preglowit was
01:29:25 Quit tvelocity ("Αποχώρησε")
01:29:51amiconnWorking on the same data from both cores on PP would be a bad idea afaiu
01:30:05stripwaxnot same data though right?
01:30:24stripwaxpreglow - did this actually get committed?
01:30:52preglowstripwax: two out of three did
01:30:59preglowthe last one was bugged, the coupling one, afaik
01:31:20stripwaxcoupling? are you thinking of a different patch?
01:31:24amiconnWith the PP multicore architecture, the best thing to do is to let both cores work on different things, as little related as possible
01:31:25 Quit ompaul (Client Quit)
01:31:26preglowi want that thing closed, btw, if markun fixed it
01:31:43 Join TiMiD[FD] [0] (n=TiMiD[FD@
01:31:49TiMiD[FD]hi there
01:32:08preglowi think those got commimted
01:32:17preglowbut got reopened because the mdct_arm thing broke gigabeat
01:32:20preglowbecause of lack of iram, i think
01:32:26stripwaxah, I see
01:32:33saratogaugh stupid perl default variables
01:32:37stripwaxcouldn't see that meta-info from the tracker..
01:32:48saratogawhere does "while (<MAP_FILE>)" store the string?
01:32:49 Quit PaulJam (Read error: 110 (Connection timed out))
01:32:59preglowsaratoga: in $_
01:33:07preglowthe default variable, as you said
01:33:17saratogai can never remember what any of them are called
01:33:43preglowi remember that one :)
01:34:02preglowi'll need to check with markun on this one
01:34:15saratogaif code uses built in perl vars i typically just give up and rewrite it properly
01:34:19stripwaxok, gotta go - good night
01:34:24preglowme too, seeya
01:34:35markunpreglow: I'm here. I didn't find a nice wat to use .text only for the Gigabeat
01:34:41amiconnpreglow, stripwax: Should be easy to make the section in mdct_arm.S conditional...
01:34:47preglowmarkun: ahh, good
01:35:06amiconn#include "config.h"
01:35:09preglowamiconn: don't we just #ifdef in other targets?
01:35:30markunamiconn: I did that, but don't remember why it didn't work
01:35:36amiconn#if CONFIG_CPU == <gigabeat_type_don't_remember>
01:35:51amiconn.section .icode,"ax",%progbits
01:36:01preglowmarkun: well, could you try again? :)
01:36:02markunI didn't see any speedup from this btw
01:36:06markunpreglow: not now
01:36:11markuncan't you?
01:36:19amiconnI do this in several places, for PP502x vs. PP5002
01:36:25 Quit waldo ("Konversation terminated!")
01:36:30preglowtheoretically, but i can't really test the result...
01:36:40preglowthere's no rush anyway, i was on my way to bed
01:36:48markunme too :)
01:36:54markungood night everyonw
01:36:59preglowyup, gnight
01:37:02amiconnNot sure whether USE_IRAM is defined there - if it is, it's even easier
01:37:10preglowamiconn: oh, indeed
01:37:20preglowi expect it is
01:37:31preglowbut yeah, i'll check it out tomorrow/later
01:37:52amiconnAnd if fact I don't think iram is important there, except for PP5002 and probably the PNX0101
01:39:21amiconn(otherwise tomal wouldn't have use iram, methinks)
01:39:24 Join Rob2222 [0] (
01:41:24 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
01:45:40amiconnoh my
01:53:28 Quit PaulJam_ (".")
01:54:33 Quit XavierGr ()
01:54:40amiconnmdct_arm.S got committed and is used on all arm targets except gigabeat
01:56:37 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
01:56:51 Nick fxb is now known as fxb__ (
02:01:28 Quit DogBoy ("Leaving")
02:03:34 Join DrDnar [0] (
02:06:44 Join z35 [0] (
02:07:34 Quit BobShield (Read error: 104 (Connection reset by peer))
02:09:04 Quit saratoga ("CGI:IRC (EOF)")
02:10:04jhMikeSamiconn: around now
02:10:40amiconnHow many 'nop's did you add in main.c?
02:11:02amiconnAnd did you ever need to change the amount of nops to get it working again?
02:11:41amiconnI made an interesting observation on my 2nd Gen - I added one single nop, and all builds ever since didn't crash
02:11:56jhMikeSIt wasn't constant at the time I was checking
02:12:23amiconnThis might just be pure luck, but it might also mean that the alignment problem theory doesn't hold
02:12:54jhMikeSsomething seemed to change recently where fewer builds crash since I haven't had to do it lately
02:13:30amiconnHmm. I have that 'nop' still in there (PP5002 only of course)
02:13:51amiconnShould probably try without - but if it crashes, a periodicity test might be helpful
02:14:45jhMikeSstill, it's completely dependent on having the cache turned on
02:14:53 Quit DerDome ("Leaving.")
02:15:14amiconnI'm not sure whether that's possible, but could it be an init race between the cores somehow?
02:15:40amiconnI mean, even if the cop isn't used, it is initialised in crt0 as well as the cpu
02:16:02jhMikeSno, because never waking the cop outside of crt0-S makes no difference
02:16:37amiconnYeah, but the cop *does* something in crt0, right?
02:16:40jhMikeSthe cores are strictly serialized in the init up to the end of the thread init
02:17:18amiconnAnd if that something isn't completely finished and the cpu starts doing something it better shouldn't do yet, it could crash
02:17:25jhMikeSjust sets up it's own stacks and such and waits for the CPU to complete. the cache is still disabled at that point.
02:17:56amiconnIs the cache write-back or write-through?
02:17:56jhMikeSmemory mapping is serialized too
02:18:13 Part Aevum ("Saliendo")
02:18:52jhMikeSnot sure. how would I tell?
02:19:56 Quit MethoS- ("Konversation terminated!")
02:22:38jhMikeSI think it's writeback
02:27:33amiconnIs the cache enabled when crt0-pp is entered (i.e. by the bootloader)?
02:27:40jhMikeSRAM is definitely not automatically coherent after writing and it has dirty flags. In any case, the COP seems to have no bearing on this at all and the problem persists even if the COP is never woken after the bootloader
02:28:08 Quit tedrock (Client Quit)
02:28:13amiconnNo, but you said not enabling the cache fixes the problem
02:28:33jhMikeSI checked and it is not enabled at that point...not until init_cache is called.
02:28:52amiconncrt0-pp disables the cache in case it was enabled before (iiuc), but it does not flush/invalidate it
02:29:06amiconnSo if it's reenabled later, it might not be consistent with ram
02:29:25jhMikeSthe bootloader makes it coherent before calling the firmware
02:29:40jhMikeS_if_ it is enabled. but it isn't
02:32:25amiconnHmm. If it were, the code running before crt0-pp line 133 would populate the cache (consistent with ram), then it's disabled, and ram contents is changed (by bss init, stack munging etc)
02:32:59amiconnAnd if the cache is enabled later, its (now inconsistent) content could be used instead of fetching from ram
02:33:14jhMikeSthe init, like retailos, invalidates all the lines before enabling it
02:33:28amiconnThis might even be a bug in the PP5002 cache controller, and explain why it doesn't hit PP502x
02:33:44jhMikeSso, no writeback at all is perfomed before enabling it
02:34:29***Saving seen data "./dancer.seen"
02:34:56amiconnYeah, writeback wouldn't be necessary for such a bug to hit. Just using cached data instead of fetching from ram would, in case ram contents changed while the cache was disabled (but not invalidated)
02:35:56amiconnBut if it's invalidated before enabling, all should be well... hrrmmmm
02:38:53jhMikeSI tried messing with other regs with no change in behavior.
02:47:32 Join kkurbjun [0] (
02:51:28amiconnA build without the nop crashes immediately. One nop and it works...
02:55:53 Join telliott [0] (
02:57:04telliottI have a question for someone who has ripped a lot of CDs.
03:01:02 Quit Nico_P (Remote closed the connection)
03:03:26amiconnjhMikeS: Ok, I now tested various 'nop' counts. The effect is periodic.
03:03:46amiconn0, 2, 4, 6 nops - crash. 1, 3, 5, 7 nops - OK
03:04:31amiconnThe crashes come in 2 different flavours. 2 and 6 nops do the familiar mix of 'data abort', 'undefined instruction' etc
03:04:49amiconn0 and 4 nops cause an 'empty lcd' crash
03:05:52 Join avh22 [0] (
03:06:30jhMikeSanother thing I hadn't looked into was exactly where in the bin does the nop have no effect at all?
03:06:48telliottDo you use VBR or CBR when encoding mp3?
03:08:02 Quit soap (Read error: 104 (Connection reset by peer))
03:08:15amiconnjhMikeS: If you want to do the same experiment - I've put them at top of app_main(), directly before the init() call
03:08:44 Join WeeJay [0] (
03:09:27amiconnI think some data isn't cache line aligned that better should. Moving the nop(s) around in the binary should help spotting it
03:09:27WeeJayanyone active in here?
03:09:51amiconnBut I won't do that right now. 3:09 am here...
03:10:04WeeJayI just purchased a sana e260. Can someone please help me figure out the easiest way to determine if its v1 or v2?
03:10:09 Join qwedsa_ [0] (
03:10:50karashatalook for a small "v2" next to the model number on the back of the device, perhaps?
03:11:38jhMikeSamiconn: The main stack is safe of course since it's IRAM and the effect seems to be incoherent code.
03:11:57WeeJaykarashata: negative
03:12:10karashataokay, check the firmware version
03:12:24amiconnYeah, but there are other stacks, and .data, and .bss
03:12:24 Join Rob222241 [0] (
03:12:24WeeJayDoes the device display it
03:12:43amiconn(which of course contain the other stacks)
03:13:07amiconnSome other stacks are also in iram, but not all of them
03:13:20jhMikeScould check which thread is current at the time
03:13:24advlaptop2019WeeJay, it is in the info in settings
03:13:42 Quit csc` ("Powering Off")
03:13:47WeeJayadvlaptop2019: thanks....i am literally just openign the box here :-D
03:13:50WeeJayvery excited
03:14:11 Quit Rob2222 (Read error: 104 (Connection reset by peer))
03:14:49 Join soap [50] (n=soap@rockbox/staff/soap)
03:15:00jhMikeSbut then COP crashes at times (when allowed to be active) and that touches no other stacks until a thread is started
03:16:00WeeJayadvlaptop2019: Woot!!!! Version 01.02.18A
03:16:19advlaptop2019WeeJay, you have a v1
03:16:30WeeJayrock box time!
03:16:39WeeJaythis was the second last unit there
03:18:52 Quit zicho (Remote closed the connection)
03:21:48 Join jcollie_ [0] (
03:22:01 Part avh22
03:22:50 Quit jcollie (Remote closed the connection)
03:23:05 Nick jcollie_ is now known as jcollie (
03:23:33 Join jac0b [0] (
03:23:53jac0bis anyone else having a problem with svn and the multifont patch
03:24:41jac0bI get hunk errors on gui/list.c
03:25:45 Quit qwedsa (Read error: 110 (Connection timed out))
03:26:38 Join tedrock [0] (
03:28:16jac0bI think I might have fixed the error but how can i tell
03:34:15 Join arken0493 [0] (
03:34:48arken0493I'm curious. I've been using Rockbox for 4 months now, and I want to give back to the community. How can I help?
03:36:24 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
03:36:52jac0byou could do this
03:36:55arken0493So is everyone afk?
03:37:01arken04931) I don't have money
03:37:13arken04932) I'm 14, I'm not allowed to have a paypal by law.
03:37:20arken0493That's what I was more interested in,.
03:37:22 Join hannesd___ [0] (
03:37:39 Quit tedrock (Client Quit)
03:37:51jac0bthe basics is to learn C
03:38:09jac0bthat is what I think rock is coded in
03:38:23arken0493I know some C.
03:38:48arken0493I'm not good from starting at scratch. What I'm more interested in, is how there are games, and how I could make my own .rock file.
03:39:22jac0byou might want to ask some others on here I know very little
03:39:31arken0493Everyone's afk
03:39:42krazykitarken0493, well, it's IRC. not everyone is on at once
03:39:45jac0bbut if you know C that is a good start
03:39:56arken0493Very little.
03:40:08arken0493I haven't coded for it in a year, and my memory is horrid
03:40:20 Join tedrock [0] (
03:40:37krazykitarken0493, you could buy a C book for beginners or something, then start looking at the rockbox source and the developer resources.
03:41:54 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
03:41:54 Quit hannesd (Read error: 110 (Connection timed out))
03:41:55 Nick hannesd___ is now known as hannesd (
03:42:15arken0493where do I find the source to this? (not the OS its self, but the applications?)
03:42:53cool_walking_arken0493: They're all under /apps/plugins/ in the source.
03:43:04arken0493Oh, so it has everything! Ok.
03:43:10 Quit conando (Read error: 110 (Connection timed out))
03:43:38cool_walking_Read this:
03:43:59arken0493That's what I was looking for!
03:45:02arken0493wait, so how do you compile in windows? These are all linux commnads
03:45:46arken0493Oh. I see. It suggest VMware.
03:46:06 Join ashes [0] (
03:46:44ashesdoes rockbox on the iriver h300 support any kind of video? google isn't telling me
03:47:12 Join miepchen^schlaf [0] (
03:47:13karashataashes: Rockbox only supports MPEG-1 and MPEG-2 video
03:47:48ashesthat's fine, and it plays on an h300?
03:48:10*karashata nods
03:48:10 Part arken0493
03:48:40 Quit jac0b ("Ex-Chat")
03:49:41 Quit Thundercloud (Read error: 104 (Connection reset by peer))
03:50:41 Join SirFunk [0] (
03:51:11ashes"missing packet start" 3E2B F51C
03:51:18 Join ctaylorr [0] (
03:51:30Lloreanashes: You aren't using a current build then?
03:52:12 Quit tedrock (Client Quit)
03:54:49 Join tedrock [0] (
04:03:04kkurbjunjhMikeS: is there anything special about the mutex locks on a per target basis, I cannot get the m:robe 500 port to boot anymore, I'm not positive that it's the mutex changes, but I'm trying to pinpoint what is causing this freeze
04:05:19 Quit midkay ("Leaving")
04:08:17 Join midkay [0] (n=midkay@rockbox/developer/midkay)
04:08:51jhMikeSkkurbjun: not really. the basic variants are: 1) single core 2) dual core w/swp 3) dual core w/peterson's algorthm
04:08:52 Quit daurnimator ("Cyas later...")
04:11:03kkurbjunhmm, oh well - I guess I'll mess with it whenever I get time again - thanks for the feedback
04:11:46jhMikeSbe sure there are no interrupts are unmasked before kernel init or are all masked before kernel init and that threading objects like queues, etc. are initialized before anything could possibly require them
04:12:11 Quit PaulPosition ()
04:14:04jhMikeSoh, never depend on yield or sleep with interrupts disabled having the results you might want. it's really not a good policy to do so since you're turning control over to unknown code at that point anyway which could do anything with cpsr.
04:16:31kkurbjungreat, thanks for that, I'll keep an eye out
04:18:49 Quit jhulst (Read error: 110 (Connection timed out))
04:30:01 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
04:34:31***Saving seen data "./dancer.seen"
04:35:08 Quit karashata (" HydraIRC -> <- IRC with a difference")
04:36:49 Join daurnimator [0] (i=daurn@unaffiliated/daurnimator)
04:41:32 Quit BHSPitMonkey (Remote closed the connection)
04:42:25 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
04:49:49 Quit JdGordon ("Konversation terminated!")
04:55:28 Quit Tyroazard ()
05:01:46 Join spongslog [0] (
05:01:59spongsloganyone answer a newb question for me?
05:05:18 Quit ol_schoola ()
05:09:37*spongslog hears crickets
05:11:07cool_walking_spongslog: read
05:11:16spongslogyes I read the FAQ
05:11:19spongslogand searched the forums
05:11:30cool_walking_especially point 9...
05:11:50spongslogI read it worng
05:11:58spongsloganways. widows is asking for a driver
05:12:07cool_walking_What player?
05:12:16spongslogsansa 260r
05:12:29spongslogit seems to have installed correctly
05:12:41spongslogon the player
05:13:06spongslogbut windows device manager lists it with the yellow question mark
05:13:36spongslog"rockbox media player" is how it shows
05:13:48advlaptop2019spongslog, reboot into OF
05:13:53cool_walking_Rockbox doesn't have it's own USB mode yet
05:13:59cool_walking_You have to use the OF
05:14:25spongslogOF stands for?
05:14:32cool_walking_Original Firmware
05:15:14spongslogand switch back to Rockbox when not transfering files
05:15:52spongslogI will figure out how to do that :)
05:18:10 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
05:18:31 Join miepchen^schlaf [0] (
05:19:47 Quit Horscht ("User was distributing pornography on server; system seized by FBI")
05:21:04spongslogok that worked. Thanks again
05:21:15 Join JdGordon [0] (
05:21:17cool_walking_No worries
05:23:27 Join DogBoy [0] (
05:25:12 Quit telliott ("Leaving")
05:26:41 Join TotallyInfected [0] (
05:32:21 Join ctaylorr_ [0] (
05:32:56JdGordonanyone around that has a build server?
05:34:26 Join toffe82 [0] (
05:37:25 Quit ctaylorr (Read error: 104 (Connection reset by peer))
05:38:35 Quit WeeJay ("Leaving.")
05:42:05 Join advlaptop2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
05:42:29 Quit advlaptop2019 (Nick collision from services.)
05:42:30 Join psycho_maniac [0] (i=psycho_m@
05:42:31 Nick advlaptop2019_ is now known as advlaptop2019 (n=advcomp2@unaffiliated/advcomp2019)
05:44:02psycho_maniacHow does "No Bookmark Found!" when automatically loading a bookmark and it is not found sound?
05:44:05 Quit fyrestorm (Read error: 113 (No route to host))
05:44:27JdGordonfound sound bound hound!
05:45:06 Quit spongslog (Read error: 104 (Connection reset by peer))
05:45:41psycho_maniacI also see theres new languages added to the FunnyLangs wiki page.
05:46:29psycho_maniacwell new for me. theres pirate, elmer fudd, elmer fudd in brazil, and more
05:51:22cool_walking_I did not know such a wiki page existed
05:51:50psycho_maniacI dont think those langs are supported thouhg. just something to pass the time.
05:55:37 Quit psycho_maniac (" quit")
05:56:50cool_walking_... a little disappointed with the pirate translation. Though I imagine it must be near-impossible to convert technical terms to pirate.
05:57:33cool_walking_Could have thrown a "ye olde" every now and then - "ye olde low shelf filter"
05:58:34cool_walking_oh wait.. do pirates say "ye olde" or is that olden-timey english dudes?
06:06:12DogBoythere's no such thing as pirates
06:13:27cool_walking_Sure there are.
06:23:48TiMiD[FD]could someone who has a target with a wheel kindly test a plugin for me ?
06:26:20cool_walking_I'm not familiar with all the rockbox targets, so... you mean a scroll wheel, like the iPods? Not some weird target with a wheel on it?
06:26:39TiMiD[FD]ipods or sansa c200, there may be others
06:26:51cool_walking_Yah I've got an iPod Video
06:26:56cool_walking_What do I do?
06:26:56TiMiD[FD]ok !
06:27:06TiMiD[FD]I compile and I send you the plugin :)
06:27:20TiMiD[FD]but it may crash \(^^
06:27:50TiMiD[FD]32mb ?
06:28:26TiMiD[FD]ok then it's compiling
06:28:39cool_walking_What's it called?
06:28:51TiMiD[FD]I just want to make sure that I won't commit code that doesn't work on the target like I did yesterday ...
06:28:51cool_walking_Wait... 32MB?
06:29:07TiMiD[FD]it asked me for how much ram to compile
06:29:13cool_walking_ah ok
06:29:32cool_walking_I have a 64MB RAM version, but 32MB builds work fine too..
06:29:45TiMiD[FD]I'm not familiar with the ipod video so ..
06:29:54TiMiD[FD]I'll compile for 64 then
06:31:43TiMiD[FD]I would have loved to test that on my iriver as well but I'm at work and I forgot the usb cable
06:33:30cool_walking_It's not gonna delete all my files is it? Do I need to back up?
06:34:23TiMiD[FD]no, it will only brick your player
06:34:34***Saving seen data "./dancer.seen"
06:34:41cool_walking_ah okay
06:35:05TiMiD[FD]no worries :)
06:35:45cool_walking_I don't know what this plugin is supposed to do... It could be one of those ones that fiddles with files. I'm paranoid.
06:35:46 Quit cool_walking_ ("CGI:IRC")
06:35:56 Join cool_walking_ [0] (n=cb3b81c3@gateway/web/cgi-irc/
06:36:17 Quit DrDnar (Read error: 110 (Connection timed out))
06:36:56TiMiD[FD]no it's only the fire plugin
06:37:46TiMiD[FD]if it was something risky, I wouldn't ask like that
06:41:11cool_walking_Okay, what am I looking for?
06:42:50cool_walking_It seems to be running fine.
06:43:07TiMiD[FD]nothing :)
06:43:26TiMiD[FD]just to see if it doesn't crash, if it runs fine and if the fire height can be contrlled with the wheel
06:43:36TiMiD[FD]thank you for the test anyway
06:43:47 Part toffe82
06:43:48 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
06:44:17TiMiD[FD]pixelma: yo ! were you searching for me ?
06:45:46cool_walking_Is scrolling anti-clockwise supposed to make the fire higher?
06:46:04TiMiD[FD]oh maybe
06:46:17TiMiD[FD]would it be more natural the other way ?
06:46:36pixelmaTiMiD[FD]: yes, I actually have no time and was just reading the logs before I leave for work - please don't use plugin actions... I still have trouble getting the controls to work correctly with at least 2 plugins that you converted (clock, jackpot on c200) and the one you changed majorly (metronome on various Archos)
06:47:08cool_walking_I was just checking it was working and not just psychosomatic (I haven't used "fire" often)
06:47:21pixelmaTiMiD[FD]: probably not (about cool_walking_'s question, you can check the manual it's quite up to date there)
06:47:30pixelmathe manuals
06:47:50TiMiD[FD]I just corrected the code to work with the action's lib
06:48:14TiMiD[FD]seems to work flawlessly now
06:49:26TiMiD[FD]I don't use jackpot or metronome, but clock seems to work pretty nicely as well
06:49:41TiMiD[FD]are there some problems with this one ?
06:50:33cool_walking_talking to me? problems with what one?
06:50:39TiMiD[FD]ah no to pixelma
06:51:16TiMiD[FD]cool_walking_: do you have problems with the controls inside the clock plugin ?
06:51:37cool_walking_ah hold on, just connected it to the computer again..
06:52:21cool_walking_It said "Error loading settings"
06:52:46TiMiD[FD]maybe the first time you used it
06:53:00cool_walking_What are the controls supposed to do? Scrolling either way does a zoom in/out.
06:53:31TiMiD[FD]zoom ?
06:53:50TiMiD[FD]I didn't knew I put a zoom function into the clock plugin
06:53:52cool_walking_Yup the error was only the first time.
06:54:12cool_walking_It's only got two levels - in and out.
06:54:39TiMiD[FD]ah maybe you're talking about the clock switching from square mode / circle mode ?
06:54:47cool_walking_Yeah that's it
06:54:51TiMiD[FD]ok :)
06:55:12TiMiD[FD]now if you press left/right it should switch betweend analog/digital/binary
06:55:33cool_walking_Yep :)
06:55:56TiMiD[FD]can you access the menu and quit the plugin without removing your player's battery ?
06:56:39TiMiD[FD]ok so no problem with the clock plugin controls on the ipod video
06:57:10pixelmaTiMiD[FD]: try a c200 sim and change try to change the skin with a short press of up and compare with a long press and then try to figure out why is this so
06:57:32pixelmaabout plugin actions in general, please read the logs
06:58:15TiMiD[FD]ah yeah it doesn't behaves very nicely
06:58:41TiMiD[FD]I read the logs, you were talking about a conflict, but that wasn't the problem in the fire plugin
06:58:57TiMiD[FD]I'll try to get what's wrong with the clock plugin
06:58:58pixelmabtw. I know why but just want to show that the action system hides things, makes them harder to understand and I don't know a nice and easy solution (using exceptions is not nice and is against the purpose)
06:59:23TiMiD[FD]well I agree on that point
06:59:53TiMiD[FD]but it also makes the code easier to read and to develop when you don't have the target
07:00:38TiMiD[FD]since you can suppose that the action will have been choosen to match a key that makes sense
07:01:00pixelmano, for me it's not easier to read because I have to look it up in 4...5 places and not 2 (max) and you lose control
07:01:05TiMiD[FD]although it seems that the key mappings are not always logical
07:01:50pixelmait's not a keymap problem there, it's a combining of contexts... (like it is pitily done too often)
07:04:59pixelmaand please don't use plugin actions - did I already say that? If you read logs, you'll see that some developers are for removing it or at least for a major think through of the current system
07:05:04TiMiD[FD]wow the c200's up button is also used for menu
07:09:10TiMiD[FD]ok so I'll try to fix that
07:10:03TiMiD[FD]would be nice to have actions only defined as generic for the directions and the quit button, and to leave the rest to the plugin
07:10:47pixelmaTiMiD[FD]: how do you want to fix that?
07:11:09TiMiD[FD]using my brain for the moment
07:11:57pixelmaif you change the keymap in the lib file it could break other plugins, exceptions are not nice etc...
07:12:04TiMiD[FD]yes I know
07:12:28TiMiD[FD]the problem is that they put button_up on the c200 for the menu
07:12:39TiMiD[FD]on the ipod they use the wheel instead
07:13:14TiMiD[FD]there are not so much plugins using the action lib
07:13:41TiMiD[FD]could be an option to swithc back to the button system as before
07:13:59TiMiD[FD]but it's so bothersome to define the same buttons for every target on every plugin
07:14:52pixelmaanyways, I gotta leave. I'd rather see the plugin meved back to per keypad defines (even if it is many lines but you have more control) that's a personal opinion though but me try to get a discussion in the ml going wasn't very successfull (somewhen in fall of last year or so)
07:15:22TiMiD[FD]I'll think about it
07:16:55pixelmaat least you were around now ;)
07:17:06 Part pixelma
07:38:53 Join karashata [0] (
07:41:05 Join xushi [0] (n=xushi@unaffiliated/xushi)
07:47:33 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:54:35 Quit TotallyInfected ()
08:03:05 Join Rob2222 [0] (
08:07:39 Join ol_schoola [0] (
08:14:51 Join conando [0] (
08:21:07 Quit Rob222241 (Read error: 110 (Connection timed out))
08:21:56 Join w23 [0] (
08:22:44 Quit karashata (" HydraIRC -> <- Chicks dig it")
08:26:09 Quit cool_walking_ ("CGI:IRC")
08:34:39***Saving seen data "./dancer.seen"
08:37:54 Join ol_schoola_ [0] (
08:38:21 Join ender` [0] (
08:38:32 Nick linuxstb__ is now known as linuxstb (
08:38:35 Quit ol_schoola (Read error: 104 (Connection reset by peer))
09:01:25 Join n1s [0] (
09:02:01 Quit tedrock (Read error: 113 (No route to host))
09:02:02 Join tedr0ck [0] (
09:02:54 Quit tedr0ck (Client Quit)
09:09:00 Quit freqmod_nx (Read error: 104 (Connection reset by peer))
09:09:07 Join freqmod_nx [0] (
09:09:35 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
09:15:14 Quit TMM ("Ex-Chat")
09:19:52 Join Axio [0] (
09:22:23 Join tedrock [0] (
09:22:43 Join Zagor [242] (
09:22:48 Join petur [50] (n=petur@rockbox/developer/petur)
09:25:26 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
09:25:44TiMiD[FD]commited some changes, hope it'll be fine that time ...
09:31:58 Join cool_walking_ [0] (
09:42:16 Join Casainho [0] (
09:46:11TiMiD[FD]probably be back in 2h if there is a problem with my commit ...
09:46:13 Quit TiMiD[FD] ("leaving")
09:54:29 Join DerDome [0] (
09:56:49 Quit n1s ()
10:07:59 Quit Axio (Remote closed the connection)
10:09:29 Quit amiconn (Nick collision from services.)
10:09:36 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
10:11:00 Join kugel|s [0] (n=54bd443b@gateway/web/cgi-irc/
10:11:54 Join Casainho_ [0] (
10:15:32 Join pixelma [0] (n=d5494d4d@gateway/web/cgi-irc/
10:16:12LinusNjhMikeS: there?
10:16:17*pixelma starts crying
10:17:19peturjust revert it again, he'll give up one day ;)
10:17:26 Join waldo [0] (
10:18:57pixelmaI can't see how making things more complicated, introducing and combining yet another context can fix issues (and/if not introducing more problems), pitily (or luckily) I can't test on any targets until the evening (forgot USB cables :\
10:19:43LinusNwhat would be the better solution?
10:20:04pixelmaanoher question: where has the identify field in the "advanced" login for the webclient gone?
10:20:42Zagorpixelma: identify?
10:21:08pixelmaLinusN: scrap plugin button actions in their current way (and/or use it much more carefully)
10:21:40peturZagor: in the past, one could login and be identified all in one
10:22:15pixelmaZagor: as petur said, some time ago it was possible to type your password in a specific field and already connect as identified (as petur said)
10:22:27Zagorhmm, ok I'll take a look
10:22:34petur(I only said it once)
10:23:35peturpixelma: you can connect to freenode on port 8000, most firewalls pass it without problems...
10:24:43Zagorpassword added back
10:24:52 Join Casainho__ [0] (
10:25:12pixelmapetur: hmm? I don't have a connection problem but don't want to type the identify line if it was possible to do it while logging in anyways. :)
10:25:17pixelmaZagor: thanks
10:25:44peturpixelma: I mean use a normal chatclient ;)
10:26:12peturZagor: it looks like 'rob' needs to be added to the svn name translation table
10:26:22pixelmawell, I don't want to install something here
10:27:06peturSome clients don't need install, just put some files there, or even on a usb stick
10:28:10Zagorpetur: done
10:28:25pixelmapetur: need to google for it... or can you recommend one? ;)
10:28:27peturgoodie :)
10:28:56peturpixelma: nettalk can do it iirc (and it is German...)
10:29:09 Quit Casainho (Read error: 110 (Connection timed out))
10:29:19 Nick Casainho__ is now known as Casainho (
10:29:33pixelmathanks, hopefully their german interface is better than CGI:IRC's ;)
10:30:21peturpixelma: it is at the bottom of this page:
10:30:39pixelma"Nettalk downloaden" - oh yeah ;)
10:33:23 Quit kugel|s ("CGI:IRC")
10:34:42***Saving seen data "./dancer.seen"
10:37:23 Join Arathis [0] (
10:38:41jhMikeSLinusN: yes
10:40:02LinusNseen this?
10:40:28LinusNlooks like your ata spinlock change caused some problems
10:42:48 Quit Casainho_ (Read error: 110 (Connection timed out))
10:43:11pixelmagotta go again, see you later
10:43:46*petur waves
10:43:48jhMikeShmm...that sounds to me especially since mine is behaving just fine but I'll look around
10:43:54jhMikeSsounds weird
10:45:01 Quit tedrock (Read error: 113 (No route to host))
10:45:16pixelmawow I can't quit... if I type "/quit" it gives me "*** Access to that command is blocked"
10:45:49jhMikeSif buffering is fragile against thread timing changes, that could explain "wierdness" otherwise this should be sound
10:46:15 Join Mathiasdm [0] (
10:47:55 Join TotallyInfected [0] (n=ebola@
10:49:11LinusNspeaking of buffering, we seem to have serious issues with buffering in general (crashing X5's, insane disk activity on ipods etc)
10:49:58 Quit pixelma ("CGI:IRC (EOF)")
10:50:22GodEaterindeed - my ipod took 1.5 minutes to change a theme last night
10:50:26GodEaterthe disk activity went insane
10:50:30jhMikeSan experiment on using DMA and interrupts for all I/O on gigabeat made playback behave strangly and play things in the wrong order even though the file reads are ok
10:52:12 Quit Siku (Read error: 131 (Connection reset by peer))
10:52:27 Join Siku [0] (
10:52:49*amiconn didn't observe any buffering strangeness so far
10:53:03GodEateramiconn is a lucky fellow then :)
10:53:45amiconnThat said, I'm using my H180 most of the time, but occasionally use my other targets too (like the H10 twice last week)
10:54:24GodEateryou do also use the least disk intensive themes do you not ?
10:55:31jhMikeSother code I know to have sound threading use showed no side effects from scheduling order changes
10:57:53 Join Casainho__ [0] (
10:57:54Nico_PGodEater: did you change themes while music was playing?
10:59:16amiconnGodEater: Of course not. I either use the default, or my own wps'es (which are bitmap free)
10:59:41amiconnMeh, misread a little, but you get the point ;)
11:00:45 Join lee-qid [0] (
11:01:37LinusNalso, we have many reports about buffer underrun problems (audio dropouts) on the PP platforms, they seem to be caused by the new disk poweroff code
11:01:47GodEaterNico_P: no, I deliberately made sure music playing was off
11:02:02Nico_Pthen it's a disk problem, not a buffering problem, isn't it?
11:02:11LinusNnot necessarily
11:02:19*GodEater goes to double check
11:02:35LinusNthe buffering is there to handle slow disks, isn't it?
11:02:35Nico_PLinusN: I mean buffering as in buffering.c
11:03:19LinusNbut it might be something else, like starving the pcm code
11:03:36amiconn"new" disk poweroff code?
11:03:49amiconnDisk poweroff was committed months ago, way before MoB
11:04:01GodEater53 seconds to change a theme, with no music playing
11:04:15GodEaterNico_P: the commit we suspect is 15444 isn't it ?
11:04:23*petur gets out his h380
11:04:41Nico_PGodEater: yes
11:05:07Nico_Psomeone reported that it was the culprit on the tracker entry about audio dropouts
11:05:30 Quit waldo ("Konversation terminated!")
11:06:09GodEaterafter that initial theme change, subsequent ones are done in about 10 seconds (depending on number of bitmaps involved, rockbox default is 3 seconds)
11:07:02GodEaterusage of AA seems to make the dropouts worse
11:07:08GodEaterthough I still get them even without using it
11:07:32Nico_PI have to go, but I'll read the logs when I get back
11:09:43LinusNamiconn: do you subscribe to the flyspray mailing list?
11:10:02amiconnGodEater: On my 2nd Gen, loading iCatcher takes ~3 seconds (which I consider too slow for actual use). iCatcher has ~50 bitmaps
11:10:11amiconnLinusN: No I don't
11:10:31LinusNi guess that's why you don't know about this then
11:10:35LinusN"Ok, I found out that the bug has started from the r15444, which doesn't introduce any change on buffer managing but on disk spinning as far as I understood. In my opinion, MoB isn't the real cause of this bug."
11:10:45jhMikeSNico_P: if yield is used inside a mutex, it should be replaced with priority_yield() (a yucky hack that works for now)
11:11:08amiconnLinusN: That's clearly a buffering issue.
11:11:13LinusNamiconn: probably
11:11:13Nico_PjhMikeS: will try to look into it this evening
11:11:21amiconnWith disk poweroff, spinup takes a little longer than without
11:11:27Nico_Pamiconn: buffering as in buffering.c or as in ata buffering?
11:11:34LinusNas in buffering.c
11:11:37amiconnThe swcodec buffering code doesn't adapt to spinup time
11:12:22amiconn...unlike the hwcodec buffering code, which does this for years. One point I complained about numerous times
11:12:48Nico_Pamiconn: I'd love directions as to where to look int the HWCODEC code
11:12:56Nico_Por I'd love patches even more :)
11:13:03*Nico_P really has to go now
11:13:51amiconnmpeg.c handles this stuff for hwcodec
11:14:12LinusNNico_P: see recalculate_watermark() in firmware/mpeg.c
11:14:40GodEatersurely the code in buffering.c isn't used when switching a theme though ?
11:15:19LinusNGodEater: that could be an issue of two threads using the disk simultaneously
11:15:28LinusNi.e a locking issue
11:15:40GodEaterbut if I see the issue whilst NOT playing music - what thread is also accessing the disc ?
11:16:27LinusNmaybe the wps loader should boost the cpu?
11:16:56 Quit Casainho (Read error: 110 (Connection timed out))
11:17:18GodEaterthe cpu is usually boosted while playing right ?
11:17:29LinusNthat depends on the codec
11:17:33GodEatermp3 ?
11:17:45LinusNmp3 doesn't boost that often iirc
11:18:02*GodEater forgets that line of thought then
11:20:09 Quit Mathiasdm ("Yuuw!")
11:21:04 Join Mathiasdm [0] (
11:28:53 Join Thundercloud [0] (
11:30:18 Join waldo [0] (
11:32:30 Nick fxb__ is now known as fxb (
11:34:18 Join TiMiD[FD] [0] (
11:40:39peturTiMiD[FD]: 10.16.17 # * pixelma starts crying
11:41:52 Join Casainho___ [0] (
11:41:54 Nick Casainho___ is now known as Casainho (
11:43:25 Join MethoS_mobile [0] (
11:44:24TiMiD[FD]ow ?
11:44:43peturyes. NEVER make a lady cry! :p
11:44:45TiMiD[FD]what's wrong this time ?
11:45:20TiMiD[FD]it's the first time I would see someone cry for a computer program
11:45:36 Quit MethoS_mobile (Remote closed the connection)
11:45:54peturTiMiD[FD]: think of all those windows users ;)
11:46:29 Join MethoS_mobile [0] (
11:46:48TiMiD[FD]oh those I know just become crazy :)
11:51:17jhMikeShcs: I'm getting the lockup with one particular SPC file on both e200 and H10 (two very different kinds of targets) but if and only if playback is started cold. I haven't tested other files. Time for the "I did something really dumb" checks.
11:52:28 Quit TotallyInfected ()
11:58:32LinusNjhMikeS: i do those a lot :-)
12:00:44*Bagder added a hidden feature in that commit by mistake
12:00:50jhMikeSwell, time to revert to before suspect revisions...always a fun time :)
12:00:57 Quit Casainho__ (Read error: 110 (Connection timed out))
12:01:38 Join FOAD_ [0] (
12:02:02 Nick fxb is now known as fxb__ (
12:06:32Bagderapps/genlang-features is now built by the build procedure
12:07:13Bagderwill simply the main (generated) Makefile
12:07:25LinusNtime to rename the sim directory name too? :-)
12:07:31BagderI thought about that
12:07:40Bagderbut decided there are N people already using it now
12:08:57BagderI think we need to do something about the build system to allow some server to just build the small things
12:09:24LinusNthe small things?
12:09:34*rasher thinks the sim directory should be called "disk" or something like that
12:09:46BagderLinusN: bootloaders, sh builds
12:10:00Bagdersh is separate already of course
12:10:32Bagderoh well, the speculative parallel build concept would of course fix that anyway...
12:11:04Bagder15 servers built now
12:12:06rasherBagder: hand out one build to multiple servers and see which finishes first?
12:12:39Bagderbasically, all servers should be busy until all builds are done
12:12:41rasherwould make sense for the period between handing out the last unique build, and recieving the last build back
12:13:35Bagdernow we can see really fast servers stop working really soon
12:13:48Bagderwhen a bunch of slower servers got the last N builds
12:14:26Bagderexactly, that's what the current systems makes
12:15:58buckoHey, badger, I plan to update your Sansa checksum.c thing within a week or so to cope with walking the file Saratoga-style. Though I don't think it'll finish the file; there's some still about 20% through that seemed to have a different header...
12:16:05Bagderright now I'll remove sh builds from some of the fastest servers
12:16:18buckoI don't have much free time at the moment, though.
12:17:12Bagderbucko: well, just post the info/source you have in a wiki page or something and we/anyone can take it from there when time permits
12:17:30buckoI think saratoga already posted the algorithm on the forum.
12:17:43Bagderyes he did
12:17:49waldogot i microsd in my sansa .. worked fine .. but since this morning i can't connect my sansa to my pc with the microsd inserted ... i guess some currupted files of so ... in the orig filmwire there is a 'format' otion ... does that format the microsd or the main memory ?
12:17:50jhMikeSr16105 seems to bring this about
12:17:50Bagderbut 1) no code and 2) the forum is bad for static info
12:18:41 Quit FOAD (Read error: 110 (Connection timed out))
12:18:41 Nick FOAD_ is now known as FOAD (
12:18:51buckoBadger: OK. But when I get around to it. Maybe when I do it'll already be done by someone else; there seems a good few active people.
12:19:07buckoMust shower and go to campus now.
12:34:40 Join fasmaie [0] (
12:34:44***Saving seen data "./dancer.seen"
12:35:33 Join tvelocity [0] (n=tony@
12:47:30 Join broom [0] (n=c27f0814@gateway/web/cgi-irc/
12:49:17broomBagder: I think you were a bit unexact in your last change to "ProjectGlossary". Even if the build has produced red marks, the build itself was OK :-))) Since it has delivered the correct result.
12:49:51Bagderit didn't deliver the correct result
12:50:06Bagderhow could it if it is red?
12:50:34Bagderalthough I may of course still be inexact
12:53:04broomThe correct result in the case that the program has (e.g. compile) errors is an error message :-P It's an expected (and correct) result.
12:53:59GodEaterit's a correct result for the *compiler* but not for the build
12:55:07Bagderthe expected result for the build is to build with no warnings and errors. other results are not expected nor correct
12:55:20 Quit conando (Read error: 110 (Connection timed out))
12:55:29 Quit linuxstb (Read error: 113 (No route to host))
12:55:51 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
12:55:57broomFor the build script, I think, it's still a correct result. It's expected that such result may come. I recently had a lengthy discussion (not RB related) on this topic. I think there's a big misconception out there.
12:56:39Bagderthe build script expects errors to happen yes
12:56:47Bagderso an error is not unexpected for the script(s)
12:56:54Bagderbut unexpected for the build
12:57:03GodEaterwe expect builds to work
12:57:40Bagderbroom: but feel free to improve the phrasing if you think I did a poor job, that's what the wiki is for
12:57:59broomBagder: ah, ok then. For me, "build" is a execution of the build script.
12:58:16GodEaterfor us "build" is a working copy of rockbox
12:58:24GodEaterfor several values of "working"
12:58:46broomBagder: I was just kidding to an extent. But the other part of my phrase is still true IMHO. But here is not the right place to discuss this.
12:59:07BagderI understand what you're saying
12:59:15Bagderbut I don't agree with it in this context
13:00:18 Nick parafin|away is now known as parafin (
13:02:51broomBagder: If build=working copy of rockbox, then I think the exact (but not readabe :-) words would be "red is the color used to mark errors that occured during the attempt to produce a build"
13:08:05Bagderthat sounds a bit weird to me
13:08:16BagderI mean the "during the attempt to produce a build" part
13:09:43broomBagder: =during the attempt to produce a working copy of rockbox
13:09:56soappotato / potato
13:12:05 Join mf0102 [0] (n=michi@
13:12:21broomWhat I want to say is that the build as a script execution was successful even if a build (in the sense of a working copy of rockbox) could not be produced. I'm not a native English speaker.
13:13:10Bagderyes, that's the point you're pushing
13:13:23Bagderso edit the wiki page to be better
13:14:01BagderI really don't care that much
13:15:23broomBagder: I think everebody will understand the wiki as it is now. Not everybody sees the differences in success vs. error in my reasoning.
13:16:45 Quit broom ("CGI:IRC")
13:23:43 Join FOAD_ [0] (
13:37:47 Quit FOAD (Read error: 110 (Connection timed out))
13:37:47 Nick FOAD_ is now known as FOAD (
13:40:50TuplanollaHello, I just got a microsd to my e280. Should I turn the player off before inserting or taking the card off?
13:42:31 Quit jcollie ("Ex-Chat")
13:46:07LloreanDoesn't matter.
13:47:38TuplanollaOk, good. But I think I shouldn't take it off when I'm listening a song from it ;)
13:48:11 Join pixelma [0] (n=d5494d3a@rockbox/staff/pixelma)
13:48:20 Quit tvelocity (Read error: 110 (Connection timed out))
13:52:58 Quit MethoS_mobile (Read error: 113 (No route to host))
13:53:46pixelmaTiMiD[FD]: what you did wrong? That you insist on plugin actions... I only had the chance to read the diffs now (because I don't have the target with me) but I'm guessing that "flame moving", "flame type" and simply exiting the plugin still won't work on C_200. And even if it does, having to use the volume buttons for increasing/decreasing the flames is awkward because it force me to twohanded operation and the controls were working ine before :(
13:54:56pixelmabesides I don't understand your "#ifdef HAVE_LCD_BITMAP" around the whole thing and your brace style (especially in some do while loops)
13:58:02pixelmabut the style is not as important to me as the controls
13:58:11 Join kugel [0] (i=kugel@unaffiliated/kugel)
14:01:37 Quit fasmaie (Read error: 110 (Connection timed out))
14:02:50 Quit SirFunk (Remote closed the connection)
14:03:15 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
14:04:18TiMiD[FD]pixelma: about the flames high, I thought it was more logical to put it on the wheel as the ipod is done the same way
14:04:43 Quit jhMikeS (Nick collision from services.)
14:04:45TiMiD[FD]it would be trivial to put it on the up/down button, only a matter of preference
14:04:49 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
14:05:13LloreanTiMiD[FD]: The c200 has no wheel...
14:05:17 Join MethoS_mobile [0] (
14:05:32TiMiD[FD]oh ?
14:05:36TiMiD[FD]shit ...
14:05:42pixelmaexactly, I don't speek about the e200
14:05:43TiMiD[FD]I confunded with the e200
14:05:44 Quit Arathis ("Bye, bye")
14:06:00TiMiD[FD]uh oh
14:06:41pixelmaTiMiD[FD]: did you try in a sim? Yesterday's problem with the controls were already visible there...
14:06:53TiMiD[FD]yes I tried in a sim and it works now
14:07:05pixelmain a c200 sim?
14:07:36TiMiD[FD]yesterday I didn't tried to change the flame type or to quit (I closed the sim instead, but the problem was not related to the mapping by itself)
14:07:44TiMiD[FD]yes in a c200 sim
14:08:09TiMiD[FD]also works on my h140 and on a ipod video
14:08:22 Quit Toki (Read error: 113 (No route to host))
14:09:38TiMiD[FD]I'll search for a picture to try to get what a c200 looks like
14:09:46pixelmaif I understand correctly the controls seem worse than before
14:10:09pixelmaand I don't understand why it was necessary
14:10:31pixelmaTiMiD[FD]: start the sim with the −−background switch
14:10:34TiMiD[FD]because I thought the c200 had a wheel
14:11:55TiMiD[FD]ah nice
14:12:07pixelmaI don't understand moving this plugin to actions at all... :\
14:12:21 Quit ctaylorr_ (Read error: 110 (Connection timed out))
14:12:29TiMiD[FD]less code ...
14:12:41TiMiD[FD]I don't think action API is bad
14:13:01LloreanLess code hardly matters if it makes it less usable too, especially with plugins
14:13:01TiMiD[FD]in the worst case just define the actions per target as it was done before
14:13:36GodEaterwhere does leave all it's temporary files when you run it as root ?
14:13:41pixelmaconfusing and easily to break controls, being afraid of breaking other plugins (invisble connections), being pleased with inferior controls
14:13:48pondlifeTiMiD[FD]: The most important thing is that the UI is entirely unchanged.
14:13:52TiMiD[FD]the problem with the fire plugin doesn't lie in the action api anyway
14:14:51Lloreanpixelma: I do think that making one plugin's controls dependent upon another is a very, very bad thing.
14:15:28 Quit idnar (Read error: 110 (Connection timed out))
14:15:50TiMiD[FD]well if the only action defined were directions and quit it would be a good thing
14:16:13LloreanNot necessarily
14:16:23LloreanSince wheels exist to cloud things up.
14:16:30TiMiD[FD]less time spent on developping a plugin when you don't have the target
14:16:34pixelmaI'm not much of a coder so most things I did was porting plugins to new devices (basically doing the graphics and controls) and already now there are so many half working things due to plugin actions
14:16:37amiconnThe plugin action api was meant as a way to make porting plugins to a new target easier, but due to the nature of many plugin, it has the opposite effect
14:16:43LloreanWhat one could do, instead of 'actions' the way they are now, is define a list of actions based on what they do in the list/menu. So "Up" "Down" "Back" "Stop" "Select" and a few others
14:17:21TiMiD[FD]there are 12 plugins using the action API
14:17:23LloreanSince most buttons, and several combinations of buttons, are always used in the list browser, for application-style plugins you have a context to imagine your actions coming from.
14:17:24amiconnAs soon as you start to chain contexts, there are interdependencies, and if you change things for one target, a dev can easily break the plugin for another target without noticing
14:18:03pixelmaTiMiD[FD]: that doesn't mean that you have suitale controls just some that work somehow and it's very very hard now if someone with the target wants to improve
14:18:22amiconnAnd if the actions are defined per plugin (no chaning) you'll end up with one context per plugin, i.e. essentially the same as not using the actions api at all
14:18:34TiMiD[FD]well so lets rework this from the basis for all plugins
14:18:43TiMiD[FD]there is a fundamentl problem here I agree
14:18:57amiconnBut using it scatters the code path across several files, making it harder to follow than the simple definitions
14:19:26amiconnI think plugin actions (not actions in general) in their current form should be reverted
14:19:31TiMiD[FD]action api is easier when you have to handle buttons combinations
14:19:31LloreanI agree.
14:19:49LloreanAgree with amiconn, in case the internet made that show up on the wrong side of Timid's statement.
14:20:00pixelmayou can guess my opinion ;)
14:20:14LloreanI do think that there's probably a more flexible solution somewhere, but I don't think Plugin Actions is it.
14:20:15TiMiD[FD]ohoh as if I was responsible for that ...
14:20:19amiconnA plugin could still use the actions api, using locally defined contexts, if doing so makes input processing easier
14:20:32TiMiD[FD]mixing buttons and actions ?
14:20:53amiconnDefining a context *for that plugin* within the plugin
14:20:56TiMiD[FD]plugins should do as they did before : define their own actions and eventually use some standard one like quit
14:21:16amiconnThe mixing is the bad thing
14:21:17pixelmaTiMiD[FD]: no but you went on implementing it for the fire plugin (which is why I started the dscussion again) even after the talk earlier today
14:21:25TiMiD[FD]I ddn't talked about mixing
14:21:33amiconnyou did
14:21:40TiMiD[FD]not in mylast statement
14:21:44amiconn"define their own actions and eventually use some standard one like qui"
14:21:49TiMiD[FD]I wasn't thinking about that
14:21:50amiconnThat _is_ mixing
14:21:53TiMiD[FD]onlyusing action api
14:21:56 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
14:21:57TiMiD[FD]to make it clear
14:22:19TiMiD[FD]define their own actions with the action api
14:22:41amiconnEither use *one single* standard action context, or *one single* private context. Or no actions at all
14:22:51TiMiD[FD]at least to save the hassle of having to code the quit action and handle the case with the remote in every plugin
14:23:42TiMiD[FD]using standards action makes the plugin dependant from external code subject to change and to brek it as you stated
14:23:45linuxstb_TiMiD[FD]: Not all combinations of plugins/targets have a quit action - if there are not enough buttons, quit is a menu option.
14:24:00TiMiD[FD]a lot still have one
14:24:15LloreanThat doesn't mean they all do.
14:24:23LloreanAnd that doesn't mean that one button is the best quit button for all plugins.
14:24:37TiMiD[FD]for most of them it is
14:24:50linuxstb_That's the problem - "most"...
14:24:56 Join tchan1 [0] (
14:24:59TiMiD[FD]in the other case nothing prevents the developper from not using the standard action
14:25:15LloreanThen they have to mix again.
14:25:22LloreanOr not use actions.
14:25:28LloreanOr rather, not use plugin actions.
14:25:30TiMiD[FD]no they define their own actions ...
14:25:39LloreanWhich is mixing.
14:25:40TiMiD[FD]it's the same as defining keys
14:25:48LloreanYes, in which case why not just do it?
14:26:10LloreanYou could very easily write a "standard" keymap for most targets, that people could copy and paste into their plugin, then edit.
14:26:14TiMiD[FD]less switch statement when checking for buttons, less ifdefs
14:26:41LloreanWith this, they could be sure in advance there were no conflicts, unless they introduced them by hand, and it could have sane defaults for up/down left/right and cancel.
14:27:06TiMiD[FD]no conflict by using the action api either
14:27:07LloreanAnd since it's hard coded per plugin, there's no concern about later changes breaking the plugin's keymap
14:27:24TiMiD[FD]if every plugin defines its own contexts
14:27:57TiMiD[FD]used like that no drawbacks
14:28:01LloreanThat's not the Plugin Action system anymore then.
14:28:08LloreanThat's simply using the normal action system, within plugins
14:28:08 Quit cool_walking_ (Remote closed the connection)
14:28:44pixelmawhat should that be good for?
14:28:56TiMiD[FD]easier to code
14:29:11TiMiD[FD]and to maintain
14:29:12LloreanYou still have to define a full context for each plugin.
14:29:22LloreanWhich means a full keymap for all targets.
14:29:24pixelmain another file?
14:29:26TiMiD[FD]same as before
14:29:55LloreanSo why exactly is that any different from including the keymap *in* the plugin?
14:30:10TiMiD[FD]but it's somehow easier to handle buttons with the action plugin in most cases
14:30:29TiMiD[FD]that way it would be included in the plugin
14:30:31Llorean"action plugin"?
14:30:50LloreanWhat is "the action plugin"?
14:30:59linuxstb_Llorean: I assume TiMiD[FD] means "the action API"
14:31:06TiMiD[FD]let me read my sentence again
14:31:18LloreanAh, I wasn't sure if he meant plugin actions again.
14:31:27TiMiD[FD]yep that's what I meant
14:31:45TiMiD[FD]action API
14:32:01LloreanYes, but the topic here is that Plugin Actions should be removed
14:32:05Lloreanthat's what we've been saying ALL along.
14:32:11TiMiD[FD]ah I'm fine with that
14:32:12LloreanNot that the action API should be removed
14:32:37LloreanThe point was that each plugin should have its keys not be dependent on other plugins.
14:32:41linuxstb_Llorean: Yes, that's what TiMiD[FD] has been saying as well... (IIUC)
14:32:56TiMiD[FD]couldn't agree more on that
14:33:16LloreanThen why is he adapting plugins to plugin actions?
14:33:28TiMiD[FD]oh to the action API
14:34:28TiMiD[FD]actually when I used the plugins predefined actions, I didn't knew there were targets on which the keymaps for the different overlapped
14:34:39TiMiD[FD]*different context*
14:34:48***Saving seen data "./dancer.seen"
14:34:50LloreanThat's why different contexts exist...
14:35:46TiMiD[FD]on my h140 directions and actions doesn' overlap and that's the only target on which I'm able to test reliably
14:36:00LloreanYou have sims.
14:36:34LloreanNot to mention the plugin actions file itself.
14:36:43TiMiD[FD]so anyway lets define per-plugin actions instead of using the global ones
14:36:58TiMiD[FD]I'll take care of that tomorrow if there is no objection ?
14:37:06Zagorno, let's do away with actions in plugins
14:37:16LloreanZagor: I think actions is okay.
14:37:24linuxstb_TiMiD[FD]: Probably best to post a patch first - I don't think we're all understanding each other...
14:37:29LloreanI mean, that's practically what the huge #define block at the beginning of plugins is anyway.
14:37:45linuxstb_TiMiD[FD]: But IIUC, I agree... ;)
14:37:49LloreanZagor: I mean, actions within a plugin, where they're all defined within that plugin, is pretty safe.
14:37:57LloreanTiMiD[FD]: I do too if we're all understanding now.
14:38:23ZagorLlorean: sure you can add lots of #defines at the top of every plugin. but what does it really accomplish?
14:39:14LloreanZagor: The same thing as hand coding keymaps per plugin, which is also a huge block of defines?
14:39:44TiMiD[FD]excepted that the action API is more powerful to express a keymap
14:40:11ZagorI'd use the words "less transparent"
14:40:11pixelmathe sims don't simulate everything though, e.g. you can use button combo with every buttons and not see the restriction that are there for the Iaudios an H100 (IIRC)
14:40:26 Quit tchan (Connection timed out)
14:40:38TiMiD[FD]pixelma: yes and still no remote buttons for the irivers for example ...
14:40:39linuxstb_pixelma: But that's the same problem with the old-style button #defines...
14:41:05LloreanI'd rather just see a KEYMAP file that's a huge block of defines for every target and valid key combination. The enterprising plugin author can then paste it, add names for all the defines he wants to use, cut out the remaining lines, and go.
14:41:09pixelmayes, it was just an addition to the "you have sims" statement ;)
14:42:05LloreanI find the old defines more readable anyway, since I at least tend to have a more clear idea what's going on in some cases.
14:42:26TiMiD[FD]depends of the cases
14:42:31pixelmaLlorean: yes, that's also what I was thinking
14:42:47TiMiD[FD]in somecases it can become tricky
14:44:28 Join n1s [0] (
14:45:04TiMiD[FD]OK so I'll patch the plugins already converted to the action API
14:45:25amiconnThere are many button combo constraints on various targets, both electrically and mechanically
14:45:43LloreanThat's why a list of them somewhere would be good.
14:46:00amiconnE.g. Left+Right won't work on ipods with clickwheel, even if it would work electronically
14:47:32 Quit phossy (Read error: 110 (Connection timed out))
14:47:52LloreanWhich ones are cilckwheel?
14:48:02LloreanI can never keep track the various names for the wheels.
14:48:44pixelmathe newer ones with "integrated" buttons
14:49:27pixelmano extra buttons either around the wheel or above
14:49:32LloreanIf I recall, someone mentioned that they *can* press Play/Menu at the same time. I personally think it's a bad idea (straining the wheel, surely) but I seem to recall it being tried in some odd unsupported build or whatnot.
14:54:58pixelmaok, thanks for discussing this once more. I should leave again and not waste much of not my money ;)
14:56:36pixelmaand I still can't /quit the webclient... weird
14:58:13 Join webclienttest [0] (n=d4efd6a6@gateway/web/cgi-irc/
14:58:28 Quit MethoS_mobile (Remote closed the connection)
14:58:48 Quit webclienttest (Client Quit)
14:58:59peturindeed, doesn't work
15:00:06 Join jcollie [0] (n=jcollie@
15:01:40 Quit pixelma ("CGI:IRC (Ping timeout)")
15:01:57 Join LambdaCalculus37 [0] (n=44a0430f@gateway/web/cgi-irc/
15:02:16LambdaCalculus37Morning, everyone!
15:19:50 Join ch4os_ [0] (n=ch4os@unaffiliated/ch4os/x-059673)
15:20:14 Join idnar [0] (
15:20:38 Quit kugel ("ChatZilla 0.9.80 [Firefox]")
15:21:22 Quit Rob2222 ()
15:21:42 Join Rob2222 [0] (
15:23:30 Quit Rob2222 (Client Quit)
15:23:41 Join Rob2222 [0] (
15:24:31 Quit pabs ("brb")
15:26:23 Nick tchan1 is now known as tchan (
15:34:04 Join zagor_test [0] (n=c2ed96ab@gateway/web/cgi-irc/
15:34:31 Quit zagor_test (Client Quit)
15:48:13 Nick fxb__ is now known as fxb (
15:48:47 Quit Rob2222 ()
15:48:57 Join Rob2222 [0] (
15:51:01 Quit TiMiD[FD] ("leaving")
15:54:06 Join BobShield [0] (
16:00:21 Join webguest [0] (n=d4efd6a6@gateway/web/cgi-irc/
16:00:32 Quit webguest (Client Quit)
16:04:54 Join linuxstb__ [0] (
16:05:04 Join scorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
16:05:48 Quit linuxstb (Read error: 110 (Connection timed out))
16:09:20 Part LinusN
16:09:46 Quit Zagor ("Client exiting")
16:10:53 Quit linuxstb_ ("Client Exiting")
16:14:58 Join jac0b-work [0] (n=jac0b-wo@
16:15:49 Quit xushi ()
16:15:59jac0b-workhas anyone got viewports working in placing the progress bar
16:16:52jac0b-workI have found when trying to use viewports to place the progress bar it doesn't work
16:19:45 Quit jac0b-work ("ChatZilla 0.9.80 [Firefox]")
16:24:31 Join FOAD_ [0] (
16:25:05 Join pabs [0] (
16:28:16 Quit hcs ("Leaving.")
16:34:53***Saving seen data "./dancer.seen"
16:41:58 Quit FOAD (Read error: 110 (Connection timed out))
16:41:58 Nick FOAD_ is now known as FOAD (
16:46:40 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
16:47:12 Join MethoS- [0] (
16:48:21 Join phossy [0] (
16:53:41 Quit ol_schoola_ ()
16:56:28 Join qwedsa [0] (
17:04:56 Join Arathis [0] (
17:06:18 Join qwedsa__ [0] (
17:07:10 Join stewball [0] (n=WTFOMGBB@
17:12:57 Quit BobShield ("leaving")
17:13:06 Quit qwedsa_ (Read error: 110 (Connection timed out))
17:15:02 Part pondlife
17:18:00 Join ackbahr [0] (
17:21:22 Quit qwedsa (Read error: 110 (Connection timed out))
17:23:23 Join qwedsa [0] (
17:25:54 Quit MethoS- (Remote closed the connection)
17:32:57 Join SiaS [0] (
17:37:55 Quit phossy (Read error: 110 (Connection timed out))
17:38:47 Join phossy [0] (
17:39:55 Quit qwedsa__ (Read error: 110 (Connection timed out))
17:47:25 Quit BRi7X (Read error: 104 (Connection reset by peer))
17:56:45 Part Llorean
18:01:29 Quit Siku (Nick collision from services.)
18:01:40 Join Siku [0] (
18:02:35 Join Domonoky [0] (n=Domonoky@
18:03:57*Domonoky has just combined a broken h120 (nokia syndrome) with a broken ipod (broken hd) to a working ipod :-)
18:07:24at0m|cDomonoky, battery or hd? :>
18:08:10at0m|ci wonder if rockbox supports 'usb on the go' on h3x0
18:08:22LambdaCalculus37Domonoky: Nokia Syndrome? =/
18:08:26Domonokynow on to restore the ipod, and put rockbox on it..
18:08:35scorche|wat0m|c: it doesnt
18:08:43 Join j0tt [0] (n=j@unaffiliated/jott)
18:08:58at0m|ci think it doesnt, but booting to original OS doesnt work - pressing record while booting pops up a quick 'booting orig os..' but then it shuts down :z
18:09:14DomonokyLambdaCalculus37: if you put a nokia charger onto a iriver hxxx its dead, burned..
18:09:15at0m|cscorche|w, cheers
18:09:49 Join MethoS- [0] (
18:10:55at0m|cscorche|w, the original problem is that my h340 doesnt boot into its original OS no more
18:11:11at0m|cso i tried to 'usb on the go' using rockbox
18:11:20at0m|cbut that doesnt work either...
18:12:02LambdaCalculus37Well, now I know a good way of destroying irivers... (but I'd never destroy one :P)
18:13:17at0m|cif i cant boot in the orig OS of my h340, i'd order some SD cards from, it's their offer of the day..
18:14:13markunWhat do SD cards have to do with not booting the into the OF?
18:14:20CasainhoDomonoky: hello :-)
18:14:49CasainhoDomonoky: did you sew the new update to TWiki page of RB player?
18:14:51at0m|cmarkun, usb on the go empties my cam's SD on the h340
18:15:08 Quit Domonoky (Read error: 104 (Connection reset by peer))
18:15:21at0m|cmarkun, so then i just need one SD card and transfer to h340 when it's full
18:15:51at0m|cnow that the dual boot has disappeared, i need more storage for in cam eh
18:16:58at0m|ctried inserting hd from my other/older h340, but that hd seems broken.
18:21:08 Quit jott (Connection timed out)
18:21:25 Nick j0tt is now known as jott (n=j@unaffiliated/jott)
18:22:11 Join Domonoky [0] (n=Domonoky@
18:23:31 Quit petur ("work->home")
18:24:07 Join Silens [0] (
18:24:49 Quit ackbahr (Remote closed the connection)
18:26:37 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
18:29:16 Join Arathis2 [0] (
18:31:02 Quit Arathis (Nick collision from services.)
18:31:05 Nick Arathis2 is now known as Arathis (
18:31:05 Part pondlife
18:31:57 Join TotallyInfected [0] (
18:32:44markunat0m|c: maybe petur or amiconn have an idea what the problem could be.
18:33:00CasainhoDomonoky: heelo :-)
18:33:10CasainhoDomonoky: are you there? I would like to talk
18:33:34 Join PaulJam [0] (
18:34:08DomonokyCasainho: yes, i just returned home, so i am a bit busy..
18:34:30 Join lazka [0] (n=lazka@
18:34:58***Saving seen data "./dancer.seen"
18:35:07CasainhoDomonoky: just to say that I did update the page - I will put a message in forum about connections - thank you for giving that info :-) - bye bye :-)
18:35:12 Quit Casainho ("ChatZilla 0.9.80 [Firefox]")
18:36:51SilensDoes enybody know if ther's any firmware hacking going on for the sony mp3-players? any chance of getting rockbox on them?
18:37:22 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
18:37:39 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
18:38:18pondlifeHmm, looks like some ipod owners can use keyclick without nasty noises...
18:38:32pondlifeMaybe the underlying issue has already been fixed?
18:38:42pondlifeAny Sansa owners tried it?
18:42:10 Join BigBambi [0] (n=alex@rockbox/staff/BigBambi)
18:43:38LambdaCalculus37pondlife: Not yet. Have to update my c240.
18:43:56pondlifeI'd protect your ears, just in case
18:47:42 Join JdGordon|w [0] (n=836b0065@gateway/web/cgi-irc/
18:49:01rasherNico_P: wouldn't it make sense to change fs#8499 into a bugreport about the database tool?
18:49:31 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:49:39LambdaCalculus37pondlife: I'll use ear plugs. :)
18:49:45Nico_Prasher: it probably would, yes
18:49:49n1sat0m|c: there is an issue where rockbox can corrupt the eeprom so that the h300 of will crash when it's trying to read its settings, there is a known good eeprom image somewhere in the forums iirc so it might be worth a serach
18:49:57 Join styleism [0] (
18:50:02at0m|cmarkun, cheers
18:50:18at0m|cn1s, tried restoring the eeprom from debug, but didnt work
18:50:30at0m|cnow doing mkdosfs -c -F 32 -n H300 -i ae1ddf3e /dev/sdb1 ...
18:51:03at0m|ci mean, the eeprom restore did work, but it didnt solve the problem
18:51:09n1sat0m|c: did you try with an image you downöoaded or just a dump-write?
18:51:27at0m|can image downloaded from wiki FAQ
18:51:42n1sah, strange problem then...
18:52:11at0m|cmkdosfs takes a while... :)
18:52:43 Quit TotallyInfected (Read error: 110 (Connection timed out))
18:53:37at0m|cif not for my new cam supporting 'usb on the go', i wouldnt care for the orig OS
18:53:47at0m|cjust to say how much i like rockbox :)
18:54:05 Join miepchen^schlaf [0] (
18:55:36PaulJamat0m|c: try holding the play button longer when booting the OF.
18:57:11at0m|cPaulJam, cheers, will try, if mkdosfs hasnt already solved it
19:02:01amiconnat0m|c: If you want to boot into the OF on iriver, you need to hold the ON button for a while. The OF needs to detect the button being pressed, otherwise it thinks the poweron was occasional and it shuts down again
19:02:33amiconnThat's all from memory; I didn't boot into the OF for ages (why should I?)
19:02:53at0m|cdidnt know about the long play press...
19:03:10at0m|ci just let it go when the screen lit up and pressed rec
19:03:22amiconnThe OF loader is pathetic.... (not uncommon with dap original firmwares...)
19:03:33at0m|canyways, formatting :)
19:04:56amiconnRegarding usbotg on h300 (or any other target with usbotg capability hardware): rockbox doesn't support it
19:06:04amiconnpetur tried to get the 'wasabi' usb stack running, but didn't succeed so far. 'wasabi' is the (open) example usb stack for the philips isp1362 usbotg chip
19:06:26at0m|ca pitty..
19:06:41at0m|cbut not such a problem if the orig OS still works eh
19:07:09at0m|cjust a slight inconvenience eh
19:07:15amiconnAmong the usbotg capable targets, the h300 is the one with the highest chance to become supported (because the isp1362 is the only usbotg chip with public documentation)
19:07:53at0m|cso usbotg's still being developped? cooles
19:07:57 Join bertrik [0] (
19:10:52 Join tedrock [0] (
19:24:41XavierGrat0m|c: so how do you turn on the OF? All I do is 1)Press and hold REC then 2) Push PLAY together with REC for 3-4 seconds (until I see the OF icon)
19:27:14 Join obo [0] (n=obo@rockbox/developer/obo)
19:28:16at0m|cXavierGr, i pressed play untill i saw the screen lit up, then pressed rec and saw 'booting OF', then screen went blank again...
19:28:48at0m|cXavierGr, cheers for refreshing the sequence eh :)
19:29:03at0m|cstill formatting...
19:29:16*ender` yawns
19:29:16at0m|cmkdosfs takes ages. formatting jfs 500gb takes like 4secs eh
19:29:48 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:31:20 Quit linuxstb__ (Remote closed the connection)
19:31:42 Join linuxstb [0] (n=chatzill@rockbox/developer/linuxstb)
19:31:46 Join PaulPosition [0] (
19:31:54Nico_PGodEater: are you available for some testing?
19:32:23 Quit Domonoky (Read error: 104 (Connection reset by peer))
19:33:44XavierGrat0m|c: the rockbox bootloader is quite failsafe for most of us, you should press the rec button first and hold them both for a considerate amount of time
19:34:31XavierGrmost of the times people have trouble after the OF has started (eternal loading of the OF etc)
19:34:57 Quit jhMikeS (Nick collision from services.)
19:35:03 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
19:35:44 Join ol_schoola [0] (
19:37:12at0m|ci see.. i thought i still knew how to do it from my old h340 in which i broke some essential wire replacing the battery like 6months ago. this one i have since 3-4 days..
19:37:47 Quit SiaS ("Nettalk6 -")
19:38:14at0m|cafter my old h340 was sent to iriver, i didnt hear from them untill i got my unit back a month or 2 later
19:38:45XavierGrin working order or still faulty?
19:38:47at0m|ciriver said i didnt reply to their repair cost estimate - which never made it to me
19:38:56at0m|cit's still faulty, the old one
19:39:02at0m|chave another newer one now
19:39:11XavierGrUS EU model?
19:39:43at0m|cthe wire /crossing/ the wire bundle for the controls iirc was broken off, it was pretty tight
19:39:44 Join Domonoky [0] (n=Domonoky@
19:39:45XavierGrI mean is it the US or EU model?
19:39:50at0m|chm not sure, checkin
19:40:11XavierGrif it has USB OTG it is EU though the US version has the connector too
19:40:20at0m|cdoes it say on the box?
19:40:21XavierGr(but for another purpose)
19:40:26XavierGrI am not sure
19:40:34XavierGrwhat is the USB OTG Port says on it?
19:40:36at0m|clooking at it, but cant tell
19:40:48at0m|ci put eu firmware on it, usbotg worked
19:40:58at0m|c.. on the old one, and on the new one
19:41:13at0m|cnot sure if that depended on the OF
19:41:15PaulPositionWould it make sense to correct a small bug in a file *now* even if we know it would probably get corrected in the next few weeks (through Buschel's experiments)? I'm not talking of precipitating commits of the patches I'm talking about an omission in a file that *maybe* made for worse battery efficiency (relative) on the h10_5gb target than the 20gb got.
19:41:43PaulPosition(ie, an if defined IRIVER_H10 that should have a || IRIVER_H10_5GB after it.
19:41:45 Join conando [0] (
19:42:12XavierGrat0m|c: on US models you can't just put an EU firmware and get the USBOTG working
19:42:27XavierGrUS models need some hardware tweaking in order to get USBOTG to work
19:42:28at0m|cah, both EU models then
19:43:23XavierGrat0m|c: are you doing quick format or full?
19:43:30at0m|cif i find another one cheap on ebay (this last one was <100eu), then i'd love to install one in the car
19:43:32XavierGrI don't think there is a reason to do a full
19:43:48at0m|cXavierGr, i think full, it takes 1hr by now - at0m@asp1r:~$ mkdosfs -c -F 32 -n H300 -i ae1ddf3e /dev/sdb1
19:43:52at0m|coki complete!
19:44:02XavierGrat0m|c: yeah I've put an H115 on my car, it is wonderful
19:44:29at0m|cthen just occasionally update it using laptop...
19:44:38at0m|cbeautifull :)
19:45:08at0m|coki, got OF back
19:45:13XavierGrit boots?
19:45:17at0m|crestoring files on the disk
19:45:22at0m|cOF, that is
19:45:31 Part pondlife
19:46:28 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
19:46:55 Join karashata [0] (
19:47:04JdGordon|wPaulJam: yeah, that would be good to fix and doesnt need to wait for the full patch
19:47:37PaulJami gess that was for PaulPosition
19:48:29Domonokydoes an ipod charge from usb if it is in the "Ok to disconnect screen" ?
19:48:37 Join FOAD_ [0] (
19:48:38 Join TMM [0] (
19:49:54PaulPositionJDGordon - I never generated a patch, do you think I could just tell someone what file / lines the OR should go ?
19:51:48PaulPosition(ie, ../firmware/target/arm/system_pp502x.c line 273... #if defined(IRIVER_H10) should become #if defined(IRIVER_H10) || defined(IRIVER_H10_5GB)
19:52:05LambdaCalculus37GodEater: Here?
19:52:50amiconnPaulPosition: Iirc I tested this on my 6GB, and saw no difference at all
19:53:02amiconn(5GB == 6GB in that respect)
19:53:32PaulPositionUhuh... Okay.. Just that when I started testing bits for Andree, nothing would change in any build I did until I saw that..
19:53:58PaulPositionNow, I wouldn't say I understand 1% of what's happening there.
19:54:33PaulPositionBut it seemed weird to me that stuff would get initialized or something on the 20gb while *nothing* would get any treatment on the 5/6gb
19:55:02PaulPositionBut if it has no effect, I guess it could wait for Andree's results. :p
19:55:03amiconnDon't forget the iriver flash loader. It already inits important stuff
19:55:27PaulPositionamiconn - I guess.
19:55:51amiconnSomeone should compare register values with and without that init. I guess they will be the same
19:56:31amiconnIt would still make sense to do the init if we're ever going for true cold-boot in rockbox, but I doubt that will happen on PP
19:56:41 Join l00kp [0] (
19:56:51PaulPositionamiconn - I *think* they are, well the io screen I don't remember changing. You may be very right.
19:57:03amiconn(unless we magically get access to the secret PP docs)
19:57:22*PaulPosition isn't even a newb at this, he's just NOT in this at all.
19:57:40at0m|camiconn, h340 dual boot fixed after format and files restored. too bad i didnt try the advise on proper OF boot before i did the mkdosfs, so we won't know the real cause of the problem :/
19:57:42l00kpI notice that rockbox has not been developed for the ipod 160gb. Do you expect support for it in the next few years, or is this a model that can't be developed for.
19:58:21LambdaCalculus37l00kp: We'll probably have support for the iPod Classic when Duke Nukem Forever comes out.
19:58:27Horschtnext week?
19:58:45l00kpWhat about the 160GB?
19:58:58Horschtthat IS the classic, l00kp
19:59:05l00kpoh. my bad
19:59:10PaulPositionl00kp, like every target : it can't be done ... until it can be done. Usually happens all of a sudden, but it might be long before 'all of a sudden' happens.
19:59:11amiconnIf people just keep asking, no new model will ever be supported. It requires someone with that model and a clue to work on it
19:59:38amiconn(preferably more than one such person per model, depending on how hard the port will be)
20:00:04Horschtwhat worries me more than ipod classic support is battery life on the current ipod video port :)
20:00:13PaulPositionamiconn - Thanks, I'll go back to testing the bits Buschel mapped for me, report to him and then trust the 'professionals' :)
20:00:20*Horscht hides
20:00:26BigBambil00kp: The firmware is encrypted, the hardware is new and undocumented, no-one with a clue has tied, ...
20:00:27amiconnThe port to the classic, or 2nd/3rd Gen nano will be hard ones due to the encryption
20:00:38BigBambiHorscht: Then get working :)
20:00:52*BigBambi replaces tied with tried
20:01:31HorschtI did, BigBambi. but covering it in nuclear waste didn't help
20:01:40Horschtso there goes my idea
20:02:31l00kpThat sucks. I was going to buy the 160GB if rockbox might be supported in another year or two.
20:02:32HorschtI am not "demanding", though. I hope for the best.
20:03:12BigBambil00kp: It may if someone does the hard work
20:03:27BigBambil00kp: Rockbox is ported by interested owners
20:03:34 Join davina [0] (
20:04:47 Quit FOAD (Read error: 110 (Connection timed out))
20:04:48 Nick FOAD_ is now known as FOAD (
20:06:05l00kpI'ld love to assist in creating a IPOD 160gb port, but I probably lack the knowlage to do so.
20:06:26at0m|cl00kp, send one to an interested dev :)
20:06:56*l00kp is ripped off by college loans
20:07:03 Join linuxstb_ [0] (
20:07:06BigBambiThen it won't happen until an owner with the skills appears, or an owner wants it enough to learn (although as amiconn says the encryption makes this much harder)
20:07:20 Join kslater [0] (
20:09:08l00kpThe 5th generation 80GB ipod had encryption, right?
20:09:36LambdaCalculus37That was easily defeated.
20:10:10LambdaCalculus37But it seems that Jobs^H^H^H^HGod wanted something a little stronger for the Classic and nanos.
20:10:12linuxstb_ No, the main firmware wasn't encrypted, just the same simple checksum that had been used by Apple on earlier ipods.
20:12:16*l00kp sighs
20:12:29l00kpI'm going to stab the dark and grab a 160gb ipod then
20:14:41LambdaCalculus37l00kp: Or you can go onto eBay and pick up an iPod video.
20:14:54LambdaCalculus37That way, you get a supported target for less money.
20:15:24l00kpWell, I would use it for voice recording and my music library.
20:15:49l00kpMy ipod microphone only works in the ipod firmware and my music is in flac.
20:15:54l00kpThe key is storage space.
20:16:06Horschtflac wont work on ipod classic...
20:16:09 Join Frazz [0] (
20:16:36l00kpIt works with rockbox!
20:17:29Horschtyeah, for a collection full of flac, 80GB might be too wimpy
20:17:30LambdaCalculus37l00kp: Buy an 80GB iPod video on eBay. There are larger hard drive sizes available at many online stores.
20:17:59LambdaCalculus37For a voice recorder, I just picked up a Sansa c240 and loaded Rockbox onto it. Works great for classes.
20:18:16Horscht1.8 inch is the size to go for, no?
20:18:28l00kpMy ipod mic has a line in function that I use for recording as well.
20:18:38 Join thgz [0] (
20:18:39BigBambiHorscht: Yes, but there are many interface options
20:18:50l00kpLambdaCalculus37, what should I google for ipod hard drives?
20:18:50Horschtoh i see.
20:18:58BigBambiYou also need to get the type of 1.8" hard drive
20:19:01 Quit linuxstb (Read error: 113 (No route to host))
20:19:05LambdaCalculus37l00kp: iFixit
20:19:12l00kpAny 1.8" hard drive?
20:19:29BigBambiSee what I said 6 lines above
20:19:30LambdaCalculus37ZIF (Zero Insertion Force)
20:19:59LambdaCalculus37They sell Apple parts.
20:21:17amiconnLambdaCalculus37: If someone's going to replace the hdd anyway, a 60GB G5 would be a better choice to start from (cheaper, and also dual platter)
20:21:54LambdaCalculus37amiconn: Oops... forgot about the 60GB.
20:22:12l00kpSo, if I'm going to replace a ipod's hdd, I need one that's 1.8" + ZIF . . .
20:22:37amiconnThat is, the G5.5 is a slightly better choice if someone's going to use the backlight a lot
20:23:16Horschtwhy is that?
20:23:22amiconnAfaiu, there are 2 main differences between the G5 and the G5.5. (1) The G5.5 backlight is brighter and needs less power
20:23:28 Part l00kp
20:23:29Horschtis the 5.5G backlight less power consuming?
20:23:38linuxstb_What is the largest available (and sensibly priced priced) 1.8" drive at the moment?
20:24:23amiconn(2) The G5.5 presents the medium as having 2048-byte sectors via USB, while the G5 uses standard 512-byte sectors
20:24:34Horscht100GB seems to be the biggest ZIF40 HD out there
20:24:53amiconnI think there's a 120GB model as well
20:25:10amiconn(by Seagate)
20:25:48LambdaCalculus37This is the biggest HD iFixit offers:
20:26:33Horschtthe HD page on the rockbox wiki says it's not for sale yet, though, amiconn
20:26:44*linuxstb_ would happily rip the 160GB drive out of a classic if any other DAP could use it...
20:28:01 Join roxfan2 [0] (
20:29:51BigBambipixelma: There?
20:31:49 Quit Mathiasdm ("Yuuw!")
20:32:40BigBambiI'm working on the radio section of the manual - how do you feel to having a seperate section for the radio rather than it all under the menu item?
20:33:09BigBambiAlong the lines of, this starts the radio (disclaimers for units, that may/may not have it), see here for full info
20:33:23thgzBigbambi: I think it would be better
20:33:44BigBambiIt is going to be a bit bigger than it is now
20:34:20BigBambiFor instance, I can't find anywhere a description of the radio settings menu (which only seems accessible from the radio screen anyway)
20:34:42pixelmaI guess I would also try to seperate it more, I thought earlier about an introduction, button table, description of the radio menu, description of the prest list (and preset concept in general...
20:35:00***Saving seen data "./dancer.seen"
20:35:03BigBambiYes, I have a Working with Presets heading :)
20:35:11 Quit ompaul (Read error: 104 (Connection reset by peer))
20:35:29BigBambiBut it seems a bit odd to have that all in the main menu chapter under a desription of what the fm radio main menu item does
20:35:36pixelmaonly one small problem with these parts was/is that for the Ondio the preset list is only available from the radio menu
20:36:05 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
20:36:14BigBambiCan you get to the preset list from else where on other targets? (I can't find a way of doing it on my H100)
20:36:19pixelmaBigBambi: yes, more structure would be nice
20:37:03pixelmaBigBambi: that should at least be in the manual (button table) now, I think I had an eye on it when updating for the c200...
20:37:30BigBambiSorry, I misunderstood you
20:37:48BigBambiDon't worry, I misread :)
20:38:13BigBambiOK, I think I'll work on splitting it a bit as well as restructuring
20:38:15pixelmaok, maybe I didn't make myself clear :)
20:39:46amiconnHorscht: The Samsung HS122JB (120GB dual platter ZIF-40) is available though. USD247 (qty 1) at
20:41:28Horschtwow, that's a lot of money
20:42:27 Quit roxfan (Read error: 110 (Connection timed out))
20:47:14thgzBtw, is this only a bug or a missing feature, but when I press long menu (going to quickmenu) I get no voice output
20:48:26BigBambipixelma: I'm looking at the ondio sim + manual - why doesn't short mode go to the preset list?
20:49:04BigBambiI see that double mode is record (although not in the table), is this why? What about a timeout?
20:50:17pixelmaearlier one short mode was record and it was to easy to accidentally start recording, so it was changed later
20:50:33 Join tvelocity [0] (
20:50:56BigBambiSo would short mode for preset list present a problem? (I've never seen an Ondio, let alone used one:) )
20:50:56amiconnthgz: The quick menu isn't voiced, on any target, because it is planned to be replaced with something more flexible, which just didn't happen yet...
20:51:23amiconnBigBambi: Yes it would. If short press would be taken, double-click would not be possible
20:51:34BigBambiamiconn: Even with a timeout?
20:51:49amiconnThat's not Ondio specific. If the first click takes you to another screen, the second click will never reach the first screen
20:52:01amiconnWhat timeout?
20:52:13BigBambiWell, I guess it isn't possible
20:52:37BigBambiI was thinking of after first press waiting x time to see if there were a second press
20:52:51pixelmahe means that it should wait if the user presses a second time and only if not go to the preset list
20:52:57amiconnThen nothing would happen after the first click for a while... very confusing
20:53:08BigBambiNot if it were short enough
20:53:30BigBambiBut anyway, it isn't important for what I'm doing now
20:53:54pixelmawhat if you think - why did nothing happen and just press a second time?
20:54:07BigBambipixelma: Oh, I can see the issues right enough
20:54:20BigBambiI was just wondering after seeing the key table
20:54:47pixelmabtw. I've never felt reaching the preset list through the menu was a hassle (and holding my good old OndioFM in hands) :)
20:55:09BigBambiFar enough ('twas idle curiosity more than anything)
20:55:25thgzAmiconn: ok
20:56:01pixelmaBigBambi: sure
21:02:17pixelmaBigBambi: hmm... the double mode is not in the table? I thought it once was... need to have a look
21:02:25 Part thgz
21:02:28BigBambiI'll add it :)
21:03:02 Join Casainho [0] (
21:03:34CasainhoDomonoky: heelo :-) I am here if you want to talk about hardware :-)
21:04:05DomonokyCasainho: hi..
21:04:52Domonokyyou / we need to figure out which connections from the devboard we can use for the external board with buttons/lcd and dac..
21:04:59 Join jgarvey [0] (
21:05:16Domonokyi think the LCD is SPI, correct ?
21:05:27CasainhoI will see that about the LCD...
21:05:43Casainholook, I don't know about the DAC for buttons, what are that tecnhic...
21:08:01 Join stripwax [0] (
21:08:10CasainhoDomonoky: looks like you are right :-) - "9-bit SPI serial interface (clock/data signals)" −−>
21:08:55Domonokyyes, so for the lcd we could use the spi on the uext connector..
21:09:51Domonokybuttons are just a resistor network, which you can read with a adc, ie different voltages for different buttons..
21:09:58Casainhoand since the audio codec also works with SPI, we could choose SPI for both :-)
21:10:31Casainholook, in future, can you please use the forum and not put questions on TWiki page?
21:10:36amiconnSPI for a colour lcd would be nasty
21:10:40DomonokyCasainho: SPI has cs signals, and only one cs is on the uext..
21:10:50 Quit styleism ()
21:11:23Domonokybetter use twi for the dac, and we need something for the i2s of the dac
21:11:39amiconnmarkun: around?
21:12:11PaulPositionDang, posted my results to Buschel five minutes after he left the forum. lol
21:12:28Domonokyamiconn: whats normally used to interface color lcds ?
21:12:39amiconnSomething parallel
21:13:00amiconnSPI is quite slow for 2-bit greyscale already
21:13:47Domonokythis nokia lcd only has spi, but the spi interface can run up to 6 Mhz..
21:14:24amiconnWhat resolution and bit depth does that lcd have?
21:14:34Domonokyis there a hardware interface for i2s on our targets, or do we do this with software ?
21:14:58Domonokylcd is 132x132x12
21:15:15amiconn12 bit only? Kinda half-baked lcd...
21:15:28Domonokybut its cheap :-)
21:15:31amiconn(but what you'd expect from nokia...) </rant>
21:16:08Domonokythe lcd itself is not from nokia, its used in nokia phones, thats why they are so cheap..
21:16:48CasainhoDomonoky: can't we take another CS form other extension port?
21:17:51amiconnUnder ideal conditions (i.e. running the SPI at full 6MHz, which might be tricky), you'll get ~28fps full screen *for plain lcd_update*. Forget mpegplayer then
21:18:53Domonokydoesnt sound good...
21:19:53DomonokyCasainho: i2s is aviable in the mcu of the devboard, but i dont know on which pins it is, (too lazy to check all those pdfs ) :-)
21:20:26CasainhoDomonoky: have you experince in draw schematics?
21:20:41CasainhoDomonoky: that LCD is just fine for this 1st version! :-)
21:21:34DomonokyCasainho: no experience in drawing shematics, i just build my circuits directy from my mind.. :-)
21:22:43CasainhoDomonoky: okok, I will draw in Kicad... however I will need to lear :-) - I plan to draw the "second board", with LCD carrier board + stereo audio codec + buttons
21:23:30DomonokyCasainho: yes, just try to make first version.. we can help out, if you have something..
21:23:34amiconnUmm, a few extra fps are lost due to the 9th bit sent along every data byte
21:23:59CasainhoI will try to use SPI for both, for to not have to write 2 diferent drivers :-)
21:24:41Casainhoamiconn: FPS are not important! this hardware will not be for video, just for an audio player, at max to show album art.
21:24:43DomonokyCasainho: this could be difficult, if they have different SPI formats, SPI is very flexible.. :-)
21:25:12Casainhookok... so I will chose maybe SPI for one and I2C for the other
21:25:47Domonokyyou need SPI or TWI *and* i2s for the dac
21:26:29amiconnFast updates are still rather useful, as they save cpu time and hence power for other stuff like decoding audio
21:26:36Casainhoyes, I know that :-) - I read here: "AT91-AN01: Using the Two-wire interface (TWI) in Master Mode on AT91SAM Microcontrollers" -
21:27:00amiconnYou could go down to 256-colour mode, which would increase frame rates by ~50%
21:27:15CasainhoI will copy the hardware from that document and from DSPdap hardware :-)
21:27:44Casainhoamiconn: there is no need for that frame rate:
21:28:08CasainhoShow RockBox running on some Free/Open hardware. After we can go to Olimex and ask them (or any others) If they are still interested on assembly and sell the hardware as a gadget on their online shop. We had contacted Olimex on 11 January 2007 and they said they will probably be interested and they even suggest a possible hardware configuration and price: AT91SAM9260 + LCD NOKIA 6610 + 8MB...
21:28:09CasainhoVersion 1 of the RockboxPlayer will be made on some development board having as base an ARM MCU. We should not seek to have exotic hardware, just simple and cheap, to try not spending time in coding drivers and spending money on some hardware that will change in future. Just do the necessary to have a functional hardware that plays audio from files stored on some kind of flash memory, and...
21:28:12Casainho...flash + 8/16MB SDRAM + nRF24L01 wireless chip + audio CODEC + Li-ion battery + charger + miniUSB + buttons, price in range of 100 €.
21:28:14Casainho...which displays files, audio information, on a display :-)
21:28:24amiconnTell that to those who e.g. want to use peakmeters in the wps
21:29:05Casainhoamiconn: please read the page about
21:29:06amiconnPerformance suffered a lot on G2 when there were peakmeters before the lcd river was optimised. That was at ~50fps (!)
21:29:34amiconn*lcd driver
21:29:41CasainhoI understand but that will be for a future 2nd version, not now! :-)
21:30:29CasainhoDomonoky: I have all that I need for draw the schematic of the second board, less the resistor with DAC...
21:30:31 Join BjoernEr1k [0] (
21:31:02Casainhowould not be simple to use digital inputs? - how does players like ipods and sansas do?
21:32:36 Join BobShield [0] (
21:33:36 Join petur [50] (n=petur@rockbox/developer/petur)
21:36:00stripwaxBagder - would the install directory for sims be renamed from "archos" to "modelname" too? :-)
21:36:23rasherstripwax: rather just "disk"
21:36:27 Join robin0800 [0] (
21:37:01 Quit tedrock (Client Quit)
21:37:49 Join testdasi [0] (n=8143906d@gateway/web/cgi-irc/
21:41:30testdasihello, I'm trying to compile rockbox (again) and installing Cygwin. Last time I remember there is a cache thing that I installed on Cygwin to make compiling Rockbox faster (between attempts). Now I completely forgot what it is so I would greatly appreciate if someone can tell me. I remember that I got the information from here.
21:42:39bertrikit's called ccache
21:43:22 Join tchan1 [0] (
21:43:24 Quit tchan (No route to host)
21:43:46 Quit BjoernErik (Connection timed out)
21:44:08 Quit Frazz (Read error: 110 (Connection timed out))
21:45:27pixelmaI thought that would have a negative effect in cygwin (slowing down instead of the speedup it gives in linux)?
21:45:35amiconnccache doesn't speed up compilation on cygwin, it even has the opposite effect (might depend a bit on disk speed though)
21:45:57stripwaxrasher - that would be better. is that rename planned?
21:46:07amiconnFor me, compiling with ccache on cygwin is ~20% slower than without
21:46:53amiconn(unlike on linux, where speed easily doubles if there aren't many changed source files)
21:47:16 Quit karashata (" I love my HydraIRC -> <-")
21:50:41 Quit testdasi ("CGI:IRC (EOF)")
21:51:38 Quit robin0800 (" HydraIRC -> <- Organize your IRC")
21:51:57 Join robin0800 [0] (
21:52:25stripwaxamiconn - is it bad that the mdct_arm.S patch got committed? (just reading your log comment from late last night)
21:52:42amiconnIt's not bad. It's good, and it works
21:52:44 Quit ch4os_ ("Lost terminal")
21:52:46stripwaxok :)
21:53:03amiconnThe only sub-optimal thing is that the gigabeat uses different code than the rest of the arm world
21:53:19 Join Zagor [242] (
21:53:30amiconnI'll soon post a patch that makes the gigabeat use the assembler code. Someone with a gigabeat should test it...
21:53:48amiconnI don't expect a major speedup, but I think unification is good
21:54:16LambdaCalculus37Good night, all!
21:55:22 Quit LambdaCalculus37 ("CGI:IRC (EOF)")
21:57:12 Join Llorean [50] (
21:59:42Leartestdasi: tip to speed things up in cygwin: run "svnversion > docs/VERSION" after updating. This way svnversion isn't run on each make, which makes for a noticeable speed difference...
22:00:32rasherLear: it's even run multiple times on each make..
22:01:51 Quit linuxstb_ ("ChatZilla 0.9.80 [Firefox]")
22:02:00amiconnOk, patch uploaded to fs:
22:02:16amiconnIt's almost trivial...
22:02:38amiconnNow we need a tester with a gigabeat...
22:02:51 Quit mf0102 (Read error: 104 (Connection reset by peer))
22:03:56amiconnZagor: Btw, is it normal that flyspray throws a ton of mime type warnings when posting a comment with an attached file before the updated page appears?
22:04:25Zagoramiconn: no, that doesn't sound right
22:04:38Learrasher: Maybe it did. Pretty sure it wasn't more than twice, if so...
22:04:48amiconnThe post itself was successful...
22:05:13rasheramiconn, Zagor: I've seen that as well
22:05:20Zagorwas the file plain source code?
22:05:38rasherThat, I don't remember
22:05:57amiconnIt was a .diff
22:06:21rasherLikely the same for me
22:06:40Zagorok. I'll look into that a bit later.
22:08:16scorche|wZagor: while we are adding things to the list, can you have a look at adding a link to the IRC guidelines on the CGI:IRC login page please? (
22:08:29Zagorgood idea
22:09:28pixelmaand also check why one can't /quit in the webclient (at least I couldn't 5 hours ago, ignore if it is already fixed)?
22:10:00ZagorI fixed that a few minutes after you said it. it simply wasn't in the list of allowed irc commands.
22:11:18pixelmaok, just wondered because I remembered doing that before
22:12:10Zagoryeah I set up a more strict configuration when I upgraded version cgiirc a few weeks back
22:13:02pixelmaI see, nevermind then
22:14:04JdGordon|wwhile we are talking about the web client... the Status tab has a screen full of "Access to that command is not allowed" warings
22:14:15JdGordon|wI have no idea which command is triggering it though
22:14:33pixelmaI had that with /quit /part and /leave
22:14:56JdGordon|wits not any of them.... i havnt used em
22:15:09preglowamiconn: i believe markun said there was no gain from the asm
22:17:25 Quit robin0800 (" HydraIRC -> <-")
22:17:40 Join robin0800 [0] (
22:17:57amiconnpreglow: Maybe, but if there's no loss, it should still be used imho
22:21:14preglowamiconn: guess i agree, yeah
22:25:49 Join Rondom [0] (
22:28:12JdGordon|wcan we do devcon at the beginging of august? say straight after wacken?
22:28:34*scorche|w googles wacken
22:28:59 Quit jhulst ("Konversation terminated!")
22:29:23scorche|woooo...type o was there last year
22:29:52scorche|wi dont know...august might be a bit late...
22:30:45CasainhoDomonoky: I have free time tomorow, I will draw the schematic and put on TWiki page :-) - please see then and tell what you think. bye bye :-)
22:30:53 Quit Casainho ("ChatZilla 0.9.80 [Firefox]")
22:31:08*n1s is planning to go to wacken too :9
22:33:07 Quit JdGordon|w ("CGI:IRC")
22:33:35 Join JdGordon|w [0] (n=836b0065@gateway/web/cgi-irc/
22:35:01***Saving seen data "./dancer.seen"
22:40:51XavierGramiconn: do you need a gigabeat F tester?
22:41:33amiconnF or X, doesn't matter
22:42:34 Join DerPapst [0] (
22:42:54*DerPapst wonders who he'll meet at wacken ;-)
22:43:09 Quit desowin (Read error: 104 (Connection reset by peer))
22:43:36pixelmaRockbox people can be recognised by their light blue t-shirts :P
22:45:17 Nick tchan1 is now known as tchan (
22:45:17XavierGramiconn: what should I test for after I apply the patch, ogg playback?
22:46:10amiconnThe problem was that just enabling the asm without adjusting the section made the vorbis codec crash
22:46:42amiconnA nice extra would be doing a test_codec comparison with/without patch
22:47:48XavierGrcompiling now
22:47:53XavierGrok I will do one too
22:49:01amiconnNot that optimising an audio codec would be important on gigabeat, but still...
22:52:13 Quit Lear ("ChatZilla 0.9.80 [Firefox 3.0b3pre/2008012204]")
22:52:44stripwaxDoes anyone know if we're using any of the Tremolo optimisations for ARM?
22:54:32 Quit Jon-Kha (Remote closed the connection)
22:57:42 Quit Domonoky ("Trillian (")
22:59:10pregloweh, no
22:59:13preglowi don't believe we do
22:59:15preglownever seen the thing
22:59:41preglowstripwax: would be most interesting to see how it works...
22:59:47 Join desowin [0] (n=desowin@atheme/developer/desowin)
23:00:13stripwax'how' or 'how well' ?
23:00:37preglowwell, i expect it to work :)
23:00:49preglowspeed test and how much faster
23:00:55preglowper component, of course...
23:02:14*stripwax builds..
23:02:24stripwaxwill just check out boost ratio first
23:02:48stripwaxehm, are the pp502x builds little endian or big?
23:03:19stripwaxjolly good, thanks
23:03:22PaulPositionHow can I 'backtrack' my svn download to an earlier revision (say, r16133).. Is there an easy way or would I need to find out the exact date/time and such?
23:03:46Nico_PPaulPosition: svn up -r X
23:03:48stripwaxsvn update -r 16133 or something?
23:04:01PaulPositionCool, thanks.. Very useful. :)
23:10:05XavierGrdo I have to compile test_codec or is it on by default?
23:10:49preglowhave to enable
23:11:41 Join uwe_ [0] (
23:13:24XavierGrdo I have to put it on the viewers.cfg too?
23:17:39 Quit jcollie ("Ex-Chat")
23:17:54soapBuschel (reading the logs?) New patch with same name as old patch?
23:19:36 Join tedrock [0] (
23:20:49 Quit tedrock (Client Quit)
23:24:35 Quit uwe__ (Read error: 110 (Connection timed out))
23:28:04stripwaxXavierGr - I think it can be run as a viewer - can't remember if it *has* to be run as a viewer
23:29:21XavierGrwell i had to add it in viewers in the end
23:29:27stripwaxok :)
23:32:59XavierGramiconn: with SVN Decode time: 233.98, 599.75% realtime
23:33:26XavierGrbah that was with SVN + your patch
23:34:28XavierGrwith SVN only it is: Decode time - 233.01, 602.24% realtime
23:34:45XavierGrso yeah your version makes it a little bit slower
23:37:26amiconnHmm. Did you measure several times?
23:37:45amiconnThere might be some measuring uncertainty. The difference is small...
23:37:50 Quit MethoS- ("Konversation terminated!")
23:38:16 Join trypt [0] (n=hi@
23:38:56amiconnBut if it's indeed a bit slower, I wonder what would be the best thing to do...
23:39:17amiconnThe asm might be tailored for arm7, and the gigabeat f/x is arm9
23:39:18XavierGramiconn: yeah of course it could be, let me try once more
23:39:29trypthey guys, anyone having problems with latest ipod video (60/80) rockbox build? disk access when starting a song is through the roof, takes 30 seconds or longer
23:40:22 Quit davina ("GNU/Linux the free alternative to Windows")
23:40:23 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
23:42:11saratogapreglow: did you look at that ARM optimized ogg decoder?
23:43:49preglowsaratoga: which?
23:44:25 Quit trypt ()
23:44:29saratogathe tremorlo
23:44:39preglowno, i'm eagerly awaiting feedback from stripwax :)
23:44:48saratogait looks really interesting
23:45:02saratogaalmost the entire mdct is in ASM
23:45:26saratogai bet its waaaaay faster then what WMA and FAAD are using
23:45:58preglowhopefully yes
23:46:06amiconnIt might be faster, but I doubt it will be much faster
23:46:11preglowi don't think so either :/
23:46:16saratogawhys that?
23:46:40preglowwell, they claim 20-30% total speedup, and there's a lot of other asm code
23:46:46preglowso it can't have yielded much
23:46:47amiconnThe limiting factor for mdct on arm seems to be the slow multiplier
23:46:55 Quit stewball (Read error: 113 (No route to host))
23:47:11saratogatrue but theres a lot of flexibiliity in how many multiplies you actually do in an FFT/MDCT
23:47:17amiconnI wonder what those 20..30% are measured on. arm7? arm9? other arm?
23:47:25saratogai suspect that some codecs do a lot more then what they need
23:47:27amiconnThey are quite different in execution timing
23:47:49 Quit BobShield (Read error: 104 (Connection reset by peer))
23:47:55XavierGramiconn: 2nd test came even slower at 599.64% and 234.02s decoding time
23:48:06amiconnAnd without adjustment, it might even be significantly slower on PP5002
23:48:17saratogai really need to profile WMA, Ogg and FAAD, they each have completely different MDCT algorithms, would be neat to see which is fastest on ARM7
23:48:18stripwaxI'm still trying to work out how to change the code to make it work with the rockbox-modified version of tremor..
23:48:27XavierGrthat's quite strange though, I remember that I fiddled with the backlight more on the first test
23:48:29 Join trypt [0] (n=hi@
23:48:40preglowsaratoga: faad author tells me the faad mdct isn't very fast
23:48:57XavierGrbut anyway the slow down is minimal
23:48:59saratogai wonder if its slower then WMA
23:49:11stripwaxfwiw tremolo makefile says -march=armv4 -mtune=xscale
23:49:52stripwaxand the makefile is a windows ce make job
23:50:53amiconnSo it builds for arm v4, but tuned for xscale which is an armv5 core, that is arm9 or higher afaik
23:53:24tryptare there known problems with latest ipod video (60/80) build? regarding playing songs in file mode, the ipod juns spins the disk drive and after waiting about 30 seconds i give up and restard the ipod then turn it off
23:54:32stripwaxtrypt - any ideas what it's doing for those 30 seconds (what does screen say?). what audio format? and do you have database and dircache enabled?
23:54:47 Quit jgarvey ("Leaving")
23:55:09tryptit goes to the normal playing song screen, with the information loaded, but it doesn't start playing, just sits at 0 seconds
23:55:12 Quit petur ("gonne")
23:55:23tryptit's mp3 (vbr)
23:55:51stripwaxtrypt - hm, could that particular mp3 be corrupted maybe? does it play on your PC for example?
23:55:53trypti've noticed it'll start playing the first song in the playlist after about 5-10 seconds, but when it tries to go to the 2nd song it takes forever
23:56:01trypti've tried many files
23:56:06tryptand yes they work on my computer
23:56:28 Quit desowin ()
23:56:29stripwaxlatest build as in revision 16151?
23:56:29trypti had about a month old version of rockbox, updated last night and this problem started happening
23:56:44tryptwell i updated about 10 minutes ago
23:57:24tryptbut it was happening on last night's version as well

Previous day | Next day