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 2005-07-04

00:00:24 Join Cassandra [0] (
00:02:33ColdtoastI'd hate to suggest your HDD's faulty....
00:02:45Coldtoastor becoming so
00:02:58stripwaxbut only faulty when using rockbox fw? seems suspicious..
00:03:08Coldtoastoh? yeah
00:03:12Coldtoastthat is weird
00:04:21Coldtoastonly thing I hear is the usual slight whirring noise of a spinnin gHDD and the head moving
00:04:30stripwaxit's a really quite beep/squeal by the way. i'd imagine i don't even notice it a lot of the time, unless the room is pretty much silent.
00:05:07stripwaxbut it only happens during first bootup / first access of the file tree / first buffering of a file. after that, it doesn't seem to happen again.
00:05:13stripwaxunless I turn off/on the player again
00:06:37stripwaxhm, any good utilities to thoroughly disk-check a usb drive?
00:06:50stripwaxor would i have to pull it out and fit a 1.8"/3.5" adapter?
00:07:18HCldoes anyone happen to know a good way to get adrenaline out of your system?
00:07:27Coldtoastgoogling doesn't offer any very promising info
00:07:35Coldtoastjust stuff about imminent failure
00:07:47ColdtoastHCl: meditation
00:07:51HClColdtoast: hm.
00:07:57HClgood idea.
00:08:01HCli'm buddhist after all.
00:08:08Coldtoastthere you go then
00:08:18stripwaxbeep,clunk. beep/clunk.. repeatedly while trying to scandisk from windoze :-(
00:08:18amiconnMrf, I *hate* that remote hold switch
00:08:35HClstripwax: sounds like a faulty harddisk... :/
00:08:41HClit shouldn't react like that in usb mode :/
00:08:49stripwaxHCi quite ... :-(
00:09:19stripwaxooooh, very noisy, it's scaring me now
00:09:26*HCl goes afk to try to meditate around the various game sounds coming from gta san andreas to the left and burnout3 on the right
00:09:51Coldtoaststripwax: and you're sure it NEVER does it with the iriver fw?
00:10:15stripwaxnever. could be a horrible coincidence tho if it's doing it in usb mode :-(
00:10:25HClyea :/
00:10:27stripwaxscandisk with 'attempt repair of bad sectors' by the way.
00:10:42HClreplace or reflash with iriver firmware, strip rockbox off it and claim warranty
00:10:44Coldtoastwhat about trying USB mode via the iriver fw?
00:10:54stripwaxColdtoast hm, worth a shot
00:11:13HClthey shouldn't differ since usb is all handled by the usb chip
00:11:18HClbut you can try, i suppose
00:11:35*HCl is still running on stress/overdrive cause of adrenaline... x.x
00:11:37Coldtoasthe says it doesn't beep with the iriver fw
00:11:39HCli'd better go afk completely...
00:11:44stripwaxerk, battery too low for iriver fw to boot. hm, wonder if that's related?
00:11:56Bagdercould very well be
00:12:14stripwaxi'll try checking again tomorrow after a full recharge
00:12:14Coldtoastyeah. there's no chance rockbox drains the batt too low is there?
00:12:23Coldtoastthat's been confirmed to be unable to happen?
00:12:35stripwaxColdtoast it definitely drains it lower than iriver fw
00:12:46Bagderhow would that hurt the disk anyway?
00:12:47Coldtoastnot too low tho?
00:13:15stripwaxand i've had it in the case before where rockbox couldn't boot: disk would spin up, stop, spin up, stop, as if there wasn't enough power for disk. hope that wouldn't hurt disk?
00:13:31Coldtoastoh yeah. that was only the other day right
00:14:10HCli'd really prefer an auto shutdown before it hits 3volts... afk now..
00:14:18stripwaxColdtoast i last noticed that maybe four weeks ago - i've been careful to not let batt get too low since then
00:14:22ColdtoastI'm down to 1 bar myself
00:14:27stripwaxalthough maybe it already is too low, dunno
00:14:45Coldtoastcan you still RMA if you need to?
00:16:09stripwaxColdtoast - not sure. probably not, got it 18months ago.. but someone says European players have a 2yr warranty? I'll see if I still have the box somewheres..
00:16:17stripwax.. or i could just buy a 30GB hdd...?
00:17:04Coldtoastyeah. you may as well upgrade it if you do have to replace the drive
00:17:07ColdtoastI would
00:17:32ColdtoastI'd also stick a 2200mAH battery in there while I was at it cos they're SO cheap :)
00:18:11stripwaxhm, am I imagining that cpu is boosted while the backlight fades out?
00:18:41CassandraCurse this digital transfer mode. Really shows up the limitations of my crappy 5.1 speakers.
00:18:46stripwaxseems a spooky coincidence that the info always shows boost when screen fades in/out
00:19:17amiconnyes it is
00:19:30stripwaxamiconn - why is that?
00:19:40amiconnThat is so just to avoid flicker, which would occur when the frequency changes while fading
00:20:03amiconnThe fading is timer interrupt based, and the timer depends on the cpu clock
00:20:51stripwaxBatt 17% apparently. Too low for iriver fw to boot
00:22:07stripwaxamiconn hmm ok. could it use TIME_AFTER or something instead and be independent of cpu clock?
00:22:08Coldtoastwhere do you get batt as a %?
00:22:24amiconnstripwax: No and no
00:22:36stripwaxhm ,just jumped to 22%. silly. Coldtoast - view IO ports in Debug
00:22:42stripwaxamiconn - why not?
00:23:27amiconnTIME_AFTER() is a macro using current_tick, which (1) also glitches a little while changing cpu freq and (2) is much too low resolution for flicker-free fading
00:23:57 Quit Aison ("( :: NoNameScript 3.72 :: )")
00:24:15amiconncurrent_tick also happens to use a timer
00:24:16stripwaxamiconn - how do you mean by 'glitches' ? but understood on second point
00:24:54amiconn...and a timer counts cpu cloks, so the count cycle has to be adapted when cpu freq changes
00:25:00*HCl feels a bit better :/ are there still reds caused by HCl ?
00:25:29BagderHCl: yes, those lacking functions in the archos version
00:25:31amiconnHCl: yes :(
00:25:43amiconnNo, the sims are red
00:25:48HClamiconn: do you think you have time to implement those functions or should i just define them for iriver only for now?
00:25:55Bagderaha, only sums
00:25:55HClor maybe non simulator only
00:26:10*Bagder types so very crappy
00:26:24HClah... in my stress i misread h300 as h100 and thought there were errors not caused by me..
00:26:50amiconnThe simulators should have stub functions instead (probably in playback.c)
00:26:52HClshould i just disable the runtime database on the simulators for now? the runtimedb registers, but will just never be called back
00:27:21HClthe archos targets don't have playback.c in sim though
00:27:29amiconnAh, yes
00:27:38amiconnThat's probably the reason
00:27:45amiconnThey don't have playback.c at all
00:27:47Bagderso add the stubs in a simulator-specific file, I say
00:27:50HClyea, since iriver sim doesn't red
00:28:01amiconn...and I wonder why the target builds are *not* red
00:28:13HClyou haven't added the functions, i assume?
00:28:16Bagdercrappy script
00:28:29Bagderthe target builds fail as well
00:28:42HClso should i add stub functions in mpeg.c or?
00:28:47HCli really don't know much about archos..
00:29:25Bagdershouldn't these function be common for either platform?
00:29:39amiconnI'd say just disable the registering in database.c for archos for now
00:29:40HCllet me correct that logf define a bit
00:29:49HClthat find function does way more than just a check
00:29:55HClits required to be called, just without the if
00:29:56amiconn...with a comment
00:29:58HClwithout a logf
00:30:07BagderHCl: sorry
00:30:09amiconn...then it can be reenabled once the funcs get implemented
00:30:38amiconn...and yes, they'll go into mpeg.c for archos
00:31:16HClits okay
00:31:22amiconnBagder: Could someone pls fix these colouring scripts?
00:31:28*amiconn hides
00:31:42HClamiconn: k, will do
00:32:06amiconnI often rely on the last status line. All green means don't worry for me, unless I see something going wrong myself
00:33:37HCljust fixed the red builds and corrected the findrundbentry thing, hopefully
00:34:07 Quit cYmen ("zZz")
00:36:05HCli'd appreciate it if more people could test runtime database functionality on iriver
00:36:22HCli haven't had much time to check it properly, i only confirmed that it worked
00:36:54 Quit matsl (Remote closed the connection)
00:37:12HClsorry about the commit earlier :/ i was far too stressed
00:37:16Bagderwell, what can we do with it?
00:37:22Bagderif I enable it I mean?
00:37:28HClyou can use the search engine
00:37:38HClto make search querys containing the playcounta
00:37:42HCland autorating
00:37:49HClwhich is a rating based on playcount ranging from 1 to 10
00:38:09HClafter enabling the database and playing some songs
00:38:21HCldatabox will be able to generate a search query file for you..
00:38:32Bagderis that the file name it asks for?
00:38:47HCli still need to put a lot of work in databox and searchengine
00:39:11HClafter the search query is created just play it and searchengine will run and create a /search.m3u with the results
00:39:17Bagder_some_ docs would help ;-)
00:39:23HCli know i know :/
00:39:27HCli'll add it to my todolist..
00:39:44HCltry recreating a blank database while keeping the old database
00:39:50CassandraBagder: do you know who did the last revisions on the Windows Installer? If it were checked in, I could try to get it working for iRiver too.
00:39:55HCland check whether it self-repairs properly, it will tell you in the logf log
00:40:16HClsorry, thats a lot of bugtesting, i'll do part of this myself as well
00:40:17BagderCassandra: it might've been Linus
00:40:33HClwhile keeping the old runtime database*
00:40:41amiconnCassandra: I did the windows installer extensions. Dunno if Linus did some further changes
00:41:01Bagderah, ok. I may be hallucinating
00:41:21BagderI blame linus for that too ;-)
00:42:09amiconnBagder: Just to make sure, I think Cassandra means the windows installer supporting all archoses so far, not the windows fw patcher for iriver
00:42:19CassandraGuess we'll have to check with Linus before we can get the most recent version into CVS.
00:42:31Bagderamiconn: yes, I meant that too
00:43:04*Cassandra thinks evil thoughts about including the fwpatcher in the installer.
00:43:10Bagderamiconn: now check the cvs table ;-)
00:43:13amiconnI still have that package laying around, but I don't think this can be checked in as-is
00:43:22stripwaxCassandra - someone was working on something like that already. Possibly 'paul'?
00:43:38Bagdernot really perfect, but...
00:43:41stripwax.. now that fwpatcher accepts cmdline args, and all...
00:43:45CassandraWell, we should check in whatever's currently being used to build them.
00:43:48amiconnBagder: linklinklinklinklinklinklink ???
00:44:09Bagderwhere they previously were green
00:44:16 Nick gromit` is now known as gromit`MOURETTE (~gromit`
00:44:28Cassandrastripwax: Cool - even less work for me.
00:44:32HCli see a few that don't seem to have errors though...
00:44:36HClthe win32 sim...?
00:44:41HClfor the player
00:44:45 Nick gromit`MOURETTE is now known as gromit` (~gromit`
00:44:49HCland the recorder
00:44:56 Join webguest48 [0] (
00:44:57stripwaxwhatever happened to 'paul'?
00:45:06BagderHCl: they failed to link as well
00:45:12CassandraI know him not.
00:45:14HClsorry, mispaste
00:45:24HClBagder: mmm, didn't see that, but ok
00:45:42CassandraIs it true that the H3xx doesn't have Optical in/out, or did I imagine that?
00:45:54HClno, i think thats correct
00:46:08*HCl sees the musquitos buzzing around him, thinks its time to move to his room
00:46:16CassandraWhat is this thing manufacturers have about removing features from their products.
00:46:31CassandraI wonder how difficult it is to retrofit a H300 screen to an H100.
00:46:37Bagdercost and/or DRM I guess
00:47:00CassandraIf it's DRM, they need to get a freaking backbone.
00:47:11CassandraWe're the consumers, not the record industry.
00:47:40Bagderyes, but then we are not forced to buy h3x0 players ;-)
00:47:47CassandraWell, quite.
00:48:02CassandraDoubt I'll be "upgrading", although a colour screen would be nice.
00:48:02 Quit webguest48 (Client Quit)
00:53:02HClwhy is it still failing..
00:53:15CassandraSod's law.
00:53:53*HCl fixes
00:54:18HClIf it stinks, it's chemistry. If it moves, it's biology. If it doesn't work, its computerscience!
00:56:09CassandraIf it's false, it's statistics.
00:58:41 Join TCK [0] (
01:00:12 Join muesli- [0] (
01:00:38stripwaxok gnight
01:00:55stripwaxi'll let yall know if my hdd really is stuffed :-(
01:01:35 Quit stripwax (Read error: 104 (Connection reset by peer))
01:01:47HClwhat does the h300 sim look like? just the same as h100?
01:01:51Coldtoastwee! look at all that red. Makes me want to go watch War of the Worlds again
01:01:52Bagderheh, the script did wrong on two of the three win32 sim builds
01:02:14BagderHCl: yes, there's no adjustments made for the colour lcd done yet
01:02:21Bagderonly for size
01:02:50BagderI haven't actually ran it though
01:02:59HClohyea, size differs..
01:03:20HClbah... gay flatmate is playing loud music while its 1 am.. :/
01:03:33muesli-kick him like bekcham :D
01:03:39Bagderok, the two existing red "link" table cells are due to script flaws
01:03:54Bagdershould be fixed in the next build
01:04:10 Quit Coldtoast (Read error: 104 (Connection reset by peer))
01:04:12HCli think i fixed archos builds till the events are added
01:04:23HCli'll bugtest a bit tomorrow
01:04:25muesli-sorry, are you talking about working h300 display?
01:04:35HClmuesli-: sim
01:04:48muesli-you mean theoretically?
01:04:53HCli was just wondering cause the build table says it compiles an h300 simulator
01:05:00Bagdermuesli-: the simulator builds and runs
01:05:18Bagderbut it isn't fully simulating a 300
01:05:26HClhardly simulating, i'd say
01:05:28HClbut anyways
01:05:43Bagderit simulates lots of the code that will run on the 300
01:05:50*HCl goes to nap/rest a bit
01:05:52HClhm, okay
01:06:21muesli-kewl :D i am seriously thinking about exchanging my 140 for a 3xx
01:06:33BagderI wouldn't
01:06:43muesli-why not?
01:07:02Bagderthe only plus it has IMHO is the LCD
01:07:04CassandraCrap plastic case, lack of optical support.
01:07:17CassandraCrap control keys.
01:07:19muesli-i love its color display
01:07:32*amiconn wouldn't have expected that
01:07:36muesli-sorry, amiconn ;)
01:08:03amiconnIt seems my greyscale drawing routines are running. This was the first try!
01:08:21CassandraShame the controller chips are different. If they weren't Rockbox could theoretically drive a 3xx display on a 1xx.
01:08:43amiconnCassandra: I wouldn't want that
01:08:59CassandraThat's OK. You wouldn't have to have it. :)
01:09:07amiconnMost colour lcds are battery eaters
01:09:17BagderI assume this is one
01:09:26muesli-yeah true..but watching jpg's would be great
01:09:26Bagderthe iriver firmware shuts off the h300 display
01:09:41amiconnThere *are* some that don't need a backlight to be readable, but so far I only know of one device series that uses them
01:09:42Bagderwhen the h100 displays without backlight
01:10:06CassandraAFAIK the 3xx series has a slightly longer advertised runtime than the 1xx.
01:10:15amiconn...and that are the Blackberry 72x0 cellphone/email push clients
01:10:28muesli-Cassandra because of the missing lcd remote probaby
01:10:29BagderCassandra: but that is at the price that you can't see what it does
01:11:01amiconnI remember someone saying the H3x0 lcd is also pitch black when not backlighted it is the bad type
01:11:12Bagderamiconn: I don't think it is without backlight
01:11:13muesli-yeah, notthin 2 xc
01:11:16BagderI think they shut it off
01:11:31CassandraI'll stcik with the 1xx.
01:11:31 Quit ghostiger (Remote closed the connection)
01:11:33muesli-yepp, u cant see anything
01:11:52muesli-but i would work with my 1x remote anyway
01:12:01CassandraUnless someone proposes to send me a 3xx so I can get the manual right.
01:12:08muesli-but would love as i said before to watch jpgs
01:12:08Cassandra(That's so likely to happen)
01:12:14muesli-taken with my cam
01:12:25*Bagder runs off to bed
01:14:01*muesli- steals it
01:15:11CassandraWish I had an Ondio and a Player. Would make my life a lot easier.
01:18:04muesli-whats so special about an ondio?
01:18:49HClondio would be way too small for my needs..
01:19:58CassandraIt's hard to write the manual without one in front of me.
01:20:04CassandraNot impossible.
01:20:05CassandraBut hard.
01:20:36amiconnHmm, good news and bad news
01:20:51amiconnThe good news is that my code is working
01:21:44amiconnThe bad news is that it seems the 4-grey mode of the lcd seems to use the same mechanism to generate the grey levels as the grayscale lib for the archos does
01:22:00amiconni.e. dithering in the time-domain
01:22:47amiconnThat means it won't be possible to add our own temporal dithering to that and get more shades of grey
01:23:12amiconnWe could still do this, but with the display itself in b&w mode
01:24:01amiconn...but that means the grayscale lib will have to take over the whole display; no overlay mode like on archos possible
01:24:21HClmaybe a mode switch?
01:24:26CassandraYou can do overlay on Archos?
01:24:28HClbeing able to switch to b&w for dithering
01:24:35HCland just native 2bit grayscale normally
01:24:38amiconnHCl: Yes
01:25:01HClokay, sorry, i didn't see that you already had that in mind :)
01:25:14amiconnOn archos we don't have to switch, so the greyscale area can take just a part of the display, with the other parts still being able to display in b&w
01:25:25 Quit thegeek (Read error: 131 (Connection reset by peer))
01:25:25amiconn...using the core graphics routines
01:25:43amiconnThat won't be possible on iriver when we switch the mode
01:26:03 Join thegeek [0] (
01:26:32amiconn...and of course my calculation of possible shades of grey is rendered invalid. It will be the same as on archos, provided the display doesn't flicker too much
01:28:46amiconnI will add a slight effect to make greyscale mode visible in the UI until more conversion is done
01:28:49HClwe could probably display the rest in black and white rather than the 2bit grayscale when switching to grayscale lib?
01:29:12HClor maybe even simulate the 2bit grayscale using the grayscale lib itself
01:29:14amiconnHCl: Yes, but that would require another set of graphics routines for b&w mode
01:29:40amiconnThe lcd framebuffer differs between the modes
01:29:40HClwhat about the second bit?
01:29:49HCloh, yea
01:29:59HClah well, never mind me
01:30:04HClyou're the expert on this :p
01:30:07amiconnBut I'd say, one step after the other
01:30:19amiconn(1) 4-grey mode for the ui and some plugins
01:30:21HCli'd better sleep
01:30:35amiconnThe grayscale lib may come later...
01:30:53HCli agree
01:32:00amiconnBtw, the slight effect will be a light grey background in the splash() boxes
01:32:21CassandraSounds good. :)
01:34:52amiconnI won't believe *how* slow the lcd is in 4-grey mode
01:35:13amiconnIt's significantly worse than in b&w mode when grey is actually used
01:36:10 Quit thegeek (Read error: 131 (Connection reset by peer))
01:37:22*Cassandra downloads Pink Floyd's Live 8 performance.
01:43:04 Join tvelocity [0] (
01:48:33***Saving seen data "./dancer.seen"
01:53:08muesli-n8 fellows
02:01:58 Quit muesli- (Read error: 145 (Connection timed out))
02:03:23*amiconn just noticed that lcd_mono_bitmap() and lcd_bitmap() are *not* identical even on b&w displays
02:15:32 Join thegeek [0] (
02:23:54 Join amiconn_ [0] (
02:28:07 Quit amiconn (Nick collision from services.)
02:28:08 Nick amiconn_ is now known as amiconn (
02:38:08 Join austriancoder [0] (~austrianc@
02:52:08RoriPink Floyd rocked
02:58:36 Quit austriancoder ("using sirc version 2.211+KSIRC/1.3.12")
03:48:35***Saving seen data "./dancer.seen"
04:05:17 Join QT_ [0] (as@area51.users.madwifi)
04:08:59 Quit QT (Read error: 148 (No route to host))
04:13:39 Quit pill (Read error: 145 (Connection timed out))
04:19:05 Join ashridah [0] (
04:43:26 Quit Sucka ("a bird in the bush is worth two in your house")
04:44:00 Quit lostlogic (
04:46:41NJoinlostlogic [0] (
04:58:46 Join pill [0] (
05:28:02HCli hate dreams in which my dad dies
05:31:51ashridahyou have them that often?
05:33:20HClsecond time in a month
05:33:49HCland i already have trouble with not sleeping enough, this isn't helping, heh
05:35:08*HCl tries to go back to sleep x.x
05:39:02*Rori is running a temperature
05:44:57Rorionly a bit though
05:48:36***Saving seen data "./dancer.seen"
06:54:21 Quit tvelocity ("Leaving")
06:59:30 Join CheeseBurgerMan [0] (
07:12:11 Quit pill (Read error: 110 (Connection timed out))
07:22:54 Nick CheeseBurgerMan is now known as CBM-away (
07:45:07 Join pill [0] (
07:48:37***Saving seen data "./dancer.seen"
08:03:19 Join Bger [0] (~Bager@
08:03:31Bgermorning all;)
08:16:38SlasheriHmm, can i enable logf for simulator?
08:23:15 Join B4gder [0] (
08:23:23Slasherior how am i supposed to output debug information, DEBUGF?
08:23:44CassandraUnder Linux, printf should work.
08:23:55amiconnDEBUGF() it is
08:24:20CassandraWhat does logf do anyway?
08:24:25amiconnCassandra: printf() is bad because it isn't conditional, i.e. such code will fail to compile for target
08:24:54amiconnlogf() is similar to debugf(), but it logs into a memory buffer instead
08:25:05amiconn(and displays on the remote for H1x0)
08:25:18CassandraThat's kind of cool.
08:25:31Slasheriamiconn: Maybe logf can be defined also to DEBUGF for simulator targets?
08:26:04amiconnI agree, it's kinda redundant for the sims
08:26:08SlasheriEh.. :D after i ran the rockbox simulator, now my keyboard repeat is totally off :D
08:26:20amiconnImho this could be done another, possibly easier way though
08:26:33amiconnSlasheri: Did you kill it?
08:26:42Slasheriamiconn: yes, it crashed
08:26:59amiconnThe sim has to disable autorepeat for its own button handling
08:27:00Slasheriah, rerunning it worked
08:27:04Slasherinow repeat works too :)
08:27:05B4gderthe fact that logf works the same in the sim is good for debugging the logf functionality
08:27:23amiconnIt's supposed to restore it, but obviously can't if it crashes
08:29:42amiconnB4gder: Agreed, but otoh it would be easier to only have one debug output macro. It's rather unlike to use logf() and debugf() at once
08:29:55B4gderI agree
08:30:32B4gderwe could have an unset define to set when we want to debug logf
08:30:57B4gderand make logf() basically do printf() in the sim
08:31:40amiconnMy idea was to drop the LOGF() macros, and make DEBUGF() == logf() for logf builds, and DEBUGF() == debugf() for debug builds and the simulator
08:31:50B4gderworks for me
08:32:06B4gderbut then we probably need to cut out a few existing DEBUGF() calls
08:32:21amiconnHmm, perhaps
08:33:02amiconnOr we could make LOGF() == DEBUGF() for the sims,
08:33:19amiconn...and leave it like it is for target
08:34:06B4gdereither way, it also makes sense to use local versions for specific debugging, like the linux kernel sources do
08:34:15B4gderlike if you want to debug a specific module:
08:34:32amiconnDifferent topic: I'm still a bit undecided about the names for the various bitmap functions
08:34:35B4gderyou add a #define MYDEBUG DEBUFG() in the top of the file
08:35:25amiconnFor all current code, lcd_bitmap() would become lcd_mode_bitmap() with my current proposal
08:35:44amiconnErm, lcd_mono_bitmap() of course
08:36:01amiconnA bit long...
08:36:33amiconnB4gder: Local debugging makes sense, yes
08:45:56ashridahhrm. the vorbis codec doesn't do resampling yet does it
08:47:09B4gdergood to see the table going green again
08:48:18amiconnB4gder: What do you think about the bitmap function names?
08:49:01amiconnI have 4-grey mode already running on target btw, optimsed. Some functions are still missing tho, and the simulator display is looking funny
08:49:10B4gderI haven't read the proposal in detail lately so I don't remember
08:49:52B4gderis the win32 sim capable of showing greyscale?
08:50:02amiconnSimulators only display the top half, stretched vertically by 2
08:50:12amiconnIt will be before I commit
08:50:54amiconnI can't do that for the X11 sim though, I'll at least adapt the x11 sim to map dark grey->black and light grey->white
08:51:25B4gderI'll have to dig up some code that does colours in x11 and see how we can do it
08:51:32amiconnI have no clue how to use more than foreground and background pixels
08:52:02amiconnI wonder why the x11 sim uses such a starnge mechanism (setting/resetting individual pixels) instead of just using a bitmap
08:52:39B4gderI made it just because I didn't know how to do a bitmap
08:52:43amiconnThe windows sim uses a bitmap object; it's quite simple to adapt to greyscale
08:52:51amiconnAh ;)
08:54:03B4gderand apparently nobody has cared enough to fix it ;-)
08:54:12amiconnThe win32 simulator bitmap will need some more adaption when it comes to >8 bit depth
08:54:31amiconnShouldn't be too hard either
08:55:15 Join Coldtoast [0] (
09:04:25Coldtoastthe " Update the wps when next track info is available." fix still doesn't work when %s is used
09:08:07Coldtoasthey B4gder
09:09:16Coldtoastwill it be possible or will next track info be classed as dynamic?
09:09:56B4gderyou want me to explain the flaw once more?
09:10:07Coldtoasterr, ok
09:12:21 Quit B4gder ("go go go")
09:13:03 Join B4gder [0] (
09:13:42 Join as [0] (
09:13:54 Join bipak [0] (
09:13:57 Join amiconn_ [0] (
09:14:13ColdtoastB4gder: I realise it's very frustrating for you and that ppl keep mantioning it
09:14:44B4gderI just don't see why the same people have to repeat it over and over
09:14:51 Join Gibbed [0] (
09:14:52 Quit Rick (Read error: 131 (Connection reset by peer))
09:14:52 Quit thegeek (Read error: 131 (Connection reset by peer))
09:14:53 Quit bipak_ (Read error: 131 (Connection reset by peer))
09:15:07 Nick Gibbed is now known as Rick (
09:15:48Coldtoasthmm. wouldn't you expect it to be mentioned after things like " Update the wps when next track info is available." are listed in the updates?
09:16:10B4gderthose are CVS commit messages
09:16:16B4gderthey are for devs
09:16:37CassandraMaybe there should be something in "Features we won't implement"?
09:17:01Coldtoastfine. I'll never mention it again
09:17:03B4gderCassandra: this is a bug/flaw, mentioned in the bug tracker
09:17:21B4gderColdtoast: I'm pretty sure you'll get to know or figure out when it works
09:17:36ColdtoastI'll just say one more thing tho, as a final thought
09:17:37B4gdersince the person doing the fix will be a happy one
09:17:52amiconn_This is a loooong-standing bug
09:18:05B4gderwith emphasis on the loooong
09:18:06amiconn_Like RLD was :)
09:18:28B4gderI think RLD is older though
09:18:56Coldtoastah screw it. I won't
09:20:01 Join Musicmad [0] (
09:21:07amiconn_Bagder: RLD *was* older :)
09:21:39*B4gder stands corrected
09:22:04CassandraI take it a fix is non-trivial then? Is it worth me having a look at it, or am I likely to get lost in a maze of twisty code?
09:22:22Cassandra(I'm thinking if I fix it, people will quit whining about it.)
09:22:32B4gderthe problem is to update a scrolling line with info, without resetting the scroll
09:22:32Cassandra(And whine about something else instead.)
09:22:33 Quit amiconn (Read error: 145 (Connection timed out))
09:22:33 Nick amiconn_ is now known as amiconn (
09:23:02Cassandra*nods* Tricky, but it ought to be feasible.
09:23:14B4gdergood summary :-)
09:23:25CassandraOK. I'll have a look at it later in the day when I'm awake.
09:23:54amiconnB4gder: I have plans for a substantial rework of the scrolling
09:24:07B4gderfor multiple fonts etc?
09:24:23amiconnMultiple fonts will come even later
09:24:37amiconnMy first steps will be pixel-wise positioning and right-clipping
09:24:41CassandraMmmm. Multiple fonts.
09:24:49amiconnSo we can have a scrolling box
09:24:52B4gderamiconn: ah, neat
09:25:04amiconnI need that for some of my ideas for menus & settings
09:25:14B4gderand scrolling in the splashes
09:25:16CassandraErm, btw, did anyone who knows something about the WPS check out that formatting patch?
09:25:35 Quit Musicmad ("Trillian (")
09:25:38B4gdernot to my knowledge
09:25:50B4gderbut the WPS using it looked nice
09:26:15Cassandra*nods* Wonder if it can do multiple formattings on the same line.
09:26:19CassandraI suspect not.
09:26:42amiconnB4gder: While at it, I could also try to add changing content without resetting the scroll pos, but this will get really tricky with proportional fonts
09:26:59B4gderbut some "flicker" must be accepted
09:28:03CassandraI'll leave it to you then ami. You know more about the display code than I do.
09:28:33B4gderamiconn is quickly becoming mr lcd ;-)
09:29:21CassandraI think he's got that one in the bag.
09:29:29amiconnCassandra: The wps code would still need fixing once scroll-text-change-without-reset is implemented actually use it
09:30:52CassandraI generally try to work on user friendliness isssues.
09:31:01 Quit ashridah ("Leaving")
09:32:44Coldtoastsorry to have been irritating B4gder. you're all doing such an awesome job on the whole rockbox project. and to be honest, rockboc is pretty much the only reason I didn't end up selling my player
09:32:57B4gderno worries Coldtoast
09:32:57 Part Coldtoast
09:33:15B4gderanother issue we should fix is to alter some paths/dirs
09:33:30B4gderto make it easier to upgrade without erasing your own customized stuff
09:34:14B4gdersince we use .rockbox for both stuff we deliver in releases and forces users to add custom thinhs in there
09:34:49B4gderbut I haven't figured out an approach I like for that yet
09:35:09CassandraI think two seperate dirs in the root is probably the best option.
09:35:20CassandraAlthough even that's clumsy.
09:35:20Bgerwhat about /.settings folter
09:35:33Bgeror other
09:35:38B4gderthe userdir should probably then be without dot
09:35:53Bgerfor example /.rockbox/settings
09:36:05Bgerand /.rockbox/settings/default <= this is overwritten
09:36:24B4gderthen plugins, codecs, fonts etc where would they go?
09:36:25CassandraBger: Don't like that. It's nice to be able to say to user "open up your jukebox and delete the foo folder"
09:36:56CassandraI think we should keep .rockbox for system data, and create a new dir for user stuff.
09:37:07B4gderI lean towards that too
09:37:07Cassandrarockbox-settings, perhaps?
09:37:15Bgeryou just say "open your /.rockbox/settings
09:37:42CassandraBger: You usually want to delete /.rockbox, but not the settings dir.
09:37:58Bgerand what about rockbox.iriver to be in the .rockboxfolder, i.e. /.rockbox/rockbox.iriver
09:38:26BgerCassandra you're right
09:38:33B4gderBger: that's a separate issue, but I could agree with that
09:39:11Cassandra*sighs* So much to fix.
09:39:55CassandraWhy does fixing one thing cause three more issues to raise their heads. It's like that scene from Fantasia.
09:41:07CassandraHmm. I think I favour "rockbox-config" over "rockbox-settings". WPS bitmaps aren't really settings as such.
09:41:39amiconnI *never* delete /.rockbox ...
09:42:00B4gderthen you have some old cruft in there by now ;-)
09:42:13CassandraI do it all the time. Clean out old plugins etc.
09:42:36CassandraAnd it often solves weird problems that users are having.
09:42:46amiconnB4gder: No, I usually delete old files from time to time
09:43:42CassandraIt's our OS directory. Definitely has bitrot problems from time to time.
09:45:23Bgeri agree with Cassandra
09:45:50 Nick Lynx_awy is now known as Lynx_ (
09:47:15B4gderbut... I guess that would force us to start checking for things in two dirs instead of the single one we do today
09:47:48CassandraScrew backwards compatibility. :)
09:47:57B4gderlike if you add a font to your config
09:48:41***Saving seen data "./dancer.seen"
09:48:54CassandraThat's not going to be a huge performance hit, surely?
09:49:14B4gderjust some added logic that isn't present today
09:50:23CassandraI think this whole thing falls in the "when we get around to it" category.
09:51:16CassandraI had a preliminary shot at converting the manual to LaTeX the other day.
09:51:35CassandraTeX choked on the output. I think that one's going to be interesting.
09:51:43CassandraIn the Chinese curse sense.
09:52:33CassandraBadger: You release manager these days?
09:53:05B4gderpossibly, but I think I can shake zagor into doing it
09:54:02CassandraBasically I'd like to have a week or two's notice (and feature freeze if possible) before release.
09:54:19B4gderyes, I like that too
09:54:24CassandraSo I can whack the manual on the head and release simultaniously with 2.5.
09:54:29B4gderand I think we should have it as soon as possible
09:54:57CassandraI agree. We're ready for a new release.
09:55:00B4gderamiconn: what's your thoughts on a 2.5 release soonish?
10:01:02B4gderI mailed the list to see what people think
10:01:42 Join LinusN [0] (
10:02:35B4gderhowdy grill master LinusN
10:03:46LinusNstill full
10:09:49 Quit Bger (Read error: 113 (No route to host))
10:10:33LinusNi have a bold suggestion
10:10:46LinusNrename ".rockbox" to "rockbox"
10:11:46LinusNmac os x owners have issues with the hidden directory
10:12:14Cassandra*nods* I think we should still have the WPS ignore it unless set to "all files"
10:12:17LinusNwe could even (as suggested before) move rockbox.iriver to the rockbox directory
10:12:28CassandraBut if we can do that, I see no problem.
10:13:04CassandraOh, speaking of which, Linus, can we have the latest version of the Windows installer checked into CVS please?
10:13:28LinusNaha, it isn't in cvs at all, is it?
10:13:41LinusNwhat files do we need?
10:13:50CassandraIf we're going to start messing about with paths, it needs to be.
10:14:20dwihnoYay! Vacation!
10:14:53LinusNCassandra: i have an "installer" dir on the buld server
10:15:13LinusNi assume the "bin" dir is pure innosetup stuff, right?
10:16:22LinusNthe installer scripts are in src
10:16:29LinusNls files
10:16:48CassandraYes. I *think* you just need /installer/and and /installer/src
10:16:52CassandraIt's been a while.
10:17:02LinusNbin and src
10:17:19LinusNdo we want the innosetup binaries in cvs?
10:17:38CassandraProbably not.
10:18:38LinusNCassandra: would it be enough if i put the installer/src in cvs?
10:19:21CassandraShould be. Move README in there as well though, since it's all the docs we have.
10:19:50CassandraI shouldn't have called that dir src really.
10:20:01LinusNwe can change that
10:20:31*HCl ponders screwing the stereo open in order to fix the turn on switch to always on so his annoying flatmate won't be able to turn it off all the time
10:20:40LinusNin fact, we need to change a lot to make the scripts usable anywhere else than on the build server
10:21:11LinusNHCl: wonderful idea, that's how you make real friends
10:21:49CassandraSuggestion. Check the src dir (+README) in as a new module "win-installer". We'll start from there.
10:22:03LinusNa long time ago, my brother was playing his stereo really loud, and my father cut the main power to the house, just to make a point
10:22:36HClLinusN: heh
10:22:40LinusNwhat he didn't know was that i had been hacking all night on my c64, without saving
10:22:53HClwell, its the stereo in the livingroom
10:22:54LinusNi was *this* close to strangling him
10:23:04HClwithout it turned on, we don't have sound on our tv
10:23:13HCland it actually costs more power to turn it off and on than to just keep it on
10:23:19HCland its very annoying when you want to watch tv
10:23:32LinusNno remote to the stereo?
10:23:39HClits just an amplifier
10:23:44 Join Bger [0] (~Bager@
10:23:49HClnothing fancy
10:23:49Cassandralinus: :(
10:24:06HCljust the crap that my flatmates didn't need anymore, hooked up to be a stereo set for the tv
10:24:31HCli'm already happy we got stereo now, heh, a while ago only one speaker worked
10:26:49LinusNCassandra: would you have time to work on making the installer scripts more generic?
10:28:43CassandraLinus: Probably.
10:29:05 Quit as (Read error: 104 (Connection reset by peer))
10:29:13 Join thegeek [0] (
10:29:46 Join [-AIR-] [0] (
10:29:52 Nick [-AIR-] is now known as west-acre (
10:30:58CassandraIf we're feature freezing, the manual is going to be my priority though.
10:40:24CassandraAnyone else having trouble building for archos?
10:41:13*B4gder tries
10:41:21B4gderworks fine here
10:41:59B4gderwhat happens?
10:42:51CassandraCC thread.c
10:42:53Cassandrathread.c:68: warning: `always_inline' attribute directive ignored
10:42:53Cassandrathread.c:69: warning: `always_inline' attribute directive ignored
10:43:12CassandraBuild firmware file
10:43:12Cassandraerror: max firmware size is 200KB!
10:43:24B4gderwhat gcc version is this?
10:43:46Cassandra[christi@slappy jbr2-build]$ sh-elf-gcc −−version
10:44:00CassandraWas working as of a couple of days ago.
10:46:47 Join bobTHC [0] (
10:46:58bobTHCmorning folks !!!
10:49:28CassandraIs gcc-3.4.4 likely to work with sh-elf?
10:51:19amiconnCassandra: Use 3.3.5 for SH1
10:52:09amiconnBoth 3.0.x and 3.4.x produce bigger binaries than 3.3.x on SH1
10:52:18CassandraOK. Will build that and try again.
10:52:26amiconn(and it seems 3.0.x doesn't know about always_inline)
10:52:37CassandraEvidently not.
10:53:10CassandraHaven't updated my sh build env since ... July last year.
10:53:19CassandraAlmost to the day.
10:54:35amiconn3.0.4 is stone-age
10:55:57CassandraMust've been what the wiki recommended at the time.
10:56:08CassandraOf course the wiki could've been out of date.
10:56:16CassandraOr maybe this was pre-wiki even.
10:57:54 Join Aison [0] (
10:58:25 Quit Bger (Read error: 110 (Connection timed out))
11:06:51 Join ashridah [0] (
11:13:28 Join markun [0] (
11:14:50*HCl bites who ever is filling up the entire wireless router
11:20:10amiconnHCl: You might be able to go without the trackchange event.
11:20:20HCli know
11:20:26HClby checking the elapsed field, right?
11:20:41amiconnJust let the playback code count the seconds played (it probably knows best) and store it in the id3 struct (if it's not already there)
11:21:02amiconnThe you can decide at the track unbuffer event whether to up th eplaycount
11:21:10LinusNi think the trackchange event should broadcast a SYS_TRACK_CHANGED message
11:21:11B4gderlunch time!
11:21:22amiconnLinusN: We won't need it
11:21:41HClyea, that would probably be bettr
11:21:49LinusNthat way a plugin can react on it and do something
11:21:51HClfor now, i want to play gta till i get the hovercraft :P
11:22:09LinusNfor example
11:22:10ashridahheh. i prefer the hydra
11:22:17amiconnThe playback code would the take care of skipping, pause etc with almost no additional effort
11:22:30ashridahdo a quick flyby over the army base, so other jets and helicopters chase you, then start dogfighting :)
11:23:37CassandraWell, that was far more effort than it ought to have been.
11:25:27ashridahfor some reason, missles don't seem to hit the news helicopter tho, it's annoying
11:32:26CassandraFly, my winged build monkeys! Fly!
11:37:14 Join Bger [0] (~Bager@
11:48:44***Saving seen data "./dancer.seen"
12:04:32 Nick QT_ is now known as QT (as@area51.users.madwifi)
12:13:12 Join cYmen [0] (
12:31:34 Join preglow [0] (
12:31:40amiconnBagder: wiki span, including changed user pages :(
12:32:59 Join Moos [0] (
12:33:08MoosHi all
12:33:53preglowwoot, wavpack encoding!
12:34:05Bgerwow ???
12:34:11Bgerreally ?
12:34:56preglowwell, yes, unless i'm hallucating
12:35:28Bgerand me too
12:37:02B4gderand the spam just poors in
12:38:42preglowcan't you block this idiot?
12:38:56B4gdernow I reject his IP
12:40:23B4gderseems to work for now at least
12:42:25 Quit preglow ("CGI:IRC (EOF)")
12:42:40markunmorning Moos
12:42:56B4gderoops, I threw out preglow ;-)
12:43:18 Join cYmen_ [0] (
12:44:15 Quit cYmen (Read error: 104 (Connection reset by peer))
12:44:33B4gder218.19.42.254 it is
12:48:12B4gderquite a speed in the log
12:48:51B4gderin June, the rockbox site server 50GB of data
12:53:43B4gderall spam erased now
13:25:42 Join courtc_ [0] (
13:25:47 Join Zagor [0] (
13:27:21B4gderg'day Z, added another deny in the .htaccess
13:29:13 Join [1]Chamois [0] (
13:37:06 Quit courtc (Read error: 110 (Connection timed out))
13:44:24 Join jlee [0] (
13:44:32 Nick jlee is now known as MrStaticVoid (
13:46:31 Quit Chamois (Read error: 110 (Connection timed out))
13:46:31 Nick [1]Chamois is now known as Chamois (
13:48:45***Saving seen data "./dancer.seen"
13:55:44 Join bipak_ [0] (
14:05:09ZagorB4gder: ok
14:09:06ZagorB4gder: i had already blocked him, or at least thought I did. using "deny from 218.19". apparently you can't block subnets that way.
14:10:22Zagorhe's using various IPs in the 218.19 and 218.20 nets
14:13:58 Quit Bger (Read error: 110 (Connection timed out))
14:14:36 Quit bipak (Read error: 110 (Connection timed out))
14:14:44amiconnZagor: In apache that would be deny from 218.19.* iirc
14:15:52Zagor says my syntax is the one to use, but apparently not...
14:16:12 Join Musicmad [0] (
14:16:19 Part Musicmad
14:17:00amiconnZagor: Hmm. I'd try then
14:18:12Zagoryeah i'll do some experimentation
14:23:25LinusNSlasheri: u there?
14:32:31 Join webguest40 [0] (
14:33:25 Join SatNav [0] (
14:33:41SatNavhi all
14:34:43SatNavI wonder if anyone could help me, i'm trying to compile the win32 iriver h120 simulator, and I'm getting a couple of errors
14:34:43 Quit webguest40 (Client Quit)
14:35:04LinusNSatNav: and the errors are?
14:36:07SatNavmake[1]: /dev/rockbox/tools/convbdf: Command not found
14:36:15SatNavis the first
14:36:20LinusNSatNav: cd tools
14:37:15LinusNquite unusual path i must say, /dev/rockbox :-)
14:37:27SatNavis it?
14:37:35 Quit MrStaticVoid ("leaving")
14:37:37SatNavit's in my cygwin directory
14:38:03LinusN /dev is the device file directory in unix/linux
14:38:14SatNavi see
14:38:29LinusNok, did it work better after running make in tools?
14:39:05SatNavI got a couple of different errors, but I'm starting from ../tools/configure again...
14:39:19LinusNdid you read this?
14:39:47SatNavyeah, Im looking at that page as we speak
14:39:53amiconnLinusN: Perhaps a rockbox device? ;)
14:40:14LinusN5. Build mentions the building of the tools
14:40:40LinusNamiconn: :-)
14:41:44SatNavok, the compiling got further along, but there are three brand new error messages now :)
14:41:48amiconnBtw, /dev is device file dir on cygwin as well, even if it doesn't exist physically. It's a virtual dir; stuff like /dev/sda1 works on cygwin
14:43:11 Join Maxime [0] (
14:43:23SatNavthey seem to be based around something called libmusepack
14:43:50SatNavsynth_filter.o Error 1 :/
14:44:50SatNavany ideas?
14:45:03B4gderwhat gcc version are you using?
14:45:11LinusNgcc −−version
14:45:29B4gderm68k-elf-gcc even
14:45:40B4gderoh it was sim?
14:45:42LinusNB4gder: he's building the sim
14:47:42amiconngcc −−version
14:48:06 Join Bger [0] (~Bager@
14:48:11SatNavit's 3.4.4
14:48:26SatNav(cygming special) ... :/
14:48:28B4gderand what is the error?
14:49:04 Quit markun ()
14:49:19SatNavmake[3]: *** [/dev/rockbox/build/apps/codecs/libmusepack/synth_filter.o] Error 1
14:49:29LinusNSatNav: no, before that
14:50:39SatNavtheres like a page full of warnings about IDATA_ATTR
14:50:55SatNavexcess elements in scalar initializer
14:51:05ashridahhave to rerun configure recently?
14:51:30SatNavyeah, like twice in the past 10 minutes
14:51:46amiconnSatNav: 3.4.4 cygming special is what I'm using as well. No errors here...
14:51:56ashridahah, nevermind then. i was getting a build issue until i ran that earlier
14:53:23SatNaverm.. synth_filter.c:346: error: 'Di_opt' undeclared
14:53:35amiconnDid you make clean?
14:54:04B4gderit seems like an include file issue
14:54:05amiconnI got some problems due to missing dependencies due to my IDATA changes and header renames
14:54:13amiconn...until I ran make clean
14:54:23SatNaver.. no. i downloaded todays bleeding edge over one from a few days ago. reckon its worth a clean start?
14:54:38SatNavoh, im not sure what that means
14:55:44amiconnIt may be that the dailiy/bleeding source packages are missing some files
14:56:25amiconn...lemme check
14:56:58SatNavamiconn: thanks
14:57:11B4gderit sounds like that, yes
14:57:18 Join markun [0] (
14:58:40markunFreeBSD's USB is so buggy.. doesn't take a thing to crash the whole OS :(
14:59:22B4gderthey have usb support? ;-P
14:59:41Bgerregarding conversation about h1x0 & h3x0 battery life last night: both series have *exactly* the same advertised runtime : 16h with 128kbps mp3, volume @ 20, no sound effects
15:00:19Bgerbut there are rumors that h1x0's battery life is better than h3x0
15:00:23Bger's one
15:00:37SatNavI would have expected that
15:01:47BgerSlasheri: some people say that they have runtime ABOVE 19h with stock firmware, 128kbit mp3, vol 20, no sound effects
15:02:00Bgerstandard battery, of course
15:02:24Bgerbut i don't remember whether this was ONE mp3 file, or directory with 128kbps files
15:04:22LinusNwhen i start playing an mp3 file, the first split second of the music is cut
15:08:53 Join webguest76 [0] (
15:09:29webguest76can anyone tell me what the battery dummy thing is on my ihxx
15:09:49Bgerwebguest76 just delete it
15:10:08webguest76it got to 30 MB
15:10:28Bgeryou have played with the plugin "Battery Test", haven't you ;)
15:10:49webguest76i was checking if it actually worked
15:10:58webguest76and was accurate
15:12:14Bgerthis plugin simulates hard disk access
15:12:20Bgerand writes to this file
15:12:45Bgerit's perfectly safe to delete it, even more - it's suggested
15:14:13SatNavok, I downloaded an older source tarball and compiled the sim, and it worked.
15:14:18SatNavYay! :)
15:14:36markunwhen I do a "umount /mnt/iriver" at the 'right' time it does nothing but when I unplug FreeBSD reboots :(
15:15:21SatNavso I tried to build the firmware, and it tells me that m68k-elf-gcc is not in the path
15:15:39SatNavdo i have to patch cywin and look for that package or something?
15:17:10ashridahyou need to compile your own version of gcc and binutils, telling them to produce cross compilers. cygwin won't provide that so far as i know
15:17:37SatNavoh bugger
15:17:46SatNavthat sounds like fun
15:17:49B4gderthis is mention in the docs
15:18:39HClmarkun: why don't you use linux?
15:18:58markunHCl: dunno
15:19:03 Join Sucka [0] (
15:19:05HClmaybe its time to switch? heh
15:19:15SatNavso that stuff wont come as part of cygwin? :(
15:19:27markunHCl: went from linux to freebsd a few years ago and sicked with it.
15:19:32markunsticked :)
15:19:39HCli switched between windows and linux tons of times
15:20:17markunI'm trying to get 24-bit output from tremor, but it's not working so well.
15:20:23HCl :/
15:20:36*HCl doesn't feel like looking at the database for now..
15:21:03Bger hm
15:23:33*amiconn found 22 files and 2 complete directories missing in the daily/bleeding source packages...
15:23:34 Quit webguest76 ("CGI:IRC (EOF)")
15:24:04SatNavwhoa. well good to know it wasn't entirely my bad :)
15:25:03amiconnWell, a number of them are irrelevant for building (stuff like README etc), but some are relevant
15:25:55amiconnE.g., a source package wouldn't allow building uart_boot because the Makefile is missing.... etc pp
15:26:30amiconnI'll try to fix that
15:26:40SatNavi see: cool
15:26:55amiconnBtw, the FILES mechanism is cumbersome. This tends to always lag behind
15:27:31amiconnBagder: Couldn't the source packages being built from CVS (using cvs export or something like that)?
15:27:33B4gderamiconn: you have a better idea on how to do it?
15:27:46B4gderyes it could
15:28:14amiconnThis way it would be automatically in sync with cvs
15:28:26B4gderit would include more files than today though
15:28:37B4gderlike firmware/test/
15:28:51amiconnYes, that's one of the dirs that are missing
15:29:00amiconnI don't see a problem including it
15:29:14B4gderno, not me either
15:29:16amiconnThe other is uisimulator/common/player_icons
15:30:28amiconnThe important files I found are e.g. apps/codecs/SOURCES, apps/codecs/lib/SOURCES ... gdb/setjmp.S ...
15:31:58SlasheriLinusN: Hmm, there is a small delay (HZ/8) before master channel is unmuted but this shouldn't be over 1s.. Maybe the first dma frame is lost
15:32:51amiconnSlasheri: I got skipping at track changes yesterday. No smooth gapless :(
15:33:26SlasheriAh, i think the low watermark without crossfade should be increased to the same point that crossfade watermark is currently
15:33:30 Quit SatNav ("Chatzilla [Firefox 1.0.4/20050511]")
15:34:03B4gderI got very weird non-functional skip-next today
15:34:17B4gderit just got silent for .5 secs and then it continued, on the same song
15:34:29B4gderseveral presses did the same thing
15:34:49SlasheriHmm, really weird
15:34:54Slasherican you reproduce that?
15:35:08B4gderI'll try more on my way home tonight
15:43:42 Quit Bger (Read error: 113 (No route to host))
15:48:46***Saving seen data "./dancer.seen"
15:51:37 Join Bger [0] (~Bager@
15:55:49LinusNSlasheri: what is this delay for?
15:58:10SlasheriLinusN: It's to prevent pops when starting playback
15:58:18SlasheriWithout the delay we have a loud pop..
15:58:20LinusNwhat are the pops coming from?
15:58:34Slasherii am not sure..
15:58:40Slasherimaybe a dc offset or random data
15:58:48B4gderSlasheri: you should add a comment explaining what the times are based on
15:59:04B4gderthere are numerous sleeps and yeilds now
15:59:06SlasheriB4gder: Hmm, good idea
15:59:20Slasheriand i think the HZ/8 is too long delay..
15:59:32LinusNwhat is it waiting for?
15:59:36SlasheriWe should try shorter delays (sleep(1) is too short)
16:00:00Slasheriit's waiting that the hardware pcm fifo has got some data
16:01:56LinusNbut that should be instantly after dma_start() shouldn't it?
16:02:53Slasheriyes it should.. i think we have to debug this more..
16:03:29LinusNalso, what's the sleep(1) in pcm_play_stop() doing?
16:03:41LinusNand why are we muting in the first place?
16:03:50LinusNquestions, questions...
16:03:58Slasherioh, that sleep is useless..
16:04:01Slasherijust feel free to remove it :)
16:04:19Slasheriwe have to mute to prevent the pops / dc offset
16:04:46zeso ~1.36 grams of sugar has as much energy as the 3.7v 1600mAh li-ion battery in my rio karma
16:05:20zethe 21 grams of sugar i put in my tea has around 91.6 watthours hehe
16:06:33LinusNSlasheri: somehow i feel that muting isn't the right solution...
16:09:27amiconnLinusN: There is a nice section in the UDA1380 datasheet: "8.12 Plop prevention"
16:09:58amiconnI think most of the playback code dealing with plop prevention is based on that...
16:10:36LinusNbut that isn't the kind of plops we have, is it?
16:10:48CassandraWonder if they knew that plop is a kids word for defecation.
16:11:09B4gderso plop prevention is diapers? ;-)
16:11:11LinusN8.12.1 Use diapers
16:11:30*B4gder runs home
16:11:32 Quit B4gder ("go go go")
16:12:03CassandraWell, we like to call them nappies in this part of the world. ;)
16:13:30 Quit Febs (Read error: 110 (Connection timed out))
16:16:04 Join Febs [0] (
16:19:07 Quit ashridah ("Leaving")
16:22:33HClwho has experience with the menu system?
16:22:49CassandraYup. Wassa problem?
16:22:51LinusNdoes it matter who?
16:23:04HClLinusN: yea, cause thats the person who'll know :)
16:23:15HClCassandra: i'm wanting to add a rating setting to the wps context menu
16:23:15LinusNjust ask the question
16:23:25HClbut i'm not sure how to go at it, in fact, i have no clue
16:23:36HCli looked at it a bit but it was with LANG_ stuff and i don't know anything about that either..
16:24:04LinusNHCl: a good opportunity to learn then
16:24:19LinusNHCl: add a LANG_ entry in apps/lang/english.lang
16:24:25LinusNlast in the file
16:24:36HCli'll look at it later
16:25:12LinusNonplay.c contains the wps context menu code
16:27:59CassandraHcl: I'm thinking perhaps we should temporarily remove visible rundb related stuff for 2.5.
16:28:34LinusNwe need a stable tag database implementation
16:29:05CassandraIt seems to me that the code is still maturing quite fast, and if we can leave it as "unofficial" that gives us more room to change things about before the next release.
16:30:27Bgerbye all
17:00:56 Quit Cassandra (Read error: 110 (Connection timed out))
17:05:29 Join Cassandra [0] (
17:12:38HCltag database is stable, i think.
17:12:48HClat least the current version doesn't need any changes..
17:13:00HClmake a snapshot of the database generators though
17:13:10HClsince the current one on the wiki is already incompatible with 2.4
17:16:35 Join kurzhaarrocker [0] (
17:18:50 Quit CBM-away (Read error: 110 (Connection timed out))
17:20:14kurzhaarrockerI want to build the song db but only get "Can't locate" from the perl script. Is that a bug or a feature? :)
17:21:28amiconnYou'll need a full perl installation
17:21:54kurzhaarrockerI installed the ActiveState thingie as described
17:22:03kurzhaarrockerIs there a trick I didn't see?
17:22:41*kurzhaarrocker finally located
17:23:57kurzhaarrockerDo I need a system variable to the lib path of perl? (c:\perl\bin is in the path already)
17:25:16kurzhaarrockerarg, Stupid me. There was another perl.exe in the local path
17:28:44 Quit kurzhaarrocker (Remote closed the connection)
17:29:06markunamiconn: Did you listen to the output of flite (festival lite) yet?
17:30:46 Quit Cassandra (Read error: 110 (Connection timed out))
17:32:16markunIn case you haven't, I made a small sample:
17:35:50 Join Cassandra [0] (
17:48:49***Saving seen data "./dancer.seen"
17:53:42 Join ghostiger [0] ([U2FsdGVkX@cf093ce9d4a6bfcf.session.tor)
17:57:45 Join TCK- [0] (
18:02:46CassandraBoomtown Rats had a few.
18:02:57CassandraHe also played the lead in the film of The Wall.
18:03:18LinusNfelt kind of off-topic :-)
18:12:43amiconnLinusN: Any news concerning SYS_SHUTDOWN?
18:13:14LinusNworks fairly ok, but i have some thing to debug first
18:13:51LinusNfor instance, you can prevent shutdown by not calling the default handler
18:14:31amiconnI can't say whether this would be a bug or a feature...
18:14:41LinusNa feature, i'd say
18:14:54 Quit einhirn (Read error: 131 (Connection reset by peer))
18:14:55Cassandralets call it a beacher.
18:15:19amiconnImo the shutdown code should then resend SYS_SHUTDOWN from time to time
18:15:52amiconn(if it is idle poweroff or sleep timer)
18:16:17LinusNi had that thought too
18:16:53 Quit TCK (Read error: 110 (Connection timed out))
18:17:28CassandraWhy would one get lost?
18:23:44LinusNwhen the button queue is cleared, or when a message is ignored for some reason
18:31:43 Quit TCK- (Read error: 110 (Connection timed out))
18:34:28 Quit Bger (Read error: 113 (No route to host))
18:47:40 Quit Chamois (" HydraIRC -> <- IRC for those that like to be different")
18:51:23LinusNtime to go home
18:51:26LinusNcu folx
18:51:28 Part LinusN
18:53:36 Join KRAK_JOE [0] (
18:54:22KRAK_JOEhelp anyone
18:55:22bobTHCnot bot here
18:56:02west-acreKRAK_JOE cretin
19:05:26 Quit bobTHC ("Smoke Weed Every Day !")
19:19:47 Nick Lynx_ is now known as Lynx_awy (
19:30:26 Join hardeep [0] (
19:33:32 Quit KRAK_JOE ()
19:37:12 Join einhirn [0] (
19:37:31 Quit einhirn (Client Quit)
19:45:23HClheh. that person made me think i was on my wrong client for a second
19:48:53***Saving seen data "./dancer.seen"
19:48:55 Join Naked6 [0] (
19:49:35Naked6anyone home?
19:49:40 Join Stryke` [0] (
19:49:40HCli wouldn't call it home
19:49:43HClbut there are people here
19:50:30Naked6I can't wait till rockbox is done for the iriver
19:50:41HCli don't think there is such a thing as done
19:50:51HClbut i'm guessing you mean you can't wait for the official rockbox iriver release
19:50:52Naked6well. running
19:50:57HClit already runs.
19:51:13Naked6but the codecs aparently need work
19:51:29HClits all still in development
19:51:48Naked6if I were to install it... is it possible to ever get it back to the normal firmware?
19:52:00Naked6very cool
19:52:02HClit doesn't remove the original firmware either
19:52:10Naked6thats incredible
19:52:13HCljust hold rec while booting or use the remote play to start and it'll boot the origina
19:52:35Naked6im so happy i bought an iriver over an ipod
19:52:46HCli never considered the ipod, heh.
19:53:00Naked6i was considering many things
19:53:00*amiconn just implemented some rather sophisticated preprocessing for constructing the bmp header for the screenshot feature
19:53:05HCli think iriver was the only 40gb player with radio at the time when i bought it
19:53:09Naked6but the iriver was the best imo
19:53:46HCli want radio for when i get bored of my own collection
19:54:04Naked6yeah, the radio is not super great though
19:54:12Hadakaprobably no hope of seeing rockbox on a small flash player that can do ogg in the near future?
19:54:15Naked6you have to have the remote hooked up to get good reception
19:54:17 Join RotAtoR [0] (
19:54:17HClmeh, it works good enough
19:54:39HClHadaka: depends on whether you can get people to put in effort into it.
19:54:46*HCl yawns
19:54:52HCli'm gonna go for a shower, bbl.
19:55:03Naked6im gonna roll out
19:55:06Naked6later fellas
19:55:14HadakaHCl: well, I was more like asking if iriver or something had already similar hardware and such so it wouldn't be a completely from scratch affair
19:55:15 Quit Naked6 ()
19:57:57 Join t0mas [0] (
19:58:06t0mas:D :D :D
19:58:28t0masback in the Netherlands
19:58:47t0masnot yet home... but still... it's good to have internet :)
19:58:57 Join rubberglove [0] (
19:58:59t0mas(I'm not IRC addicted in any way ;-))
20:00:18rubberglovehere's a question for anyone listening:
20:00:44rubberglovewhat happens to pug-ins when there is an idle timeout poweroff?
20:01:00rubbergloveahem ^plug-ins
20:01:34rubbergloveare they (the plug-ins) informed?
20:03:17CassandraI believe they get a SYS_POWEROFF event like the other running threads.
20:05:30HClHadaka: that would depend on the player...
20:06:27t0masI have an idea
20:06:33t0masscream if you don't like it...
20:06:56t0maswhy not power off after a few seconds on players with sw_poweroff if we have an unhandled exception?
20:07:19t0masit's rather irritating to search for a paperclip everytime it crashes...
20:07:22 Join TCK [0] (
20:08:04 Join merbanan [0] (
20:09:13 Join nexus3 [0] (
20:09:16 Part nexus3
20:15:00 Join Bger [0] (~Bager@
20:15:41HCli'd say half a minute
20:15:45HClnot a few seconds
20:15:52t0masas long as it powers off
20:16:02t0masas it's irritating if you don't have a paperclip...
20:18:06amiconnrubberglove: They don't get informed
20:18:22amiconnLinus is working on changing that
20:18:57amiconnt0mas: It would be better to reboot on a button press instead
20:19:16*Cassandra wonders if we could have a three finger salute somewhere in the button driver.
20:19:43amiconn(Play/Stop are the only possible buttons, since they can be checked without the button driver working
20:20:34amiconnOr we could mimic archos here: Holding Stop for 10 seconds to power off
20:21:02rubbergloveamiconn: ah. thanks. . . i'll stop searching through the code....for now.
20:21:28 Part rubberglove
20:22:41amiconn4-grey screendump feature is working...
20:25:12t0masamiconn: how do I wait for the stop button?
20:25:33amiconnYou have to check the port pin yourself (low-level)
20:25:48amiconnYou can't rely on the button driver in a crash situation
20:26:31t0masI know
20:26:39t0masthat's why I wanted to just wait 10 seconds or something
20:27:01amiconnThe 10 seconds are also a bit tricky, as you can't rely on the button tick either
20:27:13amiconns/button tick/timer tick/
20:27:18t0masthen just wait for stop and ~HOLD
20:27:54amiconnYou can use a delay loop, but that depends on the clock frequency (of course)
20:28:21t0masor do the stop button? and check for !hold?
20:29:08amiconnWe do such things in crash situations on archos, but not consistently
20:29:30amiconn...i.e. we do it in panicf(), but not in UIE()
20:30:06amiconnAnd btw, switching of isn't as good as rebooting
20:30:28amiconnThe hd might be unparked, and switching off forces it to do an emergency park
20:32:53t0masok, but does iriver support rebooting=
20:33:03t0masdidnīt find that in powermanagement code..
20:34:07amiconnRoLo might be impossible, as it requires the ata driver and file system access working
20:34:58t0masok, how do I reboot then?
20:35:53amiconnWe would need to jump into a certain address in the ROM
20:35:59amiconnLinus should know where
20:36:15t0masok, so a simple inline asm to jump?
20:36:33amiconnYes I think so
20:36:47t0masbut... how does the bootloader deal with already init-ed hardware?
20:37:20amiconnI think the bootloader does its own init it shouldn't matter
20:37:54amiconnOtherwise we would have to de-init this or that before rebooting
20:37:57t0masok, then we need Linus to tell us where to jump to
20:38:11t0masand we need some way to check the stop button... and hold switch
20:38:28t0masas just short pressing the stop button is ok with me... as long as it checks for the holdswitch
20:39:23amiconnWhy do you think checking Hold is necessary?
20:39:58t0maswell... you can accidentally press stop short when you get it out of your pocket
20:40:14t0masand then you won't see the error...
20:40:17amiconnThis check will happen only if crashed...
20:40:20dionoeais the rockboy plugin still under developement ? (great plugin btw :) )
20:40:21t0masand long press stop is difficult to check
20:40:26amiconnAh, that you mean
20:40:41t0masso combined short press hold button would be nice
20:40:57amiconnWe could check for combined Play and Stop press
20:41:03t0masas you normally set it to hold when it's in your pocket
20:41:14amiconnI don't use hold
20:41:24t0masthat would be ok too... but wouldn't stop !hold be more convinient?
20:41:36amiconnstop & play & !hold ;)
20:41:40t0mashuh?? why don't you use hold?
20:41:48t0masyes, that's ok too
20:41:56amiconnI don't need it, and it irritates me all the time
20:42:03*t0mas starts reading button.c for the ports to check Why the *** do the buttons not work???
20:42:20t0masthere is an icon in the screen... lol...
20:42:36t0masI have my iriver in my pocket a lot
20:42:45amiconnI don't
20:42:58amiconnImho it's too bulky for a pocket...
20:44:12t0mastry skater wear ;-)
20:44:26t0mashm... and it fits my swimming pants too...
20:44:28amiconnif ((GPIO1_READ & 0x62) == 0x60)
20:44:37amiconnThat's Play & Stop & !hold
20:44:49amiconnErm, wrong
20:44:58amiconnif ((GPIO1_READ & 0x62) == 0)
20:45:01t0masok, but is GPIO1_READ usable in the crashed state?
20:45:25amiconnIt's a plain i/o port, so it should always work
20:45:42 Join Lear [0] (
20:45:54t0masand GPIO1_READ is a macro?
20:46:38amiconnYes it is
20:46:47amiconnSee mcf5249.h
20:48:20t0masis it a problem to check the port in a while (1) loop?
20:48:34t0masshould I check it with some interval for example?
20:48:48t0masor can the cpu handle the continuous checking?
20:48:57amiconnNo, you can just use a simple while loop
20:49:17t0masok, then I just need to restart command from Linus...
20:49:24t0masor I need to find it out myself :)
20:49:33*t0mas starts looking around...
20:49:56amiconnWe can't do much to save energy, and we don't need to give time to other threads, so busy polling is ok here
20:50:37t0mashm... energy saving isn't an issue in the crashed state :)
20:51:10amiconnWe could try to disable the pll in order to run at 11 MHz
20:52:07t0masyes... that would be nice...
20:52:14amiconnIf the cpu happens to run at 120 MHz when the crash occurs, it will get quite warm...
20:52:21t0masI know
20:52:24t0masit did last week...
20:52:31t0masand I had no paperclip
20:52:41t0masso I had to wait for a few hours to get the battery empty...
20:53:08amiconnJust add the first 2 statements of the set_cpu_frequency() default: case before your polling loop
20:53:16amiconnDCR = (DCR & ~0x01ff) | 1;
20:53:23amiconnPLLCR = 0x00000000;
20:54:04amiconnThe first sets the correct DRAM refresh for 11 MHz, the second sets the pll to bypass (11 MHz)
20:54:23amiconnEverything else doesn't matter, resp. might cause nasty effects
20:54:28t0mashmh.. I knew the second 1
20:55:15amiconnBtw, someone had a good idea this weekend, but I don't remember who it was
20:55:25amiconnWe could set the cpu to 11 MHz while on USB
20:55:35t0masand when paused
20:55:49t0maslinus already talked about that some time... but never implemented it
20:55:49amiconnI wouldn't do this
20:55:58amiconnThe UI might get sluggish
20:56:18t0mashm... we talked about 22 Mhz as usefull for the non decoding, but still pcm filling state
20:56:36t0masSlasheri thought it was possible to feed pcm... at 22 mhz
20:56:47amiconnLinus wants to do some experiments how low we can go and still have a responsive UI
20:57:02t0masyes, that would be a good idea
20:57:03amiconnI would settle for 30..35 MHz
20:57:18amiconn...for the unboosted state
20:57:21t0masoh... and I was looking at the audio thread debug last week
20:57:26amiconnNo additional state
20:57:26t0masthe trackcount doesn't go down
20:57:42t0masbut maybe it's already fixed
20:57:59amiconnNo idea...
20:58:09amiconnI'm busy with greyscale gfx
20:58:47t0mashas Slasheri optimized playback a lot?
20:59:05t0masI'm looking at the audio thread debug... and it has a way lower boostrate now...
20:59:54Slasherit0mas: Hmm, 22 MHz might be possible
21:00:38t0masand the switching is way faster :)
21:00:38 Quit TCK (Read error: 104 (Connection reset by peer))
21:00:48 Join TCK [0] (
21:01:03t0maserm... 1 remark left :)
21:01:12t0maswhy doesn't it buffer > 10 songs?
21:01:25t0masI have a few mix cd's split up in 1 to 2 minute tracks...
21:01:29Leart0mas: dithering was disabled a while ago, maybe that's what you're seeing?
21:01:46t0masmight be... I had the version with new dsp code
21:01:50t0masthe first try :)
21:02:05LearI think it was disabled fairly soon after that change...
21:02:21t0masoh you mean prelows thing?
21:02:26t0masin the mp3 codec?
21:03:02amiconnThe 10 track limit is in the playback code; I already talked to Slasheri that this might be too low
21:03:24t0masyes, I think it should be more like 20 or something
21:03:34amiconnWe have a 16 track limit on archos, and some users even managed to hit that limit (with 2 MB RAM!)
21:03:37t0masdepending on how much more memory it costs
21:03:52amiconnI would go for 100 tracks on iriver
21:03:54t0mashm... how much extra mem does 1 more song cost?
21:04:00t0mas4 bytes? 1 pointer?
21:04:08amiconnNo, it's quite a bit more
21:04:15t0mashm... > 100 bytes?
21:04:28Lear4-500 bytes at least, I think...
21:04:40t0masthen 100 might be a but much...
21:04:41amiconnsizeof struct track_info per track
21:04:50*t0mas checks the struct
21:05:01LearWhere is the track info, btw? I don't remember... :)
21:07:14hardeepi'm looking into fixing the problem with resume info only being saved while in the wps. The mpeg/playback code seems like the ideal place for this, anyone disagree?
21:07:35t0masI agree hardeep
21:07:40t0masLear: I don't know
21:07:45 Quit Zagor ("Client exiting")
21:07:48t0masdamn... 9 days no rockbox...
21:08:02t0masand I start forgetting places :(
21:08:29amiconnt0mas: If I counted correctly it takes 1108 bytes per track
21:08:55LearThat much? I can see 600 on a quick look...
21:08:56t0masthen 100 takes a lot of mem..
21:09:13amiconnsizeof (struct track_info) = sizeof(struct mp3entry) + sizeof(struct mp3info) + 40;
21:09:35amiconnstruct mp3entry is the biggest sucker by far
21:09:39amiconn- 896 bytes
21:09:54LearDidn't do sizeof, but a second look made it more like 800. :)
21:10:07amiconnYes, but as I said users managed to hit 16 with 2 MB ram
21:10:13t0masamiconn: is it on stack now? as we can't just dump 100 * 1108 = 110.800 on the stack..
21:10:24LearCutting 100 bytes per entry should be easy though. MP3 TOC can be found in both mp3entry and mp3info...
21:10:28amiconnNo, it's a static array
21:10:39Leart0mas: not stack, static, so that's not a problem.
21:10:49amiconnWe have 16 times the ram, and 16*16 = 256...
21:11:05amiconnBut I'd limit to 100 because a CD can have a max of 99 tracks
21:11:06t0masok, maybe 50 is a good middle?
21:11:19LearBut increasing the MAX_TRACK for 32 MB platforms makes sense. I'd say 30 should be plenty. :)
21:11:31t0mas30 / 50 = 0.6 MB per file
21:11:36amiconn...and "albums" with such tiny tracks are usually rips from language learning CDs
21:12:05t0masamiconn: but we don't need to load every exceptional big album in ram fully...
21:12:30amiconnHmm, if it behaves correctly otherwise, you're right
21:12:35t0masand the next one... amiconn:
21:12:41amiconnIt might happen that it doesn't
21:12:46t0masit behaves perfectly with 10 now
21:12:52t0masand most of my albums are > 10
21:12:57amiconnOn archos we have the problem that the 16th track repeats over and over
21:13:33amiconn...until you skip once. Then all is fine again (until the next time with 16 tracks buffered
21:13:54t0masok... and my pastebin dump?
21:14:16t0mascan we jump to the address where the bootloader loads to?
21:14:40amiconnNo, this would be rockbox
21:14:51t0masyes, is that a problem?
21:15:08t0masoh wait... we want to re-init
21:15:09amiconnYes, rockbox might be partially overwritten
21:15:11t0masnevermind ;)
21:15:12Slasherihuh, over 16 tracks on 2 MB buffer? :D
21:15:18amiconn...because it behaved erratically
21:15:31t0masok, so jump to the beginning of the rom?
21:15:32amiconnSlasheri: Yes, as I said - language learning CDs
21:15:36amiconnt0mas: Yes
21:15:37Slasheriamiconn: ah, ok
21:15:55t0masamiconn: linus knows the addres... and with some searching... I might find it too :)
21:15:56amiconnThese CDs often have tracks with a few seconds length
21:15:56Slasheriamiconn: just change the track limit to something bigger if you feel that way :)
21:16:19t0masI would go to 50... not 100...
21:16:29t0masas it works fine when not everything is buffered
21:16:33amiconnt0mas: Apart from that, you copied from an old init sequence
21:16:40amiconnThe 0x30000000 is wrong
21:16:50amiconn(but works even though it is wrong)
21:17:50amiconnWe should try to correct the behavior on archos too (just don't use the whole buffer when the track number limit is reached)
21:18:03amiconnThen we could lower the maximum on archos...
21:18:16t0masamiconn: I copied it from viewcvs...
21:18:25t0masas I don't have my own files here...
21:18:33amiconnWhich file?
21:19:00 Join LinusN [0] (
21:19:53t0masah LinusN :)
21:20:10t0masamiconn: bootloader/main.c
21:20:19t0masLinusN: have you read the logs?
21:20:41t0masok, and do you have the answere?
21:20:52LinusNwhere to jump to restart?
21:21:20LinusNwithout running the boot loader?
21:21:26t0masno, with
21:21:53t0masI already found a way to jump to the loading address... but we need the bootloader to init some things
21:22:04t0masas we don't know the hardware state before we jump
21:22:08LinusNjust call system_reboot()
21:22:15t0masfrom an exception :)
21:22:22amiconnt0mas: Ah, I fixed rolo_restart(), but the routine in bootloader/main.c still uses the wrong RAM start address
21:22:22t0maserm... interrupt...
21:22:22LinusN#include "system.h"
21:22:22 Join ghode|afk [0] (
21:22:39t0masamiconn: ok
21:22:54ghode|afkquick question, i want to convert my wma files to ogg, what would the best settings be? and how do i mae sure that they have tag info?
21:23:02t0masLinusN: is it safe to do that from the unhandled interrupt handler?
21:23:24LinusNhmmm, wait
21:24:05LinusNyes, it should be safe
21:24:09amiconnLinusN: The wrong RAM start in bootloader/main.c should be fixed. However, I couldn't safely test...
21:24:36Learyou do know that converting between lossy formats is a bad idea as such? ;)
21:24:56t0masamiconn, LinusN: we rolo for the other devices if ON is pressed
21:25:02t0masshould I just add that to iriver too?
21:25:10Learother than that, it depends on the source. No point in using, say q6 for ogg for a 128 kbps wma... :)
21:25:29LinusNamiconn: yes it should be fixed, but it is safe for now
21:25:59t0masamiconn: you agree on just checking the ON button and the hold button? like archos devices do?
21:26:05t0masand rolo?
21:26:18ghode|afkLear, i know but rockbox won't support wma anytime soon, and i actually have some wma albums i want to listern to ;/ my wma's aew 192 btw]
21:26:23amiconnLinusN: We rolo from unhandled irq on archs?
21:26:31t0masamiconn: we do
21:26:35amiconnAfair we only do that for panicf()
21:26:44amiconn...and even then it doesn't work always
21:27:10t0massee UIE
21:27:44amiconnYes, I see it
21:27:46LinusNroloing is a bad idea imho
21:27:53amiconnLinusN: Yep
21:27:57t0mashmz... it does rolo load there... but we can also do system_reboot ?
21:28:03t0masah... ok, you agree :)
21:28:06amiconnWe should try to restart the ROM instead
21:28:07t0masslow webchat :X
21:28:08Learghode, then I'd use something that gives about the same bitrate, say q5.5 or 6.0.
21:28:12amiconn(also possible on archos)
21:28:24ghode|afkok, any idea on command line settings? jump to 0x00000000
21:28:32Learpossibly a little lower. Try!
21:28:35t0masso change both archos things to system_reboot() ?
21:28:44t0masand add the iriver to it?
21:29:00HClghode|afk: i'd just get a new rip of the albm
21:29:14LearFor command line, just use -q 5, or whatever you prefer. There are other options, but I wouldn't bother about them (except for the tagging ones that is).
21:29:24LinusNt0mas: each target has its own UIE()
21:29:33amiconnt0mas: Yes, imho that would be the best solution
21:29:49LinusNline 402
21:30:10t0maswithout linecounts :)
21:30:16*t0mas switches to annotate view
21:30:27ghode|afkHCl, i'd love to dot hat, but all my wma are from when a guy at my old work place bought in 20 albums, and at the time, the only thing i could use to rip them was WMP10 :(
21:31:00CassandraThen perhaps you should buy the album.
21:31:08amiconnt0mas: Choose the annotated view in viewcvs, the you'll have line numbers
21:31:29t0masyes amiconn I already did
21:31:42ghode|afkCassandra: send me the money then sure!
21:31:56t0masLinusN: I don't see why it;s different per device?
21:32:24t0masit's 1 function?
21:32:29t0masare there others for archos devices?
21:32:53HClghode|afk: can't you get them off the internet?
21:32:58LinusNUIE contains asm code
21:33:02HClwhat album do you need?
21:33:25LinusNso there's one UIE for each cpu
21:33:30CassandraGuys, can we please not discuss commiting illegal acts on this channel. There are many better places for it.
21:33:36t0masLinusN: ok, I missed that :)
21:33:43amiconnLinusN, t0mas: ...and some knowledge about the exception stack frame
21:33:47HClexcuse me, but making copies of albums you own is not illegal in my country.
21:33:57HClneither is downloading an mp3 version of that album
21:33:58HClin fact.
21:34:01HClits legal to download anything.
21:34:13LinusNnot in my country, since july 1st
21:34:27ghode|afkde[ends what country ;p, but i guess this isn't the place for such talk
21:34:41 Join as [0] (
21:34:56t0masLinusN: but for archos devices pressong ON causes it to rolo, shall I change it to ON and !hold for iriver, and call system_reboot() then?
21:35:07t0masamiconn suggested ON STOP and !hold
21:35:17t0mas* ON STOP
21:35:22LinusNstop is a bad idea
21:35:34t0mashm... the cgi chat doesn't display the plus sign here :|
21:35:36CassandraHCl: Good for you. Doesn't hold for others. However, in general, it's a good idea for the Rockbox project to have as cleae a llegal status as possible. It would be very annoying if people were to accuse Rockbox of encouraging piracy.
21:35:38t0masLinusN: shy?
21:35:49HClyea, i guess
21:35:57LinusNbecause stop is read by the button interrupt handler
21:36:05LinusNwhich might have crashed too
21:36:15amiconnLinusN: Hmm?
21:36:22t0masif ((GPIO1_READ & 0x62) == 0)
21:36:30amiconnOf course we have to check low-level, i.e. reading the port
21:36:30LinusNon and hold are the only buttons that can be read with port i/o
21:36:45t0masLinusN: stop can be read separate too afaik?
21:36:52LinusNnot on iriver
21:36:52amiconnOn (=Play), Off(=Stop=) and hold...
21:36:59amiconnLinusN: Yes it can
21:37:03*LinusN is silly
21:37:26amiconn...but On and !hold might be sufficient
21:37:33t0masmake it ON and STOP and !hold? or just ON and !hold?
21:37:37 Nick as is now known as thegeek_ (
21:37:44LinusNon + !hold
21:37:58t0masok, amiconn, can you do the bit magic again? :)
21:38:17t0masok, I'll read it myself :)
21:38:39LinusNgotta go
21:38:42 Part LinusN
21:38:43amiconnif ((GPIO1_READ & 0x22) == 0)
21:39:01amiconn...and the 62 was wrong btw
21:39:07*amiconn was silly too
21:39:16t0masamiconn: do you have some time to explain how you got that 0x22 ?
21:39:23amiconn...and Linus wasn't silly
21:39:32amiconn...we can't use Stop
21:39:47amiconnI mistook Play on the remote for Stop
21:40:26amiconnCheck button.c line 446 and line 580
21:40:54t0masBUTTON_ON 0x0001
21:41:53amiconnNot the symbolic button constant, the 'data' check is the interesting part
21:42:12amiconnYou can also check your iriver (debug -> view I/O ports)
21:42:37amiconnPress & release On, and switch Hold on & off. Watch GPIO1_READ...
21:42:45 Join bagawk [0] (1000@bagawk.user)
21:48:42 Quit ghode|afk ()
21:48:56***Saving seen data "./dancer.seen"
21:50:48 Quit thegeek (Read error: 110 (Connection timed out))
21:53:13 Quit t0mas ("CGI:IRC (EOF)")
21:54:09LearEr, why is get_mp3file_info called twice on the iriver?
22:00:45LearAnd the mpeg codec seems to need three additional fields, enc_delay, enc_padding and frame_count, thus mp3info is added to the trackinfo struct (with an extra copy of the toc)...
22:04:39 Join tvelocity [0] (
22:20:49HCli wish my gay flatmate would move, he's such a jerk
22:28:05 Join [IDC]Dragon [0] (
22:28:27 Join n0bby [0] (
22:28:45[IDC]Dragonamiconn: do you read?
22:28:55amiconnHello to Hanover
22:29:03[IDC]DragonHi Berlin
22:29:11LearSlasheri: Why a PCM watermark at 75% of buffer capacity? :)
22:29:17n0bbyhello from florida
22:29:43[IDC]Dragonamiconn: you could do the missing bootbox tests, with Player and OndioSP
22:29:49SlasheriLear: to make sure gapless works good :)
22:29:52[IDC]Dragonn0bby: nice
22:30:08Slasherii think that should not affect battery performance much
22:30:16LearAnd if you don't use gapless...? :)
22:30:30SlasheriLear: you can't "turn off" gapless =)
22:30:42Slasherionly crossfading is an option
22:30:49LearSorry, read that as crossfade first... :/
22:31:07[IDC]Dragonamiconn: shoudn't take long, with no charging test
22:31:16amiconn[IDC]Dragon: H00,
22:31:32amiconnI *hate* numlock on laptops
22:31:54n0bbyn4036c2 5s f4n 6n 3a-t-s! +-
22:32:03[IDC]DragonI hate both numlock and caps lock
22:32:10n0bby= numlock is fun on laptops! :P
22:32:23[IDC]Dragonespecially when caps lock is in fact shift lock
22:32:33LearStill, 2 seconds for gapless should be enough - unless the gap is intentional, in which case it should be kept, perhaps...
22:32:52LearOr 1 second... :)
22:32:57amiconn[IDC]Dragon: If caps lock == shift lock the you are using the wrong keymap
22:33:20*[IDC]Dragon is
22:33:48SlasheriLear: Hmm, yes.. maybe that watermark can be adjusted to little lower. But some people were reporting problems with the previous watermark setting
22:33:56amiconnI'm always changing the default keymap in windows to "Deutsch (IBM)" in the setup, and remove all others
22:34:14[IDC]Dragonthere's some registry stuff to even disable it
22:34:43 Join hubbel [0] (
22:34:45[IDC]Dragonand I want Ctrl for Fn
22:35:18amiconnMy kb has both Ctrl and Fn twice
22:35:53Slasherihuh, i am afraid if Fn even sends any key events to os when pressed..
22:36:16[IDC]Dragonremember my words, one day the shift bar will be as small as a regular key, with all the Windows, Apple and Linux special keys eating it up
22:36:34n0bbyi hate laptop spacebars
22:36:50n0bbytoo small
22:36:51LearBoost ratio increases a bit. Seems to hover at about 13% for 128 kbps.
22:37:01[IDC]Dragons/shift bar/space bar/
22:37:11SlasheriLear: Hmm, what it was before?
22:37:28Slasheribut that little increase is an expected change..
22:37:32LearIIRC, 8-10... Was a while ago though, so I'm not sure at what bitrate.
22:37:46Slasheriah, ok
22:38:13amiconnHmm, 4-grey core done, screendump adapted, x11 sim (temporarily) adapted. Win32 sim and bmp2rb still left...
22:38:31Slasherisombody who doesn't use the crossfade should replace that watermark level with a better value :)
22:39:38 Quit hubbel (Client Quit)
22:40:19LearCould be made dynamic without much problem, I'd guess, but should there really be that esoteric (sp?) options? ;)
22:41:17n0bbyami, does that mean 4 colour grey for iriver is now complete?
22:41:23n0bbyor is there more work do do?
22:41:44amiconnn0bby: " Win32 sim and bmp2rb still left..."
22:42:51n0bbyso we dont get our grey in real life until it works in the sim?
22:42:58n0bbywhat does bmp2rb do?
22:43:07n0bbyrockbox has its own bitmap format now?
22:43:23Bagderit always had
22:43:37n0bbybut the graphic wps doesnt use it
22:43:44Bagderyes it does, internally
22:44:02amiconnn0bby: I won't commit without the sims working, and I want a working bmp2rb to be able to have the boot logo in greyscale
22:44:45n0bbyis that a number of minutes, hours or days amount of work?
22:46:48 Quit [IDC]Dragon ("CGI:IRC")
22:47:50LearSlasheri: boost at 10-11% with smaller watermark, so the difference sure is small. :)
23:06:03 Join Moxon [0] (~Moxon@
23:09:20 Join LinusN [0] (
23:10:53*Moxon really likes the full seeking support in most codecs.
23:13:02 Quit Cassandra (" HydraIRC -> <- The professional IRC Client")
23:13:42 Join stripwax [0] (
23:18:28 Quit bagawk ("Leaving")
23:20:36*HCl guesses he'll go work on a rating thing..
23:27:13 Join stripwax_ [0] (
23:30:46HCli'm confused
23:30:51HClhow do i get into the wps context menu..
23:30:57 Quit stripwax (Read error: 110 (Connection timed out))
23:31:03n0bbyhold select
23:37:25stripwax_... so, how hard would it be to redirect the mas_codec_readreg stuff in vu_meter to some function that either does the mas read or reads some property of pcm_playback, depending on circumstances?
23:40:32 Quit Lear ("Chatzilla [Firefox 1.0.4/20050512]")
23:40:40HClyou'd probably wouldn't even need a function for it, just a define
23:40:54HClat least, thats what i'd say
23:41:38stripwax_yeah, a define would work, i guess, as long as it resolved to one of two different rb-> api functions (is that right?)
23:44:05Bagderyes, the plugin struct would simply have pointers to either the archos function or the pcm function
23:44:10 Part LinusN
23:44:57*HCl makes an attempt at getting a rating setting in and has high hopes
23:45:00hardeepthe peakmeter uses the same functions doesn't it?
23:45:12BagderI assume it does
23:45:23stripwax_that would be an extra indirection, no? couldn't we just #define get_left_vu rb->mas_codec_readreg(0xC) instead?
23:45:34HClwe'll probably want to move it to a song options submenu...
23:45:48stripwax_(and then rewrite that code to use get_left_vu instead of rb->mas_codec_readreg..)
23:46:06HClam i still allowed to commit this?
23:46:16HClwith the feature freeze that was suggested and all.
23:46:54Bagderthere's no freeze yet
23:48:58***Saving seen data "./dancer.seen"
23:49:37HClokay good
23:49:53HCli want to try to get these few last features in to give all our 2.5 only users a little extra
23:50:17HClthough i guess without an up to date search engine its not *that* advanced yet. but still usable
23:50:32Bagderpersonally, I want it for stats
23:50:37Bagdernot searches
23:50:48HClor you mean playcount
23:50:51HCland lastplayed
23:50:57HCli haven't implemented lastplayed yet..
23:51:07HClbut yea
23:51:25HClthats gonna be my next thing, trying to make it possible to display it on the wps
23:51:44Bagderthat'll be fun!
23:51:59HClyup :)
23:52:14HClthough i really should improve the playcount handling first
23:52:20HClbut i can do that afterwards
23:52:32HClit still counts songs that you skip
23:54:08 Join einhirn [0] (
23:55:09HClouch, thats not good :(
23:55:10 Quit TCK (Read error: 110 (Connection timed out))
23:55:12hardeephmmm, resume is off by several seconds on the iriver... looks like offset is being updated too early
23:55:12HClhits 0... why.. :/
23:55:12stripwax_hmm... what's the range of mas_codec_readreg(0xC) .. looks like 0..45 ?!
23:55:37pilljust a suggestion: would it be possible to add a "queue from" option in the playlist/queue context menu, so that when you choose a track, playback will go on to the next track after the one enqueued?
23:55:48pilli don't know how hard it would be to implement
23:56:20*HCl scratches his head...
23:56:40hardeeppill: probably not too hard, it could be implemented the same way as move to next folder
23:57:26pillyeah probably
23:57:35stripwax_ah nemmind. I'm a fool. found MAX_PEAK

Previous day | Next day