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

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

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

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

#rockbox log for 2009-01-21

00:03:04 Join solexx [0] (
00:08:47bertrikI've often wondered whether boost could be made automatic instead of explicitly triggered, something like "boost when # of ready threads > X for at least Y ticks"
00:09:17 Quit domonoky (Read error: 54 (Connection reset by peer))
00:13:57gartraldo "/" and "\" characters break ID3 info?
00:13:57 Quit parafin (Read error: 60 (Operation timed out))
00:13:57Zagorbertrik: I don't think that would work so well with our cooperative task switching
00:13:57 Join parafin [0] (
00:13:57 Quit saratoga ("CGI:IRC (Ping timeout)")
00:13:57bertrikcooperative tasks could still yield to indicate that they're not quite done yet and that would count as a ready thread I think
00:14:35Zagoryes, but then you replace explicit boosting with "speculative" yielding
00:16:14LloreanIt's also possibly lowering the average time spent boosted may not actually improve battery life.
00:16:33LloreanFor example, maybe buffering can be done unboosted sometimes, but we probably want to get it done as fast as humanly possible so we can spin down.
00:17:36gartralwhat about targets that have flash based disks, how is the "spindown" function handled?
00:19:20Zagorgartral: each storage type has its' own driver
00:20:06 Quit solexx_ (Read error: 110 (Connection timed out))
00:21:08 Quit petur ("Zzzzz")
00:22:23 Quit Zagor ("Clint excited")
00:25:52 Quit parafin (Read error: 60 (Operation timed out))
00:26:12 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
00:28:39 Quit bertrik ("Leaving")
00:29:10 Quit arohtar (Client Quit)
00:29:16 Quit jhMikeS (Read error: 131 (Connection reset by peer))
00:30:03 Join parafin [0] (
00:30:20 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
00:30:56 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
00:31:24saratogaZagor: you probably should not test this on MP3 due to multithreading
00:31:49saratogaI think boosting for decode frame will just cause mad_synth_frame to bottleneck on the other core, which is probably why the CPU load went up
00:32:13saratogathings are actually fairly well tuned right now for libmad, so its not representative
00:33:16saratogaWMA might be a better choice since its only a little slower but single threaded, you could try boosting around wma_decode_superframe_frame in wma.c
00:36:02 Quit jhMikeS (Nick collision from services.)
00:36:08 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
00:37:00 Quit gevaerts (Nick collision from services.)
00:37:11 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
00:37:34 Join itcheg [0] (i=62db4767@gateway/web/ajax/
00:43:48 Quit toffe82_ (Read error: 104 (Connection reset by peer))
00:47:18 Quit QuickStart (Remote closed the connection)
00:50:06 Quit ender` (" The three most dangerous things in the world are a programmer with a soldering iron, a hardware type with a program patch a")
00:51:25 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
00:52:18 Join toffe82 [0] (
01:00:04 Join fenugrec [0] (
01:09:47 Join marek_ [0] (
01:09:59 Join scheckler [0] (n=47cf5d0e@gateway/web/cgi-irc/
01:10:08 Join kugel_ [0] (
01:10:11marek_hi, i formatted my toshiba gigabeat F10, i have no GBSYSTEm folder now, what can i do?
01:11:08 Quit kugel (Read error: 60 (Operation timed out))
01:13:12toffe82marek_: check here
01:14:04 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
01:14:59 Quit scheckler ("CGI:IRC (Ping timeout)")
01:15:28 Join akur1 [0] (
01:17:15 Quit jhulst (Read error: 113 (No route to host))
01:18:24marek_toffe82 still "no system found"
01:18:53toffe82maryou are on windows ?
01:19:38 Quit faemir (Read error: 110 (Connection timed out))
01:19:52marek_toffe82 no
01:20:46toffe82you copy the gbsystem folder to your player ?
01:21:11marek_yup, arent theese files to small?
01:21:14marek_take a look
01:21:19marek_only a few bytes
01:21:21 Quit itcheg (" ajax IRC Client")
01:21:39toffe82you download the dummy gbsystem ?
01:22:28toffe82there are modified from the original
01:22:46marek_should i paste there also bootloader from rockbox?
01:22:59toffe82you can't run the originale firmware with them, you have to install rockbox too
01:23:30toffe82yes for the bootloader
01:23:37marek_well i dont give a sh!t about original firmware :) i love rockbox
01:23:39 Quit robin0800 (Remote closed the connection)
01:23:52toffe82if you want the full gbsystem, check the link I gave in the forum
01:25:15 Quit flydutch ("/* empty */")
01:27:18marek_ok im in :) thx toffe82
01:28:33 Quit bmbl (Client Quit)
01:29:30gartraltwiki wont save any pages i edit, i hit save, and it resets too the revision before the change, and bumps the revision number
01:32:35 Quit toffe82 ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
01:46:27 Quit jgarvey ("Leaving")
01:47:08 Join TiZ [0] (
01:48:47 Join robin0800 [0] (
01:49:28 Quit culture (Connection timed out)
01:49:34TiZHey, guys. I just used RBUtil to install the rockbox bootloader on Ubuntu Linux. I thought it might be a good idea to stop by and let you guys know that it needs to be done as root. You guys probably know that, but it doesn't seem to be on either the RockboxUtility or GraphicalInstall wiki pages. :)
01:50:33 Quit robin0800 (Remote closed the connection)
01:50:48 Join robin0800 [0] (
01:50:53TiZIt tripped me up at first, 'cos it wouldn't install the bootloader when I ran it as a normal user. When I did gksudo /path/to/rbutilqt, it worked fine. KDE users would use kdesudo instead, and those who don't have sudo would just login as root, I guess. I'm no linux guru, I'm just trying to share what I know. :)
01:50:58 Quit Thundercloud (Remote closed the connection)
01:51:07***Saving seen data "./dancer.seen"
01:51:43 Quit kugel_ (Remote closed the connection)
01:55:26TiZSo, that's really all I wanted to say. Congrats on 3.0 and 3.1 by the way!
01:55:29TiZTake care, guys.
01:55:33 Quit TiZ ("Jumping onto White Base! AAAHHHH!!")
01:59:45 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
02:03:10 Quit saratoga ("CGI:IRC (EOF)")
02:04:43 Quit akur1 (Read error: 110 (Connection timed out))
02:04:50 Join akur [0] (
02:06:41 Quit robin0800 (Read error: 104 (Connection reset by peer))
02:09:24 Quit efyx (Remote closed the connection)
02:13:19 Join break [0] (
02:15:05breakPermission for disc access denied!
02:15:12breakwhat can i do!
02:15:41breakthe internet is farrr too slow to browse any longer. been working on this since about 5:30pm
02:16:14cool_walking_more information? What gave you this message? What are you trying to do?
02:16:35breakthe rockbox utility, i'm trying to install rockbox on this refurb sansa E260
02:16:56kadobanbreak: do you have permission for disc access?
02:17:05breakhaha, of course i do
02:17:15 Quit nibbler (Read error: 110 (Connection timed out))
02:17:29breakoh yeah, vista 64
02:17:32 Quit DerDome ("Leaving.")
02:17:55 Join tajlero [0] (
02:18:08cool_walking_Is RbUtil running as administrator? I think someone said something about UAC as well.. maybe you have to disable it?
02:18:21breakUAC is disabled, and i am trying to run it as an admin
02:18:21 Join robin0800 [0] (
02:18:31kadoban"trying to"?
02:18:48breakconsidering it won't let me do anything...
02:19:12kadobanwell running it as an admin seems like a binary either are or aren't. perhaps you're leaving something out?
02:19:20cool_walking_Right-click RbUtil.exe and select "run as administrator".
02:19:36breaki've already done that
02:19:37 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
02:19:47 Quit marek_ (Remote closed the connection)
02:19:48breakabout 30 times maybe?
02:20:16breaki've tried running a compatibility mode for xp sp2, as an admin, and it does the same thing
02:20:36breakthe utility is detecting the sansa, in the right place
02:20:39breakit just doesn't work.
02:21:51gartralbreak: have you tested with a live cd os? mabey its the hardware...
02:23:16breakthen why did it work with my E280?
02:23:40LloreanAre you sure the e260 is a vanilla V1 and not an R?
02:24:15gartralhardware issue with the dap itself then? Sansa are known too go bad for no readily appearent reason, the PP hardware isnt exactly reliable...
02:24:34krazykitbreak, have you tried doing a manual install?
02:24:39Lloreanbreak: So it was named MSC and MTP, not Rhapsody and PlaysForSure in the options?
02:25:16pixelmaand is it in MSC or MTP mode? (Didn't see the question asked yet)
02:26:22breaki think MSC
02:26:27breakwhere it shows up as a drive, not a device
02:26:56Lloreanbreak: Did you actually check in the menus?
02:27:05breakits been a few hours
02:27:15breaki had to change it
02:28:15breakyes, MSC
02:28:17LloreanAnd which version of the utility are you using?
02:28:39breakfirmware 01.02.18a, rbutilqt-v1.0.9
02:29:00breaki bet i have to go back to an older firmware or something, don't i.
02:29:07breaksomething i haven't tried
02:29:25LloreanIf you're getting permission denied issues it's a host-side thing.
02:29:35LloreanThe device doesn't control your permissions.
02:29:40gartrali have had numerous problems with m1.0.9
02:30:50Lloreangartral: The problem with 1.0.9 relates to detecting Sansas though. His seems to be succeeding at that, at least.
02:31:24 Quit akur (Read error: 110 (Connection timed out))
02:31:25LloreanBut the obvious steps would be to try 1.0.7, then sansapatcher, then file a clear bug report documenting the issue.
02:31:38breakwhere can i grab that...
02:32:23 Join akur [0] (
02:32:24Llorean for the utility, the patcher should be in the manual if the utility doesn't work.
02:32:43LloreanYou will file a bug report, right?
02:32:57breakprobably not
02:33:03Lloreangee, thanks
02:33:12LloreanWe write free software for you, and you can't spare five minutes of your time to help make it better.
02:33:28breakthe page times out...
02:33:32breakis it up?
02:33:43gartralwell, it would be very helpful if you did, then we could follow this and make shure it dosnt happen too others
02:33:46LloreanThe page works fine for me.
02:33:49breakmy internet is being brutally raped
02:33:55gartral <−− thats a direct link
02:34:25gartral(i know im not supposed too deep link, but hes having probs)
02:34:50breakthis inauguration stuff is pretty annoying
02:35:13krazykitbreak, this is not a social channel. please take non-rockbox related comments elsewhere
02:36:11breakwow, mention one thing
02:36:33Lloreanbreak: The policy isn't "first one's free." We tell you when it happens so you know in the future.
02:36:39LloreanYou probably want to read the channel guidelines in the topic.
02:36:45breaki mentioned ONE THING
02:36:54breakand you get all over my ass about the rules
02:37:00LloreanNo, you were just told about them.
02:37:11gartralbreak, its a logged channel, wadeing through the non-dev talk gets annoying when were trying to trace what each other is saying :)
02:37:11LloreanNow you're making an ass of yourself when you could've just said "Oh, I'll keep that in mind."
02:37:28breaknow that makes sense, thanks gartral.
02:37:41 Nick fxb is now known as fxb__ (
02:38:17breaki'll probably be back whenever my internet works, so I can file a bug report.
02:38:25gartral(i know, but as an example) amiconn: how many months did it take for me too learn that one?
02:38:45Lloreangartral: _that_ is off topic too...
02:39:28gartrallike i said, as an example
02:39:54*scorche wonders why an example was needed
02:40:05 Part break
02:47:28 Join krazykit` [0] (
02:47:28 Quit krazykit (Read error: 54 (Connection reset by peer))
02:47:29 Quit fenugrec ("Leaving")
02:51:44 Quit Soap_ (Read error: 113 (No route to host))
02:52:49 Quit PaulJam (".")
02:54:08 Join Soap_ [0] (
02:56:45 Quit robin0800 (Read error: 60 (Operation timed out))
02:58:29 Quit XavierGr (Nick collision from services.)
02:58:42 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
03:01:20 Join QuickStart [0] (
03:01:40gartralanyway, USB full speed works fairly well on my e250 with the latest patchs from FS #8663 i get the inserted byte error once in a blue-moon with it, and it doesnt ever break anything that isnt easily replaced, especially with few files at a time
03:06:48Unhelpfulamiconn: i'm not sure that's legal - it gives me "syntax error before unsigned" (where the temp is declared). i'll just make it a static inline function
03:07:49Lloreangartral: That patch hasn't been updated since August.
03:08:05gartralbut it works...
03:08:23LloreanYou still get errors... how is that "works"?
03:09:34gartraland all that hoo-haa about the albums that didnt play, were corrupted MP3s from WiMP10 crashing, im sorry for trying to blame the firmware
03:10:07gartralyea, but its like 12 bytes per gig i transfer..
03:10:24LloreanAnd if it were working it'd be 0 bytes per gig.
03:10:40gartralok, fine, _mostly_ working
03:11:25LloreanAll it does is partially cover up a problem. It may turn out to be entirely the wrong solution, so there's no sense applying it until there's a proper fix since USB is disabled anyway, and will be disabled until it doesn't corrupt data.
03:12:35 Join internet [0] (
03:12:44 Quit timc`` (Connection timed out)
03:16:52gartralmeh, my dap was just successfully updated with the patched rb, if found if you keep transfer under 30 megs, they work without fail... its just if you exceed the buffer size it corrupts some files, and i just relised i didnt build a patched version, just stock source, and that seems too work the same...
03:20:55gartralheh, and it was plugged through a hub,which isnt even supposed too work, i have 0 explanation there
03:21:15LloreanHubs should work in many cases.
03:21:18LloreanIt's just some hubs that are problematic.
03:25:01 Join fenugrec [0] (
03:26:02 Nick krazykit` is now known as krazykit (
03:28:13 Quit Soap_ (Read error: 113 (No route to host))
03:30:19 Quit akur (Read error: 104 (Connection reset by peer))
03:36:38Unhelpfulamiconn: also, after i wrote div255 as an inline function, gcc reduces it to two ADDs, and a shift in the MOV, without repeating any of the math that went into the input value
03:38:35 Quit cool_walking_ (" ajax IRC Client")
03:38:55 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
03:44:24 Join timc`` [0] (n=aoeu@
03:51:11***Saving seen data "./dancer.seen"
03:53:39 Join akur [0] (
03:58:09 Quit tvelocity (Remote closed the connection)
04:01:57 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
04:07:58 Quit QuickStart (Remote closed the connection)
04:08:28 Join itcheg [0] (i=62db4767@gateway/web/ajax/
04:12:41 Join Barahir_ [0] (
04:18:54 Join blkhawk- [0] (
04:26:11 Quit itcheg (" ajax IRC Client")
04:27:54 Quit Barahir (Read error: 110 (Connection timed out))
04:31:33 Join Soap_ [0] (
04:35:20 Quit blkhawk (Read error: 110 (Connection timed out))
04:35:53 Nick blkhawk- is now known as blkhawk (
04:36:55 Join AndyIL [0] (i=AndyI@
04:43:27 Quit miepchen^schlaf (Read error: 101 (Network is unreachable))
04:44:20 Join yhuang [0] (
04:46:38 Join macocoy [0] (n=macoy@
04:48:07 Quit AndyI (Read error: 110 (Connection timed out))
04:54:45 Quit akur (Read error: 110 (Connection timed out))
04:54:57 Quit XavierGr ()
04:55:02 Quit fenugrec ("Leaving")
04:55:45 Join akur [0] (
04:56:18 Join itcheg [0] (i=62db4767@gateway/web/ajax/
05:01:13 Quit macocoy (Read error: 104 (Connection reset by peer))
05:08:01gartraltwiki says i no longer have edit rights too my own page >.>
05:13:44 Quit akur (Read error: 104 (Connection reset by peer))
05:25:07 Join rocko_ [0] (
05:25:47 Quit timc`` (Success)
05:26:29 Quit itcheg (" ajax IRC Client")
05:26:34 Join timc`` [0] (n=aoeu@
05:31:23 Part gartral
05:32:44 Quit Aurix_Lexico (Read error: 110 (Connection timed out))
05:34:16 Quit internet ("Leaving")
05:38:20 Quit Horscht ("Verlassend")
05:39:28 Join akur [0] (
05:50:57 Quit rocko_ ("Leaving")
05:51:13 Join rocko_ [0] (
05:51:15***Saving seen data "./dancer.seen"
05:51:59 Quit rocko_ (Read error: 54 (Connection reset by peer))
06:04:58 Quit tmzt (Read error: 60 (Operation timed out))
06:05:02 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
06:05:24 Join advcomp2019 [0] (
06:07:01 Join tmzt [0] (
06:08:38 Join daurnimator [0] (n=quae@unaffiliated/daurnimator)
06:23:15 Quit tajlero ("Leaving")
06:44:38 Join nuonguy [0] (
06:53:35 Join __lifeless [0] (n=lifeless@
06:54:11 Join Darksair [0] (n=user@
06:58:56 Join tajlero [0] (
07:07:56 Quit tajlero ("Leaving")
07:10:41 Quit _lifeless (Read error: 110 (Connection timed out))
07:17:03 Quit __lifeless (Remote closed the connection)
07:19:29 Quit jhulst (Remote closed the connection)
07:19:34 Join nibbler [0] (
07:34:52 Join _lifeless [0] (n=lifeless@
07:48:37 Quit nibbler (Read error: 113 (No route to host))
07:51:20***Saving seen data "./dancer.seen"
08:06:25 Join ender` [0] (
08:07:09 Join tajlero [0] (
08:07:30 Join akur1 [0] (
08:07:55 Quit akur (Read error: 110 (Connection timed out))
08:09:19 Quit tajlero (Client Quit)
08:24:54 Quit kachna|lappy (Read error: 104 (Connection reset by peer))
08:25:09 Join kachna|lappy [0] (
08:31:32 Quit akur1 (Read error: 104 (Connection reset by peer))
08:37:28 Join akur [0] (
08:37:49 Join nibbler [0] (n=Nibbler@
08:38:04 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
08:41:10 Quit BigBambi_ (Read error: 113 (No route to host))
08:41:12 Quit cool_walking_ (" ajax IRC Client")
08:48:35 Join kugel [0] (n=chatzill@unaffiliated/kugel)
08:48:38 Quit AndyIL ()
08:49:10 Quit BHSPitMonkey ("Ex-Chat")
08:53:56kugelJdGordon: ping
08:55:37Zagorkugel: since the general opinion about the fade settings seem to be "leave them alone", we'll do just the cleanup. I'll take a closer look at your patch later today.
08:57:34kugelZagor: okay
08:58:07 Join tajlero [0] (
08:58:26kugelamiconn said he has plans to make sw fading configurable, then it would be consistent too
08:58:31 Join AndyI [0] (i=AndyI@
09:02:25JdGordonkugel: pong... but depends what you wanna talk about :)
09:02:36JdGordongoing afk for an hour or so..
09:04:11JdGordontoo slow... bbl
09:04:25kugelJdGordon: oh, I'm just wondering if you have an idea what to do with the dead parts of the screen when leaving the wps to a custom list
09:04:35kugelit's somewhat nasty what the current patch does
09:07:05 Join Bagderr [0] (n=daniel@gateway/web/cgi-irc/
09:08:21 Quit nuonguy ("This computer has gone to sleep")
09:12:34linuxstbkugel: Where is the latest version of that patch? The version on FS #8799 seems very out of date.
09:15:29kugellinuxstb: yea, it is
09:15:46 Join petur [50] (n=petur@rockbox/developer/petur)
09:15:47kugellinuxstb: but that's the version I'm referring to
09:17:46kugelbasically, the whole display needs to be cleared upon exiting screens which certainly don't use the list dimensions
09:17:58kugelbut that causes statusbar flickering
09:30:03 Join MrDuck [0] (
09:32:13 Quit kachna|lappy (Read error: 104 (Connection reset by peer))
09:32:45 Quit akur (Read error: 110 (Connection timed out))
09:32:47 Quit GodEater_ (K-lined)
09:34:05 Join akur [0] (
09:36:25 Join webguest93 [0] (n=86ceaf40@gateway/web/cgi-irc/
09:36:34 Part akur
09:36:53webguest93jhMikeS: There?
09:37:00 Nick webguest93 is now known as BigBambi (n=86ceaf40@gateway/web/cgi-irc/
09:37:53BigBambiHas anyone got any experience in USB charging in Windows with the beast?
09:39:19BigBambiIt's the first time I've tried it on Windows (at work), and when I plug holding menu to charge only, I get the Rockbox media player popup in Windows which I cancel, after which the beast claims not to be charging
09:43:02linuxstbBigBambi: I assume you have a current SVN build? i.e. after jhMikeS's recent commits?
09:43:24BigBambilinuxstb: yes, updated just prior to coming in
09:43:45BigBambiI've never tried it on windows before, so I don't know if used to or not
09:44:04BigBambi(just prior to coming in here that is :))
09:45:42BigBambiThe odd thing is that the debug menu claims charger: absent, yet rockbox is using my backlight on charge settings
09:46:43 Nick Bagderr is now known as B4gder (n=daniel@gateway/web/cgi-irc/
09:49:44 Join tvelocity [0] (
09:51:24***Saving seen data "./dancer.seen"
09:55:00 Join Thundercloud [0] (
09:56:08 Join GodEater_ [0] (i=c2cbc962@rockbox/staff/GodEater)
10:01:08 Quit BigBambi ("CGI:IRC (EOF)")
10:01:11 Join BigBambi_ [0] (i=86ceaf40@rockbox/staff/BigBambi)
10:01:44 Nick Barahir_ is now known as Barahir (
10:02:00 Nick BigBambi_ is now known as BigBambi (i=86ceaf40@rockbox/staff/BigBambi)
10:22:26 Join gartral [0] (n=Gartral@
10:24:48 Quit petur ("gone to customer...")
10:26:11gartralin the buffer threads view, is the PCM line supposed to jump past alloc and real?
10:28:55JdGordonkugel: well... if we get to the point where we can have a WPS sort of thing going in the rest f the screen, it will update automatically in the new screen
10:29:16JdGordonfor now.. I think we should NOT clear the whole screen in preperation for that
10:29:33JdGordongartral: the pcm line is a seperate buffer
10:29:45JdGordonits not really related to the 2nd two lines
10:33:00gartralok, thanks JdGordon, my next question is, how hard would you think it would be too get either the radio to display info in the "now playing" WPS, or add a WPS too the radio?
10:34:59JdGordondepends... do you know C and actually want to stick around to implement this?
10:35:33 Quit Thundercloud (Remote closed the connection)
10:35:40gartralwell, ive been lightly tweaking the USB stack, with some succsess
10:36:14*gevaerts understands from the backlog that gartral hasn't tweaked the USB stack in any way
10:36:28gartrali have it too the point it'll take say a 30 meg block of stuff and properly transferit...
10:36:39 Join BigBambi_ [0] (i=86ceaf37@rockbox/staff/BigBambi)
10:37:02gartralwell, ive tryed with an old patch, and it gave me an idea or two, and now i kind of have it working
10:37:16JdGordonit shouldnt be all that hard to hack the wps display to show radio info.. i did it 1 or 2 years ago... doing it nicely is a different story
10:37:20gevaertsthe USB stack does not corrupt data
10:37:37 Quit Seed ("cu, Andre")
10:40:02gartralmay i upload an updated fmr that has the right name for a stationj, as it switched last month
10:41:31gevaertsBigBambi_: charge only doesn't work well on the beast, because windows doesn't enumerate the device correctly (because of the missing driver), and the code plays nice and only uses 500mA when told that it's allowed to (which doesn't happen if enumeration isn't completed)
10:41:52gevaertsSo basically you're stuck because everything works as designed and according to spec
10:42:07BigBambi_so it is a Windows issue?
10:42:37gevaertsissue yes (I think it doesn't happen on linux, but I'm not sure). Bug, no.
10:42:50pixelmareading the logs too - on my PC I have no problems transferring a bigger amount of data at once (filled an 8GB microSD in chunks of 1 or two albums at once, so 60MB and more usually) with Rocbox USB on my c200. I do have a problem with the front USB ports, they work but resetting quite often so it gets too slow to be usable.
10:42:58BigBambi_It doesn't happen on linux (i.e. it charges properly)
10:43:40BigBambi_gevaerts: Can I (me being not too skilled) easily edit something to do it incorrectly and let me charge it?
10:43:52BigBambi_As I don't want to take the carger to and from work everyday
10:45:15Zagorgevaerts: I assume it only plays nice if it gets a response but is not enumarated?
10:45:38gevaertsBigBambi_: I think changing usb_allowed_current() in usb_core.c to always return 500 will work
10:46:13BigBambi_gevaerts: Thanks, I'll try that tonight (can't build here)
10:47:21kugelJdGordon: I think I found a nice solution. btw: the statusbar is poorly viewport'fied :(
10:47:43JdGordoni know :/
10:47:44BigBambi_gevaerts: Is there anything we can do in the long run - I know it is a windows problem, but I'm sure users won't see it that way
10:47:48JdGordonwhats your solution?
10:48:10gevaertsZagor: the new code distinguishes pure chargers from hosts by looking at bus resets.
10:48:15 Quit nibbler (Read error: 113 (No route to host))
10:48:26Zagorgevaerts: aha
10:48:55gevaertsBigBambi_: yes. Implementing HID. I disagree about it being a windows problem though. Windows behaves entirely according to spec here
10:49:24BigBambi_ah, OK
10:49:56kugelJdGordon: issuing a redraw if viewportmanager_set_custom_vp is called with true, but only if it wasn't active before
10:50:06BigBambi_I thought you meant it should say OK to 500 mA even though not enumerated (spot the person who knows nothing about USB)
10:50:11B4gderdarn! we missed our 8th year version history anniversary just few days ago
10:50:13kugelthis way, the redraw only happens if a fullscreen screen is left to a non-fullscreen one
10:50:19B4gder7th even
10:50:27B4gderr1 | (no author) | 2002-01-17
10:51:25JdGordonisnt r1 just making the repo?
10:52:00B4gderhm, true r4 with some actual contents came 2002-03-25
10:52:09 Join __lifeless [0] (n=lifeless@
10:52:09 Join PaulJam [0] (
10:52:22JdGordonso we have 2 months to prepare the celebrations!
10:52:33 Quit _lifeless (Read error: 113 (No route to host))
10:52:45JdGordonkugel: hmm... yeah that sounds like it should get fixed in svn
10:53:10kugelhm, I don't understand
10:53:27 Part gartral
10:53:59JdGordoni mean.. it should be changed in svn to do that... instead of waiting for your patch to be fixed
10:56:11kugelah ok
10:56:57*JdGordon tries to rememeber the other ui tweaks which need to be done
10:58:11JdGordoncan anyone think of any screens which the statusbar appears obviously before the rest of it?
11:04:08JdGordonkugel: hmm.. no that fix isnt so good... the WPS does a full screen clear so if the statusbar is enabled there is a small time when the bar isnt shown
11:04:25JdGordonnot many WPS' show the bar though so it might not be a big deal
11:06:08 Quit tajlero (Read error: 110 (Connection timed out))
11:09:32PaulJamJdGordon: i noticed in some situations when the disk needs to spin up, that the statusbar appears too early. when leaving the WPS to the filebrowser with dircaache disabled and when leaving the virtual keyboard (when saving a file).
11:09:59 Join robin0800 [0] (
11:10:50JdGordonits not that it shows up too early... just that the rest of the ui cant keep up
11:13:22JdGordonok, i've moved the first list redraw in the tree up a bit so it happens before the disk access so that should be fixed
11:13:59kugelJdGordon: no it doesnt
11:14:43kugeland redrawing the statusbar via viewportmanager_draw_statusbar is part of the redraw (without, yes, the statusbar disappears shortly)
11:15:04JdGordonno it doesnt?
11:15:28kugelwell, if it would the dead parts wouldn't be in around the list
11:16:32JdGordonPaulJam: hm.. nope.. better not do that fix, it might crash :p
11:16:55kugelJdGordon: it doesn't in svn
11:17:45kugelwhat I do now is a full redraw (including immediate statusbar redraw) if you go from fullscreen to custom
11:17:46JdGordonof course it doesnt... it always redraws in svn
11:17:54JdGordonim saying with that change
11:17:56 Quit linuxstb (Read error: 110 (Connection timed out))
11:18:11kugel"WPS does a full screen clear" is not true for svn
11:18:35 Quit kugel (Remote closed the connection)
11:25:09 Join kugel [0] (n=chatzill@unaffiliated/kugel)
11:31:33 Join culture [0] (
11:31:33kugelJdGordon: here's what I have now:
11:31:36kugelworks good already
11:34:05JdGordonkugel: gwps-common.c:370
11:34:51JdGordonthat set_custom(true) call is redundant
11:35:02kugelJdGordon: that's on entering
11:35:06kugelnot on leaving
11:35:19kugelleaving the wps is problematic
11:35:25JdGordonit must always be reenabled on exiting... it is always assumed to be on on entering
11:35:43kugelremove the call, and see what happens
11:35:51JdGordoni know what happens...
11:36:00JdGordonlook how its handled in the current wps code
11:36:04JdGordonthats how it shold be done
11:36:24JdGordonremember.. we want to make it a PITA for things to NOT use the custom vp
11:37:41JdGordonviewportmanager_set_statusbar should be replaced with _set_custom_vp (but with a better name)
11:37:42kugelJdGordon: the wps will not set custom_vp, no matter of "making it a PITA"
11:37:49kugelunset even
11:38:02kugelin needs to be set again on exiting anyway
11:38:43JdGordonlook at the case ACTION_WPS_CONTEXT: in gwps.c (line 260)
11:38:55JdGordonspecifically the 2 calls to viewportmanager_set_statusbar
11:39:22kugelI see them
11:39:42kugelthey should be removed, since onplay is doing do_menu, and do_menu should handle statusbars
11:40:15JdGordonits the responsability to the screen which disables them to reenable them
11:40:26 Quit kugel (Remote closed the connection)
11:40:37 Join kugel [0] (n=chatzill@unaffiliated/kugel)
11:41:14kugelJdGordon: do_menu needs to set the custom_vp, not only because of the wps
11:41:24kugelalso because of nasty things with theme changes and other screens
11:41:26JdGordonit must not set it...
11:41:40kugeland again, nothing will be redrawed if custom vp was already set
11:43:42*JdGordon figuring out how to best articulate his thoughts...
11:45:05kugelyou want screens which don't use the custom vp to mess with it, instead of having each screen specifying what he's going to use, right?
11:45:23JdGordonnot mess with it.. just disable it
11:45:32kugelI've did that, and it didn't work well
11:45:37kugelespecially on theme changes
11:45:51JdGordonthat is exactly how the current status bar handling works
11:46:09kugelthe wps can't do the redraw. the redraw can happen after the list data is in the framebuffer, not before
11:46:16JdGordonsee the problem is, there are a few things which need more work before this works properly
11:46:26JdGordonone of which is sending an event on theme change
11:46:36JdGordona simple fix but noone has done the work
11:47:39kugelwhat you planning just just makes it more complex imo
11:49:01kugeland statusbar handling is different, its content doesn't change between screens
11:49:05JdGordonto implement maybe.. but once its done it makes things very simple
11:49:14 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
11:49:33 Quit kugel (Remote closed the connection)
11:49:49 Join kugel [0] (n=chatzill@unaffiliated/kugel)
11:50:30kugelno, I don't think so
11:50:32 Quit robin0800 (Remote closed the connection)
11:50:42JdGordonyour not understaning what i want to do
11:51:26***Saving seen data "./dancer.seen"
11:51:42 Join nibbler [0] (
11:52:04JdGordonok, bassically what you have to do is add a event and call that every time the statusbar/customvp is toggled which redraws the background image
11:52:34JdGordonwhen all this is finished that will be removed and the WPS will be redrawn over the dead space automatically
11:55:27kugelwhat I did now is the easiest and most uncomplex way
11:55:34kugeland it's not unproper
11:55:41JdGordoni disagree
11:56:09kugelthe list code setting custom vp avoids a dozen of nasty problems at once
11:56:27kugeland doesn't introdce any
11:56:47JdGordonit creates 15 more
11:57:42kugelwhich? with what I have, only screens that do not want to use the custom vp need to unset it, every other screen doesn't
11:58:29JdGordonthat is not what you have
11:58:32JdGordonthat is what I want
11:58:50kugelthis is what I have.
11:58:58JdGordonthe very first hunk in the patch you pastebined is doing the exact oposite
11:59:10kugelif you go from a list to any screen, it'll use the custom vp
12:00:15kugelit's not finished, that one will get removed
12:02:23JdGordonanyway the list is absolutly the wrong place to be fiddling, it is redundant there... 1) if a parent is passed to the list it wont get changed and 2) it gets the custom/full viewport if no parent is passed which means if something disabled it before and didnt reenable it that needs fixing.. no the list
12:08:48kugelJdGordon: we basically want the same, just our (current) ideas of the exact implementation differs
12:09:05kugelwhat I showed you is by no means finished so don't take that too serious
12:11:43JdGordonthe viewportmanager_* functions are all that is needed for this custom list stuff... it just needs renaming and to add that background redraw..
12:11:56JdGordonand the code to load the string -> vp
12:12:02kugelthe problems is that the wps can't do the comlete redraw
12:12:22JdGordonwhat complete redraw?
12:12:40kugelclear display and update display
12:12:45 Quit gevaerts (Read error: 145 (Connection timed out))
12:13:22JdGordoncourse it can... right now the only thing outside of the custom vp is a background image which gets redone when the custom list is reneabled
12:14:46kugelwell, redrawing the backdrop only apparently doesn't work
12:14:47JdGordonkugel: I cant tempt you away from this with multifont can i? :p
12:14:56JdGordonthen you're doing it wrong
12:15:05kugelthat also happens without all that stuff, and still shows the dead parts
12:15:12 Join robin0800 [0] (
12:15:38JdGordon"that also happens without all that stuff" .. too many "that"s
12:16:09kugelthe first that is redrawing the backdrop, the latter is full redraw
12:16:43kugeli.e. if you leave the wps, the backdrop is redrawn, but there's dead parts on the screens anyway
12:17:00JdGordonof course... what are you expecting there?
12:17:20kugelyou seem to be expecting that redrawing the backdrop will get the dead parts away
12:17:42kugel" it just needs renaming and to add that background redraw.."
12:17:54JdGordonno, i was using the redraw to get rid of the old screen bits... i have no idea what you're expecting it to display in the dead parts
12:18:10kugelthe background of course
12:18:23*kugel is confused now
12:18:31JdGordonthat makes 3 of us
12:18:38kugellol :)
12:21:11kugelJdGordon: and no, you can't tempt me to do multifont for the next weeks. I have exams soon, and I cannot contentrate on a new project until they're done
12:21:52kugeland I'd really like to get something done, not always starting something new
12:25:31JdGordonok... so going back to the confusion... "dead space" I'm talking about is the part of the display which the previous screen used but the new one doesnt. I was under the impression that only the background image would be shown here (or its not dead space)...
12:25:53kugelyea, that's how it should be
12:26:22kugelbut how do you exactly mean "redrawing the background to get that dead parts away"?
12:26:33JdGordonthen I say again... rename the viewportmanager functions, redraw the background when its reneabled and its all done
12:29:06JdGordonand if that is the only change then there is almost no extra work needed because most things were fixed when that first came in
12:30:25kugelwhat do you mean with redrawing the background? Do you mean just setting the backdrop, or clear&update the display or what?
12:30:55kugelyes to?
12:31:13JdGordonwhatever it takes to have the backdrop image redrawn
12:31:33kugelbut that's what I already do
12:31:50JdGordonthen you're doing it wrong! :)
12:32:40kugelJdGordon: see line 183+ in the paste
12:32:49kugelthat's what it takes to remove the dead parts
12:33:13kugeland that's done on toggeling the custom_vp (only if needed ofc)
12:33:48 Join Soap__ [0] (
12:34:02JdGordonwell.. you're not making sure its using the NULL viewport which could be problematic
12:34:22JdGordonand if thats coming out of the wps it needs to unload the wps backdrop and load the main one
12:34:53kugelJdGordon: clear_display() update() is using the null viewport
12:35:31JdGordonim not so sure about that
12:36:34JdGordonclear does...
12:37:30 Join bootflash [0] (n=50bb681b@gateway/web/cgi-irc/
12:37:42JdGordonthere is no lcd_update in the 16bit lcd driver so im assuming that is a #define to lcd_update_viewport?
12:38:24kugelJdGordon: there's display->update() and ->update_viewport()
12:38:35kugelupdate() is a wrapper to the update_viewport with the null viewport
12:40:09 Join Anges [0] (
12:40:30bootflashflashed H140 and load from ROM set, if I RoLo without changing the boot sequence, which version of the firmware will load upon a off and on cycle, flashed or RoLo ?
12:40:46kugelJdGordon: no wait, I confused it
12:41:02kugellcd_update is implemented in the target's lcd driver
12:41:32Zagorbootflash: flashed
12:41:33kugelas well as lcd_update_rect (and lcd_update_viewport == lcd_update_rect, and lcd_update is lcd_update
12:43:36bootflashsoory about irc lag, thanks Zagor
12:48:59 Quit bootflash ("CGI:IRC (Ping timeout)")
12:49:42 Quit Soap_ (Read error: 110 (Connection timed out))
12:53:59 Nick Soap__ is now known as soap_ (
12:57:57 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
13:01:34 Quit japc (Read error: 110 (Connection timed out))
13:26:50 Quit soap_ ("ChatZilla 0.9.84 [Firefox]")
13:37:16 Quit daurnimator ("Cyas later...")
13:51:30***Saving seen data "./dancer.seen"
13:52:44 Join kugel [0] (
13:53:46JdGordonkugel: see the patch i uploaded?
13:54:06kugelJdGordon: yea
13:54:26kugelessentially the same as I did, just with the event, instead of toggled custom_vp
13:55:02JdGordonright.. that was to show that most of the hard work needed has been done already
13:55:19JdGordonthe wps needs to be reworked so it doesnt try enableing the statusbar so often
13:55:51kugelI'm not exactly sure though if you got what I meant with wps doing the redraw/fullscreen update
13:56:12kugeldisplay->update sends the framebuffer to the lcd
13:56:33kugelif you do that in the wps (before list drawing), then the wps framebuffer data is sent to the lcd
13:57:49kugelthe wps can only do clear_display, but (I didn't test yet, but I fear it) I think the clear would be more noticeable
13:58:53kugelso, I did the list code do the redraw, so that it's instant and not noticeable
14:00:11JdGordon works great
14:00:45JdGordonI tihnk ill commit the changes to apps/gui/gwps-common.c to make things easier later
14:03:13kugelI'll check it out
14:03:46*kugel needs to answer emails first now :)
14:08:22 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
14:21:36*B4gder welcomes committer #68
14:22:51*B4gder points to kugel in case that wasn't clear ;-)
14:23:56kugelthanks :)
14:24:05jhMikeSZagor: Notice the PCM buffer doesn't seem to filly fully now on my e200 but sort of hangs around just above the boost point (75%).
14:24:39ZagorjhMikeS: with the boost patch?
14:25:10jhMikeSJust pain SVN. Wait, you didn't commit that? Somehow I got the impression...
14:25:11 Join massiveH [0] (
14:25:21ZagorjhMikeS: hehe, no.
14:25:23 Quit massiveH (Read error: 104 (Connection reset by peer))
14:25:59pixelmaSVN pain hurts ;)
14:26:15B4gderit makes you alert!
14:27:09jhMikeSWeird. 3g does it too. It's full until buffering stops then seems to only boost for a very short period.
14:28:18B4gderkugel: a good first commit ought to be yourself added in docs/COMMITTERS!
14:28:33ZagorjhMikeS: we boost during the whole buffer fill
14:28:39kugelI'm preparing. I hope I'm not going to fail horribly :/
14:28:49 Join LambdaCalculus37 [0] (i=44a04303@rockbox/staff/LambdaCalculus37)
14:29:14B4gderkugel: we promise to publicly ridicule you at first chance! ;-)
14:29:24jhMikeSZagor: Yes, during the initial buffering. Something kills the codec thread boost too soon now it seems.
14:29:28LambdaCalculus37kugel: Just saw the news... congrats! :)
14:29:30B4gderI'm sure you'll manage
14:29:48*kugel tripple checks if he wrote his name correctly
14:30:40ZagorjhMikeS: which codec are you using? I didn't see that on c200 during my testing yesterday.
14:30:47kugelwhat's the codepage of that file?
14:30:53kugelnot utf-8?
14:31:00jhMikeSheh, I find I'm not even in the COMMITTERS doc :\
14:31:07jhMikeSZagor: MPA
14:31:18B4gderjhMikeS: you are!
14:31:37 Join kachna|lappy [0] (
14:31:37B4gderkugel: it might be iso8859-1 yes
14:31:45jhMikeSwait, stupid search with typo
14:31:50LambdaCalculus37jhMikeS: I see your name in COMMITTERS.
14:31:51 Quit MrDuck (Read error: 104 (Connection reset by peer))
14:32:15B4gderkugel: primarily because we use that output in the web site outputs and that still assumes iso and not utf8
14:32:31kugelah ok, seems I nearly failed
14:32:35*jhMikeS sees himself now that he just scanned it manually
14:32:47jhMikeSor optically, whichever you prefer
14:35:06kugelhm, password incorrect?
14:35:25B4gderlemme check
14:35:34jhMikeSZagor: I'm seeing the same with WMA on e200 as well
14:35:56kugelah now :)
14:36:22jhMikeSWAV seems fine though
14:36:40jhMikeSwell, it doesn't need boost so why wouldn't it be?
14:36:41*LambdaCalculus37 hands kugel a beer for his first commit
14:36:48ZagorjhMikeS: could your r19808 commit have jinxed something?
14:37:25kugelLambdaCalculus37: thanks! tastes delicious
14:37:43LambdaCalculus37kugel: It's a new brew... "Sweet taste of success." ;)
14:37:47jhMikeSZagor: Who knows but that's so far removed from any of that.
14:37:51*LambdaCalculus37 goes back to more manual work
14:38:19jhMikeSnp, to just go back a couple revisions and check
14:39:32Zagorplayback is a rather fragile house of cards :)
14:40:04LambdaCalculus37Zagor: Doesn't playback need to be reworked quite a bit?
14:40:24ZagorLambdaCalculus37: yes it does. it's on my list of things I want to do
14:40:30jhMikeSPerhaps, but that commit is buried in the kernal and no actual behavior should change to anything outside.
14:41:29ZagorjhMikeS: I know. it's just a stab in the dark.
14:41:45LambdaCalculus37Zagor: One of the reasons why I ask is because if I can ever get an Archos AV100 or JBMM before hell freezes over, I want to get a port started on that, and that may be a hybrid HWCODEC/SWCODEC target.
14:41:49jhMikeSI'll get you your answer though, sir :)
14:42:08ZagorLambdaCalculus37: ugh...
14:42:25ZagorI won't tell you where my AV300 is then...
14:43:07LambdaCalculus37An AV300, you say? ;)
14:43:24PaulJamjhMikeS, Zagor: i think this behavior has been there since a few weeks ago (i didn't mention it because i thought the change was intended)
14:43:29 Part B4gder
14:44:23jhMikeSPaulJam: Hrm. I doubt it was intended. It should top off properly upon boosting.
14:44:42ZagorPaulJam: wow. I haven't seen it at all, and I've spent quite a bit of the last few weeks staring at that screen
14:44:55*JdGordon makes one last ditch attempt at getting hwcodec testing on 9795 and goes to bed
14:44:57 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
14:45:07 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
14:45:08jhMikeSOtherwise it boost/unboosts alot. The boost ration for MPA went from about 1.2% to around 3.5% as a consequence.
14:45:23 Join Casainho [0] (
14:45:38jhMikeS...on e200, tone controls, watching buffering screen.
14:46:17Zagorahh, I never use tone controls. maybe that's the difference.
14:46:31Casainhohello :-) −− I would like to know if working on rockbox Player couldn't be a project for Google Summer of Code 2009 ?
14:47:09ZagorCasainho: it could, if you can define a distinct project in it. i.e. a defined start and end.
14:48:10CasainhoZagor: hmmm, I don't have energies for that, and unfortunaly I am bad at english language :-(
14:48:21CasainhoZagor: thanks ;-)
14:54:19 Quit culture (Read error: 110 (Connection timed out))
14:56:41 Quit Casainho ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
14:58:51 Join nplus [0] (n=nplus@243.131.Globcom.Net)
15:00:05kugelwas there any decision about the natural sorting thing?
15:00:16kugelthe patch basically waits for it
15:00:39LambdaCalculus37kugel: Not sure. Have the logs revealed anything?
15:01:20kugelI don't think so. I remember to discussions, 1 was very controversial about what to be default, the other not so much
15:01:53Zagorhow good is the numeric sort? can it screw up sorting in unexpected ways?
15:02:09kugelI think it's good
15:02:18 Quit itcheg (" ajax IRC Client")
15:02:23kugelI've made a little test_sort folder, looked good
15:02:53Zagoreven when the artist or album contains numbers?
15:03:01 Join MrDuck [0] (
15:03:03kugelalso it doesn't seem to noticeably slow down browsing
15:03:07 Join itcheg [0] (i=41d59de2@gateway/web/ajax/
15:03:39kugeldo you have some example? I have no artists with numbers in my library
15:04:22Zagor"Front 242 - 05:22:09:12 Off - Crushed.mp3"
15:04:49Zagoror rather "Front 242 - 05:22:09:12 Off - 9 Crushed.mp3"
15:07:51 Quit kachna|lappy (Read error: 145 (Connection timed out))
15:09:58Zagoror how about this one: "Paul van Dyk - 45 RPM - 11 - Pump this 45.mp3"
15:10:59PaulJamZagor, jhMikeS:i tried some older builds and it seems as if this has started with the watermark rework (r19743/4).
15:11:42kugelZagor: what problems are you expecting?
15:11:58ZagorPaulJam: good find. do you use the equalizer?
15:12:25kugelthe patch doesn't change alphabetical sorting. Only for leading numbers, and for files which have the same non-number characters before numbers
15:13:19 Join japc [0] (n=japc@
15:13:27Zagorkugel: so those albums I listed won't be sorted numerically? (if the filenames are formatted that way)
15:14:28kugelthose would
15:14:50kugelI think, I haven tested if it sorts properly if such a large number of leading non-numbers is there
15:15:28kugel"and for files which have the same non-number characters before numbers" i.e. a2 comes before a10
15:15:30Zagorthe crux is that these albums/files use both numbers and non-numbers in the leading string
15:16:31Zagorif the patch identifies identical non-number "prefixes", surely it could do it with numeric prefixes too?
15:16:55kugelnot sure what you mean with numerical prefixes
15:17:09kugel2a is sorted before 10a, so the sort would kick in
15:17:23ZagorI mean the leading part of each filename that is identical for every track
15:18:43kugelah, yea, it would sort after the first non-identical char
15:19:09Zagorexactly. so "room202 - 10" is sorted after "room202 - 2"
15:19:31kugelso "Front 242 - 05:22:09:12 Off - 2 xxx.mp3" comes before "Front 242 - 05:22:09:12 Off - 11 xxx.mp3"
15:19:48kugelbut again, I haven't tested it so indepthly, it's just supposed to do that
15:20:03kugelI'll give it a test right now
15:20:38 Quit film42 (Client Quit)
15:20:53PaulJamZagor: usually yes (only preamp), but for testing i used the default settings (except backlight).
15:21:19ZagorPaulJam: ok. which target(s) do you see this on?
15:21:29PaulJamon H300
15:23:07Zagorah, I see it. my patch added cancel_boost when pcm goes above the watermark.
15:24:24Zagormy commit, rather
15:25:13 Join tyfoo [0] (
15:26:48PaulJamwow, this was scary... my player suddenly decided to only output white noise instead of music. and only a reboot fixed it.
15:27:07*LambdaCalculus37 should fix the wording of the blind descriptions for the Archos JBR v1/FM/v2 in the manual
15:28:36LambdaCalculus37I wonder if I should also start working on the Clip manual as well.
15:28:50*LambdaCalculus37 adds that to his TODO list
15:29:45 Join petur [50] (n=petur@rockbox/developer/petur)
15:37:10ZagorI think "numeric" is a much better name than "natural". numeric indicates it does something special with numbers, which is precisely what it does.
15:37:45Zagorsmart, natural, human etc just indicates it does something abnormal and gives no clue as to what
15:40:13BigBambiyes, things such as "natural" depends on what is natural to you
15:40:24BigBambisame with e.g. "smart"
15:40:36kugelthe thing is, I don't see how we ever agree on a name
15:41:01dionoea"obama sort"?
15:41:02kugelso sticking to the de-facto seems resonable to me. It's known as natural.
15:41:11*dionoea goes back to lurking
15:41:12Zagorkugel: is it? where?
15:41:12rasherI really don't think numeric is terribly descriptive - it seems to suggests that numbers will be assigned or something..
15:41:28 Join culture [0] (
15:41:54rasherZagor: most high-level programming languages
15:41:58kugel ;)
15:42:12Zagorrasher: it's not completely descriptive, but it indicates what differentiates that algorithm from standard sort.
15:42:20kugelthe function is even called strnatcmp
15:43:09rasherZagor: People have no idea what the so-called "standard sort" does though
15:43:30 Quit tyfoo (Read error: 104 (Connection reset by peer))
15:43:42Zagorrasher: that doesn't change the fact that "numeric" contains information that "natural" does not
15:43:50rasherIt's a black box that doesn't do what you expect
15:44:14rasherMisinformation in my opinion
15:44:23kugelZagor: there's no proper description imo
15:44:36kugelevery name proposed seems flawed to me
15:45:10ZagorI see no indication of "natural" being the de-facto name
15:45:54rasherIn computer circles it is. I've no idea how you missed tip
15:46:26kugeland in no-computer circles it's just sorting
15:46:39Zagoris that why it's mentioned a whopping 8600 times on the net?
15:46:47kugelmaybe we name it alphabetical, and rename the current alphabetical to ascii?
15:47:20kugelZagor: google has 8.370.000 hits
15:47:38Zagorkugel: for natural and sorting in the same page, yes. not for "natural sorting"
15:47:41rasherI still think human is fine - a normal human will NOT sort "Foo 10" between 1 and 2
15:47:41dionoeaalphabetical doesn't imply understanding the words, just comparing successive letters. So that wouldn't work
15:47:42LloreanWe could just name the sorting methods "Exact" (or "Explicit") for the current one and "intelligent" for the new one.
15:47:45 Join Anges1 [0] (
15:48:17dionoeanumber-aware sorting
15:48:31rasherHow can you name something that sorts incorrectly "Exact"?!
15:48:47dionoeadepends how you define "sorting correctly"
15:48:47LloreanDefine incorrectly
15:48:50kugelLlorean: well, I proposed smart, it isn't too much different to "intelligent", is it? anyway, it seems rejected as a name
15:48:59LloreanIt sorts based on the exact string, rather than by parsing the stringe.
15:49:10rasherAgainst all logical rules.
15:49:22 Join gevaerts_away [0] (i=5292654d@rockbox/developer/gevaerts)
15:49:22LloreanNot *all* or it wouldn't exist as a sorting method at all...
15:49:24rasherInternal representation is irrelevant.
15:49:39*gevaerts_away likes "number-aware sorting"
15:49:46 Join DerDome [0] (
15:50:03*kugel likes smart and simple most
15:50:09kugelsimple = ascii
15:50:35rasherIt's a string for human consumption - not a just series of bytes
15:50:46rasherAnd should be sorted accordingly.
15:51:05Lloreanrasher: But how that string is formatted is unpredictable.
15:51:23LloreanIf it were predictable, we'd never have people complaining about it when our database's intelligent sorting gets things wrongs by mis-parsing a string.
15:51:34***Saving seen data "./dancer.seen"
15:52:08kugelwell, that's due to more or less dumb filename-to-tag guessing
15:52:53LloreanYou think it's possible to do it perfectly, every time?
15:53:11LloreanThen it will always sometimes be dumb.
15:53:24kugelI just wouldn't offer that "feauture"
15:53:29LloreanThe advantage of ascii is that it's dumb in a consistent way, whereas when you add intelligence sometimes you *can't* work around the dumbness.
15:54:10kugelthe natural/whatever is pretty predictable imo
15:54:29kugeljust by judging after human logic
15:54:50pixelmabut you can't work around easily in cases where it's not - in contrast to ascii sorting
15:54:57LloreanMy point, yes.
15:55:05kugelwhere it's not?
15:55:09Zagorwikipedia calls it numerical. /bin/sort calls it numeric. perldoc calls it numerical. php calls it numeric.
15:55:17LloreanFor example, "007" (as in, James Bond" is treated as just the number 7
15:55:34LloreanPutting it between 6 and 8 without the leading zeros looks very odd.
15:56:07LloreanSometimes numbers *aren't* just numbers.
15:56:28kugelI'd predict that
15:56:48pixelmamaybe, but how would you prevent that?
15:57:04kugelprevent what?
15:57:06dionoeaif(!strcmp("007",...)) :)
15:57:22LloreanIt being put there if you don't want it there because to you it's *not* just the number 7
15:57:31kugelwhere would you put it then?
15:57:38kugelat the top, like ascii would do?
15:57:45kugelthat doesn't make any sense either
15:57:50LloreanIt's Oh, Oh, 7
15:57:52LloreanNot "seven"
15:58:10kugelit's 007, 6,60,7
15:58:13LloreanIf it were just 7, I wouldn't prefix it with zeroes
15:58:19kugelthat's rigged, not logical in anyway
15:58:32gevaerts_awayPossibly, but it's a real case
15:58:42 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
15:58:47kugelanyway, people aren't thinking about how it's going to be sorted
15:58:55*gevaerts_away can live with 007, but 008 should make things crash
15:58:59kugelthey're just expecting it to be sorted like it was in any other OS
15:59:15Zagorgevaerts_away: is it, really? will 007 ever be the actual sort key?
15:59:17gevaerts_awayAre those consistent?
15:59:46gevaerts_awayZagor: I can well imagine album names starting with that
16:00:13Zagorgevaerts_away: true, I was thinking of track names only.
16:00:14*kugel thought we discussed on the name
16:00:18Lloreankugel: My point was merely that "natural" sorting is hard to work around if it sorts wrong, whereas ascii isn't.
16:00:23LloreanNatural sorting attempt to guess "meaning"
16:00:25kugelit seems we're back on "do we even want that" :/
16:00:37Zagorwe want it
16:00:41*Llorean isn't against it.
16:00:45Zagorand an option to turn it off :)
16:00:47kugelLlorean: ascii is only predictable for engineers
16:01:02Lloreankugel: Ascii is predictable for _anyone_ as long as they know it exists and how it works.
16:01:09LloreanPlease, don't assign stereotypes.
16:01:15LloreanI know artists who are happy with ascii sorting.
16:01:17kugeland only engineers do know how it works
16:01:21*gevaerts_away denies being an engineer
16:01:23LloreanLook, don't be dumb
16:01:56*pixelma is no (software) engineer either
16:01:56LloreanIt's not the way to handle this sort of discussion.
16:01:57 Join {phoenix} [0] (
16:01:59*GodEater_ is an analyst according to his business card
16:01:59 Quit Anges (Read error: 110 (Connection timed out))
16:02:23*jhMikeS has thought perhaps language files should include _how_ to do things (voicing and sorting) as well as basic strings.
16:03:20kugelI think the majority of people can't predict ascii sort, because they don't know what it is
16:03:37kugelof course, until they read in wikipedia about it or something
16:03:44Zagorkugel: it doesn't really matter how many can do either. both options are wanted, and that's that.
16:04:19LloreanMy original point was only in justification of the phrase "Exact" in terms of strings and consistent predictability.
16:04:24 Join jain [0] (n=jainbasi@
16:05:28kugelyea, exact will work
16:05:49kugelbut that doesn't tell us how to name the "natural" sorting :(
16:06:23Zagoryou can bet we will get support questions for "exact" anyway... "how much more exact than '9' and '11' can you get?"
16:06:25LloreanI think "Intelligent" is a better word than "Smart". "Smart" implies other ones are not-smart. "Intelligent" implies the other one doesn't have an 'intelligence' behind it. A subtle difference, but kinda accurate.
16:06:51*gevaerts_away still likes "number-aware sorting"
16:07:22LloreanI like "intelligent" because I guarantee over time someone's going to want to add some more "intelligence" to it, and sometimes it's going to be a good idea to do so.
16:07:50kugelI like intelligent too
16:07:56ZagorI don't
16:08:12Zagorit doesn't say anything about how it is intelligent
16:08:14LloreanThe old sort could also just be "simple" or "string"
16:08:29LloreanZagor: Do users need to know how it's intelligent, or just that it is going to do things differently?
16:08:35LloreanEspecially if we default to it being on?
16:09:02ZagorLlorean: I would say users want to know what the different options do, yes.
16:09:18 Join nibbler_ [0] (
16:09:31LloreanThere's a manual for it, that's always going to be better than what we name it.
16:09:52LloreanSomething like "numerical" just means "using numbers." It still doesn't tell people who it uses them, or why the previous one isn't numerical.
16:10:08ZagorI disagree. I consider options that cannot be understood without reading the manual to be a failure.
16:10:14 Quit nibbler_ (Client Quit)
16:10:42 Quit nibbler (Remote closed the connection)
16:10:49 Join evilnick [0] (i=0c140464@gateway/web/ajax/
16:10:54LloreanThat's bullshit, frankly. We have terms like "replaygain" and "crossfeed" many members will just have never heard of before.
16:10:55ZagorLlorean: I agree, but it says *more* than "smart", "intelligent" or anything else. it hints that the special thing has to do with numbers.
16:10:57LloreanSometimes you have to define things.
16:11:17 Quit petur (Read error: 110 (Connection timed out))
16:11:26 Join nibbler [0] (
16:11:36ZagorLlorean: it's not bullshit. it's failure. some things we can't make intuitive, but that does not mean we should stop trying.
16:11:58LloreanI don't think we should try to make everything intuitive.
16:12:13LloreanI think that if the behaviour is complex, we should encourage people to seek a description of it.
16:12:36Zagorthen let's call then 'a' and 'b'.
16:12:42LloreanIf we have a sort that strips identical prefixes, drops leading zeroes, then tries to interpret numbers, a description might be nice.
16:13:52LloreanCalling it numerical is going to confuse people who think "01, 02, 03... 10, 11, 12" is already numerical.
16:14:23Zagorcalling it "intelligent" is not going to help them any more
16:14:34 Nick Anges1 is now known as Anges (
16:15:10LloreanIt adds the implication that the other system _isn't_ intelligent.
16:15:34Zagor"numeric" indicates "this option changes sorting in some way regarding numbers". no other name (except gevaerts') gives even that little hint.
16:15:50LloreanIt indicates "this sorts based on numbers"
16:16:02Zagorand it does
16:16:14dionoeaitsorts based on numbers and strings
16:16:18dionoeanot numbers only
16:16:25dionoeahence the "number aware sort" :)
16:16:26Zagordionoea: nobody said "only"
16:16:34 Join petur [50] (n=petur@rockbox/developer/petur)
16:16:34dionoeait's implicit
16:16:45LloreanWe could also call it "alphabetic" then..
16:16:45Zagordionoea: no :-)
16:16:47LloreanEqually valid.
16:17:05dionoeaZagor: hehe :)
16:17:05ZagorLlorean: no, because nothing in regards to the alphabet changes with that algorithm
16:17:25LloreanZagor: Changes relative to what?
16:17:31LloreanI thought it was going to be the new default.
16:17:36Zagorthe other option
16:17:47LloreanYou can't name it relative to the non-default option...
16:17:59Zagorof course we can
16:18:55LloreanIf anything, you name the non-default relative to the default. Because nothing's "changed" if you leave it on the default.
16:19:40Zagorthe options should have names that indicate what the difference is between choosing one over the other.
16:19:55Zagorwhich one is default has no immediate effect on its' name
16:20:14LloreanIt does if you're not giving the other one a name that describes what changes too.
16:20:23Zagorthat I can agree with
16:20:42Zagorthe other should also be named carefully
16:21:12LloreanI just don't feel "numeric" is a particularly descriptive name. Telling "it has something to do with numbers" is also true of the other one.
16:21:35LloreanThe simple sorting changes how numbers are sorted relative to the complex sorting in either direction.
16:21:53LloreanSo I think we need a "how it's changed" word rather than a "what it changes" word.
16:22:25LloreanWhich my word doesn't deal with either, yes.
16:22:52LloreanBut as amiconn's pointed out in the other channel, the "numeric" sort will still have asciiisms in it.
16:23:08LloreanSince the non-numeric characters will still be sorted with accented letters out of place.
16:23:44LloreanSo there's a target for that further added intelligence. :-P
16:24:10Zagoruh, no thanks :-)
16:24:51gevaerts_awayYou *can't* sort accented characters properly in a music context IMHO
16:24:52LloreanMaybe we could just name it something dumb, but clear, like "009 == 9"
16:25:29Zagorgevaerts_away: just add a locale tag in each file ;-)
16:25:46dionoea\<0*\([0-9]\)+\>==\1 ?
16:25:48LloreanOr even call the category "numeric sorting" rather than sorting, and name them "1, 10, 2" and "1, 2, 10"
16:25:55*dionoea doubts that people would understand that though
16:25:58 Join webguest95 [0] (n=5848f8a8@gateway/web/cgi-irc/
16:26:20ZagorLlorean: there's a fresh idea!
16:26:29kugelLlorean: now that's descriptive
16:26:45dionoeaLlorean: add one with a leading 0 too in the example
16:26:53dionoealike 03
16:27:15gevaerts_awayZagor: sure. Now how do you compare two files with a different locale?
16:28:04Zagorgevaerts_away: I know
16:28:24Zagor(that it can't be done)
16:29:02kugeldionoea: don't or 007 ;)
16:29:34jhMikeSeh? can't be done? /me no like "can't be dones"
16:30:13ZagorjhMikeS: one locale dictates that å sorts before b, another that it sorts after z.
16:30:52*Llorean would think sorting should be a factor of the user's locale, not the file's locale.
16:31:00 Quit webguest95 ("CGI:IRC (Ping timeout)")
16:31:06*kugel wants sort tables in .lang files :)
16:31:22jhMikeSTwo files, two different locales. I'd expect it to sort according to _my_ locale really.
16:31:33Zagorwe can sort alright, but there would always be those claiming we did it wrong
16:31:47ZagorjhMikeS: not the authors locale? you'll get the tracks in wrong order...
16:32:14LloreanZagor: Only if they're expected to be sorted by something other than their track number...
16:32:17Zagornot so fun for audiobooks :-)
16:32:45 Quit jain (Read error: 54 (Connection reset by peer))
16:33:04jhMikeSI'm sure other sticky instances that requires psychic software and/or hardware can be deal with by the user
16:33:51LloreanLike prepending track numbers for non-numbered files. ;)
16:35:13jhMikeSI always number my own files when ripping and tagging.
16:36:17kugelso, the decision on the name is delayed?
16:36:51gevaerts_awayLlorean: the user's locale might not handle the filenames
16:37:27Lloreangevaerts_away: Then we ascii(ish) sort, since we already do that anyway. :-P
16:37:50*gevaerts_away wants to know how German locale is supposed to sort カ vs री
16:38:17kugelnot at all, it segfaults :P
16:38:19gevaerts_awayAnd there's another problem with that. Transliterations
16:39:05ZagorI think we've strayed away from the topic a bit here :)
16:40:10Zagorkugel: I say go with Llorean's last idea. if that causes an uproar we can always have a new discussion...
16:40:17Llorean"03, 10, 2" and "2, 03, 10" is probably the shortest way to illustrate the sorting differences.
16:40:26kugelyea, I like that too actually
16:40:47kugelsome times illustrative examples tell more than 1000 words
16:41:08*dionoea 's sure that you could come up with a coherent explanation in less than 1000 words
16:41:09jhMikeSI would find things messing with my carefully numbered filenames exceedingly annoying :)
16:41:58kugelZagor, Llorean: and it doesn't even need to be translated ;)
16:44:55rasherZagor: PHP does *not* call it numeric sort -
16:45:27dionoeaYou could add an option to change the way it's called :D
16:45:39kugelseems php is as split about as we are? :)
16:45:45rasherZagor: That's not the same thing.
16:46:18rasher.. I'm pretty sure, anyway.
16:47:09dionoeanumeric sort ignores leading zeros while natural sort doesn't according to what I see on the php website
16:47:44Zagordionoea: yeah, that's what I see too. which means our sort is numeric. :-)
16:47:46dionoeathe site mentions sorting stuff with floats in the string like: 1.002 and 1.1
16:47:57dionoeain which cas ignoring leading zeros is bad
16:48:13*Zagor stops arguing about this now
16:49:00rasher> POSIX sort(1) has the -n option to sort numbers, but this doesn't work if there is a non-numeric prefix.
16:49:08rasherAccording to "some site"
16:51:39Lloreanis ours recursive? Like, will 001bob001.txt, 1bob001.txt and 001bob1.txt all resolve to the same (1bob1.txt) when sorting?
16:51:59*rasher knows nothing about the patch
16:54:22 Quit Zagor ("Client exiting")
16:56:43 Join flydutch [0] (
16:57:50 Join TheSphinX^ [0] (
16:59:14 Quit evilnick (" ajax IRC Client")
17:02:42 Join evilnick [0] (i=0c140464@gateway/web/ajax/
17:04:32 Join jgarvey [0] (
17:15:29 Join fredddy [0] (
17:19:55 Quit fredddy ("Verlassend")
17:25:30 Join gregzx [0] (
17:25:51 Quit gregzx (Remote closed the connection)
17:26:16 Join _lifeless [0] (n=lifeless@
17:27:01 Quit __lifeless (Read error: 113 (No route to host))
17:28:49 Join gregzx [0] (
17:32:04 Quit nibbler (Read error: 113 (No route to host))
17:34:06 Quit jfc (Read error: 104 (Connection reset by peer))
17:36:35 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
17:37:19saratogafor what its worth, when I was in school, they taught us about "natural sorting" and I have never heard another name for it, at least in the US
17:38:15 Quit nplus (Remote closed the connection)
17:45:43LambdaCalculus37saratoga: That's the only name I know by as well.
17:46:01*LambdaCalculus37 adds an "it" between "know" and "by"
17:46:51 Join n1s [0] (n=nils@rockbox/developer/n1s)
17:47:23kugelhence I said it's resonable to go with the de-facto, even if it's not 100% accurate, but at least people will now what it means
17:47:40saratogait may be specific to US english
17:48:14 Join toffe82 [0] (
17:48:21kugelbut I like the "2 03 10" and "03 10 2" suggestion quite much
17:49:26rasherThat seems good.
17:50:38kugelmaybe it should be implemented as a 3-line selection list (so that these are above each other) to make the idea behind the naming more obious?
17:51:36***Saving seen data "./dancer.seen"
17:52:09 Join RockBoxNoob [0] (n=84eb2e40@gateway/web/cgi-irc/
17:52:32rasherHow about just: "Sort: 2 03 10" / "Sort: 03 10 2"
17:52:43rasherTo make it clear that it's a sort-example
17:52:57kugelsounds good too
17:53:05RockBoxNoobI have a question about dual booting an 2nd gen. iPod Mini...
17:54:32RockBoxNoobI couldn't seem to find the info that I needed (in the manual or the boards)
17:55:43kadobanRockBoxNoob: don't ask to ask, just ask what you need. it saves time :)
17:56:28RockBoxNoobI have it booted in the Apple firmware, how do I get back to rockbox now?
17:56:39LambdaCalculus37Press and hold MENU+SELECT until the mini reboots.
17:56:46LambdaCalculus37And the info is right here in the manual:
17:57:00RockBoxNoobI have tried that to no avail
17:57:17rasherHow long did you hold it? Try holding it longer...
17:57:34kadobanRockBoxNoob: i don't have that specific device, but on others it's somewhat difficult. if you press other buttons at the same time, or yeah don't hold it long enough, it won't work
17:57:42RockBoxNoob5-odd seconds. OK
17:57:55 Join jfc [0] (
17:57:58rasherPretty sure it's closer to 15-20
17:59:14*LambdaCalculus37 makes sure that the manual says to "press and hold Menu and Select"
17:59:34LambdaCalculus37Well, it says "pressing and holding", but you get the idea. :)
18:00:03RockBoxNoobTried it... rebooted, Apple appeard, the booted back into Apple firmware.
18:00:37LambdaCalculus37You didn't click the hold switch on during the bootup, did you?
18:01:02LambdaCalculus37Is it connected to a dock or to your computer?
18:01:30LambdaCalculus37Did you install the bootloader?
18:01:30 Quit RockBoxNoob ("CGI:IRC (EOF)")
18:01:33 Join RockBoxNoob [0] (n=84eb2e40@gateway/web/cgi-irc/
18:01:50RockBoxNoobIt is not connected in any way.
18:01:51*LambdaCalculus37 tries that from the top
18:01:58LambdaCalculus37Is this a brand new install?
18:02:09RockBoxNoobYes, new install
18:02:17LambdaCalculus37Did you install the bootloader?
18:02:57LambdaCalculus37Are you sure? You used iPodpatcher or Rockbox Utility and asked it to install the bootloader for the iPod?
18:03:01RockBoxNoobI noticed some talk about issues w/ 2nd Gen. Mini's on the boards
18:03:18*LambdaCalculus37 has one that works just fine, and he had no issues installing
18:03:34RockBoxNoobI used The utility
18:04:05LambdaCalculus37Try with iPodpatcher. What OS?
18:04:14kadobanis there some way that i can bring defines in from my plugin header files to the manual files (in order to use them in \opt{}s and \nopt{}s), or is that not reasonable at all?
18:04:40RockBoxNoobI used Mac OS X
18:05:39rasherkadoban: I believe you can use the HAVE_* defines in the manual, not sure what else is available - have a look around the manual. As for bringing in defines, I don't know how reasonable that sounds
18:05:55RockBoxNoobThanks so much! :-)
18:05:58 Quit petur ("*plop*")
18:06:15kadobanrasher: alright, thanks...i guess it's probably not possible
18:06:48rasherkadoban: Can't you use the already available defines?
18:07:21kadobanrasher: yes, but if i could import my own defines the logic would be much easier, and it wouldn't need to be updated for new targets
18:08:00rasherAh true. I think you'll have to duplicate the logic from your plugin, basically.
18:08:37 Join bertrik [0] (
18:08:53LambdaCalculus37RockBoxNoob: Run iPodpatcher from the Terminal, and make sure to run it as root (e.g. sudo ./ipodpatcher)
18:10:55RockBoxNoobThanks. I used sudo. Seems to have fixed it.
18:11:38n1skadoban: the list of automatically generated opts for the manuals is in the features.txt file
18:11:40RockBoxNoobI tired rbutilqt to attempt to fix it before. It stated taht it was already installed.
18:12:28RockBoxNoobThanks for all your of help and you great work here! Keep up the good work!
18:12:48kadobann1s: ah, thank you...i was just looking for that :)
18:13:23 Quit toffe82 ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
18:13:47 Quit jfc (Remote closed the connection)
18:15:16LambdaCalculus37RockBoxNoob: Enjoy! :)
18:15:27 Quit BigBambi (" ajax IRC Client")
18:15:30 Join toffe82 [0] (
18:16:55 Join jfc [0] (
18:16:58 Quit BigBambi_ (" ajax IRC Client")
18:18:20 Quit RockBoxNoob ("CGI:IRC")
18:22:49 Join nibbler [0] (
18:33:16kugelsaratoga, Llorean: I'm going to use for a few days, to see which possibly unexpected changes the patch introduces, and commit if it's fine. You may want to give it some testing as well
18:33:30 Join toffe82_ [0] (
18:36:45 Quit robin0800 (Read error: 110 (Connection timed out))
18:36:54 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
18:38:01 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:39:52 Quit toffe82 (Read error: 110 (Connection timed out))
18:41:38 Join basetwo [0] (
18:42:40 Part basetwo
18:46:00 Quit nibbler (Read error: 54 (Connection reset by peer))
18:46:17 Quit gevaerts_away (" ajax IRC Client")
18:46:53 Quit Horscht ("Verlassend")
18:48:55 Join nibbler [0] (
18:50:07 Join bimbel [0] (n=Miranda@unaffiliated/bmbl)
18:59:49 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:03:45 Quit saratoga ("CGI:IRC (Ping timeout)")
19:05:40 Quit bmbl (Read error: 113 (No route to host))
19:05:51 Join massiveH [0] (
19:06:47 Quit massiveH (Client Quit)
19:16:01 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
19:19:25 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
19:20:11 Join Horschti [0] (n=Horscht@xbmc/user/horscht)
19:20:17saratogadoes anyone actually have floor0 vorbis files?
19:20:33 Quit Horschti (Client Quit)
19:23:04 Join Strife89 [0] (n=michael@
19:27:25preglowi saw one once...
19:28:25n1si think i have 2 for testing somewhere
19:29:57saratogapreglow: I just pulled some of their IRAM
19:30:09saratogasince it seemed to have little impact on their decoding performance, at least on PP
19:30:47saratogaand eventually I'd like to remove support for them entirely on lowmem targets
19:30:58amiconnOn coldfire things are *way* different regarding performance and IRAM
19:31:01saratogasince they seem to require an extra half MB of memory
19:31:14amiconnAlso don't forget PP5002
19:31:30preglowyeah, but seriously, how many use floor0 files?
19:31:34preglowlike 0.00001%?
19:31:40saratogathey haven't been used in 6 years by Xiph
19:31:41preglowi think this is ok, we have to be realistic
19:31:48amiconnWell, either we should support them properly, or not at all
19:31:55preglowamiconn: agreed
19:32:06saratogawell as long as they're real time i'm not too concerned
19:32:09n1si think that bad performance for floor0 can be accepted but we should play them if at all possible
19:32:15preglowthen test and see
19:32:24preglowi wouldn't be completely opposed to removing floor0 support
19:32:26 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
19:32:33saratogai think it makes sense on a lot of targets
19:32:47saratogaan extra 500KB of buffer for hd targets would be quite nice
19:32:51preglowanother option is splitting up the vorbis codec to handle each file type separately
19:32:51preglowbut we might not wanna go that way
19:32:52amiconnn1s: Unless I missed something, making things slower than for standard vorbis files will easily make them non-realtime on PP5002
19:33:05n1swhat exactly do we gain from dropping support for them?
19:33:12preglowiram, it seems
19:33:15preglowmore speed for ordinary files
19:33:19n1samiconn: ah, didn't realize they were that slow
19:33:25saratogano we don't need IRAM to decode them, at least i doubt it
19:33:54preglowsome kind of init time iram_alloc() might be very practical
19:34:08preglowso a codec could decide it's iram use some time during init
19:34:23preglowthen we could alloc full iram for both floor0 and floor1
19:34:33saratogamostly i'd like to shirnk the codec buffer to 512KB, and that will mean dropping them
19:34:52preglowhow big is it now again?
19:34:56amiconnWhy do they need that much ram?
19:35:19saratogaamiconn: it allocates a ton of buffer space by design
19:35:31 Quit bimbel ("Woah!")
19:35:32saratogafor this reason it was replaced but not depreciated by xiph
19:35:40amiconnDoes it actually *use* all that ram?
19:36:03saratogai'd be surprised if tremor malloc buffers it doesn't use
19:36:08saratogabut i can't say for sure
19:36:47 Join j8048188 [0] (n=cd7c35fe@gateway/web/cgi-irc/
19:37:20amiconnIt probably won't allocate buffers it doesn't use at all, but (1) it might use less than what it allocates for each buffer and (2) it might not need all of them at the same time
19:37:40preglowyou'll just run into the typical vorbis problem
19:37:40 Quit j8048188 (Client Quit)
19:37:41saratogaanyway its premature to worry about now, since AAC needs to be fixed first, but once thats done supporting floor0 will be the only use for that extra buffer space
19:37:51preglowit wastes memoruy, but there is no hard limit
19:38:22 Quit Horscht (Read error: 113 (No route to host))
19:38:41saratogaand 500KB is an enormous amount of RAM to spend on a feature that few if any people ever need
19:38:44n1samiconn: before the codec and malloc buffers were merged, the file in FS #5796 crashed −− it tried to malloc more than half a meg and exhaused the malloc buffer
19:38:49amiconnWell, we *could* add proper malloc for codecs and plugins. It might make debugging harder though...
19:39:14saratogawhat would we use for the malloc buffer?
19:39:43amiconnThe rest of the respective buffer: codec buffer for codecs, plugin buffer for plugins
19:40:21saratogawe'd still be wasting an enormous amount of memory just to support a beta version of the vorbis encoder
19:40:22 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
19:40:26amiconnThat's like it is now afaik, just that our el cheapo malloc doesn't have free()
19:40:44saratogatremor actually has a tmp_malloc that can free IIRC
19:41:20amiconnsaratoga: The question is whether we need nearly as much as we have now. Does the file in FS #5796 malloc everything at once, or in small chunks, calling free() inbetween?
19:41:31saratogawhen i looked at mallocs in floor0, it seemed like there was 1 enormous malloc for several hundred KB
19:41:35amiconnsaratoga: There's a nice 'bestfit' malloc somewhere on haxx...
19:42:03saratogalet me double check what it did
19:42:41amiconnIs there any vorbis encoder that can produce floor0 files?
19:42:49n1sthe problem with tremor imho is that the code is very resistant to reading :/
19:43:36saratogano i'm wrong theres 3-4 totalling several hundred KB
19:44:23saratogan1s: that and the fact that Xiph puts no limits at all on how big each buffer can be
19:44:42amiconnHmm: "No known Vorbis encoder past's own beta 4 makes use of floor 0"
19:44:50amiconnDo we need to support beta encoders?
19:44:56saratogaso we can't easily make assumptions about buffer sizes
19:45:02saratogaamiconn: Xiph says we do!
19:45:03preglowit's not really a practical problem
19:45:06preglowno one has floor0 files
19:46:04 Quit Darksair ("People who are zhuangbility want to show their niubility but only reflect their shability.")
19:46:21n1samiconn: the vorbis standard claims that floor0 should be supported and the fact that we've gotten bug reports about it suggests that some files are out there but they are probably not that many
19:46:21amiconn"Floor 1 is also considerably less expensive to decode than floor 0" <== That means floor0 might not even be realtime now...
19:46:25preglowif we were to disable floor0 support today, i doubt we'd hear a single complaint for a long time
19:46:38saratogaI'd really just like to remove it and see if anyone complains, then maybe point them to the last stable release or maybe a define int he config files that reenables it, in case theres some really old vorbis user out there
19:46:40*domonoky would think it would be okey, to detect floor0 files, and refuse to play them.
19:46:44amiconn(on some targets)
19:46:57saratogapeople who tested beta versions of vorbis can probably compile anyway
19:47:01n1sI'm not against dropping support for floor0 as long as there is any gain from doing so
19:47:32kadobanis there a way to take a screenshot in the simulator? all of the external programs i'm trying aren't working well (they keep including the "resizing handle" on the bottom right)
19:47:40domonokyn1s: there is a gain. more free buffer
19:48:16 Quit Anges (Read error: 104 (Connection reset by peer))
19:48:27saratogai just think its absurd that we fret about 20KB for image resizing while wasting 20x as much on codecs
19:48:39n1sdomonoky: we can't shrink the codec buffer until aac is fixed
19:49:01saratogatrue but fixing AAC is possible
19:49:14saratogai'm hoping to get it done next week after I return from a trip
19:49:37amiconnsaratoga: Does vorbis free some of those blocks before allocating the next?
19:49:38domonokyn1s: yes, but when its fixed, we can do it. Also the low-mem targets dont have a codec-buffer this big. So they probably just crash on such a file at moment...
19:49:43 Quit japc (Read error: 60 (Operation timed out))
19:50:18saratogaamiconn: I can't figure out what its doing, after a fair bit of effort I couldn't even figure out where the floor0 code is called from
19:50:24n1sdomonoky: as i said, if there is any gain i'm not against it
19:50:29saratogai just printed the malloc pointer instead
19:51:00amiconnHmm. DEBUGF()ing in the sim should be possible. Just DEBUGF() the malloc() and free() calls
19:51:11saratogaamiconn: i'll do free right now
19:51:40***Saving seen data "./dancer.seen"
19:52:04amiconnEven if free() does nothing but print, it should still be possible to tell what tremor is actually doing, and then we can tell whether a better malloc implementation would help
19:53:07pixelmakadoban: F5
19:53:22*amiconn wonders whether other DAPs supporting vorbis also support floor0
19:54:07bertrikI could try a floor0 vorbis file on my sansa clip OF
19:54:08preglowsaratoga: everything is called via function pointers, afaik
19:54:48pixelmakadoban: you should get a screenshot bmp in the simdisk dir when pressing F5 (long answer)
19:55:06kadobanpixelma: ah thanks, in simdisk :)
19:56:50 Join miepchen^schlaf [0] (
19:57:04 Quit evilnick (" ajax IRC Client")
19:58:30saratogaamiconn: many do not
20:00:53bertrikamiconn, if you can provide me with a sample floor0 vorbis file, I will try it
20:01:09*amiconn doesn't have one
20:03:48saratogabertrik: theres one on the tracoker in FS #5796
20:06:59 Join Anges [0] (
20:08:37bertriksaratoga, the OF on my clip goes non-responsive when selecting that file for playback ... :/
20:10:45 Quit yhuang ("Leaving")
20:12:29saratogabertrik: make sense, that file needs roughly 50% of the total memory on the clip to decode
20:14:18 Quit BigBambi (Read error: 113 (No route to host))
20:14:30 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
20:16:33 Join Rob2222 [0] (
20:19:14saratogaamiconn: theres quite a few calls to free, I'll try and figure out a way to see how much memory they try to free
20:20:18 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:22:07 Quit Strife89 ("Huzzah!")
20:22:21 Join Strife89 [0] (n=michael@
20:29:59 Quit nibbler (Read error: 110 (Connection timed out))
20:31:04saratogastandard vorbis frees 98KB, floor 0 210KB
20:32:44 Join massiveH [0] (
20:33:08 Quit massiveH (Client Quit)
20:33:47saratogaso a proper malloc would save ~40% of the memory we could save by dropping floor0
20:34:07saratogabut if floor1 vorbis is the most memory hungry codec, it might be worth doing anyway
20:36:43*LambdaCalculus37 has no objections to dropping floor0 Vorbis support... he hasn't used that in ages, anyway
20:37:53saratogai suppose a quick and dirty free implementation wouldn't be too hard anyway
20:38:21saratogamallocs only happen during initialization, so performance is a nonissue, and we only need to worry about free
20:39:08 Join kachna|lappy [0] (
20:39:26 Quit MrDuck (Read error: 131 (Connection reset by peer))
20:41:51 Join QuickStart [0] (
20:42:38QuickStartquestion when creating a wps where do the x and y coordinates start from for sansa e200 series
20:42:57 Join faemir [0] (
20:43:02 Quit jbenj ()
20:43:07QuickStartsweet thanks
20:45:17dionoeaQuickStart: that's a general rule for most computer graphics in fact. (0,0) is always the top left corner
20:47:44zeunless its some vector apps, 3D apps, cad apps... then its often bottom left
20:47:46 Quit miepchen^schlaf ()
20:48:41QuickStartI just wanted to make sure
20:48:51 Quit toffe82_ (Read error: 104 (Connection reset by peer))
20:49:01QuickStartcause I am designing the wps in GIMP and wanted to make sure the same coordinates were sound
20:53:21 Quit saratoga ("CGI:IRC (EOF)")
20:54:09 Quit TheSphinX^ ("XChat@Linux")
20:55:40 Join Xerion_ [0] (
20:56:14 Join bimbel [0] (n=Miranda@unaffiliated/bmbl)
20:56:59 Join Rob2223 [0] (
20:59:30 Join miepchen^schlaf [0] (
21:02:20 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
21:03:01 Quit Rob2222 (Read error: 60 (Operation timed out))
21:03:56 Quit XavierGr ()
21:04:11 Join toffe82 [0] (
21:07:03 Quit Xerion (Connection timed out)
21:07:04 Nick Xerion_ is now known as Xerion (
21:07:06 Join nibbler [0] (
21:08:32 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
21:09:23 Join LucasPestana [0] (n=bd3b72f4@gateway/web/cgi-irc/
21:09:39LucasPestanahey thre?
21:09:52LucasPestanais comebody online?
21:10:19krazykitmany people are around. if you have a question, just ask
21:10:19kadobanLucasPestana: there is always someone here
21:10:51LucasPestanaI'd to like to ask for some help
21:11:47LucasPestanaI'm on a Win XP, and had just installed Rockbox on my player −− a iAUDIO X5
21:12:09 Quit bmbl (Read error: 113 (No route to host))
21:12:12LucasPestana[btw, a 30 GB iAUDIO, if that makes nay difference]
21:12:40LucasPestananow, I'd like to install some patches
21:13:20LambdaCalculus37LucasPestana: You have to use patches on the source code.
21:13:45LucasPestanaummm.... right... how do I do that?
21:14:19LucasPestanaI've got some .DIFF abd .PATCH files...
21:14:21kadobanLucasPestana: first you need to set up a development envronment and get that working, then look how to deal with patches
21:14:25LambdaCalculus37Read this page for information:
21:14:25 Quit Neovanglist (Read error: 104 (Connection reset by peer))
21:14:31 Join grobkorn [0] (n=584a9d22@gateway/web/cgi-irc/
21:14:33 Quit Xerion (Read error: 104 (Connection reset by peer))
21:14:37grobkornhi channel
21:14:47kadoban(by the way, .diff and .patch are just different names for the same thing, in this case)
21:15:02LambdaCalculus37grobkorn: This isn't a social channel.
21:15:06 Join Xerion [0] (
21:15:42grobkornneed help with rockbox - isnt that the right place for asking 1 question?
21:15:47LucasPestanaoh, ok, thanx
21:15:54LambdaCalculus37Ask your question.
21:16:23 Quit jhulst (Read error: 110 (Connection timed out))
21:16:38grobkornhave done that CFMOD... there comes "loading firmware... result: -1" but i cant find any help about that error-mesaage
21:16:45LucasPestanaokay... I thin I'm gonna bother you, anyway −− how do I set the development environment on Win XP?
21:17:07n1sgrobkorn: which player?
21:17:16grobkorniriver h320, sorry, i forgot
21:17:27kadobanLucasPestana: it's on the page you were linked to. if you have specific questions we can answer them, but in general people aren't interested in holding your hand through it when there's already a document explaining it
21:17:51grobkorngot that adapter mentioned on the rockbox wiki and a adata 32gb sppedy CF card
21:18:08n1sgrobkorn: can the Original Firmware read the card? and did you install a .rockbox dir on it (with all the files of course)
21:18:26LucasPestanaok, thanks
21:19:24grobkornorigianl firmware doesnt work (like explained in the wiki) ... i moved the to the player and extracted ther in the root (the .rockbox folder exists)
21:19:35grobkorni can acess the cf card by usb
21:20:20grobkorni had rockbox installed on my HDD too
21:20:27grobkornbootloader is installed
21:20:32grobkornbut hdd crashed
21:21:16PaulJamgrobkorn: have you installed the inofficial bootloader from the CFModGuide wiki page? the official one doesn't support CF cards
21:21:17 Quit Xerion (Read error: 60 (Operation timed out))
21:21:30mcuelenaeredo bootloaders have an audio buffer allocated?
21:21:45mcuelenaere(or access to audio_get_buffer() ?)
21:21:51n1sthey shouldn't and i don't think they do
21:22:01grobkornrockbox utility does not work with that hex file
21:22:48grobkornaborts and says: not a correct firmware
21:23:31PaulJamwell, to update the bootloader you have to use the original firmware. so you'll probably have to (temporarily) install a working harddrive in order to update the bootloader.
21:23:31n1sgrobkorn: you don't need to do anything with a hex file if you already have the rockbox bootloader installed, but how does the Original firmware fail?
21:24:04n1sor is there an unofficial one required for CF modded players?
21:24:21PaulJamn1s: the official bootloader doesn't support CF cards on the H300
21:24:37grobkornin the CFModGuide they say −−> OF does not work.... OF aborts booting with: CHECK HDD CONNECTION
21:24:46rashergrobkorn: Rockbox Utility will not install unofficial firmwares
21:25:16grobkornhow can i install unofficial firmware WIRTHOUT OF?
21:25:40n1sPaulJam: oh, i was under the impression CF cards had the same interface as the HD's an should "just work", what is needed to support them?
21:26:09 Join Xerion [0] (
21:26:15n1sgrobkorn: short answer, you can't
21:26:15 Quit LucasPestana ("CGI:IRC (EOF)")
21:27:00n1sit would be possible to flash from rockbox if you could start it and adapted the h100 flashing plugin but _very_ risky and not easy at all
21:27:17mcuelenaerecan anyone with a Rockbox USB stack-compatible device test what speed they get when the RAM disk is enabled? (dd if=/dev/sdX of=/dev/null count=1024)
21:27:45grobkornwaaa? then my player is broken and i cant MOD them??? :(
21:28:19n1sgrobkorn: apparently you need a working hd
21:28:34grobkornoh noooo
21:28:58PaulJamn1s: i have no idea what was needed to support them, but the commit was done after the last official bootloader was released, so the last one doesn't contain those changes.
21:29:21*n1s wonders if we will see a new official h300 bootloader released
21:29:55n1sPaulJam: ah, i somehow got the impression that the CF fixes were specific to pp based players
21:30:10PaulJamn1s: there is still the problem that a SVN bootloader crashes for some people.
21:30:22 Quit grobkorn ("CGI:IRC")
21:30:57 Join grobkorn [0] (n=584a9d22@gateway/web/cgi-irc/
21:31:18grobkornbut what does RESULT -1 mean
21:31:35 Quit QuickStart (Remote closed the connection)
21:31:40 Quit Horscht ("Verlassend")
21:32:06 Join Aurix_Lexico [0] (
21:32:11rashergrobkorn: that it can't load the rockbox binary I believe
21:33:21 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
21:34:47mcuelenaeregevaerts: why is ramdisk_buffer = tb.transfer_buffer + BUFFER_SIZE*2; ?
21:35:23mcuelenaereah never mind
21:36:45grobkornok - thank you guys for your help
21:36:55grobkorntry to get a hdd ;(
21:38:11 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:45:43 Quit grobkorn ("CGI:IRC")
21:46:19 Quit Strife89 ("Rebooting.")
21:47:05 Quit gevaerts (Nick collision from services.)
21:47:14 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
21:49:52 Quit XavierGr ()
21:50:01 Join Strife89 [0] (n=michael@
21:50:13 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
21:50:59 Part Anges
21:51:44***Saving seen data "./dancer.seen"
21:58:00 Join japc [0] (
21:58:15 Quit LambdaCalculus37 (" ajax IRC Client")
21:58:19 Quit Strife89 ("Focusing on some video.")
22:12:39 Join tyfoo [0] (
22:17:29 Join l403 [0] (n=l@
22:21:58l403I opened up my Meizu M6SL. markun, do you need pictures of the inside? I looked on the port page and the link to the meizu forum with internal pictures seems to be broken. don;t know wether they weren't pics of SP
22:22:13 Quit DataGhost (Nick collision from services.)
22:22:21 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
22:30:27 Quit sarixe ("Connection reset by the motherfucking peer")
22:31:28mcuelenaerel403: it's always nice to have some extra pictures
22:34:04l403well, I managed to pop off the back cover but can't flip out the pcb to see the other side. I see the samsung ARM, something that might be the codec and maybe a power manager so don't know if I should try to open it.
22:41:52gevaertsIf you find an easy way to do it, go ahead, but don't break your player in the attempt...
22:42:06 Join ko [0] (n=56ddcfdd@gateway/web/cgi-irc/
22:42:23 Quit ko (Client Quit)
22:45:09 Quit bimbel ("Woah!")
22:45:30CtcpPing from gevaerts!n=fg@rockbox/developer/gevaerts
22:52:10l403hi gevaerts :) how is it going? I opened it cause it wouldn't turn on cause some moisture/water got in and I thought its dead but then I found an oxidized robon connector guess coming out of the sensor, so I cleaned it and it works again
22:52:58gevaertsl403: I hope to work on the meizu again soon :)
22:55:02 Join xnyhps [0] (n=xnyhps@2001:470:1f14:da:219:e3ff:fed7:c57c)
22:55:13l403alot of ppl do :D
22:55:48 Quit jhulst (Read error: 60 (Operation timed out))
22:56:13 Quit {phoenix} (Read error: 104 (Connection reset by peer))
22:59:05 Join sarixe [0] (n=sarixe@
23:13:42 Join pixelma_ [0] (n=pixelma@rockbox/staff/pixelma)
23:13:44 Quit pixelma (Read error: 110 (Connection timed out))
23:13:44 Quit amiconn (Read error: 110 (Connection timed out))
23:14:23 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
23:17:41l403gevaerts: where shouold I put my findings? shouold I try to insert it somwhere appropriate in the wiki or just leave a post in the forum and some of you willl organize it conviniently into the wiki?
23:18:21gevaertsl403: attach them somewhere on the wiki. If someone finds a better place later they can easily be moved
23:18:36LloreanBut you should try to be appropriate if you can.
23:18:43l403okeley dokeley
23:20:16gevaerts has some pictures already (of the M6SP I guess).
23:20:45gevaertsAt least it has links to pictures that once existed...
23:22:54l403I cant loaod the link to meizu forum foor the SL picturez
23:23:53l403argh, I don't feel like filling in my ral name in the wiki registration :-(
23:24:12 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
23:24:59l403I guess I'll have to leave it somewhere in the pastebin for you when I'm finished
23:25:31gevaertsThat doesn't really solve things. We generally don't like to not know where content comes from
23:28:03gevaertsi.e. the real name policy isn't there to be difficult, it's there to be reasonably sure of copyright status of material. Putting things on a pastebin anonymously for someone else to put them on a wiki goes directly against that
23:29:27l403I, the copyright hollder of the work that is yet to be done :D releasy it under public dumain :D
23:29:43gevaertsOK. Then add a note with your name stating that :)
23:29:55kadobanl403: that's not very helpful unless anyone has any idea who you are :)
23:30:42 Join Zagor [242] (
23:31:56l403ooh, man. first, I work, then I'll have a cigar and then think abouut how to publish it ;)
23:33:16 Join Seed [0] (
23:34:22 Quit xSlack ("Ex-Chat")
23:34:44 Join xSlack [0] (
23:38:32Zagoryikes, the vorbis code looks like it's written by someone with whitespace allergy!
23:39:18ZagorI'm surprised he wasted space indenting...
23:39:40 Join balug_ [0] (
23:43:41amiconnZagor: That means you probably shouldn't look at code written by HCl either...
23:50:49 Quit itcheg (" ajax IRC Client")
23:51:44 Quit Seed ("cu, Andre")
23:51:45***Saving seen data "./dancer.seen"
23:52:26 Quit nibbler (Read error: 110 (Connection timed out))
23:53:48 Quit domonoky (Read error: 54 (Connection reset by peer))
23:56:50Zagoramiconn: what did he write?

Previous day | Next day