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 2010-03-31

00:00:26 Quit jgarvey (Quit: Leaving)
00:01:20FlynDicepamaury: &
00:02:34pamauryFlynDice: my question was more about reducing the buffer to 32-byte and sending a sector in two times: first time with the bank and then null the buffer and send the rest
00:07:13FlynDicepamaury: I thought the card may be expecting a full sector of 512 bytes but funman seemed to think it may not matter if it gets the full 512. Noone ever tried it. You could always just try it and see if it works?
00:07:57pamauryOk, but we could just continue to send a whole sector, just use a 32-byte buffer instead of a full 512 one (I think I'm not clear :))
00:08:13 Quit bertrik (Quit: De groeten)
00:08:22 Join WalkGood [0] (~WaveRider@unaffiliated/walkgood)
00:10:05WalkGoodquick question, the old e200r build link isn't linking anything?
00:10:40WalkGoodthis one
00:11:20FlynDiceWhy don't you just try it by sending 32 bytesand see if it works. If it doesn't then send enough more bytes to make it 512? I don't see why you couldn't just use a 32 byte buffer, send the bank switching code, then zero out the buffer and send the rest of the bytes, but like I say noone tried it
00:12:01FlynDiceback later ;)
00:12:48 Quit xiainx (Ping timeout: 258 seconds)
00:15:09 Join kramer3d_ [0] (~kramer@unaffiliated/kramer3d)
00:16:10 Join webguest56 [0] (
00:16:13 Quit webguest56 (Client Quit)
00:16:44 Nick Guest62701F is now known as Status (
00:16:49 Join froggyman [0] (~me@unaffiliated/froggyman)
00:17:24 Join xiainx [0] (
00:17:51 Quit shaggy-h ()
00:20:31pamauryFlynDice: look at the current code ! It sends 32-byte at a time, and does it until the whole sector has been transfered
00:20:49 Part captainkewllll
00:22:01pamauryFlynDice: I'll have a try tomorrow and I'll tell you
00:22:28 Quit pamaury (Quit: Page closed)
00:23:38 Join saratoga [0] (~9803c6dd@gateway/web/freenode/x-ofryovwaopbtkjhn)
00:23:59saratogaWalkGood: i don't think theres a separate e200R build, so that link probably just linked to the main e200 build
00:24:36saratogaIIRC the R needed a different bootloader, but the rockbox build is identical either way
00:24:52WalkGoodthanks, forum question
00:24:55saratogai'm not sure why that link was broken then
00:25:03saratogaseems like it would be less confusing to have it keep working
00:25:13 Join RoronoaZoro [0] (~vijayss@
00:25:15WalkGoodsure would
00:25:36 Join shaggy-h [0] (
00:26:58 Join User67703 [0] (
00:28:14User67703text viewer plugin quited very long, around 5 seconds on my sansa e200. on 3.5.1 release all good
00:28:37Tornehm, did we know that check_bootfile(false); takes 1.5 seconds at boot on ipodvideo?
00:28:43Tornethat seems pretty.... long
00:28:48 Part WalkGood
00:29:06 Join Darkknight512 [0] (
00:29:31Ziphilti am having trouble with installing to an iriver h340: the automatic installer claimed to work, but only the default firmware boots. I tried patching a firmware .hex file and told the player to update with it, to no avail
00:29:34User67703on 3.5.1 quit time ~0.1sec
00:30:14 Quit phanboy4 (Quit: Leaving)
00:30:14Ziphiltthe new firmware loaded, but no rockbox
00:31:46 Nick fxb is now known as fxb__ (
00:32:06 Quit geertvdijk (Quit: ChatZilla 0.9.86 [Firefox 3.6.2/20100316074819])
00:35:02 Quit junkY_ (Quit: Verlassend)
00:36:54 Part RoronoaZoro ("Leaving")
00:37:50 Join RoronoaZoro [0] (~vijayss@
00:40:23User67703and bars mode in fft plugin available only in vertical mode
00:41:05 Join WalkGood [0] (~WaveRider@unaffiliated/walkgood)
00:44:39 Join Tomis [0] (~Tomis@
00:46:16 Quit ender` (Quit: I don't see what C++ has to do with keeping people from shooting themselves in the foot. C++ will happily load the gun, offer you a drink to steady your nerves, and help you aim.-- Peter da Silva)
00:46:23Tornewow. basically the entire boot time of my ipod is "loading the theme"
00:46:33 Part WalkGood
00:46:44Torneif you kill the theme entirely and use the compiled in font/icons/wps/sbs..
00:46:51Tornestartup takes 1.2 seconds
00:51:03 Quit kramer3d_ (Quit: Leaving)
00:55:40 Quit Stephen__ (Quit: Leaving)
00:58:15Ziphilti fixed my own problem by trying again; goodbye all
00:58:31 Part Ziphilt ("Leaving")
00:58:42TorneLlorean: it's nothing to do with dircache, as far as i can tell
00:58:53TorneLlorean: cabbiev2 just takes six seconds to load on ipodvideo 80gb
00:58:58TorneLlorean: i have graphs :)
00:59:16Torneand as a bonus if you are loading cabbie it somehow makes the step where it saves the mtime/size of rockbox.ipod take 1.5 seconds instead of no time at all
00:59:24Torneporesumably because some disk access is still going on in the background
00:59:38Tornedircache on or not makes no real difference. this is with all other settings at the defaults.
01:00:07***Saving seen data "./dancer.seen"
01:00:25TorneLlorean: so if it didn't take this long int he past then presumably the theme loading code has gotten worse.
01:00:42TorneLlorean: this is disabling the whole theme, btw, so no font, no icons, no wps, no sbs
01:00:44LloreanTorne: Well, dircache (on my player at least) is still obviously behaving weird though
01:00:54TorneYah, maybe
01:00:57LloreanIt may be that the slow theme loading is hurting dirache too somehow.
01:01:01Tornei don't have your player to test :)
01:01:08Tornebut seriously, six seconds
01:01:10Tornenot good
01:01:38Tornethe entire boot from the time the tick gets turned on until calling root_menu takes 1.23 seconds with no theme
01:01:46Torne8.7-8.8 with cabbiev2
01:01:48Torneeverything else the same
01:02:06Tornedircache makes it ~100ms slower to get to root_menu
01:02:11Tornei didn't try tagcache
01:03:58moosTorne: problem I noticed aren't related to dircache, I suspect r23258 or around it, but canot test right now (trying to fix my beast, any hand would be great:)
01:04:32moosit's even before dircache scan :)
01:04:51TorneYeah, the theme is all loaded before dircache is initialised
01:04:59Tornewhich might also be a problem ;)
01:05:23Tornei'm gonna put more charting in
01:05:26moosTorne: you know pretty well the evil beast OF, right?
01:05:28Torneat the moment i'm only charting init()
01:05:30Tornemoos: yes
01:05:40Tornei'm gonna chart settings_apply() as well to narrow it down more
01:06:02mooslike once a year, it make me crazy
01:06:12Torneoh, it reformatted itself?
01:06:21Torneyeah i haven't figured out the cause of that yet
01:06:24moosif you can check 23257 if you have all handy
01:06:25 Quit ivortheengine (Ping timeout: 258 seconds)
01:06:53Tornemoos: well it's highly likely that *is* it, but i'm just gonna see how much time it spends in each part for now
01:07:02Torneif it's sbs loading then that will show up on a chart of settings_apply
01:07:06moosI made the mistake to retry to get single boot runing (related to this loong boot we have those times)
01:07:31TorneYeah i've not discovered anything useful about that yet
01:07:50Tornewe had several working theories that have failed to be supported by what i've found so far in the disassembly but i've not been working on it for a while because it is quite tedious ;)
01:08:09moosand now things was reformated, but windows dosn't seems to recognise my beast well, then no possibility to use patchers
01:08:21Torneit shohuld always be fine to recover
01:08:27moosTorne: hence my *evil* OF :)
01:08:28Tornethe bootloader tries *really hard*
01:08:30Tornetoo hard, in fact
01:08:43moostoo much hard here? :)
01:11:10moosI still have a .zip with sendfirm, but canot have this running... In the past I recovered without pbs with vpatchers, but now... :/
01:11:45moosseverals times even, thanks Microsoft
01:13:17Tornei do know wher in the flash to poke to make it stop formatting the disk ;)
01:13:19Tornethere's a config page
01:13:22 Join clauwn [0] (
01:13:29Tornealas if you stop it formatting the disk then all it does is just hang
01:13:38Torneso while you avoid the format, it doesn't make the device any more bootable
01:13:46clauwnguys, i have to ask some questions... my rockbox is not adding some of my files into the database
01:13:46Torneyou then would have to take the disk out and repair it with something else ;)
01:14:16clauwnhow do i force that?
01:14:24Torneyou don't
01:14:33clauwnand how do i fix that?
01:14:35Torneif it finds the files and they have valid metadata they will go in the db..
01:14:51Torneso i would guess it can't read their tags
01:15:11clauwnthey do have valid metadata
01:15:22Torneyou know this because..
01:15:25moosTorne: aie, I guess it didn't formated my 120gb drive, how to check now?
01:15:42Tornemoos: hm?
01:15:46clauwnTorne: because i saw it?
01:15:46Tornemoos: it will have formatted it
01:15:50clauwni can access it
01:15:50Torneclauwn: saw it in what?
01:15:57clauwnnautilus :)
01:16:00Tornemoos: this isn't a setting you can turn off
01:16:11Torneclauwn: unless you saw it in rockbox, that doesn't count :)
01:16:19Torneclauwn: can you play the files in rockbox by selecting them manually?
01:16:21 Join komputes [0] (~komputes@ubuntu/member/komputes)
01:16:24Tornedoes the right tag info show up when you do?
01:16:57moosTorne: don't remenber the circumstances, but I already updated the nk.bin without reformat
01:17:34Tornemoos: it reformats the disk if it doesn't like *anything* about it
01:17:50Torneif it gets to the point where it's showing the screen telling you to connect to the pc and run the tool then it's already formatted it.
01:18:04moosbecause it did take a while when it did formated it few times in the past. But when it updated without reformat, thesteps was fast, like this time
01:18:23clauwnTorne: now i deleted the two folders which it didn't get and now they are in the db but not on the HDD
01:19:06Tornemoos: it doesn't format the entire disk if it doesn't have to
01:19:11LloreanTorne: That isn't true, actually
01:19:21Tornemoos: if it thinks the firmware is bad it will just format the firmware partition, which is small and thus doesn't take long
01:19:24LloreanI've had it tell me to hook up to the PC and run the tool, and managed to reboot back into Rockbox
01:19:34moosTorne: I vaguely remenber someone found something about this non reformat thing but don't remenber. But now it stuck at this "connect bla bla..." screen
01:19:41TorneLlorean: hm, odd
01:19:42LloreanThis has actually happened to me a few times.
01:20:23LloreanIt would tell me to run the tool if I accidentally rebooted into the OF, but battery poweroff + reboot into the OF would get it to working (if I tried to reboot into Rockbox, I'd get that message again).
01:20:29LloreanThen once I'd gotten into the OF, I could get back to Rockbox.
01:20:30moosTorne: ah then it could just reformat the other partition, ok
01:20:47Tornemoos: it only reformats the entire disk if it doesn't like the partition layout, or similar
01:20:54Torneif it's just unhappy with the firmware it won't erase the main data partition
01:21:33moosTorne: ok. Then now you think it didn't make the works well and that I have to reformat the drive elsewhere?
01:22:19TorneLlorean: also, i just charted settings_apply and the bit that takes all the time when loading cabbie is loading the font
01:22:22Tornenothing to do with sbs/wps
01:22:25moosbizzare that it could be that silly, since it pass good all the steps, even if I re use the OF recovery/format combo
01:22:40TorneLlorean: takes nearly the entire 6 seconds inside font_load
01:22:56moosTorne: want me to build you r23257? :)
01:23:20Tornemoos: i'd rather you tested disabling the user font on whatever build you have that's slow ;)
01:24:24TorneLlorean: likewise for you ;)
01:24:25moosI canot test things without my beast here :/ just have my H10 and beast here (and no usb cable for H10, proprietary cable)
01:24:32TorneOh, yeah :)
01:28:14clauwnTorne: solved with some trying
01:28:25 Part toffe82
01:28:28Torneclauwn: what was the problem?
01:28:48clauwnTorne: .trash-1000 and some stuff
01:28:52clauwni had some config errors
01:29:03Torneah, yes, having files in recyclebin-like places will confuse the dtabase
01:29:18clauwnit thought they were available
01:29:36clauwnbut i now deleted the trash folder and deleted the db files
01:29:41clauwnand then it rescanned
01:29:44clauwnso it works fine now
01:33:27TorneIf you are concerned about the PC recreating the trash folder (which it will), the best hting to do is probably to leave the folder there but put a database.ignore file in it
01:33:35Tornebut yeah, s'up to you
01:33:46Tornethe disktidy plugin will delete .Trash and similar things for you, btw
01:34:50saratogai wonder if the database could easily enough check for trash folders and ignore them
01:35:16Tornehm. it could reuse disktidy's config
01:35:18 Quit CGL (Quit: Saliendo)
01:35:20Torneor something?
01:37:07saratogai assume theres only a few possible names for these folders, maybe it could just strcmp a folders against the standard windows, macos, windows names
01:37:27Tornewell, yah, but that's more or less what disktidy's config is anyway, a list of those..
01:37:35saratogaah ok
01:37:36Tornewell, and the odd few other things like thumbs.db and the like
01:37:47Tornebasically, "files OSes like to create all over the shop that you don't need"
01:38:56MinatakuThe usual garbage
01:39:38 Join CaptainKewl [0] (
01:41:10 Quit anewuser ()
01:41:22saratogawell most of those will be ignored by the database anyway, i think its just thinks like a trash folder that are a problem
01:41:59Tornewell, it could be worth it
01:43:39moosTorne: how much time you said to go to root menu? I'm counting for the old version I have here and 1.6 sec
01:43:54moosHD based
01:43:59Torneyeah, 1.6 is about right
01:44:07Tornefrom main()
01:44:16Tornei guess you are probably counting by hand/clock?
01:44:21Tornestarting from when the logo comes up?
01:44:25Tornethat's close enough, for these purposes
01:44:28Tornelogo is early
01:44:32Torne<100ms into boot
01:44:48moosdidn't get those for my beast, and by far !!!
01:44:53mooshence the troubles
01:45:05moosyou could had some 10extras sec at least ;)
01:46:36Tornewell, the *absolute* times vary wildly between devices/settings
01:46:44Torneit's the proportion of the time spent in each task that i'm looking at
01:47:20moosTorne: of course tried on beast with various set of seetings (including default ones)
01:48:54 Quit CaptainKewl (Remote host closed the connection)
01:49:08 Join planetbeing__ [0] (
01:51:38 Join S_a_i_n_t [0] (S_a_i_n_t@
01:52:32 Quit planetbeing_ (Ping timeout: 246 seconds)
01:53:57 Quit S_a_i_n_t_ (Ping timeout: 265 seconds)
01:56:58 Quit Status (Read error: Connection reset by peer)
01:57:28 Part clauwn
02:00:28 Quit xiainx (Ping timeout: 264 seconds)
02:02:23 Quit Adubb (Read error: Connection reset by peer)
02:02:49 Join Status [0] (
02:02:53 Join Adubb [0] (~aldubuc@
02:04:11 Join xiainx [0] (
02:10:10 Quit Adubb (Read error: Connection reset by peer)
02:11:45 Quit adnyxo (Remote host closed the connection)
02:14:40 Join Adubb [0] (~aldubuc@
02:20:09 Join adnyxo [0] (
02:21:59 Join willyg_cos [0] (
02:26:34 Join webguest45 [0] (
02:27:35 Quit webguest45 (Client Quit)
02:34:16 Part willyg_cos
02:35:31 Part RoronoaZoro ("Leaving")
02:37:01 Quit xiainx (Ping timeout: 268 seconds)
02:38:15 Join CaptainKewl [0] (
02:43:21 Join RoronoaZoro [0] (~vijayss@
02:45:28RoronoaZorowhen i run the command 'svn co svn:// rockbox' it says svn: Unknown hostname ''
02:46:05RoronoaZorois svn down or is the command wrong
02:47:47 Join xiainx [0] (
02:49:01RoronoaZoroanyone ????
02:50:20RoronoaZoroleave a reply :-)
02:50:28 Part RoronoaZoro ("Leaving")
02:54:57saratogaRoronoaZoro: SVN is working fine, so I would check for connection problems on your end
02:58:08saratogaBagder: can you remove the e200r from this page:
02:58:27saratogawe don't have a separate release for the 'R' model anymore
02:59:25 Join walrus [0] (
03:00:08***Saving seen data "./dancer.seen"
03:04:41 Join RoronoaZoro [0] (~vijayss@
03:05:33RoronoaZorosaratoga: the connection on my side is working fine
03:06:15RoronoaZoroi am able to connect to all sites ... does it have something to do with proxy??
03:06:25RoronoaZorocauz i am behind one
03:10:34 Quit planetbeing__ (Quit: planetbeing__)
03:11:01 Quit komputes (Ping timeout: 260 seconds)
03:14:40 Quit slck (Ping timeout: 248 seconds)
03:15:03 Quit Horscht (Quit: Verlassend)
03:15:28 Join saratoga_lab [0] (~9803c20d@gateway/web/freenode/x-msjvenfmplgfjpbi)
03:15:39 Quit mt (Read error: Connection reset by peer)
03:15:51 Quit merbzt (Read error: Connection reset by peer)
03:16:07 Quit Kitr88 (Read error: Connection reset by peer)
03:16:35saratoga_labi'm not sure, but I would guess a proxy might not work with SVN
03:18:04 Quit jd (Read error: Connection reset by peer)
03:18:21 Join jd [0] (
03:18:21 Quit jd (Changing host)
03:18:21 Join jd [0] (~jd@Wikipedia/HellDragon)
03:19:11 Join funman [0] (~fun@rockbox/developer/funman)
03:20:38funmanjunky_san left, i'm just going to commit the change anyway
03:22:27CIA-5New commit by 03funman (r25405): Clipv2/Clip+ : use CHG_IN adc channel, it seems to work better than BVDD and RTCSUP ...
03:23:49 Join slck [0] (Venci@Slackware.SlackPix.Com)
03:30:32 Quit Schmogel (Ping timeout: 258 seconds)
03:32:10 Join Blue_Dude [0] (
03:33:59Blue_DudeOK, got a lot of work done on different keymaps. There are quite a few targets that can be switched on for hotkeys without any more keymapping. But there are also many that need work. I'm putting a big patch in FS #11081 that has all the keymap changes.
03:35:35Blue_DudeAll the targets that can support hotkeys out of the box are switched on. The others have placeholder lines that show where work is needed, and are switched off. Some look like they won't ever support hotkeys but that's up to someone more knowledgeable than me to determine.
03:37:04Blue_DudeThis might be a good time to figure out whether we need to overhaul the keymaps in their entirety to be more consistent across the range of targets. Even similar targets have different mappings that aren't strictly necessary.
03:40:36saratoga_labfirst two GSOC proposals are in
03:43:17funmanBlue_Dude: how would you do that?
03:43:29Blue_DudeDo what?
03:43:39funmanoverhaul the whole keymaps
03:44:10funmanreading every keymap for every target looks like the shortest path to madness
03:44:10saratoga_labpeople's heads would explode, except maybe for the newest targets
03:44:56funmansaratoga_lab: :) even then, some keymaps were just copied from old targets
03:45:33Blue_DudeWell, there's always DevCon.
03:45:35saratoga_labi often wonder about our keymaps, I can't half the time remember what buttons do even for players I've used for years
03:45:46saratoga_labthough i guess having extra buttons i never use doesn't really hurt things
03:47:57Blue_DudeOK, patch is uploaded. It's a big one. Anybody want to take a look?
03:48:55Blue_DudeTHe code is the same as last night's. The keymaps and configs are different.
03:49:34 Join aaron__ [0] (~aaron@
03:49:56saratoga_labspeaking of keymaps, did anyone ever have a good argument about making power off happen even if another key is held down with power?
03:50:05saratoga_labthis is driving me nuts on the Clip
03:51:25Blue_DudeBTW, the Clip was one of those targets that had a mapping conflict...
03:51:37saratoga_labi see
03:51:53saratoga_labis the hotkey persistent across reboots?
03:52:09Blue_DudeYeah. Same as most settings.
03:52:14funmanBlue_Dude: what conflict?
03:52:29saratoga_labit replaces the view playlist option in the wps
03:52:30Blue_DudeBUt you can patch it to make it reset at boot time.
03:52:47saratoga_labi guess you could always make the playlist the hotkey
03:53:24Blue_Dudefunman: I couldn't assign the same key to both the WPS and tree hotkeys. There was probably already an assignment that would have prevented it.
03:54:41Blue_DudeYeah, there is it. That keypress was already assigned to ACTION_SETTINGS_DECREPEAT.
03:55:34Blue_DudeSo something has to change before I can switch it on. That decision is better coming from someone more familiar with the target.
03:56:33Blue_DudeIt also goes to ACTION_STD_NEXTREPEAT. My, there are issues.
03:56:45 Quit RoronoaZoro (Quit: Leaving)
03:59:13funmani quite don't get how keymap are handled
04:00:12funmanACTION_SETTINGS_DECREPEAT in settings is the same than ACTION_STD_NEXTREPEAT in standard context ?
04:01:19 Quit TheSeven (Disconnected by services)
04:01:35 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
04:01:44 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
04:02:22Blue_DudeThat's what it looks like. But context points to standard, so wouldn't they both fire?
04:03:01Blue_DudeOr would just the first one listed in the switch/case fire?
04:03:03*S_a_i_n_t looked at messing with some keymaps...*once*
04:03:12funmani'm lost
04:03:18funmanaren't "settings" and "standard" separated ?
04:05:09Blue_Dudefunman: The keymaps can point to another keymap in a given context. In the Clip (as in most), after the context menu keys are listed, they point also to the standard keys. But both structs have the same key mapped to different values.
04:06:03funmanah right
04:06:11Blue_DudeI just don't know in a given situation which value would pop out from a given keypress.
04:07:34Blue_DudeThe first listed? i.e. the context? Or would the keypress handler keep looking and recognize the last listed? I would guess it would see the first and quit looking but I don't know.
04:07:35funmanif you want to make a change i can test it
04:08:22Blue_DudeThat's actually pretty clever if it's on purpose. If it's not on purpose, well...
04:08:38S_a_i_n_tThe keymaps are short. I wanted to try to not make it a mindblowing experience switching between OF/RB and gave up very quickly.
04:08:42funmanI pretty much copied the c200 keymap and s/REC/HOME/ iirc
04:10:26Blue_DudeI'm baffled at the differences between the e200 and c200. Sure, the c200 doesn't have a scrollwheel but some of the changes seem kinda random. I'm sure there's a good reason, but I have to check the manual to see how to make it go.
04:10:51 Quit adnyxo (Ping timeout: 265 seconds)
04:12:34Blue_DudeThat's why I wondered aloud whether there ought to be some overarching philosophy regarding the interface. Play always plays/pauses, long play always stops, etc. Just basic stuff.
04:13:33Blue_DudeBut I guess with hard buttons on some targets that probably wouldn't work.
04:14:07funmanhard buttons?
04:14:40Blue_DudeWell, bad name. Buttons that are already labeled for the OF.
04:15:19Blue_DudeSeperate play and stop keys for example.
04:16:05Blue_DudeThat would be enough to crash the "overarching philosophy" pretty quickly.
04:17:07 Quit Strife89 (Quit: Bed.)
04:17:48 Quit Darkknight512 (Quit: ChatZilla 0.9.86 [Firefox 3.5.8/20100202165920])
04:29:24 Quit evilnick (Read error: Connection reset by peer)
04:29:50 Join evilnick [0] (
04:32:17 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
04:35:24 Join Rob2222 [0] (
04:38:50 Quit Rob2223 (Ping timeout: 245 seconds)
04:47:21saratoga_labwell we do ignore button lables a lot
04:47:38saratoga_labon the fuze we have a home button but don't use it to go back to the main menus
04:47:49LloreanI don't think we ignore too many button labels.
04:47:59LloreanI don't know about the newest targets, but e200/c200 and earlier we pretty much respect them.
04:49:21saratoga_labthe problem with the fuze is that using the home button for what you would expect means it would work quite differently then the e200 even though they're very, very similar players
04:49:44 Quit Status (Ping timeout: 268 seconds)
04:49:44saratoga_labaside from the home button
04:52:50LloreanIs the home button also the power button (like the e200)?
04:53:04funmanno it 'replaces' the e200 rec button
04:53:41saratoga_labfuze has a dedicated power button, although its a slider thats switches between off, hold and power
04:54:23 Join Status [0] (
04:56:09funmanrecording keymap on clip is crap
04:56:28 Join Barahir_ [0] (
04:57:30 Quit Barahir (Read error: Operation timed out)
04:58:41CIA-5New commit by 03funman (r25406): Clip: enable recording keymap when needed (forgotten in r25390) ...
05:00:10***Saving seen data "./dancer.seen"
05:00:28 Quit evilnick (Read error: Connection reset by peer)
05:00:41 Join phanboy4 [0] (
05:00:50 Join evilnick [0] (
05:01:02 Quit shai (Write error: Connection reset by peer)
05:01:21 Quit TheSeven (Disconnected by services)
05:01:26 Join shai [0] (
05:01:35 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
05:01:45 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
05:08:46Blue_DudeLlorean: So what do you think about the new keymaps? Is that what you had in mind?
05:11:59LloreanFrom your description, it sounds like it.
05:15:16Blue_DudeMost of the Sansas, all the iPods, IRiver H100's and 300's, Archos Recorders and Players, AV300, are all fully supported. A number of other just need someone to do some remapping/deconflicting.
05:16:16Blue_DudeAnd the placeholders are already there for those targets that need some love.
05:18:48 Quit moos (Quit: ChatZilla 0.9.86 [Firefox 3.6.2/20100316074819])
05:21:37saratoga_labi wouldn't worry too much about the av300
05:22:48Blue_DudeWell, it was there, so... Anyway, it's supported because it didn't have any conflicts. So both of its users can enjoy hotkeys now.
05:23:34saratoga_labi don't think its actually a usable device
05:23:39saratoga_labnot even sure if you can compile it
05:24:26saratoga_lab"54MHz arm7tdm"
05:24:34saratoga_labthat would have been a fun target for codecs
05:24:39Blue_DudeBut the keymap works great, dude. :)
05:28:49 Quit saratoga_lab (Quit: Page closed)
05:33:26 Join hd [0] (
05:33:26 Quit hd (Changing host)
05:33:26 Join hd [0] (~jd@Wikipedia/HellDragon)
05:35:35Blue_DudeAnyway, I'm still planning to commit soon. Probably tomorrow night. I need to get this thing off my hard drive so I can work on something else. :)
05:36:19 Quit jd (Ping timeout: 245 seconds)
05:36:43Blue_DudeThinking about making a bookmark versioning system so you can change bookmark formats and add optional info without breaking backward compatibility.
05:41:16 Quit CaptainKewl (Remote host closed the connection)
05:43:40 Quit Blue_Dude (Quit: ChatZilla 0.9.86 [Firefox 3.6.2/20100316074819])
05:54:02 Quit SirFunk_ (Quit: Ex-Chat)
06:03:43 Quit User67703 (Ping timeout: 268 seconds)
06:24:21 Quit n17ikh (Ping timeout: 258 seconds)
06:30:25 Join n17ikh [0] (
06:42:14 Join CaptainKewl [0] (
06:44:09 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
06:46:25 Quit CaptainKewl (Remote host closed the connection)
07:00:14***Saving seen data "./dancer.seen"
07:12:41 Quit blairb (Ping timeout: 240 seconds)
07:15:39 Join blairb [0] (
07:16:46 Quit xiainx (Quit: Good Bye)
07:23:57 Quit S_a_i_n_t (Quit: Stupid "auto-update" restart...Bah!)
07:28:37 Join S_a_i_n_t [0] (S_a_i_n_t@
07:29:43 Join planetbeing_ [0] (
07:30:35 Quit walrus (Ping timeout: 264 seconds)
07:41:01 Join LinusN [0] (~linus@rockbox/developer/LinusN)
07:42:08 Quit planetbeing_ (Quit: planetbeing_)
07:44:07 Join planetbeing_ [0] (
07:45:29 Join arbingordon [0] (~w@unaffiliated/arbingordon)
07:46:20 Join Buschel [0] (
07:53:39 Quit Zarggg_ (Quit: Zarggg_)
07:58:37 Quit gevaerts (Ping timeout: 276 seconds)
08:02:47 Quit Buschel ()
08:05:05 Quit jordan` (Remote host closed the connection)
08:10:35 Join Zarggg [0] (~zarggg@2001:0:4137:9e74:0:fbf2:beb1:ba3d)
08:14:04 Quit advcomp2019__ (Changing host)
08:14:04 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
08:14:09 Join Bagder [0] (~daniel@rockbox/developer/bagder)
08:18:01 Join ender` [0] (
08:19:43 Quit beta2k (Ping timeout: 252 seconds)
08:19:59 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
08:20:21 Join beta2k [0] (
08:33:21 Join wodz [0] (
08:33:54 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
08:40:42 Quit gevaerts (Read error: Connection reset by peer)
08:42:29 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
08:49:19 Join mitk [0] (~mitk@
08:50:40 Quit bmbl (Ping timeout: 265 seconds)
08:57:30 Quit tmzt (Ping timeout: 276 seconds)
09:00:08 Join Zagor [0] (
09:00:09 Quit Zagor (Changing host)
09:00:09 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
09:00:15***Saving seen data "./dancer.seen"
09:00:19 Join geertvdijk [0] (
09:04:02 Join petur [0] (~petur@rockbox/developer/petur)
09:07:37 Quit froggyman (Quit: GAME OVER)
09:12:54 Join huelk [0] (~huelk@n60.informatik.FH-Dortmund.DE)
09:15:36 Quit RadicalR (Read error: Connection reset by peer)
09:15:47 Quit evilnick (Read error: Connection reset by peer)
09:29:44 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
09:29:51 Join flydutch [0] (
09:35:45 Join ivortheengine [0] (~ivortheen@
09:38:17 Quit wodz (Quit: Leaving)
09:41:59 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
09:42:02 Quit pamaury (Client Quit)
09:42:08 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
09:47:43 Quit bmbl (Ping timeout: 264 seconds)
09:49:22 Quit huelk (Ping timeout: 258 seconds)
09:49:46 Quit Connor_ (Ping timeout: 264 seconds)
09:52:10 Join Connor [0] (
09:56:57 Join efyx [0] (
10:00:58 Join huelk [0] (~huelk@n60.informatik.FH-Dortmund.DE)
10:02:14 Join ender [0] (
10:03:09 Quit ender` (Ping timeout: 276 seconds)
10:03:32topiki am looking at keymaps. is there a way/place to find all possible/defined/allowed actions for each screen?
10:03:35 Join einhirn [0] (
10:11:42Zagortopik: in the code?
10:13:39 Quit tmzt__ (Read error: Operation timed out)
10:16:05 Quit aaron__ (Quit: Leaving)
10:16:15 Quit Connor (Quit: Leaving)
10:16:32 Join tmzt_ [0] (~ircuser@
10:19:33 Join CGL [0] (~CGL@
10:20:48 Quit CGL (Client Quit)
10:25:01topikwell i'm comparing the clip and fuze keymaps and they have some totally different actions defined
10:25:15topikso i don't really know if all are used, valid or some are missing
10:29:13Zagorwhere are you comparing?
10:33:07 Quit phanboy4 (Read error: Connection reset by peer)
10:33:34topikfm radio screen
10:44:32 Join Kitar|st [0] (
10:49:47CIA-5New commit by 03user (r): logmsg
10:50:24Zagor^^^ that's just me testing
11:00:18***Saving seen data "./dancer.seen"
11:01:25 Quit huelk (Ping timeout: 258 seconds)
11:03:48 Join vaguerant_ [0] (
11:04:44 Part vaguerant_
11:04:46 Join vaguerant_ [0] (
11:05:53 Quit vaguerant_ (Client Quit)
11:06:09 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
11:07:34 Join Luca_S [0] (
11:08:39Luca_Syay! rockbox on my fuzev2! nice work guys :) I followed the instructions on the forum and I see the display, seems like I have the same version of the developers' ones
11:10:39funmanLuca_S: so far we're not even sure the other version exists
11:10:46 Join jordan` [0] (~jordan@
11:12:14Luca_Sis there any way to move the selection down in lists? all the scrollwheel buttons seems to work, even the hold and the home button
11:12:40funmanno, wait for the scrollwheel support to be added
11:12:45pamaurybluebrother: how translation works ? I edit on the web tool or with linguist and then I post on the tracker or I commit it directly ? Is someone reviewing translations ?
11:12:50 Join vaguerant [0] (~3aaf4cc7@gateway/web/freenode/x-evodgvkugialfsnj)
11:13:04 Quit vaguerant (Changing host)
11:13:04 Join vaguerant [0] (~3aaf4cc7@wikipedia/vague-rant)
11:13:31Luca_Sok :) thank you
11:13:37funmanpamaury: do you want review of the french?
11:13:42 Quit BHSPitMonkey (Quit: Ex-Chat)
11:14:17pamauryWell it's always better if several people have a look at it :)
11:15:30vaguerantIs there anywhere more specified than here to discuss emulators running on Rockbox?
11:16:11funmanvaguerant: depending what you want to discuss, here or on the forums. What is the topic?
11:16:45vaguerantWas curious if anyone had looked at Pokemon mini emulation.
11:17:55vaguerantIt's a handheld Nintendo released in the late '90s IIRC, notably the display res is 96*64, so it has potential to work on a lot of devices; on the downside, it's all custom parts so it's not well documented. There's a GPL'd Windows emulator though.
11:18:07funmannot that i'm aware of
11:21:22 Quit Rob2222 (Quit: Rob2222)
11:22:10 Quit vaguerant (Ping timeout: 252 seconds)
11:24:26 Join Rob2222 [0] (
11:24:49 Quit Rob2222 (Client Quit)
11:25:10 Join Rob2222 [0] (
11:25:58 Quit geertvdijk (Ping timeout: 276 seconds)
11:33:06 Join geertvdijk [0] (
11:36:04 Join xsteadfastx [0] (~spectrum@
11:36:28*JdGordon grumbles at people who think the sbs loading is the boot time culpret
11:36:32 Quit Rob2222 (Quit: Rob2222)
11:41:18JdGordonTorne: I'd be interested in your settings_apply() timeing graph....
11:42:22 Join Rob2222 [0] (
11:48:42funmanor patch? could it go in svn with #ifdef DEBUG for example?
11:51:21JdGordondepends how it was done, If there is a DEBUGF() call around every call then no
11:51:46pamauryfunman: I opened a task on FS for Rbutil french translation, if you want to participate/review, that would be nice as there a lots of unfinished translations
11:52:48TorneJdGordon: I dunno that my results apply to everone
11:52:57Tornethe problems moos/llorean were having may be different
11:52:59JdGordonwhich target?
11:53:04Torneipodvideo 80gb
11:53:14Tornei was just surprised by the *huge* amount of time that font_load() takes
11:53:29Tornebooting with cabbiev2 is a good 7.5s slower than booting with the builtin theme
11:53:32JdGordonso close enough... I can get from press to music in 4s on my CF ipodmini
11:53:43JdGordonfont_load() has always been slow
11:53:46TorneYes, it has
11:53:49JdGordonit reads 100K or so
11:53:56TorneI went back in time substantially and it makes no difference to my player
11:54:06Torneso, i think Llorean/moos have some other problem, maybe
11:54:14TorneI am considering how best to put the boot charting stuff into svn, though
11:54:35JdGordona seperate branch or something
11:54:36Torneat the moment it *is* spectacularly intrusive, but it needn't be
11:54:44TorneNo, I think I can make it reasonable
11:54:56Torneas a compile time option
11:55:18Tornemacro abuse may be involved :0
11:55:25JdGordonI want someone to time boot from today and a specifc build before i care at all about the boot time
11:56:00 Join junkY_San [0] (
11:56:13Torneyeah i don't think i'm reproducing the problems other people are having
11:56:24Torneso, i was gonna see if i can get this code suitable to commit so they can easily test it themselves
11:57:20JdGordonheck, I'd be happy with boot times from 3.0, 3.1, 3.2, 3.3, 3.4, 3.5.. on a consistant config.cfg
11:57:41funmancan we use the simulator for that ? :)
11:58:11TorneJdGordon: I can do that
11:58:19Tornei have been testing with no config
11:58:30Torneor with solely dircache turned on
11:58:43Torneseriously though, why is font loading so slow?
11:58:50Tornethe default font for cabbie is what, 40kb?
11:58:54JdGordonbecause font loading sucks
11:58:58JdGordonno, 60KB i tihnk
11:58:58Torneit doesn't take six seconds to read that from disk
11:59:06Torneyah, but still
11:59:10Torneipodvideo has DMA now
11:59:11 Join jeffp [0] (
11:59:18Torneit can read at 20MB/s with a following wind :)
11:59:30JdGordonit depends how the .fnt is layed out. it might be doing a crap load of seeking
11:59:49Torneon my device it *should* be 100% defragmented
11:59:50jeffphey guys, where is the web cvs??
12:00:07jeffpthankyou sir
12:00:25JdGordontheir boot problems must be a big jump between 3.3 and 3.4
12:00:37JdGordonI don't beleive that there is one anymore
12:00:53 Join dockimble [0] (~dockimble@
12:00:54JdGordonhmm, no, 3.4 to 3.5
12:01:22Torneyah, i dunno what setting or target-specific thing they are tripping over, though, cuz it doesn't happen for me
12:01:40Tornemoos said if he went back to some old revision on his h120 the boot time went down to 1.2s from much longer
12:01:49JdGordonit used to be doing settings_apply() twice in main() which was removed very recently, which *may* have been an issue... and IS the only issue I added
12:01:53Tornei went back to that same revision on my player and it was within ~50ms
12:01:59jeffpdo i have to make an account to report a bug?
12:02:04Tornejeffp: yes
12:02:11jeffpblah k
12:02:14Torneso whatever they're experiencing either doesn't happen on ipod or depends on a non-default setting
12:02:54JdGordonTorne: are you logging to a file or RAM? for proper times I'd want it to at leats not write to disk untill well after boot is finished
12:03:22Tornei'm logf'ing
12:03:36Torneso, it shoulnd't interfere
12:04:37 Quit Luca_S (Quit: CGI:IRC)
12:04:46Tornebut yeah, i'm gonna reduce the number of chart points because right now i have 100+ during boot which makes main.c rather cluttered
12:05:02Torneand change what they logf to be something less ambiguous
12:05:12Torneand then see what people think about sticking it in svn
12:05:55JdGordonI dont tihnk its worth cleaning up and going in svn unless we have very clear objectives...
12:06:08JdGordonotherwise its just as pointless as the delta graphs
12:06:28gevaertsOh, not very pointless then
12:06:32JdGordonsure its nice to know, but is juyst another place for whingers to object to features
12:07:16Tornedoes some part of theme loading happen in the background, also?
12:07:26TorneThen that's really weird
12:08:03Tornebecause having cabbie selected makes check_bootfile take 1.5 seconds
12:08:06JdGordonI think effort should be put into making main() more target independant... why dont we have a target_init() and stuff? only the apps layer init shuold be in main()
12:08:29Tornebut if you pick the compiedl in theme, check_bootfile is basically instant
12:08:38Tornethis is with all other settings identical
12:08:41Torne(and default)
12:08:54Torneit shouldn't take 1.5 seconds to stat a file
12:09:16JdGordondircache off?
12:09:31Tornebut check_bootfile bypasses dircache anyway
12:09:48TorneIt doesn't improve if dircache is on
12:10:03JdGordondoesnt opendir() use dircache if enabked?
12:10:06Tornein fact, with default settings on my ipod dircache makes no difference to boot time except for the extra 100ms or so spent kicking off dircache
12:10:16Torneall the other boot tasks take the same time with or without it
12:11:05JdGordonI ask about dircache because of its thread
12:11:21TorneYes, but like i said, it makes no difference to any part of booting
12:11:26Torneother than inside dircache_init
12:11:31Torne(which is obviously very quick if dircache is disabled)
12:12:00JdGordondoes the disk get a chance to spindown during init?
12:12:07Torneinterestingly check_bootfile is commented saying "we don't use dircache"
12:12:11Tornebut then it just calls opendir()
12:12:17Torneso, i dunno ;)
12:12:34TorneHmm, that's a good point
12:12:50Tornei would hope not :)
12:12:55JdGordonwhen is check_boot() called in relationg to the themes?
12:13:22 Join robin0800 [0] (
12:13:25Tornecheck_bootfile is the second to last thing called in init(), just before settings_apply_skins
12:13:32Torneand long after settings_apply which is what takes most of the time (font_load)
12:13:54Tornei didn't have time to experiment a great deal but it looks like not loading a font doesn't fix check_bootfile
12:14:11JdGordonand check_boot() is faster if cabbie is disabled? even though it is not yet cared about?
12:14:21JdGordonde fuck!? :p
12:14:23Torneall other settings identical :)
12:14:31 Quit ivortheengine (Remote host closed the connection)
12:14:32Tornewell most of cabbie is loaded in settings_apply
12:14:44Torneby time, at least ;)
12:14:54Tornei don't know how any of the theme stuff works, really
12:15:02Tornebut this is 100% repeatable
12:15:26Tornethis was consistent every boot for a dozen or more tries last night
12:15:37Tornetheme selected: takes 1.5s, builtin theme: takes <100ms
12:16:17JdGordonthat makes sense... cabbie is disk heavy, builtin doesnt touch it
12:16:31JdGordonbut no part of any theme should be loaded untill setting_apply_themes()
12:17:14JdGordonexcept the font which is loaded in settings_apply()
12:17:18Tornefont, remote font, keyboard, voice, icons and filecolours are loaded during settings_spply
12:17:29Torneso for cabbie on ipod that's font/icons
12:17:54 Quit shai (Quit: Leaving)
12:18:06JdGordonoh right, forgot about icons
12:18:08 Join dfkt [0] (dfkt@unaffiliated/dfkt)
12:18:22Torneyou aren't vim
12:18:31JdGordonnope :D
12:19:06Torneanyway, i think the charting stuff *is* valuable to have in svn if we can coem up with a way that's not too horrible to look at in the code.
12:19:17 Join n1s [0] (~n1s@rockbox/developer/n1s)
12:19:33funmanTorne: would a simple 1: subs r0, r0, #1\n bne 1B loop be efficient to measure frequency of a CPU ?
12:20:01Tornefunman: that's bogomips, not frequency
12:20:16funmancan't we time this loop in an exact number of cycles?
12:20:31Torneonly if you know the specifics of the CPU you are running on ;)
12:20:42funmani know it's an arm926-ejs
12:20:59Tornewell probably, then, but you'll have to do the math yourself for how many cycles that actually is :)
12:21:50funmanwhere is that described? I don't remember seeing it in the Technical Reference Manual but then I didn't read it fully
12:21:50TorneJdGordon: what's your opinion on wrapping function calls in weird macros? :)
12:22:05JdGordonTorne: a slight gotcha with timing font loading. for the main font, if there is a .rockbox/.glyphcache file the load will be slightly different
12:22:17TorneJdGordon: hm
12:22:21TorneJdGordon: i hadn't allowed for that
12:22:26Tornewhen does that get created?
12:22:30JdGordonTorne: well the settings_list.c macro mess is me so I dont mind too much :)
12:22:44JdGordonshutdown or font unload iirc
12:22:52JdGordonor just comment it out to ignore it
12:23:05 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
12:23:16TorneJdGordon: i mean something like CHART() wrapped around calls in init, like CHART(settings_apply(true);)
12:23:29Tornewhere normally #define CHART(x) x
12:23:54JdGordon#define CHART(x) do { x } while (0);
12:23:58 Quit dfkt (Disconnected by services)
12:24:02 Nick dfkt_ is now known as dfkt (dfkt@unaffiliated/dfkt)
12:24:22junkY_Sanin the clip simulator i can start the stopwatch, go back to now playing and the stopwatch keeps counting. when i do this on my clipv2 the stopwatch is reset when i run the plugin again
12:24:39TornejunkY_San: plugins exit when you exit them
12:24:46Torneif this works on the simulator the simulator is broken
12:25:16Torne(or am i missing something) :)
12:25:26 Quit robin0800 (Read error: Connection reset by peer)
12:25:33TorneAh, yes
12:25:39Tornethe stopwatch saves the time when it started and reloads it
12:25:41junkY_Sanit definetely works on the simulator, but i like that behavour
12:25:48Tornethis doesn't work on clipv2 hardware because clipv2 doesn' thave disk writing
12:25:59TorneIt doesn't really keep counting on the sim either
12:26:02Torneit just fakes it by saving/restoring
12:26:05junkY_Sani see, annother feature of disk writing
12:26:21Torneif you are seeing the wps the plugin has exited. :)
12:26:37n1sit could be quite nifty to make the stopwatch TSR so it could keep running afer exit
12:27:02Tornen1s: that might cause people to wonder why it stops when you run another plugin? :)
12:27:18n1sTorne: yes :)
12:27:54gevaertsn1s: start work on this plugin reform then :)
12:28:10junkY_Sani used the player for the first time yesterday and i like the RB volume compared to the OF
12:28:11n1sIdeally we should implement a prompt like: "Plugin foo is currently running in the background, and needs to exit before starting another plugin, proceed?"
12:28:48n1sgevaerts: plugin reform == multiple position independent plugins?
12:29:29*n1s playse the "I am backpacking in SE asia with no targets at hand" card again
12:29:54n1son hand too i guess
12:30:25n1slooks like i will be job free this summer though :)
12:30:53junkY_Sanwould it be difficult to bind "next track" to a key in the stopwatch?
12:31:26TornejunkY_San: if you can modify the code and rebuild, not difficult at all, no
12:32:35n1sgevaerts: btw i think, (hope) the hardest part of that will be the UI to handle switching between plugins and the allocation of memory of course
12:33:17n1sor are we going with "however many fits in x kB"?
12:33:21gevaertsn1s: I don't care about switching actually. I don't mind a restriction of only one UI plugin :)
12:33:30JdGordongevaerts: I'm splitting your new content from NoDo to NoDoDiscussion so there is a clear seperation between the definites and the "NoDos"...
12:33:51n1sthen how would that help my hypothetical stopwatch in the background?
12:33:52 Join robin0800 [0] (
12:34:06 Quit planetbeing_ (Quit: planetbeing_)
12:34:14gevaertsn1s: that's the next bit! Plugins can expose an API to other plugins! :)
12:34:16funman"cycles per instruction" is mentioned in the glossary of the arm926ej-s technical reference manual, but not in the manual itself
12:34:34Tornefunman: there isn't really a single source that tells you the numbers :)
12:34:42Tornefunman: you kinda have to combine a bunch of different bits of information
12:34:44gevaertsJdGordon: you mean like the HWCODEC NoDos versus things like configurable X?
12:35:19JdGordon"This is a page where arguments for and against each possible NoDo topic can be summarised (please add links to mailing lists and IRC discussions when they happen. Also you MUST add your signature to the edits or they will be removed."
12:35:26funmanTorne: hm ok, can i found all those bits in the manual?
12:35:27n1sgevaerts: sounds complicated, why do we want that? (and how does *that* help me) ;)
12:35:47gevaertsJdGordon: sounds good
12:36:08 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
12:36:16gevaertsn1s: you'd have a stopwatch plugin in background, and a stopwatch GUI
12:37:07n1sok, so a stopwatch server and a client :)
12:37:41n1sand the client silently starts the server if it isn't running and then it runs untill shutdown
12:37:58*n1s feels a bit dense
12:38:16gevaertswell, it could work!
12:38:25funmanperhaps i should just ask saratoga or Unhelpful
12:38:47 Quit dfkt (Ping timeout: 265 seconds)
12:40:35 Quit dfkt_ (Client Quit)
12:42:12*Torne can't help but think there are easier ways of doing the stopwatch ;0
12:44:18 Join shai [0] (
12:46:03 Join Luca_S [0] (
12:46:38junkY_Sancan someone give me a hint which function to call to skip to the next track?
12:47:55linuxstbjunkY_San: Look at apps/plugins/lib/playback_control.c
12:53:15 Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:00:21***Saving seen data "./dancer.seen"
13:00:39 Join moos [0] (moos@rockbox/staff/moos)
13:08:49 Join wodz [0] (
13:11:02pamaurymoos: I started a FS task about french translation of rbutil, if you want to contribute also or review what I did, I guess it would be a good thing to do it on this thread
13:11:55moospamaury: Hi, ok. I'll look at it...
13:13:54JdGordonmoos... do a boot time comparisson between 3.3, 3.4 and 3.5 please...
13:14:20JdGordonwhich all default settings (delete.config.cfg and .rockbox/.glyphcache between boots)
13:15:16moosJdGordon: I wanted to do some tests on my beast (where the thing is more noticable), but I have to fix it first (crappy OF)
13:16:54junkY_Sanyees the next and previous track work in stopwatch :D
13:17:25UnhelpfulTorne,funman: there are tables in teh ARM System Developer's Guide that, for the most part, agree with what i've been able to work out from tables in processor manuals.
13:18:10amiconnfunman: Sound setting only work for fm if they're done in hardware, i.e. it depends on target
13:18:17amiconnThe eq never works for fm
13:19:11Unhelpfulthey're not always *simple* table because there's an increasing number of instructions with variable cycles based on various stall conditions, and on pre-arm9e the multiply instruction terminate early based on the value of the multiplier.
13:20:43wodzamiconn: could You review FS #11153 ?
13:20:48 Part jeffp
13:22:21junkY_Sanmh i just got "codec failure" during playback and now "error accessing control file" when i want to select any mp3 track. it won't play anymore
13:27:22funmanamiconn: do you know how to remove the settings from the screen?
13:27:59funmanUnhelpful: hmm googling for that guide show me books, is there a pdf of it ?
13:29:07Unhelpfulfunman: i believe you can view/download it via docstoc
13:30:59funmanyep just got it, thanks
13:32:24funmani wouldn't even know what to grep for
13:32:37funman(to find sound settings in the FM screen)
13:34:00funmanhm perhaps "sound_settings"
13:34:42funmancan menu entries be dynamically removed?
13:36:18 Join patgodo [0] (
13:37:07 Join perfectdrug [0] (
13:37:12funmanJdGordon: how would we remove treble/bass from sound_settings when in FM screen?
13:37:22JdGordonwith a machete
13:37:28*JdGordon has a quick look
13:38:02funmanbut only on targets which do not have HW control for them (not sure if there's a #define for that)
13:38:12funmanHAVE_SW_TONE_CONTROLS ?
13:38:19JdGordonit would be a bit anoying, but bassicly required a callback for sound_menu
13:38:26JdGordonsound_settings i mean
13:39:33 Quit funman (Quit: free(random());)
13:43:23JdGordonsilly funman leaving....
13:44:00JdGordonanyone got an email for him?
13:45:40JdGordonfunman, see FS #11159
13:47:03GodEaterAnyone else interested in having a go at one of these for Rockbox?
13:48:20 Quit TillW (Ping timeout: 248 seconds)
13:53:56 Quit Rob2222 (Read error: Connection reset by peer)
13:54:20 Join Rob2222 [0] (
13:54:23 Quit Adubb (Quit: "The mind and body must be subjected to extreme stimulus, by means of drugs and music." - HST)
13:57:28 Quit robin0800 (Remote host closed the connection)
13:58:59 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:59:10linuxstbGodEater: Who/what are/is "doubletwist" ?
14:03:08amiconnRegarding yesterday's discussion about slow boot - I am observing the boot slowdown on all targets, and I don't use .sbs and my .wps don't have any bitmap
14:04:03 Quit dockimble (Ping timeout: 246 seconds)
14:04:24amiconnAlso I don't use dircache on several targets (lowmem hwcodec doesn't have it anyway, and on flash/ssd/cf targets and targets with only 16MB RAM I've disabled it)
14:04:51 Nick fxb__ is now known as fxb (
14:06:25amiconnFwiw, I'm not talking about boot-to-musc, but boot-to-browser. I don't use manual resume and my start screen is set to the file browser
14:06:41amiconnErr, I don't use automatic resume...
14:08:22 Join TillW [0] (
14:10:11CIA-5New commit by 03uchida (r25407): Rockbox Utility: Japanese translation updates
14:10:31Torneamiconn: everyone yesterday seemed to be measuring different things, which didn't help ;)
14:17:37junkY_Sanwhen i create a database on the simulator, what needs to be copied to the player? i copied database.log and database_tmp but the player says "database not initialised"
14:19:56S_a_i_n_t junkY_San: I don't *think* it works like that
14:20:11S_a_i_n_tI believe the sim/player handle the "database" differently.
14:21:11gevaertsjunkY_San: the config file
14:21:18gevaertsand the .tcd files
14:23:11gevaertsJdGordon: I'm not sure if "Trivial to implement" (or its opposite) should be used as an argument in a NoDo discussion
14:24:01Tornei'm inclined to agree :)
14:24:19Torneunless we actually *don't know how to do something*, how difficult it is is irrelevant
14:24:32Torneit might be too hard for anyone to bother, but that doens't mean we wouldn't take it if someone did it
14:27:31 Quit moos (Quit: See you later)
14:31:39JdGordongevaerts: well yes and no
14:32:57 Quit junkY_San (Quit: Verlassend)
14:34:34 Join junkY_San [0] (
14:35:38 Join Schmogel [0] (
14:36:16junkY_Sangevaerts: i copied all the files but the database is still not initialized :/
14:36:18Luca_Sfirst song played by my fuzev2: john denver - leaving on a jet plane :D rockbox looks awesome on this device!
14:36:47Luca_Suhm... the clock is wrong
14:37:32Luca_Sshows 05:10 instead of 14:41, but I have no way to say if rb is using 12 or 24 hrs display
14:39:52*Unhelpful could swear that database portability was exactly *why* we have endian support in the database
14:40:17Unhelpfulwe discussed this under "things we could get rid of to fix rombox" :)
14:41:35 Join hebz0rl [0] (
14:43:27 Join archivator [0] (~archivato@
14:46:31 Quit geertvdijk (Ping timeout: 276 seconds)
14:47:58 Join adnyxo [0] (
14:48:21 Join geertvdijk [0] (
14:49:34junkY_SanUnhelpful: ah so what exactly is needed to get the database "initialized" on the player?
14:50:01gevaertsjunkY_San: I'd say the tcd files and the right settings...
14:50:11junkY_Sangevaerts: that's all copied
14:50:50gevaertsJdGordon: why?
14:52:21 Quit archivator (Quit: Leaving)
14:56:35 Quit blairb (Quit: Leaving)
15:00:24***Saving seen data "./dancer.seen"
15:00:38JdGordongevaerts: to counter the delta con?
15:00:52gevaertsDon't talk nonsense
15:01:21*Torne notes that complexity of implementation is barely related to binsize/ram usage
15:01:50*Bagder agrees with Torne
15:02:13*JdGordon is 3/4 asleep and doesnt understand Torne
15:02:17 Quit JdGordon (Quit: goodnight)
15:03:11linuxstbIMO, "hard to implement" isn't a reason for "no do". The question should be "If someone implemented this as well as possible, would we accept the patch?"
15:03:55Tornebut i think jd is conflating this with his dislike of gating stuff based on binsize
15:04:07junkY_Sangevaerts: the file nvram.bin also needs to be copied to the player
15:04:33pamaurybut hard to implement things have to be useful, it's not clever to add a very complicated feature which will trigger zillions of bugs if only one or two persons use it, no ?
15:04:54 Join robin0800 [0] (
15:04:56Tornepamaury: things hav eto be useful *regardless* :)
15:05:04gevaertshard to implement doesn't have to mean hard to maintain either
15:05:45pamauryno but much more error prone and will probably trigger more bugs than an easier to implement one
15:06:32kugelgevaerts: but often?
15:07:08pamauryit pretty much depend on what hard to implement means I think :)
15:12:47gevaertsA plugin can be hard to implement, but it doesn't trigger bugs
15:15:49 Quit togetic (Ping timeout: 260 seconds)
15:18:24pamauryplugins are a bit different because there don't increase delta either
15:18:42pamaury(if binsize delta is the problem I mean)
15:19:44 Quit wodz (Quit: Leaving)
15:23:32 Quit flydutch (Ping timeout: 246 seconds)
15:24:31 Join locke [0] (
15:24:36 Nick locke is now known as Locke_Fireclaw (
15:25:25Locke_FireclawHey all, how's it all going? Got another problem with my RB. Woo. lol
15:25:35 Join flydutch [0] (
15:26:37Locke_FireclawI had a very rocky introduction into compiling my own version yesterday, just to get the 80gb classic HD to work in my 80gb Video.
15:27:08Locke_FireclawIt was a little flaky, so I started completely from scratch today on a proper Ubuntu install instead of the VB I was working from yesterday.
15:27:29Locke_FireclawWell it works, but hard locks when you adjust the volume either too high or too low, any idea's?
15:33:14 Quit tchan (Ping timeout: 265 seconds)
15:33:16linuxstbLocke_Fireclaw: Is that the only time it hard locks?
15:33:54 Quit robin0800 (Remote host closed the connection)
15:34:04linuxstbAnd what do you mean by "too high" and "too low"? And where are you adjusting the volume (i.e. which screen in Rockbox).
15:34:33Locke_FireclawSo far linuxstb.
15:35:25Locke_FireclawI'm sat at the 'now playing' screen, everything works fine with the volume set in the middle-ish of the bar, but if you try to increase the volume to say 0db, it'll just freeze & die.
15:35:30 Join togetic [0] (~togetic@unaffiliated/ibuffy)
15:36:15Locke_FireclawYou CAN set the volume if you go to the sound settings and set it to 0db before playing anything, and it'll work (and deafen you), but as soon as you try to adjust it while a song is playing, it'll lock & die again.
15:37:42linuxstbWhat changes did you make to Rockbox?
15:37:58Locke_FireclawTwo, well, three.
15:38:21Locke_FireclawJust the one's detailed there, to fix the sector size issue with the thinner 80gb HD.
15:38:30Locke_FireclawThat one is for the 240gb, but figured it'd work.
15:41:05 Join robin0800 [0] (
15:44:31 Join evilnick [0] (~0c140464@rockbox/staff/evilnick)
15:44:50 Quit robin0800 (Remote host closed the connection)
15:44:58 Quit n1s (Read error: Connection reset by peer)
15:45:15 Join robin0800 [0] (
15:46:33 Quit hebz0rl (Quit: Ex-Chat)
15:47:17 Part xsteadfastx
15:48:21Locke_FireclawAny idea's?
15:48:56TorneLocke_Fireclaw: I don't know what's causing your problem, but just for reference it's better to *undefine* MAX_PHYS_SECTOR_SIZE rather than increasing it, if this makes your hard disk work
15:49:14Tornethe MAX_PHYS_SECTOR_SIZE thing is to accomodate badly behaved drives which don't deal with this themselves
15:49:30Tornemost drives don't need it and will work faster without it
15:49:37Tornebut if it's defined it must be large enough. if that makes sense :)
15:50:02Locke_FireclawAhh, so on next rebuild, if I just comment it out? I notice what I consider 'commenting' is a #, I see those are already there, so is it a - to comment out in that config?
15:50:06 Join CGL [0] (~CGL@
15:50:20Torne# in C is a preprocessor directive
15:50:27Tornecomments are anything between /* and */
15:50:34Locke_FireclawWoosh, there it goes, right over my head...
15:50:52 Quit mitk (Quit: Leaving)
15:51:00TorneIt's possible your drive *does* require it
15:51:06Tornebut you should try without :)
15:51:15 Join jgarvey [0] (
15:51:26Locke_FireclawI literally am only doing this because I wanted my 80gb to have Rockbox. lol My 4th gen 40gb only ever needed me to poke the installer, so I'm in way over my head here.
15:51:43linuxstbLocke_Fireclaw: So I assume a standard Rockbox build doesn't work?
15:52:04Locke_FireclawOnly because of the hard drive problem linuxstb.
15:52:09Tornelinuxstb: the problem is the standard rockbox builds for ipodvideo *do* define all this, just as too small a size..
15:52:32Tornei suspect that if the normal ipodvideo builds didn't have the big disk stuff in, upgraded drives would mostly Just Work
15:52:39Torne(but then the 5.5G wouldn't work any more)
15:52:42Locke_FireclawAnd from what I could find out there's no way to just change that particular setting without going through the whole rebuild recompile business.
15:52:59TorneLocke_Fireclaw: yes, unfortuantely this is set at compile time because it affects how much memory is reserved for particular purposes
15:53:17linuxstbTorne: But this is a 80GB drive, not a "big disk". So I just wanted to ask the obvious question about the standard builds.
15:53:29 Quit kugel (Remote host closed the connection)
15:53:33Torneah, you also don't need LBA48 if it's an 80GB drive
15:53:34Locke_FireclawI can help with that one linuxstb
15:53:49TorneYou would probably be better off if you just undefined the sector size settings and left LBA48 out
15:53:52Locke_FireclawIt's a thin 80gb from a Classic, not the bulky drive from a regular Video.
15:53:56Tornethat will *probably* work on your drive.
15:54:01Torneand will be faster/better :)
15:54:35linuxstbLocke_Fireclaw: Yes, I know - you said that. What sector sizes does it use?
15:55:33Locke_FireclawI've no idea, but Rockbox threw a panic at me at stated that 4096 was an unsupported sector size.
15:55:39Locke_FireclawSo I changed it to that.
15:56:56Locke_FireclawOk, a little extra interesting stuff here, I just paused the song and I can crank up the volume all the way without it hard locking, and all the way down too.
15:57:09Locke_FireclawBut it will lock if the song is playing and I try the same thing.
15:57:21TorneWhen you say all the way up I assume you mean past 0dB?
15:57:39Tornecan you do it from the sound settings menu so you can see what numbers it locks at?
15:57:46Locke_FireclawWell, until the little bar goes red. lol But it does actually lock before then.
15:57:49Locke_FireclawSure, bare with.
15:57:49Torne(or use a WPS that displays the sound in decibels)
15:58:13Tornebecause between some low number and 0dB we are just adjusting the hardware volume setting
15:58:25Torneat very low volumes, or at volumes above 0dB, rockbox is having to actually adjust the digital audio itself
15:58:39Torneso if it's those levels where it goes wrong that's a big hint as to where the problem might be
15:59:04Locke_FireclawThat's interesting...
15:59:11linuxstbLocke_Fireclaw: Also, can you try and test other things? e.g. scrolling through long lists in the file/database browser whilst music is playing?
15:59:20Locke_FireclawUsing the volume menu in the sound goes right up, and right down, without crashing.
15:59:46Tornewhile it's playing?
15:59:48Tornethat's interesting
16:00:00Locke_FireclawIsn't it just! lol
16:00:03Torneso it's only if you are on the WPS *and* playing that it dies.
16:00:13Tornepaused, or not on the WPS, works fine?
16:00:27Locke_FireclawI only have a single album on here atm, so no long list to scroll through, I'm purely in testing stages.
16:00:31Locke_FireclawSeems to be.
16:00:34Locke_FireclawTheme problem maybe?
16:00:36TorneVery odd indeed
16:00:41TorneSet the theme to rockbox_default
16:00:43Torneand try it.
16:00:49Locke_FireclawIt's only cabbie, the default theme.
16:01:01Torneyah, but still. try rockbox_default, which is the compiled in theme
16:01:05Torne(veyr very ugly) :)
16:01:24Locke_FireclawOw, my ain't kiddin...
16:01:36Tornebut it has the advantage of not using any loaded files
16:01:43Tornewhich reduces the number of things that could be wrong :)
16:02:05Locke_FireclawYep, right up, right down, no problems.
16:02:06 Join n1s [0] (~n1s@rockbox/developer/n1s)
16:02:17Locke_FireclawExcept my ringing ears.
16:02:27Tornelinuxstb: is it really only the ipod 5.5G's disk that has the requirement to do physical sector aligned writes? have we found any other disks that actually *need* it?
16:02:55Tornelinuxstb: because it seems really unfortuante that people who upgrade drives on ipods almost universally run into this issue and it's basically caused by the 5.5G workaround
16:03:10linuxstbTorne: I think the Nano2G also needs it, but that's done by redefining SECTOR_SIZE
16:03:21Tornelinuxstb: yeah, nano2g works differently for good reason
16:03:41Tornebasically what i mean is, isn't there a way we can make these upgraded drives work without needing a recompiled build?
16:04:01Torneat the moment we assume that if MAX_PHYS_SECTOR_SIZE is defined, we *must* do the emulation ourselves
16:04:05 Quit flydutch (Ping timeout: 246 seconds)
16:04:06Locke_FireclawOh, and one more thing while your setting every problem I've ever had with anything right. lol
16:04:07Torneand thus if it's too small we fail because we can't
16:04:31Locke_FireclawAlbum art...can't seem to get it to display, no matter the picked theme. Am I just being dim?
16:04:35Tornebut if we don't really need to do emulation most of hte time this seems suboptimal..
16:04:43TorneLocke_Fireclaw: Where is your album art stored, and what format is it?
16:04:57linuxstbI would hope there's a better way. But I don't know enough about it. ramsize shouldn't really be an issue on those targets either, so we could sacrifice a little for convenience and simplicity.
16:05:15Locke_FireclawJPG, in the same folders as the music playing. I read that JPG wasn't supported, then read something else that said it was supported as of late 2009.
16:05:24Tornejpeg is supported now.
16:05:33Torneare they progressive jpegs, by any chance? we can't load those, iirc
16:06:07 Join flydutch [0] (
16:06:11Locke_FireclawThorne & linuxstb, are you two Rockbox devs perchance? Seem to know ya stuff...and be far more polite than most with newbs such as...well, me. lol
16:06:12Tornelinuxstb: well I would think all we'd need is some wya to detect the "broken" drives and set phys_sector_mult to 1 otherwise
16:06:24Tornelinuxstb: so the code to handle it would still be in there, with 1024 byte buffers
16:06:25Locke_FireclawI'll have a looksee.
16:06:29Tornelinuxstb: but it just wouldn't get used.
16:06:47Tornethis is laready what happens if you run it on a 5G whose disk has 512 byte sectors
16:06:49Locke_FireclawOh, if it helps, this is Jamendo album art...but I imagine that doesn't help at all. lol
16:06:53 Quit n1s (Ping timeout: 264 seconds)
16:07:09TorneLocke_Fireclaw: well, jpeg should work as long as it's not progressive. what are the jpeg files called?
16:07:19Tornefolder? cover? nameofalbum? :)
16:08:02Locke_FireclawAhh, there may be naming complications I guess...just a stab in the dark, but the current playing one is...
16:08:11Locke_Fireclaw[cover] Alexander Blu - May.jpg
16:08:58TorneIt needs ot be called either exactly the same thing as the album title, or folder.jpg, or cover.jpg
16:09:07Torneit doesn't just look for any images in the same folder
16:09:17Tornesee appendix C of the manual for the exact list of filenames we check
16:09:43Locke_FireclawAhh, so if I rename the file to folder.jpg it'll grab it?
16:09:46Tornelinuxstb: the hard part is deciding whether the drive needs the workaround, that's a lotta renaming. lol
16:09:57TorneLocke_Fireclaw: yes, and folder.jpg is also recognised by Windows and many other OSes, as a bonus
16:10:02Torne(that's why we support it)
16:10:04linuxstbTorne: Attempt a non-aligned write and see if it works?
16:10:05evilnickLocke_Fireclaw: There are renaming tools that will do that en masse
16:10:12Tornelinuxstb: I guess, but where can we write to safely?
16:10:32gevaertsIs it only write?
16:10:35Tornegevaerts: yes
16:10:39GodEatersomewhere in the MBR?
16:10:41Tornethe drive still emulates reads iirc
16:10:53GodEaterwait, hang on
16:10:54TorneGodEater: hehe
16:11:03GodEaterare we talking about the broken 80GB ipod drives?
16:11:05GodEateror something else?
16:11:16GodEatersingle sector reads fail on those
16:11:19GodEaterisn't that enough of a test?
16:11:24TorneOh, they do?
16:11:27TorneThat's much easier then
16:11:30GodEateryou get a DNF
16:11:35TorneWell, we could do that
16:11:50Tornei'm talking about the fact that the workaround for those drives stops us from using lots of other *non* broken drives, installed as upgrades
16:11:54GodEaterI told people not to archive that forum thread :)
16:11:59Tornebecause those drives claim a physical sector size greater than 1024
16:12:04Tornebut do emulate reads/writes happily.
16:12:14Tornei'd like to fix that i fpossible
16:12:17Locke_FireclawJust need to find a theme that supports art now. lol
16:12:20*Locke_Fireclaw fiddles
16:12:22TorneLocke_Fireclaw: cabbie does
16:12:23GodEaterI remember this well because I did all those early tests on the 80GB
16:12:30TorneLocke_Fireclaw: it rearranges the screen to make room if the art exists
16:12:34GodEaterI remember having to read the bloody ATA INFO off the LCD
16:12:36TorneGodEater: ah right
16:12:45TorneGodEater: i'm also wondering what the actual binsize impact of LBA48 is
16:12:53GodEaterlearnt more about ATA than I ever wanted to
16:12:55Tornebecause having that enabled all the time might be nice too
16:13:08GodEaterI Can't imagine it's that bad
16:13:15Torne(LBA48 does actually *kinda* give you an extra featur even for drives that don't need it: you can load more sectors in one request)
16:13:21Torne(which makes large DMAs quicker)
16:13:29GodEaterI'm all for that
16:13:37Tornethough we don't load enough at once in buffering for that to help
16:13:44Tornesince we only buffer multiples of the chunk size
16:13:49GodEateruseful to know for future reference though
16:13:52Tornemaking buffering load 1MB chunks means you waste at the end some :)
16:14:02Torne(unless you make even more invasive changes to buffering)
16:14:25Tornebut yes, if possible without breaking anything *else* and increasing binsize much I would like to make standard builds work with upgraded/replaced drives
16:14:35GodEaterbtw, if writes turn out to yes be required to detect this, I was serious about the MBR
16:14:39Torneso that we can eliminate these monster drive builds which rot and do things wrong.
16:14:47GodEaterjust attempt to re-write the magic numbers
16:14:53TorneWe could just write to sector 1
16:14:57GodEateror that
16:14:59Tornethe rest of the sectors in cylinder 0 are empty
16:15:10Torne(normally, at least)
16:15:17Torne(we might upset people with custom GRUB installs? :)
16:15:24GodEaterscrew them :)
16:15:25Locke_FireclawBut cabbie seems to die when I adjust volume Thorne.
16:15:33GodEaterLocke_Fireclaw: re-read his nick
16:15:36GodEaterthere's no "h"
16:15:37TorneLocke_Fireclaw: well, that's a bug in rockbox, not cabbie
16:15:48TorneLocke_Fireclaw: the theme should no tbe able to crash the player
16:15:56Tornebut yes, please do try other themes
16:15:57Locke_FireclawSorry, my bad. :/
16:16:00Torneand see if it's *just* cabbie that breaks it
16:16:03Torneor any graphical theme
16:16:08Torneas that's also useful data
16:16:15 Quit antil33t ()
16:16:26Torneit sounds like you've found a genuine bug in there somewhere, but it must be triggered by *something* obscure, or we'd've seen it :)
16:16:42Locke_FireclawEither that or my crappy recompile.
16:16:52TorneWell, hopefully not.
16:17:05TorneIf you want to eliminate that one, try just *removing* all of the defines mentioned in that forum post
16:17:11Locke_FireclawIn my defense I did follow to the letter.
16:17:21Tornedon't define physical/logical sector size maximums at all, and don't define LBA48
16:17:24Torneand build that.
16:17:32TorneIf that still boots and works at all then your drive doesn't need them
16:17:36Torneand will work better without them
16:18:36Locke_FireclawOk, how do I comment out again? with - +?
16:19:00Tornejust change them to #undef MAX_PHYS_SECTOR_SIZE
16:19:04Torneand similar
16:19:15Torne#undef tells it to un-define them instead of defining them as particular numbers.
16:19:18Locke_FireclawAhh, right, got it.
16:19:29Tornethis *might* not work. i've not tried it ;)
16:19:35Tornebut it should if my understanding is correct ;)
16:20:16TorneGodEater: i guess i should just write the code and post a test build :)
16:20:27GodEaterit shouldn't take long :)
16:20:32TorneGodEater: if reads are sufficient then it's trivial to just read sector 1 and see if it succeeds
16:20:43GodEaterI'm ready to guarantee it won't
16:20:49Torneand if so, ignore ATA IDENTIFY's sector size and just set the value to 1
16:21:08Tornethen post a test build and ask all the people with replaced/upgraded disks to try it
16:21:23Torne(i have a 5.5G with a stock disk to try on)
16:22:40Locke_FireclawTorne: I think I love you.
16:22:47Locke_FireclawWas has album art. :D
16:22:58Tornewell, next time read the manual and you'll have album art without help ;)
16:23:34Locke_FireclawBut alas, still a volume crash with a theme, not cabbie this time either.
16:23:48TorneRight, so that's useful information
16:24:05Tornenot sure quite how to reproduce htough :0
16:24:24GodEaterlinuxstb: apologies for missing your question earlier - did you find out who/what they/it /are/is ?
16:25:03Locke_FireclawI think it's only with extremes of volume, probably past 0db
16:25:04linuxstbGodEater: Sort of... It's an itunes-like app for WIndows and OS X?
16:25:13GodEaterpretty much
16:25:21GodEaterbut without actually having to have an ipod
16:25:25GodEaterit supports many devices
16:25:33TorneLocke_Fireclaw: right, but if it works when paused, or when not showing the wps, or when using the builtin theme, then it can't really be a bug in the code that adjusts the volume
16:25:39TorneIt must be a bug in the code that displays the WPS
16:25:43GodEaterbut Rockbox'd iPods in particular confuse it
16:28:02linuxstbGodEater: I had a quick look at the xml though, and some things seem odd. For example, you specify folder names for music, video, photos etc in there - so users don't seem to get a choice.
16:28:29GodEaterlinuxstb: yes, that annoyed me too :(
16:28:43GodEaterbut I guess for people who wanted to use it, they wouldn't really care
16:28:55GodEaterso we could specify "Rockbox_Music" or somesuch
16:30:28Tornelinuxstb: the assumption on the media players' part is that users *don't* get a choice, though
16:30:38Tornewhich is pretty standard :)
16:30:43GodEateryeah, it's pretty rare that you do
16:30:52GodEaterwe are one big edge case
16:31:12Tornethey also assume, reasonably, that players have a fixed set of formats they support which is unlikely to change :0
16:31:40GodEaterthat disappointed me a whole lot more - I thought dT supported more formats than it turns out it does
16:32:39linuxstbI don't have much experience of OFs, but I can't remember file location restrictions, apart from ipods.
16:33:46 Quit Luca_S (Quit: CGI:IRC)
16:34:00Tornelinuxstb: it tends to be more recent players that do it
16:34:12Tornethe ones that play video and stuff as well, the chinese ones with big-ish screens
16:34:46Tornetheir terrible OFs tend to have segregated modes for video/music/picture/etc
16:34:52Tornewhich look in different places :(
16:36:15BagderI also think both cowon and sandisk OFs have had such ideas
16:36:21Locke_FireclawHm, something more than volume problems methinks. :(
16:37:01Locke_FireclawAh well, back to the drawing board.
16:37:20 Join toffe82 [0] (~chatzilla@
16:40:08Locke_FireclawBBL guys, freezing my fanny off out here, that's what you get for having a shed as an office I guess.
16:40:57Locke_FireclawThanks again for all your help, it's much appreciated.
17:00:09S_a_i_n_tBeen scrolling up a little...anyone able to shed any light on the "changing volume in the WPS hardlocks the player" thing?
17:00:16S_a_i_n_t...seems *very* odd.
17:00:25***Saving seen data "./dancer.seen"
17:02:19TorneS_a_i_n_t: yeah we have no idea
17:02:28Tornebut from the sound of his last comment he has found other problems too
17:02:32Torneso maybe there is a more general issue with his player
17:02:34 Join froggyman [0] (~me@unaffiliated/froggyman)
17:02:46S_a_i_n_tI can't reproduce, *at all* if it helps.
17:05:04 Join Luca_S [0] (
17:05:07 Quit TillW (Ping timeout: 248 seconds)
17:06:03Tornewell no, not relaly
17:06:23Torneif the problem wasn't conditional on some unknown probably obscure thing then someone would've noticed by now already :)
17:06:45Tornethe fact that a user even has this issue without us knowing why already suggests it's based on something pretty obscure ;)
17:07:49S_a_i_n_tMy immediate thought is he's pissed around with something he forgot about, seeing an "svn diff" might of been a good thing to check.
17:08:16*Torne shrugs
17:08:22Tornewe'll see I guess :)
17:08:38Torneit has reminded me that I want to fix it so people don't *have* to do custom builds when they use nonstandard drives.
17:08:45Tornewhich would eliminate that as a possible cause ;)
17:09:01Torneso maybe i'll test htat later
17:09:21S_a_i_n_tyes, wouldn;t that require some pretty interesting ifdef's though?
17:10:19S_a_i_n_tI always assumed there was some level of fuckery that prevented "bigdisk" builds being offered.
17:10:43Torneunless i misunderstand entirely it's not much of a problem at all
17:11:00Tornethe problem is actually that the workaround to make the 5.5G ipod disk work *breaks* support for loads of otherwise perfectly good disks
17:11:06Tornewhich would've worked fine without 5.5G support
17:11:14S_a_i_n_tthe 5.5G bigdisk fiends will be pleased... ;)
17:11:28Tornewe just need to test the disk to make sure it really can do unaligned reads/writes
17:11:35Torneand if so we don't *need* to do the workaround.
17:11:45Torneand thus it doesn't matter if the allocated buffer for the workaround is too small for this disk.
17:12:05TorneThis will then work for the 5.5G disk and also for any old non-broken disk that emulates single sector accesses correctly
17:12:09S_a_i_n_twell, the theory seems simple enough.
17:12:26Torneit would only fail if someone has a disk which doesn't emulate single sector accesses *and* has a sector size >1024
17:12:40Torneso, yeah, i may try this out later
17:13:28Tornethen try enabling lba48 as well, since i don't think the binsize impact should be very much
17:13:40Torneand then the bigdisk builds can die and there will be a lot less people using unsupported builds ;)
17:14:03gevaertsTorne: make sure you measure the boot time impact of all this! ;)
17:14:11Tornegevaerts: Conveniently.. :)
17:14:17 Join TillW [0] (
17:14:51 Quit toffe82 (Remote host closed the connection)
17:17:56 Part patgodo
17:18:00 Join toffe82 [0] (~chatzilla@
17:18:17Tornegevaerts: do you have an opinion on committing boot charting code to svn, btw?
17:18:22Torneas a default-off #define
17:18:30gevaertsTorne: I think it'd be useful
17:18:46Tornealso I am currently doing it by calling _logf
17:18:49gevaertsMaybe just have it depend on LOGF_ENABLE
17:18:50 Join katyl [0] (
17:18:52Torneso that you don't have to define LOGF_ENABLE in each file
17:18:59gevaertshm, right
17:19:09Tornethe chart crosses several files
17:19:16Torneso, i wanted to have one global define to turn it on
17:24:54 Join Tomis2 [0] (~Tomis@
17:25:14 Join archivator [0] (~archivato@
17:25:56 Quit archivator (Remote host closed the connection)
17:26:02 Part LinusN
17:27:17 Quit Locke_Fireclaw (Quit: ChatZilla 0.9.86 [Firefox 3.5.8/20100214235958])
17:27:31 Quit Tomis (Ping timeout: 264 seconds)
17:27:31 Nick Tomis2 is now known as Tomis (~Tomis@
17:30:34 Quit Hillshum (Ping timeout: 246 seconds)
17:33:59Tornegevaerts: expected use: CHART_CALL(tagcache_init()); or CHART_STR("boot done");
17:34:49Tornewe might want __LINE__ in there as well?
17:34:57gevaertsTorne: I guess we can't use fancy linker tricks? :)
17:35:04Tornegevaerts: I thought of that
17:35:08Torneit's a bit tricky
17:35:12Torneand not all the calls are between files ;)
17:35:48 Join archivator [0] (~archivato@
17:36:28gevaertsline numbers might be good. I suspect there's no robust way to make this less visible
17:36:39TorneYeah, it's very hard to not have it intrude on the source
17:36:44TorneI wasn't going to put one on *everything*
17:36:48Torneonly the likely culprits
17:36:55Tornebecause the times *between* them still tell you about everything inbetween
17:37:04gevaertsYou could play games with #defines to redefine the function calls to some wrapper, but that's going to cause headaches
17:37:38TorneCHART_CALL to me seems less intrusive than spelling out function names again
17:37:48TorneCHART("tagcache_init"); tagcache_init();
17:37:55Torneis even *more* lines :)
17:38:06gevaertsYes, that's really bad and errorprone
17:38:20Torneat least with doing CHART_CALL you can only accidentally omit it
17:38:31Tornewhich as i say doesn't destroy the usefulness of the data, just the resolutoin
17:39:08Tornethe code i was using last night *only* used __LINE__ which is less intrusive but make sit really hard to interpret and compare results
17:39:13gevaertsWhat I was briefly considering was #define tagcache_init() _tagcache_init() and then CHART_CALL(tagcache_init());
17:39:24TorneThat's a huge pain in the ass as well though
17:39:38Torneyou have to go and redefine it in the places where the functions are *implemented* then
17:39:46TorneI want to keep most of the charting in places like main.c, settings.c
17:40:01Tornenot in the individual iplementations unless someone feels it's needed in that level of detail for a particular thing
17:40:23Tornethe BC: prefix is so you can filter it from other logf's
17:40:43Torneand comma so you can treat it as CSV and load it into a spreadsheet :)
17:49:25 Quit Bagder (Quit: It is time to say moo)
17:59:00Tornehm, CHART_CALL is ugly for the lines that do, say, rc = disk_mount_all();
17:59:36 Quit toffe82 (Remote host closed the connection)
18:00:46 Quit bluebrother (Disconnected by services)
18:00:46 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
18:01:16 Quit pamaury (Quit: Quitte)
18:01:23 Quit petur (Quit: *plop*)
18:05:21 Join komputes [0] (~komputes@ubuntu/member/komputes)
18:09:46linuxstbTorne: IMO, CHART_CALL() is ugly everywhere. I know it's harder to write, but I would find the CHART_STR() syntax far easier on the eyes when reading the code. It's more obvious what is happening, without knowing what those macros do.
18:10:39linuxstbBut that's just me...
18:10:49Tornewell, i am not exactly supe rhappy with it :)
18:11:16Tornei'd like to do it with metaprogramming but C doesn't have any worth the name ;)
18:12:04linuxstbWhat about simply adding CHART_STR() inside the function itself? But I guess that won't work if the functions have many return points...
18:12:26Torneit also means they keep getting issued, for functoins which are used after boot
18:13:00Tornewell, i am not proposing we add a lot of these. :)
18:13:08Torneright now I have, hmm..
18:13:56Torne20 in main.c, 9 in settings.c, and two in gui/skin_engine/theme_settings.c
18:14:10Tornewhich as far as i can see covers bascially everything that hits the disk or has an involved parse/setup step
18:14:31Torneup to the point where we call root_menu() which I am currently considering to be "the end of booting"
18:15:02Torneso yeah, it could be done with explicit strings seperate from the calls
18:15:22Tornei am happy to go whichever way people think best
18:15:38Tornebut at least with CHART_CALL the messages are guaranteed to be accurate..
18:21:10 Join phanboy4 [0] (
18:21:56 Join Hillshum [0] (
18:22:19 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:31:52Tornegevaerts: linuxstb: <- proposed patch
18:35:13gevaertsTorne: looks "good". I guess it will need some more discussion...
18:35:33TorneYah, i haven't actually tested it, either ;)
18:35:51Torneit compiles. i'll try it when i get home and make sure that 1) it works 2) the gaps between the chart points are small
18:36:43Tornei guess whether to keep CHART_CALL is the major thing
18:36:53Tornei can be convinced either way
18:37:07 Join pamaury [0] (~c2c7a50a@rockbox/developer/pamaury)
18:45:02 Join toffe82 [0] (~chatzilla@
18:46:42 Quit Tomis (Read error: Connection reset by peer)
18:46:55 Quit Luca_S (Quit: CGI:IRC (EOF))
18:46:59 Join Tomis [0] (~Tomis@
18:50:44 Nick fxb is now known as fxb__ (
18:51:42 Quit pamaury (Quit: Page closed)
19:00:08 Quit shaggy-h (Ping timeout: 240 seconds)
19:00:27***Saving seen data "./dancer.seen"
19:00:31 Join watto [0] (~watto@
19:01:28 Join arjunvj3 [0] (~arjunvj3@
19:09:35 Join mitk [0] (
19:11:52linuxstbTorne: I don't like it, but I also can't suggest anything better. And the feature itself is useful...
19:11:56*linuxstb decides to sit on the fence
19:13:47 Join bertrik [0] (
19:13:50 Quit bertrik (Changing host)
19:13:51 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:20:29 Join Luca_S [0] (
19:25:13 Quit kaniini (Remote host closed the connection)
19:26:07 Quit shai (Quit: Leaving)
19:28:46 Join kio [0] (~kio@
19:29:14 Quit arun (Ping timeout: 276 seconds)
19:30:12Tornelinuxstb: well you could convince me to explicitly do CHART_STR before and after each call. for 28 calls I can be bothered to do that if needed :0
19:32:58 Join Strife89 [0] (~michael@
19:33:16 Quit phanboy4 (Read error: Connection reset by peer)
19:36:41 Join planetbeing_ [0] (
19:38:00 Quit junkY_San (Quit: Verlassend)
19:38:33 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
19:40:39 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
19:42:00 Quit Strife89 (Read error: Connection reset by peer)
19:48:00 Join shai [0] (
19:48:18 Join Strife89 [0] (~michael@
19:48:57shaiHi :) So is it correct that rockbox won't work on iPod Classic 7th gen?
19:49:53gevaertsthat's extremely correct
19:50:08shaiYou recall my issue with the iPod yesterday?
19:50:24shaiWell, I don't think that my issues ended with just re-installing RB
19:50:38shaiSo I'm considering getting a new iPod... but I shouldn't get something very new then?
19:50:50shaiTouch doesn't work, new gens of Classic won't work...
19:51:01shaiSo I should only get old gens on eBay?
19:51:08gevaertsI'd recommend first reformatting the disk, and if that doesn't help just replacing the disk
19:51:31shaigevaerts, I might even have a battery issue...
19:51:55gevaertsWell, the list of supported players is in a prominent spot on our website
19:52:37shaiWhen I turn it off, for a min (or so), I see a black screen with a battery icon in the middle, with a triangle and an exclamation mark.
19:52:53gevaertsthat's expected
19:53:17shaiWhat is? The batt. issue or the black screen with the icons I mentioned?
19:53:35 Quit TillW (Quit: TillW)
19:53:43shaiAgain, "yes" to which one?
19:54:10gevaertsthe battery icon. Do you have other battery issues?
19:54:25shaiI don't think so...
19:54:29 Join arun [0] (~arun@unaffiliated/sindian)
19:54:41shaiBut I don't recall ever seeing that when I power off the iPod ...
19:54:53gevaertsIt's a side effect of an attempted solution of a problem that makes some ipods not boot every now and then
19:55:21shaiThat icon is a side effect?
19:55:40shaiThat sounds so strange :)
19:56:10S_a_i_n_tI have an .sbs fail...when I load an .sbs I've just made, it shows the "rewind" icon for the playmode. I'll pastebin the .sbs, but from what I can see. the .sbs is fine and this is some odd bug.
19:56:27 Join Stephen__ [0] (~S@
19:56:31gevaertsshai: I know. It's a fact nevertheless...
19:56:32S_a_i_n_tshai: there's a *possible* fix for that issue being tested now.
19:56:44S_a_i_n_tbut the "issue" is purely cosmetic.
19:57:03shaiS_a_i_n_t, odd ... who's working on that solution you speak of, Apple?
19:57:19S_a_i_n_tApple has *NOTHING* to do with this...
19:57:27shaiOh !?
19:57:35gevaertsshai: well, yes and no. There's been some weird things reported since that fix went in, so there might be more going on
19:57:50gevaertsS_a_i_n_t: that was for you...
19:58:02shaiSo you're saying that icon when I power it off, is from RB and not the iPod itself?
19:58:13S_a_i_n_tshai: yes.
19:58:23shaiYou guys need to get on the same page :)
19:58:40gevaertsThe icon is from the OF, but it's triggered by rockbox behaviour
19:58:58shaiWhat is OF short for?
19:59:00S_a_i_n_tI was under the understanding it was from the "shutdown "fix"", am I wrong?
19:59:16gevaertsOriginal Firmware
19:59:31shaiSo this is cosmetic. I can live with that for now.
19:59:33gevaertsS_a_i_n_t: the "fix" is to invoke some OF code at shutdown time
19:59:49shaiSo you think I should just Format the FS again and if all else fails, replace the HDD?
20:00:12shaiI've already opened the iPod today to try and see if there is an issue with the connection...
20:00:17shaiFor the bat.
20:00:22gevaertsYes. Possibly you can use the diagnostic mode to do a disk check. Don't ask me how though...
20:00:23shaiAnd there was no issue.
20:01:03shaiThing is, that after I left yesterday, and started driving to meet up with my friends, after you, gevaerts, helped me get it to load the FW, it shut off the iPod every min or two after playing the new songs.
20:01:20shaiSo I think (not sure), that I still have issues, but not sure where they lay yet.
20:01:34shaiCould be that the FS is screwed, and that I really do need to format it and start from scratch...
20:01:36shaiNot sure...
20:01:52S_a_i_n_twhat type of iPod is this?
20:01:56shaiIs it as easy as formatting drive H:\ and re-installing RB?
20:02:05shaiiPod Classic 5th gen.
20:02:22gevaertsIt should be, yes. Remember to format it as FAT32, not NTFS
20:02:33shaiOf course....
20:05:15 Quit mitk (Quit: Leaving)
20:07:37S_a_i_n_tCan anyone tell me WTF "rewind" is firing when I boot the iPod (well, the sim) with this .sbs loaded?
20:07:51amiconn[05:24:27] <saratoga_lab> "54MHz arm7tdm" <== The av300 is hybrid hwcodec+swcodec, and iirc there's also some gfx processing hardware in the lcd controller
20:07:53S_a_i_n_tI can't see *how* it is...unless I'm *really* blind.
20:09:30amiconnSo the thing should at least manage mp2/mp3, all the various pcm formats, flac and shorten
20:11:13 Join stripwax [0] (
20:11:21 Quit stripwax (Client Quit)
20:14:54 Join FlynDice_ [0] (
20:14:55 Quit FlynDice (Ping timeout: 264 seconds)
20:16:43S_a_i_n_tIf an .sbs is loaded, and the player is not playing (so isn't "stopped" yet at least), or seeking (which you can't do in the menu anyway AFAIK), then "Playmode" icon should be blank. Correct?
20:17:56gevaertsShouldn't it be "stopped" then?
20:18:34S_a_i_n_tHmmm..eitther way, "rewind" is firing...and (pastebin a few lines up) I can't see why
20:18:46S_a_i_n_tthe same playmode code works fine in the WPS
20:19:46S_a_i_n_tthe "stopped" icon is currently all "magic" colour, so I shouldn;t see *anything* if its stopped.
20:20:30 Quit FlynDice_ (Read error: Operation timed out)
20:21:21 Join FlynDice_ [0] (
20:22:53saratogaamiconn: iirc its a ti dsp like on the mrobe, so we could write code for it
20:23:09saratogabut i think that would be fairly difficult compared to carefully optimizing for the main arm core
20:25:52amiconnsaratoga: It's a MAS3587F, the same as in the recorders and the Ondio FM
20:26:30saratogaah i must be mistakebn
20:26:38amiconnThe av series is the reason why archos and micronas developed the pcm "codec" for the mas
20:26:40saratogathought it was a TI core
20:31:22 Join shaggy-h [0] (
20:31:42 Join anewuser [0] (anewuser@unaffiliated/anewuser)
20:32:15 Join stooo [0] (
20:32:21 Quit stooo (Client Quit)
20:34:41shaigevaerts, After yesterday, after I've moved the .rockbox to .rockbox.old and re-installed RB, what files should I fetch from that old dir. to keep some of my old settings that I had?
20:35:14gevaertsconfig.cfg mainly
20:36:31shaikio, 10x
20:36:52 Quit Hillshum (Ping timeout: 268 seconds)
20:38:36 Part watto
20:40:31S_a_i_n_tgevaerts: I found the problem...
20:40:55S_a_i_n_tthe .sbs needs to define (even if not displayed) all 9 playmode types
20:41:07S_a_i_n_twhereas the WPS does not have to do this.
20:41:37S_a_i_n_tand I was right, booting the player isn't "stopped"
20:41:58S_a_i_n_tit's not "stopped" until after it's played something.
20:42:12 Join webguest79 [0] (
20:42:17amiconnTorne: The correct fix for the large physical sector issue with some hdds would be probing. But implementing that needs access to such a hdd
20:42:34 Quit webguest79 (Client Quit)
20:43:15amiconnAs GodEater meantioned, you get IDNF - but iirc there is a problem in recovering from this state
20:43:47gevaertsamiconn: IIRC Torne has an 80gb ipod
20:45:19*pixelma wonders where Nico_P has gone
20:46:36*S_a_i_n_t *now* realises why past .SBSs have failed...why did I have to figure it out on *this* friggin' one?!?
20:47:05S_a_i_n_tthat playmode thing should *really* be in the manual, but then, there's no place to put it.
20:47:28S_a_i_n_tAs there is no, .sbs its assumed they all work in with the WPS tags.
20:49:55 Quit ender (Quit: The difference between 'involvement' and 'commitment' is like an eggs-and-ham breakfast: the chicken was 'involved' - the pig was 'committed'.)
20:50:30S_a_i_n_tbut I fear that "Note: Remember you have to define all nine playmode types, even if its just with "-", and even if you don't use them all...Otherwise your .SBS will fail. We won't show you all nine types in the Manual though, Oh No! That'd be too easy, go looking in "Custom WPS" to find out what you're doing wrong" is a little long winded...
20:50:49 Join captainkewlllll [0] (~2669ecc2@gateway/web/freenode/x-mlltdnlglynjchbk)
20:51:35pixelmaI'd say that should be fixed rather
20:52:42S_a_i_n_tIt does seem odd, but whats easier to fix? make the WPS need all 9 types to be defined, or make the SBS deal OK with 5 like thw WPS does?
20:53:22 Join mitk [0] (
20:53:49pixelmaas conditional/enumeration tags usually should work even if you don't define all cases, I'd expect the latter
20:54:11S_a_i_n_tA job for JdGordon then I suspect.
20:54:30S_a_i_n_tHe'll probably go insane reading the logs, I may have to wait fot him in person.
20:55:08gevaertsYes, and after he's fixed the manual, I think he's also the right person to fix the code
20:55:19 Join Barahir [0] (
20:55:22pixelmafor other enums that means "take the last defined and use it for all missing conditions" IIRC. I'm sure the codec type used conditionally works this way
20:56:21S_a_i_n_tWhat was happening was (in the .sbs) if I just had %mp<stop|play|pause|fastforward|rewind> as my playmode line, it was showing "rewind" at boot-time
20:56:22 Join ender` [0] (
20:56:58S_a_i_n_tI fixed it by adding four more "-" conditions to the end.
20:57:06S_a_i_n_tA weird fix, but it worked.
20:57:54S_a_i_n_tI can't find anywhere where it states that *all* playmode states need to be conditions in the .sbs
20:58:09S_a_i_n_ts/to be/to have/
20:58:31 Quit Barahir_ (Ping timeout: 276 seconds)
20:58:33pixelmawhat's the other modes after rewind (in the sbs)?
20:58:48gevaertsRecording, Recording paused, FM Radio playing, FM Radio muted
20:59:12 Quit ender` (Client Quit)
20:59:14S_a_i_n_tnone of which I use, and the WPS doesn't use...
20:59:59S_a_i_n_tSo, I had *no* idea that they needed to be defined in the .sbs playmode conditional.
21:00:17S_a_i_n_tIdeally, I guess "playmode" should get its own tag in the .sbs
21:00:22pixelmathough it could be that the conditional works as expected but it doesn't get correct info which mode it's in until something was played
21:00:29***Saving seen data "./dancer.seen"
21:00:54 Quit Barahir (Ping timeout: 246 seconds)
21:01:08 Quit robin0800 (Remote host closed the connection)
21:01:18S_a_i_n_tpixelma: it was returning to "rewind" instead of "stopped" even *after* something was played/stopped
21:01:19gevaertsS_a_i_n_t: please report a bug about this
21:01:29S_a_i_n_tI thought of that too, but disprooved it.
21:01:33 Join robin0800 [0] (
21:01:56S_a_i_n_tgevaerts: OK, will do.
21:04:46pixelmathen it possibly doesn't have correct info in the menu at all (the rewind could also show up if it put non-coder like "thinks" it's in the other modes too
21:05:46 Join ender` [0] (
21:05:54S_a_i_n_tthen...why would defining all 9 states make it work perfectly as expected?
21:06:28 Join pamaury [0] (~c2c7a50a@rockbox/developer/pamaury)
21:07:39 Quit robin0800 (Read error: Connection reset by peer)
21:07:48gevaertsThe wiki is wrong as well. If you have a tuner but no recording, the FM tokens will move :\
21:07:48*S_a_i_n_t thinks its just "lack of general knowledge/available information about how .SBS syntax works"
21:08:03 Join robin0800 [0] (
21:08:22S_a_i_n_tgevaerts: Same , or seperate bug?
21:08:43S_a_i_n_tor just, token parsing weirdness for .sbs?
21:08:46gevaertsWell, that's only a bug if you want WPSes to be reusable on different targets...
21:09:02gevaertsah, no, I'm wrong
21:10:58 Quit adnyxo (Ping timeout: 248 seconds)
21:11:56gevaertsIt sticks on rewind, right?
21:13:16gevaertsS_a_i_n_t: can you do a quick test and use %mp as a non-conditional? It should just give you a number then
21:14:15pixelmayes, but if you don't define something for the following states it'll take the last defined which is "rewind"
21:14:15S_a_i_n_tgevaerts: yes.
21:14:23gevaertsWell, which one?
21:14:33S_a_i_n_tI mean, it displays << at boot-time
21:14:39gevaertsAh, ok :)
21:14:47S_a_i_n_twhat do you want me to change the %mp line to?
21:14:59pixelmajust %mp
21:15:08gevaertsJust %mp instead of %?mp<...>
21:15:10S_a_i_n_tone sec
21:17:23 Join webguest980 [0] (
21:17:46S_a_i_n_tIts REALLY hard to see, but I *think* it displays a "7"
21:17:52 Quit webguest980 (Client Quit)
21:18:32gevaertsWell, that's pretty different from 5 anyway
21:18:51evilnickS_a_i_n_t: Can you do a screendump of it?
21:19:13gevaertsSo it thinks it's in record pause
21:19:14S_a_i_n_tcorrection, it *is* a 7
21:19:45S_a_i_n_tdo you want a dump of it showing 7 at boot-time, or the |<< icon?
21:19:46*archivator thinks he managed to persuade rbutil to encode files in parallel (testing atm)
21:19:47S_a_i_n_tor both?
21:20:06*gevaerts believes S_a_i_n_t on his word :)
21:20:25S_a_i_n_tI know! Amazing, right? ;)
21:22:04*gevaerts finds a place where firmware/ calls apps/
21:22:14S_a_i_n_tHmmm...if I hold "play" on the sim, it goes from 1 (after playing, of course) to 7
21:22:19pixelmaas I said that would be expected if you only have 5 icons
21:23:27pixelmahmm... no
21:24:27 Join adnyxo [0] (
21:24:43 Join ender [0] (
21:24:49pixelmabut it is "allowed" to display the last - rewind - one as the one for mode 6, 7, 8 and 9 then
21:26:32gevaertsS_a_i_n_t: is this on the sim, or also on target?
21:26:59 Quit S_a_i_n_t (Ping timeout: 265 seconds)
21:27:59gevaertshm, in the sim this Can Not Happen!
21:28:10 Quit einhirn (Ping timeout: 258 seconds)
21:28:28 Quit ender` (Ping timeout: 276 seconds)
21:29:19 Join spaax [0] (
21:30:03 Quit ender (Ping timeout: 260 seconds)
21:30:22 Quit mitk (Quit: Leaving)
21:32:09 Join mitk [0] (
21:32:30 Join einhirn [0] (
21:33:48 Join ender` [0] (
21:33:48 Quit robin0800 (Remote host closed the connection)
21:34:45gevaertsok, 7 is not record pause, it's radio...
21:38:27gevaertsThe bug was introduced in r25403
21:41:28pixelmawould that affect the inbuilt staatusbar too - or does it?
21:42:34gevaertsno idea. I haven't looked there yet
21:43:46*gevaerts wants S_a_i_n_t to come back to answer questions
21:47:29 Quit arbingordon (Quit: `)
21:47:33gevaertsI can see how this could happen in the sim
21:49:56gevaertsthe radio should go back to OFF state before the sbs is initialised, that's done in radio_init(). That one isn't called in the sim startup however
21:50:23 Quit ender` (Quit: I'm a complex person. I have a real and an imaginary part.)
21:51:35 Join arbingordon [0] (~w@unaffiliated/arbingordon)
21:51:46 Quit jgarvey (Quit: Leaving)
21:53:54 Part spaax
21:55:40 Quit arbingordon (Client Quit)
21:56:23 Quit bmbl (Quit: Bye!)
21:59:52CIA-5New commit by 03gevaerts (r25408): Fix FS #11152 in a cleaner way. At least on the sim, the state remained in FMRADIO_PLAYING forever after r25403, which broke (at least) the playback ...
22:00:30 Join arbingordon [0] (~w@unaffiliated/arbingordon)
22:04:59 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
22:07:23 Join jeffp [0] (
22:10:46 Quit Zagor (Quit: Clint excited)
22:11:14CIA-5New commit by 03tomers (r25409): rbutil: Make strings translatable
22:11:46CIA-5New commit by 03tomers (r25410): rbutil: Hebrew translation
22:12:18 Join tomers [0] (
22:15:55 Join S_a_i_n_t [0] (S_a_i_n_t@
22:16:23gevaertsS_a_i_n_t: was this in the sim or on target?
22:17:11gevaertsok. Can you check if r25408 fixes it?
22:17:23 Join Hillshum [0] (
22:17:29S_a_i_n_tsorry for going awol like that, see #rb-community
22:17:38S_a_i_n_tand yes, will do.
22:19:40 Quit froggyman (Ping timeout: 265 seconds)
22:22:22 Quit amiconn (Disconnected by services)
22:22:24 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
22:22:46 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
22:23:15 Quit tomers (Quit: ChatZilla 0.9.86 [Firefox 3.5.8/20100214235838])
22:23:27 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
22:23:28 Quit pixelma (Disconnected by services)
22:23:47 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
22:25:08 Join froggyman [0] (~me@unaffiliated/froggyman)
22:30:04 Join ender` [0] (
22:31:58archivatordomonoky: I managed to port the encoding part to QtConcurrent! :) It's not pretty but it works
22:32:28archivatorAnd the ugliness is mainly due to the fact that I can't use boost::bind (since we don't use boost!)
22:32:49domonokycan you show patch ?
22:32:58archivatorneed to clean it up first
22:33:05archivatorgive me 15 mins
22:35:29 Join Schmo [0] (
22:36:35 Quit Schmogel (Ping timeout: 265 seconds)
22:52:10CIA-5New commit by 03bluebrother (r25411): Add build folder cleanup on build errors and add option to keep it.
22:53:07bluebroth3rdomonoky: any objections to add the beast as disabled target (no bootloader installation yet)?
22:53:32domonokybluebroth3r: sure disabled target is fine for the beast.
22:53:43 Quit Hillshum (Ping timeout: 245 seconds)
22:54:08CIA-5New commit by 03bluebrother (r25412): Add Gigabeat S to Rockbox Utility. ...
22:54:13 Quit efyx (Remote host closed the connection)
22:54:17bluebroth3rgreat, so we have the same understanding about disabled targets :)
22:54:26 Quit Strife89 (Quit: Going home.)
22:58:41S_a_i_n_tgevaerts: Sorry, there was a fair bit of pissing-around (lets call it "testing") mixed up in there, but I can confirm that r25408 fixes the .sbs playmode icon issue I was having.
22:59:04S_a_i_n_twhat *was* the problem?
22:59:33gevaertsWell, I'd say the diff explains it better than I could :)
23:00:06 Quit mitk (Quit: Leaving)
23:00:31***Saving seen data "./dancer.seen"
23:01:04gevaertsBasically since 25403, the radio status was set to "playing" on boot, so the radio poweroff during init would work (it checked for "is it already off?"). This is all done before any wps/sbs code runs, so all would be fine
23:01:24gevaertsExcept that this init code isn't run on the sim, so the status remained "playing"
23:02:29gevaertsI've changed that to not check the state at init time, so initialising it to "playing" isn't needed anymore
23:02:34S_a_i_n_tyes, I was just checking out svnsearch...but thanks for the better description than I could possibly parse from the code ;)
23:03:32archivatormy vim config is messed up at the moment, so excuse the tabs
23:05:51*bluebroth3r wonders if there's a way to have a modeline for vim for a folder
23:07:50archivatordomonoky: here's one with the spacing fixed and prettier debug messages:
23:10:42 Quit pamaury (Quit: Page closed)
23:11:22 Quit Luca_S (Quit: CGI:IRC)
23:11:25bluebroth3rhmm, calling an auto variable m_something somewhat defeats the puropose of the m_ prefix :/
23:11:36 Join Luca_S [0] (
23:12:29archivatorbluebroth3r: k it does but I don't use m_ altogether. I find it kinda ugly and meaningless..
23:12:53linuxstbsaratoga: You were right - the AV300 does have a TI CPU (TMS320DSC25 - arm7tdmi plus DSP), but as amiconn said, it also has the MAS.
23:12:59bluebroth3rarchivator: you actually remove that use :)
23:13:00domonokyarchivator: looks fine on the first look. Instead of putting the pointers into TalkEntry, you could also jump from the static member to a normal member function via a static TalkGenerator pointer. But dont know if that is better :-)
23:14:42archivatordomonoky: I was thinking of making TalkGenerator a singleton but seems unnecessary for something that can be fixed with 8 (12 in future) additional bytes :)
23:15:39archivatorbluebroth3r: I blame years of wxWidgets. At some point you just get so tired of it, you want to erase it from the face of the planet :)
23:16:38domonokyarchivator: so how much speed improvement does that give for you ?
23:17:20archivatordomonoky: haven't done proper testing but it does max out both cores and is noticeably faster
23:17:31 Nick fxb__ is now known as fxb (
23:17:42bluebroth3rarchivator: I'm not entirely sure if I like those prefixes or not. There is some use in it, but people tend to want hungarian notation then (which is completely crap and nonsense IMO). So I'm somewhat split on the topic. I try to be consistent within classes :)
23:17:48archivatorhowever, bear in mind that the voicing process is more wasteful and is where most resources are wasted (or rather, left unused)
23:18:10bluebroth3ron my mac encoding takes longer than voicing
23:18:28bluebroth3rat least judging from the progress bar :)
23:18:49domonokysounds good, sure the voicing could need improvements too, but lets do one at the time :-)
23:19:39domonokyand for the voices it more difficult to make sure all engines still work..
23:19:48archivatorSpeaking of Mac, domonoky, do you think TTSCarbon can be parallelized in the same way or do we need to add a special method to check if the engine can run in parallel?
23:19:55domonokydid you try what lame does with this paralell encoding ?
23:20:20archivatorWell, TTSExes and Festival are straight-forward. TTSCarbon and SAPI is where the trouble's at.
23:20:58archivatorno, I don't use lame. No reason why it shouldn't work - it's just a QProcess wrapper anyway
23:21:10domonokyarchivator: i am not sure, ask bluebrother for TTScarbon :-)
23:21:54archivatorsorry, thought you wrote it. bluebroth3r: you heard the question :)
23:22:38domonokyfor sapi i am also not sure, we just use the sapi_voice.vbs script from rockboxs buildsystem :-)
23:23:19archivatorYeah, I think that one will work. Carbon is more interesting though
23:23:19bluebroth3rarchivator: I don't see a reason why it shouldn't work. It's creating a wav file at a time. I can simply try it ;-)
23:23:21 Quit anewuser ()
23:23:43bluebroth3r(need to give the diff a closer look to judge from that but I'm a bit too tired right now)
23:23:58archivatorbluebroth3r: well, is voice() reentrant for one?
23:24:28 Join Hillshum [0] (
23:24:42 Join Strife89 [0] (
23:26:00bluebroth3rarchivator: the apple tts can't handle multiple SpeakText() calls for the same voice channel
23:26:05 Join JohannesSM64 [0] (
23:26:16archivatorbluebroth3r: well, that solves it then
23:26:20bluebroth3rit should work if you create multiple TTSCarbon objects
23:26:28bluebroth3rhaven't tried though.
23:26:44archivatorYeah but that's not a universal solution..
23:27:24archivatorThat would mean 2 server processes for festival (initialization is slow) when 1 can serve 2 clients easily
23:28:25domonokyunless TTSCarbon would handle that internally. Ie create a second voice channel, if current already in use.
23:28:26*bluebroth3r rechecks with apple docs
23:28:34archivatorAlso, there's no way to maintain thread context using QtConcurrent::map(). For all I know, the thread is destroyed every time the entry point returns.
23:28:45 Join blairb [0] (
23:29:27archivatordomonoky: how about a tts->parallelizeMode() that's either Full or Voice ?
23:29:43archivatorAnd then handle the two different cases. Kinda ugly but it would work.
23:29:54bluebroth3rIf SpeakText is called while the channel is currently busy speaking the contents of a prior text buffer, it immediately stops speaking from the prior buffer and begins speaking from the new text buffer as soon as possible.
23:29:59archivatorOr even easier, tts->canParallelize() and just don't bother with TTSCarbon :)
23:30:12bluebroth3rok, so it's definitely not possible without using different channels.
23:30:12domonokythat would be better.
23:30:39domonokybluebroth3r: could you test if the channel is busy, and create a new one if needed ?
23:33:28bluebroth3ryes, there's a SpeechBusy() call. Would need quite a few changes though, as the channel is currently created in start() and destroyed in stop()
23:33:30archivatordomonoky: having just 2 channels is hacky - what about a quad core mac? It's kinda hard to code and misses the whole point of QtConcurrent
23:35:03archivatorIf we want parallel voicing for all engines, we should go with parallelizeMode(). If not, go with the other option.
23:35:24domonokybluebroth3r: does start and stop need noticable time ? if not we could create it on demand for every voice call, and have it working in paralell.
23:36:28bluebroth3rdomonoky: no, it's fast. I've put it there to not create a new channel on each voice() call.
23:37:19bluebroth3rthe clean solution would be to still do so, and keep a map with channel -> used elements. Then add a new element if no unused channel is available, and dispose all unused channels but one in voice()
23:38:27archivatorbluebroth3r: that's the cleanest solution, yes, but it's beyond my expertise :) I'd like to get voicing parallel tonight, possibly tomorrow :)
23:38:54 Join jgarvey [0] (
23:39:00 Quit Tomis (Quit: Tomis)
23:39:33domonokyarchivator: so i would go with a flag for the tts to indicate if it is able to work in paralell.
23:39:41bluebroth3rarchivator: it doesn't sound being too complicated for the TTSCarbon part. No idea about the stuff around though :)
23:40:17archivatordomonoky: yes/no or full/voice only?
23:40:28domonokyjust yes/no
23:40:29bluebroth3rbtw, I wanted to extend voice() a bit so it can speak for itself. That would make the "Test TTS" button more responsive, at least on Mac (as the TTS defaults to outputting to the speaker anyway)
23:41:12archivatorbluebroth3r: yeah but that's not easy to implement for the other engines
23:41:16bluebroth3rso it might make sense to introduce a capabilities() functions with a bit mask, similar to the bootloader installation classes.
23:41:17 Quit Status (Ping timeout: 264 seconds)
23:41:51bluebroth3rarchivator: sure, but with a flag the caller can delegate speaking to voice() if it supports it.
23:42:47archivatorRight, so capabilities() it is then. I'll only put one in there, you figure out the rest :)
23:43:18bluebroth3rfine. I've already looked into that a while ago so it shouldn't be much work.
23:43:23archivatorBtw, festival can be tricked to pass audio to the speakers, too.
23:43:52bluebroth3rand if we can use capabilities() for other purposes than mine it's a good thing too
23:44:51bluebroth3rsomewhat different: I'd like to get a new rbutil release done in the not to far future, especially due to that bootloader installation bug
23:45:29domonokyand we have 3 new targets in rbutil.
23:45:36amiconnlinuxstb: Oh. Do you know what the OF is using the TI DSP for?
23:45:57 Join Blue_Dude [0] (
23:46:12 Join Status [0] (
23:46:47bluebroth3rdomonoky: it's 4 now :)
23:47:05bluebroth3rthough I consider the bugfix more important than the new targets
23:47:09linuxstbamiconn: Video I expect.
23:47:39linuxstbamiconn: I can't remember what formats the OF handles though. scorche now has that av300.
23:47:58domonokybluebroth3r: sure, but its another point for a new release. :-)
23:50:00 Quit domonoky (Read error: Connection reset by peer)
23:50:11linuxstbamiconn: google tells me it plays MPEG-4 SP with MP3 audio up to 640x272@25fps. It also plays WMA up to 160kbps.
23:50:52amiconnWell MP3 is handled by the MAS. That leaves the MPEG-4 part for the TI
23:51:16*amiconn wonders what clock the DSP is running at
23:51:59amiconnThose DSP cores seem to be rather clock efficient - the MAS is running at ~24MHz internally, and it does not only manage to decode, but also encode MP3 realtime
23:53:00pixelmalinuxstb: he even brought it to last year's DevCon...
23:53:18linuxstbIIRC, there's also video-related features in the lcd controller - things like colour-conversion, overlays and scaling.
23:54:28amiconnYeah, I remembered that part
23:55:16 Join Barahir [0] (
23:55:21 Quit komputes (Ping timeout: 258 seconds)
23:58:10 Join Geert_van_Dijk [0] (

Previous day | Next day