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

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

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

#rockbox log for 2010-03-29

00:00:57 Join BHSPitMini [0] (
00:01:15FlynDiceIs it the buffer size comment above you're interested in?
00:01:18pamauryWell that's simple, in ata-sd-pp.c, in sd_select_bank, there is a big buffer (512 bytes) but from what I understand, only the first word is set (to the bank index), the rest is null and it's written 16 words at a time. That would suggest that a 32-byte buffer is sufficient no
00:02:29 Quit stripwax (Quit:
00:02:48FlynDiceLet me take a look, funman worked on the bank switching code but I studied it a bit to understand what was going on.
00:03:43pamauryok, thanks.
00:03:51 Quit MikeSch (Quit: CGI:IRC (EOF))
00:03:59 Quit BHSPitMini (Client Quit)
00:04:16 Join Barahir [0] (
00:10:58 Quit pamaury (Quit: Page closed)
00:13:17 Quit perfectdrug (Quit: Leaving.)
00:14:55 Quit bertrik (Quit: De groeten)
00:15:00 Quit adnyxo (Quit: Leaving)
00:17:36FlynDicepamaury: (logs I guess) I don't think I can give you a reliable answer to that question but I can give you an educated guess. I think the card expects to receive 512 bytes containing the bank number in the first word.
00:18:46FlynDiceIf you can send the first word and then continue to fill the FIFO with null bytes until you get to 512 I think it may work.
00:19:51 Quit killan_ (Quit: ( :: NoNameScript 4.22 :: ))
00:20:06 Part domonoky
00:20:17FlynDiceIt looks like activating bank switching take a few more words also when you turn it on.
00:21:06 Join killan [0] (
00:21:22FlynDicefunman:(logs) Would you proof ^^ to make sure I'm not blowing smoke out my ### ;-)
00:31:16 Join adnyxo [0] (
00:32:00 Quit bmbl (Quit: Bye!)
00:32:00 Quit xiainx (Ping timeout: 248 seconds)
00:36:34 Quit adnyxo (Remote host closed the connection)
00:36:56 Join xiainx [0] (
00:40:54 Join adnyxo [0] (
00:40:58 Quit xiainx (Client Quit)
00:44:02 Join xiainx [0] (
00:44:29 Quit n17ikh (Ping timeout: 246 seconds)
00:44:43TorneI presume it's not intentional or desirable that if you set Bookmark on Stop to "ask" and then wait for idle poweroff, it prompts you for it, and then when you don't answer it does an unclean shutdown after a timeout? :)
00:45:31Torne(at least that's what i presume is happening, the LCD doesn't get turned off properly)
00:47:27Torne...what would we *like* it to do in that case? always save one? never save one?
00:50:23 Quit adnyxo (Remote host closed the connection)
00:51:04 Join n17ikh [0] (
00:55:47 Quit jhulst (Quit: Read error: EOF)
00:59:19***Saving seen data "./dancer.seen"
01:00:56LloreanI'd say "never save one" but I consider "stop" a manual action.
01:01:23LloreanWhile a stop technically occurs during shutdown, the user didn't stop and whether or not a bookmark is created, there's still the option of resuming.
01:01:58TorneLlorean: This is from FS #2720
01:02:07TorneA possibly unrelated change half-implemented this later
01:02:17Torneit might not have *intended* to :)
01:02:30 Join S_a_i_n_t [0] (S_a_i_n_t@
01:02:37TorneThe user in the FS expected that since they'd set it to yes, it would save one then too
01:02:49Torneand thus when they forget to resume playback and just go play something else, they were unhappy about losing their place
01:02:54Tornewhich seems fair enough. and that case already now works
01:03:05Torneso i guess i should fix it to not prompt if we're shutting down, and just assume no.
01:03:09 Quit ender` (Quit: I like work: it fascinates me. I can sit and look at it for hours. -- Jerome K. Jerome)
01:03:16Tornesince that's clearly an actual bug
01:03:24LloreanI tend to think if you want to be sure there's a bookmark, don't set it to "ask"
01:03:37TorneYah, that seems reasonable
01:04:06Torneok, i'll see if i can fix it then ;)
01:04:09LloreanSounds good to me.
01:04:18Tornethis may not be entirely trivial, however
01:04:46 Quit S_a_i_n_t_ (Ping timeout: 252 seconds)
01:05:02*Llorean personally just does an awful lot of manual bookmarking.
01:05:15*Llorean looks forward to the day it's possible to bookmark a dynamic playlist.
01:05:23Tornepersonally i use the unrelated patch which makes bookmark on stop only when the bookmark file already exists :)
01:05:35Torneso i do one manual bookmark for an audiobook and then it remembers for me after that
01:05:42Tornewithout cluttering my fs with bookmarks for albums
01:05:55Tornebut the menu options that one adds are kinda ugly, not gonna submit it without a UI redesign ;)
01:05:56LloreanThat would be nice.
01:05:59LloreanIs that a patch in FS?
01:06:08LloreanUgly menu options?
01:06:13TorneFS #6272
01:06:15TorneWell, not ugly
01:06:20Tornebut it adds yet another thing to bookmark on stop
01:06:28Torneand the behaviour wrt. the recent bookmarks is maybe confusing
01:06:41TorneIf we're gonna add this I think we need to redesign the bookmark on stop options
01:06:54Tornethe "Yes - Recent Only" and similar are rather odd
01:07:04Torneand adding more things just makes it worse
01:07:17TorneIt might be nicer to have two options, for the two kinds of bookmark, instead.
01:07:44Tornethe patch from 6272 currently saves a recent bookmark entry always, if recent bookmarks are enabled, and a regular bookmark only if the file exists, which is *probably* the desired behaviour
01:07:53LloreanI've actually thought the "Recent bookmarks" list should never result in scattered bmark files
01:08:01TorneWell that on its own doesn't, no
01:08:18TorneIf you set bookmark on stop to "Yes - Recent Only" it doesn't make any files except the single recent one
01:08:41LloreanAh, I don't use bookmark on stop
01:08:55TorneBut yeah, I agree with the other commentors on FS #6272 that the settings menu for this is confusing eonugh already, and thus adding another option to it is probably not desirable
01:08:57LloreanI more or less want "temporary bookmarks"
01:09:00Tornei'v ejust not gotten around to thinking about it
01:09:08TorneRight, that might also be good
01:09:12LloreanI want bookmarks in the one or two spoken word things I'm listening to that vanish once I'm done.
01:09:13TorneBut again, more settings complexity :)
01:09:32Torneso, yeah, if you can think how to fit it in the UI i'l see what i can do ;)
01:09:42Tornei am reasonably sure i know how the boomarking code works
01:10:00LloreanWell, I like your idea of "don't create bookmarks unless one is manually triggered"
01:10:09pixelmaeven though I personally almost never use bookmarks (except for testing purposes and on special occasions) I still wonder why the dynamic menu defaults to off - and if it was me it wouldn't need to be dynamic at all
01:10:33LloreanBut what I'd want to see then, in the Recent Bookmarks list, is two marks for that file - the most recent manual bookmark, and the most recent automatic one (independently on whether I've created several manual onces since the automatic, or several automatic since the manual)
01:10:41Tornewell, if anyone has comments on 6272 please post them, and i may get around to working on it again and committing it ;)
01:10:49Tornei didn't write the original patch or raise the issue, btw
01:10:52Tornei just saw it and likedit
01:11:03Torneit's an old one as you can tell from the number
01:11:12LloreanI don't like automatic bookmarks when I realize I've fallen asleep with my headphones on and just hit stop and put it away.
01:11:29LloreanSo it'd be nice if there was a way to distinguish quickly which one I made and which the software did.
01:11:50LloreanIdeally a whole new bookmark UI would kinda be nice.
01:11:56LloreanDrop the idea of .bmark files entirely
01:12:13LloreanOr at least, scattered throughout the filetree
01:12:27LloreanA bookmark catalog that shows you a list of folder names / playlists you have a bookmark in.
01:12:41LloreanWhen selected you see the two most recent (one auto, one manual) and an entry to see all manual bookmarks.
01:13:01Tornewell i'm not volunteering to write *that*, probably ;)
01:13:02LloreanAuto bookmarking is always on when you resume a bookmark, but never on otherwise.
01:13:06LloreanSo no settings whatsoever.
01:13:31LloreanBut that's more or less what *I'd* want.
01:17:10Torneis there a way to know we are shutting down, btw?
01:17:17Tornewithout having to pass that information around?
01:17:51LloreanI don't know for certain
01:18:00LloreanBut I want to say "no, and we've had problems with that before"
01:19:17Tornewell, i can easily add such a parameter to bookmark_autobookmark
01:24:09 Nick fxb__ is now known as fxb (
01:27:51CIA-5New commit by torne (r25376): Fix "bookmark on stop: ask" when idle poweroff triggers. ...
01:28:19Torneok that's fixed :)
01:32:00S_a_i_n_tBah! That reminds me I was supposed to be fixing up FS #11101, I had a look at it, but IMO, I just made it more complicated...and worse.
01:32:25S_a_i_n_tAnyone else think that "Reverse USB Handling" is still pretty fucking confusing?
01:32:35S_a_i_n_t...that was the Author's "fix"
01:35:49Tornethat's terrible, yes
01:35:56Tornedidn't we discuss this the other day and decide on a much better wording?
01:36:03*Llorean proposes some wording
01:36:17S_a_i_n_t*phew* glad I'm not the only one that thought so...
01:36:19LloreanSomething like "Storage on cable attach" with the options "When button is pressed" and "When no button is pressed"
01:36:26LloreanMake it about the storage rather than the power.
01:36:33S_a_i_n_tWe *did* can't say I remember *exactly* what it was though... :/
01:36:45S_a_i_n_tLlorean: that was the idea...
01:36:55S_a_i_n_tsomething about, USB mass storage mode.
01:36:55LloreanWouldn't know, wasn't around.
01:36:57*linuxstb recalls having many discussions about this in the past...
01:37:07Lloreanlinuxstb: We seem to be down to the naming
01:37:23Tornelook in the logs, then
01:37:31S_a_i_n_t*And* needs more than yes/no for its options.
01:37:33Tornei'm pretty sure we did find something everyone present was happy with :0
01:37:41Tornelemme see if i can find it
01:37:50linuxstbLlorean: FS #11101 isn't what I remember being agreed. Although maybe there has been a more recent discussion that I missed.
01:38:01TorneIt was more recent than that
01:38:02Tornei think
01:38:06Tornemaybe not.
01:38:12S_a_i_n_tA couple of days, tops.
01:38:13Tornethe original author wasn't involved in the discussion, anyway
01:38:20Torneso he did his own thing independantly
01:38:48LloreanI don't know either.
01:38:59Tornefound it, one mo
01:39:05LloreanI'm still in favour of "the user picks what the default behaviour is, and then any button held down reverses it"
01:39:35S_a_i_n_tA little bit long for a menu entry though ;)
01:39:45LloreanWell, I meant for the behaviour
01:39:56S_a_i_n_tThat *is* the behaviour IIUC
01:40:06S_a_i_n_twell, the one I'm using now does exactly that.
01:40:27TorneHere was the discussion :)
01:40:30S_a_i_n_tCan't see any changes except naming in the new one.
01:40:53TorneI suggested "Default USB mode: Storage device, charging only"
01:41:04TorneSuch that we could add MTP as a future option
01:41:05LloreanThat patch seems to just present the option, without opening it up to "any button" from the current "a single, sometimes hard to press while inserting, button"
01:41:08TorneOr other possible USB modes.
01:41:26LloreanTorne: I'd say the storage type should be another setting entirely.
01:41:51LloreanThis setting should just be for helping out the people who just want to charge and occasionally connect vs the people who usually connect to sync and rarely want to just charge.
01:42:15Tornewell, that's a future discussion
01:42:21Tornebut I like that wording, anyway
01:42:31Torneand it seemed ok to the people around at that time :0
01:42:32*Torne shrugs
01:42:39LloreanI'd say "On USB insertion" rather than "Default USB mode" but could leave the options
01:42:41TorneI agree that it might be nice if it was any button, though
01:42:48LloreanDefault USB mode sounds like you might be picking other things too
01:42:54linuxstbTorne: There have been more discussions on the general issue further back in time.
01:43:06Torneanyway time i slept. :)
01:43:21S_a_i_n_tSeemed OK to me, any botton would be better than just menu...then you wouldn't get the damn quickscreen on the ipods.
01:43:31S_a_i_n_tWhich is a pet peeve of mine.
01:45:03Llorean"select" is probably easier to hold down while inserting the cable on iPods
01:45:48S_a_i_n_tIndeed, and wouldn;t have the quickscreen porblem, but would possibly bring up the ontext menu depending where you were in the menu at the time.
01:46:21LloreanWell, every button does something so that's just something to accept
01:46:39 Join shaggy-h [0] (
01:47:13S_a_i_n_tGood point, I had an idea that the quickscrren (if menu was used) would/should time out if the button is held for X period of time, but that's harder to implement.
01:47:39LloreanThe original point of the quickscreen was actually to only be up while you held the button
01:47:48LloreanYou'd hold the button, change the values, then release it causing the screen to vanish
01:47:50LloreanThus, very quick
01:48:04S_a_i_n_tNot possible on an ipod though...
01:48:22S_a_i_n_t(you can only hold one button at a time IIRC)
01:48:41LloreanI don't think that's actually the case
01:48:58LloreanI mean, obviously select can be combined with other buttons, but I seem to recall "Left+Right" being a combo for something at one point
01:49:12S_a_i_n_tit's be *really* difficult, if not impossible to hold menu, then for instance.
01:49:50S_a_i_n_tand holding left+right would literally bend the clickwheel.
01:50:01LloreanBut, more to the point - making the QS vanish if you held the button too long, but only on some players, just introduces even more behaviour schisms.
01:50:11S_a_i_n_tI've tried this with 2 player just doesn;t work.
01:50:27S_a_i_n_tYes, that I agree with.
01:51:06S_a_i_n_tIts just annoying that if you mess it up, you end up scrolling at light-speed through the top quickscreen selection.
01:51:31LloreanWell up in the quickscreen didn't *used* to change options
02:08:05 Quit xiainx (Ping timeout: 268 seconds)
02:21:11 Join xiainx [0] (
02:21:13 Quit m3dlg (Ping timeout: 252 seconds)
02:24:54 Join m3dlg [0] (~m3dlg@
02:27:10 Quit CGL (Ping timeout: 265 seconds)
02:30:32 Nick fxb is now known as fxb__ (
02:32:14kugelwhy is the charging button on menu anyway?
02:32:17 Quit m3dlg (Read error: Connection reset by peer)
02:33:32LloreanBecause that's the button that was used on other players before, I think
02:33:35LloreanIsn't it Select on the e200?
02:33:39S_a_i_n_tkugel: it is?
02:34:10kugelit's select on the e200 yes
02:35:39kugelthe easiest fix seems to simply change it to select on the ipods
02:36:31pixelmawhat do you gain - you'll call the context menu instead
02:38:15kugelas Llorean, every button does something so that's something you can't avoid
02:38:47Lloreanpixelma: Holding Select in the context menu doesn't easily cause unwanted effects.
02:38:55LloreanWhile holding Menu apparently can now that changes were made
02:39:07kugeland for accidental entering, the context menu seems less bad that the quickscreen
02:39:56 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
02:44:58 Join m3dlg [0] (~m3dlg@
02:47:01 Quit Barahir (Ping timeout: 264 seconds)
02:50:16 Quit m3dlg (Ping timeout: 264 seconds)
02:50:17S_a_i_n_tHmmmm...FWIW |<< & >>| do nothing on long holds on the ipods.
02:50:23 Join m3dlg [0] (
02:50:44S_a_i_n_tI'd see one of those two as a better solution to flip the charging behaviour on the them.
02:50:58S_a_i_n_tAs opposed to, menu/select.
02:52:40S_a_i_n_t*when in the menu that is, obviously this won;t work in the WPS
02:52:45Llorean|<< Should take you to the furthest up menu level
02:52:59S_a_i_n_ta short press does, long press...nothing.
02:53:05Llorean>>| doesn't do anything yet, but it has been considered for "select" if multi-select ever happens.
02:53:31LloreanBut if anything, it should be select. It's pretty much the easiest button to hold consistently while plugging something in.
02:53:32 Join anewuser [0] (anewuser@unaffiliated/anewuser)
02:53:38S_a_i_n_tShit, sorry, my mistake.
02:53:44S_a_i_n_tYou're quite correct.
02:53:52S_a_i_n_tBah, back to the drawing board.
02:56:36 Join saratoga_lab [0] (~9803c20d@gateway/web/freenode/x-fhzihnppyrxgmzms)
02:57:03JdGordonregarding the bookmarking... there is an old patch of mine whcih tries to replace the .bmark file with a copy of the .playliist_control file so everything can be bookmarked
02:59:06pixelmaisn't holding left or right a forced horizontal scroll (you need a long enough menu item though depending on font and actual screen size)?
02:59:20***Saving seen data "./dancer.seen"
03:03:16Lloreanpixelma: It used to be, actually. I don't know
03:04:10 Join adnyxo [0] (
03:04:16 Join Barahir [0] (
03:12:27kugelone of our most unknown features probably
03:13:13 Join hebz0rl_ [0] (
03:13:19LloreanI used to use it on my H100 all the time, but the extra width of my Gigabeat's screen, and the fact that I still use about the same font size, means it's less necessary.
03:14:10S_a_i_n_tMy ipod Nano1g (25296) doesn't give the "Battery low: Shutting Down" splash when the battery dies and its in the WPS. Is this expected behaviour?
03:14:23S_a_i_n_tI seem to remember it prompted wheter in the WPS or not.
03:14:45saratoga_labthat probably means it shut off before it hit the low voltage threshold
03:15:26S_a_i_n_tthe battery reads 9%, it gives the splash when its not in the WPS, I've checked this.
03:16:01S_a_i_n_tWhen it *is* in the WPS, it just dies...the screen slowly fades.
03:16:15saratoga_labwas the screen already on?
03:16:25S_a_i_n_tcome again?
03:16:39saratoga_labwas the screen already on when it faded
03:16:43 Quit hebz0rl (Ping timeout: 260 seconds)
03:16:52S_a_i_n_tShowing the WPS.
03:17:04saratoga_labincluding backlight?
03:17:05S_a_i_n_tthen, just BAM...dead.
03:17:16S_a_i_n_tbut if its in the menu, I get the splash.
03:17:26saratoga_labhmm though i guess on the 2g it probably doesn't need to bring on the backlight when it shows the splash
03:17:40S_a_i_n_tthis is the Nano1g
03:19:03S_a_i_n_tIt usually brings on the backlight for "shutting down" (which I think is odd), but in this case, while in the warning, just, dead.
03:27:21S_a_i_n_tCorrection, it *doesn't* give the Battery low: shutting down" splash while in the menu.
03:27:39S_a_i_n_tIt gives that splash if I turn it off manually when the battery is this low.
03:27:49S_a_i_n_tbut it never *quite* shuts down properly...
03:28:27S_a_i_n_tOn the main menu/menu, it just dies like I described and the screen slowly fades out.
03:31:39S_a_i_n_tI wanted to keep experimenting its failings...but the battery is good'n'dead now.
03:31:51S_a_i_n_tIts definitely not "intended behaviour"
03:38:50 Quit adnyxo (Remote host closed the connection)
04:02:59 Quit kugel (Remote host closed the connection)
04:04:26 Quit Strife89 (Quit: Bed.)
04:04:32 Quit TheSeven (Disconnected by services)
04:04:44 Join tuxbot007 [0] (
04:04:46 Join The_Seven [0] (~theseven@rockbox/developer/TheSeven)
04:04:56 Nick The_Seven is now known as TheSeven (~theseven@rockbox/developer/TheSeven)
04:07:05 Quit SpyBot (Ping timeout: 258 seconds)
04:08:01 Join SpyBot [0] (
04:13:04 Quit Darkknight512 (Quit: ChatZilla 0.9.86 [Firefox 3.5.8/20100202165920])
04:25:05S_a_i_n_tAn .icons file *should* override anything defined in viewers.config, should it not?
04:25:29JdGordonthats the thery
04:25:45 Join xavieran [0] (
04:25:50 Join panni__ [0] (
04:25:53S_a_i_n_tHmmm...somethings definitely wrong then, I *really* wish I could see what it is though.
04:29:09 Quit panni_ (Ping timeout: 264 seconds)
04:29:24S_a_i_n_tTry as I might, I just *can't* get an icon displayed for .config, .playlist_control, .tcd, .bin, ...shit, that's all I remember.
04:29:34S_a_i_n_tAnyone else experience this weirdness?
04:30:51S_a_i_n_tand *all* of those exts are in my .icons file, in fact I have ~250 exts in my icons file, so I *should* be covered for pretty much any eventuality.
04:31:46LloreanThere may be an upper limit to how many filetypes a .icons can cover?
04:32:13S_a_i_n_tI thought that, but it didn;t work when my .icons file only had ~20 entries.
04:32:26S_a_i_n_tI made it specifically to cover the exts that had no icon.
04:32:54S_a_i_n_tI've tried an .icons file with *just* the exts that don;t display an icon, but still no dice.
04:37:44 Join Rob2222 [0] (
04:41:12 Quit Rob2223 (Ping timeout: 248 seconds)
04:49:08 Quit saratoga_lab (Quit: Page closed)
04:53:16 Quit Barahir (Ping timeout: 265 seconds)
04:59:21***Saving seen data "./dancer.seen"
05:02:27 Quit n17ikh (Ping timeout: 265 seconds)
05:08:02 Join n17ikh [0] (
05:13:12 Quit xiainx (Ping timeout: 248 seconds)
05:25:26 Quit hebz0rl_ (Quit: Ex-Chat)
05:28:48 Quit __arbingordon (Quit: `)
05:39:58 Quit Horscht (Quit: Verlassend)
05:45:44 Join kramer3d [0] (~kramer@unaffiliated/kramer3d)
05:59:36 Quit panni__ (Read error: Connection reset by peer)
06:03:11 Quit tuxbot007 (Ping timeout: 276 seconds)
06:04:27 Join Barahir [0] (
06:37:50 Quit mt (Ping timeout: 240 seconds)
06:40:04 Join froggyman [0] (
06:40:06 Quit froggyman (Changing host)
06:40:06 Join froggyman [0] (~me@unaffiliated/froggyman)
06:43:36 Join mt [0] (~mtee@rockbox/developer/mt)
06:46:02 Quit froggyman (Quit: give me a sec)
06:46:34 Join froggyman [0] (
06:46:37 Quit froggyman (Changing host)
06:46:37 Join froggyman [0] (~me@unaffiliated/froggyman)
06:47:11 Join mt_ [0] (~mtee@
06:48:46 Quit mt (Ping timeout: 265 seconds)
06:48:52 Nick mt_ is now known as mt (~mtee@
06:54:12 Part froggyman
06:59:22***Saving seen data "./dancer.seen"
07:05:29 Quit kramer3d (Quit: Leaving)
07:12:28JdGordonhow do we delte wiki pages?
07:12:36JdGordonor is that not possible for regular users?
07:13:40JdGordonah, found it
07:13:58 Join merbanan [0] (
07:24:32 Quit DV (Ping timeout: 276 seconds)
07:24:44S_a_i_n_tAnyone know of a way I can screenshot/dump the images in an .ipsw file?
07:25:24S_a_i_n_tiPodwizard *says* it can do it, but my antivirus had a fucking fit when I clicked the dl I'm not going there :P
07:27:55 Quit merbanan (Ping timeout: 260 seconds)
07:28:39 Join DV [0] (~DV@
07:37:35 Join blairb_ [0] (
07:40:32 Join CGL [0] (~CGL@
07:40:58 Quit blairb (Ping timeout: 258 seconds)
07:46:57TheSevenS_a_i_n_t: which generation?
07:47:56 Quit shai (Read error: Connection reset by peer)
07:48:14 Join shai [0] (
07:48:27S_a_i_n_ti/2g...doesn't really matter.
07:48:46S_a_i_n_tI'd just like to drop the images from the ipod/.ipsw if possible
07:49:01S_a_i_n_t1/2g rather
07:49:05 Quit shai (Read error: Connection reset by peer)
07:49:21S_a_i_n_tI know its offtopic, so PM me if you'd rather
07:49:29 Join shai [0] (
07:49:52 Quit shai (Read error: Connection reset by peer)
07:50:09 Quit CGL (Ping timeout: 246 seconds)
07:54:46 Join CGL [0] (~CGL@
08:02:18 Join shai [0] (
08:09:27 Quit anewuser ()
08:11:26 Quit m3dlg (*.net *.split)
08:11:27 Quit amiconn (*.net *.split)
08:11:27 Quit jordan` (*.net *.split)
08:11:27 Quit togetic (*.net *.split)
08:11:27 Quit advcomp2019_ (*.net *.split)
08:11:27 Quit bzed (*.net *.split)
08:11:27 Quit Llorean (*.net *.split)
08:11:27 Quit logiclost (*.net *.split)
08:11:27 Quit BlakeJohnson86 (*.net *.split)
08:11:27 Quit Guest6244 (*.net *.split)
08:11:27 Quit ps-auxw (*.net *.split)
08:11:27 Quit alexbobp (*.net *.split)
08:11:27 Quit aevin (*.net *.split)
08:11:27 Quit simabeis (*.net *.split)
08:11:27 Quit ranma (*.net *.split)
08:11:27 Quit ThomasAH (*.net *.split)
08:11:27 Quit jobec (*.net *.split)
08:11:27 Quit fxb__ (*.net *.split)
08:11:27 Quit Kitar|st (*.net *.split)
08:11:27 Quit merbzt (*.net *.split)
08:11:27 Quit tmzt (*.net *.split)
08:11:27 Quit dionoea (*.net *.split)
08:11:27 Quit rasher (*.net *.split)
08:11:27 Quit linuxstb (*.net *.split)
08:11:27 Quit Galois (*.net *.split)
08:11:27 Quit Utchybann (*.net *.split)
08:11:27 Quit S_a_i_n_t (*.net *.split)
08:11:27 Quit Hadaka (*.net *.split)
08:11:27 Quit mikroflops (*.net *.split)
08:11:27 Quit kadoban__ (*.net *.split)
08:11:27 Quit GHF (*.net *.split)
08:11:27 Quit rvvs89 (*.net *.split)
08:11:27 Quit mt (*.net *.split)
08:11:27 Quit Barahir (*.net *.split)
08:11:27 Quit shaggy-h (*.net *.split)
08:11:27 Quit Adubb (*.net *.split)
08:11:28 Quit SirFunk_ (*.net *.split)
08:11:28 Quit FOAD (*.net *.split)
08:11:28 Quit DV (*.net *.split)
08:11:28 Quit Unhelpful (*.net *.split)
08:11:36 Join m3dlg [0] (
08:18:11 Quit CaptainKewl (Ping timeout: 246 seconds)
08:26:24 Join ender` [0] (
08:29:51 Nick m3dlg is now known as 17SAACD6H (
08:29:57 Join Adubb [0] (~aldubuc@
08:29:57 Join m3dlg [0] (
08:29:57 Join advcomp2019__ [0] (
08:29:57 Join DV [0] (~DV@
08:29:57 Join mt [0] (~mtee@
08:29:57 Join Barahir [0] (
08:29:57 Join shaggy-h [0] (
08:29:57 Join S_a_i_n_t [0] (S_a_i_n_t@
08:29:57 Join Kitar|st [0] (Kitr88@
08:29:57 Join amiconn [0] (quassel@rockbox/developer/amiconn)
08:29:57 Join jordan` [0] (~jordan@
08:29:57 Join togetic [0] (~togetic@unaffiliated/ibuffy)
08:29:57 Join Hadaka [0] (
08:29:57 Join SirFunk_ [0] (
08:29:57 Join merbzt [0] (~benlar@
08:29:57 Join tmzt [0] (
08:29:57 Join bzed [0] (
08:29:57 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
08:29:57 Join mikroflops [0] (
08:29:57 Join logiclost [0] (
08:29:57 Join dionoea [0] (
08:29:57 Join kadoban__ [0] (
08:29:57 Join rasher [0] (~rasher@rockbox/developer/rasher)
08:29:57 Join BlakeJohnson86 [0] (~bjohnson@2002:1876:a27b:0:227:13ff:fe65:1262)
08:29:57 Join GHF [0] (~meow@unaffiliated/ghf)
08:29:57 Join Guest6244 [0] (
08:29:57 Join FOAD [0] (
08:29:57 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
08:29:57 Join rvvs89 [0] (~rvvs89@pdpc/supporter/base/rvvs89)
08:29:57 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
08:29:57 Join alexbobp [0] (~alex@
08:29:57 Join Unhelpful [0] (~quassel@rockbox/developer/Unhelpful)
08:29:57 Join aevin [0] (eivindsy@unaffiliated/aevin)
08:29:57 Join fxb__ [0] (
08:29:57 Join jobec [0] (
08:29:57 Join ThomasAH [0] (
08:29:57 Join ranma [0] (
08:29:57 Join simabeis [0] (
08:29:57 Join Galois [0] (
08:29:57 Join Utchybann [0] (
08:35:51 Join flydutch [0] (
08:46:32 Quit DV (Ping timeout: 264 seconds)
08:47:21 Quit xavieran (Read error: Operation timed out)
08:48:27 Join LinusN [0] (~linus@gateway/web/cgi-irc/
08:48:27 Quit LinusN (Changing host)
08:48:27 Join LinusN [0] (~linus@rockbox/developer/LinusN)
08:48:57 Join DV [0] (~DV@
08:59:12 Join Zagor [0] (
08:59:12 Quit Zagor (Changing host)
08:59:12 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
08:59:23***Saving seen data "./dancer.seen"
08:59:38 Quit TheSeven (Quit: ChatZilla 0.9.86 [Firefox 3.6.2/20100316074819])
08:59:53 Quit SpyBot (Read error: Connection reset by peer)
09:15:12 Quit m3dlg (Ping timeout: 240 seconds)
09:15:16 Quit 17SAACD6H (Ping timeout: 258 seconds)
09:19:21 Join efyx [0] (
09:20:19 Quit JdGordon (Quit: Leaving.)
09:25:14 Quit liar (Ping timeout: 258 seconds)
09:25:52 Join liar [0] (
09:26:52 Join Bagder [0] (~daniel@rockbox/developer/bagder)
09:29:17 Join xavieran [0] (
09:35:41 Join Tomis [0] (~Tomis@
09:37:06 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
09:38:41 Join xsteadfastx [0] (~spectrum@
09:46:41 Quit shaggy-h (Ping timeout: 240 seconds)
09:53:34 Join wodz [0] (
09:54:40 Quit BHSPitMonkey (Remote host closed the connection)
10:03:11xsteadfastxi discovered that its pretty slow to transfer music to the microsd card in my sansa e250. pretty much slower then if i put music on the player itself
10:03:33 Quit CGL (Quit: Saliendo)
10:14:01 Quit xavieran (Read error: Connection reset by peer)
10:14:48 Join xavieran [0] (
10:22:52 Quit shai (Read error: Connection reset by peer)
10:23:02 Quit MagusG (Read error: Connection reset by peer)
10:23:17 Join shai [0] (
10:23:28 Quit niekie (Read error: Operation timed out)
10:23:30 Join MagusG [0] (
10:24:11 Quit _flow_ (Read error: Operation timed out)
10:25:00 Join m3dlg [0] (~m3dlg@
10:25:09 Join _flow_ [0] (
10:29:09 Quit efyx (Quit: Quitte)
10:30:59 Join Tomis2 [0] (~Tomis@
10:32:32 Quit Tomis (Ping timeout: 276 seconds)
10:32:32 Nick Tomis2 is now known as Tomis (~Tomis@
10:33:49 Join efyx [0] (
10:44:10S_a_i_n_tIs there a reason why the Nano1g manual doesn't show A-B as a repeat mode? The description is "Repeat mode, 0-4, in the order: Off, All, One, Shuffle"...which, obviously, makes no sense at all.
10:44:28S_a_i_n_tie. the "0-4" part.
10:49:18pixelmayou would get one of the numbers between 0 and 4 if you use the repeat mode tag unconditioinally
10:50:07S_a_i_n_tMy point is, the description is missin "A-B"
10:50:59pixelmathe reason is probably that some targets with few buttons only got A-B repeat later. First it was only possible if you had free button (combos) to set the A-B markers, only later a "enable A-B repeat, set A marker, set B marker, clear markers" mode was added
10:51:16pixelmaand the manual not updated accordingly
10:52:06pixelmamy guess with some background knowledge
10:56:28 Quit m3dlg (Ping timeout: 258 seconds)
10:58:46linuxstbpixelma: So what you're saying is "patches to the manual are welcome" ? ;)
10:59:25***Saving seen data "./dancer.seen"
10:59:45 Join mitk [0] (~mitk@
10:59:53pixelmaindeed :)
11:05:36 Join TheSeven [0] (~theseven@rockbox/developer/TheSeven)
11:11:34 Quit alexbobp (Ping timeout: 248 seconds)
11:14:10 Join kugel [0] (~kugel@rockbox/developer/kugel)
11:19:00 Join pjm0616 [0] (~user@
11:29:29 Join n1s [0] (~n1s@rockbox/developer/n1s)
11:36:10 Join alexbobp [0] (~alex@
11:41:48 Join arbingordon [0] (~w@unaffiliated/arbingordon)
11:56:50 Join archivator [0] (~archivato@
12:03:50 Join notlistening [0] (
12:05:24notlisteningI know this is a stupid question but I have rockbox dev enviroment just setup using /tools and now have to add the path to the arm-elf tool and I am struggling to add that to my path I have added .bashrc and path shows up correctly when i do echo $path but still getting error but which arm-elf-gcc is not working
12:05:35notlisteningany ideas? Ubuntu 10.04
12:06:27archivatornotlistening: what error are you getting and what exactly is in your PATH?
12:07:09notlistening/bin/sh: arm-elf-gcc: not found
12:07:12krazykitdid you log out and back in after altering your .bashrc?
12:07:28archivatorkrazykit: shouldn't he just restart the terminal app?
12:07:37archivatorit's bashrc, not .profile
12:08:07notlisteningwhen i type echo $PATH I get /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/arm-elf
12:08:37n1sappend a/bin
12:08:42notlisteningbut have just noticed the arm-elf dir now contains a /bin folder with the tools should i add that
12:08:43n1seh a /bin
12:08:47notlisteninglol thought so
12:09:10notlisteningumm that needs changing in the install script for idiots like me
12:10:28notlisteningto prevent a logout use: source ~/.bashrc
12:11:55pixelmaS_a_i_n_t: as I thought - the A-B part in the WPS tag table in the appendix is only enabled for the Archoses (that already had this mode in the beginning). It was put this way in the initial commit of this part of the manual and never touched since
12:12:39pixelmain the meantime all targets should have A-B repeat (and the Ondio does, so the most problematic regarding number of buttons)
12:12:52 Join shaggy-h [0] (
12:12:56archivatorre: porting flite : is it going to be a huge problem if I stick to ARM at the beginning (btw, I'm almost convinced it can be done now, so I'll be doing it gsoc or no gsoc :) )?
12:13:01notlisteningTrying to compile on the CLip+ and getting : make: *** No rule to make target `/home/tom/rockbox/apps/codecs/libmusepack/musepack.h', needed by `/home/tom/rockbox/build/apps/codecs/mpc.o'. Stop.
12:13:34linuxstbarchivator: Why do you need to stick to ARM?
12:14:06pixelmaS_a_i_n_t: the description how to use it is explained and got a section for all in the repeat mode settings
12:14:29archivatorlinuxstb: as a safety precaution. Basically, flite has a port to palmos, so I know it can work on ARM. There are some ugly hacks in the internal data structures that will need investigating on other archs.
12:14:53S_a_i_n_tpixelma: I know how to use it, I just thought it was odd the manual didn;tlist it.
12:15:15S_a_i_n_tBut it obviously had the intention of being there, as it listed 0-4 options.
12:15:25S_a_i_n_tThanks for looking into it though ;)
12:15:32 Quit DV (Read error: Connection reset by peer)
12:16:44linuxstbarchivator: I thought palmos was also m68k, like our Coldfire targets? Also, does that mean it won't work in the sim?
12:18:05archivatorlinuxstb: from what I gather, m68k emulation was too slow for flite to be functional, so they ported it to the underlying ARM platform. I see no reason why it shouldn't work on the sim. It compiles just fine on x86 and x86_64.
12:18:37archivatorReally, I wasn't saying I won't try to port it to every arch. All I meant was that I'll focus primarily on ARM at the beginning.
12:22:06notlisteningIs there a way to stop the Sansa E200v2 Loding into the OF when the power is turned cord is plugged in?
12:22:42 Join DV [0] (~DV@
12:25:41 Quit kugel (Ping timeout: 248 seconds)
12:33:39S_a_i_n_tnotlistening: FS #11101
12:33:59S_a_i_n_tshit, sorry...I read your post wrong. Ignore that.
12:36:41notlisteningS_a_i_n_t, thanks for looking anyways
12:37:51 Join kunal [0] (~kunal@
12:38:36S_a_i_n_tThere is a way to do it, well *I'm fairly certain there is*...
12:38:55S_a_i_n_tI don;t have that target, so I'm not sure though.
12:39:37S_a_i_n_tI think someone came up with a solution for this as it annoyed them that it booted the OF every time they started thier car...but I can't seem to find the Forum thread.
12:58:16 Join perfectdrug [0] (
12:59:27***Saving seen data "./dancer.seen"
13:01:05wodzLadies and gentelments We have sound on MPIO HD200 !
13:03:24Bagderwodz: congrats!
13:03:55pamaurycongratulations !
13:04:02wodznow I have to clean up this mess I do have in my sources
13:06:28 Join perfectdrug1 [0] (
13:08:53 Quit perfectdrug (Ping timeout: 248 seconds)
13:10:35 Join kugel [0] (~kugel@rockbox/developer/kugel)
13:11:03 Quit n1s (Quit: Lämnar)
13:11:12 Join petur [0] (~peter@rockbox/developer/petur)
13:12:27kugelno gentlement mail? :)
13:14:40kugelgood news that the hd200 has sound, the samsung yh-j70 I have has the wm8750 as well
13:15:02 Join Barahir_ [0] (
13:15:04kugelthe L version though
13:18:10 Quit Barahir (Ping timeout: 240 seconds)
13:18:48 Join m3dlg [0] (~m3dlg@
13:20:40 Quit Barahir_ (Ping timeout: 240 seconds)
13:27:20wodzhmm, the ui reacts on volume buttons but audiohw_set_master_vol is never called. Any clue why?
13:29:18 Join Strife89|PalmTX [0] (
13:38:50 Join funman [0] (
13:38:51 Quit funman (Changing host)
13:38:51 Join funman [0] (~fun@rockbox/developer/funman)
13:39:02funmankugel: hey, i've got backlight working
13:40:09funmanit doesn't work in the bootloader, i'll try setting a default brightness at init
13:40:23funmancan you send me what you've done for buttons ?
13:41:09funmanFlynDice: : you're right. bank switching/init might even work while not transferring a full sector but i didn't try it
13:41:21 Quit m3dlg (Ping timeout: 264 seconds)
13:45:43kugelfunman: yes, wait a second
13:45:45 Join Edwerd [0] (
13:46:33 Quit DV (Read error: Connection reset by peer)
13:47:27funmankugel: i don't get the comment in _backlight_on(), does it come from another mode L?
13:47:27kugelfunman: I hacked dbg_ports() to show before root_menu
13:47:43kugelwhich commend exactly?
13:47:56funman#if (CONFIG_BACKLIGHT_FADING != BACKLIGHT_FADING_SW_SETTING) /* in bootloader/sim */
13:47:59funman /* if we set the brightness to the settings value, then fading up
13:48:02funman * is glitchy */
13:48:12kugelwith that patch you can see gpio changes for all buttons except scrollwheel
13:48:29 Join avar [0] (avar@wikipedia/avar)
13:48:30kugelthat's from another target but effectively applies to all targets with FADING_SW_SETTING
13:48:43 Join literal [0] (
13:51:21 Quit Strife89|PalmTX (Quit: Building change.)
13:51:45 Join stoffel [0] (
13:56:54kugelfunman: should I give you push access on my repo so we can work on a single repo?
13:57:24funmankugel: you want to work on another tree before buttons are finished?
13:57:52kugelI surely don't plan to commit that dbg_ports hack to svn :p
13:58:21funmanif i can use the arrows+center & power that's enough for me :P
13:58:29kugeli just thought it's more convinient than sharing the patches here
14:00:53funmani'm not sure if i'll help you with the buttons so i wouldn't need write access
14:07:40kugelbut I would like to have your backlight improvements :)
14:08:10funmanbacklight_on doesn't modify the backlight setting (i think it is kept in memory when we switch the backlight on/off)
14:08:19funmanso I removed the special case, does it sound ok ?
14:09:34kugelbacklight_on should set the brightness to what the user has set if fading feature is disabled
14:10:02funmanand what else if it's enabled ?
14:10:18kugeldon't touch it, since it's handled in another function
14:10:40kugelbacklight_on is called before the fading starts (to enable the lcd, for example)
14:10:46CIA-5New commit by funman (r25377): Fuzev2: fix backlight brightness ...
14:10:49funmanok :)
14:11:54kugelwas I reading the wrong function in the disasm ?
14:12:23funmannope, it was just calling a special case of ascodec
14:12:54funman1 argument is a subregister (found this in the linux code), and they're write-only (didn't see that in the linux code but deep in the OF today)
14:14:57funmani don't know what the 8| is for
14:15:39funmani'll look at clip+ backlight later
14:16:40kugelbut *10 seems simpler that what I extracted from the OF
14:17:18funmanwhat it really did was ((setting + 1) * 10) & 0xff;
14:17:19 Quit blairb_ (Ping timeout: 258 seconds)
14:17:43funmansince ascodec only write bytes we don't need to clear the top bits, and we can skip the +1 by starting at 1 directly
14:19:38funman(((x << 2) + x) + 5) << 1) = ((x*5) + 5) << 1 = (x*10+10) = (x+1)*10
14:20:03funmanthe << 1 is mixed with & 0xff (<< 25, >> 24)
14:24:03funmanpamaury: I keep up to date as much as possible
14:32:06 Join Tomis2 [0] (~Tomis@
14:32:14 Join webguest85 [0] (
14:32:25 Quit webguest85 (Client Quit)
14:33:21pamauryyes, I was looking at it but I was wondering why it was unstable with only usb non working but then I saw that storage was read-only :) That changes things a bit !
14:33:27 Join Mateusz [0] (
14:34:27 Quit Mateusz (Client Quit)
14:34:30funmanyes, and there are still some problems with sound. It shouldn't be too hard to fix (especially with the as3543 source code in the linux patches) but I wait a bit to see if my AMS contact answers with the as3543 datasheet
14:35:22 Quit Tomis (Ping timeout: 258 seconds)
14:35:23 Nick Tomis2 is now known as Tomis (~Tomis@
14:35:44pamaurythe clip ones are brickable in theory, did that already happen/is there a real risk ?
14:36:44funmanall clips are brickable, if you don't modify mkamsboot it should be safe (i bricked a few but only with hacked versions of mkamsboot)
14:38:24kugelfunman: ah, so I sucked :p
14:38:36pamaurybut there is no risk if I don't touch mkamsboot ?
14:39:04pamauryfunman: did you recovered them or were there permanently bricked ?
14:39:11avarSounds like there's need of a Clip+ FAQ , I'm here for it too :)
14:39:14funmankugel: i just simplified the arithmetic today while thinking hard about it, i suck just like you then :p
14:39:25funmanpamaury: bricked
14:40:43funmanassuming you only try released bootloaders on tested versions of the OF (there are md5sums of tested OFs in mkamsboot) then the risk is the same than upgrading the (unpatched) OF
14:40:56avarWho's been doing most of the Clip+ work? FlynDice judging from some of the commit messages, and funman.
14:43:39pamauryfunman: what happen if the file system is corrupted on the clip ? Is it bricked ?
14:43:52CIA-5New commit by funman (r25378): Clipv2/Clip+ : turn off backlight in 0x1b register ...
14:44:07kugelfunman: maybe their compiler optimizes *10 in that way?
14:44:11funmanavar: recently yes, although i hope kugel will participate after he finishes fuzev2 buttons
14:44:13 Quit MagusG (Ping timeout: 268 seconds)
14:44:55funmankugel: gcc does something similar afaict
14:45:00 Join MagusG [0] (
14:45:29kugelwell, on armv5 *10 as a mul is probably faster than that series of instructions
14:45:33funmanpamaury: the firmware is not stored on the filesystem so it's safe. it's stored on the first sectors of the flash (see the offset applied to start sector in sd_*_sectors() )
14:51:49avarfunman: That's neat, it was on my infinite todo to try to port rockbox to my sansa clip+, nice that somebody that knows what they're doing beat me to it. Anyway if there are some testing / C coding tasks to be done I'd be happy to try to help. I didn't find anything relevant in the bug tracker.
14:52:12avarI also subscribed to rockbox-dev, I guess that'll help:)
14:52:49funmanavar: check out SansaAMS on the wiki, not working items are in red/yellow (there's nothing on the tracker)
14:53:28 Quit mitk (Quit: Leaving)
14:53:32funmanpick any items and tell us so we don't work on the same thing, FlynDice is working on SD writes, except that you're free to pick up any item :)
14:54:09CIA-5New commit by funman (r25379): as3543: name 0x1b & 0x1c registers
14:55:03kugeldo we really have 25 backlight levels?
14:55:16kugelcould we go even finer if we do *5 instead of *10?
14:55:21funmani don't know
14:55:35kugelthe of has 12 levels
14:55:37funmanwriting 5 (1*5) didn't work
14:55:50funmanfinish the buttons so we can test ! :p
14:56:13 Nick fxb__ is now known as fxb (
14:56:34kugelwell, I tried hard to find the scrollwheel bits
14:57:25kugelthe OF always calls the read function with read_gpioa_67 == false, but it calls the read_gpioa_67 separately later
14:57:27funmanif the other buttons work fine can you commit them ?
14:57:38kugelI can try later today
14:59:29kugelfunman: btw, something is wrong with amsinfo I think
14:59:30***Saving seen data "./dancer.seen"
15:00:27kugelusb_functio is detected strangely, and if I load it into ida, and then jump from the firmware part into a function that's supposed to be in usb_functio I mostly get garbage
15:00:36CIA-5New commit by funman (r25380): as3525v2: factorize writing to ascodec PMU
15:00:45funmankugel: which OF ?
15:00:54kugel2.01.17 fuzev2
15:02:25funman"detected strangely" ?
15:03:19funmanwhere is this function (in the firmware part) ?
15:06:40 Quit funman (Quit: free(random());)
15:13:36wodzthere is a bug in scramble.c I think. there is definition char modelname[5]; and later there is strcpy(modelname, &argv[1][5]); for name h10_5gb this will fail badly
15:14:57kugelwodz: modelname comes from configure, it should only be 4 chars
15:15:46wodzso what does the line else if(!strcmp(&argv[1][5], "h10_5gb")) ? in scramble.c
15:16:25wodzanyway why do we stick to 4 chars names?
15:16:45kugeltool="$rootdir/tools/scramble -mi4v2 -model=h105 -type=RBOS"
15:17:00kugelnothing passes h10_5gb to scramble actually
15:17:19Tornewodz: because 4 chars is enough :)
15:17:52wodzYou should add for everyone :-)
15:18:53kugelit should use strncpy still
15:19:03wodzwhy not extend modelname size and do strncpy for safety
15:19:30kugelbecause the former would break everything :)
15:19:40wodzkugel, how?
15:19:47kugel(unless our bootloader aren't actually checking the model string?)
15:20:42 Nick crwll is now known as crwl (
15:20:59 Quit kugel (Remote host closed the connection)
15:25:00 Join robin0800 [0] (
15:26:55 Join evilnick_B [0] (~0c140464@rockbox/staff/evilnick)
15:28:30 Join adnyxo [0] (
15:39:21 Quit evilnick_B (Quit: Page closed)
15:40:17 Join evilnick_B [0] (~0c140464@rockbox/staff/evilnick)
15:41:29 Quit wodz (Quit: Leaving)
15:41:42 Join bjoernb [0] (
15:43:09bjoernbi am searching for a small (like finger size) media player that is capable to play flac and ogg. i do not need a big display for video or sth. like that. do you have any reccomendations?
15:43:28 Quit shai (Quit: Leaving)
15:43:38Bagdersansa clip?
15:43:52bjoernbi'll have a look
15:49:14 Quit FlynDice (Remote host closed the connection)
15:53:52topik_clip+ does it well
15:54:00 Nick topik_ is now known as topik (
15:56:19 Join CGL [0] (~CGL@
15:56:27 Join dfkt [0] (dfkt@unaffiliated/dfkt)
15:56:33 Quit dfkt (Client Quit)
15:59:34 Join dfkt [0] (dfkt@unaffiliated/dfkt)
16:00:26 Part LinusN
16:06:07 Part xsteadfastx
16:15:04 Join Barahir [0] (
16:18:08bjoernbhow does sansa clip+ work with linux?
16:20:45ThomasAHbjoernb: perfectly, even special characters (umlauts, japanese characters) work with OF and rockbox if you mount it with utf8 mount option (usb mass storage)
16:20:55 Join kugel [0] (~kugel@rockbox/developer/kugel)
16:27:44literalyeah, no problems at all in Linux
16:29:11bjoernbdo i really need rockbox, or does it work out of the box?
16:29:23literalyou don't need rockbox
16:29:44literalin fact, rockbox support for the Clip+ is still experimental
16:29:50gevaertsif you don't want rockbox though, you're better off somewhere else...
16:33:22literalbjoernb: the only Linux-specific thing you need to know about is that if you use the player in MSC (mass storage) mode, which is recommended in Linux, you should put a file called .is_media_player on it so that programs like Banshee know it's a media player and not just some usb stick (see
16:33:34literalsorry, I meant .is_audio_player
16:34:21literalbut yeah, this is outside the scope of this IRC channel
16:38:55bjoernbliteral: thanks.
16:39:16 Join jgarvey [0] (
16:39:29bjoernbi was just curious to buy such device, when i have it i still can install rockbox on it. thanks so far.
16:40:09 Part bjoernb
16:44:12 Quit pamaury (Quit: Quitte)
16:44:39 Join DV [0] (~DV@
16:59:34***Saving seen data "./dancer.seen"
17:06:08 Quit DV (Read error: Connection reset by peer)
17:11:14 Quit TheSeven (Ping timeout: 265 seconds)
17:11:26 Join anewuser [0] (anewuser@unaffiliated/anewuser)
17:13:07 Quit togetic (Ping timeout: 248 seconds)
17:21:25 Join Blue_Dude [0] (~chatzilla@
17:22:42Blue_DudeJust a heads up. I plan to commit FS #11081 this evening if there are no more objections. I think I've addressed the main issues. So far only the e200 target is affected. Other targets can follow very easily, just need some work on the keymaps.
17:23:29Blue_DudeI'll come back later and check the log and the mailing list in case anyone raises another issue.
17:23:33 Quit Blue_Dude (Client Quit)
17:24:33 Join DV [0] (~DV@
17:24:34 Quit DV (Remote host closed the connection)
17:25:22*Llorean thinks keymaps for as many targets as possible should go in _before_ the commit
17:25:28LloreanRather than just hoping they'll get done later.
17:32:25 Join arun [0] (~arun@unaffiliated/sindian)
17:47:37 Nick fxb is now known as fxb__ (
17:56:57 Join mitk [0] (
17:59:48 Quit bluebrother (Disconnected by services)
17:59:49 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
18:00:37 Quit Bagder (Quit: It is time to say moo)
18:12:40 Join komputes [0] (~komputes@ubuntu/member/komputes)
18:15:51 Join pamaury [0] (~c2c7a50a@rockbox/developer/pamaury)
18:16:34 Join captainkewllllll [0] (~2669ecc2@gateway/web/freenode/x-yyuiuvxblnujmbkt)
18:26:58 Quit S_a_i_n_t (Ping timeout: 260 seconds)
18:29:35 Quit perfectdrug1 (Quit: Leaving.)
18:29:53 Join S_a_i_n_t [0] (S_a_i_n_t@
18:32:54 Join TheSeven [0] (~theseven@rockbox/developer/TheSeven)
18:41:52 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
18:46:32 Join perfectdrug [0] (
18:53:55LloreanTorne: Don't most/all Apple chargers charge at 1000 isntead of 500, actually?
18:54:02 Join toffe82 [0] (~chatzilla@
18:54:04TorneThey are *rated* for 1000
18:54:19TorneBut as far as I have been able to tell I don't think any of the PP ipods can actually draw that much
18:54:20LloreanBut the OF doesn't draw that much?
18:54:33TorneI don't htink they have a GPIO hooked up to the relevant pin of the ltc charge chip
18:54:37*Llorean 's HTC chargers are also rated at 1000.
18:54:39TorneOr on some of them it may not even have such a pin
18:54:49LloreanI'm pretty sure they actually do charge at that though, my phone charges much faster there than hooked up to a PC
18:54:52TorneThe iphone/ipodtouch/etc I am *reasonably* sure *do* draw 1000mA
18:55:06Torneand htey *require* the USB 2.0 charger id method: data pins shorted together
18:55:22Tornewhich older ipods do not, they are happy with ancient apple chargers which just pull them high/low through weird resistors
18:55:31LloreanDoes the USB 2.0 charger spec specify how much a charger should be rated for?
18:55:43Tornewell sort of
18:56:22Torneshorting D+ and D- means it's a dedicated charging port
18:56:36Tornedevices are allowed to draw 1.8A from such a port
18:56:47Torneat 5.25V if they want :)
18:57:12Tornebut the general assumption in the industry seems to be "well we'll just turn our regulators off and the charger will not allow us to explode it"
18:57:23Tornewhich is probably fairly resonable and sane in practise :0
18:58:22Torneif the charger has bothered to short the data pins it's likely it is also not going to let the attached device melt it.
18:59:36***Saving seen data "./dancer.seen"
18:59:40Torneanyway, i don't know of a way to make any of the PP ipods draw anything other than 100 or 500
18:59:41LloreanNot that I'm terribly concerned about the iPod charging situation. The only real charging problem I ever butt my head against is one that's rather unlikely to be changed.
18:59:50Torneso for now that's not that important :0
19:00:32LloreanSince we can't do anything about the beast's OF charger detection during boot.
19:00:38 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
19:01:53Tornewell maybe we will one day ;)
19:02:20LloreanOnce we can reflash it, maybe.
19:02:40LloreanIf the battery's low, it'll boot if you have the actual charger connected, but not if you're trying to power it from USB. =/
19:03:22LloreanAnd it doesn't charge from USB while off, so unlike a lot of the other devices, you can't just make sure you've got a USB cable on hand.
19:04:15Tornewell we *can* reflash it now, i'm pretty sure it's a normal CFI nor part
19:04:25Tornewouldn't take five minutes to code a flashing plugin for it ;)
19:04:35Torneand yeah, i know that one
19:04:45Torneit's on my list of "things to find out about the bootloader when i have infinite time" :)
19:08:58Tornewell, i have gotten fairly far in the beast rom disassembly
19:09:08Tornejust not been doing it for a while, it's been pushed fairly far down the stack.
19:12:19CIA-5New commit by funman (r25381): as3525: CGU_USB has only 3 bits for the divider
19:19:41 Quit komputes (Ping timeout: 260 seconds)
19:20:35 Join Horscht [0] (~Horscht2@xbmc/user/horscht)
19:20:52 Join togetic [0] (~togetic@unaffiliated/ibuffy)
19:23:18 Join Buschel [0] (
19:28:20 Quit _silentAssassin (Ping timeout: 245 seconds)
19:31:05 Quit stoffel (Remote host closed the connection)
19:35:03 Join komputes [0] (~komputes@ubuntu/member/komputes)
19:35:36Buschelpixelma, Llorean, nls: If there are no more concerns I will submit fs #11148.
19:36:40LloreanIsn't spdif by default off?
19:38:20LloreanNot that it really matters or anything, the patch looks fine I think
19:40:12BuschelLlorean: default is off. but it should be referenced in the new section
19:41:23LloreanYes, it should
19:41:27 Join SuperMouton [0] (~chatzilla@2a01:e35:2439:f6b0:a81a:8446:3528:471e)
19:41:37SuperMoutonHello :)
19:41:47LloreanI just would've written it with a "don't turn it on unless you need it" tone rather than a "you should turn it off" style statement. It's not important at all though.
19:42:11SuperMoutonI have some really newbies questions to ask, can i ? :3
19:42:54bluebroth3rdomonoky, pixelma (and others interested): I'm planning to replace the rbutil icon on linux with a svg version. Unfortunately it seems I can't use different images for different sizes with svg (works fine with bitmaps) so that would mean to completele replace it. Which one looks better?
19:43:41 Join panni_ [0] (
19:43:46pixelmaSuperMouton: don't ask to ask, just ask (if it's about Rockbox)
19:43:51bluebroth3rusing the "Rb" icons is somewhat problematic for the small display (taskbar), but the clef looks a bit "empty" on big sizes
19:43:53Lloreanbluebroth3r: I personally prefer #2
19:43:55SuperMoutonOk ! :)
19:44:49bluebroth3ralong that, we could also consider replacing the "Rb" icon for large icons with the clef on w32 and mac to have it consistent (and not change the icon between sizes)
19:45:00domonokybluebroth3r: i like the Rb icon more, at least for bigger icons :-)
19:45:27 Join S_a_i_n_t_ [0] (S_a_i_n_t@
19:45:28domonokythe clef in big looks so empty.
19:45:37 Quit jgarvey (Ping timeout: 276 seconds)
19:45:39bluebroth3rdomonoky: I wanted to show the Rb one for large icons (i.e. 48x48 and up), but it seems QIcon doesn't handle this for non-pixmap icons :(
19:45:48 Join Strife89 [0] (~michael@
19:46:16domonokyusing the same icon always is a good idea.
19:46:39 Quit S_a_i_n_t (Ping timeout: 276 seconds)
19:47:06bluebroth3rwell, I kinda liked using a different (i.e. less detailed) icon for small sizes. Like currently on w32 for the explorer icon (though that is handled by the windows resource and not related to Qt)
19:47:11SuperMoutonI've installed today RockBox for my 5.5G iPod Video (80G). Is it possible to choose between RockBox or the initial Apple system when i boot my iPod ? I search a little in the forum but english is not my native language so it's pretty diffucult for me :3
19:47:23bluebroth3rotoh I would really prefer to use svg on linux.
19:47:37domonokySuperMouton: yes, by holding a buttón on boot. See manual :-)
19:48:13SuperMoutonThank you :)
19:48:30bluebroth3rSuperMouton: the primary source for documentation should be the documentation, not the forums :)
19:48:52 Quit mitk (Ping timeout: 276 seconds)
19:48:57SuperMoutonSorry :s
19:49:23 Join jgarvey [0] (
19:50:18domonokybluebroth3r: what happens if you set two different svgs for different icon sizes ?
19:51:05SuperMoutonWhat is the 'hold switch' ?
19:51:44domonokySuperMouton: the slider on the side (top?) of your ipod.
19:51:59SuperMoutonTo lock the ipod ?
19:52:12domonokyjup, thats the "hold switch"..
19:52:19SuperMoutonOk thank you :)
19:52:29domonokymaybe try to use google translate if you dont know words :-)
19:52:38bluebroth3rdomonoky: it simply uses the last one set.
19:53:14bluebroth3rSuperMouton: scroll a bit up on that page, there is an explanation of the buttons
19:53:28SuperMoutongoogle translate get me 'commutateur Hold', doesn't mean anything in french ^^
19:53:33SuperMoutonOk, thank you again !
19:53:42bluebroth3rplease do look first ... it's even on the very same page
19:54:08domonokybluebroth3r: bad. thats probably the svg icon engine doing it. There is a note in Qt docs, that they are free to ignore additional images.
19:54:40 Join krabador [0] (
19:54:46bluebroth3rdomonoky: yes, I assume that. Maybe I should ask in #qt to make sure. It works when I use a png image instead, but using svg + png doesn't
19:55:15 Join mitk [0] (
19:55:17bluebroth3rinterestingly if I use png images the order does matter :o
19:55:18domonokyor we just use the clef everywhere, if its too much work :-)
19:55:56domonokythat definitly points to the svg icon engine. we could write our own to solve that :-)
19:56:22bluebroth3rthere is also the option to use either clef or "Rb" only on linux. Windows and OS X are not affected by this (w32 gets the icon from the resource, and OS X from the bundle)
19:58:19CIA-5New commit by Buschel (r25382): Add new section 'Optimising battery runtime' to the manual (FS #11148).
19:58:34 Quit krabador (Quit: Sto andando via)
20:01:49domonokythen i would vote for only rb on linux..
20:02:58bluebroth3rany other opinions which one to use?
20:03:49 Quit Adubb (Read error: Connection reset by peer)
20:03:53SuperMoutonIs rockbox using more battery than the original system ?
20:04:35ZagorSuperMouton: generally, no
20:04:41BuschelSuperMouton: On an iPod Video you can suspass the runtime of the Apple firmware.
20:05:14SuperMoutonOk, thanks.
20:06:25 Quit kugel (Ping timeout: 276 seconds)
20:08:40SuperMoutongot to go. thank you for you answers and have a nice and beautiful day :-)
20:08:43 Quit SuperMouton (Quit: ChatZilla 0.9.86 [Firefox 3.6.2/20100316074819])
20:17:12 Join Adubb [0] (~aldubuc@
20:19:30 Quit flydutch (Quit: /* empty */)
20:20:01 Quit mitk (Quit: Leaving)
20:22:05 Join Stephen__ [0] (~S@
20:22:08 Join {phoenix} [0] (
20:35:50 Quit Xerion (Quit: )
20:43:26 Join Makuseru [0] (
20:43:27 Quit CGL (Ping timeout: 276 seconds)
20:44:54MakuseruHi, I just got a Toshiba Gigabeat S with the intentions to put rockbox on, but I cant get my computer to recognize it. I'm running Kubuntu 9.10, and when i plug the player in it just says "connected" on the player, and "mtp-detect" returns the player, but it doesnt show up as an external drive like it should. Can anyone help me with what would be causing this?
20:45:49 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
20:46:03 Join panni_ [0] (
20:46:37gevaertsMakuseru: the gigabeat S won't show up as a drive. The original firmware on it is MTP only
20:47:02 Join xiainx [0] (
20:48:07Makuserugevaerts: well how do i go about installing it then? (This is my first Gigabeat S, but ive put rockbox on like three Gigabeat F's)
20:49:19gevaertsMakuseru: see tha manual,
20:52:15Makuserugevaerts: I downloaded beastpatcher, but it wouldnt install, it said "no devices found"
20:52:36gevaertsare you running it with root permissions?
20:52:59MakuseruShould I?
20:53:10gevaertsI suspect so, yes
20:54:01gevaertsthe manual should say this...
20:54:10*gevaerts grumbles and heads of to fix the manual
20:54:25 Quit toffe82 (Read error: Connection reset by peer)
20:55:11Makuserugevaerts: isnt it a pretty bad idea to run wine as root though/
20:55:32AlexPMakuseru: Why not run the linux binary?
20:55:46gevaertsthe wine binary isn't going to work on linux...
20:56:33AlexPgevaerts: Are you sure it needs root, I didn't think so
20:56:34MakuseruHaha, i didnt see the linux version, i just saw the windows one.
20:56:48AlexPI think the problem here is trying to use a binary for the wrong platform
20:57:00gevaertsAlexP: maybe not. I'm not sure now...
20:57:59MakuseruWell that worked just fine.
20:58:05gevaertsgood :)
20:58:20MakuseruThe Gigabeat is displaying "Bootloader USB mode" now and it just popped up as a volume.
20:58:30MakuseruI appreciate you guys and all the work you put into Rockbox.
20:58:45AlexPWait till you get it running :)
20:59:01MakuseruIve been using Rockbox on a Gigabeat F for years now.
20:59:08MakuseruI absolutely love it.
20:59:22MakuseruI just got a Gigabeat S and it not showing up as a drive was just messing with me.
20:59:36AlexPI mean until the S actually boots, it can occasionally be a bit temperamental
20:59:37***Saving seen data "./dancer.seen"
20:59:51gevaertsAlexP: "until"? You mean it stops then? :)
21:00:10AlexPIt stops booting when it has booted, yes :)
21:03:17saratogai want to tell that guy on the mailing list that adding a jump to time feature in rockbox really isn't useful enough to go into SVN
21:03:22saratogabut i'm not sure others would agree
21:03:38gevaertssaratoga: if you add "I think", you're fine :)
21:03:49saratogado we have a policy on adding new features to playback/gui that aren't generally useful to most people
21:03:54saratogabut might be to a few
21:04:28gevaertsFrom looking at the list of features we have, I seriously doubt if there's such a policy
21:04:29saratogait seems like we should probably have some agreement on what can be added so we don't keep ending up with a mess of confusing and not widely useful features (e.g. disable track skip)
21:05:35evilnick_Bsaratoga: Maybe something for the nodo chat at DCE?
21:06:03 Quit Makuseru (Quit: CGI:IRC (EOF))
21:06:11notlisteninge200v2 stop it booting to the OF when power is plugged in, what do i need to modify?
21:06:37gevaertsnotlistening: the best way is to implement USB support
21:06:41 Join wodz [0] (
21:07:55 Join User67703 [0] (
21:08:14notlisteninggevaerts, lol thats what everyone want but iknow during the development it was default to boot rockbox, and it got changed but i just want this for myself as i have an external power supply that i want to use on the move
21:08:17User67703rockbox support ntfs?
21:08:30gevaertsUser67703: no
21:08:55gevaertsnotlistening: I'm pretty sure that the default on AMS sansas on USB plugin has *always* been to boot the OF
21:09:14saratogait didn't get changed, its always been like this
21:09:36notlisteninglol it was changed i know :D
21:09:53 Join Xerion [0] (
21:10:01gevaertsYou know more than we do then. When was it changed?
21:10:20notlisteningwould it have been funman that did the work or atomik?
21:11:12 Join kugel [0] (~kugel@rockbox/developer/kugel)
21:11:51notlisteningif i find it in the forums i will let you know
21:11:53wodzhello, I would like to rise again bug in scramble.c. Well it is not a problem but rather "not good programing practise". 1) variable modelname is not initialized and 2) later we do strcpy to it. If the name of the model is longer than 4 chars bad things will happen.
21:12:50 Join jfc^2 [0] (
21:13:13wodzthe fix is simply initialize modelname (line 148) and do strncpy instead of strcpy (line 336)
21:13:59gevaertswodz: it should use strncpy(), yes. Why does it need dummy initialisation though?
21:15:43wodzgevaerts: suppose we will have 5 chars name (by accident or intentional). After strncpy(modelname, &argv[1][5],4) we may end up with not null terminated string
21:16:17wodzit is safer to initilaize modelname to all nulls
21:16:36 Quit jfc^3 (Ping timeout: 276 seconds)
21:16:38gevaertswodz: use strlcpy then :)
21:17:23kugelwodz: or (the more common practise) set modelname[4] = '\0'
21:17:29kugelafter strncpy
21:17:48gevaertsOr use sprintf
21:17:51wodzall this will do
21:18:04kugelnot snprintf? :)
21:18:10gevaertsno need
21:19:16gevaertssprintf(modelname,"%.4s",whatever) is guaranteed to be bounded
21:19:36saratoganotlistening: r18767 already had the USB detect for the e200v2 present, so if it changed it was probably early fall or maybe summer of 2008 when it happened
21:19:55saratogaalthough i think the e200v2 port did not really exist back then
21:20:12kugelwasn't it changed recently?
21:20:12 Join DV [0] (~DV@
21:20:23kugelgevaerts: ah, right :)
21:21:22wodzanyway should I craft FS patch for such trivial fix or some of You will commit it?
21:23:08kugelFS is preferred, generally
21:24:45 Quit xiainx (Read error: Connection reset by peer)
21:25:06 Join xiainx [0] (
21:26:18notlisteningI agree that it was quite some time ago and when the port was in its infancy but i know it was changed in the bootloader
21:26:39bluebroth3rgevaerts: beastpatcher doesn't need to run as root.
21:26:54*bluebroth3r wonders why people try to run windows binaries with wine :/
21:27:16 Quit DV (Read error: Connection reset by peer)
21:27:20gevaertsbluebroth3r: well, that bit is obvious. The question is why they want to run windows binaries at all :)
21:27:29bluebroth3rremembering those guys that were trying to run rbutil on wine and claimed the autodetection not working at all :o
21:28:00bluebroth3rgevaerts: we should stop distributing windows binaries. Those windows users should be able to compile themselves :)
21:28:04 Join DV [0] (~DV@
21:29:16AlexPbluebroth3r: You're a cruel man :)
21:29:25gevaertsbluebroth3r: can't you detect wine in some way?
21:29:31bluebroth3rAlexP: why? Building on Windows isn't hard at all :)
21:30:11bluebroth3rgevaerts: no idea, but if so we could add a "do not run this on wine" and exit(-1) to all out tools :)
21:30:12archivatorIs anyone actually using festival with rbutil? I tried it last night but it was acting up (typical) and did not change the voice..
21:30:45 Join funman [0] (~fun@rockbox/developer/funman)
21:30:47wodzFS #11151
21:30:57*bluebroth3r only uses the OS X TTS
21:31:12archivatorlucky you..
21:31:21archivatorIt's actually quite well done
21:31:42bluebroth3rthe OS X one? Yes, the quality is amazing (at least on 10.6)
21:31:59notlisteningfunman, you might know this changes to the e200v2 firmware to boot to the OF not rockbox was that your work?
21:32:14funmanno it's not
21:32:20bluebroth3rwhich unsurprisingly is the reason why I use it on OS X :)
21:32:42notlisteningarchivator are you intrested in better TTS on linux?
21:32:55archivatorbluebroth3r: some of us haven't paid our Apple tax :(
21:33:05archivatornotlistening: I can get pretty decent TTS on linux
21:33:12archivatorI'm interested in TTS on rockbox :)
21:33:14funmannotlistening: if you want to boot rockbox, boot if before plugging USB and keep select button pressed
21:33:19pixelmahow was it "acting up"?
21:33:28notlisteningoh i have SAPI running under Linuz that's all
21:33:39archivatorpixelma: well, it took several tries to actually start and then it hadn't changed the voice.
21:33:46domonokyand there is the opensapi patch for rbutil in the tracker :-)
21:33:57*pixelma is off though for a while... so maybe shouldn't have forced some discussion
21:34:00notlisteningfunman your the m,man as normal :P
21:34:06archivatorSAPI's no fun
21:34:15archivatorFestival is actually where the money is
21:34:20notlisteninglol domonoky just updated that for the latest svn
21:35:09notlisteningok archivator depends on your aim
21:35:41bluebroth3rindeed, its not really working as expected.
21:35:49*funman listens to FM on his Clip+
21:36:48ThomasAHfunman: wow!
21:37:04gevaertsfunman: can you save the preset though? :)
21:37:17ThomasAHgevaerts: that's what the simulator is for :)
21:37:18archivatorI didn't want to sound dismissive, so I'll clarify what I mean and will end this OT - festival allows for unit selection voices (including something called multisyn, developed by AWB and KAL), while SAPI voices are generally diphone voices. True, there are some amazing diphone voices (i.e. the OS X default voice) but generally a good unit selection voice sounds more natural and fluent.
21:39:28notlisteningwhen did the clip+ get FM?
21:39:55ThomasAHnotlistening: the "how" is more interesting :)
21:40:18ThomasAH(unless the answer is "by booting OF")
21:40:24AlexPnotlistening: Just now I imagine
21:40:30bluebroth3rarchivator: interesting, I didn't look into much details about TTS −− just implemented support for OS X TTS because it sounds great (compared to the other TTS I've heard before).
21:40:35AlexPThat is what funman was telling us
21:40:56bluebroth3rI'll try to look into festival support the next couple of days. No promises though.
21:41:53archivatorbluebroth3r: I actually wrote the festival patch so I'll try and figure it out tonight :)
21:42:26bluebroth3rarchivator: even better (was already wondering who wrote that patch but didn't look it up yet ;-)
21:42:49*bluebroth3r wants automated tests for rbutil
21:43:20archivatorqt has an entire unit test framework
21:43:27archivatorEven gui scripting
21:43:32 Join spaax [0] (
21:44:55bluebroth3rI know. The problem is just that someone needs to implement tests ;-)
21:45:14*bluebroth3r started looking into this a couple of days ago but it's just the beginning
21:45:59 Quit Buschel (Ping timeout: 268 seconds)
21:47:05archivatorgiven that some things can't be tested without a DAP, what do you feel needs unit test coverage?
21:47:31bluebroth3reverything that can get tested without a DAP :)
21:48:06archivatornot much - the extras pretty much cover it all
21:48:20CIA-5New commit by funman (r25383): as3525v2: FM playback (only tested on Clip+)
21:48:30 Quit xiainx (Quit: Good Bye)
21:48:35funmansomeone can test this commit on Clipv2 ?
21:49:00bluebroth3rOne thing I'd like to see is the bootloader installation classes that don't need a DAP. Especially file and mi4 installation methods. I'm still thinking that the installation problem the current release has (on h10, probably mr100 and similar too) was triggered by a (correct) change in Qt behaviour.
21:49:13bluebroth3rbut without any tests we won't notice that.
21:49:59notlisteningfunman u want another clip+ for verification?
21:50:13funmannotlistening: sure
21:50:16notlisteningnot doubting your work there ;)
21:50:52funmanthere's no doc for this part so i could have forgotten a bit
21:51:59kugelfunman: do you have problems connecting the fuzev2 as well?
21:52:14funmanit *seems* i can lower the FM volume much more than fore file playback, not sure since it isn't muted completely
21:52:17funmankugel: connecting?
21:52:26kugelto the pc
21:52:49*kugel gets "hub 3-0:1.0: unable to enumerate USB device on port X" 50% of the time
21:52:52funmannope but i had problems with clipv2/clip+ iirc
21:53:11kugelah the backlight is nice
21:53:23funmanit was depending from the direction of the wind, so i closed the window and now it's ok
21:53:46kugelmy window is closed
21:53:52notlisteningbuilding...... yawn
21:54:00 Join toffe82 [0] (~chatzilla@
21:54:01funmantry to open it :P
21:54:09 Join aCiD2 [0] (
21:54:15funmannotlistening: why not taking ?
21:54:33*kugel works on buttons now
21:54:46notlisteninglol because i did not know if it :D
21:54:51funmanhow can i add fuzev2 to the binsize delta table ?
21:54:54notlisteningI am so out of touch
21:55:49CIA-5New commit by bluebrother (r25384): Remove OSX deployment script, Qt now has macdeployqt.
21:55:53CIA-5New commit by bluebrother (r25385): Add Rockbox Utility SVG window icon on Linux. ...
21:55:56kugelfunman: isn't "(swap16(pixels >> 16) << 16) | (swap16(pixels & 0xffff))" == "swap32(pixels)"?
21:56:17kugelwell, when I tried that it didn't work
21:56:33funmanno, swap32(abcd) => dcba, we need badc
21:56:55kugelah, right
21:57:13 Join bieber [0] (~bieber@
21:57:17funmanperhaps there is a more efficient method to do that
21:57:40CIA-5New commit by bluebrother (r25386): Set missing svn:keywords property.
21:58:26 Quit shaggy-h ()
21:59:06 Quit S_a_i_n_t_ ()
22:00:16topikradio works nicely on my clip+, funman
22:00:55 Quit Strife89 (Quit: Leaving work.)
22:03:41 Quit bmbl (Quit: Bye!)
22:04:24 Join shaggy-h [0] (
22:04:29topikaww, pressed too many buttons now it froze
22:06:18funmannext thing to find out would be if/why charging works or not
22:07:20topikcharging _icon_ works
22:07:46Stephen__how did you get radio working ?
22:08:46funmantopik: if you check in the debug screen > battery , does the % oscillate between 75 and 81% ?
22:09:29kugelfunman: you took stuff out of the lcd init (the ssp stuff)?
22:09:30topikmine is nearly fully charged
22:10:01funmankugel: yes, it's not needed
22:10:25kugelhave we found someone with the other lcd type (if it exists)?
22:10:26funmani hope people test it and tell us if their screen doesn't work, if it's the case perhaps we'll have to put it back
22:10:58funmantopik: what's the % with USB unplugged?
22:11:29topiksteady at 92
22:12:08 Join moos [0] (moos@rockbox/staff/moos)
22:12:22funmanbtw someone on ABI forum said the Clipv2 has ~13hours runtime with rockbox
22:12:31funmanwithout CPU freq scaling
22:13:04kugelmaybe we just run very low?
22:13:10topik13 hours would be quite impressive
22:13:39funmankugel: yes i think it's quite low, according to test_codec
22:13:51archivatorbluebroth3r: can you please send me the output from "festival −−version" and "festival_client −−version" on your machine?
22:14:13archivatorIt seems that either they changed something or it worked by pure coincidence before.
22:14:15funmanCGU_PROC = 0x11 on my Clip+
22:14:40topikfunman: the battery number on the first battery debug screen seems to decline in a realistic way
22:15:11funmanthat's clkin=PLLA(384MHz), postdiv=1 => fclk = 192MHz if we interpret like as3525(v1)
22:15:12topikand was increasing when i had it plugged in
22:16:09funmantopik: it's ok when not charging (it goes down), but when charging it goes low & high
22:16:32funmanbefore I changed the ADC source we read, it went to 100% when charging
22:16:45topikyeah, somewhat erratic
22:16:54funmanI applied the same workaround used for e200(v1) and now there's this problem
22:17:02kugelworking backlight is annoying without buttons!
22:17:14funmankugel: yeah, finish the buttons first! :)
22:17:56topikthe startup time of the clip+ is quite amazing too
22:18:49bluebroth3rarchivator: festival: Festival Speech Synthesis System: 1.96:beta July 2004 and festival_client: Edinburgh Speech Tools Library v1.2.96:beta July 2004
22:19:45 Quit bieber (Quit: Leaving)
22:19:59 Quit {phoenix} (Remote host closed the connection)
22:20:21bluebroth3rarchivator: that's the F12 packages in case its relevant
22:20:35 Quit amiconn (Disconnected by services)
22:20:37 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
22:20:59 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
22:21:08notlisteningdomonoky, when using the opensapi patch and creating the voice files the yes and no entries are not created and various others are also missed any ideas on this, it reports errors but this is due to a response from the server it is not expecting but this is a valid reponse does that affect rbutil missing that entry?
22:21:16archivatorbluebroth3r: I should have guessed, i'm running F12 as well :)
22:21:37archivatorbluebroth3r: I *think* something changed in festival_client - it resets everything when it connects to the server
22:22:05 Quit pixelma (Disconnected by services)
22:22:07 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
22:22:27 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
22:22:40domonokynotlistening: yes, those errors are probably why it doesnt generate the files.
22:23:59notlisteningok i will have a look and see if i can edit to code to accpt that as a valid reponse
22:23:59CIA-5New commit by funman (r25387): pitch_detector plugin only works on color LCDs
22:24:04domonokynotlistening: if the server response is valid and the requested wav file is correctly generated, you should tweak the opensapi part in rockbox to not treat that as error :-)
22:24:04archivatorIt would be nice if someone with a distro != F12 (preferably something older) would check if festival support in rbutil works
22:24:50notlisteningarchivator, ypu wish is my command
22:25:12archivatornotlistening: It takes some trial and error
22:25:38notlisteningoh any clues of should i ask as i struggle
22:25:51archivatorActually, what are the festival[_client] versions? No point in checking if they're the same
22:26:16archivatorfestival −−version, festival_client −−version
22:26:51notlisteningfestival: Festival Speech Synthesis System: 1.96:beta July 2004
22:27:13notlisteningfestival_client: Edinburgh Speech Tools Library v1.2.96:beta July 2004
22:27:20archivatoryeah, looks like the same version..
22:27:32funmanhm weird my Clip+ doesn't turn on (2nd time i see that this evening)
22:27:45archivatorit's really weird, I swear that it worked when the patch was submitted!
22:27:58notlisteningoh ok thats what my distro ships with
22:28:34 Quit Zagor (Quit: Clint excited)
22:28:35*Unhelpful wonders why one couldn't detect pitch without a color display?
22:28:36archivatordomonoky: do you remember changing any of the functionality when you converted the TTS settings?
22:29:11domonokyarchivator: i think not. and if i changed anything with festival, i surely tested it.
22:29:55archivatorthat's too bad. I have no idea what broke it. I know how to fix it but it's not pretty
22:30:06funmanUnhelpful: no problem, one just needs to adapt pitch_detector.c to monochrome screens
22:30:23notlisteningEncoding of /tmp/rbvoice//LANG_SET_BOOL_YES.wav failed
22:31:42domonokynotlistening: can you find out why that failed ? is that wav actually generated ?
22:33:14notlisteningGenerates one file then does not do anymore
22:34:35notlisteninghaha my engine also fails on the yes clip :D
22:35:33bluebroth3rarchivator: tried it on F8. Festival reports the same version but now it works. It's a bit flaky though (need to open the configure tts box two times)
22:35:46linuxstbgevaerts: What do you (and others?) think about putting "bootloader menu" on the NoDo ? It's certainly been called a no-do in the past, but I don't know current opinion (I'm not strongly against).
22:36:23bluebroth3rthough on F12 it doesn't get a voice list. Running festival_client manually works though.
22:36:24archivatorbluebroth3r: I think I just fixed the two times thing
22:36:26*linuxstb wonders how "Alphabetical list skipping" got on that list...
22:36:31archivatorHowever, why it works is beyond me
22:36:40AlexPI'm against if it a) slows anything down at all and b) if a is not true then it should not be default
22:36:48notlisteninglinuxstb, , from an accessibility point of view umm no from me
22:37:13archivatorbluebroth3r: by "works" you mean it changes the voice successfully? Or that it just voices everything with the default voice?
22:37:35funmanlinuxstb: i think a bootloader menu would be nice (depending on how it's implemented however)
22:37:44Unhelpfulfunman: ah... drawing peculiarities?
22:37:47notlisteninglinuxstb because i suggested it :p
22:37:59gevaertslinuxstb: I think it should be on there. It doesn't matter much if you agree with it or not, if it's been traditionally considered as a NoDo, it should be listed so it can be discussed
22:38:16linuxstbgevaerts: OK. I don't want to add it though ;)
22:38:27funmanUnhelpful: probably, it might even be very easy to modify the plugin. I won't look at it before i can record from the Clip though
22:38:39wodzwhat are the objections with menu in bootloader?
22:38:47AlexPSlows things down
22:38:50bluebroth3rarchivator: open "configure tts", verify paths. Ok, then open again. Now the refresh buttons for voice list works, and I can select a voice. Ok again, then "Test TTS" -> system talks to me.
22:39:05AlexPIf it is optional, doesn't slow down boot, and is off by default then I don't min too much
22:39:07notlisteninglinuxstb, it was boohood becuase of missing components on rockbox to implement it
22:39:07*bluebroth3r goes installing more voices
22:39:29archivatorbluebroth3r: using the selected voice?
22:39:38gevaertswodz: that's actually the point of the discussion. With many of these items, nobody remembers...
22:39:43archivator'Cause it talks here as well, using kal_diphone (argh!)
22:40:13 Quit efyx (Remote host closed the connection)
22:40:14bluebroth3rarchivator: I have only one voice installed yet.
22:40:27linuxstbAs AlexP said, there's the potential to slow things down, especially on hard-disk targets. Plus also the argument that bootloader code should be as simple as possible.
22:41:04*linuxstb didn't want to start the discussion again now though...
22:41:15archivatorbluebroth3r: please do install some more. The major bug atm is that it doesn't change the voice.
22:41:51linuxstbIs it possible to see diffs on foswiki older than the first 6 or so? (without manually editing the URL, and without viewing the entire history?)
22:41:51wodzslow things down by means of timeout? or what?
22:41:54bluebroth3rarchivator: just was at it :) It speaks but doesn't change the voice. That's on F8.
22:42:03ThomasAHhow do I make "Mode: preset" the default for FM radio instead of "Mode: scan"? (on clip+, so I have to manually edit config.cfg)
22:42:06linuxstbwodz: Slow things down by needing to read a config file from disk.
22:42:28funmanThomasAH: you can edit the settings easily in the simulator
22:42:32archivatorbluebroth3r: weird and weirder. i was using Ubuntu when I wrote that patch and it's possible festival_client is different there..
22:42:35*linuxstb is just the messenger...
22:42:54wodzlinuxstb: unless You do not use it.
22:43:09notlisteningcan you not have a key combination for the bootmenu so it only read a config when it is called?
22:43:18AlexPwodz: It has to read whether you use it or not
22:43:36gevaertslinuxstb: yes, that's the problem with listing NoDos...
22:44:01ThomasAHfunman: the simulator somehow defaults to preset :-/
22:44:02wodzI written simple bootloader with menu for mpio hd200 and I do not use config at all
22:44:05AlexPnotlistening: In that case, just use the existing key configs
22:44:34funmanThomasAH: check out the manual to edit the configuration in the simulator, copy simdisk/.rockbox/config.cfg to your device
22:44:36linuxstbgevaerts: But definitely a good idea to bring them all up for discussion at the next devcon.
22:44:36bluebroth3rarchivator: the strange thing is that on F12 I don't get a list of voices at all even though I now have several voice installed
22:45:12wodzIf current bootloader supports 2-3 modes of running code why not present the choice to the user?
22:45:24CIA-5New commit by bluebrother (r25388): Set svn:mime-type property.
22:45:46ThomasAHfunman: This is what I did before, but some things behave differently ... e.g. backlight on instead of 15 seconds ... and radio preset mode seems to behave differently, too
22:45:54notlisteningfunman: FM radio working here
22:46:12archivatorbluebroth3r: open up festival and type (voice.list)
22:46:18archivatorDo you get the voices there?
22:47:09ThomasAHfunman: besides that little point radio works great, thanks
22:47:19bluebroth3rarchivator: I get a list with 7 entries
22:47:34archivatorAlso, I found another bug - festival will exist with an error if the destination file exists - and rbutil does not even check if the destination file already exists (i.e. so that it continues an interrupted process)
22:48:13archivatorbluebroth3r: keep trying various combinations of closing the config window and hitting refresh :) I think I fixed that in the local branch but can't be sure atm
22:49:51 Quit mt (Read error: Connection reset by peer)
22:50:43 Join mt [0] (~mtee@rockbox/developer/mt)
22:51:31archivatorWhat version of Qt is rbutil dependent upon? 4.5 or 4.6?
22:51:48*ThomasAH has to leave and will investigate that preset mode thing tomorrow
22:51:55notlisteningfunman after i exit radio on the clip+ it still plays and the voice will not work for the menus? Just to let you know.
22:52:46 Quit Hillshum (Remote host closed the connection)
22:53:30funmannotlistening: hm i saw that, let me look at it
22:53:36funmando you have a clipv1 ?
22:54:01archivatorDo we care if the RockboxUtility GUI freezes for a second while the festival process is starting?
22:54:17funmanperhaps it's just a keymap problem: pressing power on Clipv1 leaves the FM menu but it keeps playing
22:55:01archivatornever mind, that's not gonna happen..
22:55:40 Join mt_ [0] (~mtee@
22:57:28 Quit mt (Ping timeout: 245 seconds)
22:57:46 Nick mt_ is now known as mt (~mtee@
22:58:12notlisteningbrand new and think it is the clip+ v 01.02.09A
22:58:40 Join Strife89 [0] (
22:58:51notlisteningfunman, how else should i exit the radio properly?
22:58:59bluebroth3rarchivator: you need at least Qt 4.5
22:59:17notlisteningfunman, probably just a button config issue
22:59:38***Saving seen data "./dancer.seen"
23:00:02 Join Darkknight512 [0] (
23:01:36 Part spaax
23:01:38funmannotlistening: long press on button up (play)
23:01:39archivatorbluebroth3r: thought so.
23:01:52bluebroth3rarchivator: interestingly my debug build prints "Festival failed to start" all the time (on F12 again)
23:02:09archivatoryeah, it does that, I'm trying to debug it right now
23:02:48archivatordomonoky: I have a question regarding the life of the TTS* objects - are they created once and kept in a cache or is it constantly creating and deleting them?
23:03:42domonokyarchivator: i am not sure, have to check the code.
23:03:43 Quit mt (Ping timeout: 245 seconds)
23:04:11notlisteningRoght i will wait for the manual before i start asking stupid questions again
23:04:35archivatordomonoky: will do, thanks
23:05:34domonokylooks like they are just created when needed.
23:06:04archivatorThe thing is, I'm not 100% sure the stop() method is called all the time
23:06:13 Join mt [0] (~mtee@rockbox/developer/mt)
23:06:34domonokynot between every file, but for every complete run, or if you open the config dialog.
23:06:38CIA-5New commit by gevaerts (r25389): Fix strcpy from a user-supplied string to fixed size string. ...
23:06:48funmannotlistening: manual should be up
23:06:52archivatorI'm just worried there might be a memory leak somewhere
23:07:02 Quit jgarvey (Quit: Leaving)
23:07:04archivatorWhere a TTS* object is allocated but not deleted.
23:07:15notlisteningdomonoky, calling on your memeory now what is the false on the end of this: TTSOpenSapi::eFILENAME,"outFile",202,false
23:07:19funmanarchivator: you know about valgrind to detect/debug memory leaks?
23:07:34 Join shai [0] (
23:07:35archivatorfunman: I do. I also know of the horror of debugging Qt apps
23:07:59domonokynotlistening: cant remember.
23:08:05 Join mt_ [0] (~mtee@
23:08:11notlisteninghaha i will change it and see :D
23:08:26notlisteningi have a good idea what it is for ;)
23:10:21 Quit mt (Ping timeout: 240 seconds)
23:10:27 Join mt [0] (~mtee@rockbox/developer/mt)
23:10:39domonokynotlistening: memory management with Qt is a bit difficult, most objects are allocated with parent ptrs, and if the parent dies, the object dies too.
23:10:59domonokythe tts objects should all have a valid parent set (i hope).
23:12:21 Quit mt_ (Ping timeout: 240 seconds)
23:12:45 Join ivortheengine [0] (~ivortheen@
23:13:11archivatordomonoky: any particular reason why the path to the festival server is passed on as an argument to every single method? I don't think that was there last time I checked :)
23:13:48domonokyarchivator: i also cant remember :-)
23:14:05archivatordomonoky: I think I know why. But it's really poor design :)
23:14:37 Join stripwax [0] (
23:14:54 Quit wodz (Quit: Leaving)
23:15:15 Quit pamaury (Quit: Page closed)
23:15:18archivatorIn fact, is there any other source for the path other than RbSettings?
23:15:19 Join mt_ [0] (~mtee@
23:15:38 Quit kunal (Quit: Ex-Chat)
23:17:05notlisteningdomonoky, umm it is stange but it only fails for certain entries yes being on the them it always fails for but gives no error message werid but sure how to go about sorting that out :D
23:17:21 Quit mt (Ping timeout: 240 seconds)
23:17:32gevaertsnotlistening: it's in a negative mood?
23:17:36 Join Tomis2 [0] (~Tomis@
23:17:52domonokydoes "no" work ? :-)
23:18:07 Quit Tomis (Read error: Connection reset by peer)
23:18:07 Nick Tomis2 is now known as Tomis (~Tomis@
23:18:22notlisteningyes ;) but let me generate and check it
23:19:07domonokyso does opensapi fail for this, or only the rbutil client ?
23:20:12 Join fml [0] (
23:20:15notlisteningdo you know where yes comes in the list of generated files?
23:20:20 Join mt [0] (~mtee@rockbox/developer/mt)
23:20:21 Quit mt_ (Ping timeout: 240 seconds)
23:21:48domonokynotlistening: if you look into the system trace you should see which entrys came before yes.
23:22:08archivatordomonoky: what's the difference between getSetting() and RbSetting::subValue() ?
23:22:17notlisteningokay need to setup to debug again
23:22:30archivatorit all seems awfully convoluted
23:22:32domonokynotlistening: systemtrace also works in release
23:22:37funmanhm it seems that i can get recordign to work on Clip+ (FM at least), but can't start recording due to keymap problems
23:22:55 Quit fml (Client Quit)
23:23:08domonokywe redirect qdebug into our buffer, and you can look at it under Troubleshoot->Systemtrace (or similar)
23:23:08 Join Schmogel [0] (
23:23:48saratogaah gsoc applications are open today
23:23:49domonokyarchivator: in a tts engine ?
23:23:49 Join Kitr88 [0] (Kitr88@
23:24:02archivatordomonoky: yes
23:24:31notlisteninghow do i get systemtrace on domonoky? Sorry for being dumb
23:24:34 Join fml [0] (
23:24:45domonokynotlistening: its in the menu of rbutil.
23:24:54 Join blairb [0] (
23:25:06funmanhelp menu
23:25:06fmlgevaerts: re your last commit: doesn't strlcpy do exactly what the two committed lines do?
23:25:11*domonoky doesnt have a systemtrace, only rbutil has one :-)
23:25:11 Quit captainkewllllll (Quit: Page closed)
23:25:29gevaertsfml: is strlcpy available on all systems we want to build on?
23:26:13domonokyarchivator: RbSetting::subValue() gets a value from permanent storage. not sure about getSetting() can you point me to a place where its used ?
23:26:14funmanit doesn't exist on glibc afaict
23:26:36 Quit Kitar|st (Ping timeout: 276 seconds)
23:26:41archivatordomonoky: see TTSFestival::updateVoiceList() for example
23:26:50archivatorQStringList voiceList = getVoiceList(getSetting(eSERVERPATH)->current().toString());
23:27:07archivatorDoes that use the immediate settings (as in, not yet saved)?
23:27:39domonokyah, yes. that are the settings from/for the ui, the config dialog.
23:27:57Stephen__funman, FM working here too
23:28:19archivatorBuuuuut, if the settings have been saved, can I call getSetting() to get what RbSettings::subValue() would give me?
23:28:53archivatorIn other words, can I use just one of the two and skip the other?
23:28:54 Quit Xerion (Quit: )
23:29:50archivatorI don't see a reason why TTSBase does not provide an abstraction above RbSettings ...
23:30:18domonokydepends one where. the saveSettings() function should get the UI settings and save it to permanent storage.
23:30:32domonokyall other places on a tts should use RbSettings.
23:30:37 Join anubisfremen [0] (~58f08ec4@gateway/web/freenode/x-trximybwsxtjgymh)
23:30:42archivatorexcept, they don't
23:30:56archivatorthat's why you hacked in the path argument to every single method
23:31:06domonokycan you point me to that ?
23:31:20archivatorthat's the way to provide the currently set path for when you get the voice list/description..
23:31:43fmlgevaerts: ah, that's a "server side" tool! So we can't be sure we have all the functions we have in the core rockbox, right? Is that the reasoning?
23:32:00archivatordomonoky: updateVoiceList/Description
23:32:19archivatorI understand *why* it's that way but I don't like it one bit
23:32:19gevaertsfml: yes, and strlcpy is a nonstandard function
23:32:38fmlgevaerts: OK, got it
23:33:03notlisteningdomonoky, busy night for you :D, I would guess that becuase the server as it stands takes a long time to startup it timesout of the first entry and it can also timeout for portions of text that take longer I think more testing needed
23:33:12notlisteningit gives: voicing: "Yes" to "/tmp/rbvoice//LANG_SET_BOOL_YES.wav"
23:33:13notlistening[Opensapi] cmd: "outFile /tmp/rbvoice//LANG_SET_BOOL_YES.wav"
23:33:13notlistening[OpenSapi] serverCommand timout: "202:Setting OK:outFile
23:33:13DBUGEnqueued KICK notlistening
23:33:13notlistening[Opensapi] cmd: "speakMe Yes"
23:33:13***Alert Mode level 1
23:33:13notlistening[OpenSapi] serverCommand timout: ""
23:33:26 Nick fxb__ is now known as fxb (
23:33:41 Quit perfectdrug (Quit: Leaving.)
23:34:08 Part literal
23:34:13domonokyarchivator: now i understand. the updateVoiceList gets used by the UI, because i connect it to my ui settings object. so it reacts to new paths in the ui.
23:34:55archivatordomonoky: I understand but it's not the queryServer or any other function beside startServer's business to know what the path to the server is
23:35:02archivatorit's ugly and hacky
23:35:19domonokynotlistening: yes, that looks like it just timesout.. so we need a bit better control of server startup.
23:35:46 Join CGL [0] (~CGL@
23:36:00 Quit Adubb (Read error: Connection reset by peer)
23:36:23 Join Adubb [0] (~aldubuc@
23:36:33domonokyarchivator: if you can do it better, feel free to improve it :-)
23:36:47archivatorwill do
23:37:59domonokyi agree that is a bit hacky.
23:38:12notlisteningdomonoky, I have implmented the new version to start the server up first them get the tts running under a thread which means we can startup and response within about 3 secs
23:41:09domonokysounds good :-)
23:41:29notlisteningincreases the timeout from 1/2 sec to 1 sec and it seems happier with the old version
23:42:47notlisteningI have now implemeted getting raw audio out from SAPI direct from the memory stream can you see any use ie for the .voice file as i know that is a continuous file
23:43:14***Alert Mode OFF
23:43:53domonokynot really, we always need a wav file to encode it with rbspeex or lame.
23:44:40notlisteningok thought it might have to be done that way but would i output the full .voice file in one go?
23:45:37domonokyno it processes all strings individually. they are also cut a bit before encoded and assembled to a voicefile.
23:46:23 Quit fml (Quit: CGI:IRC)
23:47:18archivatordomonoky: on a dual core machine, can't we voice the files and use the second core to encode them as we get them? Or is the encoder and external process as well?
23:47:42notlisteningdomonoky, okay thanks for you help again, it seems to be running smoothly at the moment. Will try to release a new opensapi server soon, still bug hunting though ;)
23:48:00domonokyarchivator: depends on target. encoder can be external lame or internal rbspeex.
23:48:05 Quit dfkt (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
23:48:46domonokyrbspeex could theoretically be made to work parallel. it shouldnt be to hard to change talkgenerator.cpp to encode in parallel for rbspeex.
23:48:59archivatorThat's what I meant.
23:49:12archivatorWe can't control the affinity of child processes, can we?
23:50:02amiconnlame is also open source, isn't it?
23:50:13domonokyprobably not easily in a crossplatform way. But Qt has special classes to make parallel codeing with as much threads as cores for example.
23:50:20CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:50:20*amiconn has no idea how difficult it woul dbe to build lame
23:50:58archivatoramiconn: probably not difficult but lame is more than just a library - why reinvent the entire frontend just to get a tiny speedup?
23:51:02domonokyso theoretically it should also be possible to call lame multiple times out of different processes.
23:51:08notlisteningif i wanted to manually set rockbox to voice menus on the clip+ as i can not save setting where do i need to edit?
23:51:36amiconnvoie menus should be enabled by default
23:51:43archivatordomonoky: I thought lame was an external executable?
23:52:18domonokyarchivator: yes, but wouldnt it be possible to just call the external exe multiple times in paralell ?
23:52:35archivatornot really
23:52:46domonokynotlistening: edit config.cfg
23:52:48archivatoreven if you call from different threads, you'll just be spawning new processes
23:53:07archivatorit's up to the underlying OS to distribute those processes among the cores
23:53:21archivatoryou can't guarantee where each one would end up
23:53:28domonokyyes, and hopefully the OS shedules those to different cores to distribute the load.
23:53:29 Quit Schmogel (Read error: Connection reset by peer)
23:53:33archivatorhence, you can't guarantee they would run in parallel
23:53:37notlisteningthanks domonoky & a think i would love to see if to add multiple dirs to add .talk clips to save me doing twice for all my Sansas
23:53:59archivatoreither that, or things would be twice as slow...
23:54:41domonokyno, it would just not be faster.. (minus losses for sheduling)
23:54:57amiconndomonoky, notlistening: You shouldn't need to edit anything for voice menus, as it is enabled by default
23:54:59 Quit evilnick_B (Quit: Page closed)
23:55:01archivatorthose losses can be significant :)
23:55:29amiconnAs soon as there is a suitable voice file, it should start talking (possibly after a reboot)
23:55:54notlisteningamiconn, .talk clips i meant if i said it wrongly
23:56:18amiconnAh, that's a different thing of course
23:57:08domonokyarchivator: but are general OS so dumb to not shedule to different cores if there is much load ?
23:57:35amiconntalk dir clip: on for directory clips, talk file clip: on for file clips
23:58:13archivatordomonoky: I hope not :)
23:58:38notlisteningnow being stupid where does config.cfg live?
23:58:42domonokyso for encoders would could make it parallel, if someone codes it. :-)
23:58:45notlisteningfind did not find it

Previous day | Next day