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 2009-06-08

00:00:18amiconnNo, but arm has, and coldfire also has little deltas
00:00:21 Quit Juice^ (Read error: 110 (Connection timed out))
00:00:46amiconnWhy on earth is a macro that evaluates to 1<<(n) different from writing 1<<(n) directly?
00:02:03Mikachuhow big was the size difference?
00:02:16*mcuelenaere recommends utils/analysis/
00:02:22Mikachumaybe it was just due builds jumping between different gcc versions on different build servers?
00:03:19*amiconn has a suspicion
00:03:23BagderMikachu: no, for targets we have fixed versions we use
00:03:58 Quit HellDragon (Read error: 104 (Connection reset by peer))
00:04:01amiconnThere's still some fluctuation in binsize depending on buildservers, but this one was too uniform
00:04:15Mikachucommit a comment change and see what happens? :)
00:05:01 Join Chesteta [0] (
00:07:52bertrikkugel, ok, let's ask some other devs about their opinion on this
00:08:07 Part ademille
00:08:25bertrika free yield could be nice, meaning it can do something useful other than just waiting for the FIFO
00:08:46Chestetahey I just saw this site, which deals with computer generated algorithms and speeding up execution of code... may be worth checking out
00:08:46bertrikalthough the FIFO takes only a rather short time to empty itself
00:09:08bertrikand I think we need to make sure we're not yielding to another task that also wants to access the display
00:09:10ChestetaI do not know if it can be applied to rockbox or not :/
00:09:35kugelbertrik: drawing is only allowed in the main thread, so that shouldn't be a problem
00:10:29 Quit bertrik ("Leaving")
00:10:53peturgevaerts: already had an answer about H10 usb charging?
00:11:00gevaertspetur: no
00:11:10peturok, I'll test then...
00:11:21gevaertsI know mine seems to charge, but I'd like to have a more general overview anyay
00:17:50 Quit robin0800 (Remote closed the connection)
00:19:53 Quit ender (" You don't have to think too hard when you talk to teachers. -- J. D. Salinger")
00:20:10peturgevaerts: is there a button to hold to connect usb and not have it mounted?
00:20:32gevaertspetur: yes. The one on the right just below the LCD I think
00:21:10peturah, just found out myself by trying all combinations
00:21:33 Join Thundercloud [0] (
00:22:12peturI see the charge animation, I'll let it charge a bit to see if it actually fills the battery
00:22:36peturthis is a H10 5GB pure (no radio)
00:26:58 Quit bluebrother ("b4n")
00:28:59peturgevaerts: seems to charge fine over USB with current SVN
00:29:08gevaertsgood :)
00:30:15peturseeing the slidepad control my laptop volume after I unmounted was new to me...
00:30:37gevaertsthere's always the hold switch :)
00:31:17 Quit Mikachu (Read error: 110 (Connection timed out))
00:32:49 Quit petur ("Zzzzz")
00:34:15 Quit Chesteta ()
00:35:01 Join barrywardell_ [0] (
00:37:42***Saving seen data "./dancer.seen"
00:38:06 Join funman [0] (n=fun@rockbox/developer/funman)
00:38:44funmansaratoga: did you look at skyeye?
00:41:00*amiconn wonders why changing two loop variables from signed to unsigned increase binsize on arm
00:45:42 Quit TheDJACR ("ZNC -")
00:47:52amiconnHmm. Gcc then puts the variables on stack, while they're in registers for signed. Weird...
00:50:16CIA-38New commit by amiconn (r21206): Undo the delta on ARM but making the loop variables signed as before r21205. GCC is weird...
00:53:50 Join onlysoaa [0] (i=422480c4@gateway/web/ajax/
00:54:34 Quit domonoky (Read error: 104 (Connection reset by peer))
00:55:21 Quit barrywardell_ ()
00:56:27 Quit n17ikh (Read error: 104 (Connection reset by peer))
00:56:55 Join n17ikh [0] (
00:59:16funmanobo: still progressing on view ?
01:01:02 Join robin0800 [0] (
01:01:46 Join kkurbjun [0] (
01:04:00 Join Mikachu [0] (
01:04:00 Quit robin0800 (Remote closed the connection)
01:09:22 Quit moos ("Rockbox rules the DAP world")
01:11:30 Join HellDragon [0] (
01:11:34 Quit HellDragon (Read error: 104 (Connection reset by peer))
01:11:39 Join HellDragon [0] (
01:11:48 Quit HellDragon (Read error: 54 (Connection reset by peer))
01:11:53 Join HellDragon [0] (
01:12:04 Join soycamo [0] (
01:13:23FrankTM_what button is BUTTON_SELECT on fuze?
01:14:21FrankTM_how do i read that?
01:14:45funmanpress center and up at the same time
01:14:48FrankTM_i'm in the blackjack.c source
01:14:57FrankTM_button_select starts a round
01:15:10FrankTM_so as i see it, there's no way you can press them both
01:15:44LloreanIf it's programmed properly, it should trigger the start of the round on the release of select.
01:16:10 Quit funman (Read error: 104 (Connection reset by peer))
01:16:17FrankTM_hmm. i think it's not programmed correctly then :P
01:16:21 Quit soycamo (Client Quit)
01:16:32LloreanHave you actually tried playing it on the player to see if it works?
01:17:06 Join funman [0] (n=fun@rockbox/developer/funman)
01:17:24FrankTM_i'm in the source to fix some other controls :p
01:17:41FrankTM_to be consistant with what the game says the buttons do
01:18:09funmanfuze keymaps need some work, especially in the radio screen
01:18:15LloreanAlso, couldn't you press UP and while holding UP press SELECT?
01:18:28FrankTM_Llorean: let me check
01:18:40 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
01:18:40 Quit pixelma (Nick collision from services.)
01:18:51FrankTM_ah. that seems to work
01:18:56FrankTM_solutions are so simple sometimes :P
01:18:59 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
01:19:20 Quit amiconn (Nick collision from services.)
01:19:22 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
01:19:44 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
01:20:57FrankTM_funman: i'll try to patch some of the keymaps
01:26:56 Quit mcuelenaere ("Zzzzzzzzzzzzzzzzz")
01:29:24 Quit HellDragon (Client Quit)
01:29:41 Quit Thundercloud (Remote closed the connection)
01:31:13funmanr21195/r21196 makes my Fuze lock on the logo screen with FS #10048 applied
01:31:35funmanthe only change I see is 1LU being changed to 1UL (and i don't know what 1LU means)
01:32:02FrankTM_i think i might have crashed the radio-bit
01:33:11*amiconn just found a weird settings effect
01:33:35 Join bubsy [0] (n=Bubsy@
01:33:48CIA-38New commit by funman (r21207): FS #10273 : Power Sansa Clip screen for some models needing as3514's DCDC15 ...
01:37:41 Join funman_ [0] (
01:37:55 Quit funman (Read error: 104 (Connection reset by peer))
01:38:03 Nick funman_ is now known as funman (
01:39:44amiconnFull rebuild helped. Weird.
01:39:51 Quit gevaerts (Nick collision from services.)
01:40:03 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
01:41:05 Quit funman (Read error: 54 (Connection reset by peer))
01:43:12 Join HellDragon [0] (
01:45:24 Join itcheg [0] (i=62db4c46@gateway/web/ajax/
01:46:10 Quit tessarakt ("Client exiting")
01:46:25FrankTM_what are FLA_ buttons
01:48:14pixelmaFLA - are you sure that's typed correctly?
01:49:26FrankTM_hacking on beer :+
01:51:51 Quit beta_ (Read error: 110 (Connection timed out))
01:52:12pixelmathat is a try to unify the (button) action handling in plugins, using the plugin lib. If you ask me it's not really handy in it's current state
01:53:04Unhelpfulit's just fine if you need exactly one context. otherwise you'd best create your own context, or use a core context and map the core actions to yours, or some combination of both strategies.
01:54:25pixelmayeah, unfortunately there are plugins that combine a few contexts, creating problems on some targets - and I don't like that things are "hidden" in a second file
01:55:27FrankTM_well.. it would be nice to have one file for all controls for example
01:55:30pixelmareally only useful for very basic controls where chances of "overlapping" definitions are very very low
01:55:36FrankTM_but it's not quite working
01:55:38Unhelpfulnot all plugins need the same controls.
01:56:31pixelmaand there can be differences between targets and what's good (and possible) with their button layout
01:57:10FrankTM_so.. (I'm guessing) it's alright to patch controls in a plugin, which is using FLA controls
01:57:27Unhelpfulamiconn: that loop is called once per *file*. we should probably just take the binsize gain and make both loops var < 64.
01:57:47 Join r0b- [0] (
01:58:10r0b-ok im using the latest current build of Rockbox on my sansa e250 and when playing songs the normal buttons dont work
01:58:12Unhelpfulerm, actually, that would be if we made it one loop. :)
01:58:21r0b-like Play/Pause Next Previous and the power button to stop the playback
01:58:27pixelmaFrankTM_: exceptions are a weird thing too...
01:59:21FrankTM_pixelma: that is not the short answer i was after :P
01:59:59 Nick fxb is now known as fxb__ (
01:59:59FrankTM_just a quick brainfart here:
02:00:27pixelmar0b-: maybe you have accidentally enabled "party mode"?
02:00:28FrankTM_would be nice to have a file with default buttons (for every plugin etc)
02:00:36FrankTM_and one for every device, with alterations to that
02:00:59r0b-that might be true :P
02:01:04*r0b- disregards
02:01:38 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
02:03:44 Quit itcheg (" ajax IRC Client")
02:04:12 Join notlistening [0] (
02:04:23 Join froggyman [0] (
02:06:00 Join itcheg [0] (i=62db4c46@gateway/web/ajax/
02:06:51Unhelpfulamiconn: i think we can do some optimization of that pre-multiply, too. the second member of struct idct_entry can be removed, which brings the shift down to PASS1_BITS, which means that we can guarantee it fits in a SH hardware multiply.
02:07:23pixelmarasher: did you have time to look at how to stop this "mipper" theme from showing up on the H300 theme page? (Probably related something to do with the RWPS, since it doesn't appear on the Ipod Photo page and is actually a 160x128 theme)
02:07:48LloreanFrankTM_: There's a file with default core buttons for each device.
02:08:40FrankTM_Llorean: yeah.. well.. i'm just thinking out loud right now
02:08:53LloreanFrankTM_: Maybe look at what's available first?
02:09:11r0b-the USB-HID stuff is kinda neat
02:09:18FrankTM_well.. as i see it, there's a lot of files that need adjustments when there's a new device
02:09:32FrankTM_correct me if im wrong, please :)
02:09:41LloreanWhich is always going to be true, because there's a ton of stuff that needs to come into existence.
02:09:58LloreanPlus devices have different numbers of physical buttons, some of which may or may not be absolute on/off (such as wheels, etc)
02:10:06LloreanSo the controls for each device will always need quite a bit of customization.
02:10:29FrankTM_yes. but i think that could be in a lot less files
02:10:36FrankTM_not quite sure about that yet, though
02:10:44 Join hd [0] (
02:10:51 Quit hd (Read error: 54 (Connection reset by peer))
02:11:08 Quit HellDragon (Read error: 104 (Connection reset by peer))
02:11:10LloreanWith the exception of plugins, all the controls are basically in a single file.
02:11:15LloreanOne file per target.
02:11:21pixelmaFrankTM_: I still prefer lots of files to adjust over having half-working controls which you only find out about later and often have to work around in a weird way
02:11:37 Quit advcomp2019 ("IRC is just multiplayer notepad")
02:12:35FrankTM_pixelma: i prefer usability too ;)
02:12:52FrankTM_i'm just trying to work out if there's a better way to reach that
02:13:35 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
02:13:46 Quit stripwax ("")
02:13:56LloreanHaving all the controls for every plugin in a single file just makes it more likely something will be overlooked.
02:14:21LloreanIf you see all the controls for several different targets in the bubbles plugin, you get an idea of how the controls work for several different key layouts, and it helps you map the controls to the new hardware.
02:14:41pixelmathe PLA thing was a try and there are still some plugins that are not completely usable on some targets (e.g.)
02:14:43LloreanOtherwise, you'd have to flip through pluginkeymaps-x5.h, pluginkemaps-h100.h, etc, etc, for comparing
02:14:50LloreanWhen you're just trying to make them work for a single plugin at a time
02:15:00pixelmae.g. metronome, I meant
02:15:06FrankTM_that might be true
02:15:20pixelmaLlorean: bubbles uses PLA (with some exceptions)
02:15:35Lloreanpixelma: Replace bubbles with "hypothetical plugin 01" :-P
02:17:34LloreanThe point is, it makes it easier (in my perspective) to be able to see all the target keymap variants for a single plugin while I'm working on it, rather than seeing all the keymaps for different plugins on the same target.
02:17:51*Unhelpful thought that what he did with PF worked out fine
02:19:15Unhelpfulespecially after kugel cleaned it up a bit ;)
02:19:20CIA-38New commit by kkurbjun (r21208): M:Robe 500: Mostly complete USB driver, supports BULK mode currently and gets about 2 MB/s writes vs 1.1 MB/s on the OF. Mostly tested against Linux, ...
02:20:00r0b-im using the 21200 Sansa e200 firmware
02:20:41kugelthe current PLA is useless imo
02:20:55FrankTM_Llorean: i think you're right
02:21:27kugelthere should be a single context with 6 or 7 buttons (directionals, select, quit, maybe context), +2 for scrollwheel
02:21:43kugelif a plugins need more than that, PLA isn't the right thing anyway
02:23:21*Unhelpful could swear he used that same argument when kugel advised *against* the PF input rewrite...
02:23:41FrankTM_seems like i brought up a discussion :p
02:23:44kugelyou must be wrong
02:24:16LloreanI think the PLA context should be bare minimum. Not even scrollwheel. Just four directions, select(yes) and cancel(no).
02:24:18 Quit Lss (Read error: 110 (Connection timed out))
02:24:23LloreanEverything else can be custom coded as necessary.
02:24:40LloreanAnd wherever possible we should use standard contexts (using our normal list for menus, etc)
02:25:25FrankTM_i keep fucking up mazezam.c
02:25:50kugelI think scrollwheel in that context makes sense
02:26:33kugelespecially if we treat up/down/left/right as directional on e.g. e200 instead of left/right/scroll_fwd/scrollback or some sort of that
02:27:21LloreanI think scrollwheels are generally too much of a special-case thing to really work in a custom context.
02:27:30kugelI don't think so
02:27:52Llorean*without a custom context
02:28:50kugelPLA will be useless again if you can't access the scrollwheel with it
02:29:03LloreanHah, that's a rather bold statement.
02:29:12kugelscrollwheel is a usefull adition which doesn't cost the non-scrollwheel targets anything
02:29:32LloreanThat doesn't make PLA useless without it, or PLA would also be useless in the first place for non-scrollwheel targets.
02:29:58LloreanI'm sorry, but what do you mean by "useless" other than "without a use"?
02:30:15kugelscroll* would be an extra action, the plugins #ifdef it with HAVE_SCROLLWHEEL when checking it
02:30:19LloreanIf PLA is useless without scroll wheel support, I don't see how it can be useful on targets without a scrollwheel, as they obviously don't have scrollwheel support.
02:30:56kugelI don't understand that
02:31:10LloreanYou said PLA would be useless without scrollwheel support.
02:31:38kugelyes, but how does that mean it will be *only* useful for scrollwheel targets?
02:31:53LloreanIf it can be useful without scrollwheel support, then it's obviously not useless, right?
02:33:16LloreanAnyway, my point is that the context made available should have the absolute bare minimum of universal keys that can have expected behaviour in a device.
02:33:41LloreanDepending on how a scrollwheel is physically mounted (think of the wheels on the sides of Palm devices and such) they may not make any sense for left/right action like front-face scrollwheels do
02:33:57LloreanScrollwheels shouldn't be treated as a universal thing like directional buttons, yes, and no can be.
02:34:10LloreanThe idea of a universal context should be that it can be used without having ever seen the hardware
02:34:15pixelmaPLA for scrollwheel has one problem - for some plugins scroll fwd/back are better mapped to up/down for some others it makes sense for left/right, so you would at least need two contexts (IMO)
02:35:24notlisteningcould it be an option in the menus? or did i miss the point of the argument?
02:35:32*Unhelpful knows at least one target where PF maps a vertically oriented scroller control to left/right in album view because it is a scroll control, and the left/right buttons look terribly awkward for that purpose.
02:35:35Lloreannotlistening: You missed the point. It's a code thing.
02:35:47kugelyou would have ACTION_PLA_SCROLLFWD, not connected to directions at all. The plugin would do whatever it wants withi that
02:36:17Lloreankugel: But does "FWD" mean rotating a wheel clockwise? Sliding up or down on a touchpad scroller? Or rolling a wheel on the side of the player up or down?
02:36:50 Join Lss [0] (
02:36:57FrankTM_plugins could expect something else than what it actually does practically
02:37:01pixelmakugel: I don't understand - where's the generic context in that?
02:37:04notlisteningright i am of to bed got a headache already and you guys are on the way to headaches by the sounds of it
02:37:39 Quit notlistening ("Leaving")
02:37:45***Saving seen data "./dancer.seen"
02:37:46LloreanFrankTM_: Basically my point - the provided context should be as simple as possible, and map to explicitly predictable actions. Give users as intuitive controls as possible.
02:38:02LloreanIf you can't accomplish it in those six buttons, create a custom control scheme and actually look at the hardware.
02:38:23Unhelpfulhuzzah, core scrolling in PF works (on color, only for album title, and with some minor issues) without crashing!
02:38:39Lloreancore scrolling?
02:39:06UnhelpfulLlorean: using the core text scrolling functions to scroll the album title when it's wider than the screen
02:39:07pixelmaso that pf dsoesn't do its weird own scrolling (of text)
02:39:22saratogaslasheri: did you work on the rockbox file tree code?
02:39:25LloreanMan, I hate that dual use of scrolling
02:40:26Unhelpfulthe hacks were thus: 1) styled scroll functions return a pointer to the scroll item 2) support for a STYLE_CUSTOM and a global custom styled draw hook on the main LCD 3) support for a STYLE_NODRAW that only updates the scrollinfo
02:42:44FrankTM_hmm. for something completely different
02:42:59FrankTM_anyone here who maintains rockbox flyspray?
02:43:24LloreanDepends on what you mean by "maintains"
02:43:27Unhelpfulcolor doesn't actually need STYLE_CUSTOM. greyscale will. on color, the album view saves a strip of the rendered covers when the albums stop scrolling, and sets this as the backdrop, adding an offset for y-coordinate of the album text, and then it starts the scrolling text. when it switches to the cover-in animation it sets STYLE_NODRAW, and uses the scrollinfo to update redraw the text correctly when it draws a new animation frame. the
02:43:27Unhelpfuln it changes the scrollinfo back to STYLE_DEFAULT when it reaches the tracklist.
02:44:07FrankTM_Llorean: well.. i'm submitting a patch
02:44:23FrankTM_but Sansa Fuze isn't in the "Player Type" list
02:44:30Unhelpfulthe nifty bit is that once you stop moving the album list, and the album title starts scrolling, it doesn't stop or reset while you go to track list etc
02:44:36FrankTM_which i think at this point, should be
02:44:52LloreanAh, this requires on of the flyspray admins, yes.
02:45:02LloreanWho are possibly asleep.
02:45:39FrankTM_i have trouble creating english sentences at this point of the day :p
02:45:52 Quit kugel (Remote closed the connection)
02:47:03FrankTM_anyhoo. i did a patch for fuze:
02:47:34FrankTM_my previous one was implemented by funman
02:48:33LloreanNo need to link to closed tasks.
02:48:52CIA-38New commit by kkurbjun (r21209): Fix Creative ZVM build
02:49:11FrankTM_Llorean: fair enough
02:50:34 Join advcomp2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
02:53:55 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
02:57:09CIA-38New commit by rmenes (r21210): New translation: Lithuanian by Vytenis Sabelka. From FS #10255.
02:59:33***Alert Mode level 1
02:59:33CIA-38New commit by rmenes (r21211): FS #10293: More Tagalog translation by Melba Sitjar.
03:01:49UnhelpfulSlasheri: if you get a chance, i had a tagcache issue. starting a search on tag_filename works, but if no filters are added and the RAM cache is enabled it returns no items. it works fine otherwise. i was able to trace this to a problem with opening the index file for the current search late (it ended up trying to read the header as an item), but after fixing that, the search still takes about 6x longer with ramcache enabled than with i
03:01:49Unhelpfult disabled - so i suspect i fixed it wrong. any ideas?
03:03:29FrankTM_< off to bed
03:04:17 Join HellDragon [0] (
03:05:27***Alert Mode level 2
03:05:27CIA-38New commit by kkurbjun (r21212): M:Robe 500: Add missing file.
03:07:45togeticodd, right speaker works now
03:08:08 Quit spyder ("Lost terminal")
03:08:12togeticlove how rockbox plays mp4 files :)
03:08:34 Join spyder [0] (
03:08:51togeticso many songs that have low quality sound and wouldn't want to transcode them any further
03:10:10Unhelpful(as to why, pictureflow needs a list of albums, but also needs to find a file from each album to locate album art, and searching on tag_filename and uniquifying the albums is faster than searching on tag_album and then searching files for each album, even with this problem)
03:12:36 Quit LambdaCalculus37 (Read error: 110 (Connection timed out))
03:15:28***Alert Mode OFF
03:15:33 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
03:31:02tmztUnhelpful: rockbox is dedicated to not changing the files on the filesystem, right? outside of .rockbox of course
03:31:46Lloreantmzt: Generally, yes.
03:32:14tmztso something like folder.jpg/png being created automatically is not desired?
03:32:28LloreanCreated from what?
03:32:34LloreanAnd Rockbox doesn't support .png
03:32:38tmztfrom embedded album art
03:32:48LloreanA plugin to do that would probably be accepted
03:32:49tmztoh, yeah I know that was discussed, .bmp then
03:33:20tmztso it would be something where the user chose what song to use as the album cover?
03:33:57LloreanI'd say something like have a few options, "create images for all files with embedded art" or "use first embedded image found as cover.jpg or bmp for each folder"
03:34:24n17ikhwill jpeg album art search for any old .jpg?
03:34:30n17ikhor does it have to have a specific name
03:35:20n17ikhnever mind, found it in the manual
03:36:07saratogatmzt: plugins that change the file system are already in rockbox (the disk tidy one) so doing something like that would be fine
03:36:10saratogabut also a lot of work i think
03:36:28 Quit froggyman (Read error: 113 (No route to host))
03:36:35tmztcould be useful with pf, wasn't really before
03:44:04 Join slyyx [0] (n=slyyx@
03:44:28slyyxonlysoaa: Hey, been busy today, will be busy tomorrow too, got a presentation I am doing Tuesday...Any progress?
04:00:54 Quit efyx_ (Remote closed the connection)
04:03:46onlysoaaslyyx: I've also been busy today, and I don't feel like anything right now. D:
04:03:58onlysoaaslyyx: I might work on it a bit tomorrow.
04:04:08onlysoaaStupid LCD... xP
04:05:37slyyxYeah, I agree
04:05:44slyyxI may do some reversing of the firmware
04:06:00slyyxFind out just what makes it tick
04:06:04onlysoaaCool! So far I've only had mild success with it. You might get farther.
04:06:28slyyxyeah, maybe
04:11:45n17ikhhow does Rockbox do SD card access? I was under the impression that the SD card specs required paying royalties to the SD card association
04:11:55n17ikhdoes it just do it in spi mode?
04:15:52 Quit avacore (Read error: 110 (Connection timed out))
04:19:40 Quit perrikwp (" ajax IRC Client")
04:37:48***Saving seen data "./dancer.seen"
04:48:33onlysoaaslyyx: If you can find how to read/set the LCLK it'd be really cool. I'm thinking that might be the problem.
04:48:45onlysoaaslyyx: I gotta go now. Please post in the forum if you find anything.
04:50:30onlysoaaslyyx: Oh, looks like I'm staying. o: I'll try looking, but... I don't know if I'll be able to find anything.
04:50:56slyyxIts in the datasheet :P
04:51:50onlysoaaIt sure is, but where? :P
05:02:25 Quit fdinel (Read error: 110 (Connection timed out))
05:17:52martian67saratoga, it has stopped
05:20:36 Quit itcheg (" ajax IRC Client")
05:22:09 Join ademille [0] (
05:25:13 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
05:35:12 Join itcheg [0] (i=62db4c46@gateway/web/ajax/
05:49:31 Quit Horscht ("Verlassend")
06:20:01togeticjust curious, how would i go about playing a few games that i have on my pc using rockbox, examples of the games i want to play:
06:20:26togeticalso want to play these two:
06:22:11cool_walking_Those are PC games... they would need to be individually ported. Lots of work.
06:24:00togeticPassage is linux
06:24:42togeticokay it doesn't recognize the exectuables
06:24:43cool_walking_Okay.. still needs to be ported. Rockbox isn't Linux.
06:24:53togeticit's based on ipodlinux, no?
06:26:52 Quit itcheg (" ajax IRC Client")
06:27:28cool_walking_No. I think Rockbox has a *tiny* bit of iPodLinux code.. probably for drivers.
06:28:14Lloreantogetic: Rockbox didn't even start on iPods.
06:28:35scorchetogetic: it is written from scratch essentially...even if it was linux, it still wouldnt be able to run those games without plenty of work
06:30:58Lloreantogetic: If I may ask, where'd you get the idea it was based on iPL. We run into that assumption occasionally and it'd help us if we could track down where the confusion comes from.
06:31:58togetici read it online today.. on the ubuntu forums.. one sec
06:32:29togetici googled 'ipodlinux versus rockbox', it's the first link
06:33:06togetic"IIRC, iPodLinux was developed as a 'hey, wouldn't it be cool if we could put linux on an ipod... lets try it!' Later, Rockbox came along and, in an effort to add functionality and flexibility to the ipod, used the iPod Linux platform to develop a functional user interface."
06:33:55LloreanIf you read the rest, someone just a few posts later points out that Rockbox uses its on kernel.
06:33:57scorcheand if you read down 4 posts, you would see that someone else corrected him and said we use our own kernel
06:34:04*Llorean is amazed at the person who says iPL is *easier* to install than Rockbox
06:34:41scorcheLlorean: and if you read the second page, you will see that soemone mentioned something to that effect too =P
06:34:57LloreanYeah, I noticed there was a second page after hitting enter.
06:35:26togetici should give ipodlinux 'a spin' sometime soon when there's free time
06:35:41togeticmaybe i could play the games using it
06:35:53togeticthe gba games are practically useless
06:36:03togeticno sound, it's highly glitchy
06:36:26cool_walking_You're not going to be able to play PC games on iPodLinux without porting them, either.
06:36:35scorchetogetic: as i said, even on ipl or if rockbox was based on linux, you would have to do a lot of work...
06:36:46togetic..but it's linux?
06:36:51togetici don't understand :P
06:36:54scorcheit an ipod It's different hardware.
06:37:32scorchetogetic: there are also architechure differences, etc...lots of complex stuff beyond the kernel that comes into this ;)
06:37:50***Saving seen data "./dancer.seen"
06:38:08togeticanyone have a 1g nano ipod and have successfully played a game without glitches?
06:38:14togetic..on rockbox
06:38:17scorche"a game"?
06:38:19togetici realized it's being emulated, but..
06:38:24 Quit AndyI (Read error: 60 (Operation timed out))
06:38:25togeticyeah, any gba game
06:38:33LloreanRockbox doesn't have a GBA emulator
06:38:36LloreanSo, no, nobody has.
06:38:36 Join AndyI [0] (i=AndyI@
06:38:42togetici meant gbc
06:38:46togeticor even gb
06:38:47LloreanThen yes, I have.
06:38:57togeticwhat game Llorean?
06:39:09LloreanFinal Fantasy Legend.
06:39:14togeticah, i have that
06:39:22togeticwell, final fantasy adventure..
06:39:29LloreanDifferent games entirely.
06:39:45LloreanBut the Nano's pretty slow, so don't expect perfect performance from most games.
06:39:53LloreanAnd the emulator could always use more polish
06:40:11togeticthe mpeg movie player works flawlessly
06:40:37togeticno sound in doom
06:40:48LloreanNo music in doom. It should have sound effects
06:42:01togetichad the sound turned off
06:43:40togeticthe game is difficult to play like this, but it's so cool
06:45:48Mikachuit's almost possible with god mode on :)
06:47:03togeticany good puzzle games?
06:47:33togeticsudoku is alright i guess
06:47:42Mikachusokoban if you have a looking glass :)
06:48:05Mikachuwrong word, magnifying glass
06:49:03togetici forget what game it is w/ the helicopter, but it just crashes
06:49:18togeticthat's all, no flying, it just does a nose dive straight to the ground each time it starts
06:49:27togeticah, 'chopper'
06:49:30Mikachupress the button ;)
06:49:41togetici see :)
06:50:04Lloreanis chopper in the manual?
06:50:07togeticnow it's fun :)
06:50:21DhraakellianMikachu: but how do you activate god mode?
06:50:40MikachuDhraakellian: your favorite text editor :)
06:51:49Dhraakellianhmm... <−−could that be used to unbrick a Fuze?
06:52:59 Join Xerion_ [0] (
06:54:13togetichmm, for some reason, the podcasts from here won't play on me rockbox:
06:54:49saratogayou're probably doing something wrong
06:55:11togetici can play everything else
06:55:33LloreanWhat format are they, and how long are they?
06:55:37togeticwhat's also odd is that it won't play mp3s after attempting to play one of these
06:55:49LloreanPerhaps they have odd metadata.
06:56:28togetici listened to radio podcasts that were both ~1hr long, i listened to them... using rbox
06:56:50togeticooh, ooh
06:56:59togeticit plays
06:57:19togetici had to press play...
06:57:38togeticodd, they are all playing..
06:58:05togeticmaybe i was pressing pause/select at the same time when selecting the podcast
06:58:10togetici do have big thumbs
06:58:18togeticshush, don't say it scorche :P
06:59:02togetici suppose it's time for bed
06:59:23togeticthanks for entertaining my random thoughts tonight
06:59:57 Part ademille
07:09:03 Quit Xerion (Read error: 110 (Connection timed out))
07:09:03 Nick Xerion_ is now known as Xerion (
07:13:05 Quit jmillikin (Read error: 60 (Operation timed out))
07:32:22 Join dmb [0] (n=dmb@unaffiliated/dmb)
07:41:53 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:59:17CIA-38New commit by amiconn (r21213): Add language identifier for Lithuanian.
08:02:51 Join bertrik [0] (
08:08:02 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:13:55 Join ender` [0] (
08:21:38 Join Rob2222 [0] (
08:26:44 Join FlynDice [0] (
08:37:12 Join petur [50] (n=petur@rockbox/developer/petur)
08:37:54***Saving seen data "./dancer.seen"
08:40:08 Quit Rob2223 (Read error: 110 (Connection timed out))
08:54:44 Join avacore [0] (
08:56:56 Nick J-23 is now known as Quintasan___ (
09:00:01 Nick Quintasan___ is now known as J-23 (
09:04:17 Join DarkAnthey [0] (n=darkanth@
09:04:49 Part DarkAnthey
09:06:45 Join flydutch [0] (
09:10:43 Quit bertrik ("Leaving")
09:14:51 Join n1s [0] (n=n1s@rockbox/developer/n1s)
09:28:00 Join matsl [0] (
09:28:43 Join Thundercloud [0] (
09:28:59 Quit Bagder (Read error: 110 (Connection timed out))
09:30:17pixelmaZagor: do you remember where you got the USB VID/PIDs of the c200s from that you put on the DeviceChart wiki page? I was wondering because I got a different result when compared to the description - after asking here and Llorean confirming my results, I changed them (see the diff between the revs of the page)
09:30:42pixelmabut I still wanted to ask, in case there are differences between the actual devices...
09:31:41ZagorI think I got them from my own c200.
09:31:48pixelmait looked like only the description was wrong though
09:33:15Zagorum, which page? DeviceChart is from dec 08
09:33:16pixelmawell it had the VID/PID of my (and Llorean's) c200 in UMS mode listed as "c240 in UMS mode" and mine in MTP mode as "c250 in UMS mode"
09:33:37pixelmaerr... DeviceDetection ...something with Device ;)
09:35:23pixelmaLlorean had exactly the same results with his c240 as me with a c250
09:36:07Zagorok. I can't disagree since I don't remember the details of my check.
09:37:18 Quit n1s (Remote closed the connection)
09:37:56pixelmawould be nice if you find time to test again, maybe there are really differences...
09:38:22ZagorI'll try to remember tonight.
09:38:51pixelmathanks :)
09:39:26 Join barrywardell_ [0] (
09:39:30 Quit barrywardell_ (Remote closed the connection)
09:47:26 Join n1s [0] (n=n1s@rockbox/developer/n1s)
09:52:53 Quit Thundercloud (Remote closed the connection)
09:55:24 Quit _lifeless (Read error: 110 (Connection timed out))
10:04:06 Join einhirn [0] (
10:07:59 Quit BHSPitMonkey ("Ex-Chat")
10:11:12 Quit kachna (Read error: 113 (No route to host))
10:13:08 Join AndyIL [0] (i=AndyI@
10:24:29 Quit AndyI (Read error: 110 (Connection timed out))
10:32:22 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
10:36:24 Join _lifeless [0] (n=lifeless@
10:37:57***Saving seen data "./dancer.seen"
10:40:06 Join efyx_ [0] (n=efyx@
10:40:40 Quit fyrestorm (Read error: 54 (Connection reset by peer))
10:42:23 Join FrankTM [0] (
10:43:44 Quit FrankTM_ (Read error: 60 (Operation timed out))
10:49:05 Join mib_dq3or4tm [0] (i=792c1403@gateway/web/ajax/
10:50:48mib_dq3or4tmHey can anyone help me? I accidentally deleted some files from my gbsystem, then reinstalled the bootloader and rockbox, and it still won't boot
10:51:22mib_dq3or4tmI've formatted it multiple times to exfat and fat32, and tried to restore the original firmware and rockbox
10:51:53mib_dq3or4tmand all I get is no system found on hdd (error code 0010070)
10:54:45 Join fyrestorm [0] (n=nnscript@
10:55:23mib_dq3or4tmpls help?
10:55:27 Quit cool_walking_ (" ajax IRC Client")
10:55:33linuxstbmib_dq3or4tm: So neither Rockbox or the Rockbox bootloader is starting
10:55:55mib_dq3or4tmIt starts to load the hd
10:56:14mib_dq3or4tmthen it just goes to the error screen
10:56:16linuxstbHave you seen this?
10:56:25mib_dq3or4tmI've looked all over the internet but not there :D
10:56:42 Join mcuelenaere [0] (
10:56:45mib_dq3or4tmah wait
10:56:50mib_dq3or4tmI looked at that last night
10:56:53linuxstbThere's also a dummy gbsystem folder on that page, which you can download and copy to your gigabeat.
10:56:55mib_dq3or4tmdidn't try though
10:59:41CIA-38New commit by mcuelenaere (r21214): VX767: cleanup LCD driver a bit + add forgotten backlight-target.h
11:03:55 Join kachna [0] (
11:04:42 Join PaulJam [0] (
11:09:10 Quit ender` (Read error: 110 (Connection timed out))
11:23:40GodEateris there anything still seriously blocking FS #9708 ?
11:25:18 Join __lifeless [0] (n=lifeless@
11:25:35 Quit _lifeless (Read error: 60 (Operation timed out))
11:33:41gevaertsGodEater: right now, the freeze :) Apart from that I'd like to know as well...
11:38:52n1si always liekd the commit-and-see-if-anyone-complains strategy :)
11:51:28 Nick fxb__ is now known as fxb (
11:52:41 Quit slyyx (Read error: 113 (No route to host))
12:11:52 Quit Kupop (Read error: 110 (Connection timed out))
12:20:49 Join Kupop [0] (
12:20:52 Quit perrikwp (" ajax IRC Client")
12:21:20 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
12:25:59mib_dq3or4tmOk back again
12:26:09mib_dq3or4tmi've been trying to remove the stupid backplate
12:26:15mib_dq3or4tmtook all 4screws out
12:26:18mib_dq3or4tmcan't get it off
12:26:26mib_dq3or4tmon gigabeat x30
12:26:29mib_dq3or4tmany help here?
12:27:41mib_dq3or4tmlol nvm theres 2 more hidden
12:28:18 Join petur2 [50] (n=petur@rockbox/developer/petur)
12:28:35 Quit petur (Nick collision from services.)
12:28:38 Nick petur2 is now known as petur (n=petur@rockbox/developer/petur)
12:38:02***Saving seen data "./dancer.seen"
12:38:06CIA-38New commit by bluebrother (r21215): Add Lithuanian to rbutil voice language options.
12:38:24 Join gregzx [0] (
12:48:05 Join AndyI [0] (i=AndyI@
12:51:29mib_dq3or4tmhey, how do I remove the hard drive, I don't get it
12:57:56mib_dq3or4tmpls someone help me :(
12:58:03mib_dq3or4tmmy gigabeat is in peices
12:58:13mib_dq3or4tmand i dunno how to get the hard drive out :d
13:01:16 Quit AndyIL (Read error: 110 (Connection timed out))
13:01:33linuxstbmib_dq3or4tm: Be patient - not many people own an X, and not everyone is online every minute of the day...
13:01:57mib_dq3or4tmi got it off
13:02:06mib_dq3or4tmhow do i disconnect the cable?
13:02:27mib_dq3or4tmsorry if I seem impatient, It's late here, I have to go soon
13:02:56mib_dq3or4tmhis is a nooby thing, but i kinda don't see why you would be here for hours connected, without sayin anything?
13:03:17mib_dq3or4tmif you know how to remove tha cable, pls says so, thx in advance
13:05:25 Join wark [0] (
13:06:07FrankTMwhat is the difference between Initialize now & Update now in the Database settings menu
13:07:01mib_dq3or4tmupdate updates the database
13:07:13mib_dq3or4tminitialize means just use it or something
13:07:22mib_dq3or4tmmebe temporary updating?
13:07:27pixelmaFrankTM: is the manual not clear enough?
13:07:30mib_dq3or4tmtry it, see what happens
13:09:01FrankTMthere is (afaik) no manual for my device.
13:09:08FrankTMhmm. might aswell check some others
13:09:25SlasheriUnhelpful: hmm, interesting. Do you have a patch to show how you tried to fix that?
13:10:00SlasheriFrankTM: initialize removes and re-creates the database from scratch (i.e., removing all statistics, ratings and so on)
13:10:28SlasheriFrankTM: update does just insert or remove items from the db
13:11:02FrankTMratings etc doesn't work anyway :P
13:11:05 Join petur2 [50] (n=petur@rockbox/developer/petur)
13:11:11FrankTMbut i understand now
13:11:59SlasheriFrankTM: doesn't work?
13:12:04Slasherithose should work
13:12:44FrankTMlast played and what else more does not work for me
13:12:49pixelmaSlasheri: I don't know which device FrankTM has, but some newer ports aren't able to write to disk yet (e.g. D2)
13:12:54 Quit fyrestorm ("lamers envy me like they envy bill g -- main boot xp, just the way it should be!")
13:13:01FrankTMit's the Sansa Fuze
13:13:17FrankTMplugins do write to disk
13:13:27 Join fyrestorm [0] (
13:13:31FrankTMdatabase get's filled aswell
13:13:37pixelmaah hmm, did you enable "gather runtime data"?
13:13:58FrankTMi think "should I?" explains that
13:14:26Slasheriif you want ratings to work, that has to be enabled :)
13:15:08*FrankTM goes sit in the corner
13:15:14FrankTMhow did i miss that
13:19:40scorcheby not reading the manual? ;)
13:20:17mib_dq3or4tmdoesn anyone here know how
13:20:32mib_dq3or4tmto remove the cable from a 1.8 inch cable on a gigabeat x?
13:20:41FrankTMscorche: well.. i haven't read it all
13:20:59FrankTMbut i went through all those menu's and didn't even see it :P
13:21:38 Quit linuxstb (Read error: 113 (No route to host))
13:22:12scorchemib_dq3or4tm: please read the guidelines for this channel as linked in the topic...there are a number of them that you have to your inquiry, it is a common ZIF socket...there are likely many results in a google search for more information...
13:25:49 Quit petur (Read error: 110 (Connection timed out))
13:26:13mib_dq3or4tmsorry :D
13:26:50mib_dq3or4tmits just the hard drive is stuck in a protective cradle, so access to the connector is limited
13:29:48 Join petur [50] (n=petur@rockbox/developer/petur)
13:34:14mib_dq3or4tmah ok thats off
13:34:54mib_dq3or4tmnow, i accidentally knocked off the little plastic switch to turn on and off the battery
13:35:03mib_dq3or4tmany ideas how to fix?
13:35:17mib_dq3or4tmits just 1 tiny piece of plastic
13:35:33mib_dq3or4tmnvm answered that too :D
13:36:59FrankTMwe're glad to have helped with that
13:38:34 Quit petur2 (Nick collision from services.)
13:44:29anigavis there a way to have rockbox switch back to the now playing screen after a certain period of inactivity, like the ipod software does?
13:46:22 Join arohtar [0] (
13:47:06Torneanigav: no
13:47:22anigavshitburgers. why?
13:47:35Tornebecause software doesn't do anything unless someone programs it to?
13:47:54gevaertsIs pressing the go-to-WPS button too hard?
13:48:15anigavwhat's WPS? and which button is that?
13:48:52gevaertsWhile Playing Screen, and see the manual.
13:49:24anigavI don't think the ipod video does have a button for that ...
13:49:28anigavunless you mean the quick menu
13:49:54pixelmathe Ipod has, please read the manual
13:50:42 Join stoffel [0] (
13:54:53 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
13:55:03mib_dq3or4tmugh, does anyone else find zif connectors fidgety?
13:55:18mib_dq3or4tmI don't think its quite plugged in all the way (hard drive)
13:55:24anigavpixelma: well I can't find it, if you know which button it is please tell me
13:55:55 Quit arohtar (Read error: 104 (Connection reset by peer))
13:56:14mib_dq3or4tmtell anigav try pressing buttons till it happens :D
13:56:15FrankTManigav: probabbly arrow up
13:56:31Torneit's "play"
13:56:32FrankTMhmm.. without the arrow part
13:56:52Tornewhich shouldn't be that difficult to discover given that there are only five buttons
13:57:10anigavnope, when I press play, playback stops, only if I then press play again I get to the WPS
13:57:29Tornehow long are you pressing play for?
13:57:34Tornelong press of play is stop
13:57:43Torneshort press is play/pause on WPS, or "go to wps" on everything else
13:57:58mib_dq3or4tmhuh indeed!
13:58:00anigavwell now it seems to work
13:58:02mib_dq3or4tmall that wasted time
13:58:09anigavI tried it before and it didn't
13:58:22Torneit doesn't work from in plugins (which map buttons however they want)
13:58:25mib_dq3or4tmto find out that opening the stupid gigabeat didn't fix my problem!!
13:58:27Torneor from a few screens like the recording screen
13:58:30FrankTMhmm.. on fuze: up = play & wps; home hold = stop playpack
13:58:38Tornebut from almost all the menus, it should work
13:58:58mib_dq3or4tmgoodnight ladies, ,may you have fun solvin ipod problems
13:59:02 Part mib_dq3or4tm
13:59:06Torneand yes it's in the manual :)
13:59:43anigavnevermind, I've found it :)
14:01:06anigavit's under "file browser", but now under "main menu"
14:01:27Tornemost of the file browser controls apply to all the menus
14:01:49anigavthat's good to know
14:02:57anigavanother thing: is there a way to edit tags with rockbox?
14:03:07anigavthat would be cool as hell
14:03:20Tornedon't think so.
14:03:35Torneit's possibly the least convenient/comfortable possible interface to be doing that :)
14:03:53anigavwell, often you only see that a tag is wrong while you're playing a song, and by the time you get home, you forget about it
14:04:08anigavso it would be useful to edit tags right there
14:04:19Torneif you aren't using ratings for anything else, you could mark stuff with those
14:04:25Tornerate it as 1 star if it's taged wrong
14:04:28Tornethen you can pick the list out later
14:04:39anigavok that makes sense
14:04:52*Torne shrugs, never needed to do that, all the music on my player has been through musicbrainz already
14:05:00anigavwhat's that?
14:05:00FrankTMbe sure to remove all ratings before you start tagging :P
14:05:38anigavanyway, a tag editor would be more useful than the paint app :P
14:05:48Tornefeel free to implement a tag editor plugin, anigav :)
14:05:54 Quit stoffel (Read error: 60 (Operation timed out))
14:07:13anigavgive me half an hour
14:07:31FrankTManigav: you could draw the filenames that need work :+
14:07:40anigavyou write plugins in c right?
14:07:56anigavcause I know c
14:08:31Tornewell, feel free to learn how to write plugins, and write one for whatever you like ;)
14:09:11anigavI've been thinking of attaching a NES controller to the ipod and using that to play the games
14:09:45anigavis there a NES emu for rockbox?
14:09:55Torneit's not done, but i believe it works
14:10:01Tornelook in the bugtracker, there's a patch
14:10:12anigavfor the emu or for the controller?
14:10:58TorneFS #2911
14:10:59Tornethe emulator
14:11:19anigavok, but playing with the ipod wheel is really shit
14:11:21Tornethe controller would be possible, technically, but you'd need some intervening hardware to convert the NES pad interface into serial commands
14:11:49anigavyeah, a small attiny could do that easily
14:12:05anigavthe nes controller is just buttons attached to a shift register
14:12:25anigavyou just need to read that out and send it over the rs232 port
14:12:26Torneor just break open an ipod dock that has play/skip/etc buttons and use that :)
14:12:40anigavhow about no?
14:12:59anigavI didn't buy any accessories for my ipod whatsoever
14:13:06anigavit's fucking robbery
14:13:28 Join froggyman [0] (n=187b533e@gateway/web/cgi-irc/
14:13:41anigavanyway, gotta go to the dentist, see ya
14:15:59 Join gregzx_ [0] (
14:18:50 Join gregzx__ [0] (
14:18:54 Join mt_ [0] (n=MTee@
14:19:48 Quit gregzx (Nick collision from services.)
14:19:53 Nick gregzx__ is now known as gregzx (
14:20:12 Join barrywardell_ [0] (
14:20:19 Quit matsl (Read error: 110 (Connection timed out))
14:20:52 Quit gregzx_ (
14:20:52 Quit Zagor (
14:20:52 Quit lyngaas (
14:20:52 Quit barrywardell (
14:20:52 Quit mt (
14:20:52 Quit FOAD (
14:20:52 Quit GodEater (
14:20:52 Quit maraz (
14:20:58NJoingregzx_ [0] (
14:20:58NJoinZagor [242] (n=bjorn@rockbox/developer/Zagor)
14:20:58NJoinlyngaas [0] (
14:20:58NJoinmt [0] (n=MTee@rockbox/developer/mt)
14:20:58NJoinFOAD [0] (
14:20:58NJoinGodEater [0] (n=nochance@rockbox/staff/GodEater)
14:20:58NJoinmaraz [0] (
14:23:36togeticabout to leave for awhile, but i was wondering, how do you save readable to txt to rockbox?
14:24:10togetici put simple .txt files in the default Notes folder, but they aren't visible when I browse there with the ipod
14:24:17 Quit __lifeless (Remote closed the connection)
14:24:33 Join __lifeless [0] (n=lifeless@
14:25:05 Quit froggyman ("CGI:IRC")
14:26:24GodEatertogetic: have you set "View files" to "supported", or is it just set to "Music" at the moment ?
14:29:31togeticGodEater: i don't even know how to do that
14:30:10togeticnever heard of it
14:31:55 Join stoffel [0] (
14:32:14togeticit was on 'supported'
14:33:20 Quit FOAD (
14:33:20 Quit lyngaas (
14:33:20 Quit maraz (
14:33:20 Quit gregzx_ (
14:33:20 Quit Zagor (
14:33:20 Quit GodEater (
14:33:20 Quit mt (
14:33:47NJoingregzx_ [0] (
14:33:47NJoinZagor [242] (n=bjorn@rockbox/developer/Zagor)
14:33:47NJoinlyngaas [0] (
14:33:47NJoinmt [0] (n=MTee@rockbox/developer/mt)
14:33:47NJoinFOAD [0] (
14:33:47NJoinGodEater [0] (n=nochance@rockbox/staff/GodEater)
14:33:47NJoinmaraz [0] (
14:33:59 Quit gregzx_ (SendQ exceeded)
14:34:39togeticfound the problem
14:34:54togeticthe extensions .txt weren't on the name of the file
14:35:05GodEaterthat would present an issue then yes
14:35:49togeticheh, i just figured dolphin would do that automatically when i created the .txt file using the right click menu
14:37:45 Quit mt (Connection timed out)
14:38:06***Saving seen data "./dancer.seen"
14:39:56 Join mrgreen [0] (
14:40:47mrgreenAnyone point me too a list of working cf cards please?
14:44:59 Quit mrgreen ("Lost terminal")
14:46:52 Quit FOAD (
14:46:52 Quit lyngaas (
14:46:52 Quit maraz (
14:46:52 Quit Zagor (
14:46:52 Quit GodEater (
14:47:06NJoinZagor [242] (n=bjorn@rockbox/developer/Zagor)
14:47:06NJoinlyngaas [0] (
14:47:06NJoinFOAD [0] (
14:47:06NJoinGodEater [0] (n=nochance@rockbox/staff/GodEater)
14:47:06NJoinmaraz [0] (
14:47:29 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
14:49:47 Join daurn [0] (n=daurnima@unaffiliated/daurnimator)
14:52:11UnhelpfulSlasheri: it includes some new tagcache methods, but here's one of the fixes i tried, with the relevant hunk at line 89 in the patch:
14:55:00 Join Lynx_ [0] (
15:05:09 Quit daurn| (Read error: 110 (Connection timed out))
15:08:01 Join evilnick [0] (i=0c140464@gateway/web/ajax/
15:08:38 Quit FOAD (
15:08:38 Quit lyngaas (
15:08:38 Quit maraz (
15:08:38 Quit Zagor (
15:08:38 Quit GodEater (
15:08:50NJoinZagor [242] (n=bjorn@rockbox/developer/Zagor)
15:08:50NJoinlyngaas [0] (
15:08:50NJoinFOAD [0] (
15:08:50NJoinGodEater [0] (n=nochance@rockbox/staff/GodEater)
15:08:50NJoinmaraz [0] (
15:08:52 Join robin0800 [0] (
15:09:56GodEaterthis is getting tedious
15:10:06 Quit FOAD (
15:10:06 Quit lyngaas (
15:10:06 Quit maraz (
15:10:06 Quit Zagor (
15:10:06 Quit GodEater (
15:14:36 Join GodEater [0] (n=nnochanc@rockbox/staff/GodEater)
15:19:16 Quit stoffel (Read error: 113 (No route to host))
15:23:22NJoinZagor [242] (n=bjorn@rockbox/developer/Zagor)
15:29:45 Join daurn| [0] (
15:39:37 Join stoffel [0] (
15:45:33 Quit daurn (Read error: 110 (Connection timed out))
15:45:45 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
15:47:07 Quit Kupop (Read error: 110 (Connection timed out))
15:51:39 Join kugel_ [0] (
15:53:02 Join Kupop [0] (
15:53:09 Quit stoffel (Remote closed the connection)
15:53:39 Join stoffel [0] (
15:58:41 Quit robin0800 (Remote closed the connection)
15:59:09 Join froggyman [0] (n=187b533e@gateway/web/cgi-irc/
15:59:53 Join kugel__ [0] (
16:00:11 Quit kugel__ (Remote closed the connection)
16:00:25 Join kugel [0] (n=kugel@rockbox/developer/kugel)
16:01:44 Nick mt_ is now known as mt (n=MTee@
16:06:27 Quit froggyman ("CGI:IRC")
16:10:16 Join ED-209_ [0] (n=chatzill@
16:11:02ED-209_hello, where can I find utilities or games/emulators/etc for an iriver h300?
16:14:57 Quit linuxstb (Read error: 113 (No route to host))
16:15:01 Quit kugel_ (Read error: 110 (Connection timed out))
16:15:39evilnickED-209_: What sort of utilities/games?
16:16:47ED-209_basically anything I can try out, I don't have anything specific in mind
16:17:18LambdaCalculus37ED-209_: There are already some games in Rockbox.
16:17:21evilnickHave you read the manual? All the emulators that work are available in each build -
16:17:23LambdaCalculus37And the manual explains how to use them.
16:17:48ED-209_except for one thing, I was wondering if it's possible for a plugin to be made that will pitch shift a signal from the line in, and output it in realtime
16:18:26evilnickTo use the h300 as a sort of effects box?
16:18:33LambdaCalculus37ED-209_: Any kind of plugin can be written, provided someone is willing to work on it.
16:18:49ED-209_in a sense... I want to use it as a kind of voice modulator
16:19:24LambdaCalculus37This is an entirely volunteer-based project, so if someone comes in and says, "Gee, I wish we had such and such...", then we tell them, "Well, feel free to contribute such and such".
16:19:54LambdaCalculus37But there are also patches on Flyspray that you can look at to see if someone has maybe covered what you want.
16:20:59ED-209_yeah I've checked out every feature of the stable release and the current/daily releases for he H300, but I was vaguely hoping there might be some h300 fans in here who knew about some other things that aren't included in rockbox
16:21:09 Join funman [0] (n=fun@rockbox/developer/funman)
16:22:17evilnickED-209_: This thread might be interesting for you: (ignore the title of it!)
16:22:41 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
16:23:49 Join pyro_maniac [0] (
16:26:05 Quit HellDragon (Read error: 104 (Connection reset by peer))
16:26:17 Join HellDragon [0] (
16:28:04ED-209_interesting thread
16:28:45evilnickAnd the other one that is linked there may be useful too
16:28:56ED-209_it was just about 20 bit recording
16:29:08 Join jgarvey [0] (
16:29:08evilnickThe other one:
16:29:11ED-209_if the recording could have effects put on it in realtime
16:29:18funmanis there any ARM targets without the cp15 coprocessor ?
16:29:50Torneipod, no?
16:30:05Tornecp15 is usually missing on arm7tdmi
16:30:18 Join cmwslw_ [0] (
16:30:18Tornesince it would serve no purpose other than chip id
16:30:28 Part cmwslw_ ("Ex-Chat")
16:31:22 Nick barrywardell_ is now known as barrywardell (n=barry@rockbox/developer/barrywardell)
16:31:44 Join toffe82 [0] (n=chatzill@
16:32:00funmanI need this info for fs#10296 : ARM: Display Fault Status and Address for data aborts
16:32:20Tornearm7 doesn't have those registers, no
16:32:30 Quit at0m (Read error: 110 (Connection timed out))
16:32:33Tornethe only way to find the faulting address is to look in r14 when you hit the abort vector
16:32:54Torneand there are no abort statuses on arm7tdmi, all aborts are external (or possibly alignment?)
16:33:02 Quit onlysoaa (" ajax IRC Client")
16:33:16funmanis it the fault address, or the address where the fault occured?
16:33:29Torneoh, sorry, yes. the pc address where the fault occurred.
16:33:33Tornethe faulting address is not saved.
16:33:42Tornethe value of r14_abt is all you get
16:33:57Tornethis will be the case for any ARM that doesn't have an MMU or MPU
16:34:03funmando you have some doc, or will to explain what an "external" abort is ? (this what i'm seeing)
16:34:06Tornenot sure which chips we support
16:34:21Tornean external abort is just somethnig outside the processor pulling the abort pin
16:34:36Torneusually, it's either the bus multiplexor saying "there is no hardware at this address"
16:34:38funmanok, like the memory controller in this case?
16:34:45Torneor a peripheral saying "there is no register at this address"
16:35:01r0b-well i see USB in 21200 is working to an extent
16:35:02Torne(not all peripherals/memory controllers bother to signal aborts though, they're free to just mirror or return garbage or whatever)
16:36:08funmanthere is a lot of arm7 targets in rockbox (all PP)
16:36:18TornePP is arm7tdmi
16:36:25 Part LinusN
16:36:33funmanthen if this patch is useful we should create a new feature HAVE_CP15
16:36:35Tornethere are ARM7s with MMUs as well, but i don't know if there are any players based on that or not
16:36:44Torneit's not the presence of cp15 that makes it work
16:36:54Torneit's having an mmu/mpu. it's technically independant
16:37:09kugelisn't cpu already on board when there are caches?
16:37:46Torneiirc the portalplayer's caches are weird and controlled via proprietary methods. i'm not an expert on PP though so i could be wrong
16:38:10***Saving seen data "./dancer.seen"
16:38:27funmanyes there is a specific cache controller
16:38:45Torneso, yah, as far as I know there's no reason why teh PP chips would have cp15
16:38:50 Join shodanX_ [0] (
16:38:59Tornethat's only to control standard ARM-architecture spec MMUs/MPUs/caches
16:39:13Tornei don't know for certain but i'm reasonably sure that ARM7TDMI never has cp15
16:39:35kugelbut at least all arm9*?
16:39:52Tornecp15 is required to be present on arm9 iirc but that doesn't mean the fault status register will be
16:40:03funmanTorne: i think it's only required on armv6
16:40:05Tornefault status is only available if you have MMU/MPU, whatever ARM core it is
16:40:16Tornefunman: ..yes, that's it actually ;)
16:40:47funmanwe would need HAVE_ARM_MMU then
16:40:56Tornesurely there is already such a thing :)
16:41:10Tornesome of the ports already use the MMU
16:41:29funmani don't think so, specific ports handle themselves their memory management
16:41:39funmanor manage themselves their memory (better english)
16:41:44kugelthere's a arm7tdmi #define
16:42:04 Join hd [0] (
16:42:04funmanhum perhaps that could help
16:42:14Tornei don't know what processors are relevant
16:42:18kugelso, for current targets, we could do #if ARM_ARCH >= 4 && !defined(ARM7TDMI)
16:42:22Torne has the specs for all extant ARM cores
16:42:28funmanall arm9 don't have a mmu anyway, but we could ask an expert (amiconn?)
16:42:36funmankugel: ARM_ARCH >= 4 is always true
16:42:43kugelyep :)
16:42:45Tornekugel: no, because ARM966 and 968 don't have MMU either
16:43:03kugel"for current targets"
16:43:05Torne946 has an MPU which i forget if it populates the fault status registers or not
16:43:50Tornekugel: you'd do better defining it the other way around. i think you'll find most rockbox ARM targets lack an MMU :)
16:44:13kugelprobably, yes
16:44:33*Torne checks what you get from an MPU.
16:45:10Torneyah, the fault status/address registers are also defined for MPU, though not all fault statuses are possible then
16:45:36Torneso you need something like ARM_HAS_MEMORY_PROTECTION really
16:46:06Torneto distinguish devices that actually have *some* kind fo silicon looking after memory from ones that just splat an address out onto the bus and wait for data or an abort to come back :)
16:46:54funmanthe best would be #if defined(ARM_CRAPPY_PORT_CRASHES_A_LOT)
16:47:46FrankTMfunman: btw: i did some more fuze keybindings
16:47:58 Quit ED-209_ (Read error: 104 (Connection reset by peer))
16:48:24kugelFrankTM: I find the scrollwheel quite comfortable to use in maze
16:48:25funmanFrankTM: yep i noticed, I don't know what to do for the blackjack plugin though
16:48:42funmanisn't maze the plugin which mixes scrollwheel and left/right for directions ?
16:48:54FrankTMscrollwheel does up/down
16:48:58FrankTMbuttons do left/right
16:49:02FrankTMquite annoying imo
16:50:01kugelI like that
16:50:06 Quit timc (Read error: 60 (Operation timed out))
16:50:45Tornefunman/kugel: anyway as far as that patch goes,yes, you want a define based on the presence of a mmu/mpu, and you probably want to print the status for instruction aborts as well
16:50:59Tornenote that on armv6 the instruction abort status register is seperate
16:51:38kugelwe have jhmikes for armv6
16:51:43 Quit HellDragon (Connection timed out)
16:51:58funmanTorne: arm922 manual doesn't say FAR is updated for instructions faults, only data aborts
16:52:14TorneFAR is indeed not updated
16:52:17Tornebut FSR is still valid
16:52:20 Join timc [0] (n=aoeu@
16:52:25Tornetells yuo whether it's a translation/permission/domain/etc fault
16:52:39funmanoh it's for instruction fetch?
16:52:43kugelaren't we running in supervisor mode all the time?
16:53:03Torneit's prefetch abort
16:53:09funmanoh ok
16:53:10Torneundefined instruction is a different vector :)
16:53:32Torneso, yah. armv4/5 has FSR updated for data or prefetch abort
16:53:44Tornearmv6 has a seperate FSR for data and prefetch, so you'd need to read the right one, have ARM still not published the v6 arm arm on their site?
16:55:50Torneoh, no, they have, just not updated the link title ;)
17:00:29 Join perfectdrug [0] (
17:01:03 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
17:05:35 Join TheSphinX^ [0] (
17:08:01 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
17:10:29 Join Ubuntuxer [0] (
17:13:47 Join r121 [0] (
17:14:13 Quit wark ("Leaving... for now. But I'll be back.")
17:14:24 Join wark [0] (
17:14:25 Quit Zagor ("Don't panic")
17:16:55 Join robin0800 [0] (
17:19:41funmankugel FlynDice saratoga bertrik : FS #10048 updated
17:19:58 Part r121 ("Leaving")
17:24:28 Join orblivion-work [0] (
17:26:36 Join jmillikin [0] (
17:26:57 Quit jmillikin (Client Quit)
17:27:03 Join ender` [0] (
17:30:17 Join at0m [0] (
17:30:46 Quit shodanX_ (Read error: 60 (Operation timed out))
17:32:53 Join shodanX_ [0] (
17:35:16 Quit stoffel (Read error: 113 (No route to host))
17:42:44 Quit funman ("+")
17:44:54 Quit hd (Read error: 104 (Connection reset by peer))
17:48:01 Part orblivion-work
17:51:38 Join PaulJam_ [0] (
17:54:00 Quit petur ("*plop*")
17:55:10 Join PaulJam__ [0] (
18:04:28 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:05:36 Join bertrik [0] (
18:05:50 Quit PaulJam__ (Read error: 54 (Connection reset by peer))
18:08:24 Quit mt (Remote closed the connection)
18:08:30 Join HellDragon [0] (
18:10:18 Quit PaulJam (Read error: 113 (No route to host))
18:11:23 Join PaulJam [0] (
18:11:59 Quit PaulJam_ (Read error: 113 (No route to host))
18:13:25 Quit Ubuntuxer ("Leaving.")
18:17:21 Join stoffel [0] (
18:18:22 Join omokas [0] (n=omokas@
18:18:58 Join PaulJam_ [0] (
18:18:58 Quit PaulJam (Nick collision from services.)
18:20:22 Part pyro_maniac ("Leaving.")
18:21:20 Quit kachna (Read error: 110 (Connection timed out))
18:25:05 Join kugel_ [0] (
18:25:11 Quit kugel (Nick collision from services.)
18:25:17 Nick kugel_ is now known as kugel (
18:27:01anigavshow files > all
18:27:19 Quit kadoban (Remote closed the connection)
18:27:41 Join kadoban [0] (
18:28:22 Quit HellDragon (Read error: 104 (Connection reset by peer))
18:28:47omokasshow files > all
18:28:54 Part omokas
18:29:38 Join tvelocity [0] (
18:29:59evilnickanigav: What are you trying to do/say?
18:30:12anigavshow files > all!
18:30:54evilnickOkay, but why are you typing that here? Are you trying to show all the files on your DAP?
18:31:30anigavwhat's a DAP?
18:31:33 Join Lynx [0] (
18:31:54evilnickDigital Audio Player
18:31:59 Nick Lynx is now known as Guest40080 (
18:32:06anigavyou mean like an ipod?
18:32:27anigavI call it ipod
18:32:34anigavcause that's what it says on the rear
18:32:49Mikachurockbox runs on other daps than the ipod
18:33:04anigavyeah but those suck
18:33:30LambdaCalculus37Fine, that's your opinion. But that's not the same opinion others here hold.
18:33:57anigavwhat other players are there with >= 80gb capacity
18:34:03evilnickMost of them have better hardware than most ipods and are cheaper to buy.
18:34:40 Quit bubsy ("Mrrrrreow!")
18:34:44pixelmait's also a bit off-topic here
18:35:50LambdaCalculus37And also note that the iPod Classic (the new 6th gen models available) are not compatible with Rockbox.
18:36:14anigavgood thing I got the 5.5G 80GB
18:36:21anigavdid you know that there are 240
18:36:29anigav240GB disks that fit in there
18:36:31 Join HellDragon [0] (
18:36:38evilnickstill yes
18:37:02anigavbut yeah, a flash player with 64G would be ok for me
18:38:01 Join mk429 [0] (
18:38:11***Saving seen data "./dancer.seen"
18:39:24 Quit Kupop (Read error: 110 (Connection timed out))
18:39:27CIA-38New commit by kugel (r21216): Pictureflow: Don't show the playback control one targets that can't have playback in pictureflow.
18:41:12 Join calman_ [0] (
18:41:37 Nick PaulJam_ is now known as PaulJam (
18:45:25 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:48:14 Quit Lynx_ (Read error: 110 (Connection timed out))
18:48:14 Nick Guest40080 is now known as Lynx_ (
18:50:54 Join JdGordon| [0] (i=836b0070@rockbox/developer/JdGordon)
18:52:51 Quit markun (Remote closed the connection)
18:57:37 Part pondlife
18:59:55 Join markun [50] (n=markun@rockbox/developer/markun)
19:01:06 Quit Lynx_ (" HydraIRC -> <- Like it? Visit #hydrairc on EFNet")
19:08:29 Join funman [0] (n=fun@2001:0:53aa:64c:4c9:393c:a9b7:58be)
19:09:16saratogafunman: i've been looping your latest MMU patch on my e200v2 for 90 minutes without issue
19:09:34funmansame here (playing flac)
19:09:57*kugel too
19:10:02kugelno problems here
19:10:24saratogathough the wheel light issue is kind of funny
19:10:38kugelin fact, I'm using the recent versions since a week or so without problems
19:10:46funmani always saw it on my fuze (wheel light on SD transfer)
19:11:39bertrikI had instant data aborts with one of the recent patches on my clip
19:11:42kugelI probably failed a bit when inserting the buttonlight handling in the sd driver (not checking the timeout or so), but I didn't bother to look at it since I find it really helpful in this stage
19:11:48funmanunless someone reports again filesystem corruption i'll commit it this evening so we can all work on the same codebase, and then fix minor problems
19:12:08funmanbertrik: i commented about this on today's patch
19:12:25kugelI have buttonlight set off, so I see it only on sd transfers :)
19:12:33kugelfunman: 100% agree
19:12:36saratogayes i think it should go into SVN so that people test it more
19:13:50funmansomehow if the DMA buffers aren't aligned on cache size it will causes problems: I think this is what we saw after sme random commit broke the penultimate patch (i should have versioned them, but there were too much ..)
19:13:54kugelI think the disk led would be a valuable feature for other targets too
19:14:13kugelwhat's the penultimate patch btw?
19:14:20bertrikfunman, your patch from today 16:16 works for me
19:14:22saratogafunman: I don't know about ARM specifically, but on a lot of systems accesses to unaligned memory addresses aren't always atomic
19:14:29funmankugel: the one just before the last
19:14:34saratogapenultimate is the "one before last"
19:14:46saratogalike antepenultimate (one before the one before the last)
19:14:52saratogaor preantepenultimate
19:15:03*kugel never heard those terms
19:15:06funmansaratoga: unaligned means not aligned on word size (32bits), but in this case it means not aligned on cache line size (32*32bits)
19:15:12saratogaits one of those wonderful words that native english speakers have no idea what it means
19:15:20kugelthen it goes further with pre? preprepreantepenultimate?
19:15:23saratogafunman: yeah
19:15:38funmanoops, 32 bytes
19:15:51saratogax86 has similar issues with atomic memory access across cachelines and page boundaries too
19:16:40saratogafor instance on some intel chips a load double isn't atomic (first word can be loaded, the second changed, and then loaded) if it passes a page boundary for instance
19:17:58bertrikfunman, I found another |= operation on a write-only register :P
19:18:14saratogaonce the MMU patch is in, I'll submit release canidate bootloaders to the tracker
19:18:29saratogathen we can release whenever everyone thinks its ready
19:18:46funmanthe drawback with having to pass data through a cache line aligned buffer is only noticeable on large (1MB), and unaligned (on 32 bits) transfers, so we can live with it
19:19:05funmansaratoga: shouldn't we wait 3.4 ?
19:19:28funmansaratoga: also there is some patches we must finish (clip button, e200v2 radio for instance)
19:19:38kugelfunman: if the transfer is large and unaligned or one of those?
19:19:41funmanbertrik: oh where ? :)
19:19:57saratogafunman: the feature freeze doesn't apply to new ports, so you can commit when you like
19:20:03bertrikata_sd_as3525.c # 440, it probably doesn't any harm
19:20:13funmankugel: hum i mean if the transfer is large and aligned, because if it's unaligned we have to copy to an aligned buffer anyway
19:20:18saratogathough unless the clip issue is discovered, it might make sense to just do the e200v2 and Fuze initially
19:20:23*kugel sees no reason to wait for 3.4
19:20:28funmansaratoga: i meant shouldn't we wait 3.4 for a release
19:20:41saratogasorry i mean "supported build"
19:20:45saratoganot a 3.3 release
19:21:04saratoga3.4 will be the first release for AMS for sure, but we'll offer SVN builds and rbutil support before then
19:21:18funmanlogf() is quite handy, and could help finding what's wrong on the Clip
19:21:36funmanrbutil (and mkamsboot) also need some work before it can be supported
19:21:39saratogadid anyone try shrinking the memory on the fuze to be the same size as the clip?
19:21:58kugelwhy did we release 1.0 if it needs work
19:22:26funmankugel: i wrote on rockbox-dev@ about this : we need a way to support firmwares released after compilation
19:22:43funmanLlorean suggested we provide a downloadable list of tested OFs
19:22:45kugelI'm in favor of a −−force switch
19:22:54bertrikfunman, I thought the clip button is just a case of insufficient delay between setting the rows and reading the columns
19:23:04kugelI don't feel comfortable with just accepting
19:23:13*bertrik agrees
19:23:13 Quit perfectdrug ("Leaving.")
19:23:22funmanme too, but we need to make mkamsboot support such a list, so rbutil can use it. We could make the current list builtin mkamsboot just like it is now
19:23:54saratogathis is something thats important for 3.4 but thats still al ong ways away
19:23:59funmanbertrik: that's it, but we'll need to address it one day or the other ;) I want to test your patch before modifying again the delays though, and I already have modified them in the current FS #10048 patch
19:24:05saratogafirst i just want to get binaries and bootloaders on the download server
19:24:07kugelrbutil is supposed to show the supported OF versions?
19:24:28funmankugel: not exactly
19:24:44kugelwhat does it need the list for then?
19:25:04saratogato see if a given OF is supported
19:25:10funmanIf rbutil is built on july 1st, and a new Fuze OF is released on july 2nd, and tested by us on july 3rd : people running rbutil after july 3rd can patch this OF version
19:25:16kugelthat can be seen by the return value of mkamsboot too
19:25:29saratogawe prompt the user for the OF, so they could provide one rbutil doesn't know about, in which case it would reject it
19:25:54kugelif (OF_unknown) return EXIT_OF_UNKOWN;
19:26:08kugelexit(EXIT_OF_UNKNOWN) even
19:26:29CIA-38New commit by alle (r21217): Fix case in the constant
19:26:49kugelrbutil only needs the list if it wants to show it. unsupported OFs can be handled in mkamsboot itself
19:27:12funmankugel: we want rbutil to support only OFs verified to work with mkamsboot
19:27:39funmanand we want it to support OFs tested after a specific rbutil version have been built
19:28:03kugelwithout needing a new rbutil release? so, mkamsboot is built-in?
19:28:29funmanyes mkamsboot is built-in, and the list of tested OFs (with their md5sums and their version) is downloaded from rockbox website
19:28:34kugelmkamsboot could just be downloaded
19:28:45kugelthe bootloader need to be downloaded anyway
19:29:23funmanmkamsboot is used as a library by rbutil, not as an external progarm
19:29:40kugelthat's already set in stone?
19:29:44funmanand doing a release at each new OF sandisk releases is exagerated i think
19:29:54 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:30:05*kugel disagrees
19:30:17kugelbut well, either way is fine, as long as we get something out :)
19:30:21*domonoky likes having the md5sum lis of known firmwares in svn. and it shouldnt be much work.
19:30:40 Quit barrywardell (Read error: 104 (Connection reset by peer))
19:31:27funmankugel: also the mkamsboot 1.0 release accepts any untested OF, only printing a warning.
19:31:45funmanthis is why i removed the link from the SansaAMS wiki page, so users get it from svn
19:32:13domonokymkamsboot just needs a small modifcation to accept a different md5sum list as the builtin list.
19:32:16kugelyea, I knew that, but not about the other "more work"
19:32:57bertrikcan we do it the same for ams as we currently do for pp sansas?
19:33:25domonokybertrik: what do you means with "the same" ?
19:33:45bertrikwith a built-in list
19:33:48domonokypp sansas use sansapatcher and a completly different install method.
19:34:02domonokyit doesnt use a md5sum list
19:34:20bertrikmaybe for sansapatcher we don't really have much choice because it needs encryption keys, not md5 sums
19:34:25*kugel wonders if we can extract the currently installed OF version and recommend it it or so
19:35:08domonokyams install method is similar to hxx0 installation. there the list is builtin. but no new OF firmwares expected for them.
19:36:07domonokybertrik: ah, the encyption keys. theoretically we could also make them downloadable, if we expect many new OF firmwares for them ..
19:36:12kugelpixelma: I have a fix for the color choose & statusbar problem
19:36:39pixelmanice, want a tester?
19:36:40kugelhowever, there's a similar issue when changing themes
19:37:11kugelfixing that basically needs refreshing the list view before the "Theme loaded" splash appears
19:37:32kugelthat's part of my customlist patch anyway, so it wouldn't be much of a problem
19:37:33funmankugel: i don't think we can know currently installed OF
19:37:35domonokykugel: if you find a way to extract the AMS OF, without already installed rockbox, sure :-)
19:37:54bertrikfunman, I'm running my clip keyscan patch for a few days now and haven't seen any problems yet (but I usually don't do very demanding stuff with buttons either)
19:38:06pixelmakugel: and another thing that I would have to test though - on the c200 display the preview rectangle is gone, I think it was there at some point in time, not a 100% sure though and don't know when it's gone then
19:38:17kugelwell, it should be written somewhere in the weird "firmware partition", shouldn't it?
19:38:21funmanbertrik: when fs#10048 is in svn i'll proof-test it :)
19:38:32kugelI have no idea if it's feasable, I was just wondering if it's possible
19:38:43domonokykugel: you dont have access to that via USB..
19:38:52kugelah right
19:39:06funmannot even via rockbox
19:39:30domonokyrockbox could access it. but then its already installed :-)
19:39:32bertrikoh, we're no longer using -mlong-calls for ams sansa?
19:40:18kugelpixelma: is showing the new theme just before the theme loaded splash OK? I'd think so
19:40:33kugelthat splash is annyoing anyway, and I plan to make it cancable
19:40:49funmanbertrik: no, because the IRAM is mapped just next to DRAM (= accessible via a short call)
19:41:08funmansaratoga: do you still have a fuze?
19:41:24bertrikvery nice
19:41:34*kugel hands funman a pencil and peace of paper to finally write down who has which target :)
19:41:50 Quit JdGordon| (" ajax IRC Client")
19:42:37funmanbertrik: and -mlong-calls has never been needed in the bootloader (it only accesses IRAM - or DRAM for (not in svn yet) clipv2 - )
19:42:58pixelmakugel: wouldn't it make the splash redundant, I mean if you already see the new theme?
19:43:44pixelmathe splash has the advantage that it can be "spoken" though
19:43:58kugelthe same splash is used for any config, i.e. those which don't change the appearance at all too
19:44:16kugelthe voice thing too, indeed
19:44:37 Join Rondom [0] (
19:44:54 Quit stoffel (Remote closed the connection)
19:44:57kugelredrawing the before the splash makes it look faster too
19:45:27kugelnow it seems it would load while the splash is active, which is not true at all, the splash is prolonging by a whole second
19:45:46 Join JdGordon| [0] (i=836b004d@rockbox/developer/JdGordon)
19:47:14saratogafunman: I do but its a 4GB
19:47:30kugelJdGordon|: I have a patch for the cancalable splashes, can you look at it so we can get it in after the freeze?
19:47:56JdGordon|yes, but not now...
19:48:15funmansaratoga: i wonder if you can see any difference between the scrollwheels of e200v2 and fuze (FS #10284 - e200v2: make button driver more equal to fuze)
19:49:00kugelJdGordon|: sure, I'll upload it to fs
19:49:36CIA-38New commit by rasher (r21218): Translation updates: ...
19:49:58 Join solexx_ [0] (
19:50:22saratogafunman: what should I be looking for?
19:50:44funmansaratoga: i don't really know .. i'd say only responsiveness could change
19:50:53kugelfunman: I'll look at unifying later, I'm familiar with those
19:51:00saratogai'll try it out tonight
19:51:49funmankugel: i thought unifying the button driver could be made instead of putting work in this FS#
19:51:55kugelthe main difference is that the scrollwheel can't be read on the e200v2 during lcd updates, while the Fuze's can, the e200c2 needs a bit of special handling
19:52:22kugelthe work is done, it's ready to commit it seems
19:52:33kugelunifying isn't critical imo
19:52:40AlexPrasher: Do you have a sec?
19:52:54kugel(by the means that this patch can be committed before unifying)
19:53:17funmankugel: ok
19:56:15 Join bluebrother [0] (
19:56:19CIA-38New commit by rasher (r21219): Apply r21086 to all languages. In the future, please use tools/ for such changes (the command used in this case was: tools/ ...
19:57:04bluebrotherhas there been any discussion about the beast being a 3.3 target?
19:58:17pixelmaI thought the install method was still a deal breaker
19:59:14bluebrotheryes, plus some issues with the singleboot bootloader. At least from my point of view ...
19:59:23bluebrotherwe're currently in freeze, right?
20:00:15pixelmasee topic ;)
20:00:35bluebrotheroh. Seems I'm a bit uninformed these days :o
20:00:54bluebrotherwell, I'll assume beastpatcher not being frozen. Anyone screaming?
20:01:22bluebrothershouldn't hurt as it isn't finished anyway.
20:02:54***Alert Mode level 1
20:02:54CIA-38New commit by rasher (r21220): Used the wrong patch from FS #10062.
20:04:13***Alert Mode level 2
20:04:13CIA-38New commit by bluebrother (r21221): Major W32 beastpatcher rework (FS #10220). ...
20:05:49 Quit solexx (Read error: 113 (No route to host))
20:11:17 Join FOAD [0] (
20:12:00bluebrotherdoes anyone know if sansapatcher is supposed to work with firmware version 01.02.18a? (
20:12:41saratogait shouldn't matter
20:12:50AlexPwhy not?
20:13:06AlexPI thought sansapatcher checked stuff (technical I know)
20:13:25saratogai don't think the OF version matters, we don't even need it anymore
20:14:14***Alert Mode OFF
20:17:29bluebrotherthat's the point of the question: does sansapatcher recognize that OF version? Or is that somewhat different and would result in a "no sansa found" issue?
20:18:45 Join _lifeless [0] (n=lifeless@
20:19:47CIA-38New commit by kugel (r21222): Fix a problem that the statusbar was redrawn to late and not really in sync with the lists when changing loading a (theme) .cfg or changing the colors ...
20:23:43 Quit flydutch ("/* empty */")
20:25:11 Join petur [0] (
20:25:50domonokybluebrother: pretty unlikely. the key isnt need to detect the sansa (sansa_scan). only for patching.
20:26:23bluebrotherthen someone else needs to come up with an idea why it doesn't find the sansa in that case.
20:26:33anigavwhat kind of setup do you need to compile rockbox?
20:26:36bluebrothercan't see anything wrong
20:26:56bluebrotheranigav: easiest is linux, or you can use cygwin
20:27:05AlexPor a linux VM
20:27:12bluebrotherthat's linux too :)
20:27:17saratogasince when is lack of a beastpatcher on OSX for the gigabeat a show stopper?
20:27:24 Nick jfc^3 is now known as jfc^2 (
20:27:27anigavyeah I got a quad core linux box with 8GB ram sitting around here, will that do?
20:27:45saratogamake it a build a server
20:27:48 Join lyngaas [0] (
20:27:48 Join maraz [0] (
20:27:56saratogabuilds take too long
20:28:21anigavdo you need a special toolchain or can you install it directly from debian packages?
20:28:55AlexPanigav: un in the tools directory of the source and it'll set it up for you
20:28:58bluebrotherdoes debian provide packages for gcc crosscompilers btw?
20:29:18AlexPdunno, but especially with the patches we apply I doubt it
20:29:37kugelbluebrother: I think so, but if, then surely not the versions we recomment
20:29:50bluebrotherI'd rather think if they package it they have newer versions.
20:30:31anigavbluebrother, at least for avr it does
20:31:06bluebrotheranigav: well, fedora (which I use) provides avr-gcc too, but no arm-elf-gcc (or sh or m68k) so I wouldn't count from that.
20:31:57bertrikbluebrother, just out of curiosity, are you compiling anything with avr-gcc?
20:32:16AlexPanigav: I doubt it - look at the various requirements here:
20:32:28bluebrotherbertrik: yes and no. The latest avr stuff I did was on windows
20:32:51bluebrotherplanning to move that to linux but haven't done yet.
20:33:37 Quit mk429 ("leaving")
20:34:03 Join Thundercloud [0] (
20:34:50 Quit __lifeless (Read error: 113 (No route to host))
20:35:42rasherAlexP: shoot
20:35:44anigavyeah I just found out debian has little else except avr-gcc
20:35:53anigavno big deal
20:36:09AlexPrasher: Could you have a look at my question here and let me know what you think?
20:36:28rasheranigav: the crosscompilers are self-contained. You can decide where installs them by setting a few env variables
20:37:45rasherAlexP: yes and yes
20:38:12***Saving seen data "./dancer.seen"
20:38:27AlexPrasher: Thanks :)
20:39:04AlexPI'll commit it shortly then
20:41:51kugelAlexP: hurry (my rXXX to url plugin needs more tests :) )
20:42:01AlexPhehe :)
20:43:53anigavI gota say this box really zips through the builds
20:44:32 Join Juice^ [0] (
20:44:57*funman points rasher to r21038 - language string description only changed in english.lang
20:47:05CIA-38New commit by alex (r21223): FS #10249 by Teruaki Kawashima - "Gigabeat F/X: fix wrong key in "Set Time/Date" screen". I was able to change most languages as they don't translate ...
20:50:16 Quit JdGordon| (" ajax IRC Client")
20:52:24bertrikfunman, in GPIOC_PIN(2) |= (1<<2); the |= doesn't really do anything, right?
20:53:56anigavuhh ... that's the bitwise or assignment operator
20:54:38bertrikI mean |= as opposed to a simple =
20:55:02anigavyes it does ... it sets bit 2 and leaves the others alone
20:55:18 Quit funman (Read error: 60 (Operation timed out))
20:55:33*gevaerts assumes that bertrik has a specific context in mind...
20:56:16anigavprobably ... what does the GPIOC_PIN macro do?
20:57:04domonokyit expands to a specific GPIO register :-)
20:59:59anigavbut the name suggests it addresses only a specific pin
21:00:24anigavso you'd say something like GPIOC_PIN(2)=1;
21:00:25bertrikoh just forget my question :)
21:03:37 Join JdGordon| [0] (i=836b0049@rockbox/developer/JdGordon)
21:04:08anigavthe toplevel makefile is in rockbox/tools, right?
21:04:53anigavnah that can't be
21:05:23 Quit kugel (Read error: 110 (Connection timed out))
21:05:45domonokyanigav: the main makefile is generated by configure :-)
21:05:56anigavyeah I just found the readme :P
21:06:04 Join PaulJam_ [0] (
21:06:04 Quit PaulJam (Nick collision from services.)
21:07:27 Nick PaulJam_ is now known as PaulJam (
21:07:54 Join kugel [0] (n=kugel@rockbox/developer/kugel)
21:13:49anigavwell the build took 2 minutes 55 seconds, that's not too bad
21:14:16 Quit JdGordon| (" ajax IRC Client")
21:14:17 Quit perrikwp (" ajax IRC Client")
21:14:22 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
21:15:24gevaertsthat's *slow*!
21:15:31anigavwhat do you get?
21:15:57anigavI mean with everything clean
21:16:27 Join JdGordon| [0] (i=836b0070@gateway/web/ajax/
21:16:28AlexPI don't see what showering has to do with it
21:16:39gevaertswhat target?
21:16:46anigavipod video
21:16:50anigav64M RAM
21:17:02Mikachugevaerts: presumably without ccache too
21:17:17gevaertsnot using ccache is silly
21:17:23anigavthat goes without saying
21:17:39Mikachufor a build server yes
21:20:00anigavlol I just saw that conky is using 2GB RAM :P
21:20:24 Join funman [0] (n=fun@2001:0:53aa:64c:4c9:393c:a9b7:58be)
21:21:36*gevaerts forgot that he has logs for this...
21:21:39gevaertsanigav: 83 seconds
21:21:52anigavreal or user?
21:22:14anigavwell mine takes 59 seconds real
21:22:38anigavafter using -j 5 to load all 4 cores :P
21:23:07gevaertsusing ccache for the same build the second time?
21:23:24anigavno, just make clean and then make -j 5
21:23:40anigavthe first time it used only one process :P
21:23:40gevaertssure, but do you have ccache?
21:23:48anigavIsch don't think so
21:24:13anigavnope, the package wasn't even installed
21:24:25*gevaerts measures again
21:24:50 Join tessarakt [0] (
21:24:51anigavmy linux box > your linux box
21:25:17anigavand I have firefox and conky running in the background, going nuts, leaking memory, and using cpu cycles
21:25:21saratogai get about 45 seconds on a quad 2GHz athlon
21:25:46anigavthere ... are ... no quad athlons
21:26:46Unhelpfulamiconn: i played around a bit with gcc on arm, with a test function that returns the value of a short that it's passed a pointer to. optimization off gets me the silliness with two shifts to sign-extend after ldrh, but from -O up gets me something much shorter: ldrsh r0, [r0]; bx lr;
21:27:57gevaerts63 seconds. Maybe I really should start using all cores in this machine...
21:27:59Unhelpfulthe reason i'm looking at this is that i'm pretty sure some of the jpeg data could be kept as s16 instead of s32, which would save some space... wanted to see how much it might hurt code size and performance.
21:28:08anigavgevaerts: cpu?
21:28:52 Join calman__ [0] (
21:29:20gevaertsdual L5410, running xen
21:29:23 Join hd [0] (
21:29:40gevaertsanyway, this is getting far off topic..
21:29:59anigavwhy? you want fast builds if you want to develop stuff
21:30:02saratogaanigav: sure there are you just need lots of sockets
21:30:03 Quit HellDragon (Read error: 104 (Connection reset by peer))
21:30:18anigavsaratoga: for which target?
21:30:31anigavand from a clean build dir without ccache?
21:32:14anigavgevaerts, you should supply the -j option to make with the number of your cores +1, in your case -j 9
21:32:31 Join sko [0] (
21:32:43gevaertsanigav: I know that... Also, I only use 5 cores on that VM
21:33:03JdGordon|num cores + 1 is a stupid myth like swap should be 1.5x ram
21:33:24anigavwhy? that way you get max. load on all cores
21:33:30gevaertsthat too. It really depends on type of source and machine
21:34:10anigavwell if the number is less than your number of cores you'll certainly not load them all
21:34:21JdGordon|well for one thing, do you *really* want to use the enitre cpu for the compile? what about the other stuff you have running?
21:34:52anigavyou mean like sshd? :P seriously, there's nothing much on that box
21:35:03Unhelpfulthat depends on throughput of memory, disk, filesystem, etc. i've seen better times with -j > cores+1
21:35:21anigavUnhelpful, that certainly makes sense
21:35:40anigavif a number of processes are always waiting for i/o
21:35:50anigavbut that is unlikely if you have lots of RAM
21:35:52Unhelpful-pipe being something wonderful and magical is a myth as well, btw. :)
21:36:16 Quit calman__ (Read error: 104 (Connection reset by peer))
21:36:26kugelFlynDice: haha
21:36:34bertrikfunman, replacing |= by = when setting GPIO pins (like GPIOC_PIN(2) |= (1<<2)) should be safe, right?
21:36:38JdGordon|I have the rbclient set to -j6 (cores+2).. which bassically locks up the computer for the 10s or so the build lasts for
21:36:41kugelthat was what I'm talking about (although I didn't have numbers)
21:36:50 Join froggyman [0] (
21:37:06funmanbertrik: yes
21:37:20 Join calman__ [0] (
21:37:53FlynDicekugel: wasn't even close.... Had to look for about 30 minutes trying to see where I screwed it up, but I didn't!
21:38:10kugelI'm not really surprised
21:38:34kugelwell, I didn't think of a 3h difference either, but I almost expected that the higher clocks give better runtime
21:39:41anigavgevaerts, did you manage to get better times?
21:39:53kugelFlynDice: so, what I'd find interesting would be a test with 248/62 boosted and 31/31 unboosted
21:40:41 Join merbanan [0] (
21:41:07anigavcan you install additional rockbox plugins without reinstalling rockbox?
21:41:23bertrikhm, 15:41 h is quite nice for a target in development!
21:41:24kugelthere are no additional plugins
21:41:27 Quit calman__ (Client Quit)
21:41:34anigavkugel: I'm writing one
21:41:54kugelthen you compile it with rockbox, don't you?
21:42:02 Quit r0b- (Read error: 110 (Connection timed out))
21:42:35kugelbut yes, you can install the plugin.rock off the build folder to your player as long as it's plugin API compatible
21:42:56anigavalright, that's cool
21:43:14FlynDicekugel: What I did was 62/62 boosted, 31/31 unboosted. test_codec shows something like 32.6 MHz required for realtime in this config
21:43:50FlynDiceI thought that had to be the low power winner...
21:44:08Unhelpfulkugel: and built by the same person, it would seem. i sent gevaerts and LambdaCalculus37 some svn-based test plugins that failed horribly on their devices.
21:44:37anigavwell you probably need the same compiler/lib versions
21:44:37kugelit probably required a plugin api bump?
21:44:45 Quit calman_ (Read error: 110 (Connection timed out))
21:45:31 Join barrywardell [0] (
21:46:04Unhelpfulkugel: it didn't involve API changes. it was an entirely self-contained plugin that used the pluginlib build of the scaler. only plugin API use was open/close/fdprintf and current_tick.
21:48:34CIA-38New commit by bertrik (r21224): Sansa e200v2: use = instead of |= when setting GPIO level
21:50:15kugelUnhelpful: weird
21:51:03kugelUnhelpful: well, there's many #ifdeffing in the plugin_api struct, making the thing not very cross-target compatible
21:51:05 Quit dmb (Remote closed the connection)
21:51:44Unhelpfulany thoughts on how i might trap things writing past the end of their buflib allocations? i can't just ask gdb to watch the entire contents of the buffer, obviously there are writes to it all the time :/
21:52:53amiconnUnhelpful: -O0 is sillyness for any production code. But I've seen numerous silly things in gcc generated code at -O1 and higher, for all rockbox architectures (I don't know about mips yet)
21:53:22 Join dmb [0] (n=dmb@unaffiliated/dmb)
21:55:01CIA-38New commit by bertrik (r21225): Sansa fuze: don't configure GPIO A7 in the LCD driver, this pin is used for fm radio
21:55:35Unhelpfulanyway, i think i can take a column out of the IDCT function table, and make dequantization use single hardware multiplies exclusively on SH. that loop where you fixed the binsize gain on ARM could also be changed to just shift *all* quantization table entries, and it might even be faster if it did that as it wouldn't need to do the extra lookup to get the zig-zag order of each entry it changed.
21:55:56kugelbertrik: hmm
21:56:13kugelcould that be the reason the radio worked on fuze, and not on the e200?
21:56:36kugelwith FS #10048
21:56:44 Quit wark ("Leaving... for now. But I'll be back.")
21:56:52skowould be interesting...
21:56:54 Join foobar [0] (
21:57:03bertrikI wondered about that too, but funman tried this out on his fuze already and it didn't make a difference
21:57:23Unhelpfulthe reason i'm playing with sign-extended loads is because the quantization table could also be made s16 on all targets... and i *think* the buffer used for IDCT as well.
21:58:00kugelsomeone should try it on his e200v2, and also if what the fuze does in button_init_device does make any difference on the e200v2
21:58:01bertrikif others fuze owners now suddenly start complaining about the radio, we might be on to something :)
21:58:49kugelfunman: how long till fs#10048 is in? :)
21:59:09***Alert Mode level 1
21:59:09CIA-38New commit by mcuelenaere (r21226): Update Dutch translation of RBUtility
21:59:12 Join bigtomrodney [0] (n=tom@pdpc/supporter/active/bigtomrodney)
22:00:05FlynDiceHas anyone else had problems compiling recent checkouts? I've been struggling all morning until I found I could just type make again and the compile would finish. Here's the error where compiling stops:
22:00:35*bertrik says aye to committing the current version of fs#10048
22:01:21bertrikFlynDice, perhaps a make dep would solve that, otherwise a make clean or make veryclean
22:01:32*FlynDice finally got a finished build to try ;)
22:02:09FlynDiceHow bout a brand spankin new checkout from SVN same problem
22:02:43Mikachuyou might want to mention what target you're building for
22:02:53saratogafor what its worth i think i got the fastest rockbox builds with just make -j
22:03:06 Quit AndyI ()
22:03:17saratogaalso, my e200v2 is still playing without a lockup hours later
22:03:32saratogai would say playback is quite stable
22:06:06amiconnUnhelpful: Btw, SH1 always sign-extends on load
22:06:47amiconnOtoh, coldfire needs a separate instruction. It doesn't extend at all on load
22:06:50kugelFlynDice: what problem?
22:07:19funmankugel: today if we get no problem reports
22:07:28Unhelpfulamiconn: i was wondering, the coldfire manual was not really clear on what the top half of a register looked like after MOVE.W
22:07:32 Quit anigav (Read error: 110 (Connection timed out))
22:08:02amiconnUnhelpful: The top bits are kept (from previous content)
22:08:14kugelfunman: today ends soonish, hurry up :P
22:08:35amiconnThat gives a slight advantage when a loop needs to process zero-extended data
22:09:10***Alert Mode OFF
22:09:13amiconnThis allows to clear the whole register once, before the loop
22:09:24Unhelpfuli see what you mean... i should probably make s16 quant tables and idct workspace !defined(CPU_COLDFIRE) then?
22:10:00 Quit shodanX_ ("leaving")
22:10:42amiconnDepends. If gcc is clever enough to use muls.w, it can skip sign extension (unless the first op is something else than multiplication)
22:11:43Unhelpfulhrm. dequant is multiply->store, so gcc *should* be ablet to figure it out, although we might need an explicitly 16x16->32 multiply instruction or macro for it to "get" it...
22:15:59 Quit rasher ("leaving")
22:19:52 Nick foobar is now known as anigav (
22:20:51 Quit merbanan ("Leaving")
22:21:23funmankugel: well perhaps not today since now my Fuze doesn't boot anymore
22:21:31JdGordon|Bagder: you see my patch to clean up a bit and hopefully make releasing slighty better?
22:21:39 Join low_light [0] (i=c730180b@gateway/web/ajax/
22:22:01kugelfunman: with the latest patch?
22:22:09kugelafter bertrik's commit?
22:22:26Bagderno I didn't... will check
22:22:46Unhelpfulthe idct situation is a little different, though... the 1-point and 2-point idct never multiply, and none of them multiply the DC coefficient.
22:23:50JdGordon|Bagder: 10278
22:24:30funmankugel: yes, reverting it does help.. let me see if it's a problem with the radio or not
22:25:38 Quit dmb (Remote closed the connection)
22:25:52 Quit Thundercloud (Remote closed the connection)
22:26:31kugelfunman: works here, let me update the bootloader
22:27:18bertrikfunman, I thought it was safe to commit because it seemed to work for you initially
22:27:56 Join dmb [0] (n=dmb@unaffiliated/dmb)
22:28:35funmanit worked when i last tested it, not anymore with svn
22:29:06funmanif i revert r21225 or make fmradio_i2c_* return 0 my fuze boots
22:29:45 Quit Rondom (Remote closed the connection)
22:32:33kugel(for the logs) works fine on my fuze including radio
22:33:19 Join HellDragon_ [0] (
22:33:20kugelFlynDice: can you test if copying button_init_device from the fuze helps for the radio (and maybe also check if it makes scrollwheel reading possible if lcd_button_support fails as it works on the fuze)
22:33:26 Quit hd (Read error: 104 (Connection reset by peer))
22:34:19 Quit domonoky (Read error: 104 (Connection reset by peer))
22:36:10Unhelpfuldoes jpeg require that component 0 always use quant table 0? in that case we can skip looping over the quant tables in fix_quant_tables entirely.
22:37:08BagderJdGordon: I like it!
22:38:11 Join kachna [0] (
22:38:16***Saving seen data "./dancer.seen"
22:39:06 Quit petur ("Zzzzz")
22:39:56*low_light does the sa9200 touchpad happy dance
22:40:02low_lightPhilips sa9200
22:40:05low_lightfinally working
22:40:23 Quit bigtomrodney ("Ex-Chat")
22:40:35FrankTMkugel: do you need something tested on a fuze?
22:40:45kugelno, I have one myself :)
22:40:51 Join mt [0] (n=MTee@rockbox/developer/mt)
22:40:54FrankTMah. alrighty
22:41:03FrankTMthought you were asking for that :p
22:42:34 Quit HellDragon_ (Client Quit)
22:42:36Unhelpfulhrm, jpeg also stores only the difference for the DC value from block to block, so that the DC value will be having addition done on it. but i guess if the result is stored with MOVE.W that would not need an explicit sign extension?
22:43:06skoI have a first try of unifying ready... works on my e200v2 but tons of #ifdefs -.-
22:43:44 Join stripwax [0] (
22:44:41Unhelpfuli think the size of the stored quant table can be cut in half pretty much for free. the IDCT table can have this done as well, but on coldfire it will cost an explicit sign extension for the DC coefficient in all IDCT functions, and possibly for a few other values. :/
22:44:45JdGordon|Bagder: :) do you want to take over it? the remiang part depends on how much you want to automate the whole thing
22:45:24BagderI've wanted it to become a single command line to build everything for a specific release
22:45:36kugelsko: looks good. I was thinking if WHEEL_REPEAT_INTERVAL could be the same for both
22:45:51kugelI tried HZ/4 on my fuze but found HZ/5 better
22:46:48skohmm... ok, i'll try HZ/5 on my e200v2
22:47:31funmansko: by the way you can remove one endif (line 496 of patch) by not including the correspnding '{' in a conditional
22:48:07funmankugel: my fuze doesn't boot if i undo the modification to system-arm.c
22:49:42skofunman: so curly brackets are allowed without "if" or something?
22:50:15kugelthey mark a new block which is important for name space of variables, but doesn't apply here
22:50:56kugelyou could f.e. override _dbop_din if you declared another with within the curly brackets
22:53:10 Join kachna|lappy [0] (
22:53:57funmanthat's the same symptom i had with the penultimate fs#10048 patch, but on the clip it boots fine
22:53:57 Quit mcuelenaere ()
22:54:07 Join HellDragon [0] (
22:54:37kugeltype of yours, or a bug of my parser?
22:54:44 Quit kachna (Read error: 104 (Connection reset by peer))
22:54:44funmanrockbox is stuck on the logo screen, backlight still goes off after a delay
22:55:12funmankugel: i really typed "that's"
22:55:48 Quit stripwax ("")
22:55:49funmanthe consequence would be a changed alignement of some critical parts of the code, but where, and why does it harm are still mysteries.
22:56:21kugeli thought reverting bertrik's commit helps
22:56:32kugelmaybe that's needed like the dcdc on some clips
22:57:24 Join stripwax [0] (
22:57:52 Quit LambdaCalculus37 (" damn, I'm tired!")
22:57:53 Join Thundercloud [0] (
22:58:27funmanyeah with system-arm.c unmodified and bertrik's commit reverted my fuze boots
22:58:44funmanbut then the next random commit will break my fuze again
22:59:11funmanbecause a random modification (of functions never called in system-arm.c) broke it
22:59:26kugeldo you think the plain binsize change broke it?
22:59:52kugelso leaving bertrik's commit in, only reverting system-arm.c makes it boot again too?
23:00:01kugelmaybe it's a filesystem issue :?
23:00:37 Join courtc [0] (n=court@unaffiliated/courtc)
23:05:48funmanhum fsck.vfat corrected the filesystem (both fat differed) but it still doesn't boot
23:06:25kugelfsck.vfat never worked well on my fuze
23:07:26Mikachuevery time i've run fsck.vfat on my ipod and it found errors, it still found errors on the next run, usually have to run it twice or thrice
23:08:14 Quit HellDragon (Client Quit)
23:08:15FrankTMseems like a corrupt disk
23:08:37Bagderno, fsck.vfat simply isn't good enough
23:08:50FrankTMor that :p
23:11:27 Quit low_light (" ajax IRC Client")
23:11:51 Quit perrikwp (" ajax IRC Client")
23:12:15 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
23:12:48 Join rasher [0] (
23:16:21 Quit bmbl ("Woah!")
23:17:13 Join matsl [0] (
23:20:05 Quit stripwax ("")
23:21:02kugelfunman: just try reformating
23:21:36 Join notlistening [0] (
23:21:50kugelfsck.vfat is mostly crap
23:22:48Bagderyeah, a fresh mkfs is often faster/easier
23:23:17*Mikachu has a complete backup of his huge 2GB ipod nano disk on his computer
23:23:28bertrikfunman, the symptom I had on my clip when it didn't work was that it booted fine but gave data aborts on starting a codec
23:23:28 Quit jgarvey ("Leaving")
23:23:29*amiconn wonders whether windows chkdsk works better than fsck.vfat
23:23:44Bagderamiconn: in my experience, it does
23:24:07kugelimpressive, 2GB, don't you need raid 224 for that? :)
23:24:07Mikachukugel: ;)
23:24:07kugelchkdsk always works for me
23:24:31kugelby the means that the FS is in a working state again, not that the all the data is rescued, of course
23:24:37funmanbertrik: i noticed the same, and identified memory corruption by looking at the address content reported by rockbox and the same address in rockbox.elf
23:26:16kugelthis not-booting is what I regularly see on my clip
23:26:28funmankugel: with fs#10048 ?
23:26:39bertrikat first I thought it was perhaps because of changes in configure / compiler flags, could it also be that ccache gets confused when compiler flags change?
23:26:39kugelno, in general
23:26:40funmanby the way the fuze boots after a reformatting
23:27:06kugelmy clip's FS is always borked after a few minutes of using
23:27:26kugelbertrik: that's also possible
23:27:45kugelI remember when I tried to make mp3 play with activating IRAM
23:28:55 Join kugel_ [0] (
23:29:09kugel_codecs didn't work for me, but when funman managed it, it turned out that the deps or something didn't manage that (i,e, rm -fr * in the build dir would've solved my codec issues and we had mp3 playback a month earlier)
23:29:43 Quit kugel (Nick collision from services.)
23:29:46 Nick kugel_ is now known as kugel (
23:33:19 Join stripwax [0] (
23:34:59 Quit sko ("bye")
23:36:28 Join faemir [0] (
23:37:14 Join safetydan [0] (n=deverton@rockbox/developer/safetydan)
23:38:22kugelfunman: so it can go in?
23:39:21 Quit bluebrother (Read error: 113 (No route to host))
23:39:48funmani still don't know - if the same (perhaps corrupted, perhaps not) filesystem let rockbox boot or not depending on a random modification, that's still soemthing weird
23:42:14kugelI have no issues
23:42:28kugelat least not more than with what's in SVN
23:42:55kugelI'd definitely say go for it, so that we can receive better testing and tweaking with code in SVN
23:44:25funmanif my case is isolated then it's ok, since the majority of developers can continue working
23:48:56 Join flydutch [0] (
23:50:26kugelfunman: we can't tell if the FS was really causing it or not. It was defintily corrupted, which makes behavior generally unpredictable
23:52:34funmanwhatever a wider audience testing the code will help us to fix problems
23:54:02mtGood evening everyone. Could someone give me some pointers on on-target debugging ?
23:54:37Bagdermt: splashes, usb-serial, logf...
23:56:04mtusb-serial for an arm target ? (sansa e200 is what I have)
23:56:30BagderI've not used it myself, but I hear it should work
23:56:52 Quit matsl (Read error: 110 (Connection timed out))
23:58:15mtBagder : for logf, iirc I should #define LOGF_ENABLE then #include logf.h in the file where I want to use it, then build a debug build, right ?
23:58:33saratogamt: you should consider using the IMDCT library in rockbox
23:58:36saratogaits a trivial change

Previous day | Next day