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

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

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

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

#rockbox log for 2009-06-04

00:02:18 Quit robin0800 (Remote closed the connection)
00:03:22 Quit domonoky (Read error: 104 (Connection reset by peer))
00:03:31 Join Hillshum [0] (n=chatzill@unaffiliated/hillshum)
00:06:18 Quit n1s ("Lämnar")
00:08:11 Quit bertrik ("Leaving")
00:09:06 Join bubsy [0] (n=Bubsy@
00:09:25kugelfunman: how hard would it be to use dma for dbop transfers?
00:14:08 Join stripwax [0] (
00:14:18onlysoaaYeah, so... I changed lines 161 to 175 in this file here to match the P2's OF, but it didn't seem to change the screen output:
00:14:34 Quit Moarc (Read error: 60 (Operation timed out))
00:15:41onlysoaaI also changed the lcd_display_on() and lcd_display_off() functions in this file to match the P2's OF, but that didn't change anything either:
00:16:43onlysoaaI changed the timings in lcd_init_device(), and that seemed to change some things, but it's still off.
00:17:01funmankugel: not much I think, but we need to use an available channel (1 is used for SD and the other one for PCM playback)
00:18:32funmankugel: to test performance, you would need to also figure how much data you can transfer at a time
00:18:49funmanperhaps we could switch to interrupts to transfer PCM if that's worth freeing the DMA channel
00:18:59 Quit Russel-Athletic (Remote closed the connection)
00:19:31onlysoaaIn lines 282 and 283, a function lcd_copy_buffer_rect() is declared, but where is it actually defined?
00:19:50 Join J-23 [0] (
00:21:27funmanonlysoaa: firmware/target/arm/lcd-as-memframe.S
00:21:32kugelfunman: isn't there more than 2 channels?
00:21:55funmankugel: nope, the PrimeCell PL081 has only 2 channels :(
00:22:13kugelthat's what I was asking ;)
00:22:23kugelwell, we apparently reached the maximum transfer rate already, I was just after reducing the cpu usage
00:23:56kugelfunman: I posted some numbers for the clip fps rate to FS #10272.
00:24:55kugelI'm too having button problems on my clip with the MMU patch
00:25:19funmankugel: i think your last comment on that ticket is wrong : i get blue bars on the bottom 48 lines of my clip's screen
00:25:25funmanand yellow on the top 16
00:25:45funmankugel: </joke> with down button only, or others as well ?
00:26:53funmanI had tweaked the clip button code to fix down button not responding, but now I notice again it's sometimes not working
00:27:22onlysoaaThanks, funman. (: I'll take a look at it.
00:27:25 Quit PaulJam (".")
00:28:58funmanamiconn: is CPU_COLDFIRE timer implementation for both iaudio & iriver ?
00:30:22kugelfunman: down button only it seems
00:30:42onlysoaaYeah, seems that function doesn't have to do with the blue display issue, huh...
00:30:51kugelthe blue bars are apparently caused by the MMU patch. I tried CACHEALIGN_ATTR for the framebuffer (which is in IRAM) without success
00:31:04funmankugel: if you are in a hurry you can try adding more nops in button-clip.c
00:31:22funmanI don't think this has something to do with caches (especially if it only appears when boosted) since we don't use DMA
00:31:56 Quit matsl (Read error: 110 (Connection timed out))
00:31:56funmandid you try lowering DBOP clock? bertrik was speaking about perhaps incorrect DBOP_TIMPOL_* values
00:33:58 Quit perrikwp (" ajax IRC Client")
00:34:24 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
00:34:46 Join Horschti [0] (n=Horscht@xbmc/user/horscht)
00:35:28***Saving seen data "./dancer.seen"
00:36:05 Quit ender` (" Little expense had been spared to create the impression that no expense had been spared.")
00:37:37kugelfunman: well, it gives the same speed without blue bars if fs#10048 is not applied (now in #rockbox)
00:38:35funmankugel: perhaps the problem is too much speed (both fs#10272 + fs#10048) ?
00:38:55kugelit's the same speed
00:38:58Hillshumsounds like a nice problem to have
00:39:07kugelat least as far as fps are concerned
00:39:31funmanhm don't you get the highest number of fps with both patches applied?
00:39:33kugeland from what I've noticed, 10048 doesnt change fps much
00:39:48kugel" well, it gives the same speed without blue bars if fs#10048 is not applied"
00:40:20funmanyeah i wondered about what happens with 10048 only
00:42:57onlysoaaSo, anyone got a clue what's to change that could fix the LCD on the P2? ;P
00:43:20onlysoaaIt might have to do with three lines...
00:43:38funmanby the way we don't use the GPIOB interrupt, do you have an idea of what could the Fuze/e200v2 OF Use it for ?
00:44:18onlysoaaLines 31 to 33 in lcd_cowond2.c.
00:44:19kugelfunman: no, but in the early button days I managed to read power with it
00:44:28kugelwow, this is weird
00:44:40kugeljust 1 nop\n added makes my clip unable to boot
00:44:49kugelchanging it back solved the button problem :S
00:45:20funmanI think you'll take back "solved" if you use it for a longer time
00:45:51kugelI'm pretty sure of that, yes :)
00:52:05kugelfunman: the other OFes don't use it?
00:52:14kugelit might have something to do with the scrollwheel then
00:52:35 Quit Horscht (Connection timed out)
00:52:41funmanno i have only seen mentioned in fuze & e200v2, before i leave to the mountains and when you were working on scrollwheel
00:52:59CIA-38New commit by 03mcuelenaere (r21184): Ingenic Jz4740: should fix timer (thanks to Rafaël Carré)
00:53:10funmanwell my fuze doesn't boot at all
00:53:44kugelcan you see what the ISRs do in the disassembly?
00:53:59funmanI used 50 as a delay though, not 45
00:54:08funmanhum yes but not now :)
00:54:23kugelprobably worth a look (hint hint :) )
00:54:33funmani'm still looking at timers now
00:55:17kugel50 makes it not boot?
00:55:27kugelthat's like on the clip then :S
00:57:10 Quit Hillshum ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
00:59:33funmansame with button_delay = 45
00:59:54funmanperhaps it's my tree, will try later
00:59:58kugelhm, works for me
01:00:55 Join timc [0] (n=aoeu@
01:04:53 Join Thundercloud [0] (
01:05:35 Quit mcuelenaere ("ZzZzZzZzZzZz....")
01:14:23 Join killan [0] (
01:15:29 Quit pixelma (Nick collision from services.)
01:15:30 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
01:15:42 Quit amiconn (Nick collision from services.)
01:15:45 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
01:15:46 Join obo_ [0] (
01:15:47 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
01:16:05 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
01:16:15Unhelpfulkugel: there are *lots* of places where the tag value is used as an index into a table. i don't think pre-shifting would be a huge gain, since i'd then have to calculate the tag's index value from the bitmask value, which will be more expensive than doing a shift.
01:17:39 Quit killan_ (
01:17:39 Quit trisiak (
01:17:39 Quit Zambezi (
01:17:39 Quit obo (
01:20:46 Quit jgarvey ("Leaving")
01:22:44 Quit mt (Read error: 101 (Network is unreachable))
01:24:49 Quit JdGordon| (" ajax IRC Client")
01:29:48kugelfunman: ok, I need to correct myself, the fps are quite slower without 10048
01:30:09kugelalthough, 93.5fps when boosted which is near the 100fps border
01:30:37funmando you also get 400.0fps for 1/4 screen updates ?
01:31:11kugelno, 380
01:31:39kugelit seems so fast, the pre-calculations in lcd_update_rect() get noticeable
01:31:44kugelUnhelpful: ah, ok, I understand
01:36:16kugelUnhelpful: pf is permanentely updating the display it seems?
01:36:29Mikachudo you mean constantly?
01:37:06NJointrisiak [0] (
01:37:22Unhelpfulkugel: yes, it does update constantly. stopping updates when not scrolling is on my to-do list, unless you want to tackle it first. you'll need to save a background strip of the part under the album title, in case the album title scrolls.
01:37:54kugelcan't we use core scrolling?
01:38:14kugelthat scrolling bugs me for a long time already :/
01:38:42Unhelpfulkugel: on color, it may be possible. on greyscale, i'm not really sure how?
01:39:06Mikachupf on grayscale sounds.. not so exciting? :)
01:39:28 Quit Thundercloud (Remote closed the connection)
01:39:34 Join JdGordonn [0] (n=irchon@
01:39:38kugelMikachu: how does it sound on monochrome :> ?
01:39:58Mikachu(the value wrapped around)
01:40:03kugelUnhelpful: I see it just uses putsxy. There's putsxy_scroll_somthing for that too
01:40:26Mikachui'm not the only one who reads "putsexy" right?
01:40:39kugelalthough I think that may get tricky with the pointer stuff (the scroll engine needs a pointer that doesn't change), I don't know the details though
01:40:44UnhelpfulMikachu: PF uses greylib on greyscale and monochrome targets, to simulate about 128 grey levels. it looks pretty nice on sims.
01:40:59Mikachuso you can tell which cover is which still?
01:41:08 Quit JdGordonn (Remote closed the connection)
01:41:09kugelsometimes :)
01:42:13UnhelpfulMikachu: if they're not lacking in luma contrast, sure! if they're a lot of stuff the same brightness in different colors, obviously that just turns out grey.
01:42:30kugelthe problematic thing is probably that the texts are centered - I don't know how the scrolling engine works with that
01:42:49Mikachuif it needs to scroll, it fills the entire width anyway?
01:43:08kugelUnhelpful: on target, the greylib doesn't look as nice as on the sim :( at least on my clip
01:43:25kugelalthough it's still awesome
01:44:06kugelhm, the wps can scroll centered texts too, so that should work
01:45:05kugelUnhelpful: if I'm not mistaken, the scroll engine handles the text updates, so that would not the problem
01:45:54pixelmaI think it looks quite ok on my greyscale and monochrome targets (although mine is not of the "inverted" type like the Clip is
01:45:54Unhelpfulkugel: amiconn had already suggested setting a backdrop and using core scrolling. i think the first question would be, does the backdrop need to be the size of the whole screen? can i set one that is the size of the text area?
01:46:24kugelwhole screen. but what is the backdrop needed for?
01:46:40Mikachuso the scrolling can lcd_clear to restore what was under the text
01:47:00Mikachu(i presume)
01:47:15Unhelpfulif it only clears the area under the text, we could cheat ;)
01:47:44kugelit should. It would be insanely wasteful if it cleared the while screen
01:48:25kugelalso, you see at the wps that it only clears its part (since the album art is only drawn once per track, it doesn't re-appear if it's cleared)
01:49:30Unhelpfulhrm, we could cheat, then
01:49:42 Quit bubsy ("Mrrrrreow!")
01:50:30Unhelpfulmake a strip of background, set the backdrop to point somewhere ahead of it
01:51:06Mikachuis it too expensive to set the currently rendered frame as backdrop?
01:51:25Mikachui guess if it is memcpyd, it is
01:52:07UnhelpfulMikachu: it would be best to minimize the amount of buffer needed for it, since it's sharing space with the slide cache.
01:52:22kugelthe backdrop has it's own buffer
01:52:59kugeland no, I don't think it would be expensive to set the backdrop if it saves constant lcd_updates
01:53:25Unhelpfulkugel: how do we restore the original backdrop, then?
01:53:47Unhelpfulsave a copy of it ourselves?
01:54:00kugelloading it from the settings?
01:54:01Mikachudoesn't core restore it when you exit any plugin?
01:54:17UnhelpfulMikachu: i don't know, does it? :)
01:54:38kugelit should. but if we're overrwriting it, it's lost from memory I guess
01:54:58kugelbackdrop(NULL) doesn't actually remove it
01:55:07funmanshould I put PP timer code in target/arm/timer-pp.c ?
01:55:43*kugel points funman to Mr.Someone's TODO list: Put PP properly into the new target tree system :P
01:56:35kugelbut that should work for now
01:57:54Unhelpfullooking at the signatures for lcd_get_backdrop/lcd_set_backdrop, it looks like you provide your own buffer when you mess with the backdrop, and the functions just let you set/get the pointer to that buffer.
01:58:27kugelUnhelpful: loading it from the settings would probably work, but I think only backrops which are in /.rockbox/backdrops get saved to the settings
01:58:45kugeltemporary backdrops would be lost (as they are after rebooting)
01:59:08funmanfirmware/SOURCES line 340 : "if defined(CPU_PP)" should go away since CPU_ARM is always defined for CPU_PP
01:59:47kugelUnhelpful: we could cheat and overwrite the wps backdrop
01:59:51Mikachui think you can just void * whatever = lcd_get_backdrop(); do stuff; lcd_set_backdrop(whatever);
02:00:19kugelthe wps backdrop is reloaded everytime you enter the wps anyway
02:00:25Unhelpfuli don't think it will be a problem. we just need to provide our own buffer for the backdrop, and we can fake it being larger than it is. use buflib_buffer_out to get a fixed-address segment large enough for a strip of backdrop behind the title area, and then use lcd_set_backdrop to fake it.
02:00:36kugelMikachu: the problem is allocating memory :)
02:00:40Mikachulook at grep _backdrop apps/plugins/mazezam.c
02:00:51Mikachuwhat memory are you drawing the screen into in the first place?
02:00:56UnhelpfulMikachu: plugin.c already does that around running the plugin.
02:01:01kugelplugin buffer
02:01:11Mikachukugel: so give that memory to lcd_set_backdrop?
02:01:19MikachuUnhelpful: okay, so i guess if i were bored i could remove that from mazezam
02:01:25Unhelpfulkugel: no, the screen is drawn directly into the lcd framebuffer on color targets.
02:01:58kugeloh, I got his question wrong
02:02:06Unhelpfulgreyscale ones use greylib, so they draw into a greylib 256-value greyscale buffer, and greylib takes care of drawing to the actual screen.
02:02:14Mikachuah, okay
02:02:26Mikachuand that is where the scrolling also draws, so you can't use that memory as the backdrop
02:02:29Mikachui understand the problem then :)
02:02:31Unhelpfulgreylib will probably be the hardest case for getting core scrolling to work. i'm a little unsure what to do about the tracklist, as well.
02:03:42Unhelpfulkugel: becaues you're not allowed to do direct draws to the framebuffer while in greylib. you would need to use a greylib bitmap draw to restore the strip of backdrop, and a greylib string draw to render the text.
02:03:58kugel"i'm a little unsure what to do about the tracklist, as well." −− why I meant :)
02:04:41kugelthe greylib targets need constant updating anyway, I'd think
02:05:15Unhelpfulkugel: because the backdrop in the tracklist is not static... though i suppose we could render that gradient bar into the backdrop buffer?
02:06:22 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
02:06:39 Quit FlynDice (Remote closed the connection)
02:06:43kugelthere's no backdrop in the tracklist, do we need one?
02:06:48 Join bubsy [0] (n=Bubsy@
02:07:14Unhelpfulif we can somehow hook the scroll engine's calls to clear the background and draw the text, then we'd be pretty much set to get things going on greylib as well. maybe we should ask amiconn about that, i believe he wrote greylib :)
02:07:38Unhelpfulkugel: there is a backdrop of sorts, the current album text has a gradient bar behind it.
02:07:59kugelit has?
02:08:02kugelI can't see it
02:08:42Unhelpfulit's blue on color targets, and pulses. i suppose on your clip it's blue, as well. ;P
02:08:50pixelmaI can only see the gradient in the selected track bar on greyscale
02:08:58kugelyou mean the track titles
02:09:05kugelthe album text has nothing
02:09:26pixelmaand it looks wrong there IMO (darkest is in the centre)
02:09:36kugelgood that you said it pulses, I was about to think it was a fuze-specific issue lol
02:09:50kugelI think it looks awesome
02:10:23kugelI've never noticed the pulsation before though, and I don't think it's needed
02:10:45 Join Zambezi [0] (
02:11:09Unhelpfulsorry, yes, i meant the track text... the album text lacks a background, but i assume that the scroll engine will expect the backdrop buffer to have data for that area of the screen, as well. we'll probably need a screen-sized backdrop buffer while in the tracklist. a line-sized one will do in the album list.
02:11:53kugelI don't think the scroll engine expects anything
02:12:20kugelthe problem is probably, like Mikachu said, that the scroll engine clears the text area
02:12:27funmando all PP have a coprocessor ?
02:12:39funmanhum no, for example not in the bootloader
02:12:44kugeland backdrops aren't affected by clearing while the framebuffer is
02:13:06kugelfunman: I think so yes, but I'm not sure
02:13:22Unhelpfulkugel: i would assume the scroll engine uses the backdrop to clear the text area, so the backdrop needs to have black at the address corresponding to the album title area, and the gradient at the address where the selected title would go.
02:13:28kugelUnhelpful: you don't necessarily need a backdrop for scrolling
02:13:44Unhelpfulkugel: oh?
02:14:09 Join FlynDice [0] (
02:14:11kugelclearing (clear_display() and sorts) don't clear the backdrop
02:14:43kugelI guess amiconn was saying that, if the album text covers the album art, then the album will be blank at the text area
02:15:22kugelhence, setting a screenshot as backdrop before scrolling, would make the album art "scrolling safe"
02:16:27Unhelpfulkugel: right, and i'm saying that if the scroll engine uses the backdrop buffer to do clears, that we need to have a proper backdrop buffer for the tracklist, or funny things like the menu backdrop may be copied behind the album title.
02:17:04kugelno, I don't think it works like that
02:18:04Unhelpfuli don't see how else it could work, if setting a screenshot of the slides as backdrop is suppose to make the album title work in the album list view.
02:18:34 Part toffe82
02:18:46kugelbecause the scrolling would clear the album art
02:19:04kugelsetting it as backdrop would not clear it, because backdrops aren't cleared
02:19:30Unhelpfulkugel: right, but the framebuffer has to be "cleared" by copying a strip of the backdrop to it.
02:19:56kugelyes, and since our backdrop is NULL in pf, that will be plain black
02:20:08Unhelpfuli never said the backdrop was being cleared, i am trying to say that whatever is in the backdrop will end up in the framebuffer when the framebuffer is cleared.
02:21:08Unhelpfulkugel: right, but if we *set* a backdrop, and draw the gradient bar into it, that same backdrop will be used in the track list to clear the area behind the album title (if it scrolls)
02:21:19kugelthat's right. but it won't be the main menu backdrop. it will be the background color, as we did set_backdrop(NULL)
02:21:58Unhelpful...but we can't *do* set_backdrop(NULL) and have the gradient bar work with core scrolling.
02:22:01kugelso you would set_backdrop(NULL) again upon entering the tracklist
02:22:41kugelI've not looked at the gradient bar code
02:23:29Unhelpfulkugel: you don't need to look at it. if we let core scroll the current track title without a backdrop set, then core scrolling will clear anything else that's on that line.
02:23:48kugelthat's right, yes
02:24:01kugelhow are core bars drawn?
02:24:08pixelmahow does the core gradient bar work then?
02:24:33Mikachumagic! *oooOOooh*
02:24:43Unhelpfulpixelma: i have no idea how :/
02:25:22Unhelpfulis there a hook called by scrolling before the draw? if the clear before the draw can be hooked, we don't need to screw around with a backdrop.
02:25:48kugelwell, it has to work somehow
02:25:55Unhelpfulit looks to me as if the scroll engine itself redraws the color bar or gradient.
02:26:44 Quit stripwax (Read error: 104 (Connection reset by peer))
02:27:27Unhelpfulfirmware/drivers/lcd-16bit.c:1261 has cases for STYLE_GRADIENT and STYLE_COLORBAR that make calls to draw the gradient or colorbar. we could just as easily add a STYLE_BGHOOK case that calls a function to do the background clearing.
02:28:39kugelor port pictureflows bar to the core :>
02:28:43Unhelpfula STYLE_BGHOOK case would also let us use greylib methods and the greylib draw buffer to draw the background, and then a STYLE_TXTHOOK could let us hook the font draw as well.
02:28:59pixelmabut I guess those are mot available on greyscale or monochrome though
02:29:33pixelmaalthough I think greyscale could at least have a "colour" for the bar - e.g. use a lightgrey one
02:31:08funmanamiconn: FS #10274 Move timer code in target tree - comments welcome, especially about SH7034
02:31:12Unhelpfulpixelma: they're not, 2-bit and mono only offer STYLE_INVERT, but the idea of offering STYLE_TXTHOOK and STYLE_BGHOOK would still be sound, and would not be a huge bloat for core.
02:31:33funmanI assigned the task to him but I don't know if he will get notified
02:32:17Unhelpfulwe could even have an option for color targets to use the core list styling for the highighted item background instead of the PF gradient... but i think i'll start with getting it to work with the PF gradient.
02:32:30funmankugel: the Fuze OF has a DBOP isr as well
02:33:12kugelwhat does that doo?
02:33:48kugelI don't think it does the same as my ISR, I've looked at the disassembly
02:34:14 Quit FlynDice (Remote closed the connection)
02:34:29kugelmaybe for DMA? AFAIK the OF uses DMA for dbop
02:35:32***Saving seen data "./dancer.seen"
02:36:47*kugel wonders if the scroll wheel issues a push or pop error
02:39:49funmanwell the gpiob isr is a bit complex, read/writes pins on GPIO A & B
02:40:17funmanwe don't know if it is used at all
02:40:34kugelon the fuze, you need to set A7 or so for the wheel to work at all
02:43:49kugelfunman: would be nice if you can share that part (maybe you also find out how it is masked?), I'd like to see it too
02:45:39 Quit dmb ("Leaving")
02:46:04funmanwhich part : the disassembly?
02:46:45 Join FlynDice [0] (
02:48:23kugelfunman: yes, that part
02:48:45kugelfunman: I mean on which event the interrupt is triggered
02:49:18funmankugel: well i just started reading it, but i'll look around
02:52:02kugelgood luck, I'm off
02:52:02 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
02:52:59 Join dmb [0] (n=dmb@unaffiliated/dmb)
02:54:14 Quit msi (
02:54:14 Quit Bawitdaba (
02:54:32funmankugel: bit 8 of DBOP_CTRL : IR enable set on push empty, like you did
02:54:33NJoinmsi [0] (
02:54:33NJoinBawitdaba [0] (
02:55:35 Join hillshum [0] (n=quassel@unaffiliated/hillshum)
02:56:43 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
03:02:07 Join kugel [0] (n=kugel@rockbox/developer/kugel)
03:02:35kugelfunman: I did on push fifo full (bit 5)
03:02:56kugelbit 8 was an experiment to indicate finished transfer, but it wasn't any faster
03:03:31funmanthe "trick" to be able to read wheel from dbop also happens in the dbop isr
03:03:59funmandepending on a variable (i suppose it's our lcd_button_support)
03:04:03kugelthat could have something todo with the red pixel we need to write
03:04:32funmanOF writes 0xF8FF
03:04:44kugelwe simply set lcd_busy in the update functions, we could also use an isr for that
03:05:03funmanwhile we write 0xF000
03:05:23kugelyea, I guess it doesn't matter much, as it's outside visible area
03:05:52kugelwriting any value works, as long as it's 0xFXXX
03:06:03funmanOF doesn't use lcd_window there, so i suppose it does it in all its lcd_update funcs
03:08:12funmanby the way in FS #10272 i don't think you need to clear bits in DBOP_CTRL inside the isr
03:08:19funmanthere is clear bits only for error interrupts
03:08:37funmanbtw you should check these bits when you have blue lines
03:08:55kugelthe error ones?
03:09:17funmanyep, fifo push error
03:09:25kugelyea, I planned to, but didn't really get to code today :/
03:09:46funmanputting the irc window elsewhere sometime helps ^^
03:10:42kugelGood idea, I'll do it now. Please let me know of any findings, I suspect the ISR does some more magic if it involves GPIOA and B
03:10:48 Quit kugel ("exit(0);")
03:11:22 Quit funman ("leaving")
03:25:25 Quit hillshum (Remote closed the connection)
03:27:40 Quit onlysoaa (" ajax IRC Client")
03:33:29 Quit SirFunk_ (Read error: 104 (Connection reset by peer))
03:37:38 Join hillshum [0] (n=quassel@unaffiliated/hillshum)
04:24:37 Join chandoo [0] (
04:27:36 Quit froggyman ("CGI:IRC")
04:30:35 Join mib_3be8l3 [0] (i=42248748@gateway/web/ajax/
04:30:44mib_3be8l3I'm back. :D
04:31:20mib_3be8l3Whoa, what's up with this nick?
04:31:24 Nick mib_3be8l3 is now known as onlysoaa (i=42248748@gateway/web/ajax/
04:35:33***Saving seen data "./dancer.seen"
04:48:37 Join webguest00 [0] (n=630272c1@gateway/web/cgi-irc/
04:49:19 Quit webguest00 (Client Quit)
04:52:00 Quit chandoo ("Leaving")
05:15:35 Quit patmulchrone (Read error: 104 (Connection reset by peer))
05:24:49 Quit Zambezi (
05:25:34NJoinZambezi [0] (
05:25:43 Quit Zambezi (Excess Flood)
05:26:11 Join Zambezi [0] (
05:30:04 Quit hillshum (Read error: 104 (Connection reset by peer))
05:32:38 Quit lucent (Connection reset by peer)
05:32:45 Join patmulchrone [0] (
05:42:46 Quit Horschti ("Verlassend")
05:55:10 Part BUMBACL0T
06:03:15 Quit patmulchrone (Read error: 110 (Connection timed out))
06:24:14Unhelpfulurgh... if i want to place a scrolling line of text at some arbitrary coordinate, i guess i need to set up a viewport? it seems that unlike putsxy, the scroll functions use lines :/
06:27:38 Nick pabs_ is now known as pabs (n=pabs@
06:32:58 Join VytenisS [0] (
06:35:10 Join HonYu [0] (
06:35:34***Saving seen data "./dancer.seen"
06:40:07Unhelpfulthat's the biggest hurdle so far. maybe this is actually going to work ;)
06:44:36 Quit _lifeless (Read error: 60 (Operation timed out))
06:45:46 Join kadoban [0] (
06:57:21 Nick obo_ is now known as obo (n=obo@rockbox/developer/obo)
07:04:23 Join hd [0] (
07:04:32 Join _lifeless [0] (n=lifeless@
07:05:22 Quit hd (Client Quit)
07:06:37 Join hd [0] (i=jd@
07:12:42 Join Sn0wm4n [0] (i=4b42bb91@gateway/web/ajax/
07:13:48Sn0wm4nDoes rockbox call the rockbox OS just firmware?
07:14:32Sn0wm4nin the manuals it doesn't mention the "Rockbox OS" anywhere
07:14:40Sn0wm4nso do they just call it the firmware?
07:14:50scorcheyes, we call it a firmware, if that is what you are asking
07:15:35Sn0wm4nbut it technically isn't firmware since it is located on the hard drive and not the ROM of the device right?
07:16:02cool_walking_That doesn't mean anything.
07:16:08cool_walking_Some OFs are stored on the HDD
07:16:20scorcheOF meaning original firmware
07:16:59Sn0wm4nfirmware means that it is stored on ROM
07:17:20cool_walking_Since when?
07:17:36scorchethere isnt any sort of strict definition for the terms...firmware is just typically used to mean OSs or programs on embedded or small electronics
07:17:38Sn0wm4nsince the meaning of firmware
07:17:57cool_walking_I've never heard that before.. Most firmware I've encountered is on flash memory.
07:18:07Sn0wm4never heard of "flashing firmware"?
07:18:16cool_walking_exactly. flash isn't ROM
07:18:24Sn0wm4nflash = type of ROM
07:18:39UnhelpfulSn0wm4n: no, flash is not Read-Only Memory. it can be written.
07:18:40Sn0wm4noh nm I'm thinking of something different
07:18:54Sn0wm4nI know, I corrected myself
07:19:01scorcheSn0wm4n: is there some sort of point to this, or did you just come in here because you are trying to force your definition for a term onto us?
07:19:31Sn0wm4nI'm wondering where everything is stored and why there is no RockBox OS and just RockBox firmware
07:19:31Unhelpfuland i own several devices which don't have flash or ROM, and have to have their firmware sent to them on init - i assume they have a bit of RAM they copy it into.
07:19:43 Quit HellDragon (Read error: 110 (Connection timed out))
07:20:17scorcheSn0wm4n: it depends on the device...on some Rockbox can be flashed onto an actual chip rather than on the general storage medium
07:20:39scorcheSn0wm4n: call it the rockbox OS or call it the rockbox firmware if you want...i dont really care much...
07:20:55Sn0wm4noh yah... because flash doesn't lose integrity when it loses power
07:21:01Sn0wm4nnor does a hard disk...
07:21:21Sn0wm4nbut the bootloader is always on ROM correct?
07:21:47scorcheSn0wm4n: again, it depends on the device
07:22:59Sn0wm4nso as long as it doesn't lose integrity when it doesn't have power, it can hold the bootloader and firmware/OS?
07:23:35cool_walking_As Unhelpful said, firmware doesn't even need to have persistent storage.
07:23:46cool_walking_It can be loaded at every boot.
07:24:25Sn0wm4nbut then how would the device load?
07:25:08scorchethrough some sort of networked/serial/other external connection
07:25:58scorcheof course we dont do this with Rockbox, but the point is, these terms have no distinct definition in typical with all technology, terms are fluid and change as technology changes
07:25:58Sn0wm4nbut wouldn't the networked/serial/other external connection be a either powered or persistant storage?
07:26:24cool_walking_It's not part of the device though.
07:26:52scorcheand i am still struggling to see the point of this discussion
07:26:55UnhelpfulSn0wm4n: i have wifi and hdtv cards that both need their firmware loaded by the host computer's OS before they will work.
07:27:47Sn0wm4nso you would use another computer to store the bootloader/firmware in RAM or whatnot to then begin the boot process
07:27:57Sn0wm4nscorche: I'm just asking some questions
07:28:19 Join einhirn [0] (
07:28:20cool_walking_If this isn't about Rockbox it should go to #rockbox-community so as not to clutter the logs.
07:29:04Sn0wm4nnm then thanks for your help
07:29:09scorcheSn0wm4n: yes, but this channel is strictly for rockobx-related discussions seems you are more concerned about the term "firmware" than Rockbox itself...if so, we have another channel where other topics can be discussed, or if not, return to something about Rockbox..
07:32:31 Quit Sn0wm4n (" ajax IRC Client")
07:47:15 Join matsl [0] (
07:56:08 Quit jmillikin (Read error: 60 (Operation timed out))
08:02:36amiconnUnhelpful: The backdrop needs to be fullscreen for performance reasons. This way, the address difference between framebuffer entry and corresponding backdrop entry is the same for each and every pixel
08:03:28amiconnAlso, it seems that scrolling is more difficult to understand than it really is...
08:05:45amiconnScrolling text always requires to draw both the foreground and background pixels of the text (DRMODE_SOLID), or alternatively clear the whole rectangle first, then draw the text using DRMODE_FG
08:06:06 Join tvelocity [0] (
08:06:51amiconnThe result is the same with plain clearing, but the latter method is used for the gradient bar, where the background is not cleared to a uniform rectangle, but by drawing the gradient
08:08:15amiconnAll this is independent of whether a backdrop is set or not. The only difference between backdrop and plain background colour is what DRMODE_SOLID uses for the background pixels (either a copy of the corresponding backdrop pixel or the plain background colour)
08:08:57 Quit kachna (Read error: 113 (No route to host))
08:12:09Unhelpfulamiconn: i added hooks for drawing the background, drawing the text, and updating the LCD to lcd_scroll_fn - these can be used in color to draw a background strip saved from the rendered slides screen, and also in the greylib case to draw to the greylib buffer instead of the lcd framebuffer.
08:13:48 Quit kkurbjun (Read error: 110 (Connection timed out))
08:15:10Unhelpfulthe way i was planning on cheating with the backdrop wouldn't have broken that requirement, actually - i was going to use a screen-width strip, and set the backdrop to some number of whole rows before that. but i think the hooks are better, i can't really see any other way to use core scrolling with greylib, and maybe something else can use them later.
08:23:46 Join ender` [0] (
08:25:40 Join Rob2223 [0] (
08:32:01 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:35:35***Saving seen data "./dancer.seen"
08:37:39*amiconn doesn't like the char array solution for SH in tagcache
08:37:57amiconnIt means the same information is added twice, and needs to be kept in sync by hand
08:38:06amiconnVery error prone...
08:38:26 Quit safetydan ("Leaving.")
08:40:20 Quit martian67 (Connection timed out)
08:42:32amiconnUnhelpful ^^
08:42:39 Quit Rob2222 (Read error: 110 (Connection timed out))
08:42:56Unhelpfulyou'd prefer a 1<<N LUT? i wonder if it might be useful to have a global one on SH.
08:43:10amiconnThat's what I was thinking
08:43:17 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
08:43:19amiconnIt could even be a rockbox macro
08:44:53amiconn#ifdef CPU_SH / const unsigned bittab[32] = {.....}; / #define BIT_N(n) bittab[n] / #else / #define BIT_N(n) (1<<(n)) / #endif
08:45:16amiconnThen use BIT_N(n) instead of 1<<n wherever n is variable
08:45:52amiconnMaybe the array should have a better name than just bittab[]
08:46:11Unhelpfulif there are other places it would be used, that's superior to having an LUT for each set - it can speed up similar bit-testing operations elsewhere.
08:47:38amiconnPlugins would need their own bittab[] for linking
08:49:10Unhelpfulor a redef of the macro, and a reference to the core bittab in the plugin API
08:55:06 Join n1s [0] (n=n1s@rockbox/developer/n1s)
08:59:32amiconnThat would mean double lookup -> slowdown. The whole point of this is speed
08:59:58 Join flydutch [0] (
09:02:51 Quit VytenisS (Remote closed the connection)
09:05:57Unhelpfulfair enough, especially if use is going to extend beyond tagcache. for tagcache, even that would still be an improvement over array scanning.
09:07:24Unhelpfulugh, this is maddening... i can't figure out what's going wrong with drawing my copied background strip back into the framebuffer :/
09:08:21 Quit ender` (" HydraIRC -> <- It'll be on slashdot one day...")
09:09:33 Join ender` [0] (
09:10:03 Join kachna [0] (
09:15:38 Quit HonYu ("Ex-Chat")
09:17:29 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
09:20:19Unhelpfulalso, it looks like i need to hook lcd_puts_style_offset too, or else move the hooks there and have lcd_fn_scroll call lcd_puts_style_offset instead of lcd_putxyofs. that might even make some sense, since it looks like that duplicates most of the lcd_fn_scroll logic aside from updating scroll position.
09:27:39 Join Thundercloud [0] (
09:28:49AlexPSo are we freezing today?
09:29:55scorcheit looks like it will be a bit more than 39 degrees here today, so no
09:30:25*n1s feels the cold
09:30:25AlexPI am of course referring to freezing for 3.3 :)
09:30:45AlexPWhich seemed to be the decision on the mailing list when it came up at the end of last month
09:30:55AlexPSo that the release is out before devcon
09:31:14AlexPThread starting here:
09:33:55Unhelpfulwoohoo! core scrolling of the album title is (mostly) working on color. just need to move those draw hooks so that the initial draw doesn't wipe the background.
09:34:29 Quit r0b- (Read error: 60 (Operation timed out))
09:34:38 Join r0b- [0] (
09:40:25 Join Sl5avka- [0] (
09:43:19n1si guess we should check with pondlife if he thinks FS #8894 is good enough to commit
09:43:35pondlifeShould I read the logs?
09:43:36 Quit cool_walking_ (D-lined)
09:43:45pondlifeI think it's ok, but would like an e200 test
09:43:56pondlifeI was thinking post-freeze, maybe
09:44:07pondlifeUnless you want a new feature for 3.3...
09:44:15n1spondlife: not much in the log, just discussing freezing today and it would be a shame for that patch to miss it if it's ready
09:45:02 Quit timc (Read error: 110 (Connection timed out))
09:45:10pondlifeI won't have time to commit it this week, but feel free if you (or anyone else) wants to.
09:45:20n1swell, post-freeze is fine too :)
09:45:29pondlifeSafety first
09:46:51pondlifeAlso, more testing is always good - Beast and c200/e200 mainly, but keymaps in general
09:47:15n1si can maybe test on my beast tomorrow
09:48:15pondlifePlease do
09:48:51 Join timc [0] (n=aoeu@
09:49:42 Join HellDragon [0] (
09:50:00 Part Sl5avka-
09:55:14 Quit Thundercloud (Remote closed the connection)
10:00:20 Quit hd (Connection timed out)
10:10:51 Quit einhirn (Read error: 54 (Connection reset by peer))
10:11:16 Join einhirn [0] (
10:13:55 Join pyro_maniac [0] (
10:21:06 Nick fxb__ is now known as fxb (
10:29:38 Join ch4os_ [0] (n=ch4os@gentoo/user/ch4os)
10:30:43 Join sanuser [0] (n=63b4f6db@gateway/web/cgi-irc/
10:31:51 Quit sanuser (Client Quit)
10:32:16 Nick ch4os_ is now known as ch4os (n=ch4os@gentoo/user/ch4os)
10:32:47 Quit kachna (Read error: 60 (Operation timed out))
10:33:51 Join kachna [0] (
10:33:59 Join Chesteta [0] (n=Chesteta@
10:34:19Chestetahello, I have an idea/suggestion for the sansa ams devices.
10:35:38***Saving seen data "./dancer.seen"
10:36:12Chestetasince the original sansa firmware creates folders and files when it boots up (which is difficult to stop if one wants to transfer files) is it possible to enable 'hiding' files, or making a txt document in the rockbox directory containing directories we could hide?
10:37:33ChestetaI realize it is possible to show only supported files or just music, but there are still folders that are created by the firmware hangin around in the file browser
10:38:46scorcheChesteta: it isnt difficult to stop if you use rockbox's USB stack which is enabled in the current builds ;)
10:39:05 Join webguest99 [0] (n=45ee3317@gateway/web/cgi-irc/
10:39:36Chestetaoh hmm... when was that enabled? I am using an e200v2 so perhaps it is not activated for that yet?
10:40:01scorchethat isnt even a supported device...
10:40:01 Quit webguest99 (Client Quit)
10:40:08Chestetano... :/
10:40:25Chestetabut it was an issue on my e200v1 also ;p
10:40:57gevaertsyou know you can just ignore the music directory, right?
10:41:29Chestetalike me not paying attention to it or do something so it does not show up in the file browser?
10:43:02ChestetaI mainly mention the 'issue' because when driving and stuff it would be nice to have less to scroll through, that way i could put my artists in the root directory instead of in a seperate music folder
10:43:56tmztdo dos attributes matter?
10:44:27scorchethey do if "view all" isnt selected
10:44:44tmztand OF doesn't care?
10:44:59scorcheit is already marked as hidden
10:45:11Chestetaill try that, i think the firmware changes them back...
10:45:20Chestetaon my sansa they show up as not-hidden
10:46:46Chestetagimme a coupple mins and ill report back, im waiting on a music xfer
10:47:31 Join pyro_maniac1 [0] (
10:49:32tmztwhich sansa?
10:50:38Chestetae200v2... the e200v1 has a broken screen (why i got the v2 ;)
10:51:37tmztoh, I was going to say you can use usb mode and not boot OF, for v1
10:52:10Chestetayea, wish i could for the v2; how do you like the usb mode?
10:52:31tmztI lost my v1, that's not on-topic here though
10:53:06Chestetaif you like the usb mode or loosing your v1?
10:54:32pondlifegevaerts: Don't suppose you've had (or will have) time to try the latest FS #8894 patch?
10:55:06tmztChesteta: since I lost the v1 I don't have usb mode anymore, I did like it very much for the reason you mentioned among others
10:55:40gevaertspondlife: not yet
10:55:55pondlifeOK, just wondered
10:56:07pondlifeNo rush really.
10:59:45Chestetaoic, thanks, just tested things out; using the hidden attribute does work when setting it to Music or Supported, there are 2 folders showing up however (besides the microSD1) named ##MUSIC# and ##PORT#
11:01:52tmztthose are related to mtp?
11:01:53 Quit pyro_maniac (Read error: 110 (Connection timed out))
11:03:09Chesteta##MUSIC# and ##PORT# do not show up in windows explorer... even with show hidden files and folders (and the protected operating system one) unchecked...
11:03:23Chestetaso I am not sure where they come from
11:03:39tmztwhich mode do you use your sansa in for copying files?
11:03:48tmztin OF
11:04:10Chestetaumm I believe its MTP... the one that just shows the drives in 'my computer'
11:04:33ChestetaI will try switching it and see how it goes
11:05:52Chestetagee it'd be nice to have the rockbox usb mode... then i wouldnt have to 'refresh my database' everytime i unplug hah
11:07:12ChestetaI was in MSC actually (when the 2 drives showed up as harddrives)
11:07:36tmztright, and those folders don't appear?
11:08:09Chestetathey dont appear on the computer but in rockbox they do
11:11:40ChestetaIn MSC mode (the mode where the player shows up in my computer rather than the indiviudal drives) I notice that the contents of the device are the same directories shown within the ##Music# directory showing in rockbox so I believe MSC is related to the two folders
11:12:24gevaertsyou mean MTP
11:13:50tmztMTP is more like ftp or another protocol, MSC is a raw scsi command/block storage device protocol in which the internal memory and sd card appear to the computer as hard drives like with any other "usb stick"
11:13:51Chestetaumm MSC, MSC shows an icon of the sansa in my computer, 'within' that icon you can see battery status, and the two drives (MSC seems to be more 'consumer friendly')
11:14:15tmztis it MSC and Auto?
11:15:00Chestetaoh I am sorry, you are correct :/ I could have sworn I changed it to MSC
11:15:57Chestetaso to set things straight, MTP is related to the ##MUSIC# and ##PORT# folders
11:16:36tmztif you don't want to use MTP you should be able to delete those and just copy files like a hard drive
11:16:57tmztbut then you get the updating database screen when you unmount/eject the device
11:17:47Chestetawell I have always been using MSC actually, this is the first time since I installed rockbox I ahve gone back to MTP
11:19:03Chestetaso if I would like to delete the two 'mtp folders' where would I do that, because in MSC mode they do not show up in windows explorer, also, in MTP mode, the ##MUSIC# folder appears to BE the root directory for the player
11:19:36tmztI usually use linux for that kind of thing, maybe rockbox file manager/browser will be enough though
11:19:51Chestetabasically I am not sure where I can access the ##MUSIC# and ##PORT# to delete or hide them
11:20:13Chestetacan one delete things or change attributes in the rockbox file manager?
11:20:20tmztyou can try the command prompt in windows, and attrib +s could be enough
11:20:40n1sChesteta: you can delete things in rockbox, not change attrs though
11:21:51 Quit perrikwp (" ajax IRC Client")
11:22:00 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
11:23:09Chestetahmm, well in MSC mode I do not see the ##MUSIC# or PORT directories when using the dir /a command
11:23:53tmztsomething else is hiding them then, I would suggest trying to delete them with rockbox
11:24:15Chestetaok, how does one delete with rockbox? never done that before...
11:26:34 Quit ch4os (Read error: 110 (Connection timed out))
11:26:36 Join biengo [0] (
11:28:25 Join gregzx [0] (
11:34:52 Join stripwax [0] (
11:38:48 Join kugel [0] (n=kugel@rockbox/developer/kugel)
11:38:59n1sChesteta: you could for example look in the manual ;)
11:40:00ChestetaYea, sorry, I did that ;) I have not ever deleted and figured it would have been a quick answer
11:41:27n1ssure, if we knew which player you had and remembered which button to press to get the context menu... also we try to direct people asking questions answered in the manual to it so they know where to find info on other things they might want to know
11:42:06Chestetatrue, understood
11:44:03ChestetaWell, I managed to delete the two directories however after going into the OF (for a file transfer or something) they re-appear so I am not sure if there is a perminant solution
11:44:21tmztdid you put the usb setting in MSC then?
11:44:35biengopondlife: just testing your patch on the e200
11:44:48tmztit might recreate them no matter what, sorry if that's the case
11:45:07pondlifebiengo: Thanks, post back with results, good or bad
11:45:27Chestetayea... not your fault, thank you all for your help; atleast I got rid of most of the files/directories (with the hidden attribute)
11:45:29biengoseem fine so far; skip track and go to previous is a bit lagging, but nonetheless
11:46:40 Quit stripwax ("")
11:49:23biengobit confusing that speed above 100% actually is slower
11:50:05 Join hasmind [0] (n=hasmind@
11:50:16 Part hasmind ("Leaving")
11:51:05biengoah, no. that only happens if you change pitch and speed with the scrollwheel
11:54:08 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
11:54:15 Nick fxb is now known as fxb__ (
11:55:56 Quit Chesteta ()
11:56:07 Quit HellDragon (Success)
12:00:32biengothe control interface is a bit daunting to me. With Rec you can cycle through 3 states. state one says "pitch". although it shows arrows up/down/fw/rew, the pitch can only be altered with the scrollwheel. Play returns to wps, "down" has no effect, fw/rew 2% make the percentage flicker right back
12:01:33 Quit pyro_maniac1 ("Leaving.")
12:03:25biengoStage 2: says "Semitone up/down", button presses are same as above, but the scrollwheel steps are 6% instead of 0.1%
12:05:03pondlifebiengo: Stage 3 is the new one
12:05:37pondlifeI'm not sure how it maps to a wheel though
12:05:45biengois there a reason that there are tow symbols for rew/fw in stage 3?
12:05:51pondlifeOn a keypad, up/down control pitch and left/right control speed.
12:06:00biengoboth long and short presses just alter speed
12:06:38*pondlife checks the e200 keymap
12:06:45biengothe wheel clockwise increases pitch and decreases speed
12:07:42biengobut speed percentage is displayed less than 100% when it's fast and vice versa
12:08:02pondlifeOK, so the wheel is adjusting pitch, and left/right is adjusting pitch
12:08:32pondlifeWhen in this mode, any pitch change also adjusts speed to maintain constant apparent speed
12:08:38biengoleft/right adjustes just the speed
12:09:04pondlifei.e. making pitch higher will slow down actual speed so it sounds the same speed
12:09:22pondlifeI don't know if I'm making much sense there ;)
12:09:59biengoaah, the scrollwheel just adjusts the pitch
12:10:25pondlifeBut this also changes the % speed to keep the same BPM, say.
12:10:33biengobut the speed-percentage is also displayed adjusted; this is just relatively
12:10:50pondlifeNo Data Aborts seen ?
12:11:15pondlifeI know the sound quality isn't great, but that can be worked on after initial commit, I think.
12:11:20biengono, all seems fine, except that the concept is too new to grasp in a sec
12:11:41pondlifeI really need somebody to update the manual.
12:11:54biengosound is fine, clicking artefacts only occur while adjusting
12:12:05pondlifeThat should be someone who (a) can build manuals and (b) can express ideas clearly.
12:12:15pondlifeI am neither of those :)
12:12:40pondlifebiengo: It's really meant for spoken word more than music at the moment.
12:13:10biengoyeah, I'm testing it with the hitchhiker
12:13:42biengoit really is good, awesome work, pondlife!
12:13:57pondlifeI did very little of it. Just the UI really.
12:14:11pondlifePlus nagging people to test it ;)
12:14:31pondlifeNicolas Pitre and Bryan Vandyke did most of the code.
12:19:34pixelmapondlife: can you not build the pdf manual at least?
12:19:42biengoSo where would this belong in the manual. The Sound Setting?
12:20:04pondlifeNot at the moment, I moved to a new PC and never got back to that point in my Cygwin setup.
12:20:12pondlifeI should fix it up, but won't have time in the near term.
12:22:05pondlifebiengo: I guess the enable should be mentioned in sound settings, the rest in the pitch screen.
12:24:06pondlifeAnyway, I have to go
12:30:14 Join gregzx_ [0] (
12:35:39***Saving seen data "./dancer.seen"
12:35:50 Quit gregzx (Nick collision from services.)
12:35:59 Nick gregzx_ is now known as gregzx (
12:38:28 Quit saratoga (D-lined)
13:04:03 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:09:56 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
13:10:13 Quit linuxstb (Read error: 104 (Connection reset by peer))
13:10:43 Join pyro_maniac [0] (
13:13:54 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
13:18:10amiconnbiengo: The temporary +/-2% are there on purpose
13:18:35biengosorry, my first time on irc
13:18:49 Part biengo (" - Chat comfortably. Anywhere.")
13:19:11 Join wark [0] (
13:32:24 Join biengo [0] (
13:33:12biengoamicon: why is that so?
13:42:08 Quit kugel (Read error: 110 (Connection timed out))
13:45:46 Join funman [0] (n=fun@rockbox/developer/funman)
13:45:53 Quit linuxstb (Read error: 104 (Connection reset by peer))
13:57:47CIA-38New commit by 03mcuelenaere (r21185): Fix RoLo on MIPS targets
14:00:23 Join stoffel [0] (
14:05:56 Join __lifeless [0] (n=lifeless@
14:08:36CIA-38New commit by 03mcuelenaere (r21186): Ingenic Jz4740: fix debug messages not being readable due to backdrop
14:10:10 Quit _lifeless (Read error: 60 (Operation timed out))
14:12:53 Nick J-23 is now known as Moarc (
14:15:43 Nick Moarc is now known as J-23 (
14:29:12 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
14:35:41***Saving seen data "./dancer.seen"
14:39:40 Join HellDragon [0] (
14:40:20 Quit funman ("leaving")
14:41:17 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
14:43:47biengopondlife: I uploaded a diff that extends the manual to include Time Scaling.
14:44:40 Join hd [0] (
14:44:42 Quit HellDragon (Read error: 54 (Connection reset by peer))
14:45:35biengoAlso, is there a reason it is called Timestretch in the sound settings menu and Time Scaling in the pitch menu?
14:51:15 Quit biengo (Remote closed the connection)
15:01:02 Join kkurbjun [0] (n=kkurbjun@rockbox/developer/kkurbjun)
15:09:20 Join biengo [0] (
15:11:11 Join jgarvey [0] (
15:11:34 Join _lifeless [0] (n=lifeless@
15:20:53 Quit jgarvey ("Leaving")
15:21:08 Join jgarvey [0] (
15:22:14 Quit fyrestorm ("lamers envy me like they envy bill g -- main boot xp, just the way it should be!")
15:22:43 Join fyrestorm [0] (
15:22:46 Quit bmbl ("Woah!")
15:29:20 Quit __lifeless (Read error: 113 (No route to host))
15:32:50 Join yomamathecablegu [0] (n=4b5904fe@gateway/web/cgi-irc/
15:33:15 Quit yomamathecablegu (Client Quit)
15:33:32 Join arcticfang [0] (n=4b5904fe@gateway/web/cgi-irc/
15:33:46 Join __lifeless [0] (n=lifeless@
15:35:03arcticfangHey everyone ^_^ Just wondering, is there any way to make save games on rockboy? Cause it's getting really annoying playing something for an hour on my nano and then having to start it all over again the next time I feel like playing it. :/ If anyone can help me with this, they will be like a god. XDD
15:37:28evilnickYeah, check the manual and find out what button (or button combo) brings up the menu within Rockboy
15:37:54evilnickAnd there should be an option to save the current state
15:38:26 Join Rondom [0] (
15:39:24 Quit kachna (Read error: 110 (Connection timed out))
15:39:54Torneif the game actually has savegame support then rockboy should automatically back that up to a file when you quit, and load it when you load the rom
15:40:02Tornethat's seperate from the savestate support that evilnick describes
15:43:03arcticfangOkay 2 questions (probably both n00b questions, but oh well XP) 1) Which button is the "Switch" button (i.e. Hold Switch button) and 2) How do I quit from rockboy without hard resetting?
15:43:26Tornethe hold switch
15:43:34Torneas in turn the hold switch on
15:43:47Torneyou quit from rockboy by turning the hold switch on and selecting quit from the menu
15:44:01Tornethat will automatically back up the cart's SRAM if you saved in-game
15:44:16Torneyou only need to use savestates for games with no actual save support at all, or if you want to save inbetween save poitns or similar
15:44:44evilnickTorne: Thanks for the clarification, I should have mentioned that in my reply.
15:44:59Torneno problem. we've gotten to the root of the problem this way, anyway
15:45:08Tornehard resetting the ipod means rockboy has no chance to write out the SRAM
15:45:26Torneit doesn't actually save to a file on the ipod when you press the game's save option, you have to quit rockboy properly for that to happen
15:45:35arcticfangTorne & evilnick = gods. XP Thanks so much! It works! I'm glad I can come here for support when I need it. :) lol
15:46:00Tornethe hold switch is not the most obvious button, admittedly :)
15:46:02 Join ch4os [0] (n=ch4os@gentoo/user/ch4os)
15:46:08Tornebut all the other buttons are needed to have a semblance of sensible gameboy controls
15:46:12Torneso, hey
15:46:22evilnickTrue, but the lack of buttons on the iPods kinda prohibit using any other one
15:46:28arcticfangWelll yeahhh it was a n00b question, I admit XD I always thought it was called the lock button lol
15:46:36 Part ch4os
15:46:58Torneit might well be called that by someone, somewhere. the manual/wiki for rockbox calls it the hold switch though
15:48:26arcticfangI see. And now I know that for future reference ^_^ Again, thanks! Now those hours won't be wasted! XP See you guys later. :)
15:48:29Torneevilnick: i need to finish my Frotz port at some point just because that will have an even more hilarious lack of buttons
15:49:19gevaertsyou're working on a frotz port? That should be fun!
15:50:19Torneyeah, i haven't done it for a while but i had it almost starting up
15:50:28Tornelots of stuff i could've done with has been added to pluginlib now
15:50:40Torneso i'll go back and redo some of it that way at some point
15:50:46 Quit _lifeless (Read error: 110 (Connection timed out))
15:50:56Tornei was then going to write an inform library for scrolling-menu-driven games
15:51:07Tornewhere you just get panes of text to scroll through and choices to pick
15:51:28Torneon the grounds that most z-machine games will be unplayable on an mp3 player ;)
15:51:44Tornebut getting any zcode to run would be a good start
15:52:18gevaertswell, there is the virtual keyboard :)
15:52:21 Quit arcticfang ("CGI:IRC")
15:52:25Tornethe virtual keyboard sucks. no offense. ;)
15:52:38Tornewell, it specifically sucks on the ipod because the 2d grid really doesn't map well onto the scrollwheel
15:52:45Tornei'm sure it's much less painful on players with a 4-way pad
15:53:24Tornei'd prefer a 1d keyboard for my ipod but i've not gotten around to looking at how much effort it would be to implement
15:54:04Tornebut yes i was going to start off by having the zmachine code for "input" pop up the keyboard
15:54:12Tornejust so i can run some trivial story files to test that the port works
15:54:40Torneit's not a real computer until it can run z-machine code, after all. that's the difference between computers and overblown calculators ;)
15:55:49gevaertstrue :)
15:56:15Tornei'm being lazy and only supporting targets with large plugin buffers for now, also
15:56:26Tornebut later it could probably be fixed to work on lowmem as well
15:56:37Tornestory files are specifically constructed to allow for virtual memory
15:56:56 Join {phoenix} [0] (
15:57:12Tornenot allowed to have mor than 64k of writable memory, so teh rest can be swapped :)
15:57:22 Part LinusN
16:05:21 Quit bubsy ("Mrrrrreow!")
16:07:03 Join froggyman [0] (n=187b533e@gateway/web/cgi-irc/
16:12:39 Join saratoga [0] (i=98039dd8@rockbox/developer/saratoga)
16:14:55 Quit biengo (Remote closed the connection)
16:21:44 Join toffe82 [0] (n=chatzill@
16:22:44 Quit n1s (Read error: 60 (Operation timed out))
16:31:21toffe82I try t build for the ipod video with the svn version of yesterday and I have a problem building rockbox, the bootloader build ok but rockbox stop with an error on setting_list.c ?
16:31:38Tornewhat svn revision?
16:32:29Torneand what target?
16:32:37toffe82the one of yesterday night, let me find out
16:33:09 Quit Rondom ("Ex-Chat")
16:33:17Torneoh, sorry, you said ipod video
16:33:45 Join robin0800 [0] (
16:33:50Torneok, that's very recent
16:34:00Torneand in fact i just build 21186 thirty seconds ago
16:34:01toffe82yes, yesterday night :)
16:34:14Tornewhich has no changes relevant to ipod video
16:34:28Torneso, i would expect it to work. what's the actual error? ( if it's more than one line)
16:35:44***Saving seen data "./dancer.seen"
16:36:00Tornealso, does svn status say any files are locally modified?
16:37:22 Join mt [0] (n=MTee@rockbox/developer/mt)
16:37:38Tornei would guess that genlang has failed?
16:37:42Torneor similar
16:37:48 Quit matsl (Read error: 110 (Connection timed out))
16:37:53Tornelook in lang/lang.h in your build directory and see if LANG_SORT_INTERPRET_NUMBERS is in there
16:38:34toffe82the ata.c is modified with MAX_PHYS_SECTOR_SIZE 4096 instead of 1024
16:38:44Torneis that all?
16:40:01toffe82I try the gigabeat S to see if I have the same problem
16:40:03Tornetry deleting lang/lang.h from the build directory and running make again
16:40:11Torneto make it regenerate the header
16:40:19Tornemaybe it's managed to be out of date without realising it somehow
16:40:31Tornealternatively just delete your entire build directory and start over
16:40:33Torneif you've not done that already
16:40:39toffe82doesn't make clean clean the directroy normally ?
16:41:01 Join bubsy [0] (n=Bubsy@
16:41:04Torneit should do, yes
16:41:19Torneare you using a seperate build directory, or building into the source directory?
16:41:52Torneanyway, have a look in lang.h and see if the missing defines are actually missing
16:41:53toffe82separate directory
16:41:53 Join at0m [0] (
16:42:13Torneif they are delete lang.h and run make again, and check to see if genlang prints any errors/warnings
16:43:04*toffe82 waiting for the gigabeat S build to finish , it takes time on an ubuntu virtual machine on an acer aspire one :)
16:43:26Tornei don't think the target is going to make a difference
16:45:43Torneanyway, a completely clean build from 21186 for ipodvideo works for me and i can't see any reason why those last two changesets would be relevant so i'm guessing you either have something that's not getting cleaned/rebuilt, if you have no local changes other than that ;)
16:46:12Tornei just rm -rf and recreate my build directory every time out of general paranoia as my computer builds qutie fast :)
16:48:44 Quit froggyman ("CGI:IRC (EOF)")
16:50:13CIA-38New commit by 03funman (r21187): Sansa AMS: Revert r21177 : the timer frequency is used by the user timer (example: metronome will be twice too fast) ...
16:51:02 Join froggyman [0] (n=187b533e@gateway/web/cgi-irc/
16:51:39r0b-can mpegplayer decode mpeg 4 video or just mpeg2
16:52:36linuxstbJust mpeg2 (and mpeg1)
16:55:57r0b-well i put a full length tv show encoded in mpeg2 and aside from a little a/v sync issues it works flawlessly
16:56:24r0b-i only asked about mpeg4 because mpeg2 is bloated :P
16:58:01toffe82Torne: no problem with the gigabeat S
16:58:15Tornetoffe82: that may just be because changing the target forced it to rebuild something though
17:01:05*toffe82 tries nnow the ipod video
17:05:05CIA-38New commit by 03funman (r21188): : fix revision in git-svn trees ...
17:13:45 Quit froggyman ("CGI:IRC")
17:14:24 Quit Zagor ("Don't panic")
17:16:49amiconntoffe82: I already mentioned more than once that increasing MAX_PHYS_SECTOR_SIZE is *not* the best solution
17:17:14toffe82amiconn: I know but do you have another one :)
17:17:18 Join mrgreen [0] (
17:17:29amiconnYes, commenting out MAX_PHYS_SCTOR_SIZE
17:17:43mrgreencan anyone help me with my cf pod?
17:17:48amiconnThe beast doesn't suffer from that problem because it doesn't have that enabled
17:18:23mrgreenreally really want to put rockbox on it
17:18:41amiconnThe macro is not needed because a disk has large physical sectors - if the drive has a proper firmware, it handles that internally
17:18:44toffe82amiconn: but in the disktest it was slower
17:19:17toffe82or something changed on that part from rev 19398 ?
17:19:19amiconnIt's just that the MK8011GAH has a buggy firmware
17:19:22mrgreentried to run ipodpatcher
17:19:32 Quit evilnick (" ajax IRC Client")
17:19:34mrgreenbut it states apple firmware missing
17:19:49toffe82this one is a samsung HS081HA
17:19:58gevaertsmrgreen: does the apple firmware work?
17:20:14mrgreenafter converting it was fscked
17:20:16gevaertsyou need to fix that first then
17:20:27mrgreenitunes does not find it
17:20:33mrgreenso tried to format
17:20:39amiconnThe ata driver should be improved to query the disk instead of just assuming it needs that workaround if it has large physical sectors
17:20:41mrgreenbut it does not showe up
17:20:53mrgreenyet I can use it under Floola
17:21:01mrgreentransfer files etc
17:21:07mrgreenbut as a player nothing
17:21:17mrgreenwhich is cool, 8gb pendirve
17:21:44Lloreanmrgreen: Have you tried the ipod manual restore instructions in the Rockbox wiki?
17:22:10gevaertsitunes should find it though
17:22:17mrgreenright lets see .... itunes nothing
17:22:27mrgreenwill try manual restore
17:22:46 Quit stoffel (Read error: 113 (No route to host))
17:23:15 Join evilnick [0] (i=0c140464@gateway/web/ajax/
17:23:24mrgreenwoah! will give that go thanks very much :)
17:23:47mrgreenLlorean: that looks like what I need
17:24:32mrgreengot a 20gb greyscale
17:24:39mrgreenso 4th gen I guess
17:31:17 Quit parafin ("So long and thanks for all the fish")
17:31:25 Join parafin [0] (
17:31:34mrgreencan anyone tell me firmware for 20g greyscale
17:31:50mrgreen10/ or 4/
17:32:48gevaertsI think both work. No idea why there are two firmwares for each
17:32:57gevaertsDon't forget the unzip step
17:33:46mrgreenthanks x
17:34:42 Join funman [0] (n=fun@rockbox/developer/funman)
17:37:42mrgreenI think its toast
17:37:51mrgreenI get folder error
17:37:58mrgreenit does not like cf card
17:38:02mrgreennot a problem
17:38:26mrgreengot a 39.5mb primary partition weird
17:39:03gevaertsthere is of course the possibility that your particular CF card isn't compatible
17:39:16mrgreenis that possble
17:39:58gevaertsyes. I don't know the details, but not all CF cards work with all players
17:40:05 Join chandoo [0] (
17:40:10toffe82Torne: building ok after deleting the lang.* files
17:40:30mrgreengot a second 20g coming so will test its drive on this unit
17:40:38Tornetoffe82: good to hear
17:40:46mrgreenmeantime can use it for storage
17:40:57mrgreenthanks again for all your help
17:41:11mrgreenreally trying to put life back into my ipod
17:41:27gevaertsAlso, apparently not all CF adapters work equally well
17:42:00 Join B4gder [0] (
17:42:08 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:51:53 Join barrywardell [0] (n=barry@
17:57:07 Quit mrgreen ("Lost terminal")
17:58:12 Join jmillikin [0] (
18:03:13 Join JdGordon| [0] (i=441af2d4@rockbox/developer/JdGordon)
18:05:43 Quit bubsy ("Mrrrrreow!")
18:07:16 Quit FOAD ("I'll be back")
18:07:24 Quit Rob2223 ()
18:10:10 Join FOAD [0] (
18:11:31 Join Rob2222 [0] (
18:12:39pondlifepixelma (or anyone who knows LaTeX): Could you look at my latest patch on #8894 and work out why it breaks the Archos manuals?
18:13:24 Quit Rob2222 (Client Quit)
18:14:33 Join _lifeless [0] (n=lifeless@
18:14:45 Join Rob2222 [0] (
18:20:46 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
18:21:19 Part pondlife
18:22:52 Quit timc (Read error: 60 (Operation timed out))
18:23:31 Join VytenisS [0] (
18:25:14 Quit _lifeless (Remote closed the connection)
18:30:33 Quit __lifeless (Read error: 113 (No route to host))
18:30:41 Join evilnick_6 [0] (i=0c140464@gateway/web/ajax/
18:31:15CIA-38New commit by 03funman (r21189): mkamsboot: refuse to patch an untested OF, and provide a list of tested OF versions
18:33:25 Join stoffel [0] (
18:34:34 Quit stoffel (Client Quit)
18:34:43 Join stoffel [0] (
18:35:46 Join Dhraakellian [0] (
18:35:49***Saving seen data "./dancer.seen"
18:35:58Dhraakellianwhat would be the best way to do ABX testing in Rockbox?
18:37:34Tornecomparing codecs?
18:37:54Tornewell, it depends exactly what thing you are trying to ABX
18:39:04Dhraakellianwell, I finally have a decent pair of earphones due to arrive tomorrow, and I'm guessing that even my e200 is likely to be better/less noisy than my onboard soundcard
18:39:14 Quit JdGordon| (" ajax IRC Client")
18:39:33Dhraakellian(waiting for Rockbox on the Fuze too, of course)
18:39:47Torneso you're interested in comparing the behaviour of different audio compressors, not the rockbox implementations?
18:39:56Torneeasiest way then is to just compress samples with different stuff
18:40:00Tornethen decompress them all to wav again
18:40:04Torneand just shove all the wav files on the player
18:40:06Dhraakellianbut I want to finally do some comparisons with Ogg Vorbis and see just how high I need the quality
18:40:10Tornehave someone else rename them all for you :)
18:40:30Torneit works
18:40:44Torneif you just make N copies of the same thing and compress/decompress the wavs are all the same size/etc
18:40:50Dhraakellianwondering if there's any sort of stuff in Rockbox that would make this easier
18:40:54Tornedon't think so
18:41:16Torneyou just need a little shell script that renames them to random numbers and writes out a textfile with the mapping back to the parameters
18:41:51Tornethen follow a standard ABX type procedure by hand
18:42:10Dhraakellianor put the encoder settings in a Vorbis Comment that wouldn't normally be read from the FLAC file
18:42:39Tornebest to properly randomise the filenames if you care that much ;)
18:42:45Torneproper blinding
18:43:02Tornebut yes, you could put the things in a comment
18:43:18Dhraakellianyeah, separating the storage of encoder settings from the filename
18:44:03Tornei'll writ eyou a lil script if you like to randomise the filenames ;)
18:44:23Torneanyway, offtopic a bit. i don't think there's anything in rockbox itself that can help you, no, but it shouldn't be needed.
18:53:31 Quit B4gder ("SSCCHKLABONG")
18:54:13 Join biengo [0] (
18:57:55 Quit biengo (Remote closed the connection)
18:58:27 Join biengo [0] (
19:04:23 Join matsl [0] (
19:08:31 Join hillshum [0] (n=quassel@unaffiliated/hillshum)
19:10:42amiconnUnhelpful: ping
19:12:12 Quit barrywardell (Remote closed the connection)
19:15:43funmanpyro_maniac: do you have problems connecting your YH920 to usb or is it only mine? (both with OF and with rockbox)
19:15:45 Join bertrik [0] (
19:16:16 Quit dmb (Remote closed the connection)
19:16:22pyro_maniacfunman: i had a problem yesterday, but i didn't know why
19:22:35 Join timc [0] (n=aoeu@
19:22:45pyro_maniacfunman: i can test later. and i still have the symbol bar on the buttom
19:23:38 Join JdGordon| [0] (i=836b0070@gateway/web/ajax/
19:23:42funmani'll try to see if i find something weird in lcd driver, i'm testing greylib
19:24:14 Quit r0b- (Read error: 110 (Connection timed out))
19:24:22 Join r0b- [0] (
19:24:28 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:24:57 Quit pyro_maniac ("Leaving.")
19:37:00 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
19:40:16 Part Dhraakellian ("Meddle not in the affairs of ircops, for they are (not so) subtle, and quick to anger.")
19:43:51 Quit domonoky (Read error: 104 (Connection reset by peer))
19:45:56 Join n1s [0] (n=n1s@rockbox/developer/n1s)
19:50:05 Quit Llorean (Read error: 54 (Connection reset by peer))
19:53:06 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
19:53:17 Nick hd is now known as HellDragon (i=jd@Wikipedia/HellDragon)
19:56:16funmanamiconn: gray ipod, mrobe100, and samsung yh920 use almost the same code for graylib.
19:56:55funmando you think we should unite them in one common file, or keep them separated should extra optimizations be made?
19:57:21*amiconn doesn't understand
19:57:57amiconnmrobe100 and greyscale ipods have different pixel formats, so their code differs
19:59:31amiconnThe lib itself has no asm code for arm, and the drawing code just differs per pixel format
19:59:45funmanwell then the yh920 & ipod have the same code
20:00:18funmanexcept some device-specific parts (IPOD_MINI2G / IPOD_MINI)
20:00:30*amiconn wonders what part funman is talking about
20:02:03funmanlcd_write_data_shifted; lcd_write_data is different for IPOD_MINI2G
20:03:39funmansee : the diff I applied to ipod/lcd-as-gray.S to make it work on YH920
20:04:04amiconnAh, you're talking about the blit functions in the drivers... not the greylib itself
20:04:06funmani.e. it would work unmodified
20:04:21funmanah yes, sorry for the confusion
20:05:55funmanSo I suggest arm/ipod/lcd-as-gray.c be moved to arm/lcd-as-pp-gray.S
20:06:33amiconnWell, the iPod Mini G2 is special because the LCD is hooked up using the serial interface, which is written 16 bit at a time from the cpu
20:06:44amiconnThe other greyscale ipod use the 8 bit parallel interface
20:07:11 Join mirak [0] (
20:07:19amiconnThe mrobe100 is yet different because of its inverted display
20:08:29amiconnBoth minis are also different from anything else in that their lcd doesn't match byte boundaries in the frame buffer, hence flipped display needs bitshifting
20:08:45amiconn*their lcd *size*
20:09:44funmanthe bitshiftign function could stay in ipod/ in another file
20:11:08amiconnThe question is whether it's worth unifying the driver
20:11:13Mikachufunman: i'm not sure if it works better or worse, but you can just do git svn find-rev refs/remotes/git-svn
20:12:00 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
20:12:07amiconnThe ipods need target tree cleanup anyway... they're not adjusted to the modified scheme for arm/
20:12:32amiconnThere's another possibility: Unify the driver for most greyscale PP targets, but keep the minis separate
20:12:47funmanMikachu: looks simpler yes ;)
20:13:02Mikachufunman: if you check out an older version it will still print the latest though, so maybe it's not so good :)
20:13:28 Quit pondlife (Client Quit)
20:13:36amiconnI'd also keep the mrobe100 separate, so there'll still be 3 assembler parts: ipod G1..G4 + yh920, mino G1..G2, and mrobe100
20:13:52amiconnThat's still 3 drivers, but with less ifdefing than what we have now
20:13:57 Join Thundercloud [0] (
20:14:19Mikachugit svn find-rev `git log −−pretty=format:'%H' −−grep=git-svn-id: -1` maybe?
20:15:13Mikachufunman: it's slower than your version though :P
20:20:31funmanamiconn: ok
20:22:01funmanamiconn: now, wrt FS #10274, is the timer frequency on sh7034 equal to cpu frequency ?
20:22:32*bertrik is tempted to commit the simple (non-interrupt) DBOP FIFO patch for clip
20:24:03amiconnfunman: Yes, and timer.h *does* define it
20:24:31 Quit flydutch ("/* empty */")
20:27:11amiconnRegarding the int_prio argument - I'd happily remove it, as almost all places just set it to the same, low value
20:27:42amiconnUnfortunately the only plugin that actually sets differet priorities is alpine_cdc, and that one I can't test
20:27:51 Join hd [0] (
20:28:03amiconnIt was written by [IDC]Dragon, who is a rare guest these days :\
20:29:09JdGordon|we could probably remove it and see if anyone complains
20:29:37BagderI would bet pretty much on that nobody will
20:30:03 Quit stoffel ("leaving")
20:30:15 Quit biengo (Remote closed the connection)
20:33:41 Quit HellDragon (Success)
20:34:08 Join dmb [0] (n=dmb@unaffiliated/dmb)
20:34:11AlexPSo, if we aren't freezing today, what is the release timeline?
20:35:50***Saving seen data "./dancer.seen"
20:40:36JdGordon|there is an email with dates in the dev ml
20:40:51*JdGordon| thught he had a few more days before the freeze
20:45:06 Join moos [0] (i=mustapha@rockbox/staff/moos)
20:45:34 Join pyro_maniac [0] (
20:45:38 Join petur [0] (
20:48:16pyro_maniacfunman: i got a problem on connecting usb too. i had to change the port.
20:50:44AlexPJdGordon: And going by those emails it is today, hence why I am asking
20:52:56 Join tessarakt [0] (
20:54:00funmanpyro_maniac: same here ..
20:57:12JdGordon|can we freeze on sunday instead? or is 10094 OK to commit during the freeze period?
20:57:38AlexPJdGordon|: Don't ask me - this is the second time I've asked today...
20:57:51JdGordon|I was asking generally :)
20:57:55gevaertssunday would be great. It would mean I don't have to feel guilty about not having worked on FS #8802 yes
20:58:23JdGordon|the previous ones just started wwith a random email saying its started didnt it?
20:58:35bertrikand sansa ams fixes/updates can still be committed, right?
20:58:36pyro_maniacfunman: any idea to reproduce?
20:59:04gevaertsbertrik: only if they're not part of the release ;)
20:59:22funmanpyro_maniac: how to reproduce the usb problem?
20:59:23hillshumJdGordon: I think they were started by a random email from Bagder
20:59:52 Quit Thundercloud (Remote closed the connection)
21:00:36pyro_maniacfunman: yes. i connected and disconneted serveral times now but the problem didn't came back.
21:01:32 Join Llorean [0] (
21:04:24gevaertsJdGordon|: I wouldn't think FS #10094 is really important enough to warrant a freeze-break. I'm in favour of delaying the freeze until sunday though
21:05:13*JdGordon| looks at the calendar...
21:05:19AlexPSunday is fine IMO, I just think we ought to agree on one, and have it so that we release before devcon
21:05:23JdGordon|sunday means the whole effort of bringing it forward was wasted :p
21:05:42pyro_maniacfunman: no i got it i think.
21:06:12AlexPThe freeze could be two days shorter
21:06:33AlexPReleasing on the friday before devcon is a bit of a nodo as Bagder and Zagor will be travelling
21:06:45funmanpyro_maniac: the same for me, when i changed the port it was functioning OK for 10 plug/unplug, then started doing bad. Perhaps these devices are just crappy
21:06:59JdGordon|so maybe bugger it.. freeze today and stay on the plan
21:07:12*funman looks at what is needed for making Sansa AMS releasable before sunday
21:07:23JdGordon|gevaerts: not a freeze break.. ill just say thats all bug fixes and do it anyway :)
21:07:29pyro_maniacfunman: did you also get this:
21:08:03AlexPOr freeze sun, unfreeze next sat, release thurs?
21:08:08AlexPOr just freeze today :)
21:08:18funmanpyro_maniac: hum no, i believe this happens when the YH920 is detected, starts to be mounted, and is disconnected while it'sn ot mounted yet
21:08:31*JdGordon| doesnt see why there is any gap between unfreeze and release?
21:08:41JdGordon|ping Bagder
21:09:20hillshumJdGordon: It's a week with a release branch
21:11:23gevaertsAlexP: Bagder could release from my car :)
21:11:41hillshumHe does have 3g now
21:11:49AlexPgevaerts: Only if you are sure not to crash it again :)
21:12:09JdGordon|are we still talking about rockbox?
21:12:21gevaertsbertrik: you did a lot on the sansa charging, right? Am I right in thinking that it always draws full power over USB?
21:12:54 Join jspeter_ [0] (
21:13:17bertrikgevaerts, actually I made a good start with it but then stalled, IIRC jhmikes properly finished it
21:13:54jspeter_Does anyone know how to get rockbox to resume a playlist on startup?
21:14:25bertrikas far as I remember it does not take into account any USB state to determine charge current, but I'll have to look that up inthe source
21:14:56JdGordon|jspeter_: set the start screen option
21:15:08jspeter_k, thx
21:15:10AlexPjspeter_: Either press whatever the resume key is (see the mnanual) or set the start screen to resume (see the manual)
21:15:39 Quit timc (Read error: 60 (Operation timed out))
21:18:07gevaertsI think we should just commit FS #8802 as-is. We have lots of targets that don't check for power allowance when connected, we can't even fix that on half of them, and I don't see us getting the logic perfect fast enough
21:18:43 Quit jspeter_ (Remote closed the connection)
21:20:34gevaertsWhat do we do with though?
21:24:01 Join Blackfrosch [0] (
21:24:08 Part Blackfrosch
21:28:03Lloreangevaerts: Considering you've been able to charge iPods by holding Menu and plugging in USB for years at this point, wouldn't we have seen the kinds of problems you'd get from uncontrolled charging ages ago otherwise?
21:30:10gevaertsLlorean: ipods never really charged properly as far as I can see. I'm not worried too much about USB strictness because we've been breaking that for ages on other targets and a few more won't change much, but I'm a bit nervous about dreamlayers' doubts there
21:30:51gevaertsI'd like someone who knows ipods well to comment on that
21:31:14LloreanThey never charged properly, but they did charge. Isn't the worry about overcharging?
21:32:03gevaertshm, possibly. Let's find someone who understands charging as well :)
21:33:41LloreanWe could always just enable it, and when batteries start exploding remind people of that "without warranty" bit ;)
21:33:59JdGordon|Llorean: I would guess most people wouldnt even know about that option.. and even if they did, most people would just connect it up to the computer and leave it to charge
21:34:48JdGordon|hold menu to charge
21:34:59LloreanA lot of people use that, so they can listen to music while charging.
21:35:12LloreanAs well, for wall/car chargers, you don't need to hold menu, and people have been using those the whole time as well
21:36:05*JdGordon| grumbles at headphone detection not working when the device turns itself off :p
21:37:08 Join crashd [0] (
21:37:35 Quit crashd (Client Quit)
21:37:41 Join crashd [0] (
21:46:12 Join froggyman [0] (n=47ba06b6@gateway/web/cgi-irc/
21:50:30 Quit wark ("Leaving... for now. But I'll be back.")
21:53:43CIA-38New commit by 03bertrik (r21190): Make more effective use of the DBOP FIFO for the sansa clip.
21:58:53 Quit LambdaCalculus37 (" ajax IRC Client")
22:06:44 Join Rondom [0] (
22:10:51 Quit perrikwp (" ajax IRC Client")
22:11:49 Join ecrips [0] (
22:12:25 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
22:12:43toffe82strange the build with the MAX_PHYS_SECTOR_SIZE 4096 doens't work anymore on my ipod, it show the error invalid sector size 4096 ??
22:14:56 Quit perrikwp (Client Quit)
22:15:08 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
22:15:20 Join Thundercloud [0] (
22:16:53 Quit antil33t (Read error: 60 (Operation timed out))
22:18:03 Join kachna [0] (
22:20:18 Part pyro_maniac ("Leaving.")
22:21:54 Quit {phoenix} (Remote closed the connection)
22:24:21ecripsfunman: are you around?
22:24:21Bagderregarding release, we should have plenty of time to correct mistakes from devcon
22:26:03ecripsI've had a further attempt at getting the button light on on my Fuze v2 - but outputting 0 on GPIO doesn't work either :(
22:26:14ecripsGPIOA_PIN4 seems a deadend too
22:26:14 Join faemir [0] (
22:26:15JdGordon|Bagder: the plan was to release before devcon... (wasnt it?) why cant we release staright after the freeze which would be just under a week before devocn?
22:27:01LloreanJdGordon|: If the branch was unnecessary, why do we have it in the first place?
22:27:35JdGordon|thats my point...
22:27:43*JdGordon| was never a fan of these quick releases...
22:27:48Bagdersure we could, it would just mean slightly less time to fix bugs
22:28:07AlexPBagder: I thought the plan was to release before devcon, so that lots of guilt free hacking could go on there :)
22:28:08LloreanJdGordon|: Quick?
22:28:21AlexPEither that, or delay the release until a while after devcon
22:28:42 Join chantzguy [0] (n=60f083f8@gateway/web/cgi-irc/
22:28:42chantzguyanyone here?
22:28:59AlexPchantzguy: Yes, just ask
22:29:03BagderI don't consider us going away much of a problem, as we should be even more online than usual during that travel
22:29:44AlexPI personally think it's be nice to get the release done before devcon, friday is fine too :)
22:29:53Bagderquick in what way?
22:30:11chantzguycould someone reply to my post in the hardware section about the sansa?
22:30:39JdGordon|every 3 months
22:30:39chantzguyi really want to get rockbox on this
22:30:39chantzguyand nobody is posting
22:30:44AlexPchantzguy: Just have patience
22:31:14AlexPchantzguy: If anyone that knows the answer sees it, then they will answer
22:31:20AlexPchantzguy: Alternativly, ask here
22:31:22chantzguyand i just did
22:31:22BagderAlexP: right, and there's really not a lot of panic planned post-release...
22:32:00 Quit faemir (Read error: 54 (Connection reset by peer))
22:32:10AlexPBagder: Sure, it isn't a major issue - it would be useful to have some agreed dates though, as those on the mailing list don't appear to be :)
22:32:15BagderI'm a fan of the every 3 months approach
22:32:23LloreanJdGordon|: I don't see what frequent releases has to do with the problem we're coming up against right now, though.
22:32:28AlexPchantzguy: Ask the question here
22:32:36evilnick_6chantzguy: AlexP meant that you could ask your question here, not that you should ask people to look at the forums without providing a link
22:32:38LloreanAlexP: I think we can just make the dates on the ML official
22:32:47 Quit chantzguy (Client Quit)
22:32:50AlexPLlorean: So we freeze today then
22:32:55LloreanAll someone has to do is change the channel topic in here to include that it's started, and it's started.
22:33:01 Join chantzguy [0] (n=60f083f8@gateway/web/cgi-irc/
22:33:24chantzguyi cant get into recovery mode on my e200R sansa!!!
22:34:09Lloreanchantzguy: that really didn't need three lines. Are you following the instructions here:
22:34:29saratogachantzguy: what firmware version does your player have?
22:34:33AlexPLlorean: Best poke Bagder then :)
22:34:40*Llorean pokes Bagder
22:34:47*Bagder feels poked
22:34:54 Quit petur (Remote closed the connection)
22:35:01*Bagder declares freeze
22:35:04chantzguymy bad, and yes i am
22:35:05*Llorean applauds.
22:35:12Bagderand darnit, it has been freezing here all day!
22:35:25Lloreanchantzguy: And saratoga's question?
22:35:44Mode"#rockbox +o Bagder " by ChanServ (ChanServ@services.)
22:35:47*JdGordon| sticks his hands in the air and stops running
22:35:47chantzguyoh, lol i didnt see it...lemme check
22:35:54***Saving seen data "./dancer.seen"
22:36:29Topic"Feature Freeze period is now official! | Please read before speaking: | Please direct offtopic/social chat to #rockbox-community" by Bagder (n=daniel@rockbox/developer/bagder)
22:36:50Bagderthat's not an e200R then
22:37:05chantzguybut it has "rhapsody channels"
22:37:08saratogaman am I good at this
22:37:12saratogayou have an e200v2
22:37:20Lloreanchantzguy: Any firmware version starting with V03 means the player is an e200v2
22:37:30LloreanIt has completely different hardware inside.
22:37:33chantzguylemme guess, rockbox wont work with the v2?
22:37:42LloreanNot yet, it's in-development
22:37:58chantzguyfantastic...but good to see you guys are trying
22:38:02chantzguyif i knew programming i'd help
22:38:14Bagderthere are more ways to help...
22:39:11 Join lymeca [0] (
22:39:21JdGordon|but that article is surely out of date :D
22:39:35chantzguyand the Software on this is slow as anything. and i click on album>Tribute to blink-182 heroes of pop/punk>"song title" then when i want the next song i click the "next" button and it replays the song?!?!?
22:40:15Lloreanchantzguy: We really don't care about the original firmware in here. That's off-topic, and we ask that you keep the channel on-topic.
22:40:45 Quit lymeca (SendQ exceeded)
22:40:50 Join lymeca [0] (
22:41:19FrankTM_22:39 <@Bagder>
22:41:43chantzguyill help out in the forums i guess
22:42:06 Quit lymeca (SendQ exceeded)
22:42:09chantzguyim taking cpu math nxt year lol hopefully that will teachme something to help wiht coding?
22:42:49 Quit chantzguy ("CGI:IRC")
22:42:53HorschtUnhelpful, just giving you a headsup: the jpeg albumart issue is fixed now :)
22:43:06 Join lymeca [0] (
22:43:22Horschtsorry, "thank you"
22:43:32funmanecrips: hi
22:43:33 Quit lymeca (SendQ exceeded)
22:44:04ecripsfunman: Hi. I'm a bit lost as to what to try next, any suggestions?
22:44:10 Join lymeca [0] (
22:44:40 Quit lymeca (SendQ exceeded)
22:44:42funmandid you test power_off() function? it should prove that the i2c is working
22:44:53ecripsno I haven't yet
22:45:06 Join BryanJacobs [0] (i=80974574@gateway/web/ajax/
22:45:22funmani think that on sansa fuzev1/e200v2 you have to enable DCDC15 by i2c to power the button light (and lcd)
22:47:02ecripsdo I need to call ascodec_init before using power_off?
22:47:30funmanit should be done by system_init()
22:47:51ecripscurrently I've just been hacking dualboot.S
22:47:55 Join flydutch [0] (
22:47:57 Quit mt (Read error: 110 (Connection timed out))
22:48:13funmanhum now that you have a button for dualboot you should be hacking bootloader/sansa_as3525.c
22:48:30 Quit VytenisS ("Quit")
22:48:32bertrikhacking dualboot.S can be very dangerous w.r.t. bricking
22:48:43funmanwell you can even skip system_init (specific to as3525), and call ascodec_init() yourself
22:49:09funmandon't forget to put a while(1); to not execute sd/lcd code, or a power_off() if that works
22:49:30ecripswell it was when I tried writing my own little firmware to play with the button light that I semi-bricked it so I've found dualboot safer so far :)
22:49:50funmanif you have a working dualboot, don't touch it anymore
22:50:15funmanit's easy to make a mistake - RIP, my 2 Clips -
22:51:55ecripsthe benefit of dualboot hacking is I can run the OF afterwards, but now I know having a flat battery is a recovery mechanism I suppose I won't have lost too much
22:51:56 Join timc [0] (n=aoeu@
22:52:51funmanecrips: you can test if hardware power off works this way:
22:52:57hillshumonly to recover from that anyway
22:53:21funmanwrite a long (1 minute) loop before booting OF
22:53:58funmanrun, hold power switch for 20 seconds, wait another 40seconds, see if OF boots or if the player is down - or plug it on USB to see if it boots immediately -
22:54:24funmanhaving to wait for the battery to be empty is not exactly friendly
22:55:28funmanecrips: perhaps you touched a pin you were not supposed to, and that's what disabled the hardware power off
22:55:28ecripsok I'll have a go
22:56:12ecripsfunman: I suspected that - but it seemed a bit of a coincidence
22:56:18funmantry that in sansa_as3525.c: kernel_init(); ascodec_init(); ascodec_write(AS3514_DCDC15, 0x10); GPIOA_DIR |= (1<<4); GPIOA_PIN(4) = (1<<4); delay=?; while(delay−−); power_off(); while(1) ;
22:56:52funmanpin A4 is modified in Fuzev2 OF nearly at the same place than D7 (led) in the Fuzev1 OF
23:00:13 Quit mirak (Read error: 104 (Connection reset by peer))
23:03:18 Quit bmbl ("Woah!")
23:04:57ecripsfunman: ok, well I've got to go now, but I'll try that code in sansa_as3525.c
23:05:22ecripsI've confirmed that you can (normally) use the 20sec hold to turn off, just obviously not when you've poked the wrong GPIO register :)
23:05:31funmanahah, nice to hear :)
23:05:33 Join mirak [0] (
23:05:53ecripsthanks for your help
23:06:06 Quit ecrips ("Ex-Chat")
23:06:06funmantry to set DCDC15 and to get light working, if it's still not lighted i'll try to see if something extra is needed
23:07:20 Quit tessarakt ("Client exiting")
23:08:08 Quit CaptainKewl (Read error: 110 (Connection timed out))
23:12:12 Quit evilnick_6 (" ajax IRC Client")
23:12:38 Nick fxb__ is now known as fxb (
23:14:14 Join faemir [0] (
23:21:49 Join wincent [0] (
23:22:23 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
23:23:12pixelma would it be better if the arrows on the scrollwheel would point in the other direction? (I think that would be a bit more consistent with the other scrollwheel target drawings too)
23:26:17 Quit faemir (Read error: 104 (Connection reset by peer))
23:27:19bluebrotherpixelma: I agree. Also, at least the e200 has the arrows in the upper part of the ring, not the right
23:28:34pixelmaneeds a little rearranging of all the other button labels... other than that, I'll commit it soon
23:30:50 Quit killan ("( :: NoNameScript 4.22 :: )")
23:31:20 Join safetydan [0] (n=deverton@rockbox/developer/safetydan)
23:32:49 Quit moos ("Rockbox rules the DAP world")
23:33:14 Join killan [0] (
23:36:20 Part BryanJacobs
23:37:28 Part wincent ("Kopete 0.12.7 :")
23:37:59pixelmajust noticed that the e200 drawing still says "scroll up/down" (and so does the Fuze one now) because the code did before, later it was changed to "scroll ffwd/back" like on the other scrollwheel targets but the drawing was not. Some more to do...
23:38:10 Join CaptainKewl [0] (
23:50:33 Quit BlakeJohnson86 (Remote closed the connection)
23:52:43 Quit flydutch ("/* empty */")

Previous day | Next day