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

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

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

#rockbox log for 2010-08-31

00:00:03Buschelkugel: from my understanding it does not fully reset and refill the buffer, but only loads 1 chunk. but the data is not buffered anymore.
00:00:32LearNo, rebuffer_handle is only called if it isn't buffered yet, but it should be soon. And the queue_send is to make sure it is.
00:02:02kugelBuschel: you describe that it needs to seek from the end to the beginning; it seems the if() is wrong if that case goes through it
00:03:02LearBut it looks like ringbuf_add should be used there, and find_handle should be called again, since queue_send do sleep.
00:03:04kugelshouldn't newpos be smaller than h->offset here?
00:03:54Buschelkugel: The codec seeks to somewhere near the file-end. Then it seeks back to somewhere near file-begin using reset/buffer. The very first rebuffer attempt afterwards enters the if()-statement −− and that's when the codec exits with gargabe data.
00:04:12 Quit pamaury (Remote host closed the connection)
00:04:52*S_a_i_n_t doesn't understand:
00:05:38 Quit jgarvey (Quit: Leaving)
00:06:55LearBuschel: I'm not sure if it is the seek back to the beginning that is the problem; in this case it is to well beyond the first chunk.
00:09:03BuschelLear: Well, the error occurs with the first call of rebuffer_handle() after the seek back to the beginning.
00:09:36BuschelLear: The seek back is not entering the if()-statement within rebuffer_handle(). So, I assume the seek back is fine.
00:10:09LearSo it is the first seek after the rebuffer_handle back to the beginning, so to speak?
00:10:09BuschelLear: I assume the code within the if()-statement is somehow faulty
00:10:30kugelrebuffer_handle is only called from bufseek(), why should there be another bufseek() after the one that goes to the beginning?
00:10:41BuschelLear: exactly
00:11:52 Quit Zarggg (Quit: Zarggg)
00:11:58Buschelkugel: you should really try to use the problematic file. the seek is not seeking the file-beginning, but to a special position for a sub-header within the bitstream. it is reasonable that another seek is happening afterwrda when starting to decode the file
00:12:36kugelI have no idea about audio formats, listening to that file doesn't help me
00:12:37*Buschel gets another (and the last!) beer
00:13:04Buschelkugel: no, not listening. it does not play without my hotfix
00:13:34Buschelkugel: at least on iPod Video and sansa e200 (w/o dircache)
00:14:15*preglow sips his gin&tonic
00:14:29preglowpretending it still feels like summer
00:14:48Buschelpreglow: it is pretty damn cold here in southwest germany :/
00:15:27preglowisn't exactly stellar here either, rain and grey skies
00:15:40 Join h3po [0] (
00:16:55LearThere's a seek to locate the first audio frame, and that is within the first chunk after offset (which now is at mdat start).
00:18:54 Quit h3po1 (Ping timeout: 276 seconds)
00:18:56kugelLear: ringbuf_add isn't appropriate here (IIUC) because h->ridx, h->offset and newpos aren't representing the position in the buffer, but the position in the file
00:19:15kugelbut you might be right about calling find_handle() again
00:19:17BuschelLear: When I printf-debugged I saw the the problematic rebuffering was used to seek +8 byte (relative to mdat start)
00:19:52LearBuschel: Makes sense, yes.
00:23:17Learkugel: h->ridx and h->data is a buffer index. Newpos is an absolute file position, so that math looks really bad to me.
00:25:04 Join krazykit [0] (~kkit@
00:25:05kugelyes indeed
00:25:07LearBuschel: Maybe a proper fix would be something like "h->ridx = ringbuf_add(h->data, newpos - h->offset);"?
00:26:13kugelthat could be worth a try
00:27:20Buschellooks good :)
00:29:25Buschelseems to solve the issue as well
00:29:31 Join krabador [0] (
00:29:59LearThen there's all the find_handle calls that probably need to be added. :)
00:30:11LearBut I I _really_ need to go to be now...
00:30:42BuschelLear: I will submit yuir proposal to fix the issue.
00:31:07 Quit Lear (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
00:31:33kugelanother find_handle within that if() shouldn't be needed, buffer_handle() doesn't change it
00:32:20funmani'm trying to make voice work on clipv1 without success so far
00:34:42clone4crwIs anyone interested at looking at a new plugin?
00:34:47 Join Zarggg [0] (
00:35:34CIA-81New commit by Buschel (r27950): Fix FS #11586. Corrects rebuffering behaviour which did not allow to play several m4a files. Thanks to Magnus Holmgren.
00:37:05 Join kaiscene [0] (
00:37:22CIA-81r27950 build result: All green
00:38:18 Quit bmbl (Quit: Bye!)
00:38:20kugelclone4crw: #define RES_SILVER_MUL 1/100 isn't going to work as expected
00:38:46funmanah clips are queued and played asynchronously
00:39:02kugelalso remove PLUGIN_HEADER
00:39:11kugelbtw, why does your plugin not work in the sim?
00:40:26clone4crwremove plugin_header? ok
00:41:10clone4crwI really don't know why it doesn't work in the sim. i start rockboxui from the command line, and after a while it reports a bunch of SDL errors
00:41:18kugelwhy are all your functions indented?
00:41:35clone4crwshould they not be?
00:41:43kugelrun it with gdb, you've probably make it crash somewhere
00:42:28funmanhm .. voice works but not playback
00:42:34funman(with current build)
00:42:38kugelthe functions itself not, but what's in the functions (i.e. |int foo()\n|{\n|−−->int i;)
00:42:49funman-> audiobuf size is - (minus) 198K
00:43:42 Part froggyman
00:43:54clone4crwi'm recompiling an e200 sim right now - i'll try gdb
00:43:57 Quit shai (Ping timeout: 245 seconds)
00:44:56 Quit evilnick (Quit: Page closed)
00:46:10 Join shai [0] (
00:46:15clone4crwkugel: anything else?
00:47:16 Quit {phoenix} (Remote host closed the connection)
00:48:19 Join froggyman [0] (~4845cf9e@unaffiliated/froggyman)
00:48:20 Quit froggyman (Client Quit)
00:51:37kugelclone4crw: well, it's not quite finished yet
00:52:49 Quit h3po (Read error: Connection reset by peer)
00:53:15 Quit bluebrother (Disconnected by services)
00:53:19 Join bluebroth3r [0] (~dom@rockbox/developer/bluebrother)
00:54:08clone4crwYeah, I know. I really wanted to put it out there after sitting on it for so long
00:54:50***Saving seen data "./dancer.seen"
00:54:59 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
00:57:00S_a_i_n_tclone4crw: I was actually looking at that plugin right now ;)
00:57:05CIA-81New commit by funman (r27951): audio_reset_buffer: check audiobuffer size not only in DEBUG builds ...
00:57:09S_a_i_n_tI cleaned up one of your bitmaps for you.
00:57:14S_a_i_n_tJust about to post it.
00:57:46clone4crwSounds great.
00:57:52gevaertsclone4crw: the biggest issue I see is that you're trying to store fractions in ints
00:58:11clone4crwyes, i'm still a bit unsure of how to do some things in C
00:58:17 Join Guest50782 [0] (
00:59:47CIA-81r27951 build result: 0 errors, 71 warnings (funman committed)
01:00:25 Quit Jerom (Read error: Connection reset by peer)
01:00:34 Nick Guest50782 is now known as froggyman (
01:00:46 Quit froggyman (Changing host)
01:00:46 Join froggyman [0] (~seth@unaffiliated/froggyman)
01:00:49kugeldo we really need such a useless check?
01:01:01funmanwhy useless?
01:01:27 Join saratoga [0] (9803c22e@gateway/web/freenode/ip.
01:01:44funmanadding support for %zd/%zu in format.c costs 124 bytes on clipv1
01:02:01saratogawe really need some kind of unit testing for the playback engine
01:02:37CIA-81New commit by funman (r27952): panicf() is declared in panic.h
01:03:17gevaertsclone4crw: also, you can't just declare a string constant and then write to that. It might work with some runtimes and compilers, but it *will* break eventually
01:04:41CIA-81r27952 build result: All green
01:04:59gevaertsI wouldn't be surprised if that's where your sim segfaults come from
01:08:01kugelfunman: btw, I meant to ask you, why did you add lcd_putsxyf to the plugin api instead of to the multi screen api (you asked to remove lcd_putsf from the plugin api minutes before)
01:08:11CIA-81New commit by funman (r27953): format: supports %zd (ssize_t) and %zu (ssize_t)
01:08:14CIA-81New commit by funman (r27954): don't cast to int when not needed
01:09:09funmankugel: i don't know, what's different between lcd_putsf() and screens->putsf() ?
01:09:23kugelno difference
01:09:52funmanmostly i didn't know about the screens API before starting this work
01:09:58gevaertsclone4crw: Another thing, I'm pretty sure you don't mean "(BUTTON_UP || BUTTON_SCROLL_BACK)
01:10:02CIA-81r27953 build result: All green
01:10:12kugelyou can loop over rb->screens[] instead of calling lcd_putsf and lcd_remote_putsf separately
01:10:28 Join dfkt [0] (dfkt@unaffiliated/dfkt)
01:10:39kugelwell, it's just a set of variables and function pointers for lcd_[remote_]* functions
01:11:06funmanthere's also a similar api for greylib
01:11:40CIA-81r27954 build result: All green
01:12:19S_a_i_n_tclone4crw: The new image is in a comment I just added to your patch on the tracker.
01:12:54S_a_i_n_tWhen I get more time, I might clean up the other one too...I just did that one because it was the image appropriate for my screen size :D
01:13:15kugelfunman: did va_arg(ap, ssize_t) not work=
01:13:50funmanoops typo, thanks
01:14:36clone4crwgevaerts: yep. i got that mixed up. i'll fix the indentations, too.
01:14:55clone4crweven though most of those button declarations aren't being used yet
01:15:02CIA-81New commit by funman (r27955): typo in r27953 spotted by kugel: s/long/ssize_t/
01:16:44CIA-81r27955 build result: All green
01:18:36funmankugel: the idea was to use the same case than long but for some platforms sizeof(long) != sizeof(size_t)
01:19:09kugelyea, I understand that (IIRC gevaerts tried the reuse-long-case approach alredy)
01:19:15clone4crwS_a_i_n_t: Think you'll add a keymap for the Nano?
01:19:34kugelbut ssize_t might also not work on all platforms (it's a unix type)
01:20:21 Quit krabador (Quit: Sto andando via)
01:20:59funmanwe could remove it if it cause problems
01:22:12S_a_i_n_tclone4crw: Not presently...but, in time, yeah. Unless you beat me to it.
01:22:37kugelor calculate it from size_t
01:23:19S_a_i_n_tadding a keymap for the Nano kills multiple birds with one stone....pretty sure all iPods use the same keymap.
01:23:24S_a_i_n_tclone4crw: ^
01:23:46kugelclone4crw: how many keys do you need? consider pluginlib_actions if you only need 5 or so
01:23:48clone4crwgood point
01:24:09clone4crwgaverts: Yeah, stupid me. fixed the constant char thing, the sims now work
01:24:40clone4crwThe only reason I defined so many was for later when we could have a calculator-like keypad for input
01:24:42kugelwith pluginlib_actions you kill *all* birds at once
01:24:57S_a_i_n_tkugel: Good point ;)
01:27:51kugelclone4crw: "char * the_resistance_is = "The resistance is "" doesn't work, while "char the_resistance_is[] = "The resistance is "" does. also you don't need to reserve 35 chars for the result, you can simply use snprintf()
01:28:43kugelI mean, you need to reserve some space, but it doesn't need to be an array of spaces
01:28:56S_a_i_n_tclone4crw: Give me a yell if you need more images done, I'll probably re-do & post the 220x176x16 later on tonight...if you need other images for other targets, just let me know the dimensions.
01:29:10clone4crwis char total_resistance_str [35] ok?
01:29:13S_a_i_n_tPS: I'm terrible with greyscale/monochrome images ;)
01:29:25kugelbuf[35]; snprintf(buf, sizeof(buf), "The resistence is %d", resistence);
01:29:40S_a_i_n_tBut...this plugin won't really work on non-colour targets, so, I suppose that isn't a problem.
01:30:01clone4crwYeah, I was thinking in that case it would just be textual output
01:35:09S_a_i_n_twell, when you know what sizes you'll need...gimme a yell.
01:35:26S_a_i_n_tI'll re-do the 220x176x16 later on as I said earlier.
01:35:49S_a_i_n_t(not that I think you didn't do a good job...however)
01:36:15clone4crwI had a worse one to start out with, then whipped up the current one in GIMP
01:37:38clone4crwI almost want the bands themselves to be BMPs with a slight curve to them, giving the whole thing a more 3D effect, but I'll focus on getting the whole thing finished first :p
01:39:12 Join bunnyboi [0] (
01:39:33 Quit bunnyboi (Read error: Connection reset by peer)
01:44:59funmani have it semi-working now
01:49:50 Quit lostlogic (Quit: leaving)
01:51:24 Quit MagusG (Read error: Connection reset by peer)
01:51:35 Join Transformer [0] (
01:52:00 Join MagusG [0] (
01:53:26 Quit Transformer (Excess Flood)
01:55:36clone4crwI just put a fixed version on the tracker
01:56:10funmanit seems to work fine but spelling fails after 10 letters or so
02:07:06funmanFS #11587 - Voice for SWCODEC and low memory
02:10:13 Quit funman (Quit: free(random());)
02:11:08 Quit lestatar (Ping timeout: 240 seconds)
02:12:56 Join lestatar [0] (
02:15:03 Quit dfkt (Read error: Connection reset by peer)
02:15:04 Join dfkt_ [0] (dfkt@unaffiliated/dfkt)
02:20:37 Quit lestatar (Ping timeout: 258 seconds)
02:21:30 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
02:21:43 Join fdinel [0] (
02:21:52 Part JdGordon|
02:29:32 Join captainkewllllll [0] (
02:30:09 Quit captainkewllllll (Client Quit)
02:33:11clone4crwwhere does it fail?
02:33:50 Quit MagusG (Read error: Connection reset by peer)
02:34:39 Join MagusG [0] (
02:35:03 Quit kugel (Remote host closed the connection)
02:48:24 Quit komputes (Remote host closed the connection)
02:54:52***Saving seen data "./dancer.seen"
03:04:39 Join G4Oblivion [0] (
03:06:54G4OblivionI've done testing and found that r27020 is what has been causing the "*PANIC* stkov nand" errors on nano 2g.
03:07:31 Join madalu [0] (~matt@unaffiliated/madalu)
03:08:32G4OblivionI only have one nano I can test this on so if anyone would like to test this I have uploaded r27019 and r27020 here:
03:08:53G4OblivionUsers that would also like to revert back to r27019 can use that.
03:09:37 Quit G4Oblivion (Quit: CGI:IRC (EOF))
03:10:49 Part madalu
03:29:26 Join powell14ski_ [0] (
03:37:50 Quit dfkt_ (Quit: -= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.)
03:39:03 Quit GeekShadow (Quit: The cake is a lie !)
03:40:42 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
03:41:00JdGordon|is "ew" from the forums in here?
03:41:10JdGordon|S_a_i_n_t: ping me when you come online
03:41:19 Part JdGordon|
03:51:33 Quit simonrvn (Quit: bbiab)
04:03:46S_a_i_n_tJdGordon: (though I know you're not here...) Pong.
04:04:21 Join huffpuff [0] (
04:04:56S_a_i_n_tTheSeven: Pingeroony?
04:05:30 Join simonrvn [0] (
04:06:16S_a_i_n_tTheSeven: G4Oblivion has done some testing apparently, and seems to think that the Nano2G Stkov NAND panic was introduced in r27020
04:08:40S_a_i_n_tI lost all my testing results in "The Great HDD Failure of Last Week", but, I seem to remember my testing telling me that there were good revisions well after that....but, I also got different results on different, I don't know what to expect.
04:11:11S_a_i_n_tRight now I'm trying to bisect the "skipping multiple tracks *really* fast causes Undefined Instruction" bug, and not having a bunch of luck there either.
04:14:13 Join steve|m1 [0] (
04:14:13 Quit steve|m (Disconnected by services)
04:14:46 Nick steve|m1 is now known as steve|m (
04:19:51CIA-81New commit by jdgordon (r27956): Should fix the line alternator issue described in
04:22:22CIA-81r27956 build result: All green
04:22:45 Join G4Oblivion [0] (
04:23:42 Join timccccc [0] (~lisa@
04:24:08G4Oblivion@ S_a_i_n_t: I just built the newest revision of rockbox and manually changed the nand-nano2g.c to match r27019's and I am not getting the crashes anymore.
04:24:09 Quit amiconn (Disconnected by services)
04:24:09 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:24:24 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:24:49G4OblivionI know little to nothing of programming though. I have no idea why it caused the crashes
04:24:57 Quit pixelma (Disconnected by services)
04:25:00 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:25:02 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:25:54 Quit TheSeven (Ping timeout: 258 seconds)
04:28:00G4Oblivionr27020 is definitely whats causing the crashes.
04:28:28S_a_i_n_tG4Oblivion: Could you please post a diff of your changes?
04:28:36S_a_i_n_t(assuming you know how to diff)
04:29:06S_a_i_n_tMy solution was to more than double the NAND stack.
04:29:13S_a_i_n_tWorks, but, hardly ideal.
04:29:34 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:30:39S_a_i_n_tI _think_ the default NAND stack size is ~0x400, at 0x800 I could still get it to panic in some circumstances...but at 0x1000 its fine.
04:30:39G4OblivionI have no idea how to post a diff
04:30:57S_a_i_n_tagain, hardly an ideal "fix" ;)
04:31:14S_a_i_n_tG4Oblivion: You know how to make a diff?
04:31:17G4Oblivionall I did was revert the changes made in r27020 with the newest revision
04:31:30S_a_i_n_t"svn diff > diffname.patch"
04:31:46 Join Barahir [0] (
04:31:49S_a_i_n_t(without quotation marks)
04:32:04G4OblivionBarely know linux =/
04:33:53G4OblivionI ran that in my rockbox folder and got a diffname.patch
04:33:59G4Oblivionis that what you want
04:34:23S_a_i_n_tyou ran it in the source directory? and, yes.
04:34:33S_a_i_n_tpastebin it, please.
04:34:45clone4crwS_a_i_n_t: Just added better LCD support for the Nano in the resistor plugin, FYI
04:35:35 Quit Barahir_ (Ping timeout: 276 seconds)
04:35:57S_a_i_n_tclone4crw: Awesome, I shall have a look. I'll also get on with "prettying up" the other bitmap, maybe do some more on the other one I started initially also.
04:36:22S_a_i_n_tI know it's irrelevant, but, I like things to look good.
04:36:38S_a_i_n_tIt could look like crap and still be a cool, functional plugin ;)
04:36:57S_a_i_n_tG4Oblivion: Thanks.
04:37:24G4OblivionNo problem. Hope it gets fixed soon.
04:38:14S_a_i_n_tAs do I, I'm really glad you've managed to narrow it to one particular revision. I couldn't do that myself.
04:38:25clone4crwS_a_i_n_t: Well, of course. But if we were all for utilitary, and not looks, we'd all be using window managers that look like Windows 98
04:38:28S_a_i_n_tMy good/bad revisions kept changing on me.
04:39:00S_a_i_n_trevisions I knew to be good, randomly paniced on boot, ones I knew to be bad...booted fine. was really very frustrating.
04:40:02S_a_i_n_tG4Oblivion: TheSeven can't reproduce it *at all*, which is *really* weird.
04:40:15G4OblivionI just started doing it to things that affect/might affect the nano
04:40:20S_a_i_n_tOut of interest G4Oblivion, what capacity is your Nano2G?
04:40:43G4OblivionI just got ubuntu the other day. First time using a linux system.
04:41:14S_a_i_n_tI have 2 and 4GB Nanos, and TheSeven's is an 8GB.
04:41:36S_a_i_n_tI can reproduce the panic on both, but, TheSeven can't on his 8GB
04:41:50S_a_i_n_twhy the capacity would change anything...I have no idea, but, apparently it does.
04:42:19G4OblivionApple isn't one to change something thats already working. I don't know why it would be different.
04:44:13S_a_i_n_tIt puzzles me also, appart from the NAND capacity, there *shouldn't* be anything different between my Nanos, your Nano, and TheSeven's Nano...but, apparently there is.
04:44:43S_a_i_n_tAnd I can't see how the NAND capacity is a factor in this.
04:44:50*S_a_i_n_t shrugs
04:46:48 Quit BlakeJohnson86 (Ping timeout: 245 seconds)
04:46:54G4OblivionWell, I hope it gets figured out. If TheSeven or someone can't find a fix soon, might want to revert 27020. The nano is probably considered unusable right now.
04:47:32G4OblivionIts getting late so I'm going in for the night. 'night
04:47:40S_a_i_n_tG4Oblivion: A simpler "fix" is to change the default NAND stack size to 0x1000
04:48:42 Quit G4Oblivion (Quit: CGI:IRC (EOF))
04:49:47 Join BlakeJohnson86 [0] (
04:49:48S_a_i_n_tfirmware\target\arm\s5l8700\ipodnano2g\nand-nano2g.c, line 94, change from "static long nand_stack[DEFAULT_STACK_SIZE];" to "static long nand_stack[0x1000];"
04:49:52S_a_i_n_tawww, crap. too late.
04:50:03 Quit timccccc (Ping timeout: 258 seconds)
04:54:53***Saving seen data "./dancer.seen"
04:59:33 Quit clone4crw (Quit: leaving)
05:04:16 Join clone4crw [0] (
05:06:48 Quit clone4crw (Client Quit)
05:07:30 Quit BlakeJohnson86 (Ping timeout: 240 seconds)
05:08:57 Join BlakeJohnson86 [0] (
05:09:48 Join anewuser [0] (anewuser@unaffiliated/anewuser)
05:17:24 Quit kaiscene ()
05:18:59 Join [Saint] [0] (S_a_i_n_t@
05:19:26 Quit S_a_i_n_t (Ping timeout: 255 seconds)
05:22:15 Join bunnyboi [0] (
05:22:26 Quit bunnyboi (Read error: Connection reset by peer)
05:23:56 Nick [Saint] is now known as S_a_i_n_t (S_a_i_n_t@
05:25:19 Quit Horscht (Quit: Verlassend)
05:28:46 Quit Zarggg (Quit: Zarggg)
05:33:45 Quit ps-auxw (Ping timeout: 258 seconds)
05:37:26 Quit S_a_i_n_t (Ping timeout: 255 seconds)
05:45:08 Join ps-auxw [0] (
05:56:09 Join sweetandy [0] (~sweetandy@unaffiliated/sweetandy)
05:56:12 Join milz [0] (
05:57:46 Quit Judas_PhD (Ping timeout: 265 seconds)
06:07:23 Quit milz (Remote host closed the connection)
06:08:44 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
06:09:37 Quit huffpuff (Quit: leaving)
06:09:56*JdGordon| thinks it should be possible to scan a sublines tokens for a %t() hiding inside conditionals (up to some max depth) and not waste too much cpu time
06:10:15JdGordon|that would fix the issue with %t(0) inside conditionals
06:10:57 Join Judas_PhD [0] (
06:14:20 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
06:15:55 Join milz [0] (
06:18:15 Quit milz (Remote host closed the connection)
06:21:11 Part sweetandy ("This space intentionally left blank.")
06:26:16 Quit MagusG (Ping timeout: 272 seconds)
06:54:55***Saving seen data "./dancer.seen"
06:55:52 Join funman [0] (~fun@rockbox/developer/funman)
06:56:36funmanhi, i got FS #11587 (voice for clipv1) working, i'll waita bit before committing someone wants to comment
06:56:51funmanalso perhaps it would make sense to use it on other targets to reclaim a bit of memory
06:58:29saratogawhats the downside, more disk spin ups?
07:00:34funmanyes, so i think it would make sense only on flash targets, where memory isn't that important .. so i guess a blind user has to try
07:01:54funmanthere are 64 entries buffered so it depends how frequently you go in the menus i think
07:04:36funmanafter this goes in we can move the clipv1 to stable and focus on the last issue: battery life
07:05:25saratogai think testing the amsv1 targets again with boosting between 31 and (approximately) 100Mhz might give interesting battery life
07:05:56saratogasince 62 is too high for low boost, but 250MHz results in very poor performance per clock due to memory
07:06:25funmanmemory clock is (nearly) at the max on amsv1 already
07:06:50 Join lostlogic [0] (~lostlogic@rockbox/developer/lostlogic)
07:06:51funmanso it would be 62 (unboosted) and 124 (boosted) ?
07:08:57saratoga32 unboosted, 124 boosted
07:09:00saratogawould be interesting I think
07:18:08funmani'll try with 32/128 & pclk @ 32MHz
07:19:51 Join Zarggg [0] (
07:20:21funmanclipv1 boots noticeably slower, might not be worth to do it on fuzev1/e200v2
07:26:35 Join Chronon [0] (
07:33:45saratogafunman: thats with 32 pclk boosted too?
07:36:30 Join stoffel [0] (
07:43:36 Join l403 [0] (
07:45:04 Join fyrestorm [0] (
07:46:05funmanjust started the bench
07:52:04funmanhm it just crashed
07:52:45funmani'll try with no voice loaded
08:03:30 Quit solexx_ (Ping timeout: 260 seconds)
08:05:14 Join solexx [0] (
08:05:54 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
08:10:43 Quit Not-iH8sn0w (Ping timeout: 240 seconds)
08:13:47 Join Not-iH8sn0w [0] (
08:16:43 Quit Chronon (Ping timeout: 240 seconds)
08:19:15 Join S_a_i_n_t [0] (S_a_i_n_t@
08:24:25 Join Kitar|st [0] (
08:26:24 Quit Rob2223 (Read error: Connection reset by peer)
08:26:57 Join Rob2222 [0] (
08:30:29funmananother crash with same symptoms: backlight goes on and i see the wps, pressing a button reboots
08:31:29funmansaratoga: iirc you made a change to enable the backlight on panics?
08:31:44saratogafunman: no i gave up when I couldn't figure out how to do it
08:33:03 Quit r0b- (Ping timeout: 264 seconds)
08:33:39funmanah it's ranma in r26917 (only for AMS), but it misses something
08:34:46funmanwe need lcd_update() after enabling backlight else it has the old content
08:37:12 Quit S_a_i_n_t (Ping timeout: 276 seconds)
08:41:16 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
08:44:11funmansegfault in lcd_puts_scroll_style_offset()
08:44:26funmanalthough screen was off
08:44:56 Join ender` [0] (
08:46:08saratogathis is just with the clock changed?
08:47:28funmanyes but it i didn't try with current build. i just changed the theme to cabbiev2 to check if it was my theme
08:49:11CIA-81New commit by funman (r27957): AMS: update the LCD on system exception to be sure the content is up to date ...
08:50:14 Join LinusN [0] (~linus@rockbox/developer/LinusN)
08:50:42amiconnfunman: For voice on lowmem you might want the same mechanism as on archos, i.e. not a separate voice buffer, but loading voice to the audio buffer after stopping playback
08:51:10CIA-81r27957 build result: 0 errors, 14 warnings (funman committed)
08:51:21amiconnThis means no voice during playback, but you don't need a separate voice buffer
08:51:28 Quit scorche (Disconnected by services)
08:51:38 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
08:53:30CIA-81New commit by funman (r27958): include lcd.h , forgotten in r27957
08:55:00***Saving seen data "./dancer.seen"
08:55:02funmanmpegplayer isn't realtime on the fuze with the change, 15 fps if i read correctly
08:55:13CIA-81r27958 build result: All green
08:56:19 Quit Kitar|st (Ping timeout: 240 seconds)
08:56:43Stummiis there a id3-tag-editor for rockbox?
08:59:45 Quit anewuser (Quit: FAMICOMPO 7 - August 15 Through August 31 JST -
09:01:46 Join Kitar|st [0] (
09:06:33 Join S_a_i_n_t [0] (S_a_i_n_t@
09:12:36 Quit funman (Quit: free(random());)
09:18:32 Join [Saint] [0] (S_a_i_n_t@
09:19:15 Quit S_a_i_n_t (Ping timeout: 264 seconds)
09:19:27 Nick [Saint] is now known as S_a_i_n_t (S_a_i_n_t@
09:26:33 Join Rob2223 [0] (
09:28:32 Join giovanni_ [0] (~giovanni@
09:28:56 Join petur [0] (~petur@rockbox/developer/petur)
09:30:33 Quit Rob2222 (Ping timeout: 276 seconds)
09:32:56giovanni_I just updated the theme 'FullTouch' for ondavx777, please take a look, I am very pleased with the result.
09:34:22giovanni_And, if there is someone who ca do it, please, rename the theme from 'FullTouch 0.1' to 'FullTouch', thanks
09:37:30S_a_i_n_tgiovanni_: why didn't you just replace the theme?
09:37:42S_a_i_n_twhy rename it with +0.1?
09:41:43 Quit TheSeven (Ping timeout: 240 seconds)
09:47:07 Quit S_a_i_n_t (Ping timeout: 240 seconds)
09:48:30giovanni_because I didn't understand well how to update a theme
09:49:42 Join S_a_i_n_t [0] (S_a_i_n_t@
09:49:43giovanni_S_a_i_n_t: Now I've learned
09:54:39 Join T44 [0] (
09:58:54 Quit Topy44 (Ping timeout: 265 seconds)
10:00:31 Quit JdGordon| (Quit: leaving)
10:02:51 Join Jaykay [0] (
10:05:08 Join rob [0] (
10:05:47 Quit kadoban (Remote host closed the connection)
10:06:51 Join kadoban [0] (
10:07:00 Quit user890104 (Read error: Connection reset by peer)
10:07:06 Join user890104 [0] (
10:15:05 Quit lostlogic (Read error: Connection reset by peer)
10:15:25 Join lostlogic [0] (~lostlogic@rockbox/developer/lostlogic)
10:52:03 Quit sasquatch (Quit: WeeChat 0.3.2)
10:52:27 Join sasquatch [0] (
10:55:04***Saving seen data "./dancer.seen"
11:10:01 Quit user890104 (Ping timeout: 272 seconds)
11:12:11 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:21:05 Join TheSphinX^ [0] (
11:24:03 Quit simabeis_ (Ping timeout: 264 seconds)
11:26:38 Join simabeis [0] (
11:29:54 Join hebz0rl [0] (
11:40:44 Quit l403 (Ping timeout: 240 seconds)
11:41:26 Join l403 [0] (
11:45:58 Quit jae (Ping timeout: 272 seconds)
11:46:36 Join jae [0] (
11:57:50giovanni_I've optimized my theme for the ondas DAP, this theme is touch screen friendly so probably will be good also on the android port.
12:06:40 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
12:15:04 Join einhirn [0] (
12:16:16 Quit bmbl (Ping timeout: 258 seconds)
12:16:18 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
12:16:43 Join MagusG [0] (
12:19:36 Join _s1gma [0] (~d.d.derp@
12:35:11 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
12:41:57 Join user890104 [0] (
12:55:05***Saving seen data "./dancer.seen"
12:59:58 Part LinusN
13:18:13 Quit fxb__ (Ping timeout: 276 seconds)
13:29:21 Join LinusN [0] (~linus@rockbox/developer/LinusN)
13:29:23 Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:36:59 Join esperegu [0] (~quassel@
13:56:46 Join h3po [0] (
14:01:00 Quit S_a_i_n_t (Ping timeout: 245 seconds)
14:10:58 Quit parafin (Ping timeout: 246 seconds)
14:12:12 Join parafin [0] (
14:26:33 Join efyx [0] (
14:29:27 Quit Zarggg (Quit: Zarggg)
14:49:28 Quit balintx (Ping timeout: 265 seconds)
14:55:09***Saving seen data "./dancer.seen"
14:56:28 Quit bmbl (Read error: Connection reset by peer)
14:58:04 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
14:58:38 Quit giovanni_ (Quit: Sto andando via)
14:59:33 Quit togetic (Ping timeout: 265 seconds)
15:00:52 Join balintx [0] (
15:06:43 Join user890104_ [0] (
15:08:13 Join bug2000 [0] (~bug@unaffiliated/bug2000)
15:08:31bug2000Where do I get Sansa Clip boot loader v2.0? The file linked in the manual seems to be v1.0
15:08:48 Quit user890104 (Ping timeout: 272 seconds)
15:15:20 Join S_a_i_n_t [0] (S_a_i_n_t@
15:16:16 Quit TheSphinX^ (Quit: XChat)
15:16:16 Quit dfkt (Read error: Connection reset by peer)
15:16:21 Join dfkt [0] (dfkt@unaffiliated/dfkt)
15:19:25 Nick user890104_ is now known as user890104 (
15:21:30 Join DexterLB [0] (
15:21:43DexterLBhow do I disable the database?
15:22:52 Quit S_a_i_n_t (Ping timeout: 264 seconds)
15:22:56 Quit krazykit (Quit: Connection reset by beer)
15:24:18 Quit DexterLB (Quit: So long and thanks for all the fish!)
15:24:29 Join S_a_i_n_t [0] (~st.lasciv@
15:25:49 Quit h3po (Quit: Leaving.)
15:29:40 Join {phoenix} [0] (
15:34:25 Join [Saint] [0] (S_a_i_n_t@
15:34:31 Join anewuser [0] (anewuser@unaffiliated/anewuser)
15:34:31 Quit S_a_i_n_t (Ping timeout: 276 seconds)
15:35:56 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
15:36:05JdGordon[Saint]: ping?
15:37:08JdGordonhey, free to test a fix for that %t(0) issue?
15:38:21[Saint]Not right this very second...but, if you pastebin it, I could do it at some point overnight.
15:38:48JdGordonok, will do
15:39:04[Saint]...sorry, I would, but...kinda busy with IRL stuff :/
15:39:14JdGordonI'm not exactly happy with this fix though :/
15:39:15JdGordonno worries
15:41:00 Quit hebz0rl (Quit: Ex-Chat)
15:41:24 Join Strife89 [0] (~Strife89@
15:41:52 Join krazykit [0] (~kkit@
15:42:58 Join giovanni_ [0] (~giovanni@
15:47:06 Join n1s [0] (~n1s@rockbox/developer/n1s)
15:48:55 Nick [Saint] is now known as S_a_i_n_t (S_a_i_n_t@
15:50:22 Part LinusN
16:03:34bug2000mc2739: It's 1.0
16:06:23mc2739bug2000: hmm, you are correct. I guess they ever got uploaded to the server.
16:06:44mc2739ping funman and he should be able to help you out
16:07:27bug2000Yey memoserv! ^_^
16:15:51 Quit T44 (Quit: Leaving)
16:22:52 Quit jae (Ping timeout: 264 seconds)
16:23:08 Join jae [0] (
16:29:02 Quit jae (Ping timeout: 265 seconds)
16:29:38 Join jae [0] (
16:39:25 Join fip [0] (
16:40:31 Quit giovanni_ (Quit: Sto andando via)
16:41:29 Quit pamaury (Remote host closed the connection)
16:42:30 Quit jae (Ping timeout: 265 seconds)
16:42:50 Join jae [0] (
16:51:01 Part fip
16:54:40 Quit bmbl (Quit: Bye!)
16:55:11***Saving seen data "./dancer.seen"
17:00:05 Join [sko] [0] (~sko]
17:03:15 Join jgarvey [0] (
17:03:23 Join bmbl [0] (~Miranda@unaffiliated/bmbl)
17:03:56 Part Zagor
17:07:38 Quit [sko] (Ping timeout: 265 seconds)
17:08:00 Join [sko] [0] (~sko]
17:08:42 Join pamaury [0] (
17:08:42 Quit pamaury (Changing host)
17:08:42 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
17:19:18 Join G4Oblivion [0] (
17:20:55G4OblivionRemoving "call_storage_idle_notifys(false);" in nand-nano2g.c line 726 and its brackets fixed the *PANIC* stkov nand crashes.
17:21:47S_a_i_n_tG4Oblivion: I haven't caught up with TheSeven yet, but, I suspect he'll find an elegant solution.
17:22:10G4OblivionOk. Just letting it be known that thats whats causing the problems.
17:22:10S_a_i_n_tIt's all his code in nand-nano2g.c anyway IIUC
17:22:14TheSevenG4Oblivion: that's what I suspected
17:22:15*docgnome sighs
17:22:20docgnomei hate getting stuff from usps
17:22:22docgnomeso sloooow
17:22:26TheSevenmy question is just what has changed with that callback
17:22:38TheSevenit used to work some hundred revisions ago, and it still works for me
17:22:46TheSevenif anyone can bisect it, i would be very glad
17:22:54S_a_i_n_tg4 did
17:23:01S_a_i_n_tI posted about it earlier
17:23:04TheSevendid he get an exact revision number?
17:23:09*TheSeven apparently missed it
17:23:10G4OblivionYes. r27020
17:23:10S_a_i_n_tapparently, yes.
17:25:43S_a_i_n_tG4Oblivion: You left before I could finish telling you, but, my "fix" for the panic was/is to change DEFAULT_STACK_SIZE to 0x1000 ;)
17:25:51gevaertsDo people who suffer from this use the database or the scrobbler feature?
17:26:13S_a_i_n_tgevaerts: I use database, but, I've never scrobbled in my life
17:26:47G4OblivionI saw that. My temporary fix is working fine for me though.
17:26:47S_a_i_n_tG4Oblivion: Can you please see if you can confirm a bug I'm seeing?
17:26:59TheSevennow this is strange
17:27:02gevaertsThere aren't *that* many places that register idle callbacks, and if commenting out the call_storage_idle_notifys() call "fixes" the issue, one of them could be the problem
17:27:22TheSevenso this has been happening all the time since we added that callback?
17:27:33TheSevenwhy is it using more stack for some people and not for others?
17:27:41S_a_i_n_tG4Oblivion: Try skipping traks (whilst playing) as fast as you possibly *should* crash with an undefined instruction.
17:27:49TheSevenapparently this happens when something called by the callback tries to actually write to a file
17:28:06TheSevenwhich points towards the stack size increase actually being the right fix for it, not just a workaround
17:28:11G4OblivionShould I do this with default rockbox or 0x1000?
17:28:22S_a_i_n_tG4Oblivion: Doesn't matetr.
17:28:31S_a_i_n_tit's unrelated AFAIK
17:28:38G4OblivionOk. let me go grab my nano
17:29:06S_a_i_n_tTheSeven: 0x1000 seems a little bit extreme though.
17:29:20TheSeventry 0xc00
17:29:39TheSeven0x800 is needed for the fsinfo sector buffer alone, and the must be some bytes left for other things
17:29:41gevaertsYou shouldn't have to increase DEFAULT_STACK_SIZE. Just the nand_stack in nand-nano2g.c should be enough
17:29:45S_a_i_n_tit's what I had to push it to to guarantee it wouldn't panic, but, I'll try that.
17:29:48TheSeveni'd probably go for default + 0x800
17:30:14TheSevengevaerts: let's get rid of those damn on-stack sector buffers
17:30:18S_a_i_n_tisn't default 0x400?
17:30:34S_a_i_n_tso...1200 total?
17:30:43gevaertsit is, yes
17:30:56G4OblivionI started a song in database and tried skipping tracks as fast as I could and did not get a crash.
17:31:09S_a_i_n_tI mean..good.
17:31:10TheSevengevaerts, pamaury: i have proof of concept code for that in embios, which uses a central static (and properly synchronized) scratch buffer in all places that used to use on-stack buffers
17:31:18S_a_i_n_tbut, weird I'm seeing it on all my 2Gs
17:31:34G4Oblivioncould it be your config?
17:31:53S_a_i_n_tG4Oblivion: I can reproduce it with default settings
17:32:06S_a_i_n_ton multiple different buikds, aslo.
17:32:09*TheSeven can't
17:32:15G4OblivionI gotta go. I will test it more later.
17:32:22S_a_i_n_tTheSeven: Really? Aaaawww...shit.
17:32:23gevaertsG4Oblivion: if you trigger lots of activity, it will be less likely to crash I think
17:32:31S_a_i_n_tthe Nano2G is just a barrel of laughs no?
17:33:09S_a_i_n_tgevaerts: With my Nanos, skipping say 6~10 tracks in a row will almost guaranteed crash the player.
17:33:10 Quit G4Oblivion (Quit: CGI:IRC (EOF))
17:33:30gevaertsS_a_i_n_t: do you do ratings?
17:33:43S_a_i_n_tthis isn't doing "skip..think about it for a bit...skip another, listen to that for a short time...skip again"
17:33:43gevaertsI mean, "gather runtime data"?
17:33:53S_a_i_n_tbut rather "skip, skip, skip, skip" etc.
17:34:02S_a_i_n_tgevaerts: yes.
17:34:13S_a_i_n_tbut, that's not a default setting, is it?
17:34:16gevaertsS_a_i_n_t: can you make it crash if you disable that?
17:34:22gevaertsno, I think it's off by default
17:34:22S_a_i_n_tI can reproduce with a default .cfg
17:35:49S_a_i_n_tgevaerts: The Nano2G is some cruel bitch...really. This NAND bug that's been plagueing many a user, is completely unreproducable by TheSeven
17:36:11S_a_i_n_tso it kinda doesn't surprise me that I have a weird unreproducable bug also ;)
17:36:18gevaertsS_a_i_n_t: there's a call to register_storage_idle_func() in apps/tagcache.c, and one in apps/playlist.c. If you comment those out, does it still crash? What if you comment them out separately?
17:36:34TheSevenS_a_i_n_t: did you confirm that it happened for you with the very same build i am using, which i sent you some days ago?
17:36:53S_a_i_n_tgevaerts: I'll have to get back to you on that one...doing IRL work, not RB stuff at the moment.
17:37:09gevaertsThere are also registrations from battery_bench, scrobbler, and settings, but settings should be irrelevant I think, and the other two aren't used by default
17:37:10TheSevengevaerts: my suspicion is that everything triggering an fsinfo update, so probably everything that allocates a new cluster, will trigger it
17:37:16S_a_i_n_tTheSeven: Yeah...that build failed for me.
17:37:24TheSevennow that's interesting
17:37:24gevaertsTheSeven: but then why aren
17:37:29S_a_i_n_tI suspected it would.
17:37:29gevaerts't some people seeing it?
17:37:50TheSevenit must be related to some files on the player
17:38:06TheSevenor maybe cluster size
17:38:29gevaertsS_a_i_n_t: note that if you comment those out, some bits will not save data, so don't use those builds for too long
17:38:31S_a_i_n_tTheSeven: That's what I thought...but, I tried it with *just* the OF and the RB installation...still panic'd
17:39:09S_a_i_n_tTheSeven: perhaps you don;t see it, as you have no OF?
17:39:10*TheSeven has a superfloppy-formatted 8k cluster size ipod
17:39:30TheSeveni doubt that stuff outside of .rockbox is related to it
17:39:37TheSeven(besides maybe music files)
17:39:41S_a_i_n_tAh, hmmm. I said, I tried it with just RB and the OF, and still it panic'd at boot.
17:40:30TheSevenhow is your ipod formatted/partitioned?
17:40:36S_a_i_n_toccasionally I can get to the menu, but as soon as I scroll, or select something...boom.
17:41:01S_a_i_n_tTheSeven: They're all "as it comes" from iTunes/Windows
17:41:41TheSevenaha... that may be related then
17:42:01TheSevenat least it would explain why i'm not experiencing it with iloader
17:42:11TheSevendidn't use an rb bootloader for ages
17:43:18S_a_i_n_tThe GF has temporarily misplaced (for an extended period of time) the iLoadered Nano2G I have floating around.
17:43:34S_a_i_n_tmethinks it to be in the car, somewhere.
17:43:44 Quit Strife89 (Read error: Connection reset by peer)
17:43:58 Join Strife89 [0] (~Strife89@
17:45:44gevaertswell, it's pretty clear. update_fsinfo() uses an on-stack sector, i.e. 0x800 bytes. The idle callback thread uses a 0x400 bytes stack. That *can't* work
17:46:21gevaertshm, no
17:46:24*gevaerts is confused
17:46:34gevaertsIt's a 0x400*sizeof(long) stack
17:49:40TheSeven* or /?
17:50:04gevaertswell, long[DEFAULT_STACK_SIZE]
17:50:11TheSeven/ would make way more sense
17:50:34gevaertswhich means char[DEFAULT_STACK_SIZE*sizeof(long)] if you want bytes
17:50:42TheSevenah, so it's just 0x400?`
17:50:53gevaertsyes, but it's 0x400 longs
17:50:56*TheSeven thought you meant that DEFAULT_STACK_SIZE == 0x400*sizeof(long)
17:51:20TheSevenhmm, so it's apparently two sector buffers then
17:51:45TheSevenwriting a direntry that causes a cluster allocation? that might have two of them
17:52:27 Quit rob (Disconnected by services)
17:52:41gevaertsupdate_fsinfo() is usually called at the end of functions that don't have big stack usage themselves, so it *should* be safe against that if I read the code correctly
17:53:34gevaertsah, no
17:54:35gevaertsfat_create_dir() and fat_rename() are suspect
17:55:00gevaertsThey use a sector buffer of their own, and call fat_flush() which calls update_fsinfo() which also has a sector buffer
17:55:53 Quit anewuser (Quit: FAMICOMPO 7 - August 15 Through August 31 JST -
17:56:49TheSeventheir own sector buffer isn't needed at that place any more, is it?
17:57:23TheSevencould adding a set of braces to reduce its scope help? or will the compiler still keep it allocated until the end of the function for performance reasons?
17:57:25gevaertsno idea
17:57:38gevaertsah, yes, a new block might help
17:59:47TheSevenwhat do you think is more likely to happen in a storage idle callback? creating a dir or renaming something?
17:59:53 Join r0b- [0] (
18:00:25gevaertsOne of them is saving the playlist control file, the other is saving some database things. Neither even involves directories
18:00:32TheSevenhmm, things being written to a temp file and then swapped around for safety?
18:00:42gevaertsyes, that's what I'm thinking
18:02:00 Quit petur (Quit: *plop*)
18:04:05 Quit Strife89 (Quit: Departure.)
18:04:12 Join komputes [0] (~komputes@ubuntu/member/komputes)
18:08:08 Quit bug2000 (Ping timeout: 276 seconds)
18:08:47 Join Lear [0] (chatzilla@rockbox/developer/lear)
18:11:57*TheSeven should probably add some logging
18:12:11 Quit [sko] (Read error: Connection reset by peer)
18:12:26 Quit lostlogic (Quit: leaving)
18:13:48TheSevenactually we should add a way to dump the logf buffer, or just the entire memory, after a panic via usb or something
18:17:21preglowbetter to just set up a console via usb
18:17:31preglowyou can't count on anything working after a panic has happened
18:17:38saratogai wrote a patch ages ago that just continuously dumped the logf buffer to disk
18:17:55TheSeventhat's not very helpful if the storage code is broken :)
18:19:33TheSeveni think there are ways to do usb-based post-mortem memory dumps that are more likely to succeed than the printing of the panic message
18:20:21TheSevenand a usb console will either slow down things dramatically or lose the last (and probably most important) part of the log
18:24:58 Quit l403 (Ping timeout: 265 seconds)
18:25:22 Join fxb__ [0] (
18:26:11 Join lostlogic [0] (~lostlogic@rockbox/developer/lostlogic)
18:27:08 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:34:51 Join [Saint] [0] (S_a_i_n_t@
18:35:06[Saint]man, what a time to fall off the net :/
18:35:06 Quit S_a_i_n_t (Ping timeout: 252 seconds)
18:35:27[Saint]fxb__: I can't reach , is it known broken?
18:35:44 Join l403 [0] (
18:35:52[Saint]Or, can you suggest anywhere else I can pull a .ipsw file?
18:42:39n1s[Saint]: i can acces that page, but the list of ipod firmware is empty
18:43:08n1sah, works now
18:47:00 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
18:47:59[Saint]n1s: I can't acess it, nor can "downforeveryoneorjustme"...weird.
18:48:18[Saint]I guess it has something to do with my locality.
18:48:46 Quit Barahir (Ping timeout: 240 seconds)
18:49:31pixelma[Saint]: I can access it now but couldn't earlier - did you try a forced refresh or a different browser?
18:50:28[Saint]pixelma: Yes, I tried different browsers...and tried deleting my caches, etc. :/
18:50:44 Join Barahir [0] (
18:50:46[Saint]downforeveryoneorjustme can reach it now...but I still can't :/
18:50:58[Saint]I assume I'll be able to in time, not important I guess.
18:51:32 Nick [Saint] is now known as S_a_i_n_t (S_a_i_n_t@
18:55:15***Saving seen data "./dancer.seen"
19:01:50 Join h3po [0] (
19:09:28 Join Jerom [0] (~jerome@
19:11:04 Join Horscht [0] (~Horscht@xbmc/user/horscht)
19:20:41 Quit {phoenix} (Remote host closed the connection)
19:24:51pamauryTheSeven: I just notices you ping me, I don't understand what you mean
19:25:40pamauryAh yes I get it. You would like to copy their model/code ? Or it's just a though to justify the approach ?
19:26:13pamauryI guess it would be easier to debug if the simulator could emulate a driver with a file, I begun this but did not finished it
19:26:41amiconnTheSeven: How about a gdb stub for usb?
19:27:19amiconnFor archos we have a gdb stub for serial, and iirc one arm target has one for usb already
19:27:37TheSeventhat's roughly what i'm thinking about, but i haven't investigated how such a thing works yet
19:28:14 Join fragilematter [0] (~barbu_dor@
19:28:17 Quit tchan (Ping timeout: 264 seconds)
19:28:23 Join bunnyboi [0] (
19:32:25 Join toffe82 [0] (~chatzilla@
19:35:01 Quit andash (Read error: Operation timed out)
19:35:24 Quit S_a_i_n_t (Ping timeout: 240 seconds)
19:38:39 Quit guymann (Quit: -uDNav world time)
19:43:00fragilematterhello everyone
19:43:19 Join Buschel [0] (
19:43:51fragilematterI've got a quick question: have any of you experienced ogg vorbis decoding problems, especially on the e200v2?
19:44:30BuschelCan anyone with a clip+ or a fuze v1 check if FS #11476 is reproducible? (musepack sv8 seek instability)
19:44:41BuschelI cannot reproduce this at all...
19:45:39Buschelsv8 files are availabel here ->
19:46:20 Join Lear_ [0] (chatzilla@rockbox/developer/lear)
19:48:04 Join Lear__ [0] (
19:48:36 Quit Lear (Ping timeout: 272 seconds)
19:48:45 Nick Lear__ is now known as Lear (
19:50:02 Join evilnick [0] (
19:50:23gevaertsbluebroth3r, domonoky: I suspect we can do FS #11580 (the ipod video merge) more or less independently from a new rockbox utility release
19:51:22gevaertsIf the merge is first but we leave both build for a while, the transition will be transparent. If the new rockbox utility configuration is first, people will just get a 32mb build for a while.
19:51:34 Quit Lear_ (Ping timeout: 265 seconds)
19:55:14gevaertsI'd like one or two more test reports. Right now I only have two. One from creposucre on a 5G 64G, and one on my 5G 30G, but I don't really use mine, so if random things break (which I don't expect), I wouldn't notice
19:57:22 Join Buschel_ [0] (
19:57:36 Join andash_ [0] (~andash@unaffiliated/andash)
19:58:49 Quit Buschel (Ping timeout: 265 seconds)
19:58:49 Nick Buschel_ is now known as Buschel (
19:59:42bluebroth3rgevaerts: good point
20:01:03gevaertsbluebroth3r: if I commit in the next one or two days, I guess I'll just leave both builds around until rockbox utility is ready
20:02:28 Join Transformer [0] (
20:03:08 Part Transformer
20:03:54 Quit bunnyboi (Quit: IRC is just multiplayer notepad)
20:11:02 Join S_a_i_n_t [0] (S_a_i_n_t@
20:12:21 Quit Buschel (Ping timeout: 240 seconds)
20:12:25 Quit evilnick (Quit: evilnick)
20:23:09 Join Buschel [0] (
20:26:19 Quit Jerom (Quit: Leaving.)
20:26:36 Quit andash_ (Read error: Operation timed out)
20:27:11 Join Jerom [0] (~heidi@
20:27:20 Quit kadoban (Ping timeout: 265 seconds)
20:27:50 Quit Dreamxtreme (Quit: I was raided by the FBI and all I got to keep was this lousy quit message!)
20:28:42 Join fip [0] (
20:28:56 Join kaiscene [0] (
20:29:42 Join kadoban [0] (
20:30:41 Join andash [0] (~andash@unaffiliated/andash)
20:32:41 Quit andash (Read error: Operation timed out)
20:33:41 Join andash [0] (
20:33:41 Quit andash (Changing host)
20:33:41 Join andash [0] (~andash@unaffiliated/andash)
20:34:35 Quit kadoban (Ping timeout: 265 seconds)
20:38:49 Join evilnick [0] (
20:39:30 Quit stoffel (Remote host closed the connection)
20:47:20 Quit andash (Read error: Operation timed out)
20:50:03AlexPgevaerts: You have a couple of reports in the test builds forum, no?
20:50:17 Join andash_ [0] (~andash@unaffiliated/andash)
20:50:46gevaertsAlexP: well, soap's report seems more like "The initial numbers look OK"
20:51:50AlexPNothing has gone wrong in my extensive testing :)
20:52:16 Quit andash_ (Read error: Operation timed out)
20:52:31*gevaerts accuses AlexP of not having an ipod video!
20:52:44AlexPThat's a minor detail :)
20:53:47 Join andash [0] (
20:53:47 Quit andash (Changing host)
20:53:47 Join andash [0] (~andash@unaffiliated/andash)
20:53:52gevaertsIt does qualify what you mean by "extensive" though :)
20:54:10 Part evilnick
20:54:21gevaertsOn the other hand, the 64mb build *really* shouldn't see any changes
20:55:16***Saving seen data "./dancer.seen"
20:55:37 Join kadoban [0] (
20:58:15gevaertsMaybe I should just learn to live dangerously every now and then...
20:59:26 Join [sko] [0] (~sko]
21:00:53 Quit n1s (Ping timeout: 276 seconds)
21:03:25AlexPgevaerts: It is obviously good to ask for testers first, but I suspect something like this won't see widespread testing until it is committed
21:06:36*gevaerts kicks CIA-81
21:07:50gevaertsCIA-81: get on with it!
21:08:12 Join kugel [0] (~kugel@rockbox/developer/kugel)
21:08:14gevaertsWe don't need lazy bots! We can be lazy on our own if needed!
21:08:51CIA-81New commit by gevaerts (r27959): remove the memory specification from ipod video builds in preparation of r27960. The 64mb build will be removed later when supporting infrastructure ...
21:08:51CIA-81New commit by gevaerts (r27960): Unify 32mb and 64mb ipod video builds - FS #11580 ...
21:08:51CIA-81r27960 build result: All green
21:10:56 Quit andash (Read error: Operation timed out)
21:11:31 Part fragilematter
21:11:49gevaertsYay, md5sum tells me that the ipodvideo and ipodvideo64mb builds are now exactly identical :)
21:14:59Torneit would still be nice to handle the default battery capacity
21:15:55gevaertsYes, but I don't see a nice (and clean) way to do that right now
21:19:32gevaertsOne problem is that the settings code really isn't made to support non-constant default values
21:19:41AlexPGood work chaps
21:19:42Buschelgevaerts: what was the status regarding ata.c and (MEM == 64)? with your change this special treatment is also used for the 32MB variant of te ipod video
21:19:53Buschelgevaerts: is this special treatment still needed?
21:21:20gevaertsBuschel: I have no idea if it's still needed or not. The svn logs show that it's been enabled and disabled several times. The only thing I know is that I believe that *if* it's needed, not enabling it on the 32MB builds is a bug, since those can run with the 80GB large sector drives too
21:21:58Buschelgevaerts: true, the fix was introduced for large HDDs...
21:24:02gevaertsYes. The commit message says "for iPod 5.5g 60GB and/or 80GB", and the code checks for 64MB...
21:27:19gevaertsBuschel: actually, you should know all about this. You're the one who opened FS #8568 :)
21:28:03Buschelyes, but I own a 30GB variant. The user who actually had problems was Horscht.
21:28:35Tornepresumably the problem was that the locking required for the large sector stuff was really slow
21:28:50Tornehence replacing it with a weird custom spinlock
21:29:14Tornein which case it's probably really slow any time the large sector code is actually used..
21:29:54BuschelIs this stuff still needed for the large HDD's? Or is obselete with the large-HDD-support changes (r25646)?
21:30:31Tornewas that me?
21:30:43Buschelyes :-)
21:30:50Tornei just hacked it so it only actually uses the emulation code if it has to
21:31:10Tornethe "large hdd support"w as mostly just turning lba48 on by default on everything
21:31:26Torneit still needs to do all the same stuff for the drive that has the problem
21:32:22Buschelok, at least we might remove the "&& MEM == 64" then. as this is true now.
21:32:41Tornedidn't you already remove that, gevaerts?
21:32:59gevaertsno. I decided to leave that out of this patch, since it's really unrelated
21:33:13Tornesomeone should test it without that code and see if it actually makes any difference
21:33:27Torneit seems weird having a single weird custom spinlock thing that's only used by one target in one place
21:33:39Torne if we can throw it away entirely we can drop those odd macros out of kernel.c
21:34:09gevaertsYes, so we really need to find someone with that sort of ipod :)
21:34:13gevaerts(and time)
21:34:40BuschelHorscht: you there?
21:34:57TorneBuschel: gevaerts probably is hinting at me ;)
21:35:11*gevaerts wouldn't dare!
21:35:21AlexPhinting is a waste of time
21:35:25AlexPTorne: test it!
21:35:30AlexPSee, much better :)
21:35:41Tornecan't be bothered at the moment
21:35:43Tornetalk is cheap :)
21:35:45AlexPhehe :)
21:35:59gevaertsTorne: send me your ipod. I'll test :)
21:36:11Tornei need my ipod for french electro house music
21:36:24Torneapparently not anything else any more, according to my history
21:37:54 Quit TheSeven (Ping timeout: 265 seconds)
21:41:28 Join stripwax [0] (
21:42:14 Quit stripwax (Client Quit)
21:43:16 Join stripwax [0] (
21:44:46 Join captainkewllll [0] (2669ecc2@gateway/web/freenode/ip.
21:45:09 Quit gevaerts (Ping timeout: 240 seconds)
21:49:05 Join AndyI [0] (~pasha_int@
21:52:16 Quit [sko] (Quit: Leaving.)
21:52:23 Join yelped [0] (
21:53:00yelpedNew Fuze+ Released :
21:53:22 Quit Buschel (Ping timeout: 265 seconds)
21:53:47 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
21:55:34kugelI suspect funman will have some lengthy nights in the following weeks :)
21:58:50saratogai saw theres a new phillips mp3 player running android, so in theory we could run on that too
21:59:39froggymanyay! it uses a non-proprietary port
22:01:19yelpedNot so yay is the lack of full reviews so far.
22:01:52yelpedWhen the Clip+ was released ABI had their review up the same day.
22:02:16yelpedHowever they still didn't review Cowon J3 yet.
22:02:47yelpedThe're very behind on a lot of things, however they did promise to get better.
22:10:03Kohlrabiwhat's new in Fuze+?
22:10:06KohlrabiJust µSD?
22:10:09Kohlrabiand new firmware?
22:10:58 Join h3po1 [0] (
22:11:23 Quit h3po1 (Client Quit)
22:13:50 Quit h3po (Ping timeout: 276 seconds)
22:14:09yelpedµSD isn't new.
22:14:56yelpedMore Formats; AAC for audio, WMV, H.264, Flip, for Video.
22:15:11yelpedNew design.
22:15:17yelpedMore Capacity.
22:15:31yelpedBetter, Bigger OLED screen.
22:16:06kaisceneuses a micro USB connection like the clip
22:16:46pixelmaI thought it was mini USB on the Clip
22:16:49kugelyelped: where does it say oled?
22:16:50yelpedI guess that means they gave up on docks.
22:17:26bertrikit says LCD right on the page yelped linked
22:17:58 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
22:18:08AlexPKohlrabi: Bigger screen, touch stuff. Fuze already had µSD.
22:19:44AlexPvery disappointing in other words
22:20:01 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
22:20:02KohlrabiMost certainly no upgrade for existing users
22:20:04bertrikkaiscene, the clip doesn't use micro USB, it has mini USB
22:20:14*gevaerts doesn't really like power cuts
22:20:15KohlrabiBut looking at Clip+ vs. Clip, probably the UI got a facelift, too
22:20:28kaiscenei always confuse the two. micro/mini usb
22:21:44 Quit leavittx_ (Ping timeout: 240 seconds)
22:22:06 Quit leavittx (Ping timeout: 240 seconds)
22:23:09yelpedThe screenshot from Dapreview says OLED.
22:26:50 Join krabador [0] (
22:27:22 Quit bmbl (Quit: Bye!)
22:30:42 Join andash [0] (
22:30:42 Quit yelped (Quit: CGI:IRC (EOF))
22:30:42 Quit andash (Changing host)
22:30:42 Join andash [0] (~andash@unaffiliated/andash)
22:31:24 Quit stripwax (Quit:
22:35:28 Join leavittx [0] (~leavittx@
22:35:31 Join guymann [0] (
22:35:40 Quit Lear (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
22:38:58 Quit Jaykay (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716])
22:46:49 Join evilnick [0] (
22:47:21 Join [Saint] [0] (S_a_i_n_t@
22:48:14 Join fyre^OS [0] (
22:49:21 Quit fyrestorm (Ping timeout: 245 seconds)
22:49:50*Horscht returns
22:50:05Horschtwhat did Buschel want earlier?
22:50:23 Quit S_a_i_n_t (Ping timeout: 272 seconds)
22:50:52 Quit evilnick (Client Quit)
22:51:02gevaertsHorscht: do you remember FS #8568 ?
22:51:23gevaertsWe'd like to find out if the workaround introduced back then is still needed
22:52:21Horschtah, the "let-my-harddrive-spin-insanely-long"-issue?
22:53:03gevaertsthat's the one!
22:53:14 Quit bertrik (Ping timeout: 265 seconds)
22:53:18Horschtok, I do remember it, yes
22:53:46 Quit jae (Read error: Operation timed out)
22:53:52Horschtso, what do I need to do in order to help you?
22:54:03Horschtcompile a build with a certain patch?
22:54:08gevaertsYes, basically
22:54:15 Join evilnick [0] (
22:54:24gevaertsAnd compare that with a standard build
22:54:47HorschtI just hope that my virtual machine still works...
22:55:07Horschthaven't compiled rockbox in a while...
22:55:11gevaertsI can provide a build if that helps
22:55:17***Saving seen data "./dancer.seen"
22:55:27Horschtprobably faster that way...
22:55:47 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
22:57:46kugelI wonder if/how the PP caches work with the dual memory mapping in the ipods
22:58:11Horschtcurrently, I am running an "ancient" build anyways.... a mid july build
22:58:47gevaertsHorscht: gevaerts/">
22:59:13gevaertsThat's r27960 with that workaround disabled
22:59:21Horschtthat was quick, my VM hasn't booted by now
23:02:15 Join hebz0rl [0] (
23:03:10Horschtoh, great... now my wps wont work anymore :p
23:03:32gevaertsYou could try the upgrade tool
23:04:00Horschtdid you chamge the wps code again?
23:04:07 Join jae [0] (
23:04:15gevaertshm, wait
23:04:22Horschtanyways: "regular" build = 1:20 seconds until spindown
23:04:24*gevaerts doesn't remember when the last change was
23:04:54Horschtgonna test your build now
23:07:36 Part fip
23:09:24 Quit krabador (Quit: Sto andando via)
23:11:19 Quit evilnick (Quit: evilnick)
23:12:34Horschthm... just above 2 minutes... but hardly any consistent disk activity, just a small "read" every 3 - 4 seconds
23:12:48 Join G4Oblivion [0] (
23:13:20gevaertsMaybe try copying a few files using both builds?
23:13:30gevaerts(preferably the same files each time)
23:13:48gevaertsAnyway, two minutes is a lot more than 1:20...
23:13:48Horschtwait, copying?
23:14:14gevaertsYou can copy files in the file browser :)
23:14:32HorschtI think we might be mixing up issues here. The issue I though we were talking about was heavy and long disk activity during music buffering
23:14:47gevaertsIt is, yes
23:15:06gevaertsBut I'd expect a file copy to show that as well
23:15:46captainkewllllmake fails when I use -j2. current build from svn.
23:15:53G4Oblivion[Saint]: I could not reproduce your track skipping bug.
23:16:41gevaertsHorscht: what is your spin down timeout set to?
23:17:53gevaertshm, default is five seconds. If we're talking about two minute times, five seconds is negligible
23:17:54Tornekugel: the caches undoubtedly *don't* work with the dual memory mapping :0
23:18:05 Join clone4crw [0] (
23:18:07 Join huffpuff [0] (
23:18:10Tornekugel: so it's important that any 16byte part of memory is only accessed via one mirror or the other
23:18:22Horschtgevaerts, default. 5 seconds
23:18:41Horschti retested with the current build, 1:20 minues
23:18:54Horschtyour patched build: 2 minutes
23:18:56gevaertsok, so it's really a 40 second difference
23:19:06gevaertsThat means it's *not* gone away
23:19:14 Quit G4Oblivion (Quit: ChatZilla 0.9.86 [Firefox 4.0b4/20100818120848])
23:19:40Horschtnot entirely, no. It seems to have gotten less severe, though
23:20:04gevaertsYes, but I don't think we can remove the workaround
23:20:39Tornewell the next question is what's the actual issue :)
23:20:41Horschtmy initial reports from the flyspray say 1:20 v 2:30 and iirc it was *heavy* disk activity for the whole 2 minutes back then
23:21:02TorneHorscht: we use DMA now, though, so timings won't be comparable
23:22:25Horschtand now... i have to find a way to get my wps running again
23:24:12 Quit Rob2223 (Read error: Connection reset by peer)
23:25:51 Join Rob2222 [0] (
23:26:18gevaertsTorne: could this be somehow vaguely related to the extremely slow font loading issue?
23:26:36Tornegevaerts: quite probably
23:26:41Tornewell ok
23:27:22gevaertsNot very directly of course, but they're both about disk IO slowness on this drive
23:28:25 Join Om3ga [0] (
23:28:38gevaertshm, since the filesystem uses 2K blocks, *most* reads should be sector-aligned I suspect, so the partial sector thing shouldn't be needed very often
23:28:44 Quit Not-iH8sn0w (Ping timeout: 240 seconds)
23:28:59Om3gaproblem with my sansa fuze
23:29:12 Join Bug2000 [0] (~bug@unaffiliated/bug2000)
23:30:14gevaertsOm3ga: we'll need more details to be able to help :)
23:31:01 Join Dreamxtreme [0] (~Dreamxtre@
23:31:10 Join stephen__ [0] (~stephen@
23:31:22 Join evilnick [0] (
23:31:40Om3gamy sansa..
23:31:47Om3gascreen is blank
23:32:11Om3gathe screen would only appear after 5 seconds of pressing a button
23:32:21Om3gajust for a few seconds
23:32:32kugelTorne: doesn't the PPs have magic memory regions which are the uncached pendant of the normal region?
23:32:46stephen__pixelma, ping
23:33:04Tornekugel: only because the cache controller is set up that way
23:33:37kugelright, I'm thinking the mirror of the 32MB could be some uncached variant
23:33:43Om3gaproblem with my sansa fuze
23:33:47Om3gamy sansa..
23:33:50Om3gascreen is blank
23:33:54Om3gathe screen would only appear after 5 seconds of pressing a button
23:33:59Tornekugel: naw, we are just setting up the cache controller to think there is 64MB
23:34:21 Quit jae (Ping timeout: 255 seconds)
23:34:31gevaertskugel: I suspect an uncached codec buffer would be *quite* noticeable
23:34:49 Join Not-iH8sn0w [0] (
23:35:25gevaertsOm3ga: there's no need to repeat your question every two minutes. As soon as someone who can help sees it, they'll tell you
23:36:18Om3gai'll wait.
23:36:51 Join stripwax [0] (
23:37:38stephen__Om3ga, what were you doing that made the screen go blank ?
23:39:03 Join jae [0] (~jae@
23:39:56stephen__pixelma, (for the logs) I updated my comment on flyspray, you were correct with the revision numbers I got mixed up.
23:40:00Om3gai was charging my sansa with rockbox on it
23:40:08stephen__I also think it is to do with the remotes
23:40:11Om3gathen my screen just went out.
23:40:21gevaertsTorne: if I read ata.c correctly, phys_sector_mult!=1 just means that we have to be careful with aligning transfers, but actual sector sizes and thins like that are the same?
23:40:23stephen__have you held the power button for 30 secs ?
23:40:30stephen__Om3ga, ^^
23:40:33Om3gahmm, i think i tried that
23:40:43Om3gai'll try it again
23:40:46Tornegevaerts: er
23:40:50Om3gado it while its on ?
23:40:54Om3gait will turn on
23:40:58 Quit stripwax (Client Quit)
23:41:14gevaertsTorne: I mean for the ata registers and things like that
23:41:19Tornegevaerts: yes
23:41:21stephen__is it just the scrren thats blank ? does the click wheel light up ?
23:41:28Tornegevaerts: the interface both to rockbox and hte hardware is exactly the same
23:41:34Torneyou jhust have to be careful not to do stuff that's invalid
23:41:36 Join stripwax [0] (
23:41:37Om3gathe wheel lights up
23:41:49Om3gaits a fuze v2
23:41:58Om3ga4 gb
23:42:46stephen__hmmm not sure what you could try then, sounds like the screen is gone ? your best bet is to wait till funman arrives.
23:43:07Om3gaoh no, not completely
23:43:23Om3gaif my battery is 25% or lower, it will return
23:43:41Om3ga5 seconds after i press a button
23:43:47 Quit stripwax (Client Quit)
23:43:59Om3gaif my battery is low, my screen will work normally.
23:44:06stephen__is this just in rockbox ?
23:44:08 Quit huffpuff (Quit: leaving)
23:44:32stephen__in the of too ?
23:44:39Om3gaoriginal firmware
23:44:42stephen__hardware problem then by the sounds of it
23:44:57 Join evilnick_ [0] (
23:45:01Om3gayea, i installed rockbox on my v2
23:45:16Om3gathen charged my sansa with rockbox operating
23:45:17pixelmastephen__: thanks for the info
23:45:26 Join Zarggg [0] (
23:45:49stephen__just hope it gets sorted soon.
23:46:02stephen__wanted to get some x5 themes updated
23:46:03 Join huffpuff [0] (
23:46:42 Quit evilnick (Ping timeout: 258 seconds)
23:47:42gevaertsHorscht: is this 1:20 or two minutes *just* buffering, or does that include e.g. boot time, dircache scan, wps load, ...
23:47:42 Quit stephen__ (Quit: Leaving)
23:48:32 Quit BlakeJohnson86 (Ping timeout: 245 seconds)
23:48:32 Join stripwax [0] (
23:48:40 Join BlakeJohnson86 [0] (
23:48:43 Nick evilnick_ is now known as evilnick (
23:49:33 Quit jgarvey (Quit: Leaving)
23:50:28 Quit bertrik (Read error: Connection reset by peer)
23:50:46 Quit stripwax (Client Quit)
23:53:42 Quit domonoky (Read error: Connection reset by peer)
23:58:33 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
23:58:57 Join stripwax [0] (

Previous day | Next day