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

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

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

#rockbox log for 2009-10-25

00:00:16TheSeveni consider indenting them altogether ok. but not indenting them *between* the # and the ifdef
00:00:31saratogaIMO something like #IF \t \t CONDITION might be a nice compromise, so that the # is left aligned but the check is lined with the code
00:01:26TheSeven"# ifdef WHATEVER" looks very weird though
00:01:35TheSeven(and that's what I just found)
00:01:49saratoga#ifdef X>Y looks ok to me
00:02:04saratogasince you can find it easily while skimming, and then see what level it applies to
00:02:07 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
00:02:17 Join StealthyXIIGer [0] (
00:02:34TheSevenand in files like config.h where there are *only* preprocessor directives, I'd indent them like regular code
00:05:33SlasheriDavide-NYC: that might be bad/risky
00:06:30Davide-NYCwell I just flashed to the latest (SVN) BL without the patch and now I can only boot from ROM
00:06:30Slasheriit's a safety feature to boot from disk after having used the bootloader usb mode (if flash images are corrupt)
00:06:51Slasheriah yes, but try_flashboot should check the bootmethod before that
00:07:21Slasheriwell, it could be safe. But ideally the bootloader should be fixed to allow booting directly from CF
00:07:43bertrikTheSeven, do you configure the SDRAM controller for the nano 2g?
00:07:58Davide-NYCI agree. And I'm willing to be the test pilot (once). So if you'd like to help I'm here to test for the next 30 minutes or so.
00:08:09SlasheriDavide-NYC: that's great to hear that the svn bootloader is working
00:08:36TheSevenbertrik: No, I rely on norboot to set that up.
00:08:39Davide-NYCwell it's NOT working. Only ROM boots.
00:09:13Davide-NYCHD and RAM do not boot. In other words, SVN BL for iriver H1x0 targets is borked.
00:09:13Slasherioh.. does it give any error message when it tries to boot from the CF?
00:09:21Davide-NYCata -80
00:09:31Slasheriouch.. HD wont boot either?
00:09:45Davide-NYCI don;t have any HDDs anymore. (LOL)
00:10:06Davide-NYCI can however still boot from ROM and access the filesystem on the CFCard no problem.
00:10:21Slasheriah, ok. hmm, i have not much idea about CFs because my HD is still working :)
00:11:13SlasheriDavide-NYC: just remember to keep a working ram and rom image :)
00:11:47Davide-NYCSlasheri: Certain CF cards seem to work fine while others do not. The idea is to have the BL work with all CF cards on the market. My other (tiny) CF card give no problems.
00:11:55Slasheriso it says: "ATA error: -80"?
00:12:16Slasherihmm, interesting
00:12:26 Quit bmbl ("Bye!")
00:12:32Davide-NYCI'm going to reflash with the semi-official bootloader v7pre4
00:14:36Davide-NYCOh crap. The iriver flash utility just gave me a "NOT from ROM" error.
00:15:47SlasheriDavide-NYC: you have to boot using the RAM image
00:15:54Slasherithen it works
00:16:23Davide-NYCYeah, but with my new SVN bootloader RAM booting does not work!!!
00:16:39Slasheriouch, really? that's bad
00:16:42Davide-NYCOnly ROM. So now what?
00:16:56TheSevenDavide-NYC: Don't say you hadn't been warned. ;-)
00:17:06Davide-NYCI'm not saying anything of the sort.
00:17:06 Quit mt (Read error: 131 (Connection reset by peer))
00:17:09Davide-NYCI'm smiling.
00:17:19kugelamiconn: this gives me a slight speed up, but still slower than the asm
00:17:26Davide-NYCSee? :-)
00:17:35 Join mt [0] (n=mtee@rockbox/developer/mt)
00:17:35Slasheriwell, that's a little more difficult issue then..
00:17:48Slasheridid you try re-flashing the RAM image (rockbox.iriver)?
00:18:07Davide-NYCCan I do that if I boot from the ROM image?
00:18:34Slasheriah, yes! there is a way
00:18:44Slasheritry ROLO'ing directly to the rockbox.iriver
00:18:51Slasheri(short click navi on it)
00:21:04Slasheridid it work?
00:21:07Davide-NYCStrange, I was able to reflash the bootloader, but the same problem persists. I'm going to reflash the RAM image to see if that canges anything.
00:21:37Davide-NYCBoot to ROM −−> ROLO from CF −−> flash RAM −−> reboot
00:21:49Slasherihmm, interesting. maybe the ram image has become corrupted
00:22:07Davide-NYCThat's what I'm thinking
00:22:50Slasheri(but i don't think that should happen)
00:23:49Davide-NYCI also cleared all of my settings for good measure (twice)
00:24:00Slasherigreat! so the ram image stopped working immediately after flashing the new bootloader?
00:25:08Davide-NYCI had V7pre4 on there originally. CF booting to my RiData card didn't work. After flashing the BL to the current SVN image my RAM booting also failed to work.
00:25:30Davide-NYCI had today's SVN compiled with RTC flashed to both RAM and ROM.
00:25:37*Davide-NYC living on the edge!
00:26:28Slasheriok, i will have to check that. In fact re-flashing bootloader and testing ram/rom boot after that was missing from my test sequence. So it's very possible there was a bug
00:27:00SlasheriDavide-NYC: great :)
00:28:04Davide-NYCFor my other card (different brand) everything works as it should. Petur mentioned that it may be a timing issue with the different cards. (I have no clue what that actually means0
00:28:59 Quit reydarz ("CGI:IRC")
00:29:30Slasheriindeed. I was looking for ata error code -80 but couldn't find it yet, amiconn?
00:30:25Davide-NYCMaybe I just buy a Kingston Ultimate or a Sandisk Extreme III and call it a day.
00:31:03 Join T44 [0] (
00:31:26Davide-NYCOr better yet, the Kingston Elite Pro cards are even cheaper and I bet have no timing issues. Thoughts?
00:32:12*TheSeven found the difference between the slow and fast menus
00:35:33 Join stoffel [0] (
00:36:18saratogaI don't understand the coldfire libmad assembly
00:36:33saratogait looks like it doesn't use the same mdct algorithm as the arm assembly
00:38:14Davide-NYCSlasheri: I'm also noticing that my "HDD" light stays on my RiData unit but no on the other one.
00:40:13SlasheriDavide-NYC: i am just trying to find the correct error code.. but it looks rather hard, there may be several possible places because that -80 isn't very accurate
00:41:45Davide-NYCSlasheri: Please confirm that rockbox.iriver is the name of the bootloader. When I try to flash it from a ROLOd RB it asks if I want to reflash the RAM image.
00:41:59Slasheriok, it might be the set_features() function in ata.c where the problem occurs
00:42:11kugelSlasheri: -80 is possibly the 8th error that can happen in the storage driver (those return usually -N*20)
00:42:26saratogaunless i'm missing something the coldfire libmad mdct is much less efficient then the one used on ARM
00:42:59SlasheriDavide-NYC: that depends what target you are building
00:43:02 Join German [0] (
00:43:10Slasherinormally rockbox.iriver is the actual rockbox ram image
00:43:14Germanhallo =D
00:43:20Slasheribut when building a bootloader, it will be the bootloader
00:43:21saratogaunless maybe its unrolled somehow, it does seem to have roughly 2x the multiplications as the arm version
00:43:30 Nick German is now known as Guest44680 (
00:43:41Slasheri(you can determine that from the file size, bootloader is below 64 KiB)
00:43:56Davide-NYCGuys, when I compile a BL I configure (H120, BL) then I make. I end up with bootlader.bin, bootloader.elf, and rockbox.iriver.
00:44:24Slasheriyep, rockbox.iriver is the bootloader (you have to rename it to bootloader.iriver)
00:44:41 Quit stoffel (Remote closed the connection)
00:45:09AlexPSlasheri: Is there any reason that is is named rockbox.iriver and not directly bootloader.iriver, or is it just legacy?
00:45:27Davide-NYCSlasheri: quick commit!
00:46:00 Quit JackWinter (Read error: 131 (Connection reset by peer))
00:46:12SlasheriAlexP: it's just nobody has not "fixed" that in the Makefile yet :)
00:46:23AlexPSlasheri: I suspected as much :)
00:46:55Slasheribecause not so many people are building those bootloader either
00:47:05 Join JackWinter [0] (
00:47:05 Quit Guest44680 (Client Quit)
00:49:15AlexPSlasheri: Looks quite an easy fix actually, just testing
00:49:22 Quit JackWinter (Remote closed the connection)
00:49:28Davide-NYCOK, I had originally simply overwritten my RAM file with the newer bootloaders.
00:49:43 Quit Topy44 (Read error: 110 (Connection timed out))
00:49:59Davide-NYCI guess the name of the file is what the iriver_flash utility uses to determine what the user is trying to do.
00:50:26SlasheriAlexP: that sounds good :)
00:50:51Davide-NYCNow when I try to reflash the bootloader it says "Incompatible/Untested bootloader" and it does not let me flash the BL.
00:50:52SlasheriDavide-NYC: indeed, that would explain it then
00:51:14Davide-NYCDo I need to modify the iriver_flash utility?
00:51:30Slasheriyep.. you have to override the safety check from iriver_flash.c. How did you flash the SVN version just a moment ago?
00:51:42***Saving seen data "./dancer.seen"
00:52:17Davide-NYCI never did. I only *thought* that I had tested it when in fact I only borked the RAM image.
00:52:31Davide-NYCAm I nuts to even attempt this?
00:52:33Slasheriah :D
00:52:49Slasherilets hope it works, nobody knows
00:53:12Slasheriand, be sure to keep a working rom image..
00:53:35Slasheribootloaders have some simple safety mechanism to load than rom image if bootloader fails
00:53:39AlexPSlasheri: - if bootoutput isn't specified it uses output
00:53:53 Join Bob_C_ [0] (
00:54:42SlasheriAlexP: ah, that looks good. Did you test it? Just go and commit :)
00:54:43AlexPI've checked with and without that and rockbox.iriver and bootloader.iriver have identical md5sums
00:55:03Slasheriok, then it should be fine
00:55:34 Join JackWinter [0] (
00:56:02CIA-85New commit by alex (r23336): Change the name of the H100 bootloader output from rockbox.iriver to bootloader.iriver (as required by the iriver_flash plugin).
00:57:10Davide-NYC If I just comment out the "if (!detect_valid_bootloader(audiobuf, len))" section (to the endbrace) is that all I need to do in iriver_flash.c?
00:58:11 Quit elinenbe_ (Read error: 113 (No route to host))
00:58:23Slasheriyep, that should be enough (from the flash_bootloader-function)
00:59:55 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
00:59:58 Join StealthyXIIGer [0] (
01:00:41Davide-NYCNow I can actually test the SVN bootloader. Then (if my player's not bricked) I'll test the patched version.
01:01:17Davide-NYCmake reconf
01:01:17Davide-NYC?me doh! wrong window!
01:01:17kugelit reads "audiobuf" directly? nasty stuff ;/
01:02:16Slasherikugel: well, it uses audiobuf directly for other purposes (to load the flash image in memory for flashing)
01:02:48kugelit was probably made when we didn't have a useful buffering api
01:03:29Slasheriindeed. And anyway, the number of api calls should be minimized when flashing critical code (like bootloader)
01:04:32Slasheribest way would be to disabled interrupts entirely during the flashing process
01:06:05Davide-NYCme/ Wonders if there is a way to globally remove all games and demos from the build process. Like a "lean" build...
01:06:05*Davide-NYC learns the me switch
01:06:05kugelit's impressive which powers plugins actually (could) have when looking at these old plugins
01:06:53kugelDavide-NYC: you can do make `pwd`/apps/plugins/iriver_flash.rock
01:07:11kugelto only build that and no other plugin
01:07:27Davide-NYCkugel: thanks
01:08:28AlexPDavide-NYC: If you want to do it permanently, remove the ones you don't want (or comment them out) from /apps/plugins/SOURCES
01:09:39Davide-NYCAlexP: I know that method but I compile so infrequently that I'm usually redownling the source and my patches are broken. :-P
01:10:43AlexPThe build system seems to be checking for rockbox.iriver specifically, so is erroneously being red
01:11:03AlexPIs that something I have access to, or do I need Bagder / Zagor?
01:12:58TheSevengevaerts: did you see that weird beast build?
01:13:35Davide-NYCW00t! I just flashed to the SVN build and booted from CF card!
01:13:41kugelAlexP: I *think* the error-regex is somewhere in SVN, but I'm not sure
01:14:59AlexPI think I need to change /www/buildserver/builds
01:15:04SlasheriDavide-NYC: that's great! maybe it's a time for next bootloader release :)
01:15:10AlexPBut I'm a little wary of doing it blindly
01:15:23Davide-NYCRAM boots OK.
01:16:23Davide-NYCSlasheri: the real test is if USB diskmode exits gracefully.
01:16:24kugelAlexP: m68k:nozip:h100boot:iriver H100 - Boot:rockbox.iriver:−−target=h100,−−type=b:690 to m68k:nozip:h100boot:iriver H100 - Boot:bootloader.iriver:−−target=h100,−−type=b:690 looks like the way to go
01:16:36AlexPkugel: Yes, that's what I think
01:17:26AlexPI'll try :)
01:17:32kugelif it's not right then you discovered a useless field at least :)
01:17:40AlexPIf it still moans I'll revert and wait :)
01:18:13Davide-NYCI am confirming that USB disk mode exits gracefully to disk boot (default)!
01:18:28AlexPDavide-NYC: Is that with any changes to SVN?
01:18:42 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
01:18:46Davide-NYCno patch. Just whatever is in SVN
01:19:15CIA-85New commit by alex (r23337): Hopefully fix the buildserver to take account of the change of name of the H100 bootloader in r23336.
01:19:47Davide-NYCCan we get anyone else that has a CFModded H1x0 series player to test this new BL *before* we make it official?
01:20:45AlexPIs the build server in SVN live, or does it need manually updating like the website?
01:22:38 Quit bertrik (Read error: 113 (No route to host))
01:23:20 Quit Horscht ("Verlassend")
01:23:29mc2739AlexP: I thought it needed a manual update, but i could be wrong
01:23:30TheSevenAlexP: We'll see on the next commit
01:23:38 Quit Bob_C (Read error: 110 (Connection timed out))
01:23:41*TheSeven hunts for something to commit :-)
01:23:46AlexPThanks :)
01:24:01 Quit DerPapst ("Leaving.")
01:26:19 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
01:27:12 Quit Strife1989 ("My number of files is OVER 9000!")
01:27:14CIA-85New commit by theseven (r23338): Adjust the iPod Nano 2G CPU clock freq again, this time to an exact value. Also add hardware defines for the 9th DMA channel we discovered on the ...
01:27:44AlexPaha :)
01:27:48TheSevenlet's see what happens...
01:28:39 Quit faemir ("Leaving")
01:28:41 Quit ender` (" "There is no reason anyone would want a computer in their home." -- Ken Olson, president, chairman and founder of DEC")
01:29:47 Join elinenbe_ [0] (
01:30:41 Quit Horscht (Client Quit)
01:31:26AlexPah good :)
01:31:36 Join mt_ [0] (n=mtee@
01:32:22Davide-NYCMy HDD LED will not shut off. Is there a way to find out what is causing this?
01:34:56 Join faemir [0] (n=faemir@
01:35:34*TheSeven will have another shot at doing a nano2g battery bench tonight
01:38:19 Join reydarz [0] (
01:40:41pixelmamc2739: hope I didn't scare you off
01:41:28Davide-NYCnight all...
01:41:36 Quit Davide-NYC ("ChatZilla 0.9.85 [Firefox 3.5.3/20090824101458]")
01:42:08 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
01:42:12mc2739pixelma: not at all, just been busy with work
01:42:52amiconnsaratoga: The coldfire libmad mdct is very likely *more* efficient than the arm version
01:43:08amiconnThat is, if the mdct is the major contributor to decoding time
01:44:17 Quit faemir (Read error: 104 (Connection reset by peer))
01:44:32 Join faemir [0] (n=faemir@
01:44:41amiconnColdfire has emac, and it really pays off to make use of it
01:47:03mc2739pixelma: I will rework my patch after reading your comments. I'm not trying to supercede any work you have done, just trying to learn the manual stuff better.
01:48:22 Quit mt (Read error: 110 (Connection timed out))
01:49:41 Quit esperegu (Remote closed the connection)
01:52:18pixelmamc2739: just read your comment - the \nopt{scrollwheel,IRIVER_H10_PAD}{} is really there for complete "lines" in the finished table, so everything from there to the next \\ . It's decided there whether you need one line in the table (with the description "Scroll through albums / track list
01:53:03CIA-85New commit by mt (r23339): Add support for ID3v1 metadta tags in Real Media container.
01:53:13pixelma") or if you need two line because the targets use different buttons - one with the description "Scroll through albums" and the second line "Scroll track list" or something
01:53:29 Join saratogalab [0] (i=9803c264@gateway/web/freenode/x-tigowkcugvhspnup)
01:53:57saratogalabamiconn: coldfire mp3 is fast becaue the EMAC is fast, but the ASM code seems to do quite a few more multiplies per IMDCT then the ARM version
01:54:19pixelmamc2739: since I *believe* the M3 belongs to the latter, you wouldn't put the M3 in this \opt - but in a later one for the \opts in the columns
01:54:47amiconnsaratoga: Yes, and I'm saying that this is very likely okay, since reducing the number of multiplies there will probably make things slower
01:54:59saratogalabARM uses a fast algorithm that takes ~46 muls per 36 point IMDCT (lowest known count is about 43), whereas Coldfire seems to need more then 100
01:55:17mc2739pixelma: I see, and thats why that \\ was inside the {}, otherwise you would have a blank line at the top of the table
01:55:23pixelmamc2739: for excluding it from the first column and then \opting in the second
01:55:58AlexPIsn't it nopt where you can't have multiple things in one \nopt ?
01:56:54pixelmamc2739: yes, but it wouldn't be blank because the description would probably there
01:57:06mc2739oh, right
01:57:27pixelmaAlexP: there is some limits with \nopt but this kind of "adding" works
01:57:42AlexPI always forget and just check it :)
01:59:45pixelmait doesn't work if you want both or more things to be true then you would have to use nested nopts, if I remember correctly
01:59:58 Quit petur ("*plop*")
02:00:09*kugel wonders if pixelma ever saw my question/problem
02:00:23kugelabout the manual entry for the goto playlist feature
02:00:30pixelmacan't remember seeing it
02:00:52kugelit was 1 day before 3.4 or so
02:01:51pixelmawhat was it?
02:01:57kugelI don't know a way to properly opt that thing. some targets that don't have this feature due to lack of buttons shouldn't have the manual entry
02:02:57AlexPkugel: If there is no existing way of doing it, you could add it to features.txt
02:03:21kugelAlexP: that would mean introducing a #define for it (there's none) ;)
02:03:23pixelmaor \(n)opt it manually by *_PAD since that's the restrictive thing
02:03:40AlexPkugel: yep, but it is that or by each individual pad
02:03:53AlexPIf it is only a few pads, that'd be better I think
02:04:07kugelI basically would like to know if it's possible to make an opt from the existence of the following command "\newcommand{\ActionWpsPlaylist}{\ButtonRec}"
02:04:45AlexPkugel: Just do \opt{pads_with_lots_of_keys}{blah}
02:05:02AlexPor \nopt{pads_with_no_keys}{blah}
02:05:09AlexPdepending on which is the bigger group
02:05:38pixelmayes, you could add this command to the platform files that have it and use it like e.g \nopt{ONDIO_PAD}{\ActionWpsPlaylist}
02:06:08pixelmalater in the table or description
02:06:41kugelhrm, that would work, I more hoped though that something like \opt{\ActionWpsPlaylist} would be possible
02:06:49saratogalabmt_: can RM files have ID3v2 tags?
02:07:02kugeli.e. whether that command is defined or not
02:07:17pixelmakugel: I sometimes wished there was something like this too
02:07:31AlexPIt'd be nice, but no :(
02:08:09pixelmawould be useful for the ActionRC* in texts I think
02:08:14gevaertsTheSeven: yes. I've seen similar failures in the past. It seems to be caused by apps/features not being generated correctly, or at least I can get similar errors by tweaking that file
02:08:24gevaertsNo idea why this happens though
02:09:20pixelmamc2739: did you get what I wanted to say about using IAUDIO_RC_PAD vs. IAUDIO_M3_PAD in the second column parts?
02:09:58pixelmanot sure if I expressed myself clearly...
02:10:00kugelI asked google, and some pages suggest it was possible but I couldn't make it to work
02:10:06 Join Strife89 [0] (
02:10:17mc2739pixelma: yes, I understand that now, thanks for the explanation
02:11:22pixelmait will probably mostly apply to plugin button tables (currently)
02:13:56mt_saratogalab: According to this :, no. They have their own structuring for extended metadata tags however.
02:14:04 Nick mt_ is now known as mt (n=mtee@
02:14:05TheSevenis there really nobody who wants to give FS #10714 a shot?
02:14:15saratogalabmt: ok was wondering if it made sense to use the MP3 parser to save some space
02:14:32saratogalabbut if its ID3v1 only maybe thats more effort then 500 bytes is worth
02:14:51mtit's 128 bytes even :)
02:16:34*TheSeven 's battery benchmark is running now, so he'll go to bed
02:18:09*mt heads off
02:20:00 Join peter-b [0] (
02:20:14 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
02:20:37 Join StealthyXIIGer [0] (
02:26:04 Join petur [50] (n=petur@rockbox/developer/petur)
02:28:30 Quit peter__b (Read error: 145 (Connection timed out))
02:33:41reydarzI'm testing it now TheSeven
02:49:33 Quit Rondom (Nick collision from services.)
02:49:44 Join Rondom [0] (
02:51:46***Saving seen data "./dancer.seen"
02:56:50 Quit kugel (Remote closed the connection)
02:57:10 Quit petur (Read error: 110 (Connection timed out))
02:59:13 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
02:59:42 Quit GeekShadow (Read error: 104 (Connection reset by peer))
02:05:52 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
02:06:33 Quit amiconn (Nick collision from services.)
02:06:35 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
02:06:40 Quit saratogalab ("Page closed")
02:06:55 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
02:08:47 Quit pixelma (Nick collision from services.)
02:08:49 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
02:09:09 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
02:10:53 Quit mc2739 ("Leaving")
02:14:00 Quit Horscht (Nick collision from services.)
02:14:06 Nick Horschti is now known as Horscht (n=Horscht2@xbmc/user/horscht)
02:16:55 Quit GeekShado_ (Read error: 110 (Connection timed out))
02:26:25 Quit shaggy-h (Read error: 110 (Connection timed out))
02:33:38 Join mc2739 [0] (
02:34:10 Join peter__b [0] (n=peter_b@
02:34:58 Quit mc2739 (Client Quit)
02:35:28 Quit peter-b (Read error: 60 (Operation timed out))
02:37:14 Quit Lss (Read error: 54 (Connection reset by peer))
02:51:01 Join mc2739 [0] (n=mc2739@rockbox/developer/mc2739)
02:53:28 Join packetdeath [0] (
02:53:56packetdeathhey, have a dumb question
02:54:11mc2739packetdeath: go ahead and ask
02:54:41packetdeathhow would i go about backing up my ipod library before installing rockbox
02:55:17 Join _0x90u [0] (
02:55:57packetdeathHAW HACKE IPOD!
02:56:25 Quit packetdeath ("leaving")
03:06:21 Quit dmb (Read error: 104 (Connection reset by peer))
03:16:02 Quit dockimble ("Leaving")
03:16:55 Quit mc2739 ("leaving")
03:18:13 Join mc2739 [0] (n=mc2739@rockbox/developer/mc2739)
03:30:31Unhelpfulkugel: the only work-around i ever found for that problem was a pseudo-eabi build - a compiler that passes the flags for eabi to the assembler by default, but does not itself generate eabi code
03:31:30 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
03:31:52 Join StealthyXIIGer [0] (n=stealthy@
03:32:56 Quit adiroiban (Read error: 145 (Connection timed out))
03:51:48***Saving seen data "./dancer.seen"
03:59:52 Nick Ypsy is now known as YPSY (n=ypsy@
04:00:34 Join elinenbe [0] (
04:03:16 Quit TheSeven (Nick collision from services.)
04:03:37 Join The_Seven [0] (n=theseven@rockbox/developer/TheSeven)
04:03:49 Nick The_Seven is now known as TheSeven (n=theseven@rockbox/developer/TheSeven)
04:08:47 Quit Strife89 ("Bed.")
04:14:14 Quit StealthyXIIGer (Read error: 131 (Connection reset by peer))
04:14:25 Join StealthyXIIGer [0] (
04:18:02 Quit elinenbe_ (Read error: 113 (No route to host))
04:18:02 Nick elinenbe is now known as elinenbe_ (
04:33:46 Quit reydarz ("CGI:IRC")
04:48:57 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
04:49:19 Join StealthyXIIGer [0] (
04:52:03 Join reydarz [0] (
04:53:37 Quit reydarz (Client Quit)
05:11:51 Part toffe82
05:24:18 Quit efyx_ (Remote closed the connection)
05:44:50 Join ademille [0] (
05:46:39 Part ademille
05:50:46 Join Llorean [0] (
05:51:52***Saving seen data "./dancer.seen"
05:58:41 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
05:58:44 Join StealthyXIIGer [0] (
06:10:31 Quit Llorean ("Leaving.")
06:20:56 Join _0x90u[Nap|AFK] [0] (
06:23:59 Quit _0x90u (Read error: 60 (Operation timed out))
06:38:41 Quit panni_ (Read error: 60 (Operation timed out))
06:49:37CIA-85New commit by FlynDice (r23340): Add define for SD_SET_CLR_CARD_DETECT 42 /* acmd42 */ to sd.h for use with 4 bit WIDEBUS.
07:09:20 Quit alexbobp (Read error: 110 (Connection timed out))
07:20:05 Join ck2000 [0] (
07:21:06ck2000ok since i dont have an mp3player i use the simulators but i want an apple ipod 2nd gen nano
07:21:25ck2000where would i get 1?
07:21:53ck2000and i wish this would be ported to the ipod touch =)
07:23:00ck2000i am downloading kubuntu
07:23:32ck2000ok what is the programing language for this?
07:25:15ck2000alexP? you here?
07:25:49 Quit ck2000 ("CGI:IRC (EOF)")
07:29:56 Quit StealthyXIIGer (Read error: 110 (Connection timed out))
07:36:26 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
07:36:44topikrockbox is mostly C
07:37:00topikand try ebay for a 2nd gen nano
07:41:20 Join jfc^3 [0] (n=john@
07:46:49 Join Crackerizer [0] (
07:51:56***Saving seen data "./dancer.seen"
07:55:29 Quit Horscht (Read error: 110 (Connection timed out))
07:58:05 Join stoffel [0] (
08:01:13 Join robin0800 [0] (
08:05:34 Join mt_ [0] (n=mtee@
08:07:03JdGordonmt: hey, are rm really supposed to have id3 tags?
08:07:05 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
08:07:18JdGordonor is that stupid people doing non standard stuff?
08:13:20 Quit mt (Read error: 145 (Connection timed out))
08:21:33 Join mt [0] (n=mtee@rockbox/developer/mt)
08:24:04 Quit phanboy4 ("Leaving")
08:26:00 Quit mt_ (Read error: 60 (Operation timed out))
08:29:31 Join mt_ [0] (n=mtee@
08:33:05mt_JdGordon: :) .. No they could have id3v1 tags, see the end of the this page
08:34:02mt_I didn't knew about it until that guy on the forums put some samples with id3 tags
08:43:07 Join Rob2222 [0] (
08:43:21 Join bertrik [0] (
08:45:05 Quit mt (Read error: 113 (No route to host))
08:45:40bertrikThe meizu m6 framebuffer is located in system memory, do I still have to do something in lcd_update in that case?
08:50:28 Quit mt_ (Read error: 60 (Operation timed out))
08:50:42amiconnYou mean the lcd controller reads from that buffer by itself?
08:53:03bertrikyes, it does master DMA to get the data
08:53:44bertrikI vaguely remember that e200 does this as well, but this has a copy step, I'll check
08:54:36amiconnYes, that copy step is needed, because the application needs a way to decide which changes go to the lcd and when
08:55:01amiconnWell, unless you have a way to control the dma
08:55:05bertrikto prevent flicker?
08:56:35*amiconn still wonders whether nobody tried to find out how this dma is set up in the e200
08:57:12bertrikok, for the meizu m6 I plan to commit it without the extra framebuffer for now, I'll add this issue in a comment in the top of the file
08:57:32 Join n1s [0] (n=n1s@rockbox/developer/n1s)
08:57:44amiconnhmm :\
09:01:03 Quit Rob2223 (Read error: 110 (Connection timed out))
09:01:20 Quit BHSPitLappy (Remote closed the connection)
09:02:11markunbertrik: did you get the display working?
09:02:26bertrikmarkun, more or less
09:02:42tmzts3c framebuffer?
09:02:47bertrikthe LCD is in portrait mode and there's sometimes a line at the bottom belonging to the top of the screen
09:03:21bertrikthe 16-bit colors work out fine (I was afraid it could only handle 32-bit pixels)
09:03:27tmztthe timing is wrong
09:03:45bertrikI don't know that yet
09:03:55markunbertrik: will you add the other LCD inits from the SP and SL firmwares as well?
09:04:16tmztthat's the only problem with it?
09:04:22 Quit bluebrother (Read error: 104 (Connection reset by peer))
09:04:31bertrikmarkun, what do you mean?
09:04:43tmzttry changing all the values (vertical) by one one at a time
09:04:43 Join bluebrother [0] (n=dom@
09:05:20bertriksometimes is not just a single line. I've seen once that the bottom contained a kind of vertically mirrored version of the top
09:05:57tmzthave you located all the lcdc writes?
09:06:08markunbertrik: I mean the values that are written to the LCD driver over SPI, I assume you had to copy them into rockbox, right?
09:08:18bertrikthe m6 sp driver basically contains code for two separate LCDs now, I copied the controller init, lcd init/enable etc. from a reverse engineered OF
09:09:50bertrikthe weird thing is that one of the two meizu m6 drivers cannot even work in an m6 sp, because it selects the LCD driver based on an ID read back from the LCD, and the pin to read this is not even connected
09:10:19bertrikI haven't really looked at the meizu m6 sl driver yet
09:11:46bertriktmzt, yes I think I've located most of them. The OF does seem to make much more use of the special features of this LCD controller (OSD overlay for example). I'm currently only using the "background buffer"
09:12:05tmztwhat chip?
09:12:37tmztlcd controller/lcdc I mean the framebuffer registers, not lcm or driver
09:12:37 Quit bluebrother (Read error: 131 (Connection reset by peer))
09:12:56 Join bluebrother [0] (
09:13:57bertrikthe LCD controller is part of the s5l8700 SoC and handles the framebuffer part (for example conversion of packed pixels into 24-bit RGB towards the LCD), the LCD is probably a samsung S6DXXXX type
09:15:06tmztso those are the registers I mean, the type of corruption you seem to be describing is related to that
09:15:16bertrikoh and one giant issue is that the player seems to hang after initialising the display .. :\
09:15:48tmztI believe incorrect pixclock would have different symtoms depending on lcm
09:16:12bertrikwhat do you mean by lcm?
09:16:12tmztokay, that's probably dma or EN bits of some kind
09:16:23tmztlcd control module
09:16:40tmztthe samsung s6dxxxx you were referring to
09:16:57tmztis that the part number of the lcd?
09:19:57bertrikyes I think so, the protocol for talking with the lcd (to set the gamma curve for example) is identical to the protocol used in the s6d0139
09:27:32tmztkug: I think your the one to ask, ams sansa has a two stage rockbox bootloader?
09:29:59bertrikmarkun, the "aha" moment to get this driver working was the realization that the framebuffer was located in DRAM and DRAM access has not been properly set up yet ... :)
09:31:03bertrikso setting up DRAM properly is probably the next thing I'll work on
09:31:54tmztmissing clock?
09:32:48tmztare you working on bootloader now?
09:36:59bertrikyes, I'm compiling it as a bootloader and writing it into the internal NOR flash
09:37:36 Join esperegu [0] (n=quassel@
09:38:05tmztso you need full dram init including timings?
09:38:16tmztor just an enable bit and the proper clock?
09:39:02bertrikI don't know, I think I need both
09:39:25bertrikI guess this is probably easy to locate in the OF
09:39:36tmztthere should be uboot patches for that generation
09:40:36bertrikthe SDRAM is a K4M28163PF
09:41:31tmztand the controller?
09:41:46tmztis it in the package with flash?
09:42:11bertrikyes, the SDRAM controller is part of the s5l8700 SoC
09:42:55tmztwhat is this soc? it's for mp3 players or it's more general purpose like the s3c/s5c
09:47:12bertrikyes, it seems to be aimed at mp3 players (it has IIS and a unique chip id for example)
09:47:48 Join liar [0] (n=liar@
09:49:22tmztfor drm?
09:49:44bertrikI think so
09:51:18tmztwhat arm core?
09:51:59***Saving seen data "./dancer.seen"
09:52:16tmztwhich is what isa
09:52:30bertrik+ CalmADM2E and a CalmRisc16
09:52:53tmztinteresting, never heard of those
09:58:07 Join ender` [0] (
10:00:35 Quit fyrestorm (Read error: 54 (Connection reset by peer))
10:00:53 Join fyrestorm [0] (
10:09:21 Join |DaMaGeD| [0] (n=|DaMaGeD@
10:09:24TheSeventmzt: samsung proprietary stuff
10:09:45TheSevenand it's an ARMv4T, I think
10:12:23TheSevenARM9TDMI core, which is ARMv4T
10:13:26TheSevenhey my battery bench is running fine, we've passed 8 hours now and it only ate 2 columns of pixels on the battery icon in the status bar
10:13:44TheSevenso i won't be able to play with it today, i think :-/
10:15:39 Quit _0x90u[Nap|AFK] (Read error: 113 (No route to host))
10:17:48 Join DerPapst [0] (
10:31:00 Join skyhunter [0] (n=quassel@
10:34:24bertrikmarkun, do you know how the left/right buttons on the meizu are connected? possibly through the touch strip?
10:39:46 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
10:54:25CIA-85New commit by bertrik (r23341): power-meizu.c: fix usage of USB_TR name
10:54:53n1sam i just blind or does talked_tick in main.c:tagcache_init() not need to be static (with the init moved out of the loop of course) ?
10:58:11bertrikn1s, that does indeed look silly
11:00:36bertriksvn blame says you last touched that line :)
11:00:36 Quit fyrestorm (Read error: 104 (Connection reset by peer))
11:00:58bertrikit was a patch from FS #6159
11:02:57 Join fyrestorm [0] (
11:03:41n1sbertrik: heh
11:08:51n1sok, i'll fix then
11:20:17CIA-85New commit by nls (r23342): No need for this to be static
11:22:42markunbertrik: on the SL it's connected to GPIO pins, on the SP probably the touchpad driver
11:24:40markunbertrik: I don't have the M6SP schematics with me right now, can you see it there?
11:27:26 Nick Horschti is now known as Horscht (n=Horscht2@xbmc/user/horscht)
11:28:02bertrikmarkun, no, I don't see any left/right or fwd/rev buttons on the m6 sp schematic
11:29:14bertrikit shouldn't be too hard to figure out if they are indeed connected through the touchpad once the display driver is fully stable
11:37:21markunbertrik: yes, like we did for the m3
11:37:43markunwhen we found out that the <> button is touch sensitive :)
11:41:30 Join adiroiban [0] (
11:42:23CIA-85New commit by nls (r23343): Little bit of const correctness
11:52:00***Saving seen data "./dancer.seen"
11:56:48 Quit gevaerts (Nick collision from services.)
11:57:00 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
11:59:55 Quit |DaMaGeD| (Read error: 131 (Connection reset by peer))
12:01:48 Join funman [0] (n=fun@rockbox/developer/funman)
12:04:00CIA-85New commit by funman (r23344): buffering: leave a comment to explain what is broken in the code and link to FS #10605
12:23:40bertrikTheSeven, I don't understand the s5l8700 file. Why is the bss data moved by 12 MB?
12:24:46TheSevenprobably because the firmware will be loaded to the addresses below
12:25:37 Join ChelseaToner [0] (
12:25:57TheSevenno idea why linuxstb chose 12MB for that, because an apple firmware is around ~7MB, and one could have chosen an even thing like 16MB
12:27:22ChelseaTonerHi there. I'm confused about installing Rockbox onto the Ipod Nano 2nd Generation. It says in the PDF that it can be done manually, yet when I went to the page to get the .zip, there was no option for the 2nd generation.
12:27:47TheSevenwhich page?
12:28:11TheSeventhe zip is at:
12:28:39ChelseaTonerThank you very much. I'll try this.
12:29:30bertrikI will need to work around that for the meizus
12:29:43TheSevenbertrik: why?
12:30:30TheSevenah right you only have 16MB of DRAM...
12:31:15CIA-85New commit by bertrik (r23345): Meizu M6 SP: update lcd driver, it is now possible to display a rockbox logo (with some limitations, see comments in lcd-m6sp.c)
12:31:54 Quit ChelseaToner ("Leaving")
12:33:07bertrikWhy would you have to skip over the apple firmware?
12:34:13TheSevenbecause we need to use the bss while loading the apple firmware there
12:43:32 Quit Zarggg (Read error: 110 (Connection timed out))
12:55:30linuxstbTheSeven: I chose 12MB so the same value could be used for targets with 16MB RAM
12:57:10linuxstbbertrik: What needs changing for the Meizu?
12:58:42linuxstbmt: (if you read the logs) Why not use the existing id3v1 parser for rm?
13:01:30 Quit antil33t (Read error: 104 (Connection reset by peer))
13:01:35 Join antil33t [0] (n=Mudkips@
13:04:26 Quit Horscht ("Verlassend")
13:13:26 Join mitk [0] (
13:18:06mitkHi. I prepared patch introducing PLAYLIST_INSERT_LAST_SHUFFLED functionality. Can one who can commit take a look on FS #10717?
13:19:12 Quit JackWinter (Remote closed the connection)
13:19:59 Join JackWinter [0] (
13:26:53 Join fdinel [0] (
13:29:54funmansaratogo: mt: before starting any work on the Clip+ we need a copy of the OF. The 2 possibilities are : wait for the first firmware upgrade for Clip+, or try to put it in recovery mode just like the e200v2 so we can read the OF directly from the device
13:30:28funmanor 3/ unsolder the nand chip and read it with a custom device
13:31:22 Join teru [0] (
13:32:37 Quit funman ("free(random());")
13:39:23 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
13:39:50 Join efyx_ [0] (
13:42:52 Quit Crackerizer ("CGI:IRC (Ping timeout)")
13:43:28 Join DerPapst1 [0] (
13:52:02***Saving seen data "./dancer.seen"
13:54:13bertriklinuxstb, but don't we waste 3/4 of SDRAM that way?
13:54:23 Quit DerPapst (Read error: 110 (Connection timed out))
13:54:47 Nick YPSY is now known as Ypsy (n=ypsy@
13:55:12bertrikoh, it's just the bootloader
13:55:29bertrik4 MB should be enough for that I guess
13:55:59 Join Crackerizer [0] (
13:58:48 Quit skyhunter (Remote closed the connection)
14:00:16 Join ALdbeign [0] (
14:01:41ALdbeigni have a potentially stupid question if anyone is around to answer it :D
14:02:12TheSeventhen just ask it
14:02:41ALdbeignis a device (such as an ipod) required to use rockbox as a player?
14:03:11TheSevenwell, there are simulators :-)
14:03:56ALdbeignhmm...ok thanks, more work than i intend to go through for what im looking for
14:04:03ALdbeignthanks for your help though :)
14:04:26*ALdbeign waves
14:04:37 Part ALdbeign
14:12:47*TheSeven is going to get a 4g btw
14:14:32pixelmagreyscale or Color/Photo or something entirely different? ;)
14:16:34TheSeven4g nano
14:17:02n1sTheSeven: \o/
14:17:49TheSevenwell, I'm a little concerned that this may not be too good for the nano2g port
14:20:59 Join phanboy4 [0] (
14:22:04 Join elinenbe [0] (
14:32:02*bertrik tries to configure meizu m6 as a target without storage at all to prevent compile errors
14:40:26CIA-85New commit by teru (r23346): Correct Id keyword and add svn:keywords property.
14:41:37bertrikWhat was the policy again on svn:eol-style properties?
14:42:59 Quit elinenbe_ (Read error: 113 (No route to host))
14:42:59 Nick elinenbe is now known as elinenbe_ (
14:52:48CIA-85New commit by bertrik (r23347): Meizu M6 SP: make bootloader compile and run again ...
14:55:49 Quit mitk ("CGI:IRC")
15:03:45CIA-85New commit by bertrik (r23348): Fix ipod nano 2g bootloader link problem, caused by my previous commit
15:08:59 Quit JackWinter (Remote closed the connection)
15:11:20 Join JackWinter [0] (
15:15:04n1sbertrik: svn:eol-style native i think
15:15:29n1sTheSeven: that's mostly working now anyway ;)
15:15:46TheSevenn1s: I don't think so
15:16:07TheSevenit's obviously still screwing people's FTLs and there's still quite a bunch of things to fix
15:16:20TheSevenespecially the very slow reaction of the menus, which is really annoying
15:17:06 Join mt [0] (n=mtee@rockbox/developer/mt)
15:18:14mtlinuxstb: I just looked into metadata_common.c and how metadata/ape.c implemented id3 parsing.
15:25:12 Join flydutch [0] (n=flydutch@
15:25:35 Quit efyx_ (Remote closed the connection)
15:28:27 Quit JackWinter (Remote closed the connection)
15:29:13 Join JackWinter [0] (
15:39:09 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
15:39:22 Quit liar (Read error: 148 (No route to host))
15:40:50 Quit stoffel (Read error: 113 (No route to host))
15:45:56 Quit Crackerizer ("CGI:IRC")
15:52:05***Saving seen data "./dancer.seen"
15:59:34 Quit teru ("Quit")
16:04:05 Join rvvs89_ [0] (n=ivo@pdpc/supporter/base/rvvs89)
16:04:36elinenbe_TheSeven: quick question... what's the status with the 4G nano?
16:04:45 Quit rvvs89 (Nick collision from services.)
16:05:15 Nick rvvs89_ is now known as rvvs89 (n=ivo@pdpc/supporter/base/rvvs89)
16:05:35 Nick elinenbe_ is now known as elinenbe (
16:06:21TheSevenwe're trying to work out a way to dump the bootrom
16:27:33 Nick Ypsy is now known as YPSY (n=ypsy@
16:30:00 Join petur [50] (n=petur@rockbox/developer/petur)
16:30:10linuxstbbertrik: Yes, the bootloader needs about 300KB BSS (but that could probably be trimmed a lot if it ever needed to be)
16:33:23linuxstbmt: I can't see any id3 parsing in metadata/ape.c. You should just be able to call setid3v1title() (in apps/metadata/mp3.c) - that is taking account of the current codepage, which I don't think your version is doing.
16:39:45 Join reydarz [0] (
16:40:46 Quit petur ("later")
16:40:55 Join panni_ [0] (
16:42:02*bertrik is slightly confused by the polarity of the hold button in ams sansa hardware
16:42:41 Join gtkspert [0] (
16:50:00 Join stoffel [0] (
16:51:03 Quit reydarz ("CGI:IRC (EOF)")
16:56:19 Join StealthyXIIGer [0] (n=stealthy@
16:58:04 Quit gtkspert_ (Read error: 101 (Network is unreachable))
17:04:16bertrikideally I should have a fuze, e200v2 and a c200v2 to really make the button reading robust ...
17:05:56 Quit Horscht ("Verlassend")
17:13:10 Join toffe82 [0] (
17:13:42mtlinuxstb: stupid question, should I just copy setid3v1title ?
17:16:28FlynDiceDoe anyone interested in the AMS Sansas have any concerns about changing the IDECLK to 62 MHz under the assumption that MCLK == IDECLK for the internal SD card as I posted here:
17:16:38FlynDiceer does...
17:19:00 Quit DerPapst1 ("Leaving.")
17:20:14bertrikI don't see a problem with that, what is the current value?
17:21:14FlynDiceright now we're doing 82 MHz
17:22:18bertrikFlynDice, I could test stuff on a clip (internal only) and a c200v2 (internal + uSD) if you want
17:24:12FlynDiceI've tested on my e280v2 but perhaps a clip test would be prudent. Pastie coming up
17:28:25 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
17:29:07bertriksorry, I like to test things, but I don't have much time right now
17:31:42FlynDicebertrik: No problem , thought you wanted to do it, I'm not too concerned, it should be fine
17:39:20 Quit StealthyXIIGer (Read error: 131 (Connection reset by peer))
17:39:34 Join StealthyXIIGer [0] (
17:52:07***Saving seen data "./dancer.seen"
17:55:48toffe82bertrik: for the meizu, to install rockbox how do you do ?
18:04:50 Nick uflops is now known as mikroflops (
18:04:58 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
18:05:07 Join StealthyXIIGer [0] (
18:05:15TheSevenmy nano 2g is running for 17 hours now, and the battery icon is still at around 2/3
18:05:50TheSeven(i know this is badly calibrated and especially the lower end of the scale is far off, but i'd nevertheless say that it's around half full)
18:07:01 Join petur [50] (n=petur@rockbox/developer/petur)
18:12:21mtShould a clip+ page go under "Sandisk" or under "Sandisk->SansaClip"
18:13:14 Join liar [0] (
18:13:31mc2739clipv2 is on so maybe that would be a good place for clip+ also
18:14:26 Join Tuplanol1a [0] (
18:18:12 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
18:20:00 Join kugel [0] (n=kugel@rockbox/developer/kugel)
18:21:56kugelbertrik: what do you mean by "different pre-charge level"?
18:23:24 Join froggyman [0] (
18:28:49 Quit stoffel (Read error: 113 (No route to host))
18:31:28bertrikkugel, I mean the the value you write to the DBOP, before doing the DBOP read, so for example 0xE0FF instead of 0xF0FF (where bit 12 for the hold button is different)
18:32:20bertrikI'm having trouble to really understand how the buttons are connected to the DBOP lines in the various ams sansas
18:32:31kugelI tried that, and it didn't work. without precharging it would always detect whether it's pressed or not
18:32:48*Unhelpful takes another stab at building gcc-4.4.2 :/
18:33:03bertrikOn the c200v2 for example, all buttons seem to pull down a line when pressed, but this is different apparently for the hold button on the e200v2/fuze
18:33:07bertrikkugel, so it always detected hold, or never detected hold?
18:33:09kugelUnhelpful: did you have the unwind exceptions thing also?
18:33:20Unhelpfulafter disabling libiberty, and patching out the direct dep on libiberty.a for fixincludes
18:33:28kugelbertrik: it things it was pressed if it wasn't
18:34:29kugelbertrik: see my very first comment on that task
18:34:50kugelUnhelpful: was that a yes?
18:34:56 Join Zarggg [0] (
18:36:20Unhelpfulkugel: yes, i did. one way to address it is by building a non-eabi compiler that passes the eabi flags to the assembler. this will create non-standard objects, not that we care, which will have non-eabi struct and calling conventions, but which have the eabi-style jump relocations which will be stubbed if needed at link time.
18:36:23 Join Zarggg_ [0] (
18:36:26 Quit Zarggg (Connection reset by peer)
18:36:26 Quit Zarggg_ (Read error: 54 (Connection reset by peer))
18:37:43kugelUnhelpful: do you think it's a bug or did we miss something? unfortunately nobody on the gcc ml answered yet
18:38:11Unhelpfuli think it has to do with things that eabi expects of the environment in which code is executed
18:38:25 Join Zarggg [0] (
18:38:59kugelIIUC this unwind stuff is only needed for non-C languages which can do advanced exception handling
18:40:06kugelI can live with that I have to have stub sections for extab/exidx, but not that unwind makes require linking against a C lib implementation
18:41:27kugelI haven't tried patching gcc to #if 0 the unwind related code though
18:41:58Unhelpfuli was discarding extab/exidx. i *believe* it was a div0 handler that pulled in the unwind stuff. i don't have the depth of knowledge to figure out where to stub out a handler to avoid the unwinding...
18:43:04TheSevendoes anybody know what the "synopsys otg" is?
18:43:52kugelUnhelpful: a div0 handler?
18:44:25bertriksynopsis designs logic circuits, so the synopsis otg is an USB otg controller designed by synopsis. I looked for more information on it but I couldn't find it.
18:44:36Unhelpfuli *think* that was what it was. it was a while ago. :P
18:48:03 Quit robin0800 (Remote closed the connection)
18:49:15linuxstbmt: Why would you want to copy setid3v1title? Can't you just call it directly? (after making it non-static of course)
18:49:49bertrikkugel, yeah that's the weird thing about the hold button I don't understand: it seems that if you precharge the hold data line with 0, and wait only a short while it reads back as 1, but if you wait longer it reads back as 0 again ... :S
18:50:22mtlinuxstb: Yeah ok .. didn't want to modify mp3.c :)
18:50:23 Join einhirn [0] (
18:50:33n1slinuxstb, mt: maybe it should move into the metadata_common.c file if used by several parsers
18:50:41kugelbertrik: I haven't noted that behavior
18:50:49mtn1s: Was just going to ask about that
18:51:11kugelor a id3v1.c :)
18:52:42Unhelpfulyou can try looking in the mapfile to see what you need to fix...
18:52:50 Join _0x90u [0] (
18:56:07kugelUnhelpful: was that for me?
18:57:14mtlinuxstb: So what do you think about moving setid3v1title to metadata_common ?
18:58:12linuxstbmt: I don't think it fits there. I would just leave it in mp3.c, as that's where all the id3 parsing code lives.
18:58:31Unhelpfulkugel: indeed
18:58:42linuxstbAlthough that could be split into mpeg audio parsing and id3 parsing, but that doesn't seem important...
18:58:43kugelbertrik: I'm not sure about the wrong detection of hold with 0xE0FF anymore actually, it's very possible that it just didn't work (i.e. never detected hold)
18:59:23kugelI rather think it was this way since your initial patch caused the false positives IIRC
18:59:36kugelUnhelpful: what map file?#
18:59:52kugelalso, you're having a go right now so I won't do anything :P
19:00:35saratogais there some script that parses the test_codec output files and converts them or did everyone make those tables by hand?
19:00:43mtlinuxstb: It's ok to put its prototype in metadata_common.h though right ? :)
19:01:07 Join stoffel [0] (
19:01:33linuxstbmt: Maybe metadata_parsers.h - that's where other functions from mp3.c are exported - e.g. getid3v2len()
19:01:57linuxstbipods are not mp3 players?
19:02:43Bagder"If you get stuck you'll find a manual specifically written for your device on the site with full instructions, an impressive documentation effort."
19:02:51Bagdera fine quote
19:03:13n1slinuxstb: no! they are so much more! thay can... uh, have notes on them!
19:03:49Unhelpfulkugel: i'm still trying to compile gcc-4.4.2... i think it's time to give up :)
19:04:05linuxstb"Rockbox can be extended through plugins and many of these are available on the site" is a bit misleading though.
19:04:10n1swhich btw was a nice thing added by apple but for entirely different reasons :)
19:04:21bertrikyay, we just got ogg support \o/
19:05:21linuxstb"Sadly it doesn't work on newer iPods, as changes to the newer products have meant that you need iTunes, which is still not available for Linux" is also a bit odd...
19:05:56 Join quidam [0] (
19:06:08quidamhi, quick question
19:06:41quidamis there a way to remove the original apple firmware from my ipod and use just rockbox?
19:07:35kugelUnhelpful: I wonder if it's worth trying a 4.3.0 snapshot
19:08:36krazykitquidam, no, not completely.
19:08:47linuxstbquidam: There are two parts to the Apple firmware - the part stored in flash ROM, and the part on disk. What do you want to remove, and why?
19:09:19quidamI'd like to get rid of any non-free software piece
19:11:57CIA-85New commit by mt (r23349): Revert r23339 and use setid3v1title() from metadata/mp3.c instead.
19:12:23quidamlinuxstb: can I remove at least the part on disk?
19:14:26 Join reydarz [0] (
19:15:38mtFrom ClipV2's components, I can identify the different chips on the clip+'s board, but how do I double check ? searched but couldn't find something useful.
19:15:58 Quit Bob_C_ (Read error: 104 (Connection reset by peer))
19:17:27 Join Bob_C [0] (
19:17:32mt(I've searched using the chip markings btw)
19:19:11 Join tomers [0] (
19:21:27linuxstbquidam: You can if you want to, but there's no real point - it's not run unless you select it. See for various options.
19:21:27 Quit reydarz ("CGI:IRC (EOF)")
19:22:10 Join tresh [0] (n=tresh@
19:23:42quidamlinuxstb: thank you, that is interesting
19:23:45 Quit stoffel (Remote closed the connection)
19:25:48 Join liar|netbook [0] (n=liar@
19:26:09 Quit Bob_C (Read error: 104 (Connection reset by peer))
19:26:27 Quit liar|netbook (Read error: 131 (Connection reset by peer))
19:28:22treshHi. I just dropped by to ask if i can install rockbox on my sansa e260v2 using the automated way of installation. The e200v2 installation entry and the wiki are kinda contradicting. Right now i tried the current stable version of rockbox utility and that seems to be unable to flash e200v2 players. tia
19:28:36 Quit liar (Read error: 110 (Connection timed out))
19:30:06 Quit einhirn (Read error: 104 (Connection reset by peer))
19:30:07 Join TopyMobile [0] (
19:30:16JdGordonkugel: seen 10721?
19:31:46CIA-85New commit by FlynDice (r23350): AMS Sansa: Assume IDECLK is MCLK for the internal SD Disk. Reduce IDECLK to 62 MHz for now to be consistent with MCLK for uSD which is PCLK. ...
19:34:27kugelJdGordon: yes, now :)
19:34:49JdGordonlooks like the custom sb broke it :)
19:40:15 Join Lss [0] (
19:46:44kugelJdGordon: I wonder how :/
19:48:08JdGordonyeah, no clue either :p
19:48:36JdGordonmaybe its the same thing that I was asking about where the default vp gets set to 0,0,0,0? although if other text is being displayed probably not
19:49:19 Part quidam ("WeeChat 0.2.6")
19:52:08***Saving seen data "./dancer.seen"
19:52:28kugelJdGordon: it seems to try to load a "-.bmp"
19:54:44*JdGordon builds a h100 sim to have a looky
19:56:38JdGordonyep... tris reading -.bmp!
19:57:42JdGordonhmm... umm
19:58:05JdGordonit doesnt if you manually load the boxes.wps
19:58:39*kugel didn't touch %pb parsing
19:58:54JdGordonand the parsing code still looks fine
20:03:14pixelmathe fallback default WPS doesn't seem to display a progressbar either
20:04:21JdGordone200 also
20:04:26JdGordonso presumably all..
20:04:39pixelmac200 here
20:06:17 Quit TopyMobile ("Ex-Chat")
20:06:29 Join einhirn [0] (
20:06:34 Quit StealthyXIIGer (Read error: 104 (Connection reset by peer))
20:07:02 Join StealthyXIIGer [0] (
20:08:34kugelJdGordon: if %pb||....| works it might be that this "-" slips in from somewhere
20:08:48JdGordonthat shouldnt work
20:09:16JdGordonoh, that does work
20:09:51JdGordonI'm just looking at the code
20:10:49 Join TopyMobile [0] (
20:14:18 Quit faemir ("Leaving")
20:15:32 Join pamaury [0] (
20:16:51kugelI think I know the problem
20:17:51kugelboxes.wps %pb is different due my changes (probably the ones)
20:18:48kugelso if plain %pb was ever supposed to work it was likely broken before
20:19:52pixelmait worked after the last big %pb rework
20:20:14pixelmaafter some fixes, if I recall correctly
20:20:20kugelarg that wouldn't explain the builtin wps breakage
20:24:09 Quit pamaury ("exit(*(int *)0 / 0);")
20:26:06 Quit Rondom (Nick collision from services.)
20:26:09pixelmacould it be something with bitmap or no bitmap handling (or the parameter). I can confirm that non bitmap progressbars still work if they have all the parameters set (%pb|-|...)
20:26:17 Join Rondom [0] (n=Rondom@
20:30:18 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
20:34:02bertrikFlynDice, I've been playing audio on my sansa clip since your ide clock update and I've seen no weird effects yet
20:34:45bertrikBTW, did saratoga's clip stability work already go in? I haven't had a crash yet ...
20:34:56FlynDicebertrik: Yes, it should be fine, we're still overclocking but we're overclocking less.... Thanks for letting me know ;)
20:35:17FlynDiceI don't think so
20:35:39bertrikoh maybe I just got lucky then
20:36:21 Quit n1s ("Lämnar")
20:36:47saratogabertrik: its mostly matsch's work
20:36:59saratogahe found the bug somehow
20:37:12saratogaso far all thats gone in is a comment by funman today explaining the bug
20:37:31pixelmakugel: or the fallbacks are off if the parameters ar not set? Like e.g. zero height or so that it is just invisible
20:41:20pixelmaI already get a weird result if I replace the height with - in my WPS...
20:42:47pixelmathe progressbar is a lot higher than it should be (that should be dependent on font size IIRC or at least something reasonable like 7 pixels or so
20:45:38pixelmahmm... no. It looks correctly was bislead by the fact that the viewport was limiting the size
20:46:10 Join liar [0] (n=liar@
20:46:34saratogahuh is it just me or has wavpack decoding gotten slower on arm
20:47:01amiconn[00:32:19] -*- TheSeven found the difference between the slow and fast menus <== What is it?
20:48:16TheSeveneverything that goes through tree.c seems to be slow
20:48:35TheSeveni.e. everything that goes through browse() in root_menu.c or whatever it's called
20:48:43pixelmakugel: oh, but it disappears if I replace the width with "-". This is with a test WPS now that only contains %pb|1|3|-|-| not even disabling the statusbar. If I put a number into the width field the progressbar shows up
20:48:47JdGordonbecause of disk access?
20:49:05pixelmakugel: no viewports defined in that WPS either
20:49:52kugelpixelma: what about the filename argument?
20:49:58saratogaaccroding to test_codec, wavpack has actually gotten 10% slower
20:49:58kugel%pb|1|3|-|-| lacks one
20:50:10TheSevenno idea what actually slows it though
20:50:44JdGordonkugel: that one should fail the parse.... or 1 with be the filename
20:50:49pixelmakugel: sorry typoed %pb|-|1|3|-|-|
20:51:04JdGordoneither way... %pb should be working :)
20:51:12kugeldoes omitting parameters work the same way as for viewports?
20:51:23JdGordonthey get given default
20:51:32*kugel remembers some line height magic
20:51:46JdGordonthats only with %pb though
20:51:51JdGordonnot if it has any params
20:51:52JdGordoni tihnk
20:51:54pixelmawell height seems to work but width doesn't
20:52:19JdGordonno.. im wrong.. it happens if the y position is -
20:52:56pixelma%pb|-|1|3|130|-| works so I assume I'm allowed to ommit parameters, and I'm already allowed to do so for the bmp
20:53:34pixelmaI'm convinced it worked once
20:53:53kugelJdGordon: what happens with y="-"?
20:54:07JdGordonpb->y = -line_num - 1; /* Will be computed during the rendering */
20:54:41JdGordonI want to say I dont tihnk the default viewport is getting initialised corectly... but that doesnt seem to make sense
20:58:28 Quit phanboy4 (Read error: 104 (Connection reset by peer))
20:59:13saratogaWMA is now 1.5x as fast as it was 2 years ago
20:59:31saratogaMP3 2x
20:59:33saratogaon PP
20:59:37saratogaVorbis is a lot faster too
21:00:17saratogaAAC is still slow, but faster then MP3 was in August 2007
21:00:32saratogawavpack got slower
21:02:18saratogaslower on both ARM and Coldfire
21:02:22kugeldid any hybrid wav code land in svn?
21:02:31saratogai don't think so
21:02:34 Join Strife89 [0] (
21:02:35saratogabut it happened a long time ago
21:02:42saratogaprobably fall 2007 or early 2008
21:04:47saratogai updated this:
21:06:11saratogathe coldfire results are surprisingly poor for most transform codecs
21:06:20saratogai suppose the current imdct lib isn't all that fast on coldfire
21:07:02saratogahopefully the new one will do something about that
21:12:33kugel"the new one"?
21:13:18 Quit TopyMobile (Remote closed the connection)
21:15:30saratogamt and i are working on a better one (hopefulyl better)
21:24:21 Quit adiroiban ("Leaving.")
21:27:05 Join toffe82_ [0] (
21:27:38 Join reydarz [0] (
21:27:40CIA-85New commit by tomers (r23351): Use ARRAYLEN macro in usb_keymaps.c
21:31:56 Join Gosc411 [0] (
21:34:09 Quit Gosc411 (Client Quit)
21:41:08 Join phanboy4 [0] (
21:44:21 Join phanboy_iv [0] (
21:47:55 Quit niekie (Read error: 131 (Connection reset by peer))
21:50:34 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
21:52:12***Saving seen data "./dancer.seen"
21:56:42 Quit toffe82 (Connection timed out)
21:57:57 Quit bluebrother (Nick collision from services.)
21:57:59 Join bluebroth3r [0] (n=dom@rockbox/developer/bluebrother)
22:01:47 Quit crashd ("leaving")
22:02:48 Join _0x90u[Nap|AFK] [0] (
22:03:00 Join crashd [0] (
22:03:06 Quit phanboy4 (Read error: 110 (Connection timed out))
22:03:07 Quit reydarz ("CGI:IRC (EOF)")
22:06:09 Quit n17ikh (Read error: 104 (Connection reset by peer))
22:12:46 Join stoffel [0] (n=quassel@
22:14:34 Join n17ikh [0] (
22:15:02 Join AndyI [0] (n=pasha_in@
22:17:30 Join faemir [0] (n=faemir@
22:17:57 Quit _0x90u (Read error: 113 (No route to host))
22:17:59 Join adiroiban [0] (
22:19:05 Join _0x90u [0] (
22:24:00kugelbertrik: I'll try the dbop patch later today. is there anything else I can test? I don't think 0xe0ff is going to work
22:27:09kugelbtw, is writing to dbop_dout safe during lcd updates? couldn't that won't that mess up the coordinates? I don't understand it fully
22:29:13 Join stoffel_ [0] (
22:29:28 Quit esperegu (Read error: 131 (Connection reset by peer))
22:31:23bertrikkugel, the lcd only really looks at the DBOP data lines around the time a control line pulses. The control lines are disabled by the dbop read routine (and restored to their old value at the end of the dbop read routine), so this should be safe.
22:32:15bertrikIn the old situation we actually communicated with the LCD controller instead, so it was possible that a command was interrupted by the write-red-pixel command.
22:32:35bertrik(at least, that is the theory :) )
22:33:52bertrikkugel, I'm a bit out of ideas what next to try, I'd really like to get rid of the extra delay for reading the hold button on the fuze, but maybe we just can't get around that.
22:34:33bertrikOne final thing to try is to write something into bits 8-11 of the pre-charge and see if that has any effect on the hold button reading
22:36:44 Quit Horscht ("Verlassend")
22:36:55 Quit faemir ("Leaving")
22:37:24 Quit _0x90u[Nap|AFK] (Read error: 113 (No route to host))
22:37:30 Quit stoffel (Read error: 148 (No route to host))
22:38:10mtThis is what I could identify so far on the clip+'s board :
22:39:38 Join DerPapst [0] (
22:40:10bertrikgreat, please put it on the wiki. The FM chip looks like another si4702 just like the clip v1, and the "SD" in the memory marking looks like a memory chip with SD card interface (which is easier to interface than a raw NAND chip)
22:40:10 Quit bmbl ("Bye!")
22:40:11mtShould I save it like that ? I wanted to double check the chips' markings but searching with any part of the marking didn't get me any good result.
22:41:01tomersCan somebody check the build server?
22:41:16tomersBuild should have been done 68m 57s ago, at 22:32:15...
22:41:56TheSevenipod playing for 21.5h now, battery indicator still above half full...
22:42:03mc2739tomers: I think it will correct itself the next time something is committed
22:42:28 Nick YPSY is now known as Ypsy (n=ypsy@
22:43:17TheSeveni think this can't be that terribly off... breaking the 24h apple manages should succeed then!
22:43:22kugelmt: the soc has the same markings as clip and fuze?
22:43:24 Join alexbobp [0] (n=alex@
22:43:39mtkugel: Yes
22:43:48TheSeveni wouldn't be surprised if it's still running when i get up tomorrow
22:43:53amiconnkugel: Do you have a patch for testing lcd_bitmap_transparent_part() in test_gfx somewhere?
22:44:21kugelamiconn: yea, but I noticed my tests are pretty nonsense
22:44:32*amiconn got some C code that's only 1 instruction longer than the asm; speed should be about the same, depending on the amount of transparent pixels
22:44:49 Join kyle6513 [0] (n=kyle@
22:44:58kugelor rather they only tested the loop performance, as the bitmap didn't have the transparent colors
22:45:26amiconnIf there are many transparent pixels, it should be a bit faster than the asm, if there are none or almost none, it should be a little slower (but still measurably faster than the C in current svn)
22:46:50 Quit flydutch ("/* empty */")
22:47:25kugelI basically made a test calling show_logo() (and modified show_logo() itself a bit), but I didn't exchange the rockbox logo
22:48:48 Quit kyle6513 (Client Quit)
22:49:32saratogamt: so basically you're stuck waiting for a sandisk firmware update?
22:50:07mtOr way to get into recovery mode :)
22:50:21kugelwith my paste from tonight I got 68fps on my fuze (again, only the loop performance). That does not do "int i = width" and then counting towards zero because that was actually slower it (the asm version does this)
22:51:06 Join faemir [0] (n=faemir@
22:51:25saratogamt: you cuold try shorting the pins on the flash memory and see if you can get into recovery mode like the e200v2
22:52:17mtI'll look into it tonight
22:54:19kugelI think I'm gonna do better tests as I also didn't cut enough out of show_logo() (i.e. it does more stuff that just doing the lcd_bitmap_transparent() call) :/
22:55:30linuxstbMaybe choosing some different-sized bitmaps from apps/plugins/bitmaps/native/ and drawing those
23:01:57linuxstbTheSeven: Wow, that's an impressive runtime...
23:02:32TheSevenSomebody needs to do one without boosting/undervolting then
23:03:17TheSeventhe battery runtime so far is far better than expected, the ADC measurements are probably way too high because of the current the ADC itself and the CPU (which will be busy waiting during the measurement) consumes
23:04:03 Quit stoffel_ (Remote closed the connection)
23:04:11TheSevenand remember that this ipod battery is 3 years old and very heavily used, several times deep discharged, etc.
23:04:47linuxstbIs boosting/undervolting in SVN?
23:04:56TheSevenno, but in flyspray
23:05:40TheSevenIf the battery indicator is right, we'll hit ~50 hours, that's ~8mA if the battery would still have its full capacity, so probably even less!
23:05:43linuxstbI could run some tests - my Nano2G has hardly been used, apart from the last couple of months for Rockbox development.
23:06:16TheSevenif what I think right now is true, the boosting patch may almost double the run time
23:06:43TheSevenso yours is likely to end within a day at least, while i'll have to wait for another day...
23:07:44linuxstbI could try with SVN, and then again with your patch. It makes sense to do them both with the same ipod.
23:08:01bertrikTheSeven, that's a really good runtime! many targets seem to consume something on the order of 30 mA or so.
23:08:05linuxstbSo is battery bench working, or are you just watching it "manually" ?
23:08:14TheSevenlinuxstb: if you don't need your ipod, go for it!
23:08:37TheSevenit's working after some fixes like stack size increases
23:08:56TheSevenhm, the battery bench stack is probably the only one I haven't committed
23:08:59bertrikcan you turn the ADC measurement somehow in a wakeup/signal wait?
23:09:13TheSevenit's I2C
23:09:29TheSevenso i need to poll the PMU for a finished conversion
23:09:39TheSevenbut yes, i could wait a bit less busy
23:10:08 Join niekie [0] (
23:10:15TheSevenmy ADC readings are at ~16mA
23:18:16 Quit Strife89 ("She sells C Shells by the seashore....")
23:21:58 Join Thundercloud [0] (
23:25:37 Join Strife89 [0] (
23:26:21linuxstbBagder: Can you poke/kick/tickle/whatever the build system?
23:30:53 Part froggyman
23:31:20 Join FuRom [0] (
23:31:25 Join Topy [0] (
23:31:53 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
23:31:53BagderI restarted the server, but I guess it doesn't regenerate the table until the next commit
23:32:41FuRomCan I use rockbox with the ipod touch?
23:33:39saratogaBagder: can you upload this to the webserver in the testfiles directory:
23:33:56 Part FuRom
23:34:12 Quit tresh (Client Quit)
23:34:46saratogaon a side note if anyone wants to encode a copy of that track in Apple lossless, Cook, ATRAC3, etc it would be great
23:35:00saratogai'm not really sure how to do most of those oddball formats though
23:35:38saratogaspeex and shorten too
23:36:06linuxstbApple lossless is hardly "oddball"... ;)
23:36:24saratogawell odd for me since I don't have itunes :)
23:36:35saratogathougyh i've got a lab machine with it somewhere so maybe i can make a file tonight
23:37:03saratogaalso i should dig out my gigabeat F and run a test on it
23:37:03linuxstbDoesn't ffmpeg have an alac encoder now? But I guess using itunes would make a better test file.
23:37:07 Nick Ypsy is now known as YPSY (n=ypsy@
23:37:09saratogayes it does
23:37:21saratogai think its actually supposed to be quite good so maybe that would be good enough
23:37:50Unhelpfulkugel: looking at maps from attempting to build on e200 with 4.4.1 eabi, after discarding the exception-related segments, i see references to a number of division-related (per naming convention) routines with __eabi prefixes. my guess is that the eabi compiler emits calls to different division routines than the "standard" one... perhaps aliasing these to more "normal" routines would help?
23:39:29 Quit faemir ("Leaving")
23:40:04kugelUnhelpful: maybe, remember we still need working division :/
23:40:58Unhelpfulkugel: obviously, but if we supply our own division routines with the expected names, or alias the eabi ones to "standard" ones, we can perhaps avoid the other cruft that they pull in.
23:41:01kugel__eabi_* ain't the problem, is it? the unwind relateted functions are
23:41:47kugelone should probably have a look why the eabi pulls unwind exceptions in even for C code :S
23:42:18kugelin the gcc sources, that is
23:42:59Unhelpfulgood luck, libgcc is a bit icky.
23:43:02 Join mt2 [0] (n=chatzill@
23:43:17kugelhehe, I can imagine
23:43:41 Join faemir [0] (n=faemir@
23:43:57 Quit ender` (" If you're going to do something tonight that you'll be sorry for tomorrow morning, sleep late.")
23:44:16mt2Tried shorting the flash's pins; 17 & 18, while and connecting .. here's what I get :
23:44:40mt2er .. *while connecting to usb (!!)
23:45:15mc2739kugel: FS #10723
23:46:09Unhelpfuli recall seeing things in libgcc, when i tried to trace the source of various functions, like defines in function names and in types being used to build the same function several times for different types, by compiling the same source file repeatedly with different flags. it's an ugly mess in there.
23:46:22kugelmc2739: grr, I'll have a look
23:46:40mt2(following funman's post here :
23:46:50 Join explore [0] (
23:47:20 Quit explore (Client Quit)
23:47:45kugelmt: interesting, that's more than what you get on the other devices
23:47:59 Join casainho [0] (
23:48:37casainhojoim emc
23:48:52casainhoui, sorry - my mistake
23:49:15mt2kugel: It reads the size as 4MB though ?
23:49:20 Quit T44 (Read error: 110 (Connection timed out))
23:49:52 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
23:50:41kugelmt: yea, 3,5MB more than on my fuze
23:51:56mt2oh ok.
23:52:14***Saving seen data "./dancer.seen"
23:53:10Unhelpfulkugel: a function that does nothing but return int a / int b calls __aeabi_idiv on arm-elf-eabi, __divsi3 on arm-elf
23:53:57mt2what are those files for : MTABLE.SYS, RES_INFO.SYS, SYS_INFO.SYS ?
23:54:27kugelUnhelpful: I remember that divsi3 somehow
23:55:00Unhelpfuldividing long long is somewhat more tricky, eabi uses __aeabi_ldivmod instead of __divdi3 - these quite likely have different calling conventions
23:56:06kugelcan you look what __aeabi_*div* calls related to unwind excetopns?
23:56:09bertrikmt2, I think one of those holds the configuration (like FM enabled or not, language, volume limit, etc)
23:56:23kugelalthough these are probably asm
23:56:46 Join evilnick_ [0] (n=evilnick@
23:58:46Unhelpfulneither seems to refer to those symbols :/

Previous day | Next day