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 2009-01-11

00:00:27 Quit solexx_ (Read error: 60 (Operation timed out))
00:00:29 Join matsl [0] (
00:00:51matslZagor: !
00:01:43*Zagor goes to grep the logs :)
00:06:02 Join MarcGuay [0] (
00:07:50MarcGuayjhMikeS: Please disregard my earlier comments about beast battery charging, I was running an old build. Current problems are much worse, though. The single-boot loader seems to fail.
00:12:47BigBambi_MarcGuay: I have been using the single boot for ages with no problems, but haven't updated for a long time. I'm a bit reticent about trying a new one as with a 120 GB disk it'd take ages to recopy the music over if it reformatted it for me
00:12:53 Part tusjen
00:12:57kugelwhy is (S)imulator under advanced options in configure too?
00:14:16MarcGuayBigBambi: Understood. I'm a bit peeved meself that my collection will probably need to be tranfered back on.
00:15:02BigBambi_Does it make any difference if you make it dual boot or not?
00:15:45 Join tyfoo2 [0] (
00:15:48 Quit tyfoo (Read error: 60 (Operation timed out))
00:15:51 Nick tyfoo2 is now known as tyfoo (
00:15:54MarcGuayBigBambi: I'm still trying to recover. Just found the OF CD in a drawer...
00:16:07 Quit bluebrother ("leaving")
00:20:53 Quit {phoenix} (Remote closed the connection)
00:21:39 Join z35_ [0] (
00:23:18MarcGuaytoffe82_: :) Thanks. Just realized that my CD is the Apps one, not the firmware. Cheers.
00:23:43 Quit z35 (Read error: 110 (Connection timed out))
00:26:57MarcGuayHmm. The updater crashes....
00:28:19 Quit nplus (Remote closed the connection)
00:28:26MarcGuaytoffe82_: "Updater has encountered a problem and needs to close".
00:28:36MarcGuayHere's hoping a reboot is sufficient...
00:28:40toffe82_you are on winxp ?
00:29:30MarcGuayI'll be back..
00:29:31 Quit MarcGuay ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
00:30:30 Quit archivator (Read error: 104 (Connection reset by peer))
00:33:44 Join funman [0] (n=fun@rockbox/developer/funman)
00:38:57 Join Mordechai [0] (
00:41:06 Quit bertrik ("Leaving")
00:42:22 Part XavierGr
00:44:55 Join MarcGuay [0] (
00:45:37funmankugel: int 00 watchdog seems to be an alarm that something is wrong somewhere (i already seen that)
00:48:37 Join mcuelenaere_ [0] (n=mcuelena@rockbox/developer/mcuelenaere)
00:52:50***Saving seen data "./dancer.seen"
00:54:05 Quit PaulJam (Read error: 113 (No route to host))
00:54:29 Join MOHDMACH [0] (
00:54:32MarcGuaytoffe82_: Hmm. The updater is still crashing.
00:54:59 Quit mcuelenaere_ (Read error: 104 (Connection reset by peer))
00:55:05 Join mcuelenaere_ [0] (n=mcuelena@rockbox/developer/mcuelenaere)
00:55:09MarcGuayDavide-NYC: Not sure if you read the logs but how did you recover your beast?
00:55:15 Quit mcuelenaere (Read error: 60 (Operation timed out))
00:55:42MOHDMACHWhich revision of CustomWPS list what %mo tag stands for or represents?
00:55:50 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
00:56:00toffe82_when you start it ?
00:56:45MarcGuaySendfirm is bombing out as well. Maybe a problem with the MTP protocol... toffe82_ : It bombs when I connect the player.
00:57:17toffe82_problem on your usb prot
00:57:28toffe82_port ^^
00:58:15kadobanis there documenation at all for lib/picture.h used in plugins? it's sort of hard to understand what's going on just by looking at the code...
00:59:02MarcGuaytoffe82_: The MTP icon disappears from the taskbar when the program dies as well..
00:59:35MarcGuayErr, maybe not realted, actually..
01:00:53 Quit domonoky (Read error: 110 (Connection timed out))
01:01:38robin0800jhMikeS: it would be loverly to see charging patch commited I can confirm it works fine for me now on my c240
01:01:39 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
01:01:39Lloreankadoban: Probably not. You're welcome to improve the comments, etc, once you learn. I'm sure it would be appreciated.
01:01:44 Quit mcuelenaere_ (Read error: 104 (Connection reset by peer))
01:03:22kadobanLlorean: is pluginbitmaps and picture.h and those what i should be using? i'm trying to have...multiple size sprites for mono/greyscale/color targets. will that work easily?
01:03:43toffe82_MarcGuay: your player is in recovery mode ?
01:04:07Lloreankadoban: I'd suggest just looking at how other plugins do it. A good example is Jewels as it has different images for nearly every target.
01:06:21MarcGuaytoffe82_: Yes, error #2.
01:06:56toffe82_MarcGuay: the display says connect to your pc ?
01:08:21 Join tim__ [0] (n=aoeu@
01:08:28 Quit timc`` (Connection reset by peer)
01:08:30 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
01:09:23MarcGuaytoffe82_: Update or restore Portable Media Center using PC application. I run the program, connect it, and it crashes,.
01:10:01toffe82_reset the palyer , disconnect he battery reconnect and try again
01:10:19kugelfunman: I figured ;)
01:10:20MarcGuaytoffe82_: "Disconnect" meaning flip the switch?
01:10:33 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
01:11:41MarcGuaytoffe82_: Same issue.
01:12:03 Join Failrar [0] (
01:14:19MarcGuaytoffe82_: The USB cable and port are fine, just tranferred some photos with it.
01:14:46toffe82_if you run the update without connecting the player, it crashes also ?
01:14:52MarcGuaytoffe82_: No.
01:15:03toffe82_problem on you player
01:15:07jhMikeStoffe82_: Does this sort of trouble begin after using the updater you found?
01:15:24toffe82_I don't think , mine is working
01:15:29 Quit flydutch ("/* empty */")
01:15:29MarcGuayI never used it before trying to recover from the single-boot loader.
01:16:08jhMikeSI was just wondering if that updater perhaps rewrote any code in flash that might makes things incompatible.
01:16:46toffe82_I have it running in dual boot, I didn't try single boot
01:16:56 Join gartral [0] (
01:17:13gartralthe sim for the e200s wont load any plugins...
01:17:44jhMikeStoffe82_: That's strange it would even care.
01:17:47MarcGuayjhMikeS: I didn't use it. Had v1.0 on when I tried to install the RB single-booter.
01:18:00 Quit robin0800 (Remote closed the connection)
01:18:42JdGordonamiconn: oh? how?
01:18:57MarcGuayI find it particularly concerning that sendfirm crashes as well.
01:19:03amiconnIt doesn't work anymore...
01:19:52amiconnJdGordon: Enable 'follow playlist'. Start playback in a folder, last track. Stop & shutdown, reboot, resume. Let it play to the end
01:20:13amiconn-> you will end up in the browser root, not on the track played last
01:20:43gartralok, make that the sm is copletly busted, and wont play anything
01:21:05gartralthe SIM is COMPLETELY busted, dangit keyboard
01:21:35toffe82_MarcGuay: your last option is to write direct on the disk, take off the disk from the player and connect direct to a computer
01:21:57MarcGuaytoffe82_: I'm going to see if sendfirm crashes in Linux first..
01:22:05amiconnIt works when using 'Stop', but it's also broken when going to the browser while playing
01:22:30amiconnAll 3 cases definitely worked in December
01:24:40*toffe82_ finally has a hdd1630 running rockbox, the touchpad is really sensitive
01:27:07Zagorgartral: the sim works fine for me. it plays both music and games.
01:28:48*toffe82_ wonders how to switch it off ?
01:31:52Zagortoffe82_: #define POWEROFF_BUTTON BUTTON_POWER
01:32:19ZagorI have no idea which button that is though :-)
01:32:32toffe82_I try to find out
01:33:02toffe82_the other problem is that if I connect the usb, the system reboot until I disconnect
01:33:37Zagorsounds like the bootloader doesn't detect usb? it boots into rockbox and then immediately reboots?
01:34:24gartralits been doing that for a few days now
01:36:32Zagorgartral: what has been doing what?
01:37:03gartralUSB reboots into RBFW then reboots AGAIN into RBFW and freexes
01:37:32Zagorgartral: are you too using a hdd1630?
01:37:43toffe82_it boots rockbox, when I connect I see the logo usb and then it reboot
01:37:44gartralnopem sansa e250 v1
01:38:26Zagorgartral: when did you install a new bootloader last?
01:38:28toffe82_I think ther is only 3 person with a hdd1630
01:38:45gartraland now, with a fresh build of the sim firmware option, the sim STILL crashes
01:39:44 Join itcheg [0] (i=62db4767@gateway/web/ajax/
01:39:46Zagorgartral: I am listening to mp3 and playing bubbles on the e200 sim right now
01:40:44 Quit mcuelenaere ()
01:41:19 Quit funman ("leaving")
01:42:55gartralwell then, what could i be doing wrong?
01:43:36toffe82_how do you boot on the OF on the sansa ?
01:43:49Zagortoffe82_: hold "left"
01:46:26 Quit fyrestorm (Read error: 104 (Connection reset by peer))
01:48:10*gartral tries sim with a clean build folder
01:48:35 Quit Zagor ("Clint excited")
01:50:17 Quit tim__ (Read error: 110 (Connection timed out))
01:51:17 Join tim__ [0] (n=aoeu@
01:51:44gartralclient excited? o.o
01:52:58 Nick blkhawk- is now known as blkhawk (
01:56:59 Quit matsl ("Leaving")
01:57:19gartralso why does a "homebuilt" sim not want too load anything, and a "boxed sim" from the site work perfectly?
01:58:08Unhelpfuldid you make install?
01:58:31Unhelpfulactually, fullinstall would be a good idea, to get the fonts as well
01:59:21 Quit itcheg (" ajax IRC Client")
02:00:56Unhelpful <- are the LCD_PIXEL_ASPECT_{WIDTH,HEIGHT} defines decently named? and what do we think about the ALIGN_BUF macro... it seems to be reasonably proof against attempts to make it fail compilation, etc, and the align_buf function provided on swcodec doesn't exist for many of the targets that are now getting the scaler.
02:02:36 Quit BigBambi_ ("Leaving")
02:03:45 Join itcheg [0] (i=62db4767@gateway/web/ajax/
02:06:25*toffe82_ didn't build the good version to have dual boot on the hdd1630 :(
02:10:58gartralthis will be portabe from the linux host too the windows sim, right?
02:12:40gartralthe ./simdisk/.rockbox dir
02:13:50Unhelpfuli would think so
02:14:18gartralok, im just easily confused, thank you for helping me
02:16:29linuxstbgartral: No, the .rockbox dir contains the plugins and codecs - you'll need to make a new one for Windows (or at least replace the binary portions)
02:17:38kadobando LCD_DEPTH == 1 have any analogues to the simple_resize_bitmap and smooth_resize_bitmap functions? (or really any way to scale a bitmap)
02:17:50kadobanthat should be LCD_DEPTH==1 targets
02:19:35gartrallinuxstb: and ide do that from I at configure?
02:20:07Unhelpfulkadoban: i've made the core scaler available in pluginlib on mono targets. look at apps/plugins/test_greylib_bitmap_scale.c
02:20:33linuxstbgartral: ?
02:20:48kadobanUnhelpful: oh thanks, i missed that one somehow :)
02:21:22Unhelpfulthe *_resize_bitmap stuff only works on already loaded bitmaps, and scaling mono bitmaps is basically pointless
02:21:53gartralhow would i go about compiling a windows SIM? its not covered in the tutorial...
02:22:11Unhelpfulif you *want* scaled mono output for a plugin, a format plugin could be written to produce such bitmaps... wouldn't be of much use, though.
02:22:19Unhelpfullinuxstb: silly me... thanks for catching that :)
02:23:21linuxstbgartral: On Windows (i.e. in cygwin), the same as in Linux. If you want to cross-compile a Windows sim from Linux, then there should be some instructions somewhere on the wiki, but I don't know exactly where.
02:23:39 Quit MOHDMACH ()
02:24:02gartralok, nvm, the tut DOES have them, and im being a blind lame man again
02:25:41*gartral bows and faceplats the table, "i forgot to make the makefile before i did make fullinstall"
02:26:01Unhelpfulgartral: if you ran configure, you've "made" the makefile
02:26:19 Join int21hexster [0] (
02:26:39gartralerr... typed MAKE
02:26:39kugelUnhelpful: did you have some bitmaps for me to try out?
02:27:03gartrali did "make fullinstall" with the prior "make"
02:27:15Unhelpfulwhere i'm going with the inline stuff is looking for a better replacement for MEM_FUNCTION_WRAPPERS in apps/plugin.h
02:27:19gartralwithout* rather
02:27:45Unhelpfulideally, i'd like something exactly equivalent to replacing a call to memfoo(...) with a call to rb->memfoo(...)
02:28:21Unhelpfuland a macro can't solve this, because the problem is not calls to the functions in source, but compiler-generated calls to them :/
02:28:40 Join Thundercloud [0] (
02:35:04Unhelpfulkugel: working on it.
02:35:26kugelok :)
02:47:18 Quit moos ("Rockbox rules the DAP world")
02:50:27 Join Darksair [0] (n=user@
02:52:52***Saving seen data "./dancer.seen"
02:59:53 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
03:03:01 Nick z35_ is now known as z35 (
03:06:55 Quit int21hexster (Read error: 60 (Operation timed out))
03:07:43 Quit tessarakt ("Client exiting")
03:21:10 Quit Thundercloud (Remote closed the connection)
03:26:56 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
03:38:02 Quit AndyI (Read error: 104 (Connection reset by peer))
03:39:09 Join kugel [0] (n=chatzill@unaffiliated/kugel)
03:40:50kugelfunman: there's apparently a problem with the SD driver. Writing a config.cfg on my clip causes it to not boot the next time (OF works). I looked at the created config.cfg (because that's loaded at boot), and it contained invalid chars (my text editor couldn't open it). Additionally, deleting it, made the clip booting again. This is reproducable
03:41:15 Quit kugel (Client Quit)
03:43:30 Join AndyI [0] (i=AndyI@
03:43:47 Quit killan (Read error: 104 (Connection reset by peer))
03:44:40 Join killan [0] (
03:45:13 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
03:50:33 Quit tim__ (Remote closed the connection)
03:53:37 Join timc`` [0] (n=aoeu@
04:11:33 Join rotarychainsaw [0] (
04:15:26rotarychainsawAnyone try using the rockbox installer gui on ubuntu
04:16:18krazykitrotarychainsaw, do you have a more specific question?
04:17:39rotarychainsawOh, yeah just seeing if anyone was looking in here. It claims to not auto detect, and it won't let me use the normal mount point of /media/ipod
04:17:57rotarychainsawBut if I look at the command line out put it says it finds it at /dev/sdc
04:18:30Lloreanrotarychainsaw: You need to run it with sudo for it to work properly
04:19:12rotarychainsawyeah, I am. If I tell it that the path to the ipod is /dev it doesn't complain, but it sounds dengerous to me
04:19:33LloreanAutodetect should work.
04:19:51LloreanIt works for everyone else. What model iPod, and what version of RBUtil?
04:20:54rotarychainsaw$g grayscale ipod
04:21:14LloreanYou're connecting with USB, and don't have any other modifications installed to it?
04:21:26rotarychainsawand the RButil on the front page was... 1.0.7 since I run 64 bit
04:21:53rotarychainsawYeah, usb, looks plain jane, just got it from a co-worker and replaced the battery
04:21:56 Part aurix_lexico
04:22:04 Join aurix_lexico [0] (
04:22:10LloreanRBUtil uses the same code to detect an iPod as the iPodPatcher tool
04:22:27LloreanTry using the latest RBUtil with 32-bit compatibility if you can.
04:22:59rotarychainsawK i'll check it out. Is there source for the util somewhere? Maybe I could just compile a newer version
04:23:08LloreanIt's just in the rockbox SVN along with everything else
04:23:21rotarychainsawmakes sense haha
04:23:57 Part gartral
04:24:04LloreanAnd the device node is used for bootloader install, not the mount point. The mount point is only used after that to extract the binary to.
04:26:23rotarychainsawah, I think that is where I am getting messed up
04:26:32rotarychainsawcause the node is /dev/sdc
04:26:45rotarychainsawbut I can only pick /dev in the menu
04:27:25LloreanYou shouldn't need to pick anything...
04:27:55rotarychainsawoh well thats true too I guess
04:33:03 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:44:25 Quit Bensawsome ("The awsome is gone :(")
04:44:33 Join Barahir_ [0] (
04:44:42 Join Bensawsome [0] (n=Bensawso@unaffiliated/bensawsome)
04:46:36 Part toffe82_
04:48:20rotarychainsawLlorean, found the problem, its a mac ipod.
04:48:29rotarychainsawI'm converting now
04:48:29 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
04:49:45Lloreanrotarychainsaw: Ah, the manual should mention the necessity that it be converted. But I thought the utility could detect mac formatted ones and tell you.
04:50:23rotarychainsawyeah I was hoping it was windows, I guess not. It told me it was a mac one after I just told it to go to work on /dev
04:50:58saratogai think i understand how to efficiently parse an mp4 file
04:51:11Lloreansaratoga: That would be excellent news.
04:51:18LloreanSince that's the main flaw in our AAC support
04:52:56***Saving seen data "./dancer.seen"
04:54:25 Quit Darksair ("People who are zhuangbility want to show their niubility but only reflect their shability.")
04:55:22 Join MOHDMACH [0] (
04:57:07 Join blkhawk- [0] (
04:57:38Lloreansaratoga: Improved AAC support _would_ be great with iTunes dropping DRM (assuming that includes audiobooks and not just music)
05:01:44 Quit Barahir (Read error: 110 (Connection timed out))
05:03:02saratogaLlorean: yeah cleaning up AAC has been on my todo list for ages
05:03:32saratogai spoke to one of the ffmpeg people tonight, apparently they're well along the way to getting working AAC-HE, so maybe we can steal that from them and have fast AAC-HE support too
05:05:00LloreanWhat do they use for AAC-LC then?
05:05:16LloreanI mean, don't we have -HE on our very fast targets?
05:05:27LloreanOr is what we have missing either SBR or PS?
05:05:41 Quit tyfoo ("Carpe diem")
05:08:03 Quit rotarychainsaw ("Leaving")
05:08:05saratogaLlorean: they've written their own AAC-LC decoder from scratch, but not HE quite yet
05:08:16saratogawe have faad, which has all aac profiles, but its quite awful
05:08:35LloreanNow I remember, their -LC wasn't fixed point when we looked at it, right?
05:08:45saratogaits all floating point unfortunately
05:08:57saratogabut compared to fixing faad, converting the SBR stuff is probably easier
05:09:20saratogathough i still only half understand SBR so I could be completely wrong about that
05:09:42LloreanIs there any open source MP3Pro stuff? That's SBR too, right?
05:09:55saratogai want to try combing faad for LC and ffaac for HE
05:10:21saratogamp3pro is SBR, but i don't know anything about it, i've never seen an encoder or decoder for it
05:11:50saratogaif we ever merge in the ffaac stuff, i think we'd be justified in dropping and pretense of being able to resync libfaad changes, in which case we could clean out all the stupid mallocs and enormous amount of ifdef'ed code for things like SSR, DRM, etc
05:12:06saratogathen we'd have something like an understandable codec
05:12:44 Part Bensawsome ("The awsome is gone :(")
05:14:11 Quit blkhawk (Read error: 113 (No route to host))
05:15:05 Nick blkhawk- is now known as blkhawk (
05:15:50 Part hobbs ("Leaving")
05:22:18MOHDMACHI finally setup a viewport and volume to make it look like the way I want it but the none of the text would display.
05:22:51MOHDMACHI have the viewport sized so it would only cover the progress bar.
05:23:59 Join blackthunder [0] (
05:24:08MOHDMACHThe WPS works but the text will not display only the progress bar, play status, volume and battery.
05:24:16MOHDMACHHere is the paste bin.
05:25:58 Quit MarcGuay ("ChatZilla 0.9.84 [Firefox 3.0.4/2008102920]")
05:26:14blackthunderI am having trouble with rb util
05:26:18Lloreansaratoga: Is faac even maintained any more? Do we really *want* to try to be syncable?
05:26:40LloreanEr faad
05:27:12blackthunderIt won't detect my sansa to install the bootloader
05:27:24 Join efyx_ [0] (
05:34:10saratogaLlorean: i think its actually maintained, but we're quite badly out of sync already
05:34:37saratogait would make sense to go through and see if theres anything else we want from them i guess
05:34:40Lloreansaratoga: I thought that since the licensing issue, there really weren't anyone working on the older code. But I'll admit I never really looked at it.
05:35:04saratogathey resolved the licenseing issues, and continue to maintain it
05:35:16LloreanI thought the later code wasn't GPL compatible?
05:35:41saratogasome people complained that the license wasn't in the spirit of the gpl or whatever, so they changed it
05:36:11saratogabut ahead also sells a closed source decoder too, so i think they're mostly interested in making faad compliant, not efficient
05:36:30saratogaanyway, going to bed
05:36:44 Quit saratoga ("CGI:IRC (EOF)")
05:39:49 Quit MOHDMACH ()
05:41:29 Join MOHDMACH [0] (
05:42:24 Quit blackthunder ()
05:49:14 Join fyrestorm [0] (
05:52:00 Quit ze ("blah")
05:57:20kadobangoto isn't banned in rockbox, right?
05:57:24kadoban(for sane use)
05:59:46 Quit amiconn (Nick collision from services.)
05:59:49 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
06:05:33 Join nibbler__ [0] (
06:12:27 Join Darksair [0] (n=user@
06:16:46 Join pixelma_ [0] (n=pixelma@rockbox/staff/pixelma)
06:17:11 Join gfather1 [0] (n=g@
06:17:41 Quit pixelma (Read error: 110 (Connection timed out))
06:19:09 Quit gfather (Read error: 60 (Operation timed out))
06:21:17 Quit nibbler_ (Read error: 110 (Connection timed out))
06:22:13MOHDMACHI have a .wps file that has the "%mo" tag but it says nothing about it in the CustomWPS guide and version 3.1 comes standard with support for the tag.
06:22:58MOHDMACHHas someone just removed the tag recently and revised the guide?
06:23:42MOHDMACHThe only information I got from quickly looking at the old patch for it was that it was a view mode tag.
06:24:22 Join Davide-NYC [0] (
06:25:01 Join TheSkunkMan [0] (
06:27:07Davide-NYCI am considering atttempting a dual-boot install of Rockbox on my Gigabeat S (v1.2) after the problems with the rockbox only bootloader. Has anyone successfully installed a dual-boot nk.bin to an updated Gigabeat S?
06:27:41TheSkunkManWhere can I obtain some bootloader source code to view as examples?
06:28:41Davide-NYCAFAIK it's in SVN in the bootloader folder. But I amnot a dev.
06:38:41LloreanMOHDMACH: 3.1 shouldn't support that tag. Are you sure you're using 3.1 without any patches added in?
06:39:04MOHDMACHYes positive.
06:40:07MOHDMACHThe wps file itself is the only "advanced" one I've seen work with a current build of rockbox.
06:40:50LloreanDoes the WPS actually do the mode changes properly, or are you saying it "works" just because it's not rejected?
06:41:12LloreanThat wasn't a yes/no question.
06:41:18LloreanWhich of those two statements are you yessing?
06:41:46 Quit z35 ("Leaving")
06:41:57MOHDMACHIf by rejected you mean "broken" then no 3.1 does not break the .wps file.
06:42:10MOHDMACHIt displays as it should.
06:42:10LloreanDoes the mode change work?
06:43:51MOHDMACHWhen Jdgordon was discussing that I did not exactly understand what he meant by that.
06:43:56MOHDMACHIn flyspray.
06:44:26MOHDMACHBut the .wps file itself should be broken if that tag should not work.
06:44:38LloreanYes, but it could be a bug that the tag isn't rejected
06:44:52LloreanIf you press "Select" while in the WPS while using that .wps, what happens? Does the WPS change, or does it send you back to the browser?
06:45:12MOHDMACHI'll try it right now.
06:47:15MOHDMACHWhen I quickly press it while the WPS is displaying it sends me back to "Files".
06:47:32LloreanSo then, the bug is that it does not reject WPSes with the %mo tag.
06:47:42 Quit AndyI ()
06:47:43LloreanThe tag isn't supported, but it's not treating it as unsupported properly
06:47:51LloreanPlease, file a bug report.
06:48:41JdGordonwait... what?
06:48:54LloreanJdGordon: %mo isn't rejected.
06:49:03MOHDMACHWhat he said.
06:49:05JdGordonneither is %zz
06:49:21LloreanUnrecognized tags are supposed to be rejected, right?
06:49:32MOHDMACHBoth of which are not in the CustomWPS guide.
06:49:34LloreanThat's kinda the whole point of the stricter parser.
06:49:35JdGordonI dont know
06:50:47MOHDMACHSo this is a bug I should file right?
06:51:30 Quit fyrestorm ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
06:51:41 Quit itcheg (" ajax IRC Client")
06:53:00***Saving seen data "./dancer.seen"
06:53:33 Join AndyI [0] (i=AndyI@
06:53:44LloreanPossibly it's a more general bug (WPS parser doesn't reject unknown tags) rather than one pertaining to that specific tag.
06:55:07MOHDMACHSo I should file this right?
06:56:02LloreanIt's a bug, there's not a bug report filed about it. So, yes.
06:56:29MOHDMACHAny details you want me to give.
06:56:58MOHDMACHYou will need the .wps itself to reproduce the bug, right?
06:57:32MOHDMACHThe .wps file was made for testing view modes.
07:01:50 Quit aurix_lexico ("Leaving.")
07:03:26Davide-NYChow do I make the mknkboot tool?
07:03:35JdGordondo we definetly want to reject any wps with a unknwon tag?
07:03:49JdGordonDavide-NYC: make in the mknboot folder
07:04:40Davide-NYCJdGordon: where is that? I only find a mknkboot.c file in tools.
07:05:10JdGordonoh, sorry... im confused with mkamsboot
07:06:43 Quit efyx_ (Remote closed the connection)
07:07:02 Quit Davide-NYC ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
07:07:29LloreanJdGordon: We want to reject any WPS that won't display properly, I think, so yes. Maybe I've misunderstood, but I thought "tags that don't exist" are also syntax errors.
07:18:49JdGordonok, well its easy enough to reject any not found tokens.. but the error message isnt very useful
07:19:01JdGordonERR: Failed parsing on line 1 : ERR: Invalid parameter list for token 0: "No token"
07:19:24JdGordononly the 2 numbers in that will change... it will always be tha same error
07:22:26TheSkunkManHey I'm andrewahlers is this the right place to ask for write access to the wiki?
07:23:21lucentI think Bagder can configure this for you
07:23:28lucentwhat part of the wiki do you need to update?
07:23:50TheSkunkMani was thinking of starting a page for the iriver lplayer
07:24:01JdGordonanyone can give write permissions... hang on
07:24:26JdGordonwhats your wiki name?
07:26:15 Join ze [0] (i=ze@
07:28:47JdGordonlucent: for future reference.... just add the name to the TWikiUsersGroup page...
07:29:17lucentJdGordon: oh, I didn't know it was like that, thanks
07:32:27MOHDMACHI have created a bug for it.
07:33:15MOHDMACHThe tag itself doesn't cause me any inconvenience as for other it might.
07:36:03MOHDMACHAlthough in the next version the .wps file will probably not work anymore. :(
07:38:14MOHDMACHThe "%mo" tag had a lot to do with the volume effect it created with conditional viewports as you will soon see.
07:44:01 Join lee321987 [0] (
07:44:37lee321987Did anyone ever fix the "2 bytes inserted" bug on the Sansa c200?
07:45:07 Part MOHDMACH
07:45:49lee321987Did anyone ever fix the "2 bytes inserted" bug on the Sansa c200?
07:46:39lee321987(sorry if that posted twice −− I forgot to identify with NickServ)
07:46:57lucentgood, I was almost confused
08:04:41 Join midkay [0] (n=midkay@rockbox/developer/midkay)
08:27:36 Join Rob2223 [0] (
08:29:38 Quit jhulst (Read error: 104 (Connection reset by peer))
08:32:58 Nick Barahir_ is now known as Barahir (
08:34:54 Quit Rob2222 (Read error: 60 (Operation timed out))
08:42:22 Join fyrestorm [0] (
08:44:18 Quit midkay ("Leaving")
08:49:35JdGordonsomeone forgot to bump one of the many api min version values :/
08:50:38 Quit fyrestorm (Remote closed the connection)
08:53:04***Saving seen data "./dancer.seen"
08:53:15 Join Lss__ [0] (
08:53:15 Quit Riku (Read error: 104 (Connection reset by peer))
08:56:16 Quit lee321987 (Read error: 110 (Connection timed out))
09:01:09 Join at0m [0] (n=a548c80b@gateway/web/cgi-irc/
09:10:49 Quit at0m ("CGI:IRC")
09:11:11 Join at0m [0] (n=a548c80b@gateway/web/cgi-irc/
09:15:29 Join gartral [0] (
09:17:05gartralall codecs fail in RB, build r19746m any attemp too access any pluging returns with error "plugin array full"
09:18:18gartraland menus are all kinds of scrambled
09:19:55 Join wojtek [0] (n=4dfdf542@gateway/web/cgi-irc/
09:20:28 Nick wojtek is now known as Guest41679 (n=4dfdf542@gateway/web/cgi-irc/
09:21:09Guest41679I have a guestion about the sansa c250
09:21:39gartralplease dont ask to ask, just ask
09:21:58 Quit Guest41679 (Client Quit)
09:22:09 Quit tvelocity (Remote closed the connection)
09:23:00 Join webguest67 [0] (n=4dfdf542@gateway/web/cgi-irc/
09:24:03webguest67is there any big changes between rockbox v.3.0 and 3.1 on sansa c250??
09:24:35gartralumm, look through the changelog
09:24:36scorchewebguest67: i suggest that you have a look at the MajorChanges wiki page
09:25:38scorchethere is also the ReleaseNotes31 wiki page
09:25:44gartralone moment, ill get the link
09:26:01webguest67web client works so slowly ,sorry
09:27:05webguest67can i get here from emule irc client?
09:27:38gartralno, but ill point you too a faster webclient, one sec
09:27:44DBUGSent KICK webguest67 to server
09:27:44Kick(#rockbox webguest67 :Stop repeating yourself) by logbot!n=bjst@gateway/web/cgi-irc/
09:30:46gartralall plugins are failing, including most core plugins....
09:30:47 Quit at0m ("CGI:IRC (EOF)")
09:38:03JdGordongartral: we dont support custom builds.... nfi what causes plugin array full though
09:38:42gartralok, ill see if im haveing the same trouble with a stock
09:39:31gartralthough, i should mention, its not really a costom build, as all i did was replace the bootlogo image >.>
09:39:52amiconngartral: You cannot exchange parts of the sim between windows and linux. The pligins and codecs are shared objects resp. dlls, so won't work on the other os
09:40:49amiconnUnhelpful: The mem function wrappers must not be inline.
09:42:19Unhelpfulamiconn: so, they either have to be defined in the file that will need them, via the current macro, or they can be "regular" functions in the pluginlib?
09:43:06Unhelpfulwhy can't they be inlined? surely it would "work" if gcc used a call to rb->memcpy instead of memcpy, if we had a way to get it to do that.
09:43:24gartralamiconn: this my physical e250, not the sim thats doing this
09:44:12amiconnUnhelpful: The warppers aren't meant for use by the plugin programmer. (s)he should always use the proper rb-> call, directly or via macro
09:44:39amiconnBut in some cases gcc uses the mem* functions itself, e.g. if you assign structures
09:45:04Unhelpfulright, and i wanted something that it could inline in that case... but it appears that that is not possible
09:45:07amiconnThe wrappers are there to provide the necessary fucntions, as demanded by gcc
09:45:34amiconnNo it *is* not possible. Gcc wants a proper function symbol, as per manual
09:45:55 Join flydutch [0] (
09:47:17Unhelpfulgood to know, i won't waste any more time on it. should the wrappers stay the way they are? it would be nice if plugins didn't have to "know" when gcc might insert calls to them, but i don't want to hide code that accidently uses memfoo instead of rb->memfoo
09:49:26 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:49:26amiconn"GCC requires the freestanding environment provide memcpy, memmove, memset and memcmp."
09:49:38amiconn(from the gcc manual)
09:50:42amiconnThe wrappers could indeed be part of the pluginlib though. This would need testing, as the linking order might matter
09:51:38gartralok... im being retarded, my entire problem was i wasnt paying attention to wich build folder i was in, and DLed the wrong
09:52:21*JdGordon sends gartral an invoice for the wasted bandwidth!
09:52:39amiconngartral: I was referring to your question from 02:10
09:53:00*amiconn sees it was already answered... :\\
09:53:01gartralahh, i have already figured it out...
09:53:17Unhelpfulare CACHE_FUNCTION_WRAPPERS and ALIGN_BUFFER_WRAPPER strictly conveniences? it seems that things using those functions should be calling the api
09:54:18*gartral deserves a good thwack in the head after that last one
10:02:00 Quit timc`` (Remote closed the connection)
10:03:08*amiconn doesn't know
10:03:42amiconnThe buffer align mascro could be made available for all plugins, and the repective function removed from the core, as it isn't used therer
10:07:26Unhelpfulit seems to me that a macro for this belongs with ALIGN_UP/ALIGN_DOWN, etc.
10:10:50 Join bertrik [0] (
10:11:27UnhelpfulCACHEALIGN_BUFFER is only used one place, and it really seems like it ought to be replaced with a call to a generic alignment macro, using the cache alignment, instead of having a special macro for aligns in that size
10:12:22Unhelpfulshould all of these changes wait until *after* a decision is made about global api pointer, though?
10:14:44 Join archivator [0] (i=foobar@
10:16:19 Quit archivator (Client Quit)
10:16:28 Join archivator [0] (i=foobar@
10:21:26 Join timc`` [0] (n=aoeu@
10:22:22amiconnUnhelpful: These changes should be done in a separate commit, but they're essentially independent and could be done either before or after.
10:23:45Unhelpfulmem* in pluginlib kind of depends on global API, otherwise it's just replacing the wrappers macro with memfuncs_init to copy the pointer
10:23:49amiconnIt happens quite often that preparing one largeish change turns up a bunch of other ugly things
10:24:19amiconnAh, yes
10:24:47amiconnDid you try to apply the global api stuff to codecs as well?
10:25:01amiconn(rb -> ci but essentially the same thing)
10:26:26Unhelpfulno, i did not, yet. i don't see much else to do that would belong in the same commit, and the patch is pretty huge already... maybe i should start a discussion on the ML regarding it?
10:29:33amiconnDid you try your idea to make __header a union? Does that allow to just write 'rb' without any reference to an element of the union?
10:31:04 Join gfather [0] (n=gg@
10:31:08gfatherhello guys
10:31:12Unhelpfuli couldn't find a way to remove reference to the union, no, unless the pointer is the first member of it.
10:33:32Unhelpfulan anomymous union will inject its named members into the scope in which it is defined, but from what i can find, global anonymous unions are forced to be static, and can not be initialized.
10:34:36gfatherCasainho u there ?
10:35:54amiconnThe pointer could be the first member, aliasing the "magic". But how do you get rid of the member reference?
10:37:56gfatherwhats the subject guys ?
10:38:53scorchegfather: we have public logs for a reason...
10:39:26bertrikWhat's (supposed to be) the status of mp3 playback on clip? I only get noise...
10:40:01gfatherscorche i said whats the subject , not give me a preife on whats going on
10:40:31Unhelpfulamiconn: it would also overwrite target_id and api_version, on 64bit sim - not sure if that's a problem. and i don't think that an anonymous union can be (ab)used that way, after all.
10:41:23 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
10:42:29Unhelpfulin particular, anonymous unions not in the scope of a struct or union declaration would appear to be a c++ extension - g++ will compile "static union {int a; int b;};", gcc only thinks that i'm declaring a union type without a name or any instances.
10:43:16Unhelpful...and obviously a and b should have different types *smacks forehead*
10:43:54*domonoky wonders how you would access this, as it has no name ?
10:44:43Unhelpfuldomonoky: in c++, that will inject a and b into the namespace in which the union is declared
10:45:36domonokysounds dangerous :-)
10:45:46Unhelpfulyou can't as far as i can tell, initialize them. you can't make the union non-static. and you can't compile this source for C, only C++. we need to do all three of those things for this to be a solution we can use
10:46:51amiconnOverwriting any other member of the struct after loading is no problem except the entry point of course
10:47:19Unhelpfulamiconn: right, but we still can't use an anonymous union in that manner, anyway
10:53:05***Saving seen data "./dancer.seen"
10:56:49 Part gartral
10:57:11gfatheranyone thinks the themeing system needs update ?
10:59:37 Join fyrestorm [0] (
10:59:55JdGordongfather: depends what updates you're think
11:00:02 Quit Darksair (Remote closed the connection)
11:00:39gfatherJdGordon improve the visual api and or change it
11:01:32 Join merbanan [0] (n=banan@
11:01:56JdGordonlike how?
11:02:18gfatherim thinking of using enlightenment
11:02:43gfatherill search a littel and then inform u with something solid ;)
11:03:11 Join gregzx [0] (
11:04:09scorche"using enlightenment"? would you do that?
11:04:14domonokygfather you are sure, you know what you are talking about ? :-)
11:05:09 Join {phoenix} [0] (
11:05:49gfatherwell i said , im gonna search a littel , and then get some solid info here
11:05:57gfatherbut i have to change the visual api ,
11:06:09*scorche awaits a patch
11:06:19JdGordonyou probably want to change the visuals.. not the api
11:06:56archivatorI was thinking, could it be possible to bring FS #8523 to the driver. i.e. move all screen updates to a buffer and blit to the screen only when the backlight is on?
11:07:24archivatorI really have no idea how the lcd drivers work right now..
11:07:30gfatherno , the api , becouse its great on the playing screen , but not on the menues and stuff
11:07:37gfatheri have something in mind :)
11:07:54JdGordonit is pretty unliekly that the themeing will get any better in the menus...
11:10:15scorcheJdGordon: we can add transition effects!
11:14:07archivatorI guess I'm in the wrong time zone - everyone seems to be sleeping..
11:15:18 Join HDDtoCF [0] (n=50bb6031@gateway/web/cgi-irc/
11:15:30gfatherim not :)
11:15:43 Join Jaykay [0] (
11:16:43 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
11:17:14amiconnarchivator: fs 8523 is not about the driver, but about the wps code. The LCD driver already does that on targets where the LCD is unreadable without backlight, and hence gets disabled
11:17:50archivatoramiconn: define "unreadable"
11:18:33amiconnMost colour LCDs are unreadable without backlight (among our targets, the only exceptions are iPod Video and Nano)
11:18:57amiconnMonochrome/greyscale LCDs stay readable, so 8523 makes no sense
11:19:43archivatoramiconn: I see. Well, what do I change to make the Video "unexceptional"? :)
11:19:52 Join nuonguy [0] (
11:19:54amiconnYou can't
11:20:09amiconnFor the Video's LCD controller, we don't know how to disable it
11:20:30bertrikIIRC, for sansa c200, the display is disabled when backlight is off to save power, but screen updates are still written
11:20:45JaykayjhMikeS: is the battery charging for e200/c200 a candidate for major changes? and is your real name mike or michael?
11:20:46archivatorCan't we at least stop updating it? Simple physics says that will save power.
11:21:50jhMikeSJayKay: Either name really. I guess its a major change. I didn't really think about it.
11:21:59linuxstbamiconn: Is your ipod color's LCD unreadable without backlight?
11:22:18 Quit HDDtoCF ("CGI:IRC (Ping timeout)")
11:22:29archivatorAlso, can't we brute force it? :) I mean, get a webcam pointed at the screen, rockbox trying to brute force the controller and then just examining the captures :) (i'm only half joking)
11:22:49Jaykayjhmikes: i didnt get that with your name, sorry... and should i add it to MajorChanges?
11:23:56pixelmalinuxstb: from what I have seen (and remembered about Linus' Video and Nanos I saw at DevCons), the Ipod Color screen is not as readable as the Video's and Nano's
11:24:02jhMikeSJayKay: My birth certificate says "Michael". And yes, I think it's a major change.
11:24:33pixelmalinuxstb: maybe not "unreadable" though
11:25:07jhMikeSI'm guessing FM Recording on Gigabeat S and Charging from Main/USB for Gigabeat S should be there if the FM addition is there.
11:25:07JdGordonI know.. settings bloat... but maybe the option should be added
11:27:26linuxstbpixelma: I have both a video and a Color, and I can't remember there being that much difference (but I haven't compared them side-by-side) - neither are comfortable to use without backlight, but they can be.
11:29:07*pixelma wonders whether there are differences in readability without backlight for the two different panels (if remembering correctly that the Color had different ones)
11:30:33linuxstbpixelma: Yes, I was wondering the same - that's why I was asking amiconn about his Color. Although maybe the panel is equivalent, and it's just a different controller.
11:30:48archivatorSo, just to clarify, disabling lcd updates when the backlight's off is not possible atm?
11:30:52 Join CFmod [0] (n=50bb6031@gateway/web/cgi-irc/
11:31:34linuxstbDepends what you mean by "lcd updates", and what device you are talking about.
11:31:59archivatoriPod Video - I mean exactly that - updating the screen. Not turning it off, just disabling the updates.
11:32:36_Auron_as the screen is still updated, just not lit
11:32:43_Auron_which is a bit of a waste of power :P
11:33:35linuxstbarchivator: In Rockbox "lcd_update()" is the function which blits the Rockbox framebuffer to the LCD hardware. Are you talking about disabling that update, or about not even drawing into the framebuffer when the backlight is off?
11:33:56CFmodHi, with regards to r19743 commit "Calculate watermark from bitrate and harddisk spinup time", how will that affect DAP's which have been modded with Compact Flash cards ?, flash cards don't have a spin up time, or do they ?.
11:34:01 Join casainho [0] (
11:34:04Jaykayas youre discussing about it... is now working correctly?
11:34:22casainhohello :-)
11:35:00archivatorlinuxstb: we need to keep the framebuffer to have something to draw once the light's on. But I see absolutely no reason why we should blit the buffer when the light's out. Disabling updates, yet keeping the buffer is what I think is best.
11:36:01casainhoĨ would like to start working on my hardware SD card ATA drivers... does anyone knows a player that have SD card, or can help me looking at his code? −− The rockbox firmware will be stored on SD Card on my player...
11:36:40scorchecasainho: look at the sansa devices...
11:36:55linuxstbarchivator: Well, you can still see the contents of the screen when the backlight is off on the Video...
11:37:22casainhoscorche: I have a sansa v1 :-) okok - any other
11:37:53casainhoany other player? - because sansa do not stores firmware on sd card...
11:38:01archivatorlinuxstb: how often do you find yourself doing that? The contrast is so low that you need an external flashlight to see anything.
11:38:11Jaykaylinuxstb: how about adding a setting?
11:38:43domonokycasainho: sansav2, and the sansa does store the firmware on sd, as the internal storage also is a sd card :-)
11:39:25 Quit CFmod ("CGI:IRC (Ping timeout)")
11:39:34casainhodomonoky: eheh - I didn't know ;-) - thanks :-)
11:39:38domonokycasainho: just take a look into the target tree, you will see several sd-drivers where you can get inspirations.
11:40:43 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
11:41:01casainhoon target/arm/sandisk/ I can't see any drivers for sd card...?
11:41:47_Auron_linuxstb: you can, but only if you have an extreme amount of light on it, at a certain angle (as the screen is rather reflective). without the backlight on it's next to impossible to see what's actually on the screen
11:42:04_Auron_there's really no point in updating the LCD if the backlight isn't on
11:42:09linuxstbarchivator: I don't use my Video any more, but if I was in conditions (i.e. well-lit) where I can see the screen without turning the backlight on, I didn't. That ability also depends on you using a high-contrast theme. Whether it's worth a setting probably depends on how much extra runtime it will actually gain you, rather than just speculation.
11:42:24casainhohmmm, maybe on arm/ata-sd-pp.c ....
11:46:11archivatorlinuxstb: I'll try and patch my rockbox and see what happens then..
11:46:11gfatherJdGordon can we overlay the text with images ?
11:46:26JdGordonlinuxstb: supposedly pretty good runtime... there are benchs from in the WPS and in the menu
11:46:36casainhoso, my sd card is wired on SPI bus of my MCU, I just need to develop drivers for spi_sd card or I will need more?
11:46:41JdGordongfather: you mean draw text over images? sure...
11:47:40domonokycasainho: the sd-driver needs to be able to read and write sectors. just take a look what the various ata-sd-xxx drivers do.
11:47:48 Join tvelocity [0] (
11:50:10casainhodomonoky: there is just one file, ata-sd-pp.c , right?
11:50:25domonokycasainho: no there are more sd drivers.
11:51:04linuxstbJdGordon: I know - but how much is the lcd_update() responsible for that, compared to the drawing?
11:51:58casainhodomonoky: ah, ok, for example, I saw now the driver for as3525 :-)
11:52:06 Quit nuonguy ("This computer has gone to sleep")
11:52:30JdGordonShould I do a bench with the backlight on in the menu and wps on my h300?
11:52:40Jaykaylinuxstb: extra runtime of the old patch was 1h 15min...
11:53:31Jaykaycan i do the benches? as you may have noticed i did a few others (i wrote in the dev-mailing list)
11:54:14scorcheyou both can ;)
11:56:06Jaykayjust say me what i should do (exactly)
11:59:05 Join HBK- [0] (
12:01:28 Quit HBK (Read error: 110 (Connection timed out))
12:02:33 Nick synergist is now known as _synergis (
12:02:40gfatheranyone know how to change the text with images
12:04:33kadobanthe actual targets don't support %f in snprintf :(
12:04:48JdGordonwe dont have floats.. so no need for it
12:05:27kadobanaren't floats ever used in settings and things like that?
12:05:49kadobanis there some nice way to print out a fixed point? :)
12:06:00JdGordonyes.. the manual way :p
12:09:41 Join Schmogel [0] (
12:12:15jhMikeSlet's see, the m200v4 wiki says it does NOT run, yet it's in the build table with a bootloader?
12:12:45domonokyjhMikeS: :-) it does run, but its not stable.
12:13:42jhMikeSI have one, that's why I'm looking.
12:14:31 Quit Schmogel (Client Quit)
12:14:49 Join Schmogel [0] (
12:15:50*domonoky also has one :-) At the moment it has a strange bug. If you turn the volume too high, it will just shutdown. :-)
12:16:34 Quit {phoenix} (Remote closed the connection)
12:16:34jhMikeSThat is pretty strange, it must be hardwired for ear protection :)
12:16:51linuxstbkadoban: Use something like printf("%d.%02d",x/100,x%100)
12:16:54 Join {phoenix} [0] (
12:17:25linuxstbkadoban: And stop using floats ;)
12:17:45domonokyjhMikeS: :-)
12:19:27 Quit Lss__ (Read error: 104 (Connection reset by peer))
12:19:30 Join ender` [0] (
12:19:50kadobani'm only using them because i thought that they were easy to print out...sim lies :( and it's in like, one place for a setting...i don't see the impact
12:20:08 Quit BigBambi (Remote closed the connection)
12:20:09jhMikeSdomonky: how do you get the firmware on the device or is everyone expected to figure it out themselves? :)
12:20:32kadobanbut i might as well change it if %f isn't supported i guess
12:20:39domonokyjhMikeS: take a look at the SansaV2 page
12:21:29Jaykayshall i do some benches? if yes please say me...
12:22:17jhMikeSSansaV2Firmware ?
12:22:34 Join BigBambi [0] (
12:22:41jhMikeSwoops, parent topic
12:27:30kadobanis there any way to tell if rb->set_int is cancelled by the user? i thought the first callback allowed it, but it turns out that that's only a minor bug in the simulator
12:27:31bertrikjhMikeS, I think it's a bit of a dilemma: we should make the installation clear enough for new ams sansa developers to get started, but obscure enough to prevent non-devs from installing experimental stuff and bricking their device/coming up with complaints about the port now being ready yet
12:29:43linuxstbkadoban: The impact is that your code will require the gcc floating point library to be linked - something I'm pretty sure doesn't happen in core Rockbox at the moment.
12:29:48 Quit BHSPitLappy ("Ex-Chat")
12:30:32kadobanlinuxstb: really? that would be bad...
12:31:09 Join at0m [0] (n=a548c80b@gateway/web/cgi-irc/
12:31:42 Join moos [0] (i=Mustapha@rockbox/staff/moos)
12:34:45jhMikeSbertrik: If it's so dangerous then perhaps it should be obfuscated for now
12:37:08*domonoky detects something strange on e200v2. at the moment when i enter the cpu-freq debug menu the first time, the boost_counter is rising and rising and rising...
12:37:39 Join mib_bhkiz6 [0] (i=c14b3728@gateway/web/ajax/
12:37:48 Join Darksair [0] (n=user@
12:48:13archivatorIs the iPod 5.5G LCD BCM 16-bit?
12:53:06***Saving seen data "./dancer.seen"
12:54:10 Quit casainho ("ChatZilla 0.9.84 [Firefox 3.0.5/2008121622]")
12:55:02bertrikdomonoky, it doesn't do that on clip. I suspect it's a incorrect button reading (e.g. up button) that causes it
12:55:37domonokybertrik: that may be it, the button driver for e200v2 is still not really working.
12:55:53 Join tyfoo [0] (
13:02:53 Join PaulJam [0] (
13:09:24archivatorIf I want to have the backlight controls in arm/ipod/video/lcd-video.c, do I include "backlight.h" or backlight-target.h"? In fact, is that considered breaking the encapsulation?
13:10:56*jhMikeS wonders where info is on bootloadering the sa9200 is as well
13:14:38 Join miepchen^schlaf [0] (
13:15:35 Quit Jaykay (Read error: 104 (Connection reset by peer))
13:16:01 Join Jaykay [0] (
13:18:21linuxstbarchivator: Why do you want the backlight controls in lcd-video.c? You could just use the same "lcd_enable" mechanism used on other targets, but with the actual disabling doing nothing (apart from preventing lcd updates). i.e. add HAVE_LCD_ENABLE to firmare/export/config-ipodvideo.h and implement the lcd_enable function (e.g. copy firmware/target/coldfire/iriver/h300/lcd-h300.c)
13:20:08 Quit bmbl ("Woah!")
13:20:11 Join bimbel [0] (n=Miranda@unaffiliated/bmbl)
13:20:36 Quit bimbel (Remote closed the connection)
13:21:24archivatorlinuxstb: Well, that just shows how little I know about the other targets... Wouldn't that require an additional flag somewhere that would be checked by lcd_update_rect? I.e., where should I put that flag?
13:21:28 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
13:21:48 Quit ender` (Read error: 54 (Connection reset by peer))
13:22:20 Quit _Auron_ ("Infinity repeatedly denies rumours of plotting with zero to bring down the Universe.")
13:22:21*bertrik vaguely remembers about having to install a callback function when disabling lcd updates
13:23:47 Quit bmbl (Client Quit)
13:24:13 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
13:24:43 Quit Horscht ("User was distributing pornography on server; system seized by FBI")
13:27:17archivatorOkay, I did the simplest thing possible - included "backlight.h" and return'd from lcd_update_rect if the light's off. It works as expected. It is basically FS #8523's older brother :) The only thing is, now plugins are supported as well.
13:28:39 Join AndyIL [0] (i=AndyI@
13:30:28linuxstbarchivator: Look at lcd-h300.c - it should be clear... There is just a local "bool display_on" variable in that .c file.
13:31:58archivatorlinuxstb: just out of curiosity, does this have any chance of being accepted? I don't think any of the other targets "fake" behavior in such a way..
13:32:36amiconnImo this should not go into trunk
13:33:00amiconnShowing outdated content is worse than showing nothing. Users won't expect this
13:34:35 Quit sbhsu (Read error: 60 (Operation timed out))
13:34:37archivatoramiconn: not even if it's a setting?
13:34:52 Join Thundercloud [0] (
13:34:54 Join _Auron_ [0] (
13:35:04 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
13:37:00amiconnThere really is no reason.
13:37:49archivatorWell, the WPS patch showed improved battery life. I think the generalized version can only do better.
13:38:52 Quit advcomp2019 (Read error: 60 (Operation timed out))
13:39:27kadobanarchivator: with your patch, doesn't the back-buffer still get drawn to? as i understand it, the patches are complementary, not one a generalized version of the other
13:39:27 Join ThaCrip [0] (
13:39:46ThaCripi was curious as to how this "2009-01-11: Battery charging on Sansa e200v1/c200v1." works?
13:40:02 Quit AndyI (Read error: 110 (Connection timed out))
13:41:14archivatorkadoban: correct. The lcd-video.c version has the same effect as the WPS patch. The mechanism, however, is different. That's what I meant when I said "generalized".
13:41:44ThaCripanyone know anything on that at all?
13:41:56ThaCripcause it appears it was added into recent build according to that
13:42:00archivatorkadoban: as for the back-buffer, yes, it's still drawn to to allow immediate recovery. That can easily be changed, though.
13:42:10 Join pyro_maniac [0] (n=jens@
13:42:23jhMikeSThaCrip: it's new. it should charge the battery.
13:42:27kadobanarchivator: but has it been tested to see if the effect on runtime is the same? i somehow doubt that it is
13:42:35ThaCripso it's 'safe'?
13:42:57ThaCripand will rockbox boot... or are they not planning on changing boot loader etc?
13:43:03jhMikeSThaCrip: It's safe, yes. It basically turns the charger on and off and the rest is hardware controlled.
13:43:15ThaCripcause right now if player is off and u plug in USB cable it will boot to default of Sansa firmware
13:43:27archivatorkadoban: I frankly do not have the nerves to do the benchmarks. I don't see any reason why the effects shouldn't be the same, though. It's the actual drawing that takes power, not the CPU cycles..
13:43:31ThaCripahh i c... so it would be identicle as the sansa firmware
13:43:40jhMikeSThaCrip: That's avoidable by holding the correct button for the player
13:43:54ThaCripwas that the center button?
13:43:56ThaCripi cant recall
13:44:04jhMikeSOn e200, yes
13:44:12ThaCripok that's what i have
13:44:18ThaCripe250 v1
13:44:42ThaCripi just wonder if there going to change the bootloader in the future now that it supports charging?
13:45:14ThaCripsince we dont have a real need for original firmware now if the charging works... o wait, mevermind ikeep forgetting about file transfers lol
13:46:22 Quit mib_bhkiz6 (" ajax IRC Client")
13:46:32ThaCripbut once the file transfers get sorted out.. i would assume then it would make more sense to boot rockbox over the standard sansa when connecting it to the pc
13:46:37ThaCripbut thanks for the information ;)
13:46:57ThaCripill probably just hold off til March 23rd when the next stable rockbox will be availible ;)
13:48:02 Quit BigBambi (Read error: 104 (Connection reset by peer))
13:48:32JdGordonwhat makes you think rockbox usb will be working by then?
13:48:34archivatorIs there something one abstraction level above lcd_update_rect that allows me to change the behavior on multiple targets at once?
13:49:02ThaCripim not saying the USB will be... i was just saying since the only major update is charging ill probably just stick with v3.1 final for now ;)
13:49:21ThaCripsince u still need to boot original firmware in general anyways for file transfers
13:49:32ThaCripso i might as well just hold off on updating then
13:50:15ThaCripi assume adding in USB support (i.e. file transfers) is pretty hard since they aint did it by now
13:51:25ThaCripbut if the file transfers support ever gets added in... then there will be pretty much no reason to boot to original sansa firmware ;)
13:52:25 Quit gfather1 (Read error: 110 (Connection timed out))
13:53:33ThaCripbut thanks for everyone's time. later ;)
13:53:38 Quit ThaCrip ()
13:54:14archivatorNever mind, I'll just do the benchmarks on my iPod first, If it turns out it saves power, I'll go with the idea...
13:54:21 Join stripwax [0] (
13:56:06 Quit archivator ("""")
13:58:57 Join massiveH [0] (
13:59:13 Quit massiveH (Read error: 104 (Connection reset by peer))
14:00:51PaulJamlittle question concerning the earlier discussion about rejecting unknown WPS tags: wouldn't that make WPS incompatibele across targets that share the same display resolution, but don't have the same WPS features? for example the H300 has a remote hold button (%mr) but the ipod color doesn't. so wouldn't a WPS using %mr be rejected by the ipod color in that case?
14:07:49 Quit Schmogel (Read error: 104 (Connection reset by peer))
14:14:52 Quit bmbl ("Woah!")
14:17:24 Quit Darksair ("Emacs = ESC-Meta-Alt-Ctrl-Shift")
14:19:06*domonoky finds the e200v2 wheel very strange. it seems to only work when you press and hold "<<" or ">>" while turning the wheel. And also that only works, when i dont printout dbop_in. *scratches head*
14:19:15JdGordonPaulJam: ah yeah.. good point
14:19:53JdGordonPaulJam: although... apart from RTC (which is enabled for all targets anyway) which tags might be effected?
14:20:20JdGordondo hwcodec share lcd size with any swcodec targets?
14:21:08 Join ender` [0] (
14:30:16pixelmanot that I know of. Other tags might be RTC and remote hold (already mentioned) or the virtual LED tag
14:30:56 Quit japc (Remote closed the connection)
14:31:02pixelmae.g. the H100 does not have it because it has a real LED
14:31:28JdGordonrtc is handled on non rtc targets... I noticed remote hold and it is there for all targets also (as a NO_TOKEN)
14:31:28 Join TheSphinX^ [0] (
14:31:43PaulJambut the tag seems to be still enabled, because the h100 has a LCD remote.
14:31:55pixelmaand shares screen specs with the M5 and the greyscale Ipods (except Mini)
14:32:02dionoeajhMikeS: is there something special to enable at compile time to have charging on an e260v1?
14:33:04JdGordonpretty easy to figure out which might be problems... bc (battery charging), bp (charger connected)
14:33:23JdGordonSp (sound pitch on all but 1 mas target)
14:33:43dionoeaI also have to figure out the hotkey to prevent it from switching to USB mode
14:33:54JdGordonvled, peakmeters (although not really.. its bitmap lcd or charcell)
14:34:47JdGordonother than that, its AA, and database tags which are only disabled on custom builds
14:35:10JdGordonno.. missed 2 swcodec tags... rg and xf
14:35:35pixelmaah yes, %X - but that breaks WPSs already (try loading a backdropped 160x128 WPS in an M:Robe100 sim
14:35:47 Join B4gder [241] (n=daniel@rockbox/developer/bagder)
14:36:52pixelmathe one MAS target that does not have sound pitch is the Player, so no WPS sharing anyways
14:36:59JdGordonso, the question is... do we really want to make the parser this strict? and should that %X be fixed?
14:37:00jhMikeSdionoea: it's enabled now if it's a version 1
14:37:24pixelmaJdGordon: how would you fix %X?
14:37:37JdGordonignroe the tag
14:37:49JdGordonskip the line
14:38:01JdGordonif thats the only thing that makes those wps' not work
14:39:18pixelmadepends on whether the WPS has album art - if it does there would be a second thing...
14:40:12dionoeajhMikeS: and what key do I need to hold to keep if from going into USB mode?
14:40:35dionoeaok, thanks.
14:42:12pixelmaI don't know, haven't thought about very deeply but I don't think it's necessary to make the parser reject those kind of tags if they don't make the WPS look differently. I'm not sure if they do (show up as text in the WPS) or if they are just ignored currently.
14:42:20JdGordonpixelma: AA is easy to fix also... we can make it always think there is no aa
14:44:11 Join kugel [0] (n=chatzill@unaffiliated/kugel)
14:53:09***Saving seen data "./dancer.seen"
14:58:19 Join ajb [0] (
14:59:12JdGordonis there much benefit on making sure all tags "work" on all targets?
15:00:38 Quit stripwax (Read error: 54 (Connection reset by peer))
15:00:39domonokyJdGordon: the benefit is, that we can keep ordering the themes by screensize and not by target.
15:00:46gevaertsAs long as WPSes are classified (only) by screen size, yes. I'm not convinced that this is a good idea though
15:01:57gevaertsA theme that makes full use of e.g. remote related tags is likely to be confusing on targets without remote
15:02:37JdGordonare there any remote related tags?
15:02:55JdGordonremote hold is the only one
15:03:10*gevaerts isn't sure
15:03:35gevaertsHowever I think just one is enough to make it confusing
15:04:03 Join stripwax [0] (
15:04:34JdGordonit looks like there are 5 which definetly need fixing, and another 4 maybes
15:06:11bertrikThe cabbie v2 128x64x1 wps is used by the iriver remote but also by the clip. It doesn't look good on the clip because some text crosses the yellow/cyan border. So this wps may need to get 2 versions for a single screen size (one for iriver remote and another for clip).
15:07:08gevaertss/may need/needs/ imho :)
15:09:04 Quit gfather (Read error: 104 (Connection reset by peer))
15:09:23JdGordonyeah, well the clip is hopefully going to be by itself with that wierd screen :)
15:13:03pixelmabertrik: yes, but you need to make the script able to tell the two apart and pack the right thing into the zips
15:14:09JdGordonor.. could it be fiddled with so it fits correctly into the clips screen and still looks good on the remotes?
15:14:54pixelmathe 128x64x1 WPS is made for the 12-Adobe-Helvetica font, not for the 8-pixel tall one that's used on the Clip
15:14:55gevaertsMaybe, but the same problem is (potentially) there for all included wpses
15:16:53pixelmathe script is already able to pack different 128x64 WPS into the H300 or H100 zip depending on the main screen's resolution, then the WPS file bears the two resolutions in its name (see iCatcher for example)
15:20:35JdGordonhang on... why is the %X a problem for the mr100? isnt that screen size issue not a problem because the bpp is different?
15:20:44ajbUnhelpful: I've tweaked FS #9758 so a SVN rNNNNN is reported if HEAD is a git-svn commit
15:21:15pixelmaJdGordon: no backdrop on a monochrome screen
15:21:50pixelmanot a problem for the shipped themes but in case someone wants to try H100 themes or so
15:22:22 Quit stripwax (Read error: 104 (Connection reset by peer))
15:24:59 Join stripwax [0] (
15:25:19 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
15:27:03 Quit bmbl (Client Quit)
15:27:09 Quit Thundercloud (Remote closed the connection)
15:27:15 Join Thundercloud [0] (
15:27:43 Join gfather [0] (n=gg@
15:29:01 Quit Zambezi (Remote closed the connection)
15:29:51 Join DerDome [0] (
15:30:09bertrikpixelma, I wasn't initially aware of the font size difference. What do you think is the best solution for the two different 128x64x1 wps's, update the packing script?
15:30:17 Join xorAxAx [0] (n=xorAxax@moinmoin/coreteam/alexander)
15:30:30 Quit B4gder ("It is time to say moo")
15:30:42xorAxAxhmm, has anybody ever figured out why self-destruction runs on my ipod 1st gen device if the battery ran empty and i need to reset the whole thingie?`
15:31:17xorAxAxor better, did apple fix that in newer devices? :)
15:31:26xorAxAxs/ipod/& nano/
15:32:34 Join Zambezi [0] (i=stolgfor@
15:32:38 Join itcheg [0] (i=62db4767@gateway/web/ajax/
15:33:09 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:34:16 Join __lifeless [0] (n=lifeless@
15:34:21 Quit stripwax (Read error: 104 (Connection reset by peer))
15:34:39pixelmabertrik: I'd hope so, yes
15:35:27 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
15:36:09 Nick fxb__ is now known as fxb (
15:36:23 Quit itcheg (Client Quit)
15:37:39 Quit gfather ()
15:39:28 Join itcheg [0] (i=62db4767@gateway/web/ajax/
15:41:28 Quit {phoenix} (Read error: 110 (Connection timed out))
15:42:55 Join stripwax [0] (
15:49:44 Quit Thundercloud (Remote closed the connection)
15:50:55 Quit _lifeless (Read error: 110 (Connection timed out))
15:50:55 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
15:51:27 Join {phoenix} [0] (
15:53:17 Part pyro_maniac
16:03:08jhMikeSI wonder what was broken on a beast single-boot bootloader since the one I just made and tested worked.
16:08:30 Join pyromaniac [0] (n=jens@
16:10:08 Quit pyromaniac (Nick collision from services.)
16:13:22 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
16:15:45 Join pyro_maniac [0] (n=pyro@
16:16:04 Join mirak [0] (
16:24:59 Quit pyro_maniac (Remote closed the connection)
16:26:40 Join aurix_lexico [0] (
16:26:53Unhelpfulajb: it's probably going to need to report the most recent git-svn commit regardless of whether it's HEAD or not. people want it to always show where the build fits in the svn repo, and at least some are against even appending a git commit ID.
16:30:35 Join MarcGuay [0] (
16:35:41 Quit gromit`` (Read error: 104 (Connection reset by peer))
16:37:31 Join gregzx [0] (
16:38:42 Join pyro_maniac [0] (n=pyro@
16:48:10 Join gregzx_ [0] (
16:48:43 Quit gregzx (Nick collision from services.)
16:48:45 Nick gregzx_ is now known as gregzx (
16:49:04 Quit nibbler__ (Read error: 113 (No route to host))
16:49:27 Quit miepchen^schlaf ()
16:53:13***Saving seen data "./dancer.seen"
17:04:45 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
17:07:07 Quit bmbl ("Woah!")
17:18:32 Join tessarakt [0] (
17:19:09 Join nuonguy [0] (
17:20:25 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
17:21:29 Quit pyro_maniac (Remote closed the connection)
17:21:38 Join gfather [0] (n=g@
17:21:51 Quit gfather (Client Quit)
17:21:55 Join gfather [0] (n=g@
17:23:07 Join pyro_maniac [0] (n=jens@
17:23:56 Nick gfather is now known as mjfahmi (n=g@
17:25:43 Nick mjfahmi is now known as gfather (n=g@
17:25:48 Join miepchen^schlaf [0] (
17:28:27 Quit at0m ("CGI:IRC 0.5.9 (2006/06/06)")
17:30:56Unhelpfulamiconn: what about unifying the plugin and codec headers structures? i notice that is shared with codecs, and the codecs and plugins have most of the same header structure already. the anonymous union method works fine inside structures, so they could have the same header structure, with codec_entry_point called for codecs and plugin_entry_point for plugins.
17:31:40gfatherstrange that rockbox in not linux kernel ?
17:32:03bertrikno, it's not strange
17:32:08 Join Thundercloud [0] (
17:32:37gfatherif rockbox had linux kernel it would have more support
17:33:26 Quit merbanan (Read error: 110 (Connection timed out))
17:33:53gevaertsMore support for what?
17:34:01gfatherfor rockbox
17:34:08PaulJamas much support as ipodlinux?
17:34:30gfatherno , im not saying we should do ipod linux
17:34:38*gevaerts has no clue about what gfather means
17:34:45gfatherim only saying it should be linux kernel
17:34:57Unhelpfulyes, but you can't seem to give a good reason why?
17:34:57krazykitgfather, perhaps you don't understand what rockbox is, exactly?
17:35:14gfatherbecouse im thinking about rockbox player project
17:35:41gfatherif it linux kernel based , worrying about driver and such for embeded devices is littel
17:36:04gfathermostly becouse all teh drivers and such is already available and/or can be easlly done
17:36:15Unhelpfulthe rockbox devs seem to do just fine writing their own drivers - and on the devices where that's a problem, linux would be no help, either.
17:36:33Unhelpfulthere is no linux driver for the buttons and wheels on the AMS sansa
17:37:20gfatherim not saying about that , im talking about , cpu , chip , ram , and those stuff ,
17:37:46gfatherit can have better preformance for already tested and developed drivers
17:37:55gevaertsWhy do you think linux has more or better drivers for this sort of hardware?
17:38:23gevaertsAnd why would linux drivers have better performance? I'd expect the opposint...
17:38:24gfatherand any unix programer , could easlly do allot of stuff , that can be helpfull later , new ideas , improvments and such
17:38:47Unhelpfulbut most of the rockbox API imitates POSIX already, where it's appropriate
17:39:07 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
17:39:43gfathergevaerts dont think as it the same kernel that runs ubuntu or debian , the linux kernel is clean and not patched with anything , that kernel should be running rockbox
17:40:43gfatherUnhelpful but also when u want to represent rockbox with its own kernel for a venture to invest in it , its not gonna give the same effect when presenting it as a liinux kernel
17:40:46gevaertsgfather: obviously. We'd be talking about uclinux here...
17:41:04bertrikthe clean linux kernel requires an MMU if I recall correctly, which most rockbox targets don't have
17:41:37gfatherbecouse a venture would think of what his future invesment can be
17:42:02gevaertsBuilding a custom player is not in any way the primary goal for rockbox...
17:42:15Unhelpfulrockbox is not about rockbox player. it's primarily about alternate firmware for existing devices.
17:42:29gfatherim not thinking of it as for the normal players we are using now , but for future builds and stuff
17:42:59 Join control_your_pla [0] (n=549a5159@gateway/web/cgi-irc/
17:43:12gfatheryah its not only about the rockbox player
17:43:26gfatherbut it would be better for rockbox future
17:43:48gevaertsSo you want to use linux for marketing reasons, and you're trying to come up with technixal sounding arguments for that
17:44:02Unhelpfulpretty much, it sounds like.
17:44:09gfatherand yes , rockbox is the best palyer i ever wished for , and its great as it is
17:44:18gfathergevaerts , no
17:44:22gevaertsIt is that partly because it doesn't use linux...
17:44:28gfatherur talking about linux kernel as its a bad idea
17:44:32gevaertsIt is
17:44:44gevaertsAlso, please use real words
17:44:54Unhelpfulit's not the right solution for all problems.
17:45:14ThundercloudLinux is not the right solution for my pet rabbit.
17:46:09gfatheryes , its not , but its great for embeded devices , and if ur guys here dont even care to think for 5 minutes about a suggestion to rockbox , then its a waste of time to argue with u
17:46:24gevaertsPlease use real words
17:46:36gfatheryah yah
17:46:38ThundercloudRockbox has been going for quite a while, I seriously SERIOUSLY doubt that the team hasn't took linux into consideration at least once.
17:47:38jhMikeSgevearts: I think it's about the "ooh, ahh, we have linux on this little thing." fetish but I could be wrong. I doubt it though.
17:47:51gevaertsCan you run linux on a device with 2MB of RAM and have between 1 and 1.5MB left over for buffering?
17:47:53Unhelpfulseriously, do you really think that rockbox devs have never though about linux? i use it exclusively as a development platform, and i believe many others as well.
17:48:06gfatherThundercloud as someone said before here , that the team is doing a great job with it , but thats the problem , becosue its a custom kernel , not that much of people could give it driver support and such
17:48:40gfatherso the team is mostly doing every thing , or most of the jov
17:48:50gevaertsAs we would with linux...
17:49:06ThundercloudJust because you're using linux does not magically entitle you to loads of driver support.
17:49:21gevaerts"linux" is not a magic word that makes drivers appear out of thin air
17:49:24gfathergevaerts yes , its done with 2 mb of ram , and with lots of stuff
17:49:34jhMikeSWhy do we need driver support when most drivers ain't that hard anyway
17:49:46*_Auron_ smells an extremely ignorant person
17:50:00*gevaerts decides to leaves the discussion
17:50:11*jhMikeS wonder who's being called ignorant
17:50:26gfatheryes , it seems everyone here thinks they are very smart ]
17:50:34 Quit itcheg (" ajax IRC Client")
17:51:14*Unhelpful wonders if a linker script could be used to define a custom section and to allocate a few symbols, while still creating a platform dynamic library... without having to duplicate the existing linker script for such libraries entirely...?
17:52:18gfatherwell after i talked to the people here , i dont even care what happens to rockbox , seems people here now everything ,
17:52:30 Part gfather i have to continue restraining the laughter, now that he's gone? ;)
17:52:56_Auron_the smell went away, I can breathe again. :X
17:52:59ThundercloudIs it just me or does this channel regularly get these noobs?
17:53:47 Join fdinel [0] (
17:53:58ThundercloudAnd also, for the record, is one of the main reasons that a custom kernel was built was for extreme speed given the low specs of a lot of DAPs? :P
17:54:21Unhelpfulactually, memory is probably a larger concern than CPU usage, on many of them.
17:54:32ThundercloudWell yes
17:54:44ThundercloudI doubt a kernel, even a uclinux kernel, could boot and run on 512kb of RAM
17:57:51ThundercloudBesides, if a switch to linux was made, wouldn't the emphasis all of a sudden switch from getting rockbox's kernel to work on a given processor to getting linux to work on a given processor? :P
17:58:19ThundercloudGiven the huge array of processors, input devices, screens, etc, it's sheer folly to expect drivers to exist for the majority of them.
17:58:22domonokyalso a linux kernel would waste too much ressources for thing we dont need.
17:58:25ThundercloudYes precisely
17:58:28Thundercloudlinux kernel is multi-purpose
17:58:41ThundercloudSo in the end the entire argument boils down to: "Well it'll achieve precisely nothing"
17:58:43jhMikeSThe kernel hardly needed anything code wise to get onto the MIPS even though it wasn't supported.
18:01:06MarcGuayjhMikeS: Any thoughts on my beast problem? It's still crunked at error#2.
18:01:48jhMikeSMarcGuay: I just built and ran single-boot just fine.
18:01:49MarcGuayI don't imagine replacing the nk.bin could screw up their recovery mode?
18:02:18Unhelpfuli wonder if it's something that the official updater causes...
18:02:26jhMikeSRecovery doesn't reform the firmware partition. I've recovered with all my extra files in tact there.
18:02:29Unhelpfuli've never used a windows updater on mine
18:02:33MarcGuayUnhelpful: I haven't run it yet.
18:03:02jhMikeSI keep a bunch on nk.bin version on it.
18:03:04MarcGuayI've been trying to use it to recover from whatever happened when I put the RB bootloader on, but it crashes.
18:03:09Unhelpfuloh, i thought you had. davide-nyc did, and he gets error #2 after booting rb bootloader once. :/
18:03:45MarcGuayUnhelpful: I think he's managed to recover his at least... :)
18:03:46jhMikeSUpdater is corrupting firmware partition?
18:04:08 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
18:04:17 Quit Xerion (" ")
18:04:28UnhelpfulMarcGuay: maybe you should talk to him... perhaps the solution is the same, since we don't really know what the problem is in either case. :/
18:04:31*jhMikeS wonders what steps went into initiating the problem exactly
18:04:39 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
18:05:15MarcGuayI never updated the OF and had the same problem as him. I simply replaced the nk.bin while connected via USB in rockbox, updated the build, and restarted.
18:05:15saratogafinally sansa charging support! only took a year
18:05:53 Part xorAxAx ("I am a partmsg virus! Copy me into your configuration! X5O!P%@AP[4\PZX54(P^)7CC)7}$EICAR-STANDARD-ANTIVIRUS-TEST-FILE!$H+H*")
18:05:54Horschtbtw: I started a conversation in here a few days ago regarding the tag cache. having dircache enabled reduces the updating of the tagcache (ipod 5.5G 80GB) from ~30 minutes to ~10 minutes
18:05:59Unhelpful"same problem" but we really only have a symptom. we don't know the cause. :/
18:06:04jhMikeSMarcGuay: I dump nk.bin on there quite often, from Windows anyway.
18:06:06 Join merbanan [0] (n=banan@
18:06:30HorschtIt would still be cool to have a small commandline app that can build the tagcache while the player is connected to the PC
18:06:30MarcGuayI'm using Windows as well.
18:06:34mooshi there, I encountred this weird recovery mode bug, it annoyed me few times.
18:06:58MarcGuaymoos: How did you recover it?
18:07:01moosI installed via windows btw
18:07:03jhMikeSmoos: Then you take it from here :). I'm not sure how you got out of it and I haven't gotten it.
18:07:10saratogahas anyone familar with buffering looked at FS #9775 ?
18:07:21moosjhMikeS: mhehe :)
18:07:28Horschtbecause 10 minutes is still kinda long :)
18:07:35MarcGuayjhMikeS: Don't forget your life-jacket. :)
18:07:42 Quit tessarakt (Read error: 110 (Connection timed out))
18:07:49moosMarcGay: the vupdater caused me problems and file not found at the end
18:08:08 Join tessarakt [0] (
18:08:26moosfor some reason I managed to use vpatcher just for recover the OF nk.bin
18:08:44moossendfirm fixed the issue
18:08:47domonokyHorscht: there is the beginning of such a tagcach pc tool in svn, but nobody finished it *hint hint* :-)
18:08:52MarcGuaymoos: I have an old dual-bootloader nk.bin I could use. Crap, sendfirm crashes for me...
18:09:03Horschtdomonoky, i can "finish" it.
18:09:05jhMikeSmoos: I think even after the .tar fix, and after a boot cycle the ROM loader starts complaining all of a sudden.
18:09:13Horschtbut it would probably break the PC :D
18:09:19Horschtcause I can't code
18:09:28UnhelpfuljhMikeS: you were one of the others looking at #9770... is there anything else you'd add on top of Jens' version of it, aside from possibly cleaning up remaining uses of passed API pointer?
18:09:52jhMikeSUnhelpful: I'll have a look. What remaining issues btw?
18:10:03moosMarcGayI noticed that when the bootloader failed to found rockbox file when OF recovery missed to creat OF files on the partitions
18:10:20moosi.e: no content dir et all
18:10:42Horschtbut it's nice to hear that there are beginings of such a tool already. Here's hoping someone with the skills will pick it up :D
18:10:56mooswhen I saw those files again, I was sure rockbox will work again :)
18:11:28Unhelpfuli don't have any in specific. if somebody comes up with a cleaner solution for setting the API pointer, that'd be a nice-to-have, but i don't think it's a blocker. the only things left taking API pointers as arguments right now are macros.
18:11:30MarcGuaymoos: Hmm. I can't get past the error#2, hence no MSC mode...
18:11:31moosMarcGay: note that I use a recent singlebootloader
18:11:51moosMarcGay: aie, weird :(
18:12:04saratogaif you screw up the nk.bin on a gigabeat, is it recoverable without removing hte hard disk?
18:12:08jhMikeSplugin_iram_init in pluginlib? hmmm. It might have gone in the core because there didn't seem to be a place to stash it at the's been like two years or more. :\
18:12:18jhMikeSsaratoga: indeed
18:12:21MarcGuaysaratoga: I'm learning the hard way that it isn't.
18:12:38jhMikeSeh? I've messed it up plenty.
18:12:44moosMarcGuay: you canot recover the OF then?
18:12:59saratogajust wondering since I picked up one of the flash based gigabeat Ts, which also uses the nk.bin and mx31 cpu
18:13:03MarcGuaymoos: Nope. sendfirm and update utilities crash when I connect the player.
18:13:12MarcGuayCould be a problem with MTP, of course.
18:13:24Unhelpfulwell, again, that's probably another patch. moving that, creating pluginlib functions to replace function wrapper macros, they're all trivial, and can quickly follow global API. i don't think i have any issues left with global API itself.
18:13:45moosMarcGuay: tried on an other PC?
18:14:04MarcGuaymoos: Not yet.
18:14:17mooslet's hope that will be better then :(
18:14:19saratogawas considering trying the rockbox bootloader to see if I could get UMS mode
18:14:25jhMikeSUnhelpful: Well, there is the audio stopping procedure and such that if a change is needed there, all plugins need to be incompatible.
18:15:00Unhelpfuloooh... yes, that's a good point.
18:15:15jhMikeSWe can also be sure we're not messing around in the .bss sections while doing that.
18:15:35MarcGuaymoos: What version of WMP do you have installed?
18:16:38Unhelpfulperhaps just mention Jens' last version of the patch to the ML, and maybe get some comments from more than 3 people before making a decision?
18:17:29moosMarcGay: I'm using vista, WMP 11 I think
18:21:54jhMikeSUnhelpful: Perhaps. This is getting more involved than I expected it would. Yeah, you have four bytes for a pointer, but you save extra code passing/reading api parameter and saving it.
18:23:37Unhelpfulwell, i think you're likely right about sim/target disparity, so i've been going forward with jens' version. i haven't found anything to add to it, and any broken plugins i had are fixed since fixing the iram issue.
18:24:02Unhelpfulthe overlay loader still needs to be tested, and right now, that means somebody with an archos.
18:24:28jhMikeSUnhelpful: I think killing the parameter from the wrappers and macros is reasonable though. They do expect to call core functions anyway.
18:24:33Unhelpfulmoos: mind risking life and limb, or at least risking crashing chessbox? ;)
18:25:03UnhelpfuljhMikeS: i'll add that and turn out a new patch. at least that way, the one we're talking about is at the bottom of the task ;)
18:25:50moosUnhelpful: depending... :)
18:26:12 Quit tessarakt (Success)
18:26:19jhMikeSUnhelpful: I must admit I'm not familiar with this "overlay loader" for archos...why it's needed and what it does for a plugin.
18:26:39 Join tessarakt [0] (
18:26:44UnhelpfuljhMikeS: it allows you to use the audio buffer as a substitute for the plugin buffer, as i understand it.
18:27:21Unhelpfulnot just for more data, but for code as well
18:27:39Unhelpfulrockboy, chessbox, and zxbox support linking for loads into the audio buffer
18:29:47 Join nibbler__ [0] (
18:30:12jhMikeSsort of a plugin within a plugin then (sub-plugin?)? :)
18:30:59Unhelpfulbasically, yes, it's a stub plugin that stops playback, grabs the entire audio buffer, and duplicates the plugin loader to load a plugin into it... or at least, that's how i understand it to work.
18:35:56 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
18:36:44Unhelpfulmoos: basically, it changes how plugins access the rockbox API - the passed pointer is replaced with a loader-initialized global pointer, that all plugins and pluginlib functions are converted to use. it works with everything i've tried, but the overlay loader is an archos-only feature.
18:39:19moosUnhelpful: don't have my ondio with me here (the only archos target I own) sorry I'm Unhelpful for this :(
18:39:53Unhelpfulah... well, thanks anyway. i can always pester somebody else. ;)
18:40:59moosno problem, sure you will can catch pixelma or amiconn for this
18:42:52jhMikeSUnhelpful: I can dig out my player to check it
18:43:23Unhelpfulplayer's charcell, isn't it? do any of those run on it?
18:44:02*domonoky checked his archos player, but non of this plugins are available :-)
18:44:31jhMikeSDon't really know, I guess I should see first
18:44:42Unhelpfuli guess i could just crank out a build and see ;)
18:46:31Unhelpfulno good
18:53:16***Saving seen data "./dancer.seen"
18:53:33amiconnNo overlay plugin on the player so far
18:54:08amiconnAlthough I think that chessbox could work.
18:54:26amiconnI know there are chess cracks which don't need a board for playing. It would just display a list of moves
18:55:53amiconnThe overlays are just plugins which are linked to a different address, as far below the normal plugin buffer that it ends as close before that buffer as possible
18:56:48amiconnAll overlay plugins have a corresponding loader plugin, which gras the audio buffer, and then loads the overlay the same way as an ordinary plugin is loaded by the in-core loader
18:57:11amiconnIt checks memory boundaries and refuses to load if the necessary memory area is not fully available
18:57:56amiconnRight now overlays have their own .lds files, which are archos only.
18:58:18Unhelpfulright... the only change in there is to set the API pointer via the plugin header, as the core loader does. i just want to make sure that it actually works correctly.
18:58:22 Join Schmogel [0] (
18:59:00amiconnBut since there are now other lowmem targets which might want to use overlays too, I think it's worth changing it so that can also be used for overlays
18:59:28amiconnThen the plugin specific .lds would just define the overlay size (per target as necessary) and #include
19:00:05amiconnA side effect would be that the overlay mechanism could be tested on large-mem targets if desired
19:00:12Unhelpfuldoes the overlay .lds need to be plugin-specific? they all load into the same buffer on each target, right?
19:00:15 Join nibbler_ [0] (
19:01:38 Quit nuonguy ("This computer has gone to sleep")
19:02:46amiconnIt needs to be plugin specific, at least if the plugin itself needs further audio buffer memory
19:03:25amiconnAn unnecessary large overlay buffer lowers the amount of available audio buffer, so that the buffer should be just large enough for the plugin to fit
19:05:29Unhelpfulthanks, i was wondering why we had 3 that explains it quite nicely.
19:06:18 Join Xerion [0] (
19:07:17jhMikeScouldn't a global variable be declared inside the lds that gives the start of available remaining buffer?
19:08:35Unhelpfulsimilar to how we mark the start and end of ibss for plugins using iram?
19:09:34amiconnThe remaining buffer is *before* the overlay start address, not after
19:10:01jhMikeSwhat does that do?
19:10:04amiconnAnd the overlay plugins cannot be put at the start of the audio buffer, because that start is variable. The end is fixed.
19:10:50jhMikeSor the bufalloc stuff?
19:11:09Unhelpfulcan the .lds support a fixed end address? i'm guessing no?
19:11:17jhMikeSeeek, nvm firmware size sets that.
19:11:51dionoeahum, rockbox policy is no tabs right?
19:12:25amiconnjhMikeS: The start of the free buffer varies at runtime, depending on what limits are set, whether .cue support, scrobbler support etc are enabled ...
19:13:21bluebrotherdionoea: yes (except Makefiles of course ;-)
19:15:15 Join itcheg [0] (i=62db4767@gateway/web/ajax/
19:15:30linuxstbdionoea: See docs/CONTRIBUTING
19:16:12 Join BigBambi [0] (
19:16:13 Quit nibbler__ (Read error: 110 (Connection timed out))
19:16:24jhMikeSamiconn: well, yeah, it's fw + other stuff grabbing buffer mem at runtime
19:17:08amiconnAnd overlays cannot be loaded if the start of free audio buffer is higher than that overlay's load address
19:17:27amiconnAnother reason to keep them located as high as possible
19:17:40 Quit miepchen^schlaf ()
19:18:06Unhelpfuland the lds can't grow sections "down" from a fixed end address?
19:18:44amiconnnot that I know
19:19:09amiconnIt would be nice if it could...
19:20:04Unhelpfulthat would give us one overlay lds for all plugins - although it might no longer be feasible to include the plugin lds any more
19:20:08amiconnPerhaps we could use a trick: first linking to a fixed start address, and then analyzing the .map with a script which calculates the start address for the second linker run
19:21:08Unhelpfulwhat if the plugin needs to ask for buffer to use? or does that come from the audio buffer, which is in front of it?
19:21:44amiconnIt can use the audio buffer, but has to make sure it doesn't overwrite itself
19:21:51 Join nibbler__ [0] (
19:22:05amiconnIt can also use plugin_get_buffer, which returns the rest of the standard plugin buffer
19:22:23amiconnThe standard buffer isn't used much - the loader plugins are less than 1KB
19:22:32Unhelpfulwhich will be nearly empty, since it contains only the overlay loader stub
19:23:28 Quit aurix_lexico (Remote closed the connection)
19:23:28 Quit CaptainKewl (Read error: 104 (Connection reset by peer))
19:23:30 Join CaptainKwel [0] (
19:26:41 Quit nibbler_ (Read error: 113 (No route to host))
19:29:36 Quit itcheg (" ajax IRC Client")
19:37:35 Join amdgoon [0] (n=amdgoon@
19:39:14 Quit jon-kha (Read error: 131 (Connection reset by peer))
19:40:49 Join jon-kha [0] (
19:41:08Unhelpfulkugel: i've uploaded various-width test bitmaps to">, which should allow you to test various read sizes without rebuilds, by opening them in rockpaint. you'll need to use an unpatched build, of course, as the BM_MAX_WIDTH=8 patch hides the problem.
19:42:59 Join miepchen^schlaf [0] (
19:44:06 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
19:50:42MarcGuayIs there a linux binary of sendfirm available?
19:53:08 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:06:33 Join karashata [0] (n=karashat@
20:17:03 Quit tyfoo ("Carpe diem")
20:22:23 Join petur [50] (n=petur@rockbox/developer/petur)
20:37:09 Quit control_your_pla ("CGI:IRC")
20:42:58 Join midkay [0] (n=midkay@rockbox/developer/midkay)
20:48:14 Join Seed [0] (
20:51:16 Join casainho [0] (
20:51:31casainhohello :-)
20:52:22casainhoI would like to know if for reading and lanching rockbox fimrware on bootloader, is just need a function to read a block of 512MBytes?
20:52:31UnhelpfulMarcGuay: you should be able to build one, if you install a recent libmtp...
20:53:18***Saving seen data "./dancer.seen"
20:53:30casainhoI assume that FAT driver needs 512MBytes blocks... my question is if Rockbox firmware will work just with read or I will need to also implement the write function, on my case, on the SD card...
21:01:33 Join robin0800 [0] (
21:02:10 Quit robin0800 (Read error: 104 (Connection reset by peer))
21:02:23 Quit TheSphinX^ ("XChat@Linux")
21:04:11linuxstbcasainho: Rockbox itself needs read/write access. The bootloader just needs read access.
21:04:20 Quit Horscht (Read error: 104 (Connection reset by peer))
21:05:09 Join nibbler_ [0] (
21:05:14 Quit BigBambi (Remote closed the connection)
21:05:19casainholinuxstb: ah, because I remember that Rockbox write a file for configurations...
21:06:03gevaertsYou can worry about write access later though. It's needed for proper working, but you can get quite fat without (see e.g. the D2 port)
21:06:06 Quit Schmogel ("Miranda IM! Smaller, Faster, Easier.")
21:06:46casainholinuxstb: will rockbox run without write acess? or will just refuse to run? or will give some warning? - just for know If I should implement or not the write, at this stage...
21:06:58 Join toffe82 [0] (n=chatzill@
21:07:43casainhogevaerts: thanks :-) −− what do you mean by "get quite fat without"
21:07:54casainhogetting fat with less code? ?
21:08:04gevaertsoops, "quite far without"...
21:08:48casainhogevaerts: so, how will Rockbox reacts without writing?
21:08:54 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
21:09:34gevaertsYou'll get error messages now and then, but you can play music, run plugins, ...
21:10:33 Quit saratoga ("CGI:IRC (Ping timeout)")
21:11:40 Join robin0800 [0] (
21:11:42casainhogevaerts: ah, okok :-) thats nice and is what I was expecting :-)
21:12:47 Join gregorovius [0] (
21:15:41 Quit nibbler__ (Read error: 113 (No route to host))
21:26:45 Join massiveH [0] (
21:26:53 Quit casainho ("ChatZilla 0.9.84 [Firefox 3.0.5/2008121622]")
21:30:52 Quit massiveH (Client Quit)
21:35:21 Quit jhulst (Remote closed the connection)
21:38:19 Join gromit` [0] (
21:38:26 Quit mirak (Read error: 104 (Connection reset by peer))
21:40:48 Part pyro_maniac
21:44:52 Join aurix_lexico [0] (
21:46:20 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
21:46:45 Join gregzx [0] (
21:48:19 Join t0mas [0] (n=tomas@rockbox/developer/t0mas)
21:49:43 Join Zagor [242] (
21:52:17 Join Genbu [0] (
21:54:00GenbuIs there any way to get TTA+CUE support on my RockBoxed 4G?
21:56:03 Quit stripwax (Read error: 104 (Connection reset by peer))
21:58:16t0masGuess who has an ipod video :D
21:59:31 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
22:00:19linuxstbGenbu: No, Rockbox doesn't support TTA
22:00:32Zagorwhat is tta?
22:00:42Genbua lossless file format
22:00:48Genbua lot of the music i get from share is in it
22:01:16Genbulinuxstb: is there any plans to support it/any way to compile a build that will decode it?
22:01:46linuxstbNo-one has expressed any interest in it...
22:01:56linuxstbNo unless you know C programming, then no.
22:02:03*Genbu shakes fist at japan for encoding stuff in TTA
22:02:12Genbui do know C somewhat
22:02:18Genbuprobably not enough to do it myself though
22:02:33*Genbu is only a first year computer engineering student :\
22:03:42ZagorGenbu: why not just transcode them? they're lossless.
22:03:43t0masmight be a nice exercise
22:03:49domonokyif there is a opensource fixed-point TTA decoder implementation somewhere, it shouldnt be to hard to port it :-)
22:04:16Genbu its open
22:04:19t0masI am looking for it
22:04:41Genbuand come to mention it, the TTA decoder for foobar doesnt work anymore either
22:04:58Genbuhahhhhh might give me reason to install VS again
22:05:02linuxstbThey have a "hardware player" optimised decoding lib, which looks very straightforward to port to Rockbox.
22:05:35linuxstbNo floats, and BSD license...
22:05:46t0maseeh.. indeed
22:06:39t0masGenbu: they also offer to help with hardware support
22:06:53t0masso maybe they consider Rockbox "hardware support" for all our targets? :-)
22:07:00GenbuTTA's website hasnt been updated in forever.
22:07:05Genbui sorta wouldnt count on that ;)
22:08:20t0mastry to find a living IRC channel? or mailinglist?
22:08:28linuxstbOr a user...
22:08:52Genbui really wanna update to foobar too, but the dll component they have for that player is compiled for older foobar and doesn't work anymore :(
22:09:22GenbuI'll ask some of my upper classmen friends if theyd be interested in helping me out i guess
22:09:26Unhelpfulstill not sure why you can't transcode...
22:09:59linuxstbUnhelpful: Don't discourage him - new codecs are always welcome...
22:10:03Genbubecause i'd rather not?
22:10:36t0masUnhelpful: the sport of coding Rockbox..
22:10:36linuxstbGenbu: Does TTA actually have any advantages over more common formats like FLAC?
22:10:41Genbuits convenient for me to keep things in one format that I can use across everything and its irritating to have to transcode everything i grab off of share or perfect dark
22:10:55Genbunot sure.
22:11:08t0maslinuxstb: they claim 30% compression..
22:11:13t0masjust like FLAC
22:11:18Unhelpfulif you want a lossless format to use everywhere, flac is probably the best choice...
22:11:31*linuxstb thinks we're straying off-topic..
22:11:45*domonoky thinks for rockbox, more formats are always welcome :-)
22:11:47Unhelpfulother channel?
22:12:51*t0mas always forgets to join the other channel :)
22:13:12linuxstbZagor: That's comparing against some very old versions of other codecs though
22:13:15Genbuanything i rip myself i put to flac, but i cant help what other people rip. and i have an enormous amount of music in TTA already its a bore to transcode it all
22:13:25ender`linuxstb: in my tests, TTA was very fast to compress, and achieved quite good ratio, however it also needed more CPU to decompress. it was a while ago though
22:14:01Genbuender`, yeah thats true, i wonder if the ipod could handle the decompression?
22:14:03Zagorlinuxstb: yeah. looks like most of the site is old.
22:19:21 Join pyro_maniac [0] (n=jens@
22:24:49GenbuUnhelpful: what do you reccommend for a program to use in transcoding the tta? if i update foobar i wont be able to use it to do that anymore :\
22:26:01Unhelpfulyou should join #rockbox-community, because this is really getting to be not very much about rockbox... and for windows, i don't really know, i use linux for pretty much everything.
22:31:11 Part Genbu ("Leaving")
22:33:29 Join archivator [0] (i=archivat@
22:35:20 Join casainho [0] (n=chatzill@
22:36:07 Quit nibbler_ (Read error: 113 (No route to host))
22:36:14casainhohello :-) - I am looking at the source and looks to me that is no possible to have a device with SD card and Nand Flash, for example... or I am wrong?
22:36:58gevaertsRight now that's not possible, but I'm (slowly) working on that
22:40:21casainhogevaerts: okok :-) −− because I was trying to understand how SD card drivers are called, by who... and I saw a define for "storage_read_sectors"... eheh, and that name can't be the same for a device with SD card and Nand Flash, for example ;-)
22:41:01gevaertsIndeed it can't. Have a look at FS #9545 for the current status of my work
22:43:41kugelcasainho: have you looked at the as3525 sd driver? I think that one is quite readable
22:43:49Unhelpfulamiconn: really, the main difference with an overlay build is the start address, right? perhaps we could add a symbol at the beginning of the first section, and end of the last, in the linker script, as well as an option to define a start address outside the script proper.
22:44:29Unhelpfulbuild once, use the begin and end symbols to get the size of the plugin when loaded, then build again with an appropriate start address?
22:46:45amiconnhmpf! :(
22:46:45 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
22:47:09*amiconn made the mistake to quickly build current svn and put it on his H180 before going home
22:47:18Unhelpfulbroke it? :/
22:47:21amiconnThis rendered rockbox unusable, as playback doesn't work
22:47:33casainhokugel: I got a simple and quick tutorial about SD cards:
22:47:34amiconnIt pretends to play, but there's no sound output
22:47:36 Quit merbanan ("Leaving")
22:47:43amiconnVoice works though
22:48:00*amiconn suspects Zagor's watermark patch
22:48:08kugelcasainho: I really recommend to get in touch with some SD docs
22:48:33casainhokugel: and then I think that drivers on Rockbox are confuse, for a starter as me ;-) - and I found other simple code, here the explanation:
22:48:36 Join Virindi [0] (
22:48:48 Quit miepchen^schlaf ()
22:48:50Zagoramiconn: I'll test on my h140 now
22:49:11kugel"Simplified_Physical_Layer_Spec.pdf" is a good start
22:50:14kugelcasainho: btw: Why blocks of "512 MByte"?
22:50:21 Quit karashata ("G'bye everyone!")
22:50:25kugela block is supposedly 512 bytes afail
22:50:26Zagoramiconn: works for me
22:50:27bluebrotherI guess he means 512 bytes ...
22:50:45casainhokugel, my mistake, blocks of 512 bytes :-)
22:51:11VirindiDoes anyone know about 'data abort' errors? I just installed RB and I keep getting them, I have the bootloader that the GUI installer installs and my iPod works just fine with the apple firmware...
22:51:13 Quit XavierGr (Read error: 60 (Operation timed out))
22:51:25amiconnZagor: Voice doesn't matter, it also fails with voice disabled
22:51:28bluebrotherVirindi: what ipod?
22:51:35Virindi5g 60gb
22:51:37amiconnAudio format I'm testing with is 256kbps mp3
22:51:38kugelok, I just saw it in your patch (i.e. in code), so I was just curious
22:51:48kugelin a comment actually, but still
22:51:56casainhokugel: I got now that pdf file :-)
22:52:23kugelamiconn: I also had problems with playback, but only on my ams sansas (not on my e200v1), so they don't really count
22:52:33Virindibluebrother: if it helps most of my music is flac...
22:52:42 Quit jhulst (Read error: 60 (Operation timed out))
22:52:57Zagoramiconn: 204 kbps vbr here, and various others
22:52:58amiconnZagor: It is format dependent. FLAC works, vorbis also works. MP3 is broken.
22:53:00casainhokugel: what do you think of that code for the SD card? is simple, it will work? at least I understand it after reading that tutorial...
22:53:01domonokyVirindi: and which build have you installed ? the 30GB version or the 60/80Gb version ?
22:53:20bluebrotherVirindi: what _ipod_ do you have? I'm not interested in the type of music files (yet)
22:53:23***Saving seen data "./dancer.seen"
22:53:28Virindidomonoky: the 60/80 version
22:53:45domonokyVirindi: then try the 30Gb version...
22:53:58Virindidomonoky: okay, thanks, I'll do that
22:54:22 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
22:54:45ZagorI can't get it to fail on any track
22:54:51kugelcasainho: I haven't looked much
22:55:12kugelcasainho: I think if you don't move too far away from the existing sd drivers you'll be fine
22:55:32amiconnZagor: APE and musepack also work. MP2 (also mad.codec) fails. AC3 crashes with AddrErr...
22:55:54bluebrothercasainho: accessing a SD card isn't that hard −− that website should hold everything that's required for the basics.
22:55:55amiconnI'll delete the codec dirs in my build dir and rebuild the codecs
22:55:57casainhokugel: well, I found even a simpler code, however I don't know if it will work with all of sd cards... I will discover then ;-)
22:56:05ZagorI did a completely clean build
22:56:09amiconnZagor: Maybe it's a dependency problem in the codecs?
22:56:12bluebrotherIt can be a bit tricky to get the timing right (at least it was back the time I did some SD stuff)
22:56:47Zagoramiconn: keep your build dir and make a new one
22:56:58Zagorso we can compare
22:57:11amiconnHmrmmm, that will take a lot longer... :\
22:57:14Virindidomonoky: Thanks, it appears to be working with the 30gb least, it didn't crash in the place it always crashed with the other one
22:57:44domonokyVirindi: then you ipod is really a 30Gb version with a upgraded harddrive :-)
22:58:10Virindidomonoky: Hah, that's funny...thanks! :)
22:58:12Zagoramiconn: yeah, but it might help us find the issue.
22:58:30*amiconn got that
22:59:26Virindidomonoky: It was one of the very first 60gb video ipods so I guess that's why
22:59:29amiconnUnhelpful: We already have those start and end symbols
23:00:22Unhelpfulfor the start and end of the total space used by the loaded plugin? so, then we need to add the build system and linker script changes to make it work?
23:00:32 Quit dionoea (Remote closed the connection)
23:00:56 Quit domonoky (Read error: 104 (Connection reset by peer))
23:03:12casainhobluebrother: tricky getting the timming right? it isn't a communication with clock?
23:03:21casainhobluebrother: what timing can be wrong?
23:03:57amiconnUnhelpful: Ever checked a plugin .map file? There are plugin_start_addr and plugin_end_addr
23:04:24 Join fenugrec [0] (
23:04:43amiconnBoth are put into the plugin header as well
23:05:28amiconnSo a script could either read them from the .rock (attn target endianess) or parse the .map
23:05:33 Quit petur ("Zzzz")
23:05:43amiconnReplace .rock with .ovl for overlays
23:06:15Unhelpfulthe .map seems easiest... python is good at doing hex, but i'm guessing we'd want perl? i know there are some other perl scripts in the build process, already.
23:06:42*amiconn would vote for perl as well
23:06:56ZagorUnhelpful: definitely perl
23:07:30Unhelpfulthe votes are, in. not much i can do, then. ;)
23:08:16amiconnThe "preliminary" linker script needs to define a large enough memory area so it is guaranteed to not overflow. The linker barfs if an area overflows (but strangely not for a slight overflow...)
23:09:24 Quit casainho ("ChatZilla 0.9.84 [Firefox 3.0.5/2008121622]")
23:09:53Unhelpfulif we're passing the start address from outside, why not assume the entire audio buffer being used for the prelim. link? then we do the relink at the "right" address
23:10:22 Quit {phoenix} (Remote closed the connection) there anything to change besides the start/end addresses?
23:14:24 Quit bluebrother (Nick collision from services.)
23:14:29 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
23:17:57Bagderthe iriver lplayer uses a MIPS core...
23:22:29 Join miepchen^schlaf [0] (
23:23:32amiconnZagor: The fresh build fixed the problem. Now I need to compare...
23:23:42*amiconn thinks this is quite odd...
23:24:30 Quit pyro_maniac ("Leaving.")
23:25:47 Part Jaykay
23:27:06*Bagder learns that MIPS has a thumb-line mode for 16bit instructions, called MIPS16e
23:28:07amiconnZagor: a52.codec, adx.codec, mpa.codec, nsf.codec and speex.codec differ between my normal build and the fresh build. All other codecs are identical
23:28:38Zagorhmm, what could be the common factor for them
23:31:10 Quit bluebrother ("leaving")
23:34:20Unhelpfuli was afraid that might be a problem :/
23:35:00 Join dionoea [0] (
23:35:11Unhelpfulwell, the generic loader might be possible, even if it's not a viewer...
23:35:35 Join advlaptop2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
23:36:16amiconnZagor: It looks like some things there werent rebuilt, so that the argument enum values for ci->configure are off by one (you removed one so the subsequent entries changed)
23:36:33 Quit fyrestorm ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
23:37:49amiconnThe main .o files of these codecs weren't rebuilt. The ones of the other *decoder* plugins were rebuilt
23:37:52 Join nibbler_ [0] (
23:37:56Zagorright. the affected codecs were not modified by the patch, and due to unsufficient dependencies were not rebuilt
23:38:09amiconnThe ones of the _enc codecs also weren't rebuilt
23:38:15 Quit fenugrec ("Leaving")
23:38:20amiconnAh, they didn't use ci->configure to set watermark
23:38:53amiconnYou should have bumped codec api and min api versions, as changing the meaning of ci->configure arguments is an incompatible change...
23:39:02Zagorindeed I should
23:39:27amiconnOkay, at least that solves the mystery
23:39:53ZagorI should probably bump them now, since everyone svn-uping from a revision before that will get the same breakage
23:40:09amiconnYes, unless he does a full rebuild
23:40:19 Join akur [0] (
23:40:46 Part akur
23:41:56 Quit advlaptop2019 (Read error: 104 (Connection reset by peer))
23:42:23*amiconn wonders whether there are any functions in the codec api waiting to sort
23:42:57Zagor(luckily, since I committed before you said that :)
23:43:52 Quit tessarakt ("Client exiting")
23:45:01amiconnBtw, I think that the default spinup time for watermark calculation (if measured spinup time isn't available yet) might be a bit short
23:46:24 Quit flydutch ("/* empty */")
23:47:00 Join advlaptop2019 [0] (n=advcomp2@unaffiliated/advcomp2019)
23:48:42 Quit archivator ("Leaving")
23:48:51amiconnIt uses measured time + 1 second, but just 3 seconds when measured time is unavailable. But measured time is typically around 3 seconds...
23:48:57Zagorindeed. my h140 says 3120ms.
23:49:16amiconnSO I think 5 seconds would be a better safety default
23:49:22ZagorI agree
23:50:04amiconnIirc on archos we use some extra safety margins
23:50:08*amiconn goes checking
23:51:26amiconnAh, yes. It assumes spinup time is always >3.5 seconds if measured value is lower, *and* multiplies the measured value by 3
23:52:10*amiconn thinks that a safety factor makes sense, as the actual spinup time varies, and can easily be extended if the device is shaking
23:52:34*amiconn also thinks that mpeg.c isn't properly adjusted for Ondio wrt margins
23:52:58Zagorthat's a lot of safety. but it is also more critical on archos. on swcodec the pcm buffer keeps the music playing even if the rebuffering is delays a few seconds.
23:53:22amiconnAh, yes. That is, if pcm isn't low at the same time...
23:53:28Zagorit never is
23:53:57Zagorthat's why I'm working to shrink the pcm buffer. which will introduce the need for more spinup safety margin...
23:54:01*amiconn thinks that this assumption of spinup time >3.5 seconds is quite strange
23:55:01amiconnThose 2.5" hdds tend to spin up fster than 1.8" hdds
23:56:22amiconnHmm, in fact about the same
23:57:24Zagorthe pcm buffer is used rather strangely in my view. it is constantly topped up, so it is always full. on my h140 it toggles between 14 and 15 full buffers in use.
23:57:57 Join solexx_ [0] (
23:57:59amiconnIt's only like that if the codec doesn't need to boost, or only very little

Previous day | Next day