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

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

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

#rockbox log for 2010-11-24

00:08:12 Quit evilnick_B (Quit: Page closed)
00:08:57 Quit Kupop (Ping timeout: 272 seconds)
00:14:44 Join JdGordon| [0] (
00:14:44 Quit JdGordon| (Changing host)
00:14:44 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
00:16:14 Quit GeekShadow (Read error: Connection reset by peer)
00:16:32 Join GeekShad0w [0] (
00:17:59 Quit liar (Ping timeout: 255 seconds)
00:26:03 Join Llorean1 [0] (
00:26:21 Quit Llorean (Disconnected by services)
00:26:28 Nick Llorean1 is now known as Llorean (
00:26:31 Quit Llorean (Changing host)
00:26:32 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
00:28:35 Quit xavieran (Ping timeout: 272 seconds)
00:37:56 Quit TheSeven (Ping timeout: 240 seconds)
00:39:40 Join xavieran [0] (
00:55:11***Saving seen data "./dancer.seen"
00:56:29 Join Llorean1 [0] (
00:58:44 Quit Llorean (Ping timeout: 245 seconds)
01:02:41 Quit JesusFreak316 (Ping timeout: 245 seconds)
01:04:24 Join Llorean [0] (
01:04:39 Quit Llorean (Client Quit)
01:06:28 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
01:07:53 Quit Llorean1 (Ping timeout: 264 seconds)
01:16:06 Quit Keripo (Read error: Connection reset by peer)
01:17:55 Join Keripo [0] (
01:18:09 Quit pamaury (Remote host closed the connection)
01:31:33 Join fdinel [0] (
01:34:35 Quit Keripo (Quit: Leaving.)
01:52:39 Quit LambdaCalculus37 (Quit: Fwump)
01:55:08 Join Guest9567843 [0] (
02:06:32 Part Guest9567843
02:08:40 Join Guest9567843 [0] (
02:10:24 Join Gnea [0] (~gnea@unaffiliated/gnea)
02:14:32 Quit DerPapst (Quit: Leaving.)
02:15:20Guest9567843Trying to install on Sansa Clip v2: Using Utility after selecting the firmware i get - [ERR] Model name "clv2" not found in C:/Users/USER NAME/AppData/Local/Temp/qt_temp.da4776 and using the manual install it says it cant load - bootloader-clipv2.sansa any ideas?
02:17:02 Quit BlakeJohnson86 (Quit: Leaving.)
02:19:40 Join anewuser [0] (anewuser@unaffiliated/anewuser)
02:23:00 Join BlakeJohnson86 [0] (
02:24:04 Join Keripo [0] (
02:25:22 Quit Guest9567843 (Quit: Guest9567843)
02:27:23 Quit krabador (Quit: Sto andando via)
02:28:37 Quit GeekShad0w (Quit: The cake is a lie !)
02:35:13 Quit Zarggg_ (Quit: Zarggg_)
02:39:00 Join Keripo1 [0] (
02:41:14 Quit Keripo (Ping timeout: 265 seconds)
02:45:37 Quit InsDel (Read error: Connection reset by peer)
02:55:13***Saving seen data "./dancer.seen"
03:00:08 Join eWill [0] (
03:01:11eWillIf RB had an XVID decoder, would two video files of comparable quality, one MPG and one XVID, be very different in size?
03:02:12 Join antil33t [0] (
03:02:27krazykityou could find out be encoding two test files
03:03:51 Quit eWill (Client Quit)
03:40:27 Quit panni_ (Read error: Connection reset by peer)
04:01:51 Join binaryhermit [0] (
04:10:41 Join InsDel [0] (
04:10:46 Join S00row [0] (
04:14:26 Join fyrestorm [0] (
04:14:28 Quit Barahir_ (Ping timeout: 255 seconds)
04:15:52 Join Jagalives [0] (
04:16:02 Join Barahir [0] (
04:16:19 Part Jagalives
04:17:42 Nick mortalscan is now known as Mortalscan_away (~mortalsca@
04:21:56 Join Llorean [0] (
04:22:13 Quit Llorean (Changing host)
04:22:13 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
04:43:10 Nick tmzt_dg2root is now known as tmzt_g2root (~tmzt@
04:47:01 Quit TheSeven (Ping timeout: 252 seconds)
04:51:24 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:55:09 Quit amiconn (Disconnected by services)
04:55:10 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:55:17***Saving seen data "./dancer.seen"
04:55:30 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:55:57 Quit pixelma (Disconnected by services)
04:55:59 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:56:02 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
05:01:35 Quit dys (Ping timeout: 276 seconds)
05:01:53 Join dys [0] (
05:19:00 Join wastrel [0] (~wastrel@nylug/member/wastrel)
05:19:25wastrelhey, when i have "repeat: one" set , is there a way to advance to the next track?
05:19:33wastrelpushing next just starts over
05:26:12 Quit anewuser ()
05:27:59Lloreanwastrel: View the playlist, select the next track.
05:29:45wastrelyeah that's what i've been doing, was hoping for a quicker way
05:29:48wastrelthanks :]
05:30:43LloreanNo quicker way at the moment, though if "repeat" is in your quickscreen it might be quicker to turn it off and back on.
05:31:10 Quit shai (Quit: Leaving)
05:32:26 Join LambdaCalculus37 [0] (~rmenes@rockbox/staff/LambdaCalculus37)
05:34:13 Quit ps-auxw (Ping timeout: 265 seconds)
05:44:17 Quit Horscht (Quit: Verlassend)
05:45:15 Join ps-auxw [0] (
05:47:41wastreli think it's basically the same amount of effort
05:47:45wastrelbut yeah i've done that too :]
05:51:52 Quit InsDel (Read error: Connection reset by peer)
05:56:24 Join shai [0] (
06:02:08 Quit binaryhermit (Quit: Leaving)
06:19:17 Quit MethoS- (Remote host closed the connection)
06:22:18 Quit factor (Remote host closed the connection)
06:29:06 Join freddyb [0] (
06:29:08 Join factor [0] (
06:29:25 Quit factor (Read error: Connection reset by peer)
06:31:47 Join factor [0] (
06:34:23freddybAlexP: any chance of un-reverting r28000 for release? (fixed in r28616: FS #11608) I have two guys that say it fixed their problem.
06:34:27 Join kish_ [0] (~o2@unaffiliated/spice)
06:37:17 Quit kish (Ping timeout: 255 seconds)
06:44:02 Quit freddyb (Ping timeout: 255 seconds)
06:45:32 Quit fdinel (Quit: Miranda IM! Smaller, Faster, Easier.
06:50:32 Quit LambdaCalculus37 (Quit: This computer has gone to sleep)
06:55:18***Saving seen data "./dancer.seen"
06:55:36 Join BHSPitMonkey [0] (~stephen@unaffiliated/bhspitmonkey)
06:58:56 Quit Llorean (Ping timeout: 240 seconds)
07:00:39 Join Llorean [0] (
07:00:46 Quit Llorean (Changing host)
07:00:46 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
07:17:14 Join Llorean1 [0] (
07:17:14 Quit Llorean (Read error: Connection reset by peer)
07:17:45 Nick Llorean1 is now known as Llorean (
07:17:50 Quit Llorean (Changing host)
07:17:50 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
07:23:07 Join Llorean1 [0] (
07:23:28 Quit Llorean (Disconnected by services)
07:23:30 Nick Llorean1 is now known as Llorean (
07:23:32 Quit Llorean (Changing host)
07:23:32 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
07:30:55 Quit shai (Ping timeout: 260 seconds)
07:43:05 Quit JdGordon| (Quit: leaving)
07:52:46 Join Buschel [0] (
08:14:46 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
08:17:17 Quit BHSPitMonkey (Quit: Ex-Chat)
08:33:59 Join GeekShadow [0] (
08:33:59 Quit GeekShadow (Changing host)
08:33:59 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
08:37:17 Quit Buschel (Ping timeout: 250 seconds)
08:43:45 Join ender` [0] (
08:51:29 Join LinusN [0] (~linus@rockbox/developer/LinusN)
08:55:21***Saving seen data "./dancer.seen"
08:55:33 Quit jepler (Ping timeout: 245 seconds)
08:58:49 Join CaptainKewl [0] (
09:02:29 Join jepler [0] (~jepler@emc/developer/pdpc.professional.jepler)
09:22:22 Join Rob2223 [0] (
09:25:44 Quit CaptainKewl (Ping timeout: 265 seconds)
09:25:52 Quit Rob2222 (Ping timeout: 252 seconds)
09:31:41 Quit rasher (Read error: Operation timed out)
09:32:28 Join utanapischti [0] (
09:35:26 Quit sasquatch (Read error: Operation timed out)
09:46:00 Join esperegu_ [0] (~quassel@
09:46:02 Quit esperegu_ (Read error: Connection reset by peer)
09:47:36 Quit esperegu (Ping timeout: 276 seconds)
09:57:49 Join xxcv [0] (
10:03:22 Join petur [0] (d408b802@rockbox/developer/petur)
10:15:09 Quit dys (Remote host closed the connection)
10:19:59 Join swilde [0] (
10:20:54 Join rasher [0] (~jonas@rockbox/developer/rasher)
10:34:42 Join DerPapst [0] (~Alexander@
10:37:13AlexPfreddyb: I don't think we should, we don't want to risk breaking it
10:38:21gevaertsBattery life improvements are of course nice, but for a stable update they're not essential
10:40:19gevaertsZagor: do you think it would be a lot of work to set up a parallel build system for stable?
10:40:20AlexPAnd it can be a 3.8 feature :)
10:40:55Zagorgevaerts: ummm, yes :)
10:41:02gevaertsok :)
10:41:18Zagorat least if we want to use the same clients
10:41:35Zagora completely separate farm isn't that much work
10:42:30gevaertsThe main problem I see is that we'd need to remember to switch to another branch every major release
10:42:35*AlexP accuses Zagorof lacking a basic understanding of the challenges of an agricultural business :)
10:42:37 Join Loto [0] (
10:43:56 Quit Loto (Client Quit)
10:46:11 Join Loto [0] (
10:46:57 Quit Loto (Client Quit)
10:47:28 Join Loto [0] (
10:47:30 Quit Loto (Client Quit)
10:55:23***Saving seen data "./dancer.seen"
10:56:58 Join Buschel_ [0] (
10:57:04Stummiscorche, *ping*
10:57:54 Nick Buschel_ is now known as Buschel (
11:06:27 Join marines [0] (
11:06:33marineshi there
11:06:51marinesi hope somebody's here :D
11:06:56CIA-7New commit by stummi (r28654): thanks for admitting me to the team :)
11:07:13marinesi've just received package with sansa c250... v2
11:07:22Stummimarines, 132 people can be "somebody" :)
11:08:06marinesnot in every channel ;]
11:08:19marinesdoes anybody use this player with rockbox?
11:08:29Stummii am using an fuze
11:08:45marinesi'm not sure if it is stable enough to use v2 version
11:08:46ZagorI have several c200s
11:08:55CIA-7r28654 build result: All green
11:09:01Zagorah, not the v2 though
11:09:13marinesthat's the problem
11:09:19marinesv1 worked great
11:09:22pixelmathe c200v2 doesn't seem to be used very often. As far as I know, it's a little "behind" the other AMS Sansa ports
11:10:00marinesrbutil doesn't autorecognize this player
11:10:51gevaertsmarines: are you using the latest version of rbutil?
11:10:57gevaertse200v2 should just work
11:11:08marinesi've downloaded it a second ago
11:11:15gevaertsStummi: welcome!
11:11:23Zagorgevaerts: c200, not e200
11:11:33gevaertsoh, right
11:11:39Stummigevaerts, thanks
11:11:53gevaertsyes, c200v2 isn't fully stable yet
11:13:09pixelmathe DeviceDetection in the wiki has no USB PIDs/VIDs for the c200v2 so I guess that's just not implemented in the Rockbox Utility either
11:14:09Stummiat first a question as commiter: When I have seval changes here bzw want to add serval features one time, i shouldn't do this with one commit but with serval commits (one for each feature), is this right?
11:14:20gevaertsStummi: preferably, yes
11:14:49marinesi'll try svn build of rbutil
11:15:18gevaertsmarines: I suspect you'll be better off trying the manual install
11:15:31gevaertsNot sure though
11:15:49 Join casainho [0] (
11:17:23marinesok, package build failed so i'll install it manually ;]
11:18:49Stumminext point on checklist: getting this cool rockbox-developer-cloak
11:21:06 Quit TheSeven (Ping timeout: 255 seconds)
11:22:12AlexPStummi: Welcome by the way :)
11:22:33AlexPStummi: Do you have a forum account?
11:24:09Stummiyes, but didn't made a post yet
11:24:21Stummiusername is Stummi
11:25:57AlexPYou get a cool badge
11:26:11AlexPWhich I can do (but you'll need scorche for the cloak)
11:26:21 Join PurlingNayuki [0] (~PurlingNa@
11:27:01StummiAlexP, i know. I already tried to reach him here a few times.
11:27:31PurlingNayukiAre Rockbox considering moving OndaVX747 to unstable target?
11:28:09AlexPPurlingNayuki: I don't think many people really know much about it
11:28:16Zagorthe hive mind has not decided yet ;)
11:28:16marinesrb launched ;]
11:28:52PurlingNayukiAlexP: However, most parts of Rockbox works well on it.
11:29:47AlexPPurlingNayuki: Cool - I just meant that I don't think many people know enough enough about it to consider it, it needs the interested devs to say it is unstable then it can be :)
11:29:52 Join dfkt [0] (dfkt@unaffiliated/dfkt)
11:30:44marinesare there any information i can provide to help you, dear devs? :)
11:31:14Zagormarines: you can start with adding the usb id to the DeviceDetection wiki page
11:34:59marinesit;s the id in lsusb?
11:35:38Zagorand note that sandisk likes to use different IDs for ums and mtp mode
11:36:05 Quit PurlingNayuki (Ping timeout: 272 seconds)
11:36:13marinesah, right, i didn't scroll down ;]
11:38:14marinesi guess i can't edit wiki by myself so if you could do it, here is id: ID 0781:7452 SanDisk Corp.
11:39:22 Join dantje [0] (
11:40:46Zagormarines: is that in ums or mtp mode?
11:42:32marinesi followed instructions on c200v2 wiki to change mode to MSC mode before installation
11:42:40marinesi don't know if it's permanent setting
11:43:04marinesi've done lsusb after player reboot
11:43:46Zagorsounds like ums/msc then. can you do the same in mtp mode and see what you get?
11:44:19marinesthe process is the same?
11:44:38Zagorboot the OF, set it to mtp mode. plug in usb, run lsusb.
11:47:44marineskind off kernel panic i guess :>
11:47:53marinessandisk splash screen doesn't go off
11:49:01marinesoh, works, i ejected memory card
11:49:34marinesbut there is not such setting like usb mode
11:52:14marinesid changed
11:52:21marinesID 0781:7453 SanDisk Corp.
11:52:38marinesafter following usb mode change instructions
11:53:07Zagorso 7453 is mtp?
11:53:39Zagornearly all other sansas have mtp on the lower id
11:55:00Zagoryou can run lsusb -v and look for "bInterfaceClass 8 Mass Storage"
11:55:02marinesinstructions tell about changing mode to MSC
11:59:53marinesfor 0781:7452 there is "bInterfaceClass 8 Mass Storage" entry
12:01:20marinesi'll paste lspci -v for both id's
12:04:47marinesi don't know which one is mtp mode but i guess 7453 is MSC because it appears after "changing mode" (as described on wiki)
12:07:21 Quit xavieran (Ping timeout: 276 seconds)
12:13:04 Join T44 [0] (
12:16:24 Quit Topy (Ping timeout: 245 seconds)
12:17:17 Join n1s [0] (
12:17:17 Quit n1s (Changing host)
12:17:17 Join n1s [0] (~n1s@rockbox/developer/n1s)
12:19:23 Join PurlingNayuki [0] (~PurlingNa@
12:23:08marineswhat's the difference between those two modes?
12:23:40PurlingNayukiWhich two?
12:23:55marinesmtp and msc
12:24:00Stummimarines, in msc the device appears as an normal volume
12:24:23Stummiin mtp it uses a special protokoll for file transfert
12:28:01 Part marines
12:28:19PurlingNayukiSo did you agree to move OndaVX747 to unstable?
12:32:14 Quit xxcv (Ping timeout: 240 seconds)
12:36:43AlexPPurlingNayuki: No, there has been no discussions
12:37:30AlexPPurlingNayuki: If I were you I would post to the dev list saying what the status it is and that you think it should move (showing that it meets the criteria for unstable)
12:39:43CIA-7New commit by stummi (r28655): Adding an reset option to the rockboy menu
12:39:54 Join marines [0] (
12:40:10Stummifirst productive commit :)
12:41:45CIA-7r28655 build result: All green
12:45:30 Quit Buschel (Ping timeout: 240 seconds)
12:45:39PurlingNayukiThx, I'll try.
12:47:14 Join LambdaCalculus37 [0] (~rmenes@rockbox/staff/LambdaCalculus37)
12:52:11 Part marines
12:53:56 Part PurlingNayuki
12:54:58n1shmm, the test_codec menu seems broken
12:55:25***Saving seen data "./dancer.seen"
12:55:59 Nick kish_ is now known as kimjongil (~o2@unaffiliated/spice)
13:01:11CIA-7New commit by nls (r28656): codeclib: there is no need to limit the outputs of the coldfire X(N)PROD_R macros to data registers, no significant speed diff
13:02:49 Quit antil33t (Read error: Connection reset by peer)
13:02:53n1sit's running checksum dir when i want to run just checksum, this used to work fine
13:02:59CIA-7r28656 build result: All green
13:03:00 Join antil33t [0] (
13:04:19 Quit Dreamxtreme (Ping timeout: 245 seconds)
13:04:31n1sah, i see it now
13:05:00*LambdaCalculus37 looks at the Gigabeat MEG50JS specs
13:06:34 Join marines [0] (
13:08:49LambdaCalculus37The Neomagic Mimagic 3 NMS7210 SoC is an ARM720T. Wonder how performance will be...
13:10:10 Join xxcv [0] (
13:10:49n1sLambdaCalculus37: wow, that's without the "fast" multiplier then?
13:12:09LambdaCalculus37n1s: Want the datasheet?
13:12:46n1sok, is that in Gigabeat MEG50JS?
13:13:31n1sah, are you planning a port?
13:13:44LambdaCalculus37Some of the specs on the datasheet: "32-bit ARM V4T instruction set with 16-bit instructions (THUMB), 8KB four-way set-associative unified cache, MMU with 64 entry TLB, up to 110MHz in Asynchronous Mode"
13:14:10LambdaCalculus37n1s: toffe82 gave me one, and I had been looking at a firmware update to see how to get firmware on the device.
13:14:32n1swell, at least it has a cache :)
13:14:33LambdaCalculus37n1s: It's pretty interesting, because the entire firmware update is a tiny Linux system.
13:14:53n1swhat kind of clock frequency does it run at?
13:15:08LambdaCalculus37n1s: According to the datasheet, it can clock up to 110MHz.
13:17:11 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
13:17:18LambdaCalculus37n1s: It can do dynamic frequency switching, at 18/36/74 MHz in synchronous CPU mode and 110 MHz in asynchronous mode.
13:19:03n1si think i'm wrong and that core is arm7tdm so it has the "fast" multiplier
13:20:06n1ssould be similar performance to a PP then i think
13:20:26 Join einhirn [0] (
13:21:31 Join Dreamxtreme [0] (~Dre@
13:24:39n1shmm, it has a "write buffer" that sounds like it would help somewhat
13:27:51LambdaCalculus37n1s: Seems that it only has a 16MB buffer.
13:30:59n1s16MB of ram should be enough, a number of our old coldfire targets get by with 16MB
13:32:15n1safter reading a bit the arm720T seems nice, basically an arm7tdmi with cache, mmu and the write buffer, i had forgotten that the standard arm7ydmi doesn't have a cache (but the PP cores do)
13:32:41n1sLambdaCalculus37: do you have sorce for the linux firmware?
13:34:32 Quit Llorean (Ping timeout: 245 seconds)
13:37:13 Quit n1s (Quit: Lämnar)
13:37:17 Join Llorean [0] (
13:37:31 Quit Llorean (Changing host)
13:37:31 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
13:40:17 Quit xxcv (Ping timeout: 240 seconds)
13:40:33 Join Llorean1 [0] (
13:42:06LambdaCalculus37n1s: I haven't come across it yet.
13:43:08 Quit Llorean (Ping timeout: 240 seconds)
13:43:57LambdaCalculus37n1s: However, I once asked markun about how he ported Rockbox to the Gigabeat F, and he had mentioned that he was able to modify the firmware updater shell script and add some commands to dump the contents of the flash ROM.
13:47:20LambdaCalculus37Sure enough, there's a /dev/from in the firmware update for the MEG50JS.
13:48:05LambdaCalculus37n1s: Back in a bit.
13:53:28 Join PurlingNayuki [0] (~PurlingNa@
13:53:39 Part PurlingNayuki
13:56:14 Quit Llorean1 (Ping timeout: 240 seconds)
13:58:04 Quit TheSeven (Ping timeout: 265 seconds)
14:01:45LambdaCalculus37I posted a complete listing of the files in the firmware here:
14:08:25CIA-7New commit by stummi (r28657): rockoy: boosting the max frameskip option up to 20 Max instead of 6 ...
14:10:11CIA-7r28657 build result: All green
14:10:23markunStummi: congrats on the commit access
14:11:46Stummihehe, thanks :)
14:14:05LambdaCalculus37I think playing with the firmware update shell script might yield interesting results:
14:15:46 Quit antil33t (Read error: Connection reset by peer)
14:15:57 Join antil33t [0] (
14:37:04 Join Llorean [0] (
14:37:18 Quit Llorean (Changing host)
14:37:18 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
14:37:21markunLambdaCalculus37: I'll try to keep the rockbox talk in here :)
14:38:30LambdaCalculus37markun: Okay. :)
14:38:44LambdaCalculus37I don't see any /mnt references in the firmware so far.
14:41:27 Quit cjcopi (Ping timeout: 264 seconds)
14:41:31 Quit Llorean (Ping timeout: 240 seconds)
14:42:33markunLambdaCalculus37: perhaps you can print something on the screen with /fwup/err-disp, but I don't know the syntax.
14:45:06LambdaCalculus37markun: There are three HD-ROMFS.{_* files in /FWUP; those look interesting.
14:45:44markuncheck the −−LINUX.. files for the real names :)
14:45:48LambdaCalculus37Combined, they add up to 4,063,328 bytes.
14:47:15 Quit Gnea (Remote host closed the connection)
14:55:29***Saving seen data "./dancer.seen"
14:59:57 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
15:03:33 Join panni_ [0] (
15:05:03*LambdaCalculus37 will come back to some hacking later... gotta bounce!
15:05:11 Quit LambdaCalculus37 (Quit: BOUNCE!)
15:08:31 Join InsDel [0] (
15:08:57 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:09:37 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
15:10:11 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:23:14 Join CaptainKewl [0] (
15:25:42 Join InsDel1 [0] (
15:27:32 Quit InsDel (Ping timeout: 260 seconds)
15:30:10 Quit InsDel1 (Ping timeout: 245 seconds)
15:31:03 Quit domonoky (Read error: Connection reset by peer)
15:40:17 Join InsDel [0] (
15:55:01 Quit evilnick_B (Ping timeout: 265 seconds)
15:59:20 Quit InsDel (Ping timeout: 245 seconds)
15:59:24 Quit Mortalscan_away (Ping timeout: 265 seconds)
16:01:23 Join kevku [0] (~kevku@2001:7d0:0:f000::135d)
16:02:11 Quit dantje (Quit: Ex-Chat)
16:02:57 Join Mortalscan_away [0] (~mortalsca@
16:03:05 Join MethoS- [0] (~clemens@
16:03:49 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
16:08:53 Quit MethoS- (Remote host closed the connection)
16:09:16 Join MethoS- [0] (~clemens@
16:09:57 Quit MethoS- (Remote host closed the connection)
16:10:24 Join MethoS- [0] (~clemens@
16:10:31 Join InsDel [0] (
16:10:44 Join liar [0] (
16:13:01 Quit TheSeven (Read error: Connection reset by peer)
16:13:43 Nick Mortalscan_away is now known as mortalscan (~mortalsca@
16:16:17 Quit petur (Ping timeout: 265 seconds)
16:16:49 Join MethoS-- [0] (~clemens@
16:17:22 Quit GeekShadow (Ping timeout: 255 seconds)
16:18:15 Quit evilnick_B (Ping timeout: 265 seconds)
16:20:11 Join The_Seven [0] (
16:20:11 Quit The_Seven (Changing host)
16:20:11 Join The_Seven [0] (~TheSeven@rockbox/developer/TheSeven)
16:21:19 Quit MethoS- (Remote host closed the connection)
16:21:19 Join MethoS- [0] (~clemens@
16:21:19 Quit MethoS- (Remote host closed the connection)
16:21:26 Quit The_Seven (*.net *.split)
16:21:26 Quit panni_ (*.net *.split)
16:21:27 Quit markun (*.net *.split)
16:21:27 Quit advcomp2019 (*.net *.split)
16:21:27 Quit Galois (*.net *.split)
16:23:48 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
16:24:28 Join u42p [0] (
16:24:32 Join The_Seven [0] (
16:24:32 Join panni_ [0] (
16:24:32 Join markun [0] (~markun@rockbox/developer/markun)
16:24:32 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
16:24:32 Join Galois [0] (
16:24:56 Quit mortalscan (Ping timeout: 240 seconds)
16:24:57 Join mortalscan_ [0] (~mortalsca@
16:31:13 Quit The_Seven (Ping timeout: 245 seconds)
16:34:02Zagorgevaerts: I'll see if I have time to whip up a release branch buildserver tonight or tomorrow. it's a good way to test the new rockbuild version...
16:35:11 Part Zagor
16:35:28gevaertsZagor: sounds good. Probably not too useful for 3.7.1 any more, but we'll do 3.7.2 :)
16:36:06u42pi just noticed my rockbox on sansa clip+ only plays vorbis if it is in .ogg not in .oga
16:36:15 Quit MethoS-- (Remote host closed the connection)
16:39:58gevaertsu42p: is that the same file, renamed?
16:41:10 Join bmbl [0] (
16:41:10 Quit bmbl (Changing host)
16:41:10 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
16:41:44u42pah, sorry. i was transcoding with ffmpeg2theora. the .ogg have no "skeleton" (which is metadata about the whole track) while the .oga ones do.
16:42:14*gevaerts leaves this to the codec people
16:42:17u42pmy experience seems similar to
16:42:40u42piirc if i tried to play one of those .oga ones, it would not freeze but not play anything until i rebooted
16:43:26 Part LinusN
16:43:29u42phm, trying it now it skips to the next track fine
16:44:21 Join hebz0rl [0] (
16:51:00 Join GeekShadow [0] (
16:51:00 Quit GeekShadow (Changing host)
16:51:00 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
16:55:33***Saving seen data "./dancer.seen"
16:57:26 Quit casainho (Read error: Operation timed out)
16:59:08 Join casainho [0] (
17:04:10 Quit GeekShadow (Ping timeout: 255 seconds)
17:05:01 Join krabador [0] (
17:10:38 Join anewuser [0] (anewuser@unaffiliated/anewuser)
17:17:28 Quit riotz (Ping timeout: 272 seconds)
17:18:38 Join riotz [0] (riot33r@gateway/shell/
17:22:15 Quit riotz (Excess Flood)
17:22:48 Join kugel [0] (~kugel@rockbox/developer/kugel)
17:23:08 Join riotz [0] (riot33r@gateway/shell/
17:30:15 Join n1s [0] (~n1s@rockbox/developer/n1s)
17:40:44 Quit tchan (Read error: Connection reset by peer)
17:41:48CIA-7New commit by stummi (r28658): Rockboy: Added an short stats option which only shows the two numbers
17:42:43CIA-7New commit by nls (r28659): test_codec: fix menu broken in r28637
17:42:52 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
17:42:53 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
17:43:40CIA-7r28658 build result: All green
17:44:21CIA-7New commit by stummi (r28660): Rockboy: Changed the time of showing splash messages to 0 to decrase loading time. ...
17:45:26CIA-7r28659 build result: All green
17:47:13CIA-7r28660 build result: All green
17:47:40 Quit antil33t (Read error: Connection reset by peer)
17:47:51 Join antil33t [0] (
17:50:35pixelmaStummi: changing time a splash is shown to 0 sounds weird - either you want the splash then it needs to be shown for longer or you don't, then why not disable the splash completely?
17:51:50 Quit panni_ (Read error: Connection reset by peer)
17:52:12 Quit mc2739 (Ping timeout: 260 seconds)
17:52:15Stummipixelma, for example the "loading rom file"-splash will be showed as long the romfile is realy loading now
17:52:24 Join Kitr88 [0] (
17:52:26 Join pamaury [0] (
17:52:26 Quit pamaury (Changing host)
17:52:26 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
17:53:03Stummioh, it won't. Was my failure
17:53:25pamaurygevaerts: any thought about FS #11774 ?
17:54:24 Quit phanboy4 (Quit: Leaving)
17:54:32pixelmahmm, didn't know that, was just going by the commit message which I didn't fully understand but it sounded weird
17:54:55 Quit Kitar|st (Ping timeout: 265 seconds)
17:55:12gevaertspamaury: I haven't looked at it very deeply yet, but I think it's a good idea
17:56:34pamauryI thought is was already unmounting disks
17:56:38pamauryat leats on slave mode
17:56:39 Join Buschel [0] (
17:56:51 Quit Kitr88 (Ping timeout: 252 seconds)
17:57:01pamauryfurthermoer, native usb is the contrary of slave mode which is hardware usb no ?
17:57:26gevaertsIt doesn't unmount
17:58:04gevaertsAnd usb_slave_mode() is implemented for both the hardware USB mode as for the usb stack mode
17:58:06pamaurythen it could be a good idea if all handles are properly released on usb connection
17:58:45gevaertsThey should be, but enforcing this via unmount is probably a good idea
17:59:02gevaertsThe alternative is possible disk corruption, which isn't nice at all
17:59:43pamauryyes but unmount doesn't ensure proper handle release
18:00:09pamauryif rockbox keeps a handle, say a font handle, and reuse it after connection, it might well be a disaster
18:00:24gevaertsthis implementation does
18:00:42pamauryit invalides the handles ?
18:00:46CIA-7New commit by stummi (r28661): correction from previous commit. Completly removing the splashes is better
18:00:46gevaertsIt calls release_files() and release_dirs()
18:01:00gevaertswell, the new disk_unmount_all() does
18:01:17 Join MethoS- [0] (~clemens@
18:02:24 Join Kitar|st [0] (
18:02:27CIA-7r28661 build result: All green
18:02:44 Join afk [0] (~Dre@
18:02:52 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
18:03:20 Quit MethoS- (Remote host closed the connection)
18:03:40 Quit TheSeven (Disconnected by services)
18:03:44 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
18:03:57gevaertsI'm not sure if the patch does it the right way though. I'd have disk_unmount_all() call disk_unmount(), not duplicate all code from disk_unmount()
18:04:16 Quit afk (Client Quit)
18:04:20pamauryyes me too
18:04:38pamauryI don't have the code in memory but it seems that it duplicates features
18:05:24 Join afk [0] (~Dre@
18:05:35 Quit Dreamxtreme (Ping timeout: 245 seconds)
18:21:21 Join JesusFreak316 [0] (
18:26:19 Join bertrik [0] (
18:26:20 Quit bertrik (Changing host)
18:26:20 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
18:26:40 Quit utanapischti (Quit: magick reboot)
18:30:47kugelI think I have a solution for the database on RaaA
18:30:56 Join sasquatch [0] (
18:31:14kugelcurrently we scan the root filesystem, so it should be OK to not follow symlinks
18:31:23 Part marines
18:32:23kugelthis way the database builds just fine on my phone
18:34:22 Join TheLemonMan [0] (
18:35:31 Join clone4crw [0] (
18:35:50kugelthe other solution could be to keep track of every directory already scanned (which can be a lot)
18:36:02kugelI can't think of any other
18:36:02 Quit factor (Read error: Connection reset by peer)
18:37:22 Quit DerPapst (Quit: Leaving.)
18:38:38 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
18:40:14pamaurykeeping the list in memory would be difficult
18:40:42pamauryexcept if you are sure that it fits, but even this way it would make the code even more harder to read :)
18:41:47kugelnot following symlinks might be undesirable if we ever get a "start database scan here" option
18:42:17 Quit clone4crw (Remote host closed the connection)
18:43:06pamauryyeah, we could have a list of starting points for the db, perhaps the user don't want to scan the whole fs on RaaA
18:45:29 Quit swilde (Quit: ERC Version 5.3 (IRC client for Emacs))
18:47:39 Join clone4crw [0] (
18:50:15 Join Strife89TX [0] (~cstrife89@
18:50:51 Quit afk (Quit: Ex-Chat)
18:50:57kugelall other media players handle this better than we, perhaps we should just copy one of the existing solutions
18:51:18 Join Dreamxtreme [0] (~Dre@
18:52:17pamaurywhat do they do ?
18:52:42pamaurythey also target this specifically
18:52:55pamaurywhereras our code is designed for something quite different
18:53:19pamaurybut we need to do better then :D
18:53:26kugelno, our database search has the same purpose
18:54:19 Join factor [0] (
18:54:38pamaurywe don't have siminks
18:54:51gevaertssymlinks aren't the only possible cause for loops in linux
18:55:18pamaurylike ?
18:55:32gevaertsAlthough maybe (I should check) bind mounts that make loops are disallowed
18:55:35***Saving seen data "./dancer.seen"
18:56:01pamaurygood question
18:56:29Torneno, they're fine :)
18:56:40Tornewell, for linux
18:56:41Tornenot us :)
18:57:11kugelbind mounts?
18:58:51*Torne tries again with more clarity: you can construct loops with bind mounts, actually maybe you can't.
18:59:21Torneok ignore me
18:59:25Tornei don't know :)
18:59:40gevaertsI don't seem to be able to
18:59:46Torneyou'd need to −−rbind
18:59:46gevaertsThey don't inherit sub-mounts
18:59:48kugelmainting a list and checking against it adds quite a lot memory consumption and speed penalty I'm afraid
18:59:50gevaertsIf that's clear
18:59:52Tornethey do if you −−rbind
18:59:59Tornebut i think you still can't have a cycle
19:00:17gevaertsIndeed. I still didn't manage with rbind
19:00:51pamauryeven without cycle, you can have redundant entries, but I think there handles by the db, aren't they ?
19:01:20gevaertsRedundant entries if you do weird things are user error I'd say
19:01:45 Quit benedikt93 (Quit: Bye ;))
19:02:49gevaertsThe problem at hand is how to avoid cycles if people do stupid things with symlinks, right?
19:03:04kugelit's not necessarily user error
19:03:19kugelnot if you start scanning in / anyway
19:03:54bertrikAnyone here familiar with the meaning of LCD bias? in particular in relation to the multiplex ratio
19:04:21 Quit casainho (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101027124101])
19:05:07gevaertskugel: a filesystem with actual useful content (from our point of view) that's mounted (not just symlinked) in more than one place without user intervention sounds pretty unlikely
19:05:32kugelfrom that point of view you shouldn't start in / :)
19:05:36 Join DerPapst [0] (
19:06:03kugelon RaaA
19:06:15pamaury starting in / sound a bit akward, you will scan lots of unecessary files on RaaA, giving a speed penalty
19:06:22*gevaerts doesn't follow
19:06:41kugelpamaury: it was faster on my phone than scanning on my daps
19:06:59kugelbut it was "only" 26k objects to scan
19:07:07kugelon a pc it's more I guess
19:07:24kugelgevaerts: / has no useful content for rockbox
19:07:25pamauryon a pc you can easily have 100k objects
19:07:32kugelbut it has recursive symlinks
19:07:36gevaertskugel: yes, and?
19:08:03pamaurythus it is stupid to start the scan search in /
19:08:03gevaertsThat means there is a filesystem somewhere that does have useful content
19:08:21kugelare you talking about loops or multiple entries?
19:08:35gevaertsWhat I'm saying is that *that* filesystem won't be mounted multiple times, unless the system is owned by someone who's trying to break things
19:08:54kugelcyanogenmod mounts the sdcard twice
19:08:59gevaertsSo you won't find the same content twice if you don't follow symlinks
19:09:05kugeldon't ask me
19:09:17kugelwell, it's not mounted twice, but /sdcard links to /mnt/sdcard
19:09:33pamaurythat's different and quite common
19:09:39gevaertsok, so what I'm saying still applies
19:09:41kugelso yes, I had multiple entries without user intervention
19:09:49gevaertsIt has symlinks
19:09:52gevaertsThat's different
19:10:24gevaertsSo the problem at hand is how to avoid cycles if people do stupid things with symlinks
19:10:49gevaertsAnd if possible also to avoid multiple entries for the same file
19:10:56kugelI don't worry about FSes mounted multiple times, in fact I didn't even think about that until now
19:10:57pamauryyou need to mark objects or keep them in memory/db
19:11:41pamaurya nasty way would be to "touch" symlink
19:12:08pamauryerr, symlink targets more likely
19:12:58 Quit DerPapst (Quit: Leaving.)
19:13:12gevaertsWhat if you *only* follow symlinks that point outside your (configured) search root?
19:13:38pamauryit won't avoid redundant entries
19:13:51pamaurybut will avoid cycles
19:13:56pamauryerr no
19:14:00pamaurymy mistake
19:14:37gevaertsWhat if you *only* follow symlinks that point outside your (configured) search root(s), and whenever you do that you add the target of that symlink to your list of search roots?
19:15:18gevaertsIf your search root is /, that means "don't follow symlinks"
19:15:31pamaurythis way you process at most each file twice
19:16:02gevaertsin which cases?
19:16:14pamaury"if you *only* follow symlinks that point outside your (configured) search root(s), and whenever you do that you add the target of that symlink to your list of search roots?"
19:16:22kugelgevaerts: that would work I guess
19:16:29pamauryyes might work
19:17:21kugelwhat if the targets have recursive links in them again?
19:17:44gevaertsThen they'd point back to one of the search roots, at which point you stop
19:18:20pamauryHum, you might process files more than twice in fact, but a finite number of times :)
19:18:30 Join marines [0] (
19:18:52kugelgevaerts: they would?
19:19:34gevaertsIf you have /a/b/c/d -> /e/f, and /e/f/g-> /a/b, and you start in /a, you add /e/f to your search roots (which now are /a,/e/f), then you see /e/f/g which points to /a/b which is inside /a so you ignore it
19:19:36pamaurykugel: if you mark the target of a symlink, since there is a finite number of symlinks, there is a finite amount of targets, you at some points will have marked then all
19:20:47gevaertspamaury: the only ways I see to add a file more than once still involve somewhat interesting symlinking
19:20:53kugelah you add the outside targets to the search roots
19:21:10 Quit Buschel (Ping timeout: 240 seconds)
19:21:22pamaurygevaerts: that's true, but in theory you can be in that case
19:21:39kugelsounds good, I can't see a disadvantage right now
19:22:02gevaertsI'd only do that if those targets are directories, in which case symlinks to files can still provide duplicates
19:22:25kugeloh, except you can still get duplicates
19:22:35gevaertsI think that if you also add all files that get pointed at, you might get a huge list of search roots
19:22:42pamauryit will be really hard to eliminate duplicates
19:22:45 Join DerPapst [0] (
19:22:56gevaertsYou can *always* get duplicates if you're creative enough
19:23:00kugelif /sdcard/ points to /mnt/sdcard, then it'd scan /sdcard/* and the /mnt/sdcard/* no?
19:23:02pamaurythere aren't so many symlinks I think
19:23:19pamaury(ot directories I mean)
19:23:24pamaury*to :)
19:23:27gevaertskugel: only if /sdcard or /mnt/sdcard is the search root, and you have a symlink to the other *inside* the first
19:23:58gevaertsIf / is your search root, you'll ignore /sdcard (because it points to a target inside /)
19:24:15factorcool fuze has line out enabled in the 3.7
19:24:40gevaertsAnd if people have /mnt/sdcard/something pointing to /sdcard, I'd say bad luck
19:25:35 Quit JesusFreak316 (Ping timeout: 240 seconds)
19:25:48*DerPapst did that once. was fun :)
19:26:23kugelok, it should work
19:26:51kugelyou don't follow symlinks, but instead you add the targets to the search roots if they're outside
19:27:15kugel(that's how I understand it now)
19:27:16gevaertsThis solution avoids cycles and I think it probably avoids all duplicates caused by symlinks set up by the distribution (or manufacturer), and caused by *most* symlinks set up by the user
19:28:00gevaertskugel: yes
19:29:09kugelI think that could work, and shouldn't even be too hard to implement
19:29:30gevaertsIf the symlink points to a directory at least. If it points to a file inside the search tree I'd say ignore as usual, and if it points to a file outside the search tree I'd include that file
19:29:52gevaertsBut not add it to the search roots, to avoid ending up with thousands of search roots
19:30:37gevaertsThe case I'm thinking of is people who set up all sorts of database-like result sets using symlinks
19:33:18gevaertsWhat does readlink() return if the symlink you feed it points to a path that also includes symlinks?
19:34:15 Part marines
19:38:05 Quit Stummi (Ping timeout: 245 seconds)
19:41:18pamauryI think it just dereference one symlink
19:41:37gevaertsit does, yes
19:41:42*gevaerts tested it now
19:44:13 Join Buschel [0] (
19:44:22gevaertsI think ideally we'd want to store a fully dereferenced path in the database. That way we'd eliminate every remaining weird duplicate. I'm not sure if that's easy or cheap though
19:45:39factorWill rockbox utility grab the latest firmware for your chipset?
19:46:06factorlooks like it is
19:46:10 Join shai [0] (
19:46:25factoroh just said to dl it myself
19:46:51factoroh just for the original boot opt, cool
19:47:49pamaurygevaerts: intermediate links are useless, and I doubt there are really long chains, 2 or 3 at most
19:51:32 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
19:56:31kugelgevaerts: seems to work for the /sdcard / /mnt/sdcard case
19:57:31kugelchecking in the sdl app where I can more corner cases :)
19:58:02gevaertsWell, you can create corner cases in android too, but there cleaning up may be trickier :)
19:59:33 Join kazaik [0] (
20:00:07kugelhm, it shouldn't scan /home/kugel/sources
20:01:17*gevaerts agrees
20:01:26gevaertsThere can't be anything interesting in there ;)
20:03:04kugelthat's not in my search roots
20:03:35wastreli'm charging my sansa
20:04:13kugelhe starts in my search root but then proceeds to /usr :\
20:04:56kugelwithout adding a root
20:05:47gevaertsSounds like time for a debugger
20:07:16 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
20:08:32factorWhen I pluin my Sansa fuze with rockbox it does not detect my device. while booted up into default firmware it does.
20:09:57 Quit mystica555_ (Remote host closed the connection)
20:12:04domonokyfactor: try to disable the USB Hid function and try again. Some OSs doesnt seem to like that.
20:12:32clone4crwi think the USB HID feature should be disabled by default.
20:12:35factorahh ok will try that later, just wanted to know why
20:13:11gevaertsfactor: v1 or v2?
20:13:13factorKDE4 picks up the device as its default fuze v2 firmware
20:13:29gevaertsin that case that's normal
20:13:46factordoes the mpeg player in rockbox work with v2
20:13:51kugelhm it doesn't detect my simlink
20:14:26n1sthat reminds me, my fuzev1 only connects over usb if i plug in while it'
20:14:29n1ss off
20:14:33gevaertsRockbox doesn't have USB on the v2 yet, and we don't reboot to the OF as we do on other players in similar situations because there's a bug that causes serious problems for some people if we do reboot at that point
20:14:42kugelno, readlink fails
20:15:05 Join mystica555_ [0] (
20:15:13factorahh drats
20:15:14 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
20:15:16factorthat is why
20:16:05factorThanks , that is what I wanted to know, saw the line out and usb enabled but that is for v 1
20:16:08kugelTrying to add /home/kugel/rbdev/rockbox-git/build-sdlapp/Musik to roots
20:16:08kugelreadlink: No such file or directory
20:17:33Buscheln1s: do your latest changes to CF asm in codecs/lib fix the slow down when using gcc 4.4?
20:17:49n1sBuschel: partially
20:19:15n1sasm fft4/fft8 helps a bit more so the slowdown for vorbis is just 0.8MHz with those, although i want to reduce the codesize a bit so it will fit in iram with O3 again as that also helps a bit
20:19:22n1sand that fft8 is a monster
20:19:42kugelgevaerts: why does readlink not see the link?
20:19:49kugelshell's readlink works
20:20:29Buscheln1s: is -O3 any faster than -O2?
20:20:46n1syes, for vorbis it is (on coldfire)
20:21:09n1snot so much but ~1% or 0.3MHz
20:22:16gevaertskugel: That "Musik" part is the symlink I presume?
20:22:44kugel$ readlink /home/kugel/rbdev/rockbox-git/build-sdlapp/Musik
20:22:58kugel /media/XP/kugel/Musik/
20:23:44Buscheln1s: how much does gcc 4.4 slow down vorbis when using -O2?
20:24:41Buschel(just out of interest)
20:24:46 Join afk [0] (~Dre@
20:25:00 Quit afk (Read error: Connection reset by peer)
20:25:37n1s(when i use my asm fft functions)
20:26:23n1s~1.7MHz without them i think
20:28:07 Quit Dreamxtreme (Ping timeout: 272 seconds)
20:29:15kugelgevaerts: oops, my bad. didn't read the manpage carefully enough
20:29:22gevaertsAh, good :)
20:32:22Buscheln1s: fft8 is built within the codec context? is it possible to build with different options for the mdct part?
20:32:52 Join Dreamxtreme [0] (~Dre@
20:32:56kugelgevaerts: success \o/
20:33:04n1sBuschel: no, it's built as part of the codec lib which has it's own compilation settings
20:33:06*gevaerts starts a party
20:33:47 Join JesusFreak316 [0] (
20:34:15*kugel wonders what the "if (strcmp("/", this->path) != 0)" check is supposed to do
20:34:22 Quit evilnick_B (Quit: Page closed)
20:34:35kugelonly copy the path into the buffer if it's not / ??
20:34:50Buscheln1s: hmm... I somehown don't understand. we build it with -O3, but this is only faster for vorbis? shouldn't it be faster for the other codecs using the lib as well?
20:35:05kugelit looks plain wrong to me
20:35:17n1sBuschel: ah, no, building vorbis with O3 is faster
20:35:32n1scodeclib is at O2 currently
20:35:34Buschelok :)
20:35:55*Buschel thought he had one beer too much
20:37:02 Join dantje [0] (
20:37:34Buschelbut as this isn't the case I'll just take another one :)
20:37:42n1smost codecs that have a lot of code in iram don't fit in iram when built with O3 with gcc 4.4 since the code usually gets much bigger
20:38:37Buscheln1s: from what I saw (on ARM) -O3 was seldom faster as -O2. and if it was faster, it was only a minor speed change
20:39:15n1sBuschel: yes, the difference is usually not that big but it varies with codec, cpu and compiler version :)
20:41:07Buschel0.8 MHz is quite a lot
20:41:23 Quit FOAD (Quit: I'll be back)
20:42:37n1sit's about 2-3% for vorbis so if i don't find something that closes it further i think i'll just live with it as the new compiler speeds up other codecs IIRC (will need to rerun that benchmark some day)
20:43:26Buschelhow much does it speed up other codecs (roundabout)? also a few percent?
20:43:52n1si've lost my spreadsheet so i don't know :/
20:44:33n1sbut afair there were some pretty big speedups (although this was before the latest rounds of optimizations so might well have changed)
20:44:40 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
20:44:48Buschelhehe, seems like we all make the same mistakes ;)
20:45:46n1si do have some logs left though so i can take a quick look
20:47:00 Join FOAD [0] (~dok@
20:48:18kugelperhaps you can think of more obscure cases where it could/would fail
20:49:32 Join Stummi [0] (
20:49:52 Join Kupop [0] (
20:50:42gevaertskugel: looks good, except I'd move the ifdefs a bit to have the entire search_roots_ll handling inside
20:51:19kugelthen I need an #else for normal builds
20:51:25n1sBuschel: seems like an overall win but these logs are old, flac and aache got big speedups, regular aac and wma got speedups of a couple of MHz and most others were in the +-1MHz range
20:51:25kugelbut what's there is equivalent
20:51:38n1sI'll rerun them soonish though
20:51:52gevaertsah, yes
20:52:23kugelyou need to remove that line which has my build dir hardcoded :)
20:52:33gevaertsWell obviously
20:52:40Buscheln1s: makes sense. does the gcc change reduce binsize like it did for ARM?
20:53:10n1sno, it grows a bit (the arm change was more due to the abi switch i think)
20:53:21kugelrockbox has a good time scanning multiple linux source trees :P
20:53:39n1smain binary grows about 10k
20:55:38***Saving seen data "./dancer.seen"
20:59:48 Join Strife89 [0] (cf90c980@gateway/web/freenode/ip.
21:01:32 Quit guymann (Ping timeout: 276 seconds)
21:03:07 Join guymann [0] (~charles@
21:03:10 Join Horscht [0] (~Horschti@xbmc/user/horscht)
21:04:12 Quit Stummi (Ping timeout: 245 seconds)
21:04:52n1sthere's one function that grows from 1.2k at O2 to over 5k at O3 and that is what's causing the iram to overflow...
21:10:57 Join whorerizen [0] (
21:11:01whorerizenJust got a new Fuze v2. Is it ok to update the Sansa firmware to the latest, before I install Rockbox? Also, will future Sansa updates be installable?
21:11:19kugeloh, readlink doesn't resolve relative paths, that should be fixed I guess
21:14:11 Quit whorerizen (Remote host closed the connection)
21:14:21 Quit mt (Ping timeout: 245 seconds)
21:21:14 Quit u42p (Quit: Leaving)
21:21:57 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
21:22:00clone4crwif anyone is interested, i've done some more work on the resistor calculator plugin in FS #11733 and #11734
21:23:23 Join saratoga [0] (442c8cf8@gateway/web/freenode/ip.
21:27:20 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
21:29:26 Join whorerizen [0] (
21:30:14whorerizenJust got a recertified Fuze v2. Wondering if you want me to run any kind of tests before and after I install RB...?
21:33:04*TheSeven found the base address of the ATA/MMC/whatever controller that's dealing with the HDD, but has no idea what to do with it
21:34:13 Quit Strife89TX (Quit: Now to test it.)
21:35:32 Quit JesusFreak316 (Read error: Connection reset by peer)
21:37:02 Join stephen__ [0] (~stephen@
21:38:18 Join JesusFreak316 [0] (
21:38:57 Quit stephen__ (Client Quit)
21:39:24whorerizenTrying to use RB Utility to install on Fuze v2. When I click "Autodetect" I get, "Could not detect a Mountpoint Select your Mountpoint manually. Is this OK?
21:39:25 Quit Horscht (Quit: Verlassend)
21:41:24 Quit Strife89 (Quit: Need to take the Passport with me.)
21:42:55domonokywhorerizen: jup, thats normal. Rb Utility can not detect the mountpoint on some players. Just set it manually.
21:44:47 Join Horscht [0] (~Horschti@xbmc/user/horscht)
21:48:13 Join stoffel [0] (~quassel@
21:54:58 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
21:57:24 Quit kazaik (Quit: Leaving)
22:03:22 Quit whorerizen (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
22:04:02TheSevendoes anyone have a copy of the MMC standard?
22:04:19TheSeven(or some good paper about it, listing the commands etc)
22:04:59 Quit T44 (Quit: Leaving)
22:06:17pamauryisn't mmc an open standard ?
22:08:43 Join Topy44 [0] (
22:08:49TheSevenif you are a JEDEC member, yes :/
22:09:03TheSeven"open" in terms of "no royalties"
22:09:48*TheSeven might check if he can access that through the university
22:10:07 Join lixxus99 [0] (
22:10:39 Part lixxus99
22:10:58pamaurylet me see if I have access to something
22:11:51 Quit JesusFreak316 (Ping timeout: 245 seconds)
22:13:40pamaurythere is open access to eMMC standard on google iic
22:23:03 Quit bmbl (Quit: Verlassend)
22:27:19saratogaMMC is the same as SD right?
22:32:25 Join niekie [0] (~niek@CAcert/Assurer/niekie)
22:32:52n1shmm, so it seems like gcc 4.4 sometimes thinks the stack needs to be accessed with halfword accesses this probably is the reason for code size increases in a lot of places and likely some slowdowns... i wonder if gcc 4.5 is better at this
22:36:03 Quit Keripo1 (Read error: Connection reset by peer)
22:37:10 Quit GeekShadow (Ping timeout: 240 seconds)
22:37:26 Join Stummi [0] (
22:38:01 Join eWill [0] (
22:38:17pamaurysaratoga: is it ? I think there are different standard
22:38:59eWillIs there some file on the Fuze v2 I can set to 'read only' to prevent the Database Refresh, when I have to use the OF? (that's what I did on my e200 and c200)?
22:39:00 Join Keripo [0] (
22:39:12saratogawikipedia says they're largely compatible with some differences
22:39:25saratogaeWill: Nope
22:41:05kugelrealpath sucks on android
22:41:07 Quit benedikt93 (Quit: Bye ;))
22:41:26 Quit dantje (Quit: Ex-Chat)
22:42:04 Quit niekie (Ping timeout: 245 seconds)
22:42:29*TheSeven found a function that's apparently sending an MMC command and possibly also reading a response, but can't seem to figure out the args
22:43:23 Quit FOAD (Quit: I'll be back)
22:43:33 Join FOAD [0] (~dok@
22:44:04pamaurywhere ?
22:44:24 Join GeekShadow [0] (
22:44:24 Quit GeekShadow (Changing host)
22:44:24 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
22:45:30 Quit eWill (Quit: ChatZilla 0.9.86 [Firefox 3.6.12/20101026210630])
22:49:29TheSevenpamaury: ipod classic 1g disk mode, 8009380
22:49:33 Quit riotz (Changing host)
22:49:33 Join riotz [0] (riot33r@unaffiliated/riotz)
22:49:33 Quit riotz (Changing host)
22:49:33 Join riotz [0] (riot33r@gateway/shell/
22:49:47 Quit user890104 ()
22:49:54 Join user890104 [0] (
22:49:57pamauryah, I don't have such a beast, sorry ;)
22:53:41TheSevenhm, the MMC protocol is just weird...
22:53:56 Join kazaik [0] (
22:55:40***Saving seen data "./dancer.seen"
23:04:48n1sah, 4.5 is doing this properly again
23:05:33 Quit ender` (Quit: Light a fire for a man, and he's warm for a day. Light a man on fire, and he's warm for the rest of his life.)
23:07:20 Quit Horscht (Quit: Verlassend)
23:08:28 Quit kevku (Quit: KVIrc 4.0.2 Insomnia
23:10:23n1sbuuut, it's slower anyway?!
23:11:19 Join JesusFreak316 [0] (
23:12:08 Part Zagor
23:16:08 Quit JdGordon (Ping timeout: 250 seconds)
23:16:51 Quit Stummi (Ping timeout: 240 seconds)
23:17:26CIA-7New commit by bertrik (r28662): Sansa c200/c200v2: limit allowable contrast value to 0-127, it wraps at 128.
23:17:45TheSevendoes anyone want to join my ipod classic efforts, especially reversing the hdd driver?
23:17:51TheSevenideally someone who has dealt with MMC before
23:18:37gevaertsTheSeven: is it all MMC?
23:18:50bertrikDoesn't amiconn know about MMC?
23:19:02TheSeveni see loads of mmc/ce-ata, and i haven't seen any traces of regular PATA yet
23:19:16CIA-7r28662 build result: All green
23:19:22TheSevenand the code is full of debug_printf() calls in various error conditions :)
23:19:52gevaertsMaybe the PATA code worked the first time? :)
23:20:41TheSevenjust some examples:
23:20:42TheSeven"mI: SET_RELATIVE_ADDRESS failed\n"
23:20:53TheSeven"mC: timed out waiting for DSTA_RESEND\n"
23:21:20 Quit TheLemonMan (Quit: Help me, i got shot! *DIES*)
23:21:21TheSeven"mC: DSTA_CMDRDY not set after setting command/argument (DSTA = 0x%08x)\n"
23:21:53TheSeven"IAD: failed to write ATA task file\n"
23:22:32TheSeven"cI: Set drive to MMC high speed failed\n"
23:22:50TheSeven"cI: CE-ATA signature missing (%x,%x)\n"
23:22:58amiconnThe Ondio uses MMC, but in SPI mode
23:23:08amiconnThat's somewhat different from MMC mode
23:23:38TheSevenonce we figure out which register bit is what, things should be rather easy
23:24:01amiconn(the bridge uses it in MMC mode, but there's nothing to do for us except enabling the bridge (and clocks) with some GPIOs)
23:25:24TheSevenand if I see something like sub_8009380(0x230041, 0xFF8000, NULL, hdd_driver.command_timeout), I'm pretty sure that the lower bits of the first arg are the MMC command number. but what's the rest?
23:25:46TheSeventhe 0x30000 bits seem to be related to the response to be expected
23:25:54TheSeven0x200000 is almost always set
23:26:07TheSeventhe first argument is directly written into some register of the MMC controller
23:26:19 Quit user890104 ()
23:28:11 Quit GeekShadow (Quit: The cake is a lie !)
23:28:18TheSevenimmediately after that command there's a string saying "mI: SEND OP COND failed\n"
23:28:39 Join Stummi [0] (
23:29:52bertrikOP COND is also an SD command to ask for capabilities of SD cards IIRC (e.g. supported voltage range)
23:30:43TheSevenyes, and it's command ID 0x01 (for MMC at least), matching the first argument
23:31:07TheSeventhere are also lots of similar calls with different commands
23:31:27CIA-7New commit by bertrik (r28663): Update Russian translation - FS #11770 by Alexey Polkhirev
23:31:53 Quit stoffel (Remote host closed the connection)
23:33:01CIA-7r28663 build result: All green
23:33:10 Quit Xerion (Quit: )
23:35:11 Quit kazaik (Quit: Leaving)
23:37:29 Quit Stummi (Ping timeout: 245 seconds)
23:37:51 Join lixxus991 [0] (
23:38:43 Part lixxus991
23:43:56 Join kazaik [0] (
23:48:06 Quit domonoky (Read error: Connection reset by peer)
23:48:13 Quit n1s (Quit: Lämnar)
23:49:49 Quit komputes (Remote host closed the connection)
23:50:19 Quit JesusFreak316 (Read error: Connection reset by peer)
23:50:49 Join JesusFreak316 [0] (
23:51:14 Quit Kupop (Ping timeout: 245 seconds)
23:53:02 Quit bertrik (Quit: :tiuQ)
23:55:46 Join Horscht [0] (~Horschti@xbmc/user/horscht)

Previous day | Next day