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 2009-12-17

00:00:33YoYoFreakCJbut I think that even a quick formatting goes so far to change the byte sector formatting accordingly
00:01:16gevaertsThe problem is that the disk itself doesn't tell you which sector size to use, and neither does the MBR, so you have to guess, and rockbox tries 512 bytes first. If there happens to be a valid FAT32 boot sector at that offset, rockbox tries that one first
00:01:48gevaertsAnd this offset is still far enough on the disk not to be overwritten by itunes
00:02:41gevaertsthe easiest solution for this is to fully wipe the firmware partition (or the entire disk, but that will take ages), and then restore using itunes again
00:02:45YoYoFreakCJso a firmware recoverz bz iTunes might not fix it_ is that what zou|re sazing_
00:03:08gevaertsif at least this is what happened
00:03:12YoYoFreakCJwoopsy sorry
00:03:48YoYoFreakCJI only see the one big "FREAKPOD" drive in my workstation. until a few minutes ago I didnt even know that there is a firmware partition "^^
00:04:58*gevaerts thinks
00:05:45YoYoFreakCJeven in the administration management thingy in the control panel I dont see any other partitions. is there a way to format it using windows?
00:06:05***Saving seen data "./dancer.seen"
00:06:10gevaertsprobably. I don't know much about windows...
00:06:26YoYoFreakCJwhen I get home I plan on using DBAN on it. if that doesnt kick it, I don't know how to find that partition
00:06:55gevaertsoh, if I'm right, dban will definitely fix it for you
00:07:16YoYoFreakCJsounds good :3
00:07:42YoYoFreakCJI guess Ill be cleaning up and going home then to try it out
00:07:50*gevaerts thinks that rockbox should be more robust against this case, but he can't think of a good way to achieve that
00:08:23YoYoFreakCJisnt there a way to determine the sector size and act accordingly?
00:09:16notlisteningis there a way to query the disc about that information?
00:09:44gevaertsnotlistening: the disk *always* uses 512 byte sectors. It's a purely software thing
00:10:01notlisteningoh ok :P
00:10:20notlisteningteach me to come in half way trough
00:10:28gevaertsthere are two ways. One is to look at the FAT bootsector, but that's a bit chicken-and-egg because you need to know the sector size to find that bootsector (this is what rockbox does, by trying 512, 1024 and 2048)
00:11:34gevaertsThe other way would be to look at all partitions and at the total disk size, and then pick the sector size that matches best. That would break if you don't use all disk space for partitions though (and there are some reasons for doing that)
00:12:14gevaertsOr we could change the bootloader to try the next size if it finds a FAT but no files, but that requires some pretty invasive work...
00:12:52YoYoFreakCJanyways, I'm going home now and try DBAN on it. thanks for the explanations so far
00:13:14YoYoFreakCJbyebye :)
00:13:45gevaertsnotlistening: basically, the 5.5G uses 2048 byte sectors because the 80GB variant of it uses a disk that can't handle single-sector transfers, so the minimum transfer size is 1024 bytes, so using FAT with 512 byte sectors would not work well. I don't know why they then went to 2048 bytes though
00:14:10 Quit ps-auxw (
00:14:10 Quit linuxguy3 (
00:14:10 Quit elcan (
00:14:10 Quit fish_ (
00:14:10 Quit crashd (
00:14:10 Quit Unhelpful (
00:14:10 Quit reid02 (
00:14:10 Quit sdubois (
00:14:10 Quit tmzt (
00:14:10 Quit Topy (
00:14:10 Quit gibbon_ (
00:14:10 Quit rvvs89 (
00:14:10 Quit scorche (
00:14:10gevaertsnotlistening: over USB, the disk appears as 2048-byte sectors, so host OSes don't have this problem
00:14:11notlisteningi know there are not to many users of speech but if the svn version is out of sync with the build version and you generate the voixe file using rbutil o seem to get problem with the voicing on the menu system
00:14:55linuxstbgevaerts: What about on the 80GB ipod - can/do we detect that feature of the disk?
00:15:23*linuxstb also wonders if the hwrevision (stored in a fixed location in flash ROM) could give a clue
00:15:37gevaertslinuxstb: no, it's all handled in the same way. However, since that one doesn't handle 512 byte transfers, I don't think you can actually format it wrong
00:16:22linuxstbgevaerts: Isn't it possible for a host OS to work around that?
00:18:10NJoinps-auxw [0] (
00:18:10NJoinsdubois [0] (
00:18:10NJoinreid02 [0] (
00:18:10NJoinrvvs89 [0] (n=ivo@pdpc/supporter/base/rvvs89)
00:18:10NJoinscorche [50] (n=scorche@rockbox/administrator/scorche)
00:18:10NJoinTopy [0] (
00:18:10NJoingibbon_ [0] (
00:18:10NJointmzt [0] (n=tmzt@
00:18:10NJoinelcan [0] (
00:18:10NJoinlinuxguy3 [0] (
00:18:10NJoinUnhelpful [0] (n=quassel@rockbox/developer/Unhelpful)
00:18:10NJoincrashd [0] (
00:18:10NJoinfish_ [0] (
00:18:10linuxstbAlso, Apple have a different firmware for the 5.5G ipods (but just one), so maybe that's why all disks use 2048 byte sectors.
00:18:23 Quit YoYoFreakCJ ("CGI:IRC (Ping timeout)")
00:18:32gevaertsNot using a 5.5G ipod. If you use something else, maybe, although I suspect that any usb thing will give you IO errors. Possibly it will work using an ATA adapter
00:18:52 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
00:18:54 Quit ps-auxw (Remote closed the connection)
00:18:56gevaertsI mean, 1024 byte sectors would have worked fine
00:19:04 Quit PaulJam__ (Read error: 110 (Connection timed out))
00:19:14 Join ps-auxw [0] (
00:19:29gevaertsbut I guess nearly nobody uses those, so maybe HFS doesn't work on that
00:20:46gevaertsIf we can reliably detect that we're on a 5.5G, I guess we could make use of that
00:21:00 Quit ps-auxw (
00:21:00 Quit linuxguy3 (
00:21:00 Quit elcan (
00:21:00 Quit fish_ (
00:21:00 Quit crashd (
00:21:00 Quit Unhelpful (
00:21:00 Quit reid02 (
00:21:00 Quit sdubois (
00:21:00 Quit tmzt (
00:21:00 Quit Topy (
00:21:00 Quit gibbon_ (
00:21:00 Quit rvvs89 (
00:21:00 Quit scorche (
00:21:18NJoinps-auxw [0] (
00:21:18NJoinsdubois [0] (
00:21:18NJoinreid02 [0] (
00:21:18NJoinrvvs89 [0] (n=ivo@pdpc/supporter/base/rvvs89)
00:21:18NJoinscorche [50] (n=scorche@rockbox/administrator/scorche)
00:21:18NJoinTopy [0] (
00:21:18NJoingibbon_ [0] (
00:21:18NJointmzt [0] (n=tmzt@
00:21:18NJoinelcan [0] (
00:21:18NJoinlinuxguy3 [0] (
00:21:18NJoinUnhelpful [0] (n=quassel@rockbox/developer/Unhelpful)
00:21:18NJoincrashd [0] (
00:21:18NJoinfish_ [0] (
00:21:23linuxstbAnyone with a 5G or 5.5G ipod around?
00:21:25*linuxstb plugs in his 5G ipod to charge...
00:21:33*gevaerts has a 5G too, but it also needs charging...
00:21:41*gevaerts also needs sleep :(
00:22:13gevaertsGodEater has a 5.5G
00:22:28*DerPapst raises hand
00:23:00 Quit petur ("Zzzzz")
00:23:00linuxstbI'm pretty sure Rockbox reports the hardware revision (a 32-bit number) in the debug menu somewhere. We should check if the 5.5Gs have a different value there.
00:23:16 Quit GeekShadow (Read error: 54 (Connection reset by peer))
00:23:42 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
00:24:35DerPapst5.5G HW rev. 0x000B0011
00:24:54 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
00:25:36linuxstbMy 5G is 0x000B0005
00:26:07linuxstb(Debug -> View HW info)
00:29:10gevaertsMy 5G is 0x000B0005
00:30:55linuxstbSo based on a sample of 3, it would work...
00:31:02 Join Topy44 [0] (
00:31:14 Quit ps-auxw (
00:31:14 Quit linuxguy3 (
00:31:14 Quit elcan (
00:31:14 Quit fish_ (
00:31:14 Quit crashd (
00:31:14 Quit Unhelpful (
00:31:14 Quit reid02 (
00:31:14 Quit sdubois (
00:31:14 Quit tmzt (
00:31:14 Quit Topy (
00:31:14 Quit gibbon_ (
00:31:14 Quit rvvs89 (
00:31:14 Quit scorche (
00:31:16 Join fish_ [0] (n=fish@
00:31:17gevaertsIf we use the information just to reverse the search order, it shouldn't actually break anything even if we get the check wrong
00:31:17NJoinsdubois [0] (
00:31:20 Join Unhelpful [0] (n=quassel@
00:31:22 Join crashd [0] (i=foobar@
00:31:42NJoinps-auxw [0] (
00:32:13NJoinscorche [50] (n=scorche@rockbox/administrator/scorche)
00:32:14 Join reid02 [0] (n=reid02@
00:32:33linuxstbgevaerts: Is there any need to search? If I'm understanding correctly, that number will tell itunes what firmware to install (5G or 5.5G), and it's that choice of firmware which determines the sector size.
00:32:42 Join rvvs89 [0] (
00:32:44NJoinlinuxguy3 [0] (
00:32:51 Quit GeekShadow (Read error: 54 (Connection reset by peer))
00:32:53 Nick Ypsy is now known as YPSY (
00:33:14linuxstbgevaerts: I'm sure people will soon shout if that assumption is wrong ;)
00:33:16 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
00:33:52 Join sustineo [0] (
00:34:00 Join Guest23293 [0] (
00:34:12gevaertslinuxstb: for normal unmodified ipods, there's no need to search. However, the code is shared with other devices that might need this (the 240GB disks with 4096 byte sectors are one example), and I don't know what happens if you use a 5.5G ipod with a different disk and Torne's boot method...
00:34:14sustineois there a way to rotate album art in the wps?
00:34:25sustineo(im making a grunge theme)
00:34:37 Quit n17ikh (Nick collision from services.)
00:34:48 Nick Guest23293 is now known as n17ikh (
00:34:54 Quit n17ikh (Client Quit)
00:35:03linuxstbDerPapst: Can you remember how IPL handled this? Was anything special needed, or did the kernel "just work" ?
00:35:16 Join n17ikh [0] (
00:35:49gevaertssustineo: no
00:36:04DerPapstlinuxstb: iirc the driver worked "out of the box" but it assumed the wrong sectorsize.
00:36:07NJoingibbon_ [0] (
00:36:14NJointmzt [0] (n=tmzt@
00:36:24sustineothat sucks
00:37:14 Quit sustineo (Client Quit)
00:38:42 Quit bertrik ("De groeten")
00:39:18gevaertsok, they also try all of them until they find one, except that they start at 2048
00:40:59linuxstbgevaerts: Could the case of other disks be handled by querying the disk's capabilities? Do you know anything about the changes in the unofficial "big disk" build that means it's not in svn?
00:41:01gevaertsThat's also not 100% foolproof unfortunately. I suspect it's vulnerable to the same remains-of-old-filesystem problem, as well as a filesystem image that happens to be at the right place
00:41:34gevaertslinuxstb: I don't know the details. That's actually the main reason I'm hesitant
00:42:21linuxstbCurrent SVN doesn't appear to support those disks anyway, so we're not breaking anything...
00:43:27gevaertshm, in extreme cases (or weird coincidences), trying 2048->1024->512 can also break with real valid filesystems, if you use exactly the right partition layout and multiple FAT partitions
00:44:43 Quit matsl (Remote closed the connection)
00:45:01gevaertslinuxstb: you could be right. I suspect amiconn knows though
00:46:21 Join ZooYork2007 [0] (
00:48:16*linuxstb wonders how to filter flyspray searches by player type
00:48:19 Join JdGordon1 [0] (
00:49:05 Part ZooYork2007
00:52:46 Join grndslm [0] (
00:56:27 Quit Omlet^study ("( :: NoNameScript 4.22 :: )")
01:00:22 Join froggyman [0] (
01:05:44 Quit efyx_ (Remote closed the connection)
01:07:20 Quit notlistening ("Leaving")
01:09:28 Quit grndslm (Connection timed out)
01:09:49 Quit GeekShadow ("The cake is a lie !")
01:16:13 Join grndslm [0] (
01:18:15 Quit JdGordon1 (Read error: 113 (No route to host))
01:19:21 Quit grndslm (Operation timed out)
01:29:54 Join grndslm [0] (
01:35:10 Join funman [0] (n=fun@rockbox/developer/funman)
01:42:43 Quit ps-auxw (
01:43:06NJoinps-auxw [0] (
01:43:58 Quit Hillshum (Read error: 60 (Operation timed out))
01:46:38 Join Hillshum [0] (
01:47:20 Join Tomis2 [0] (n=Tomis@
01:47:28 Quit killan (Read error: 54 (Connection reset by peer))
01:51:55 Quit Tomis (Read error: 60 (Operation timed out))
01:51:55 Nick Tomis2 is now known as Tomis (n=Tomis@
01:56:45 Quit robin0800 (Remote closed the connection)
02:03:04 Quit liar (Read error: 60 (Operation timed out))
02:03:25 Quit ps-auxw (
02:06:07***Saving seen data "./dancer.seen"
02:08:32NJoinps-auxw [0] (
02:10:20 Quit ps-auxw (
02:13:00NJoinps-auxw [0] (
02:13:01 Part toffe82
02:34:27 Part froggyman
02:39:51 Join DerPapst1 [0] (
02:41:03 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
02:42:25 Quit Hillshum (Read error: 60 (Operation timed out))
02:43:22 Quit rasher (Read error: 104 (Connection reset by peer))
02:43:51 Join killan [0] (n=nnscript@
02:53:20 Join rasher [0] (
02:56:44 Quit DerPapst (Read error: 110 (Connection timed out))
03:02:49 Quit arohtar (Client Quit)
03:06:05 Quit advcomp2019__ (Read error: 110 (Connection timed out))
03:07:18 Join Hillshum [0] (
03:13:17 Quit DerPapst1 ("Leaving.")
03:44:32 Quit zu (Remote closed the connection)
03:46:11 Quit nosa-j (Read error: 54 (Connection reset by peer))
03:49:22 Nick fxb is now known as fxb__ (
03:51:56 Quit BlakeJohnson86 (Read error: 104 (Connection reset by peer))
03:52:24 Join BlakeJohnson86 [0] (
03:52:53 Join zu [0] (
03:56:23 Nick lol2 is now known as martian67 (
03:58:56 Quit panni_ ("( :: NoNameScript 3.81 :: )")
04:01:13 Join Dhraakellian [0] (
04:06:10***Saving seen data "./dancer.seen"
04:06:19 Quit ps-auxw (
04:06:19 Quit rasher (
04:06:44 Join rasher [50] (n=rasher@rockbox/developer/rasher)
04:06:44NJoinps-auxw [0] (
04:06:48 Join froggyman [0] (n=sopgenor@
04:07:56 Quit ps-auxw (
04:07:56 Quit rasher (
04:11:20NJoinrasher [50] (n=rasher@rockbox/developer/rasher)
04:11:20NJoinps-auxw [0] (
04:13:50 Join JdGordon1 [0] (
04:14:47 Quit ps-auxw (
04:14:47 Quit rasher (
04:15:08NJoinrasher [50] (n=rasher@rockbox/developer/rasher)
04:15:08NJoinps-auxw [0] (
04:15:24 Join Strife89 [0] (
04:33:37 Part froggyman
04:46:19 Quit funman ("free(random());")
04:46:36 Quit antil33t1 ()
04:50:04 Quit Hillshum (Read error: 60 (Operation timed out))
04:51:46 Join Hillshum [0] (
05:41:25 Quit JdGordon1 (Read error: 60 (Operation timed out))
06:06:12***Saving seen data "./dancer.seen"
06:31:51 Quit Strife89 ("Bed.")
06:32:43 Join MikeJB [0] (n=Michael@wesnoth/umc-dev/developer/aethaeryn)
06:33:38MikeJBAccording to the iPod FAQ: it says 6G (Classic) isn't supported? What's the problem?
06:40:18 Join Tekno_Cowboy [0] (n=Tekno_Co@
06:41:37*Dhraakellian thinks he got some corruption on his r24000 Fuze
06:41:50*Dhraakellian glances at since-4weeks.html
06:42:02Dhraakellianfixed in r24019?
06:43:37Tekno_CowboyHello, I have an iPod Video 5.5G, and I'm wondering if it's possible to use the themed USB mode that shows immediately after plugging in the USB cable. I have yet to find a setting that will prevent my ipod from rebooting into the black and white disk mode.
06:48:38 Quit GodEater (Remote closed the connection)
06:48:48 Join GodEater [0] (
07:06:11 Part MikeJB
07:11:09 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
07:19:17Unhelpful"themed usb mode"?
07:21:21Tekno_CowboyWhen I first plug in the USB cable, a USB cable appears on the screen with the current theme applied. It then reboots into a very basic monochrome disk mode. I'd prefer to use the one that doesn't require a reboot.
07:22:01UnhelpfulTekno_Cowboy: what version or revision are you using?
07:23:29Tekno_CowboyI'm using version 3.4, I'm not sure what revision.
07:25:27amiconngevaerts, linuxstb: Current svn doesn't support those 4096-byte sectored disks, but I consider this a bug
07:26:29amiconnIt's because our code simply assumes it has to use the disk's sector size, instead of probing whether the disk handles split sectors
07:26:55Unhelpfulhrm, is usb enabled on ipod 5.5G in 3.4? MajorChanges seems to say that pretty much everything has rockbox usb enabled in 3.4, but i seem to recall otherwise :/
07:27:18amiconnThose 4096-byte sector disks do proper split sector handling, only the original 80GB G5.5 hdd doesn't
07:28:09Unhelpfulah, 5.5G is not among the PP chips mentioned in that list according to DeviceChart...
07:29:15amiconnSo in order to use such a disk, one can either *disable* large disk sector support in rockbox, or increase the limit to 4096 bytes. The former method yeilds better performance iirc
07:29:42Tekno_CowboyI see. That's too bad. Hopefully a future version will support it. Might I ask where you found the chart? I looked all over for it before trying IRC, with no luck.
07:29:43UnhelpfulTekno_Cowboy: it looks to me as though USB is not supported on your device in 3.4, although i'm not absolutely certain. have you tried a current build instead of the 3.4 release?
07:30:26 Quit Horschti (Read error: 110 (Connection timed out))
07:30:28Tekno_CowboyI have not tried a current build yet. How stable do the current builds tend to be?
07:30:36UnhelpfulTekno_Cowboy: the info i'm going by is here: and here: but i should point out that neither list is absolutely exhaustive.
07:31:56 Quit dmb (Read error: 104 (Connection reset by peer))
07:32:08UnhelpfulTekno_Cowboy: they're generally safe unless something broken has just been committed, and we try not to commit broken things and to fix them reasonably quickly. i've never had a release version installed on my devices, and they've pretty much only broken due to things i had changed myself.
07:35:04Tekno_CowboyI'll try the current build and see what happens. If it doesn't work out, I can always revert to 3.4 and wait for the next release. Thanks for all yor help!
07:39:25 Quit Tekno_Cowboy (Remote closed the connection)
07:41:04 Quit maraz ("...")
07:46:35 Join PaulJam__ [0] (
07:56:57 Quit PaulJam__ (Nick collision from services.)
07:57:00 Join PaulJam_ [0] (
07:57:11 Quit sdubois (Read error: 60 (Operation timed out))
07:58:39 Join sdubois [0] (
08:02:42 Join JdGordon| [0] (
08:03:41 Quit PaulJam_ (Nick collision from services.)
08:03:46 Join PaulJam [0] (
08:05:38 Quit JdGordon| (Client Quit)
08:06:13***Saving seen data "./dancer.seen"
08:06:21 Quit sdubois (Read error: 60 (Operation timed out))
08:06:46 Join sdubois [0] (
08:17:40 Join hebz0rl [0] (
08:29:46bluebroth3rB4gder: can you check the manuals? It seems they haven't been built at all, but they build fine for me.
08:47:10CIA-6New commit by jdgordon (r24037): code police and make do_menu() obey the hide_bars param which was never meant to be removed
08:48:03JdGordondo we have a stop scroll funciton which only stops the scrolling in a viewport? not the full display?
08:49:05 Join Kitr88 [0] (
08:52:06CIA-6New commit by jdgordon (r24038): only stop scrolling in the menu's viewport not the whole display (Fixes scrolling lines in sbs's being stopped when moving around menus)
08:52:24JdGordonapparently we do :)
08:54:39 Quit Kitar|st (Read error: 60 (Operation timed out))
08:55:27 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:55:28 Quit Kitr88 (Read error: 60 (Operation timed out))
08:55:34 Join Kitar|st [0] (
09:02:37 Join petur [50] (n=petur@rockbox/developer/petur)
09:06:01 Quit gevaerts (Nick collision from services.)
09:06:14 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
09:09:06CIA-6New commit by jdgordon (r24039): from FS #10771 - fix the clear backdrop option so it clears the full screen
09:09:49*GodEater sees no HWCODEC testing was done last night :(
09:10:48JdGordonwhat was the FS#?
09:11:15 Join niekie [0] (i=quasselc@CAcert/Assurer/niekie)
09:11:22 Join flydutch [0] (
09:12:29JdGordonGodEater: I really dont like the idea of playback.c being compiled for hwcodec...
09:12:54JdGordonthe better solution is to move that new code to another file all together, or duplicate it in mpeg.c
09:13:26GodEaterwhich I believe is what pondlife said he'd do over a year ago and never did :)
09:13:26JdGordonespecially when its 2 speerate code paths anyway
09:14:47*JdGordon is usually not one to object to a feature... but this one needs more work...
09:16:04JdGordonit could be cleaned up a fair bit.. the events arnt needed... they were only there because when it was first done mpeg.c was in firmware where events were required for it to work
09:16:20*GodEater would welcome help on it.
09:17:11*JdGordon will have a quick play
09:20:41JdGordonactually.. no I guess thats fine
09:20:54JdGordonbut the code for the event should be in playlist.c or a new file
09:22:18 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
09:22:32*GodEater senses someone's ears were burning ;)
09:22:37 Part pondlife
09:22:43GodEater maybe not....
09:22:43JdGordontoo hot
09:23:01GodEaterhow did that get into my copy buffer
09:23:07JdGordonthe other chan :)
09:23:10JdGordoncould be worse...
09:23:13GodEateryep, but I didn't copy it!
09:30:57JdGordonhmm.. does that patch work in the sim?
09:32:08GodEaterit *compiles* into the sim, but I've not tested it. LinusN says there's a noticeable lag.
09:32:16GodEaterbut as he says, "who cares"? :)
09:32:27JdGordonit needs to work in the sim..
09:32:38JdGordonand it shouldnt lag.. the wps can pause fine
09:32:48GodEaterin that case it probably needs a lot more work to find out why it lags
09:33:09GodEaterand I'm not an expert by any means in the difference between the sound code in the sim vs. real rockbox
09:34:46JdGordonin swcodec.. there is no difference untill the pcm is ready to be dumped i think
09:35:37GodEaterI've basically got not time left to look at it this week. Have too much paid work to do :(
09:36:21JdGordonah yuck... this feature is a bit nasty
09:37:01JdGordonit works on target?
09:37:14*GodEater is running it on his ipod right now
09:39:54 Join stripwax [0] (
09:40:37*GodEater has to run to a meeting....
09:43:39*JdGordon wonders why audio_pause() blocks
09:45:30 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
09:55:35 Join funman [0] (n=fun@rockbox/developer/funman)
09:56:27 Quit goffa (Read error: 110 (Connection timed out))
10:05:42 Join Bagder [0] (
10:06:14***Saving seen data "./dancer.seen"
10:08:56 Join maruk [0] (
10:11:42 Join thegeek_ [0] (
10:11:57 Join jfc^3 [0] (
10:12:45 Join fyre^OS [0] (n=nnscript@
10:13:09 Join AlexP_ [0] (n=alex@rockbox/staff/AlexP)
10:14:08 Join jordan`` [0] (n=jordan@
10:14:13 Join The3_14ed|r [0] (
10:15:07 Join FlynDice_ [0] (
10:16:09 Quit markun (
10:16:09 Quit flydutch (
10:16:09 Quit crashd (
10:16:09 Quit Utchybann (
10:16:09 Quit JdGordon|| (
10:16:09 Quit martian67 (
10:16:09 Quit thegeek (
10:16:09 Quit FlynDice (
10:16:09 Quit Galois (
10:16:09 Quit Res1 (
10:16:09 Quit niekie (
10:16:09 Quit jordan` (
10:16:09 Quit Dhraakellian (
10:16:09 Quit AlexP (
10:16:09 Quit chaos (
10:16:09 Quit Zaba (
10:16:09 Quit fxb__ (
10:16:09 Quit bzed (
10:16:09 Quit tchan (
10:16:09 Quit fyrestorm (
10:16:09 Quit avacore (
10:16:09 Quit scorche|sh (
10:16:09 Quit jfc^2 (
10:16:09 Quit jds2001 (
10:16:09 Quit alexbobp (
10:16:14NJoinbzed [0] (
10:16:35NJoinchaos [0] (n=chaos@gentoo/user/ch4os)
10:16:38NJoinjds2001 [0] (n=jds2001@fedora/jds2001)
10:16:45NJoinmarkun [50] (n=markun@rockbox/developer/markun)
10:17:19NJointchan [0] (n=tchan@lunar-linux/developer/tchan)
10:17:28NJoinniekie [0] (i=quasselc@CAcert/Assurer/niekie)
10:17:51 Join Zaba [0] (
10:18:41NJoinfxb__ [0] (
10:20:01 Join avacore [0] (i=nobody@
10:20:05 Quit JdGordon (Read error: 60 (Operation timed out))
10:20:20 Quit FlynDice_ (Read error: 60 (Operation timed out))
10:20:36 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
10:21:01 Join FlynDice_ [0] (
10:22:50NJoinmartian67 [0] (
10:25:35NJoinalexbobp [0] (n=alex@
10:28:35NJoinUtchybann [0] (
10:31:43 Quit stripwax ("")
10:31:44NJoinscorche|sh [50] (n=scorche@rockbox/administrator/scorche)
10:32:06 Join einhirn [0] (
10:35:46 Join crashd [0] (
10:40:51 Join Res1 [0] (
10:48:58gevaertsUnhelpful: USB is still disabled for releases on ipod because of the charging issues
10:50:12Unhelpfulgevaerts: thanks, i couldn't recall the specifics. :)
10:52:11gevaertsfor some reason we forgot that in the release notes though
10:55:13 Quit Zarggg (Read error: 104 (Connection reset by peer))
11:04:50 Quit PaulJam (Nick collision from services.)
11:04:57 Join PaulJam_ [0] (
11:06:35 Nick crwll is now known as crwl (
11:13:46 Join Zarggg [0] (
11:15:14 Join goffa [0] (n=goffa@
11:16:38gibbon_i was here yesterday for my build server being disabled due to disk quota problems
11:17:12gibbon_i saw mc2739 do a commit to unblock it, but it still gives me an error, that the client is blocked
11:20:55 Join stripwax [0] (n=Miranda@
11:22:49*gevaerts looks at Zagor
11:24:34Zagorgibbon_: a commit to unblock?
11:24:45gevaertsgibbon_: blockedclients
11:24:51gevaertsZagor I mean
11:26:21Zagorah, looked in the wrong place..
11:27:10stripwaxanyone seen mt or saratoga recently?
11:27:49gibbon_Zagor: r24018
11:28:36gibbon_the quota problems are resolved and since build clients are the only way i can currently contribute, i would be glad, if that machine would work again
11:30:20Zagoryes. I'm looking at the build server to see why it bugs
11:31:03gibbon_thank you very much
11:37:45amiconnZagor: Btw, the false duplicate client bug keeps showing up every now and then
11:38:00gevaertsah, true. I also have a bug...
11:38:56Zagoramiconn: right, that was never resolved properly
11:39:21CIA-6New commit by zagor (r24040): Clear blocklist before reading file
11:39:34gevaertsZagor: I get "Server message: Fatal build error: Missing log file. You have been temporarily disabled" regularly
11:39:39amiconnHmm, and I keep getting *** Subversion error! in my log now...
11:40:10amiconnHmm, Repository locked... how do I fix this?
11:40:29funmansvn unlock ?
11:40:37gevaertssvn cleanup I think
11:40:50Unhelpfulcleanup should usually do the trick
11:45:19amiconnThat helped, thanks
11:46:14*amiconn would like an active monitoring method
11:47:06gibbon_Zagor: thanks a lot, i am properly connected again :)
11:48:17 Quit stripwax ("")
11:50:39 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
12:00:25 Join stripwax [0] (
12:06:17***Saving seen data "./dancer.seen"
12:10:56 Join Galois [0] (
12:12:05 Join watto [0] (n=watto@
12:14:35 Join DerPapst [0] (n=DerPapst@
12:26:55 Join maraz [0] (i=maraz@
12:27:58 Quit stripwax (Read error: 104 (Connection reset by peer))
12:40:12 Join roolku [0] (
12:41:00*roolku gets that as well: "Server message: Fatal build error: Missing log file. You have been temporarily disabled"
12:42:01roolkuwould it be possible to re-enable it from the server side? It get's annoying having to restart the servers (and to monitor them all for that matter)
12:42:41roolkuthat was for Zagor −−^
12:49:12 Nick fxb__ is now known as fxb (
12:50:48 Join stripwax [0] (
12:51:58 Quit PaulJam_ (Nick collision from services.)
12:52:06 Join PaulJam__ [0] (
12:55:29 Join Xerion_ [0] (
12:57:28 Quit stripwax (Read error: 104 (Connection reset by peer))
12:57:34 Quit Xerion (Read error: 60 (Operation timed out))
12:57:34 Nick Xerion_ is now known as Xerion (
13:02:57 Join stripwax [0] (
13:04:22Zagorroolku: the temporary block now only lasts 10 minutes
13:08:47 Quit funman ("free(random());")
13:12:53 Join Sajber^1 [0] (
13:13:41 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:17:11 Quit Sajber^ (Read error: 60 (Operation timed out))
13:18:45 Join teru [0] (
13:24:48roolkuZagor: thanks I'll keep an eye on it
13:25:20 Join AlexP__ [0] (i=86ceaf37@gateway/web/freenode/x-epmyovuftsajleyk)
13:25:31 Nick AlexP__ is now known as AlexP (i=86ceaf37@gateway/web/freenode/x-epmyovuftsajleyk)
13:26:01 Nick AlexP is now known as Guest52169 (i=86ceaf37@gateway/web/freenode/x-epmyovuftsajleyk)
13:26:11 Part Guest52169
13:26:21 Quit roolku ()
13:26:43 Join fml [0] (
13:27:52 Join xavieran [0] (
13:28:02 Nick AlexP_ is now known as AlexP (n=alex@rockbox/staff/AlexP)
13:28:18fmlGodEater: Hello. Now that I read keymap tables not horizontally (row by row, as a sentence), but rather as tables (key: abc, action: xyz) I think that it might be better to use the do-form :-)
13:29:58AlexPI would prefer that
13:30:04AlexPAs would linuxstb
13:30:25AlexPI was too late to say so yesterday, as what discussion there was happened when I wasn't there
13:30:45fmlAlexP: I wouldn't object too. Now we have to convince GodEater
13:31:14fmlAlexP: No sorries. We can still change that :)
13:32:49 Quit PaulJam__ (Read error: 110 (Connection timed out))
13:34:02GodEaterI don't take much convincing
13:34:39fmlShall we strat a poll on that?
13:35:05GodEaterI think just do what AlexP /linuxstb said
13:39:41 Quit Zaba (Read error: 60 (Operation timed out))
13:43:24 Join Zaba [0] (
13:46:24 Quit BHSPitMonkey ("Ex-Chat")
13:47:11fmlGodEater, AlexP, linuxstb: should the actions end with a full stop? I'd unify this as well.
13:47:38GodEatersince I got the answer wrong yesterday, I'll pass the buck to linuxstb / AlexP ;)
13:48:13fmlI feel like they should not since these are not full sentences, but... ^^ :-)
13:48:33linuxstbfml: haha, I was going to say the opposite...
13:49:13 Join Xerion_ [0] (
13:49:50linuxstbfml: From a consistency point of view, it's probably easier to always have a full stop. Otherwise, you leave it to the author's judgement to decide if a sentence is "full" or not.
13:49:57fmllinuxstb: there are some cases where a full stop would be needed because there are some additions to the action description. And that should be separated somehow.
13:55:12GodEaterA semi colon.
13:55:19GodEaterit's underused in written english!
13:55:29GodEaterI think we should make it our mission to remedy that.
13:56:23fmlI also saw something like "Move about the virtual keyboard." SHouldn't it be "move _around_ ..."?
13:57:36 Quit Xerion_ (Read error: 104 (Connection reset by peer))
13:59:09 Join Xerion_ [0] (
13:59:36CIA-6New commit by teru (r24041): Update virtual keyboard manual.
14:05:16 Quit Xerion (Connection timed out)
14:05:59 Join JdGordon| [0] (
14:06:21***Saving seen data "./dancer.seen"
14:10:31 Join AlexP_lab [0] (i=86ceaf37@gateway/web/freenode/x-byawmbwmiusosppf)
14:11:24AlexP_labI'm not sure about the full stop :)
14:11:59AlexP_labI would probably say not, but I'm not too bothered
14:12:26AlexP_labfml: And yes, I saw that too - it should be move around
14:12:42linuxstbWhat about this page for example - what should not have a full stop?
14:13:05AlexP_labmove about the keyboard could suggest that it is the keyboard that you are moving
14:13:38AlexP_lablinuxstb: yes, I agree it is tricky
14:13:46AlexP_lablinuxstb: Let's go for full stop
14:14:16AlexP_labOn reflection I think that those that don't have it that should will look more odd than those that do have it but some people think might not need it
14:17:30stripwaxAlexP_lab - re "move about the keyboard" - didn't teru's change a few minutes ago fix that?
14:17:55stripwaxnow reads "Moves the cursor on the virtual keyboard"
14:19:00AlexP_labah, didn't see that
14:25:42fmlOk, the patch is in the FS (FS #10861). Please look at it. I did it in a half automatic mode :-)
14:26:38 Quit JdGordon| (Read error: 113 (No route to host))
14:28:16AlexP_labheh, will do
14:28:28AlexP_labIt might have to be a bit later though
14:29:20Zagorwhy is recorderv2 and fmrecorder sharing manual when ondiosp and ondiofm doesn't?
14:31:32linuxstbZagor: Don't some recv2's have radio?
14:31:41Zagoraha, do they?
14:32:31*linuxstb seems to recall they do, but wouldn't bet his house on it
14:33:27linuxstbYes, seems they do - the config file for the recv2 defines CONFIG_TUNER
14:33:40Zagorok, thanks
14:36:18 Quit xavieran (Client Quit)
14:37:54linuxstbZagor: I don't know if you have any time/desire to fix it, but the old releases are not available via the website. The "old releases" link on goes to very old releases (and not in a nice way...). I guess at the very least, that link could go to
14:39:29Zagoryeah I saw that just the other day too
14:44:11CIA-6New commit by mcuelenaere (r24042): Enable backlight brightness in simulator
14:44:45Zabathat's a nice revision number
14:49:18 Quit markun (Read error: 60 (Operation timed out))
14:49:22 Join markun [50] (n=markun@rockbox/developer/markun)
14:49:39 Quit fml ("CGI:IRC")
14:52:41 Join flydutch [0] (
14:53:30CIA-6New commit by zagor (r24043): Renamed special archosrecorderv2fm manual build to standard archosfmrecorder.
14:57:58 Join Jonathan__ [0] (
14:58:31Jonathan__Hey, I'm thinking about installing rockbox for my Fuze V1, what are the chances of it messing up and destroying my player?
15:00:08Jonathan__and also, if the firmware does mess up, are there ANY chances of reinstalling the old firmware and being able to use it again?
15:00:13Zabawell, it's not a bomb, so the odds are that it lacks proper circuity to explode anyway.
15:00:24Jonathan__lol... j
15:00:41Jonathan__um, it can still brick my player though... right?
15:01:05krazykitif you follow the instructions for installation it's very unlikely that it will brick your player
15:01:24krazykithowever, there's no recovery mechanism like the e200 series
15:01:54Jonathan__if I do a duel boot and the firmware does mess up... can I still boot into the old one?
15:02:10 Quit Jonathan__ (Client Quit)
15:03:36 Quit teru ("Quit")
15:03:56 Join Jonathan__ [0] (
15:04:08Jonathan__ok, I'm back, sorry about that
15:04:23Jonathan__if I do a duel boot and the firmware does mess up... can I still boot into the old one?
15:04:57krazykitit's already set as a dual-boot. so no, if it gets bricked, it's now a crummy paperweight
15:05:01linuxstbIn theory yet - the dual boot feature is very reliable (but we offer no guarantees).
15:05:23Jonathan__hmm, conflicting answers...
15:05:38Zabain short, if the bootloader doesn't get messed up, you should be fine..
15:05:45*linuxstb doesn't understand krazykit's answer
15:06:06Jonathan__if I get it installed and it's working, is there a chance its going to die later?
15:06:09krazykiti should clarify. if the bootloader installation gets messed up, then you're in trouble
15:06:13Jonathan__or is it mostly during the install process
15:06:16Jonathan__ok, thanks
15:06:28CIA-6New commit by mcuelenaere (r24044): Only do BACKLIGHT_FADING_SW_SETTING in simulator (fixes red)
15:06:46linuxstbJonathan__: Short answer is that 100s (possibly 1000s) of people have installed Rockbox on a fuze without problems, and we've designed it to be as safe as we can, but we don't guarantee anything.
15:07:06krazykitif a certain revision of rockbox doesn't work, you can easily boot into the original firmware and unzip a working build over it without issue
15:07:16Jonathan__ok, thanks guys
15:07:30Jonathan__I just don't have the money to replace it right now lol
15:08:21Jonathan__I suppose I will try it then. one more question. How big is it installed? I only have a 2GB mp3 player... will it fit ok?
15:09:51krazykitit's only a couple MB, so it'll fit no problem
15:09:58Jonathan__ok, cool. thanks
15:12:44 Quit maraz ("hyvinp tmkin toimii.")
15:15:01 Join |DaMaGeD| [0] (n=nikantan@
15:17:54Jonathan__thanks for your help guys, I suppose I will probably be on later with more problems lol. cya guys
15:18:04 Quit Jonathan__ ("ChatZilla 0.9.86 [Firefox 3.5.6/20091201220228]")
15:22:18CIA-6New commit by mcuelenaere (r24045): Don't do fading in simulator on targets with BACKLIGHT_FADING_PWM or BACKLIGHT_FADING_TARGET (should fix red)
15:22:23 Nick YPSY is now known as Ypsy (
15:27:49 Quit BHSPitLappy ("Ex-Chat")
15:33:38 Quit bzed (
15:33:38 Quit ps-auxw (
15:33:38 Quit rasher (
15:34:11 Nick FlynDice_ is now known as FlynDice (
15:40:32 Part LinusN
15:41:15 Join panni_ [0] (
15:43:22NJoinps-auxw [0] (
15:43:22NJoinbzed [0] (
15:43:22NJoinrasher [50] (n=rasher@rockbox/developer/rasher)
15:44:52 Quit |DaMaGeD| ("Ухожу я от вас (xchat 2.4.5 или старше)")
15:46:40 Quit Ypsy ("Oh noes! My ZNC obviously just shut down :(")
15:52:52 Join Ypsy [0] (
16:00:19 Join PaulJam__ [0] (
16:03:27 Quit Zagor ("Don't panic")
16:03:44 Join fml [0] (
16:05:01fmlmcuelenaere: I'm playing with a freshly build sim for sansa e200. After the backlight goes off it doesn't go on anymore, even if I press a button.
16:06:25***Saving seen data "./dancer.seen"
16:15:23 Join kugel [0] (i=kugel@rockbox/developer/kugel)
16:16:26kugelfml: that's because of sim_backlight(old_val); old_val is 0 since backlight fading uses _backlight_set_brighthness()
16:16:44 Join Ubuntuxer [0] (
16:17:18kugeltry to normalize backlight_brightness (that's a global variable) and use that in _backlight_on()
16:17:32 Quit Ubuntuxer (Client Quit)
16:19:46 Join Ubuntuxer [0] (
16:21:29kugelIIRC _backlight_on() shouldn't even change any brightness. it should just enable needed registers which doesn't apply for the sim
16:22:03kugelso, removing that line entirely (or rather putting it into an #if as the e200's target backlight driver does) should do it
16:31:12 Quit PaulJam__ (Nick collision from services.)
16:31:20 Quit Ubuntuxer ("Leaving.")
16:31:21 Join PaulJam_ [0] (
16:32:52 Quit Ypsy (
16:33:58 Quit PaulJam_ (Nick collision from services.)
16:34:05 Join PaulJam__ [0] (
16:37:53CIA-6New commit by Ubuntuxer (r24046): Hide menu entries "Resume Game" and "Quit without Saving" if not available and clean up the code
16:38:20 Quit tchan ("WeeChat 0.3.1-dev")
16:39:36NJoinYpsy [0] (
16:39:40 Join toffe82 [0] (n=chatzill@
16:39:47 Nick Ypsy is now known as YPSY (
16:40:19 Join tomers [0] (
16:40:20 Nick YPSY is now known as Ypsy (
16:41:25 Quit stripwax (Read error: 104 (Connection reset by peer))
16:43:44 Join LambdaCalculus37 [0] (n=LambdaCa@rockbox/staff/LambdaCalculus37)
16:45:31kugelUbuntuxer: I think your change is buggy
16:45:54kugelI also don't like how you ignore the comment and just make the delay caused by the splash longer
16:46:44 Join Ubuntuxer [0] (
16:46:45 Join stripwax [0] (
16:47:10kugelUbuntuxer: see logs :)
16:48:56 Quit PaulJam__ (Nick collision from services.)
16:48:58 Join PaulJam_ [0] (
17:01:18kaniiniis there a GUI front end to code rockbox?
17:03:18 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
17:04:24 Quit DerPapst ("Leaving.")
17:05:19 Join FOAD_ [0] (
17:07:16kaniiniin windows enviroment
17:07:30Bagderkaniini: your favourite editor
17:07:31AlexP_labkaniini: Sure, notepad, etc
17:07:55*AlexP_lab is not suggesting for a minute that notepad should be used by the way :)
17:08:26kaniiniso, the source can be opened in notepad and it's not gibberish?
17:08:28kaniinicool :)
17:08:42AlexP_labOf course - source is just text
17:08:53Bagderparts of it will probably be considered gibberish by some... :-)
17:09:57kaniiniwith gibberish i meant that it's readable w/o a dedicated software
17:10:21 Quit Lss__ (Read error: 104 (Connection reset by peer))
17:10:26AlexP_labRockbox is just C (bits of assembly)
17:11:10 Join Lss__ [0] (
17:16:31 Quit stripwax (Read error: 104 (Connection reset by peer))
17:20:57 Join stripwax [0] (
17:22:21 Quit FOAD (Read error: 110 (Connection timed out))
17:22:21 Nick FOAD_ is now known as FOAD (
17:25:34 Join Tomis2 [0] (n=Tomis@
17:26:52 Quit Tomis (Read error: 60 (Operation timed out))
17:26:52 Nick Tomis2 is now known as Tomis (n=Tomis@
17:28:45 Quit kugel (Read error: 110 (Connection timed out))
17:29:19 Join dfkt_ [0] (i=dfkt@unaffiliated/dfkt)
17:35:32 Quit dfkt (Read error: 60 (Operation timed out))
17:37:29 Join notlistening [0] (
17:37:47 Part Bagder
17:38:09 Quit reid02 ("Coyote finally caught me")
17:38:27notlisteningFlynDice, hey formatted full buld from svn last night and it started to work again now after a night being left on same issue with the Panic message
17:39:29 Join Omlet^study [0] (
17:46:29 Join cas_ [0] (
17:47:48 Nick cas_ is now known as Casainho (
17:47:55 Quit fml ("CGI:IRC")
17:53:48 Quit Ubuntuxer ("Leaving.")
17:54:27tomersHey, I have a question regarding manual: Should it say "Menu or Power", or "Menu, or Power". And "Menu, Down or Power" or "Menu, Down, or Power" ?
17:56:13linuxstbYou don't use ", or" (IMO)
17:57:07 Quit petur ("work->bike+snow->home")
17:57:23 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:57:27tomersIf I have 3 different buttons that does the same thing, should I use "X or Y or Z", or "X, Y or Z"?
17:58:22*linuxstb agrees
17:59:00pixelmain the text or in tables?
17:59:29tomerspixelma: Good point. Should I use it differently depending on the context?
17:59:32linuxstbpixelma: Does it matter?
18:00:21linuxstbBut where do three buttons all do the same thing? That doesn't sound sensible...
18:03:11 Join JdGordon| [0] (
18:03:14tomerslinuxstb: Just asking... I'm reviewing something in the manual. But for touchscreen targets, it could be like: Menu, Minus, Center or BottomMiddle
18:03:42tomersBecause the two options (two buttons) becomes 4 options (2 buttons, 2 touch regions)
18:04:45pixelmain the tables those lists of buttons are usually seperated by /
18:05:34linuxstbpixelma: I was just looking at that, and that convention seems to be used for things like "Scroll forward / Scroll back" - i.e. two different actions listed in the same row of the table.
18:05:49linuxstb(not two buttons doing the same thing)
18:05:59pixelmahmm... true
18:06:13linuxstbe.g. here -
18:06:23linuxstb"or" is used for two alternative buttons, and "/" is two different actions
18:06:28***Saving seen data "./dancer.seen"
18:06:46JdGordon|does anyone know why audio_pause() blocks? I cant of any reason it needs to...
18:06:53tomersAlso, for this cause, in LaTex, should I use \ButtonX{} / \ButtonY, or somehow else, as there's in some places - \ButtonX\ / \ButtonY?
18:07:05 Nick fxb is now known as fxb__ (
18:07:29pixelmaI think I rememenber seeing "or" a lot too
18:07:49pixelmaremember too
18:08:13AlexP_labDefinitely "x, y or z"
18:08:33tomerspixelma: There's a lot of 'or' in the Cowon D2 manual, and I'm not sure it is correct
18:09:12 Quit AlexP_lab ("Page closed")
18:09:23 Join kugel [0] (n=kugel@rockbox/developer/kugel)
18:09:24linuxstbtomers: This looks OK -
18:09:38linuxstb"or" is used for alternatives, and "/" is used for different actions
18:10:08linuxstbAlthough some way to differentiate touches from button presses would make it clearer.
18:10:31 Join MethoS- [0] (n=clemens@
18:10:39tomersThe first sentence in Minesweeper manual page looks totally obsolete. Am I wrong?
18:11:53pixelmaI think the second sentence is obsolete as this is now set through a menu
18:12:34 Join fml [0] (
18:13:03pixelmatomers: also be careful - not all touchscreen devices also have additional physical buttons
18:13:05tomerslinuxstb: Regarding the link you gave, shouldn't the slash have to be spaced? i.e. '\X{} / \Y', not just '\X{}/\Y' or even '\X{}\ \Y'?
18:13:39fmltomers: even better, use ButtonA~/ButtonB, i.e. a ~ instead of space before the /. This will prevent a line break before / which looks strangely in tables.
18:13:40tomerspixelma: I'm only working on the Cowon for now. Therefore using COWON_D2_PAD, not 'touchscreen'
18:14:15tomersfml: Good tip! I guess you mean \ButtonA{}~/~\ButtonB ?
18:14:17linuxstbtomers: To be honest, I'm not sure. What about the "/" in "previous/next" ?
18:14:29tomerslinuxstb: I'm not sure either
18:14:59tomersthat's why I'm asking (in contrary to how I behaved two days ago :- )
18:15:26 Quit hebz0rl ("Ex-Chat")
18:16:18tomersSo everyone agrees \ButtonA{}~/~\ButtonB for alternative buttons? (i.e. left/right)
18:17:43tomerspixelma: So in Minesweeper, should I delete the sentence: "Use Scroll Forward/Scroll Backward to select the required percentage of mines to set the difficulty then press the Select key to begin."?
18:18:27fmltomers: I think the second ~ is not needed because table columns are rather narrow. Tex will have problems with breaking lines or will have high penalties.
18:19:09fmlAlso I think we should change our macros so that a line break is possible within a button (I'm not sure if it's possible now)
18:19:18tomersfml: \ButtonA{}~/ \ButtonB ? If there's a consensus here, I like to go over the manual and change...
18:19:40tomersfml: The button macro only changes the font
18:19:54fmltomers: I would do it like that but I'm not god (yet ;-)
18:20:10tomersbut maybe I'm wrong here
18:21:16fmltomers: if you use ~ you don't need the {} after ButtonA. But they do no harm either.
18:22:18fmltomers: I put up a patch that reverts the yesterday's commit. Are you aware of that? (Before you start making your changes)
18:22:37tomersfml: You mean the 'does' patch?
18:23:06*linuxstb reads that a "/" should have spaces if the things it is separating have spaces - so "Scroll forward / Scroll backwards" and "previous/next" and think's that's reasonable.
18:24:02 Quit Lss__ (Read error: 54 (Connection reset by peer))
18:24:06fmltomers: yes, I mean that patch. Not everybody (including me) was happy about that
18:24:21tomerslinuxstb: So for example \ButtonMinus/\ButtonPlus, but \TouchMidLeft~/~\TouchMidRight ?
18:24:26 Join Lss [0] (
18:24:30 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
18:24:46tomersfml: It was a big mistake of me to commit it :-( I'll behappy if you can fix it
18:24:50linuxstbtomers: It depends if those macros expand to include spaces.... ;)
18:25:34fmllinuxstb: I'd always use spaces to be on the safe side
18:25:55linuxstbfml: Although maybe if you post a patch and you don't think it's ready for immediate commit, it would be a good idea to say that on the tracker?
18:25:56tomerslinuxstb: Indeed. It translates to Minus/Plus, and Middle Left / Middle Right. Is that OK?
18:26:11linuxstbfml: Yes, perhaps...
18:27:01fmllinuxstb: technically it was ready to commit. But it was a question of style and taste.
18:28:29CIA-6New commit by tomers (r24047): Mosaic: Add Cowon D2 keymaps, and update manual
18:29:04 Quit maruk ("Leaving.")
18:31:00CIA-6New commit by tomers (r24048): Manual: Remove obsolete instructions for Mine-Sweeper
18:32:58tomersCan you please approve that the 'Menu Control' section for Wormlet is obsolete?
18:33:27tomersit is obsolete since now this game has a menu
18:35:19 Quit toffe82 (Read error: 104 (Connection reset by peer))
18:39:00CIA-6New commit by tomers (r24049): Manual: Fix spacing
18:39:40CIA-6New commit by tomers (r24050): Manual: Fix missing key mapping for Cowon D2 in stats
18:41:43 Nick Ypsy is now known as YPSY (
18:44:38 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-jwjccgezxusitvrn)
18:44:57 Quit linuxstb (Remote closed the connection)
18:47:23 Quit MethoS- (Remote closed the connection)
18:48:33 Join toffe82 [0] (n=chatzill@
18:54:15 Quit fml ("CGI:IRC")
18:54:25CIA-6New commit by tomers (r24051): Manual: Add missing Cowon D2 key to Wormlet section, and change phrasing
19:04:19CIA-6New commit by tomers (r24052): Cowon D2: Map ACTION_WPS_BROWSE to BUTTON_MINUS|BUTTON_POWER because you sometimes hit it by accident when turning off hold feature. ...
19:05:12 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:11:00 Join Jonathan__ [0] (
19:11:09Jonathan__Hey, I'm back with another question or two
19:12:01Jonathan__I read that installing rockbox will destroy my ability to play DRM protected music. I do two things that I think may be DRM protected.
19:12:27kugeltomers: what is the "hold feature"?
19:12:42kugelthe second entry in the keymap file seems unneeded
19:13:01Jonathan__one, I use the overdrive media console to check out audio books on the more system.
19:13:38kugelJonathan__: target?
19:13:53Jonathan__two, I might buy a couple songs off of itunes. I would burn them to a CD then rip them back on to the player. are either of these things DRM protected?
19:14:01Jonathan__not sure what you mean
19:14:21Jonathan__ this is where I would get them from
19:14:22topikrockbox doesn't break drm on the fuze anymore
19:14:26kugelwhich player do you use?
19:14:45Jonathan__sweet, I am definetly going to get rockbox
19:14:57kugelwell, I don't think we can guarantee that it doesn't break DRM
19:15:08Jonathan__oh, but it most likely won't?
19:15:08tomerskugel: When you release to hold button (not 'feature', oops, taken wording from the guy in the forum thread), you might slide the button too much, and accidentally press hold
19:15:11topikrockbox won't play those drm files though, you need to go back to the original firmware for that
19:15:37kugelwe had that problem indeed, but IIRC it went away after fixing the internal clock. I'm not too sure since we haven't much users that use DRM'd files
19:15:38Jonathan__so the overdrive media console stuff is out on rockbox?
19:15:38topikthere are few guarantees in life. rockbox shouldn't break the original firmware's ability to play drm files any more
19:16:06topikpeople on the sansa forums reported it (drm files) works fine for them now
19:16:09kugeland 0 developers
19:16:27 Quit Zarggg ()
19:16:32Jonathan__do you know if burning Itunes music to a CD and ripping it back will destroy the DRM?
19:16:39topikJonathan__: yes it will
19:16:45Jonathan__thanks topik
19:16:55kugelit will remove it
19:17:28topikJonathan__: holding the 'left' button will boot your fuze into the original firmware even when rockbox is installed
19:17:36topikyou can still play your infected files there then
19:17:54topikholding left while you turn it on that is
19:17:55Jonathan__ok, thanks. This rockbox thing is going to be so much fun lol
19:18:24tomerskugel: The second line is needed, since you might press Minus for too long, before sliding the power switch to press Power
19:18:40tomerskugel: I've just tested it...
19:19:40kugelwhat an awkward combo anyway
19:20:17*kugel would be annoyed if he needed to do this
19:20:25tomerskugel: Indeed, but it's usable. We have only 3 buttons (-/+/Menu), and a sliding button which serves both as Power and Hold
19:21:08tomerskugel: Using the D2 in your pocket, you'll realize it is needed :-)
19:21:47kugelgoing to the filebrowser isn't something you do from the pocket, is it?
19:21:53Jonathan__I can't wait to try doom on a fuze lol... its gonna be weird.
19:22:26 Join Zarggg [0] (
19:23:21kugelJonathan__: and short, only the first level works IIRC
19:23:37Jonathan__what's wrong with the other ones?
19:23:53kugelnot enough RAM
19:24:32Jonathan__awww, do other wads work better?
19:24:36Jonathan__like free doom?
19:24:53tomerskugel: You got a point... But still you get it out of your pocket, look at the screen, realizing it is on hold (black screen), and put it out of hold. You may accidentally press Power this way, thus going to latest browsed dir. Going back to WPS requires hitting the WPS combo, or going up the path to the main menu, and press 'Now Playing' list item...
19:24:55Hillshumkugel: I was able to get to higher levels on my e200v2
19:25:04kugelI was talking about freedoom, I don't know how other wads work
19:25:04topikdoom on your fuze is a nice gimmick but not something you'll be dedicating your life to
19:26:14kugelHillshum: maybe it depends on what core features you enable (some take additional RAM after enabling)
19:27:13Jonathan__I'm just looking for something to waste time on on a car trip
19:27:22Jonathan__I can disable features to play it, that's no problem
19:27:24topiki hope you won't be the one driving
19:27:31Hillshumkugel: Although one of the higher levels doesn't quite work
19:28:50Jonathan__lol... I'm 13
19:31:40 Join pixelma_ [0] (i=quassel@rockbox/staff/pixelma)
19:31:40 Quit pixelma (Nick collision from services.)
19:31:57 Nick pixelma_ is now known as pixelma (i=quassel@rockbox/staff/pixelma)
19:32:05 Quit amiconn (Nick collision from services.)
19:32:07 Join amiconn_ [0] (i=quassel@rockbox/developer/amiconn)
19:32:27 Nick amiconn_ is now known as amiconn (i=quassel@rockbox/developer/amiconn)
19:32:43AlexP"Shutdown" or "Shut down"?
19:34:09Jonathan__alright guys, thanks for your help, I'll be back :)
19:34:12 Quit Jonathan__ ("ChatZilla 0.9.86 [Firefox 3.5.6/20091201220228]")
19:35:58 Quit Casainho ("ChatZilla 0.9.86 [Firefox 3.5.3/20091020122601]")
19:40:50 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
19:40:57stripwaxsaratoga/mt (for the logs) - I updated my patch (on top of mdctexp branch), vorbis and wma now using the new mdct. vorbis about 25% slower (~10MHz) on ipod 5g. Take a look, see what you think. Few changes to the actual new mdct (mostly cosmetic, sharing tables across scopes, etc)
19:41:19stripwax(and currently running a test_codec comparison to check)
19:42:00 Join Omlet [0] (
19:42:07 Quit Omlet^study ("( :: NoNameScript 4.22 :: )")
19:42:17 Quit Omlet (Client Quit)
19:42:24 Join Omlet [0] (
19:43:52 Join saratoga [0] (i=9803c6dd@gateway/web/freenode/x-tkqsbrcvjxdoogng)
19:44:20saratogastripwax: have you looked in much detail at how the tremor IMDCT works now?
19:45:04 Quit tomers ("ChatZilla 0.9.86 [Firefox 3.5.5/20091109125225]")
19:46:33stripwaxsaratoga - in what sense? I have a 'vague' feel for it, given that it's based on an algorithm I haven't read about much. is useful for understanding the algo a bit
19:47:39saratogastripwax: i'm trying to understand how the trig tables it uses work without much success
19:47:46saratogathey seem really odd to me
19:48:06stripwaxaren't they just cos tables, used with offsets for different values of N ?
19:48:43stripwaxsaratoga- in mdct.c, why is revtab0 created/setup ? doesn't seem to be being used at all
19:48:45notlistening FlynDice, hey formatted my e200v2, full build from svn last night and it started to work again now after a night being left on same issue with the Panic message
19:48:58saratogastripwax: i think its ffmpeg baggage
19:49:13FlynDicenotlistening: Thanks for doing the format thing and checking. I left mine on all night also and even though it is working just fine I have found some more corrupted playlist_control files. I'm going to revert the 4 bit widebus for now and see if I can do some investigating
19:49:17saratogastripwax: the tremor imdct does some neat tricks with the trig tables
19:49:36FlynDicelol was just writing to you before you posted ;-)
19:49:38saratogafor instance they use the same trig tables for both rotation and DFT
19:49:51saratogawhich i didn't think was possible without a large reduction in accuracy
19:49:57stripwaxneat-ish, I guess. I thought it was just one table of cos(iPI/n) for largest n, and then the loops step through it at different speeds
19:50:18stripwaxnot sure how many other trig tables you would even need .. ?
19:50:33stripwaxdoes ffmpeg have multiple different trig tables?
19:50:46saratogastripwax: it does
19:51:40stripwaxso ... maybe I've changed something in my patch, but there's only one fsincos function afaict.
19:52:01saratogatheres a table of cos/sin (2 * M_PI * i / n ) and another of cos/sin (2 * M_PI * (i + 0.125) / n)
19:52:24saratogathat factor of 0.125 is really annoying
19:55:23stripwaxisn't that just the same as 2*M_PI*i /n+ 2 * M_PI*0.125/n = 2*M_PI*i/n + M_PI/4n ? Which could be implemented as an offset into the same table, I think? (provided table is of size 4n..)
19:55:36stripwaxoh but I guess that's kinda wasteful
19:56:48saratogai think theres some way to fold those two together
19:57:07saratogai just can't see how Tremor does it
19:57:17stripwaxsaratoga - actually vorbis appears to have two tables too.
19:57:43saratogayeah but they're really just two halves of the same table
19:57:58saratogayou could merge them into one table sampled twice as densely
19:58:12stripwaxSuppose that's true
19:58:22saratogabut i think thats somehow the trick
19:58:30saratogathey've massively oversampled those two tables
19:58:32notlisteningFlynDice, seems that it might be the playlists issue as well here for me based on the files that are corrupt
19:58:43stripwaxwhere in the new mdct does it use that 0.125 offset?
19:58:48 Quit AlexP (Remote closed the connection)
19:58:51 Quit ehntoo_ (Remote closed the connection)
19:58:57saratogastripwax: mdct.c
19:58:59 Join AlexP [0] (n=alex@rockbox/staff/AlexP)
19:59:25stripwaxhm, I can't see it
19:59:51saratogacos(2*i*PI/4096) and cos((2*i+1)*PI/4096) together are equal to cos(2*i*PI/8192)
20:00:00saratogawhich seems way oversampled
20:00:05saratogastripwax: fixed32 ip = itofix32(i) + 0x2000;
20:00:13saratoga0x2000 in fixed is 0.125
20:00:45saratogaso its computing cos(2*(i+0.125)*PI/4096), although the code is quite hard to see that from
20:00:46stripwaxOh. But there's still only one table no? tsin0 and tcos0 adding up to 2048 entries in total. Or am I missing the other tables?
20:00:50CIA-6New commit by alex (r24053): FS #10861 - Use the 'do' form everywhere in key descriptions, and that with a full stop. ...
20:00:56FlynDicenotlistening: I though it was just playlist_control files but now I'm not sure. I'm also missing several of the database_X.tcd files which seem to be overwritten with bad files
20:01:00saratogastripwax: yes but they're not shared with fft.c
20:01:05saratogait has its own trig tables
20:01:09stripwaxah.. gotcha
20:02:10saratogathe weird thing is that for an N point transform, we do cos(2*i*PI/N)
20:02:31saratogatremor does something like cos(2*i*PI/(N*4))
20:02:53 Join ehntoo [0] (
20:02:59stripwaxyeah, it's symmetrical if there's no offset
20:03:02saratogaso actually i'm not even sure if we use more memory then Tremor since i think our tables are smaller
20:03:55 Quit LambdaCalculus37 ("Leaving")
20:04:00saratogawell right now we're bigger, but thats because we compute 2x as many points instead of only computing the parts of sin and cos that aren't symmetric like Tremor does
20:04:07saratogabut thats easy enough to fix
20:04:27saratogaha just noticed this: // PJJ optimize
20:04:30stripwaxAnyway, would be great if you could take a look at the patch I added to FasterMDCT page
20:04:33saratogasome of pauls old WMA code crept in
20:04:45saratogastripwax: i skimmed it, anything in particular?
20:05:09stripwaxsome rearrangement of code between mdct.c/h and fft-ffmpeg.c/fft.h
20:05:17stripwaxplus changes to wma and vorbis codecs to use the new mdct
20:05:24notlisteningWhat i see as i have speech enabled is that progressively more and more becomes innaccessible on my unit. Trying any playback by selecting a file then choosing play causes a panic. Speech works menu entries and then breaks when it tries to use the .talk files maybe related to
20:05:38stripwaxI don't know anything about wma; my changes were largely uncommenting existing commented-out code
20:06:26saratogastripwax: you're asking how to clean it up to commit it?
20:06:29***Saving seen data "./dancer.seen"
20:06:45stripwaxNo, if you don't want to look at it I'll happily commit it anyway :)
20:07:11saratogawell i'm getting ready to head out to europe for the holidays, so i'm a little busy
20:07:19notlisteningI will add an entry on flyspray on the exisiting patch and give what i have seen in detail
20:07:50stripwaxsaratoga - anywhere in particular in europe
20:07:50saratogabut if you want to keep working on this, a neater way to call the init functions would be nice
20:08:04notlisteningbut a thai meal is calling me now :D
20:08:10saratogaNaples, Rome
20:08:19notlisteningsaratoga, nice
20:08:21saratogaunfortunately i'll have limited internet access
20:08:27notlisteningi am off to egypt ;)
20:08:47 Quit flydutch ("/* empty */")
20:09:05notlisteningsaratoga, leave the internet behind its a holiday right?
20:09:09topiksaratoga: does that mean taking a break from trying to educate people on the sansa fuze forum?
20:09:12topikthey need you!
20:11:52 Quit notlistening ("Leaving")
20:14:00saratogaheh the helix code is little help here
20:14:16saratogathey do the 3 mul/ 3 add complex mul trick, which means they need even more tables then we do
20:14:20saratoga" cps2 = (cos+sin), sin2 = sin, cms2 = (cos-sin)"
20:17:03saratogastripwax: helix does this for their mdct rotation stage: "angle = (i + 0.25) * M_PI / nmdct"
20:17:05saratogasame as ffmpeg
20:17:12 Join shai [0] (
20:17:25stripwaxhrm, 0.125 or 0.25 ?
20:17:55CIA-6New commit by FlynDice (r24054): Sansa AMS: Revert 4 bit widebus ...
20:18:39saratogastripwax: i think thats because they use PI instead of TWOPI like us
20:19:17saratogahuh they use a completely different set of tables for the post rotation "angle = i * M_PI / 1024;"
20:19:25saratogai guess they don't have any targets with IRAM
20:19:57stripwaxI can't help feeling it should be completely symmetric.. 2PI/8 = 45 deg; everything's symmetrical around that
20:20:17stripwaxsaratoga - maybe 1024 is their max nmdct and, again, they just step through it at different rates?
20:20:37saratoga didn't turn out to be useful - I think he has three separate trig tables, for different steps of the process
20:22:31saratogayeah helix looks awful with memory usage
20:22:42saratogabut their code is extremely well commented and they explain their math
20:23:31stripwaxIt seems the reason they have two different trig tables is two different kinds of precision? Q30 and Q31
20:24:16 Join bluebrother [0] (
20:24:49saratogastripwax: that and I don't think the bigger table is an exactly oversampled version of the smaller
20:25:20saratogaright now in mdct.c we assume smaller transform sizes can just step through the bigger table in multiplies of 2, 4, 8 ...
20:25:28saratogabut thats actually not right
20:25:34saratogaso it hurts the snr and should be fixed
20:27:19saratogasomehow the version of the code used for the mdct lib lost this commit:
20:29:13saratogastripwax: basically theres a million things that need to be fixed in the mdct lib before its going to be fast, useful or accurate
20:29:30saratogai would just pick things that look wrong and work on them without worrying about the other things
20:29:36stripwaxbut wait, I thought you/mt already checked for distortion in the mdct?
20:29:44saratogayeah thats fixed
20:29:55saratogabut accuracy is still poor i think due to the above issue
20:30:22saratogawell poor is relative
20:30:29saratogaprobably something like 70dB SNR
20:30:45saratogaso i doubt anyone can hear the difference, but without that commit it was something like 150dB SNR
20:31:03stripwaxyou mean 'with' ?
20:31:09saratogayeah sorry
20:31:17saratogabefore I reverted it, we got 70dB or so
20:31:34saratogathats even more confusingly worded
20:31:52kugellol :)
20:31:53stripwaxI think that makes sense because of the PI/4n offset thing. If the offset is done away with (by starting at a different place in the lookup rather than element 0), I think all the trig tables can be shared
20:32:21saratogastripwax: i sat around trying that in matlab all day and got nowhere
20:32:25markunthe commit was pretty careless of me :)
20:32:58saratogamarkun: i did the same thing but double checked in matlab right before I commited it!
20:33:11saratogathen a couple months later checked the logs and saw you did it too
20:33:33saratogabut it looks like Tremor does it now anyway!
20:33:41saratogai wonder how accurate tremor is, i should test
20:33:46stripwaxsaratoga - But really, isn't it just cos(2PI*(i+0.125)/n) = cos( (2PI*i + PI/4 ) /n) = cos(2PI*(i+N/8)/n) ?
20:34:20saratogalet me get a pen and paper i suck at algebra
20:34:24stripwaxand sin(2PI*i/n) = cos(2PI*(i+N/4)/n) ?
20:35:25stripwaxTremor's quite cute in this regard, because it basically iterates until it falls off the trig table; then reverses direction and iterates until it falls off the other end
20:35:53stripwaxas long as you have the right number of iteration steps, it sorta doesn't matter what your initial starting point is (so long as you get it right ..!)
20:37:30stripwaxFor N=8, i=0..8, then: cos(2PI*i/N) = [1,s2,0,-s2,-1,-s2,0,s2,1] (where s2= sqrt(2))
20:37:48saratogastripwax: i was thinking that the reason tremor used such a large N (basically 8192 between the two tables) was to minimize error while stepping through the table while neglecting the N/8
20:38:42stripwaxcos(2PI*(i+0.125)/N) = cos((2PI*i+2PI/8)/N) = [s2,0,-s2,-1,-s2,0,s2,1] - i.e. shifted by one elem
20:39:34stripwaxsin = [0,s2,1,s2,0,-s2,-1,s2,0]; shifted sin = [s2,1,s2,0,-s2,-1,-s2,0,s2]
20:39:43stripwaxI think/hope, anyway
20:39:55saratogai plotted in in matlab and that was not the case IIRC
20:40:02stripwaxwish I had matlab :)
20:40:06saratogathey've very close but not the same
20:40:26stripwaxhm.. can't see why not. was N a power of 2 and a multiple of 4?
20:40:48saratoga128, 256, 512 and 1024 are the commonly used values
20:40:58stripwaxI mean when you plotted it in matlab
20:41:42 Quit bluebroth3r (Read error: 110 (Connection timed out))
20:41:43saratogai plotted all of them, and they're close but not overlapping lines
20:41:47saratogahold on, i'll do it again
20:42:21*stripwax downloads Octave..
20:42:52saratogaheres what i used:
20:42:59saratogai did a bunch of sizes
20:43:37 Join pamaury [0] (
20:43:57stripwaxhang on .. what are you comparing with what there?
20:44:23stripwaxI'm saying adding theta is equivalent to shifting the x-axis
20:44:51saratogastripwax: I plotted the argument to cos computed for each size
20:45:00saratogaso i could measure the error in trying to resuse one table for all sizes
20:45:03stripwaxwhereas I think you're saying, for different N, cos(2PI*(I+theta)/N) is different. Which is true.
20:45:14 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
20:45:56saratogastripwax: i thought you meant you could shift the table values and reuse them?
20:46:11stripwaxI'm saying cos(2PI*(I+0.125)/128) for I+=1 is identical to cos(2PI*I/1024+2PI/8/1024) for I+=4
20:46:14stripwaxor something
20:48:10 Join solexx [0] (
20:48:51saratogastripwax: so basically, that I can take an N point table, take every other element, and then shift which element is zero and get the N/2 table exactly?
20:49:28 Quit Hillshum (Read error: 54 (Connection reset by peer))
20:49:43 Join Hillshum [0] (
20:49:56 Quit Omlet (Connection reset by peer)
20:54:22 Quit kugel ("exit(0);")
20:54:51stripwaxyeah, something like that. So for N=8, the table with 0.125 offset starts one element in from the start of the table without the offset; for N=8, the 0.125 offset is two elements in from the start of the table without the offset; etc
20:54:58stripwax(again, I think/hope)
20:55:27stripwaxSo make N as big as you need, and you end up with all tables for both with+without offset versions for all n <= N
20:55:38stripwaxprovided N>=8, which it always is here
20:57:10CIA-6New commit by FlynDice (r24055): Sansa AMS: Fix Red. Write delay is not included for non HAVE_MULTIDRIVE.
20:57:35saratogastripwax: [x,y]=(min(abs(alphas512(1)-alphas4096)))
20:57:51saratogawhere alphasXYZ are the arguments to cos for that size table
20:58:08saratogai get that the smallest difference between them is 1.9175e-004
20:58:26saratogaat an offset of 1 sample
20:59:16saratoga[x,y]=(min(abs(alphas128(1)-alphas4096))) gives an offset of 5 samples, but no better accuracy
20:59:49saratogaRMS error in that case is -76.2102 dB
20:59:56saratogamaybe thats good enough?
21:01:38 Join n1s [0] (n=n1s@rockbox/developer/n1s)
21:03:33*gevaerts thinks that r24055 illustrates the need for more thinking about storage. Should HAVE_MULTIDRIVE really be used to see if the second slot on a controller is used? HAVE_MULTIDRIVE will also be true if someone uses that controller, one sd card, and one other storage device...
21:05:38 Quit solexx_ (Read error: 110 (Connection timed out))
21:06:13stripwaxsaratoga - sorry, didn't follow that at all. I'm not a matlab kinda person
21:07:55*JdGordon| volanteers gevaerts to fix that :)
21:07:55stripwaxOk so specifically for N=128: alpha[0]=2*PI*(0+1/8)/128 = (2*PI/8)/128.
21:09:36stripwaxFor N=1024: for a table without offsets, call it costab, costab[i] = 2*PI*i/1024
21:09:50stripwaxSo alpha[0] for N=128 = costab[8] for N=1024
21:09:59gevaertsJdGordon|: I know :) There are more important issues, and the only place where this particular one would ever bite is if someone mods a clip to add an ATA interface or something like that, so it's probably not very urgent
21:10:05stripwax(or costab[1] for N=256, etc)
21:10:44stripwaxWhereas you're comparing alpha for one N against alpha for another N. You need to compare alpha for one N against unshifted costab for some N.
21:13:43stripwaxActually I think my logic is all screwed up.
21:13:57stripwaxWish I had matlab :)
21:14:21stripwaxYou actually need an unshifted table of size N*8 for that to work
21:15:25stripwaxi.e. alpha[0] for N=128 = costab[1] for N=1024 (not costab[8] - no idea where that came from)
21:17:50stripwaxBut you only need 1/8th the data anyway due to being able to reflect the trig data. So for N=2048, to get only unshifted cos/sin you need only 256 points (either cos or sin) ; and to get alpha func you need N'=2048*8=16384 BUT you still only need to store 1/8th data for the same reason => hence total 2048 points (either cos or sin) should get you everything you need
21:18:05stripwaxand I think *that* is exactly what tremor is doing :-p
21:26:59 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
21:29:57 Quit domonoky ("Leaving.")
21:31:29 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
21:40:11Llorean JdGordon|: in regard to that flyspray task you just closed, mike's comment on the thread also simply says >>| doesn't work (it doesn't say that seeking *does* work) so it's probably just party mode enabled, whereas this other person seems to get seeking when he tries to skip, so probably just has some hardware failure goin' on.
21:40:22 Join froggyman [0] (
21:41:53 Join faemir [0] (n=faemir@
21:42:02gevaertsany opinions on FS #8936?
21:42:15JdGordon|its a rather rounded number :)
21:42:24gevaertsyes, but apart from that :)
21:43:10JdGordon|fix bugs.. not requests :)
21:43:33 Part watto
21:43:49LloreanClose with "Works for me" or "Out of date"?
21:44:52saratogastripwax: sorry had to go out, basically what that code does it take the difference of one argument to sign in the 128 entry table against all the arguments to sign in the 4096 table, and find the closest match
21:45:14saratogaand it finds that with the predicted offset you get almost but not exactly the correct value
21:45:23gevaertsI think so ("Out of date" anyway. "Works for me" isn't really the right response for rare hard to trigger bugs)
21:45:24saratogaand I estimated that you'd get about 14 bits accuracy if you used that method
21:45:34saratogaalthough its entirely possible I'm making some mistake
21:48:10 Join qwertz [0] (
21:48:41gevaertshm, FS #9093 looks like a simple patch for a stupid problem. Any database experts around to comment?
21:49:41qwertzHi all ... got a problem with "could notopen Ipod" "konnte ipod nicht öffnen"
21:49:45linuxstbgevaerts: I would close 8936 as out of date - the reporter was asked to provide more info if it happened again, and he never did...
21:49:55 Join Tomis2 [0] (n=Tomis@
21:50:00linuxstbqwertz: With Rockbox Utility?
21:50:27qwertzlinuxstb: hi, yes
21:50:41linuxstbAre you running it as root/Administrator?
21:50:50qwertzi do start with gksudo on karmic
21:51:27qwertzfunny is that I can do other actions like installing themes
21:51:28JdGordon|gevaerts: haha... sorry. I looked at the wrong task :) apparently im semi dyslexic
21:51:56JdGordon|8963 != 8936
21:52:02qwertzlinuxstb: But I will nit find the results when starting the Ipod
21:52:24gevaertsJdGordon|: very true!
21:52:58JdGordon|amiconn: do you know why audio_pause() blocks?
21:53:23qwertzlinuxstb: I do have a quite old version (manually installled) of rb on ipod 5gn 60GB
21:53:57gevaertsbluebrother, domonoky: is FS #9177 still valid? That's quite old, and rbutil has moved a lot since then
21:55:16qwertzI have two partitions on the ipod and I am not very sure which is the real system partition for rb. Maybe I did some craps whe installtin a year ago
21:55:20linuxstbqwertz: Are you using the "autodetect" feature?
21:55:31domonokygevaerts: i am not sure, as it is a sansapatcher issue, it could still be valid. But noone has a e200r to test.
21:55:42bluebrothergevaerts: I guess it is still valid
21:55:44linuxstbqwertz: If you are specifying the device for your ipod, you specify the "whole-disk" device - e.g. /dev/sdb
21:56:18linuxstbdomonoky: sansapatcher doesn't work on the e200r
21:56:34qwertzlinuxstb: yes, atodetect finds the the small partition
21:56:39bluebrotherqwertz: can you post the output of the About / System Trace dialog somewhere?
21:56:48linuxstbdomonoky: Ah, I've read the task now...
21:56:49domonokylinuxstb: does it also reject the e200r and give a error ?
21:56:59linuxstbdomonoky: No idea, I don't have an e200r ;)
21:57:12linuxstb(and it was originally written before they existed)
21:57:14domonokylinuxstb: thats the problem :-)
21:57:36bluebrotherRockbox Utility wants the mount point, not the device node. It resolves the mountpoint to the device node internally
21:57:54qwertzlinuxstb: the auto found path is /media/rockbox_root
21:59:07qwertzhi bluesbrother; Yes, if you tell me how to get the => About / System Trace dialog
21:59:10 Quit Tomis (Connection reset by peer)
21:59:10 Nick Tomis2 is now known as Tomis (n=Tomis@
21:59:41JdGordon|doesnt sansapatcher work with the e200r once you manually do the first step to unlock it?
21:59:59linuxstbJdGordon|: No.
22:00:01amiconnJdGordon|: Afaics it doesn't block
22:00:07domonokyqwertz: look in the in the about menu at the top of rbutil :-)
22:00:19linuxstbJdGordon|: The firmware partition is hidden via USB (IIRC)
22:00:29amiconnIt just sends an event (using queue_send() on swcodec, queue_post() on hwcodec)
22:00:49JdGordon|queue_send() blocks
22:00:50linuxstbdomonoky: Looking at sansapatcher.c, the is_sansa() function checks the partition layout, and then the contents of the firmware partition, so I can't see how it can detect an e200r. Maybe rbutil isn't using that function?
22:01:08JdGordon|I'm wondering if that can be changed to queue_post which doesnt block
22:01:25amiconnI have no idea
22:01:52JdGordon|:( I dont think any active devs do
22:02:29bluebrotherlinuxstb: is_sansa() is called in bootloaderinstallsansa.cpp, so if the partition layout is different on the e200r it should fail
22:03:24JdGordon|jhMikeS was the last to fiddle with that, havnt seen him in ages
22:03:34linuxstbbluebrother: But then what happens? The complaint seems to be that there is no error message.
22:04:11 Join arohtar [0] (n=faemir@
22:04:21amiconnr13361: SWCODEC Playback: Seriously mess with it and get rid of (at least some ;) race conditions.
22:04:25amiconnBy jhMikeS
22:04:44qwertzdomonkey, linuxsstb, bluebrother:
22:04:52bluebrotherwell, according from the code it should return an error
22:05:24bluebrotherdo we know the partition layout of the e200r? Faking it might help to test the issue
22:05:50 Quit Tomis (Read error: 104 (Connection reset by peer))
22:06:33***Saving seen data "./dancer.seen"
22:07:36linuxstbbluebrother: It just has a single FAT32 partition.
22:08:02 Quit dfkt_ ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
22:08:25JdGordon|oh thats right, the partition it exports to usb isnt the full disk
22:08:36bluebrotherhmm, then I don't see why rbutil should install the bootloader on the e200r
22:09:24qwertzdomonkey, linuxsstb, bluebrother: The Dropbox link is the "About / System Trace dialog"
22:09:28linuxstbbluebrother: The bug report says that rbutil _says_ it does... But it's from July 2008 - was the code different back then?
22:09:30JdGordon|anyone have an email addy for jhMikeS?
22:11:19 Join kugel [0] (n=kugel@rockbox/developer/kugel)
22:12:23 Join Tomis [0] (n=Tomis@
22:12:30bluebrotherlinuxstb: that's quite possible. The bootloader installation was reworked in October 2008
22:12:56linuxstbbluebrother: I guess we can close it as out of date then...
22:13:09 Quit Lear ("ChatZilla 0.9.86 [Firefox 3.6b4/20091124213835]")
22:13:10 Quit faemir (Read error: 110 (Connection timed out))
22:13:17bluebrotherfair enough.
22:13:51linuxstbIf sansapatcher itself didn't reject an e200r, I'm sure we would have heard by now...
22:13:53qwertzlinuxstb bluebrother: Are you talking about my prob? Should I follow? Thanks!
22:14:03linuxstbqwertz: Sorry, no we're not ;)
22:14:47bluebrotherqwertz: no, but from looking at the trace resolving the mountpoint didn't seem to work :/
22:15:28qwertzbuebrother: hmmmm :'(
22:15:45bluebrotherbtw, you're running a localized system?
22:16:10qwertzbluebrotehr: German
22:16:39linuxstbqwertz: If you unplug your ipod, do you still have a /media/rockbox_root directory?
22:16:54bluebrotherah. I've spotted another issue I fixed a few days ago which hits on localized machines. But thats a minor issue
22:17:15 Join petur [0] (
22:17:41stripwaxsaratoga - here's the thing. alpha for one N and alpha for another N will not match. But they will both be present in a table of unshifted cos/sin for N*8.
22:17:59qwertzbluebrother;;: Nope, will be unmounted and dissapear from Nautilus
22:18:16stripwax(just reading your message from 20:44 - I also disappeared for a bit)
22:18:33linuxstbbluebrother: What do the "[BootloaderInstallIpod] ipodpatcher: overriding scan, using " lines mean, and why are they truncated?
22:18:41linuxstb(are they truncated?)
22:18:43bluebrotherqwertz: if you plug the ipod and then type "mount" in a terminal window, does the mountpoint appear in the output?
22:19:26bluebrotherlinuxstb: that line isn't trunctated, that's the problem. It should print the device node at the end, but resolving the mountpoint didn't work, so that message contains an empty string.
22:19:38bluebrotherthere's definitely room for improvement here ...
22:19:52bluebrotheri.e. handle that case with a proper error message
22:20:41qwertzbluebrother: No, does not appear in mount list
22:20:59 Nick The3_14ed|r is now known as Dhraakellian (
22:21:27bluebrotherwell, that's interesting. How does Nautilus access the Ipod if it isn't mounted? If it is mounted it has to appear in the mount output
22:21:30linuxstbqwertz: But you can see your ipod in /media/rockbox_root/ ?
22:22:10qwertzbluebrother, linuxstb: sorry it does * lalala* /dev/sdb3 on /media/rockbox_root type ext3 (rw,nosuid,nodev,uhelper=devkit)
22:22:25gevaertstype ext3?
22:22:31linuxstbtype ext3?
22:22:42gevaertsalso, /dev/sdb3?
22:22:48bluebrotherqwertz: do you have ipod linux installed?
22:22:51gevaertsshould be fat and /dev/sdb2
22:23:05linuxstbqwertz: Is /dev/sdb2 mounted anywhere?
22:24:45saratogastripwax: yeah thats what I thought, but i searched for an entry in the 128 table and it wasn't in the 4096 table
22:24:49qwertzahhh, got it: /dev/sdb2 on /media/476A-AB95 type vfat (rw,nosuid,nodev,uhelper=devkit,uid=1000,gid=1000,shortname=mixed,dmask=0077,utf8=1,flush)
22:25:48linuxstbqwertz: But we're assuming you have/had ipodlinux installed?
22:25:53qwertzI did an experiment wit ipodlinux abot 1 or two years ago
22:25:58bluebrotherok, then you need to use /media/476A-AB95 as mountpoint
22:26:43bluebrotherdid Rockbox Utility autodetect the mountpoint?
22:28:34qwertzoki doki, should I do complete install to erase all old stuff?
22:28:37stripwaxsaratoga - the alpha for 128 won't be in the alpha for 4096. But the alpha for 128 will be in the unshifted cos/sin table for 4096. And the alpha for 4096 will be in the unshifted cos/sin table for 32768
22:28:52qwertzbluebrother: Auto took the rockbox partition
22:29:07linuxstbbluebrother: How does rbutil find the mountpoint? The first one it finds for that device in /proc/mounts ?
22:29:17stripwaxSo here we have alpha for N=2048, which will be in the unshifted cos/sin table for 16384. But we only need to store 1/8th of it
22:29:27bluebrotherqwertz: "rockbox partition" means /media/rockbox_root?
22:29:30gevaertsJdGordon|: re FS #9640: I think we reached a consensus, but I can't remember what it was
22:29:36stripwaxunless I've completely screwed up
22:29:47JdGordon|gevaerts: which was 9640?
22:29:56qwertzbluebrother: Yes
22:29:58gevaertsJdGordon|: the wps line ending vs comments thing
22:30:17JdGordon|ah yes. I distincly remember the conversation happened...
22:30:29qwertzBluebrother: A complete installation should not affect my music, I guess?
22:30:55qwertzbluebrother: If you say roger I will give it a try ...
22:30:59bluebrotherqwertz: does /media/rockbox_root/.rockbox exist, especially rockbox-info.txt in that folder?
22:31:13bluebrotherno, installation does not delete any data
22:31:25stripwaxyou should be able to check that in matlab; build cos(alpha), sin(alpha) for alpha in N=128, and build cos(2*PI*I/N),sin(2*Pi*I/N) for N=1024
22:32:03JdGordon|rasher: pixelma: do either of you remember the outcome for the comment bug?
22:32:07saratogastripwax: i don't understand
22:32:15saratoga"the alpha for 128 won't be in the alpha for 4096. But the alpha for 128 will be in the unshifted cos/sin table for 4096."
22:32:50qwertzbluebrother Yes: /media/rockbox_root/.rockbox exist, especially rockbox-info.txt does exist
22:33:00stripwaxsaratoga - when you said you searched for an entry in the table for 128, and it wasn't in the table for 4096. By "table" do you mean your tcos2=cos(alpha) tables , in both cases?
22:33:20saratogastripwax: i computed all the alphas for 4096
22:33:29saratogathen i looked for the first alpha in the 128 table
22:33:35stripwaxright so try computing regular cos(2*PI*i/N) for 4096 instead..
22:33:44bluebrotherqwertz: and no .rockbox on the vfat partition? That would explain the wrong detection
22:34:16saratogastripwax: if alpha doesn't match, then cos(alpha) won't match . . .
22:34:18qwertzbluebrother: So rockbox_root is maybe a legacy of experiments with ipodlinux. should I rease it
22:34:30 Join p3tur [50] (n=petur@rockbox/developer/petur)
22:34:49stripwaxright, that's why you shouldn't compare alpha with alpha, or cos(alpha) with cos(alpha) - you must compare cos(alpha) to cos(2*PI*i/N)
22:35:05stripwaxi.e. compare the thing with the 0.125 to the thing without the 0.125 but N a factor of 8 larger
22:35:30stripwaxBecause: for N=128 and i=0, alpha = 2*PI*(0 + 1/8) / 128 = 2*PI/(128*8)
22:35:46saratogastripwax: ohhhh
22:35:57stripwaxWhich is the same as 2*PI*(1)/1024
22:36:01linuxstbqwertz: Yes, delete .rockbox from the rockbox_root partition
22:36:02saratogaif its really that simple i will feel really stupid
22:36:02domonokybluebrother: its the silly autodetection code, it stops at the first rockbox-info.txt/rbutil.log it finds.. (which is probably the worng one)
22:36:18 Quit arohtar (Client Quit)
22:36:19stripwaxWhich will be in the regular (cos(2*PI*i/N) *not* cos(alpha)) table for N=1024
22:36:25bluebrotherdomonoky: yes, definitely :(
22:36:33 Join faemir [0] (n=faemir@
22:36:47qwertzbluebrother: .rockbox and rockbox-info.txt exist on vfat partition, too. As well as a lot of database_x.tcd
22:36:56stripwaxeither that or I'm going crazy. I have been off sick actually so who knows :)
22:37:04bluebrotherfortunately I'll be on a couple of longish train trips the next weeks ...
22:37:09saratogastripwax: you're right it works
22:37:14stripwaxwoohoo! :-)
22:37:29saratogaso we can combine the mdct tables with the fft tables and be more accurate then we are now
22:37:33bluebrotherreworking the autodetection is on my todo list since quite a while
22:37:42stripwaxeveryone's a winner :)
22:37:42domonokybluebrother: my autodetection patch in flyspray would find the correct one (or at least present two options) .. :-) but unfortunatly that still needs more work.. and i didnt found the time for it.
22:38:09stripwaxRewiring the fft to use smaller tables and some kind of modulo addressing might be a bit messy though. but I'm sure we can nick ideas from tremor for that also...
22:38:46bluebrotherdomonoky: yeah, time is this problem that's always around :o
22:39:05qwertzOK - will erase the wrong partition ;o) Where is th original Ipod OS found. Do not want to delet it. But I guess it wont be on ext3 *lol*
22:39:27bluebrotherqwertz: the ipod firmware is on the first partition, which is of type 0
22:39:41domonokybluebrother: so if you use your train time to improve the patch/ write your own better autodetection, it would be great ! :-)
22:40:23qwertzbluebrother: can I see it or is it some kind of hidden flash memory, or so
22:40:53saratogastripwax: error is less then -150dB with your scheme
22:40:54bluebrotherqwertz: it's a regular partition. It doesn't have a filesystem, so you can't mount it.
22:40:56linuxstbqwertz: There's no filesystem there, but you can read/write it by accessing /dev/sdb1 directly... (probably not recommended though).
22:41:04saratogawe're probably limited by the x86 cos() accunracy
22:41:37 Join bertrik [0] (
22:41:44gevaertsOpinions on FS #9395 ?
22:42:57qwertzbluebrother: OK, I gonna erase the wholw "wrong partition" and rename the right wone to make a proper unit. Then try install. Thanks. CU Later
22:43:57saratogastripwax: changing it in mdct.c should be easy, since we can just change revtab around
22:44:04saratogai don't know about fft-ffmpeg.c
22:44:31JdGordon|gevaerts: I'd close that with a message "reopen if its still happening".. but im sure someone will disagree with that :/
22:44:42n1sgevaerts: i'd say close with "fixed"
22:44:54AlexPI'd close it
22:45:25gevaertsThree people who would close it. Now who will get to it first? :)
22:45:40n1swow, bug count has dropped nicely, good work people :)
22:46:14AlexPfrom what to what?
22:46:49n1sfrom300 a few days ago to 280 now
22:47:08AlexPI was expecting a bigger drop from what you said :)
22:47:17n1swell, *open* bug count
22:47:22gevaertsAlexP: we're still working on it!
22:47:36AlexPI'm not criticising!
22:47:46n1sAlexP: the trend has been a slight increase so any drop of more than 1 is nice :)
22:47:56AlexPvery true :)
22:49:20JdGordon|if we are more agressive with closing very old and abandond bugs we should easily get under 200 again
22:49:44JdGordon|bugs which we think are fixed and havnt had any comments for months
22:50:04linuxstbgevaerts: Do you understand FS #10808 ?
22:50:08gevaertsone problem I'm having is that some bugs are a bit vague, list several related things, and might be mostly fixed
22:50:38JdGordon|I'd say depending on what the unfixed things are either create new ones for them, or ignore them
22:51:10gevaertslinuxstb: I'd need to check, but this seems to indicate that windows 2000 doesn't handle composite devices properly
22:51:42gevaertsin other words, reassign to microsoft :)
22:52:03n1sgevaerts: windows 2000 has been out-of-support from MS for quite a while, hasn't it?
22:52:29gevaertsn1s: probably, but we should still be able to confirm that this is the issue
22:53:35n1smaybe some way to disable the hid functionality and do just msc would be good for such cases
22:53:50gevaertsthat's there in svn
22:54:08CIA-6New commit by bluebrother (r24056): Remove scanning for Ipod from Ipod bootloader installation and clean it up a bit. The Ipod is specified by its mountpoint, so there is no need to scan ...
22:55:07*AlexP closes one!
22:55:45n1sgevaerts: as a setting? or some hidden magic button combo thing?
22:55:55gevaertsn1s: a setting
22:56:03n1sgevaerts: cool
22:57:08gevaertshm, when did we enable USB in the svn builds on ipods?
22:57:17JdGordon|AlexP: we should do a prize for the person that closes the most :)
22:57:24linuxstbgevaerts: About a year ago?
22:57:24JdGordon|that should add some incentive
22:57:32AlexPJdGordon|: :)
22:57:36JdGordon|for 3.1 wasnt it?
22:58:40linuxstbgevaerts: You enabled it in Feb 2009 - r20105
22:59:34gevaertsThat means it was after FS #9793, which means I'll close that one
23:00:13 Quit bmbl ("Bye!")
23:02:05JdGordon|while people are awake... look at the dev mail thread about the postponing.... what starte are we trying to get to? and where are we in relation to that?
23:03:48gevaertsuhm, FS #9810?
23:03:56gevaertsDon't we just measure the voltage?
23:05:53JdGordon|the percent comes from the voltage based on target iirc
23:05:56*bluebrother wonders how r23943 should work −− #define QUIT 0 but later if(pressed & QUIT) ?
23:06:09gevaertsso if this is a bug, it's a hardware bug
23:06:11n1sgevaerts: yeah, and we convert voltage to percent
23:06:18*gevaerts closes
23:06:39n1sor a adc driver bug, conversion bug but that seems unlikely
23:07:00bertrikthe last 10 percent is on a quite steep part of the voltage vs time curve
23:07:00n1syeah, close
23:07:16n1saccuracy tends to be bad at low voltages
23:11:52gevaertsHow about FS #9868 ?
23:12:50 Quit faemir (Read error: 110 (Connection timed out))
23:13:28AlexPyeah, 2 down!
23:13:32 Join DerPapst1 [0] (
23:13:34*AlexP is getting into this
23:14:45AlexPgevaerts: I'd close that - it is likely an accent issue, and your own voice files can be generated
23:14:55gevaertsAny chance of FS #9952 still being present today?
23:15:01AlexPIt isn't wrong per se
23:15:05bertrikI do think it's a bit odd that the percentage lingers at 9% so long, but nearly impossible to investigate further because the exact DAP model is not mentioned
23:15:34bertrikoh, sansa e200
23:17:04JdGordon|it could very well be the calibration on that target is screwed up
23:17:07JdGordon|or was at the time
23:17:21*JdGordon| doesnt like these nano2g bugs in the tracker yet
23:17:32JdGordon|unless they get closed they will stay forever with this same problem
23:17:54*gevaerts agrees. Maybe we should downgrade nano2g to "unusable" again...
23:18:39JdGordon|or we be more agressive closing bugs as they become out dated :p
23:19:16gevaertsJdGordon|: can you quickly look at FS #9868?
23:19:29gevaertsJdGordon|: oops no, FS #9952
23:19:30HillshumAre they all of one particular thing?
23:20:39JdGordon|gevaerts: i had to do a double take on the date there.... that cant still be open!
23:20:53JdGordon|it didnt even click that 20102 was months ago :p
23:21:09bertrikit indeed looks like the 9% percentage is caused by a different battery curve, the calculation of battery percentage is correct
23:22:24bertrikat around 3.68V, our calibration curve expects the battery voltage to drop a lot faster than was the case with the battery from the FS bug
23:23:08 Nick YPSY is now known as Ypsy (
23:23:11linuxstbJdGordon|: What do you mean by "unless they get closed they will stay forever with this same problem" ?
23:24:06JdGordon|I mean the actual bug could be fixed, but because the reporter disappeared, or forgot the bug just stays there, and then months later we wonder if its indeed fixed or not
23:24:21JdGordon|9420? it seems like hardware problem?
23:25:14JdGordon|unless we actually reset the radio frequency every couple of minutes?
23:25:48linuxstbJdGordon|: For example?
23:26:35JdGordon|this battery curve bug
23:26:53gevaertsJdGordon|: one comment says that the OF doesn't do this
23:27:21JdGordon|the last comment says it does showup in the OF
23:27:37gevaertsyes. A bit confusing...
23:28:37gevaertsJdGordon|: see also FS #10405
23:28:53JdGordon|yeah, saw it
23:28:56*linuxstb wonders what JdGordon| is talking about - he is talking about the Nano2G bugs...
23:29:38gevaertslinuxstb: people going through bug lists tend to have confused minds :)
23:31:01AlexPWhat about things like 5072 - isn't that just not implemented rather than a bug?
23:31:15JdGordon|linuxstb: i dont have a specifc bug in mind.. and maybe im getting the wrong impression..
23:31:45gevaertsAlexP: even JdGordon| said that it's a bug!
23:31:54AlexPMmmm, but is it?
23:32:01rasherJdGordon|: I think there was a camp of "Well you have strange tags" and then little old me saying "Please try harder, the info is all there"
23:32:09linuxstbJdGordon|: The main problem (IIUC) on the Nano2G seems to be the NAND driver, but only on some devices. On most, it seems to work fine.
23:32:15AlexPgevaerts: I mean, it is functionality we'd like, but it isn't intended to work currently
23:32:27*linuxstb wishes you could search by player type on flyspray...
23:32:42gevaertsAlexP: does the manual state this properly?
23:32:50AlexPI don't know
23:32:57AlexPbut I doubt it :)
23:33:06JdGordon|rasher: :(
23:33:14JdGordon|the usual lack of agreement then?
23:33:26bertriklinuxstb, yeah I miss that too
23:33:47JdGordon|5072 shuold probably be closed, or changed to a feature request
23:34:12 Quit n1s ("Lmnar")
23:34:17bertrikand then closed because we don't do feature requests :| ?
23:34:29gevaertsor rephrased into a manual bug
23:34:39AlexPI'm just going to put a note in the manual, then close it
23:35:11*JdGordon| wishes someone would work on the database so it can be more like the file browser from the codes POV
23:36:39JdGordon|then bookmark, playlist resume, follow playlist would all just magically work
23:37:50pamauryBut isn't a database completely different from a file hierarchy ?
23:38:32JdGordon|yes and no
23:38:38gevaertsyes. I'm not convinced that this would work well
23:38:47AlexPSo to make sure - bookmarking doesn't work at all when the file is launched through the database?
23:39:10JdGordon|. /all/genres/metal/bandname would be the "file location"
23:39:25JdGordon|which would always generate the same database query
23:39:33JdGordon|AlexP: iirc, yes
23:39:59JdGordon|just like /music/albuum/ always gets the same filesystem loookup
23:40:09gevaertsJdGordon|: a major point of the database is that you're not limited to that...
23:40:27AlexPSo does "Bookmarking only works when files are launched from the file browser, and does not currently work for files launched via the database." suit people?
23:40:28JdGordon|of course... you use the database browser to get that path
23:40:30 Quit domonoky (Read error: 104 (Connection reset by peer))
23:40:52linuxstbJdGordon|: What about playlists generated on playcount?
23:41:01gevaertsJdGordon|: you can change the path halfway and then add more tracks to the playlist
23:41:09JdGordon|the query would be the same.. the value would be different
23:41:33JdGordon|well, yeah, its not a perfect solution
23:42:00gevaertsHow about FS #10103? Just close?
23:42:39linuxstbBut I agree the database needs work. IIUC,when you create a playlist, every filename is added individually to the .playlist_control file, which is what Rockbox's playlist code was designed to avoid.
23:43:38linuxstbSomeone should have paid attention at the time and shouted at Slasheri...
23:43:45stripwaxsaratoga - sorry, back now
23:43:55stripwaxso, mdct.c doesn't use revtab..
23:44:15stripwaxoh, fft.revtab
23:44:35JdGordon|even a resumed file browser playlist stuffs up if you modify the file contents
23:45:12JdGordon|which is why I wouldnt have a problem with a playlist generated from the DB having a "path" like the filebrowser has
23:45:46gevaertsor a set of such paths
23:45:55stripwaxoh - excellent - so only the unrolling at the end of ff_imdct_calc makes the assumption that input & output buffers are physically distinct. (the thing that prevents ffmpeg imdct operating in place). that should be easy to fix
23:46:09gevaertsbut yes, that would solve some issues
23:46:15Hillshumpossibly using a UUID or something for the path
23:47:46stripwaxsaratoga - I'm going to merge the r17783 you spotted into my patch; and commit to branch svn. shout if you have any questions/reservations ...
23:48:15Hillshumgevaerts: Might make figuring out what the paths should be easier? I don't know much about this
23:48:31JdGordon|sounds like over-engineering...
23:48:41JdGordon|speaking of which.. has anyone thought about using sqllite for the db :)
23:48:51CIA-6New commit by alex (r24057): Add note to Bookmarking sections that bookmarking only works from the file browser and not the database.
23:48:52gevaertsHillshum: I'd expect such a path to be "genre=xxx/artist=yyy"
23:48:53gevaertsJdGordon|: yes
23:49:12 Join captain_kewl [0] (i=2669ecc2@gateway/web/freenode/x-zkegqyabsvmptqza)
23:49:14JdGordon|did it get anywhere?
23:49:46 Join GeekShadow [0] (n=Antoine@reactos/tester/GeekShadow)
23:49:46gevaertsit's not feasible. You need half a megabyte of RAM basically, where the current code uses 100kb or less
23:50:07gevaertsbut there's a patch that allows you to use it frokm plugins :)
23:50:25gevaertsalso, this is version 2 IIRC. 3 is even bigger
23:53:33AlexPDown to 260!
23:53:46*JdGordon| prods AlexP for that one
23:55:06qwertzbluebrother linuxstb: *WOW* *EEK* everything works fine now. Rockbox 3.4 looks much mor better than the stone age version i ran before!!
23:55:50Hillshumqwertz: What version was that
23:56:04JdGordon|the dos 1 I bet :D
23:56:15bertrikprobably some version before the cabbiev2 default theme
23:56:25qwertzbluebrother linuxstb: BIG BIG thank you and thank you to the coders .. I think I should invest a few hours to contribtute
23:58:21qwertzHillsum: I think it was 1.5-2 years old. C`ant tell version number

Previous day | Next day