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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2011-02-11

00:02:48 Quit guymann (Ping timeout: 265 seconds)
00:05:48 Join JdGordon1 [0] (
00:09:32 Quit pamaury (Remote host closed the connection)
00:17:04 Join sideral [0] (~sideral@
00:17:04 Quit sideral (Changing host)
00:17:04 Join sideral [0] (~sideral@rockbox/developer/sideral)
00:18:11 Nick evilnick_B is now known as sillynick_B (0c140464@rockbox/staff/evilnick)
00:19:41 Quit robin0800 (Ping timeout: 240 seconds)
00:34:00 Quit sillynick_B (Quit: Page closed)
00:37:18*amiconn wonders about the weird binsize effect of r29275
00:42:49JdGordon1S_a_i_n_t: ok, if we had a proper "current screen" system... what would you expect the current screen to be when you enter the (example) WPS context menu?
00:47:06 Nick milk_ is now known as milk (
00:54:43 Join guymann [0] (
00:57:29 Join qurvel [0] (
00:59:13jhMikeSamiconn: I think because the targets that went up didn't have the static struct before
01:06:43 Join DerPapst [0] (
01:07:33JdGordon1S_a_i_n_t: OY!
01:07:53 Join markun_ [0] (~markun@rockbox/developer/markun)
01:11:07 Quit markun (Ping timeout: 272 seconds)
01:17:03JdGordon1RaaaAA really should be have an AA scraping activity for when it isnt found locally
01:17:45***Saving seen data "./dancer.seen"
01:18:30 Join krazykit [0] (
01:20:09CIA-70New commit by 03sideral (r29280): autoresume: Match full directory path names only in autoresumable() ...
01:25:22CIA-70r29280 build result: All green
01:26:42 Quit DerPapst (Quit: Leaving.)
01:30:27jhMikeSis thumb code being used on some targets?
01:35:14 Quit mudd1 (Read error: Operation timed out)
01:39:57 Quit JdGordon| (Ping timeout: 265 seconds)
01:42:53jhMikeSaha, didn't know it was actually being used, explains weird deltas (r29275 notwithstanding) compared to other targets
01:43:59 Quit dfkt (Read error: Connection reset by peer)
01:44:30 Join Xerion_ [0] (
01:45:45 Quit Xerion (Ping timeout: 265 seconds)
01:45:46 Nick Xerion_ is now known as Xerion (
01:56:12kugelgevaerts: there's a problem :(
01:56:33 Quit T44 (Read error: Connection reset by peer)
01:56:35 Join Xerion_ [0] (
01:57:24 Join Topy44 [0] (
01:57:46kugelgevaerts: somwwhy your commit breaks eclipse
01:57:59kugelI'm currently fighting some random problems
01:58:09gevaertsEclipse uses make now?
01:58:29 Quit Xerion (Ping timeout: 240 seconds)
01:58:29 Nick Xerion_ is now known as Xerion (
01:58:33 Quit liar (Quit: Leaving)
01:58:49kugelno but the mapke apk classes seems to be a problem for it
01:59:27 Quit sideral (Ping timeout: 246 seconds)
01:59:31gevaertsin what way?
02:00:23kugel“Conversion to Dalvik format failed with error 1”
02:00:32kugeland my code doesn't work in an unexpected way
02:01:05kugelI listen to music, but the java layer isn't aware that rockbox is running
02:01:52jhMikeSbuffering_handle_finished_callback is dangerous because it sends an event on the buffering thread which can result in calls that could conceivably demand buffering (deadlock if data isn't ready)
02:02:42*gevaerts is confused
02:02:52gevaertsIs that eclipse breaking?
02:04:01kugelah the latter was unsurprisingly a problem in my code
02:04:59kugelah uh, I managed to recover the lcd speed in gingerbread by using the new native apis
02:05:29kugelgevaerts: eclipse reports the first error as soon as I hit make apk
02:05:56gevaertsWell, I don't have eclipse here, so I can't try
02:06:27kugelhm, the lcd speed is back to normal, but the code is incompatible with <= froyo :(
02:09:00*jhMikeS isn't sure long as the handle really stays finished as isn't asked to go anywhere
02:09:28JdGordon1jhMikeS: don't you want to just replace the whole system with buflib? :D
02:09:57jhMikeSwhat's buflib? :)
02:10:22*JdGordon1 might be confusing people and parts
02:10:42jhMikeS?? lulz ??
02:10:53JdGordon1sure, why not
02:11:58*jhMikeS gets confused for his parts (but buflib makes it easy)
02:14:28 Quit Xerion (Read error: Connection reset by peer)
02:15:17 Join Xerion [0] (
02:15:21jhMikeSat least it looks compatible with concurrent access
02:22:23*jhMikeS wonders if any codecs yield while still using the same direct pointer
02:23:18JdGordon1anyone interested in testing FS #11925 for me? i have no toolchains vailble now so dont even know if it builds :/
02:26:04jhMikeSin what way?
02:31:09kugelhm, the new code has some advantages
02:31:46jhMikeSSCREEN_UNKNOWN undeclared :\\
02:31:57JdGordon1prob a typo
02:32:53jhMikeSmisc.c "SCREEN_UNKNOWN" (copied)
02:33:54jhMikeSsupposed to be in a header somewhere?
02:34:05JdGordon1misc.h at the bottom
02:34:37JdGordon1SCREEN_* is a really bad name though, need to change that
02:35:03jhMikeSmisc.h "SCREEN_UNKNONW" :)
02:35:57JdGordon1s/SCREEN_/ACTIVITY_/ maybe?
02:39:30jhMikeSthat was a question, not a suggestion
02:39:48JdGordon1what would ? do?
02:40:02jhMikeSMicrosoft (instead of Jesus)
02:40:46jhMikeSwell, got it to, what should I see (if anything)?
02:41:12JdGordon1just really if it crashes when you go in and out of the browsers and quickscreen
02:45:11jhMikeSseems ok so far
02:46:54 Quit mystica555_ (Read error: Connection reset by peer)
02:48:38jhMikeSI've been in and out of the various screens dozens of times by now
02:49:30JdGordon1ok thanks
02:49:56JdGordon1if you're keen you can test the actual change with a sbs?
02:50:47jhMikeSyou have suitable one?
02:51:31JdGordon1what target you on?
02:54:22JdGordon1i tihnk that will work
02:55:25 Join mystica555_ [0] (
02:57:37jhMikeSwhawt do you mean "<- should change"?
02:58:01JdGordon1save that to a .sbs and load it :)
02:58:03 Join Dreamxtreme [0] (~Dre@
02:58:21JdGordon1it should show a number and "<- should change" on the top line
02:58:25JdGordon1the number should change
02:58:54JdGordon1only for the browsers and quickscreen though
03:03:21jhMikeSeither i'm doing something wrong or I see no change
03:04:22jhMikeSI should just put that in an .sbs file and load that?
03:05:31jhMikeSI don't see anything different
03:06:07JdGordon1ok, thanks anyway
03:07:37JdGordon1ah there is a typo
03:07:52JdGordon1the 3rd line has a . instead of a ,
03:08:01JdGordon1after the 8
03:09:46jhMikeSmain menu: 0, browser: 2, quick screen: 8
03:10:36JdGordon1gigeddy gigeddy
03:11:11jhMikeSmajor lag in the update though (1s or so)
03:11:58JdGordon1yeah, that's no surprise
03:12:24 Quit kugel (Remote host closed the connection)
03:15:29 Quit Kitar|st (Read error: Connection reset by peer)
03:17:49***Saving seen data "./dancer.seen"
03:27:16 Quit qurvel (Ping timeout: 240 seconds)
03:36:09 Quit Raptors (Ping timeout: 255 seconds)
03:48:53 Join Raptors [0] (
04:00:14 Quit mordocai_afk (Quit: Leaving)
04:04:30 Quit pixelma (Disconnected by services)
04:04:32 Join pixelma_ [0] (quassel@rockbox/staff/pixelma)
04:04:34 Nick pixelma_ is now known as pixelma (quassel@rockbox/staff/pixelma)
04:04:49 Quit amiconn (Disconnected by services)
04:04:50 Join amiconn_ [0] (quassel@rockbox/developer/amiconn)
04:05:08 Nick amiconn_ is now known as amiconn (quassel@rockbox/developer/amiconn)
04:06:49 Quit TheSeven (Ping timeout: 265 seconds)
04:07:05 Join Barahir [0] (
04:10:36 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
04:10:40 Quit Barahir_ (Ping timeout: 276 seconds)
04:18:11 Quit Raptors (Ping timeout: 240 seconds)
04:23:57 Quit simonrvn (Quit: ZNC -
04:28:07 Quit MethoS- (Read error: Connection reset by peer)
04:30:21 Quit thegeek (*.net *.split)
04:30:22 Quit sasquatch (*.net *.split)
04:30:23 Quit Battousai (*.net *.split)
04:30:58 Join Raptors [0] (
04:43:26 Join Mikeb0ok [0] (
04:43:28 Join MikeH__ [0] (
04:43:28***Alert Mode level 1
04:43:28DBUGEnqueued KICK mystica555
04:43:28DBUGEnqueued KICK mystica555_
04:43:28***Alert Mode level 2
04:43:28DBUGEnqueued KICK Mikeb0ok
04:43:28DBUGEnqueued KICK MikeH__
04:43:28***Alert Mode level 3
04:43:37 Quit mystica555 (Ping timeout: 240 seconds)
04:44:38 Quit mystica555_ (Ping timeout: 240 seconds)
04:53:29***Alert Mode OFF
04:57:33 Join simonrvn [0] (
05:01:07 Join thegeek [0] (
05:01:07 Join sasquatch [0] (~username@
05:01:07 Join Battousai [0] (~bryan@gentoo/developer/battousai)
05:16:04 Quit panni_ (Quit: ( :: NoNameScript 3.81 :: ))
05:17:53***Saving seen data "./dancer.seen"
05:24:59 Quit milk (Read error: Connection reset by peer)
05:35:13 Join Judas_PhD [0] (
05:38:13 Join Rob2223 [0] (
05:39:44 Quit Horscht (Quit: Verlassend)
05:42:11 Quit Rob2222 (Ping timeout: 250 seconds)
05:59:59 Quit Keripo (Quit: Leaving.)
06:03:54 Quit Judas_PhD (Quit: This is a quitting message)
06:05:42 Join Keripo [0] (
06:12:06 Join T44 [0] (
06:14:41 Quit Topy44 (Ping timeout: 250 seconds)
06:26:32JdGordon1S_a_i_n_t: so, do you reckon I can leave it up to you to add all the screens to make FS #11925 work (with a bit of guidance)?
06:26:52JdGordon1it involves adding 2 calls per screen (one on entry, one on exit)
06:28:47 Quit Keripo (Quit: Leaving.)
06:34:04 Join Keripo [0] (
06:47:36 Quit Keripo (Read error: Connection reset by peer)
06:49:34 Join Keripo [0] (
06:51:03 Join liar [0] (
07:07:25 Quit JdGordon1 (Quit: leaving)
07:17:56***Saving seen data "./dancer.seen"
07:18:44 Quit eGen_ (Quit: ... gettin' screew my wife ....)
07:22:10 Join Guinness` [0] (
07:25:11 Quit Guinness (Read error: Connection reset by peer)
07:25:22 Join eGen_ [0] (
07:26:37 Join sideral [0] (~sideral@
07:26:37 Quit sideral (Changing host)
07:26:37 Join sideral [0] (~sideral@rockbox/developer/sideral)
07:32:19 Nick MikeH__ is now known as mystica555 (
07:32:19DBUGEnqueued KICK mystica555
07:35:50 Quit sideral (Quit: Leaving.)
07:39:41 Join JdGordon| [0] (
07:39:41 Quit JdGordon| (Changing host)
07:39:41 Join JdGordon| [0] (~jonno@rockbox/developer/JdGordon)
07:54:42 Quit sasquatch (Ping timeout: 240 seconds)
07:58:37 Join sideral [0] (~sideral@
07:58:49 Quit sideral (Changing host)
07:58:49 Join sideral [0] (~sideral@rockbox/developer/sideral)
08:02:26 Quit soap (Ping timeout: 240 seconds)
08:02:50 Quit krazykit (Ping timeout: 240 seconds)
08:03:49 Join krazykit [0] (
08:05:54 Join soap [0] (~soap@rockbox/staff/soap)
08:11:39 Join sasquatch [0] (~username@
08:14:16 Join PurlingNayuki [0] (~PurlingNa@
08:14:57PurlingNayukiI have some problems when compiling Rockbox for Android.
08:18:54S_a_i_n_tWell, no one can help you unless you say what the problem(s) is/are.
08:19:12PurlingNayukiI'm typing.
08:19:31PurlingNayukiYou know, my English is poor
08:20:11PurlingNayukiI tried to compile 1024x600, but it tell me that there isn't a file called rockboxlogo.h.
08:20:42PurlingNayukiBut when I tried to compile the default definition (320x480) it was OK
08:20:58PurlingNayukiI don't really know why.
08:22:03S_a_i_n_t perhaps there is some hardcoded artificial limit, but that would only be speculation on my part, I would have expected it to work also.
08:22:51PurlingNayukiSo now I'm trying to compile 400x845.
08:24:05CtcpIgnored 5 channel CTCP requests in 2 hours and 48 minutes at the last flood
08:24:05*PurlingNayuki is willing to see the compiling goes correctly.
08:25:50PurlingNayukiI think I should change the file name of the Rockbox's boot image.
08:37:07 Join {phoenix} [0] (
08:39:25PurlingNayukiIt's true that Rockbox has some limit for higher definition.
08:42:00 Join kugel [0] (~kugel@
08:42:01 Quit kugel (Changing host)
08:42:01 Join kugel [0] (~kugel@rockbox/developer/kugel)
08:44:53 Quit Dreamxtreme (Read error: Connection reset by peer)
08:49:35 Quit Quazgaa (Ping timeout: 265 seconds)
08:50:51 Join Quazgaa [0] (
08:55:26 Quit PurlingNayuki (Remote host closed the connection)
08:55:55 Join pamaury [0] (
08:55:56 Quit pamaury (Changing host)
08:55:56 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
08:56:19 Quit TheSeven (Ping timeout: 240 seconds)
08:56:20 Join Dreamxtreme [0] (~Dre@
08:57:38 Quit factor (Ping timeout: 240 seconds)
08:59:13 Join ender` [0] (
08:59:30 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
08:59:36 Join petur [0] (d408b802@rockbox/developer/petur)
08:59:36 Join efyx [0] (
09:03:03 Quit kugel (Ping timeout: 240 seconds)
09:05:04 Quit sideral (Quit: Leaving.)
09:17:29 Quit JdGordon| (Ping timeout: 276 seconds)
09:17:59***Saving seen data "./dancer.seen"
09:22:13 Join n1s [0] (
09:22:13 Quit n1s (Changing host)
09:22:13 Join n1s [0] (~n1s@rockbox/developer/n1s)
09:27:30 Join mudd1 [0] (
09:36:07 Join einhirn [0] (
09:54:52 Quit n1s (Quit: Lmnar)
09:55:01 Join n1s [0] (~n1s@rockbox/developer/n1s)
09:57:50 Join afk [0] (Dre@
09:57:50 Quit Dreamxtreme (Read error: Connection reset by peer)
09:58:04 Quit afk (Client Quit)
10:02:32 Join kevku [0] (~kevku@2001:7d0:0:f9af:feed:feed:feed:feed)
10:06:14 Join Dreamxtreme [0] (~Dre@
10:20:25 Quit linuxstb (Read error: Connection reset by peer)
10:20:53 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
10:27:17 Quit Keripo (Quit: Leaving.)
10:32:04 Join vnl [0] (
10:32:07 Part vnl
10:33:24 Quit antil33t (Read error: Connection reset by peer)
10:40:24 Join antil33t [0] (
10:41:30 Join DerPapst [0] (
10:51:18 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
10:54:11 Join francesco_ [0] (
10:56:00francesco_my clip+ reboot to OF every time it's plugged in to USB. is there a way to avoid that? i've being pointed to FS #11664. is there the right bug?
10:59:50 Join pondlife [0] (~Steve@rockbox/developer/pondlife)
11:03:36pondlifeHi all. I'm trying to debug a reproducible-ish queue overflow in the database generation, using the sim.
11:03:41 Join sideral [0] (~sideral@
11:03:42pondlifeCurrently it results in a KERNEL_ASSERT which calls exit(). Is there an assertion available that merely breaks execution in gdb?
11:03:45 Quit sideral (Changing host)
11:03:45 Join sideral [0] (~sideral@rockbox/developer/sideral)
11:05:51n1spondlife: can't you set a breakpoint on it in gdb?
11:06:25pondlifeOf course! Thanks
11:06:44n1sfrancesco_: rockbox usb suport isn't working stably on that target yet and that patch is a workaround that doesn't seem to work 100% either
11:06:55Zagoroh, we have native usb for ams. sorry for misleading you yesterday, francesco_!
11:07:07n1sZagor: it's diabled though
11:07:31Zagoris it? then I was right and rebooting is the expected behaviour?
11:07:51sideralPondlife: Could the problem you're debugging have something to do with FS #11821?
11:08:04Zagorn1s: ok. thanks.
11:08:12n1sbecause it could cause these weird crashes that made a hard reset not work so people had to wait for the batteries to drain
11:09:00pondlifesideral: Maybe, I haven't done much looking into it yet
11:09:04sideralfrancesco_: Yes, that's the right tracker item. It's a bit difficult to track the current status, though, as there's no complete summary in that item.
11:09:16pondlifesideral: Will apply that patch and repeat
11:10:13sideralfrancesco_: There are patches in there for enabling USB on the Clip+ and for an USB-bus-reset-related workaround. There are also hints what various people have done to make it work
11:11:29sideralfrancesco_: For me, the USB support is completely stable without the bus-reset workaround, but only if I back out a recent change to the SD driver (r29169)
11:11:39sideral(using a ClipV2)
11:12:20 Join casainho [0] (
11:12:49sideralpondlife: There may be a synchronization problem in the code that patch disables. It would be great if you could find out what it is. Unfortunately I cannot reproduce the crash.
11:13:19 Quit casainho (Remote host closed the connection)
11:13:24sideralIf you have a data set that reproduces it reliably with a simulator, I could help debugging it.
11:18:00***Saving seen data "./dancer.seen"
11:19:58 Join esperegu [0] (~quassel@
11:20:07 Join qurvel [0] (
11:21:55pondlifesideral: Data set is about 80MB of MP3s... :/
11:22:24pondlifeBut thanks, will report on FS if this is also my case
11:24:46sideralpondlife: perhaps you can truncate all those files to a few hundred bytes and still reproduce the crash? ID3v2 info typically is at the beginning of the file IIRC
11:28:02 Quit kevku (Quit: KVIrc 4.0.2 Insomnia
11:37:50 Quit sideral (Quit: Leaving.)
11:40:41 Quit TheSeven (Disconnected by services)
11:40:42 Join [7] [0] (~TheSeven@rockbox/developer/TheSeven)
11:45:46 Quit qurvel (Quit: Lost terminal)
11:50:46francesco_sideral: thanks. can you confirm that clip v2 is 8Gb only?
11:58:14 Join Kitar|st [0] (
11:58:59 Quit user890104 (Ping timeout: 272 seconds)
11:59:09 Join user890104 [0] (
12:17:15 Quit mudd1 (Ping timeout: 240 seconds)
12:18:53 Join MethoS [0] (~clemens@
12:19:20 Nick MethoS is now known as Guest64118 (~clemens@
12:28:24CIA-70New commit by 03gevaerts (r29281): Make sure there is a logo for LCD_WIDTH>480 && LCDWIDTH<600 (useful for 600x1024 LCDs)
12:31:03n1sgevaerts: typo in the message?
12:31:14gevaertsah, probably
12:31:35gevaertsThe code itself is fine I hope :)
12:31:43n1sah s/00/0/
12:31:50n1sah s/00/40/
12:32:41n1sdidn't make much sense since 600 !< 600
12:33:13CIA-70r29281 build result: 1250 errors, 488 warnings (gevaerts committed)
12:34:37n1sweird error
12:35:31 Quit eGen_ (Quit: ... gettin' screew my wife ....)
12:36:47gevaertsDoesn't happen here
12:36:53*gevaerts claims a random client problem
12:38:08n1sthe build log is supicioulsy short
12:38:22pondlifesideral: For the logs - I put a backtrace (for vanilla SVN) up at - looks like a sim fault, not a database fault to me..
12:39:26 Join eGen_ [0] (
12:46:58 Join PurlingNayuki [0] (~PurlingNa@
12:48:27PurlingNayukiI found that there is a update for 600x1024 Rockbox for Android.
12:50:39 Quit {phoenix} (Ping timeout: 255 seconds)
12:50:39 Join {-phoenix-} [0] (
12:57:17 Quit DerPapst (Quit: Leaving.)
12:58:51 Join sideral [0] (~sideral@
12:58:52 Quit sideral (Changing host)
12:58:52 Join sideral [0] (~sideral@rockbox/developer/sideral)
13:00:15sideralpondlife: Good to know, thanks
13:01:06pondlifeAm now running the test with your patch - but it takes 30 mins to crash (init big database over a slow network link)
13:02:51sideralfrancesco_: I don't know for sure in which sizes the ClipV2 is sold. I guess that it comes in all three sizes (2, 4 or 8 GB).
13:03:28sideralMine is an 8 GB model
13:06:09 Join dfkt [0] (dfkt@unaffiliated/dfkt)
13:08:05 Quit sideral (Remote host closed the connection)
13:08:22 Join sideral [0] (~sideral@rockbox/developer/sideral)
13:12:23 Quit user890104 ()
13:18:04***Saving seen data "./dancer.seen"
13:18:20 Quit sideral (Remote host closed the connection)
13:18:58 Quit S_a_i_n_t (Disconnected by services)
13:19:00 Join [Saint] [0] (S_a_i_n_t@
13:19:06 Join sideral [0] (~sideral@rockbox/developer/sideral)
13:20:38 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
13:28:35 Quit sideral (Remote host closed the connection)
13:28:55 Join sideral [0] (~sideral@
13:28:55 Quit sideral (Changing host)
13:28:55 Join sideral [0] (~sideral@rockbox/developer/sideral)
13:30:03 Part Zagor
13:31:51 Join mudd1 [0] (~cmertes@2001:638:504:20e0:221:70ff:fe83:655e)
13:35:53 Quit liar (Read error: No route to host)
13:36:30 Join TheLemonMan [0] (~lem0n@
13:41:25 Join liar [0] (
13:42:01 Quit jhMikeS (Ping timeout: 240 seconds)
13:42:49 Join jhMikeS [0] (
13:42:49 Quit jhMikeS (Changing host)
13:42:49 Join jhMikeS [0] (~jethead71@rockbox/developer/jhMikeS)
13:43:41pondlifesideral: Same crash happens with the patch from FS #11821
13:44:36sideralpondlife: Thanks for the update! Sorry for you, good for me :) It means that the code this patch removes hasn't triggered your crash
13:46:01 Quit TheLemonMan (Quit: free(me))
13:53:22 Join factor [0] (~factor@
13:54:23 Part pondlife
14:03:35 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
14:09:45 Join runner [0] (
14:11:14 Quit runner (Client Quit)
14:12:45 Part PurlingNayuki
14:27:00 Join pokey [0] (
14:28:01pokeyanybody active this time of day?
14:30:04 Part pokey ("ok. i'll try again later")
14:39:03 Join pondlife1 [0] (
14:39:13 Part pondlife1
14:39:14 Join coc00n [0] (
14:42:45 Quit krazykit (Quit: awe yeeeeeee)
14:44:25 Join {phoenix} [0] (
14:45:45 Quit {-phoenix-} (Ping timeout: 276 seconds)
15:03:20 Join TheLemonMan [0] (~lem0n@
15:04:03 Part Zagor
15:06:55 Join DerPapst [0] (
15:06:55 Quit francesco_ (Remote host closed the connection)
15:08:27 Join Zagor [0] (~bjst@rockbox/developer/Zagor)
15:09:36 Quit sbhsu (Quit: leaving)
15:09:57 Join sbhsu [0] (
15:11:42 Join komputes [0] (~komputes@ubuntu/member/komputes)
15:15:04 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
15:16:47 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
15:17:05 Join t0rc [0] (~t0rc@unaffiliated/t0rc/x-5233201)
15:18:05 Join Xerion_ [0] (
15:18:08***Saving seen data "./dancer.seen"
15:18:42 Quit Xerion (Ping timeout: 255 seconds)
15:18:42 Nick Xerion_ is now known as Xerion (
15:20:10 Quit petur (Quit: Page closed)
15:44:51 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
15:49:24 Join markun [0] (5571f93d@rockbox/developer/markun)
15:50:33 Quit einhirn (Ping timeout: 245 seconds)
16:05:07 Quit GeekShadow (Ping timeout: 260 seconds)
16:14:12 Join Xerion_ [0] (
16:14:14 Join tchan1 [0] (
16:14:57 Quit thegeek (*.net *.split)
16:14:58 Quit Battousai (*.net *.split)
16:15:00 Quit markun (*.net *.split)
16:15:00 Quit evilnick_B (*.net *.split)
16:15:01 Quit domonoky (*.net *.split)
16:15:03 Quit esperegu (*.net *.split)
16:15:03 Quit linuxstb (*.net *.split)
16:15:03 Quit efyx (*.net *.split)
16:15:04 Quit T44 (*.net *.split)
16:15:05 Quit yosafbridge (*.net *.split)
16:15:06 Quit logiclost (*.net *.split)
16:15:08 Quit liar (*.net *.split)
16:15:09 Quit n1s (*.net *.split)
16:15:09 Quit Raptors (*.net *.split)
16:15:10 Quit Zarggg (*.net *.split)
16:15:11 Quit froggyman (*.net *.split)
16:15:11 Quit Strife89 (*.net *.split)
16:15:11 Quit jfc (*.net *.split)
16:15:12 Quit kkit|sh (*.net *.split)
16:15:12 Quit dionoea (*.net *.split)
16:15:12 Quit pikytcus (*.net *.split)
16:15:12 Quit scorche|sh (*.net *.split)
16:15:12 Quit MagusG (*.net *.split)
16:15:14 Quit ChanServ (*.net *.split)
16:15:14 Quit jhMikeS (*.net *.split)
16:15:14 Quit [7] (*.net *.split)
16:15:15 Quit Quazgaa (*.net *.split)
16:15:15 Quit mystica555 (*.net *.split)
16:15:16 Quit timccc (*.net *.split)
16:15:17 Quit ehntoo (*.net *.split)
16:15:17 Quit feisar- (*.net *.split)
16:15:19 Quit Xerion (*.net *.split)
16:15:20 Quit antil33t (*.net *.split)
16:15:21 Quit tchan (*.net *.split)
16:15:21 Quit zu (*.net *.split)
16:15:22 Quit n17ikh (*.net *.split)
16:15:23 Quit Dhraakellian (*.net *.split)
16:15:24 Quit Farthen (*.net *.split)
16:15:25 Quit komputes (*.net *.split)
16:15:26 Quit Dreamxtreme (*.net *.split)
16:15:26 Quit soap (*.net *.split)
16:15:26 Quit Guinness` (*.net *.split)
16:15:27 Quit bluefoxx (*.net *.split)
16:15:27 Quit powell14ski_ (*.net *.split)
16:15:28 Quit linuxguy3 (*.net *.split)
16:15:29 Quit jepler (*.net *.split)
16:15:29 Quit Kohlrabi (*.net *.split)
16:15:30 Quit amee2k (*.net *.split)
16:15:31 Quit [Saint] (*.net *.split)
16:15:32 Quit rasher (*.net *.split)
16:15:32 Quit Galois (*.net *.split)
16:15:33 Quit ps-auxw (*.net *.split)
16:15:34 Quit bzed (*.net *.split)
16:15:34 Quit tintamarre (*.net *.split)
16:15:35 Quit ranmachan (*.net *.split)
16:15:35 Quit Elfish (*.net *.split)
16:18:04 Join jhMikeS [0] (
16:18:29 Join ehntoo [0] (
16:18:43 Join francesco [0] (
16:19:04 Join zu [0] (
16:19:46francescodoes anyone have the SKU for a Clip+ v2 ?
16:21:19 Nick francesco is now known as 5EXAB626W (
16:22:14 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
16:22:14 Join Dhraakellian [0] (
16:22:14 Join feisar- [0] (jljhook@
16:22:14 Join n17ikh [0] (
16:22:14 Join lostlogic [0] (~lostlogic@rockbox/developer/lostlogic)
16:22:14 Join domonoky [0] (~Domonoky@rockbox/developer/domonoky)
16:22:14 Join komputes [0] (~komputes@ubuntu/member/komputes)
16:22:14 Join liar [0] (
16:22:14 Join [Saint] [0] (S_a_i_n_t@
16:22:14 Join [7] [0] (~TheSeven@rockbox/developer/TheSeven)
16:22:14 Join esperegu [0] (~quassel@
16:22:14 Join linuxstb [0] (~linuxstb@rockbox/developer/linuxstb)
16:22:14 Join antil33t [0] (
16:22:14 Join Dreamxtreme [0] (~Dre@
16:22:14 Join n1s [0] (~n1s@rockbox/developer/n1s)
16:22:14 Join efyx [0] (
16:22:14 Join Quazgaa [0] (
16:22:14 Join soap [0] (~soap@rockbox/staff/soap)
16:22:14 Join Guinness` [0] (
16:22:14 Join T44 [0] (
16:22:14 Join Battousai [0] (~bryan@gentoo/developer/battousai)
16:22:14 Join thegeek [0] (
16:22:14 Join Raptors [0] (
16:22:14 Join tintamarre [0] (
16:22:14 Join rasher [0] (~rasher@rockbox/developer/rasher)
16:22:14 Join bluefoxx [0] (
16:22:14 Join Zarggg [0] (
16:22:14 Join timccc [0] (~timccc@
16:22:14 Join Galois [0] (
16:22:14 Join linuxguy3 [0] (
16:22:14 Join froggyman [0] (~seth@unaffiliated/froggyman)
16:22:14 Join Strife89 [0] (~Strife89@
16:22:14 Join jfc [0] (
16:22:14 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
16:22:14 Join kkit|sh [0] (
16:22:14 Join bzed [0] (
16:22:14 Join jepler [0] (~jepler@emc/developer/pdpc.professional.jepler)
16:22:14 Join dionoea [0] (~dionoea@videolan/developer/dionoea)
16:22:14 Join yosafbridge [0] (
16:22:14 Join Kohlrabi [0] (
16:22:14 Join Elfish [0] (amba@2a01:4f8:100:90a1:abc:abc:abc:abc)
16:22:14 Join amee2k [0] (
16:22:14 Join pikytcus [0] (
16:22:14 Join scorche|sh [0] (~scorche@rockbox/administrator/scorche)
16:22:14 Join MagusG [0] (
16:22:14 Join ranmachan [0] (
16:22:14 Join Farthen [0] (
16:22:14 Join ChanServ [0] (ChanServ@services.)
16:22:14Mode"#rockbox +o ChanServ " by
16:22:34 Join mystica555 [0] (
16:23:36 Quit TheLemonMan (Quit: free(me))
16:24:31 Join powell14ski_ [0] (
16:25:32 Quit [Saint] (*.net *.split)
16:25:33 Quit rasher (*.net *.split)
16:25:33 Quit Galois (*.net *.split)
16:25:34 Quit ps-auxw (*.net *.split)
16:25:34 Quit bzed (*.net *.split)
16:25:34 Quit tintamarre (*.net *.split)
16:25:36 Quit ranmachan (*.net *.split)
16:25:36 Quit Elfish (*.net *.split)
16:25:41 Join linuxguy4 [0] (
16:25:59 Join amee2k_ [0] (
16:26:14 Join feisar-_ [0] (
16:26:16 Join logiclost [0] (
16:27:06 Join [Saint] [0] (S_a_i_n_t@
16:27:06 Join tintamarre [0] (
16:27:06 Join rasher [0] (~rasher@rockbox/developer/rasher)
16:27:06 Join Galois [0] (
16:27:06 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
16:27:06 Join bzed [0] (
16:27:06 Join Elfish [0] (amba@2a01:4f8:100:90a1:abc:abc:abc:abc)
16:27:06 Join ranmachan [0] (
16:29:13 Nick tchan1 is now known as tchan (
16:29:18 Quit tchan (Changing host)
16:29:18 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
16:29:32 Quit komputes (Excess Flood)
16:29:54 Join kevku [0] (~kevku@2001:7d0:0:f9af:feed:feed:feed:feed)
16:30:12 Quit amee2k (Write error: Broken pipe)
16:30:12 Quit linuxguy3 (Write error: Broken pipe)
16:30:21 Quit lostlogic (Remote host closed the connection)
16:30:22 Quit feisar- (Remote host closed the connection)
16:35:38 Quit n1s (Quit: Lmnar)
16:39:12sideralfrancesco: Is there a Clip+ v2? I only know of these three Clip revisions: Clip v1, Clip v2, Clip+
16:42:58 Quit Zambezi (Changing host)
16:42:58 Join Zambezi [0] (Zulu@unaffiliated/zambezi)
16:43:595EXAB626Wsideral: the clip+ v2 is the one with an OF v.2
16:44:135EXAB626Wit also has a different case/hardware
16:44:27 Nick 5EXAB626W is now known as francesco_ (
16:44:58gevaertsThe clip v2 case is the same as the clip v1 case
16:45:46sideralI haven't yet heard of a Clip+ with a 2.x OF
16:45:52francesco_gevaerts: apparently not
16:46:23Zagorthat's the non-plus clip
16:46:27sideralThis is a Clip v2, not Clip+
16:47:02francesco_ah ok
16:47:07gevaertsfrancesco_: ok, "slightly different logo"...
16:47:22gevaertsThat doesn't make it a different case to me :)
16:47:53francesco_gevaerts: the clip is different :)
16:48:02francesco_..the actual clip
16:48:04gevaertsfrancesco_: in what way?
16:48:40 Join Keripo [0] (
16:49:40francesco_and the case too
16:49:46francesco_that's clip+
16:50:01francesco_the slide in front of it change quite a bit
16:50:35gevaertsYes, I know the clip+ is different
16:50:40francesco_anyway, sideral: you said that on your clip v2 (not the +) you dont' have the usb OF issue, right?
16:51:30francesco_gevaerts: ah ok.. no, clip and clip v2 are identical.. you're gith
16:51:34francesco_you're right
16:54:08sideralfrancesco_: More or less −− there were some qualifications to my statements
16:58:03 Quit Keripo (Quit: Leaving.)
16:58:41 Join jae_ [0] (
16:58:44CIA-70New commit by 03Buschel (r29282): Submit FS #11926, calibrate charge curve for e200v1. Thanks to Martin Ritter.
16:58:55 Join komputes [0] (~komputes@ubuntu/member/komputes)
17:00:44 Nick jae_ is now known as jae (
17:01:52 Join Buschel [0] (
17:01:53 Join pondlife1 [0] (
17:03:26CIA-70r29282 build result: All green
17:03:27francesco_sideral: got it. thanks
17:06:40 Join Keripo [0] (
17:06:44 Part Zagor
17:09:24 Nick amee2k_ is now known as amee2k (
17:15:03 Quit factor (Read error: Connection reset by peer)
17:15:14CIA-70New commit by 03Buschel (r29283): Use menu layout for sleep timer that is consistent to other timers. Closes FS #11923.
17:18:09***Saving seen data "./dancer.seen"
17:19:35CIA-70r29283 build result: All green
17:21:46 Quit Llorean (Read error: Connection reset by peer)
17:26:18 Quit [7] (Ping timeout: 265 seconds)
17:27:10 Quit [Saint] (Ping timeout: 272 seconds)
17:30:17 Join [Saint] [0] (S_a_i_n_t@
17:32:54 Join factor [0] (~factor@
17:34:48 Quit t0rc (Quit: Give someone code, help them with one project. Teach someone to code, help them rule the world.)
17:36:08 Quit mudd1 (Read error: Operation timed out)
17:38:06 Join kugel [0] (~kugel@rockbox/developer/kugel)
17:42:58sideralI have a draft for the new manual section describing autoresume ready. Is there any proofreading process I need to go through before it can be checked in?
17:43:02sideralIf not, is anyone interested / willing to have a look at it?
17:45:55pamauryAny GNU linker script expert ? I would like to do something a bit weird. I want to divide my code into two parts: one part at VA 0x8c0b0000 and the other one at VA 0x300b0000. The point is that these two alias (after memory setup) so what I really want is to FIRSt write the code at 0x8c0b0000 until 0x8c000byyyy and THEN write the rest of the code at 0x300byyyy. Is this possible ?
17:52:06linuxstbsideral: I don't think anyone would be upset if you just committed it. But if you want comments from others, then post a link here, or (probably better) put it on flyspray.
17:57:09 Quit Keripo (Quit: Leaving.)
18:00:32siderallinuxstb: OK, thanks. FS is probably a bit too heavy-weight for this, so I'll go ahead and commit and then ask for volunteer reviewers when the manual appears on the website :)
18:02:29sideralpamaury: that should be possible with the AT directive and a sum or difference involving ADDR(<currentsymbol>) as AT's argument
18:03:57 Join Llorean [0] (~DarkkOne@rockbox/user/Llorean)
18:13:19francesco_if i place a .m3u file in the microsd card of a clip+, what should be the base path to the songs also stored in the microsd?
18:14:21 Quit Buschel (Quit: ChatZilla 0.9.86 [Firefox 3.6.13/20101203075014])
18:15:08pamaurysideral: yeah I finally figured out this using thr manual, but thanks for your answer :)
18:17:43CIA-70New commit by 03sideral (r29284): Describe autoresume in the manual.
18:18:29 Join TheLemonMan [0] (
18:18:58evilnick_Bfrancesco_: If the .m3u is in the root of the sd card you should be able to use a relative path to the music files
18:19:32evilnick_Be.g. music/artist/album/track (if that's the directory structure that you've got)
18:21:56CIA-70r29284 build result: All green
18:22:34*kugel has re-written his yesterdays lcd speed up to be backwards compatible
18:25:06francesco_evilnick_B: it used to be an absolute path, but i'll try yourt suggestion as absolute don't work anymore (it seems)
18:29:41francesco_evilnick_B: nope, not working with the relative either
18:30:17evilnick_BCan you give some examples of both the entry in the .m3u and also where the file is on the sd card?
18:30:27evilnick_BWell, one example should do
18:34:23 Join Horscht [0] (~Horscht@xbmc/user/horscht)
18:37:25 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
18:39:33francesco_evilnick_B: both .m3u and .mp3 are on sd; mp3 are in MUSIC/hiphop/. m3u is in root folder of sd.
18:39:47francesco_the path i use in .m3u is MUSIC/hiphop/mp3_file
18:41:06evilnick_BCan you try making a playlist by playing an item in the /MUSIC/hiphop folder, then save that playlist and open it in a text editor - that should let you see what RB is expecting to see
18:41:46evilnick_BIIRC there might need to be a <sd_card> (but likely to be different, as I can't recall exactly what that should be) at the left of that path
18:42:05francesco_evilnick_B: ok i'll try
18:43:30 Join Keripo [0] (
18:45:08 Join Horschti [0] (
18:48:06 Join feisar- [0] (
18:51:32 Join robin0800 [0] (
18:51:35 Quit feisar-_ (Ping timeout: 255 seconds)
18:51:36 Quit evilnick_B (Ping timeout: 245 seconds)
18:51:37 Quit Dhraakellian (Ping timeout: 255 seconds)
18:51:37 Quit Horscht (Ping timeout: 265 seconds)
18:51:39 Join mystica555_ [0] (
18:51:40 Quit mystica555 (Ping timeout: 255 seconds)
18:51:40 Join The3_14ed|r [0] (
18:51:41 Quit Horschti (Changing host)
18:51:41 Join Horschti [0] (~Horscht@xbmc/user/horscht)
18:53:52 Quit pondlife1 (Quit: Leaving.)
18:55:37 Quit antil33t (Read error: Connection reset by peer)
18:55:46 Join antil33t [0] (
19:01:16 Quit tchan (Ping timeout: 240 seconds)
19:02:47 Join thomasjfox [0] (
19:03:18francesco_ evilnick_B: ok , so path must be /<microSD1>/folder/file.MP3
19:05:22 Join bmbl [0] (
19:05:22 Quit bmbl (Changing host)
19:05:22 Join bmbl [0] (~bmbl@unaffiliated/bmbl)
19:13:08pixelmasideral: did you testbuild some manuals? Unfortunately we only have daily manuals and it's sometimes not easy to spot if they are broken (that's also not part of the build system). I could imagine that the \ButtonLeft and \ButtonRight doesn't always work, I believe there was one pad that doesn't define those (I believe it was the H10) and seeing as it has to do with skipping I guess the \ActionWpsSomething is better suited anyway and solves this
19:13:08pixelmaissue. Other than that I'm not too sure I like the checkin as the feature is still questioned, on the other hand... well I'm really undecided
19:14:30 Join evilnick_B [0] (0c140464@rockbox/staff/evilnick)
19:15:59 Nick Horschti is now known as Horscht (~Horscht@xbmc/user/horscht)
19:18:13***Saving seen data "./dancer.seen"
19:18:17sideralpixelma: Yes, I did testbuild. I had strange problem because apparently my LaTeX tool chain is not complete, but I did get OK-looking PDF and HTML version.
19:18:17sideralI thought I'd seen \ButtonLeft and \ButtonRight used in generic parts of the manual, but now that I'm looking for it, I can't find it any longer :)
19:19:13sideralSo you're possibly right
19:19:31pixelmayes, ButtonLeft and ButtonRight do almost work everywhere but I thought there was one or two exceptionsm using the Actions is still better suitable I think
19:19:50pixelmaeverywhere = with the different keypads
19:20:23pixelmadid you testbuild one or different manuals?
19:22:45 Quit Mikeb0ok (Ping timeout: 246 seconds)
19:22:56sideraltwo different ones
19:23:14sideralone with hwcodec, one with swcodec
19:25:45 Join Mikeb0ok [0] (
19:26:29sideralpixelma: "make manual-pdf" yields a nice pdf manual, but the build ends with "make[2]: *** [rockbox-build.pdf] Error 1"
19:26:52 Join petur [0] (~petur@rockbox/developer/petur)
19:27:10pixelmaif you want the pdf "make manual" is enough
19:27:55pixelmanot sure if that's the reason for your error though, never tried that command
19:28:34sideral"make manual" yields the same error
19:29:43sideralhmm, in another build dir it works...
19:30:07pixelmathen there may be something else broken, the manual currently builds in non-stop mode which means it'll go all the way through the end even on an error - and the build output is very noisy so the error is not easy to spot
19:30:22gevaertssideral: that's with current svn?
19:30:27*pixelma dislikes non-stop mode
19:30:30gevaertsWhich target did you try?
19:30:39 Join GeekShadow [0] (~Antoine@reactos/tester/GeekShadow)
19:30:40pixelmaand which... what gevaerts said
19:30:41sideralyep. but let me try make clean first before raising a flag..
19:31:25 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
19:31:31sideralah, that helped :)
19:32:05sideralwhat's a target without a Left button? Any touch-wheel or touch-screen target?
19:32:19 Join bertrik [0] (~bertrik@rockbox/developer/bertrik)
19:33:38pixelmaI don't know off the top of my head (and maybe I'm misremembering), you would need to check the keymap-* files in the platform folder - but really \ActionWpsPrev/Next is cleaner and safer anyway
19:34:06AlexPYeah, just use the actions - they should exist as in the code, and if they don't they should be added
19:34:36AlexPThat way also if buttons change (unlikely to here), it is much easier to just change a few platform files than search all through all the files
19:35:20sideralJust wanted to check whether "\ActionWpsPrev{} button" expands to something silly on such a target, such as "swoosh to the left while holding upside down button" :)
19:37:06pixelmayou won't need the button
19:37:19pixelmajust "\ActionWpsPrev"
19:37:41AlexPCheers for doing the manual though
19:37:48AlexPNot everyone does :)
19:38:06sideralpixelma: but "press \ActionWpsPrev{}" also sound awkward as "press swoosh to the left while holding upside down" :)
19:39:08sideralAlexP: I figured it's necessary, given that the feature needs some explanation, but thanks for the kudos :)
19:39:25AlexPIt is necessary, but that doesn't always mean people do it :)
19:39:54AlexPAnd then someone else (such as pixelma or I) ends up doing it, which takes much longer as I have to work out how the feature works first :)
19:41:12pixelmato me it's also hard to write good-sounding English explanations, so I mostly do code "cleanup"/filling out missing info
19:42:02sideralAlexP: I should have let you do it, so that you get a chance to fall in love with my feature :)
19:42:15AlexPhehe :)
19:42:35pixelmait might have had the opposite effect
19:45:10CIA-70New commit by 03sideral (r29285): Use generic keyboard actions rather than the device-dependent ...
19:49:31CIA-70r29285 build result: All green
19:49:44sideralThank you pixelma!
19:50:06 Join leavittx [0] (~leavittx@
19:50:21 Quit DerPapst (Quit: Leaving.)
19:55:07sideralpixelma: Re the possibly contended commits: I'm naturally uneasy about this as well. As I don't know what the nature of the steering board request is, it's hard to anticipate what the board's decision might be and rush ahead and implement it.
19:55:35sideralBut as I think completing the work is better than being stuck with a half-finished state in SVN, I decided to finish it up speculatively.
19:56:04sideralAll this is easy enough to revert if necessary.
19:56:35sideralThis is specifically not to create facts that cannot be reconsidered.
19:58:25 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
20:02:28 Quit Galois (Quit: Leaving)
20:04:00 Quit robin0800 (Quit: Leaving)
20:04:49 Nick The3_14ed|r is now known as Dhraakellian (
20:07:46 Quit thomasjfox (Remote host closed the connection)
20:09:10Lloreansideral: While I'm against the feature, I'd say keep working on it until such time as any decision's made on it.
20:09:50gevaertsTaking into account of course that if you add any feature after Sunday evening, AlexP will come after you
20:09:57Lloreansideral: From my perspective, until the RSB makes *any* statement, it's still "good." If the RSB decides it's worried about it being hard to back out later, they can always say "please halt commits to it until we've made our decision" and at this time they haven't said anything like that yet.
20:10:12LloreanAs far as I know, at least.
20:10:38sideralgevaerts: naturally. It's one of the reasons for completing it now :)
20:10:44AlexPUnless I'm told to postpone the freeze :)
20:11:05AlexPor give special dispensation :)
20:11:14 Join vnl [0] (
20:11:22sideralLlorean: Agreed
20:11:23 Join stoffel [0] (
20:11:56sideralAlexP: If you come after me, I'll give you some cookies!
20:12:21AlexPWho told you?
20:12:36AlexPOr did you just look at devcon photos and make an informed guess? :)
20:12:53sideralAlexP: everyone loves my cookies!
20:13:10LloreanWhat do people feel about having the RC builds be offered *instead* of the "Current Build" during the branch week?
20:13:39gevaertsLlorean: that might be a good idea
20:13:48gevaertsNot sure how much it work it is to implement though
20:13:57sideralAlexP: Now excuse me while I look up those photos ;)
20:14:39Lloreangevaerts: Yeah. It just seems like an idea that might be good to discuss - the more people we can harness to test RCs, the less likely things are to slip through the cracks.
20:15:09AlexPLlorean: would be nice - just for people to install manually
20:15:51LloreanAlexP: I was thinking that we could go as far as having RBUtil not offer the typical "current build" during that period, but having the button say "release candidate" instead, and download that.
20:16:04AlexPI could do some building, would need somewhere to host them
20:16:21LloreanIt's just a short period, and for *many* purposes to the user they're similar enough, while a dev will be able to build current builds, and anyone who wants current builds can still manually download 'em from the site
20:16:23AlexPLlorean: yes, possibly - would have to be for the next release though
20:16:28LloreanYeah, not this time
20:16:50LloreanI was more testing the idea of *replacing* the availability of the current build for the "easier" installs, to encourage use of the RC, than suggesting it being done now.
20:16:52 Quit Stummi (Quit: Bye!)
20:17:15 Join Stummi [0] (~Stummi@rockbox/developer/Stummi)
20:17:21LloreanI'm sure there are reasons not to 'hide' the current build (at least within the util, I don't think any change to the website need be made) for a week, but I haven't thought of them yet.
20:17:46LloreanAnd I know the surest way to find out what's wrong with my opinion on something is to speak it out loud. ;)
20:19:11 Join Galois [0] (
20:19:20kugelaren't we talking about RC builds at every upcoming release?
20:19:58gevaertsYes, but usually nobody makes an actually practical suggestion :)
20:20:02LloreanI'd suggest we build every change after the branch and consider them RCs.
20:21:51AlexPWe haven't really thought about RC builds enough
20:22:00AlexPe.g. we need some way for people to report
20:22:04gevaertsWe'll need to catch Zagor to discuss what's possible in the build system
20:22:07AlexPA questionnaire or something
20:22:45LloreanYes, I think there was once proposed a sort of "checklist" of features to ask people to confirm work on their hardware
20:29:25 Join robin0800 [0] (
20:29:27 Join Rim13 [0] (5f1b282e@gateway/web/freenode/ip.
20:33:57 Quit Rim13 (Ping timeout: 245 seconds)
20:41:07 Quit TheLemonMan (Quit: free(me))
20:47:32 Quit francesco_ (Remote host closed the connection)
20:51:32CIA-70New commit by 03dave (r29286): FS #11924 - Use separate targets for RaaA devices. This changes tools/configure to present the (currently four) possible RaaA targets as separate menu ...
20:52:43 Part vnl
20:52:43*linuxstb now tries to understand all the #defines used to build RaaA...
20:56:04CIA-70r29286 build result: All green
20:56:35sideralSpeaking of features not to commit after Sunday... Anyone care to give a brief assessment of the FS-tracked features I like?
21:01:25 Quit Keripo (Quit: Leaving.)
21:03:54sideralFS #9448 Rewind on fade
21:04:44sideralFS #11769 Change USB charge mode detection key / FS #10198 USB charge only setting
21:05:05sideralFS #10343 Resume playback even if it reached the end and stopped / FS #11644 Playlist isn't loaded after reboot if it had ended
21:05:12sideralFS #11605
21:05:45sideralFS #11605 Clip+ new keymap
21:05:55sideralFS #11767 Meier crossfeed
21:06:01Lloreansideral: I don't know what exactly has stopped the 11769, but it's been pretty contentions in the past. I don't even know exactly why though
21:06:08sideralFS #11891 Add mp3 gap skipping support to improve gapless playback
21:07:03sideralThese are the ones I have in my private build
21:07:36 Quit petur (Quit: back in a few days ;))
21:07:48Lloreansideral: I really don't like that keymap, judging from the last patch. It seems too different from other standard Rockbox keymaps (making a menu button stop, and the power button bring up the menu, etc)
21:08:34sideralLlorean: Yeah, I have heavily modified that keymap as well. But the stock Clip keymap lacks Hotkey and Quickscreen
21:08:36LloreanAlso, the quickscreen should be on the same button in the lists as the WPS
21:08:43LloreanPutting it on "down" makes that impossible.
21:08:50LloreanMy Clip+ has the quickscreen just fine.
21:08:53LloreanAnd it's SVN
21:09:25LloreanHold "Home"
21:09:36LloreanJust like it is on many (most?) other Rockbox players
21:09:59sideralMaybe I just never found it :) I have it on Long-down now
21:10:09LloreanLong down doesn't work in lists very well.
21:10:15AlexPFS #10343 has lead to many an argument, long and winding
21:10:32LloreanInstead of changing many things around in the keymap, maybe a separate task should be brought up for "where can we put hotkey?"
21:10:41LloreanI'm not certain there's a button free for it, at least in the list
21:10:57AlexPConsistency across targets for keymaps (where possible) is an absolute must for me
21:10:59LloreanI thought I'd heard that hotkey was on a button combination on Clip+, but being a feature I don't use, I wouldn't know.
21:11:05sideralLlorean: hotkey is long-Home for me. works very well
21:11:17LloreanAlexP: The Clip+ is pretty consistent right now, and as I use it as my everyday player, I feel the keymap is just fine at the moment (better than some, even)
21:11:40AlexPLlorean: Pretty much my point - it shouldn't be made less consistent in SVN
21:12:09LloreanI'm not sure how I feel about the MP3 gap skipping. Last I'd read, the silence detection wasn't that great yet?
21:12:17linuxstbFS #11891 looks like it needs more discussion, rather than a quick commit before release.
21:12:53sideralLlorean: that patch works fine for me. It removes all gaps from commercially purchased MP3s, which often come without the LAME tags
21:13:19siderallinuxstb: I'm afraid none of them is ready for a quick commit :)
21:13:24LloreanI think rewind on pause as a setting with identical options to rewind on stop would likely be relatively non-contentious at least
21:13:42Lloreansideral: I thought someone commented that the "silent" values were still quite loud.
21:13:56LloreanToo aggressive, rather than not aggressive enough. But I don't have the task watched, so maybe that was improved?
21:14:33linuxstbsideral: Yes, I guess if any were ready for commit, they would have been...
21:14:38sideralLlorean: No, but there was an attempt to refute this objection AFAIR
21:15:11gevaertssideral: how can one refute that?
21:15:12sideralI like the Meier crossfeed patch. In my view, that could completely replace the stock crossfeed mechanism
21:15:25gevaertsEither it deletes some real audio, or it doesn't
21:15:41LloreanIt's non-customizable though (the meier crossfeed) and the patch already implements it as coexisting, right?
21:16:23sideralLlorean: right. but the coexistence as done now is a bit confusing because Meier renders all other settings useless
21:16:56sideralgevaerts: Someone argued that the "aggressive" level was OK, and from my experience, it indeed is
21:17:01 Join saratoga [0] (9803c6dd@gateway/web/freenode/ip.
21:17:13gevaertssideral: surely that very much depends on the music you listen to?
21:17:22gevaertsHave you tried it with e.g. Ravel's Bolero?
21:17:39gevaerts(or any other piece that starts with very nearly silence)
21:17:56sideralgevaerts: that patch removes only quasi-silence from the first and last frame
21:18:00saratogasideral: FS #11769 can probably go in, as far as I know no one had da problem with it
21:18:02Lloreansideral: We have plenty of options that aren't used if their related feature is turned off, or is in a non-configurable mode, such as stereo width when it's not set to "custom"
21:18:06saratogai have mixed feelings about the mp3 patch though
21:18:15gevaertssideral: that first frame can be a very quiet note
21:18:17***Saving seen data "./dancer.seen"
21:18:36LloreanIn a song that fades in / lifts up quietly...
21:18:57LloreanShould there even be a (non-intentional) gap at the start?
21:18:58sideralgevaerts: Right. But the setting is clearly useful for some (me ;) )
21:19:12LloreanI thought the idea of the patch was to correct for errors in MP3's gaplessness, which should mean it should only affect the last frame?
21:19:18gevaertssideral: sure, but that doesn't mean it's perfect right now :)
21:19:28TorneLlorean: no, mp3 is imperfect at both ends
21:19:31saratogamp3 frames are so short i doubt cutting a silent or quiet one would be noticable, but its very hacky
21:19:55Tornedoesn't hte patch only do it if there's no lame tag or equivalent, also?
21:19:55sideralLlorean: that's what I thought as well, but mp3 sucks and introduces gaps in the first frame during both encoding and decoding :)
21:19:59LloreanTorne: Isn't the imperfection at the beginning related to decoder/encoder (rather, something we can know by means other than trying to "guess" at silence, or perhaps, by detecting only pure silence?)
21:20:16TorneLlorean: that's both ends
21:20:19saratogaall that does is try to guess the pregap if theres no gapless info
21:20:21Torneand you can only know if the encoder put it in the tag info
21:20:30Tornecrap encoders don't
21:20:42LloreanTorne: I thought the one at the end of the song could also be because MP3 pads the rest of the frame if the song doesn't end on a perfect boundary?
21:20:47saratogaso its pretty safe since essentially all modern encoders have gapless info, and in the worst case you only lose 24 ms of audio
21:20:58saratogayes theres gaps at start and end
21:21:20sideralRe playlist restart: at the very least, the playlist should not get lost on poweroff. And an option to make it restartable (in place of the "Nothing to resume" message) would be nice
21:21:22saratogathe patch removes both
21:21:27saratogaup to one full frames worth
21:21:35Lloreansaratoga: Are there other MP3 players that do this sort of thing? Perhaps a common algorithm?
21:21:57saratogafoobar does something similar, but does it as a DSP without knowledge of teh mp3 frame boundaries
21:22:06saratogaas do many winamp plugins and i think a few hardware players like the rio karma
21:22:30saratogathis is a little safer though since it cannot cut past a frame boundary
21:22:33Lloreansideral: I'd at least be very strongly in favor of not losing it at poweroff. I think the "resume" issue was the more contentious part.
21:22:57saratogaIMO that patch should not have a setting
21:23:08sideralLlorean: yes, some people never want to listen to the same stuff twice −− a recurring theme ;)
21:23:13saratogait should be always on if the file has no gapless info
21:23:14Lloreansaratoga: Is there some reason why the padding isn't absolute silence (again, I just remember someone mentioning it had a high threshold)
21:23:27saratogaLlorean: i'm not sure
21:23:30saratogamaybe rounding error?
21:23:44sideralsaratoga: agree
21:23:49Lloreansaratoga: It seems that would still be very, very, very quiet. But then, I don't know where the comment was about it being set quite loud.
21:24:06saratogaGAP_THRESH 5000000
21:24:12saratogathat actually looks pretty loud
21:24:13*linuxstb also agrees there is no need for a setting (someone should mention that on the task...)
21:24:16sideralif you don't want the heuristic to kick in, properly tag your files :)
21:24:30TorneLlorean: because the FFT parameters "stick out" of the edge of the frame that does have some sound in it
21:24:37gevaertssideral: if you don't want that gap, properly tag your files :)
21:25:15saratogaIMO we should just check to see if theres a signal above 16 bits
21:25:21saratogaif there is, keep it, otherwise toss it
21:25:29LloreanTorne: But that's no longer silence, although it may not be something you want. But you can't really safely distinguish between "not silence that we want" and "not silence that we don't"
21:25:30saratogastuff below 16 bits won't be passed to teh DAC anyway
21:25:55Lloreanlinuxstb: No setting for the gap skip?
21:25:56saratogaoh i guess thats what that is meant to do, but i think its off by one bit?
21:25:58TorneLlorean: to be honest i think *both* sides of this argument are pointless ;)
21:26:16sideralLlorean: remember that this only removes silence from the first frame. If there's silence you want, it will be in 2nd frame as well
21:26:25linuxstbLlorean: Yes. As saratoga said, it's just active if there is no lame header in the file, in which case the playback will be "wrong" anyway...
21:26:34gevaertssideral: but you'll have lost some of it!
21:26:36TorneLlorean: There's no point including it because people who are going to notice the tiny gap should jus have properly encoded files. There's no point resisting its inclusion because people with properly encoded files won't be affected by it ;)
21:26:45Lloreanlinuxstb: Yes, I agree it doesn't need a setting. Just wanted it to be explicit since we've sorta talked about a few things here in the last couple minutes.
21:26:59saratogathing is, on average even if its wrong you'll only lose 12 ms of silence, and thats not going to be audible
21:27:08LloreanTorne: I just feel it should err in favor of not throwing things we might want out.
21:27:34TorneLlorean: *shrug* It is far more likely thta someone will notice 12ms of extra silence instead of 12ms of missing silence
21:27:37sideralgevaerts: I think there's no guarantee what the decoder delay will be like, so it's a pretty safe assumption that it's unintentional if not tagged properly
21:27:52LloreanTorne: I didn't think we were throwing out a whole frame at a time?
21:28:10saratogaif the track begins with real silence, you could throw out an entire frame
21:28:15gevaertssideral: I don't see that
21:28:21saratogaof course some or all of it would have been added by the encoder
21:28:26Lloreansaratoga: If it begins with real silence, and our goal is to throw out silence, that's fine.
21:28:58 Join Keripo [0] (
21:29:29LloreanI'm just suggesting I'd like it to not be particularly aggressive about declaring "quiet, but noisy" as "silence"
21:29:35gevaertsAnd anyway, I thought the purpose of "gapless" was to make sure that if two tracks happen to not have complete silence between them there would be no audible artefacts? If you remove silence, you actually *add* artefacts in some cases
21:29:55sideralI think the goal of this patch is to find only unintentionally added silence. If it is intentional, then there will be more silence in the neighboring frame, and the loss of silence will be inaudible
21:30:19gevaertssideral: it's still more than half a second! That's audible
21:30:53gevaertsIt's the bit that you cut of that may be inaudible
21:31:34sideralDidn't someone say 24 ms per frame? 50 ms is not half a second
21:32:34LloreanDo we check for silence in the following / preceding frame?
21:32:58LloreanIf it's also silent, or very close to silent, then we'd know it might be unwise to drop the frame?
21:33:16sideralLlorean: Don't think the patch does that; nice idea
21:33:17gevaertsI don't think the exact length matters
21:33:53gevaertsIf you're arguing that a gap of a certain length is too long because it's detectable, you *can't* argue that wrongly dropping a gap of the same length is not detectable
21:34:35Lloreangevaerts: To be fair, it shouldn't be able to wrongly drop a gap in a situation where it's detectable. A sudden silence between noises stands out. But a tiny bit less silence in silence doesn't.
21:35:05*linuxstb wonders why CONFIG_CPU == 0 for RaaA
21:35:16LloreanI just want the lowest likelihood of wrongly dropping a sample that's not actually "silent", so a goal of "most silence, and no noise" rather than "all silence, and maybe the occasional bit of noise"
21:35:28gevaertsLlorean: if you're using heuristics to detect silence between (especially matching) noises, you're never going to get it right anyway, and you'll always have audible artefacts
21:36:04gevaertsAnd if the tracks are not matching, I really don't see how this short bit matters anyway
21:38:06 Quit Mikeb0ok (Ping timeout: 250 seconds)
21:38:37 Quit mystica555_ (Ping timeout: 265 seconds)
21:39:00sideralYes, this is a heuristic, but one that likely leaves listeners better off than not attempting to kill the silence artifact
21:39:19*gevaerts disagrees
21:39:44siderallooks like a candidate for a config option, then :)
21:39:52linuxstbgevaerts: Do you know why CONFIG_CPU == 0 for RaaA?
21:39:58gevaertslinuxstb: sorry, no idea
21:40:33linuxstbkugel: ping
21:40:55sideralLlorean: Re rewind on pause: did you mean it should share one set of options with rewind on stop, or have a similar but independent set of options?
21:40:59gevaertssideral: it seems it's a fixed number of *samples* per frame. I suspect that I noticed the half second when working with a 22050 Hz file
21:41:02 Quit TheSeven (Ping timeout: 265 seconds)
21:41:05kugelit's 0 for simulator and all hosted targets
21:41:56linuxstbkugel: That's what I'm reading, but why?
21:42:00sideralbut 2*
21:42:16linuxstbkugel: I can understand for the sim, but why RaaA?
21:42:26kugellinuxstb: I don't know how to answer, what would you expect?
21:42:58Lloreansideral: Similar but independent, I think.
21:43:26sideralgevaerts: 2*48 ms is still only a fifth of 0.5 s ... but perhaps noticable
21:44:02saratogamp3 frames are 1152 samples IIRC, so 1152/44100 = 26 ms (rather then 24 as I said above)
21:44:03gevaertsOh, wait
21:44:10*gevaerts sees
21:44:21linuxstbkugel: I'm trying to understand why you decided to do that, instead of defining it to be the target's CPU type. By defining it to be 0, there are no CPU-specific codec optimisations, or have you enabled those in some other way?
21:44:36gevaertsThe cutting tools I used were not per frame, but per GoP. That makes a pretty big difference...
21:44:54gevaertsSo ignore those numbers
21:44:55sideralgevaerts: oh. :)
21:45:25kugellinuxstb: I based it off sdl raaa, which is based on the sim. it wasn't not much of a decision
21:45:47saratogagory details:
21:46:02kugelI thought CONFIG_CPU hasn't a meaning on RaaA so 0 kinda fits
21:46:44gevaertsAnyway, I think my point doesn't depend on that :). If the tracks are not encoded together, you'll have audible mismatches even if you drop the exact needed number of samples. If you need heuristics, I'd say the chance of the tracks being encoded together is near zero
21:47:17sideralgevaerts: Would you also oppose making this configurable?
21:48:04linuxstbkugel: OK, thanks - I'm just trying to catch up on what's been done, as I was otherwise occupied last summer...
21:48:11gevaertssideral: I'd argue against it, and assume that whoever commits it has thought about it and not ignored the arguments :)
21:48:29gevaertsIf it's *not* configurable, I'm very strongly against it
21:48:49*linuxstb should perhaps withdraw the comment he hastily added to FS #11891...
21:49:24kugellinuxstb: you're welcome
21:49:24AlexPWe have far to many damn options
21:49:28sideralgevaerts: good to know :) fortunately I'm not qualified enough in the area to commit it, but I'd love to see this in as an option.
21:49:44*gevaerts does not want any unconditional feature that automatically drops actual audio data
21:49:47saratogawait what, i thought we agreed not to have an option
21:50:02linuxstbgevaerts: If you feel strongly, please post to the task...
21:50:10saratogai absolutely do not want options that let people configure at run time how the decoders work
21:50:25gevaertsI'm not insisting that there be a setting. I'm insisting that the feature doesn't go in without a setting. That's not the same thing
21:50:25kugellinuxstb: I guess CONFIG_PLATFORM is the nearest equivalent. IIRC I thought it about repurposing CONFIG_CPU but that'd would've been a lot of changes and I think it's cleaner with CONFIG_PLATFORM
21:51:57saratogagevaerts: what about just changing the threshold so that dropping samples only happens if they have 16 or more leading samples (and thus would produce digital silence on an actual device)?
21:52:07saratoga"16 or more leading 0s)
21:52:18gevaertssaratoga: what if I want silence?
21:52:20linuxstbkugel: I'm currently looking at all the #defines etc used in RaaA to try and see if they can be simplified in any way. I think PLATFORM and CPU are very different though - e.g. the SDL app could be built for many CPUs.
21:52:27sideralAlexP: I for one like all the options :) but I repeat myself.
21:52:27sideralWe could start cutting options, though. Replace the 40 or so crossfeed options with the Meier crossfeed from FS #11767 :)
21:52:35gevaertsSilence *is* valid audio
21:53:17saratogagevaerts: to be clear, even if the detection here is completely wrong, you're only going to lose something like 15-20 ms worth of samples here, thats going to be below the temporal response of your ear
21:53:42kugellinuxstb: I'm not sure if we're going to use CONFIG_CPU for any RaaA in the future
21:53:42gevaertssaratoga: as I said, that's a very strong argument against this feature
21:53:56saratogai don't follow you
21:54:00linuxstbkugel: Why not? How else would you enable CPU-specific optimisations?
21:54:30kugelcurrently CONFIG_PLATFORM is used
21:54:32gevaertssaratoga: if it's negligible, why do you need to drop it?
21:54:53kugelso, while the RSB is busy with autoresume we can have the next controverse feature?
21:55:05sideralgevaerts: it's negligible in length of silince, not in drop of amplitude
21:55:19saratogagevaerts: because adding silence (that you can't hear) to the middle of a track can create an artifact
21:55:30kugelgevaerts: I think he means it's negible if it's actual audio, but not if it's a gap of silence
21:55:33saratogathe reverse (removing silence that you can't hear from silence) is not true
21:55:38gevaertssaratoga: where does this middle of a track come into it?
21:55:53linuxstbkugel: I'm delaying my opinion on that until I look at things more... ;)
21:56:15gevaertskugel: silence *is* *actual* *audio*
21:56:20saratogagevaerts: if you split a recording in half and make 2 mp3s from it, the pregap (which is too short for your ear to hear) will still create an artifact in the form of a click (due to gibbs effects and such)
21:56:30kugelgevaerts: not in my book
21:56:30saratogathe idea is to avoid that artifact
21:56:33linuxstbkugel: But it seems to me that in the long-term (or even now, with the SDL app), one platform can be running on multiple CPUs.
21:56:47gevaertssaratoga: if you do that sort of thing with an encoder or tools that don't handle this properly, you're an idiot
21:57:00saratogawell a lot of people have
21:57:08saratogawhat do you care then if you're not an idiot and haven't done this?
21:57:09sideralgevaerts: I'm not an idiot for buying MP3s from Amazom
21:57:22saratogayou have non-gapless mp3s from amazon?
21:57:30 Join krazykit [0] (
21:57:32gevaertsAnd if you do that, samples won't line up anyway, so you're quite likely to get clicks
21:57:52saratogayes, but you can make the glitch much harder to notice
21:57:58saratogaand in some cases eliminate it
21:58:10sideralgevaerts: In fact, they will line up nicely in the typical case
21:58:24gevaertssideral: only if you use a lossless encoder
21:58:37saratogai mean as linuxstb point out to me years ago, theres a glitch even when you have gapless info (since the phase won't be continuous) its just subtle enough that most people don't notice it
21:58:45saratogathats really the idea here, to make the artifact harder to hear
21:58:49sideralsaratoga: Yes, I have
21:59:00saratogawhat did they encode them with?
21:59:54sideralsaratoga: Can't check right now
22:01:04sideralor wait, actually can
22:03:54*gevaerts also doesn't see why people with such broken files can't add the necessary metadata to them to fix the issue properly
22:08:14kugelthat would be the first setting that alters the behavior of a (specific) codec, isn't it?
22:08:14saratogaIMO it comes down to trying to decide which is better: decode with length you know is wrong, or decode with a length you know is wrong but use what information you have to make it less wrong
22:08:14saratogakugel: as far as I know yes
22:08:14sideralkugel: well, it would fix a deficiency that a particular format has
22:08:14saratogasideral: does foobar report them as having lame gapless info?
22:08:14sideralsaratoga: That's the thing, they don't have those tags
22:08:14DBUGEnqueued KICK saratoga
22:08:14saratogaits possible they edited the mp3 in something else that removed the lame header
22:08:14DBUGEnqueued KICK sideral
22:08:14sideralI think I've seen this in multiple albums I bought on Amazon
22:08:14 Quit Zambezi (Read error: Connection reset by peer)
22:08:14 Join Zambezi [0] (Zulu@
22:08:14sideralsaratoga: Why are you so much against a config option for this? I can understand people like gevaerts who don't want us to fiddle with their audio at any cost...
22:08:14saratogaweird, never seen it with ones i've bought, but i guess some labels are better then others
22:08:26kugelsaratoga: I don't want to say it's a problem (codecs have no gui to provide their own settings so there's no other way). I just wanted to note this world premiere :)
22:08:50CtcpIgnored 2 channel CTCP requests in 0 seconds at the last flood
22:08:50*sideral will be thrown out of this building in a few minutes
22:08:54saratogabecause I don't want settings for codecs, and because I cannot understand why people would care, its an internal decision made by the decoder not something people should know about
22:09:10LloreanMaybe a "gaplessfix" plugin that does the gap detection and adds the tags, instead?
22:09:24gevaertsThat would be the proper solution, yes
22:09:33gevaertsWell, that or a host-side tool
22:10:36LloreanSo that it's simple to do on-player still, but doesn't require either settings or questionable behaviour/heuristics.
22:10:36kugelI'm very sure if the codec had it initially by upstream nobody would argue (or even notice)
22:10:37gevaertsWe once had a bug that dropped a few miliseconds at the end of playback. We *did* notice eventually
22:11:01kugelI'm indifferent about a setting; on the one hand it wouldn't be noticeable, but on the other hand people that have proper gapless shouldn't suffer
22:11:43gevaertsIf files are broken, fix the files
22:11:43kugelyours is a bug, but we talk about a feature :P
22:11:45sideralCatch up with you later. Security is comingfqwpofhn%R56rR&%(5AR NO CARRIER
22:11:47saratogaif we can't agree on how to handle this, we should just ignore it
22:11:51 Quit efyx (Quit: Quitte)
22:12:42linuxstbLlorean: I guess an option in vbrfix...
22:13:39saratogai dislike the idea of adding fake gapless tags to mp3 files
22:13:53saratogasince they're just estimates, but they make it look like the file is actually gapless . . .
22:14:41kugelso they potentially don't even remove the gap?
22:15:41saratogathey probably won't get it exactly right, so theres likely to be a few samples error
22:15:42gevaerts"I dislike the idea of detecting gaps in mp3 files, since they're just estimates, but they make it look like the file is actually gapless"
22:15:44 Quit GeekShadow (Remote host closed the connection)
22:15:53saratogahey thats my post
22:16:01saratogagive it back
22:16:03gevaertsNot entirely :)
22:16:35saratogagapless is always only approximate anyway, since even if you have the tag you can still get a slight artifact, but at least if you have the tag you know the length is right
22:16:47saratogaif you just make up a tag thats kind of ugly
22:16:53 Quit sideral (Ping timeout: 276 seconds)
22:16:59gevaertsWell, if the encoder knows about it, it's *possible* to have it artefactless
22:17:09saratogai'm not sure it is actually
22:17:27saratogabecause i don't think you can make the last and first samples continuous
22:17:38gevaertsYou don't have to
22:17:51saratogawell you do if you want to have no possibility of a glitch
22:17:54gevaertsYou can still use the tags to skip samples
22:18:02saratogayeah but that doesn't mean there won't be a glitch
22:18:19gevaertsSo if you can match halfway the last frame and halfway the first frame (more or less), you're fine
22:18:29saratogayou can get the number of samples exactly right and still have a problem if the last sample of the first track doesn't line up with the first sample of the second track
22:18:53saratogaplus its ackward to make filterbanks give the right value for the very first sample
22:19:04saratogai'm not even sure if the mp3 spec does teh right thin in this regard
22:19:13 Quit [Saint] (Disconnected by services)
22:19:15 Join S_a_i_n_t [0] (S_a_i_n_t@
22:20:12saratogabut regardless, i think the real question is how should a decoder responsibly handle a file if its not given gapless info, should it use what it knows to reduce the error (e.g. try and detect the encoder delay) or should it just give up and assume the user doesn't care?
22:20:25gevaertsSure, but these tags say to drop the first (or last) N samples, so if you know you can't get the first or last 100 samples right, use whatever encoder magic to be right starting at the 100th sample, and set the tag
22:21:05gevaertsI'm not saying this is easy to do :)
22:21:08saratogayeah but how do you make the samples have a specific value? mp3 only encodes the general energy per subband, not the exact time domain value
22:21:44saratogai'm not even sure if its possible to exert that level of control over samples in the time domain (should probably ask Buschel about that actually)
22:21:48gevaertshm, does this depend on the decoder?
22:22:11saratogaprobably not, at least not for a decoder thats going for full accuracy decoding
22:22:33gevaertsThen in the worst case you could just try again until it looks right
22:23:32saratogabut if you can't control where a given sample falls in the filterbank, and you only have 2 filterbank sizes to choose from, how do you set a specific sample to a specific value using only the (limited) quantization levels available to you?
22:24:04gevaertsAh, that's where the exercise for the interested reader appears :)
22:24:19saratogaiteratively encode and decode the file until you manage to minimize the error (and hope that theres even a solution)?
22:24:45saratogaand how do you tell the next file you encode what the solution you picked was
22:25:10 Join mudd1 [0] (
22:25:17saratogayou probably can't just pick zero everytime, since then you're introducing another artifact by forcing 0 samples that shouldn't be there
22:25:36gevaertsYou might have to encode the entire album as one unit
22:25:48saratogathats what the nogap feature in lame does
22:26:05saratogait probably does work a little better, but it can be ugly since it messes up the track lengths a bit
22:26:30 Quit stoffel (Remote host closed the connection)
22:26:39evilnick_Bgevaerts: Yes, everyone should use .cue files
22:27:07saratogai think in the end encoders and decoder just do as well as they can and usually no one noticing if theres some minor error
22:27:23saratogawhich is in keeping with the general spirit of the patch i think
22:31:56linuxstbkugel: Is this change correct? It looks like it would prevent Rockbox doing any housekeeping when shutting down?
22:32:45 Quit mudd1 (Ping timeout: 265 seconds)
22:35:29 Join mudd1 [0] (
22:36:29 Join TheSeven [0] (~TheSeven@rockbox/developer/TheSeven)
22:37:41kugellinuxstb: thomasjfox is on this, I would talk to him
22:38:01kugelthat change is part of his current state
22:38:36linuxstbWhat do you mean? That change is from your commit 7 months ago (sorry, I'm a bit late reviewing it... ;) )
22:39:20kugelah uh, then reverting that change is part of his current state :)
22:47:04 Part coc00n
22:59:20 Join Suit_Of_Sables [0] (
23:01:15 Quit evilnick_B (Quit: Page closed)
23:01:54Suit_Of_SablesHaving a hard time mounting my sansa fuve v2 under linux. I use 'sudo mount -f /dev/sdc /mnt/fuze' which doesn't return any error and repeating the command tells me it is already mounted there. but /mnt/fuze is empty. I chowned it so It was writable and with the mount point and device set in rbutil the installation seemed to go fine. even had the .bin for the original bootloader. but after running rbutil my fuze is no longer mount
23:02:24 Quit esperegu (Read error: Connection reset by peer)
23:04:23linuxstbSuit_Of_Sables: Why "-f" ?
23:04:47 Quit mudd1 (Ping timeout: 250 seconds)
23:05:24 Quit liar (Read error: No route to host)
23:06:17 Join liar [0] (
23:06:48Suit_Of_Sableslinuxstb: otherwise I get 'unknown file system type 'vfat' '
23:07:25linuxstb"-f" fakes the mounting according to the manpage, so it doesn't actually mount it.
23:07:34TorneYeah, -f is not force, it's fake :)
23:07:43BagderSuit_Of_Sables: "grep vfat /proc/filesystems" gives?
23:07:44linuxstbYou probably want /dev/sdc1 (but I don't know the fuze's disk layout)
23:07:58linuxstb(or not... maybe ignore me)
23:08:37Bagder"unknown file system" sounds like a bad thing
23:08:40 Join mudd1 [0] (
23:08:48Torneeither you're mounting the wrong device or your kernel deosn't have vfat support
23:08:55Tornewhat linux distribution is this?
23:08:59Suit_Of_SablesBagder: no output
23:09:02Suit_Of_Sablesarch llinux
23:09:07Tornedid you configure your own kernel?
23:09:09BagderSuit_Of_Sables: you don't have vfat support then
23:09:22Bagderadd it, rebuild, reboot and retry
23:09:37Bagderor possibly load the module
23:09:56 Join stripwax [0] (
23:10:51AlexPI can confirm that (perhaps not surprisingly) Arch does have vfat suport by default :)
23:11:06Tornetry "modprobe vfat"
23:12:45Suit_Of_SablesTorne: not there
23:13:02Tornedid you build your own kernel?
23:13:09Suit_Of_SablesI installed the dosfstools package which is supposed to add vfat support...
23:13:13Suit_Of_Sablesnot my own kernel
23:13:26TorneThen it sounds like your system might be broken in a rather unfortunate way
23:13:32Torneperhaps your modules have gone missing :)
23:13:48Tornedosfstools is a bunch of userspace tools for managing FAT filesystems
23:13:58Tornethe module to actually mount one is part of the kernel
23:14:16 Join DerPapst [0] (
23:15:35Torneif it's neither built into the kernel nor available as a module, and you have a regular distribution packaged kernel, then i would guess something bad has happened, since it's pretty much unthinkable at this point that anyone would ship a kernel without FAT
23:16:20AlexPThey definately do ship FAT support in the kernel, so something else is buggered
23:16:35AlexPwell, modules
23:16:42AlexPIt is there anyway :)
23:17:10Torne(have you *ever* been able to mount USB devices formatted as FAT?)
23:17:28Torne(and have you done anythig interesting to the machine since last time you did? :)
23:18:21***Saving seen data "./dancer.seen"
23:18:51AlexPSuit_Of_Sables: try "pacman -Ql kernel26 |grep fat"
23:19:11AlexPIt should show any files containing "fat" in the kernel package
23:19:17 Quit Suit_Of_Sables (Ping timeout: 240 seconds)
23:19:22AlexPor don't :)
23:24:39 Join Suit_Of_Sables [0] (
23:24:57Suit_Of_Sableshmm so my kernel must have changed because after a reboot I can mount vfat
23:24:58 Join mlt- [0] (
23:25:07Suit_Of_Sablesbut rbutil still can't auto detect my device
23:25:17Torneare you running it as root?
23:25:46mlt-I wonder if anyone tried to build UIsimulator under MS Windows recently?
23:26:52AlexPnatively you mean?
23:26:56 Quit domonoky (Read error: Connection reset by peer)
23:27:02mlt-yes, not cross from linuyx
23:27:08AlexPrasher provides cross compiled sims for Windows
23:27:13AlexPI doubt it
23:27:51rasherLogic says it should be possible
23:28:15Tornemlt-: do you ask because you tried and it didn't work?
23:28:20Torneor because you are worried that it might not work?
23:28:24Tornebecause if the latter then it probably will
23:28:28Torneso just try it
23:28:35Torneand if the former then, er, tell us what happened :)
23:28:54mlt-I hardly see any reason to cross compile simulator from linux. AFAIK most ppl use it for development and debugging and not just to play with it, so it makes sense to build and debug on host system.
23:29:22TorneMost people use it for testing themes, i suspect
23:29:22mlt-AlexP, Torne: Yes, it didn't work for me. I added some notes into Wiki. And I wonder if anyone tried )
23:29:35Tornemost people who use it for development and debugging use linux as their host system
23:29:40mlt-Torne: oh.... that is a good reason. I forgot about it
23:30:01Torneso it doesn't surprise me that nobody has tried for a while :)
23:30:35Tornetheme developers don't build it, and rockbox developers rarely use windows :)
23:31:18Suit_Of_SablesTorne: I haven't been running it as root. I can a chown on the mount point for this user but I'm being told now I can't write to it
23:31:21rashermlt-: are you trying to build in cygwin with the regular gcc, or using mingw32 to get a (more?) native binary?
23:31:38rasherNot sure what the proper way to express this is..
23:31:51mlt-mingw gcc
23:32:25TorneSuit_Of_Sables: i mean running rbutil as root
23:32:56rashermlt-: Yeah, I wouldn't be surprised if the configure script just isn't ready for that
23:33:16mlt-I don't have linux at work :( that is why I tried to build under windows so I can just debug from Eclipse CDT under Windows.
23:33:21mlt-configure script is fine
23:33:35Tornejust becuase it doesn't fail doens't mean it configured it correctly :)
23:34:04mlt-there are other problems
23:34:36rasherI think you may be the first to try this
23:34:47rasherIn fact, I'm fairly sure you are
23:35:09mlt-oh.. well.. Here is my experience
23:35:28rasherYeah, I'm looking at it
23:35:34kugelrasher: nah I gave it a quick try ages ago
23:35:46rasherShows what I know
23:36:07kugelI didn't make it public
23:40:47mlt-I guess most of the stuff can be cleaned up, but I'm not sure why some make targets don't build as mentioned at the bottom of wiki page. Perhaps not all dependencies are generated correctly.
23:42:02rasherUsing the alternate perl could be done by just setting PATH, right? Or do you only want to use the alternate one in that one place?
23:42:48mlt-strawberry perl comes with its gcc as well
23:43:06mlt-I don't want to override those
23:43:43mlt-though... it is in separate folder :-)
23:43:53mlt-yeah.. i'll correct that
23:44:54Suit_Of_Sablesugh... well I ran rbutil as root and it installed to the fuze. after the 'firmware updating' all seemed well. but now my machine doesn't see it when I run fdisk -l
23:46:55mlt-rasher: lang.make still has to be altered a bit... genlang uses #!/usr/bin/perl .... if perl was used explicitly in lang.make , then path override for perl would work
23:47:58rasherAh right, we're out of luck then. If only /usr/bin/env wasn't so goddamn useless (it won't pass arguments... what the hell guys?)
23:49:02rasherWell genlang needs /usr/bin/perl -s right?
23:49:36rasherotherwise we could do #!/usr/bin/env perl (which would use whichever perl was first in PATH)
23:51:15linuxstbkugel: You know Android isn't just ARM?
23:51:48kugelthere's no official non-ARM distribution AFAIC
23:52:04mlt-rasher: oic... but we could just use perl explicitly in lang.make
23:52:11linuxstbWhat does "official" mean? It's open-source...
23:53:12rashermlt-: you're right
23:53:15kugelAOSP only supports ARM currently
23:54:49kugelI tried an x86 port in VM recently and it's not very usable
23:57:26 Quit bertrik (Quit: :tiuQ)

Previous day | Next day