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 2009-06-23

00:00:54Zagoryes that would be a lot better
00:01:08rasherAny reason to not put runclient and rbclient in tools (or some other place in trunk)?
00:01:19rasherwould make the barrier to entry even smaller
00:01:42Zagorthe reason was simply that we didn't want to trigger builds for every commit...
00:02:05Bagderwe can put it there later on
00:02:15linuxstbIsn't there a script to prevent that?
00:02:15gevaertsalso, this shouldn't be very rockbox specific. It could be useful for other things
00:02:34rasherMakes sense. With the added bonus that it ensures things are in the right location
00:02:40 Quit funman ("free(random());")
00:02:53kugeljust wanted to talk to him
00:03:06Bagderwell, it might not be that wise to always run the script off svn though
00:04:37rasherI suppose
00:07:25 Quit __lifeless (Read error: 104 (Connection reset by peer))
00:07:42 Join __lifeless [0] (n=lifeless@
00:08:16 Quit evilnick ("Page closed")
00:08:54 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
00:09:45CIA-71New commit by zagor (r21471): Uh, don't run waitpid non-blocking... This should fix the "leftover build dirs" bug.
00:10:47rasherIronically, I now have a leftover build dir
00:12:17CIA-71New commit by bagder (r21472): match build names exactly, not just prefix, and added some basic logging
00:15:25 Join saratoga [0] (i=9803c6dd@rockbox/developer/saratoga)
00:15:53 Join n00b [0] (n=1824bf02@gateway/web/cgi-irc/
00:17:22CIA-71New commit by rasher (r21473): We don't want this.
00:18:16*JdGordon| wonders what we dont want
00:19:11 Quit LambdaCalculus37 ("Fwump")
00:19:52*kugel hands JdGordon| a rXXXXX parser
00:19:57 Join Hillshum_ [0] (n=chatzill@unaffiliated/hillshum)
00:20:08 Quit icebrian ("Leaving")
00:20:39ZagorI'm off to bed, and tomorrow morning I'm heading to greece for a week. I'll be back here july 1st. Have fun!
00:20:49gevaertsHave fun!
00:21:08 Quit Zagor ("Clint excited")
00:21:53 Quit n00b ("CGI:IRC (EOF)")
00:25:01CIA-71New commit by kugel (r21474): FS #10284 - "Sansa e200v2 & Fuze: Merge the button driver into 1 file" by Dustin Skoracki ...
00:28:14 Part CaptainKwel
00:39:07 Quit bmbl ("Bye!")
00:41:24CIA-71New commit by kugel (r21475): Actually adapt SOURCES too.
00:42:01*rasher commends kugel on not using "fix red" as the commit message for that
00:42:34Mikachu"fix red" is a pretty bad commit message :)
00:42:57rasherI twitch a little each time it happens
00:43:14 Quit efyx (Remote closed the connection)
00:44:26Mikachu0-length pipe msg from 4!
00:44:33Mikachuoops, zagor left :)
00:44:58 Part wincent ("Kopete 0.12.7 :")
00:45:51***Saving seen data "./dancer.seen"
00:46:08Bagderyeah, zagor did most of the client parts and I did most of the server parts
00:46:14kugelbad news: I have build-* folders left
00:46:23Mikachukugel: with the update?
00:46:33kugelit's supposed to autoupdate
00:46:35Mikachuie r21471
00:46:55Mikachucheck if it didn't autoupdate or if the fix didn't work :)
00:46:56 Quit agaffney (Read error: 104 (Connection reset by peer))
00:47:15MikachuBagder: i guess that just means the socket was closed?
00:47:22kugelindeed, the timestamp is quite old
00:47:47BagderMikachu: I didn't check that yet
00:47:59 Quit shotofadds ("Leaving")
00:48:33rasherI'm guessing it didn't autoupdate?
00:48:41rasherrevision 6 here
00:48:42Bagderit didn't
00:48:44kugelBagder: the autoupdate didn't work this time
00:48:55Bagderwell it doesn't just update on any change
00:49:12kugelI see, so someone forgot to trigger it?
00:49:24Bagderjust ignored it I'd say
00:50:05Hillshum_the "Project News" thing on the homepage should be updated to state that devcon is over
00:51:08bertrikgevaerts, since all meizus use the same pin for the backlight, I think I can make a backlight pwm driver that is common for all meizus. How about I call it firmware/target/arm/s5l8700/backlight-meizu.c ?
00:51:21gevaertssounds goof
00:51:25gevaertseven good
00:52:05 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
00:52:17kugelI wish the AMSes had still backlight issues, so bertrik was looking at them again :(
00:52:56 Join stripwax [0] (
00:53:05saratogawheres are since 3.2 SVN link?
00:53:17 Join Jmclark [0] (
00:53:18saratogaor perhaps a since 3.2 and a since 3.3 one
00:53:25bertrikthere's just very few people owning a meizu, but the documentation is pretty good. It's the other way around with the ams sansas ...
00:53:42saratogayou mean the AMS docs?
00:54:03 Quit tessarakt ("Client exiting")
00:54:21saratogaalso, i'm thinking of posting candiate bootloaders for the AMS targets, I don't expect we'll need to change anything before supporting them, and I dislike the idea of encouraging people to build bootloaders for targets, even if it should be safe
00:54:35 Quit nibbler_ (Read error: 110 (Connection timed out))
00:54:56bertriksaratoga, yes, well they mostly ok, but e.g. the codec communication part was hard to figure out and we're still not sure about USB
00:55:07 Quit Jmclark (Client Quit)
00:56:12kugelsaratoga: there's this sd driver issue...
00:56:55saratogakugel: does that matter for the bootloader?
00:57:23kugelno, but it's still a blocker IMO
00:57:35kugelthough I didn't experience any SD problems so far
01:01:21 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
01:01:31saratogaso you want to wait until thats fixed to post canidate bootloaders?
01:02:52kugelavailable candidate bootloaders seem to me like: we're releasing soon if we don't find a major issue. We know a major of a issue
01:03:05kugelWe know of a major issue*
01:03:25gevaertskugel: you don't have to announce them officially. Just upload them to the wiki
01:03:33saratogapatch tracker
01:03:47saratogawe've had the canidate PP bootloaders on their for 3 or 4 months now
01:04:03 Quit jgarvey ("Leaving")
01:05:10kugelis mkamsboot already fixed?
01:05:38Hillshum_kugel: it just worked for me, what was the problem?
01:05:48kugelwe should at least also make it 1.1 and put it on again if we're having bootloaders available
01:06:14kugelHillshum_: the version which was once online accepted unknown OF versions, which we considered bad
01:06:32Hillshum_not sure if it was fixed
01:07:00CIA-71New commit by dave (r21476): Add an explicit rule to compile bin2c - r21272 broke compilation of sansapatcher on at least Linux because Make's implicit rule for compiling bin2c ...
01:07:25kugelr21189 fixed that
01:07:45 Quit bertrik ("De groeten")
01:08:11linuxstbgevaerts: I don't like the idea of binaries in the wiki (where they can be changed by anyone).
01:08:24gevaertsgood point
01:08:28 Quit jordan` (Read error: 60 (Operation timed out))
01:09:02 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
01:09:09linuxstbDon't we have a "testing builds" forum now?
01:09:14 Join ash002 [0] (
01:09:45linuxstbAlthough the description says "When a new feature needs user testing in a custom build before going into SVN, a thread will be started here."
01:09:58kugeldoes the fuze manual already have the installation instructions?
01:10:17kugelI think we could actually put something into that test forum
01:10:59 Join jordan` [0] (i=gromit@
01:11:20 Nick ash002 is now known as ash (
01:11:26kugelMaybe change the description (like adding "or new targets in the pre-release phase")?
01:11:50 Nick ash is now known as Sarang (
01:12:17 Quit jordan` (Read error: 104 (Connection reset by peer))
01:12:30SarangHi. I was reading through the chan logs, and I noticed that C200V2 is almost useable... thanks to funmen
01:12:46SarangI would love to compile/test it out...
01:12:55kugelHe'll gladly read that :)
01:12:59Hillshum_though were all fun
01:13:12SarangBut am hesistant for c200v2 not having any recovery mode (yet).
01:13:27saratogatheres probably no recovery mode for the c200v2
01:13:38linuxstbSarang: It's probably as safe now as it will ever be...
01:13:55SarangI rockboxed my E200v1 already.. and bricked numerous times.. and including overwriting the recovery partition twice..
01:14:16linuxstbBy definition, you can't brick something numerous times...
01:14:29SarangBut thanks to Dan Haxxes? I was able to use linux +firmware uploader to erestore partitions.
01:14:43SarangHas anyone here bricked C200v2 yet?
01:14:56saratogai don't think so
01:14:58Sarangsorry English is not my native language
01:15:03 Quit BeChris (Ping timeout: 180 seconds)
01:15:08Sarangic... I just have bad luck with bios/firmware upgrades...
01:16:23SarangThe secret mode on c200v2.. does anyone know what it actually do?
01:16:39saratogaif you overwrote the firmware partition, it sounds like you didn't use our installer
01:17:03 Quit lilltiger (Read error: 60 (Operation timed out))
01:17:28SarangOn E200v1, I accidentally formated the recovery partion...
01:17:33Sarangbut it was recoverable
01:17:59kugelthat can't happen on the c200v2
01:18:00SarangThere was this utility that lets you create new blank partitions through USB connection in linux
01:18:18Sarangkugel, so not possible to format the recovery partition?
01:18:24 Quit BlakeJohnson86 (Remote closed the connection)
01:18:27kugelthere's no such partition
01:18:47kugelDon't use such abbreviations please.
01:19:22kugelNot here at least, we're having guidelines on proper English (as proper as you can)
01:19:26SarangKugel, btw, nice work on your previous custom firmware builds.
01:19:43SarangI mean Okay
01:19:52kugelthanks, but that's past
01:20:17 Join BlakeJohnson86 [0] (
01:20:32Sarangkugel, what are you working on?
01:20:45SarangI notice a lot of AMS units are now supported.. esp. Fuze...
01:20:49SarangI see some where on View asd well.
01:21:04SarangI mean some work on Sansa View..
01:21:14kugelWhere do you have that information from (because it's wrong)?
01:22:17linuxstbSarang: "supported" has a specific meaning in Rockbox - the AMS devices are not supported.
01:22:40SarangRockbox-dev newgroup.. Saturday post.
01:23:05SarangSorry. Some initial works (like working lcd) on most AMS units, and now Sansa view
01:24:18Hillshum_BTW, do you mean the email list or an actual newsgroup somewhere?
01:24:26 Quit HellDragon (Read error: 104 (Connection reset by peer))
01:24:37SarangActual newsgroup
01:25:00 Quit JdGordon| ("Miranda IM! Smaller, Faster, Easier.")
01:25:18linuxstbSarang: That's just the rockbox-devel mailing list.
01:25:32SarangI see.
01:26:31Hillshum_Does the clip have album art?
01:28:10 Join Sen [0] (n=4b8ff2ac@gateway/web/cgi-irc/
01:28:48 Quit Sen (Client Quit)
01:29:50 Quit Thundercloud (Remote closed the connection)
01:32:03 Join HellDragon [0] (
01:32:18linuxstbHillshum_: No
01:33:28kugelhillshum: only in pictureflow
01:33:36kugelthat's not really worth it though :)
01:38:13 Join JdGordon| [0] (i=ae914f7d@rockbox/developer/JdGordon)
01:39:16 Quit kugel (Remote closed the connection)
01:44:57JdGordon|rasher: why is english.lang being selected if no language is set? seems unnessacary and could lead to misunderstandings
01:46:09JdGordon|also, none of that will work if the used file (.fnt, .wps etc) is not in the expected folder
01:46:44 Join waran [0] (
01:49:39rasherJdGordon|: No language set means (as far as I know) first run, in which case the language is in fact english.
01:49:57 Quit Hillshum_ ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
01:50:47rasherAnd yes, if the active file is not in the relevant dir, it won't work, and it will simply pick the first file. I don't think that's so terrible
01:51:43JdGordon|yeah the 2nd point was just a fyi.... english.lng is not the language used if none is set... the internal one is and they *could* be different
01:51:59 Part heftig
01:52:03JdGordon|sure it wont happen often, but I dont see the need to special case that one
01:52:38rasherIt can't be different unless people have modified Rockbox
01:52:53rasherIn which case this is just something else they should modify
01:53:11rasherJdGordon|: The special case is there so that the language selection is correct on first run
01:53:57JdGordon|anyway, its not really a big deal, just I dont see the need for it.. but ok
01:54:19rasherWhy not? If the user has not selected a language (ie. first run), arabic will be selected, which is clearly wrong
01:55:12JdGordon|which is why I tihnk it would be better to change the icon for the item instead of the selection
01:55:21rasherAnd we *know* in that case that the builtin language is used. And we *know* that that is English.
01:55:52JdGordon|someone could change their build to use a different one as the base
01:56:14rasherIf they go to that lengths, that's not our problem at all
01:56:46rasherChanging the icon won't work if icons are turned off by the way
01:56:47JdGordon|I know.... ok...
01:57:01SarangPersonally, I would like to strip GUI on Rockbox of a lot of more advanced features.
01:57:45krazykitSarang, the code is available. you're more than welcome to edit it as you like
01:57:47Sarangor Hide them, as it is easy for non-technical users to mess up the configuration.
01:58:31SarangFork it and create a "rockdummies"?
02:00:28 Quit JdGordon| ("Page closed")
02:03:16SarangOr is it possible to hide the features through themes?
02:03:23 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
02:09:28waranWhile the topic rests on themes: I made some minor changes to the old Clix_Brownstone to make it run under newer Rockbox-Versions (Progress-Bar was slightly off). However the tick-marks used by the A-B-repeat-mode suffer from heavy graphical glitches (they are basically huge [top of the bar down to the corner of the window] inverted triangles. When both ticks are set they are displayed as thin inverted lines)
02:10:35waranHas anyone an idea how to fix it / what could cause these glitches?
02:10:42 Join jordan` [0] (i=gromit@
02:12:39 Quit ze ("drive installation")
02:15:22 Quit martian67 (Read error: 54 (Connection reset by peer))
02:16:21 Join martian67 [0] (n=xP@2001:470:b:356:221:91ff:fe8c:d8a7)
02:22:37 Quit martian67 (Read error: 54 (Connection reset by peer))
02:23:07 Join martian67 [0] (n=xP@2001:470:b:356:221:91ff:fe8c:d8a7)
02:25:11 Quit martian67 (Read error: 54 (Connection reset by peer))
02:25:42 Join martian67 [0] (n=xP@2001:470:b:356:221:91ff:fe8c:d8a7)
02:27:04JdGordondoes anyone have an idea what to call the actual logic behind the wps system? or a better name to call the actuall wps screen?
02:27:37stripwaxwhat's wrong with calling the wps screen the wps screen?
02:28:17krazykit"while playing screen screen"?
02:28:23*JdGordon slaps krazykit
02:28:26stripwaxhm, good point
02:28:44stripwax'the playing screen' ?
02:28:57Mikachuwhy not call it "the wps"?
02:29:01JdGordonI want to move the wps specific code into apps/gui/wps<or something> and have the actual screen code seperate from the buisness logic of the wps engine
02:29:23JdGordonthinking ahead to it being used by other screens like fm and maybe rec
02:30:19*JdGordon actually contemplates using it for drawing simple screens like the top of the time/date screen... save a bunch of code by pushing that logic off to code that exists already
02:30:33stripwaxok so there's the playing screen, the recording screen, the radio screen; all implemented using the "wps logic(tm)" ?
02:31:07stripwaxadmittedly the wps acronym no longer makes sense in that context but it doesn't really need to be renamed
02:31:26SarangJDGordon, that's a great idea!
02:31:53SarangDoing what you suggest would enable a rockbox GUI for technically challenged..
02:31:57 Quit robin0800 (Remote closed the connection)
02:32:26 Join robin0800 [0] (
02:32:28JdGordonSarang: right, execpt it would be hard coded :) and a pipe dream which probably wouldtn ever happen
02:32:31 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/freenode/x-7081e96a4ee27d27)
02:32:37stripwaxcould the main menu even be implemented as a wps-logic-enabled screen? <ducks>
02:32:46JdGordonin theory anything could....
02:33:18 Quit faemir (Remote closed the connection)
02:33:25*stripwax wonders what a configurable wps tag for graphical lists would look like..
02:33:28 Quit BHSPitMonkey (Remote closed the connection)
02:33:30SarangJdGordon, i was thinking stripping out more complex config settings and put that into advanced setting page...and make it harder to access
02:33:52JdGordonbut I also want to have the a very clear division in the related files, I want apps/gui/wps/ and in that parser.c, interpreter(?).c and tag_values.c or something so its obvious what each does
02:34:10JdGordonright now gwps-common.c and gwps.c are a mix of the last 2 and its a nightmare
02:34:35*Mikachu agrees there, took an hour to make it always draw the statusbar in a specific color while in the wps and the theme one outside
02:34:47 Quit stripwax (Read error: 104 (Connection reset by peer))
02:35:12Sarangalso, a simple setting page with shuffle, simple equalizer, etc.
02:37:42 Quit martian67 (Read error: 60 (Operation timed out))
02:39:10 Join martian67 [0] (n=xP@2001:470:b:356:221:91ff:fe8c:d8a7)
02:40:30 Join robin0800_ [0] (
02:42:00 Part toffe82
02:43:38 Join stripwax [0] (
02:44:16 Quit stripwax (Client Quit)
02:44:30 Quit JdGordon ("Leaving.")
02:45:55***Saving seen data "./dancer.seen"
02:46:55waranOk, my little A-B-problem is fixed. Seems like somebody set a ridiculously huge pb-height in the wps file -.-
02:48:32 Join webguest07 [0] (n=befe8501@gateway/web/cgi-irc/
02:49:38 Quit martian67 (Read error: 60 (Operation timed out))
02:49:59 Quit webguest07 (Client Quit)
02:50:56 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
02:57:15 Quit robin0800 (Read error: 110 (Connection timed out))
03:00:24 Quit robin0800_ (Remote closed the connection)
03:04:38 Join _lifeless [0] (n=lifeless@
03:04:55 Quit waran (Read error: 104 (Connection reset by peer))
03:13:00 Quit hillshum ("Leaving")
03:13:57 Nick fxb is now known as fxb__ (
03:15:43 Quit Chesteta (Read error: 60 (Operation timed out))
03:18:01 Quit __lifeless (Read error: 110 (Connection timed out))
03:18:04 Join AlexP_ [0] (n=alex@rockbox/staff/AlexP)
03:19:22 Quit AlexP (Read error: 104 (Connection reset by peer))
03:28:35 Join Llorean [0] (
03:37:12 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
03:37:43 Join lilltiger [0] (n=lilltige@
03:46:44 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
03:46:45 Quit amiconn (Nick collision from services.)
03:47:02 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
03:47:12 Quit pixelma (Nick collision from services.)
03:47:13 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
03:47:32 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
04:06:30 Quit BlakeJohnson86 (Remote closed the connection)
04:08:55 Join BlakeJohnson86 [0] (
04:12:16 Join kperri [0] (i=18ac0c41@gateway/web/freenode/x-bfc7a42991a6bc21)
04:15:19 Quit kperri (Ping timeout: 180 seconds)
04:15:54 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
04:17:40 Join ash003 [0] (n=ash004@
04:19:29 Quit Sarang (Read error: 110 (Connection timed out))
04:25:17 Nick ash003 is now known as Sarang (n=ash004@
04:28:50 Quit Sajber^ (Read error: 104 (Connection reset by peer))
04:29:21 Quit Sarang ("Leaving")
04:32:37 Quit kkurbjun (Read error: 110 (Connection timed out))
04:33:31 Join Sarang [0] (
04:45:56***Saving seen data "./dancer.seen"
05:41:15 Quit Horscht ("Verlassend")
05:58:21 Join evilnick_home [0] (
06:08:28 Join evilnick_home1 [0] (
06:08:33 Join kkurbjun [0] (n=kkurbjun@rockbox/developer/kkurbjun)
06:14:25 Quit jkl (Read error: 110 (Connection timed out))
06:17:18 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
06:17:54 Join Xerion_ [0] (
06:19:49 Quit kadoban (Read error: 104 (Connection reset by peer))
06:20:09 Join kadoban [0] (
06:27:38 Quit evilnick_home (Read error: 113 (No route to host))
06:34:37 Quit Xerion (Connection timed out)
06:34:38 Nick Xerion_ is now known as Xerion (
06:35:39 Quit advcomp2019 (Read error: 110 (Connection timed out))
06:39:56 Join antitrons [0] (n=Mudkips@
06:40:02 Quit antil33t (Nick collision from services.)
06:45:59***Saving seen data "./dancer.seen"
06:56:35 Join n1s [0] (n=n1s@rockbox/developer/n1s)
06:57:14 Quit jfc (Read error: 104 (Connection reset by peer))
06:58:04JdGordonwe need to get a date for the dev con.... its already $340 seattle<->nyc... dont want to pay too much more for flights...
07:00:16 Join jfc [0] (
07:06:14 Quit jfc (Read error: 104 (Connection reset by peer))
07:06:45 Join __lifeless [0] (n=lifeless@
07:06:53 Join jfc [0] (
07:07:05 Quit jfc (Read error: 104 (Connection reset by peer))
07:11:57 Join jfc [0] (
07:17:34 Nick fxb__ is now known as fxb (
07:19:37 Join ze [0] (i=ze@
07:23:12 Quit _lifeless (Read error: 110 (Connection timed out))
07:29:08 Join einhirn [0] (
07:36:55 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
07:46:28 Quit kkurbjun (Read error: 110 (Connection timed out))
07:47:50 Quit n1s (Read error: 60 (Operation timed out))
07:50:43 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:50:46 Quit timc (Read error: 110 (Connection timed out))
07:53:04 Quit jordan` (Read error: 54 (Connection reset by peer))
07:53:42 Nick fxb is now known as fxb__ (
07:54:44 Quit kachna (Read error: 113 (No route to host))
07:54:47 Quit BHSPitLappy (Remote closed the connection)
07:57:31 Join jordan` [0] (i=gromit@
08:04:47 Join jkl [0] (
08:20:01 Join planetbeing [0] (
08:20:20 Join Rob2223 [0] (
08:21:11 Quit Sarang (Read error: 60 (Operation timed out))
08:22:51 Join soap [50] (n=soap@rockbox/staff/soap)
08:23:00 Quit BHSPitMonkey (Remote closed the connection)
08:23:46 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
08:31:23 Join CaptainKwel [0] (
08:33:13 Join ender` [0] (
08:37:25 Quit Rob2222 (Read error: 110 (Connection timed out))
08:39:16 Quit CaptainKewl (Read error: 110 (Connection timed out))
08:42:14 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
08:42:29 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
08:43:29 Quit martian67 (SendQ exceeded)
08:44:23pondlifeBryanJacobs: If you read this, another way to look at buffering (my original intention) is that it's a layer over file i/o. i.e. it pre-reads data from which ever files you bufopen(). Think SMARTDRV on DOS! This isn't quite how it's ended up, but might be a useful way to look at it...
08:44:38 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
08:45:17pondlifeI think it's inevitable that it will end up with non-contigiuous space, but that's why we have the guard buffer and memcpy.
08:46:01***Saving seen data "./dancer.seen"
08:46:29pondlifeCodecs and other users can only assume one guard buffer (32K?) will be returned. i.e. a read of less that you wanted doesn't indicate an error or EOF.
08:46:43 Join bertrik [0] (
08:47:14pondlifeAlso, buffering ought not to be metadata specific really, but I understand why it's had to have some specialties added.
08:51:22 Join flydutch [0] (
08:51:27 Quit martian67 (Read error: 104 (Connection reset by peer))
08:51:55 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
08:52:16 Join timc [0] (n=aoeu@
08:59:08 Quit planetbeing ()
09:01:42 Join AndyI [0] (i=AndyI@
09:03:05 Quit martian67 (Read error: 104 (Connection reset by peer))
09:05:08 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
09:06:19 Join petur [50] (n=petur@rockbox/developer/petur)
09:06:46 Part safetydan ("Leaving.")
09:07:01 Join Grahack [0] (
09:10:39 Join tmzt_ [0] (
09:14:19 Quit AndyIL (Read error: 110 (Connection timed out))
09:16:29GodEatercool_walking_: is that your first official Expert response ?
09:20:38 Quit martian67 (Read error: 104 (Connection reset by peer))
09:21:09cool_walking_Why, is it broken?
09:22:23 Quit flydutch ("/* empty */")
09:22:25 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
09:23:40GodEaterno it's great :D
09:23:55 Quit Grahack (Read error: 60 (Operation timed out))
09:23:59cool_walking_oh good :)
09:24:22*GodEater is wondering about the "UserCon" email he sent to the dev list this morning
09:24:53 Quit bertrik ("De groeten")
09:26:40 Quit tmzt (Read error: 110 (Connection timed out))
09:27:47 Part Ashex ("Leaving")
09:30:00 Join Thundercloud [0] (
09:31:51 Join CaptainKewl [0] (
09:32:19 Join robin0800 [0] (
09:33:12 Part pondlife
09:38:28 Quit scorche (Nick collision from services.)
09:39:15 Join scorche [50] (n=scorche@rockbox/administrator/scorche)
09:39:37 Join mt [0] (n=MTee@rockbox/developer/mt)
09:41:01 Join nibbler_ [0] (
09:43:20 Quit CaptainKwel (Read error: 110 (Connection timed out))
09:44:55 Join funman [0] (n=fun@rockbox/developer/funman)
09:50:55 Quit Thundercloud (Remote closed the connection)
09:55:43mtAre there any functions like (time.h/clock() ) which I could use to time some portion of a code in a sim build ?
09:57:07funmanyou can use current_tick (10ms precision)
09:57:34funmanelse any function available on UNIX (and OSX) & windows
09:57:37 Join Grahack [0] (
09:58:26mtThanks. Will try.
09:58:27funmanmt: aren't clock_gettime/gettimeofday more suited?
09:59:35mtfunman : I don't want to get the current time, I want to determine how much time is needed to execute a certain part of the code.
09:59:56funmani would get the current time before and after the code and make the difference
10:00:18mtI tried in the test program (rm-wav converter) to measure the average time needed for decoding.
10:00:29mtIt's in the order of microseconds
10:00:42funmanclock() is a C89 function so it's more suited i believe
10:01:10funmanand its precision is 1µs
10:02:02mtYes, that's what I used in the test program. I was just asking for something equivalent to it in rockbox, because I couldn't find it in firmware/time.h.
10:02:57mtsorry , firmware/include/time.h
10:03:05funmanthere is nothing that much precise on target, except target-specific timers. If you only need it in the sim then clock() is a good choice
10:06:47mtI'm a bit confused, when using time.h for example in a sim, isn't it the one in firmware/include/ ?
10:09:59funmanperhaps you need to use <time.h> ? firmware/include/time.h has a special case for CHECKWPS
10:11:30mtOK. Thanks :)
10:15:09 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
10:15:52 Join efyx [0] (
10:20:02 Join kugel [0] (n=kugel@rockbox/developer/kugel)
10:21:21 Nick zitune[afk] is now known as zitune (n=zitune@bearstech/zitune)
10:23:48 Quit Grahack (No route to host)
10:38:57 Quit BHSPitMonkey (Remote closed the connection)
10:45:21 Join kachna [0] (
10:46:05***Saving seen data "./dancer.seen"
10:51:06 Join PaulJam [0] (
10:51:56 Join CaptainKwel [0] (
10:52:31 Join ocean [0] (i=d59c23e8@gateway/web/freenode/x-3856c406dd361f52)
11:01:02 Quit cool_walking_ ("Page closed")
11:02:01 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
11:02:20 Join einhirn [0] (
11:02:23 Quit CaptainKewl (Read error: 110 (Connection timed out))
11:03:06mtfunman : it seems that <time.h> includes the one in firmware/include/ ..
11:03:59funmanput an #error in it to be sure
11:05:08GodEater#error I_CAN_HAZ_SYSTEM_TIME.H?
11:08:50 Quit Zarggg ()
11:08:59 Join Nick [0] (
11:09:27 Nick Nick is now known as Guest59953 (
11:09:52 Part Guest59953
11:09:54 Join Guest59953 [0] (
11:11:23 Quit Guest59953 (Client Quit)
11:16:37mtyes, it does include rockbox's time.h
11:20:51kugelrockbox doesn't use standard lib
11:21:10kugelbut there's a way to get those for the sim
11:21:15kugelnot sure which
11:22:02 Quit ocean (Ping timeout: 180 seconds)
11:22:28 Join Zarggg [0] (
11:23:28 Join kachna|lappy [0] (
11:23:47 Quit kachna (Read error: 104 (Connection reset by peer))
11:29:22mtI included "/usr/include/time.h" which seems to work.
11:30:13 Join flydutch [0] (
11:31:19funmans/.$/around your problem./
11:32:03Mikachus,$/a,$/ a,
11:32:55funmanisn't "workaround" a real verb?
11:33:27Mikachuworkaround is a noun, work around is a verb (and another word) :)
11:33:38mtfunman : :)
11:34:38funmanMikachu: i'll ignore your regexp anyway since you didn't escape the '$'
11:36:09 Quit einhirn (Connection reset by peer)
11:36:28 Join einhirn [0] (
11:36:53 Quit einhirn (Client Quit)
11:37:12 Join einhirn [0] (
11:38:21 Join grdxyxy [0] (n=chen@
11:42:25 Nick zitune is now known as zitune[afk] (n=zitune@bearstech/zitune)
11:44:11Mikachuzitune[afk]: please don't use away nicks on this channel
11:46:59 Quit kugel (Read error: 110 (Connection timed out))
11:49:10 Join pyro_maniac [0] (
11:49:58pyro_maniacis it possible to define two usb ids for one model in device detection?
11:50:11Mikachudetection in what?
11:50:39funmanpyro_maniac: no
11:51:16pyro_maniacfunman: so i have to define a second device for the second 920 id?
11:52:16funmanwell we have to use one id, i don't really see the problem if the id reported by rockbox is different from the one reported by the OF on some models
11:53:25 Join Sajber^ [0] (
11:53:39pyro_maniacbut i have to detect for the second id in rbutil
11:56:05funmanah i don't know rbutil, since you didn't precise i thought you were speaking about the id reported by rockbox usb mode
11:57:10pyro_maniacno i want to detect both ids in rbutil
11:58:00funmanlooks possible, search rbutil.ini for iaudiox5
11:58:10pyro_maniacfunman: whats the name of your firmware file on the yh-920? PP5020 or FW_YH920?
11:59:35pyro_maniacthants the next problem
12:01:10linuxstbWhat's the next problem?
12:02:21pyro_maniacthere are two versions of the yh-920 with different usb ids and different firmware file names.
12:02:42linuxstbBut the same Rockbox build I assume?
12:03:12pyro_maniacso its the same like the usb ids right?
12:03:18linuxstbI think you should try and catch bluebrother or domonoky...
12:04:54 Quit einhirn (Connection reset by peer)
12:05:13 Join einhirn [0] (
12:05:39 Quit einhirn (Read error: 104 (Connection reset by peer))
12:05:58 Join einhirn [0] (
12:06:17 Quit einhirn (Read error: 104 (Connection reset by peer))
12:06:31 Join einhirn [0] (
12:08:04 Quit grdxyxy ("Leaving.")
12:15:28linuxstbpyro_maniac: Do the two versions of the yh-920 look any different to a user?
12:19:22 Join funman_ [0] (
12:21:12 Quit funman ("Reconnecting")
12:22:07pyro_maniacthe branding on the cover is different too.
12:22:25pyro_maniacthere is a napster and a yepp branding
12:24:22pyro_maniacand maybe there are two different display colors, but i am not sure in this case
12:25:48 Quit nibbler_ (Read error: 110 (Connection timed out))
12:31:08linuxstbpyro_maniac: Maybe it would make sense to have them as two separate targets in rbutil, but sharing the same build? That way, a user can double-check the detection has worked, and override it as necessary.
12:31:41linuxstbAlso, the manual install instructions will need to differentiate between the two, so the user gives the bootloader the correct name.
12:31:59 Quit evilnick_home1 (Read error: 104 (Connection reset by peer))
12:32:39pyro_maniaci can't compile rbspeex
12:33:15linuxstbWe're going to need more info than that ;)
12:35:09pyro_maniacsorry. i missed that again. i tried to compile rbutil but it failed because rbspeex. "/bin/sh: cannot create build/dep-speex: Directory nonexistent"
12:35:52linuxstbDo you have a full svn checkout, or just the rbutil directory?
12:36:35pyro_maniacfull checkout
12:37:51 Join kugel [0] (n=kugel@rockbox/developer/kugel)
12:38:13 Join nibbler_ [0] (n=Nibbler@
12:46:08***Saving seen data "./dancer.seen"
12:49:06pyro_maniaci tired the source tarball from the wiki but that failed to. "base/autodetection.cpp:28:17: error: usb.h: No such file or directory"
12:49:28Mode"#rockbox +o gevaerts " by ChanServ (ChanServ@services.)
12:49:33Mode"#rockbox -b *!*n=perrik*@* " by gevaerts (n=fg@rockbox/developer/gevaerts)
12:49:39Mode"#rockbox -o gevaerts " by ChanServ (ChanServ@services.)
12:49:45 Nick funman_ is now known as funman (
12:49:56Mode"#rockbox +o gevaerts " by ChanServ (ChanServ@services.)
12:50:24Mode"#rockbox -o gevaerts " by ChanServ (ChanServ@services.)
12:55:20GodEaterpyro_maniac: you need to create an empty "build" directory under rbspeex
12:55:27GodEaterthis is a bug in the Makefile apparently
12:55:37GodEater(and was detailed by bluebrother in a recent forum post)
12:57:16 Quit kugel (Nick collision from services.)
12:57:24 Join kugel [0] (n=kugel@
12:57:52pyro_maniacGodEater: thanks, the rbspeex problem is solved.
12:59:17*rasher thinks it might make sense to select the most recently activated theme (like the font, lang etc thing of yesterday)
13:00:16rasherIt will usually be completely correct or almost correct. If not, the user should know
13:15:58 Join FlynDice [0] (
13:16:20 Quit kugel (Nick collision from services.)
13:16:26 Join kugel [0] (n=kugel@
13:26:40linuxstbrasher: You mean keep track of the .cfg filename?
13:31:09rasherYep. Technically it's not accurate, but I think it'll be good enough. And better than doing nothing.
13:31:52linuxstbSo you would save this to the main .cfg file?
13:34:29rasherYeah. But not use it for anything else.
13:36:00 Quit gevaerts (Nick collision from services.)
13:36:10 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
13:38:59 Join evilnick_home [0] (
13:40:24 Quit nibbler_ (Read error: 110 (Connection timed out))
13:44:07mtPeople, I want some opinions on this please : I'm trying to measure the speedup gained by using our mdct code vs the current cook mdct code. what I'm doing is just measuring the average decoding time using both mdct's, timing is done by calling clock() before and after decode_frame(), adding up the time till the sample is finished and then dividing by the number of frames.
13:44:51linuxstbmt: Speedups in the sim are not that interesting - you want to know the speedup on target...
13:45:26mtlinuxstb : I thought it would give a rough figure of the speedup on target.
13:45:34linuxstbAlso, this seems like optimising too early - do you have the codec working on target yet?
13:47:18mtNo not yet. I was just changing the code to use rockbox's mdct and thought I'd do this comparison too.
13:47:54mtBut now I'm reminded I shouldn't be really focusing on this now. ;)
13:49:06linuxstbWhat's more important (IMO) is checking that the output is accurate - e.g. comparing WAVs produced with the two versions.
13:49:11mtI'm going to make a patch soon so you and saratoga (and anyone else who wants to) could be able to see what's going on.
13:49:28mtI see.
13:49:34Torneit's not outside the realm of possibility for the results to be reversed between sim and target performance-wise :)
13:50:05Tornehilarious microoptimised loops behave weirdly when run on radically different architectures to the ones they were written for
13:50:43mtThe results where pretty close on the sim anyway, so I'll just be patient till I have something on target. :)
13:51:34linuxstbIt's also easier to measure things on target, as there's no pre-emptive multi-tasking getting in the way. You just time how long it takes using the system tick.
13:51:45linuxstbtest_codec does that for you...
13:52:31funmanwell there is still the tick_tasks
13:53:18funmanbut test_codec is still accurate
13:53:19linuxstbYes, but I guess they're predictable. i.e. two runs of test_codec should give the same time.
13:54:37mtlinuxstb : by the way, the test program might probably have its copies of rm.[c/h] back. (I don't know if you have seen that in the logs)
13:54:59 Quit robin0800 (Read error: 104 (Connection reset by peer))
13:55:47funmanmt: couldn't you use #ifdef TEST_PROGRAM or something in these lines?
13:55:52mtfor codecs librm needs to implement only the get_packet function. The rest of the code (rm.c specifically) is just needed by the test program
13:56:00funman(inside common rm.[c,h])
13:56:07mtfunman : That's what I did first.
13:56:14linuxstbmt: Have you seen BryanJacobs' patch that allows him to run the Rockbox wavpack codec on his PC? (i.e. outside Rockbox).
13:56:39 Quit funman ("free(random());")
13:57:26mtyes I've seen it, it's very good, and I might use it for cook too.
13:58:48mtfunman : But I found out I'm #ifdef'ing almost the whole file. (except for that one function needed by the codec)
14:02:01mtlinuxstb : What do you think about copying rm* files ? Or are you with the conditional compilation solution ?
14:02:36 Join robin0800 [0] (
14:03:17 Quit robin0800 (Client Quit)
14:03:40 Join robin0800 [0] (
14:04:34linuxstbmt: Looking at the rm.[ch] currently in svn, they don't seem useful for codecs - it's reading from a file descriptor. So yes, that looks like test-only code. But you need to rewrite that code to work in your codec(s).
14:04:36 Quit robin0800 (Client Quit)
14:05:54 Quit ch4os (Read error: 60 (Operation timed out))
14:05:58 Join robin0800 [0] (
14:06:11 Join ch4os [0] (n=ch4os@gentoo/user/ch4os)
14:08:35mtrm.h is needed by the codecs and the parser for the different decalrations. (like RMContext ), but, the only part of rm.c that is actually used by cook (or, later, any rm codec) is the get_packet function. So it's actually just a matter of removing some unneeded code than rewriting it.
14:11:06linuxstbI think I would just the test code into your main.c
14:11:15linuxstb^move the test code
14:11:45linuxstbAnd if possible, use that get_packet function from rm.c
14:12:08 Join barrywardell [0] (
14:14:29 Quit barrywardell (Remote closed the connection)
14:14:43 Join barrywardell [0] (
14:17:46mtJust to make sure I'm following, librm will remain intact (except for the part that should be moved from rm.c to main.c ) ?
14:19:31linuxstbmt: I think so, yes.
14:20:56linuxstbmt: Although why is there cook-specific data in the RMContext? Is that still needed?
14:23:47mtin cook's init function, it needs to copy some parameters from RMContext.
14:24:14linuxstbI thought those were stored in the "id3" metadata struct?
14:26:15mtI wanted the test program and the codec to use the same cook library. So I copy back RMContext from id3 into a new instance of RMContext and send that to cook's init()
14:27:51mtlinuxstb : I'll try not to do any more 'heavy' modifications and get the patch ready asap for the sake of an easier discussion. :)
14:27:56linuxstbMaybe this test program is just confusing things, and it would be better to forget it (for now at least), and just concentrate on getting it in Rockbox.
14:29:55 Quit __lifeless (Remote closed the connection)
14:30:11 Join __lifeless [0] (n=lifeless@
14:31:44 Join barrywardell_ [0] (
14:32:29 Quit barrywardell_ (Client Quit)
14:32:58 Quit barrywardell (Read error: 60 (Operation timed out))
14:33:20 Part pyro_maniac ("Leaving.")
14:46:05 Join Grahack [0] (
14:46:12***Saving seen data "./dancer.seen"
14:55:47 Join barrywardell [0] (n=barrywar@
15:00:29GrahackHi, I'm playing with the lua plugin on my iPod video and I'm having problems to separate clicks on SELECT and RIGHT. I'm reading the spacerocks.c file which uses rb->button_get(false); and doesn't use the ACTION_STD_... constants, but on the lua side, rb.button_get(false) always returns 0. How can I get distinct codes?
15:01:14Torneif it returns 0 then there's no keys pressed..
15:02:03Torneare you trying to wait for akeypress or just get what keys are pressed at that instant?
15:02:17linuxstbbutton_get() returns a bitmask of the different buttons currently pressed - which is completely different to the "action" API.
15:02:19 Quit barrywardell ()
15:02:39Torneif you want to *wait* for a keypress you need to do rb.button_get(true)
15:02:55Tornepassing false means it returns immediately, so if no button is pressed at that exact moment it will return 0
15:03:06GrahackHere is the lua code I have to test this: I commented the ACTIONS test and tried rb.get_button
15:03:16Grahackok I'll try with true then
15:03:46Grahackworks with true thanks
15:03:49Torneif you're going to use button_get you don't want to be calling any of the action stuff
15:03:56Torneone or the other
15:04:20Tornespacerocks polls for buttons while it's doing other things (moving the rocks around :)
15:04:23Torneso it doesn't want to block
15:05:25Tornenext time, check what the parameters to functions mean :)
15:07:40 Join evilnick [0] (i=0c140464@gateway/web/freenode/x-56558c6518369968)
15:09:45GrahackBut apart from the code (the source luke), where can I find this info. I once found">
15:10:06GrahackI'm not a C master and the code is huge.
15:12:37GodEaterit's not huge
15:12:46GodEatertry reading the linux source some time. *That's* huge :)
15:13:23GodEaterfinding out how to call something in the rockbox code should be as simple as grepping for the function definition
15:14:02linuxstbGrahack: plugin.c and plugin.h would be the first places to look. And then, as GodEater just said, you can always look for where that function is actually implemented to see exactly what it is doing.
15:15:13 Join fml [0] (n=4fd3ceea@gateway/web/cgi-irc/
15:15:33GodEaterbutton_get() is one of the functions which *is* documented in docs/API for example :)
15:16:08Grahackok sorry guys, once again :) and thanks for the answers and explanations
15:16:17GodEaterno problem :D
15:16:28linuxstbNo need to apologise - we're here to help ;)
15:18:50fmlHello. I have a little problem with the "goto wiki page" field (in the top right corner). I visit some wiki pages from time to time. To do that, I just enter the page name into the field and press Enter. If I start to enter the name of a page I visited before, the browser shows a list of visited pages with matching names. I select the desired name in the list (using UP and DOWN) and press TAB to let the list disappe
15:19:46GodEatertry again - we only saw up to "the list disappe"
15:20:07fmlA remedy would be to have a "Go button" right to the input field. Pressing TAB while in the input field would move the focus to that button which, when pressed, would submit the HTML form.
15:20:28fml...ear. But then the input focus seems to have moved somewhere else so that pressing Enter does not take me to that wiki page.
15:20:52GodEaterI would say that's a bug in the browser
15:20:59linuxstbCan't you press other buttons, such as the right cursor, or enter, instead of TAB?
15:21:35GodEaterworks with enter
15:21:59fmllinuxstb: yes, Right works! I'm just used to press TAB in such cases
15:22:36fmlPressing Right leaves the focus in the input field. While pressing TAB not only hides the list but also moves the focus.
15:23:12GrahackAbout the lua plugin, it could be nice to have a kind of repository page for rb-lua apps/demos. Does someone here have some code like this already ?
15:23:29GodEaterdon't think so
15:23:34GodEaterfeel free to start a wiki page for them :D
15:24:03 Join kkurbjun [0] (n=kkurbjun@rockbox/developer/kkurbjun)
15:24:27*GodEater wonders if kkurbjun has come to tell us off for being so fearless with our gigabeat flashing
15:24:32LloreanIs target-agnostic lua code easy/likely?
15:24:50kkurbjunNo, technically it should always work
15:24:54GodEaterLlorean: good question
15:24:58kkurbjunthe only problem might be with dual booting
15:25:06GodEaterwho wants to do that ? :)
15:25:08LloreanGodEater: Otherwise more organization than a single page is probably wise.
15:25:53linuxstbI guess it could just start as a single page, and then split if needed - something like the WpsGallery used to be.
15:25:58LloreanIdeally I'd have a single page for simple demo scripts that must be target agnostic, and then anything more complex that's useful rather than just a scripting demo should be submitted for inclusion like normal plugins.
15:26:19linuxstbBut I just imagine it being a big table, with one line per lua script, so a single page may be easier to use.
15:26:41linuxstb(with perhaps separate tables for different types of scripts or target groups)
15:26:58gevaertskkurbjun: it does dual booting?
15:27:22GodEatergevaerts: hush. No-one in their right mind would want that. :D
15:27:25linuxstbAre we talking F/X?
15:27:33GodEaterlinuxstb: we are
15:27:38GodEaterwell, F at least
15:27:51linuxstbThen no need to think about or mention dual-boot ;)
15:27:52*GodEater is well impressed with the boot speed of his F now
15:28:16kkurbjunGodEater: the flasher modifies the reset vector which immediately jumps to the rockbox bootloader, as long as there is nothing wrong with the bootloader in memory you should always be able to boot back to rockbox, the only problem might be is in the dual boot. if that original reset vector points to a different address than is coded in the rockbox bootloader then the doal boot function won't work right
15:28:36kkurbjunso far I have not seen a flash image that uses a different reset vector location
15:28:45fmlShouldn't we strive to incapsulate the RB API (for Lua) in such a way that all code is runnable on all targets? Lua should be platform independant IIUC.
15:29:03kkurbjunand with everyone turning up different flash versions I'm wondering if there is something unique to each player stored in flash
15:29:11markunkkurbjun: I also have the bootloader in flash now :)
15:29:20linuxstbfml: Yes, but it would still need the script author to take account of differences. Although #ifs will be runtime ifs.
15:29:31gevaertsmarkun: you still need to upload your backup.bin and md5!
15:29:38kkurbjunIt should work on the X too, but I've never seen a flash image for one
15:29:39markungevaerts: can do that now
15:29:43kkurbjunmarkun: nice
15:30:12fmllinuxstb: I mean, all this stuff should be #if'ed in the wrappers.
15:30:18kkurbjunYou can dual boot by holding menu while pressing power
15:30:19Lloreanfml: Hardware's too different to be able to ensure all scripts work everywhere without developers doing some of the work.
15:30:27kkurbjunand it should follow the OF boot routine
15:31:01GodEaterkkurbjun: assuming you didn't do as gevaerts did, and delete the GBSYSTEM folder ? :)
15:31:05kkurbjunthe main reason I wanted to dual boot is so that I could use the gigabeat to copy pictures from my camera
15:31:13fmlLlorean: couldn't we agree on some standard Lua API and implement that API for all platforms? I.e. drawing color lines on b/w platforms would produce grey lines
15:31:18kkurbjunGodEater: :), yes
15:31:25Lloreanfml: And what about buttons? Do we just limit the number of buttons they get?
15:31:28GodEaterfml: what about different screen sizes ?
15:31:36LloreanLines can only be within a 112x64 range because that's our smallest screen?
15:31:42gevaertsGodEater: I was using the dummy gbsystem before anyway :)
15:31:51GodEatergevaerts: I think I was too :)
15:31:59LloreanYou *could* make sure Lua scripts would be universal, but it would be very restrictive of them./
15:32:01GodEaterlike I said, I'm not in any way interested in dual boot
15:32:06markungevaerts: where should I upload my files?
15:32:16fmlLlorean, GodEater: procentual coordinates? Or virtual screen (i.e. always 1000x1000)?
15:32:17gevaertsmarkun: FS #7505
15:32:29Lloreanfml: Doesn't work for use of bitmapped graphics. So we ban those?
15:32:33 Quit kachna|lappy (Success)
15:32:46PaulJamyou forget the charcell displays. lines shouldn't be possible at all ;)
15:32:50fmlLlorean: we can scale those
15:33:13GodEateryou still didn't cover buttons
15:33:14fmlPaulJam: that calls would translate to NOP on those platforms
15:33:20Lloreanfml: So we force bad graphics and slowdown. But even ignoring that. Buttons?
15:33:50LloreanPaulJam: Does hwcodec even run Lua?
15:33:55LloreanSomehow I guessed it'd be a larger plugin.
15:33:58fmlI just thought that we have enough hell with #if in C and could get rid of that in Lua.
15:34:34Lloreanfml: I think if it were possible to get rid of it in Lua, we'd have gotten rid of it for the C plugins too.
15:34:51LloreanBy providing a library that makes it unneeded in the actual plugins, for example.
15:35:45markungevaerts, kkurbjun: ok, uploaded my files. Don't know if they are the same as some of the other ones
15:35:54PaulJamLlorean: i must admit i havent really looked at the lua stuff. i have no idea if it runs on hwcodec.
15:36:06fmlLlorean: maybe. How does it work now? Are there some predefined magic objects (e.g. LCD) for testing the HW capabilities?
15:36:10kkurbjunLlorean: I think with some clever thought quite a few of the target specific stuff could be removed in the plugins
15:36:15GodEaterI don't think it does run on hwcodec
15:36:59kkurbjunbrickmania is one good example, scaled graphics on the targets would not loose much information for the user, they are just blocks that could be linearly scaled for most of the targets
15:37:07Lloreankkurbjun: Yes, but keymaps and use of the screen are always going to be target specific (though the screen can be generalized if performance and graphics quality isn't an issue)
15:37:14kkurbjunand then the positions of the text and menus could be relatively placed
15:37:28Lloreanfml: "testing"?
15:37:59fmlLlorean: like "if LCD.width > 1000 then blah"
15:38:40kkurbjunyes, keymaps will be specific more plugins could use the pluginlib buttons though which would make supporting new targets and screens less of a PITA
15:39:24Lloreanfml: Isn't that essentially the same as #ifs anyway, just doing it runtime instead?
15:39:32kkurbjungetting 640x480 and 480x640 working on the M:robe 500 takes alot of time to scale the graphics and make sure the hard coded positions are right, etc.
15:40:00kkurbjunso at the moment I'm looking at potentially changing brickmania around, but there's a ton of plugins that are similar
15:40:05fmlLlorean: yes, it's the same. Is this the way it must be done now?
15:40:34kkurbjunmarkun: thanks for uplaoding those
15:40:40fmlI mean, is the advantage of Lua just in avoiding write-compile cycle?
15:40:55Lloreanfml: As far as I know, there are defines for LCD width and LCD height, plugins usually test those early on to determine which images to use, as well as where to position them
15:40:56kkurbjunI need to take a look at the different images and see where the differences are
15:40:59markunkkurbjun: do you think we'll ever have a sleep mode?
15:41:20Lloreanfml: That's always going to be the advantage of lua. That and being another language people can learn instead of C.
15:41:34kkurbjunthat would be nice, but it will take quite a bit of code for rockbox to support it
15:41:52LloreanIdeally any problem we could solve in the Lua wrappers to make programs need less ifs is a problem we could also solve in the plugin libraries to make plugins need less #ifs
15:42:03fmlLlorean: so the goal was not to avoid the target specific code? Then I'm ok with it.
15:42:07kkurbjunsince you need to account for conditions where other devices need to be re-initialized, but rockbox itself can just resume from where it left off
15:42:09Grahackplease where can I find constants like BUTTON_LEFT (rb.buttons.BUTTON_LEFT?) as there is rb.actions.ACTION_NONE ? I think they are not exposed to lua.
15:42:17linuxstbkkurbjun: sleep mode is also needed for the 1g/2g (original big bars of soap) ipods as well, so it's generally useful...
15:42:25kkurbjunamiconn had some ideas on how to do a sleep mode
15:42:26linuxstbBut yes, it's far from trivial.
15:43:57kkurbjunsleep mode is something I would like to look at someday, but I have enough to do already with the M:robe 500 and eventually getting that flash writer in SVN
15:44:18linuxstbGrahack: It looks like they're not.... They're defined in a target-specific way in the button-target.h file for your device (under firmware/target/)
15:47:07 Quit fml ("Bye")
15:47:19Grahackmmf, here is the first lua cross-target issue I meet
15:48:13linuxstbYes, buttons will be tricky. The cross-target way will be to use actions, but they're very limited.
15:51:52Tornei don't think lua makes any effort to export target-specific constnats at all atm
15:53:01linuxstbThere are some - LCD_WIDTH and LCD_HEIGHT
15:53:33Tornethat actually exists on all targets, though, i guess :)
15:54:09CIA-71New commit by funman (r21477): Sansa AMS debug: small fixes ...
15:54:37linuxstbTorne: Yes. What kind of things do you mean then?
15:55:13linuxstbI guess we need things like:
15:55:16linuxstb#ifdef BUTTON_MENU
15:55:54linuxstband then define that to be a constant containing the value of BUTTON_MENU or (e.g.) 0 if it's not defined.
15:56:09Torneeurgh. :(
15:56:35Tornebutton mapping appears to be a perpetual issue for everything :)
15:57:15LloreanIf your plugin can survive with just up, down, left, right and select it shouldn't be a problem. :-P
15:57:35LloreanDisclaimer: May not hold true on Ondio
15:57:49linuxstbAnd things like "up+left" may work on some targets, but not others...
15:58:05gevaertsOr on the meizu m3 (once that one actually boots rockbox that is)
15:58:45TorneLlorean: except if you know nothing about some targets you can even get that wrong
15:59:14Tornee.g. i hate things that think up and down should be the scrollwheel on ipod :)
15:59:18Tornerather htan using menu/play
15:59:22 Join Lss [0] (
15:59:24LloreanIt's still usable though, just not ideal
15:59:33Tornewell, yah
15:59:36GodEaterwhat does the H10 use for left and right ?
15:59:44LloreanDoesn't it have buttons for those?
15:59:51*GodEater has no idea, which is why he asked
16:00:28GrahackOn iPod left and right are ACTION_STD_CANCEL and ACTION_STD_OK, and the select button is also ACTION_STD_OK, so these two last are coupled. That's why Iinvestigated thorough the buttons code
16:00:38LloreanGodEater: I think it has prev/next buttons on the side. I'm not *certain*
16:00:51 Join TBOL3 [0] (i=180ac800@gateway/web/freenode/x-ceb671376f6ac8d4)
16:01:17Torneyah, the actions lib tries to map buttons to what they generally do in the main rockbox ui
16:01:26Tornethe above is what they mostly do there. :)
16:01:38TBOL3Dumb question, how can I unsubscribe to the rockbox mailing list? I've tried going to the unsubscribe button, and it claims it sends me a conformation email, but I don't get it.
16:01:52TBOL3The next day, I get the rockbox updates, on schedule.
16:02:01GodEaterTBOL3: did you check your spam folder ?
16:02:18TBOL3Also, when clicking on send me my password (I don't remember making one), I still don't get anything.
16:02:26TBOL3Thank You
16:02:29GodEaterTBOL3: again, did you check your spam folder ?
16:03:23TBOL3It's not in there
16:03:29*Bagder detects the user list now having 1016 subscribers
16:03:38TBOL3But maybe I had it sent to a different address, and forwarded to me, let me check...
16:03:50gevaerts50 more and we can invade England!
16:03:59*GodEater detects that Bagder's status is a fib
16:04:02Tornehey, don't invade england
16:04:40GodEaterTorne: they have to, DevCon is here next year.
16:04:40TBOL3Nope, there's nothing.
16:04:46Torneooh, it is?
16:04:51TBOL3Although, I think that might have helped me, thank you.
16:05:00gevaertsGodEater: only if we get these 50 users :)
16:05:01*GodEater wonders how
16:05:07GodEatergevaerts: hehe ok
16:05:41GodEaterI still need to work out where to host it as well. It'll take me the best part of a year to work that out.
16:05:51GodEaterIt might have to be a hotel room job as like DevConWest
16:06:18TBOL3Allright, I think I'm off of it.
16:06:23TBOL3Thank you very much.
16:06:28GodEaterno problem
16:06:34*GodEater wonders how it helped mind you
16:06:39gevaertsGodEater: try to poke the other UK folks. One of them might have a location (maybe in Glastonbury?)
16:06:44 Quit TBOL3 ("Page closed")
16:06:50*GodEater is not going to DevCon in a field
16:07:52*GodEater thinks linuxstb is more than aware now. The only other UK peeps I know are pondlife, who is not London based, so not a big help, and shotofadds, and I have no real contact details for him.
16:08:04*Torne is in the UK, though he probably doesn't count.. :)
16:08:25GodEaterTorne: how close to London ?
16:08:31Torne'waterloo'? :)
16:08:40gevaertsTorne: Waterloo is Belgium!
16:08:41Torneapparently i work right around the corner from linuxstb
16:08:41GodEaterthat's in Belgium isn't it ? :)
16:09:04*GodEater wonders why the bloody hell Torne hasn't attended a single DevConPub then
16:09:13Tornebecause i don't pay attention
16:09:27*GodEater docks marks from Torne's end of term report
16:09:27Tornealso i only realised linuxstb was so close the other day
16:09:43Torneheh, my salary is already frozen :(
16:09:49Tornemy marks don't matter any more :)
16:10:30GodEaterwell that's one more person who knows it's in London then
16:10:36GodEaterso you can think of a venue too!
16:10:38 Join jgarvey [0] (
16:11:07scorche|shTorne: how about a DevConPub sometime in the next week or so? =P
16:11:09GodEatersafetydan is in the UK too. He traditionally promises to turn up to DevConPubs and never does.
16:11:13Torneif you like
16:11:25Torneif someone hilights me on the day i'll show up :)
16:11:29GodEaterscorche is arriving tomorrow night
16:12:03GodEaterscorche|sh: btw, I have the day off on Friday
16:12:07GodEaterwhat do you want to do ? :)
16:12:33 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
16:12:38 Join nibbler_ [0] (
16:14:37*Torne wonders if anyone would notice at work if he just booked a meeting room for it :)
16:14:59*GodEater proposes Friday night for an impromptu devconpub
16:15:44Grahackis there a mean to know on which target I am in this mysterious rb lua object?
16:16:21GodEaterTorne: which lucky establishment pays for your services then ?
16:18:22LloreanGrahack: Why do you need to know specifically what target it's being run on?
16:21:33 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
16:21:55GrahackI could map the values found in button-target.h to lua constants (#define BUTTON_SELECT 0x00000001)
16:22:03*pondlife hears beertalk from afar...
16:22:25GodEateryou'll be at Glastonbury
16:22:30GodEaterso pretend you didn't hear
16:22:31Grahackbut maybe I can play target-specificly waiting for the lua plugin to grab those constants
16:22:33pondlifeYep, in the thunder
16:23:06pondlifeI'm certainly up for a DCP soon after though
16:23:34GodEateryou'll miss scorche though
16:23:39GodEaterhe's going home on Tuesday
16:24:14pondlifeUnless he gets lost in beer and misses the flight.
16:25:20pondlifeI'l raise a recyleable paper cup of cider to y'all anyway.
16:25:54*pondlife has to pack
16:25:56 Part pondlife
16:26:27Torneif someone hilights me on frida ywith a location and time i'll come along anyway :)
16:26:44 Part LinusN
16:27:00 Join Hillshum [0] (i=cd7ae8fa@gateway/web/freenode/x-0c664167408ba5b1)
16:31:15 Join kachna|lappy [0] (
16:40:04 Join lymeca [0] (
16:40:31 Part Hillshum
16:40:39 Join Hillshum [0] (i=cd7ae8fa@gateway/web/freenode/x-0c664167408ba5b1)
16:40:46 Quit kkurbjun ("Leaving.")
16:44:59 Join dberg918 [0] (n=ae6a45d6@gateway/web/cgi-irc/
16:46:13***Saving seen data "./dancer.seen"
16:55:20 Join funman [0] (n=fun@rockbox/developer/funman)
16:56:43 Join faemir [0] (n=faemir@
16:58:13 Nick funman is now known as funman_ (n=fun@rockbox/developer/funman)
16:58:45 Nick funman_ is now known as funman (n=fun@rockbox/developer/funman)
17:05:23LambdaCalculus37 <−− Keymaps coming along for SA9200.
17:05:59HillshumIs the Clip keymap set in stone?
17:06:33 Nick fxb__ is now known as fxb (
17:06:51LloreanHillshum: What problem are you having with it?
17:07:06funmanHillshum: you're welcome to improve it if you have ideas
17:07:35HillshumWhat's the home button supposed to do? It's confusing me
17:08:51funmandepends the context, i think the keymap is not really consistent across the multiple screens/plugins
17:09:06LloreanIt should usually be what the "Menu" button does elsewhere, right?
17:09:47funmanthe current exit/goback button is power
17:09:56HillshumYes, and stop
17:10:03 Quit Grahack ("Leaving.")
17:11:35LloreanSo, is there a specific problem somewhere you wish to address?
17:11:36 Join Xakh [0] (n=xakh@
17:11:58 Quit Cory` ("I was raided by the FBI and all I got to keep was this lousy quit message!")
17:12:20 Join Cory` [0] (
17:13:33HillshumLlorean: I havn't yet identified anything specific
17:13:51mtlinuxstb, saratoga and anyone else who's interested : I've added a patch in FS #10182 for rm playback in the sim.
17:14:17LloreanHillshum: Keymaps are unlikely to be changed unless there's a pretty solid reason for change, though the time to do it is now before it gains "supported" status
17:15:00Xakhso how goes the port over to the sansa view?
17:15:01HillshumLlorean: yeah. Are plugin keymaps easier to get changed?
17:15:13funmanLlorean: the clip keymap is quite inconsistent so i believe it's welcome to do so :)
17:15:29funmanXakh: check the SansaView wiki page
17:15:38funmanHillshum: easier than what?
17:15:57Hillshumfunman: then core keymaps
17:16:01mtXakh : you could check the dev mailing list (weekly updates go there) or check the project's wiki.
17:16:14LloreanHillshum: Plugins, yes.
17:16:33Lloreanfunman: Inconsistent in what way? It's very limited in terms of buttons anyway
17:18:01 Quit Xakh (Read error: 104 (Connection reset by peer))
17:18:15linuxstbmt: I'll try and look at it later. So it doesn't work on target?
17:18:20funmanLlorean: searching a relation between button and function can be quite difficult - i do not have a precise exemple on mind.
17:19:35dberg918is there a way to dump readings from the scrollpad on the H10 to a text file? I looked on the website and couldn't find anything about the debug menu...
17:19:36mtlinuxstb : No it freezes when on target.
17:19:47linuxstbmt: At what point?
17:20:05Lloreanfunman: I think the main thing is that "home" should take you to the menu from the WPS (possibly duplicating the functionality)
17:20:08 Quit kugel (Read error: 110 (Connection timed out))
17:20:13mtlinuxstb : Once a rm file is opened.
17:20:29linuxstbmt: Can you tell if it's in get_metadata() or in the actual codec?
17:21:21funmanLlorean: currently home is the record button
17:21:35funmanthe keymap was mostly copied from c200 keymap, replacing rec button by home button
17:21:36LloreanFrom the WPS?
17:21:37mtlinuxstb : Not really, I was focused on fixing the bugs I see on the sim first so I didn't try any on-target debugging.
17:21:37 Join Hillshum_ [0] (i=cd7ae8fa@gateway/web/freenode/x-f43af4405f9490be)
17:21:44 Quit Hillshum (Ping timeout: 180 seconds)
17:21:50funmanyes - just tested
17:22:00linuxstbmt: OK. So is it reliable in the sim now, or are there still bugs?
17:22:29*GodEater is shortly off to visit stripwax's place of work
17:22:49Lloreanfunman: Since there's no record button, and the player has so few buttons total, it's probably best to limit recording access to using the menus
17:22:55linuxstbmt: Also, it may be easier if you don't call all your patches "rm_in_sim.patch" ;) e.g. use something like rm_in_sim-v1.patch
17:23:24 Nick Hillshum_ is now known as Hillshum (i=cd7ae8fa@gateway/web/freenode/x-f43af4405f9490be)
17:23:25funmanLlorean: good idea
17:23:30mtlinuxstb: hehe, sure :)
17:23:43funmanyou could also add "since i can't get recording to work on Sansa ams" :)
17:24:03 Join Xakh [0] (n=xakh@
17:24:05mtlinuxstb: That bug about id3v1buf isn't fixed yet.
17:24:15XakhAlso, when can I buy a Lyre?
17:24:34funmanXakh: contact the lyre project about that, not rockbox
17:25:07Xakhwhere's their IRC?
17:25:16GodEaterdon't know that they have one
17:25:43Torneor build your own hardware? :)
17:25:49XakhI don't have much cash
17:25:53Xakhat least not for that
17:25:59XakhI was hoping it was almost ready
17:26:02Xakhsince it's mid june
17:26:27GodEaterI don't think it's going to be a cheap player at all
17:26:30funmanXakh: they don't tell about any deadline
17:26:35Hillshumwouldn't building it yourself be cheaper?
17:26:49Xakhon the rockbox website it talks about the lyre project
17:26:57Xakhthere's a link to it on that sf page
17:27:03TorneHillshum: depends what size you are aiming for
17:27:10 Quit AlexP_ (Read error: 60 (Operation timed out))
17:27:21 Join AlexP [0] (n=alex@rockbox/staff/AlexP)
17:28:33LloreanXakh: Just because they link to Rockbox doesn't mean Rockbox is responsible for them.
17:28:52XakhI know, but I figured since I couldn't find much more info you guys might know
17:28:53 Quit Cory` ("Going!")
17:28:59LambdaCalculus37lowlight: (for the logs) Want to discuss some SA9200 keymaps for the plugins?
17:29:12 Join Cory` [0] (
17:29:12 Quit LambdaCalculus37 ("Fwump")
17:29:33LloreanXakh: A good rule of thumb here is "just because they might know doesn't mean this is the right place to ask the question"
17:29:57 Quit Cory` (Client Quit)
17:30:20 Join Cory` [0] (
17:30:24GodEaterXakh: the lyre guys don't come here much, so we don't have much of a clue about what they're doing
17:41:21 Quit Sajber^ (Read error: 104 (Connection reset by peer))
17:43:00*preglow reappears
17:43:14*linuxstb slams the lid back on the box
17:43:33*preglow thumps the lid and makes muffled noises
17:43:54*linuxstb locks the door and walks away
17:44:01*Hillshum shoves linuxstb away and tries to pry open the lid
17:44:20linuxstbpreglow: Welcome ;)
17:44:21preglowthat might not be wise!
17:53:03preglowthis new build system sounds cool indeed
17:54:28 Join Rondom [0] (
17:55:04preglowoh and btw, i read paul's post, am i on the steering commitee now, or has something else happened that i'm not aware of?
17:56:20gevaertspreglow: I think you are
17:56:25preglowyeah, looks like it
17:56:32preglowdamn, i missed shotofadds again :/
17:56:36obobut the vote is due again early July?
17:56:37linuxstbI don't think anyone has said anything regarding that. But being on it has so far been the same as not being on it...
17:56:44gevaertsproblem is, I don't know if the other members remember that they're members :)
17:56:50preglowlinuxstb: yeah :P
17:57:09preglowcan't remember ever having seen anyone invoke it's judgement
17:57:12preglowbut i guess that's all good anyway
17:57:19preglowits its its its its!
17:57:26linuxstbNo, people just seem to get worn down in arguments, then give up...
17:57:42preglowwhich is also, in a way, good :>
17:57:44preglowcode > talk
18:01:57dberg918I apologize for repeating my question, but I don't think anybody saw it the first time...
18:02:02dberg918is there a way to dump readings from the scrollpad on the H10 to a text file? I looked on the website and couldn't find anything about the debug menu...
18:02:38preglowsounds like the kinda thing you need to code yourself
18:02:56preglowprimarily since i can't ever think of needing such a feature :>
18:03:24dberg918I was hoping to graph the dump to see what it looks like
18:03:25Hillshumdebugging new drivers maybe, but the H10 isn't old
18:04:03preglowbut nah, it might not even be trivial, depending on if the scrollpad driver is polling or interrupt based
18:04:07preglowi dunnoes
18:05:12dberg918I also saw FS #6041
18:06:09saratogathe svn server seems unhappy today, i've been downloading the fonts folder forever
18:06:09dberg918it looks like someone has figured out a filter to make the scrollpad readout more manageable
18:06:28dberg918but the last post there was in October, not sure if anyone is working on it
18:07:17saratogawe have too many fonts
18:07:44preglowof varying quality
18:08:02 Part Hillshum
18:08:07preglowbut our art direction is kinda fuzzy, so it's to be expected
18:08:12 Join Hillshum [0] (i=cd7ae8fa@gateway/web/freenode/x-f43af4405f9490be)
18:08:58saratogaare the 35 point fonts actually usable on any target?
18:09:21 Join bertrik [0] (
18:09:54preglowno idea...
18:10:04linuxstbThat m:robe thing?
18:10:48gevaertssaratoga: define usable. Some people can accept just two lines, provided they can actually read those
18:11:00Lloreanpreglow: Some people like huge fonts for use in-car
18:14:28 Quit Zarggg (Excess Flood)
18:14:53 Join Zarggg [0] (
18:15:22funmanbertrik: i'm looking at recording support on AMS, i'm looking at the OF and i see it mentions a 0x55 register in as3514, does that tell you something?
18:15:52bertrikI'd have to look that up in the datasheet, doesn't ring a bell right now
18:16:19funmanthe datasheet only goes up to 0x3f
18:16:36funmanit might be an invalid value however
18:16:48 Quit bmbl ("Woah!")
18:17:52bertrikhm, I think I did see some aliases for the ascodec registers when looking into them for the e200v1
18:18:03bertrikso it could be that is simply aliases back to 0x15 for example
18:18:53bertrikdoes the value written to reg 0x55 make sense if it were meant for reg 0x15?
18:19:03 Part Xakh
18:21:10 Join kugel [0] (n=kugel@rockbox/developer/kugel)
18:21:35bertrikis there a way to find all rockbox wiki edits done by me, by date?
18:22:16bertrikyes, I found out how
18:23:45bertrikfunman, on the as3514 (used for example in the e200v1), the registers repeat every 0x40, so 0x55 would map to register 0x15
18:25:00 Quit Rondom ("Ex-Chat")
18:26:54 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
18:27:08preglowLlorean: sure, i can see them being used by people with poor eyesight too
18:30:01 Quit HellDragon (Client Quit)
18:30:19 Join HellDragon [0] (
18:31:41funmanbertrik: seems to be the case on AMS SoC as well ('as3514+'), thanks
18:38:13 Join notlistening [0] (
18:39:11 Quit wark ("Leaving... for now. But I'll be back.")
18:39:26 Join `VL [0] (n=vl@
18:42:23 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
18:43:37 Join planetbeing [0] (
18:43:56 Quit petur ("now sports")
18:46:18***Saving seen data "./dancer.seen"
18:47:00HillshumDoes battery_bench fully drain the battery or end with some left to prevent damage?
18:47:42linuxstbI would expect it to drain until Rockbox's normal low-battery shutdown happens.
18:48:10gevaertsbattery_bench doesn't care. It just logs things
18:48:19Tornegenerally the hardware makes it impossible to drain the battery to the point of damage anyway
18:48:26Torneunless you explicitly go and program the power controller wrong
18:48:40bertrikTorne, that depends on the specific target I think
18:48:54Tornewell, yes.
18:49:05HillshumOn the Clip, when's the shutoff point?
18:49:06bertrikok, you said "generally"
18:49:15Tornebut generally it's true :)
18:49:31bertrikHillshum, I'll look it up
18:49:35Tornepower controllers rarely need software control to prevent li-ion fires
18:49:50Tornethat's a bit of an extreme tradeoff between cheapness and safety :)
18:50:18bertrikHillshum, it shuts down at 3.3V
18:50:50HillshumDo you know (how to figure out) what percent that is?
18:52:01bertrik0% :P
18:52:38 Join JdGordon| [0] (n=Miranda@nat/microsoft/x-bd23820e17c7e540)
18:52:47HillshumCan that damage the battery?
18:52:49bertrikthe discharge curve is from 3.3V (0%) to 4.14V (100%)
18:53:07Hillshumthe other direction though
18:53:29LloreanHillshum: The whole point of cutoffs like that is to insure the battery *isn't* damaged
18:53:56HillshumSo it won't go too low?
18:54:46LloreanIsn't that answered by my previous sentence?
18:55:45Hillshumjust making sure (yes, it was implied)
18:55:48bertrikHillshum, the sansa hardware shuts down when the battery voltage is < 2.7V
18:56:46bertrikthe website of the battery manufacturer mentions a cut-off voltage of 3.0V, so the 3.3V software limit that rockbox uses is still above both of those
18:57:27 Quit Hillshum ("Page closed")
18:57:29bertrikIn battery benches I've seen so far, the discharge voltage goes down pretty linearly downto 3.6V then drops very quickly
18:57:59 Quit robin0800 (Remote closed the connection)
19:00:51 Quit kugel (Read error: 113 (No route to host))
19:02:17bertrikmarkun, for the interrupts on the meizu m3, we can keep using the "default_interrupt" macro and don't need to create a "handled_interrupt" macro or something
19:03:05bertrikI think the "weak" attribute maps the interrupt to the default UIRQ, but allows other modules to override it
19:04:15 Join robin0800 [0] (
19:04:24 Join robin0800_ [0] (
19:06:05 Join barrywardell [0] (n=barrywar@
19:07:04 Quit robin0800 (Client Quit)
19:07:35 Join robin0800 [0] (
19:07:37markunbertrik: ah, nice
19:07:47markunhow do you assign another function to it?
19:08:14bertrikI didn't, I just declare a function called TIMERA_INT
19:08:34bertrikINT_TIMERA rather
19:09:59markunbertrik: what do you want to get working now? USB? Then maybe we can dump the NAND to a file and start on the FTL driver
19:11:00bertrikI thought about doing the RTC driver, it should be very easy to get that running
19:11:13bertriknot very useful though
19:11:53markunbertrik: I remember someone wrote code for it, might be in the tracker or maybe on my PC somewhere
19:12:07bertrikradio and communication with the codec would be nice too, to see if we can get some audio out of it
19:13:07markunI don't know why communication over i2c didn't work. Probably something stupid.
19:13:39bertrikUSB just looks quite complicated to get working to me
19:14:25markunaccording to gevaerts it shouldn't be too much work. Most of the code is there.
19:15:00markunbtw, the meizu's have 2 RTC. Don't know why.
19:15:16markunMaybe the internal one doesn't get power when the device is off
19:16:06bertrikor maybe it took too much power, IIRC the dedicated external ones generally use less current
19:16:09 Join robin0800__ [0] (
19:16:12 Join stripwax [0] (
19:16:17 Join robin0800___ [0] (
19:16:17***Alert Mode level 1
19:16:17DBUGSent KICK robin0800_ to server
19:16:17DBUGSent KICK robin0800 to server
19:16:17***Alert Mode level 2
19:16:17DBUGsent MODE #rockbox +b *!*n=quassel@*
19:16:17DBUGSent KICK robin0800__ to server
19:16:17DBUGEnqueued KICK robin0800___
19:16:17***Alert Mode level 3
19:16:17Kick(#rockbox robin0800_ :*bang* too many joined users) by logbot!n=bjst@rockbox/bot/logbot
19:16:17***Alert Mode level 4
19:16:17Kick(#rockbox robin0800 :*bang* too many joined users) by logbot!n=bjst@rockbox/bot/logbot
19:16:17***Alert Mode level 5
19:16:17Mode"#rockbox +b *!*n=quassel@* " by logbot (n=bjst@rockbox/bot/logbot)
19:16:17Kick(#rockbox robin0800__ :*bang* too many joined users) by logbot!n=bjst@rockbox/bot/logbot
19:16:17***Alert Mode level 6
19:16:18DBUGQ-Sent KICK robin0800___ to server
19:16:19Kick(#rockbox robin0800___ :*bang* too many joined users) by logbot!n=bjst@rockbox/bot/logbot
19:16:19***Alert Mode level 7
19:22:06 Join Blue_Dude [0] (
19:23:53saratogai compiled cook for the e200 but it crashes with an undefined instruction, i think do to problems with the ffmpeg bitstream.h/.c files
19:23:59Blue_DudeAll right. I've been working on this thing for a couple of weeks and it looks ready for public criticism: FS #10375. It adds a transparent lookahead buffer to the DSP functions to allow a process to look into the future and process current samples referring to future samples. Any comments gratefully accepted.
19:24:07saratogai had to hack those up to get them to work with wma
19:26:20***Alert Mode OFF
19:26:21 Join Sajber^ [0] (
19:29:47 Join Setup [0] (n=chatzill@
19:32:04 Nick Setup is now known as hillshum (n=chatzill@
19:32:36 Nick hillshum is now known as Hillshum (n=chatzill@unaffiliated/hillshum)
19:40:13 Join kugel [0] (n=kugel@rockbox/developer/kugel)
19:44:05 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
19:44:56Blue_Dudekugel: Hi! Got time for a look at my latest spare time killer? FS #10375
19:46:46JdGordon|Blue_Dude: please dont bug the dev ml as soon as a patch in on flysrapy... just about all of us subscribe to the FS emails anyway...
19:46:55 Join robin0800 [0] (
19:47:25Blue_DudeSorry. I didn't know there was a FS mailing list to notify anyone of new submissions...
19:48:05AlexPBlue_Dude: You can bug people once it hasn't had any responses for a bit :)
19:48:16Blue_DudeChalk it up to too much enthusiasm.
19:48:48Blue_DudeFor all I knew, things just sat in FS unless you told someone about them.
19:48:57AlexPThe often do :)
19:49:16LloreanBut they'll sit there even if you do tell someone about them, since the key isn't "people knowing" but "people having time to do"
19:50:44CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
19:50:44*JdGordon| shuold have sent the email off he wanted to last week
19:51:20LloreanSo the attack buffer comes before the PCM buffer?
19:52:04 Join Rondom [0] (
19:52:12Blue_DudeThe attack buffer is integrated with the DSP process, and holds back 512 samples during processing. So yes, it's before the PCM buffer.
19:52:21saratogais it legal to include string.h in plugins?
19:52:26saratogasorry codecs
19:52:51linuxstbsaratoga: I don't think so, no.
19:53:15 Quit Lss (Read error: 54 (Connection reset by peer))
19:53:20saratogawell that probably explains why cook works in the sim but not on target
19:53:36linuxstbBut I'm not sure what harm it would do....
19:53:53 Join Lss [0] (
19:53:59saratogai see stdlib.h, string.h, math.h, stddef.h all over the place
19:54:01linuxstbDo you get compile errors if you remove it?
19:54:17saratogaso far no, but i probably haven't gotten rid of all of them yet
19:55:32stripwaxBlue_Dude - in what cases do you need to look at future samples? (is it equivalent to holding back past samples and outputting only when the future samples become present?)
19:55:55saratogaunfortunately, cook uses the new staticly defined huffman decode that ffmpeg added after i forked wma, while wma uses the old dynamic way but hacked to allocate into fixed buffers
19:56:12saratogaso i can't just drop cook's files into wma without changing how huffman decoding works
19:59:29Blue_Dudestripwax: I intend it for a DSP process I'm writing. I need to implement an attack function and the only way to do that is look ahead for a bit. And you're exactly right, it only let's through samples when the attack buffer is full and more samples come in. But the buffer is so much smaller than the PCM buffer that it is imperceptible.
20:02:54saratogalinuxstb, mt: heres a hacked up version of cook that takes some code from WMA and removes most of the includes that I wasn't sure about:
20:03:07saratogait still crashes on target though with an undefined instruction
20:03:34saratoga0x10141D30 on Sansa v1, which i don't even think is in memory
20:03:48saratogapatch is against current svn
20:04:04 Join BryanJacobs [0] (
20:04:15funmansansav1 has no MMU/MPU ?
20:06:24saratogai don't think so
20:06:31saratogaarm7tdmi doesn't have them i believe
20:10:19 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
20:11:08CIA-71New commit by bertrik (r21478): Implement PWM backlight driver for the Meizus. Update Meizu M3 bootloader to control brightness with the touch strip.
20:15:15 Quit flydutch ("/* empty */")
20:25:33 Quit kugel (Read error: 110 (Connection timed out))
20:25:40 Quit `VL (Read error: 60 (Operation timed out))
20:27:25 Join kugel [0] (n=kugel@rockbox/developer/kugel)
20:28:56 Join Thundercloud [0] (
20:33:16 Join `VL [0] (n=vl@
20:43:47bertrikmarkun, do you know what pins the codec is on in the meizu m3?
20:44:02bertrikI guess P10.0 and P10.1
20:46:21***Saving seen data "./dancer.seen"
20:55:12 Join mirak [0] (
20:56:09Tornedo most targets use the hold switch for dual boot and/or resetting settings? (assuming they have one)
20:56:28Torneor is that just ipods?
20:56:33linuxstbTorne: No, most don't. The ipods only do because that's the only button we can reliably detect at boot.
20:57:03Tornei was wondering because i'd rather like it not to do that ;)
20:57:22Tornethe ipod bootloader goes to OF if you hold down menu to power on..
20:57:31Tornei've not noticed that being unreliable but i've not done it often :)
20:57:38Mikachui have a patch that doesn't reset settings with hold on if the phones are plugged in
20:57:53Tornewell yah i can do a patch for it myself pretty trivially :)
20:58:00Tornei was just wondering why it was like that
20:58:29linuxstbTorne: I think it depends on the user... Lots of people complained that MENU wasn't working for them, so hold was added as an option
20:58:39linuxstb(and on the ipod - Nano users seemed to complain the most)
20:58:47 Join n1s [0] (n=n1s@rockbox/developer/n1s)
20:58:48Tornei guess i'll just patch it locally then
21:00:16Tornethen i can hit the button and put it in my pocket and nothave to wait :)
21:00:31Tornei think that's why i ws getting the boot-back-to-OF thing
21:00:43 Join CaptainKwel_ [0] (i=2669ecc2@gateway/web/freenode/x-094e771dae7343bc)
21:00:46Tornethe alarm timer that you mentioned, combined with me accidentallybooting OF by turning hold on too early
21:01:34Torne..talking of that, does rockbox already know about the alarm timer?
21:01:51Torneif so it would be fairly easy to check if it was already set to test if that theory was correct..
21:03:20Mikachuyou can set a wakeup time from rockbox, if that's what you mean
21:03:34soapthat's a neat trick, Mikachu, (headphones + hold = no reset)
21:04:02Mikachusoap: yeah, since you often want those two properties to be true while having it in your pocket
21:04:19TorneMikachu: yah, so maybe i'll try that out later
21:04:42Torneactually i say "you mentioned" but iirc that was dreamlayers actually
21:04:54Torneso i'm probably talking to the wrong person
21:05:06soapan issue I see, though, (perhaps a minor one) is that I would consider the settings reset to be a basic core function, and complicating it makes it less reliable. How sturdy is the headphone-detect switch?
21:06:01Tornewell, i was just going to patch my build to not have settings reset at all
21:06:07Tornesince i have no use for that
21:06:08soapthough, in counter, I guess you can always mount the DAP and delete the cfg file if you have hardware issues AND software issues.
21:06:13Tornewhich saves any usability issue
21:06:14 Join merbanan [0] (
21:06:40Mikachui'm not endorsing it for inclusion in svn :)
21:06:41Torneand i'd want it to not reset without headphones either, so hey
21:06:59MikachuTorne: can you write that last sentence again, differently? :)
21:07:08Tornei don't want it to reset settings ever
21:07:19Torneif i need to do that i can mount it and delete cfg
21:07:40Tornebut yah, for general use it's a desirable feature
21:07:43Mikachuif you always have a computer handy, why do you need an ipod? :)
21:07:54 Quit Blue_Dude ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
21:07:54Torneoh, i don
21:08:04Tornebut i can't imagine caring enough that i couldn't wait to fix it
21:08:38Tornei'd rather be able to not care about whether hold was on at boot. :)
21:08:53Mikachui can't seem to boot the of at all, i wonder if i did something strange (last time i fiddled with my nano was >2 years ago :)
21:08:57 Quit dberg918 ("CGI:IRC (Ping timeout)")
21:08:57Torneso, yah. i'll just do it in my build, job done
21:09:13TorneMikachu: oh? poke at the new instructions i added to the ipodpatcher wiki page perhaps
21:09:16Mikachui know i fiddled a lot with this, for example if hold was on, but the usb is in, it booted OF instead
21:09:29Mikachui haven't updated my bootloader in those two years :)
21:10:05Tornei only boot of because i haven't gotten around to testing the dma patches yet
21:10:16Torneand thus rockbox usb is slower
21:10:21Mikachui don't particularly want to boot OF :)
21:10:28Torneoh, yah
21:10:29Mikachuthere's always the rescue disk mode
21:10:34Tornerescue disk mode is slower too
21:10:37Tornethough not as much
21:10:41Mikachunot by much, since the 1.1 update
21:10:49Tornei have an ipodvideo so it's probably different
21:10:54Mikachuwith the 1.0 firmware, it did take a few minutes to copy over rockbox :)
21:10:56Tornebut it's noticable if you are copying 60GB
21:11:06soap1.1? 1.0?
21:11:12Mikachuof the ipod nano firmware
21:11:14Torneit's worth a little bit of fiddlingto boot OF first
21:11:22Tornesaves quite a few minutes :)
21:11:40soapahh, emergency disk mode is faster in Nano firmware 1.1 than 1.0?
21:11:53Mikachuthe numbers might not be 1.0 and 1.1 though :)
21:11:54stripwaxDid someone mention that the fully overlapped usb/disk io gives a better benefit than the DMA patch, or did I imagine that?
21:12:13Horschton my 80GB Video (5.5G) OF, EDM and Rockbox have huge differences
21:12:15stripwax^the fully overlapped io patch, that is
21:12:15Tornesounds like something i read somewhere
21:12:23Tornebut i've not tried any of those yet
21:12:30HorschtOF writing at 3 times the speed of the EDM and stock Rockbox
21:12:40Torneyah, that's what i have too, Horscht
21:12:46stripwaxHorscht - EDM is USB1.1 for me I think, but more recent OF versions improve that (but I never upgraded by ipod OF)
21:13:00stripwaxBut I find rockbox >> EDM
21:13:04Tornefor me EDM is still faster than rockbox
21:13:08Tornebut not as fast as OF
21:13:13Horschti have a 1.3 based firmware on my ipod
21:13:17Horschti.e. the latest
21:13:35HorschtEDM performs about the same as Rockbox for me (4Mb/s)
21:13:46Horschtstock rockbox that is
21:14:12Torneah well
21:14:31Tornekeeping the OF around for that is why i went to the trouble of patching the bootloader for dual booting from OSOS :)
21:14:48Mikachuiirc i even formated the hibernation area as a third partition
21:14:57Mikachu(when you only have 2GB, 64MB makes a difference :)
21:16:04 Quit HBK ()
21:16:35Horschtah. I remembered my results wrong:
21:16:43HorschtRockbox is indeed faster than EDM
21:17:11Horschtbut it drains the battery, i.e. I can't do a "full sync" to my ipod because it would take too long
21:17:37Torneah, the usb charging patch fixes that
21:17:42Tornei do have that one in my build
21:18:17TorneFS #8802
21:18:42 Join tessarakt [0] (
21:18:54HorschtI need to make my own build with 10239, 8802 and 9708 (and blackjack and hookers)
21:19:06 Quit tessarakt (Read error: 104 (Connection reset by peer))
21:19:09 Join tessarakt2 [0] (
21:19:50Tornewell when i get around to it mine will have those in. you can have it if you like. :)
21:19:55Torneit also has a bunch of others, though
21:20:02Horschtwell, Torne. my ipod still charges, but the drive drains the battery faster than the usb port charges...
21:20:12Horschti can make my own build, thanks :)
21:20:18Tornei've got the current measuring patch as well and tested it in various combinations
21:20:23Torneso i'm pretty sure it works
21:20:55Horschti think i might give it a shot... later... bookmarked, though
21:20:55Tornethe concern about it being committed is potentially overloading accessories and htings i think
21:21:15Torneapple have *some* kind of way for a device attached to the dock port to signal how much power it can deliver
21:21:19Mikachucould you not throttle disk writes when battery level is under some threshold?
21:21:22Torneand the patch doesn't pay any attention to that
21:21:34TorneMikachu: the problem is that current ipod builds only charge from usb at 100mA
21:21:34Horscht"overloading accesories"... that sounds bad. I have an alarm clock i plug my ipod into every night (via dock)
21:21:48TorneHorscht: tat would be a good data point for the patch
21:21:53Mikachuwhy does what it does with usb affect what it does with accessories?
21:22:06Tornethe usb-power-present thing is what's used to trigger charging
21:22:14Tornesome accessories might activate that
21:22:28Tornebut not have sufficient regulation to prevent themselves being damaged by the ipod pulling 500mA
21:22:30Horschtwell, my alarm clock charges the ipod, that's for sure.
21:22:38Mikachuah, accessories that also provide the ipod with power?
21:22:50TorneMikachu: accessories that are not *intending* to provide the ipod with power
21:22:56Mikachuwhen you use the word accessory, i think of something that draws power, not gives it :)
21:22:57Tornebut may be doing so somewhat unwillingly :)
21:22:58Horschtso... my alarm clock might break/melt/catch fire?
21:23:05Torneit's just a theory
21:23:10Tornenobody has any evidence that this will happen
21:23:14Torneit's just a concern :)
21:23:23Mikachucan't you make it an option for now, with some appropriate big flashing red warning?
21:23:27HorschtI will give it a shot then. I hope it won't burst into flames.
21:23:31Torneit is an option in the patch
21:23:40Tornei only ever connect mine to PCs via USB
21:23:44HorschtI don't think i'll find a nice ipod-alarm clock for 10€ again :(
21:23:46Mikachuso from rockbox we can't tell if it's usb for a real computer or some accessory?
21:23:54 Join HBK [0] (
21:24:00TorneMikachu: well you can, but the problemis you still want to charge fast sometimes when it's *not* a computer
21:24:05Tornenamely, when using the apple power brick
21:24:15MikachuHorscht: you don't need an accessory for that, you can set wakeup from inside rockbox, have it plugged into usb, and connect your speakers to the ipod
21:24:17Tornewhich provides loads of current on the usb pins, but no data connection
21:24:29Horschti don't have speakers...
21:24:46Mikachuoh, weird :)
21:24:50Tornei think apple's power brick provides particular inputs on the usb data pins to tell the ipod that it's a mains charger and it can draw as much as it wants
21:24:57Torneand other accessories may do similar things
21:24:59Horschtdid anyone ever measure what power the OF pulls?
21:25:01Tornebut we don't have a spec for that
21:25:11Tornei think so, ther eare various numbers on the FS#
21:25:11Horschtbecause I doubt it only pulls 100mA
21:25:15Torneno, it doesn't
21:25:21Mikachuif you want to make an ipod accessory, do you get specs from apple or do you just guess?
21:25:21Tornebut the point is they might be using additional logic to decide
21:25:28Tornethere are specs you can buy from apple
21:25:38Tornesome manufacturers just guess, though, undoubtedly
21:25:51Horschtwell... can't one just measure it?
21:26:02Tornewhat do you mean?
21:26:24Horschtby some voltmeter inbetween the charger and the ipod?
21:26:29Torneit's been measured
21:26:37Tornebut my point is that's not enough to know for sure we're doing the right thing
21:26:48Horschti see
21:26:51Tornethe OF charges at 500mA after negotiating USB with a host pc, or when plugged into a mains plug
21:27:00Tornebut we don't know *how* it decides that that's ok in the latter case
21:27:13Tornethe suspicion is the mains plug sets certain voltages on D+ and D- to tell it what to do
21:27:16Mikachuand we don't know for sure it's a host pc in the first case either?
21:27:29TorneMikachu: it's some kind of usb host in that case
21:27:36Tornebut it may not be a PC with proper regulators :)
21:27:40soapHorscht, I believe the questing is: "Is there an Apple commandment Rockbox doesn't know which says 'As an accessory manufacturer, if you don't want the iPod to pull 500ma you can drop xOhm across pins y and z and the iPod will only pull 100'"
21:27:45Torneit might be a smart dock that uses the ipod as a disk
21:27:53Tornewhich may be cheaply made and not regulate properly, assuming the ipod will do it for it
21:28:01Horschtah, i see
21:28:05Tornesoap: yes. that's much better put. thanks ;)
21:28:27Mikachudoes rockbox stop charging when the battery is full, or does the hardware handle that?
21:28:41AlexPDepends on which target I think
21:28:47Mikachuipods :)
21:29:04soapdifferent ipods have different charging circuits.
21:29:12Tornethe ipodvideo is hardware regulated
21:29:15Tornedunno about others
21:29:15soapIIRC they are all HW chargers, though.
21:29:20Torneyou just set bits to turn things on and off
21:29:29Tornethere's just a bit you set that controls 100mA vs 500mA, for this patch
21:29:46Mikachusomeone was concerned the other week that rockbox would overcharge the battery and destroy it
21:30:00soapcould happen on SW charging devices.
21:30:09HorschtTorne, can I set that "on the fly", i.e. is there a setting in the debug menu for example?
21:30:35Horschtor is it a bit in the source?
21:30:40Tornethere's a setting in the regular battery options that says "enable full power usb charging" or similar
21:30:48Torneif you turn that off it behaves as without the patch
21:30:52Horschtah, i see
21:30:56Horschtthat's nice.
21:30:57Torneso you could do that
21:31:01 Quit stripwax (Read error: 104 (Connection reset by peer))
21:31:03Torneand just remember to turn it off if you were worried
21:31:19Horschti will turn it off when my alarm clock caught fire, ok?
21:31:45Mikachuyou would probably feel it getting hot for a while first
21:32:03Horschtit's ok. it's right next to my head
21:32:16Horschti will notice when it's suddenly all bright in the room
21:32:32Tornewell have fun with that
21:32:40Torneyou probably want the current measurement patch as well
21:32:47Torneso you can check what it draws when plugged into the alarm clock ;)
21:32:57Torne..that should really get committed, tbh :)
21:33:51 Join stripwax [0] (
21:34:07 Quit stripwax (Read error: 104 (Connection reset by peer))
21:35:45Tornedreamlayers' graph shows his scaling factor is almost certainly correct
21:35:56 Join slam_ [0] (
21:35:57Torneso it looks done to me
21:38:53 Join Hendrik__ [0] (
21:40:00 Quit tessarakt2 ("Client exiting")
21:46:11HorschtI wouldn't know, I am just a clueless user
21:46:26BryanJacobsthis is very interesting, I always wondered why my iPod slow-charged from an off laptop
21:48:44 Quit n1s (Read error: 110 (Connection timed out))
21:50:11Horschtyour off laptop charges?
21:50:27Horschtmine doesn't. If I turn it off, it's off. No power from USB
21:50:50BryanJacobsmine has a BIOS option to continue to provide power
21:51:13BryanJacobsbut obviously it doesn't negotiate USB
21:53:11 Quit Sajber^ (Read error: 54 (Connection reset by peer))
21:53:13TorneBryanJacobs: it will slow-charge from an on laptop too...
21:53:20Torneat least without the above patch :)
21:53:36 Join tessarakt [0] (
21:55:42 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
21:55:45BryanJacobsTorne: under OF too
21:55:51Torneoh, yes
21:56:21Tornethe OF definately will do that ;)
21:56:35Tornethat's how you are supposed to do USB :)
21:56:47Torneforbidden to draw more than 100mA from USB without negotiating it with the host.
21:57:31gevaertsunfortunately, if you do that, people complain
21:57:34Torneand the OF keeps the lcd on while charging too, so it draws most of that just to power itself and barely charges
21:57:39BryanJacobsI never read the USB specs, I just reasoned "if the computer can provide fast charging while it's on it should provide fast charging while off too"
21:58:02BryanJacobsit doesn't have to keep the LCD on... you can click "sleep" and then it doesn't
21:58:22Tornei must be dumb :)
21:58:29Tornei never managed to get it to go to sleep while plugged in..
21:58:42Tornebut, hey. i can charge properly in rockbox now so who cares :)
21:58:43BryanJacobswell, the other day I discovered how to turn the thing off without using a menu option under Rockbox
21:58:55BryanJacobsI had missed that for years
21:59:11BryanJacobsapparently you hold the menu button
21:59:18BryanJacobsor was it the play button?
21:59:20Torneno, the play button
21:59:23BryanJacobsyeah, it was the play button
21:59:30Horschtaha wait...
21:59:50Tornesame as OF uses for sleep
21:59:50Horschtyou are talking about actualy shutting the player off... yeah i remember that conversation
22:00:01BryanJacobsHorscht: ^_^
22:00:14HorschtI thought you found a way to turn of the lcd while charging
22:00:22BryanJacobsI have one for the OF
22:00:31BryanJacobsyou just click "sleep" on the menu when you're plugged in
22:00:56*Horscht tries
22:01:33LambdaCalculus37bertrik: Nice work with the Meizu M3 so far. :)
22:01:37Horschtwaiting for of to boot...
22:01:53bertrikLambdaCalculus37, thanks, you have an meizu m3 I suppose?
22:02:40Horschtwell... I gues that works
22:03:05BryanJacobsHorscht: why so skeptical? That always worked for me
22:03:24BryanJacobskeeps the charge info on the screen but turns off the backlight
22:03:36LambdaCalculus37bertrik: Sadly, no.
22:03:49 Quit Hillshum ("ChatZilla 0.9.85 [Firefox 3.5/20090616224221]")
22:03:54 Quit barrywardell ()
22:03:59Horschtbecause I never cared to try... :)
22:04:27bertrikLambdaCalculus37, I borrowed the meizu m3 from gevaerts and an atmt from AlexP at devcon
22:05:17LambdaCalculus37bertrik: markun told me about that.
22:06:17LambdaCalculus37 <−− Now, I know these numbers are possibly outright wrong, but can someone help me get good screen dimensions setup in rockblox for the SA9200 screen?
22:06:25TorneBryanJacobs: aah, that's what i meant
22:06:29TorneBryanJacobs: it still has the LCD on
22:06:37Tornejust not the backlight.
22:06:53Tornewhich means it charges slower than rockbox, once you fix rockbox's charging with the patch :)
22:07:03Tornesince rockbox will turn the whole lcd off
22:07:30*gevaerts thinks that this belongs in #apple :)
22:07:41Torneyah, ok.
22:07:55Tornejust test that patch so we can get it committed instead, folks, and charge in RB :)
22:07:55BryanJacobsTorne: killer feature :-P
22:08:28BryanJacobswhat will this do to things like car chargers?
22:08:35Tornewe're not sure
22:08:37BryanJacobstheir inverters might not like high current
22:08:38Tornethat's why it needs testing
22:08:57Torneit works fine for PCs and for the Apple mains adapter.
22:09:00Torneanything else is unknown
22:09:06BryanJacobshmm, sounds like danger pay - give me double my normal salary and I'll start next week!
22:11:23saratogapretty sure the inverter in a car can handle a 500mA ipod if it can start a 100 Amp car engine
22:11:37Horschtzero times two is how much nowadays?
22:11:38saratogaerr not start, but sustain
22:11:38BryanJacobsI meant the AC/DC transformer
22:11:51BryanJacobsHorscht: exactly :-D
22:12:14BryanJacobsthat's still called an "inverter" these days, right?
22:12:24BryanJacobsthe thing that takes car power DC and converts into wall-plug AC
22:14:02soapevery charger I've tried works fine with the unrequested 500ma.
22:14:41saratogathe car's engine drives an alternator which is then rectified to give DC
22:14:51 Join stripwax [0] (
22:15:14soapthe concerns I've heard raised are regarding accessories (a big unknown) and situations like hubs, where one 500ma device might not break the current bank, but multiple devices on the hub and you're talking real juice.
22:15:17saratogausually a car charger then has a voltage regulator that steps that down to 5v
22:15:20gevaertsI'm not worried about dedicated chargers being unable to supply 500mA. The problem is other devices that do USB host, and hubs
22:15:49saratogai doubt any accessory can't suppy 500mA safely
22:15:53saratogathats not very much power at all
22:15:58BryanJacobsI suppose I could test against my old Zaurus, a USBotG host
22:16:15BryanJacobsif anything's going to fail that'd be it
22:16:19Tornewell the problem is not *testing* so much
22:16:31Tornethere almost certainly *is* a device out there, somewhere on earth, which it will asplode.
22:16:32soapwhat would be the point of failure? The voltage regulator?
22:16:42Tornethe proble is knowing if apple have a magic method to prevent this that we should've implemented
22:16:44saratogai think the only place you really have a risk is when you have multiple devices on an unpowered hub
22:16:55Torneand the only way to find that out is really to reverse it or to get a copy of the accessory spec ;)
22:16:55saratogahow does the USB spec intend to deal with that situation?
22:17:15Tornesaratoga: USB hubs are generally expected not to have enough ports for that to go wrong
22:17:17soapwon't a PC shut off the port if it pulls more than spec?
22:17:20Torneif they are unpowered.
22:17:31gevaertsIf we find a USB data connection, we ask. Easy. If we don't, it gets complicated. This could be a charger, but it could also be something that can only supply 100mA and a broken cable
22:17:40BryanJacobsTorne: you can chain hubs and get problems, I knew a guy who did that
22:17:49TorneBryanJacobs: not if the hubs are implemented properly
22:17:57Tornean unpowered hub should only really be giving its children 100mA each
22:18:02Torneand the second hub should know that the first hub said that
22:18:07Torneso it should in turn probably just refuse to operate :)
22:18:21gevaertsexactly. I don't see chained hubs as a problem
22:18:42Tornebasically we should be fine doing the obvious thing for things that actually have a USB data connection
22:18:44gevaertsWhat if you have hubs that don't have power switching?
22:18:49BryanJacobsthis guy had a Pertelian that would brighten or dim based on how many things were plugged into the upstream hub
22:18:58Torneif the host said we can draw 500mA and we do and itt fails it's not our fault really.
22:19:05Torneit's things *without* the data connection that's the issue..
22:19:41Horschthow much are the specs to buy from apple? :p
22:20:04gevaertsTorne: we could take the easy way, and do what the iriver h300 firmware does (and also rockbox on h300), i.e. make it a setting
22:20:17Tornewell the patch does make it a setting
22:20:20gevaertsHorscht: apple can't solve this problem either. They also have to guess
22:20:38Torneer, but we have reasont o believe they don't guess, no?
22:20:52Tornethere was some suggestion somewhere that the D+ and D- pins are used to signal it
22:20:59Tornebut we don't know exactly how
22:21:04LloreanTorne: But that would only work with Apple accessories, right?
22:21:08saratogagevaerts: i am skeptical that powered devices exist that cannot supply 500mA
22:21:09LloreanSo they still have to guess everywhere else.
22:21:11soapapple can solve the problem if they have a spec which says "thou shall set dock pin X high if you can't give 500 or suffer the consequences."
22:21:16Torne'everywhere else'?
22:21:21gevaertssaratoga: can the h300 supply it?
22:21:22Tornethe probelm is already solved for real usb hosts
22:21:25TorneUSB itself solves it
22:21:36Torneand if people don't implement their hosts properly that's not apple's problem any more than it is ours
22:21:37saratogagevaerts: ?
22:21:46gevaertssaratoga: it does USB host
22:22:03Torneand if apple's system defaults to 100mA then it's unlikely that some random device would accidentally do the right magic with D-D+ to trigger fast charging
22:22:07saratogadoes it supply any power at all?
22:22:09BryanJacobsthis is why I was going to test my Zaurus - it's about the smallest device I know that does USB host
22:22:12Tornethis is all speculation, though
22:22:20Tornei have no personal knowledge of this, just going by random stuff from the internet :0
22:22:21saratogai thought most of those USB host devices assumed that the peripheral had its own power
22:22:35Tornesaratoga: not if they implement USB correctly
22:22:44 Join wincent [0] (
22:22:45 Quit funman (Read error: 60 (Operation timed out))
22:22:46TorneUSB always expects devices to negotiate for power :)
22:22:53BryanJacobsUSB keyboards would never work if people assumed all devices were powered
22:22:59saratogaare there actually power pins on an H300?
22:23:07gevaertssaratoga: it should provide 5V. It's allowed to restrict that to 100mA, but the problems start when negotiation fails for whatever reasons
22:23:09saratogai don't think usb keyboards are supposed to work on the H300
22:23:35BryanJacobssaratoga: still talking about the Zaurus, another tiny USBotG device
22:23:50gevaertsBryanJacobs: actually, is it really OTG?
22:24:03TorneBryanJacobs: the point is that it doesn't matter really :) if rockbox or the OF manage to kill the zaurus like that, then the zaurus is faulty. :)
22:24:33TorneBryanJacobs: the problem is with things that are *not* usb hosts, but still supply power on the USB 5v pin
22:24:33 Join Sajber^ [0] (
22:24:33BryanJacobsgevaerts: the Zaurus says it is
22:25:31saratogafor USBOTG devices, can't we just ask if they support power?
22:25:43saratogai thought the problem was non-USB devices like chargers and docks
22:26:10BryanJacobsTorne: I'm not sure that can be assumed - if Rockbox kills the Zaurus and the OF doesn't, Rockbox is from a user's perspective faulty
22:26:10BryanJacobsI was planning on plugging in with the USB host functionality disabled
22:26:10DBUGSent KICK BryanJacobs to server
22:26:10BryanJacobsand just providing power
22:26:11Kick(#rockbox BryanJacobs :No flooding!) by logbot!n=bjst@rockbox/bot/logbot
22:26:13gevaertsBryanJacobs: I'm asking because many people seem to confuse OTG with host support
22:26:20 Join BryanJacobs [0] (
22:26:26BryanJacobsarr, flood protection
22:26:34 Join n1s [0] (n=n1s@rockbox/developer/n1s)
22:27:04TorneBryanJacobs: we're probably not doing anything the OF doesn't in the case where there's a usb data connection, though
22:27:12TorneBryanJacobs: with or without the patch
22:27:18Torneso that *shouldn't* happen
22:27:24BryanJacobsTorne: again, no data connection, just providing power
22:27:33Torneah. then yes :)
22:28:00 Join funman [0] (n=fun@rockbox/developer/funman)
22:29:04saratogathe wikipedia page implies that hooking a device up to a USBOTG device that can't supply enough power should result in the device failing to work
22:29:51BryanJacobssaratoga: "the device" being the client?
22:30:00gevaertsthat's still for USB things. Charging really wasn't considered when these specs were written
22:30:05saratogadevice being they keyboard or whatever
22:30:26saratogai don't see how drawing 400mA to charge and drawing 400mA to run a usb disk drive is much different
22:30:55 Join Grahack [0] (
22:31:13gevaertsif you run a USB disk, you *ask*
22:31:36BryanJacobswhat we're talking about here is just *taking*
22:31:49gevaertsIf you connect a usb disk to a pure charger, *nothing* happens
22:32:00 Quit tchan (Remote closed the connection)
22:32:05saratogai'm pretty sure they still spin up
22:32:34 Quit slam_ ("Leaving")
22:32:36 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
22:32:45Tornesaratoga: they are not supposed to, unless they can spin up in under 100mA
22:32:53Tornea correctly implemented usb disk will *not* spin up :)
22:32:57 Quit stripwax ("")
22:32:57BryanJacobssaratoga: mine doesn't spin up when it's connected to the same off laptop I was talking about
22:33:04 Quit `VL ("happines is a positive cache flow")
22:33:22saratogato be clear, the problem here is when you recieve power from a device that can't talk over usb right?
22:33:41 Quit Horscht ("Verlassend")
22:33:48saratogathese are chargers, which we agree can provide 500mA, and what else?
22:34:07Torneer, not all devices that give power but don't talk usb are chargers
22:34:21Tornethat's exactly the problem :)
22:34:22saratogaread last clause of sentence
22:34:24gevaertsThere are two main cases there I think. Chargers, and 500mA really shouldn't be a problem at all, and normal USB ports that don't negotiate for whatever reason (software issues, slightly broken hardware,...)
22:34:41LloreanA lot of chargers with USB-style plugs provide 1000mA
22:34:44LloreanDo any provide less than 500?
22:35:06*Llorean seems to also recall 6V ones.
22:35:09Tornesaratoga: and the evidence for this? :)
22:35:19saratogathey'd blow up if anyone tried to use them
22:35:27gevaertsThose would be broken by design
22:35:29Tornenot if they were ipod accessories
22:35:34Lloreansaratoga: Unless they came with something, and said "Don't use them with anything else"
22:35:35Torneand thus would only plug into ipods
22:35:46Torneand signalled to the ipod in a way we don't have docs for
22:35:49saratoga"USB-style plugs"
22:35:49LloreanI mean, people regularly plug the Nokia charger into the H100 and break it
22:36:00Tornesaratoga: but they might not be
22:36:10Torneit might be anything that attaches to the dock connector and powers the usb 5v pin
22:36:12gevaertsLlorean: most people only do that once though
22:36:18LloreanI guess the real question boils down to "do we try to be safe with everything you can throw at it, or make the assumption that people won't try to use it with strange chargers"
22:36:39gevaertsI vote for using the h300 way. Provide a setting, and default to 100mA
22:36:47Lloreanif it's the latter, the proposed solution is already good enough because we don't really need to worry about non-USB <500mA device, or at least shouldn't.
22:36:47saratogaa user can always disable any overcurrent protection you try to design if they really want
22:36:54gevaertsexcept of course for real USB connections
22:36:55saratogawe should assume they know what they're doing when they plug in a charger
22:37:06Tornegevaerts: i'm not sure if the patch currently does do that
22:37:13Lloreangevaerts: So default to 100, but try to negotiate, and do 500 if we can regardless of setting?
22:37:25TorneLlorean: the slight issue there is it's not really a negotiation
22:37:27BryanJacobsgevaerts: "most people only do that once"?!
22:37:32 Part Grahack
22:37:33Torneyou say "i want 500mA"
22:37:38Tornethe host says "sod off then"
22:37:40Torneyou now can't connect
22:37:43gevaertsLlorean: yes. If the host tells us that 500mA is OK, and it isn't, well, there's nothing we can do
22:37:43BryanJacobshow many people break their H100 twice the same way? :-P
22:37:51LloreanTorne: Well, by "negotiating" I mean "doing whatever it is USB devices are supposed to do to get 500 from a real host"
22:37:52Torneyou have to actually tear down the usb connection and then start over to re-request with a lower current
22:37:57Tornewhich we don't have code for
22:37:59gevaertsBryanJacobs: there is such a thing as getting anew one :)
22:38:11Torneand thus if someone's host port was *actually working* and reporting it can't provide enough, USB mass storage stops working
22:38:13saratogait breaks the charger or the mp3 player
22:38:14LloreanTorne: But we're open source, so such code can be invented.
22:38:16Tornee.g. if they have a laptop
22:38:17gevaertsTorne: no, but conceivably we could add it
22:38:18Torneyah, it can
22:38:21Tornebut it's not in the patch atm, i mean
22:38:35 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
22:38:40LloreanBut we're discussing what the patch needs too, since that's part of "can it be used now?"
22:38:50Torneok. so it'd probably be nice if it did do that. :)
22:39:18Tornebut it's not fatal if it doesn't
22:39:19gevaertsit would be nice, but not required I think
22:39:22HorschtTorne, you mentioned a setting which woul allow me to select the current...
22:39:22Tornebecause you can just disable the charging setting
22:39:23Torneand plug in again.
22:39:26Horschti can't find it
22:39:30*Llorean is pretty sure apple claims the iPod won't work on low-power ports on their website.
22:39:41TorneLlorean: rockbox currently does work on low power ports, though
22:39:55*gevaerts wishes that people would stop talking about ipods. This is a general problem!
22:40:04LloreanTorne: Rockbox currently doesn't charge at all.
22:40:06saratogaback in the day apple would provide you with a dock connector that could plug into a USB port and an AC wall socket
22:40:08TorneLlorean: yes it does
22:40:16TorneLlorean: it charges at 100mA minus the power tit takes RB to idle
22:40:18LloreanNot on the iPod.
22:40:19Tornewhich is about 60mA or so
22:40:22saratogathus side stepping the issue neatly
22:40:24Tornei've measured it
22:40:29HorschtLlorean, it does, but not if you transfer files
22:40:30Torneit charges extremely slowly if the ipod is idle
22:40:36Torneand discharges if you are using the disk.
22:40:49LloreanIt depends on your settings
22:40:56Tornewell yes.
22:41:00LloreanMany people have it set to turn the backlight on if plugged in, which I imagine negates that.
22:41:02Tornebut it *tries* to charge
22:41:04LloreanSo perhaps "it doesn't reliably charge"
22:41:18 Quit Thundercloud (Read error: 60 (Operation timed out))
22:41:32*gevaerts thinks that the patch is fine as it is, in the sense that more can be added later, except for the questions about suitability for all ipods
22:41:34BryanJacobsLlorean: it does charge, it just usually discharges more
22:41:37Lloreangevaerts: I thought the problem was really just iPods (in terms of the patch under discussion)
22:41:39 Join Thundercloud [0] (
22:41:40Tornegevaerts: yes, i would agree
22:41:46Torneit should be fine for ipodvideo, at least
22:41:50LloreanBryanJacobs: "Charging" is a net effect. It draws power, but still discharges.
22:42:11BryanJacobsI viewed "charging" as a gross effect
22:42:16TorneHorscht: settings, general, system, battery, charge during usb connection
22:42:24*soap has had issues getting the patch _and_ battery bench to work at the same time.
22:42:26gevaertsLlorean: yes and no. The patch is about enabling charging on ipods. It doesn't address behaviour when no USB connection is there, and that's a general issue
22:43:01LloreanBryanJacobs: You could view it as "the iPod draws 90% of its power from teh cable, and 10% from the battery" mentally as well, as opposed to "100% goes into the battery, but then a bit more than that comes right back out"
22:43:02Torneor the general issue of what if the host says no to 500mA
22:43:22Lloreangevaerts: Well, the Sansa's apparently getting away with whatever our current behaviour is without problems? Can't we just mirror that?
22:43:22Horschthm... no menu entry there... Gonna rebuild later
22:43:29TorneHorscht: oh, er
22:43:35TorneHorscht: you need to edit config-ipodvideo.h as well
22:43:42Torneto add the "i have usb charging" feature
22:43:46Tornethat's not in the patch
22:43:47BryanJacobsLlorean: except that view doesn't describe what's actually happening
22:43:51gevaertsLlorean: that will happen once the patch is in
22:43:51Torneit says in teh FS# though
22:44:05Lloreangevaerts: I didn't think the Sansas had a menu option.
22:44:06BryanJacobslaptop batteries heat up when the laptop is plugged in
22:44:36*gevaerts turns off the music. He can't think properly
22:44:55LloreanBryanJacobs: It depends on how the hardware is hooked up. Sometimes current goes into the battery first, then out and into the player. Sometimes it doesn't. I don't know what order it happens in the iPods. Can you turn yours on with the battery disconnected?
22:45:00gevaertsThere are about seventeen different behaviours by different players right now
22:45:48BryanJacobsLlorean: dunno, but after running dry it won't turn on for a while plugged in, so survey says all power flows through the battery
22:45:59gevaertsand I think all of them are wrong in some way
22:46:21 Quit Hendrik__ ("ChatZilla 0.9.85 [Firefox 3.0.11/2009060215]")
22:46:23***Saving seen data "./dancer.seen"
22:46:50gevaertsThe beast charges when USB is connected and the host OKs the 500mA, but not in any other conditions (i.e. not on a pure USB charger)
22:47:03gevaertsThe sansas always charge (I think)
22:47:07LloreanBryanJacobs: Have you tried menu+select reset to try to boot it? I've never had problems booting it, at least up until Rockbox checks the battery level, decides it's too low, and immediately shuts back down.
22:47:09gevaertsThe ipods never charge
22:47:40BryanJacobsLlorean: never tried that, but it does go on on its own after a while in OF
22:47:51TorneBryanJacobs: yah, you may be encountering the *other* problem with booting the ipod, the whole hard-reset-needed thing
22:48:01Tornewhich there is also a patch for which needs testing/work, incidentally ;)
22:48:40Lloreangevaerts: Speaking of the Beast - can we do anything about the USB charging capabilities while off or is that something we'd have to reflash for (if you even know)?
22:48:53*Llorean just wonders if it's about the state we leave the hardware in at shutdown, or something else.
22:49:00 Quit n1s (Success)
22:49:45saratogawhen the CPU is powered down i'd imagine the register settings in the charger are lost
22:49:46gevaertsI have no idea
22:49:59saratogasince i'm 90% sure its an on board charger like on the sansas
22:50:37LloreanThat's my least favorite thing about the Beast. If it's "low" (which can mean at 70% on some days) you can't boot without the real charger attached
22:51:07saratogasomeone send karl a beast with jtag attached
22:53:47 Quit funman ("free(random());")
22:54:11 Quit BryanJacobs ("Java user signed off")
22:56:15 Quit LambdaCalculus37 ("Python time is now")
22:56:54HorschtTorne, do I simply have to add "#define HAVE_USB_CHARGING_ENABLE" before the last #endif?
23:00:04 Quit Cory` ("There are two major products that come out of Berkeley: LSD and UNIX. We don't believe this to be a coincidence.")
23:04:38 Join Thundercloud_ [0] (
23:05:22 Quit Thundercloud (Read error: 54 (Connection reset by peer))
23:07:26UnhelpfulLlorean: the flash shuts it off if it's "too low". i think our only solution is RB in flash. my WinMo phone has the same problem, incidentally.
23:07:53LloreanUnhelpful: Yeah, but the flash stuff will recognize a normal charger and ignore the battery state.
23:08:08LloreanI was wondering if there was a way to leave it "waiting" for a USB charge too when we shut down, or something.
23:08:16Unhelpfulyes, it will.
23:08:18 Join Cory` [0] (
23:08:48*Llorean likes the fact that the Gigabeat F can even USB charge while turned off, and finds some disappointment that the S has such disappointing USB charging.
23:13:40AlexPThe S is disapointing in many ways
23:13:47AlexPEspecially as it has such promise
23:22:12CIA-71New commit by funman (r21479): fix a typo in comments (audio CD duration is not 74 or 80 hours)
23:22:19 Quit merbanan (Remote closed the connection)
23:23:03 Join stripwax [0] (
23:31:06 Quit ender` (" The reward for work well-done is more work.")
23:32:10*GodEater coughs
23:32:21GodEaterapparently we all missed Lamba's other expert nomination
23:32:28GodEaterfrom god knows how long ago
23:32:31scorche|shwhich was?
23:32:52scorche|shah...another furry..
23:33:07 Quit stripwax (Read error: 60 (Operation timed out))
23:33:37LloreanDoesn't really have anything to do with their ability to be an expert though.
23:33:41*Llorean can't really check any more.
23:33:45LloreanDid I comment on Karashata at the time?
23:33:50GodEaterno-one did
23:34:16GodEaterhe seems very full of clue
23:35:44LloreanI did see the nomination. I'm not sure why I didn't comment. I remember at one point not entirely feeling confident in him, but looking back through his posts I can't seem to see what would've made that happen.
23:35:59AlexPSeems fine to me
23:36:19GodEaterhe seems clueful about the H10
23:36:22GodEaterwhich is useful
23:36:27GodEaterI know very few people that even own one
23:36:30AlexPNobody else is
23:36:42LloreanAt the time, I wanted experts to be people who were trying to help about everything, when they had time.
23:36:45LloreanThat may have been it.
23:37:08*GodEater thinks that the time for generalists, if not already over, is rapidly approaching
23:37:14GodEaterwe have SO many targets coming up now
23:37:25LloreanThe goal (for me) wasn't to have "many" but to have a very high degree of confidence when they got involved.
23:37:30GodEaterI wouldn't care to try to advise on anything AMS based at the moment
23:37:31AlexPI think that as long as they don't talk rubbish about others, it isn't so important if they just stick to their area
23:37:34LloreanBut definitely now, generalists are less useful.
23:37:56LloreanThe key should always be, they need to be people who won't give advice unless it has a high degree of confidence.
23:37:58GodEaterindeed, and I see nothing there where he's jumped into something he clearly knows nothing about
23:38:05AlexPLlorean: yep
23:38:18AlexPLooking through his posts, looks good tom e
23:38:30AlexPAny objections to a PM?
23:38:31*GodEater will drop him a PM then
23:38:36GodEateror you can :)
23:38:38AlexPThere we go :)
23:38:44AlexPWill do :)
23:38:54GodEaterthen you can work out how to promote him :D
23:39:03AlexPGood practice :)
23:39:06GodEaterassuming scorche didn't already show you
23:39:10LloreanGodEater: Go to view their profile, click their "Account Settings" link, and there's a dropdown for their primary group
23:39:23GodEaterLlorean: I know, I did evilnick already ;)
23:39:26LloreanAh, good.
23:39:33LloreanI hear you tried a ban earlier.
23:39:48GodEateryes, less successfully
23:39:52GodEatergot on the second try
23:40:02 Join safetydan [0] (n=deverton@rockbox/developer/safetydan)
23:40:05LloreanI personally only ever banned username and email address for spammers.
23:40:12GodEaterseems most useful
23:40:18GodEaterI'll stick to that as they turn up
23:40:21Lloreanit seemed unlikely the host would be repeated, so it was really just killing off the account.
23:40:48*GodEater waves to safetydan
23:40:55GodEaterseems the lua code is proving popular already :)
23:41:05 Quit Cory` ("( :: NoNameScript 4.22 :: )")
23:41:08LloreanOne of my favorite games was written in Lua. Want to port it?
23:41:14GodEateryeah, we've had someone in here today asking all sorts of difficult questions about it
23:41:20scorche|shLlorean: yeah...i mentioned a few things like that to him..
23:41:24scorche|sh(re: banning)
23:41:39GodEatermentioned it?
23:41:41GodEateryou mocked me!
23:41:50GodEater"you suck you suck"
23:41:53safetydanLlorean: if I actually knew how to write lua... :)
23:42:10GodEaterLlorean: daurn seems quite proficient
23:42:14GodEatermaybe you could ask him
23:42:22scorche|sh"nice reaction for the ban...too ban it isnt actually a ban on anything..."
23:42:40scorche|shs/too ban/too bad
23:42:48 Quit mirak ("Ex-Chat")
23:44:20LloreanIt only required a 133mhz processor and 32MB of RAM. It should be no problem!
23:44:24*GodEater is still chuckling at cool_walking_s intro in the experts thread
23:44:47*Llorean will admit he misses seeing a few parts of the forum sometimes already.
23:45:01LloreanHabits may take a little while to relax out of.
23:45:04GodEateryou can ask for a cut and paste anytime
23:45:08*scorche|sh can make a "Llorean" usergroup with read-only permissions..
23:45:28GodEaterwith population one
23:45:33GodEaterto go with the "user" cloak :)
23:45:39scorche|shmuch like @rockbox/user/
23:45:40LloreanNah. I'm happy.
23:46:08scorche|sh(which i havent mentioned still allows op rights here, but i dont have to say anything, really..)
23:47:42GodEaterLlorean: for your peace of mind, it wasn't karashata you had reservations about, it was cool_walking
23:48:02LloreanGodEater: No, I had reservations about a lot of people. :)
23:48:18*scorche|sh still has reservations about a lot of people..
23:48:21GodEater"People. I am suspicious of them"
23:48:35GodEaterwell lots of people have reservations about you scorche|sh
23:48:38GodEaterand rightly so!
23:48:49scorche|shyes, but i already have a shiny badge =)
23:48:55*Llorean has now places all the responsibility in other hands, so doesn't need those reservations any more.
23:49:37GodEaterLlorean: my view is really, "what's the worst that could happen?"
23:49:47GodEaterif any of the new peeps screw up bad, they lose the badge
23:49:49GodEatereasy enough
23:50:09GodEaternot that I expect that to happen of course
23:50:25AlexPGodEater: I'll remove yours forthwith :)
23:50:41GodEaterAlexP: thanks
23:50:55AlexPalways happy to help :)
23:50:56GodEatercan I have that tarnished rusty badge lying in the corner over there instead ?
23:51:05*scorche|sh isnt sure if AlexP can or not
23:51:19scorche|shor it might be equal level and below, so you could..
23:51:27AlexPscorche|sh: I haven't tried :)
23:51:34AlexPNor am I planning to
23:51:36*GodEater can't be bothered to look
23:51:48GodEateranyway, it's now past my bedtime
23:51:51GodEaterso nn all
23:51:53notlisteninghas anyone suggested a increasing pitch for the FM tuner feature when manual tuning the radio for blind users?
23:52:02Lloreanscorche|sh: If I recall, the "Membergroup" permission is automatically equal level or below, but the levels are only "Admin" "Global Mod" and "other membergroups"
23:52:35scorche|shLlorean: so AlexP could remove GodEater and restore him again
23:52:59LloreanI think
23:53:41JdGordon|notlistening: that would be difficult because i dont think the fm audio goes through the software dsp at all... it goes stright to the headphones
23:54:11LloreanJdGordon|: I don't think that's what he means.
23:54:18 Quit nibbler_ (Read error: 110 (Connection timed out))
23:54:22LloreanJdGordon|: I think he just means playing a tone that increases as you increase the station
23:54:27Lloreanincreases in pitch
23:54:46JdGordon|well in that case.... /me shuts up
23:54:55*Llorean isn't sure though, as it is a bit unclear
23:55:29notlisteningyeah an increase in pitch to indicate you position in the frequency range
23:56:19LloreanIt'd still be wildly inaccurate. I'm not sure it'd be terribly useful.
23:56:37LloreanA button to "read" the station name to you is probably better.
23:56:48LloreanJust tune until something is coming in, then press the button to read it to find out what you've landed at.
23:57:06Mikachuyou would have to have perfect pitch to know what it is :)
23:57:29notlisteningyeah but to have an idea where you are helps a lot trust me, blind users are not so much into precision at times in the ball park usually does
23:57:41Lloreannotlistening: What does it help with?
23:57:59Mikachulike you know something is around 104?
23:58:18notlisteninga rough idea where you are in the range
23:58:25LloreanYeah, but wouldn't an exact idea help more?
23:58:28 Join Cory` [0] (
23:58:40LloreanIf you tap and hear "nine six" it's a lot better than trying to gauge from the tone whether you're in low, mid, or high 90s
23:58:45LloreanIf you're looking for, say, 95.
23:58:56CIA-71New commit by bagder (r21480): better upload path, fixed logging timestamp, lowered to 3 parallel builds per client by default, counts build/kill time, counts kills, renames uploads ...

Previous day | Next day