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

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

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

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

#rockbox log for 2009-06-21

00:00:45 Quit Ubuntuxer ("Leaving.")
00:00:56CIA-70New commit by 03alex (r21423): Much cleaner version of r21411 - add three column button tables capability to the manual. To use, add remote_keymap to the platform file, then button ...
00:03:35 Quit ender` (" Working computer hardware is a lot like an erect penis. It stays up as long as you don't fuck with it.")
00:04:53CIA-70New commit by 03zagor (r21424): Fixed zip name. Fixed core count.
00:05:57 Quit robin0800_ (Remote closed the connection)
00:14:29fmlAlexP: can't we define btnmap with three params in the remote case? If the third param is given then the table would still be defined with two columns thus allowing case by case adjusting
00:15:37 Part grimman
00:15:40AlexPfml: I'm not sure I undertand - If the table has two columns how would we get the three for Main, Remote, and Description?
00:15:52CIA-70New commit by 03alle (r21425): Fix a typo
00:16:37 Quit perrikwp (Read error: 110 (Connection timed out))
00:16:54bluebrotherAlexP: \opt inside of the btnmap environment works? Nice to know :)
00:17:02AlexPbluebrother: yup :)
00:18:00fmlAlexP: the btnmap env would accept three params. The third, if given, would set the table with two columns −− even in the "remote" case. It's just a hack to enable table by table adjustment, not all in one shot.
00:18:53Mikachuis it that much work to fill in this third column for all targets?
00:18:55fmlThe value of the third param would be ignored, just its presence would be tested
00:19:15AlexPfml: Sounds interesting, by all means add it :)
00:19:18bluebrotherfml: the btnmap environment doesn't get its data as as parameter. Changing the environment to use some kind of bool value would require changing all useage of it
00:19:26AlexPMikachu: Just very tedious
00:19:47AlexPMikachu: You have to go through every table adding the remote button for every action
00:19:58AlexPfml: Or don't :)
00:20:06Mikachuand look it up from somewhere too i suppose
00:20:15*bluebrother would consider it easier to create a new environment for that −− like having a 2column and 3column in the remote case
00:20:24AlexPMikachu: How could you do it automatically?
00:20:34Mikachuwhen did i say you could? :)
00:20:46bluebrotherAlexP: you know about sed? ;-)
00:20:54AlexPMikachu: Well I assumed you meant that by looking it up :)
00:21:00AlexPbluebrother: Indeed so :)
00:21:07Mikachuno i meant finding out what buttons to fill in
00:21:21bluebrotherwell, but doesn't make a 3rd row the tables look cramped like hell?
00:21:24fmlAlexP: is there something like \else for the TeX's \if\given?
00:21:28AlexPbluebrother: I started off making a seperate environment, but \end{...} doesn't work inside an \opt
00:21:38AlexPfml: bluebrother is the expert here :)
00:22:00AlexPMikachu: Of course, but looking them up then filling them in is the tedious I was referring to
00:22:01 Join funman [0] (n=fun@rockbox/developer/funman)
00:22:12Mikachuokay :)
00:22:31Mikachubut the work is finding the button in the code and putting it in the manual then, not just moving it from one place in the manual to another?
00:22:44funmanbertrik: big up for fixing the e200v2 FM :)
00:22:48AlexPIt isn't in the manual, so there is no moving to be done
00:23:20Mikachuyes, that was my question :)
00:23:28AlexPah right :)
00:23:38bertrikfunman, tested it already on your c200v2?
00:23:55bertrikfunman, yeah, a lot of work for a tiny fix
00:25:43matslBagder: These guys on the floor, are they building the tower or what?
00:25:43bertrikBTW your lcd patch brought fps back up on my c200v1 (49.something IIRC), can't test now, my c200v1 is currently being added to the tower of rockbox
00:26:03martian67TOWER OF POWER
00:26:18fmlAlexP: where should I add the remote_keymap macro?
00:26:53Bagdermatsl: they're working hard on the art of towering
00:27:20*matsl remembers a stack of just archos players, those were the days ;-)
00:27:32GodEaterthey need to use proper blueprints these days
00:27:37AlexPfml: It goes in the manual/platform/xxx.tex in the form of \edef\UseOption{\UseOption,remote_keymap}
00:27:43GodEaterand a lot of civil engineering math to calculate wind shear
00:30:27Bagdermatsl: last year's:
00:31:41Horschtsee: no docks
00:31:44Horschtdocks are cheating
00:32:32matslBagder: good to see those archos units at the base making it stand on solid technology
00:33:20funmanbertrik: i just tried: no HW detected on 2 boots, i'll try with an higher delay in fm i2c
00:33:37funmanbertrik: ok i'll commit it then
00:33:59matslBagder: I can't wait tonight for this years tower but I look forward to seeing the result later. Good work everyone.
00:34:33pixelmafml: do you want to start with filling out one of those columns for a specific target?
00:35:09*matsl needs some sleep
00:35:13 Join perfectdrug [0] (
00:35:15 Quit matsl ("Leaving")
00:37:56 Quit tessarakt ("Client exiting")
00:38:25fmlAlexP: I think about something like but it doesn't seem to work for me.
00:39:04fmlpixelma: no, I just want to make it possible to do it gradually. And now I just want to sleep! :-)
00:39:53fmlShould I close FS #10356 or should I let it open for some time to let the people to comment?
00:40:13Mikachustable base at least
00:40:57Bagderwe need that part in slow-motion later
00:41:06Mikachuis it recording too?
00:41:08Bagderto properly assign blame
00:41:29Mikachuis there any way to access the video stream directly (for use in mplayer)?
00:41:41BagderI don't think so
00:42:05Mikachuthose two flash players take up ~70% cpu.. :)
00:43:35 Quit fml ("CGI:IRC 0.5.9 (2006/06/06)")
00:43:44Horschtand now the cam broke down
00:43:56Horschtsomeone is trying to destroy evidence
00:44:41CIA-70New commit by 03pixelma (r21426): Remove the 'hand-made' MASCODEC UseOption from the platform files and replace it with other options which automatically parsed from features.txt. Also ...
00:44:53***Saving seen data "./dancer.seen"
00:45:29 Quit petur (Read error: 104 (Connection reset by peer))
00:45:49 Join petur [50] (n=petur@rockbox/developer/petur)
00:48:41bertrikfunman, hmm, too bad it doesn't fix the c200v2
00:48:55funmanbertrik: it's different : the radio isn't detected
00:49:07CIA-70New commit by 03funman (r21427): Sansa c200v1 & c200v2 LCD: higher performance when writing pixels ...
00:49:21CIA-70New commit by 03zagor (r21428): Added config reading, and UPDATE.
00:51:50 Quit DarkDefender ("Leaving")
00:51:50funmana higher delay makes the menu reappear but the radio still doesn't work, but with your work i know where to look :)
00:52:37bertrikfunman, oh I think I know what is going wrong
00:53:02bertriklook in firmware/drivers/tuner/si4700.c line 36
00:53:19bertrikI think we need to add the c200v2 to the defines
00:53:51 Join nibbler_ [0] (
00:55:13funmanyes it's fine, i'll tweak the delay in fm i2c and make it conditional for c200v2
00:55:52bertrikfunman, what do you mean, you can actually make it work for the c200v2 by increasing the fmradio i2c delay?
00:56:55CIA-70New commit by 03zagor (r21429): Add UPDATE to accepted commands.
00:58:05funmanbertrik: hm only with the diff in si4700.c fm works fine
00:58:06CIA-70New commit by 03zagor (r21430): Hardcoding Daniels laptop as buildmaster during testing.
01:00:58kkurbjunwhat does lowering the bar mean for supported players?
01:01:10CIA-70New commit by 03zagor (r21431): clientver = 2. Added small startup banner.
01:01:23AlexPkkurbjun: Call them supported with fewer conditions
01:02:33kkurbjunyes, well, I guess I wasn't exactly clear what the bar is, the m:robe 500 is pretty useful for general music listening and hte like
01:02:51Horschtalso, the gigabeast seems very "supported"
01:02:54CIA-70New commit by 03funman (r21432): c200v2 : fix FM radio, the c200v2 uses the internal oscillator like others Sansa AMS
01:03:06Horschtat least from the target status wiki page
01:03:25AlexPHorscht: There are a couple of issues with that one
01:03:42AlexPHorscht: The install is somewhat unreliable
01:03:47AlexPHorscht: But yes, in general
01:05:39CIA-70New commit by 03bagder (r21433): support for updating the client if found too old
01:05:41CIA-70New commit by 03zagor (r21434): Runclient wrapper shell script.
01:06:35martian67commit storm
01:06:42martian67go Zagor ;)
01:07:35pixelmawould there be a way to find out button or action macros in the manual that are not inside a buttontable (or platform file (if yes, how)?
01:07:46kkurbjunAlexP: so is there a target in mind to add as a supported target under the lower guidlines
01:08:04bertrikfunman, cool
01:09:02CIA-70New commit by 03zagor (r21435): Replacing config file with parameters instead.
01:09:59AlexPkkurbjun: There are a few I think (or approaching) - some AMS Sansas come to mind
01:11:35 Quit krazykit (Read error: 113 (No route to host))
01:11:37*funman puts a veto against supporting AMS Sansas - storage driver is buggy
01:11:49 Join krazykit [0] (
01:12:51gevaertsfunman: not *now*...
01:14:03funmanwell i put a veto for tomorrow as well if it's what you mean .. :o)
01:14:15kkurbjunso it it just a relative call for when the people working on the port feel that the target is stable enough?
01:15:25jklspeaking of AMS Sansas, what is the bleeding edge status of USB support, and is there a branch in svn for it?
01:15:35 Quit nibbler_ (Read error: 110 (Connection timed out))
01:15:53funmanjkl: undocumented and unstarted
01:16:04Mikachuthat's quite a compact tower
01:16:58jklso the undocumented part... about that, wouldn't the Synopsys Linux driver be "almost" documentation?
01:17:37 Quit jordoex (Read error: 113 (No route to host))
01:17:52funmani think someone posted a link to a synopsys driver but the registers didn't match at all with the registers documented in AS3525 datasheet
01:18:31bertrikoh, I do remember a register match
01:18:37funmanthere is belief that the usb controller is very similar to a (AMD) controller for which there is a linux driver - the details should be on SansaAMS wiki page
01:21:05funmani have started disassembling the OF usb code but didn't get very far
01:21:12bluebrotherscorche|sh: last years tower was better :)
01:21:27jklfunman: yeah i saw that... i kind of didn't look at that AMD document too carefully but I have it
01:21:57funmanmost of the registers do match but there is some specificities
01:22:13funmannow let's hope the registers content match as well :)
01:24:03gevaertsbluebrother: get on a train and show us you can do better!
01:24:05jklright... but are the current (for the current IP cores that is) linux driver sources not available from Synopsys? (Solvnet, their support portal seems to be down)
01:26:02bluebrothergevaerts: I would arrive tomorrow evening, plus I have to leave for a training at work monday afternoon ... a bit too tight
01:26:36kugelfunman: you don't need to put a veto
01:27:17kugelbut new release blocking bugs should probably be added to the SansaAMS page
01:28:01funmanjkl: do you want to help with the usb driver?
01:29:22jklyeah, i am interested in the sansa ams port
01:30:15funmanthat would be really appreciated :)
01:31:07jkli kinda looked at a while ago... i think i even send an email to AMS for an NDA and they responded and I didn't get back to them.. but i notice the datasheet is downloadable without one now
01:32:52funmanyes the AS3525 datasheet is freely available, but they also had given it to us before it was. But they only mention synopsys and give the list of registers and their addresses, nothing else
01:33:24funmansince mpegplayer needs at least 2MB of RAM, and the c200v2 only has 2MB, I think lcd_blit_yuv support is not needed ;)
01:33:32kugelfunman: how much fps do you get on the c200v2 now?
01:33:35AlexPThere have been 83 commits since the start of devcon
01:33:43funmankugel: 181.5 at 62MHz
01:33:56kugelfull screen update?
01:34:12kugelhow many MB/s are that?
01:34:55kugelI'm curious if it would also show blue bars when hitting the maximum
01:35:49kugelhmm, that's weird. why's the PP display port faster for e200, while the as3525 one is faster for c200?
01:36:00funmanhm .. when cpu is boosted the lcd goes off .. only backlight
01:36:05jklright... i was trying to get on Synopsys Solvnet, but their registration seems to be down... the linux driver datasheet says that the source is basically under a BSD style license so I was thinking it might be available without a Site ID
01:36:19funmankugel: isn't the controller different between e200 & e200v2 ?
01:36:30kugelI don't think it is
01:36:43kugelthe driver is mostly a plain copy from e200v1 too
01:37:40funmankugel: 3.66MBytes/s
01:37:50jklI do have a site ID too though
01:38:29 Join W00fer [0] (
01:38:42W00ferso i'm watching the rockbox webcams :)
01:38:55W00fergreat that there is a euro devcon
01:39:21kugelfunman: huh
01:39:23W00fertoo bad i didn't send my iriver E100 in for a loan. So you guys could have a look at it.
01:39:31W00ferif you want, ask me next time.
01:39:55kugelfunman: the 181,5fps was at boosted or unboosted?
01:40:16jklI have that old Synopsys DWC linux source you mentioned, but it obviously doesn't seem to be similar to the reg layout in the AMS
01:40:32kugelseems there's something seriously wrong on the c200v1 if the controller and lcd can actually handle more
01:40:34funmankugel: just read my last messages ;)
01:40:55funmanwhen boosted the lcd turns down and doesn't recover so i can't tell what is the performance here
01:41:05funmanthe lcd also turns down in the bootloader occasionally
01:41:19AlexPW00fer: I think JdGordon already looked at the iriver E100
01:41:23kugelI guess there's some delay missing
01:41:39funmandidn't see anything in the OF
01:41:42AlexPW00fer: There should be a forum thread or wiki page somewhere on it
01:42:04kugelmaybe a missing bit in the dbop init?
01:42:52jklwhat tools do you use to assist in reversing the OF?
01:43:12 Join mc2739_ [0] (
01:43:21 Quit mc2739 (Nick collision from services.)
01:43:22 Nick mc2739_ is now known as mc2739 (
01:43:25funmanarm disassembler
01:44:31jklfunman: anything in particular though (other than just objdump)?
01:45:00funmani use ida and sometimes objdump
01:45:02W00ferAlexP: at this devcon ?
01:45:03jkli was just wondering if there were some good tools in a similar vein to IDA
01:45:55funmanan iphone hacker is working on a disassembler :
01:46:19funman but i don't know the current state (still quite a new project)
01:47:20GodEaterW00fer: not at this devcon, no. He's not here.
01:48:54funmanhappy devcon :)
01:48:57 Quit funman ("free(random());")
01:50:25AlexPW00fer: A few months ago now
01:51:25 Quit n1s (Read error: 110 (Connection timed out))
01:51:55W00feris there a change rockbox will run at smartphones (based on windows mobile nowadays). Since a lot of activity is going on on developing Andriod (based on linux) onto HTC's Samsungs and other brands. They already have a sample of Android running on an Samsung i900 Omnia. Possibly it could run rockbox too.
01:52:49W00fersome of todays smartphones have 3,5mm headphone jacks and 1000mAH or more batteries
01:53:07W00fertouchscreen and detailed hardware support with SDK's or diagrams available
01:54:48W00feralso many thanks on rockbox 3.3 !! great progess. Too bad no USB on ipod 4G photo. Hope for it next time
01:56:20ZagorW00fer: yes rockbox can run on it. but these machines are really too powerful to only run rockbox.
01:56:37Zagorwe only use half of the features they have
01:56:50W00ferno problem, if anyone has a smartphone running around
01:56:52jklw00fer: if it can hold win mobile it certainly has enough flash for linux and stuff
01:56:52 Join jordoex [0] (
01:57:23W00fermost OS'es like windows mobile or symbian have bad media player coding, so bad battery performance
01:57:42 Nick fxb__ is now known as fxb (
01:57:51 Join BryanJacobs [0] (
01:57:52W00feronly difference is the Nokia N97, which is said to play 37 hours of mp3 on one charge
01:58:44Zagorthe problem is rather that they have to do lots of other things too, like keeping a radio constantly broadcasting
01:59:04jklseems odd a smartphone manufacturer wouldn't be able to tend to the power issue better
01:59:05W00ferand if you don't use the bootloader for windows mobile.
01:59:12Zagorstill, very few people want to turn their phone into a media player. rockbox-as-application is more interesting.
01:59:15W00ferONLY load rockbox or ipodlinux
01:59:31*BryanJacobs has an N85 and is extremely happy with the battery life playing back MP3s
01:59:56W00ferbut rockbox as application is way more work than loading linux onto a smartphone and rockbox i think
01:59:58BryanJacobsbesides, if you JUST ran Rockbox you'd lose features - like A2DP
02:00:16Zagorlike calling...
02:00:16W00ferdon't need that
02:00:17W00feri want a cool media player
02:00:28W00ferwith 1000+ mah battery
02:00:28BryanJacobsI like my stereo headphones
02:00:45BryanJacobshave you looked at the iPod battery packs?
02:00:46ZagorW00fer: then get to work. I don't expect many other owners of such hardware to agree with you.
02:00:48jkli did windows mobile integration for devices that were similar to smartphones... data terminals with GSM, bluetooth, and 802.11b/g
02:00:51BryanJacobsthere are some quite large ones
02:00:57W00ferfor calling i have a second phone
02:01:17BryanJacobsor you could get a small USB charger pack and keep the media player plugged in
02:01:26jklRF kill was a major thing for aviation purposes of course
02:01:27BryanJacobset voila, cheaper solution
02:01:49W00feranyone succesfully launched playstation 2 linux ?
02:01:53 Quit Thundercloud (Remote closed the connection)
02:02:42W00ferdoes the WPS battery saving in rockbox 3.3 also saves battery on pp5 targets (ipod photo like i have) ?
02:02:55 Quit perfectdrug (Remote closed the connection)
02:03:10 Quit mc2739 (Read error: 110 (Connection timed out))
02:03:46CIA-70New commit by 03rasher (r21436): Properly match IDs - require a perfect match, not just a substring match. Oops.
02:04:39CIA-70New commit by 03rasher (r21437): Deprecate LANG_REPLAYGAIN_ENABLE in all languages, which r21415 failed to do.
02:07:48 Quit jordoex (Remote closed the connection)
02:09:17 Join ash001 [0] (
02:11:27jklBryanJacobs: are you U of R alumnus?
02:12:26soapOhh Ohh Ohh. On BuildServerRemake. Soap's idea is "Build on a RAM drive by default".
02:13:32Bagdersince you run the client you can do that easily yourself
02:13:46Bagderjust make a ramfs or something
02:13:52 Join saratogahome [0] (n=41becb3b@gateway/web/cgi-irc/
02:14:02soapI see how it is - make _me_ do the work. ;)
02:14:05Mikachutmpfs is good for that, since it doesn't allocate the ram statically
02:14:11CIA-70New commit by 03zagor (r21438): Do blocking connect.
02:14:21Bagderright tmpfs was the one I thought about
02:14:38Bagderit really sped up the builds when I did binsize builds
02:14:44saratogahomehow is the new build server coming?
02:15:08Bagderwe're about to run a test with... 6-7 laptops building here
02:15:21kugelare we? :)
02:19:04CIA-70New commit by 03zagor (r21439): Throttle connection retries.
02:20:12W00ferlol that cola bottle at the webcam was filled alot more an hour ago wasn't it ;)
02:20:37AlexPThere is still a coke bottle infront of cam 1
02:20:41AlexPI'll wave now
02:20:51AlexPSo you can see what the lag is
02:23:25W00ferdidn't see it
02:23:44AlexPYou watching cam 1?
02:23:45 Quit ocean ("Page closed")
02:23:54W00fercam 2 only shows robodog
02:24:07AlexPI just waved again
02:24:09CIA-70New commit by 03zagor (r21440): Fixed connection throttling.
02:24:31W00ferhaha the lag is almost 30 seconds
02:24:59W00ferhows the beer, i just finished my dutch Dommelsch
02:25:15AlexPThe beer is great thanks :)
02:25:17W00ferdrunk 8 bottles with my mate, he left already
02:26:06peturyou wouldn't drink 8 of these beers and still type/talk ;)
02:26:10W00ferhaha alexp still takes a nip of his glass of beer
02:26:24W00ferno i think that ones have 5+ % alcohol
02:26:41W00fermines only have 2,5%
02:26:53AlexPThese are all over 6%
02:27:32petur2,5% doesn't qualify as beer
02:27:46W00ferwhy don't you guys use a beamer to show a roadmap or something ?
02:27:56W00feror plans ;)
02:27:58peturwhat roadmap?
02:27:59Zagorroadmap? what's that?
02:28:07AlexPW00fer: A BMW?
02:28:16W00ferno a projector
02:28:31 Quit bluebrother ("leaving")
02:28:51W00fer2,5 % qualifies as pils, not beer
02:29:21W00ferah a laptop just shut down, the grey one.
02:29:22AlexPWe ought to take this to #rockbox-community :)
02:29:50Mikachui meant to suggest that too
02:29:59W00feroh it's night
02:30:08W00fer2.30 am here
02:30:29W00ferwho is the guy with the beard, he looks like grandpa smurf haha :D
02:31:55BryanJacobsjkl: student
02:33:23BryanJacobsdid you recognize my mail server's IP range or something?
02:33:25Bagder10 build clients
02:33:35Bagderseems to run
02:33:50Mikachuon the first try?
02:34:20AlexPMikachu: Let's say yes
02:36:49 Quit XavierGr ()
02:38:42CIA-70New commit by 03zagor (r21441): Removed some verbosity.
02:41:14*Mikachu looks at his athlon xp 2ghz
02:42:06Bagder"Still 46 builds left to complete with 11 in progress"
02:44:50Horschtis there some way i can make a "screenshot" on the actual hardware?
02:44:54***Saving seen data "./dancer.seen"
02:45:03Mikachuwith the screenshot function in the debug menu?
02:45:36AlexPEnable it, then connect USB to take one
02:46:15jklBryanJacobs: no just noticed your IP
02:46:34jkli'm '10 ECE
02:46:47BryanJacobsjkl: cool
02:46:50Horschti noticed that setting before, but I was not sure how to actualy dump the screen :D
02:48:53CIA-70New commit by 03zagor (r21442): Remove killed builds. Bumped version.
02:50:49 Quit faemir (Read error: 60 (Operation timed out))
02:50:59 Join ocean [0] (i=d59c23e8@gateway/web/freenode/x-f16afb38a96fdbbf)
02:51:41oceansvn isn't working for me if i do svn up or co says connection refused :(
02:51:51 Join faemir [0] (n=faemir@
02:52:06Mikachuworks here
02:52:29oceanyes strange thing is that it doesn't work only on my notebook
02:52:44Horschtcan you access the web from your notebook? ;)
02:53:02oceanHorscht: yes and i can also access
02:53:33oceanmaybe it could be the slackware svn package i try to revert it to an older version
02:53:37Mikachumaybe Zagor is overloading the svn server :P
02:53:44oceanmaybe :P
02:53:48Zagorwohoo, auto update works beautifully
02:54:29Mikachuso now you have root access to all the client computers?
02:54:39Zagoryesssss ;-P
02:54:45Horschtand now the plan unfolds
02:54:55Zagorno, we've turned the clients around. no ssh anymore.
02:55:21kugelah, no ssh at all? awesome
02:55:25Zagorthe client owner is much more in control with the new system
02:57:43oceanmmm strange :\ keeps not connecting, while from another notebook here works
02:57:56oceanmaybe it doesn't like my pc :P
03:02:40 Join Thundercloud [0] (
03:04:39 Join perrikwp [0] (n=perrikwp@
03:06:14 Quit ocean ("Page closed")
03:06:26 Join ocean [0] (i=d59c23f7@gateway/web/freenode/x-05ea3554bdf8fe0f)
03:06:45oceanok done instead of using the wireles i got a lan wire and it works :P
03:07:26 Quit BryanJacobs ("Java user signed off")
03:11:47 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
03:14:14Mode"#rockbox +o GodEater " by ChanServ (ChanServ@services.)
03:15:26 Join fdinel [0] (
03:22:52CIA-70New commit by 03bagder (r21443): hush up, don't listen to localhost only, wait 5 seconds before starting the ...
03:23:46 Quit W00fer ()
03:24:14CIA-70New commit by 03zagor (r21444): Cleanup after killed builds.
03:25:02 Quit Domonoky1 ("Leaving.")
03:25:30 Nick fxb is now known as fxb__ (
03:26:39 Quit Bagder ("It is time to say moo")
03:26:47 Quit Zagor ("Leaving")
03:26:56 Quit petur ("Zzzzz")
03:27:40 Quit markun ("good night")
03:27:54oceani've seen lots of IBSS_ATTR as variable name in sources, what does it mean?
03:28:24 Join Bugsley [0] (n=ad631c1d@gateway/web/cgi-irc/
03:28:58saratogahomeocean: those tell the linker to put that variable into fast internal memory, rather then external dram, if any is available
03:29:31saratogahomeICONST does the same thing, but for constants, and ICODE does it for code sections
03:31:17oceansaratogahome: thanks, are all these things documented somewhere? i had bad luck trying to find them on google, only results i have are rockbox sourcecodes :P
03:33:16 Quit Thundercloud (Remote closed the connection)
03:33:54 Quit jkl (Read error: 110 (Connection timed out))
03:34:33saratogahomeocean: no they're rockbox internal defines, not something you'll find on google
03:34:44saratogahomeyour best bet is to grep the source code and see what they're defined for on your system
03:34:55 Quit bertrik (Read error: 60 (Operation timed out))
03:35:01saratogahomei believe you'll find that the IRAM defines map to various GCC c extensions
03:35:17oceani was giving a look at how to write a codec :)
03:35:21saratogahomeor to nothing if you don't have IRAM on your system
03:35:32saratogahomeocean: which codec are you interested in?
03:35:53oceani've seen it hasn't been done
03:36:38martian67sounds like a turkish male hooker :)
03:36:55BugsleyHey - sorry to take everyone off topic, but I've searched google and the Rockbox site for this issue. During install on a e250, I get "Dowloading bootloader file", "Downlad finished", "Installing Rockbox bootloader", then I get "Could not open Sansa". I'm on OSX Leopard
03:38:08saratogahomeocean: have you seen FS #8806
03:39:55saratogahomeBugsley: i don't know anything about OSX, but are you sure its a V1 sansa, and that its in MSC mode?
03:39:57oceansaratogahome: not seen it, but i don't like mikmod howewer it plays wrong some mods/xms
03:40:46saratogahomeocean: its probably a good place to start though
03:41:03oceani would like to write a xm codec to learn something and to get a good quality i'm one of those mods nostalgic :P
03:41:05BugsleySaratogahome: yes it's in MSC. I am not sure it's V1, checking
03:42:21saratogahomeif you're just interested in codecs, might i interest you in WMA pro?
03:42:43oceani've seen something in the mighty quotes about wma :P
03:42:59BugsleySaratogahome: it's Version 01.02.18A... There's nothing on the exterior that says what Version
03:43:06saratogahomethats a V1
03:43:09ocean"The problem with WMA is that the existing open source WMA implementation is using floating point arithmetics, something that the CPU in the iriver can't deal with. We need to convert all calculations to fixed point, a rather tedious and complicated task. Someone with a clue needs to step forward and do this. The problem is that people with a clue don't use WMA."
03:43:19saratogahomethats an old quote
03:43:26saratogahomei ported WMA a couple years back
03:43:28oceanyeah but have a good point :P
03:43:35oceani love ogg :P
03:43:41Mikachusaratogahome: so you have half a clue?
03:43:49martian67why would anyone use floating point
03:43:53martian67when its not required?
03:44:07saratogahomemartian67: virtually everything is floating point
03:44:14Mikachufixed point is much more annoying and less precise usually
03:44:19saratogahomefixed point is much more difficult and generally much slower on modern computers
03:44:23Mikachuunless the format is integer to begin with of course
03:45:06BugsleySo Anyone have an idea on the "Could not open..." error? I seriously doubt it's an OSX issue. Perhaps a setting on the Sansa?
03:45:06martian67Mikachu, well everything is fixed point to a computer, even floating point :)
03:45:29 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
03:45:29 Quit amiconn (Nick collision from services.)
03:45:46saratogahomemartian67: modern computers do not deal with everything in fixed point, they have FPUs, in which the decimal point floats during computation
03:45:50 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
03:46:08 Quit pixelma (Nick collision from services.)
03:46:09 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
03:46:11oceansaratogahome: i think it would be a really hard work and i would not be able (as knowledge and for free time) to get involved in such a difficult task :)
03:46:25 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
03:46:26martian67saratoga, my point was it ultimately does get delt with in a fixed point manner inside the fpu, its not terribly relivant though
03:46:41saratogahomemartian67: it does not
03:46:55saratogahomei suggest you look at the code for fixed point multiplies and FP emulated multiplies
03:47:04saratogahomethe operations are fundimentally quite different
03:47:23 Quit kugel (Read error: 110 (Connection timed out))
03:48:04martian67floating point has some rather un-expected side effects if you arnt aware how it operates
03:49:12*ocean thinks of the 40000000000-39999999999 behaviour in google
03:50:19Mikachuyou can't do 10^90 + 0.5 in fixed point
03:50:42saratogahomeif you think side effects in fp are difficult, you've probably not done much in fixed
03:51:02martian67saratoga, just dont divide :)
03:51:46martian67or wrap
03:52:01 Quit Bugsley ("CGI:IRC (EOF)")
03:52:16martian67but division and wrapping are easier to comprehend than what floating point does imo
03:52:40martian67anything involving binary radixes is a little hairy :P
03:53:14oceansaratogahome: just given a look to some sources of your libwma... that's a LOT of work
03:54:45 Quit froggyman (" Want to be different? Try HydraIRC -> <-")
03:55:08saratogahomei think wma pro would be easier, a lot of the code is already written
03:59:29oceanit has multichannel, those maybe should be mixed
03:59:53oceanhowewer it's too much harder for me i think i'll start with something simpler :P
04:00:08saratogahomeyeah they'd need to be down mixed to stereo, but theres code to do that in AC3
04:00:23saratogahomeit could be borrowed, or removed to the codeclib and shared
04:00:43oceanxm is an integer format with simple samples compression (from what i remember of this format)
04:00:59oceanso that could be a simple thing to start with :)
04:01:16saratogahomei don't konw anything about those emulated formats
04:01:36saratogahomethough i suppose if theres already a patch for XM on the tracker, correcting it for your files should be pretty easy
04:03:43oceani could take a look on the code of mikmod, but i was more focused on getting some basic code information from xmp, and then trying to get the best possible implementation from fast tracker II format document
04:11:00oceantime to go to sleep good night
04:11:02 Quit ocean ("Page closed")
04:11:20 Quit intrados (Read error: 104 (Connection reset by peer))
04:17:58 Join Strath [0] (
04:33:08 Quit PaulJam_ (".")
04:35:41 Quit saratogahome ("CGI:IRC (EOF)")
04:44:55***Saving seen data "./dancer.seen"
05:08:24 Quit martian67 (Read error: 113 (No route to host))
05:13:49 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
05:18:55ash001Is rockbox on C200V2 at least bootable without risk of bricking it?
05:19:02ash001I seems funman got a lot of work done on it..
05:21:44 Quit Cory` ("Ex-Chat")
05:26:09 Join dmb [0] (n=dmb@unaffiliated/dmb)
05:26:21 Join Cory` [0] (
05:45:02 Quit faemir (Remote closed the connection)
05:47:14 Join n17ikh [0] (
05:58:26 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
06:00:50 Quit LambdaCalculus37 (Client Quit)
06:12:03 Join Rob2223 [0] (
06:21:18 Part toffe82
06:27:16 Nick beta_ is now known as Beta2K (
06:29:52 Quit Rob2222 (Read error: 110 (Connection timed out))
06:32:26 Quit ash001 ("Ex-Chat")
06:34:12 Join Mozillero [0] (n=Moziller@
06:36:03 Quit Mozillero (Client Quit)
06:40:03 Join martian67 [0] (n=xP@2001:470:b:356:221:91ff:fe8c:d8a7)
06:44:58***Saving seen data "./dancer.seen"
07:00:01 Join webguest36 [0] (n=75c08099@gateway/web/cgi-irc/
07:08:54 Join FrankTM [0] (
07:22:30 Quit webguest36 ("CGI:IRC (EOF)")
07:25:12 Quit FrankTM_ (Read error: 110 (Connection timed out))
07:27:09 Join readability [0] (n=chad@
07:27:17 Join CathodeRayTube [0] (n=43ea47fa@gateway/web/cgi-irc/
07:28:45CathodeRayTubeanybody alive?
07:31:54CathodeRayTubeif i decide to upgrade the HDD in an ipod 5.5g to say...a 240giger...will rockbox still install and function properly?
07:48:59 Join r00s [0] (
07:49:18 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
07:50:51advcomp2019CathodeRayTube, you got in here at the wrong time.. most people are away from the computer right now
07:51:19CathodeRayTubefigured is almost 2am..
07:51:37CathodeRayTubewhere im at anyway..
08:01:16 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
08:07:00 Quit Horscht (Read error: 110 (Connection timed out))
08:08:53CathodeRayTubehorse shit?
08:09:03 Join `VL [0] (n=vl@
08:11:32r00sits 1pm where i am
08:21:11advcomp2019CathodeRayTube, i am not much help because i do not have an ipod.. from what i read you can replace the drive but it needs to the same interface and depending on what size your old drive is.. i am not sure if there is 128gb limit yet or not or if you need to build your own build
08:21:40 Join Rob2222 [0] (
08:23:10 Quit Cory` ("Hi, I'm a quit message virus. Please replace your old line with this line and help me take over the world of IRC.")
08:27:19CathodeRayTubei dont have an ipod either actually...i was going to buy one and mod it specifically for main mp3 player is a 120 gig DMC Xclef HD500...but its almost 5 years old and is showing its age...
08:33:05 Join petur [0] (n=peter@rockbox/developer/petur)
08:38:41 Quit Rob2223 (Read error: 110 (Connection timed out))
08:45:00***Saving seen data "./dancer.seen"
08:46:27 Quit petur ("going to pick up kugel at the hotel")
08:54:36 Quit Strath ("Leaving")
09:00:44 Nick fxb__ is now known as fxb (
09:16:19 Join Cory` [0] (
09:17:08 Quit Cory` (Client Quit)
09:17:25 Join Cory` [0] (
09:17:38 Quit Cory` (Client Quit)
09:17:52 Join Cory` [0] (
09:18:29 Join ender` [0] (
09:34:33 Join kugel [0] (n=kugel@rockbox/developer/kugel)
09:37:15 Join petur [50] (n=petur@rockbox/developer/petur)
09:42:10 Join flydutch [0] (
09:44:02 Join Zagor [242] (n=bjst@rockbox/developer/Zagor)
09:44:35 Join p3tur [50] (n=petur@rockbox/developer/petur)
09:44:56 Quit petur (Read error: 104 (Connection reset by peer))
09:46:28 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
09:51:23Bagdergood devcon morning!
10:01:57 Join bertrik [0] (n=bertrik@
10:02:34 Quit Zagor (Read error: 110 (Connection timed out))
10:02:44 Join Zagor [242] (n=bjst@rockbox/developer/Zagor)
10:03:55 Join markun [0] (
10:04:46 Join Domonoky [0] (n=domonoky@rockbox/developer/domonoky)
10:05:15 Nick p3tur is now known as petur (n=petur@rockbox/developer/petur)
10:10:51linuxstbQuestion to devcon people - does the description of "supported" here still meet what you decided?
10:12:21kugelit seems like that what we agreed to
10:12:44gevaertsit probably needs some review
10:16:51 Quit Cory` (Remote closed the connection)
10:20:21 Join DarkDefender [0] (
10:26:32 Nick Horschti is now known as Horscht (n=Horscht2@xbmc/user/horscht)
10:27:26 Quit BHSPitLappy ("Ex-Chat")
10:34:57 Nick fxb is now known as fxb__ (
10:37:31rasherAny forum guys around? Baffmyiloven needs banning.
10:37:44rasherI don't know if I can do that?
10:41:28 Quit saratoga (Remote closed the connection)
10:45:03***Saving seen data "./dancer.seen"
10:49:42GodEaterrasher: he's gone
10:50:01 Quit CathodeRayTube ("CGI:IRC (Ping timeout)")
10:50:54rasherCompletely, or just the post?
10:50:56gevaertsGodEater: you don't need to be op on irc to ban people on the forums!
10:53:21 Join Grahack [0] (
10:54:03scorche|shand by "banned" he means that his name was added to the ban list without any sort of ban at all =)
10:54:38Mode"#rockbox -o GodEater " by ChanServ (ChanServ@services.)
11:08:40 Join Domonoky1 [0] (
11:12:22 Join ch4os [0] (n=ch4os@gentoo/user/ch4os)
11:22:59 Quit Domonoky (Read error: 110 (Connection timed out))
11:33:34 Quit ender` (Read error: 110 (Connection timed out))
11:35:23 Join mcuelenaere [0] (
11:42:30 Join bmbl [0] (
11:59:24 Quit mcuelenaere ()
12:03:39 Quit __lifeless (Remote closed the connection)
12:06:04 Join _lifeless [0] (n=lifeless@
12:09:46 Join robin0800 [0] (
12:12:50 Quit _lifeless (Remote closed the connection)
12:13:43 Join faemir [0] (n=faemir@
12:15:07 Join _lifeless [0] (n=lifeless@
12:19:36 Join nibbler_ [0] (
12:34:09 Join ender` [0] (
12:35:17 Join AndyI [0] (i=AndyI@
12:36:41bertrikdoes anyone have the tcc773 datasheet?
12:37:14bertrikI have the tcc76x, is it similar to the tcc773?
12:37:18 Quit faemir (Read error: 110 (Connection timed out))
12:43:12 Quit bmbl ("Bye!")
12:43:27 Join bmbl [0] (
12:44:32 Join ocean [0] (i=d59c23f7@gateway/web/freenode/x-07327e29f91f74f0)
12:45:07***Saving seen data "./dancer.seen"
12:46:22 Quit AndyIL (Read error: 110 (Connection timed out))
12:48:17 Quit bmbl (Client Quit)
12:48:31 Join bmbl [0] (
12:49:18 Quit Sajber^ (Read error: 60 (Operation timed out))
12:50:51 Quit bmbl (Client Quit)
12:51:04 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
12:54:50 Join Sajber^ [0] (
12:57:00 Part Grahack
12:57:15bertrikI think I found a bug in telechips i2c, the i2c mutex seems to be never initialised
13:02:51linuxstbbertrik: The 76x is different to the 77x
13:12:31 Join Rondom [0] (
13:13:56 Join Buschel [0] (
13:14:17 Quit killan (Read error: 54 (Connection reset by peer))
13:18:11 Quit ocean ("Page closed")
13:21:10 Join jkl [0] (
13:23:52 Join killan [0] (
13:29:11 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
13:29:25 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
13:30:40 Join Thundercloud [0] (
13:32:26 Nick zitune[afk] is now known as zitune (n=zitune@bearstech/zitune)
13:40:34 Quit Sajber^ (Read error: 110 (Connection timed out))
13:44:21 Join Rob2223 [0] (
13:45:13 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
13:45:13 Quit amiconn (Nick collision from services.)
13:45:31 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
13:46:28 Join lilltiger_ [0] (n=lilltige@
13:46:51 Join kachna [0] (
13:46:55 Quit Horscht (Read error: 110 (Connection timed out))
13:47:07 Quit gevaerts (Nick collision from services.)
13:47:16 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
13:47:29 Quit jkl (
13:47:29 Quit nibbler_ (
13:47:29 Quit DarkDefender (
13:47:29 Quit Rob2222 (
13:47:29 Quit n17ikh (
13:47:29 Quit dmb (
13:47:29 Quit kkurbjun (
13:47:29 Quit MrDuck (
13:47:29 Quit BlakeJohnson86 (
13:47:29 Quit Ridayah (
13:47:29 Quit evilnick_home1 (
13:47:29 Quit lilltiger (
13:47:29 Quit CIA-70 (
13:47:29 Quit tchan (
13:47:29 Quit alexbobp (
13:47:29 Quit redfox (
13:47:44NJoinjkl [0] (
13:47:44NJoinnibbler_ [0] (
13:47:44NJoinDarkDefender [0] (
13:47:44NJoinn17ikh [0] (
13:47:44NJoindmb [0] (n=dmb@unaffiliated/dmb)
13:47:44 Join kkurbjun [0] (n=kkurbjun@rockbox/developer/kkurbjun)
13:47:44NJoinBlakeJohnson86 [0] (
13:47:44NJoinRidayah [0] (
13:47:44NJoinevilnick_home1 [0] (
13:47:44NJoinCIA-70 [0] (n=CIA@
13:47:44NJointchan [0] (n=tchan@lunar-linux/developer/tchan)
13:47:44NJoinalexbobp [0] (
13:47:44NJoinredfox [0] (
13:49:33 Quit petur (Read error: 54 (Connection reset by peer))
14:08:51 Join Sajber^ [0] (
14:09:54 Quit martian67 (Read error: 60 (Operation timed out))
14:11:35 Join martian67 [0] (n=xP@2001:470:b:356:221:91ff:fe8c:d8a7)
14:12:24LloreanIs there any real potential for improving seeking in VBR MP3 from its current state?
14:23:11BuschelLlorean: in terms of precision?
14:24:55LloreanBuschel: Primarily, though realistically I'd settle for "taking steps to ensure you're always at an earlier time than you sought to." Right now in very large files, the 2 second rewind on unplug can take me 1-2 minutes *forward* in the file
14:26:13LloreanKnowing nothing about MP3 I'm honestly surprised that we can tell I'm 2:45:30 into the file after seeking (when the seek bar said 2:43:00) but we can't hit more accurately (maybe by using this knew knowledge of where we ended up to try for a second approximation?)
14:26:34LloreanI'd think if we couldn't seek to a specific time, we shouldn't be able to determine what time I'm at *after* the inaccurate seek either.
14:26:49LloreanBut again, I don't know anything useful, so this just seems odd from my perspective
14:30:38BuschelLlorean: seeking in mp3 is quite unprecise as long as you want to have fast seek. for full precision you'll need to parse through the stream from the beginning (like mpc does). doing so needs lots of buffering and cpu time of course.
14:32:02LloreanBuschel: How can we know what the elapsed time is after a seek then? Or rather, once we know we've missed the target by 30 seconds, can we use that information to try a second approximation at least? I know it'd slow seeks somewhat, but seeks are so fast now that I think even a fair amount of slowing when they go *really* bad would be preferable.
14:32:16BuschelLlorean: the current code estimates (there are no timestamps in the mp3-stream itself) to which byte to jump based upon either 0% runtime = byte 0 and 100% runtime = last byte or sokme more detailed info from vbr-headers (if present)
14:32:33LloreanI have vbr headers, but it can still be pretty bad.
14:33:05LloreanBut how do we get the elapsed time? I mean, if we have to estimate where to seek to, don't we have to estimate the elapsed time after the seek as well?
14:33:17LloreanWhy don't they match?
14:33:33Buschellet me check the code
14:34:19LloreanIt just seems to me, if I seek to 2:50 and when the seek finished it says 2:52 as the elapsed time, if it knows the elapsed time is wrong it should've known the seek was wrong too. ;)
14:34:31Llorean2hrs 52 minutes that is
14:34:41*Llorean is not quibbling over a second or two
14:36:21 Join mt [0] (n=MTee@rockbox/developer/mt)
14:38:26BuschelLlorean: it seems like the following: 1) you want seek to let's say 2:50 -> the algorithm calculates this must be at byte no. 123.345.789 in the stream and jumps to it. 2) the time pos is calculated from the byte pos (123.456.789) via a similar algorithm which results in a slightly different time pos 2:52. the differences may come from rounding errors −− but i cannot tell from a short view into the source.
14:39:39Buschelquestion is: what is more correct -> the time you wanted to seek to or the time from the byte position? ;)
14:41:00LloreanWell, the current problem is "if I seek to 2:00, I end up at 2:02, if I seek to 2:02, I end up at 2:04"
14:41:11LloreanIf you seek to what the time displayed just 2 seconds ago, you end up several minutes in the future
14:41:26LloreanThis makes "Rewind on unplug" fail spectacularly.
14:41:50LloreanAt the very least, rewinding should *never* take you to a later place in the file than you are now.
14:42:05Buschelyes, that's another story.
14:42:20LloreanWould it be possible to seek from "now" instead of the beginning of the file?
14:42:25Buschelso, it's not only the displayed time but also the real time in the stream?
14:42:47LloreanI don't know accurate real times in the stream.
14:42:49BuschelLlorean: that would also require parsing.
14:44:18LloreanThe algorithm takes 2hrs 50 minutes and calculates an approximate number of bytes per second / millisecond / time to figure out what by to guess, right?
14:44:49LloreanIf it does that, why can't it take the current byte, then use that algorithm to calculate the *difference* between "current time" and "seeked to time" and so do that approximation on a much smaller number of bytes?
14:45:08***Saving seen data "./dancer.seen"
14:45:09Buschelyes, a vbr header contains 256 time positions and their corresponding byte positions. a 3h-file has a seek-resolution of ~40 seconds.
14:45:24Buschelall positions in between are extrapolated
14:45:54LloreanI'm just saying, when you seek you go X bytes into the file. Why can't you go Y bytes from where the seek started instead?
14:46:22LloreanIsn't it the same math, on the same data set?
14:46:54 Join jordan` [0] (i=gromit@
14:46:55BuschelLlorean: hmm, in the end the algorithm does what you describe.
14:47:07LloreanNo, it doesn't, or rewinding would never take you to a point *after* the one you're listening to.
14:47:13UnhelpfulLlorean: as i'm understanding this, part of the trouble is that when you go to an approximated seek point in the data you don't know what time you have found.
14:47:21BuschelLlorean: how long you are playing with?
14:47:41LloreanUnhelpful: Yeah, but if you were inaccurate relative to "now" rather than inaccurate relative to 0, you'd at least never rewind to a position later than the one you're listening to, right?
14:47:47LloreanBuschel: 8-11 hours.
14:48:39Unhelpfulcould we go to the latest seek point that's before the time we want, and parse forward from there? if we're only parsing frames, and not decoding, would it really cost that much to seek, say, 40-120s "the hard way" after going to the seek point?
14:48:41LloreanUnhelpful: That's what I think I'm failing to describe. If you're using approximate bitrate to approximate the number of bytes over time, why can't you approximate the number between "now" and "destination" then add or subtract that to the current file position, rather than trying to approximate an absolute position in the file and just jump there?
14:50:01UnhelpfulLlorean: ah... i understand you now. you're suggesting jump back N bytes, then find the next start-of-frame?
14:50:29LloreanUnhelpful: Something like that yeah. Relative seeking rather than absolute seeking so that even in the worst case, you at least go in the right direction in the file.
14:50:49BuschelLlorean: your main point is to avoid jumping forward when seeking backwar, right? and not the wrong total time position?
14:50:54LloreanSince we know they're seeking *from* somewhere, we should be able to at least use that to make sure we're not going someplace completely wrong.
14:51:15LloreanBuschel: Yeah, I know the time can never be absolutely accurate, but I'd rather it fail in the right direction. :)
14:52:33Unhelpfulif you jump to one of the marked seek points and parse forward, you'd also have the time right...
14:53:27LloreanI'm a little confused by that statement.
14:53:41LloreanOr do you mean, with slow seeking forward from the seek point?
14:54:21LloreanI mean, the problem is that I'm using *very* large files, so they have a pretty large-grained seek table as it is, which doesn't help
14:55:09Buscheldamn, just recognized i made a mistake when interpreting the code. it's even worse: even with a vbr-header the time resolution for seek is 100 entries only (~36 seconds in a 1h file, *10
14:55:25Buscheloops, wasn't finished :)
14:56:09LloreanMy understanding is that 100 entries was the maximum for the VBR header to have.
14:56:22*linuxstb suggests putting mp3 in a container (.mp4?) and supporting that in Rockbox...
14:56:35Lloreanmkv? :-P
14:56:48Buscheldamn, just recognized i made a mistake when interpreting the code. it's even worse: even with a vbr-header the time resolution for seek is 100 entries only (~36 seconds in a 1h file, 6min in a 10h file). each time mark has a byte position saved with a resolution of 256 (256=eof).
14:57:28LloreanUnhelpful: Does "parsing" require decoding, or is it somewhat faster than that?
14:58:30UnhelpfulLlorean: i would imagine we'd just scan forward in the bitstream for frame headers. i can't see any reason to do full decode...
14:59:04Unhelpfulis there an id3 chapters or seekpoints tag?
14:59:38LloreanHow slow could this potentially be in say, a large file with 6 minute gaps between vbr seek entries?
15:00:27linuxstbLlorean: It would mean reading the whole file from disk, and then doing some relatively easy parsing of it. The disk reading would be the main limitation.
15:01:20linuxstbOr rather, reading from the start of the file to the seekpoint. But I guess there could be some optimisations - e.g. the codec caching its own seekpoints if the file has been played from the start.
15:01:23Unhelpfullinuxstb: "the whole" as in "starting from the seek point?
15:02:05Lloreanlinuxstb: Well, as I said, I'd just be satisfied if rewind didn't ever take me forward. :)
15:02:16 Join moos [0] (i=mustapha@rockbox/staff/moos)
15:04:04linuxstbLlorean: But if you seek relatively to the current position, you'll get ever more inaccurate seeking IIUC, as the current position may be the result of a previous guess.
15:04:33linuxstbSplitting long mp3s would also solve your problem...
15:05:10LloreanWell, "I rewind and end up further in the file instead" is a bug either way though
15:08:57Unhelpfulit sounds like it may be possible for the position specified in the seek table to be off by several frames, too, though. :/
15:10:16 Nick Horschti is now known as Horscht (n=Horscht2@xbmc/user/horscht)
15:10:38 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
15:12:34 Quit _lifeless (Read error: 60 (Operation timed out))
15:13:47 Join mcuelenaere [0] (n=mcuelena@
15:16:20BuschelLlorean: you may solve the jumping forward when seeking backward issue with some small coe change in codecs/mpa.c. in function get_file_pos just comment the 4 code lines below /* Use the remainder to get a more accurate position */
15:17:38 Join petur [50] (n=petur@rockbox/developer/petur)
15:17:44 Quit kugel (Nick collision from services.)
15:17:50 Join kugel [0] (
15:18:47 Join ZincAlloy [0] (n=d9eed5b0@gateway/web/cgi-irc/
15:22:00 Quit CaptainKewl (Read error: 54 (Connection reset by peer))
15:22:20 Join CaptainKewl [0] (
15:26:37CIA-70New commit by 03bertrik (r21445): Fix missing i2c mutex initialisation for telechips.
15:28:16CIA-70New commit by 03zagor (r21446): Start next build as soon as 'make' is done. Upload while new build is running. Added svn revision and repo check.
15:31:11CIA-70New commit by 03bagder (r21447): Now the server makes sure all clients sent builds so that they always have ...
15:32:33 Join n1s [0] (n=n1s@rockbox/developer/n1s)
15:32:42CIA-70New commit by 03zagor (r21448): Add Id keyword.
15:33:12 Quit martian67 (Read error: 54 (Connection reset by peer))
15:34:07 Join martian67 [0] (n=xP@2001:470:b:356:221:91ff:fe8c:d8a7)
15:37:17 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
15:37:29 Join Zagor_ [0] (
15:38:05 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
15:38:19 Quit kugel (Read error: 60 (Operation timed out))
15:39:06 Quit Zagor (Read error: 60 (Operation timed out))
15:40:53 Quit Horschti (Client Quit)
15:41:33 Join Horschti [0] (n=Horscht2@xbmc/user/horscht)
15:43:26 Join markun_ [0] (n=markun@
15:43:59 Quit markun (Nick collision from services.)
15:44:01 Nick markun_ is now known as markun (n=markun@rockbox/developer/markun)
15:44:04CIA-70New commit by 03zagor (r21449): Adjusted revision check.
15:44:23 Nick Zagor_ is now known as Zagor (n=bjst@rockbox/developer/Zagor)
15:45:41 Join Domonoky [0] (n=domonoky@rockbox/developer/domonoky)
15:46:25 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:47:12 Quit robin0800 (Remote closed the connection)
15:47:13 Join funman [0] (n=fun@rockbox/developer/funman)
15:48:49linuxstbZagor, Bagder: Will your build-system work include using svn revisions more in the website - e.g. the build tables currently use timestamps, rather than revisions. (which I'm assuming is left-over from CVS)
15:50:03Bagderthe date stuff is a leftover yes, but it's actually usign revs only for what particular rev that builds etc
15:50:18 Quit Domonoky1 (Read error: 110 (Connection timed out))
15:51:06funmanwhat about making lcd_blit_yuv dependent on MEMORYSIZE > 2, or create a HAVE_LCD_BLIT_YUV ?
15:51:49linuxstbfunman: I assume mpegplayer doesn't run if MEMORYSIZE <= 2 then?
15:52:17funmanright, it allocates 2MB of memory at runtime. I disabled mpegplayer building for MEMORYSIZE <= 2 already
15:52:21CIA-70New commit by 03zagor (r21450): Removed svn revision check again. It doesn't work when downloading the file...
15:52:56 Quit bertrik (Read error: 110 (Connection timed out))
15:55:16linuxstbBagder: So you're saying nothing really needs changing?
15:55:32 Quit Horscht (Read error: 110 (Connection timed out))
15:57:45Bagderthe parsing and the table and things on the server should basically work unmodified
15:58:18Bagderwe'll just start this new server, and all participants run build clients and it should work
16:05:58 Join bertrik [0] (n=bertrik@
16:07:26CIA-70New commit by 03bagder (r21451): Updated "score" or "weight" of all the curreny builds based on gevaerts ...
16:09:22CIA-70New commit by 03bagder (r21452): back to the previous port number, make sure that clients only get each build ...
16:09:51Unhelpfuli've a decently-modern amd64x2 linux box to volunteer, if more build servers are wanted... since it sounds like this new system will not be too difficult for clients :)
16:09:52CIA-70New commit by 03funman (r21453): Do not use lcd_blit_yuv() if MEMORYSIZE <= 2, since mpegplayer won't run anyway (and won't build) ...
16:11:37LambdaCalculus37Unhelpful: I think all build servers are welcomed. :)
16:12:14*LambdaCalculus37 wants to contribute one himself, but that has to wait for a bit until he has the space to set it up
16:12:35Bagderyeah, all servers (now called clients) will be welcome
16:13:41LambdaCalculus37Bagder: I was considering buying a PlayStation 3, installing a Linux on it, and setting it up as a build server.
16:13:49ZagorLambdaCalculus37: gogogo! :)
16:14:10LambdaCalculus37Zagor: Actually, with that horsepower, I can also help rasher run binsize graphs on it. :)
16:14:30UnhelpfulLambdaCalculus37: as i understand it ps3 is really nothing special if you're not running code specialized for its vector units :/
16:14:52rasherLambdaCalculus37: that's not really needed. gevaerts is more than keeping up with that on his own
16:15:11LambdaCalculus37rasher: Okay.
16:15:51Unhelpfuland he might get mad if somebody shoved him out of his ranking ;)
16:16:00LambdaCalculus37Hehehe. :)
16:16:41LloreanSince builds are now scored, we could give out points for each build completed and score the buildclients.
16:18:30Bagderyes, the plan is to get proper logs for everything so that we can do things like that
16:18:43Unhelpfullong run, that's more meaningful than bogomips
16:18:55Unhelpfuli don't think bogomips accounts for dual cores, either ;)
16:18:56Bagderbut dead hard
16:19:01funmanamiconn: about FS #10274, perhaps it would be possible to move interrupt priority setting inside alpine_cdc plugin? I don't know if IPRD can be modified after timer has started.
16:19:04Bagdersince we use ccache
16:19:05LambdaCalculus37Unhelpful: What about 8 cores? ;)
16:19:11LloreanUnhelpful: It also just encourages non-technical people to volunteer too (see Folding@Home scores)
16:19:12Bagderand systems have different loads all the time
16:19:15Unhelpfulfor *multiplie* cores
16:19:27LloreanI wasn't referring to scores for technical reasons. More just for "fun"
16:19:32Bagderah yes
16:19:40BagderI'd like highscores too
16:20:04UnhelpfulBagder: over a long average, raw seconds/build should do just fine, seeing as all clients are using ccache ;)
16:20:34Bagderbut really, the exact speed is not that important
16:20:54BagderI want a rough estimate just to make sure the toughest builds get to the proper builders
16:21:13Bagdertime to close down here
16:21:23*Bagder exits
16:21:24 Quit Bagder ("It is time to say moo")
16:21:33Unhelpfulin the long run, is it going to continue handing out the same build to more than one client? killing half-finished builds seems like it might be wasting effort... although it will at least be populating ccache on the clients.
16:22:09 Quit Domonoky ("Leaving.")
16:22:31kugelUnhelpful: but you can just do bogomips*cores to reflect multicore in a basic way
16:22:52funmanabout the build table: since some asian scripts (japanese for example) can be written vertically, there must be some way to draw vertical text in web pages?
16:23:05 Quit dmb (Read error: 113 (No route to host))
16:23:19Unhelpfulfunman: i doubt that would allow one to rotate the characters...
16:23:58funmancan't unicode come to the rescue here, or we can only write 180° rotated characters ?
16:24:54Mikachujapanese is fixed width, so you can just transpose the characters
16:25:40 Quit Zagor ("Leaving")
16:26:16UnhelpfulMikachu: i thought the orientation of the characters remained the same, whether the "lines" are read top->bottom or left->right?
16:26:25Mikachuthey are
16:26:31funmanwell i'm not even sure that non rotated characters would be harder to read than current text
16:26:34Mikachuwhich is why you don't need any special html support
16:26:46Mikachu(for japanese)
16:27:11Unhelpfulah... transpose the block of text, not the glyphs :)
16:27:31 Quit BlakeJohnson86 ("Leaving.")
16:27:33Mikachuyes, instead of A B \n C D you would put C A \n D B
16:27:48Mikachu(it is right to left when it is top to bottom ;)
16:27:56funman < using css+svg
16:28:16 Join BlakeJohnson86 [0] (
16:28:37funmandoesn't work in plain IE6/7 though
16:29:24Mikachuyou can do anything with js+svg:
16:29:26 Join dmb [0] (n=dmb@unaffiliated/dmb)
16:29:35Mikachu(you need ff 3.1 for that demo though)
16:30:04funmanwe don't need diagonal text :)
16:30:43Unhelpfulfunman: are there *really* devs using ie6/7? :P
16:31:43funmanUnhelpful: not sure, i hestiated to add "(but I don't care)"
16:32:06 Quit kugel (Read error: 60 (Operation timed out))
16:32:23CIA-70New commit by 03amiconn (r21454): Change the X5/M5 remote keymapping to match the M3 way, improving consistency with the button labels. The most important change is that VOL+/VOL- are ...
16:33:39linuxstbLambdaCalculus37: I don't think a PS3 would make a good build client - at least mine is very noisy and can get very warm, so I assume it's not cheap to leave on 24/7... And it's not that fast.
16:34:41Unhelpfulhrm... the ARM asm 8-point IDCT is about 500B. perhaps it would be wise to build it both with and without the zero-row test, and select an appropriate version based on some heuristic (average quantizer, perhaps?)
16:35:18LambdaCalculus37linuxstb: I'll go for something else, then.
16:36:07Unhelpful360! ;)
16:36:16funmangame boy @#!
16:37:18LambdaCalculus37funman: That's just silly. :)
16:37:27Unhelpfulgame boy with a tetris-based visualization of the build progress. ;)
16:37:45 Quit DarkDefender (Read error: 104 (Connection reset by peer))
16:39:48 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
16:40:58 Quit Thundercloud (Remote closed the connection)
16:42:50 Quit markun (Read error: 110 (Connection timed out))
16:45:09***Saving seen data "./dancer.seen"
16:47:41 Quit mcuelenaere ()
16:47:57 Quit Sajber^ (Connection timed out)
16:48:14 Quit petur (Remote closed the connection)
16:59:58 Nick fxb__ is now known as fxb (
17:01:34 Quit CaptainKewl (Remote closed the connection)
17:06:33 Quit bertrik (Read error: 110 (Connection timed out))
17:08:41 Join CaptainKewl [0] (
17:19:08 Join Thundercloud [0] (
17:20:39 Quit funman ("free(random());")
17:20:50 Join faemir [0] (n=faemir@
17:23:32 Quit LambdaCalculus37 ("Fwump")
17:29:23 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
17:35:58 Join Sajber^ [0] (
17:36:08 Part n00b81 ("Leaving")
17:38:38 Quit ZincAlloy ("CGI:IRC (EOF)")
17:53:18 Join Lss [0] (
17:55:01 Nick lilltiger_ is now known as lilltiger (n=lilltige@
17:57:24 Quit fyrestorm (Read error: 104 (Connection reset by peer))
18:00:41 Join fyrestorm [0] (
18:03:22 Quit mt (Read error: 113 (No route to host))
18:11:24 Quit Thundercloud (Remote closed the connection)
18:14:07 Quit n1s (Read error: 60 (Operation timed out))
18:23:18 Join webguest61 [0] (n=ad0b4d0d@gateway/web/cgi-irc/
18:29:23 Join n1s [0] (n=n1s@rockbox/developer/n1s)
18:30:18webguest61do any of the rockbox targets have an implemented software usb target mode?
18:31:45 Join saratogahome [0] (n=41becb3b@gateway/web/cgi-irc/
18:32:06saratogahomeregarding walking the mp3 bitstream between seek points, i think this could be done very rapidly
18:32:23saratogahomei've walked it with inverse filterbank disabled and that doubles the speed
18:32:40saratogahomei think you could disable imdct, windowing, etc and probably walk it at 10x normal decode speed
18:33:29saratogahomesince i think the frames are fixed length, you might not even need to run the decoder at all, just the frame parser
18:33:58Buschelwell, when one frame was recognized you can calculate the length from the bitrate of this frame and just jump.
18:34:02saratogahomehmm but you still need to know how many PCM samples you're getting, so you'd have to at least get the block length information
18:34:19 Join schrottplatz [0] (
18:34:37saratogahomei'm thinking about this backwards, frames are fixed in time, not bytes right?
18:34:38Mikachusaratogahome: do you still know where in the stream each seek point is? someone (you?) said something about byte resolution
18:34:50schrottplatzcan i map the record button to an application without compiling
18:35:07saratogahomeMikachu: its in the VBR header at the start of the file
18:35:10saratogahometheres a seek table
18:35:16Mikachuyes that is the one i'm talking about
18:35:40 Part schrottplatz ("o.O")
18:35:46Mikachu"<Buschel> [...] even with a vbr-header the time resolution for seek is 100 entries only (~36 seconds in a 1h file, 6min in a 10h file). each time mark has a byte position saved with a resolution of 256 (256=eof)"
18:36:03Buschelseeking backwards via parsing seems difficult. we could use the latest byte position from the vbr-header and seek forwards again from this point
18:37:23saratogahomein ASF we estimate the frame position based on the bitrate, and then parse packets until we find the time stamp we need
18:37:37saratogahomea similar system should work here, but theres no time stamp, so you'd have to count frames
18:38:48Buschelyes, but if the vbr-encoder worked fine when creating the header, it should be no problem to count the frames when parsing from the last vbr-header's seeking point to the desired position
18:41:44Buschelhmm, but, in long files even the precision of the byte position within the stream is inaccurate (256 steps resolution). this will equal several dozen seconds on long files... so, seeking is inaccurate even with this approach.
18:41:55saratogahomeyes looking online it seems that given the start position of a frame from the VBR header, we could easily parse a frame, calculate its length, and then jump to the next frame
18:42:18saratogahomewow, if the seek table doesn't give you the byte the frame starts on, how do you find it?
18:42:28saratogahomesearch for the sync word?
18:42:40Buschelyes, parsing
18:43:40saratogahomeso we're back to needing to walk the entire file
18:44:06Buschelyes, that's the _only_ way to have precise seeking.
18:44:28saratogahomewe need MP3 in MP4 support :)
18:45:14***Saving seen data "./dancer.seen"
18:46:04 Quit webguest61 ("CGI:IRC")
18:49:57Buschelexample: we have a file ~100min playtime and 256MB large (~360kbit/s). then the resolution of the vbr-header's timestamps is 1min, the resolution of the byteposition is 1MB. so, if you by chance want to seek to exactyl 22min 00sec you will have a perfect vbr-timestamp, but the accuracy of the byte position is only 0-1MB off the desired real one (equals ~23sec accuracy). even the starting point of the search is pretty inaccurate.
18:52:38saratogahomelinuxstb: regarding the supported targets requirements, perhaps the explicit requirement for Windows, Linux and Mac should be relaxed a little? I'm thinking of devices such as the Gigabeat S which are more or less WIndows only in the OF and therefore are unlikely to have many users on Mac OS
18:53:54saratogahomeperhaps something like "install directions for Linux, Windows and Mac, or else relatively easy to install by LiveCD" would be better?
18:54:24 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
18:56:28 Join Cory` [0] (
18:56:28 Quit lilltiger (Read error: 104 (Connection reset by peer))
18:56:44 Join lilltiger [0] (n=lilltige@
18:59:49 Quit prona (Read error: 110 (Connection timed out))
19:00:09 Part n00b81 ("Leaving")
19:05:32 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
19:05:44 Part n00b81 ("Leaving")
19:07:53 Join BradfordDS [0] (
19:08:04 Quit BradfordDS (Client Quit)
19:12:15amiconnsaratogahome: You can find the exact position in an mp3 file by just parsing the frame headers, but you have to count from the beginning
19:13:25amiconnDepending on the mpeg version and layer, each frame represents a fixed amount of pcm samples. The number of bytes can be calculated from that, the bitrate, and the padding bit
19:20:16 Join tchan1 [0] (
19:20:26 Join Rosalina [0] (
19:21:43 Part Rosalina
19:22:36 Quit tchan1 (Client Quit)
19:23:43linuxstbsaratogahome: I see those requirements as being flexible. e.g. if everything else was in place for the Beast, but no-one had got beastpatcher working on OS X, I wouldn't be against calling it supported. But I still think the normal expectation would be to develop a cross-platform installation tool.
19:24:12saratogahomeok sounds reasonable
19:25:32 Quit bubsy ("Mrrrrreow!")
19:26:03 Join tchan1 [0] (
19:26:19 Quit nibbler_ (Read error: 110 (Connection timed out))
19:27:01GodEaterdid petur already upload the recording of the meeting somewhere ?
19:27:09GodEaterI forgot to ask before I left
19:28:06 Quit moos ("Rockbox rules the DAP world")
19:28:07 Quit tchan1 (Client Quit)
19:29:52pixelmaif he did I didn't notice either...
19:30:15GodEaterguess we'll just have to wait then :)
19:32:35 Join petur [0] (n=peter@rockbox/developer/petur)
19:35:34 Quit tchan (Read error: 110 (Connection timed out))
19:35:39 Join bubsy [0] (i=Bubsy@
19:43:18GodEaterpetur: ping
19:44:56 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
19:46:59JdGordonare the devcon discussion recordings online yet?
19:51:20 Join cmwslw [0] (
19:52:29peturGodEater: pong
19:53:00pixelmabluebrother: for \includegraphics - is there some kind of "margin" parameter I pass it, similar to height and width?
19:53:03GodEaterjust wanted to say thank you again for hosting. It was excellent fun.
19:53:04peturactually, there's only one and I gave it to the swedes to put it online
19:53:39*pixelma agrees with GodEater
19:53:40peturGodEater: I hope it reached the required/expected standards ;)
19:53:40 Join Thundercloud [0] (
19:53:52GodEatervery much so !
19:54:20peturthe room was a bit small and crowded, same for the bar/dining room
19:54:59 Part cmwslw ("Ex-Chat")
19:55:29peturGodEater: are you home already?
19:56:04peturand the chocolates...?
19:56:10GodEateryep, got back hours ago
19:56:17GodEaterchocolates have generated much love for me :D
19:56:31peturgood for you
19:57:08*AlexP chimes in with thanks too, both for petur and the fund :)
19:57:50GodEateryes, the weekend turned out a lot cheaper than I was budgeting :D
19:58:19bluebrotherpixelma: not that I'm aware of. What exactly is the problem?
19:59:02pixelmaI want to have some spacing between the player and the remote svg :)
19:59:32bluebrotherah, so you're putting two images in an image float?
19:59:57 Quit XavierGr (Read error: 104 (Connection reset by peer))
19:59:58peturJdGordon: I'll upload the recording to a temporary spot so you guys can start listening before it's on the rb site. I might redo the post-processing because I lost my patience with it on my slow laptop (was only a 1.4GB WAV file)
20:00:06 Nick Cory` is now known as Cory (
20:00:19 Nick Cory is now known as Cory` (
20:01:01pixelmabluebrother: sorry... float? Currently I just put two \includegraphics below each other
20:01:20pixelmaerr,, one below the other ;)
20:01:50bluebrotherah, ok. Well, you could add some \vspace{amount} between those.
20:02:29pixelmathanks, I'll try
20:03:08bluebrotherbut thinking about it, is there any target where we couldn't put the two images next to each other?
20:03:18peturaargh... Bagder deleted the recording I had put on my NAS for him :(
20:03:31*petur fires up his laptop to get the other copy
20:04:27peturthat will teach me, giving Bagder write access... stupid!
20:04:41pixelmabluebrother: you mean one on the left and the other on the right side? Not sure it won't look too crowded
20:05:10bluebrotherwell, that's the question :)
20:06:14 Quit saratogahome ("CGI:IRC (EOF)")
20:06:18pixelmaI think for the Iaudio (M3 at least) it could work out if the pictures are bottom aligned because the main unit's button labels are more at the top
20:06:31pixelmanot sure about the H300 etc.
20:06:41 Join cspotcode [0] (
20:08:45peturheh, the file was still there, couldn't see it due to access rights. Uploading now (28 minutes left)
20:16:40pixelmabluebrother: somehow I can't make it insert the vspace between the 2 pictures, somehow it's either above or below the 2
20:18:56 Join notlistening [0] (
20:19:24bluebrotherpixelma: can you pastebin the code somewhere?
20:23:22pixelma - I won't be here for long though
20:25:42*petur needs faster uplink speed :/
20:25:52bluebrothergnah. Why has been renamed?
20:27:24*petur looks forward to the new buildsystem. His new quadcore will be able to contribute in the buildsystem now :D
20:30:03bluebrotherpixelma: try adding a \newline before and after the \vspace
20:32:49 Quit n1s (Read error: 110 (Connection timed out))
20:32:51*GodEater will also be able to donate his work PC as a build server
20:33:13GodEaterit'll have to be cygwin or a vm on it tho...
20:33:48petursadly, no svn at my work, and I'm not allowed to use much bandwidth for personal uses during the day
20:34:08GodEatertechnically neither am I
20:34:13 Join d3v14710n [0] (i=d3v14710@unaffiliated/d3v14710n)
20:34:20GodEaterbut they don't monitor it, and it's not a massive amount of b/w in the scheme of things
20:34:26amiconnCygwin as a build server? Forget about that...
20:34:29petur(at my work = on the regular network behind the real firewall)
20:35:23peturGodEater: well, you've seen the bandwidth we have at work, so it would count more here... and that was already 3 links bundled.
20:35:55peturlet's hope cityconnect (ADSL2+) gets here soon, then it will be less of an issue
20:36:53GodEaterfingers crossed for you!
20:39:05peturalright... temporary location for the meeting recording:
20:39:17pixelmabluebrother: it works but still not a 100% as I like it: now I have spacing between the two and below (whiich I wouldn't like) - and it had an influence on the horizontal alignment (before both were centered (seems to break the center environment then, I guess?)
20:39:21peturJdGordon: ping ^^
20:39:48 Quit linuxstb (Remote closed the connection)
20:43:10bluebrotherinteresting. Shouldn't break the center environment AFAIK. Though you could try to use two center environments, maybe that works better?
20:45:18***Saving seen data "./dancer.seen"
20:46:46*petur would like to hear from people listening to the recording: is it clear enough or should there be more/better post-processing?
20:47:05*GodEater gets curl to sucking on the download to find out
20:47:10peturmy lazy side is showing up :)
20:47:16 Join guesticloe [0] (n=621cc6c1@gateway/web/cgi-irc/
20:47:25peturGodEater: what speed do you get?
20:47:54peturjust interested in how good this hosting company is...
20:48:06GodEater237K a second
20:48:18peturhmmm not really fast...
20:48:27 Join bertrik [0] (
20:48:28GodEaterit's about half what my line can do
20:48:36 Quit guesticloe (Client Quit)
20:49:01*petur waves at bertik. Back home?
20:49:47*bertrik waves at petur
20:59:36 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
21:00:22 Join stripwax [0] (
21:02:24 Quit stripwax (Client Quit)
21:04:03pixelmabluebrother: that helped for the centered alignment but with vspace I still have space
21:04:22pixelma ... added above and below the second image
21:09:48JdGordoncheers petur
21:11:15 Quit bubsy ("Mrrrrreow!")
21:23:23 Quit Buschel ()
21:23:50 Quit flydutch ("/* empty */")
21:27:44 Join cH360po [0] (n=57a6c8a9@gateway/web/cgi-irc/
21:28:23 Join saratoga [0] (i=9803c6dd@gateway/web/freenode/x-6d570fc1ce6748d6)
21:31:46cH360poHi there, I'm a newbie and I found a bug on the rb v3.2 - current build. When i release the hold button the display of my H340 display after several minutes of playback, the display fades in and out continously. That bug does not appear if "fade in" is deactivated.
21:32:39cH360po*When i release the hold button of my H340 after several minutes of playback, ... (should be readable then)
21:33:07bluebrothercH360po: Rockbox 3.3 is out. Have you considered updating first?
21:33:07 Join markun [0] (
21:33:29MikachucH360po: "3.2 - current build" is sort of an oxymoron too
21:33:41bluebrotherplus, what are you referring to as "3.2 - current build"? We have a release 3.2 and current builds, which are ocmpletely different
21:33:42cH360pobluebrother: yes it's the current nightly
21:33:49bluebrotherthen it's not 3.3
21:34:07cH360poi tried 3.2, 3.3 and the current nightly
21:34:16cH360poit appears on all of 'em
21:34:24gevaertsahm you mean, "from 3.2 to current"?
21:34:33cH360pogevaerts: yes
21:35:48 Join CathodeRayTube [0] (n=43ea47fa@gateway/web/cgi-irc/
21:36:00gevaertshm, I've never seen or heard that
21:36:22*gevaerts grabs his h320
21:40:33gevaertscH360po: it sort of flickers a bit here (i.e. seemingly random brightness changes). Is that what you're seeing as well, or do you have full fades to black?
21:41:03cH360pogevaerts: full fades to black ...
21:41:55gevaertsThis is an old build though (r19926, from february). I'll try a newer one
21:42:49cH360pothe bug did not appear till 3.1
21:43:46gevaerts3.1 was in december
21:45:14cH360pohmmm could it be that the wps-theme is not compatible?
21:45:22peturwoohoo.. 24049 bogomips (4 cores) :D
21:45:39gevaertscH360po: no. That should never have this sort of result
21:46:41gevaertsdoes you have a minimum time you need to keep it on hold?
21:47:13cH360poyeah, feels like it's several minutes
21:48:30 Quit markun ("leaving")
21:48:38cH360poand if i go to the settings menu then sometimes it stops blinkin
21:49:48rasherI'm seeing the same on my sansa, I think
21:50:16rasherMaybe related to caption backlight?
21:51:59*gevaerts leaves this to rasher and goes to sleep :)
21:54:08rashercH360po: do you have caption backlight enabled?
21:56:13*rasher is thinking the backlight changes might be queued up somehow
22:01:12cH360porasher: yes
22:01:37cH360pocould be that
22:01:52 Quit faemir (Remote closed the connection)
22:02:05rasherAnd also backlight off when on hold?
22:02:56rasherI'm almost willing to bet that combination causes backlight changes to get queued up
22:03:14rasherProbably requiring backlight fading as well
22:03:27 Join _Auron_ [0] (
22:03:54cH360porasher: i tested it and the longer it stays on hold, the longer it blinks
22:04:34rasherMakes sense
22:07:56cH360poi dunno how to create a bug report and what to write
22:08:22rasherI'll do it
22:08:59 Join stripwax [0] (
22:09:13cH360pothx a lot, i'm uploading a video where one can see the bug
22:10:05 Join faemir [0] (n=faemir@
22:11:19rasherkugel's the last person to touch backlight fading, wasn't he?
22:11:44bluebrotherIIRC, yes
22:13:08cH360poDo you need any further information rasher? Otherwise I would log off then :)
22:13:36rasherI don't think so, no.
22:13:59cH360pothanks a lot and bye bye
22:14:08 Quit cH360po ("CGI:IRC 0.5.9 (2006/06/06)")
22:14:34 Join nibbler_ [0] (
22:17:57 Join n1s [0] (n=n1s@rockbox/developer/n1s)
22:18:21 Join BradfordDS [0] (
22:18:41rasherAnyone knowledgeable about backlight (fading), please check out FS #10362
22:22:17 Nick Bagder_ is now known as Bagder (
22:24:10 Quit petur ("ZzzzzzzzzzzzzzzzzzzzzzzzZzzzzzzzzzzzzzzzzzzz")
22:24:58 Join bubsy [0] (i=Bubsy@
22:26:46 Quit Lss (Read error: 104 (Connection reset by peer))
22:29:30 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
22:34:12 Join Zagor [242] (
22:38:34 Quit BradfordDS (Read error: 110 (Connection timed out))
22:38:45Xerionwhat does a flashing charging light on h140 mean?
22:38:59Mikachuit means the matrix has you
22:39:07saratogacan we release the new PP bootloaders that do not boot the OF on plugin?
22:39:07Xerioni knew that already ;)
22:39:20Bagdersaratoga: yes we should
22:39:27saratogathey seem to be ready, i think all thats left is to post sansapatcher binaries for them
22:39:48saratogaBagder: the binaries are on FS waiting to go to the download server
22:40:13BagderI'm not up to it today, but will fix soonish
22:40:30saratogaok as long as you're aware
22:40:37saratogathere is no need to hurry
22:40:41MikachuBagder: i think you typod GodEater on your blag entry
22:40:54Xerionwell it fell to the floor, didn't work at all untill i connected the charger, now the light flashes, it keeps working even without charger though
22:41:02Xerionso I guess the battery is just very empty
22:41:16Tornecan we also release a new bootloader for iPod, even though there's no dualboot behaviour change for that?
22:41:17Xerionbecause it got shorted when it fell
22:41:23Xerionis that possible?
22:41:29Torneso that the released bootloader will have my OSOS boot patch in it? :)
22:41:32Bagderand petur as well :-/
22:42:14 Quit Sajber^ (Read error: 104 (Connection reset by peer))
22:42:48 Join Sajber^ [0] (
22:43:26bertrikI have trouble flashing the m3 gevaerts borrowed me, but I can't ask them since they are both probably asleep now
22:43:48Bagderboth gevaerts? ;-)
22:44:01bertrikgevaerts and markun I mean :)
22:44:24Mikachuand "lent"!
22:44:28*Mikachu tries to leave proof reading mode
22:45:22***Saving seen data "./dancer.seen"
22:46:21bertrikah I think I got it
22:47:10 Quit bluebrother ("leaving")
22:47:14 Join Grahack [0] (
22:53:08 Join Thundercloud_ [0] (
22:54:25 Quit Thundercloud (Read error: 101 (Network is unreachable))
22:58:53 Quit Grahack ("Leaving.")
23:06:00 Quit Rondom ("Ex-Chat")
23:06:59 Part cspotcode
23:10:03 Join markun [50] (n=markun@rockbox/developer/markun)
23:10:35 Quit markun (Client Quit)
23:11:22 Join markun [50] (n=markun@rockbox/developer/markun)
23:15:07 Join Blue_Dude [0] (
23:15:44Blue_DudePlease check out FS #10364 and FS #10365. They are short patches to cut down on debug output clutter. Thanks.
23:18:04 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
23:24:00 Join matsl [0] (
23:24:37CIA-70New commit by 03zagor (r21455): Exit on INT signal.
23:25:16 Join kugel [0] (n=kugel@rockbox/developer/kugel)
23:25:53CIA-70New commit by 03zagor (r21456): Changed rmtree() to 'rm -r', hopefully fixing the bug. ...
23:26:24 Nick Horschti is now known as Horscht (n=Horscht2@xbmc/user/horscht)
23:27:47 Join cmwslw [0] (
23:32:09notlisteningany know whee the instuction on how to creat your own .talk files is on the wiki i am being stupid
23:32:23 Part cmwslw ("Ex-Chat")
23:32:24kugelBlue_Dude: ah yea, those are annoying
23:39:36 Join safetydan [0] (n=deverton@rockbox/developer/safetydan)
23:39:49bertrikmarkun, gevaerts, I flashed the OF of the meizu M3, it does indeed support voice recording, so the flat cylindrical thing must be the microphone
23:42:43Zagorif anyone wants to test the new build system "live", follow the wiki instructions:
23:42:47markunbertrik: wow, cool
23:42:51ZagorI'll be here another hour at least
23:43:14bertriknow I just need a way to delete the silly voice recording I made ...
23:43:56Blue_Dudekugel: Sorry, I was off for a few minutes. I would appreciate a commit on those debug patches. Fewer messages are better unless you really need to see them.
23:45:10kugelBlue_Dude: how does that work now? it seems you need a logf build for those now?
23:45:16Mikachuhm, it's not trivial to find the svn checkout uri
23:45:25bertrik\o/ I can now control the backlight brightness with the touch strip
23:46:00markunbertrik: what did you change?
23:46:17bertrikmarkun, the problem with the interrupt was that I also needed to clear the interrupt in the interrupt controller, not just in the timer
23:46:19markungevaerts: we should have given him a meizu much earlier :)
23:47:03bertrikso what probably happened is that the interrupt kept on firing, starving the main "thread" of CPU time
23:47:08ZagorMikachu: link added
23:47:28markunbertrik: so, what's next now?
23:47:56bertrikI don't know, sleep I guess :P
23:48:03Mikachuthe "which" command on line 400 doesn't seem to work for me
23:48:08kugelmarkun: I'm still angry that you take him away from the AMSes, you shouldn't have given him one at all :P
23:48:20ZagorMikachu: what happens?
23:48:32MikachuZagor: it just runs "which" without any arguments
23:48:33bertrikby the way, I thought the meizu m3 interrupt controller was the same as the "standard" arm interrupt controller, but it's slightly different
23:48:49ZagorMikachu: what is your -archlist param?
23:49:14Mikachuprint $archlist; just above gives "1"
23:49:16Zagorthat's odd
23:49:27Zagorwhich system is this?
23:49:42Mikachuif i change "arm" to "1" it works
23:49:46Mikachuin the %which array
23:49:54rasher> You specified arch sdl but don't have in your path!
23:49:56Mikachuwell, it prints HELLO on stdout
23:49:56markunZagor: will you add some nice (and useless) text messages to broadcast to the clients?
23:49:57Zagorwell it's not supposed to be 1 ... :)
23:50:06kugel1 isn't a supported architecture though :P
23:50:12rasherZagor: Is it not sdl?
23:50:17Zagormarkun: yeah that's planned
23:50:23MikachuZagor: gentoo linux
23:50:35Zagorsdl targets are not built yet
23:50:35markunZagor: looking forward to it :)
23:50:40Mikachui am guessing it's not supposed to print to stdout either
23:50:53ZagorMikachu: uh, no
23:51:03rasherZagor: Ah, maybe it should detect if the archlist has something that is not in %which?
23:51:24Zagorrasher: yeah, I noticed that too just recently
23:51:29MikachuZagor: lines 27 and 18 look like they might be in the wrong order btw
23:51:40Blue_Dudekugel: no, DEBUGF still works fine. But if you use logf, it's easy to make all the debug messages in your file conditional with a #define statement. So if you need to debug the module, remove the comment markers around the #define. Otherwise, it's silent.
23:51:44Zagoryes, darn I forgot to commit that :(
23:52:05saratogai get a "curl: (6) Couldn't resolve host 'http:'" after running for a while
23:52:12CIA-70New commit by 03zagor (r21457): $upload must be set after $buildmaster...
23:52:25Zagoryes, I'll schedule an update. hang on.
23:52:30kugelBlue_Dude: you mean you don't need to enable logf in the advanced build menu?
23:52:42CIA-70New commit by 03zagor (r21458): Bumped revision.
23:53:29Mikachuif i don't change anything, it does this
23:53:30MikachuYou specified arch 1 but don't have in your path!
23:53:44kugelBlue_Dude: alright, I see: /* built without logf() support enabled, replace logf() by DEBUGF() */
23:53:49rasherZagor: That worked nicely
23:54:15Zagoryeah I'm quite happy with the update feature
23:54:41Blue_DudeKugel: Nope. All you need to do is #define LOGF_ENABLE, then #include logf.h. You need a few more #defines to make the code non-simulator friendly. Then replace all the DEBUGF's with LOGFQUEUE's or whatever.
23:54:48kugelif we're going to have templates for the commit messages, we surely also need the tracker to generate those for us ;)
23:55:06ZagorMikachu: I'm guessing your perl does the -s flag differently. but I don't know for sure.
23:55:07Mikachuuncommenting the use strict; is fun too
23:55:18Mikachuthis is regular perl 5.8.8
23:55:30Zagorunfortunately we only had debian boxes at devcon so the testing was a bit limited
23:55:31MikachuZagor: and also, you're not using -s
23:55:39Zagorrbclient is
23:55:46Mikachuif i add that, it works
23:55:49saratogaZagor: is it doing mulitple builds in parallel?
23:55:52Mikachulook at again
23:56:07 Quit matsl (Read error: 110 (Connection timed out))
23:56:20ZagorMikachu: ah, that's the difference. good catch!
23:56:31kugelBlue_Dude: ok, I understand, I'm gonna commit them
23:56:35Mikachuit seems to do something now
23:56:48Zagorsaratoga: not really. it "only" starts the next build while the previous is uploading
23:57:17ZagorMikachu: which shell is /bin/sh on your box?
23:57:22 Quit markun ("good night")
23:57:28Blue_Dudekugel: thanks!
23:57:29MikachuZagor: bash 4.0p17 or so
23:58:04kugelBlue_Dude: I think the vorbis one was by accident anyway, not sure about the albumart one
23:58:16Zagorperl 5.8.8 sounds a bit old for being gentoo? I have 5.10.0 on my debian testing.
23:58:53Mikachusomeone told me to stay away from 5.10

Previous day | Next day