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 2009-06-17

00:05:38 Join stripwax [0] (
00:08:57 Join FOAD_ [0] (
00:10:39 Quit AlexP ("No Ping reply in 90 seconds.")
00:11:25 Join AlexP [0] (n=alex@rockbox/staff/AlexP)
00:12:56 Join framo4 [0] (
00:14:40lilltigerthe slowdown migh be cos of the memory having poor random write access, so when it have filled an continious empty block and finds data it will take quite a bit longer for it, i seem to have just thoes issues with my SSD drive in my computer :(
00:16:25 Quit tvelocity ("Αποχώρησε")
00:18:26saratogai don't think write speeds have anything to do with it
00:18:50Lloreansaratoga: They might if it's trying to populate all the lists at once.
00:19:05 Quit evilnick_7 (" ajax IRC Client")
00:19:17saratogawhy would it be writing at all during the scan?
00:19:43LloreanWell, it almost certainly has to on some players due to RAM constraints.
00:19:56 Join HBK [0] (
00:20:57 Join moos [0] (i=mustapha@rockbox/staff/moos)
00:23:07 Quit FOAD (Read error: 110 (Connection timed out))
00:23:08 Nick FOAD_ is now known as FOAD (
00:24:09 Quit domonoky (Read error: 104 (Connection reset by peer))
00:24:45saratogai thought the reason we have to reboot when enabling the database was to clear up space in RAM for it
00:24:48 Quit perrikwp (" ajax IRC Client")
00:25:11 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
00:25:56Lloreansaratoga: Which would suggest very little RAM is used during the scan, which would suggest even more writing.
00:26:18saratogaif so then thats probably why its so slow
00:26:41LloreanMaybe a dedicated database plugin generation that can't be run during playback would make sense.
00:32:52 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
00:35:13 Quit ender1 (" Variables won't. Constants aren't.")
00:35:52 Join fdinel [0] (
00:38:53 Join nibbler_ [0] (
00:40:32 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
00:42:37***Saving seen data "./dancer.seen"
00:42:40 Join tapiocapudding [0] (n=chatzill@
00:42:56stripwaxsaratoga - hey. so, thanks for the mibbit pastebin yesterday −− I still can't see how that saves any multiplication (nor in lowmem tremor). Seems to me the best you get to save one pass over N/2 to do the unrolling (i.e. you save a loop, loads, and stores, but no mults). What am I missing ..?
00:43:06 Nick fxb is now known as fxb__ (
00:43:47tapiocapuddingOk, I followed the guide to recompile lang files after changing a string but when I do all the strings off the main menu are wrong. I noticed that the source .lang files were missing the proper line breaks, could this be the problem?
00:44:22linuxstbWhat text editor were you using to view the .lang files?
00:44:58 Quit Thundercloud (Remote closed the connection)
00:45:11linuxstbAnd what do you mean by "wrong"?
00:46:22stripwaxIn that pastebin we do 4[N/2] mults, and in tremor we do N mults on both prev and next frame, so both have 2N mults for windowing. I can't see a way around it
00:47:06 Quit chandoo (Read error: 110 (Connection timed out))
00:48:15tapiocapuddinglinuxstb: notepad, and they don't match up, i.e. the settings menu reads "Volume, Random, Auto-Change Directory, Theme Settings, Buttons Locked, Error Accessing Playlist File"
00:48:57linuxstbtapiocapudding: The line-ending problem is a limitation of notepad - it doesn't recognise unix-style line-endings in files.
00:49:55linuxstbIt sounds like something went wrong generating the .lng - e.g. you generated it for a different device than the one you are using.
00:50:14linuxstbWhat guide did you use?
00:50:20tapiocapuddingthe one in the wiki
00:50:33tapiocapuddingprobably just grabbed the wrong target id out of the config script
00:51:05linuxstbSounds likely.
00:52:15tapiocapuddingok, so in the config script it says "target_id=14", is that the id i should be using for the -i arg?
00:52:41 Quit nibbler_ ("Ex-Chat")
00:52:46 Join d3v14710n [0] (i=d3v14710@unaffiliated/d3v14710n)
00:53:27linuxstbYes, I think so. Is that what you used?
00:53:55 Quit perrikwp (" ajax IRC Client")
00:54:08tapiocapuddingi think so but i might be wrong
00:54:15 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
00:54:18tapiocapuddingi closed the cygwin terminal so i'll try again
00:54:43linuxstbCan you give me a link to the wiki page you're reading?
00:54:47saratogastripwax: i need to look at this more carefully
00:56:16 Quit framo4 ("muuh")
00:57:04saratogabut i think the savings should come because 1/4 the samples are identical to another 1/4, and 1/4 are the negation of the last 1/4, so unless tremor already uses this symmetry, 1/4 the multiplies could become negations, and 1/4 can be removed entirely
00:57:32stripwaxI don't think so: agree that 1/4 of the samples are identical to another 1/4, but the window isn't symmetrical
00:57:52stripwaxi.e. you need to multiply 1/4 by w and the other 1/4 by v, say
00:58:33saratogawhat are w and v ?
00:59:09stripwaxthe regions of the window applicable to the first 1/4 of those samples and the second 1/4 of those samples. the relationship between those parts of the window function is something like w^2 = 1-v^2
00:59:09saratogathe previous window and the current window?
00:59:56stripwaxlet's assume everything is long-long blocks so everything has same window. you still need to multiply all samples by the window function. so 1/4 of the samples in each frame might be the same, but the window function you multiply those 1/4 of samples by, is not
01:00:16tapiocapuddinglinuxstb: lemme get back to you in 1 second
01:00:19 Quit tapiocapudding ("ok bye")
01:02:09 Join tapiocapudding [0] (n=chatzill@
01:02:17tapiocapuddingok, sorry, i had to switch computers
01:02:20 Quit bertrik (Read error: 113 (No route to host))
01:03:05stripwaxAt least I think that's what going on. Wikipedia has been my guide.. also this has been pretty helpful:
01:03:05perrikwpHey all, I have a few questions about the timer in Rockbox: An HZ is equal to 1 second, correct? How many ticks are in one HZ? Does this depend on the target?
01:03:24linuxstbHZ is the number of ticks per second - defined to be 100 on all current targets
01:03:36perrikwpoh ok thanks
01:03:48tapiocapuddinglinuxstb: i was using the wiki file at
01:04:13saratogai was assuming the window functions woudl be the same up to a sign for the symetric parts but maybe thats not the case
01:04:18saratogai'll look into this when i have more time
01:04:47linuxstbtapiocapudding: So you're trying to run this line? "../../tools/genlang -e=english.lang -t=yourtarget:features -i=targetid -b=yourlang.lng yourlang.lang"
01:04:55saratogabut at least i think i was wrong about saving memory for wma
01:05:04saratogawhich is unfortunate, more IRAM would be useful
01:05:16tapiocapuddingexcept it won't work if unless you put "perl" in front of all the other stuff
01:05:17stripwaxsaratoga - ah, right. Cool. MDCT windows are *not* symmetric except for a sign, the relationship is more complex
01:05:25tapiocapuddingi should probably add that to the wiki
01:05:39saratogawell they're symmetric about N/2
01:06:00stripwaxwell, only if the prev frame and the current frame are the same size
01:06:19linuxstbtapiocapudding: So what did you use for the -t option? i.e. the list of features?
01:06:27stripwaxbut then the unrolled MDCT output is not symmetric about N/2 so still no symmetry benefit
01:06:34tapiocapuddingi just put -t=ipodnano
01:06:48saratogaso you're saying the most you save is the extra loads?
01:06:52tapiocapuddingso what should i put after the colon
01:07:00 Quit PaulJam (".")
01:07:33stripwaxok so I think in tremor right now, if this were implemented, you would save the loads+stores used to implement the unroll plus some adds/rsbs instructions
01:07:46linuxstbtapiocapudding: A long list of features... It's not easy to do manually - it's normally done automatically during a normal Rockbox build, which will probably be the best way for you to do what you want.
01:08:10tapiocapuddingok, so i should just edit the lang file and compile the whole rockbox build for my target?
01:08:22stripwaxthat corresponds to mdct_backward_step4 in my old profile, something like 2% of time spent. so maybe still an ok saving in time even if no memory benefit
01:08:24saratogastripwax: i have to go, but i noticed that ffmpeg recently updated their vorbis decoder to use the imdct symetry, so perhaps its worth looking at how they did it
01:08:33linuxstbtapiocapudding: Yes, that's what I would do.
01:08:38 Part d3v14710n
01:08:42tapiocapuddingok, thanks for the help.
01:08:55Mikachui think the languages are compiled first so you don't have to do the whole compile probably
01:08:56stripwaxsaratoga - interesting, yep sure thing.
01:09:43stripwax[for reference the windowing is something like 6% of time spent]
01:10:41perrikwpgoing along with my last question how many ticks are in a millisecond?
01:11:16pixelma"Prevent Track Skipping" will prevent skipping at all so you stay at the same track all the time, correct? I just wanted to make sure and spell that out, because the German translation in the tracker confused me a bit there, it sounded like it did something slightly different (and it looks like I can't use it anyways)
01:14:30 Join barrywardell [0] (
01:15:37JdGordon|perrikwp: HZ ticks per second..
01:15:41JdGordon|HZ usually being 100
01:16:36stripwaxsaratoga - I don't see an ffmpeg change (at least by eyeballing the shortlog), got a link?
01:17:31stripwaxperrikwp - so HZ/10 ticks per millisecond (i.e. usually less than one tick per millisecond..)
01:18:04perrikwpok thanks
01:18:53perrikwpsleep(x) is defined in ticks right?
01:19:15Mikachuif you sleep(HZ) you sleep one second, yes
01:20:46perrikwpthanks a lot just trying to make sure the plugin I'm writing is pretty accurate
01:22:23 Quit r0b- (Read error: 110 (Connection timed out))
01:23:33 Join chandoo [0] (
01:23:37 Join evilnick [0] (i=620ec27e@gateway/web/ajax/
01:24:35 Quit amiconn (Nick collision from services.)
01:24:37 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
01:24:45 Quit chandoo (Client Quit)
01:24:56 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
01:24:57 Quit pixelma (Nick collision from services.)
01:24:57 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
01:25:15 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
01:27:19 Join Thundercloud [0] (
01:35:36 Quit Thundercloud (Remote closed the connection)
01:35:40 Quit CIA-37 ()
01:36:40 Quit Genesis ("parti fabriquer des IP")
01:38:20 Quit stripwax ("")
01:38:27 Quit barrywardell ()
01:43:29 Join tvelocity [0] (
01:46:35 Join CIA-70 [0] (n=CIA@
01:47:29 Quit Rondom ("Ex-Chat")
01:48:08 Quit BlakeJohnson86 ("Leaving.")
01:48:51 Quit gevaerts (Nick collision from services.)
01:48:59 Join gevaerts_ [0] (n=fg@rockbox/developer/gevaerts)
01:49:47 Quit efyx (Read error: 104 (Connection reset by peer))
01:49:48 Join BlakeJohnson86 [0] (
01:49:56 Quit moos ("Rockbox rules the DAP world")
01:59:57 Quit tapiocapudding ("ChatZilla 0.9.84 [Firefox 3.0.11/2009060215]")
02:00:32 Join efyx [0] (
02:03:26 Quit robin0800 ("Leaving")
02:06:39 Quit bmbl ("Woah!")
02:08:27 Join robin0800_ [0] (
02:09:09 Quit Horscht (Read error: 110 (Connection timed out))
02:10:33 Quit robin0800_ (Client Quit)
02:10:55 Join robin0800 [0] (
02:11:27 Quit robin0800 (Client Quit)
02:13:27 Join dandin1 [0] (
02:14:22 Join robin0800 [0] (
02:15:13 Join r0b- [0] (
02:16:17 Part toffe82
02:20:40 Quit perfectdrug_ ()
02:27:20 Quit perrikwp (" ajax IRC Client")
02:27:45 Quit dandin1 (Read error: 60 (Operation timed out))
02:27:59 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
02:30:44 Join dandin1 [0] (
02:40:45 Quit jgarvey ("Leaving")
02:42:39***Saving seen data "./dancer.seen"
02:43:56 Quit robin0800 ("Leaving")
02:44:15 Join robin0800 [0] (
02:51:43 Quit evilnick (" ajax IRC Client")
02:58:08 Quit efyx (Remote closed the connection)
03:00:44 Quit robin0800 ("Leaving")
03:01:03 Join robin0800 [0] (
03:05:04 Quit robin0800 (Client Quit)
03:07:08 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.11/2009060215]")
03:11:54 Quit amiconn (Read error: 104 (Connection reset by peer))
03:11:57 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
03:12:08 Quit pixelma (Nick collision from services.)
03:12:10 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
03:12:17 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
03:12:28 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
03:50:32 Quit JdGordon| (" ajax IRC Client")
03:57:24 Quit martian67 (Connection timed out)
04:00:55r0b-how does one launch helloworld.lua
04:07:56safetydanr0b-: stick on your player somewhere than "play" it
04:17:27r0b-and where mighti find it? in the rockbox source or is it part of the build for like the e200
04:39:21 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
04:42:43***Saving seen data "./dancer.seen"
04:45:35 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
04:46:16 Quit martian67 (SendQ exceeded)
04:47:03 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
04:47:45 Quit martian67 (SendQ exceeded)
04:48:36 Join martian67 [0] (n=martian6@about/linux/regular/martian67)
04:52:31 Quit dandin1 ()
05:02:13safetydanr0b-: has a link to it
05:05:35r0b-i found it :)
05:15:58 Quit bubsy (Read error: 60 (Operation timed out))
05:28:03 Join antil33t [0] (n=Mudkips@
05:44:35 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
05:45:12 Quit Horscht (Client Quit)
05:48:59 Join x72_l [0] (
05:49:04 Quit lilltiger (Read error: 104 (Connection reset by peer))
05:49:11 Join lilltiger [0] (n=lilltige@
06:04:42 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
06:09:23 Quit perrikwp (" ajax IRC Client")
06:09:52 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
06:40:23 Nick intrados_ is now known as intrados (
06:42:46***Saving seen data "./dancer.seen"
07:14:29 Join matsl [0] (
07:37:48 Quit perrikwp (" ajax IRC Client")
07:38:21 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
07:44:11 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:46:33 Join markun [50] (n=markun@rockbox/developer/markun)
07:50:52 Quit timc (Read error: 110 (Connection timed out))
07:55:16 Join AndyI [0] (i=AndyI@
07:57:11 Join jordoex [0] (
08:05:22 Join jordoex_ [0] (
08:07:00 Quit AndyIL (Read error: 110 (Connection timed out))
08:09:37 Join kyp [0] (n=kyp@
08:11:06 Part kyp
08:12:32 Join evilnick_home1 [0] (
08:12:49 Join Rob2223 [0] (
08:18:10 Join Xakh [0] (
08:19:53 Quit Rob2222 (Read error: 60 (Operation timed out))
08:20:33 Quit jordoex (Read error: 110 (Connection timed out))
08:21:49 Join thecoolgman [0] (n=446e5e7d@gateway/web/cgi-irc/
08:22:51 Quit thecoolgman (Client Quit)
08:25:02 Quit evilnick_home (Read error: 110 (Connection timed out))
08:26:12 Quit Dhraakellian ("reboot! shock! horror! (well, 62 days ain't *too* bad)")
08:26:36 Join Dhraakellian [0] (n=ntryon@
08:26:42 Quit Dhraakellian (Read error: 104 (Connection reset by peer))
08:27:55 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:28:02 Join ender` [0] (
08:31:20 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
08:38:28 Quit safetydan ("Leaving.")
08:42:47***Saving seen data "./dancer.seen"
08:48:03 Join flydutch [0] (
08:48:13 Quit tessarakt ("Client exiting")
08:50:00 Join timc [0] (n=aoeu@
08:52:03 Quit cool_walking_ (" ajax IRC Client")
08:52:05 Join stripwax [0] (
08:53:45 Quit stripwax (Client Quit)
08:54:00 Part Xakh
08:58:34pondlifeHmm, can we license something yet say it's from "an unknown author"... ? See
08:59:04LloreanNo, we can't.
09:00:08LloreanAt least, we'd need to know enough about the original author to know if he licensed it CC-by-SA or permissively enough that CC-by-SA could be applied to it
09:00:18pondlifeI thought as much..
09:00:29LloreanAlso, I like your idea of dismissing the splash by simply being able to ignore it.
09:00:42pondlifeThat's how it works for the blind, really.
09:00:42LloreanAssuming I'm reading it right.
09:01:00pondlifeYes, basically it's there until a key is pressed.
09:01:10LloreanAnd that key function is executed as normal
09:01:29pondlifeProbably best to keep a 4s timeout for those who don't realise they can dismiss it.
09:01:40LloreanI'd say leave the splash timeouts as they are already.
09:01:42pondlifeBut either way there's no impact on usability.
09:01:56 Join petur [50] (n=petur@rockbox/developer/petur)
09:02:02 Join bertrik [0] (
09:02:04pondlifeBut have a timeout too
09:02:14*linuxstb likes that idea too
09:02:44LloreanKeep the timeouts as they are now, but make splashes "ignorable" by having any button press immediately clear the splash *and* do what that button's supposed to do?
09:03:37pondlifeI think that achieves the aims (1) don't annoy Jd (2) don't have any timing dependencies (3) simple
09:03:57LloreanAnd (4) any accidents that can happen are the least harmful sort (missing a splash)
09:04:14 Quit perrikwp (" ajax IRC Client")
09:04:27 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
09:04:33pondlifeI'd certainly think there's no need to have both a splash() and an asplash() call ;)
09:04:42LloreanI think most of the work for that is actually already done by Kugel's patch. IIUC it makes splashes dismissable by any key and the splash function returns the keypress used to dismiss them
09:05:40pondlifeIdeally, the calling code should not need to do any special processing - the splash should return with the button press still buffered.
09:05:47pondlifei.e. no return value
09:06:05pondlife(We do have a button buffer, don't we?)
09:09:42pixelmait's cool that JdGordon removed his first post entirely in this splash thread though
09:09:42pondlifeIt only needs to be one button (including combos!)
09:09:42LloreanI also don't see anything wrong with hiding "Resume Playback" when resuming is impossible, but I think a better option is that resuming after the end of a playlist restarts it (I think there's already a patch for this too)
09:09:42pondlifeYes, make the playlist index auto return to 0 at the end of a playlist (when not repeating).
09:09:42pixelma(actually replaced it with something else, that doesn't explain why the thread is there)
09:09:42pondlifepixelma: ?? It's on the ML, isnt' it?
09:09:42 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
09:09:42pondlifeAh - you mean the forum?
09:09:42pondlifeThat must be a recent edit
09:09:42*Llorean notes that there's not actually a patch attached, hrm.
09:12:30pixelmayou mean the "resume" patch?
09:21:04pixelmaor ask Mikachu to upload the patch himself
09:22:21 Quit perrikwp (" ajax IRC Client")
09:22:54 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
09:26:56 Join Thundercloud [0] (
09:36:48 Quit intrados (Read error: 104 (Connection reset by peer))
09:36:58 Quit bluebrother (Read error: 113 (No route to host))
09:37:11 Join nibbler_ [0] (
09:37:58 Join intrados [0] (
09:43:09 Quit bertrik (Read error: 113 (No route to host))
09:43:18 Join safetydan [0] (n=Dan@rockbox/developer/safetydan)
09:50:34 Quit Thundercloud (Remote closed the connection)
09:55:53CIA-70New commit by dan (r21315): Commit FS #10324 from Jeffrey Goode. ...
09:58:42safetydanwow... it's been a year since my last commit
09:59:02safetydanmust be something about winter that makes coding easier
10:00:52*JdGordon goes to bed, but first says the making splash() return the button is a bad idea for a few reasons.. 1) its a lot of work, 2) it means passing in a context to the splash also
10:01:53JdGordonalso, I tinhk passing the button back will lead to confusion
10:02:43*gevaerts_ also doesn't really like that idea at first sight
10:02:47 Nick gevaerts_ is now known as gevaerts (n=fg@rockbox/developer/gevaerts)
10:08:58MikachuLlorean: 10343?
10:09:18Mikachuoh it didn't attach the file
10:09:46linuxstb"it" didn't ? ;)
10:09:48Mikachuah i know why
10:10:04Mikachui filled in the box, forgot to fill in the summary and got the red text, filled in the summary and then the file box was emptied for me
10:10:12Mikachuthe same thing happened now when i left the comment box empty
10:10:47 Quit BHSPitMonkey ("Ex-Chat")
10:19:13 Quit r0b- (Read error: 110 (Connection timed out))
10:19:28 Join r0b- [0] (
10:23:03 Join robin0800 [0] (
10:24:02 Join efyx [0] (
10:25:04 Join PaulJam [0] (
10:28:53pondlifeLlorean (and JdGordon if you're still around): non-button-eating abortable splash patch needs testing - see
10:29:18pondlifeVery simplistic - probably won't work on touchscreen, and need to see how it affects USB.
10:33:43 Quit robin0800 ("Leaving")
10:34:01 Join robin0800 [0] (
10:38:34 Join Blue_Dude [0] (
10:39:51Blue_Dudesafetydan: Thanks for r21315 / FS #10324!
10:40:07safetydanBlue_Dude: no probs, looking at committing your other one as well
10:40:15safetydanjust want to run it on my h120 first as a smoke test
10:40:33Blue_DudeNo prob.
10:42:49***Saving seen data "./dancer.seen"
10:45:48 Quit Rob2223 (Read error: 110 (Connection timed out))
10:47:54CIA-70New commit by dan (r21316): Commit FS #10335 from Jeffrey Goode. ...
10:48:45 Quit perrikwp (" ajax IRC Client")
10:48:48safetydanhrm, Blue_Dude, I just realised that I assumed that was your real name. I'm hoping it is.
10:49:07 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
10:49:23Blue_DudeIt is. I didn't cheat in FS. Thanks for asking.
10:49:35Blue_DudeAnd thanks for the commit.
10:49:45*pixelma is a bit annoyed by fml hijacking the splashes thread, but doesn't really want to add more fuss to it by sending a mail (so hopes he reads logs)
10:49:58 Quit kachna (Read error: 113 (No route to host))
10:50:51 Join Grahack [0] (
10:51:30 Part Grahack
10:52:07*Llorean thinks customizable button hold times is a bit excessive.
10:52:38LloreanEspecially since some players have hardware times (power button on H100, etc) that we can't do anything about anyway.
10:55:07pondlifeHow many places does the "very long press" apply? (i.e. targets/contexts)
10:55:33LloreanPower on iPods, power on e200 *if* the player is locked up and you need the hardware poweroff.
10:55:49LloreanBut the problem he's trying to address is people being unable to short-press.
10:56:04LloreanI'd imagine the same people need to adjust keyboard repeat timings then, though.
10:56:17 Join Genesis [0] (
10:56:50LloreanAs far as I know, the only "normal" use of very long press vs long press is the shutdown on iPods. A mere long press stops playback, and is the only time a long press is also used on that button afaik.
10:57:15LloreanIf you release the button when you see the WPS vanish you're fine. Also, I imagine few people with iPods even stop playback since the OF didn't have a "stop" concept anyway
10:57:26*Mikachu changed shutdown to very long select+play on his ipod (shut down in the jpg viewer one too many times :)
10:57:38pondlifeSo, aside from hardware/startuip cases we only have that one? Could it not be remapped?
10:57:57Lloreanpondlife: There's nothing free in the WPS
10:58:15LloreanMenu = menu, long menu = WPS, left/right skip and seek, select selects or context menu
10:58:26LloreanSo stop needs to be on long play, as that's the only button free
10:58:35LloreanAnd that leaves very long *something* for powerdown, or we drop stop as a concept.
10:58:41Mikachumenu+left/right and left+right are free, but somewhat tricky to press :)
10:58:42LloreanAnd you need "stop" to be able to start new playlists.
10:58:50LloreanCombos are pretty bad on an iPod
10:59:14pondlifeOK, so maybe the shutdown menu option is a better solution for iPods?
10:59:34pondlife(I don't have one, so I don't really know how it would "feel".)
10:59:35LloreanI don't think it's *really* a problem.
10:59:41pondlifeAh, ok.
10:59:42LloreanI've never had any difficulty differentiating the two
11:00:01pondlifegevaerts: ping
11:00:28pondlifegevaerts: So non-button events go into the button queue?
11:00:44pondlifeEven if hold is on ;)
11:00:49Mikachua few years ago i had this applied, but didn't bother fixing it when it conflicted this time;a=commitdiff;h=7fecbe153ee3
11:00:59Mikachuadds some shutdown menu entries
11:01:04gevaertseven if hold is on, yes :)
11:01:39pondlifegevaerts: I guess I need to peek the queue then and ignore some things...
11:02:16gevaertsthat's hard as well. What if you first get a few SYS_EVENTs and then the button you expect?
11:03:36pondlifeI'm only checking, not changing the queue
11:03:48pondlifeIf there are any buttons in there, return.
11:03:57pondlifeThe queue will then be processed.
11:04:16pondlifeSYS_EVENTs will not be processed while the splash is displayed though
11:04:21 Quit nibbler_ (Read error: 110 (Connection timed out))
11:04:31pondlifeThe dequeuing happens as normal, outside of the splash.
11:04:35gevaertsthat's still a lot better than having them thrown away
11:04:56Mikachuyou might have to guard against scrollwheels on ipods
11:05:18pondlifeIs there a non-destructive way to read queues at the moment?
11:05:28pondlifequeue peek (position), I guess
11:05:47gevaertsah yes, scrollwheels... What should those do in this system?
11:06:08pondlifeWhen moved, they cancel the splash
11:06:17pondlifeSame as any button, I'd think
11:06:25pondlifeOr do they continuously send?
11:06:53gevaertsI know the ipod scrollwheel tends to send move events whenever you breathe in its general direction
11:07:01pondlifeAh, ok
11:07:19pondlifeWell, let's try for a sompler solution then restrict that if it proves necessary
11:07:30gevaertsi.e. if you do that, people will dismiss a lot of splashes accidentally
11:08:11pondlifeIndeed, but we can add that later (but before commit)
11:08:35pondlifeI just want to see if the basic idea is feasible first
11:11:05*gevaerts thinks he won't like easy-to-dismiss splashes at all...
11:11:26gevaertsbut yes, maybe trying it could make me change my mind
11:16:16 Join nibbler_ [0] (
11:20:30 Quit robin0800 ("Leaving")
11:20:40 Quit Blue_Dude ("ChatZilla 0.9.84 [Firefox 3.0.11/2009060215]")
11:20:48 Join robin0800 [0] (
11:28:40 Join ucchan [0] (
11:32:02 Quit robin0800 ("Leaving")
11:32:22 Join robin0800 [0] (
11:36:53pondlifegevaerts: Updated patch on - hope I got the queue stuff right :)
11:40:50gevaertspondlife: I'm not sure how queues work internally, so I don't know :)
11:41:53pondlifeheh, that makes 2 of us
11:42:14pondlifeWell, it seems to work on the sim, maybe someone could try it with a wheel?
11:42:46 Quit robin0800 (Read error: 104 (Connection reset by peer))
11:43:00 Join robin0800 [0] (
11:43:54ucchanFor Release Notes for Rockbox 3.3, FS #9383 is solved by applying my patch. Then could anyone commit it?
11:44:55 Quit robin0800 (Client Quit)
11:45:14 Join robin0800 [0] (
11:46:28ucchanIf it is possible, could you commit FS #9707 too ?
11:47:47 Quit robin0800 (Client Quit)
11:47:48LloreanYou're supposed to bring up such patches during the freeze, not after the branch. At this point only absolutely critical patches should be applied to the branch.
11:48:06 Join robin0800 [0] (
11:48:35pixelmaI didn't know that (though applying to both branchses is a bit harder but)
11:49:04Lloreanpixelma: The whole point of the branch is that it's supposed to be a release candidate. Unless some severe bug escaped our notice, it's not supposed to change once we branch.
11:49:08gevaertsI think bugfixes are still fine after the branch, but preferably *soon* after
11:49:32Lloreangevaerts: Very very safe, or very very important ones mainly
11:49:54LloreanSince the branch isn't likely to get much testing, realistically
11:50:41pixelmaso a German language update shouldn't go in anymore?
11:51:01LloreanLanguage updates aren't the same thing.
11:51:06 Quit perrikwp (" ajax IRC Client")
11:51:29LloreanThey should be fine all the way up until the release.
11:51:40ucchanIf FS #9383 applies, then known bugs disappears from one.
11:51:45 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
11:55:33 Join kachna [0] (
11:59:13pondlifeIf we think that #9383 is serious enough to mention in the release notes, it should be serious enough to patch, IMHO...
12:00:22pondlife(I'm not sure how serious it is, but then I don't use the chessclock plugin.)
12:01:52pixelmaI agree that it should get some testing before (I mean potentially it could cause another bug we couldn't describe in the release notes), on the other hand it's "only" a plugin
12:03:22gevaertsI'm not against applying it. The bugs in that particular plugin are pretty serious I think
12:03:49 Quit robin0800 ("Leaving")
12:04:08 Join robin0800 [0] (n=robin080@
12:04:23LloreanIt's been in the release notes since 3.0
12:05:33gevaertsIIRC my opinion then was to actually drop the plugin from the release...
12:05:35LloreanThe patch really should've been applied much earlier if it was to be in this release. Is it very unlikely to have new problems?
12:05:53 Join plx140 [0] (
12:05:56Llorean"it" being the patch
12:07:35plx140hello everyone. i have a some strange behaviour going on with my h320 - theres a slight rumbling noise of the HD while recording from radio. plu, the display is flickering a bit every time the hd is accessed... any idea anyone?
12:08:37LloreanI'm not even sure what 9707 does. It says it "fixes" several problems but I can't tell from the description what the fixed new behavours are. I'm not even entirely sure all of them are problems.
12:08:54Lloreanplx140: Old battery, at first guess.
12:09:56plx140the battery is qujite new. its a cameron sino 2200mah, a replacement for the dead stock battery.
12:10:10plx140happens with ac adaptor as well, btw
12:10:25LloreanIs the HD stock?
12:10:29plx140it is
12:11:03LloreanSo you've only replaced the battery? It's still quite possible the battery was a fake, it's not too uncommon.
12:12:07plx140hmm... i ordered it at a reliable dealer...
12:13:22 Quit ucchan ("Leaving...")
12:16:26plx140now, this is what happened: i first changed the battery and everything worked fine. a few weeks later i repaired the "hold" button (the knob on the case), NOT on the pcb.... and after that, in a "magical" way, the screen started flickering - i opened the h320 again to see if ive done something wrong or whatever. but (since i only replaced the hold-button ) everythings was ok
12:19:30plx140its s slight flickering when the hd is accessed and i decided to take it for usual - until i heard the noise during radio recordings. i mean i want to replace the hd with a CF card anyway, but i fear that theres somnething basically wrong or broken.... dont havea clue :(
12:22:08pondlifeplx140: Probably best if you post on the forum.
12:25:41 Join gregzx [0] (
12:29:34 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
12:30:12plx140tried that already
12:31:26plx140ok, another question: doea anyone have a tipp for a 32gb cf card that is (as) fast (as the stock hd) and not too expensive?
12:32:48Lloreanplx140: Which forum did you post on?
12:33:00LloreanAlso, is speed that particularly important? Do you transfer a lot back and forth?
12:33:01plx140rockbox forum
12:34:19plx140hmm... well, it depends - basically i just wanna go sure to NOT have a slow transfer. use it for file xchange sometimes
12:34:36 Join einhirn [0] (
12:35:08plx140rockbox forum -> hardware
12:35:45 Quit einhirn (Read error: 104 (Connection reset by peer))
12:36:13 Join einhirn [0] (
12:37:56plx140Llorean: it shouldnt be slower than the HD. i mean, i searched a bit, read forum threads and so on but i thought it would be useful to ask HERE. i took a look at the kingston elite pro 32gb 133x, ridata lightning 32gb 233x, topram 32gb (no idea of the speed), transcend 32gb. the kingston is the most attractive as its the cheapest
12:38:49 Quit J-23 (Remote closed the connection)
12:38:55 Join J-23 [0] (
12:42:53***Saving seen data "./dancer.seen"
12:44:54markunI'm trying out the timestretch feature on my Gigabeat. I've turned it on, rebooted, but when I change the speed in the pitch menu, I don't hear any sound until I change the speed back to 100%. Anything I'm doing wrong?
12:46:51gevaertsIs there a reason why the "rewind duration on pause" setting is only for headphone unplug? I'd expect that to be a useful setting for everyone
12:47:03gevaertss/everyone/all sorts of pause/
12:47:52markungevaerts: would be useful indeed
12:48:19LloreanPauses are usually intentional, headphone unplugs are considered likely to be an accident I guess.
12:48:50LloreanAlso, the pause on headphone unplug takes a second to respond, so the rewind is in part to compensate for the delay before it pauses.
12:49:46markunstill, I would like the feature. If I pause and it's in the middle of a word, it's difficult to understand when I resume.
12:50:15markunbut how to add it without turning it into a settings nightmare..
12:51:08LloreanI've never really had a problem with it. I have a lot of spoken word stuff I use, and it's pretty rare that I have a problematic word that I don't know from context, and yet badly need to know.
12:51:39plx140Llorean: any idea concerning my post about the CF stuff?
12:52:34 Join bubsy [0] (i=Bubsy@
12:52:34 Quit bubsy (Remote closed the connection)
12:52:41 Join webguest97 [0] (n=3ba72487@gateway/web/cgi-irc/
12:52:52Lloreanplx140: It's really sorta off-topic here. Hardware mods like that don't really have anything to do with Rockbox, they're not even Rockbox-specific
12:53:18 Quit safetydan ("Leaving.")
12:53:19 Join bubsy [0] (i=Bubsy@
12:53:28 Join asdf2 [0] (
12:54:31plx140right, but i hoiped that maybe people here have experiences with that and could gimme a hint, even as pm
12:54:54asdf2is there an RSS feed *just* for releases, or is the Major Changes feed all there is?
12:56:29LloreanReleases happen every 3 months
12:56:39LloreanThere's not really a need for an RSS feed for a scheduled event.
12:56:52 Part Llorean
12:57:30 Join Llorean [0] (n=DarkkOne@
12:57:41asdf2Oh I see, thanks
12:57:44 Quit webguest97 ("CGI:IRC (Ping timeout)")
12:57:46 Join mib_q0co51 [0] (i=5ed6497d@gateway/web/ajax/
12:59:09markunLlorean: any idea what I'm doing wrong with the timestretch feature?
12:59:38 Join pondlife1 [0] (
13:00:15pondlife1gevaerts: I've long thought the rewind-on-pause should have 3 options - never, headphone removal, always
13:00:27 Quit perrikwp (" ajax IRC Client")
13:00:35 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
13:01:09 Quit mib_q0co51 (Client Quit)
13:01:11pondlife1gevaerts: You might want to look at FS #9448
13:01:25Lloreanmarkun: I seem to remember someone mentioning that you have to switch to the third mode (the one that says "Pitch" and "Speed") and use it first, before the other one will work properly
13:02:09pondlife1That should not be required
13:02:15markunLlorean: I only tried it in that mode. Let me try again. The file I'm testing with is 48kHz, could have something to do with it.
13:02:27pondlife1markun: Very possibly
13:03:11pondlife1markun: Now I've read the logs, sounds like you're seeing the same as kugel was - try the patch on ?
13:05:35 Join TheSphinX^ [0] (
13:06:03markunLlorean, pondlife1: you guys are right, after changing the pitch first a bit it works.
13:06:22pondlife1markun: Does the patch fix it? I'd like to commit if so
13:06:33markunstill have to try it
13:06:46pondlife1OK, would be appreciated if you can.
13:07:13pondlife1Or I guess I could commit anyway ..
13:07:17markunI will
13:07:30amiconnLlorean: There are more uses of long press and very long press on the same button than just ipods
13:09:17LloreanWhat other players use them, where they're not a hardware feature?
13:11:25markunpondlife1: yes, that patch fixes it
13:12:00pondlife1Thanks - will commit
13:12:15 Part pondlife1
13:14:41 Quit pondlife (Read error: 110 (Connection timed out))
13:16:59 Part asdf2
13:18:39 Join pondlife1 [0] (
13:19:54CIA-70New commit by pondlife (r21317): Rework internal timestretch enable/disable code - fix FS #10341(hopefully).
13:24:35 Quit linuxstb (Remote closed the connection)
13:27:32LinusNpondlife1: "Assume timestretch will not not be used" - double negation ;-)
13:27:50pondlife1hehe, single typofication
13:28:23 Part pondlife1
13:29:29 Join Rob2222 [0] (
13:30:36CIA-70New commit by pondlife (r21318): Fix silly typo in comment, but nothing more.
13:31:11 Quit PaulJam (Nick collision from services.)
13:31:16 Join PaulJam_ [0] (
13:42:32 Join Shadz [0] (
13:43:16 Join zxczxc [0] (n=aa@
13:53:42 Join genesis_ [0] (
14:02:13 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
14:02:45 Quit perrikwp (" ajax IRC Client")
14:03:03 Quit zxczxc ()
14:03:32 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
14:05:26 Quit Genesis (Read error: 110 (Connection timed out))
14:05:54 Nick genesis_ is now known as Genesis (
14:06:18*Genesis tries to compile rockbox with openembedded
14:07:27 Join fml [0] (n=4fd3c74f@gateway/web/cgi-irc/
14:10:18fmlpixelma: I sent the email by just pressing "Reply" in my mail program and changed the subject. I think I've read somewhere this were the correct way to do that. Since then the displaying program can recognize that the posts are connected but the subject has been changed (by placing " (was: orig.subject)"). Or how am I supposed to post a reply with another subject?
14:16:51 Join perfectdrug [0] (
14:18:40 Quit robin0800 ("Leaving")
14:19:29 Join robin0800 [0] (
14:24:50 Quit Zarggg (Read error: 113 (No route to host))
14:28:14GodEaterfml: you're not supposed to post a reply with a different subject, since if it's about a new subject, it's not a reply.
14:31:54fmlGodEater: ok, now I know. But what about the automatical detection of subject change? Or is it only common case in news groups?
14:32:15GodEaterI cannot parse the question, sorry.
14:33:55 Join LambdaCalculus37 [0] (i=44a0430d@rockbox/staff/LambdaCalculus37)
14:35:27fmlGodEater: in the news groups it happens rather often that the subject is changed as the discussion goes further. If you want to change the subject you're not supposed to start a new thread. Instead, you just reply with a changed subject. But all the message references (which are automatically generated and maintained) are still correct so that it's possible to see the connection between the old and the new discussio
14:36:54GodEaterRight, that works in email too, I just fail to see it's value (and it's why I don't bother with usenet anymore). If you're talking about something that requires a subject change, how is it related to the old topic at all ?
14:39:09fmlGodEater: it's related just technically. But you can display it however you like. If you use technical references as "thread id" you'll get all the messages in the same thread. If you use the subject line you'll have two separate threads.
14:41:06GodEaterBut the point is that if you've changed the subject, the message should NOT be related to the prior thread. It's not about the same thing. It should be a new thread.
14:41:48 Quit matsl (Read error: 110 (Connection timed out))
14:42:56***Saving seen data "./dancer.seen"
14:45:25fluxgodeater, say thread begins with "there are no programs to do X", a new thread with subject "there is a program to do X: Y" should go to its own separate thread?
14:46:07GodEaterwhy you change the subject there ?
14:46:46fmlGodEater: if you have technical references you can use them or ignore them. It's up to the displaying. The RB site seems to use the technical references.
14:47:14GodEateryes, but why should we change our site to follow your broken logic ?
14:48:14fmlGodEater: I didn't suggest to change the site. I just tried to explain possible logic behind it. From now on, I won't use reply function if I want to change the subject.
14:48:21 Join zitune [0] (n=zitune@bearstech/zitune)
14:48:28zitunehi there
14:48:44GodEaterthere is no logic behind that imo
14:51:24fmlGodEater: the logic is to store as mucg data as possible and to decide at display time how to show it.
14:53:18GodEaterso by that "logic" you should have every email ever belonging to the same thread.
14:54:30 Join evilnick_7 [0] (i=0c140464@gateway/web/ajax/
14:55:42fmlPlease do not exaggerate
14:55:59 Join barrywardell [0] (
14:57:16fmlIt's about how a message is written. By using "reply" I automatically get e.g. "XYZ wrote:" as a header. By changing the subject I clearly state that it's a different subject. So there is logic.
15:03:44 Quit barrywardell ()
15:04:25zitunei've a big problem with a cowon s9, i brick it during a firmware update
15:04:40zituneis there a tcc tools that can help me ?
15:05:09LambdaCalculus37No, the S9 doesn't even work with Rockbox.
15:05:22LambdaCalculus37You're best off asking at somewhere like Anything but iPod.
15:05:46zituneLambdaCalculus37: i just did
15:06:04evilnick_7zitune: Check the New Ports forum thread to see if anyone has managed to use an unbricking tool but I don't know of one offhand
15:06:26dfkti think he means that the s9 could be supported by tcctool
15:06:41dfktdoes it need more than the VID&PID to be recognized?
15:06:44zituneeven by recompiling tcc tools with the correct VID&PID
15:07:04zituneyes dfkt, thx :)
15:07:57GodEaterdfkt: I would imagine it needs to be using a SOC that tcctool works with ;)
15:08:03dfktit does
15:12:27 Quit fml ("CGI:IRC")
15:12:40 Join funman [0] (n=fun@rockbox/developer/funman)
15:17:14funmanalright i got lcd working on c200v2 : commands are 16 bits and i was confused by the fact that c200v1 use 8 bits (transfer one command, then potential values)
15:17:54 Quit TheSphinX^ (Read error: 110 (Connection timed out))
15:18:33GodEaterawesome work
15:20:10funmanthe 'optimized' lcd_update_rect which does vertical filling when better doesn't work though
15:25:04 Quit robin0800 ("Leaving")
15:25:21 Join robin0800 [0] (
15:30:52 Join barrywardell [0] (
15:36:03 Join markun_ [0] (
15:37:17fluxgodeater, the subject would be changed as the key assumption of the thread would have been changed
15:37:19 Quit robin0800 ("Leaving")
15:37:37 Join robin0800 [0] (
15:37:45fluxgodeater, you know, the subject would sort of describe the current subject and state of the thread.
15:39:31 Quit funman (Read error: 60 (Operation timed out))
15:40:16gevaertsflux: the point is, if you're going to talk about something else, why do you want it in the same thread?
15:41:09 Quit robin0800 (Client Quit)
15:41:29 Join robin0800 [0] (
15:41:48 Quit markun (Nick collision from services.)
15:41:58 Nick markun_ is now known as markun (n=markun@rockbox/developer/markun)
15:42:33fluxgevaerts, because it is heavily related to it? the conclusions drawn on its parent thread are assumed still even after the subject has been changed?
15:42:43 Join funman [0] (n=fun@rockbox/developer/funman)
15:42:59fluxI might even quote a person when I change the subject, and how strange would that be if I don't have a References-field for the message I'm quoting?
15:44:47 Quit robin0800 (Client Quit)
15:45:05 Join robin0800 [0] (
15:45:23pondlifeThis chat about subjects and threading belongs in -community.
15:50:27 Quit Shadz ()
15:52:17GodEaterno it doesn't
15:52:20GodEaterit doesn't belong anywhere
15:52:23GodEaterthem's the rules
15:52:54pondlifeI thought this channel was Rockbox and beer, and -community was everything else ;)
15:53:07gevaertsflux: in this case I disagree that they were heavily related
15:54:36fluxgevaerts, oh, quite possibly so, I was arguing merely against the idea that changing subject and continuing thread is fundamentally wrong :)
15:58:09 Quit barrywardell ()
16:10:59 Quit PaulJam_ (Nick collision from services.)
16:11:02 Join PaulJam [0] (
16:17:30 Nick fxb__ is now known as fxb (
16:28:46 Part LinusN
16:30:16 Join biengo [0] (
16:30:49 Quit GodEater ("Permission revoked - dircproxy 1.0.5")
16:35:02 Quit plx140 ()
16:36:31 Quit perrikwp (" ajax IRC Client")
16:37:01 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
16:38:03 Join jgarvey [0] (n=jgarvey@
16:42:59***Saving seen data "./dancer.seen"
16:58:28 Join prona [0] (
16:58:45 Join evilnick_230 [0] (i=0c140464@gateway/web/ajax/
17:01:01pronaquestion: on sansa e280v2, is it necessary to rebuild bootloader when creating new rockbox.sansa? I'm finding that when the image increases in size, the unit freezes before the initial menu
17:01:17 Join faemir [0] (n=faemir@
17:01:52 Quit Zagor ("Don't panic")
17:03:19 Join TheSphinX^ [0] (
17:04:08 Quit faemir (Client Quit)
17:04:14 Quit funman (Remote closed the connection)
17:04:26 Join funman [0] (n=fun@rockbox/developer/funman)
17:04:59 Join GodEater [0] (n=bibble@rockbox/staff/GodEater)
17:05:32 Join toffe82 [0] (n=chatzill@
17:06:10 Join FrankTM_ [0] (
17:07:42 Join faemir [0] (n=faemir@
17:09:39r0b-is there any way to access the virtual keyboard while doom is loaded
17:10:45 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
17:12:21 Join Zarggg [0] (
17:13:02GodEaterr0b-: none that I'm aware of. Why would you want to ?
17:13:42r0b-to enter cheat codes :P
17:15:07pronais it ever necessary to rebuild the e200v2 bootloader?
17:15:19GodEaterprona: only when we add new features to it
17:15:40GodEaterand given the possiblity of bricking the v2s, we don't recommend you do it yourself.
17:15:42pronato the bootloader you mean, not rockbox
17:15:42funmansince you speak of doom, has anyone looked at FS #10271 ?
17:15:52GodEaterprona: correct
17:16:23pronaok i have working bootloader, it must be my rockbox build that is bad. thanks
17:17:36funmanprona: if you have not updated the bootloader since 10 days you should do it if you want to run recent, completely broken, rockbox on the v2
17:18:42 Quit FrankTM (Read error: 110 (Connection timed out))
17:18:53pronano the bootloader was built yesterday. i started with 21305
17:21:05funmancan you see the loaded rockbox logo ? (it shows "Ver. xxx" and not "Boot. Ver. xxx")
17:23:02 Quit robin0800 ("Leaving")
17:24:16pronatwo rockbox screens display, the first i believe is the bootloader, then the rockbox version screen, then you should get the top menu. it sticks on the 2nd screen
17:24:40r0b-there must be some way to get rockdoom to allow keyboard access
17:24:53gevaertsr0b-: yes. Code it
17:25:00r0b-im no good at C
17:25:11funmanprona: ok the bootloader works fine, you shouldn't need to update it in the future
17:26:33pronayea, it was all working up til my last build. I needed to rule out the bootloader
17:28:22 Quit bluebrother ("leaving")
17:33:13 Join robin0800 [0] (
17:35:54pronadoes the rockbox-info.txt have some function, my problem goes away if I copy that as well as the rockbox.sansa
17:37:35 Part pondlife
17:37:40evilnick_230prona: I believe that that is for reference in seeing that the build has the correct options (in terms of screen dimensions, memory size etc) for your player
17:38:59pronawell it may be that a mismatch can cause the rockbox to fail to initialize. not enough evidence, but could be
17:39:58pronai noticed the failure after the rockbox image grew a little bit, it worked fine before that
17:40:46 Quit Ridayah (Read error: 60 (Operation timed out))
17:41:06gevaertsprona: rockbox itself doesn't care about rockbox-info.txt
17:41:48pronawell I'm only guessing. I'm new :)
17:43:38 Join x73 [0] (
17:46:26 Quit x72_l (Nick collision from services.)
17:46:31 Nick x73 is now known as x72_l (
17:49:07 Quit kachna (Remote closed the connection)
17:49:45 Join kachna [0] (
17:50:18pronais the rockbox website down?
17:50:55 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
17:52:15LambdaCalculus37The site is working fine here.
18:01:07Genesisis rockbox libc "rocklibc" ?
18:01:31 Quit andy` (Remote closed the connection)
18:03:03funmanrockbox has no libc, only a subset
18:10:40 Quit prona (Read error: 110 (Connection timed out))
18:12:34 Quit petur ("work->home")
18:22:15 Join JdGordon| [0] (i=836b0065@gateway/web/ajax/
18:24:02 Join funman_ [0] (
18:24:25funman_can someone with a c200v1 test please ? i'm looking at merging the drivers for c200v1 & c200v2
18:25:00 Quit funman (Nick collision from services.)
18:25:03 Nick funman_ is now known as funman (
18:29:02 Quit perrikwp (" ajax IRC Client")
18:30:22 Quit tvelocity ("Αποχώρησε")
18:31:38 Quit BlakeJohnson86 (Read error: 110 (Connection timed out))
18:34:28 Quit PaulJam (Nick collision from services.)
18:34:33 Join PaulJam_ [0] (
18:37:50 Quit kachna (Read error: 110 (Connection timed out))
18:38:18obofunman: c200v1 seems to work fine - tested with a WPS with jpeg AA and mpegplayer
18:38:19 Quit PaulJam_ (Nick collision from services.)
18:38:22 Join PaulJam [0] (
18:40:14 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
18:42:12 Quit PaulJam (Nick collision from services.)
18:42:17 Join PaulJam_ [0] (
18:43:00***Saving seen data "./dancer.seen"
18:43:36 Quit PaulJam_ (Nick collision from services.)
18:43:39 Join PaulJam [0] (
18:45:52 Join bertrik [0] (
18:46:42 Quit PaulJam (Nick collision from services.)
18:46:47 Join PaulJam_ [0] (
18:46:49bertrikhi funman nice to hear about the c200v2 display
18:47:11 Quit perfectdrug ()
18:47:15bertrikdo buttons, audio playback and radio work too?
18:47:17funmanstill some work to do : mainly for buttons and lcd to interact correctly, and for mpegplayer
18:47:45funmanno i can't get buttons working : if i print the status in a loop I can see the buttons have an effect but they don't work in rockbox, so i didn't test further
18:49:12bertrikhm, too bad nobody is bringing an c200v2 to devcon
18:49:51bertrikfunman, oh, buttons are shared with DBOP?
18:50:18funmanat least the left button (which we use in mkamsboot)
18:50:42funmana mechanism similar to fuze & e200v2 is used (set the 4msb of DBOP_DOUT)
18:51:25bertrikI don't quite understand that mechanism, I read something about a red pixel being required but it's a bit vague
18:51:46funman'red pixel' just means the high bits are set
18:52:11funmankugel will know the details :)
18:52:26bertrikso the button switches are each between two DBOP data pins or something like that?
18:53:50funmanno idea
18:54:30funmani also have problems reading the buttons linked to GPIOC, and i couldn't find the code for them in the OF
18:55:22toffe82what is this message while running : whether the C compiler (gcc -U_FORTIFY_SOURCE ) works... no
18:55:23funmanby the way the code for c200v1 & c200v2 use illegal lcd commands (or the controller model is wrong)
18:55:24toffe82configure: error: installation or configuration problem: C compiler cannot create executables.
18:55:37funmantoffe82: check config.log
18:55:47bertrikmeh, I'd really like to see an explanation for it, rather than some "magic red pixel required" :)
18:56:00toffe82funman: where ?
18:56:08funman/tmp/something i believe
18:57:07 Quit PaulJam_ (Read error: 113 (No route to host))
18:57:48 Join petur [0] (
19:00:19bertrikfunman, the left button isn't in the SansaAMSHardwareMappings wiki yet
19:01:28funmani found it today and it's on dbop input, not on a gpioa
19:01:43bertrikC2? (just a guess)
19:02:43funmani didn't try, i first want to get access to known buttons :)
19:04:39funmaneven the power button which is easy to read show no effect in rockbox
19:06:50bertrikI vaguely remember seeing some mechanism where some bit is pulled high for a short time, then some delay and then a read, to read the power button
19:07:16funmanright it's for the clipv1 (not necessary afaik) and clipv2 (necessary)
19:09:57 Join PaulJam [0] (
19:12:03 Quit martian67 (Connection timed out)
19:19:17 Quit nibbler_ (Read error: 110 (Connection timed out))
19:21:57 Quit robin0800 ("Leaving")
19:26:05 Quit TheSphinX^ ("XChat@Linux")
19:29:12 Quit bmbl ("Woah!")
19:30:26 Quit timc (Read error: 110 (Connection timed out))
19:30:52funmanhum now i have no lcd output in the bootloader ><
19:32:31 Join nibbler_ [0] (
19:33:47*LambdaCalculus37 thinks that the SA9200 toffe82 sent him is in serious need of a build/bootloader update
19:33:50 Join BlakeJohnson86 [0] (
19:34:22toffe82LambdaCalculus37: go... update it ...
19:35:10LambdaCalculus37toffe82: Are there any significant updates to the GoGear SA9200 bootloader that I should know about since.... dayum, r19746!?
19:36:12 Join toffe82_ [0] (n=chatzill@
19:37:15toffe82_LambdaCalculus37: I can't say ...
19:39:06 Join Horscht [0] (n=Horscht2@xbmc/user/horscht)
19:41:15LambdaCalculus37Hmmm... according to the GoGearSA9200info wiki page, the touchpad is a Synaptics LightTouch, and the m:robe 100 has a driver for its Synpatics touchpad using the 3-wire specification.
19:41:30LambdaCalculus37*interface specification, even.
19:41:41*LambdaCalculus37 wonders if it would work on the GoGear SA9200
19:43:14 Quit fxb (Remote closed the connection)
19:44:35 Join tvelocity [0] (
19:45:04 Join bmbl [0] (
19:48:18Genesisi'm compiling a crosstool chain but i wonder if it's safe to test it compiling the bootloader for cowon D2
19:48:45Genesisi think it's better to test with rockbox itself
19:49:07 Join Thundercloud [0] (
19:50:23Genesisperharps there is safe method ? tcctool ?
19:50:25toffe82_LambdaCalculus37: the problem is to find the connection
19:53:40 Quit bmbl (Client Quit)
19:53:57 Join bmbl [0] (
19:54:19 Join robin0800 [0] (
19:54:48 Quit toffe82 (Read error: 110 (Connection timed out))
19:54:54 Nick zitune is now known as zitune[afk] (n=zitune@bearstech/zitune)
19:55:12 Quit nibbler_ (Read error: 110 (Connection timed out))
19:55:49LambdaCalculus37toffe82_: Yes, that's true. But lowlight is about the only one who really knows it, AFAIK.
19:55:56saratogaGenesis: afaik sending something with tcctool is completely safe
19:56:06toffe82_LambdaCalculus37: yes
19:56:18Genesissaratoga, : even the bootloader itself ?
19:56:43saratogaisn't the bootloader what most people are running on the D2?
19:57:46LambdaCalculus37Genesis: Even the bootloader. tcctool just sends code that is executed from RAM on the device.
19:57:58Genesisoki thanks
20:00:06 Quit PaulJam (Nick collision from services.)
20:00:10 Join PaulJam_ [0] (
20:05:03bertrikI might have another go at checking the clip memory map and trying to put the stack into iram again tonight
20:05:05 Join _lifeless [0] (n=lifeless@
20:05:14 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
20:05:16 Quit tvelocity (Read error: 60 (Operation timed out))
20:05:39 Join tvelocity [0] (
20:05:40 Quit r0b- (Read error: 110 (Connection timed out))
20:05:52 Join r0b- [0] (
20:09:18 Quit r0b- (Client Quit)
20:09:40 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
20:13:27funmanhum .. button_read_device() reads some button but they have no effect on navigation ..
20:16:03JdGordon|fixed up keymaps?
20:16:50funmanc200v2 uses the c200 keypad
20:16:53bertrikI guess he uses the same as c200
20:24:16 Join KBH [0] (
20:24:42 Join fxb [0] (
20:29:15bertrikfunman, for 2MB ams sansa we do not define USE_IRAM for codecs it seems (plugin.h line 640), is this intended?
20:29:27bertrikconfig.h I mean
20:29:59funmanyes, the codecs are fully in iram already so we don't need it
20:30:53bertrikeh? what do you mean we don't need it?
20:32:17bertrikcodec_start in codec_crt0.c depends on this define for its bhaviour
20:32:31funmanyes, for copy needed data from dram to iram, but everything is in iram already
20:33:29funmanUSE_IRAM is for support of a normal memory region, and a faster (iram) memory region
20:34:10funmanbut Sansa AMS with 2MB of RAM only use a normal memory region (which happens to be the fast iram, but the codecs don't need to know this detail)
20:34:21toffe82_funman : this is the config.log
20:36:15bertrikfunman, ok Im confused
20:36:51funmantoffe82_: install libc6-dev / build-essential
20:37:22toffe82_funman: ok I completely forgot this one :)
20:38:32bertrikIRAM can run at full processor speed and the 2 MB embedded 1-T DRAM runs at max 65 MHz or so, right
20:39:18 Quit funman ("Reconnecting")
20:39:30 Join funman [0] (n=fun@rockbox/developer/funman)
20:39:31bertrikand most of the code/data on 2 MB ams sansas is in the DRAM, with the IRAM containing some specialised routines and the codec buffer, right?
20:39:41 Quit evilnick_230 (" ajax IRC Client")
20:40:25funmanbertrik: i think what we call IRAM is the 1-T DRAM
20:40:50funmanbertrik: you speak the truth
20:42:33 Quit HBK (Read error: 110 (Connection timed out))
20:43:03bertrikah ok, indeed 1-T DRAM is the 320 kB internal RAM, the 2 MB DRAM is not mentioned in the as3525 datasheet
20:43:05***Saving seen data "./dancer.seen"
20:43:38funmanno it's just custom RAM connected to the PL172 multi port memory controller
20:44:58funmanobo: can you test please ? i want to commit it without breaking c200v1 ;)
20:46:30bertrikfunman, oh SRAM then?
20:47:31funmanno it's low-power SDRAM for the Clip/c200v2/m200v4 and high-performance SDRAM for fuze/e200v2 (you can check the value written in MPMC_DYNAMIC_CONFIG_0 in the pl172 datasheet)
20:49:17bertrikok I see the difference in system-as3525.c
20:49:19 Join evilnick_230 [0] (i=0c140464@gateway/web/ajax/
20:51:03 Join midgey|w [0] (i=836b0056@gateway/web/ajax/
20:52:07obofunman: blank LCD only
20:52:21oboplus shouldn't the lcd-c200 file be under the sandisk folder?
20:52:39funmanno i moved it backwards since i made it common to c200v1 & c200v2
20:53:07oboyou mean the c200v2 isn't made by sandisk? ;)
20:53:36funmanyes but it's in arm/as3525/ not arm/sandisk/ (because it's a different SoC)
20:53:51oboah, okay.
20:54:00funmanideally sandisk/* should move into a pp502x/ folder but that's something on MrSomeone's TODO list ;)
20:54:13toffe82_funman: building now ... thanks
20:55:21 Quit PaulJam_ (Nick collision from services.)
20:55:23 Join PaulJam [0] (
20:55:35 Quit flydutch ("/* empty */")
20:56:01 Quit PaulJam (Nick collision from services.)
20:56:05 Join PaulJam_ [0] (
20:56:19funmanobo: can you change lcd_send_command(R_N_LINE_INVERSION, 0xd) to .., 0xdf) ?
20:56:35funmanerr.. 0xdf -> should become 0xd
20:57:36oboline 237?
20:58:00 Quit midgey|w (" ajax IRC Client")
20:59:11obofunman: still blank
21:00:24funmanand if you invert setting of horizontal & vertical dimensions ~line 245 ?
21:00:45funmanhm .. and lcd_delay(100) , not lcd_delay(10) after this setting
21:01:47 Join kugel [0] (n=kugel@rockbox/developer/kugel)
21:01:50oboby invert - you mean moving the two height commands before the width commands?
21:02:01kugelfunman: nice work on the c200v2 so far!
21:02:13kugelfor your gpio button problems
21:02:24funmankugel: hey, i think you can help making it better ;)
21:02:52 Join midgey|w [0] (i=836b0056@gateway/web/ajax/
21:02:53kugelwe had them too on fuze/e200v2, until domonoky added a small delay after clearing bits and found a/the correct sequence of reading them
21:03:20funmankugel: had you find the sequence in the OF ?
21:03:45obofunman: no change
21:03:49kugelimo the common file should be named lcd-c200-c200v2 (for clarity). And putting it into arm/ seems nasty. We should probably have something like firmware/drivers/misc
21:04:05kugelfunman: domonoky found it
21:04:07funmanobo: :( still reading
21:04:30kugelI was getting insane for 2 weeks or so because the buttons worked fine in the bootloader but not in the main binary
21:04:42funmankugel: kugel there is some common drivers in arm/ already
21:04:49kugelthat is also nasty :)
21:05:35bertrikkugel, how does the thing with the red pixel work for reading the buttons?
21:05:54 Join Akiyuki [0] (
21:06:02AkiyukiDoes rockbox work on the 120GB ipod classics?
21:06:06kugelideally (yes, MrSomeone!) we should only have very arm specific stuff in their, properly put things into subdirs per soc (and shared drivers in the firmware/drivers folder)
21:06:20kugelbertrik: see lcd_button_support()
21:06:41JdGordon|Akiyuki: is it listed on the website?
21:06:49Akiyukii can't find the hardware support
21:07:24bertrikkugel, weird ... :)
21:07:31kugelit just writes a single red pixel (fbdata data = 0xf<<12) outside visible area
21:08:08JdGordon|Akiyuki: you cant find the big list of supported devices at the very top of the website?
21:08:10kugelIIRC funman found the OF also doing some of this kind
21:08:25funmanthe c200v2 of writes -1 = 0xffffffff
21:08:30AkiyukiApple: 1st through 5.5th generation iPod, iPod Mini and 1st generation iPod Nano
21:08:30Akiyuki(not the Shuffle, 2nd/3rd/4th gen Nano, Classic or Touch)
21:09:02kugelI just notice we send 2bytes, but only pass 1 as length
21:09:48kugelthat could cause the fifo to be messed up maybe the reason we weird lcd actions sometimes?
21:10:10bertrikmaybe it just needs a specific value on the DBOP lines to "precharge" the buttons
21:10:13kugelah no, it should be fine
21:10:15funmansizeof(fb_data) = 2 (16 bits)
21:10:48kugelfunman: I know :)
21:10:59 Part Akiyuki ("Leaving")
21:11:23kugelI was confused for a moment by the 1 as length, but a pixel is two bytes, so it's ok
21:11:42funmanobo: last change : can you remove lcd_delay() for c200v1 and make it #define lcd_delay(x) udelay(x*1000) ?
21:12:24funmanafter the previous changes it is the last functional difference between the previous patch you had tried
21:12:45kugelbertrik: I don't know really
21:13:56kugelfunman: as long as the higher bits are set, any value works
21:14:11kugelso we could also use -1, but it doesn't make any difference
21:15:42bertrikok, I now have the stack in iram (instead of dram) on my clip, and test_codec still shows the exact same speed :|
21:16:14 Quit PaulJam_ (Read error: 113 (No route to host))
21:16:27 Quit jfc^2 (Read error: 54 (Connection reset by peer))
21:16:52 Join jfc^2 [0] (
21:16:55bertrikI guess most of the speed comes from the MMU because "iram" on ams sansa isn't really that fast
21:18:41funmani remember something like iram being 25% faster without caching, but i may be wrong
21:19:07bertrikehm, yes cache I mean, not MMU
21:20:20 Quit KBH ()
21:22:26kugelbertrik: did you play around with putting codecs into DRAM to get the playback stops away?
21:22:45kugelI'm still thinking the IRAM could be full
21:22:50 Join HBK [0] (
21:22:59 Quit Thundercloud (Remote closed the connection)
21:23:40funmankugel: that can only happen if malloc is used, right?
21:24:12kugelsome codecs do weird stuff, I don't know
21:24:13bertrikkugel, no haven't tried that
21:24:40funmanif they use static memory it can't be full, and the stack is checked by the threads code
21:24:50obofunman: blank again
21:25:10funmanobo: :/ something weird happens here
21:25:18 Quit bmbl ("Bye!")
21:25:36kugelfunman: well, codecs in iram is the only real difference. I can't see any other reason for the clip to fail
21:25:52funmankugel: i think of a bug in buffering
21:25:55kugeldifference between the 8MB models
21:27:18 Join ExoUY [0] (
21:27:27kugelfunman: hm, in pcmbuf.c there's a specific define for <= 2MB (to get a smaller pcmbuff), maybe that's not correctly used throughout?
21:27:35bertrikI can't really reproduce the playback stops quickly and reliably
21:27:51 Part ExoUY
21:27:55kugelbertrik: another thing, i get the blue bars with your lcd speed up patch too
21:28:11kugelI checked the error bits, they seem to be all fine
21:28:35bertrikwe could experiment a bit with the TIMPOL settings maybe
21:29:25bertrikhave you seen any specification anywhere of the timings for the LCD controller?
21:30:06kugelwhich lcd controller? :)
21:30:08 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
21:30:16kugelwe don't know the controller I think
21:30:39bertrika good thing to do would be split up the TIMPOL magic words into individual bits, so we can more easily change timings
21:30:53funmankugel: the watermark is at 1second, same size than the whole pcm buffer
21:31:31funmanthe size is defined in pcmbuf_get_next_required_pcmbuf_size() (not sure if the name is really explicit ..)
21:31:35bluebrotherdomonoky: do we have anything left that would block rbutil 1.2.2?
21:32:06 Quit Genesis (Connection timed out)
21:32:11domonokybluebrother: no, i think we are fine to go.
21:32:29 Join Genesis [0] (
21:33:10kugelbertrik, funman: do we have any idea if FlynDice's core voltage patch has any influence on our boosting scheme. We can do freebosting without delay right now, but I guess that could change if the voltage needs to be changed
21:33:18 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
21:33:20funmanobo: can you try (ok, last one..) ?
21:33:38bertrikhm, I thought you usually have something like three parts: LCD data port (inside the SoC) -> LCD controller (outside the SoC) -> LCD
21:34:13kugelI mean we don't know the model
21:34:31funmanthe diff between the current file and the one patched by (which worked) is at : the only functional difference i see is that i made the PP specific init a static inline function
21:34:46funmankugel: well switching time needs to be measured
21:35:21funmanit's using i2c so perhaps it brings additional delays
21:35:24kugelfunman: use IRAMSTATIC instead. gcc does weird stuff with static functions that are in iram and called by dram code and similar
21:35:41funmanit's not in iram
21:36:15kugelI think the framebuffer is
21:37:08kugelthere's probably a reason why it is not static :)
21:37:44bluebrotherdomonoky: ok, then I'm going to release now.
21:37:51 Join bigtomrodney [0] (n=tom@
21:37:53funmanobo: well just try this:
21:38:02funmans/SANAS/SANSA/ line 196
21:38:09funmanindeed the function was not inlined :P
21:38:27bertrikan i2c write to change the voltage takes about 75 us theoretically
21:42:13obofunman: changed, but no difference when compiling. Where is SANSA_C200 defined? I can't see it...
21:43:13CIA-70New commit by bluebrother (r21319): Bump Rockbox Utility version number.
21:43:53CIA-70New commit by bluebrother (r21320): Tag rbutil 1.2.2 release.
21:44:30funmanobo: hm and c200v1_init -> c200v1_lcd_init
21:45:00JdGordon|kugel: so whats stopping the custom list vp patch being commited as is?
21:45:01bertrikkugel, oh so the fuze display was made working without having any datasheet at all?
21:45:59kugelJdGordon|: a few local changes that are not on the tracker, if at all. I think shortly after release (after/at devcon) is a good opportunity
21:46:25kugelbertrik: yes, purely by disassembly
21:46:31JdGordon|well, the release is branched, so getting it in soon would be great
21:46:33 Join ExoUY [0] (
21:46:41JdGordon|plenty of time to fix bugs which pop up before 3.4
21:46:50kugelbertrik: but I also get the blue bars with 31MHz DBOP, obviously only 50fps then
21:47:30kugelI'll get around it next week (hopefully)
21:47:43bertrikweird that you also get them at 31 MHz, I didn't expect that
21:48:02bigtomrodneyHi guys, apologies if this has been asked recently - I loved running Rockbox on my iRiver iHP140. I've moved over to an iPod Classic now and was just wondering if there has been any interest in porting to this platform?
21:48:17bertrikwhat makes the difference between blue bars and no blue bars at 31 MHz? just the FIFO?
21:48:40kugelapplying your or my patch
21:49:06bluebrotherbigtomrodney: you've read the front page?
21:49:17kugelI can show you the bars at the devcon
21:49:52bigtomrodneyI see it's not currently supported, I was wondering if anyone had showed interest or made an attempt to port
21:49:59bertrikkugel, ok good idea
21:50:10kugelbiengo: interest yes, effort to a port no
21:50:20linuxstbbigtomrodney: Interest - yes. But the port is hard due to Apple's encrypting and signing of the firmware, and no-one has solved that problem.
21:50:33bigtomrodneyah right, I understand
21:50:34bigtomrodneythanks guys
21:50:46biengoyep. although, who knows, i might get into it
21:50:57 Join caffeina [0] (
21:51:05biengo...after med-school
21:51:14bertrikI'll try to see if I can find a display datasheet anywhere that matches the fuze code
21:51:37ExoUYhi everyone, sorry to bother, i would like to know if theres any firmware suitable to run on a player using a Sigmatel STMP3505 SoC, the player is memorex brand. sorry for my bad english, i talk in spanish
21:51:55obofunman: success
21:51:59funmanExoUY: perhaps other firmwares but not rockbox
21:52:12funmanobo: great! :) thanks for helping me fixing my typos ...
21:52:23kugelbertrik: there's a few registers that don't match the e200v2/v1 registers, most are the same though (e.g. I, for fun, tested if setting contrast worked by copying the related write_data calls, and it worked)
21:52:27ExoUYfunman: thank you!
21:52:31obofunman: no worries :)
21:52:48 Quit ExoUY ()
21:53:17 Join mborus [0] (n=5490ac21@gateway/web/cgi-irc/
21:55:29CIA-70New commit by funman (r21321): Sansa c200v2 : lcd & backlight support, using the c200v1 lcd driver ...
21:56:10CIA-70New commit by lenzone10 (r21322): Updated italian translation.
21:56:12funmanlinuxstb: can you check if subversion detects the move of sandisk/sansa-c200/lcd-c200.c ? I used git mv and the move is showed in git log
21:57:16 Quit intrados (Read error: 104 (Connection reset by peer))
21:57:42 Quit biengo (Remote closed the connection)
21:57:59bertrikkugel, interesting, maybe we could copy the #defines too
21:57:59funmanlinuxstb: shows that it was renamed :)
21:58:00kugelfunman: setting GPIO in backlight is weird
21:58:09funmankugel: why?
21:58:16kugelRockRabbit got backlight without it
21:58:46funmanhum perhaps it's the same issue than with Clipv1 : 2 different screens ?
21:58:58obofunman: I see history for that file going back to r14625, the initial c200 commit
21:59:16 Quit LambdaCalculus37 (" ajax IRC Client")
21:59:20funmanbut i don't remember seeing reference to i2c/DCDC15 in the OF
22:01:00domonokyBagder: or anyone else with access to the download server: please move this to it:">
22:01:04 Join intrados [0] (
22:01:51*bertrik tried but failed to reproduce the playback stop problem on his clip
22:03:31funmanbertrik: sometimes it can take hours
22:03:33*bluebrother needs a faster machine for rbutil builds :)
22:03:53bluebrotheror maybe set up cross compiling the w32 binary ... MinGW is sloooow.
22:04:34domonokywe need automated rbutil builds :-)
22:05:43 Quit kugel (Nick collision from services.)
22:05:51 Join kugel [0] (
22:06:00bertrikfunman, I did get a playback problem on an m4a file but it seems different from the stops I get with mp3. The m4a problem is more like looping a tiny part of audio instead of a stop
22:06:19JdGordon|domonoky: ah good... that comp is still giving stupid /home problems so if you can copy it locally incase it doesnt get moved quickly that would be good.... im trying to get my brother to reinstall osx on it to fix it for good
22:06:32funmanbertrik: ah i only use mp3 and occasionally vorbis
22:06:50funmanbertrik: did you have a look at my logf dump?
22:07:13funmanthere is a repeated error message from playback.c but i can't tell if it's a real error
22:07:20kugelfunman: and you didn't get backlight without that gpio at all?
22:07:31funmankugel: no
22:07:36Bagderdomonoky: transfer in progress
22:07:46kugelI remember the e200v2 needed to do buttonlight_init in the early days for backlight
22:07:59domonokyJdGordon| will do. please notice me when you have reinstalled this machine, then i will have to redo the rbutil build enviroment :-)
22:08:03funmankugel: or was it that the backlight needed to be active for the button light to work ?
22:08:37mborushi - does anyone know who works on the timestrech feature?
22:08:42funmanbertrik: on that one, playback never worked : the first song i played after boot stopped at 0:0
22:08:48bertrikfunman, ok saved it, not sure if I can make something out of that, maybe I should just compare it with anothe rplayer
22:09:07 Quit perrikwp (" ajax IRC Client")
22:09:09Mikachumborus: maybe pondlife
22:09:32 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
22:09:38 Quit kugel (Remote closed the connection)
22:09:42JdGordon|domonoky: ok... just giving you a heads up.. of course, its also possible it wont happen untill i go home in sep :p
22:09:44funmanbertrik: i extended MAX_LOGF_ENTRY to 100 and enabled logf in pcm.c & playback.c
22:10:15mborusthanks. mikachu
22:12:40pixelmamborus: generally just ask your "real" question, people who know might answer (and it can be more than one), other people may read logs...
22:15:16mborusmy "question" is more a suggestion to have a dedicated "speed" screen in addition to the pitch-only and pitch/speed-combo screen
22:15:59bertriklcd framerate on my c200 is now 39.3 fps (at 30 MHz), according to LcdFrameRate is was 50.7 fps
22:16:27 Join strife89 [0] (n=a80a90fa@gateway/web/cgi-irc/
22:17:24funmanbertrik: hmm ..
22:17:29Genesis says my TCC7801 is dual core
22:18:00bertrikI could revert to 21319 and compare
22:18:03Genesisdoes the firmware build use the 2 cores or the second have a propriatary software use by the first ?
22:18:06funmancan you make sure it really was 50.7 ? in fact i expected that it could be a little tiny bit faster with the lcd_send_data() change
22:19:01mborusthis way the speed could also be set in 0.1% steps (like the pitch is now) and it would be easier to use in situations where you want to pitch to stay at 100% and just change the speed
22:20:41bertrikfunman, rev 21319 does give me 50.7 fps
22:21:59 Join p3tur [50] (n=petur@rockbox/developer/petur)
22:22:35funmanbertrik: can you revert only the change to lcd_send_data() ?
22:23:00 Join matsl [0] (
22:23:17 Quit strife89 ("Best wishes.")
22:26:42bertrikfunman, ok I'll try
22:27:01bertrikmove the while out of lcd_send_data, right?
22:27:31 Quit funman (Read error: 60 (Operation timed out))
22:30:48 Join funman [0] (n=fun@rockbox/developer/funman)
22:31:31funmanright, and correct the addr increment
22:31:56 Join low_light [0] (i=c730180b@gateway/web/ajax/
22:32:14 Quit midgey|w (" ajax IRC Client")
22:35:04 Quit low_light (Client Quit)
22:35:13bertrikfunman, back at 50.7 fps ...
22:35:41 Part bigtomrodney ("Ex-Chat")
22:35:48funmanbertrik: hmm .. weird
22:36:03funmanthanks for the test, i'll look at what happens
22:36:22bertrikyeah, I would expect your code to be faster too
22:38:35bertrikfunman, this is what I did:
22:41:45 Join n00b81 [0] (n=taylor@unaffiliated/n00b81)
22:41:47funmanbertrik: what if you change the loop inside lcd_send_data to do { } while(−−width); ?
22:43:04 Part n00b81 ("Leaving")
22:43:07***Saving seen data "./dancer.seen"
22:44:24 Join Thundercloud [0] (
22:45:33funmanthe branch is faster and there's one less register used : only lr is saved on the stack, while in the previous version (r21318 and before) didn't use the stack
22:46:54 Quit mborus ("CGI:IRC (Ping timeout)")
22:50:13 Join nibbler_ [0] (
22:51:13 Join PaulJam [0] (
22:55:33 Join Ridayah [0] (
22:56:19 Quit evilnick_230 (" ajax IRC Client")
22:58:02amiconnfunman: Could you check with different cpu clock settings?
22:58:29funmanamiconn: well i can't check, i have no c200v1 and the v2 is not functional yet ;)
22:58:36 Quit efyx (Remote closed the connection)
22:58:45amiconnThis effect (faster lcd transfer code sometimes leads to slower updates) can be explained
22:58:53 Join efyx [0] (
22:58:54amiconnI observed it on ipods as well.
22:59:18funmandid you fidn the cause ?
22:59:39amiconnThe lcd transfer needs to wait for the lcd to become ready. Now it depends on whether the lcd becomes ready just before the check or just after the check
23:00:11amiconnIf it becomes ready just after the check, the loop performs one more cycle. This often happens with faster transfer code
23:00:40amiconnChanging the cpu clock will change this
23:01:00funmanreading from hardware memory can be blocking ? because i expect one loop to be very fast (while(*reg & 1<<bit) ; )
23:01:04amiconn(i.e. you will find some cpu clock frequencies where the faster code is actually faster)
23:02:22amiconnReading from hardware needs whatever number of waitstates is set for this region. Since memory-mapped i/o is uncached, this can take quite long
23:02:45 Quit p3tur (Remote closed the connection)
23:03:24 Join toffe82 [0] (n=chatzill@
23:03:38 Join pat_mulchrone [0] (
23:07:22 Quit perrikwp (" ajax IRC Client")
23:07:29 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
23:09:00 Quit toffe82_ (Read error: 60 (Operation timed out))
23:15:13saratogai'd still like to try running the fuze with only 2MB of RAM, but i dind't get that to work before, could someone walk me through the needed changes?
23:15:17bertrikfunman, sorry was afk, I'll try again with some other CPU speeds
23:15:44bertriksaratoga, maybe just define AMS_LOWMEM?
23:15:55 Quit patmulchrone (Connection timed out)
23:16:43bertrikI think that's part of it, but probably not all
23:16:51funmansaratoga: just set memory size to 2 in tools/configure
23:17:15funmanthe code is common to m200v4/clip/c200v2 so there is no model specific check, it only looks at mem size
23:19:23bertrikfunman, it's slower at 80 MHz too (105.5 fps)
23:20:04funmanbertrik: does changing the loop into a tail loop have any effect ?
23:20:25bertrikI'll try that
23:23:07saratogahuh wma fails to link due to full IRAM
23:23:27 Quit perrikwp (" ajax IRC Client")
23:23:59funmanremove the framebuffer from iram in config-fuze.h
23:25:15 Join fxb__ [0] (
23:27:46bertrikfunman, the do {} while (−−width) is 40.5 fps at 30 MHz (so slightly faster but still slower than before)
23:28:27bertrik108.5 fps at 80 MHz
23:29:59funmani have no other idea so i think we should revert lcd_send_data() to the previous way
23:30:40funmanand perhaps test this way on sansa ams?
23:30:43saratogamy e200v2 won't even play with 2MB of memory
23:31:04bluebrotherBagder: can you move rbutil 1.2.2 from
23:31:09funmansaratoga: what's the bug?
23:31:43saratogaplayback never begins, it just sits there at 0 seconds
23:31:53saratogabuffering shows negative amounts for everything but PCM, which stays at zero
23:32:09funmanhmm negative means no audio buffer?
23:32:50saratogawatermark is also about 2 billion
23:33:20funmana bit high imo
23:33:32funmancan you paste the .map ?
23:33:41saratogamaybe theres not enough memory available due to extra use for the color screen
23:33:44saratogasure one second
23:34:04funmanit should have failed to link then
23:34:07saratogawhich map is needed?
23:34:14 Quit AlexP (Read error: 60 (Operation timed out))
23:34:16saratogaah rockbox,map & perhaps some
23:34:55 Join AlexP [0] (n=alex@rockbox/staff/AlexP)
23:36:05 Quit fxb ()
23:36:06 Nick fxb__ is now known as fxb (
23:36:56funmanaudiobuffert is 402908 bytes
23:37:22funmanthe codec use dram, so something didn't work
23:37:25funmandid you make clean ?
23:38:01funmanhm sorry
23:38:45 Nick fxb is now known as fxb__ (
23:38:46funman1456212 < 2MB so it's dram
23:38:47 Nick fxb__ is now known as fxb (
23:38:55saratogaisn't there a seperate define to put codecs in IRAM?
23:39:06funmanit should be auto if memorysize is <= 2MB
23:39:16saratogahmm you're right
23:39:24funmanin build/apps/codecs/ i have PLUGIN_RAM : ORIGIN = ((0x30000000 + (2 * 0x100000)) + 0x50000 - 0x48000), LENGTH = 0x48000
23:40:19 Quit caffeina (Read error: 60 (Operation timed out))
23:43:10saratogawhy are they linking to 3015xxxxx then?
23:44:06 Quit Rob2222 (Read error: 104 (Connection reset by peer))
23:44:37funman3015xxxxx is in the 2MB of DRAM
23:46:13 Join Rob2222 [0] (
23:48:24 Quit Thundercloud (Remote closed the connection)
23:48:30domonoky~400k might be too small. i remember silent failing of playback if the audiobuffer is too small. (my m200v4 shows 800k free buffer)
23:49:01saratoga#define CODEC_SIZE 0x100000
23:49:04saratogai bet its that
23:49:24funmanah possible
23:51:18 Part domonoky
23:54:04 Join midgey|w [0] (i=836b0055@gateway/web/ajax/
23:54:20saratogaerrr now plugins not linking due to LCD_COLOR

Previous day | Next day