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 2008-05-17

00:03:22 Quit cbr|w (Read error: 104 (Connection reset by peer))
00:03:35 Join CyBergRind|w [0] (n=cbr@
00:05:33 Quit mcuelenaere ()
00:06:48 Join cbr|w [0] (n=cbr@
00:07:49 Quit CyBergRind|w (Connection reset by peer)
00:08:00 Join bertrik_ [0] (
00:08:21 Quit Nico_P (Remote closed the connection)
00:08:42 Quit Thundercloud (Remote closed the connection)
00:10:09 Quit gevaerts ("bedtime")
00:12:59 Join Thundercloud [0] (
00:14:27soap(I know it is an older discussion) but I think you could rework the who paragraph to be more like "There are two ways Rockbox can attempt to use USB connection. The default way is to try to mount your \dap{} as a hard drive. If you don't want Rockbox to mount your \dap{} as a hard drive hold \ActionStdUsbCharge{ while inserting your USB cable.
00:14:37soap*use a
00:16:36soap"You do not need to hold \ActionStdUsbCharge{} while inserting the USB cable if you are attempting to charge from a wall or vehicle charger. Only if attempting to charge with a computer's USB port.
00:16:41 Quit mib_zshxol (" ajax IRC Client")
00:19:37 Quit DerPapst (Read error: 113 (No route to host))
00:20:05 Quit XavierGr ()
00:20:28pixelmathat's mentioning charging again... maybe it's really worth it to devide that into two: one for the Ondio and the other for the rest that actually try to charge via USB
00:21:09 Quit spiorf ("Non so cosa voglio, ma lo riconosco se lo vedo")
00:24:42soapYes, I hadn't looked but assumed a division would need to be made between USB powerable and USB chargeable in the manual definitions if it didn't already exist.
00:26:18 Quit bluebrother ("leaving")
00:26:30 Join DerPapst [0] (
00:28:40 Quit davina (Remote closed the connection)
00:28:43pixelmaI hoped somehow that this is avoidable with a generic description (maybe it doesn't work)
00:29:04BigBambisoap: I like the idea though, with a charge/power split
00:31:01soapA generic description defeats someone using (the 1%er) the search command for "charge"
00:31:32BigBambiYes, it may be better to just split them
00:31:57soapSomething I think is important (but not like I've gotten off my ass to do it myself) is to rework many passages with search phrases in mind.
00:35:15 Quit bertrik (Read error: 110 (Connection timed out))
00:36:15pixelmaif someone searches the Ondio manual for charger (s)he'll find too much :(
00:41:45 Join webguest84 [0] (n=55e00336@gateway/web/cgi-irc/ my ipod supposed to reboot into disk mode when i plug it into usb? o_O
00:42:10Biiarui assume that's not... default behavior.
00:43:57pixelmait is default behaviour since Rockbox doesn't have own USB support yet
00:44:05Biiaruokay then
00:44:26 Quit webguest84 (Client Quit)
00:45:49 Join nedd1 [0] (n=john@nat/sun/x-c31cb234825a1392)
00:45:55 Quit goffa (Remote closed the connection)
00:46:02 Join shinta42 [0] (
00:46:17 Quit Seed ("cu, Andre")
00:49:17 Quit domonoky (Read error: 104 (Connection reset by peer))
00:49:41 Quit roman (Remote closed the connection)
00:52:38 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
00:52:41 Quit nedd1 ("Leaving.")
00:52:49 Join Seed [0] (
00:58:39 Quit Seed ("cu, Andre")
00:59:03 Quit DerDome (Read error: 110 (Connection timed out))
00:59:21 Join Davide-NYC [0] (
00:59:26 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
00:59:37 Join DerDome [0] (
00:59:55 Quit DerDome (Client Quit)
01:01:31 Join Seed [0] (
01:01:51 Join bughunter2 [0] (
01:04:56 Part shinta42
01:14:21Davide-NYCanyone with a gigabeat FX want to test a patch?
01:14:42Davide-NYCI *think* I did this correctly but I wanted to make sure
01:14:52jhMikeSwhat's it do?
01:15:41Davide-NYCIt checks 2 sensors per touch-cross on the gbfx instead of just the outer most.
01:16:07Davide-NYCI think it works fine but I'm not entirely sure
01:16:18Davide-NYCIt *should* make the touch-cross less "finicky"
01:17:15jhMikeSWasn't aware you had an FX (or why you'd care to change this if you don't :)
01:17:35Davide-NYCI have one.
01:17:43Davide-NYCIt's the wife's
01:18:01Davide-NYCI just got a GBS yesterday, haven't tey RBed it.
01:18:09Davide-NYCTHat's next on the agenda for this evening.
01:18:28jhMikeSonce the BL is in place, it's drag and drop from there
01:18:56Davide-NYCjhMikeS: do you have a GBFX?
01:19:21Davide-NYCCan you test this trivial patch and tell me if it is actually doing anything?
01:21:27jhMikeSthe point is to detect the direction is either line registers?
01:21:40jhMikeS*direction if
01:21:50 Part toffe82_
01:22:05 Quit AndyI (Read error: 113 (No route to host))
01:22:44 Quit Mouser_X (Read error: 110 (Connection timed out))
01:24:21 Join AndyI [0] (n=pasha_in@
01:24:24jhMikeSDavide-NYC: I've got something going here I've gotta wait for to finish so right after that.
01:24:50Davide-NYCThanks. BTW: I grabbed the info from the wiki as pointed out to me by markun.
01:30:17*Davide-NYC thinks he can "feel" the difference but does not know how to confirm it.
01:32:59 Quit Zarggg ("Pursued by a bear")
01:35:01 Quit Seed ("cu, Andre")
01:37:59 Join Seed [0] (
01:41:21 Quit gregzx ("ChatZilla [Firefox]")
01:42:26***Saving seen data "./dancer.seen"
01:43:54 Join linuxstb__ [0] (
01:46:14 Join Zarggg [0] (
01:49:15 Quit leox (Read error: 104 (Connection reset by peer))
01:50:46 Quit Seed ("cu, Andre")
01:50:49jhMikeSDavide-NYC: about to test
01:50:58kugel|awayanyone familiar with the smooth_resize stuff?
01:52:28 Nick kugel|away is now known as kugel (
01:52:59 Join Seed [0] (
01:53:30soapDavide-NYC, you could always modify half the buttons to more easily compare "feel"
01:54:01soapmy F is not in my possession right now :( I'd love a "less finicky" pad.
01:54:08jhMikeSDavide-NYC: It seems a little more responsive to the press
01:54:22Davide-NYCjhMikeS: right? That's what I thought.
01:54:58Davide-NYCSince it's very subtle and many people have complained (by many I mean three or four) maybe you should just commit this tiny patch?
01:55:05 Quit ompaul (Client Quit)
01:55:36Davide-NYCmarkun: ping
01:55:42jhMikeScomplaining about have to be too precise?
01:56:26Davide-NYCAlso take into account that many people (mostly women) have longer nails and find the crosspad "finiky"/
01:56:50Davide-NYCMy wife was the main reason for me pursuing this change.
01:56:57jhMikeSnow that's just sexist! :p
01:57:27mooshehe :)
01:57:43*Llorean is apparently the only person who doesn't find the pad finicky. :(
01:58:00 Quit linuxstb_ (Read error: 110 (Connection timed out))
01:58:08*Davide-NYC still things Llorean is a robot.
01:58:09*moos really prefer the beast's pad, by far
01:59:59LloreanBut since that patch seems to still leave a deadzone between center and the directions, I doubt it's going to be problematic.
02:00:30LloreanIf all the sensors were active I'd be worried about the potential for people with "larger" fingers accidentally hitting directions when trying for center, but this still leaves a deadzone if I understand correctly?
02:00:37Davide-NYCI left the innermost ring of sensors dead by design. I'm a genius! ;-)
02:01:05jhMikeShere's a cleaner way:
02:01:07Davide-NYCFalse center positive is absolutely not a problem with this
02:01:45jhMikeSI was checking for that and didn't have a problem
02:02:03LloreanjhMikeS: I was in the middle of typing that very thing up to ask if it'd work the way I expected it to. :)
02:02:18Davide-NYCWhat is the point of every first IF in your example? (I don;t understand it)
02:02:47LloreanDavide-NYC: They're on - lines.
02:03:25*Davide-NYC give back his genius remark
02:06:37 Quit ender` (" The last time someone listened to a Bush, a bunch of people wandered in the desert for 40 years.")
02:07:19Davide-NYCjhMikeS: please use your magical commit powers!
02:07:20 Join nuonguy [0] (
02:08:58 Join leox [0] (
02:09:44 Quit bertrik_ (Read error: 104 (Connection reset by peer))
02:10:44kugelcan someone look at this? (especially the FIXME part)
02:11:20moosDavide-NYC: let's the beast play rockbox, and you will see all the jhMikeS magical powers ;)
02:11:33Davide-NYCam doing
02:11:57kugelIt's a function from pictureflow btw
02:12:40Davide-NYCIs there / has there been any talk of scalable fonts?
02:13:05 Quit Falafel (" HydraIRC -> <- \o/")
02:13:38 Join webguest31 [0] (n=4aaa3bc7@gateway/web/cgi-irc/
02:14:14*kugel got it
02:15:42 Join fdinel [0] (
02:15:42webguest31Hi all, I am having problems registering on the forums. It won't play the whole audio file for the audio verification. Someone help please?
02:15:44LloreanDavide-NYC: This has been talked on many, many times.
02:16:33Lloreanwebguest31: In my tests it played the whole file. Are you certain the file isn't completing?
02:16:42*kugel doesn't need scalabe fonts, anti-aliased would be enough ;)
02:17:07*kugel wonders if any dev has seriously looked at the patch
02:18:05Davide-NYCLlorean: What's the consensus on scalable fonts? (please save me the 15 minutes)
02:18:13LloreanDavide-NYC: They aren't wanted.
02:18:31webguest31Llorean: I've tried both the popup window and the direct link, the popup window only gives me three letters, and when I try it throuh windows media player 11, it gets to four, says an error has occured and stops.
02:18:53Lloreanwebguest31: I've just confirmed it works for me. This means, unfortunately, that it's most likely a problem with the software on your PC>
02:20:04webguest31Llorean: ah. Can you register me?
02:20:16LloreanI'm the forum administrator. If you'd like to send me a private message with your desired username, email address, and password, I can create your account for you.
02:20:28*kugel got pictureflow to actually resize \o/
02:20:40jhMikeSDavide-NYC: i have no editor on hand that like the credits file though :\
02:20:59 Quit moos ("good night")
02:21:00kugelIsn't he allready in it?
02:21:16Davide-NYCjhMikeS: huh?
02:21:40webguest31Private message? Is this an IRC thing?
02:21:59jhMikeSCheck it out
02:22:15Biiaruis the duration bar not moving a known issue when playing spcs? or does that work for everyone else?
02:22:50Biiaruor does that have to do with seeking?
02:23:28Lloreanwebguest31: Yes.
02:24:13jhMikeSsomeone else should fix that then, I don't want to corrupt it
02:24:19webguest31Llorean: Ah. How do I do it? There's a combo box abouv the form, is it in here?
02:25:04jhMikeSI try to open as UTF-8 and it's still whines about it not being 1252
02:25:13Lloreanwebguest31: I really wouldn't know how to talk you through it. You would be better served by using a real IRC client rather than the web-based one.
02:25:46Biiaruaha, it's working now
02:25:52Biiaruit wasn't working for the first few tracks, but now it is
02:26:06Biiarui'd also like to say that i'm loving rockbox so far :)
02:26:10webguest31Llorean: if it helps the box has querey, whois and kick
02:26:40 Join webguest52 [0] (n=459abb5a@gateway/web/cgi-irc/
02:26:55 Join fyre^OS [0] (
02:27:22 Quit webguest52 (Client Quit)
02:27:23Lloreanwebguest31: Well, first you'll need to actually register. You can't send private messages without registering.
02:27:54LloreanAfter which you'll wish to query me.
02:28:33webguest31Llorean: ah. Can I have your email address then?
02:29:47jhMikeSIf anyone would be so kind as to fix up the credits...:)
02:30:33LloreanjhMikeS: What name do I need to add?
02:30:50 Join _caf [0] (
02:30:52jhMikeSDavide Gentile (I'm pretty sure)
02:31:12jhMikeSDavide-NYC: spell check that please :)
02:31:37Davide-NYCaw man, are you giving me credit for that tiny patch? (the spelling is correct)
02:31:44*Davide-NYC blushes
02:32:01Davide-NYCI also made some icons for the recording statusbar back int he day
02:32:24jhMikeSyes, I used those and forgot to add then
02:32:28webguest31Llorean: Please repeat if that was you, it played the new message sound, but nothing new came up. Or is it not in the fmain frame?
02:32:42Davide-NYCWhich set of instructions should I follow for the GBS installation?
02:33:00Davide-NYCI use windows
02:33:13LloreanjhMikeS: Okay, should be done.
02:33:29Lloreanwebguest31: it will show in another tab, if you can read it.
02:33:49_cafhello, quick question: does anyone knows what is te normal utilization of the CORE0 in a sansa when rockbox is idle ????
02:34:40Llorean_caf: I assume you're the person who just posted on the forum?
02:34:56jhMikeSLlorean: thanks
02:35:25LloreanjhMikeS: I hope I didn't break formatting. I just decided to trust the text editor I've always used.
02:35:26_cafops... yes.... im lookin in the source code....
02:35:48_cafand it seems a bit obsessive to me :S
02:36:50*jhMikeS 's very first-ever patch that LinusN committed:
02:37:03LloreanjhMikeS: You might be able to answer _caf's question.
02:37:03webguest31Llorean: replied to your private message
02:37:20Lloreanwebguest31: You can't send private messages until you register. Your reply did not come through.
02:37:45jhMikeS_caf: not much. I can measure sleeping 9999 out of 10000 microseconds per tick
02:38:21jhMikeSit's 9995-9999 more normally depending but it's quite small
02:38:26webguest31Llorean: oh. Can you send me your email address then?
02:38:37Lloreanwebguest31: Why don't you just register?
02:39:39LloreanPick a nick other than webguest31 by typing "/nick Name" where name is the name you wish to be called
02:39:48LloreanThen follow the registration instructions you were given when you connected.
02:40:02 Quit Seed ("cu, Andre")
02:40:03webguest31Llorean: Why register? I probably wouldn't login very much, and all this does is make it so I can send one private message
02:40:34_cafjhmikes: thanks for the reply, if so.... what is the meaning of Idle CORE(0) 0.0 ????? doesnt this mean 100% usage ???
02:41:06Lloreanwebguest31: Because you registering means that I don't have to give out my email address to a complete stranger.
02:42:08LloreanI prefer not to give it out unless it is necessary, and I don't see it as being necessary in this situation.
02:42:14webguest31Llorean: If I register, what difference does it make?
02:42:26 Quit fyrestorm (Read error: 113 (No route to host))
02:42:39Lloreanwebguest31: If you register you can PM me the information necessary for me to create your account on the forums.
02:42:50jhMikeS_caf: you mean in the "View OS Stacks" screen?
02:43:35jhMikeSthat's just the largest % usage of a stack used for idling the core at particular times
02:43:55webguest31Llorean: But I can give you my info here. Or are messages with the intended recipients name at the beggining still logged?
02:44:03 Join Seed [0] (
02:44:27Lloreanwebguest31: You are in a channel. When you type my name in front of it the message is still sent to the channel, it just causes my client to notify me that someone spoke to me in the channel
02:44:31LloreanIt is visible to everyone, and logged.
02:44:58_cafjhMikeS: LOLOLOLOLOL , thanks for your answer, im hacking zxbox a little so that it doesnt lose keys in the games, and that was making me crazy
02:46:55webguest31Llorean: oh. I don't terribly mind my email address being visible. I have 2 email addresses, and would change it to the one I primarily use, which I of course wouldn't put here
02:47:36LloreanWell you'd also have to give me the password you desire to use for the account, although you can immediately change it after you first log in.
02:48:02LloreanOr you can ask me to pick one for you, and I can choose a few random letters, private message it to you, and you can sign in and change it.
02:48:04kugelLlorean: Couldn't you just go into a random, empty channel?
02:48:30webguest31Llorean: Yes. See?
02:49:00Lloreanwebguest31: I have no idea what statement your "yes" responds to.
02:49:41Lloreankugel: It takes just as little time for him to register. He could've done it a dozen times in the amount of time we've talked. I really don't understand what his objection to it is.
02:50:17webguest31Llorean: Sorry. I was agreeing to your last two statements.
02:51:28kugelLlorean: I understand. I just wanted to add another solution (which whould save time for both of you)
02:52:29 Quit Biiaru (Read error: 110 (Connection timed out))
02:53:49webguest31Llorean Kugel: I don't as much prefer the IRC page. I have to use another screen reader to get on here, primarily because of the dynamic content, and I don't install much anymore, I recently had a lot of adware attacks and stuff.
02:59:49 Part Llorean
03:04:17 Quit daven (Read error: 110 (Connection timed out))
03:06:36Davide-NYCAFK for 30min
03:07:46*flyback is really fed up with tragic dreams and being tormented over events 16 yrs ago :/
03:08:50 Join m0f0x [0] (
03:08:57 Quit _caf ()
03:09:46jhMikeSHorscht: around?
03:11:17 Quit linuxstb__ (Read error: 113 (No route to host))
03:11:25 Join linuxstb__ [0] (
03:30:04 Join goffa [0] (n=goffa@
03:30:35 Quit Thundercloud (Read error: 110 (Connection timed out))
03:32:16 Join Thundercloud [0] (
03:38:50Davide-NYCjhMikeS: can you help clarify the installation procedure for me under windows? Should I follow the instructions on the GigabeatSInfo page (windows section) or the GigbeatSIntallation page?
03:41:41webguest31Also, any progress on cracking the zune encryption? I've heard it's similar to the gigabeat s?
03:42:30***Saving seen data "./dancer.seen"
03:43:05 Quit DerPapst (Read error: 113 (No route to host))
03:44:43krazykitwebguest31, any progress would be in the Zune thread in the forums.
03:44:53Davide-NYCArgh, Nevermind. I'm just going to wait until the installation procedure is a bit more "dumbed down".
03:46:19webguest31Which I can't register for... Why, windows media player?!?!?!
03:46:28 Nick fxb__ is now known as fxb (
03:46:33webguest31Davide-NYC: What's the problem?
03:46:41 Quit BlakeJohnson86 (Read error: 110 (Connection timed out))
03:47:41Davide-NYCwebguest31: the windows installation instruction in the wiki involve formatting my HD which I would like to avoid. The other intructions are linux only I think.
03:48:18kugelis it possible that a tagcache search gives wrong results when searching tracks for a specific album?
03:48:29webguest31Davide-Nyc: Oh! You can't just upload the files? Strange...
03:48:37kugelwhich e.g. causes pictureflow to display the tracklist in the wrong direction
03:50:54kugelan example: pf searches for track titlles with this command: "rb->tagcache_search_add_filter(&tcs, tag_album, album[slide_index].seek);"
03:51:17kugelbut tcs.result returns the titles backwards
03:52:02kugelI guess there's only 1 person who knows that :/
03:53:19 Join BlakeJohnson86 [0] (
03:56:27webguest31I'm out all, thanks. I may try copying the audio file to my mobile device and see if it will play it, for whatever reason
03:57:37 Quit webguest31 ("CGI:IRC (EOF)")
04:00:07 Quit kkurbjun ("Leaving.")
04:03:26 Join Veto [0] (n=45466b86@gateway/web/cgi-irc/
04:04:22VetoCan somebody help me ?
04:08:31krazykitonly if you say what you need help with
04:09:17advcomp2019and sending private messages does not help too
04:10:08VetoSorry I don't really know how to use the iirc
04:11:04VetoWell... this i my question : Why when I try to install rockbox on my sansa view the installer can't find my device?
04:12:12advcomp2019rockbox has not been ported to the view
04:13:14advcomp2019ported = written
04:13:51Vetook.... weird.. because somebody says he can on Cnet.. well ok. Is there something else to do if i want to add somes games on my sansa view?
04:15:21krazykitVeto, this is an on-topic channel about #rockbox. that question is off-topic.
04:16:30Vetowell.. sorry..
04:17:22VetoIs there a version that will come out for the View? (hope this one is on-topic)
04:18:28advcomp2019you can look at the new ports in the forums for that info
04:18:39Vetook thanks..
04:20:47 Quit borges ("ChatZilla [Firefox]")
04:22:29VetoPretty hard to understand off what there talking about but what this firmware ( ) will add to my mp3 ?
04:26:06VetoIs the firemware i needed to install the rockbox?
04:28:43 Part pixelma
04:29:52soapRockbox for the Sansa View does not exist, Veto
04:30:24advcomp2019Veto, they are basically getting the info right now in the forum
04:30:47VetoSo they will have a rockbox soon?
04:30:54Vetofor the view?
04:30:59soapMaybe, maybe not
04:31:07soapyou can't predict these things.
04:31:42soapIt will happen (IF it happens) when someone with the skills, the knowledge (if said knowledge is available) and the time does it.
04:31:48soapthat someone could be you.
04:33:01VetoI'm so begging and for me reading english is pretty hard .. that's not my language actuaally it's pretty hard too talk with you xD
04:51:59 Join miepchen^schlaf_ [0] (
04:54:12 Quit Thundercloud (Remote closed the connection)
04:57:58 Join Mouser_X [0] (
05:05:27VetoIs tthe Rockbox could broke my MP3?
05:06:25 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
05:06:30 Quit Davide-NYC ("ChatZilla [Firefox]")
05:15:38 Quit Veto ("CGI:IRC")
05:19:04*kugel finished btw
05:20:13 Join z35 [0] (
05:21:05 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
05:21:08 Join Llorean [0] (
05:32:48 Quit Mouser_X (Read error: 104 (Connection reset by peer))
05:35:53 Quit linuxstb__ (Read error: 110 (Connection timed out))
05:40:55 Quit nuonguy ("This computer has gone to sleep")
05:42:32***Saving seen data "./dancer.seen"
05:43:52*flyback doesn't think he's ever going to find a girl he can open up to again :/
05:48:35 Quit kugel ("ChatZilla [Firefox 3.0b5/2008050509]")
06:02:15 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
06:14:49 Nick fxb is now known as fxb__ (
06:18:19 Join mackes [0] (
06:18:41 Join tdtooke [0] (
06:19:44tdtookelinuxstb_, you there?
06:20:42tdtookewell, if you are I apparently signed off just as you were trying to tell me something before and what you were saying is cut off
06:38:08 Join FOAD_ [0] (
06:44:58 Quit jhulst ("Konversation terminated!")
06:46:44 Join goffa_ [0] (n=goffa@
06:52:26 Quit FOAD (Read error: 110 (Connection timed out))
06:52:26 Nick FOAD_ is now known as FOAD (
06:56:49 Quit m0f0x ()
06:59:58 Quit goffa (Read error: 110 (Connection timed out))
07:11:26 Join linuxstb__ [0] (
07:14:04 Join nuonguy [0] (
07:17:03 Quit tdtooke (Remote closed the connection)
07:25:26 Quit linuxstb_ (Read error: 110 (Connection timed out))
07:27:13 Quit flyback (Remote closed the connection)
07:35:22 Quit Hidden (Remote closed the connection)
07:41:55 Quit BHSPitLappy (Remote closed the connection)
07:42:36***Saving seen data "./dancer.seen"
08:04:59 Quit mackes ("Mackes is out")
08:09:29 Join Rob2223 [0] (
08:21:28 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
08:23:36 Quit BlakeJohnson86 (Read error: 110 (Connection timed out))
08:24:17 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
08:27:45 Quit Rob2222 (Read error: 110 (Connection timed out))
08:30:48HorschtjhMikeS, you called?
08:33:52 Join n1s [0] (n=nils@rockbox/developer/n1s)
08:50:27 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
08:52:47 Join Buschel [0] (
08:52:56 Join spiorf [0] (
08:53:02Buschelmoin moin
08:54:23BuschelHorscht: as you've got an iPod Video, too: does your iPod mute when reducing the volume to -58dB?
08:54:38Llorean It didn't used to.
08:54:44Horschthm.... just a moment
08:54:52LloreanI think that's always been broken.
08:54:55pixelmaI adapted sokoban to the c200's screen and needed to invent a new colour scheme because that greenish one didn't work at all at 5x5 tilesize. Then I decided to make new graphics for all (including some more sizes so every screen gets the biggest graphics possible). Would anyone object to change them for all - comparison picture can be found at
08:55:24HorschtBuschel, any specific rev?
08:55:35BuschelLlorean: the behaviour i alwas had was -57dB = audible, -58dB = mute.
08:55:41Horschtor will a week old one do?
08:55:48BuschelHorscht: no just some current one (the last days)
08:56:14LloreanBuschel: I seem to recall that the lowest setting was still audible, but maybe it got fixed when I wasn't looking
08:56:28Lloreanpixelma: I assume the new graphics are the ones on the right? (Haven't played it in ages). If so, I quite like them.
08:56:42HorschtBuschel, r17404 does mute at -58
08:57:09Horschtit even shows a mute logo in the main menu
08:57:46pixelmathe m:robe100's one isn't bigger but there's a slight change in appearence (the box on the spot) and monochromes screens also use external bitmaps for it, currently the graphics are drawn with lines, rectangles etc.
08:57:53Buscheland 17507? which i used (with some lokal changes, which is the reason for asking)
08:58:18pixelmaLlorean: yes, right's the new on (the c200's doesn't exist in SVN yet)
08:59:17Lloreanpixelma: That factored into my assumption. :)
08:59:31pixelmaI wonder how the current style with the green and red works for colour bllind
09:00:31HorschtBuschel, I don't have that one at the moment, checking out
09:00:47pixelmaLlorean: and thanks for your opinion :)
09:00:55Lloreanpixelma: I seem to recall someone who's around regularly being colourblind, but I can't remember who.
09:01:18pixelma(wake up call) ;)
09:02:51Buschelamiconn: could spend a small amount of time of fs#9000?
09:03:29amiconnBuschel: I just took a quick look. The patch looks like it won't compile on targets without large iram
09:04:16amiconnIt's missing #define ICODE_ATTR_MPC_LARGE_IRAM in the #else part
09:04:26 Quit linuxstb__ (Read error: 110 (Connection timed out))
09:04:40Buscheloops, was sure i didn't miss it...
09:04:46*amiconn wonders whether Buschel tests compilation at all
09:04:54Buschelnevertheless you could test it for your x5?
09:05:31*Buschel did this late last night after a long working day...
09:05:43pixelmaLlorean: and now I realise that this also needs changes in the manual... in the c200's of course, maybe new screenshots for other manuals as well, with cabbiev2 style?
09:07:28LloreanToo bad there's not a "manual screenshots" build target that builds a special case of the sim, 'eh?: )
09:08:19HorschtBuschel, plain 17507 also mutes on -58
09:08:38Horschtbut it seems like it does on purpose
09:08:47BuschelHorscht: interesting, gotta research
09:09:17Horschtbecause going to the main menu with -58db, you can see a small mute-symbol in the status bar
09:09:32Horschtwhere the volume level used to be
09:09:36BuschelHorscth: but can you "hear" it muting?
09:09:51pixelmaLlorean: I have almost all sims here (and prepared) that I'd need for that...
09:09:52Horschtas in?
09:09:55 Join Jay [0] (
09:09:59Horscht-57 has sound, -58 not
09:10:13amiconnBuschel: The patch slows down decoding on coldfire
09:10:32 Quit Ragnor (Nick collision from services.)
09:10:42 Join Ragnor [0] (
09:10:47amiconnSVN: 617% realtime, patched: 590% realtime
09:11:43Buschelamiconn: is this caused by the use of iram or the re-ordering within the synthesis? you could check via not enabling the ICODE_ATTR...
09:12:07Horschti can hear it mute from -57 to -58 if that's what you're asking, Buschel
09:12:14pixelmaLlorean: but your kind of suggestion sounds nice :)
09:13:32 Quit goffa_ (Remote closed the connection)
09:13:39 Join goffa [0] (n=goffa@
09:14:02BuschelHorscht: that was exactly what I meant :o) so, I really gotta check what's going on with my target
09:14:35Horschtyours doesn't?
09:15:14Buschelno, but i've got local changes which i didn't thought would have direct effect on this behaviour −− but they have
09:15:28amiconnBuschel: Patch without ICODE: down to 577% realtime
09:17:16Buschelamiconn: so, i better make a patch without the re-ordering (which doesn't have any performance effect on PP5022) and with iram-usage.
09:17:30Buschelamiconn: thanks for your time!
09:18:37 Join gregzx [0] (
09:21:20amiconnThat would need testing as well
09:21:48amiconnOn coldfire, optimisations involving iram for code are rather tricky.
09:21:54Buschelamiconn: of course, that will be the time when i kindly ask for your support again ;o)
09:22:23amiconnFirst, using iram for code has much less impact than using it for data, because coldfire has an instruction cache but no data cache
09:23:25amiconnThen, moving functions to iram, or even shuffling around functions, can change performance (positive or negative), due to cache aliasing of the remaining functions in sdram
09:23:38amiconnThe coldfire instruction cache is direct mapped.
09:24:43Buschelthe internal removal was an experiment which lead to smaller bin size −− as i wasn't sure whether the code fitted into the coldfires iram i thought it is a good idea to use the "smaller" code
09:24:46amiconnThe PP cache is set associative so aliasing effects are much less likely
09:25:30n1samiconn: could we do that to the coldfire cache too?
09:25:51amiconnDo what?
09:26:14n1sset it to an associative mode instead of direct mapped?
09:26:36amiconnYou can't set the cache mode. "set associative" is a property
09:27:01amiconn(as opposed to the two extremes "fully associative" and "direct mapped")
09:27:58n1s(there's an issue with vorbis performance jumping +- 20% with unrelated changes in the core thought to be related to the cache aliasing)
09:28:10amiconnVery possible
09:28:50n1samiconn: any ideas on how to find such an aliased function and how to make it not be aliased (using some align attribute perhaps)?
09:30:29amiconnWe can't avoid such aliasing across everything.
09:31:37amiconnWithin a codec, it's possible to carefully shuffle functions around, but we can never guarantee that calls from codecs to the core or vice versa won't cause cache aliasing
09:32:15amiconnBuschel: I don't think that you need to save binsize. Coldfire code is usually more compact than arm code
09:32:54amiconnBesides, we could switch to the 32/80 distribution on MCF5250 as well (but I don't see the need for that just now)
09:33:43amiconnErm, 48/80 of course
09:37:30amiconnRight now there's 5.5KB of spare IRAM in mpc.codec on MCF5250
09:38:40Buschelamiconn: if the data in iram is so much faster we could test another thing in mpc -> 1st) 48/80 distribution, 2nd) reduce MPC_DECODER_MEMSIZE in decoder.h to 4096, 3rd) using IBSS_ATTR for the array Speicher[] in mpc_decoder.c
09:39:37Buschelamiconn: this is the buffer from which the demuxing is done (huffman-decoding etc.)
09:42:32 Join parafin|away [0] (
09:42:37***Saving seen data "./dancer.seen"
09:44:09 Quit parafin (Read error: 113 (No route to host))
09:49:26*amiconn notices that musepack is already more efficient on coldfire than on PP
09:50:20Buschelyes, since the latest changes
09:51:01amiconn...even taking cpu clock into account (single core PP) on MCF5250 vs. PP5022
09:51:18*amiconn needs to compare with MCF5249 and PP5020
09:51:53amiconnI get 358% realtime on mini G2 for the same file that I used on X5
09:54:59Buschelwhich is ~22,3MHz on PP5020 and ~20,1MHz on MCF5250. With my testfile i get ~22.6MHz on PP5022 −− but this of course scales with the file tiself
09:56:51amiconnmini G2 is PP5022
09:57:12*amiconn should probably also test PP5002
10:04:07 Nick JdGordon|afk is now known as JdGordon (i=jonno@rockbox/developer/JdGordon)
10:05:47amiconnBuschel: ipod G2 (PP5002): 241% realtime, H10 (PP5020): 301% realtime, H180 (MCF5249): 484% realtime
10:07:26 Nick parafin|away is now known as parafin (
10:09:03 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
10:12:39 Quit Jay ("Leaving")
10:13:47 Join Llorea1 [0] (
10:14:31 Quit Llorean (Nick collision from services.)
10:14:34 Nick Llorea1 is now known as Llorean (
10:15:41Buschelamiconn: the PP-family scales quite a lot (PP5022: 355% realtime with lates changes, ~330% before)... is the mcf5249 the same as mcf5250 except the iram size?
10:16:25amiconnRegarding performance it is
10:17:05amiconnOther than that there are several differences, e.g. MCF5249 core voltage is 1.2V, while MCF5250 uses 1.2V
10:17:43amiconnPP5020 is also the same as PP5022 performance wise except the iram size
10:17:49LloreanYou mean 1.8, right?
10:17:53LloreanFor the first?
10:18:01amiconnerr, off course
10:18:42amiconnPP5002 is slower when using sdram because of its broken cache. If you test with iram code, it's also identical to PP502x
10:18:53amiconnIram code + data of course
10:19:25amiconn^^ *of* course
10:24:46 Join Thundercloud [0] (
10:25:35Buschelamiconn: i did some quick
10:26:32Buschelamiconn: i did some quick-n-dirty "pimping" of the fs#9000-patch. but i cannot doe test-compilations on this pc... could you give a quick check on your x5?
10:26:33 Join nplus [0] (n=nplus@141.25.Globcom.Net)
10:26:59*Buschel slaps the touchpad on his notebook
10:27:35Buschelalways hitting it when i do not want to
10:27:39 Part Llorean
10:27:49 Join Llorean [0] (n=DarkkOne@rockbox/administrator/Llorean)
10:30:44 Join bertrik [0] (
10:33:58markunLlorean: which style with green and red do I have to look at?
10:35:05markundo I -> should I
10:36:08Lloreanmarkun: Pixelma linked an image with several screenshots of Sokoban and was concerned about the original green background contrasting with the red.
10:37:33pixelmayou could also check on your gigabeat, the green/red is the current SVN version
10:37:39 Join davina [0] (
10:37:43markunthe ones with white background are much clearer, but I can also see the ones with red and green
10:38:10pixelmaah interesting, thanks
10:38:25Lloreanmarkun: What about the bottom two in the leftmost column?
10:39:05markunthey look the same as all the other color version. Is there something different?
10:39:45LloreanNothing particular, I suppose.
10:39:46pixelmaon the gigabeat it's currently also smaller than possible because no-one bothered to invent some in-between sizes
10:40:29markunsome sokoban games use different size tiles depending on the game, that would also be cool
10:40:45markundepending on the size of the level I mean
10:40:59JdGordondrawn on the fly?
10:41:10LloreanVector Graphics!
10:41:33markunJdGordon: no, just store a few sizes
10:41:51pixelmayeah, I know. But for now I just wanted to get them as big as can be with fixed sizes (and add a c200 version)
10:42:34markunpixelma: nice improvement on the grayscale ones
10:42:53markunyou can't save a level currently, can you?
10:43:00pixelmathat was harder than it looks like
10:43:05pixelmayou can
10:43:30bluebrotherpixelma: slightly different, but I noticed the default backdrop being dithered on 4bpp targets several times. Wouldn't it be better to have that use native greys?
10:44:09amiconnbluebrother: The backdrop is dithered using all native greylevels
10:44:23markunpixelma: perhaps we could rotate the levels on the portraid targets, then we can use the same tile size..
10:44:27amiconnAnd I don't know of any 4bpp targets...
10:44:32markunon the gigabeat as the ipod for example
10:44:34pixelmabluebrother: the default backdrop in the menus? It uses the native grey, but I dithered them before so I could make it lighter
10:45:05 Join Mathiasdm [0] (
10:45:23bluebrotherah, ok.
10:46:09pixelmamarkun: I don't think it's worth it - e200 already can use the same tile size as H300 (with the stats box at the bottom) and gigabeat gets quite close now (12x12 vs. 14x14 tiles)
10:46:13*JdGordon thinks the new disktidy is done!
10:46:38markunI'm emailing a bit with the iriver t10 hacker. He's also interested in working together to port rockbox to the samsung based irivers.
10:46:48*amiconn wants persistent options in reversi
10:46:57amiconn...and a better AI
10:47:19JdGordondoes disktidy have a manual entry? or only the wiki?
10:47:50pixelmayes, I see a disktidy.tex
10:48:19pixelmavery short but it's there
10:48:57JdGordonany chance you can fix it up for the new version please?
10:49:01 Join dan_a [0] (n=dan-mirc@
10:49:50JdGordonpatch is at 8637
10:50:36pixelmaI need to prepare a manual update for my sokoban changes too
10:51:04markunLlorean: did you really not understand what the benefit for him would be?
10:52:37markun1: not to damage your ears, even though they would get used to a higher volume. 2: no need to buy a preamp. 3: no need to replaygain all your music.
10:52:40pixelmamarkun: I don't understand it either, especially how that should work?
10:53:09Lloreanmarkun: How would a plugin work though?
10:53:20LloreanYou have to manually launch a plugin. You can manually adjust volume.
10:53:37LloreanAnd setting an equalizer pre-cut (no replaygain necessary) *is* a volume limit.
10:53:41bluebrothermarkun: I don't understand how this should even work as a plugin. Plus, if you use ReplayGain you can quite decrease the risk of damaging your ears
10:53:50LloreanNo need to use replaygain. Pre-cut exists.
10:53:54markunpixelma: how I imagine it is that you choose a volume (on a test signal maybe) which with your current earphones shouldn't damage your hearing too much, the it keeps analyzing the digital audio and when it goes over a threshold the volume gets lowered.
10:54:03markunthat's how I imagine it
10:54:16markunLlorean: yes, could be implemented with the pre-cut of course
10:54:26*amiconn doesn't understand this either
10:54:54amiconnThere's no fixed limit that will suddenly damage your ears when exceeded
10:55:02markunmaybe I just have a strange mind which is trying to make sense from incomprehensible ramblings ;)
10:55:06LloreanSo yes, I really don't understand the purpose of it as a plugin as compared to existing options.
10:55:11amiconnAny high volume, applied over some time period, can degrade hearing
10:55:33LloreanEven volumes most people don't consider "high" can damage your hearing if you listen for very extended periods.
10:56:05amiconnSo it's generally recommended to listen at low to moderate volume, and then you need no limiter
10:56:09markunyes, but I think the assumption he (and me too) made was that higher volumes damage more
10:56:17pixelmamarkun: I usually do so too but this was beyond me
10:56:21LloreanHigher volumes do damage more.
10:56:50LloreanBut if he's just talking about "limiting like the iPod", you just set precut to say, -12, and you have a set limit.
10:57:03LloreanThat's all the iPod is, a "maximum" that's lower than the hardware cap.
10:57:39markunamiconn: but if you have a low volume track and then a high volume track you would need to listen to the first one at a very low level not to get damage during the second one
10:57:41bluebrotheralso, if the limit is adjustable you'll always run into a situation (like a soft song) that's not loud enough, will increase the limit and forget to lower it afterwards
10:57:58markunbluebrother: or have it dynamic (which I also hate)
10:58:07amiconnmarkun: Huh? How much difference (in dB) are we talking?
10:58:24markunamiconn: max 10 dB I think, but still
10:58:36bluebrotherdo we want mature users that can look after themselves or do we want to take every responsibility from them?
10:58:40markunold kraftwerk album and then system of a down
10:59:05markunamiconn: let me check
10:59:07bluebrotherthere could still a bug causing a loud volume
11:00:17bluebrotherso I think it's much better to keep the user in responsibility
11:00:19markunamiconn: the replaygain values of the albums I was thinking of was actually 16dB appart!
11:00:49markunamiconn: I know you have no use for replaygain, but maybe you can see now that I do ;)
11:00:54 Quit homielowe ()
11:01:18amiconnThat's rather much... I can't tell exactly as I don't use replaygain at all, but I think my albums (spanning 28 years) differ by only ~6dB
11:02:05markunamiconn: autobahn: REPLAYGAIN_ALBUM_GAIN=+4.59 dB toxicity: REPLAYGAIN_ALBUM_GAIN=-10.39 dB
11:02:09markunah, just 15 dB
11:02:59 Nick gregzx is now known as gregzxafk (
11:03:00markunbut autobahn would clip at such a gain, so it's actually +3.90 dB
11:03:32pixelmayou have to listen to that song on your way to DevCon ;)
11:03:58markunpixelma: yes, if I drive with petur and gevaerts that will be a lot of fun ;)
11:04:21bluebrotherand a lot of autobahn :)
11:07:22pixelmaand in the end tell us how many km went by during this song
11:08:14JdGordonamiconn: do you know if the ribbon cable can be easily pulled off the mini2g's main board?
11:08:44 Join ender` [0] (
11:09:00amiconnWhich one? The one connecting the microdrive?
11:10:25JdGordonbah, missed a word.. yes
11:10:59amiconnI tried to pull it, but I didn't succeed
11:11:15amiconnYou don't need to pull it in order to replace the drive
11:11:18 Join DerDome [0] (
11:11:50JdGordoni know.. just trying to get a bit more room
11:11:59JdGordonim farily confident this will fit :p
11:12:10JdGordonI'd like to be 100% certain though :p
11:15:52*amiconn wonders what JdGordon is trying to do
11:17:06JdGordonsurprise :p
11:17:47bluebrotherputting a 2.5" disk into the mini?
11:18:26*JdGordon was pretty sure everyone was sick of hearing about this already...
11:18:35JdGordonputting a bluetooth transmitter in it
11:18:57*amiconn doubts this will work at all
11:19:05amiconnThe mini's case is all-metal
11:19:17JdGordonyeah, i checked..
11:19:21JdGordonits still works
11:19:39JdGordonfortunatly.. that didnt occur to me untill after i forked out for it :p
11:20:34n1show much did it cost?
11:20:55JdGordontoo much
11:21:32*bluebrother wonders about the use of that
11:22:18JdGordonof what?
11:23:10bluebrotherof bluetooth in a dap
11:23:38JdGordonwell, I always use BT headphones anyway, so one less battery to keep chardeg, and one less thing in my pocket
11:23:55JdGordonnot being wired to the bloody thing is wonderful
11:24:15*n1s would love to go wireless too :)
11:24:17Horschtarent those expensive, though?
11:24:45Horschthow much?
11:24:56Horschtfor a decent set of BT headsets
11:24:59JdGordonphones were $100, xmitter was 50
11:25:15JdGordon6 months ago though..
11:25:15markunaustralian dollar?
11:25:21JdGordonwhen i wasnt a poor uni bum :D
11:25:43Horschtbut is the quality any good?
11:26:07JdGordonmy ears are tin anyway so i dunno.. supposedly they are the BT version of decent sonys
11:27:53Horschtthinking about it... i prefer earplugs anyways
11:30:12markunHorscht: and they have a cable on them anyway!
11:31:24JdGordononly around the neck though
11:31:46Horschtyes, but that's just as annoying, i'd say
11:32:00Horschtno point in using a wireless technology
11:32:26n1shmm, then you wold have another analog<->digital conversion...
11:32:55n1sunless JdGordon is doing something really fancy?
11:33:02JdGordonno, im not that smart
11:33:16JdGordonand yes, if you care about your music quality then this may not be for you
11:33:22*JdGordon cant tell the difference though
11:36:00 Join davina^ [0] (
11:40:56 Join BlakeJohnson86 [0] (
11:41:03 Quit Thundercloud (Remote closed the connection)
11:42:40***Saving seen data "./dancer.seen"
11:48:44 Quit nuonguy ("This computer has gone to sleep")
11:50:32 Quit davina (Read error: 110 (Connection timed out))
11:51:46Buschelamiconn: you'd say it's better not to touch the ICODE-stuff for coldfire but only for the PP5022/24? just to make sure there is no loss of performance when doing further changes on the decoder?
11:52:21amiconnI just said that iram is much more important for data than for code on coldfire
11:53:01Buschelok, then it would fine to make a short test of
11:53:44amiconnCache aliasing can hit everytime you change code layout in sdram, be it due to moving functions into/ out of iram, shuffling functions around, or changing functions in size
11:54:06 Join linuxstb__ [0] (
11:55:01 Quit linuxstb_ (Read error: 110 (Connection timed out))
11:55:30JdGordonany objections to commiting the new disktidy?
11:55:33 Join Lynx_ [0] (
11:57:41 Join einhirn_ [0] (
11:59:45markunJdGordon: what did you change? no longer made it a viewer?
12:01:26amiconnBuschel: 565% realtime...
12:02:15Buschelamiconn: slower. so, i will do the change for PP5022/24 only
12:02:37bertrikJdGordon: can you give a link for the new version?
12:02:57JdGordonFS #8637
12:03:07JdGordonthe bug in the last comment is fixed
12:05:44 Join Vorador [0] (n=tomas@
12:07:02 Join homielowe [0] (
12:08:24*Buschel made a simple profiling on mpc-decoder: 25% demux/requant, 25% dct32, 50% dewindowing
12:08:34Buschelwill do the same for mp3 now
12:09:00*amiconn will encode his test album to mp2 now
12:09:00Buschelthe bad thing is that dewindowing cannot be further optimized −− all asm'ed
12:09:13amiconnAlso for coldfire?
12:10:00Buschelbut the dct32 isn't yet. i'll check whether the mad dct32 can be adapted to be used in mpc
12:10:35 Quit einhirn_ (Read error: 104 (Connection reset by peer))
12:14:59amiconnEncoding will take a while (about twice the playtime) as I'm using Pegase on Amiga
12:15:57Buschelmpc dewindowing does 512 mul's and 480 add's, mpc dct32 does 80 mul's >100 add's, >100 sub's and several shifts per frame
12:17:02Buschellibmad's dct32 claims to use less add's and sub's. need to compare though.
12:17:56*amiconn thought mpc wouldn't use dct
12:18:04bertrikJdGordon: I don't use disktidy a lot, but I hated how it had an if () for every file type. It seems you fixed that so I like it already. Does the total number of lines increase or decrease with the patch?
12:18:11Buschelcalc_new_V() is the same as dct32
12:18:32amiconnI thought mpc is a subband codec, like mp1 and mp2
12:18:37pixelmahmm... seems it's not possible to play the solution back in sokoban, or am I missing something? I know there was a big overhaul with saving solutions/progress/unlimited undo etc., maybe it was removed during this but the manual still lists buttons for solution playback...
12:18:44Buschelthe polyphase filterbank is done via dct32 followed by dewindowing
12:18:47*amiconn doesn't know the details though
12:18:55Buschelit is a pure subband coding
12:21:13Buschelamiconn: the filterbank is nearly identical to mp1/mp2, the implementation of the synthesis filter was tweaked to take further advantage out of symmetries in the synthesis itself
12:23:02Buscheldct32 inputs 32 samples and outputs 32 samples, mpc's dct32 inputs 32 samples and outputs 64 (symmetric) samples. that is what i need to adapt for a performance comparison
12:23:32 Quit BlakeJohnson86 (Read error: 110 (Connection timed out))
12:26:14 Join moos [0] (
12:26:27Buschelmaybe libmad's dct32 has advantage over mpc's implementation regarding internal scaling as libmad does only multiply with factor <1 whereas mpc multiplies with up to 10
12:30:13JdGordonbertrik: its a plugin so big deal re LOC
12:32:08*Buschel will switch to his dev-PC now
12:32:11 Quit Buschel ()
12:38:37bertrikJdGordon: ok
12:40:01 Join Buschel [0] (
12:49:01 Nick gregzxafk is now known as gregzx (
12:50:53 Quit bertrik ("Ex-Chat")
12:54:45 Join desowin [0] (n=desowin@atheme/member/desowin)
12:59:11 Join |obo| [0] (
13:00:15 Nick fxb__ is now known as fxb (
13:01:38 Nick |obo| is now known as obo (
13:04:26 Quit Lynx_ (Read error: 110 (Connection timed out))
13:05:26 Quit krazykit (Remote closed the connection)
13:05:30 Join krazykit [0] (
13:17:43 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
13:31:57Buschelinteresting: mp3 needs ~6MHz for dct32 (asm) + 13MHz for dewindowing (mostly asm)
13:32:37Buschelmpc needs ~5 MHz for dct32 (little asm) + 12MHz for dewindowing (mostly asm)
13:32:48Buschelboth measurements on a PP5022
13:33:37Buschelthat is comparing the mp1/2 synthesis part to mpc synthesis
13:34:34*amiconn wonders how Buschel is measuring this
13:34:44bluebrotherhmm. Does charging on the m:robe need booting into the OF for full speed like on the ipods?
13:35:21Buschelamiconn: via evil uncommenting parts (i know that's not the best way, but it works fine)
13:36:54 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
13:40:31JdGordonmcuelenaere: hey, you were looking for me yesterday?
13:40:31amiconnBuschel: You can only test on PP5022?
13:40:54*amiconn thinks that codec optimisers need access to several architectures
13:41:52*Buschel just recognizes he compared 64bit-multiplies (mp3) with 32bit-multiplies (mpc)
13:41:56Buschelwill retest mpc
13:42:20*amiconn wonders whether he will be allowed access to a beast @devcon
13:42:43***Saving seen data "./dancer.seen"
13:43:44mcuelenaereJdGordon: yes, but uhm I forgot why .. :)
13:43:47mcuelenaerelet me check the logs..
13:43:54JdGordonamiconn: I thought you were thouroly uninterested in the beast?
13:44:20amiconnI don't want one myself, but I'd like to tackle libdemac optimisation for armv6
13:44:45 Join MethoS [0] (
13:44:45 Quit mcuelenaere (Read error: 54 (Connection reset by peer))
13:45:26 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
13:45:32mcuelenaereah I remember; JdGordon: could you boot into Rockbox and go to the debug menu and say me what values you get for IO_CLK_*?
13:46:15JdGordonwith a svn build?
13:46:30mcuelenaeredoesn't matter
13:46:55mcuelenaereI am trying to figure out why the DSP doesn't interrupt my ZVM while it does on the m:Robe 500..
13:47:11mcuelenaereor better why I'm not receiving those
13:47:24JdGordondoesnt seem to want to turn on?
13:48:07mcuelenaereyou mean your m:robe or the DSP?
13:48:38mcuelenaeremm didn't I break yours last time?
13:49:13JdGordonyeah, but it should have been chargning since thne
13:49:19JdGordonlooking for a power cable for it
13:50:56JdGordonyay! ok, batt must be dead... ill do a build and let you know..
13:51:16mcuelenaereok thanks
13:51:53Buschelok, when comparing 64-multiplies it's dct32 (mp3: 6MHz (asm), mpc: 6MHz (few asm)) and dewindowing (mp3: 13MHz (asm), mpc: 16MHz (asm))
13:51:58 Join MethoS- [0] (
13:55:20 Part pixelma
13:59:19JdGordonmcuelenaere: crap.. its not booting rockbox atm.. ill play a bit later and let you know :/
14:02:25mcuelenaereJdGordon: no problem
14:05:35 Quit Llorean (Read error: 104 (Connection reset by peer))
14:07:04mcuelenaereJdGordon: if you get the results, just say them here; I read the logs
14:07:11 Quit mcuelenaere ("bye")
14:10:36 Quit MethoS (Read error: 110 (Connection timed out))
14:12:47 Join MethoS_ [0] (n=clemens@
14:16:16 Quit spiorf (Remote closed the connection)
14:17:57 Join kugel [0] (n=chatzill@unaffiliated/kugel)
14:18:18 Nick Horscht is now known as Horsch86 (n=Horscht@xbmc/user/horscht)
14:18:32 Nick Horsch86 is now known as Horscht (
14:28:34 Join DerPapst [0] (
14:28:47 Quit Vorador ("Ex-Chat")
14:28:54 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
14:29:38 Part kharo ("Kopete 0.12.7 :")
14:30:30 Join kharo [0] (
14:32:18 Quit MethoS- (Connection timed out)
14:38:38amiconnBuschel: Now this is strange: I compared mp2 and mp3 performance on coldfire - and MP2 wasn't faster than MP3 (!)
14:39:09Buschelamiconn: ?
14:39:35amiconnSame track: vbr mp3 ~231kbps: 422% realtime, cbr mp2 256kbps: 415% realtime
14:40:14Buscheleffective imdct ;o)
14:40:19 Quit nplus (Remote closed the connection)
14:43:00 Quit linuxstb__ (Read error: 110 (Connection timed out))
14:47:41amiconnOn PP5022, MP2 is faster, as expected: MP3: 202% realtime, MP2: 245% realtime (same track again)
14:50:34 Quit MethoS_ (Read error: 110 (Connection timed out))
14:54:27Buschelhmm, if i declare an array static const A[] ICONST_ATTR; it is not showing up in the map-file
14:56:45 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
14:58:54markunBuschel: do you want to work on more code sharing between the codecs?
14:59:25Buschelmarkun: i am more interested in tweaking the speed to a maximum
15:02:26 Join Genre9mp3 [0] (n=yngwiejo@rockbox/contributor/Genre9mp3)
15:05:51Buschelcan the location of data influence the speed of loading?
15:05:57Buschel(within iram)
15:11:30 Join urmo [0] (
15:12:44Buschelhmm, when loading data in a loop from the same array. it is faster, if the adresses are near to each other?
15:13:06Buschellike a[4], a[16], ... instead of a[4], a[576], ...
15:13:08urmowhy do you call yourself Buschel?
15:13:25Buschelbecause was my nick since i use irc
15:13:33n1sBuschel: if you have a dcache i think close loads should hit the cache more often
15:13:36urmowhat does it mean?
15:13:43Buschelurmo: nothing
15:13:44n1sscorche: ping
15:13:50gevaertsBuschel: don't bother. our friend urmo is a regular troll
15:14:07Buschelokido :)
15:14:16urmogevaerts sucks
15:15:15Buschelif PP-targest have such cache it might be the reason for mpc being slower in dewindowing
15:16:23Buschelmaybe i need to restructure the loops a bit
15:17:01 Quit parafin ("So long and thanks for all the fish")
15:17:02 Join parafin [0] (
15:17:18urmowhy do you call yourself parafin?
15:18:35 Join crash91 [0] (n=crash@
15:18:57 Quit parafin (Client Quit)
15:19:03 Join parafin [0] (
15:21:03 Join rbirc [0] (n=4fd3c1cc@gateway/web/cgi-irc/
15:21:35parafinurmo, it's transliteration from russian of character's name in game Vangers ( That's why it's got spelling error
15:21:57krazykiturmo, he's a troll. just ignore him
15:23:19urmointeresting, parafin
15:23:56 Join Dave2 [0] (i=Dave2@freenode/staff/dave2)
15:24:54parafinadn yeah, parafin in russian is paraffin :)
15:26:08 Part urmo
15:28:06 Join K4rP4D [0] (n=KrPD@unaffiliated/krpd)
15:29:01 Join linuxstb__ [0] (
15:30:50 Part Dave2 ("Leaving")
15:32:08 Quit crash91 ("Ex-Chat")
15:35:33 Join miepchen^schlaf [0] (
15:39:11amiconnBuschel: If you load data using small *constant* offsets, gcc might generate more efficient code
15:39:29amiconnThere's a way to definitely find out...
15:39:46Buschelamiconn: and for asm?
15:40:29rbircHello. A couple of days ago I reported (here in irc) a 'bug' in the boot screen of sansa c200. There is a line there showing model id as e200. I've now taken a picture of that screen. Should I file a bug report in flyspray?
15:40:51amiconnAccess locality might also help the cache, as n1s mentioned. That doesn't apply to iram though
15:40:56 Nick rbirc is now known as fml (n=4fd3c1cc@gateway/web/cgi-irc/
15:40:57 Quit linuxstb_ (Read error: 113 (No route to host))
15:41:17JdGordonfml: typo in the bootloader?
15:42:29JdGordonfml: show us the pic...
15:42:35fmlJdGordon: I don't know. BarryWardell couldn't spot an error
15:42:44***Saving seen data "./dancer.seen"
15:43:31fmlJdGordon: what is a good site for sharing pics?
15:45:39fmlJdGordon: here:
15:46:14fmlI see this when I connect the player to the PC via USB (to trasfer files)
15:47:12JdGordonany idea what svn revision your bootloader is?
15:47:51fmlJdGordon: no, but I freshly installed everything yesterday via RbUtil.
15:48:40JdGordonwhich version of rbutil?
15:48:59JdGordondomonoky: bluebrother: any idea what c200 bootloader version is shipped with rbutil?
15:49:26domonokyrbutil, doesnt ship any bootloaders, it uses those on ..
15:49:50JdGordonah ok
15:50:02fml Jdgordon: 1.0.5, I think. It's the windows version available at
15:50:44JdGordonlooks like its a fairly old version
15:51:06JdGordonits definatly not still in svn builds.. so if it really worries you install a svn bootloader
15:51:20fmlI must add that the two or three top lines on the screenshot are overwritten. I.e. first I see something else, then the screen becomes full and wraps. The three top lines get overwritten, and then I took the picture.
15:52:19domonokyJdGordon: which means the c200 bootloader on are not uptodate ?
15:52:23 Quit miepchen^schlaf_ (Connection timed out)
15:52:43fmlJdGordon: no, it doesn't bother me really. But back then I couldn't reproduce this "bug" until i realised today that I see this when connecting to PC and not (as I assumed then) when booting up.
15:52:45JdGordonmore, not updated... I doubt there has been anything major (worth updating) since then
15:52:48JdGordonoct 07
15:52:57gevaertsFrom what I see that type is read from the OF, but I could be wrong
15:53:21JdGordonstrings firmware.mi4 suggests it should say c200 though ?!
15:54:04fmlI've just reported this since I thought that this could be a sign of an error somewhere.
15:57:33 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
15:57:42gevaertsCould this be something that's wrong in certain OF versions ?
15:58:21fmlgevaerts: I'm not really qualified to answer such questions. IMHO it could be anything! :-)
16:00:40JdGordongevaerts: I dont think so.. thats rockbox code running..
16:02:48gevaertsJdGordon: yes, but that code prints out data that come from the firmware partition
16:03:19JdGordonprintf(MODEL_NAME); is that line...
16:03:27JdGordonwell before any OF stuff is read
16:04:09gevaertsI don't think it's that line. It looks more like printf("Model id: %.4s", mi4header.model);
16:04:29 Quit crope` (Read error: 110 (Connection timed out))
16:04:41 Nick Horscht is now known as Horscht86 (n=Horscht@xbmc/user/horscht)
16:05:03 Join crope` [0] (
16:05:30JdGordonbah, yep your right
16:07:02gevaertsfml: are you using linux ?
16:07:52 Quit Genre9mp3 ()
16:08:18 Join Biiaru [0] (n=b@unaffiliated/Biiaru)
16:09:46kugelfml: I recommend updating your bootloader, the latest version has a nice bootlogo :P
16:10:40 Quit gregzx ("ChatZilla [Firefox]")
16:11:19*gevaerts is starting to suspect that bugs in certain versions of sansapatcher might be involved
16:11:59kugelHmm, I just noticed, the bootloader doesn't stop the OF from refreshing the database (I replaced the OF bootloader with the rockbox one - sansapatcher -bl bootloader.bin)
16:13:28JdGordonyeah, something funny is going on with sansapatcher
16:14:22 Quit Ragnor (Nick collision from services.)
16:14:25fmlgevaerts: no, I use WinXP
16:14:32 Join Ragnor [0] (
16:17:04kugelThe database refresh seems only to be stopped when the OF is loaded from the firmare partition
16:17:12 Nick linuxstb__ is now known as linuxstb (
16:19:22fmlkugel: just run the freshly downloaded sansapatcher −− same result. As for DB refresh: pixelma said that it works only for e200. In c200 the DB is refreshed immediately and not on the next boot.
16:19:49kugelfml: Sorry, I mean a svn bootloader. The bootloader in sansapatcher is pretty old
16:22:00 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
16:22:04 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
16:22:18 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
16:22:19fmlkugel: I wouldn't like to compile bootloader myself.
16:22:25*gevaerts found it...
16:22:53gevaertssansapatcher.c line 609 : memcpy(buf+0x1f8,"RBOFe200",8);
16:22:53fmlgevaerts: what?
16:23:22gevaertsNo wonder it says e200 then
16:23:35 Join tessarakt [0] (
16:26:10kugelgevaerts: Isn't that only a string put into the header?
16:26:42gevaertskugel: yes, but that's also only what the bug was (wrong string shown)
16:27:03kugelgevaerts: I mean, that string doesn't show up in the bootscreen, why would this cause this? Or does the bootloader identify the model by that string?
16:27:34fmlgevaerts: exactement!
16:28:24gevaertskugel: fml "complained" that the bootloader shows "Model id: e200" on his c200 when booting the OF.
16:28:59kugelgevaerts: I know
16:29:25gevaertsNothing actually uses that string as far as I can see
16:29:50kugelAh I see, sansapatcher updates the mi4header with that string, and the bootloader reads that string (mi4header.model), right?
16:30:06gevaertsyes, that's it.
16:31:09 Join DaCapn [0] (
16:32:51fmlWhat's about the study mode? Will it be reverted or retained?
16:35:42kugelgevaerts: That means every bootloader shows e200 on every c200? I wonder why nobody noticed it yet
16:36:32gevaertsMine doesn't :)
16:36:50gevaertsThat text doesn't stay long
16:37:12kugelI mean why doesn't your bl say e200
16:37:30mcuelenaereare there 64-bit machines in the build chain?
16:37:30gevaertsActually, it does, but not on plugging in
16:39:27kugelgevaerts: And are you allready having a fix?
16:39:33obomcuelenaere: yes
16:40:28gevaertskugel: no. I leave that to people who know sansapatcher better. I wouldn't want to break it while fixing something that is at most a slight cosmetic bug
16:40:46 Quit Zarggg ("Pursued by a bear")
16:40:54kugelgevaerts: Probably a good decision ;)
16:41:27mcuelenaereobo: does a pointer->int conversion on 64-bit machines requires some special types in C besides var = (int)ptr_othervar; ?
16:41:58JdGordonuse intptr_t
16:42:56JdGordonmcuelenaere gets the prize for the biggest target delta ever!
16:43:06mcuelenaereJdGordon: so like this: int_var = (intptr_t)ptr_othervar?
16:43:16mcuelenaereyes, I know there's something wrong with my last commit
16:43:30mcuelenaereI don't get that filesize at my local tree :/
16:43:43*JdGordon not sure.. too sleepy
16:43:55*JdGordon -> zzz
16:43:58 Nick JdGordon is now known as JdGordon|zzz (i=jonno@rockbox/developer/JdGordon)
16:44:02 Quit fml ("CGI:IRC")
16:44:45 Quit desowin (Read error: 110 (Connection timed out))
16:45:41gevaertsmcuelenaere: you can't reliably store a pointer in an int, so int_var already has to be of type intptr_t
16:45:41 Quit Bagder (Read error: 104 (Connection reset by peer))
16:46:31 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
16:48:14mcuelenaeregevaerts: will something like this work?
16:48:17kugelmcuelenaere: Well done ;)
16:48:30kugelregarding bin size
16:48:46mcuelenaereI know...
16:48:57gevaertsmcuelenaere: the return needs to cast to int, not intptr_t
16:49:26mcuelenaerethere's no implicit casting?
16:49:30kugelgevaerts: Is there allready a bug filed for the bootloader thing?
16:49:38kugelbug report rather
16:49:44gevaertsNo idea
16:50:09 Quit culture (No route to host)
16:50:47 Quit tessarakt ("Client exiting")
16:50:53kugelsearching for "bootloader c200" didn't give a usefull result.
16:51:08 Join Zarggg [0] (
16:51:15kugelgevaerts: I'll just do it, ok?
16:51:36Buschelwrong window
16:51:50kugelWe are not your compiler :(
16:51:51bluebrothermake: *** No targets specified and no makefile found. Stop.
16:53:22 Nick Biiaru is now known as Makefile (n=b@unaffiliated/Biiaru)
16:53:28 Nick Makefile is now known as Biiaru (
16:53:31Biiarusomeone else owns that nick
16:53:58n1smcuelenaere: since just the binsize grew and not the ram usage it is probably zero padding
16:54:30Biiarualso, for a moment there i thought i wasn't on freenode
16:54:31mcuelenaeren1s: I think it has something to do with the linker somewhere
16:56:26 Join tessarakt [0] (
16:59:28 Join spiorf [0] (
17:00:47 Join Xqtftqx [0] (i=4cbd373c@gateway/web/ajax/
17:01:34XqtftqxHey guys, im having a problem getting this video to run, ive converted it using the vlc method, and it doesnt work, this only happens when converting avi files
17:07:00kugelSlasheri: ping
17:07:12soapavi is a container - could be anything in there. In my experience ffmpeg (you can use the WinFF front end) is the most reliable method for converting videos.
17:07:26XqtftqxAlright ill try that way
17:07:47kugelXqtftqx: WinFF has rockbox presets
17:08:02BigBambiFor some devices
17:08:21BigBambiBut the Gigabeat F/X is there, that you can also use for the S
17:09:49XqtftqxAs i know, is WinFF faster then VLC?
17:10:05BigBambidon't know
17:10:10 Quit spiorf (Remote closed the connection)
17:11:05XqtftqxIll do some benchmarks
17:11:25 Quit miepchen^schlaf ()
17:11:40XqtftqxThis will probly be later though,IM ripping some dvd's for rockbox
17:11:42 Join miepchen^schlaf [0] (
17:12:17BigBambiXqtftqx: As with the other day when you were here, you do not need to tell us everything you are doing
17:12:17Slasherikugel: hi
17:12:29XqtftqxWheres the current status in this command prompt?
17:12:38BigBambiXqtftqx: eh?
17:12:42Slasherikugel: when tagcache filters are applied, search results are not supposed to be in sorted order
17:12:50XqtftqxLike the percent
17:13:05BigBambiXqtftqx: This isn't really related to rockbox
17:13:16kugelSlasheri: Ah, nice you allready know what I'm talking about
17:13:25XqtftqxSorry, nvm then.
17:13:26Slasherikugel: yes, i read that :)
17:13:49kugelSlasheri: So it's not supposed to be in order, but (in pf) it seems to have an order (the inverted one though)
17:13:54Slasherikugel: only unfiltered results are always sorted
17:15:14Slasheriinteresting.. it might have some order but it can be anything. anyway, filtered results needs to be sorted afterwards if necessary (tagtree.c does that also)
17:15:52 Join toffe82 [0] (
17:15:56kugelSlasheri: From what I've seen with my own albums and the bug report someone filed, the order is indeed inverted
17:16:59kugelSlasheri: Ok, I've done a sorting in the patch (FS #8335), obviously I only inverted the sorting of the result
17:17:25 Quit miepchen^schlaf ()
17:17:53Slasherikugel: hmm, i will check
17:18:13kugelSlasheri: Can you point me where tagcache.c sorts? 4.5k lines of code :/
17:18:39 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
17:18:55Slasheritagcache sorts only when committing something to the db. it never sorts runtime (tagtree.c does that when necessary)
17:20:02Slasherikugel: hmm, in fact adding filters shouldn't break the sorting..
17:20:04 Join Horschti [0] (
17:20:17Slasherionly adding clauses needs re-sorting after retrieving the results
17:20:26 Nick Horschti is now known as Horscht (
17:20:32Slasherithat's how retrieve_entries in tagtree.c works also
17:20:32 Join miepchen^schlaf [0] (
17:20:43Slasherireally weird if that result list is _really_ inverted..
17:20:46kugelSlasheri: When you say tagcache sorts at committing, doesn't that mean any tagcache search should result ordered?
17:21:00Slasherikugel: yes
17:21:08Slasherialmost any
17:23:13kugelSlasheri: pf searches also for albums. The results are in alphabetical order (and not inverted here)
17:23:26kugelat this search there's no filter added though
17:23:30 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
17:24:07Slasherihmm.. in fact iirc, any alphabetical tagcache result really should be ordered because of that pre-sorted data
17:24:38Slasherikugel: which result was inverted?
17:24:59kugelSlasheri: the track search, with tag_album filter added
17:25:31 Quit ali_as ("Shutdown!")
17:26:11Slasherihmm.. i have no idea how that is possible. tracks should have been sorted also
17:27:17 Join dabujo [0] (
17:27:29Slasheriunless data is somehow sorted twice and got inverted during commit
17:28:12XqtftqxHey guys, just want to say, WinFF worked
17:28:15kugelSlasheri:, this is the part where the track search happens
17:28:26kugelI couldn't see anything wrong there
17:29:22 Quit tessarakt ("Client exiting")
17:30:14kugelWhen I put "DEBUGF("tcs.result: %s\n", tcs.result) into line 42 (in the pastebin), the result was allready wrong-ordered
17:31:29Slasherikugel: interesting.. probably a tagcache engine bug then. did you try re-creating the db and see if that still happens?
17:31:55 Part Buschel
17:32:03kugelSlasheri: The database was pretty "fresh"
17:32:29kugelSlasheri: I didn't actually, but I can't remember the tracklist being in the correct order.
17:33:04kugelalso see:
17:33:05 Quit Xqtftqx (" ajax IRC Client")
17:34:10Slasherikugel: did you try if that bug happens with and without the db loaded in ram?
17:34:42 Join mackes [0] (
17:34:53 Quit linuxstb (Read error: 113 (No route to host))
17:35:03kugelSlasheri: Note, the bug only (seems to) happen in the pictureflow plugin. The database works as expected
17:35:42Slasherikugel: if you go directly to the track list in the db browser, is that sorted correctly?
17:35:55Slasherihmm, weird
17:36:41 Quit Horscht86 (Read error: 110 (Connection timed out))
17:37:54Slasherikugel: tagtree shouldn't sort that top level tracklist, so it should be displayed as received from the tc engine
17:39:21kugelSlasheri: In tagtree, I see a bool: sort_inverse
17:39:42kugelNot how meaningful that is
17:40:08Slasherithat is a configuration option in tagnavi.config that can be used to invert the result list (and if used, involves tagtree.c to re-sort the results inverted)
17:41:25Slasherikugel: please try to disable the qsort from tagtree to be sure it doesn't sort the list
17:41:34Slasheriand then check again if the tracklist is really in order
17:42:48***Saving seen data "./dancer.seen"
17:45:04Slasherikugel: ah, yes
17:45:08mcuelenaerewhy is there a .ALIGN(0x200) at the beginning of DRAM in
17:45:14 Join ali_as [0] (
17:45:32mcuelenaere.= ALIGN(0x200) *
17:45:57n1smcuelenaere: i think that is used for the usb stack or something
17:46:43Slasherii didn't remember at first, but now i do. Using any filter, makes tc to go through the master index to find matching results. And master lookup isn't ordered, so the results should be in fact random
17:46:54kugelSlasheri: I commented 'if (sort) qsort(..);' out and it appears to be still in correct order
17:47:21Slasherikugel: now try going from album -> tracklist
17:47:28Slasherithat should be in random order now
17:48:12kugelcorrect order, also correct when I go through albumartist/artist before
17:48:38kugelSlasheri: The order doesn't appear to be random, it's indeed ordered inverted
17:48:48kugel(in pictureflow)
17:49:11Slasherijust a moment, i am checking the code
17:49:15Slasheribut that shouldn't be possible..
17:50:38Slasherido you have tc loaded in ram?
17:51:58Slasheriprobably it's just pure luck if the results are ordered, they shouldn't be
17:53:24Slasheriwhen filters are used, tc engine doesn't anything about the actual tags when doing the filtering. that means a sorting is necessary after the finished retrieve
17:54:08jhMikeSHorscht: still around? I wanted to know if you experienced any dropout trouble with latest revisions (ata hack disabled). I see no FS reports or any other issues and want to just remove that code entirely.
17:54:45HorschtI was using a build about 2 weeks old until today
17:54:52HorschtI am currently compiling a new one
17:55:21jhMikeSHorscht: ok, give word.
17:55:23kugelSlasheri: It's definetly not random, I looked through several albums in pictureflow now, and it's allways inverted order
17:55:32Horschtwill do, jhMikeS
17:56:12Slasherikugel: that might happen that the file system returned files in inverted order when adding albums so it looks apparently ordered
17:56:40Slasherikugel: anyway, a full sorting is required atm to fix the ordering
17:58:17Slasherikugel: tc engine was designed to be high performance and low memory hungry, so that is one tradeoff it has
17:58:41 Quit DerPapst ("*plop*")
17:58:46 Join gregzx [0] (
17:59:00kugelSlasheri: So how would I order the tracklist?
17:59:18Slasherikugel: you have to implement a qsort in the pf
17:59:45kugelSlasheri: Ok
18:00:59kugelSlasheri: Wait! You could be right with the filesystem returns the file in inverted order. I have named my files with the track number in the front
18:01:14HorschtjhMikeS, no issues so far. The playback stuttering on the first buffering is gone
18:01:15kugelCould that be the reason it appears to have an order?
18:01:36Horschtplus it takes just over 1 minute, as opposed to 2.5
18:01:37Slasherikugel: yes, that is exactly what i was thinking of :)
18:01:56Horschtrev 17555
18:02:36jhMikeSHorscht: Nice.
18:02:48Slasherikugel: because the order tc returns is what are inside master index (i.e., the order files were added to the db)
18:03:07 Join Buschel [0] (
18:03:30HorschtjhMikeS, so, what did you do, or to be more precise: what was wrong before?
18:03:36*Buschel now know where the main difference between mpc's and mp3's dewindowing is :o)
18:03:56 Join l403 [0] (n=l403@
18:04:03l403hi guys
18:04:04jhMikeSHorscht: amiconn's commit to fix the dropout on nano was the same cause for 5.5g problems it turns out.
18:04:25jhMikeSor rather nano + 5.5g had the same problem
18:04:31Buschelmp3 saves several ldr's or ldm's plus some mul's via using additional symmetry of the window itself −− gotta try to implement this for mpc, too
18:04:59Slasheritc could be changed to keep the right order in results, but that would mean tc would have to generate a lookup list/hash of every song first and then go through the requested index and check for every entry if the lookup hash contains it too
18:05:03 Join DerPapst [0] (
18:05:04jhMikeSHorscht: they needed longer delays before using the drive after powering it up
18:05:25kugelSlasheri: Ah yes.
18:05:31Slasheribut that would need a lot more memory to do runtime (or otherwise performance would be degraded)
18:06:37Buschelsaves 8*36*2 multiply-adds as well as 248*36*2 ldr's per frame
18:06:59l403i was wondering how is it possible that daasheet for some chips are not available like samsung S5L8700. it doesnt make sense to me
18:08:52Slasherikugel: hmm.. in fact! that doesn't require that much of memory if done properly (only 1 byte/entry). Could be done in future to enhance the tc engine. The only tradeoff would be that UI couldn't display any progress when calling tagcache_search and tagcache_get_next() wouldn't iterate that much any longer
18:10:59kugelSounds good :)
18:11:04soapwhat happened to 0x0065? Did he remove his own account?
18:11:52Slasherikugel: i will definately add that in todo. It would be a slight UI tradeoff, but would simplify the code using tc a bit
18:13:03kugelSlasheri: BTW, you are indeed right
18:13:26kugelI removed the leading tracknumber in the filenames of 1 album, and the order is completely different in pictureflow now
18:13:44Slasherihehe, good to hear that :)
18:14:03Slasheriotherwise it would have been really weird
18:14:10kugelSlasheri: I didn't think of that, that my filenames are causing the order :/
18:19:42kugelSlasheri: Implement that sorting asap, I'm too lazy to implement that in pf now :/
18:19:54Slasherikugel: in fact, correcting that sorting order would also enhance two other funtions of the tc engine
18:20:06Slasheriso i will try to do that soon if i have some time
18:20:24 Join jaczehack [0] (n=d572f7f1@gateway/web/cgi-irc/
18:20:43Slasherikugel: i would like to. i might try to see today what it needs to be done :)
18:22:49 Join Davide-NYC [0] (
18:23:26Slasherihmm.. what others think, would it be ok to set a hard limit to the maximum track count in the database? for example 64k tracks (it would be very impractical to have that many tracks either)
18:23:36Davide-NYCJdGordon|zzz: ping
18:27:10 Quit HellDragon (Client Quit)
18:28:08 Join miepchen^schlaf_ [0] (
18:28:22kugelSlasheri: Not sure. I just know that pictureflow has such a gap
18:29:04 Quit l403 (Read error: 113 (No route to host))
18:29:19 Quit Davide-NYC (Read error: 104 (Connection reset by peer))
18:29:21 Join HellDragon [0] (i=jd@unaffiliated/helldragon)
18:34:06 Part Buschel
18:35:43 Join herrwaldo [0] (
18:36:57n1sSlasheri: why would it be impractical?
18:39:49 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
18:43:29kugelAnyone wants to commit ? :)
18:43:30 Quit miepchen^schlaf (Connection timed out)
18:45:09 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
18:46:34Slasherin1s: tagcache doesn't handle very well that many tracks
18:47:00 Join nuonguy [0] (
18:47:23Slasherikugel: anyway, i found i couldn't do it that way. But! I can do it so that if there are for example not more than 32 tracks in the results, then results are sorted. Otherwise, they are not
18:47:30n1sSlasheri: now that dap harddrives have crossed 100GB i don't think it's unrealistic for someone to have that many tracks...
18:47:57Slasheriindeed.. hopefully those daps have a more powerful cpus too
18:48:42Slasherikugel: but by sorted tracks, do you mean ordered by the track number or alphabetically sorted?
18:49:12kugelSlasheri: I'd rather sort by tracknumber if tracknumber info is available
18:49:20kugeljust like the tagtree does
18:49:32kugelSlasheri: In any way, I think it would be nice if the api would offer some function which handels the ordering
18:50:44Slasherikugel: ok, that would be slightly harder (but not impossible) to do
18:51:15Slasherithe problem is that sorting a large number of entries, requires extra buffer space the tc engine doesn't has
18:52:13Slasheribut limiting the sorting (and even configuring criteria) to the SEEK_LIST_SIZE number of results would be possible to do in the engine also
18:53:04 Quit mcuelenaere ("Leaving..")
18:53:52kugelSlasheri: I think it doesn't necessarily need to be done by tc. Maybe there could be an API provided by tagtree
18:54:14Slasherithat could be a good idea indeed
18:54:51Slasheriin fact it could be the best way to do that
18:55:08kugeli.e. tagTREE_search instead of tagcache_search in the plugin api
18:56:03 Join Buschel [0] (
18:57:46 Quit jhMikeS (Nick collision from services.)
18:58:02 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
19:09:01kugelNico_P: ping You wanted to commit some time ago. I think now is a good moment. I'm pretty much finished
19:09:12 Join AceNik [0] (n=AceNik@
19:09:42AceNikhey uys can i make my own bootloader from the current source, is there any harm ?
19:10:43kugelAceNik: There *should not* be a harm. Even though SVN bootloader are not tested deeply, thus you shouldn't expect it to work
19:11:00kugelAceNik: If you are running an e200, I can tell you the SVN bootloader is safe to use
19:11:25 Join Llorean [0] (
19:11:51AceNikkugel: let me compile one for h10 [20GB] & see, what kind of changes can i expect, i mean anything significant
19:13:34kugelAceNik: Do it on your own risk. The most noticeable change is the bootlogo.
19:14:36AceNikkugel: how, you mean the version number ?, secondly the only consequence is it would not work right
19:16:11kugelAceNik: I just say: Never change a running system. A bugged bootloader can mess up your dap very bad. I recommend not to update the bootloader if it's not necessary
19:17:05AceNikkugel: ok if thats the case i wont try
19:17:49AceNikkugel: but really that bad, i mean its just once that it has to boot to proove its working, h10 i think is pretty stable now
19:19:25n1sAceNik: it depends on how easy it is to recover from a bad bootloader
19:20:10AceNikn1s: usually what kind of cases do you face when you try to upgrade botloaders, i mean its gonna happen this devcon as it is
19:20:49 Join BlakeJohnson86 [0] (
19:22:07AceNikn1s: i would be able to boot into disk mode right ? even if the bootloader is messed up
19:22:24AceNiki can always re flash with my irivers local botloader
19:23:55n1sAceNik: i don't know for the h10 but for example an ipod is easy just needs to start in diskmode and write to the hd, an h100/h300 however requires special hardware and soldering to recover so you better be sure...
19:24:15 Quit gregzx ("ChatZilla [Firefox]")
19:24:50AceNikok im trying it lets hope things go normal
19:25:28n1sbut why do you wan tto update the bootloader?
19:25:57AceNikim just curious
19:27:10kugelAceNik: Please inform on how to recover the H10 before you do that
19:27:41AceNikas in ?
19:28:13AceNikwell if it gets into disk mode fine otherwise ill have to hack it through linux
19:28:25AceNikforce mount it or something
19:29:04kugelAceNik: See, I nearly bricked my e200 recently because of updating the bootloade
19:29:17kugelalthough it was my fault, the bootloader was fine
19:29:51 Join inigomontoya [0] (n=621cdcfd@gateway/web/cgi-irc/
19:30:01AceNikkugel: lets see, i mean i really want to try this, rockbox has reached a stage i think with the h10 it cant be so bad
19:30:04kugelthe e200 has 4 boot modes (normal, recovery, manufactoring and preboot mode), I needed to go into preboot
19:30:27inigomontoyakugel: have you tried the patch that fixes doom when rotated yet?
19:30:33kugelAceNik: And some players don't even have a recovery mode
19:30:34AceNiki have jus a ums boot mode
19:30:59AceNikwell i do hav a disk mode & a mtp mode relax ill get back to you on this in 10 mins
19:31:21kugelinigomontoya: No, do you?
19:31:58inigomontoyakugel: i am the one that made it, so yes :D
19:32:03AceNikok just tell me once the bootloader is compiled, after the make process, it will be saved as the final file right, i need no type any other command after make ?
19:32:06 Join linuxstb__ [0] (
19:32:17 Join Arathis [0] (
19:32:33kugelinigomontoya: Oh, sorry. I think I tried it in the simulator and didn't experience a crash
19:34:08kugelinigomontoya: Although, I just see, there*s no crash in the sim w/o that patch to
19:34:38*gevaerts isn't surprised. simulators are very different from real targets
19:34:39 Quit kharo (Read error: 110 (Connection timed out))
19:35:24kugelgevaerts: That's why I said no in the first place :)
19:36:16inigomontoyakugel: I have noticed no bad effects or changes other than doom not crashing with the patch
19:38:33kugelinigomontoya: sounds good
19:40:57 Join Zarggg_ [0] (
19:41:11 Quit Zarggg_ (Client Quit)
19:41:15 Join Wolfy|Sajber [0] (
19:41:37Wolfy|Sajberhey people :)
19:42:42 Quit amiconn (Nick collision from services.)
19:42:48 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
19:42:51***Saving seen data "./dancer.seen"
19:43:44 Quit inigomontoya ("CGI:IRC (EOF)")
19:44:10 Quit linuxstb_ (Read error: 113 (No route to host))
19:51:35advcomp2019hello Wolfy|Sajber.. have a question?
19:53:29Wolfy|Sajberadvcomp2019: any news on the ZVM port? how close is it to a real "build"?
19:54:37advcomp2019i dont know both
19:54:57Wolfy|Sajberhow about the gigabeat s? :P
19:55:37kugellinuxstb__: just curious, what's your progress on the clip?
19:55:47AceNikhow about the zune CPU encryption has it been cracked ?
19:55:55 Join Zarggg_ [0] (
19:56:01toffe82Wolfy|Sajber: the S is working but still a lot to do
19:56:24Wolfy|Sajbertoffe82: ok, so no "real" build for awhile?
19:56:51 Join weatherundergnd [0] (
19:57:03kugelWolfy|Sajber: Afaik, the main issue is the installation which is rather complicated and AFAIK requires linux
19:57:15 Quit Mathiasdm ("Yuuw!")
19:57:17kugelfor the ZVM you even need to take out the HD
19:57:17 Join DayRider [0] (n=dodo@
19:57:39toffe82Wolfy|Sajber: check here
19:57:46Lloreankugel: The main issue is that not all the necessary hardware is supported to the level we're comfortable releasing yet.
19:57:59Lloreankugel: A linux-only installation would not preclude release for the S.
19:58:22Wolfy|Sajberkugel: so the installation for the ZVM will allways requiere to take out the harddrive? :/
19:59:11kugelWolfy|Sajber: That's not my business. I don't have a ZVM or something. I just told you what I recall
19:59:18Wolfy|Sajberkugel: ok
19:59:27*DerPapst installed rockbox on the s without leenux
19:59:35kugel(same applies for Gigabeat S)
19:59:52LloreanDerPapst: Modified firmware updater?
20:02:10 Join Mathiasdm [0] (
20:02:17 Quit Zarggg (Read error: 113 (No route to host))
20:04:28weatherundergndi tried to play one .SID song on my iRiver H140, but the player jammed. is there any way to do hardboot?
20:04:38weatherundergndi can't seem to poweroff the player.
20:05:53n1sweatherundergnd: push a straightened paperclip into the reset hole in the bottom of the player
20:07:44weatherundergndhey i've never noticed that before :)
20:08:58weatherundergndn1s: you have iRiver player too?
20:09:06n1si have a h320
20:09:40weatherundergndi noticed that if i have the remote controller between the player and my headphones i get strange static noise
20:09:50weatherundergndsomekind of static interference from HD etc, maybe?
20:11:00BigBambiThere is an issue on some H100s about a remote ticking noise, and there is an option to try and help that. It doesn't happen on all of them however, I don't get anything on my H140
20:11:16BigBambiI don't know if that is the same as your issue, but you can try and see
20:11:25weatherundergndyeah it's the ticking noise
20:11:32weatherundergndbut it comes only when the remote is connected
20:11:44n1sweatherundergnd: there is a known hardware defect in some players/remotes that cause ticking, try enabling "reduce ticking" under general -> display -> remote lcd
20:11:47BigBambiHave you enabled the option to try and reduce it?
20:13:34weatherundergndit makes it little less audible, but doesn't really remove it :(
20:13:42kugelvmware isn't open source is it? shouldn't rockbox encourage to use open source software (like virtualbox) and thus offering a vbox developement image instead?
20:13:49BigBambiweatherundergnd: hence reduce
20:13:58BigBambiAs n1s, it is a hardware issue
20:14:05n1skugel: feel free to make one...
20:14:17kugeln1s: hehe
20:14:20weatherundergndi haven't noticed that ticking noise with the original firmware
20:14:25kugelthat answer was predictable ;)
20:15:32weatherundergndone more stupid question.. how do you boot back to original firmware?
20:15:59 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
20:16:02n1sweatherundergnd: hold rec when booting iirc
20:16:31 Quit Mathiasdm ("Yuuw!")
20:17:01BigBambiweatherundergnd: checking out the manual would be a good idea
20:21:59weatherundergndi seem to have the remote ticking also on the original firmware.. hmm. i'm using EU firmware on this player. i wonder if trying US firmware would do any difference.
20:22:21weatherundergndi don't remember having that remote ticking problem last when i used the player
20:22:23LloreanIt's a _hardware_ flaw.
20:23:09weatherundergndLlorean: i'm out of warranty. :( i wonder if it's still possible to get it fixed.
20:23:16weatherundergndis the flaw in the remote or the player itself?
20:24:09LloreanI don't think you're likely to be able to get it fixed. As far as I know it happens in a large number of the players, and is more of a design flaw that some people just happen to avoid than anything else.
20:24:46jhMikeS16-ohm headphones seem much more susceptible to ticking problems than 32-ohm (i've have checked this out myself)
20:25:04LloreanThe simplest solution is just not to plug your headphones into the remote.
20:25:45 Part DayRider
20:28:01 Quit linuxstb__ (Read error: 113 (No route to host))
20:33:08 Join guest45 [0] (n=864c034d@gateway/web/cgi-irc/
20:33:52 Quit guest45 (Client Quit)
20:34:01 Join guest45 [0] (n=864c034d@gateway/web/cgi-irc/
20:34:48guest45Hi, how often does the since-4weeks page update?
20:34:58 Join Falafel [0] (
20:35:53Bagder4 times/day
20:36:38guest45thank you
20:36:39AceNikhey guys im using the new booloader in the iriver h10 i just compiled it, i wanted to point out, there was a patch which barry had included which showed the booloader version first then the rockbox version, well it works with the new booloader , but now in the old one
20:39:10AceNikanything else i should be checkin for while using the new bootloader ?
20:39:11 Join spiorf [0] (
20:39:53AceNiklike the first version was 50KB, this one is 70KB, do yu think there would be code that would be useless in it somewhere ?
20:43:09n1sthe new one has a logo included, right? that probably explains the size delta if so...
20:45:03AceNikn1s: ok, but there is one problem, while booting if the hold button is on, it freezes on " hold button is on shutting down" & there is a lousy click
20:46:54 Join ido_sh [0] (
20:47:02 Quit ido_sh (Client Quit)
20:47:41 Join ido_sh [0] (
20:49:16 Join [1]Falafel [0] (
20:49:19AceNikn1s: is there anythin else that needs to be checked when testing a bootloader ?
20:49:44 Quit ido_sh (Client Quit)
20:49:51n1sdoes it boot rockbox/of and do everything else it's supposed to?
20:49:56LloreanAceNik: The "Hold button is on" message is from the OF though, isn't it?
20:50:39AceNikllorean: i dont think so ? is it ususally in rockbox when rockbox is booting hold button n resets the settings right ?
20:50:47 Join culture [0] (
20:51:26 Quit guest45 ("CGI:IRC (EOF)")
20:51:28LloreanAceNik: I can't decipher the grammar of that sentence. If you wait and turn the hold button on shortly before Rockbox boots, yes, hold will reset the settings.
20:51:43LloreanWhich doesn't involve shutting down at all.
20:52:10AceNikllorean: i figured it out, after the booloader logo shows, i tried to put the hold button on now, & it cleared the settings
20:52:40*linuxstb_ points to line 460 of bootloader/main-pp.c
20:52:52AceNikif the hold button is on before the booloader logo shows, then it gives an error messae that " hold button is on" "shutting down"
20:54:25 Join ido_sh [0] (
20:55:05*linuxstb_ thinks the H10 should use a real button for that
20:55:13AceNikotherwise the new botloader does everything perfect 1.reboots on clicking the .mi4 file, 2. boots into OF 3. boots into disk mode even through rockbox 4. clears settings after the boot logo is shown
20:55:23 Quit spiorf (Remote closed the connection)
20:55:31LloreanAceNik: Rebooting on clicking an MI4 file is RoLo not the bootloader
20:55:49LloreanClearing settings is also not handled by the bootloader
20:57:28AceNikso anything else ? i think the h10 bootloader is having no trouble then
21:00:13 Join spiorf [0] (
21:00:24 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
21:02:08 Quit weatherundergnd ("Lost terminal")
21:02:25 Quit AceNik ("Trillian (")
21:05:12 Quit Falafel (Read error: 110 (Connection timed out))
21:05:12 Nick [1]Falafel is now known as Falafel (
21:06:23 Quit Arathis ("Bye, bye")
21:08:28 Quit perrikwp (" ajax IRC Client")
21:08:37 Quit Bagder (Read error: 104 (Connection reset by peer))
21:09:20*jhMikeS hasn't done much with e200 recently and now finds all directories sort in reverse order (Z-A) :\
21:09:42 Join Bagder [241] (n=daniel@rockbox/developer/bagder)
21:12:27n1sjhMikeS: there's a bug that sounds really crazy at FS #8949 that might be the cause?
21:14:27 Join davina [0] (
21:17:47jhMikeSn1s: I'll check. It's just directories, not files.
21:18:00linuxstbDoes dircache affect that bug?
21:19:17 Join petur [50] (n=petur@rockbox/developer/petur)
21:19:30jhMikeSI don't have it on but I'll try with it on then
21:20:07jhMikeSnope, still backwards
21:21:00jhMikeSwtf, I went to turn it off but it said it was already off yet showed it as initialized in the debug screen
21:22:59jhMikeSit won't actually turn on for real here
21:23:57jhMikeSI had "Cache Initialized: No", "Entry Count: 20027"
21:25:25 Quit davina^ (Read error: 110 (Connection timed out))
21:26:01jhMikeSIndeed, you have to switch the setting and then set it back to "Alphabetical"
21:26:59 Join m0f0x [0] (
21:32:06 Quit obo ("bye")
21:32:59 Quit XavierGr (Nick collision from services.)
21:33:12 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
21:38:17 Quit spiorf (Remote closed the connection)
21:42:55***Saving seen data "./dancer.seen"
21:43:45 Quit Wolfy|Sajber ("")
21:44:25 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:48:25bluebrotherBagder: could we get easily accessible icons to note the OSes for rbutil? I.e. an icon for w32, linux and one for os x?
21:50:18 Join saratoga [0] (n=62e14d04@gateway/web/cgi-irc/
21:51:00saratogaBuschel: i noticed your profiling remarks, i'm curious if you figured out where MAD spends most of it's time
21:51:30Buschelsaratoga: not really, i am fighting with the mpc-synthesis now
21:51:35saratogai've not been able to profile it since last winter, and unfortunately i didn't save the map file when i last manged to hack up the profiler well enough to work, so i don't know what function the memory addresses refer to
21:51:39*Buschel seems to loose this fight
21:52:02saratogaBuschel: mpc is already pretty fast, maybe time to find a new project?
21:52:23saratogai think if you put the clock speed as low as it could go for mpc, you'd probably get a slide show from the gui
21:53:04Buschelsaratoga: well, at least i found out mpc is not as efficient as mad in the synthesis
21:53:42 Quit jhulst ("Konversation terminated!")
21:54:20*n1s wonders how to make items in a menu conditionally appear, macro hell isn't very much documented...
21:56:34 Quit bughunter2 ("Leaving.")
21:56:39 Quit goffa (Read error: 104 (Connection reset by peer))
21:58:03 Join mf0102 [0] (n=michi@
21:59:13 Join desowin [0] (n=desowin@atheme/member/desowin)
22:03:41 Join espn [0] (n=d1f5163c@gateway/web/cgi-irc/
22:04:13espnAnyone know how to install Pacbox onto a 30
22:04:20espnGB iPod Video?/
22:05:24BigBambiespn: It is already installed. To play it you need the ROMs, which legally you need to extract from the arcae machine you own. do not ask here where to download them.
22:06:19bluebrotherespn: read the second paragraph:
22:06:19espnOkay, but what does this all mean?
22:06:30BigBambiespn: As I said
22:06:38BigBambiThe emulator is already installed
22:06:42BigBambiIt comes with Rockbox
22:06:46espnunder where it says Md5 checkup
22:06:53bluebrother... the manual tells this ...
22:07:14BigBambiespn: Those are the MD5 checksums of the roms
22:07:24bluebrotherespn: well, if you have that files obtained legally those sums allow you to check the files
22:07:33espnOh okay.
22:07:44BigBambiThe legal way of getting the ROMs is to extract them from an arcade machine you own
22:08:24espnWhat would be a type of arcade machine?
22:08:36BigBambiThe big things you play in an arcade
22:10:11 Join bertrik [0] (
22:14:31 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
22:16:58 Quit espn ("CGI:IRC")
22:22:53 Join Xqtftqx [0] (i=4cbd373c@gateway/web/ajax/
22:23:44XqtftqxGuys is it just my player or does scrolling in text files not work on the S?
22:27:42saratogawhen i run make, where the makefiles for the individual codecs stored?
22:27:59 Quit linuxstb (Read error: 113 (No route to host))
22:28:22jhMikeSXqtftqx: no, not just yours
22:28:40XqtftqxAlright, thanks
22:28:42jhMikeSVol +/- will page up/down though
22:29:04Bagdersaratoga: apps/codecs/[codec]/Makefile
22:29:07XqtftqxIt will? Ill have try
22:29:14XqtftqxThanks jhMikeS
22:29:18 Join goffa [0] (n=goffa@
22:29:42 Join obo [0] (n=obo@rockbox/developer/obo)
22:32:05XqtftqxjhMikeS, if page up and down works, cant you just edit the settings and remap the keys?
22:32:49jhMikeSXqtftqx: Whoever is interested in doing that should go ahead :)
22:33:03XqtftqxAlright, ill see if i can take a look
22:34:06 Quit mackes ("Mackes is out")
22:35:22 Quit Xqtftqx (" ajax IRC Client")
22:35:51 Join Xqtftqx [0] (i=4cbd373c@gateway/web/ajax/
22:38:29XqtftqxjhMikeS: Wheres the txtviewer plugin?
22:39:25bluebrotherXqtftqx: apps/plugins/viewer
22:39:57 Join dabujo_ [0] (
22:39:58bluebrother*viewer.c to be exact.
22:40:26 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
22:41:02XqtftqxAlright, im looking here and i see the keys for gigabeat F/X but no S, wheres the keys for S?
22:42:01bluebrothersearch for GIGABEAT_S_PAD
22:42:02XqtftqxNVM, i found it, (Somebody needs to add a commet)
22:43:13n1shmm, 'someone' should do a s/GIGABEAT_PAD/GIGABEAT_F_PAD/g
22:46:33XqtftqxAlso Autoscroll works
22:48:59XqtftqxHmmm, theres no maps here for lineup and line down
22:49:07Xqtftqxwhere are those keymaps?
22:57:04 Quit dabujo (Read error: 110 (Connection timed out))
22:57:07 Join fdinel [0] (
22:57:23bluebrotherthe keymap is at the beginning of the file. Just add the missing line up / down definitions
22:58:21XqtftqxI see the keymaps
22:59:34Xqtftqx but only for page up and down, on all targets
22:59:38 Quit daskhweg (Read error: 110 (Connection timed out))
22:59:49 Join daskhweg [0] (
23:00:10XqtftqxI cant find anything with line up&down
23:01:10BigBambiYes, because it isn't defined for the S
23:01:17BigBambiThat is what you are trying to add
23:02:06Xqtftqx, theres none for any other target
23:02:34 Nick linuxstb_ is now known as linuxstb (n=linuxstb@rockbox/developer/linuxstb)
23:02:39XqtftqxThats for the F, which line up and down works
23:02:51BigBambiI see them for loads of targets
23:03:18BigBambiH1x0 for a start
23:03:26bluebrotherXqtftqx: just search for VIEWER_LINE_UP and you'll see ... ;-)
23:03:41XqtftqxAlright, thanks
23:03:49bluebrotherit's also around for the sansas
23:04:21XqtftqxAlright, i see it now
23:04:26BigBambiall you need to do is pick a button and add e.g. #define VIEWER_LINE_UP BUTTON_WHATEVER
23:05:38Xqtftqx#define VIEWER_PAGE_LINE_UP BUTTON_UP
23:05:47BigBambifor example
23:06:06BigBambihave you checked what the S buttons are 'called'?
23:06:18XqtftqxI havent where are they?
23:07:49n1sXqtftqx: you know there are tools for searching...
23:07:59*domonoky thinks about improving touchstrip driver of the m:robe100.. its still is too sensitiv, and scroll support would be nice..
23:08:00*BigBambi should read whilst typing
23:08:06 Quit jaczehack ("CGI:IRC")
23:09:42BigBambijhMikeS: Are any of the buttons able to be detected simultaneously on the beast?
23:10:09*BigBambi is guessing power maybe
23:10:47jhMikeSBigBambi: any combo
23:10:54BigBambiah, cool
23:10:59BigBambithat's nice
23:11:05 Join mackes [0] (
23:11:14saratogareverting to the revision i used last winter to profile mad doesn't seem to work
23:11:19saratogai wonder what i'm missing
23:13:24n1shaha, i love this: while booting the h300 pressing rec will do one of 3(!) things, if pressed early boot OF, if pressed a little later, clear settings and usually enter rec screen, and pretty late only enter rec screen :)
23:16:06peturand if the usb cable is inserted, only charge....
23:17:28n1sah, right never used that
23:19:59peturwell....vonly charge *and* enter rec screen :(
23:21:05BigBambihehe, the target list in configure doesn't fit in my normal terminal window any more :)
23:22:42linuxstbWhat's the difference between the 30GB and 60GB ZVM builds?
23:22:55*Bagder doesn't know
23:23:21BigBambilinuxstb: That was discussed in the logs a couple of days ago but I can't remember the answer
23:23:32n1sBagder: could you take a quick look at a license ~ 20 lines, and say if you think it's gpl compat?
23:23:40Bagdersure, hit me!
23:23:56*BigBambi readies his rhythm stick
23:25:14Bagderyes, I claim it is gpl compatible
23:25:31n1sso ok to commit to rockbox svn?
23:25:40n1sgreat, thanks :)
23:26:51*Buschel prays for his commit working fine
23:26:58*Buschel never added files yet
23:27:27*Bagder polishes up the "i broke the build" badge in preparation
23:28:13Buschellet's see
23:28:39 Quit Xqtftqx (" ajax IRC Client")
23:28:48BigBambiFor the S text viewer, would people prefer previous and next to shift one character left and right, or a combo with something + left and right (and what would that something be)?
23:29:00BigBambiPersonally I prefer non-combos
23:29:47amiconnBuschel: Should that ICONST_ATTR improve speed on coldfire?
23:29:55Llorean Why do you need to select individual characters in a text viewer?
23:29:58n1sBigBambi: yes, non-combos are better
23:30:04BigBambiBut the prev and next buttons are on the side, and therefore feel a bit like up and down
23:30:22Buschelamiconn: it will be, if the file uses PNS. this should only be used for low bitrates
23:30:24BigBambiLlorean: It is if you have wide mode on, and want to scroll left and right by characters not by pages
23:30:45LloreanBigBambi: Wouldn't Left/Right be better for scrolling horizontally, and Prev/Next be better for line or even page downs?
23:31:31BigBambiLlorean: Page left/right, scroll left/right, page up/down and scroll up/down all exist, I'm just adding some definitions for missing ones
23:31:50LloreanBigBambi: Just because ones exist doesn't mean they can't be changed so the overall scheme makes more sense, though...
23:31:57BigBambiLlorean: I agree
23:32:04 Join bughunter2 [0] (
23:32:06BigBambiAt the moment, page up/down is vol up/down
23:32:17BigBambiscroll up down (by lines) is up/down
23:32:18LloreanI'd strongly, *strongly* suggest trying to come up with a scheme that leaves the volume free.
23:32:32BigBambiand page left/right and character up right is not set
23:32:41LloreanI'd very much like to see the volume buttons actually controlling volume one day on targets with explicit ones.
23:32:45BigBambi*character left/right
23:33:01BigBambiLlorean: So I think up/down should be line up/down
23:33:16BigBambiAnd prev/next should be page up/down
23:33:20LloreanSince audio is the primary function anyway, I'd rather see a few combos go in leaving volume free, than using it for a function just so we can take that function away later.
23:33:24 Join faemir [0] (
23:33:30BigBambicharcter left/right should be left/right
23:33:41BigBambiand then page left/right a combo + left/right
23:33:55*Llorean doesn't use text viewer, so he probably shouldn't be listened to though. :-P
23:34:01BigBambiAny thoughts, + preferences for the combo button left/right?
23:34:13BigBambiThis would leave volume up/down free
23:34:22*Buschel is really satisfied with green after adding files for the first time ;-)
23:34:38faemirCan someone tell me the formats that that text reader in rockbox supports?
23:34:42LloreanI have an inbuilt preference for my combo button to be whatever's used as Play/Pause elsewhere, just out of long standing habit.
23:34:45faemir(Besides .txt obviously)
23:34:57BigBambithat's about it AFAIK
23:35:01n1sBuschel: couldn't that Parity array be encoded in 8 32 bit ints instead of 256 chars?
23:35:18BigBambifaemir: you can open others with the open with on the context menu
23:35:43faemirBigBambi: so no .docs? :P
23:35:52SeedBuschel: stop optimizing MPC. it'll become too fast ;)
23:35:55BigBambiLlorean: I can't see any obvious choices for combos on the S that can be done one handed
23:36:03BigBambifaemir: :)
23:36:21faemirDarn :(
23:36:29faemirpdfs? XD
23:36:40Buschelseed: no speed up this time
23:36:43bluebrotherfaemir: is .doc a text format?
23:36:44saratogaso i've defined RB_PROFILE in the makefile, and again in profile.c to make sure its getting compiled, and i see it in
23:36:58saratogayet vorbis.c still fails with an undefined reference to the profile functions
23:37:15saratogai don't understand how something could be compiled but not visable to the linker
23:37:21faemirbluebrother: I dunno.
23:37:39BigBambiLlorean: I could use volume - you can adjust audio via a menu - not as handy I know, but using a combo on the S pretty much requires two handed use)
23:38:26bluebrothersaratoga: I could imagine every compile unit being fine but some exported stuff is missing upon linking
23:38:30linuxstbBigBambi: I almost never use the text viewer, so my preference would be the same as Llorean - keep volume for volume
23:38:38amiconnCombos almost always require two handed use, if only for holding the thing when pressing two buttons at once with the other hand
23:38:51BigBambilinuxstb: looking at the S, could you suggest a combo then to do page left/right?
23:39:11LloreanBigBambi: Personally, I'd like the "page" movement to be without combo, the "fine" movement to be with combo, and Volume to be explicit. A loud song can come suddenly, and having to quit or hassle with menus to adjust it is frustrating. :0
23:39:14amiconnThe only exception I know so far are Mode+Left and Mode+Down on Ondio (can be pressed with just the thumb, as they're adjacent)
23:39:34 Quit K4rP4D ("Leaving")
23:40:00BigBambiLlorean: I agree about the volume, but I prefer fine movement left and right to without combo - if the entire screen jumps left or right I find it hard to kep my place
23:40:20BigBambiLlorean: But notwithstanding that, what key would you use for the combo?
23:40:20saratogabluebrother: but if it compiles the c file, it got the function definition out of the header, and if its in the map file it compiled the function too, so why wouldn't it be able to put the two together?
23:40:26LloreanBigBambi: Left/Right I suppose I could go either way.
23:40:52BigBambiLlorean: Up down and page up/down we have enough keys for no combo needed
23:40:59BigBambiIt is just left/right
23:40:59LloreanBigBambi: Play/Pause. It's in a good enough place anyway, I think
23:41:13BigBambiseems reasonable
23:41:14bluebrothersaratoga: hmm −− maybe that function has been made static?
23:42:00bluebrothercan you pastebin the error somewhere? Even if I might not be helpful I'm kinda interested ;-)
23:42:59***Saving seen data "./dancer.seen"
23:43:33amiconnBuschel: This is just a clean split of the arm asm, right?
23:44:22Buschelbasically yes.
23:44:46Buschelthe asm routines do the same than before.
23:47:02n1s saratoga do you call them through the codec api?
23:48:20saratogan1s: they're only exposed as function pointers?
23:49:09n1syes, they are in codecs.[ch] like any other function you would call
23:52:12bluebrotherhmm −− "undefined reference to `__cyg_profile_func_enter'". Does "cyg" stand for cygwin?
23:52:12 Join nplus [0] (
23:54:07 Quit dabujo_ (Connection timed out)
23:54:20n1swhat should the argument for profile_thread be?
23:54:30saratogan1s: no arguements
23:54:36saratogabluebrother: I have a patch for that problem
23:54:42saratogalet me post it
23:54:55n1sah, I can't read :)
23:57:12BigBambiLlorean: Nope, using BUTTON_PLAY | BUTTON_REL means it still starts auto scrolling when you use play as the combo button, just on release rather than at the start
23:57:56BigBambiI could move quit to power, and use back as the combo button
23:58:13n1sBigBambi: you can set the precodition to avoid that
23:58:23BigBambin1s: In the text viewer?

Previous day | Next day