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

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

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

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

#rockbox log for 2007-03-07

00:00:07 Quit x1jmp (Remote closed the connection)
00:00:08Pepswell, there are lots of artist out there, how do you make that short?
00:00:12Lloreanmcphail: None of the players have a button just for paged scrolling.
00:00:34Lloreanmcphail: In most cases we added the option to hold down a button while scrolling to make it do pages instead of individual entries.
00:00:49Pepsand titile are about 20 times that :)
00:00:50LloreanPeps: Divide by letter.
00:01:00mcphailLlorean: i'm replying to bluebrother and linuxstb_ as above.
00:01:29PepsBut I am talking about the database "Artist" list
00:01:30Lloreanmcphail: Yes, they were referring to the fact that the way the iPod has its buttons set up, there's not a good one to use for such a purpose.
00:01:31jhMikeSpreglow: this manual reads much more clearly than a coldfire doc esp. with examples of instruction use and what it does
00:01:38Lloreanmcphail: Not that there wasn't a button specifically designed for it.
00:01:59LloreanPeps: You can edit your tagnavi.config to divide by letter.
00:02:14pixelmaLlorean: forum question - so you know why some forum syntax issues (for example "<") gets ignored by the preview?
00:02:17mcphailLlorean: i'm not arguing with that fact
00:02:22amiconnjhMikeS: It *is* backward (and it's not consistent - the 'str' operation has the destination on the right)
00:02:37mcphailLlorean: in fact, i wasn't arguing at all
00:02:39 Quit Id2ndR ("Parti")
00:02:40Lloreanpixelma: I have no idea why the preview displays it properly but the actual post doesn't, unfortunately
00:02:55pixelmayes - I lerned that know
00:02:58jhMikeSamiconn: as long as they say so in the first place, haven't gotten to that :)
00:03:05linuxstb_Peps: I don't use the database.
00:03:19bluebrotherpixelma: s/know/now/?
00:03:22*pixelma joins rockbox typing class
00:03:30scorchepixelma: i think i have heard before that it is a prevention for someone trying to hack into the forum
00:03:41PepsI think you are all taking the developer attitude. I am a developer too, but think a "user view" will be that a bad scroolwheel performance means she may abandom rockbox altogether
00:03:52bluebrotherwe have a typing class? Cool, I should do some practice in touch typing :)
00:03:56LloreanPeps: Rockbox hasn't been released yet on the iPods
00:04:34mcphailPeps: there seem to be a few scrollwheel patches on the website. Don't know if they will help.
00:04:41pixelmascorche: thanks for the info
00:04:44bluebrothera user may abandon Rockbox because of the bad battery runtime.
00:04:46LloreanPeps: I assure you, by the time there is a release version for 'users' there will be better scroll performance, it is a prerequisite.
00:04:47preglowthere's a rockbox typing class? i sign on!
00:04:52linuxstb_Peps: There _is_ someone working on it...
00:05:23PepsLlorean, that is great. I will say no more about that then
00:05:24scorchepixelma: something to do with injection and how once something is posted, it goes into the database for easy searching i believe
00:05:49jhMikeSwell, they seem to tell you a lot more in advance anyway...not sure what suprises I'll get
00:05:58LloreanPeps: But remember, nobody pays for Rockbox, and nobody is getting paid for developing it. If you want it to happen sooner, do it yourself. Otherwise, have patience. Calling attention to 'users' doesn't work if there's no incentive to make them happy, anyway.
00:06:09LloreanPeps: It pretty much depends on who's interested in getting specific things done.
00:06:37PepsI am a developer too, and have/had several open source projects myself
00:06:48preglowwould someone who cares about the state of firmware/sound.c please take a look at this patch? it makes sim use more of the code targets use:
00:07:17PepsI do not have the time mow to go into rockbox development, and really appreciate all the work you guys are doing
00:07:17preglowthe patch name pretty much reveals my reason for doing it
00:07:18mcphailTalking of which, can anyone give me some constructive criticism of my 3g scrollwheel patch:
00:07:35linuxstb_Peps: Something else to consider is that Rockbox runs on about 20 different MP3 players, only a small number of which are ipods. Most developers prefer their non-ipod targets...
00:08:08Pepsipod marginalized. I never thought that day will come :)
00:08:11LloreanPeps: Or at the very least, spending their limited time working on things that benefit everyone.
00:09:04bluebrotheripod maginalized, at least for the devs. I guess it's the biggest user base still
00:09:09Pepsbut think about a scenario where rockbox is so good 50% of ipod's in the world are running it
00:09:20*bluebrother hides
00:09:29LloreanPeps: That'll never happen. Rockbox can never support iTunes music store music.
00:09:57preglowLlorean: unless they drop drm like steve jobs say they might
00:10:05mcphailLlorean: DRM'd music appears to be on the way out. There is hope yet
00:10:11bluebrotherRockbox isn't "cool" enough for the average end user
00:10:28Pepswell, I bought an ipod beacuse all other players in NZ of the same functionality are more expensive, and most non available. I was looking at alternatives for a long time
00:10:48scorchebluebrother: actually, that is the problem...they like us because we have DOOM!!!
00:10:56Peps\read what cory doctorow has to say about jobes and drm
00:10:56*scorche sighs
00:11:09Lloreanpreglow: If that ever happens, then yes, I suppose we could
00:11:22bluebrotherscorche: yes, and a lot of people install RB to get doom on their ipod and that's all. We really should consider removing doom
00:11:28bluebrotheror at least make it a paid option
00:11:31preglowmcphail: this might overlap slasheri's work
00:11:39LloreanPeps: On average here in the US, equivalently functional players don't exist. The iPod is fairly limited. More powerful players tend to cost less than it, though.
00:11:45scorcheand rockboy
00:11:47 Quit Ribs ("eh eh ehhhh!")
00:11:52mcphailpreglow: i haven't seen any patches for the 3g
00:12:00bluebrother"rockdoom is shareware. Please pay $$$ to the RB fund to unlock it"
00:12:05preglowmcphail: there is no patch yet, but he's going to commit it
00:12:16Lloreanmcphail: Also, it's not just for the 3G.
00:12:26Lloreanmcphail: It's general across the scrollwheels, I believe.
00:12:41preglowwell, it should be
00:13:08mcphailpreglow: Llorean: excellent. But the source code suggests the 3g scrollwheel is handled differently
00:13:08Pepsanother Q. Any chance of the cover art pathces and WPS stuff making it into the mainline? I got stung by several themes on the rockbox site that are using tags which require patched versions
00:13:27 Quit ender` (" It's tough to make predictions, especially about the future.")
00:13:34jhMikeSWhen amiconn starts yelling "slasheri!" it's in cvs :)
00:13:40preglowPeps: covert art: no, we want it done another way
00:13:48Lloreanmcphail: At a driver level, possibly, but from an application level they should both be treated the same. If they aren't, that's a problem.
00:14:02amiconnjhMikeS: The latest version I tested made things worse than svn for me
00:14:06pixelmajhMikeS: if it's only in CVS ;)
00:14:19preglowamiconn: i thought it worked nice. how did it make things worse?
00:14:27jhMikeStrue enough...still auto typing that lol
00:14:33amiconnIt was faster, but it also was incredibly jumpy and hard to get to where I want
00:15:03preglowmaybe he's done some more work since last time anyway
00:15:10amiconnYes, possibly
00:15:19preglowi'd love to see it in, the scrollwheel behaviour in rockbox annoys the hell out of me
00:15:22preglowit's laggy and slow
00:15:26PepsIt would help new users if the themes using non standard feature are clearly marked as such
00:15:26mcphailLlorean: there is a separate file for the 3g scrollwheel function. I don't know if that qualifies as driver or application level
00:15:45amiconnpreglow: I never experienced lagginess before cop support was committed :\
00:15:46preglowPeps: where did you find these themes?
00:15:53preglowamiconn: i did, all the time
00:16:03amiconnIt's just a little slow but it's reliable
00:16:13LloreanPeps: It's a wiki, which is user edited/created. Unfortunately that means it's user responsibility for the most part.
00:16:25preglowamiconn: i think it's somewhat reliable, but that doesn't mean it should be slow
00:16:40preglowamiconn: browsing a big list is sooooo much slower than it should be
00:16:49amiconnIt's a little slow but not too slow imho
00:16:54preglowit's far too slow
00:17:00bluebrotherPeps: there is also, which also list the required patches
00:17:10preglowamiconn: tried retailos?
00:17:13 Join Hoffmann [0] (
00:17:19preglowamiconn: that does it fairly well, if you ask me
00:17:25preglowamiconn: at least for nano
00:17:25Pepsthis is where I took thjem from, I think
00:17:30amiconnpreglow: Impossible to test long lists in retailos
00:17:39amiconn(even if I would put retailos back on)
00:17:53bluebrotherrockbox-themes is not a wiki. The WpsGallery page in the Rockbox wiki is ;-)
00:17:59LloreanPeps: Well, if it was, you need to talk to the administrator of that site.
00:18:10linuxstb_preglow: I agree, the retailos performance is what Rockbox needs to mimic.
00:18:12Pepswhich is?
00:18:19preglowamiconn: well, yeah, if you're not willing to put some tracks on it
00:18:25preglowlinuxstb_: agreed
00:18:36preglowit's fast when you want to be fast and precise when you want it to be that
00:18:41amiconnpreglow: I would need to put tracks on it *using itunes*, and I won't install that. Never.
00:18:48linuxstb_Although it's been a while since I used it - before we had sound in Rockbox.
00:18:50preglowamiconn: there are tons of alternatives to that
00:19:04LloreanPeps: Go to the site and look for contact information. It's a 3rd party site right now.
00:19:06amiconnBesides, there isn't enough space left on my mini to add enough tracks for a long list
00:19:16preglowyou could add small tracks :)
00:19:25linuxstb_Or empty files...
00:19:38bluebrotherPeps: look at the site for an address. He was around in this channel more frequently about a year ago but I haven't seen him for quite a while
00:19:45linuxstb_Maybe just putting a lot of files into the Notes directory would be enough to test the scrolling.
00:19:48amiconn...and in rockbox, the longest list is the all tracks listing in the db. Still bearable with rockbox svn scrollwheel handling
00:19:52bluebrotherhis nick was RedBreva
00:20:06scorchei have seen him around in the forums
00:20:21preglowamiconn: if you like scrolling through 5000 entries at that speed, you've got special tastes
00:20:24amiconn559 music files on my mini according to stats.rock
00:20:34PepsI don't see a contact anywhere there
00:22:19*linuxstb_ only has 1499 music files on his 60GB color...
00:22:45amiconn...and I can scroll the whole list in less than 30sec
00:23:05mcphailamiconn: /me has 2752 and needs a decen scrollwheel
00:23:43preglowamiconn: point is that scrolling thirty seconds to find a song in a list of 500 is abnormally slow
00:23:50mcphailamiconn: in "retailos" i can scroll through in a few seconds
00:23:51preglowit should be ton in ten seconds, max
00:24:12preglowspending two minutes to find a track among 2000 is just not acceptable at all
00:24:17amiconnIf it would scroll faster, how would I be able to spot the line where I want to stop?
00:24:27amiconnIt's already quite blurry
00:24:36preglowamiconn: you know how they're sorted, so you do a rough scan
00:24:45preglowamiconn: and slow down when you're close
00:24:48mcphailpreglow: i find the problem comes with bands named "The <whatever>". Getting to letter T is hideous
00:25:20PepsI think the ipod strips the "The". is there such a rockbox option?
00:25:26*amiconn wouldn't like the scrolling to skip entries
00:25:32 Nick JoeBorn is now known as JB_Away (
00:25:44mcphailamiconn: it _doesnt_ skip. That's why it is so nice.
00:25:46LloreanPeps: There's a patch for skipping 'the' in the database.
00:25:55amiconnThat's why I practically never use paged scrolling - I'd rather live with somewhat slower scrolling than overshoot
00:26:01preglowamiconn: i would, i don't need a precise seek always, i need a very fast seek until i see i'm close, then i slow down
00:26:04 Quit blueworm ("Leaving")
00:26:16preglowamiconn: there's absolutely no point in seeing all the entries between the start point and the song i want
00:26:31preglowi only need to see some entries to see where i am relative to where i want to be
00:26:56preglowamiconn: and if i for some reason want to see all the entries, i just scroll slower, which should also be possible
00:27:58amiconnIf you know exactly what I want - why go through ultra-long lists? Just go there via the fs browser
00:28:20*linuxstb_ creates some random filenames on his ipod, and discovers "con" and "aux" aren't valid filenames on FAT.
00:28:26 Quit CriamosAndy ("( :: NoNameScript 4.03 :: )")
00:28:45amiconnI see the all-tracks listing in the db as a way to get an overview - if I don't know exactly what I am looking for
00:28:55amiconn...and for this to work, I need to see all entries
00:28:58preglowamiconn: because everyone should be able to use rockbox the way they want without hacks like that
00:29:15Lloreanamiconn: Also, scrolling is much slower on the 5G than on your Mini, isn't it?
00:29:20mcphailamiconn: i have 428 artists. Getting a decent fs hierarchy would be a lot of work for little gain
00:29:45Lloreanmcphail: It'd take about 20 minutes with a decent tag tool, and you wouldn't even have to be at your computer while it ran.
00:30:02amiconnlinuxstb_: con, aux, lpt1...lpt3, com1..com8 etc are all reserved dos names
00:30:04preglowamiconn: can you think of a single good reason why _not_ to implement scrolling like i explain? the behaviour you want should still be possible
00:30:24mcphailLlorean: i shouldn't have to be at my comp at all
00:30:30amiconnpreglow: Well if it doesn't disturb slower scrolling I wouldn't be against it
00:30:48amiconnIt's just that I havent seen an implementation that doesn't
00:31:02preglowamiconn: scrolling speed should be entirely dependent on the speed at which your finger is scrolling
00:31:22preglowor nose, toe, or whatever you want to use for scrolling
00:31:27Lloreanmcphail: So, your statement is "I should gain without having to do anything"?
00:31:56 Join safetydan [0] (i=cbca159f@rockbox/developer/safetydan)
00:31:58mcphailLlorean: no, "I shouldn't lose for using the software"
00:32:01Lloreanmcphail: You can fix the problem by improving scrolling, which takes work, or you can fix the problem by setting up a proper tagnavi, which involves less work, or you can fix it by having a decent filesystem, which involves no work other than starting the tool.
00:32:29preglowlinuxstb_: yeah, creating files in Notes works
00:32:56amiconnpreglow: Afaik that is what Slasheri's patch does
00:33:06linuxstb_preglow: Yes, I've just tried it as well... But it takes a long time for the Notes viewer to scan the directory with 1000 files in it...
00:33:13Lloreanmcphail: But saying "I shouldn't have to do anything" is somewhat strange. You'll lose *something* for installing Rockbox no matter what, the question is solely if the gains outweigh the costs.
00:33:24Pepsbut please do not forget that it would be nice for some users to keep their ipod compatible with apple firmware as well
00:33:31preglowlinuxstb_: it took a long time with only 10 files in it
00:33:36preglowand this is flash, goddamnit
00:33:37|Rincewind|If anyone wants to test my record button patch, It now includes "Insert" in file browser.
00:33:39preglowstupid, stupid apple
00:33:52mcphailLlorean: exactly. so why compromise on a potential gain, by forcing people to organise files rather than using a scrollwheel properly?
00:34:04preglowusing scrollwheel properly is coming anyway
00:34:05Lloreanmcphail: Nobody's arguing against improving the scrollwheel.
00:34:11preglowwe're talking what you can do _now_
00:34:18Lloreanmcphail: But it's VERY important that it be done properly and well.
00:34:22 Quit entheh ("^~")
00:34:38Lloreanmcphail: There's absolutely no sense in rushing a temporary solution that will then have to be undone later.
00:34:38mcphailLlorean: and i'm not arguing about your preference to organise your files. But the software can and should do both
00:34:49Lloreanmcphail: Again, I never said it shouldn't.
00:34:54 Quit Hoffmann (Read error: 110 (Connection timed out))
00:35:19mcphailLlorean: i would look upon abandoning the database for a filesystem hierarchy as a temporary solution
00:35:22amiconnBut at least the version I tested felt really jumpy. (1) The cursor bar didn't stay at the same line in the browser while scrolling in one direction. (2) Scroll speed varied a lot even without changing finger speed (at least as perceived by me)
00:35:28 Quit midkay ("Leaving")
00:35:50amiconn(1) was probably just a bug. (2) might be solvable by silencing the speed value
00:36:00Lloreanmcphail: That's your prerogative, but it's a temporary solution that never has to be undone, and lasts forever. Unlike a temporary faster scrolling which would have to be removed if a proper faster scrolling were finally done.
00:36:01preglowamiconn: yeah, (1) annoyed me a lot as well
00:36:26amiconn(1) annoyed me far more than even a multi-minute scroll session with svn would
00:36:48Lloreanmcphail: Remember again, you're using unreleased software. The focus is on FINISHING the software, not on patching up a temporarily more comfortable version.
00:37:07safetydanpreglow, that sim tone controls patch looks fine to me. But then again, I've only read it, not tested it :)
00:37:10 Join barrywardell [0] (
00:37:18*safetydan wonders about moving the sim code to the target tree
00:37:19 Join Hoffmann [0] (
00:37:27preglowsafetydan: i've tested it on a fair amount of targets, and it seems to at least not crash any
00:37:30mcphailLlorean: patches are easily undone if a better solution is available
00:38:05preglowamiconn: retailos seems to use an acceleration scheme. the longer you keep scrolling at the same speed, the faster it does
00:38:10Lloreanmcphail: Not always, and also if the behaviour is 'better' but not implemented the right way, the incentive for someone to do so decreases drastically
00:38:20Lloreanmcphail: What exactly is the _advantage_ of doing things the wrong way?
00:38:23preglowamiconn: when you're over a certain speed threshold
00:38:29linuxstb_amiconn: I agree with you about Slasheri's first patches - I didn't like the feel of them either.
00:38:47linuxstb_preglow: And I can scroll from the start to the end of a 1000 file list in about 3 seconds.
00:38:53preglowi can scroll a thousand entries in 3-4 seconds
00:38:59preglowno blur
00:39:01preglowi always see where i am
00:39:14amiconnpreglow: The of definitely skips entries
00:39:15mcphailLlorean: it allows solutions to evolve
00:39:23preglowamiconn: indeed, but like i said, i want that
00:39:36Lloreanmcphail: That statement doesn't make sense.
00:39:56Lloreanmcphail: A hacked bad solution essentially endcaps it so that less people are likely to work on it, because it's 'good enough' for them.
00:40:10Lloreanmcphail: Patches can continue to be worked on without committing just fine.
00:40:13 Quit Peps ("ChatZilla 0.9.61 [Mozilla rv:1.7.12/20060216]")
00:40:55amiconnLlorean: Or even worse, the original author might stop working on it before it's really finished, and no one else understands it enough to continue the work efficiently
00:41:03preglowamiconn: using retailos, i can find any entry in about ten seconds, and this is the first couple of times i'm trying as well, so i'm not very skilled with it and overshoot a lot
00:41:04amiconnHappened more than once...
00:41:15Lloreanamiconn: And then you have to write a whole new system to replace it, yes. Ugh.
00:41:24linuxstb_I would like the acceleration to depend on both the size of the list, as well as the speed you're scrolling.
00:41:42Lloreanlinuxstb_: Somewhat similar to the fast-forward?
00:41:47linuxstb_i.e. no point jumping to paged mode if the list only has 20 items...
00:41:50linuxstb_Llorean: Exactly.
00:41:52mcphailWe're talking about code that is archived in svn... It doesn't take a genius to roll back
00:42:19mcphaillinuxstb_: the retailos version doesn't page at all
00:42:29Lloreanmcphail: That doesn't mean Rockbox shouldn't page.
00:42:38 Join Ranbee [0] (
00:42:41amiconnmcphail: It has to skip entries one way or the other
00:42:48linuxstb_mcphail: By "page" I mean jumping a large number of items at a time.
00:42:49mcphailLlorean: no, but it is an excellent and durable model to follow
00:43:06mcphaillinuxstb_: it doesn't jump at all. It just scrols faster
00:43:08Lloreanmcphail: And clearly files can evolve enough that rolling back to the original version could lose other valuable changes, or result in a file unusable in the current state of the code. It can, and likely will, take work to remove a change that isn't *needed* anyway.
00:43:17linuxstb_mcphail: No it doesn't, it pages.
00:43:24 Quit bluebrother ("night!")
00:43:24amiconnIf you can scroll a 5000 entries list in 3 seconds, that'd otherwise mean ~1600 lcd refreshes/second, which is plain impossible
00:43:27mcphaillinuxstb_: my 3g doesn't
00:43:27Ranbeehi, i want to try a video with rockbox, im going to re-encode it now to *.mpg. what size should i make it for my ipod video?
00:43:40preglowamiconn: of course, but i've never said it doesn't skip anything
00:43:44preglowi _want_ it to skip things
00:44:00amiconnpreglow: I was referring to mcphail
00:44:03linuxstb_Ranbee: 320x240 and probably no more than 10-15fps.
00:44:09 Quit muesli__ ("ich will Khe!!!")
00:44:12preglowah, ok
00:44:44linuxstb_mcphail: As amiconn says, the number of LCD refreshes required to scroll a long list in a couple of seconds is physically imposssible.
00:44:57Ranbeethanks, linuxstb_ . i looked at the PluginMpegplayer page and i didn't see a installation guide, so does that mean it's already installed?
00:45:17linuxstb_Ranbee: Yes, all plugins/viewers listed in the wiki come pre-installed.
00:45:20mcphaillinuxstb_: i couldn't scroll 5000 items in 3 seconds on my ipod
00:45:22LloreanRanbee: Official plugins never need installed.
00:45:36 Quit lee-qid ("aufwiederbyebientotsayonara")
00:45:49 Join JdGordon [0] (i=82c20d66@gateway/web/cgi-irc/
00:45:52Ranbeethanks linuxstb_ and Llorean :) i love rockbox so far
00:46:16linuxstb_mcphail: I can scroll a 1000 items in 3 seconds on my ipod Color. With 7 items on a page, that makes 142 frames per second.
00:46:16JdGordonstupid 3 hour prac only taking 90min :'(
00:46:36linuxstb_Sorry, that's makes 999 frames per second if you want smooth scrolling...
00:46:47linuxstb_I mean 333...
00:46:55*linuxstb_ can't count this late at night
00:47:06*amiconn wonders what linuxstb_ is calculating ;)
00:47:20*linuxstb_ wonders that as well.
00:47:35linuxstb_I'm just quoting random numbers.
00:48:03linuxstb_But anyway, on my Color, it is definitely skipping large numbers of items without displaying them - which is how it should work.
00:48:27*mcphail is off to grab his ipod and a stopwatch...
00:48:33JdGordonwithout trying to get tor working, does anyone know how to get svn to cooperate with a http firewall?
00:51:27preglowlinuxstb_: by the sound of the ticker, it definitely sounds like retailos uses discrete step sizes
00:52:08preglowthen again, the ticker might not be coupled too closely to the actual scrolling
00:53:19JdGordonpreglow: how do you feel about adding a setting like the start screen setting but for when the alaram triggers the boot?
00:53:31 Quit funky ("leaving")
00:53:56preglowJdGordon: would be great, but can't the default action be inferred from the start screen?
00:54:25preglowbut then again, what you want to do on wakeup might not be what you always want to do
00:55:12JdGordonI wouldnt want the alarm to start me in the file browser...
00:55:20 Quit atsea- (Remote closed the connection)
00:55:29JdGordonalthough, my start screen is the wps... but thats besides the point
00:55:35linuxstb_Only playback and FM would seem to be sensible options - until timed recordings are possible.
00:56:05 Quit jhMikeS (Read error: 104 (Connection reset by peer))
00:56:19preglowJdGordon: i'd pretty much say let "resume playback" be the default unless you have something like "radio" or "recording" which clearly infers something else
00:56:36 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
00:56:48linuxstb_Is the alarm a one-off alarm, or will it repeat?
00:56:55safetydanJdGordon, I would have thought SVN would respect the http_proxy environment variable
00:57:00preglowlinuxstb_: one-off
00:57:38*linuxstb_ has no target with both an FM radio and an RTC...
00:58:06LloreanThat would be the RTC-modded H120s, right?
00:58:31linuxstb_H300, X5, probably the Archos FM recorder,
00:58:40preglowJdGordon: and btw, don't know if you got it, but recording screen as startup screen now doesn't work
00:58:41linuxstb_Ondio FM?
00:58:44preglowJdGordon: possibly others too
00:58:54pixelmaOndioFM has no RTC
00:59:20JdGordonpreglow: that setting should have been removed, the start screen handles that now
00:59:31 Part Llorean
00:59:35markunJdGordon: did you get my PM?
00:59:37preglowJdGordon: that's what i'm talking about
00:59:53linuxstb_Maybe an alarm startup setting only makes sense if there is the choice between FM and file playback.
00:59:56pixelmaand I believe while the FM Recorder has an RTC it needs a mod for alarm but I don't _know_
01:00:03preglowJdGordon: when i set recording screen as start screen, nothing happens and the setting is forgotten when i start up
01:00:10linuxstb_So on ipods for example, just always make the alarm resume playback.
01:00:17JdGordonmarkun: I did, not identified so dont know if i can reply
01:00:19JdGordonbut thanks
01:00:40JdGordonpreglow: ok, ill fix that
01:01:04*mcphail may have to admit defeat on the scrolling issue...
01:01:20JdGordoncan anyone check their config.cfg file and let me know if any settings are repeated?
01:01:28 Join atsea- [0] (i=ariel@gateway/tor/x-d77f479b9c22cc36)
01:01:54jhMikeSJdGordon: I observed that once, will check which
01:02:05preglowJdGordon: disk spindown
01:02:19JdGordonok, we might have a bug then :p
01:02:48JdGordonhaha, no... just an extra line in the settings_list.c file
01:03:37jhMikeSJdGordon: Don't know if you saw but I lost voice settings that survived several reboots intact after a rolo. Will double check this though.
01:04:14markunJdGordon: and can you fix the buttons in the playlist viewer?
01:05:25markunJdGordon: thanks, sorry for giving you all this work :)
01:06:21JdGordonthats ok.. I can pass uni next semester :p
01:06:51scorcheJdGordon: you have a patch in your mail
01:07:44 Join stripwax_ [0] (
01:07:57JdGordonI'll do that one in the arvo, want to fix these other ones before I forget
01:08:21***Saving seen data "./dancer.seen"
01:09:12JdGordonafternoon... a few hour
01:09:32scorchealright...aussie slang i assume?
01:10:13JdGordonHAVE_RTC_ALARM right?
01:12:42 Part zeth
01:16:46 Part Ranbee
01:17:08 Quit robin0800 ("We be chillin - IceChat style")
01:17:42preglowamiconn: can you see something immediately stupid in the patch i posted?
01:18:23 Quit stripwax (Read error: 110 (Connection timed out))
01:18:26JdGordonpreglow: I want to add some #defines so I dont need to keep checking for TUNER and RECORDING for the alarm, wold they be OK in alarm_menu.h?
01:18:44 Quit stripwax_ ("Miranda IM! Smaller, Faster, Easier.")
01:19:30preglowJdGordon: *shrug*, i just hacked alarm support onto ipod, am not any kind of authority
01:19:44JdGordoncan we always record from the radio if both are present?
01:19:51preglowdoubt it
01:20:03preglowit's not an assumption which should be in apps/ code anyway
01:21:17JdGordonjhMikeS: HAVE_FMRADIO_IN means we can record from the radio right?
01:21:27JdGordonit wont be defined for some other reason?
01:22:51*linuxstb_ spots some IPL kernel activity - support for the 80GB
01:23:24jhMikeSJdGordon: ummm...yes
01:23:37jhMikeSthough AUDIO_SRC_FMRADIO can exist either way
01:24:14jhMikeSIf it is defined for another reason, it's a bug
01:26:47JdGordonoh fark.. I got confused, im doing this all wrong :'(
01:28:35 Part toffe82
01:29:29|Rincewind|jdGordon: do you think that my record button config patch has a chance to get commited and would you mind looking at it if there is something I can do to help get it commited? It is #5555 in Flyspray
01:30:00preglowdamn, that number alone is alone to merit a commit!
01:30:09|Rincewind|I think my code should be ready for inclusion now
01:30:28JdGordoni think we are fairly against customizable buttons, but I like the idea
01:30:37JdGordonwont be able to look at it for a while tho
01:30:55preglowi've heard plenty of people say they don't want customisable buttons, yes, but that doesn't change the fact that the record button is still unused
01:31:11JdGordonvery true
01:31:13|Rincewind|it gives the user only a few choices, so I don't see why it is "problem to support" in this case
01:31:50*jhMikeS needs to know if he's overlooking any other coldfire IRQs: DMA, timer, tick, sometimes PMU ... anything else?
01:33:59|Rincewind|we have a customizable wps, lots of options that change the way how rockbox works substantially, so one slightly customizable button shouldn't be such a big deal
01:34:35JdGordonpreglow: is it possible to test the alarm behavious in the sim? or can you try out a patch?
01:34:44preglowJdGordon: can try a patch, sure
01:35:01JdGordonok, ta
01:35:42JdGordonwhich targets actually have the alarm without a mod?
01:36:04|Rincewind|preglow: If you want to test my patch, it now has "insert" in file browser
01:36:04preglowall ipods and one of the recorders
01:36:08jhMikeS...and DISABLE_IRQ_LEVEL just doesn't right nicely
01:36:18preglowjhMikeS: none i can think of
01:36:35 Join midkay [0] (n=midkay@rockbox/developer/midkay)
01:36:35preglowJdGordon: why does it matter?
01:36:53JdGordonbecause i wanted to know which target to do a test compile on
01:37:01*jhMikeS is doing the IRQ shuffle so DMA can be synced with cause level 7 can't be disabled
01:38:35preglowJdGordon: just grep for HAVE_RTC_ALARM in firmware/export/config-*.h
01:40:12linuxstb_My two cents about customisable buttons is that yes, in general they're bad, but I can't see any harm in making one or maybe two buttons customisable on targets with lots of buttons. As long as they don't affect the core navigation.
01:40:15*linuxstb_ sleeps
01:40:29 Join hardeep [0] (i=c044e404@gateway/web/cgi-irc/
01:41:41|Rincewind|my patch doesn't affect core navigation at all
01:44:02 Join Thundercloud__ [0] (n=thunderc@
01:45:57JdGordonwoops... ok, take it again if you already got it
01:48:57preglowgotta be one of the noisiest patches i've ever tried, are you sure the fanciness you do around settings.h line 703 is ok?
01:49:14preglowi thought you couldn't do #ifdef LOLLERSKATES && something else here
01:49:32JdGordonyour right... bah
01:49:40JdGordonsettings.h is the only one right?
01:50:28JdGordonmy logic for all the messy #ifs is that if the rec or fm isnt present there is no need for the setting.. sound ok?
01:52:06Nico_Psafetydan: i've been trying to continue work on the wps tokenizer... did you ever attempt to integrate it into the existing wps code ? because i tried that and had some trouble
01:53:34JdGordonarking hell... I can get the ultimate edition of office 2k7 for $75 instead of $1,150! anyone want me to get them a copy?
01:54:51 Join l2eM1x [0] (n=l2eM1x@
01:55:02jhMikeS...SUPREME_IRQ_LEVEL ?? :P
01:59:35JdGordonpreglow: that patch work ok?
02:02:45 Join Llorean [0] (
02:03:40preglowJdGordon: waiting for wakeup
02:03:49ampleyflyJdGordon: halv that price, divide it by 10 then subtract the rest, and you've got what it costs to get
02:04:18JdGordonI know :D
02:06:38preglowJdGordon: woke up fine
02:06:40preglowstarted playing
02:06:41jhMikeSpreglow: WAKE UP!
02:06:43safetydanNico_P, no I never tried to integrate it.
02:06:51JdGordongreat :)
02:06:51 Part safetydan
02:06:53 Quit Thundercloud_ (Read error: 110 (Connection timed out))
02:07:16preglowJdGordon: didn't try recording, tho
02:07:32Nico_Pwow safetydan seemed in a hurry to leave...
02:07:54JdGordonpreglow: as long as the setting is there... it should work
02:08:03preglowJdGordon: but get rid of that noisy-ass compile problem, though
02:08:04ampleyflyoh god, I though I was in #ada, then i looked up and saw #ifdef LOLLERSKATES && and all was well
02:08:07JdGordonnicp: you brought up a touchy subject :D
02:08:10preglowi got a warning for every settings.h include
02:08:24JdGordonwhich was?
02:08:50l2eM1xlove the new menu guys
02:11:08JdGordonmarkun: I want to backout the MENU > previous screen commit from yesterday.. its causing those bugs you meniotned :'(
02:12:04JdGordonpreglow: i get no warnings here... was the the naughty #ifdef?
02:12:17 Part l2eM1x
02:13:26 Quit midkay (Read error: 104 (Connection reset by peer))
02:13:32 Join midkay [0] (n=midkay@rockbox/developer/midkay)
02:14:04preglowJdGordon: indeed
02:14:37JdGordonok, so then fixed...
02:14:41*JdGordon goes to commit
02:16:16|Rincewind|if you want to have recording start automatically on wake up alarm, you could use code from my patch
02:16:38|Rincewind|that part is only 4-5 lines
02:17:36JdGordonif (global_settings.alarm_wake_up_screen == ALARM_START_REC) ret_val = GO_tO__RECORDING
02:17:40JdGordon2 lines :D
02:17:56JdGordonexcpet, it doesnt have th typos in the code..
02:18:25|Rincewind|I have a look when I see your commit if you mean the same thing as me
02:20:46|Rincewind|what I mean is, it wakes up, goes to recording screen and automatically starts a recording
02:21:19JdGordonah, mine doesnt do that yet...
02:21:42|Rincewind|If you wait a few minutes I make a small patch
02:23:47 Join shnee_ [0] (
02:23:52JdGordonput it on FS and ill hcekc later.. im startving... yunch time
02:24:13scorchei thought you got a new keyboard
02:24:33JdGordonim at uni.. this kb is horrible
02:26:35ampleyflywhat keyboard did you get?
02:26:56JdGordonibm thinkpad r40 (laptop)
02:27:01JdGordonhurry up you sdtupd build table
02:27:21JdGordongreen.. awesome.. cya
02:27:42 Quit JdGordon ("CGI:IRC (EOF)")
02:28:46|Rincewind|damn, too late. I just finished it :(
02:33:20 Join bombi [0] (
02:33:51bombihi anybody here?
02:34:19 Quit shnee (Connection timed out)
02:35:16 Quit bombi (Client Quit)
02:37:10|Rincewind|jdGordon: You can find the Patch at
02:38:23 Join bombi [0] (
02:39:01bombigood day anybody in here?
02:39:34|Rincewind|if you have a question, just ask
02:42:21 Quit z0de (Read error: 104 (Connection reset by peer))
02:44:33 Join bombi_ [0] (
02:44:42 Join perldiver [0] (
02:45:27 Quit bombi (Read error: 104 (Connection reset by peer))
02:45:45 Quit bombi_ (Client Quit)
02:47:32 Join z0de [0] (
02:48:30 Quit barrywardell ()
02:49:38 Quit hardeep ("CGI:IRC (EOF)")
02:50:27 Quit inversions (Read error: 110 (Connection timed out))
02:51:16 Join Plouj [0] (
02:51:21 Join barrywardell [0] (
02:51:28Ploujis there a tool that checks rockbox code for problems automatically?
02:51:41preglowany tool that does so for c code should work
02:51:48Ploujor am I thinking of some other project
02:52:02 Quit barrywardell (Client Quit)
02:52:24 Join JdGordon [0] (i=82c20d69@gateway/web/cgi-irc/
02:52:43JdGordonthe library with the comics is closed :'( so your stuck with me
02:54:46JdGordon|Rincewind|: that patch works correctly?
02:55:27|Rincewind|I don't know if it works directly after startup, but it does work in my patch
02:55:48|Rincewind|I don't have RTC to test it
02:55:59 Part Plouj ("oh, never mind then")
02:56:10JdGordonit looks like it should :p
02:56:45JdGordonare you in tthe credits file?
02:56:45|Rincewind|not yet
02:57:36 Quit Domonoky_ (Read error: 104 (Connection reset by peer))
02:58:50 Quit pearldiver (Read error: 110 (Connection timed out))
02:59:09|Rincewind|maybe you can commit little parts of my patch one at a time and at one time we have record button config without amiconn realizing it until it is too late ^^
02:59:52hachione letter commits, buried inside false diff data
03:00:16hachioooh, mark the source files as 'binary' before committing, then it will just omit the diff output
03:00:32hachithen just change the content type when done
03:00:39JdGordonor not :p
03:01:01hachiif you ever need bad ideas, I'm your guy
03:01:23LloreanI still don't see why we don't just pick a good, static use for the record button and be done with it.
03:01:49 Join bombi [0] (
03:01:52JdGordonbecause we cant agree on one?
03:02:01|Rincewind|Llorean: because nobody can agree what is a good static use
03:02:06preglowwe can't agree on lots of things
03:02:08JdGordonwhy only have 180 settings when we can have 300!
03:02:18preglowsomeone just needs to commit a button map for it and be done :>
03:02:24Llorean|Rincewind|: Just because there's not a global consensus doesn't mean we can't pick one.
03:02:28bombican i play videofiles with rockbox on an ipdo video 30 gig ?
03:02:37LloreanLong Record, obviously, should be the recording 'shortcut'
03:02:53Lloreanbombi: Only MPEG1/2 files, see the PluginMpegplayer page in the wiki.
03:02:57|Rincewind|Llorean: sure, but I would really like to have it this year
03:03:02bombithx !
03:03:19LloreanThe main question is just 'what should short record do on various screens?'
03:03:23 Quit Nico_P (Remote closed the connection)
03:03:27JdGordonsomething different in each
03:03:31preglowqueue would be nice in the files/database screen
03:03:31LloreanAnd I think 'View Playlist' is pretty agreeable from the WPS
03:03:56LloreanAnd Queue or Insert is pretty agreeable for the filetree, the question pretty much seems to boil down to permanent or temporary addition to the playlist.
03:04:04|Rincewind|my patch does all that, what you suggest (and a little bit more)
03:04:05preglowi'd say temp
03:04:12LloreanI'm actually for "Queue" with "Right" becoming insert, instead of "Right" duplicating "Select"
03:04:21LloreanQueue for "Record"
03:04:31Llorean|Rincewind|: The problem is that yours add the configurability, doesn't it?
03:04:51|Rincewind|I'm all for insert, I never ever used queue
03:04:52JdGordonthats a _problem_?
03:05:02preglowwell, yes
03:05:09LloreanJdGordon: In general, we don't like configurable buttons. You know this.
03:05:26JdGordoni know
03:05:36bombisorry to ask again - does rockbox support some of the ipod toys like the FM transmitter etc ?
03:05:37 Join ryanpg [0] (n=ryanpg@
03:05:45|Rincewind|we have a configurable startup screen and nobody has something against that. Very strange opinions
03:05:51Lloreanbombi: Accessory support is very limited.
03:06:04Llorean|Rincewind|: Other than the word 'configurable' how do those relate at all?
03:06:08ryanpgyay! how exciting! the iaudio M5 is getting support! thanks!
03:06:30preglowbuttons are a bit more basic, aren't they
03:06:34preglowpeople don't expect them to be configurable
03:06:40|Rincewind|they both make a part of the UI configurable
03:07:08preglowi really don't mind "surplus" buttons being configurable, though, i think
03:07:09Llorean|Rincewind|: The option to record or play music on startup has been around for a long, long time. The configurable start screen basically just consolidated those two options and cleaned it up.
03:07:12|Rincewind|almost every linux or windows program has configurable keybord shortcuts
03:07:14preglowi just don't want all the buttons to be
03:07:16pixelma|Rincewind|: but you can tell the user easily how to change back (if you can name the buttons they have to use)
03:07:20LloreanI don't think "Record" is a surplus button.
03:07:41preglownah, probably isn't
03:07:49LloreanMaybe the 'A' button on the Gigabeat.
03:07:50hachiisn't it all configurable anyways? since the sourcecode is available, and I'm betting you guys use constants all over the place
03:07:53preglowit does have a rather specific meaning
03:08:25***Saving seen data "./dancer.seen"
03:08:38preglowbut then again, you'd never expect the record button to queue a track
03:08:53|Rincewind|pixelma: my patch doesn't change any existing buttons. It is only meant for buttons (or button combos) that are not used
03:09:00preglowand the fact that we're considering to make it do that makes it kind of surplus in my mind
03:09:00JdGordonhachi: yes, with the source it wont take long to redo the whole button config
03:09:13LloreanHonestly, I wouldn't mind giving the user the choice between "Insert" and "Queue" since in the end, those are almost identical behaviour.
03:09:32XavierGrjgordon: does your commit enable the rtc wake up features on H300 too?
03:09:36preglowXavierGr: no
03:09:43LloreanBut I don't think the button should be able to change to other functionalities.
03:09:45XavierGrthought so
03:09:51preglowXavierGr: bootloader needs fixing and the pcf drivers need to be merged before that happens
03:09:59preglowXavierGr: you'll surely notice that from the svn log
03:10:22XavierGrI knew about the pcf drivers but not about the bootloader
03:10:44preglowXavierGr: if the alarm were to work right now, the bootloader would switch itself off when it triggered, since it doesn't check for alarm currently
03:10:46XavierGrI thought Linus have made the rtc changes on bootloader and only the pcf driver was nedded
03:10:54preglowXavierGr: might be, if so i haven't noticed
03:11:18XavierGrwell I can't test it anyway, so I will have to wait for the pcf driver
03:11:26preglowshould be pretty easy
03:11:32 Quit Mouser_X (Nick collision from services.)
03:11:41 Join Mouser_X [0] (
03:12:07XavierGrjdgordon: there are still quite some quircks with "menu button to previous screen"
03:12:21XavierGrsome entries seem not to follow that behaviour
03:12:30bombiDo you know if the ipot FM Transmitter is supported?
03:12:42 Quit JdGordon ("CGI:IRC (EOF)")
03:12:47XavierGre.g the Playlist menu entry
03:13:16XavierGrgah he left
03:13:29XavierGrand I just reproduced a bug that is related to it that crashed my H300
03:13:32ryanpghas anyone here tried the M5 rockbox port? as it's almost the same as the X5 I'm hoping it would be safe to test it...
03:13:53ryanpgsomeone here told me the X5 is the best current production platform for rockbox
03:14:12XavierGrJdGordon: (if you see the logs) try the following: with an uninitialized database, move into the database menu then push 2 times the Main menu button
03:14:20Lloreanryanpg: The M5 port should work fine at this point.
03:14:34XavierGrjdgordon: it should show the "database is not ready" splash box and the player will hang
03:15:07ryanpgLlorean, thanks... that means it's not likely that I'll loose my entire music collection, right?
03:15:46Lloreanryanpg: As far as I'm aware, there's never been a case of an official build of Rockbox destroying a music collection.
03:15:59 Join JdGordon [0] (i=82c20d69@gateway/web/cgi-irc/
03:16:20 Join pearldiver [0] (
03:16:26ryanpgLlorean, awesome, well time to grab my M5 and start the fun then :)
03:16:29JdGordonXavierGr: yeah, Im thinking about reverting the commit... something isnt wrking properly
03:17:45ryanpgLlorean, I was going to try the nightly build
03:17:52Lloreanryanpg: Why not the 'current' build?
03:18:27ryanpgLlorean, dunno, if you're suggesting that then ok
03:18:36*ryanpg tends to like the bleeding edge
03:19:06JdGordoncurrent build is more bleeding edge than the nightlies
03:19:29Lloreanryanpg: The 'current' build is the bleeding edge. That's why it's 'current' as opposed to 'as much as 23.9999 hours of of date' as the daily might be.
03:19:51ryanpgLlorean, k, downloaded and copying over now
03:20:02ryanpgLlorean, thanks for the hand-holding btw :P
03:20:11ryanpgI'm pretty psyched :D
03:20:37LloreanRockbox should work quite well for you.
03:20:46 Quit JdGordon (Client Quit)
03:21:00LloreanAs a vaguely related note though, portable HD based players aren't that reliable for storage in the long term and it's never really a good idea not to have a backup of your music somewhere else.
03:21:07ryanpgLlorean, yeah I played with it on an iPod, but I reverted back, I think it'll really enhance the MF though
03:21:27ryanpgLlorean, good point about the backup, too bad I don't have a DVD burner, I've got quite a large collection of course
03:21:33 Quit perldiver (Read error: 110 (Connection timed out))
03:22:32|Rincewind|ryanpg: If your music collection is only on your portable, then I would suggest to get a larger HDD as soon as possible
03:23:09 Join perldiver [0] (
03:24:19ryanpg|Rincewind|, wise advice... I do have one thing going for me, most of my music is from emusic and I can always re-download
03:24:43ryanpgk, in my nerdy geeked out excitement I just forgot to unmount the m5 before unplugging it
03:25:21*|Rincewind| won't distract ryanpg again
03:25:34|Rincewind|good night all!
03:26:10 Quit |Rincewind| ("Cya")
03:26:22ryanpgheh, I'll brick this thing yet
03:27:19 Quit XavierGr (Nick collision from services.)
03:27:22 Join XavierGr [0] (
03:28:27ryanpgkickass... works
03:28:32ryanpgtime to explore
03:30:16 Join tkloppel [0] (
03:31:42tkloppelhey, anyone here usin rockbox on a gigabeat f20?
03:32:12Lloreantkloppel: Did you have a specific question?
03:32:57tkloppelLlorean, yeah... Im wondering what the requirements are for video files. I know it's still very unsupported...
03:33:17SoapX5 can't be the best port until it Dual Boots, or supports everything the original firmware did, like USBOTG
03:33:40Lloreantkloppel: See the PluginMpegplayer page in the wiki.
03:34:05 Join tvelocity [0] (
03:34:09LloreanSoap: Well, I think it was with the qualification of 'on still produced / available new hardware'
03:34:20LloreanThough that's iffy with the Gigabeat around, depending on what the user needs from the hardware.
03:34:20pixelmaSoap: but the USBotg-support is not an M5 issue...
03:34:31Soappixelma: touche
03:34:39 Join jhulst [0] (
03:34:44tkloppelLlorean, thanks.
03:35:25 Quit pearldiver (Read error: 110 (Connection timed out))
03:39:32ryanpgthis is really great, love the games
03:39:41ryanpgI don't see any M5 specific themes though
03:40:15ryanpgwould it be safe to try an X5 theme? the M5 has monocrome display where the X5 is color
03:40:32pixelmaryanpg: since the display has the same resolution as the H100/ 4G greyscale Ipod you can use those
03:40:53ryanpgpixelma, H100/4G Ipod, cool will browse those
03:40:55pixelmaX5 themes _should_ work too
03:41:11 Quit tkloppel ("Chatzilla 0.9.77 [Firefox]")
03:41:47LloreanBut because of the color choices, how readable they are will depend on which shades end up which grays.
03:49:52 Quit shnee_ ("Konversation terminated!")
03:50:26ryanpgsearching the docs for gapless playback
03:50:57Soapit isn't an option.
03:51:17SoapIf you have ripped and encoded properly it will work.
03:51:38Soapthere is a wiki page which discusses ripping and encoding properly, if I recall correctly, one sec.
03:52:29pixelmamarkun just added it today
03:55:23 Join shnee [0] (
03:57:00ryanpgtoo bad, it's the one feature the rio karma has over rb
03:57:16Soapwhat feature?
03:58:19scorcheyou dont need the nogap for gapless....only for archos is that needed
03:58:40ryanpggapless playback of any format
03:58:51scorcherockbox has that..
03:59:12Soapif the karma could make anything "gapless" it was faking it.
03:59:24Soapyou can fake it too with rockbox if that is what you want.
03:59:29 Quit XavierGr (Nick collision from services.)
03:59:32 Join XavierGr [0] (
03:59:33Soapjust do a short crossfade.
04:00:30SoapBut, (oh. my. god. I'm like of of /them/. Posting one thought per line. I will stop) But Rockbox will play gapless any audio codec which can be played back gapless without faking it.
04:00:31ryanpghaven't gotten to crossfade yet in the manual
04:00:37 Join Pavel_ [0] (
04:00:52Pavel_can somebody help me put linux on my ipod
04:01:21Lloreanryanpg: Rockbox plays true gapless. That means it plays exactly how much gap is in the file, under the assumption that you've properly encoded your files and thus any gaps are intentional.
04:01:30Lloreanryanpg: Playing all files without gaps, even when they're supposed to have them, would be silly.
04:01:42LloreanPavel_: This is the wrong channel for that.
04:01:59ryanpgLlorean, yes I understand... I'm sure a lot of debate is generated by the topic
04:02:02Pavel_what channel do i go to?
04:02:26Pavel_whats the difference between ipod linux and rockbox firmware
04:02:27Soapryanpg: If you want to remove intentional gaps, crossfade is what you want.
04:02:50Lloreanryanpg: Not really. I've not yet met someone who thinks it's a good idea to artificially remove gaps that are supposed to be there.
04:02:55SoapPavel_: IPL is linux. Rockbox is a music player.
04:03:02ryanpgSoap, ok I'll dig through settings with an eye for "crossfade"
04:03:05 Join Seed [0] (
04:03:24Pavel_what?? i thought you put rockbox on your ipod
04:03:35LloreanPavel_: You do put Rockbox on your iPod.
04:03:44Pavel_so what does it do?
04:03:53LloreanPavel_: But Rockbox is not linux, is not based on linux, and is completely different from iPod Linux.
04:04:02LloreanPavel_: Rockbox is a music playing firmware.
04:04:22LloreanIt plays music, and incorporates a very large range of options relating to music playback functionality and format support.
04:04:36Pavel_can u have both
04:04:56Pavel_cool thx
04:05:57 Nick Pavel_ is now known as linuxipod (
04:06:30 Quit linuxipod ("Chatzilla 0.9.77 [Firefox]")
04:06:44 Join linuxipod [0] (
04:07:18ryanpgbummer... rb doesn't fix the annoying bug where umounting on linux doesn't return the M5 to play mode
04:07:59Soapwhat happens to your M5 when you unmount?
04:08:08Soap(with Rockbox)
04:08:57ryanpgnothing... the big USB icon remains on the screen
04:09:06ryanpguntil I yank the usb cable, then it shuts off
04:09:19LloreanThat's probably the USB hardware itself then.
04:09:20ryanpg(I check to be sure it's not mounted before the "yank" part)
04:09:34pixelmaryanpg: which USB port are you connecting it to?
04:10:28 Quit linuxipod (Client Quit)
04:10:46ryanpgpixelma, err... well there's only one on the M5, and I'm connecting to one of two on the laptop
04:11:02ryanpgpixelma, do you want some specific kind of information?
04:11:24pixelmano... you would have the possibility to connect via the subpack too (if you have one)
04:11:26ryanpglsusb can provide some details
04:11:37ryanpgahh... forgot the M5 takes a subpack
04:12:09*ryanpg actually tried to reverse engineer a subpack
04:12:19pixelmaI just try to get some information for someone who could look into it (someone with a clue ;) )
04:12:38ryanpgheh, well I guess that's not going to be me then
04:12:50pixelmait's a fairly new port and these specific things aren't known yet, I think
04:13:11 Join zorroviejo [0] (i=40d4cdb8@gateway/web/cgi-irc/
04:13:11LloreanI thought most of the players were pretty much dependent on the USB hardware itself to say it was done.
04:13:18ryanpgwell, it occurred even with the official iaudio firmware
04:13:19zorroviejohi people
04:13:24bombiwhat mp3 player whould you guys recommend? is the apple ipod video 30 gig a good choice?
04:14:04zorroviejoi play any mp3 files en rockbox but soy no codec in screen
04:14:15Lloreanbombi: The Rockbox port on iPod isn't the best. There's still a lot of things that don't work well (poor performance and poor battery life) because of unknown hardware.
04:14:28 Join shnee_ [0] (
04:14:36zorroviejoi have ipod video 30gb
04:14:48Lloreanzorroviejo: Has it ever worked?
04:14:49pixelmaLlorean: just thought that there might be a workaround later if that's a general problem - but not that there will be
04:15:52Lloreanpixelma: I'm pretty sure that Rockbox has to just wait for the USB->ATA bridge to stop being busy. But I don't really know.
04:16:26bombithx llorean! what machine do you have?
04:16:59Lloreanbombi: Archos Jukebox Recorder, iHP-120, iPod Nano, Toshiba Gigabeat, iFP-799 and Sandisk Sansa e200. :)
04:17:06LloreanThe last one doesn't play music yet, though
04:18:01bombiargh - i'm backpacking since 2 years and have now enough money to buy an mp3 player, but i left my geeklife behind and so I'm not realy up to date with all the tec stuff
04:18:16 Join SirFunk [0] (
04:18:16ryanpgsheesh, with rb out there, I can't even imagine why hardware manufacturers even bother with writing OSs
04:18:58ryanpgI mean, seems like they could 1) fund development 2) hire developers 3) release open spec devices and let the community give them an OS
04:19:32Lloreanbombi: If you just want decent battery life, good storage, and audio playback, look at the Gigabeat F series.
04:19:38 Join lukaswayne9 [0] (
04:19:38bombiso do you say that RB effects the batterylife of certain players?
04:19:38bombie.g. ipod?
04:19:57 Quit lukaswayne9 (Read error: 104 (Connection reset by peer))
04:20:13Lloreanbombi: On well supported players, it extends battery life by doing things more efficiently. On the iPod we just don't have enough hardware knowledge, there's a high probability we've not properly initialized or suspended certain hardware.
04:20:20 Quit Seedy (Read error: 110 (Connection timed out))
04:20:32Lloreanbombi: Whatever it is, most iPods get 2/3 to 1/2 of the regular battery life while running Rockbox.
04:22:01 Quit zorroviejo ("CGI:IRC (EOF)")
04:22:05bombiwell thats not realy what i want - gigabeat F series you say . all right thank you very much!
04:22:35bombididnt consider gigabeat yet. didnt like it with running windows on it ....
04:22:57SoapLlorean: until I have my gigabeat in my hands - stop driving up the prices!
04:24:07 Join zorroviejo [0] (i=40d4cdb8@gateway/web/cgi-irc/
04:24:49 Join hc1 [0] (n=agashlin@nat01-silvers-ext.Rutgers.EDU)
04:25:03SoapSoon there will be proper runtime tests on Gigabeat! Stock vs Rockbox! In a televised Lithium Cage Match!
04:25:04 Quit hc1 (Client Quit)
04:25:48LloreanWe do need a proper discharge curve for Gigabeat.
04:26:05LloreanI can play mine for 3 hours and go from 100% to about 96%, but if I play it for 8 hours, it's down around 67%.
04:26:09bombiwanted to buy iAudo x5l but cant find it here in australia
04:26:56SoapWhat do you need to create a discharge curve? Lots of battery_bench.txts or something else?
04:27:09LloreanI don't really know.
04:27:27 Quit zorroviejo (Client Quit)
04:27:43LloreanI've never been _that_ interested in the battery. If it lasts long enough to get me through the day, it's 'good enough', and I don't really look at how much I have left most times. :)
04:28:00LloreanThough I do share the frustration of people who complain about their battery life, but refuse to compare it to the original firmware.
04:29:30SoapIt is the first thing I plan on doing, since there is no proper dual-boot function.
04:29:30pixelmathe battery_benches on the GigabeatRuntime page seem to be all made with different settings etc...
04:30:07bombithe gigabeat F series is quite old isnt it?
04:30:08bombifair enough
04:30:25Lloreanbombi: It's not that old.
04:30:42 Quit shnee (Connection timed out)
04:30:48Lloreanpixelma: On almost any of the battery bench pages there is a wide variety in settings.
04:31:09LloreanSoap: An effort is made, yes.
04:31:40LloreanIdeally one day someone will get even more fed up and put up a playlist and .cfg file to test with. :-P
04:31:48bombitrying to find out more about gigabeat f series. sitting on a very slow inet connection so takes me some time
04:31:51pixelmaI just thought it would help especially for calibrating if there were a few that were made under nearly the same conditions
04:32:34SoapIf those pesty developers stop mucking with the config version maybe someone would! ;)
04:33:18LloreanWell, the default settings aren't all bad. The config could just be a single line turning repeat on. :-P
04:33:27bombihow comes that there is no RB 4 creative at all?
04:33:39Lloreanbombi: Because nobody who owns one has done it.
04:33:57 Quit Thundercloud__ (Remote closed the connection)
04:34:02Lloreanbombi: Rockbox happens when someone or a group of someones who own the hardware, and is interested enough, does the work to port it.
04:34:09pixelmaLlorean: I seldom use repeat... actually almost only for battery_benches! :P
04:34:12Soap(perhaps says something about Creative owners?)
04:35:00bombitrue - maybe creative software is already good enough
04:35:11Lloreanpixelma: Other than battery benches, I do have one long playlist that I playback shuffled since it's all singles anyway.
04:35:35Lloreanbombi: No, there's definitely a group of them that things Rockbox is better.
04:36:06LloreanThey just, so far, have done a lot of talking, and a decent amount of "I don't really want to open up _my_ player" and a little bit of "Well, I don't really know C, but I'll help however else I can"
04:36:21 Quit pixelma (" bedtime")
04:36:39LloreanUnfortunately they really need someone who knows C, and probably the appropriate assembly, to get started.
04:37:55bombiwell that shouldnt be to hard.
04:37:56bombiargh - missing the times at home where i was able to do stuff like that.
04:39:42 Join bombi_ [0] (
04:43:05bombi_I wanted to say that the RB project is pritty cool !
04:44:33ryanpgg'night all
04:44:36 Quit ryanpg ("Leaving")
04:45:21 Join aliask [0] (
04:49:03 Quit perplexity_ (Read error: 104 (Connection reset by peer))
04:49:30 Join perplexity_ [0] (
04:51:47 Join Brunellus [0] (n=luigi@unaffiliated/brunellus)
04:52:45 Part Brunellus ("Ex-Chat")
04:53:44jhulstTo be fair for the ZVM port, there are a few us who know programming and are going to try and work on it
04:56:49Lloreanjhulst: I just meant overall. People have been saying there 'should' be a ZVM port for quite some time now.
04:57:20Lloreanjhulst: In the amount of time people have been saying there should be one, not much has gotten done yet. I didn't mean to speak unkindly of what could be done in the future though.
04:57:28jhulstLlorean: Understood, hopefully that will change
04:57:37LloreanI hope so.
05:07:41 Quit bombi (Read error: 110 (Connection timed out))
05:08:26***Saving seen data "./dancer.seen"
05:15:35 Quit bombi_ (Read error: 110 (Connection timed out))
05:18:47 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
05:19:15JdGordonhey aliask
05:19:33aliaskhey there jd
05:19:49JdGordonwhy havnt you been on irc?? I had 4 hours to kill... i was so bored :p
05:20:09aliaskI did the first half of my driving test today, 57% pass :D
05:20:45aliaskI bet you cant wait till they let me drive.
05:20:48 Quit Hoffmann (Read error: 110 (Connection timed out))
05:21:46JdGordonwas the the persption or driving test?
05:21:57aliaskhazard perception test
05:22:03aliaskShit test anyway.
05:22:05JdGordonha, thats utter BS
05:22:09JdGordonwhens the drive?
05:22:26aliask1st of may, they're booked up quite a while at the moment cos of the changing laws
05:23:56 Join ryanpg [0] (
05:24:10ryanpghi again, just real quick before I run off to bed... where do I find crossfade settings?
05:24:49JdGordonunder sound settings
05:25:19ryanpgJdGordon, k
05:26:04ryanpgi see crossfeed but not crossfade
05:26:36ryanpgnm, found it
05:26:48JdGordonoh.. you wanted fade.. ye in playback
05:26:55ryanpgGeneral Settings -> Playback -> Crossfade
05:27:08ryanpgthanks, g'night again :)
05:27:11 Quit ryanpg (Client Quit)
05:32:05JdGordonnice to see the deutsche post website is as bad as ausposts '
05:34:43 Part Llorean
05:37:57aliaskHrm, an X60 just came up on ebay... tempting
05:39:34JdGordonstill tihnking of getting rid of the S?
05:40:54aliaskYeah. I'd be more than happy to sell it to a developer for mates rates
05:47:37aliaskHowever personally I think it might be a bit of a dead end. There are a LOT of obstacles.
05:50:24JdGordonno such thing as a dead end
06:00:29 Quit bonbonthejon (Read error: 54 (Connection reset by peer))
06:01:02 Join bonbonthejon [0] (
06:06:01axion1aliask, so if the light goes red it means floor it?
06:06:40aliaskand always turn into oncoming traffic.
06:11:37 Quit desowin ("use linux")
06:13:09aleskawhat do I do to get rockbox to recreate the root playlist root.m3u8, which includes all songs, now that I've changed the music that is loaded. Does it happen automatically?
06:14:19 Join linuxstb__ [0] (
06:14:51 Quit Mouser_X (Nick collision from services.)
06:14:59 Join Mouser_X [0] (
06:15:46aliaskaleska: No, you'll have to create a new playlist and save it again.
06:19:24aleskaaliask: thx
06:21:45 Join Aaron4 [0] (n=sup@
06:25:20axion1if I can figure out how the heck to save it
06:25:35axion1on my x5 heh. I have to create it in winamp and copy it to the player
06:26:29aliaskIt's probably faster that way too
06:26:47aliaskIt's in playlist options though so you know for next time.
06:28:37 Quit linuxstb_ (Read error: 110 (Connection timed out))
06:29:28axion1yeah it's easy. WHen im done creating the playlist in winamp I just change D:\mp3\ to \music\ using find and replace then resave
06:29:32axion1after im done I copy to the player
06:29:35 Part tvelocity ("Αποχώρησε")
06:30:00aliaskYep, I did that too when I had rockbox.
06:30:36aliaskThere might actually be a way to get the database to make a playlist with all songs for you, but I never used the database then.
06:43:38 Quit toer (Remote closed the connection)
06:44:49JdGordonyou can make a playlist with every song on your dap easily
06:45:35axion1it can creatre a root playlist but I can never access it.
06:45:47JdGordonwhy cant you access it?
06:45:53*JdGordon cant find the damn option :p
06:46:29axion1unless it's in the files section. Only in there do I go to my playlist I created in winamp
06:47:27 Quit Aaron4 ()
06:49:02 Quit rotator ()
06:59:07 Join Lear [0] (i=chatzill@rockbox/developer/lear)
07:01:26 Quit webguest32 ("CGI:IRC (Ping timeout)")
07:05:58 Quit Mouser_X (Nick collision from services.)
07:06:09 Join Mouser_X [0] (
07:07:12 Join webguest80 [0] (i=d554be96@gateway/web/cgi-irc/
07:08:28***Saving seen data "./dancer.seen"
07:16:55 Quit Mouser_X (Nick collision from services.)
07:17:04 Join Mouser_X [0] (
07:19:28 Join webguest60 [0] (i=80c211c7@gateway/web/cgi-irc/
07:22:04 Quit webguest60 (Client Quit)
07:22:16 Join kaaloo [0] (
07:22:27 Quit Mouser_X (Nick collision from services.)
07:22:38 Join Mouser_X [0] (
07:23:18 Join toer [0] (
07:25:40 Quit Mouser_X (Nick collision from services.)
07:25:49 Join Mouser_X [0] (n=mouser_x@
07:29:55 Quit kaaloo ("Leaving.")
07:32:19 Quit Mouser_X (Nick collision from services.)
07:32:28 Join Mouser_X [0] (
07:46:23JdGordon:'( I cant get this stupid menu to work properly
07:46:31 Join Shaid [0] (
07:49:57LearWhat menu?
07:50:16hcsSystem -> Tools -> Stupid
07:51:18JdGordonthe root one... i keeep getting stuck in a wps > STOP > wps loop
07:53:30LearYes, looks like GO_TO_PREVIOUS could be a problem...
07:54:25JdGordonyeah, but any changes to that causes the MENU button to not work correctly
07:54:33JdGordoni.e goto the previous screen from the root menu
07:56:32LearHm.. Easy to reproduce, yes.
07:58:05 Join lee-qid [0] (
08:00:21 Join perplexity [0] (n=joust@
08:02:52*amiconn wonders why the alarm wakeup fix increases codse size on targets with no rtc
08:03:21amiconnJdGordon: Btw, making the menu entry for fm radio only depend on CONFIG_TUNER isn't enough
08:03:42amiconnThis entry needs to be dynamic depending on whether the tuner is actually detected
08:04:08amiconnOtherwise you confuse *most* V2 recorder users and *all* X5V users
08:04:09*JdGordon wonders also
08:04:36JdGordonthe entry in the alarm screen menu?
08:05:35amiconnHmm, the size increase is just due to the .lang string
08:06:10JdGordonhmm... :'( the setting screen doesnt handle non−−const item numbers...
08:07:52LearWhat about going to root if "GO_TO_PREVIOUS" would return to the same screen? Seems to fix the WPS loop at least, though it might break other stuff...
08:09:13 Quit perplexity ("* sigh *")
08:10:47JdGordonamiconn: I'm not sure how to fix this... I agree its a farily big problem tho
08:11:45JdGordonyes I do... we'll just have to call do_setting() manually :(
08:12:51LearOh well, off to work...
08:13:12 Quit Lear ("Chatzilla 0.9.77 [Firefox]")
08:21:14 Quit webguest80 ("CGI:IRC (EOF)")
08:21:30 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
08:23:59JdGordonthe radio really is a PITA sometimes :p
08:25:02 Join Llorean [0] (n=Llorean@rockbox/administrator/Llorean)
08:27:01 Join Rob222241 [0] (
08:27:13 Quit lee-qid (Read error: 110 (Connection timed out))
08:34:21jhMikeSamiconn: does M5 use the same keymap as x5?
08:34:58jhMikeSsame remote too?
08:36:30JdGordonamiconn: ok, fixed
08:36:41amiconnThe remote is the very same for M3 (where it's standard, because it's the only display), M5 and X5
08:37:50jhMikeSI'm just adding a keymap for the keyboard on x5 so I guess it'll make sense for all then
08:39:01 Join petur [0] (i=d4efd6a6@gateway/web/cgi-irc/
08:40:53 Join webguest40 [0] (i=c023111e@gateway/web/cgi-irc/
08:41:05 Join bluebrother [0] (i=gl8YZaNe@rockbox/staff/bluebrother)
08:41:18amiconnThe button driver is slightly different though, because Play and Rec are swapped electronically. I wonder why cowon did that
08:42:10jhMikeS?? Routing lines perhaps?
08:42:20peturengineering error probably ;)
08:45:12 Quit Rob2222 (Read error: 110 (Connection timed out))
08:45:14*petur now has his 80GB disk and 2300mAh battery, just waiting for spare time to assemble
08:45:43 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
08:45:58aliaskpetur: Where'd you get the 80gb disk from? 1.8" i'm assuming...
08:46:12*amiconn already has a "H180" :)
08:46:25amiconnaliask: I got mine from ebay
08:46:25peture-bay... ICC I think they're called... $260
08:46:37amiconnEUR 156
08:47:52peturthere's an almost new one from Canada now:
08:48:11petur'as is'
08:48:58jhMikeSWell, I mapped it ff=right, rw=left, volup=down, voldown=up, repeat play=ok, mode=cancel, rec=select/backspace depending on mode
08:49:37 Join Ribs [0] (n=ribs@
08:50:11amiconnWhy not select=select?
08:50:58jhMikeSworks faster...natigate with one hand, select chars with the other
08:51:07jhMikeSnavigate even
08:51:45*amiconn would much prefer select to be select
08:52:02amiconnI value one-handed operation
08:52:07LinusNme too
08:52:33LinusNit is a pain to have to use both hands
08:52:47amiconnIt could be both select and rec unless we have a better use for one of them
08:52:48 Join HBK [0] (i=4a46cb46@gateway/web/cgi-irc/
08:52:56jhMikeSplay is usually ok in the standard context
08:53:12HBKI got a problem with my iPod 5G
08:53:17amiconnAnd why long Play = accept? long something should only be used if short something is needed for something else
08:53:19HBKanyone can help?
08:53:30LloreanHBK: We won't know until you've asked the question. ;)
08:53:33KasperleHBK: not until you state the problem
08:53:34jhMikeSwhat should be ok? menu or could that be select? it's not assigned to anything at all
08:54:18HBKoh yeah,I just installed Rockbox on my 30g ipod video,everything goes fine until i tried to play mp3 on it
08:54:19jhMikeSthere's no assignment for the remote "menu" button
08:54:32amiconnAh, you're talking about the remote
08:54:39*amiconn somehow didn't get this
08:54:48jhMikeSthat's why I was asking about it :)
08:54:50HBKwhen i started to play an mp3 file,it said data abort 000307F8
08:55:11HBKthen I had to restart it,but the same thing again
08:55:26 Join ender` [0] (
08:55:31jhMikeSI wanted to keep "play" as ok since that's the standard button
08:55:41LloreanHBK: What color is the background during the bootloader portion (the words right when you turn it on)?
08:55:45amiconnYes, but why long?
08:56:18jhMikeSto help avoid a mistake in case it gets pressed when navigating
08:56:19HBKlight blue I think
08:56:27LloreanHBK: You need to upgrade to the new bootloader then.
08:56:43*amiconn is a bit surprised that -wf'ing the bootloader doesn't actually save much of the boot time
08:56:53amiconnIt feels more than it actually is
08:56:54LloreanHBK: Follow the instructions here for installing the bootloader:
08:57:13HBKAh,I download the one from CVS session,but dunno if it's the lates one
08:57:13*amiconn tested of boot with the latest bootloader
08:57:18amiconnWorks fine
08:57:20LinusNLlorean: i like how we have adapted the ipod bootloader version numbering scheme to that of Pokmon :-)
08:57:34LloreanLinusN: The bootloader has a numbering scheme?
08:57:47LloreanOh, the color
08:57:50LloreanBootloader Black
08:58:36HBKI'm at the page of
08:58:58LloreanHBK: I gave you a link to what page and everything. Please follow the directions at the page I sent the link to.
08:59:19HBKyeah,I followed that instruction
08:59:44 Join pondlife [0] (
09:00:16HBKI think the one I got is the latest one from that page
09:00:21pondlifejhMikeS: How's H300 recording today?
09:00:34LloreanHBK: The page I gave you has never installed one with a light blue background.
09:00:42LloreanHBK: Are you sure you didn't follow the IpodInstallation page instead?
09:01:18peturpondlife: quite ok
09:01:34pondlifeHmm, I'm getting a reproducible crash on exit now
09:01:54HBKthe version is 12652-070306,is this the latest one
09:02:08LloreanHBK: I told you, the blue background is not the latest one.
09:02:15jhMikeSpondlife: I'm gonna vanquish h300s from existence then
09:02:18LloreanHBK: The version number confirms it. Is there some reason you won't believe me?
09:02:37LloreanHBK: Wait misread the version number
09:02:43LloreanHBK: Is that your BOOTLOADER version or your ROCKBOX version?
09:02:43pondlifejhMikeS: Sorry to bear bad news, just wondered if you considered that finished or not.
09:02:58LloreanHBK: Specifically, what screen is that number from?
09:03:02GodEater_070306? that's ancient!
09:03:13LloreanGodEater_: 2007, Mar, 06
09:03:22GodEater_ah - wrong way round
09:03:25jhMikeSThat bug's fixed if the screen isn't slowing anymore
09:03:25LloreanYeah, I did that too.
09:03:26GodEater_I forgot - we do iso
09:03:37*GodEater_ goes and sits at the back of the class
09:03:50LloreanGodEater_: Still, the bootloader at the page I linked doesn't display text.
09:03:54pondlifejhMikeS: The slowing has gone
09:04:00pondlifeAnd I can record ok.
09:04:03jhMikeSbut you must be screwing with me about this crash on exit hoax
09:04:05HBKit's at the bottom of the version menu
09:04:14pondlifeIt locks when I attempt to exit the recording screen
09:04:21pondlifeAfter recording only
09:04:22LloreanHBK: That's your Rockbox version.
09:04:31LloreanHBK: I told you you needed the newest Bootloader, not the newest Rockbox build.
09:04:37GodEater_I don't think any of the bootloaders display a version number in that format
09:04:38LloreanHBK: Please, just follow the instructions at the page I linked you.
09:04:46jhMikeS?? petur? Confirmation?
09:05:02pondlifejhMikeS: Mic is still enabled, meters not moving. I'll try disabling voice...
09:05:04peturone moment, will try
09:05:14peturI do have voice disabled
09:05:30*jhMikeS did actually anticipate H300s doing always
09:05:57HBKso should I delete all the file or uninstintall rockbox before I do that?
09:06:12LloreanHBK: No. I already told you exactly what to do.
09:06:18LloreanHBK: Follow the instructions here for installing the bootloader:
09:06:20pondlifeOK. Still locks up if I disable voice without rebooting...
09:06:25 Quit linuxstb__ (Read error: 113 (No route to host))
09:06:33 Join linuxstb__ [0] (
09:06:42LloreanGodEater_: The black-screen bootloader should show the date in that format.
09:06:50pondlifeStill locks without voice, even after a reboot.
09:06:54HBKOK,I'll try it
09:07:05peturpondlife, jhMikeS: no lockup here
09:07:10jhMikeSnow that just doesn't make a lick of sense...the problem was voice not unboosting the thread
09:07:13GodEater_Llorean: including the svn revision number though ?
09:07:26LloreanGodEater_: Mine has an SVN revision number, at least.
09:07:30pondlifeActually it's not a hard lockup in this case, but it seems to be waiting forever for something
09:07:33jhMikeSor rather...boosting a thread it had no business boosting
09:07:42GodEater_it's been so long now since I looked an ipod bootloader, I've forgotten
09:07:50pondlifei.e. backlight is going on/off
09:07:59pondlifeBut long stop won't shutdown
09:08:03LloreanGodEater_: I don't normally look at it, and the newest one doesn't show anything unless you hold "Right" apparently.
09:08:16pondlifeDisk icon is lit too
09:08:18GodEater_I think the "no text" is a good idea
09:08:28jhMikeShmmmm....something with reloading the voice file probably
09:08:32***Saving seen data "./dancer.seen"
09:08:42pondlifeNo reason to reload that
09:08:45pondlifeIt's all disabled
09:09:01peturvoice loads if the file is available
09:09:14pondlifeShouldn't do. I changed that ages ago
09:09:23pondlifeIt only loads if any of the voice options are enabled
09:09:25jhMikeSit actually loads the file on first use of a clip
09:09:25HBKOh,yeah,i worked,thanks!
09:09:36 Join decayedcell [0] (i=3ba79d47@gateway/web/cgi-irc/
09:09:39HBKI think I know why I got wrong.
09:09:41pondlifeAh, it also locks if I stop and restart recording
09:10:07*petur shuts up and minds his payed work
09:10:07pondlifeSame symptoms - backlight functions working, but nothing else
09:10:07jhMikeSTwo H300s in total disagreement...hmmm
09:10:07 Join roolku [0] (
09:10:14HBKbe installation instruction I followed is for G4
09:10:46pondlifeThis is the very latest build (07:45)
09:10:57peturmine is from yesterday
09:11:11pondlifeI'm recording to WAV - default recording settings all round I expect
09:11:20jhMikeSpondlife: did you try one from right after the fix?
09:11:24peturand I _need_ reliable recording tomorrow
09:11:32pondlifeNot yet anyway
09:12:01jhMikeSI ran my stuff forever and not a peep...will check some more.
09:12:05pondlifeWhich SVN revision would be good?
09:12:19pondlifeJust before the IRQ rework?
09:12:25jhMikeS05:23- see if it changes
09:12:45 Quit HBK ("CGI:IRC")
09:12:51 Join HBK [0] (i=4a46cb46@gateway/web/cgi-irc/
09:12:53jhMikeSthough any problem there should realy hard lock
09:13:15HBKWhere's the instruction you gave me on the website?
09:13:46HBKI only found the old one in the documentation section on the website
09:13:55LloreanHBK: I don't understand the question. The link I gave you is on the website.
09:14:18LinusNHBK: for future reference, the IRC channel is logged:
09:14:36 Quit decayedcell ("CGI:IRC (Ping timeout)")
09:15:28LinusNLlorean: please tell me how to click my way to that wiki page
09:15:39LloreanLinusN: There's a big red link at the top of the IpodInstallation page.
09:15:49HBKhmm?here's the link i got
09:15:50LinusNLlorean: not anymore
09:16:29LloreanIf that's been changed, is the manual up to date?
09:16:31*Llorean goes to check
09:16:35GodEater_HBK: Wiki->IpodPort->IportInstallationBeta
09:16:46LloreanLinusN: The manual is up to date.
09:16:49GodEater_excuse me, IpodInstallationBeta
09:16:53HBKOh,yeah,I got it
09:16:56LloreanLinusN: Same instructions as the Beta page.
09:17:03LinusNLlorean: ok then
09:17:27LloreanLinusN: So the big red link is valid, it must've been changed when the Beta instructions were put into the manual because it used to link to the beta page.
09:17:37LloreanEither way, people ignore that red line regularly.
09:17:51LinusNLlorean: it's not a link
09:17:59LloreanWell, when it was a link
09:18:26LloreanI'm going to fix the IpodInstallation page now.
09:18:29LloreanUnless there's a reason to leave it.
09:18:58LinusNpondlife: latest svn works fine for me with voice enabled
09:19:04HBKI got the manual this afternoon,but it's still the old version,did you just update it?
09:19:21LloreanHBK: I just checked, it's definitely in the manual.
09:19:33LloreanHBK: It probably happened earlier today.
09:19:36pondlifeLinusN: Maybe I should put my config somewhere?
09:19:44HBKyeah,i got the pdf file of it
09:19:59pondlifeJust trying a small rollback, will report back shortly
09:20:09peturpondlife: put it in the recycle bin ;)
09:20:13GodEater_tell me you don't think this is pretty :
09:21:20*petur always sorts the 1's and 0's before throwing his digital waste away
09:21:24LinusNpondlife: it also worked when i turned off voice menus
09:21:32LinusNpondlife: and also after rebooting
09:21:34pondlifeIt's not voice related
09:21:45pondlifeI have all voice stuff off, and have rebooted
09:22:12 Quit amigan (Read error: 104 (Connection reset by peer))
09:23:16jhMikeSI don't know...things started misbehaving on the H120 where voice doesn't seem to even start. But that was before changing anything in the voice arena.
09:23:22pondlifeNope rev 12657 still locks
09:23:29pondlifeWill go back further...
09:24:33jhMikeSthe interrupt changes were tested all night so I don't really anticipate anything bad from that
09:24:50pondlifeJust building 12648.
09:25:15pondlifeWhat would it be waiting on before exiting, or before starting a second recording?
09:27:00jhMikeSjust starting a second recording and not leaving?
09:27:14LinusNpondlife: have you reset the settings, and have you set your RTC to a valid time?
09:27:18jhMikeSyou waiting for the flush on the first to finish?
09:27:25pondlifeLinusN: Not yet
09:27:34pondlifeMy RTC time looks ok though
09:27:41LinusNi have tried to stop and record again, and also starting a new file with Rec, both work just fine
09:28:11LinusNpondlife: and the disk is not full?
09:28:13jhMikeSThe only thing recording would wait on is for the next second to get a filename
09:28:22pondlifeUSB bootloader mode is working like a champ though!
09:29:17pondlifeOK, that still locks. Could it be caused by me not configuring a recordings directory?
09:29:56jhMikeSthe default should be just fine
09:30:05LinusNdefault should be /recordings
09:30:11pondlifeOK, cleared settings
09:31:05jhMikeSayef, wivepack, beglayerthree? lol
09:32:11pondlifeWorth a CHKDSK too...
09:34:01jhMikeS2 of 3 are in my favor so far
09:34:12HBKI found something wrong with manual in the uninstallation part
09:34:50jhMikeSpondlife: which voice file? I'll try the same myself.
09:35:05pondlifeNot sure, but it shouldn't be related.
09:35:09HBKit says when we uninstall rockbox we should use command line "ipodpatcher N -d"
09:35:14pondlifeI'll delete ENGLISH.VOICE next
09:35:29HBKbut the number N is not mentioned before in the manual
09:36:13HBKI think this is explained in the old version because the N is also used when we install rockbox
09:37:05jhMikeSI'm getting all strange menu behavior...I can't back out of the settings just goes to the first item over and over. Had a thing where the WPS wouldn't quite either in the same way.
09:37:17GodEater_I'm not sure the N is required in an uninstall either
09:37:28pondlifejhMikeS: OK, it's working now I cleared my settings.
09:37:30GodEater_it's *only* required if you've hooked up more than one iPod to your computer at once
09:37:40HBKbut in the new version,we don't have to scan the device to get the number it seems a little bit confusing to put this instruction which uses number N here.
09:37:49jhMikeSpondlife: con o sin voce?
09:37:51pondlifeWill re-enable stuff and see what stops it.
09:37:54pondlifeWith voice, yes
09:38:30pondlifeDefault settings and some ENGLISH.voice from that 7zip archive
09:38:40*jhMikeS totally gets in UI loops all the time now
09:39:03pondlifeOK - In at the deep end, reloading my previous config...
09:39:05GodEater_HBK: agreed - I'll check with linuxstb, and if, as I suspect there's no requirement for the N, I'll arrange a change to the manual
09:39:47*jhMikeS needs to stop getting all the bugs eyes and points to JdGordon ;P
09:40:19 Join barrywardell [0] (
09:41:08 Join webguest06 [0] (i=86223347@gateway/web/cgi-irc/
09:41:32LloreanGodEater_: There's no requirement for the N.
09:41:52HBKso what's the new method of uninstallation?
09:41:54 Quit webguest06 (Client Quit)
09:42:02 Quit barrywardell (Client Quit)
09:42:11LloreanHBK: The same as listed, with or without the N.
09:42:30 Join alexj [0] (i=86223347@gateway/web/cgi-irc/
09:42:32HBKjust use the patch.exe?
09:42:56LloreanWhat patch.exe?
09:43:01LloreanIf you mean ipodpatcher, yes.
09:43:30GodEater_HBK: just run ipodpatcher -d
09:44:23HBKyeah,i got it,but it still says the N thing in the manual.probably gotta make some change to it.
09:44:41LloreanHBK: The N thing is still valid, it's just not *necessary* any more.
09:45:03pondlifejhMikeS: Hmm, could it just be waiting a long long time to flush if database is updating?
09:45:08GodEater_Llorean: for consistency in the manual, I think we should remove the N though.
09:45:25jhMikeSyou're updating database and recording at the same time? wtf
09:45:27LloreanGodEater_: Oh, I agree wholly.
09:45:40pondlifeWell, I have update automatically turned on
09:45:41jhMikeSabsolutely...I wouldn't do that.
09:45:44pondlifeShould work!
09:45:46GodEater_having multiple iPods hooked up all at once is a pretty fringe case
09:45:58LloreanGodEater_: Indeed.
09:45:58jhMikeSnot gonna happen
09:46:15pondlifejhMikeS: Shouldn't lock up at least
09:46:48pondlifeI'll leave it 10 mins and see what happens
09:46:51jhMikeSi'm not sure it is ... will have to try it out though.
09:47:27jhMikeSdoes it take a long time if you just start recording and stop it immediately? a couple seconds should flush out fast
09:47:28pondlifeMaybe another bigfoot hiding in the database code...?
09:47:37pondlifejhMikeS: Yes
09:48:08pondlifeIt was working fine with cleared settings, and I could see the disk indicator for about 3 seconds after
09:48:23jhMikeSthat's about right
09:48:29pondlifeBut with the database enabled it's been sat there for 4 mins now
09:48:34pondlifeCan't be right!
09:48:55pondlifeMic is still enabled, would that make sense?
09:49:05pondlifei.e. flush then turn off mic
09:50:11JdGordonjhMikeS: ?
09:50:27pondlifeLinusN: On another topic, my rearranged bootloader here fixes the H300 remote issue. Would you like me to send you a copy, or just commit it?
09:50:41pondlifeI have no idea if it would work with the LCD remote though
09:50:51JdGordonLinusN: MENU > previous is causing some very odd bugs :p
09:51:00jhMikeSJdGordon: I'm running into strange infinite menu loops on my H120 when going left to go just goes to the same menu over and over
09:51:03LinusNpondlife: send me the patch and i'll test ot with all remotes
09:51:42jhMikeSBut the strangest one was the WPS restarting the track when you press stop
09:51:49LinusNjhMikeS: yes, i see the same problem - that loop was added when we aded the "menu goes to last screen" feature
09:52:03JdGordonyeah, ive got it all fixed except a very easy wps looping bug
09:52:17jhMikeSTHough I did just recording with database updating in the bg
09:52:33 Quit petur ("worrrk")
09:52:34 Quit HBK ("CGI:IRC (EOF)")
09:52:51 Nick linuxstb__ is now known as linuxstb (
09:53:41linuxstbGodEater_, Llorean: ipodpatcher now prompts for (i)nstall, (u)ninstall or (c)ancel - so to uninstall, you just double-click ipodpatcher and enter u. I'll update the manual...
09:54:02GodEater_linuxstb: okay
09:54:03jhMikeSpondlife: had no trouble with recording and updating db, in ram... and all that.
09:54:16pondlifeHmm, maybe it's not that
09:54:35GodEater_svn revert . seems not to work for me =/
09:54:36pondlifeDid you try shutting down and going straight into recording
09:54:44pondlifei.e. while the disk icon is still lit
09:54:46JdGordonGodEater_: it needs the -R
09:54:54 Join thewho [0] (i=c27f0812@gateway/web/cgi-irc/
09:55:18jhMikeSRecording screen on startup? Can't get any more immediate than that.
09:56:12GodEater_JdGordon, even if you're in the directory you wish to revert ?
09:56:14pondlifeHmm, I don't trust that quite yet!! No, I'm going to Root menu first
09:56:14GodEater_that seems silly
09:56:32JdGordonGodEater_: umm...
09:56:32pondlifeOK, it just worked, so it's not every time...
09:56:43jhMikeSThough if I try flushing a large amount of data...the db update does interfere afterall...and the flush boosts itself.
09:57:08JdGordonisnt the build table colours really bad for the colour blind?
09:57:10pondlifeNope, just locked again. Filesize 912k
09:57:11LinusNGodEater: -R is needed to revert the subdirs
09:57:17 Join webguest00 [0] (i=d554be96@gateway/web/cgi-irc/
09:57:35LinusNJdGordon: indeed :-)
09:57:50GodEater_LinusN: I realise that, but I'd dropped down into the directory I'd made the change at - so I expected it to only revert changes made to files in that directory
09:57:51JdGordonwe obviously dont have any devs with that problem then :p
09:58:00jhMikeSI think it does interfere...I have an odd blinking HD LED
09:58:10LinusNGodEater: and what happened without the -R?
09:58:14 Nick bun-bun_ is now known as bun-bun (n=bun@unaffiliated/bun-bun)
09:58:20GodEater_LinusN: absolutely nothing at all
09:58:32LinusNGodEater: i know why
09:58:35linuxstbamiconn: Can you build an amd64 version of the current SVN ipodpatcher with the bootloaders from ?
09:58:47pondlifejhMikeS: The lockup seems to happen whenever I have had the disk indicator lit throughout....
09:58:59LinusNGodEater: use * instead of .
09:59:16pondlifeIf I wait for it to go out beforehand, it works fine
09:59:19GodEater_I think the svn help is misleading in that case - it doesn't mean PATH
09:59:21LinusNif you specify a directory as an argument, you will need -R
09:59:23thewhoSorry for nagging, but: is the #ifndef in radio.c:428 needed? It causes the main menu item to be "Now playing" when the main menu is called from the radio screen IMHO
09:59:34 Quit z0de (Read error: 104 (Connection reset by peer))
10:00 wants too many disk resources or something else is wrong that busts the disk access
10:00:03*pondlife thinks that menu item should just be called "Playback" at all times
10:00:16webguest00Hi, is there a easy way to compile (a newly created) plugin. I'm trying to compile a simple hello world plugin, with directions found on the however when i run the arm-elf-gcc compiler it cannot find several headers,
10:00:56LinusNthewho: change it and submit a patch
10:01:02pondlifejhMikeS: Shouldn't lock up though...
10:01:14pondlifeIt's a pin-needed job.
10:01:31Lloreanwebguest00: Are you using the normal ../tools/configure, make, make zip method of compiling Rockbox?
10:01:50thewhopondlife: this is another question. The sim (possibly) still has an error
10:01:58pondlifeIt has many
10:02:05webguest00Llorean: just basically running arm-elf-gcc ./plugin.c ..:)
10:02:11thewhoLinusN: but why was it made that way? Isn't there something I don't understand?
10:02:41LinusNthewho: it's probably there because radio_stop() didn't exist in the sim in the older days
10:02:47Lloreanwebguest00: I would suggest following the proper compiling instructions, it actually works that way's doing nothing but disk writing there though and db is doing whatever it's doing. I can't have recording change things for database updates...It'll have to go the other way.
10:03:15pondlifeMust be a bug in the disk handling stuff
10:03:34pondlifeIt's been locked for 10 mins now. It's not coming back!
10:03:37jhMikeSI had a blinking led...I don't know if that's in the driver or what.
10:03:42webguest00Llorean: ok. but does it mean i need to recompile the whole of rockbox to just compile a single plugin ?
10:03:43thewhoLinusN: I mean: if someone wrote that code (which is an effort) this must have had a reason
10:04:01pondlifeI have a steady indicator on the status bar
10:04:03jhMikeSBlinking on/off once a second
10:04:15thewhoLinusN: sorry, missed your msg
10:04:21pondlifeDo you have an H3x0 available?
10:04:28Lloreanwebguest00: After the first compile, as long as you don't remove the build directory, it'll only recompile the files that have changed.
10:04:37linuxstbwebguest00: Yes. Add your plugin to apps/plugins/SOURCES, and then rebuild Rockbox.
10:04:41JdGordonthewho: does moving the ifndef under the audio_stop() fix the bug?
10:04:44jhMikeSX5 has a virtual LED though...not that it should matter
10:04:50thewhoLinusN: err... do you mean "audio_stop"?
10:04:51LinusNthewho: if it isn't too much trouble, please fix it and submit a patch
10:05:02LinusNthewho: of course
10:05:16jhMikeSI didn't hear the HD doing anything at all once it locked ... all actual activity ceased
10:05:40webguest00Llorean and linuxstb: ok thanks again
10:06:01pondlifejhMikeS: I don't feel HD vibration, but the indicator stays lit
10:06:18thewhoJdGordon: I dont't know, but probably yes. I removed the whole line (and #endif also of course) and that did fix the bug. But it tested only on H120 sim. The nested #if has to do with SW_CODEC which may still be importatnt for other platforms
10:06:50LinusNthegeek: try with an archos recorder sim too
10:06:59LinusNsorry, archos fm recorder
10:07:03jhMikeSpondlife: this points to something more nasty I think. Try checking with the access check turned on.
10:07:05JdGordontry just moving the #ifndef SIMULATOR to under the audio_stop(), i that fixes it then thats all that should be changed
10:07:18pondlifejhMikeS: Access check?
10:07:21 Quit alexj ("CGI:IRC")
10:07:30pondlifeIs that part of the debug build?
10:08:47pondlifeStill locks up with dircache disabled FWIW
10:08:50thewhoJdGordon: I'm quite sure it will fix it. The "problem" is that on the sim, the audio is not stopped when FM is started. And that's because of that #ifndef
10:12:17 Quit jhMikeS (Nick collision from services.)
10:12:21thewhoJdGordon: but I can't test atm
10:12:23 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
10:13:13 Quit GodEater_ ("CGI:IRC (Ping timeout)")
10:14:30jhMikeSfrom the sound of the drive, database has had it not long after recording proceeds...the disk just spins even before a flush. guess it's time to look at that code.
10:14:56pondlifeSeems ok as long as database isn't loaded into RAM too.
10:15:12pondlifeBut I'm just re-initialising, so a bit too early to say for sure.
10:15:54JdGordonthewho: seems to work, ill commit it
10:16:56jhMikeSrecording might be clobbering its memory ...
10:17:41jhMikeSI don't even know the file that contains that code :p
10:18:43jhMikeSstill tagcache eh?
10:18:56pondlifejhMikeS: I have about 14000 tracks if that's relevan
10:19:13thewhoJdGordon: dit you shift the line or remove it?
10:19:31JdGordonshouldnt really make much difference
10:19:46pondlifeHmm, the initialisation progress sticks at -1% now...
10:20:00pondlifeI'm sure it used to work
10:20:06jhMikeSpondlife: recording should not have a problem, you've got 32MB of RAM there
10:20:27jhMikeSit runs on 16MB just fine
10:20:47pondlifeIf you enable tagcache in RAM, and make sure it's updating when you attempt recording, can you repro?
10:21:02pondlifeSorry, I mean DATABASE
10:21:18thewhoJdGordon: ok, the rest is only for HW_CODEC, so not relevant for me :-p
10:21:22jhMikeSthat's what was doing it. just recording kills the database activity on disk. it doesn't even need to start flushing audio
10:21:50pondlifeOK, so the problem occurs when database attempts to continue?
10:22:47jhMikeSit's doing it's thing...recording clobbers it. database has to stay clear of it for sure.
10:23:18pondlifeSo a recording bug then?
10:23:33*pondlife likes to lay the blame out first :) gets it's buffers properly
10:23:52pondlifeHmm, so low level disk stuff?
10:24:31jhMikeSmaybe database is grabbing ram...if that's not it...could be low level.
10:25:10linuxstbI've just removed all ipod installation instructions from the wiki (replacing with a line telling people to follow the instructions in the manual). So now there are no visible old install instructions.
10:25:17pondlifeAh, have a look at allocate_tempbuf() in tagcache.c...
10:25:28pondlifeaudiobuf += tempbuf_size;
10:26:01pondlifeI wonder if that's being called?
10:27:00jhMikeSyeah, that's just wrong and only safe before audio init
10:27:30linuxstbLinusN: I would like to break the links from third-party tutorials to the old installation files on - could you create an "ARCHIVE" subdirectory in bootloader/ipod, and do "mv bootloader*.bin ipod_fw* disk_dump* ipodpatcher*.zip ipodpatcher.exe ARCHIVE/"
10:28:12jhMikeSonce audio is initialized, there's rules to obey :)
10:28:51pondlifeShouldn't there always be rules about modifying the value of audiobuf too?
10:29:11pondlifeThis should go through buffer_alloc and hence your proposed locks...
10:29:25jhMikeSIt can go anywhere before audio init
10:29:31jhMikeSit frees it too though
10:29:51 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
10:30:56pondlifeIt should just be claiiming the audio buffer, and using that maybe?
10:31:10pondlifeIf not possible, then delay commit until the next reboot as now.
10:31:29pondlifeIf you knacker part of the audio buffer after init then surely you may as well.
10:34:24 Join barrywardell [0] (
10:34:43 Join barrywardell_ [0] (
10:34:58jhMikeSnot sure...I want to check real quick if that function is called during record
10:36:27pondlifeTrigger an Update Now before you go into Recording
10:37:06jhMikeSNever had to since I never let it finish panicf yet
10:42:34jhMikeSstrange it worked just fine that time
10:42:48 Quit dmdfan- ("ircN 7.27 + 7.0 for mIRC (2002/01/10 00.00)")
10:43:22jhMikeSwill try update now
10:46:14jhMikeSI see the LED blinking and can hear the disk in the mic: click click
10:46:43jhMikeSbut it's not making calls to that
10:47:11 Join hannesd__ [0] (
10:47:21 Join safetydan [0] (n=dan@rockbox/developer/safetydan)
10:48:03 Quit shnee_ (Read error: 104 (Connection reset by peer))
10:48:54 Join Shaid` [0] (
10:50:32 Quit thewho ("CGI:IRC")
10:51:25jhMikeSyup...taking forever to flush out again
10:52:18 Quit webguest00 ("CGI:IRC")
10:54:09*XavierGr walks in with a box on his hands :D
10:54:33hcsbox of chocolates?
10:55:49*jhMikeS was thinking of some naughty kind of box
10:55:50LinusNlinuxstb: done
10:56:56*XavierGr adds himself to the decimated archos (ondio) population and joins the old family :P
10:56:56hcsbox of live grenades?
10:57:20XavierGrnah I just received my ondio FM
10:57:37amiconneuhm, for some reason my tools/ipodpacther/* files are gone (execpt ipodpatcher.exe), and 'svn up' doesn't bring them back ?!?
10:58:03*JdGordon didnt do it!
10:58:03hcsrbutil/ipodpatcher now
10:58:11XavierGrgah this thing is very small!!
10:58:23XavierGrsomewhat a little bit fatty though :P
10:58:25jhMikeSummm...svn rm can do that :)
10:58:55amiconnHmm, but why is the dir still there? Or is my working copy fooling me?
10:59:32amiconnAnd on my amd64 box the files still exists after 'svn up'
10:59:58jhMikeSI had something like that where I just had to blast the directories and do reverts but got somewhere
11:01:15amiconnlinuxstb: 'make clean' is abit aggressive in the ipodpatcher dir...
11:01:19JdGordoncrap :'( I got another 3 hour break tomorow at uni
11:01:24XavierGrsweet the box has 8 aaa batteries too
11:01:39JdGordonXavierGr: to add to the shipping cost? :p
11:03:20 Quit hannesd_ (Read error: 110 (Connection timed out))
11:03:21 Join GreyFoux [0] (
11:03:31XavierGrhe included in total 11 aaa batteries and even an rca cable
11:03:39jhMikeS8 aaa? that would get expensive to run
11:03:58XavierGrit needs 3 to run
11:05:03linuxstbamiconn: What does it delete that you think it shouldn't?
11:05:19 Quit GreyFoux (Client Quit)
11:05:20linuxstbLinusN: Thanks for cleaning the ipod directory.
11:05:23XavierGrnow time to get rid of the OF!!!
11:05:35XavierGrI don't even know how to put rockbox on this thing...
11:06:23jhMikeSI wonder if the ATA driver or file functions bite it with all the multithreaded writing going on
11:06:31 Quit Shaid (Read error: 110 (Connection timed out))
11:06:32 Nick Shaid` is now known as Shaid (
11:06:43*jhMikeS wonders if pondlife bailed out
11:07:32*JdGordon gives up... has anyone got a few min to look at the new root_menu loop to figure out why you can get stuck in wps loops easily?
11:08:35***Saving seen data "./dancer.seen"
11:09:49jhMikeSwhat about the other loops?
11:09:51XavierGris it possible to brick the ondio if I do something wrong when installing rockbox on it?
11:10:12 Part Llorean
11:10:48JdGordonhaha freeking typical... as soon as I ask for help I get it working :p
11:11:08XavierGrhmm ondio should be in fat32 filesystem correct?
11:11:37LinusNXavierGr: fat16 iirc
11:11:37jhMikeSJdGordon: you always do just before I go trying to find the relevant code
11:11:52amiconnlinuxstb: Everything...
11:11:59 Quit Farp (Read error: 104 (Connection reset by peer))
11:12:04*LinusN got as far as opening root_menu.c
11:12:05 Join Farp [0] (i=Farp@
11:12:17pondlifejhMikeS: Nope still here-ish
11:12:33amiconnjhMikeS: The ata driver is mutexed
11:13:04amiconnlinuxstb: Everything except the Makefile, to be precise
11:13:05LinusNi wonder if the ata callback thingy might be the problem
11:13:15pondlifejhMikeS: So you reproduced it...
11:13:22jhMikeSpondlife: over and over
11:13:27pondlifeGood ;p
11:13:58pondlifeI'm back on Paid Work at the moment, but will try to keep an eye on this here channel if you want anything trying out.
11:14:12jhMikeSrecording is very simple with the writing but it writes megabytes
11:15:10linuxstbamiconn: That's odd... Is the "*~" (which I use to clean emacs backup files, so I should probably have not committed it) matching files on your system?
11:15:11jhMikeSwhen they're both going, something collides and it just sits there twitching.
11:15:28amiconnlinuxstb: No it's not
11:16:23jhMikeSplayback doesn't seem bothered but it's reading
11:16:23linuxstbamiconn: Then I can't see why it would delete the .c and .h files...
11:16:30XavierGromg it takes ages to put the build to ondio....!
11:17:38 Join fejfighter [0] (
11:17:59jhMikeSthe other bigfoots are after me for bagging the other one :(
11:18:17pondlifeNever get into fights with bigfoots
11:18:29*pondlife wonders what the collective noun is
11:18:42LinusNamiconn: 'make clean' in rbutil/ipodpatcher works fine for me
11:18:45Shaida hairball of bigfoots?
11:19:24 Quit barrywardell (Read error: 110 (Connection timed out))
11:19:58XavierGryippie just booted my ondio to Rockbox! :D
11:21:53 Join _kclaf [0] (
11:22:12 Quit kclaf (Read error: 104 (Connection reset by peer))
11:22:28amiconnWhere should I throw this amd64 ipodpatcher? LinusN?
11:22:44LinusNsomewhere where i can download it
11:25:21jhMikeSLinusN: the ata callback you're talking about is the idle callback?
11:25:25XavierGricacther looks neat on the ondio
11:25:42LinusNjhMikeS: yes
11:25:54LinusNbut i realize that it shouldn't be an issue
11:26:29JdGordonyou guys really like blaming my code for everything dont you? :'(
11:26:37 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
11:26:48LinusNjhMikeS: but i can imagine that a huge write that is interrupted by another could slow down quite a lot, since the drive would seek a lot
11:26:53LinusNJdGordon: hahaha
11:26:58 Join miepchen^schlaf [0] (
11:26:58jhMikeSWho's #1? JdGordon or Slasheri? :P
11:28:05XavierGrondio doens't has a quickscreen button?
11:28:24 Join pixelma [0] (i=pixelma@
11:28:25jhMikeSLinusN: It starts out quite normal with both then just halts
11:28:31JdGordonXavierGr: iirc you press long menu twice
11:28:34LinusNjhMikeS: nasty
11:28:36amiconnXavierGr: No it doesn't
11:29:56linuxstbamiconn: Did you work out why make clean deleted all the files?
11:30:55jhMikeSmemory conflicts don't seem an issue
11:31:38 Join dpmarsh [0] (
11:31:58amiconnlinuxstb: In rbutil/ipodpatcher it works
11:32:38dpmarshi just signed up on TWiki and need access to load up my WPS theme for ipod5g, can someone help me?
11:32:55JdGordonpondlife: should <Untagged> be the last item in the list always and <All tracks> be first?
11:33:05 Quit GodEater_ ("CGI:IRC (Ping timeout)")
11:33:14JdGordonor anyone..?
11:33:21pondlifeJdGordon: I'd prefer it with <All tracks> then <Untagged>
11:33:41pondlifePity '<' is >'9'
11:33:49amiconnThe link to the ipodpatcher amd64 binary in the manual is wrong
11:34:08JdGordonSlasheri: do you know which function sorts the tree items?
11:35:41JdGordonthats what i thought... but im not getting my debug output
11:35:54pondlifeOnly when you rebuild the database perhaps?
11:36:12JdGordontheir orders are stored in the db?
11:36:38pondlifeNo idea, just a guess
11:37:04amiconnYes, the db needs to have the items pre-sorted
11:37:08bluebrotherIMO everything starting with < should be at the start
11:37:20bluebrotheras it's some kind of special entry
11:37:20JdGordonwhat about when the order is reversed?
11:37:45pondlifeJdGordon: I've got a fix for that here...
11:37:52bluebrotherif the sorting order is reversed then it's at the end
11:38:08JdGordonpondlife: fix for which?
11:38:08amiconn...otherwise segmented browsing couldn't work
11:38:12preglowamiconn: i did some measuring yesterday, and the wm8975 bass/treble controls do indeed seem to work, but they also adjust volume in addition
11:38:26preglowamiconn: do you think we should attempt to fix by prescaling volume, or just use sw tone controls?
11:38:46JdGordonpondlife: to fix your bug report? or..?
11:39:10 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
11:39:17pondlifeAlthough a simple check for a leading < would be faster and maybe better
11:39:26jhMikeSmaybe sorting rules should be part of lang files?
11:39:49pondlifeAh, that's a can of worms... The/Die/.... etc.
11:39:58 Join inversions [0] (
11:40:08pondlifeJdGordon: I can commit it if you think that's ok
11:40:12JdGordon<Untagged> is hardcoded tho isnt it?
11:40:23pondlifeYes, #defined
11:40:30JdGordonpondlife: you got rid of the inverse check?
11:41:02pondlifeOnly the commented out "The" check (that was buggy anyway)
11:41:10jhMikeSnot the The/Die/ ... etc. stuff just basic sorting
11:41:25JdGordonare we looking at the same function?
11:41:37pondlifestatic int compare(const void *p1, const void *p2)
11:41:53JdGordonoh, i was looking in tagtree.c
11:42:37pondlifeAha - that might be a better place to put it (if it's used)
11:43:11*JdGordon doesnt know
11:43:15JdGordonSlasheri: you around?
11:43:30pondlifeCertainly the tagcache compare method works upon rebuild, but I suspect it would kill the update on existing databases. Sort stability is good, huh?
11:43:51JdGordonwell, i think if its slow it should be there instead of in the display code
11:43:56pondlifeThis is a presentation issue, so should really be handled in the UI
11:44:04JdGordonyou can export then reimport on db changes
11:44:11*pondlife likes UI and logic to be seperate
11:46:26Slasheriyep, compare() in tagcache.c is used when building the db
11:46:59JdGordonwhich is the better compare() to put this in tho?
11:47:03pondlifeSlasheri: Would you be happy with my paste above?
11:47:17JdGordonand why isnt <Untagged> xlateable?
11:47:37Slasheridb files needs to be correctly sorted or low memory targets such an archos might be unable to display the entries correctly
11:48:03pondlifeOK, strict charsort?
11:48:27pondlifeSo it needs to be done in tagtree, not tagcache.
11:48:29Slasheripondlife: doing the sorting in tagcache.c is the best place, then it works for sure :)
11:48:52pondlife:~/ confused
11:49:07pondlifeDefine "correctly sorted"
11:49:18Slasherior doing it in tagtree.c needs that all entries can be loaded in ram at once and sorted then
11:49:39Slasheriwith chunked browsing that method does not work
11:49:54 Quit datachild ("deny the day of towmorrow")
11:50:12Slasheribut still work to do, back in evening ->
11:50:19pondlifeOK, I'll commit this for now!!
11:50:25preglowmarkun: testet the cf patch?
11:50:27JdGordonbut.... but....
11:50:47JdGordonis there anyway to see if it will show <Untagged> without checking the string?
11:50:59pondlifeNot as far as I can see.
11:51:09pondlifeThe string is put in there early on.
11:51:15JdGordonoh well...
11:51:22JdGordonguess thats why its not translatable..
11:51:42pondlifeIt should be translatable still, that's just another job.
11:51:45JdGordondoing char comparisson would be faster tho :p
11:52:38pondlifeWell, you can always update it some more!
11:56:06XavierGramiconn: any page on the wiki about the flashing procedure on ondio fm? I am looking bootbox right now, but isn't this outdated?
11:57:09*amiconn still didn't update & merge the flashing instriuctions :(
11:58:40XavierGrso I shouldn't try to flash it unless you update it, right?
12:00:55 Quit inversions (Read error: 110 (Connection timed out))
12:01:55 Quit fejfighter ()
12:03:02*preglow glances at e200tool and blinks
12:03:11preglowdoes this guy have usb going on portalplayer, then?
12:05:37barrywardell_although it's only a basic usb driver
12:05:57preglowwell, yes, but it works!
12:10:47 Quit aliask ("Chatzilla 0.9.77 [Firefox]")
12:13:55 Join iwantanimac [0] (
12:14:22 Join Criamos [0] (
12:15:34LinusNi am constantly amazed how many people register in the Wiki and fail to read the notices about having to ask for permission to write
12:16:01iwantanimacer.. like what exactly?
12:16:04GodEater_I'm constantly amazed about all sorts of things that people feel to read
12:16:19GodEater_especially the big red stuff in bold
12:16:26LinusNit is mentioned here: ("...and then ask for write permission")
12:16:41barrywardell_preglow: are you interested in working on a USB driver for Rockbox?
12:16:43Kasperleexcept when it says "DON'T PUSH THE BIG RED BUTTON"
12:16:50LinusNand here: (important note #1)
12:17:00preglowbarrywardell_: i might be, but i doubt i'll have time
12:17:04LinusNand then also in the registration confirmation email
12:17:08preglownot with the dsp stuff on the todo list
12:17:18iwantanimacoh lol that...
12:18:17barrywardell_preglow: ok, well let me know if you're interested. I've done a bit of work on it
12:18:21LinusNwe mention it *three* times, and so many people still miss that information
12:18:35LinusNwhat's with people and their reading abilities these days
12:18:44barrywardell_dionoea: while we're on the subject, how's the USB stuff going for you?
12:18:46Kasperlebarrywardell_: is there any documentation on the USB hardware?
12:19:14preglowLinusN: you need bigger letters
12:19:16preglowalso redder ones
12:19:20preglowif they can blink, all the better
12:19:28LinusNpreglow: they need bigger brains
12:19:38preglowwell, if you can supply that as well...
12:19:44GodEater_if they could somehow reach out of the screen and hit people too - that would be good
12:20:07LinusNi mean, the note on the registration page is in red letters
12:20:35iwantanimacyeah. impossible to miss unless you're colour blind and you can't tell the difference between light blue and bright red...
12:20:45iwantanimacwhich makes a lot of people colour blind.
12:20:46GodEater_judging from a lot of recent posts on the forums, we're getting an influx of total illiterates
12:20:48LinusNand the guy yesterday failed to read the important notice in red on the ipod installation page
12:21:23LinusNsometimes you just feel that writing good docs is futile
12:21:28*linuxstb suggests we translate pages via
12:21:50LinusNlinuxstb: lol
12:22:57GodEater_we need to answer more forum posts like that
12:23:19iwantanimacright. bookmarked. =P
12:23:45GodEater_I find it odd that google doesn't include AOL speak in their language tools
12:23:46linuxstbShame the translator doesn't work the other way round...
12:23:53 Join barrywardell [0] (
12:23:57GodEater_they do pig latin and elmer fudd for god's sake
12:24:02preglowlinuxstb: hahahaha
12:24:10scorchebork bork?
12:24:24GodEater_search me
12:24:32preglowall the random "wtf"s really bloody crack me up
12:25:37iwantanimacHi! = "HI!!1!1 WTF"
12:25:50 Join amigan [0] (i=dcp1990@unaffiliated/amigan)
12:26:11GodEater_yeah, it randomly drops in WTF, OMG, and LOL
12:26:31iwantanimacclick two more times and Hi! = "HI!!1!1 OMG WTF LOL"
12:27:40 Quit iwantanimac ("iMac, out.")
12:28:24*preglow wonders what's up with the pletora of "rescue" modes on sansas
12:28:56markunpreglow: no, forgot, and now I have to go again..
12:29:03linuxstbYou think 4 different USB modes is excessive?
12:29:29GodEater_preglow: I think it just shows Sansa understand their target demographic ;)
12:30:10preglowmarkun: i think i'll just go ahead and commit it
12:30:24preglowif anyone complains, i'll just revert. as far as i know, it sounds the same
12:31:01markunpreglow: ok, I'll check it then when I get back
12:32:48pondlifeLinusN: Did you get my mail?
12:34:23LinusNpondlife: found it in my spam box :-)
12:37:05barrywardell_Kasperle: sorry. i was afk for a minute there. There are docs and even a linux driver for the USB hardware
12:37:22XavierGrlinuxstb: man, you made my day!! This link is "da" best!11! :D
12:38:25barrywardell_Kasperle: see the i.MX31 reference manual on the freescale website and this linux driver:
12:41:34 Join muesli__ [0] (n=muesli_t@
12:44:35*Shaid commits a patch that translates all code comments into AOLspeek
12:45:29 Quit safetydan ("Ex-Chat")
12:45:33pondlifeHmm, code comments eh? Good idea
12:46:15Shaidthen we could tell everyone to RAED TEH COMM3NTS!!11!!oneoneone
12:47:57XavierGrthat aol translator really made me cry at one moment
12:48:14LinusNpondlife: works perfectly!
12:48:20LinusNwith all 3 remote types
12:48:37pondlifeWasn't rocket science, just a little rearrangement
12:48:55pondlifeWhat remains before v7 then?
12:49:04pondlifeMore RTC stuff? Car mode?
12:49:32pondlifeOr is that about it until we get the EEPROM settings
12:49:33LinusNeeprom settings might be nice, but it seems the h300 OF is a little sensitive to eeprom changes
12:49:48pondlifeIndeed. Might be best to release at this point.
12:50:19LinusNwill you commit?
12:51:10 Join muesli- [0] (n=muesli_t@
12:51:30 Quit Shaid (Read error: 131 (Connection reset by peer))
12:51:34Kasperlebarrywardell_: that looks very encouraging
12:54:07JdGordonLinusN: is the pcf stuff for the h300 to get alarm working fairly trivial? or do you need to know what your doing for it?
12:55:17pondlifeDid we fix the occasional H300 EEPROM corruption problem yet?
12:55:21 Join Nico_P [0] (
12:56:00amiconnIf we had usbotg in rockbox on the H300, there would be n need for the of
12:56:13amiconnLinusN: How's the tracing going?
12:56:17*XavierGr points to petur :P
12:56:48 Join mahi [0] (i=3a444a53@gateway/web/cgi-irc/
12:57:00mahihi all,
12:57:11mahii am using rockbox on my ipod mini, its damn slow, any suggestions
12:57:59mahianybody around :(
12:58:02pondlifeGet a Gigabeat ;)
12:58:04linuxstbamiconn: Rockbox is still missing WMA support and (completed) video playback compared to the H300 OF.
12:58:22mahiany suggestions guys
12:58:24pondlifeWMA without DRM
12:58:32mahiany tweaks to speed rock box
12:58:46pondlifeDisable the EQ and crossfeed
12:58:53mahioh is it
12:58:54pondlifeDon't use peakmeters on your WPS.
12:58:56mahilemme try
12:59:39pondlifemahi: Which file format are you playing?
12:59:42 Quit barrywardell ()
13:00:06mahiyeah disabling equalizer worked
13:00:10preglowlinuxstb: i did measurements on wm8975 treble/bass yesterday, and they do indeed work, but there's also always a gain boost/cut
13:00:27mahipondlife thanx a lot dude
13:00:42pondlifeAnyone know if EQ isn't disabled by default?
13:00:47pondlifeNo prob
13:00:58mahiyeah it is disabled default
13:01:02mahii enabled it manually
13:01:12pondlifeOK, that's good
13:01:26mahiluv Rockbox :)
13:01:26pondlifeNeeds optimisation for iPod still..
13:01:36mahido you have tips ?
13:01:53pondlifeNope, or I'd do it.
13:02:54pondlifeSo... LinusN: did we suss out the H300 EEPROM corruption yet? I've only ever seen it once. Recording was suspected wasn't it?
13:04:06 Quit muesli__ (Read error: 110 (Connection timed out))
13:05:06mahihey pondlife, can i change rockbox welcome logo
13:05:19 Quit GodEater_ ("CGI:IRC (Ping timeout)")
13:05:30preglowthis is just too silly for words
13:05:43mahiam not a geek :D
13:05:44preglowaudiohw_set_mixer_vol isn't bloody implemented on wm8975
13:06:02preglowno bloody wonder the volume changes
13:06:04mahi hey pondlife, can i change rockbox welcome logo
13:06:09pondlifemahi: If you want to.. .You'll need to look at the code
13:06:27mahi#-o, am a lazy dog :D, what ever, i wll check it out
13:06:30pondlifeNo idea how, but the source is all there for us to play with
13:06:34JdGordondont even need to look at the code
13:06:41JdGordonthere is a bmp you need to change
13:06:45mahiwhy not if there is a tool to edit splash screen
13:07:04mahipath?? JdGordon
13:07:50JdGordonthen depends on your target
13:07:53pixelmaand then you to compile your own
13:08:02pixelma*need to
13:08:38***Saving seen data "./dancer.seen"
13:09:05pixelmafor Mini: rockboxlogo.138x46x2.bmp
13:09:37 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
13:13:35 Quit actown (Read error: 104 (Connection reset by peer))
13:13:41 Join actown [0] (n=actown@unaffiliated/actown)
13:19:01preglowamiconn: it looks like the wm8975 mixer is after the tone controls, won't that make it unsuitable for prescaling?
13:19:05preglowsince it's not really pre anymore
13:20:49 Join austriancoder [0] (n=austrian@
13:22:45JdGordonaustriancoder: tomorows the big day, aye? :)
13:22:45 Quit jhMikeS (Read error: 104 (Connection reset by peer))
13:23:42 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
13:23:52 Join Id2ndR [0] (
13:24:54austriancoderJdGordon: no.. its on the 9th :)
13:25:56JdGordonwell.. umm... its almost the 8th here... so close enough to the th being tomorow :p
13:26:12austriancoder*g* okay
13:27:37preglowJdGordon: so, all the start screen options work now?
13:27:59 Quit GodEater_ ("CGI:IRC (Ping timeout)")
13:28:23JdGordonnoone has mentioned anything else which needs attention.. (or reminindg if i've forgotten)
13:29:03 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
13:29:23 Join miepchen^schlaf [0] (
13:31:17 Quit Ribs ("eh eh ehhhh!")
13:33:22 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
13:36:23pondlifeJdGordon : Shall I get testing .... ? ;p
13:37:21JdGordonanyone but you :D
13:37:30JdGordonim busy with other parts of the code now :p
13:37:31pondlifelol... I'm just compiling now...
13:37:58*pondlife wonders if jhMikeS has a headache yet?
13:39:17 Part jhulst ("Kopete 0.12.4 :")
13:40:53*JdGordon is scared he knows the answer already... _BUT_...
13:41:08JdGordonanyone not like this patch?">
13:41:26JdGordonit replaces the old menu api with wrappers around the new api
13:42:02linuxstbWhy is that needed? Shouldn't we just try and change all the old menus?
13:42:24JdGordonincluding all the plugin menus?
13:42:38JdGordonsure, but untill then, this shuold go in
13:42:46JdGordonshold save a few bytes...
13:43:00amiconnpreglow: It seems that the WM8975 tone controls are meant to be used independently from the volume, similar to how it's done in the MAS3587F / MAS3539F
13:43:07linuxstbJdGordon: It just seems a waste of time to me...
13:43:13amiconnSo no pre- or post-scaling should be applied
13:43:23JdGordonprobably.. but its finished and working.. so that time is gone anyway :p
13:44:03 Quit mahi ("CGI:IRC")
13:44:23GodEater_JdGordon: I like that argument ;)
13:44:40preglowtenthdb2mixer() seems to be just taken from the uda1380 driver....
13:44:54 Join Thundercloud [0] (n=thunderc@
13:46:20JdGordonLinusN: amiconn ?
13:46:53JdGordondrops 800 bytes from the v2 rec
13:47:04JdGordonso thats almost good enough reason to do it i tihnk :p
13:47:56pondlifeGo for it, but don't go to bed until you're sure it's working properly :)
13:49:03*JdGordon better test voice :p
13:49:20pondlifeAha, a guilty conscience?
13:49:43 Quit GodEater_ ("CGI:IRC (Ping timeout)")
13:49:47 Quit dpmarsh (Remote closed the connection)
13:50:05 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
13:50:07amiconnpreglow: Basically, sound.c: set_prescaled_volume() should be ifdefed out for WM8975, and volume, treble and bass set directly in their respective set_* functions, without any interdependency
13:50:43JdGordonpondlife: no.. but tired
13:51:34preglowamiconn: then it looses the possibility to use sw tone controls
13:51:44preglowi hate that mess
13:52:17pondlifeAren't sw tone controls only for players with no hardware tone control?
13:52:42preglowi use them for wm8975 too
13:52:50preglowsince the bass/treble controls there are fairly limited
13:53:11preglowthe real hardware values don't even map well onto the settings range rockbox currently uses
13:53:22pondlifeOK, then why bother with the hardware controls?
13:53:40*pondlife is probably asking stupid questions
13:54:09amiconnpreglow: +dB/-6dB is not sufficient?
13:54:16amiconn+9dB even
13:55:21preglowpondlife: they don't use as much battery, and are sometimes quite sufficient
13:55:38preglowamiconn: it's -6 db -> 9 db, and no, i don't really think that is sufficient
13:55:46preglowbut i don't care much, i almost never use them
13:56:09preglowbut it's my code, and i'd just like it if it could be enabled fairly easily on all platforms
13:56:22preglowbut i should probably spend the time on something better
13:56:26 Quit Id2ndR ("Parti")
13:56:27preglowbut but but
13:56:27pondlifeI'd go with hardware only. Having 2 sets of tone controls (plus EQ) is only going to confuse users.
13:56:35preglowpondlife: not two sets, just one set
13:56:41preglowpondlife: if you use sw tone controls, that's all you have
13:56:57pondlifeOK, so using more battery then...
13:56:59preglowpondlife: they already in svn, so nothing new there
13:57:12pondlifeThat's ok, I thought you were proposing both being available.
13:57:13preglowpondlife: well, if you value sound quality more than battery life
13:58:07pondlifeDepends entirely. I would personally always rate battery life above sound quality
13:58:16amiconnpreglow: Still no biggie to allow for sw tone controls, just some more ifdefs
13:58:17pondlifeBut I use my DAP on the move
13:58:47*amiconn thinks -6 -> +9 dB is sufficient
13:59:06preglowhow many of the other wm codecs should behave the same?
13:59:14amiconnI just need tone controls to account for fletcher-munson, because I use moderate listening volumes
13:59:39amiconnpreglow: Don't know about the 8758; the others don't have hw tone controls anyway
14:00:26preglow8731l seems to be mostly identical to 8975
14:00:59amiconnThe 8731 doesn't have tone controls
14:01:50preglowit's still listed in set_prescaled_volume, though
14:01:52preglowi'll just gut it out
14:02:51preglowgigabeat's wm8751 is supposed to be the same, so i'll gut that out too
14:03:17amiconnSame for 8721 (and 8711), i.e. the minis
14:03:50preglowonly 8758 left, then
14:04:34preglowand i'm willing to bet that's the same
14:04:49 Nick _kclaf is now known as kclaf (
14:08:33preglowlinuxstb: you know what datasheet you based the 8758 on?
14:09:05 Join ackbahr [0] (
14:15:07preglowamiconn: the wm codecs that have tone controls seem to use 1.5 db increments, resulting in most of the values in rockbox being wrong
14:15:15JdGordonoh bluddy helll I even remebered to use the right type cast :'(
14:15:54JdGordonah, what cast do we use for int > void* ?
14:16:38preglowthe bass/treble settings only seems to use integers, though
14:17:05preglowi could extend them to use tenth db for all targets to fix
14:18:13preglowperhaps it would be just as good to pretend the the increment is 2db
14:18:41preglowit'll be completely wrong, of course...
14:19:16 Quit ackbahr (Remote closed the connection)
14:19:22pondlifeJdGordon: Crash!
14:19:29JdGordonn burn!
14:19:45pondlifeIn the H300 sim, go into the debug menu and press PLAY
14:20:21pixelmaJdGordon: 64-bit warnings
14:20:27pondlifeAlso, I can get stuck in the menus still :(
14:20:35JdGordonpixelma: I know... what cast is needed there?
14:20:54pixelmadon't ask me! :P
14:21:15JdGordonpondlife: dont press play there then :p
14:21:59amiconnpreglow: Maybe we should go for tenth dB and have appropriate step sizes
14:22:01pondlifeSEGFAULT. #0 0x6f696475 in ?? ()...
14:22:01pondlife#1 0x004069d9 in menu_run (m=0) at menu.c:727
14:22:01pondlife#2 0x00403c69 in debug_menu () at debug_menu.c:2395
14:22:21 Join ackbahr [0] (
14:22:24 Join barrywardell [0] (
14:22:49*JdGordon is sleepy
14:22:55JdGordonI typed and instead of && :'(
14:23:11pondlifeNight night, maybe it'll be better in the morning..
14:23:32amiconnpreglow: This would also finally allow to get rid of the val2phys business.
14:23:46preglowamiconn: yeah
14:23:54amiconnIt was just introduced to keep the number of settings bits small, but now we don't use settings bits anymore
14:24:36JdGordonpondlife: fixed...
14:24:43JdGordoncan someone please fix my warnings?
14:24:47*JdGordon falling asleep
14:25:32preglowamiconn: wouldn't you also say that setting lineout volume to 0 db belongs in the codec driver itself?
14:25:36 Quit JdGordon ("Konversation terminated!")
14:25:45preglowright now it's done each time in set_prescaled_volume, apparently
14:27:05preglowbah, the balance logic is also in set_prescaled_volume
14:27:09preglowthere's got to be a better way to do this
14:27:15preglowthat treats all targets the same
14:27:18 Join BigBambi [0] (
14:27:37BigBambiAny gigabeat devs around?
14:30:12BigBambiA general question about button assignments - is the idea to be more per target or more cross platform consistent?
14:30:23amiconnpreglow: Umm, you're right.
14:30:46amiconnBut you can't handle all targets the same.
14:30:52BigBambiI ask as I ave had a H1x0 for ages, and have just got a gigabeat - they have almost the same buttons in the same place, but function is swapped releative to each other
14:30:59amiconnMost have separate left volume and right volume, but the mas does not
14:31:35preglowamiconn: yeah, know, i was talking more code path wise, so that the same functions get called no matter how things need to be exactly called
14:31:53preglowthis would also make it easier to treat the sim as just another case
14:38:13 Join perplexity [0] (n=joust@
14:39:40 Join ackbahr_ [0] (
14:39:47 Join muesli__ [0] (n=muesli_t@
14:40:12 Join Moos [0] (i=5950bebb@gateway/web/cgi-irc/
14:40:42MoosHello there !
14:41:46Moospreglow: KISS for your software tone controller, that's really rocks
14:43:57 Quit jaebird ("Leaving")
14:44:12 Join jaebird [0] (
14:44:31Moospreglow: and about the range -24/+24, I'm with you, that's sound reasonnable for me too, just a little thing to note, when you changes values during playback, there are strange volume up/down, but that's not important, the result sound really good, thanks again
14:44:41preglowMoos: nothing i can do about that
14:44:54preglowMoos: it's a direct result of the mixed hardware/software result
14:44:59preglowehh, s/result/approach/
14:45:17Moosok, that is nothing :)
14:45:33preglowMoos: jhMikeS is talking about implementing low-latency dsp. if that happens, the glitch will pretty much vanish
14:45:59Mooscrazy jhMikeS ;)
14:46:09preglowlike a fox
14:46:54 Join Arathis [0] (
14:49:39 Quit muesli- (Read error: 110 (Connection timed out))
14:49:44 Join muesli- [0] (n=muesli_t@
14:52:07 Quit ackbahr (Read error: 110 (Connection timed out))
14:53:12 Quit ackbahr_ (Remote closed the connection)
14:54:10 Join ackbahr [0] (
14:57:02*amiconn thinks +/-12dB for tone controls are more than enough
14:58:37preglowbut a greater range seldom hurts
14:59:03preglowat least if anyone ask for more than +/-24 db, i'll just tell them to go away, heh
14:59:18pondlifeCan I have +/- 30dB
14:59:23preglowgo away!"¤
14:59:52 Quit muesli__ (Read error: 110 (Connection timed out))
14:59:58preglowamiconn: i used +/-24 dB pretty much just because i could, it didn't change the fixed point format
15:00:15preglowfrom what it would have been with +/- 12 dB, that is
15:00:16 Quit barrywardell ()
15:01:09Moosmore precision (choice) we have, better that is imho
15:01:48Moos+/-24 is perfect here for a big user of basses
15:02:55preglowamiconn: i'll just hack sound.c to set prescaler to 0 for those targets for now
15:03:05preglowand see if i can think up a better way to do things
15:05:37preglowargh!!!! 3 usb hangs in a row
15:07:23preglowsweet lord, this is becoming so annoying
15:08:43***Saving seen data "./dancer.seen"
15:10:06preglowthe bass control really sucks
15:10:16preglowi think i'll use the high cutoff instead
15:12:51LinusNpreglow: usb hang?
15:13:03preglowLinusN: when inserting usb, it just hangs on the logo
15:13:29LinusNplatform? h120?
15:13:44 Quit perplexity (Read error: 110 (Connection timed out))
15:14:05 Join OgMaciel [0] (n=omaciel@ubuntu/member/gnukemist)
15:14:19preglowLinusN: ipod *, i think
15:14:31pondlifeLinusN: Committed that bootloader mod. Time for a new release?
15:14:40preglowLinusN: amiconn thinks it's a thread that doesn't behave
15:14:54preglowLinusN: i vaguely recollect having the same on h120, but might not be
15:15:16pondlifeMaybe related to the disk access problem me and jhMikeS have seen?
15:15:16LinusNpreglow: it usually is a misbehaving thread
15:16:12LinusNpondlife: perhaps
15:17:01pondlifeGo on, make a new bootloader release, then we can close 2 bugs on Flyspray, including a 3.0 one!
15:17:36LinusNpondlife: i think we want to turn off the text output, don't we?
15:17:52preglowamiconn: with no prescaling the bass boost is bloody unusable
15:17:54pondlifePersonally I prefer it.
15:18:06 Quit Hadaka (Read error: 110 (Connection timed out))
15:18:12pondlifeBut if it'll speed up boot then go for it
15:18:56pondlifeTime for lunch, back in 20
15:19:34MoosLinusN: while you are around, any chance you fix the LCD glich that we have at X5 start up?
15:19:54Mooscausing by ane Gigabeat commit iirc
15:23:16LinusNi believe that is fixed in cvs
15:23:40Moosunfortunatly that isn't here with fresh clean svn :(
15:24:08Moosand I don't remenber to saw a commit in this way, but maybe I missed it?
15:24:25LinusNMoos: you see the glitch with the svn bootloader?
15:24:48Moosno, the V2 one, I never build bootloader myself
15:24:54amiconnThe lcd glitch isn't fixed in svn. And it's not only the X5, but also e.g. the H300
15:25:02Moosouch :(
15:25:25Moosisn't just few commit lines to revert?
15:26:03LinusNamiconn: i'm pretty sure i fixed it when i split the bootloader source files
15:26:25MoosLinusN: then a V3 is comming?
15:26:32 Join muesli__ [0] (n=muesli_t@
15:27:02*amiconn isn't sure whether we're talking about the same glitch
15:27:10Moosat start?
15:27:22Moosduring rockbox logo?
15:27:27LinusNhmmm, maybe i'm silly, it's not about the bootloader......
15:27:44Moosisn't about firmware.c?
15:27:59LinusNit's about the startup order in main.c iirc
15:28:27LinusNah yes
15:28:31MoosI'm trying to remenber which commit precisely caused this
15:28:37LinusNso it has little to do with the bootloader
15:28:40Moosfind out?
15:28:59LinusNit's caused by slasheris cosmetic fix for the flashed iriver
15:29:47 Join lucas42 [0] (i=proxyuse@
15:30:00LinusNwhich works fine for the h100 series but bad for those targets that turn off the lcd when the backlight is off
15:31:09 Quit webguest40 ("CGI:IRC (EOF)")
15:31:29 Join ackbahr_ [0] (
15:32:30amiconnThe backlight also behaves ugly on my mini g2
15:32:42amiconnThat's since the same commit iirc
15:32:44 Part Moos
15:33:25 Join Moos [0] (i=5950bebb@gateway/web/cgi-irc/
15:33:37preglowit behaves ugly during shutdown on nano, but i guess that's something else
15:34:35 Join perplexity [0] (
15:35:38 Join soundavinci [0] (i=436769f2@gateway/web/cgi-irc/
15:36:06 Quit muesli- (Read error: 110 (Connection timed out))
15:36:13 Part soundavinci
15:36:30 Join webguest61 [0] (i=436769f2@gateway/web/cgi-irc/
15:36:33LinusNlet's revert that main.c change and let Slasheri do it correctly
15:36:46 Quit ackbahr_ (Read error: 104 (Connection reset by peer))
15:37:08LinusNcommitting in a sec
15:37:27amiconnpreglow: The shutdown ugliness is also present on the mini
15:38:12webguest61Hi, I know it is boring you, but I'm standing just before the question to buy me a 60GB or the 80GB ipod and I want to know that is the actual status on the 80GB version of RockBox?
15:38:46*amiconn points at LinusN
15:38:49LinusNwebguest61: the 80GB is not functional yet
15:38:54preglowamiconn: backlight fades on and off several times here
15:39:30webguest61not yet is OK, but will it or nether
15:40:08preglowit's in the works
15:40:15XavierGrpreglow: didn't you say that in order for rtc to work on H300 Linus has to fix something on the bootloader?
15:40:26preglowXavierGr: yeah, but that may have been done, i can't remember
15:41:09preglowafter that it should just be a case of merging the already very similar drivers and adding a sprinkling of #define HAVE_RTC_ALARM
15:41:57MoosLinusN, amiconn: wa don't have yet RTC_ALARM on X5 yet, right?
15:42:07preglowthe same applies there
15:43:45 Quit ackbahr (Read error: 110 (Connection timed out))
15:44:57LinusNMoos: no
15:45:44LinusNwebguest61: it will be fixed, once i manage to find the time and will to continue working on it
15:46:07MoosLinusN: is it an hardware issue?
15:46:17LinusNMoos: the 80GB?
15:46:38Mooshehe, no the rtc alarm for iaudios :)
15:46:53LinusNit is a bootloader thing
15:47:02Moosah ok
15:47:12LinusNand a simple one too
15:47:31Moosgood to know :-)
15:48:59 Join web-taz [0] (
15:57:47 Quit web-taz (Read error: 104 (Connection reset by peer))
15:57:56LinusNamiconn: ?
15:58:02 Join web-taz [0] (
15:58:53 Join PaulJam [0] (
16:02:52*amiconn wonders whether people will ever learn that sizeof(int) == sizeof(void*) is not always true
16:03:23LinusNah that one
16:03:25LinusN14.24.43 # <JdGordon> can someone please fix my warnings?
16:03:25LinusN14.24.47 # * JdGordon falling asleep
16:03:35 Join desowin [0] (
16:06:12 Join Ribs [0] (n=ribs@
16:06:27 Join Ribs2 [0] (n=ribs@
16:06:29 Quit Ribs2 (Read error: 104 (Connection reset by peer))
16:08:31 Part LinusN
16:11:22preglowamiconn: well, how would it be best to convert from int to void* ?
16:11:50amiconnEither don't use int but intptr_t, or double cast
16:12:37preglowintptr_t would be the best, i don't think i can be bothered to do that big a fix when it's not my bug
16:12:46amiconnThe warning us there to make sure the programmer knows what he is doing
16:13:27preglowdoing a (void*)(intptr_t) works fine
16:13:43preglowshould i jsut commit that for now?
16:14:53 Join petur [0] (i=d4efd6a6@gateway/web/cgi-irc/
16:15:33peturcan somebody confirm that h1x0 recording from digital can do 16bit/48kHz
16:15:37pondlifeIs there any way to perform a 64-bit style compile in a 32-bit environment (i.e to check for these things)
16:15:58preglowpetur: i've done it
16:16:05 Quit GodEater_ ("CGI:IRC (Ping timeout)")
16:16:07peturpreglow: thanks
16:16:10preglowpetur: shouldn't spdif recording do whatever you throw at it?
16:16:31austriancoderpetur: hi.. its me Christian from the mail
16:16:37preglowpondlife: well, i think you could do parts of it
16:16:37peturI wasn't sure, but I thought it should
16:17:10bluebrotherhmm. What's the exact use of intptr_t anyway?
16:17:18pondlifepreglow: I was wondering if there was any compiler option to assume 64 bit?
16:17:21preglowpetur: spdif should be able to do whatever you throw at it, since the clock is taken from the spdif signal itself. the tagged sample rate might be wrong if you use something too esoteric or the word clock is too far away from where it should be, though
16:17:29preglowpondlife: -m64
16:17:39pondlifeWooh, I'll check that with the borked menu.c.
16:17:42 Join perl|work [0] (
16:17:53preglowpondlife: don't expect to be able to link, but compiles should work
16:18:00peturhi austriancoder
16:18:01pondlifeThat's all I need
16:18:01 Quit webguest61 ("CGI:IRC (EOF)")
16:18:37pondlifeNope. My gcc just gives "sorry, unimplemented: 64-bit mode not compiled in"
16:19:00 Quit BigBambi ("Leaving")
16:19:01peturaustriancoder: Sorry to turn the request down, I'm just not up to it atm
16:19:17*petur runs away again - damn work
16:19:23 Quit petur (Client Quit)
16:19:49 Join Nishadham [0] (
16:20:05 Quit lucas42 (Read error: 110 (Connection timed out))
16:22:19 Join GodEater_ [0] (i=c2cbc95c@gateway/web/cgi-irc/
16:23:27preglowamiconn: why do some of the mas chips have bass and treble defaults set to anything else than 0 ?
16:23:44preglowas a matter of fact, it looks like that's the case for all of them
16:24:36amiconnyes it is
16:24:49XavierGryeah I just encountered that with my ondio
16:25:19XavierGrI checked the newly rockboxed booted ondio and it had other values instead of 0 on the bass and treble settings
16:26:23preglowamiconn: but why?
16:27:17amiconnThis is ooold, and afaik it is to make rockbox sound 'better' by default
16:27:36amiconnI guess it just mimics the retailos default
16:27:56amiconnIt's been like that even before I joined the project
16:28:05preglowhm, ok
16:28:22preglowbut yeah, i'm going to make the values tenthdb by default
16:28:34amiconnI have no strong opinion on these defaults, maybe they should just be 0
16:28:36 Join |Rincewind| [0] (
16:29:05preglowi'd say we just keep them as it is
16:29:20preglowif that's what people are used to and all
16:29:23pondlifeNo-one has complained about them...
16:29:26 Quit Nishadham ("Chatzilla 0.9.77 [Firefox]")
16:33:11*GodEater_ wonders if dan_a is paying attention
16:35:02amiconnpreglow: Rather colourful size diff line...
16:36:20GodEater_what's a good EAC equivalent on *nix ?
16:36:26preglowamiconn: heh, yeah, i already knew it'd gain a bit for coldfire but forgot to check arm
16:36:48 Join lucas42 [0] (i=proxyuse@
16:37:13preglowdepends on HAVE_SW_TONE_CONTROLS too, yes
16:37:14amiconnpreglow: Looks like it's a gain for sw tone control targets
16:37:35preglowyeah, they have one filter design routine less
16:37:58preglowi mainly just did that commit for cleaner code
16:38:08preglowhaving two almost identical functions bothers me
16:38:47preglowif i didn't have to keep the crossfeed sound the same as before, it'd be a bit smaller on all targets
16:39:35*amiconn thinks that this delta table is very useful even for targets where we don't need to squeeze out every byte
16:40:16preglowi think the table is great
16:40:29amiconnIt also shows whether only expected changes happen, e.g. when adding swcodec code, archos bin sizes should stay the same etc
16:40:32preglowit's always interesting to see how your code affects each target
16:40:45preglowyeah, even
16:41:47amiconnIt also shows that ucl compression sometimes reacts funny on size changes; sometimes the delta for recorders is bigger than e.g. for fm recorders, or it goes in the opposite direction
16:48:55preglowwhat's the SOUND_LEFT_GAIN and SOUND_RIGHT_GAIN defines in sound.c do? input gain?
16:49:40 Quit scorche (Read error: 104 (Connection reset by peer))
16:50:27amiconnSOUND_LEFT_GAIN, SOUND_RIGHT_GAIN and SOUND_MIC_GAIN are for recording
16:50:30preglowthey and mic gain is everything that's done in val2phys anyway
16:51:07preglowyou're probably going to need to test this patch a lot before i commit
16:51:13preglowcan't test any of this mas stuff
16:52:47 Join topbloke [0] (
16:56:58 Quit austriancoder ("Lost terminal")
16:57:00Moosis someone with svn access can commit this francais.lang update please: ?
16:57:10 Join ackbahr [0] (
16:57:23 Join illriginal [0] (
16:58:06 Join toffe82 [0] (
16:58:57|Rincewind|GodEater: you could try "abcde" I think it has a paranoia mode
16:59:33GodEater_|Rincewind|: In the gap between asking the question and your answer, I've done a little research, and it seems cdparanoia is the way to go for the rip
16:59:48|Rincewind|I haven't done any cd ripping on linux, yet. One of the few reasons to boot into Win again - EAC for ripping and foobar2k and mp3tag for organizing
16:59:57preglowamiconn: we're going to need phys2val if we remove val2phys, won't we?
17:00:01amiconnIf you want it graphical, you could use grip. It has cdparanoia built in
17:00:25GodEater_I'm not bothered about the front end
17:00:36GodEater_I'd personally use extrackt if I was
17:01:13bluebrother|Rincewind|: abcde simply uses cdparanoia (or cdda2wav), so you have everything cdparanoia supplies
17:01:15|Rincewind|GodEater_: Have you find out if cdparanoia runs fast with Plextor drives?
17:01:35GodEater_|Rincewind|: I haven't because I don't own one
17:01:40bluebrotherit does.
17:02:02bluebrotherat least it did with my old plextor burner, but I haven't used that drive for quite a while
17:02:06|Rincewind|in EAC there was this nice option to use the internal plextor C2 checking for secure rip
17:02:12GodEater_this will by first foray into speex - going to be encoding the complete Hitchiker's Guide radio scripts this evening
17:02:33GodEater_|Rincewind|: is that the Q sub-channel stuff to keep track of the TOC ?
17:02:33amiconnpreglow: I don't think so. We should rather add set functions for these, like those we have for the other settings
17:03:03GodEater_|Rincewind|: if so - that won't be in cdparanoia until paranoia IV - due anytime now
17:03:45 Quit ackbahr ("Chatzilla 0.9.75 [Firefox]")
17:03:47|Rincewind|never heard of Q-sub-channel
17:04:03bluebrotherMoos: done.
17:04:31GodEater_|Rincewind|: read the rant at the bottom of the answer
17:05:16amiconnAnd the sound_set_* functions should imho be part of the respective audio codec driver if it's a hwcodec target
17:05:17GodEater_what that all boils down to is "reading CD Audio as data is not easy" (tm)
17:05:18preglowamiconn: yeah, but those set functions will need to convert from physical value to device value anyway, so we're not actually getting rid of any code
17:05:21Moosbluebrother: merci ;)
17:05:53amiconnThis should also allow to redirect them to the sw tone controls for the sim, but with the respective hw tone control ranges
17:06:25amiconnpreglow: Maybe, but at least it gets cleaner on the app layer
17:06:51|Rincewind|reading the rant, I assume that EAC uses Q-subchannel for the Detect Gaps feature, not the actual ripping
17:08:15|Rincewind|"Can cdparanoia detect pregaps? Can it remove the two second gaps between tracks
17:08:15|Rincewind|Not yet. This feature is slated to appear in Paranoia IV." That is almost a show-stopper for me.
17:08:31preglowamiconn: why isn't the mas code moved to the target tree, btw?
17:08:34GodEater_stick to EAC then =/
17:08:46***Saving seen data "./dancer.seen"
17:09:07amiconnpreglow: That would be wrong. The mas code is chip dependent, not taret dependent
17:09:41XavierGramiconn: is there any known limitation as to why other grayscale targets don't have the videoplayer plugin?
17:09:55bluebrotherI ripped all my CDs using cdparanoia and didn't discover any problems
17:10:23GodEater_bluebrother: I'm also not blessed (or cursed) with golden ear drums, so I fully expect it to be fine
17:10:31amiconnXavierGr: The limitation is that swcodec needs libmad linked to the video player, and no-one ever looked into this to make it happen
17:10:52amiconnWett started, but then vanished from irc and didn't publish his wip
17:11:09bluebrotherthe pregap is something the recorder decides to put into the CD structure, so I don't see a reason of removing that −− if they decide to put those gaps in they usually have a reason.
17:11:49GodEater_I'm not sure I even understand what a pregap is
17:11:57XavierGramiconn:what about only rendering the video and ignore the audio stream?
17:12:26bluebrotheryou can even use the pregap to embed hidden tracks ;)
17:12:38|Rincewind|bluebrother: true, on most cds the gaps are intentional, But there are some cds where I prefer to rip without gaps
17:13:01bluebrotherthe CDs I want to rip without gaps are mastered to not include those gaps.
17:13:03GodEater_|Rincewind|: just take them out afterwards with audacity or something ;)
17:13:21bluebrotherSo I don't see a reason changing something the guys that made the CD added intentional
17:13:33 Quit lucas42 ("Leaving.")
17:13:38|Rincewind|well, postprocessing with audacity is overkill and results in much work
17:13:59GodEater_pansy, where's your commitment ? ;)
17:14:18amiconnXavierGr: Should be relatively easy; but the toolchain for making rvf also needs adjusting
17:14:38amiconnDifferent resolutions, and more important, different internal lcd refresh rates
17:14:54|Rincewind|bluebrother: on compilation cds I like to leave the gaps out, because they are not part of the song. And then there is the case when I get a burned cd from a friend with added gaps
17:15:07amiconnIf the refresh rate doesn't match, it starts to flicker horribly
17:15:21preglowwould you say range checking in the set_* functions is necessary?
17:15:33amiconnFor the greyscale ipods the pixel format also needs to be different
17:15:35preglowthey really shouldn't ever be outside the range in the settings struct
17:15:49bluebrother... and possibly these compilation CDs are made from bad mp3 files or whatever ... I prefer the original CDs for maximum quality :)
17:15:56amiconn(horizontal packing instead of vertical)
17:16:35amiconnXavierGr: And btw, greyscale video with no audio isn't really fun.
17:16:36bluebrothertoo much work for audio material that might be of doubtable source quality
17:16:43|Rincewind|true, original is always better, but not always affordable
17:17:52 Join Lynx [0] (
17:18:50bluebrotherbut if I don't know if the source is good I don't want to spend time on quality that might not be present at all
17:19:57|Rincewind|well, putting the cd in the drive, starting EAC, hitting F4 and then ripping isn't much time spent
17:20:38 Quit topbloke ("bye")
17:20:45preglowamiconn: audiohw_set_bass() and co should take tenthdb arguments, or? currentlysome take native args, some take db
17:22:58bluebrotherbooting windows only for the purpose of ripping a CD is wasted time :P
17:24:23GodEater_bluebrother: there's always vmware
17:24:43bluebrotherGodEater_: sure, but that won't work with direct hardware access.
17:25:03bluebrotherfor everything else I have a VirtualBox image running w2k
17:25:09 Quit Lynx_ (Read error: 60 (Operation timed out))
17:25:09 Nick Lynx is now known as Lynx_ (
17:29:16GodEater_I looked at ReactOS today
17:29:25GodEater_nice idea, but far from mature
17:29:57preglowamiconn: everything looks a bit uglier with all the .0 dB now, though :/
17:31:15 Join Febs [0] (
17:32:05Moosbye all, have a good day or whatever
17:32:08 Quit Moos ("CGI:IRC")
17:32:27preglowamiconn: also, wps makes assumptions about how the format of the volume variable, that swine
17:34:19 Quit illriginal ("Leaving")
17:37:33 Join Domonoky [0] (
17:38:47 Quit GodEater_ ("CGI:IRC")
17:43:11 Quit blithe (Remote closed the connection)
17:43:20 Join blithe [0] (
17:48:22preglowvolume fading is in gwps-common.c ...
17:48:49 Join webtaz [0] (
17:56:12 Join CriamosAndy [0] (
17:59:21 Quit desowin (Read error: 113 (No route to host))
17:59:58 Join ackbahr [0] (
18:00:39 Quit PaulJam (".")
18:00:48 Join blueworm [0] (
18:04:41 Quit Criamos (Read error: 145 (Connection timed out))
18:06:17 Quit web-taz (Read error: 110 (Connection timed out))
18:11:13 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
18:11:35 Join desowin [0] (
18:12:52 Part toffe82
18:13:42 Quit ackbahr (Remote closed the connection)
18:19:05 Join Rondom [0] (
18:21:10 Quit blueworm ("Leaving")
18:21:56 Join ackbahr [0] (
18:22:01 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:22:41 Quit linuxstb (Read error: 113 (No route to host))
18:24:54 Quit ackbahr (Remote closed the connection)
18:27:23 Join ackbahr [0] (
18:27:55 Join web-taz [0] (
18:28:08 Quit ackbahr (Remote closed the connection)
18:30:05 Join ackbahr [0] (
18:36:29*amiconn wonders how people are able to apply that kind of raw force to an electronic device :/
18:37:05*amiconn received the defective Ondio he got from ebay
18:37:44amiconnIt's not the kind of defect I expected, but a very different one that leaves raw force as the only possible cause
18:37:44preglowwhat happened to it?
18:38:14 Join datachild [0] (
18:38:15amiconnThe flash rom (which is located "under" the MMC slot) was stripped from the board
18:38:41amiconnNot the mass storage flash but the firmware flash
18:39:07amiconnThe only explanation is that someone tried to force an SD card into the slot, while the manual clearly says it's an MMC slot
18:39:14amiconn(and MMC *only*)
18:39:31pondlifeGo on, push harder, you know it'll fit
18:39:57 Join inversions [0] (
18:39:57amiconnIt's also not the kind of Ondio as advertised on ebay, but a better one
18:40:11pondlifeamiconn - the digital coroner
18:40:12amiconnThe auction said it's an SP, but in fact it's an old-type Ondio FM
18:40:39amiconn(the one with the Samsung tuner and just "Ondio 128" written on the front)
18:41:12preglowamiconn: you think i should make the settings formatter drop the .0 if it's not needed? the volume settings look uglier with all the x.0 dB
18:41:12amiconnIf the board is not too damaged, I might be able to salvage it by resoldering the flash
18:41:25amiconnSomewhat tricky because it's a TSOP32...
18:41:51amiconnEven if that's not possible, it's still a resource for spare parts
18:42:12amiconnOnly cost me EUR1.50 (plus 4.50 for shipment)
18:42:15 Join dpmarsh [0] (
18:42:51amiconnpreglow: Hmm...
18:44:22preglowamiconn: i also hack all the audiohw_set_* to take tenthdb format variables now so most of the conversion doesn't need to take place in sound.c
18:44:55amiconnHmm. Then you could directly link the audiohw_* functions in sound.c for hw controls
18:44:58dpmarshis there anyone here who authorize my TWiki account to be able to upload a WPS?
18:45:08preglowamiconn: no, prescaling
18:45:23amiconnAh, ok
18:45:24preglowamiconn: i probably can for some targets, though
18:45:27 Join Juice^ [0] (n=Juice^^@
18:45:29preglowbut i'll see later
18:45:44amiconnYes, at least for MAS35xxF
18:45:54amiconn(not the D)
18:45:57preglowit's not very sim friendly, though
18:46:01 Quit webtaz (Read error: 110 (Connection timed out))
18:46:29amiconnWe could probably go with a (target-only) #define for certain targets
18:47:47amiconnRe the Ondio: This must be the explanation, because the metal "hood" of the MMC slot is bent as well
18:49:32preglowi bet the seller feels really clever
18:51:51preglowamiconn: dropping .0 for integers looks better when all entries are integers, but kind of messy for mixed lists, like how bass/treble is for wm8975 now
18:52:04preglowamiconn: but i guess that's better than forcing .0 on everything else just because of that one codec chip
18:52:36amiconnYes, probably
18:52:53amiconnWe could go for 0.5dB steps on MAS35xxF
18:52:58amiconn(treble & bass)
18:53:02preglowkinda useless, if you ask me
18:53:05preglowbut why not
18:53:20amiconnIt can even do 0.125dB steps, but that'd be overkill
18:53:46preglow0.5 dB is the very edge case for overkill
18:54:01preglowmost people can't hear 0.5 db changes in treble or bass
18:54:23amiconnBy definition, 1dB changes are "barely noticeable"
18:54:29pondlifeDoes gui_syncsplash display on all screens (i.e. the remotes too)?
18:55:10preglowamiconn: there is something called adaptive bass for 8975, i wonder if that might be better than the db control
18:55:15preglowi think i'll try it out
18:57:15preglowit doesn't clip as much, no
18:57:27preglowthink i'll stick with the other ones, easier to document
18:59:18Learpondlife: that's what the "sync" part means.
19:00:44pondlifeOf course
19:00:47*preglow wonders what the intval parameter to get_tag() in gwps-common.c is for
19:03:37preglowamiconn: stuff like all the tables and tendth2reg() for mas3507d doesn't really belong in sound.c, imo
19:04:35amiconnNo, it would belong to a mas3507d driver
19:04:46amiconnBut such doesn't exist yet
19:05:45Learpreglow: primarily for the "enums".
19:06:21preglowLear: intval?
19:06:50preglowthat's what i have so far
19:07:03preglowonly done volume/bass/treble related stuff yet
19:07:03Learpreglow: yes.
19:07:22preglowwhat are these enums? :>
19:07:49dpmarshis there anyone here who can authorize my TWiki account to be able to upload a WPS?
19:08:48***Saving seen data "./dancer.seen"
19:09:21LearIn a .wps file, you can write something like %?bl<Unknown|Empty|Not empty> (or use images...).
19:09:38LearThat would display "Empty" or "Not empty" for the battery level...
19:09:44preglowthen i wonder what the volume tag uses it for
19:09:49preglowah, right
19:09:50preglowget it
19:10:50 Quit ackbahr (Remote closed the connection)
19:11:44dpmarshVolumne: %?pv<0|1|2|3|4|5|6|7|8|9|10>
19:11:56dpmarshthe returned value is spread across the values between < >
19:12:09 Join nls [0] (
19:12:49pixelmadpmarsh: I guess you are DavidMarsh?
19:14:40pixelmatry now - should work but I've never done that before....
19:14:51dpmarshoh cool. thx. i'll give it a try
19:15:28pondlife"Committing database: [8/7]"
19:15:32 Quit perplexity ("* 'night *")
19:15:53pondlifeThe 7 is hard-coded, is there somewhere else I could get that from?
19:16:05pixelmapondlife: I've already told Slasheri
19:16:26pondlifeYes, but I'm reusing that code for a patch.. just wondered
19:16:45pondlifeGot a nice database auto-making thingy going on here for root_menu.c.
19:17:25pondlifeAnd I'm making these non-LANGed strings LANGed.
19:18:24pixelmaah ok
19:18:41pondlifeActually that one is langed... bad example.
19:19:13dpmarshi think that worked i can edit now. thx
19:20:31pixelmanp - and don't spam ;)
19:23:10nlspondlife: in FS #6691 you asked me to run the sim under gdb. Care to guide me or opint me to some simple instructions, I have never used it before.
19:25:10pondlifeOK. Have you built a sim yet?
19:25:36pondlifeI only have experience under Cygwin too, note.
19:25:43nlssoon :-) building now
19:26:53nlsok, done
19:27:02pondlifeOK, do you now have a rockboxui.exe ?
19:27:16nlsnope, i'm on linux
19:27:31pondlifeAha, so I don't know from here... ;)
19:27:53nlsI have a rockboxui executable tho...
19:28:00pondlifeOK, try this..
19:28:07pondlifegdb rockboxui
19:28:56nlsok, it says (gdb) in the terminal window, but the sim didn't pop up
19:29:03pondlifeNow type run
19:29:19nlsok, seems to work
19:29:28pondlifeOK. Now learn gdb!
19:30:06pondlifeSome useful commands to look up: bt, break, continue, next
19:32:18amiconnNot repairable. 11 out of 32 pads were ripped off the board, there is nothing I could solder 11 of the chip leads to :(
19:32:38pondlifeHmm, violence.
19:33:52dan_aGodEater: You were after me earlier?
19:39:13XavierGramiconn: how much did you paid for it?
19:39:22amiconnEUR 1.50
19:39:37amiconnIt was explicitly sold as "defective, doesn't power on anymore"
19:39:37XavierGrThen I guess it is a "Rest in piece Onio".
19:40:07amiconnI wanted to have a look whether it can be revived, and if not, it's still good for spare parts
19:40:33amiconnI could also take the tuner board and try to convert my SP into an FM
19:40:39pixelmaXavierGr: you mean "pieces" ;)
19:40:53XavierGrdamn I typed piece?
19:41:16 Join Fullmetal [0] (i=551a0c32@gateway/web/cgi-irc/
19:41:26pondlifeIf I add a LANG block to english.lang, I don't need to add it to the other .lang files, do I?
19:41:32amiconnXavierGr: Btw, did you get an old style or new style FM?
19:41:44 Quit web-taz (Read error: 110 (Connection timed out))
19:42:35amiconnHmm, SP->FM conversion isn't that trivial. I would have to swap the MAS as well
19:42:56amiconnI could make a special "Ondio SP with radio"
19:43:24amiconnI could even make it record, but only to wav, not mp3
19:43:24 Quit Fullmetal (Client Quit)
19:43:36|Rincewind|I am trying to use gdb with rockbox sim under Emacs. Does anyone have experience with this?
19:46:26XavierGramiconn: what do I need to look in order to understand if it is an old style FM?
19:47:03amiconnInfo->Debug->View HW info, then check the mask value
19:47:19amiconn0x0708 would be old style, 0x0F08 new style
19:48:08dpmarshpixelma: whatever you did worked, my theme has been uploaded. thanks
19:50:23DataGhosthey how is 5.5G 80GB support for rockbox going?
19:51:04 Join webguest86 [0] (i=5593026f@gateway/web/cgi-irc/
19:52:12 Quit Nico_P (Remote closed the connection)
19:52:12 Join funky [0] (n=repulse@unaffiliated/funky)
19:53:36 Join ackbahr [0] (
19:53:50 Quit ackbahr (Remote closed the connection)
19:54:03preglowisn't 1% resolution for stereo width a bit overkill?
19:54:06preglow5% sounds better
19:54:32 Join ackbahr [0] (
19:54:52 Join z0de [0] (
19:55:58LearSounds resonable, but then I don' use it.
19:56:00 Quit ackbahr (Remote closed the connection)
19:56:38 Quit webguest86 ("CGI:IRC (EOF)")
19:59:32 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
20:00:38pondlifeAnyone mind if I commit a bit of code to initialise the database automatically if the user selects Database from the main menu?
20:01:05Kasperlethink that's a good idea?
20:01:09preglowamiconn: what do you say? is 5% good enough?
20:01:10Kasperlethat takes a couple of minutes
20:01:20pondlifeIt's a background process
20:01:37Kasperlepondlife: but it will not be useful to the user
20:01:39pondlifeJust like the current stuff, but automated
20:01:51pondlifeIt leads them through the process
20:01:56Kasperleunless you output a message "initializing now, check back later"
20:02:24pondlifeI put a counter up displaying progress (# of files) and allow them to exit with the stop key
20:02:48Kasperlethat sounds quite excellent to me, then :)
20:02:55pondlifeThey can come back in and it will show them the counter, then prompt them to reboot if needed.
20:03:08 Join toffe82 [0] (
20:03:24pondlifeI'm hoping we can then scrap the seperate initialise option later, so it won't be pure bloat ;)
20:05:26amiconnSeparate initialise option isn't bloat. It's necessary if you want a clean rebuild
20:05:42|Rincewind|but it should ask "would like to init database now?" fist
20:05:59amiconnAuto-update doesn't remove entries when database is not in ram
20:06:04|Rincewind|I don't want the database created if I accidentally select the wrong entry in root menu
20:06:15pondlifeAs a non-database user you mean?
20:06:15amiconnAnd on some targets, putting it in ram isn't available
20:06:38ompaulpondlife, what is that DB for?
20:06:44*ompaul knows where it is
20:07:10amiconnIn fact, removing stale entries in auto-update only works with both dircache enabled and the db loaded into ram
20:07:21ompaulit is not obvious what I should be using it for - I only have the root menu a few days (iaudio) and no idea what it is for
20:07:32pondlifeDon't worry, I'm not going to remove the init option.
20:07:55pondlifeJust make the process a bit simpler for users, hopefully.
20:08:08|Rincewind|pondlife: yes, I don't use the database at the moment, so I wouldn't want to have it build accidentally
20:08:43Kasperleompaul: it's for browsing your music by artist/album etc, regardless of your filesystem layout
20:08:59ompaulahh nice
20:09:04Kasperleompaul: the information is collected from id3 (and other types of?) tags in the music files
20:09:32Kasperlevery useful if you have an ipod and the music is still in that horrible itunes db format
20:09:39ompaulokay, so if I have the disk space it would be interesting for it to auto gen
20:09:50pondlifeAfter a prompt, yes
20:10:21preglowamiconn: do you think 5% step for stereo width is enough? 1% isn't really necessary, imo
20:10:47amiconnThe database doesn't take that much disk space
20:10:52amiconnpreglow: perhaps
20:10:54 Join bluey- [0] (
20:11:53ompaulis there a "plan" for features on rockbox?
20:11:59ompaulor is it more adhoc
20:12:02XavierGramiconn: old style
20:12:24XavierGramiconn: and btw there is no longer info->debug path
20:12:31XavierGrinfo is replaced with system
20:13:03linuxstb_pondlife: What you're doing sounds good, but I agree with |Rincewind| that it shouldn't be 100% automatic - I also don't use the database, and don't want to accidentally start it initialising.
20:13:20XavierGramiconn: with the root menu patch the name changed
20:13:25amiconnAH, hmm, deutsch.lang still calls it Info
20:13:45XavierGrso is there any drawback that I have the old style ondio FM?
20:14:48amiconnThe new style has a significantly better fm tuner
20:15:34amiconnOther than that, there are no user-noticeable differences
20:15:43XavierGrah that's why the reception is bad on my ondio?
20:15:52|Rincewind|amiconn: In the current svn build I run I have it called "System" in Deutsch.lang
20:16:01 Join Naked [0] (i=naked@
20:16:41amiconn|Rincewind|: Okay, the build on my Ondio is 5 days old, maybe it was changed in between
20:16:54pondlifelinuxstb_: No, I'm putting a yesno up to prompt before init
20:20:26amiconnXavierGr: The old style Ondio FM has the same Samsung tuner as the FM recorder, which is more sensitive to electronic interference.
20:20:57amiconnArchos tried to improve the situation by applying lots of brass shielding
20:21:25pondlifeBrass, not quite as dense as lead
20:21:26 Nick Naked is now known as Hadaka (i=naked@
20:21:31amiconnThe new style has the same philips tuner chip as the irivers, and that delivers clear reception without any special shielding
20:21:47XavierGrok then all that is left for my ondio to be perfected is to flash it :P
20:22:41 Quit Ribs ("eh eh ehhhh!")
20:24:14amiconnYours is like the right one in the 3 side-by-side pcb comparisons
20:26:20nlspondlife: I thik gdb is trying to tell me the sim craches in the synth_full function in synth.c in libmad...
20:26:41nlsline 984
20:26:57amiconnnls: are you on amd64?
20:27:04nlsamiconn: yes
20:27:09amiconnThat's the reason then
20:27:35amiconnRockbox' libmad has an optimisation that is known to break on 64 bit systems
20:28:07amiconnI know where it breaks and why, but fixing it will take some work
20:28:08nlsamiconn: could you add what you know to FS #6691?
20:28:16 Join Llorean [0] (n=Llorean@rockbox/administrator/Llorean)
20:28:23amiconnFixing it without reverting the optimisation, tht is
20:29:08pondlifenls: Did you see that bug where audio output runs at half speed after MIDI has been played?
20:29:41nlspondlife: haven't tested midi placyback, was it on target or in the sim?
20:29:45LearYay! make voice seems to work.
20:29:51pondlifeOnly does it on target.
20:29:59linuxstb_I think midi sets playback frequency to 22.050KHz...
20:30:09nlsok, will try to test later but gtg now
20:30:27pondlifeSomething should reset it to 41kHz ;)
20:30:53preglowor even better, 44.1khz!
20:31:01pondlifeDOH ! Yes :)
20:31:07 Join rp- [0] (n=rp@
20:31:31 Quit rp- (Client Quit)
20:31:38preglowlinuxstb_: yoes, what data sheet did you base the wm8758 driver on?
20:32:08XavierGrhmm my ondio says ONdio 128 on it but not FM
20:32:11 Join rp- [0] (i=rp@
20:32:16 Join gu_ [0] (
20:34:27amiconnXavierGr: Old style FMs are like that
20:34:39amiconnAs long as it doesn't say SP it's an FM
20:35:09linuxstb_preglow: It says in the comment at the top of wm8758.c :)
20:35:26preglowreally now ...
20:35:42linuxstb_Right underneath the lovely ascii-art.
20:35:43preglowbut anyway, if the middle bands don't work, why don't you disable the entire hw eq and set them as bass/treble controls?
20:36:22 Join lee-qid [0] (
20:36:28linuxstb_Yes, that sounds sensible. I think we were initially hoping that they could be made to work, but that time seems to have passed.
20:37:14preglowwell, i'd do a quick disassemble of retailos to see if it touches those regs at all
20:37:17preglowif not, just forget it
20:37:34linuxstb_Although I think the current UI allows more configurablility for each band than just gain.
20:37:49linuxstb_But maybe that's only for the middle 3 (non-working) bands...
20:38:26 Quit XavierGr (Nick collision from services.)
20:38:30 Join XavierGr [0] (
20:38:36XavierGramiconn: when my ondio boot it says archos firmware version 1.31f on the page you linked on the bottom there is something about bricking it if I format it.
20:38:58preglowlinuxstb_: that very same configurability is just hard coded for current bass/treble controls, so i wouldn't see that as a good reason to keep it as is
20:39:02amiconnYes that might happen. Just don't format it before flashing
20:39:50amiconnThe archos firmware < v1.32 has the problem that if it can't read the filesystem, it also doesn't allow usb access to fix it
20:39:54preglowlinuxstb_: looks like the better codec of the wm bunch anyway
20:40:04preglowat least it's capable of prescaling
20:40:06linuxstb_preglow: I never use it, so I'm relaxed about it.
20:40:11amiconn...and it can only boot from FAT16, so if you format it FAT32 -> boom
20:40:18 Quit Domonoky (Remote closed the connection)
20:40:30XavierGrargh why developers are so braindead sometimes?
20:40:30pondlifeIs there a standard way to clear displays but leave the status bar displayed?
20:40:42XavierGrand which is the last firmware version for ondio?
20:41:09amiconn1.32 fixes that problem, but there's no point in flashing that if you're going to fully flash it with rockbox anyway
20:41:41XavierGrit boots slower than H100 with the archos firmware
20:43:00linuxstb_pondlife: I don't think so. But if you want to display the statusbar, shouldn't you call it frequently anyway to update?
20:43:00amiconnYou have 1.31f, the latest is 1.32c
20:43:32pondlifelinuxstb_: It's just that the screen looks odd without it.
20:43:34rp-hi all! stupid question, but what source files are normally involved for working audio on a target?
20:43:55 Quit datachild (Read error: 104 (Connection reset by peer))
20:44:04preglowrp-: alot
20:44:05rp-i'm trying to prepare files for the sansa port...
20:44:30rp-i guess wmcodec is a good start?
20:44:32amiconnXavierGr: AFter flashing rockbox, make sure you'll *never* RoLo archos 1.32b
20:44:55XavierGrI like the legend text on that link amiconn: it says N.F. : New feature
20:45:06XavierGrand there isn't one on all the firmware versions :P
20:45:14 Quit bluey- ("Leaving")
20:45:16amiconnIt's a flash update, and will overwrite the flash without warning. After that, the Ondio wouldn't boot anymore
20:45:55pondlifeEnd Of Line product? ;)
20:46:20XavierGramiconn: I won't flash it until I find somewhere specific instructions, especially if it is brickable if I do something weird
20:46:36amiconnRoLo'ing other versions is okay. I have the latest on on my Ondio in case I want to show someone how the archos fw looks like
20:47:05XavierGrRoLo'ing? you can RoLo from rockbox the original firmware?
20:47:08amiconnXavierGr: I could walk you through if you want. The intsructions itself didn't change, but the files to use did
20:47:33XavierGrwell if you have the time I would most certainly like to
20:47:52amiconnThat's why the rockbox boot files have those "strange" names on archos - the archos in-flash firmware looks for an on-disk update of itself
20:47:53XavierGrjust point me to the page and tell me which files I need
20:48:23amiconnThat's how rockbox works when not flashed - it "mimics" an on-disk archos firmware update
20:49:02amiconnThat's also where the 200KB/400KB limit comes from - the in-flash archos firmware loader won't load firmwares larger than that
20:49:25XavierGrcan't you replace the firmware loader :P
20:49:27amiconnSome v1 recorders and many players aren't flashable
20:49:57amiconnXavierGr: If you flash rockbox, the loader will of course be replaced. But in order to do that, you need rockbox running first
20:50:37rp-dan_a: here?
20:50:40XavierGrso when the ondio is flashed there is no size limit?
20:51:21XavierGrwell lower than the flash size I mean
20:51:39amiconnThe rombox size limit is an entirely different thing
20:51:51preglowlinuxstb_: at least the code looks to be according to the datasheet
20:51:54XavierGrah so you meant the firmware file size limit
20:53:12amiconnWith flashed rockbox, there are 3 parts in the flash: The (tiny) flash loader, bootbox (which is compressed to save space and will be decompressed into ram when selected), and main rockbox, which can either be compressed (then it's decompressed into ram at boot) or uncompressed and then runs directly from rom - rombox
20:53:54amiconnThe flash is 256KB in total, flash loader + bootbox take away 28KB of that (on all archoses except the player)
20:54:02amiconnSo the rombox limit is 228KB
20:55:08 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
20:55:11XavierGrso you use rombox mainly to have more audio buffer?
20:55:14 Join miepchen^schlaf [0] (
20:55:51 Quit bonbonthejon (Remote closed the connection)
20:57:20amiconnAudio buffer size isn't important for battery runtime on Ondio (no spinups), but there are other reasons why more buffer is better
20:57:49pixelmafor example working voice...
20:58:55LloreanThere's a minimum buffer length for working voice?
20:59:15pondlifeSize of english.voice?
20:59:26preglowseems tagcache is a veritable bottomless well of size gains for rockbox :P
20:59:30amiconnLlorean: Sure, the voice file needs to fit in the buffer...
20:59:45pondlifepreglow: Sure is :)
20:59:57pixelmapondlife: which one are you referring too?
21:00:02XavierGramiconn: so to flash it ,I just need a bin file and run it with the flash plugin?
21:00:07pondlifeWait a minute...
21:00:27pondlife...and some more HAVE_TAGCACHE red will appeaer
21:01:03amiconnXavierGr: Yes. And you should dump your rom contents before, in case you want to restore it to factory state later
21:01:04pondlife(Hopefully not too much, and not in the main build table.)
21:01:50XavierGramiconn: I don't want to do that later, and I am sure that if something comes up another rom file could do the job right (from another ondio)
21:02:07amiconnOnly if it's the correct hw mask
21:02:43amiconnXavierGr: The instructions in are correct, just the actual files are not. These should be taken from
21:02:44Lloreanamiconn: Ah, that makes sense. For some reason I'd assumed it could be used in parts.
21:02:57amiconnAll the jabber about testing bootbox can be ignored, it is tested
21:03:32XavierGrso I just download the bootbox
21:03:34XavierGrI see
21:03:53XavierGrwell I will back up the rom just in case
21:03:55amiconnXavierGr: I.e.: Get; , put the 2 .bin files on your Ondio
21:04:07XavierGrbut I don't think there is a chance to revert back, not in a million years
21:04:37XavierGrexcept if archos updates the firmware to make coffee or something like that
21:04:51amiconn...make sure your batteries are in good shape, boot into rockbox, go to System->Debug->Dump ROM contents
21:05:37XavierGryeah I know that part with the flashing procedure on H100
21:05:41amiconn...go to the plugin browser, run firmware_flash.rock, follow the instructions, *don't reboot* afterwards, but browse to /.rockbox and "run" rockbox.ucl,
21:05:48XavierGrI still have the file to revert the H100 to OF
21:06:03XavierGrbut that won't happen again, I am pretty sure
21:06:06preglowi like the fact that i find "OpenGL ES 2.0" listen in the 5g firmware file...
21:06:15amiconn(which will in fact run rockbox_flash) and again follow the instructions
21:06:22 Join linuxstb [0] (i=5343d4aa@rockbox/developer/linuxstb)
21:06:35 Quit kubiix (Read error: 104 (Connection reset by peer))
21:07:07 Join kubiix [0] (
21:07:44XavierGramiconn: If I reboot without updating the flash with the build it will be bricked?
21:08:27preglowkeep this up and even langv2 and new bootbox can't save us
21:08:46amiconnNo it won't.
21:08:47preglownah, it's a nice enough feature
21:08:51***Saving seen data "./dancer.seen"
21:09:12amiconnWhat happens exactly depends on whether the flash image contains a main rockbox image or not; don't remember if it does
21:09:21pondlifeMaybe we could remove the tagcache debug screen? That's less useful now (for users).
21:09:33pondlifeMake it a debug-build only option.
21:09:40preglowpondlife: well, as long as functionality is covered
21:09:49amiconnIf it does, you would boot this (probably horribly outdated) rockbox, if not, bootbox would kick in and load your ajbrec.ajz from disk
21:09:56pondlifepreglow: Have a play and see wha you reckon.
21:10:13preglowi don't really use the database
21:10:18preglowbut perhaps that's an even better reason to try
21:10:23preglowi can play the noob now, heh
21:10:28pondlifeIndeed. Database For Dummies
21:10:44 Quit Rondom (Read error: 113 (No route to host))
21:10:45pondlifeOr maybe you can look at my code and size-optimise it?
21:11:07preglowgot more than enough else to code on, i'm afraid
21:11:07XavierGrok step one done, rom contents are saved
21:11:15XavierGrnow bin file is on root
21:11:19pondlifeJDgordon's warning?
21:11:23XavierGrI will change the batteries and then flash
21:12:16Slasheripondlife: interesting, will try out that now :)
21:12:35dan_arp-: Here now!
21:12:43Slasheripondlife: i was also planning to do something like that but you were faster ;)
21:12:44pondlifeYou can probably improve it. Look at the new routine in tagcache.h too...
21:13:07amiconnXavierGr: Further updates of rockbox will just consist of first roloing the new build (rockbox in flash does *not* check for an on-disk update), then "playing" rockbox.ucl (or rombox.ucl if that will become available again)
21:13:08pondlifeThere's a hard-coded 8 that should probably be calculated, but I've no idea how
21:13:22pondlifei.e Committing 1/8..
21:13:50Slasheripondlife: that is the number of string tags + 1
21:14:13XavierGrhmm I did something wrong, the plugin reported no image found when I opened the bin with it
21:14:17 Quit ender` (" Good programmers treat Microsoft products as damage and route around them.")
21:14:36amiconnRoloing isn't strictly required when not running rombox, but it serves 2 purposes which might hit you otherwise. (1) It's a simple check whether the new build works. (2) The plugin api might have changed, and (new) rockbox_flash.rock might refuse to run with the (old) core
21:14:51Slasheripondlife: the number of entries in sorted_tags[]
21:14:52amiconnXavierGr: firmware_flash is not a viewer
21:15:09amiconnYOu just run it and it should find the bin it needs in the root
21:15:19pondlifeOK, can you update that.. .I need to eat
21:15:23amiconnIf not, you copied the wrong bins
21:15:35amiconn...or renamed them, which you should never do
21:15:38XavierGron iriver h100 it is a viewer
21:15:42 Quit pondlife ("disconnected has pondlife")
21:15:57amiconnrockbox_flash.rock *is* a viewer, firmware_flash.rock is not
21:16:09XavierGryeah different plugins
21:16:16XavierGrbut the habbit remained
21:16:28amiconnThis is on purpose, the plugin needs to decide which of the 2 .bin files is needed
21:16:53amiconnThere are archoses with and without boot rom, and both need different init code
21:16:56XavierGrok programming
21:17:19 Join pondlife [0] (
21:17:27 Quit kubiix (Read error: 104 (Connection reset by peer))
21:17:34XavierGrok flashed the bin, flashed the ucl
21:17:37pondlifeSlasheri: There's a bug in there...
21:17:38XavierGrdo I restart now?
21:18:01amiconnYes you can restart now
21:18:08pondlifeMy test for database not initialised is triggered if you go into the option too soon...
21:18:25amiconnShould boot in ~4 seconds
21:18:39XavierGrnice and clean
21:18:43XavierGrworks like a charm
21:18:43pixelmawith iCatcher it'll take noticably longer...
21:18:50pondlifeI think I should defer the prompt if tagcache_is_initialised is false, right?
21:19:08preglowSlasheri: anything new on the clickwheel stuff?
21:19:11XavierGramiconn thanks for your time, now my ondio is now complete
21:19:58XavierGr5-6 seconds boot
21:20:10Slasheripondlife: don't use that at all
21:20:14Slasheriyou should just check if db is ready
21:20:16XavierGrbut there is no annoying archoslogo screen
21:20:28Slasheriit's always initialized if the thread has been started
21:21:07pondlifeSlasheri: I need to detect 3 states.... (1) Not ready because thread not done enough work (2) Known not to be initialised and (3) Ready for use
21:21:24pondlifei.e. initialized is yes/no/don't know
21:21:26amiconnXavierGr: Boot time depends on what features you're using. E.g. a wps with multiple bitmaps will be clearly noticeable
21:21:28Slasheripreglow: not really, but the gui stuff should be almost commit ready (a few refresh bugs still needs fixing)
21:21:49XavierGramiconn: yes I know, well if you want icatching result you will have to wait in the end
21:21:50linuxstb_Slasheri: Does that include CPU boosting when using the UI?
21:21:52Slasheripondlife: hmmm
21:21:53 Join tucoz [0] (i=5430aebd@gateway/web/cgi-irc/
21:21:59Slasherilinuxstb_: sure
21:22:04XavierGrnot so noticable in everyday usage though
21:22:18Slasherilinuxstb_: but not the clickwheel acceleration itself
21:22:19XavierGrof course a timed benchmark is a different story
21:22:28tucozSlasheri: did you get my message? The change i made to apps/main.c still makes it beep on boot
21:22:35linuxstb_Slasheri: That's fine. I think the acceleration is going to prove hard to please everyone with....
21:22:39XavierGrit is really a shame that only archos supports rvf for now
21:22:51XavierGrit plays really well even on the tiny screen of the ondio
21:22:53Slasheritucoz: ah, yes. that's weird
21:23:06Slasheritucoz: so where did you insert the audiohw_init() call?
21:23:14tucozI have only flashed by running iriver_flash though. You can flash in the bootloader as well right?
21:23:39tucozIn init, right before lcd_init
21:24:08Slasheriso you put it between these two:
21:24:10Slasheri| set_irq_level(0);
21:24:10Slasheri| lcd_init();
21:24:19tucozhmm, let me check
21:24:30Slasheriif not, then it's the wrong init function used for simulators only
21:24:36Slasheriand that shouldn't work :)
21:24:43tucozcould be
21:24:54Slasheritry put it right after kernel_init()
21:25:10tucozah, right
21:25:19tucozi missed the #ifdef simulator :D
21:25:23Slasherihehe :D
21:25:24 Join decayedcell__ [0] (n=decayed_@
21:25:29Slasheriyou are not the first ;)
21:25:47 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
21:25:53tucozOk, then i can try again
21:26:03 Join ender` [0] (
21:26:46tucozI´ll let you know what happens. Thanks for the help so far
21:27:26 Part tucoz
21:27:41pondlifeBack later... got to eat, really
21:27:41 Quit Juice^ (Connection timed out)
21:31:40preglowsurprisingly enough, it seems retailos and bootloaders use the same i2c routines
21:33:42bluebrotherdamn, missed tucoz
21:34:03 Join zomglol [0] (i=54bd7102@gateway/web/cgi-irc/
21:34:31 Quit zomglol (Client Quit)
21:35:25 Quit decayedcell__ (Remote closed the connection)
21:38:29 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
21:38:50 Join decayedcell__ [0] (n=decayed_@
21:38:52 Join miepchen^schlaf [0] (
21:42:06 Part decayedcell__
21:44:48 Join datachild [0] (
21:49:14 Join ques [0] (
21:49:28 Part toffe82
21:50:55 Part ques
21:53:21preglowlinuxstb: unless i'm very wrong, reatilos seems to write to wmcodec addresses that are far higher than any in the datasheet
21:53:30preglowlinuxstb: in addition to lots that are listed
21:54:30linuxstbFor example? (which registers?)
21:54:54preglowi've seen stuff as high as 0x70
21:55:31preglowi've found about fourty calls to wmcodec_write in retailos thus far
21:55:34 Quit Farp (Read error: 104 (Connection reset by peer))
21:55:38preglowamong them are the entire init sequence, i think
21:55:42 Join Farp [0] (i=Farp@
21:58:13 Join entheh [0] (
21:58:27linuxstbWhat are you looking for, the EQ?
21:59:13preglowcould see any of the eq regs touched
21:59:34preglowbut the high addresses i saw from time to time makes it seem something is a bit wrong
21:59:51linuxstbI would be surprised if retailos use it - afaik, it doesn't have any different options to the earlier ipods, so the EQ is most likely just exactly the same sw one.
22:00:40linuxstbThe datasheet was just a pure guess - we (aegray and I) just tried random ones until we got sound...
22:00:46preglowyeah, remember
22:02:06preglowwhat else is on the i2c bus than the pcf and the wm?
22:02:14preglowi think i saw something with address 0x2a
22:05:53DataGhost <- is that a test for 'new wannabe devs' to see if they can fix that?
22:07:05linuxstbpreglow: Isn't the clickwheel accessed via i2c?
22:07:21preglowlinuxstb: ahh, yes, that's probably it
22:08:06 Quit Lear ("Chatzilla 0.9.77 [Firefox]")
22:08:11 Join ackbahr [0] (
22:10:01 Join PhilippeBoineau [0] (i=527be18d@gateway/web/cgi-irc/
22:12:35preglowhaha, wtf, there's a function for 1 << x :P
22:14:49 Quit aleska ("Ex-Chat")
22:16:50 Join stripwax [0] (
22:17:46 Quit ackbahr (Remote closed the connection)
22:19:31PhilippeBoineauHello, my name is Philippe Boineau, I live in France (Paris). I've got an iPod 3G and I use RocBox on it since the 2006 year. As I understand that some help is needed in order to help RocBox to fly on iPod 3G, I would like to be able to contribute to the RockBox Twiki.
22:20:17 Part gu_
22:20:57desowinPhilippeBoineau: have you already registered on wiki ?
22:21:15PhilippeBoineauYes I have
22:21:23desowinwhat user name ?
22:23:11PhilippeBoineauMy username is PhilippeBoineau (as requested) : see PhilippeBoineau">
22:23:33PhilippeBoineauThank you very much desowin
22:23:50desowinyou're welcome
22:24:59preglowlinuxstb: setting one channel inverted and the other non-inverted doesn't exactly seem very intuitive, so either i've got the wrong stuff here, or the datasheet isn't as nice a match as i'd hoped
22:25:09PhilippeBoineauPlease could you tell me how I can check that I'm able to contribute to the wiki ?
22:25:35linuxstbClick on "edit" on any page (and then click on "cancel")
22:26:24desowinPhilippeBoineau: go to page you want add info, and click on edit on the bottom
22:26:52PhilippeBoineauThaks. It seems to be ok. See you later.
22:28:35preglowno, this doesn't make sense at all
22:28:48GodEaterdumb question of the evening : how does rockbox know which codec to use to decode a file with ?
22:29:09hcsfile extension
22:29:22GodEaterspeex should be .spx?
22:29:29linuxstbThe file extension is just the first check.
22:29:44 Quit PhilippeBoineau ("CGI:IRC (EOF)")
22:29:45LloreanCan't it catch speex in .ogg now?
22:29:48hcsthe further checks are up to the metadata reading and then the codec itself, no?
22:30:07linuxstbThe get_metadata() function parses the file (based on the extension, which is used to identify the container) and works out the codec.
22:30:16GodEaterLlorean: I've just tried speex for the first time, and speexenc seems to indicate .spx as the default extension
22:30:28LloreanGodEater: That is the default, yes.
22:30:31linuxstbAnd yes, I think speex in .ogg is supported in Rockbox.
22:30:44GodEaterhowever I'm getting nothing but silence back on rockbox :(
22:30:55 Join ackbahr [0] (
22:31:12linuxstbpreglow: What isn't making sense? You think we've got a wildly wrong datasheet?
22:31:50LloreanGodEater: With a build from two days ago, my speex files still play.
22:31:54LloreanI can't update right now though.
22:31:55linuxstbGodEater: Is the progress bar moving in the WPS?
22:32:10preglowlinuxstb: i think i've misunderstood something and gotten something wrong in my disassembly here
22:32:16GodEaterand time expired is counting up too
22:32:22pregloweven though i don't see what, it seems pretty straight forward
22:32:45preglowlinuxstb: i fell over some mmapping in retailos here, you wouldn't happen to know the layout of the mmap regs?
22:33:14preglowlinuxstb: three out of four mmap reg pairs are set up
22:33:15GodEaterthey definitely decode okay as well - I've checked them
22:33:19linuxstbYou know that the register is in the high 7 bits of the first byte?
22:33:26linuxstbi.e. it's shifted.
22:34:13linuxstbpreglow: Yes, I've looked at the mmap register init in retailos, and don't understand it...
22:35:12preglowlinuxstb: that's done internally in the i2c_write i got, so it shouldn't matter
22:36:02preglowi found three uses of i2c, one writing to address 8, pcf, one writing to address 1a, wm codec, and one writing to 0x2a, probably clickwheel
22:36:35linuxstbAll the PP targets seem to setup the mmap registers the same way, apart from the ipods, which seem to be doing something with the flash. Here are my notes -
22:36:42preglowand the function calling i2c_write with address 0x2a just takes an array to two bytes as argument, almost certainly register address and data, so i don't see how i could have gotten it wrong
22:37:05 Join toffe82 [0] (
22:37:33pondlifeSlasheri: Just committed a mod to check if the tagcache is initialised or not. That's probably not good enought though, can you have a look please?
22:39:10GodEaterLlorean: which version of speex did you use to do your encoding ?
22:39:10preglowah, hmm, one weird thing about the wmcodec_write i think i found is that it only writes bytes, whereas the wm often has more than one byte of data in a register
22:39:37linuxstbThe data is 9-bits IIRC - the second byte and the low bit of the first byte.
22:39:46LloreanGodEater: It was just the official encoder compiled whenever I decided to test it.
22:39:57LloreanGodEater: So, like, whatever day speex entered SVN.
22:40:18LloreanGodEater: Lemme go check version
22:40:33preglowlinuxstb: things are starting to fall in place now
22:40:36GodEater1.1.12 here
22:40:36preglowlinuxstb: i didn't know that
22:40:45preglowthat would explain the high addresses
22:41:15preglowlinuxstb: btw, how do you know 5g does something with the flash?
22:41:25LloreanGodEater: Mine's an old one apparently, 1.0.4
22:41:37LloreanGodEater: I just went and found a precompiled binary, I thought it was up to date, but apparently it's VERY old
22:41:50GodEatermaybe mine's too new
22:42:04GodEaterthe audio thread claims it's doing stuff
22:42:10GodEaterbut it's not making any noise :(
22:42:12 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
22:42:50GodEater1.0.x isn't even available for my distro
22:42:57GodEaterit starts at 1.1.5
22:43:38 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
22:43:54linuxstbpreglow: Did you see that pp-mmap.txt file I just posted a link to?
22:44:18LloreanGodEater: Encoding with a 1.1.12 now.
22:44:40LloreanGodEater: It's a very long file so it'll be like, 45 minutes though. I want to see if the other problems I'm having happen with a new speex encoder.
22:45:05GodEaterwould you care to try one of mine
22:45:11GodEaternot one of them is larger than 2MB
22:45:16GodEaterand most are much smaller
22:45:41preglowlinuxstb: yes
22:45:46linuxstbpreglow: MMAP1 is being used on the video to do the same remapping Rockbox does (0x10000000 -> 0x0). Based on that, I'm guessing MMAP2 is mapping flash (0x0) to 0x20000000. MMAP0 seems to be doing something else with the flash - some kind of map from 0 to 0...
22:46:04 Quit pondlife ("disconnected has pondlife")
22:46:22preglowand here's me hoping for some kind of iram magic :/
22:46:58LloreanGodEater: I'd rather try my own encode. It's not like I'm useful for fixing the problem if there is one, but if it doesn't work at all that'll also show it's an encoder version issue. Since our speex codec is supposed to be up to date with SVN speex, it shouldn't be speex-in-Rockbox being old.
22:46:58preglowdamn, i could have SWORN i had some descriptions of those registers somewhere
22:47:14 Quit ackbahr (Remote closed the connection)
22:47:26GodEaterI just checked, the version in RB is 1.1.14-svn
22:47:33GodEaterso it should decode my stuff :(
22:47:45LloreanWhat sample rate is your file?
22:47:56GodEaterwell it was ripped straight from CD
22:48:05GodEaterso 44.1KHz I guess
22:48:07LloreanSo, 44.1khz?
22:48:14LloreanSpeex isn't really intended about 32khz.
22:48:35GodEaterit clearly worked though
22:48:40LloreanSo, try downsampling to 8, 16, or 32 and see if that works.
22:48:42GodEaterI can decode the files again successfully
22:48:59LloreanYeah, but it might give us some clues. ;)
22:49:19GodEatershould I do the resample directly with cdparanoia ?
22:49:23GodEateror some post processing tool ?
22:49:50LloreanI just used foobar's DSP to convert the .wav to 16khz mono before encoding, back with the one that had worked for me.
22:49:59GodEaterah hang on
22:50:00linuxstb_sox would do the job.
22:50:05GodEaterI seem to recall someone having this problem
22:50:14GodEaterit was because the spx file was stereo
22:50:15LloreanStereo should work now though, at least.
22:50:24GodEaterbang goes that idea
22:50:27LloreanI haven't tested.
22:50:36LloreanPlus, the stereo file played, it just sounded REALLY weird.
22:50:48GodEaterI'm without foobar
22:50:48LloreanIf I recall it was just playing at the wrong sample rate.
22:50:54GodEaterso I'll have to find something to do the job
22:51:05 Quit amiconn (Read error: 60 (Operation timed out))
22:51:06 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
22:54:08 Quit barrywardell_ (Remote closed the connection)
22:54:11GodEaterif anyone has any tips on such a toy for linux - please shout!
22:54:33*linuxstb_ mentions sox again
22:55:00dan_alinuxstb: You say on the IpodFlash page that the Nano flash has 1MB of data in a 512k chip - is the second 512k exactly the same as the first, by any chance?
22:55:32GodEaterlinuxstb_: sorry - missed that the first time
22:55:54linuxstb_dan_a: I'm almost certain it isn't, but let me double-check...
22:56:52dan_alinuxstb_: That's what happens on the YH-820, which also has 512k flash. PortalPlayer's wonderful aliasing in action
22:57:01GodEaterdan_a: while you're here - a quick git question - do you use git-svn yourself ?
22:57:14linuxstb_dan_a: Good call...
22:57:22dan_aGodEater: At the moment I do, but I'm not 100% happy with it
22:57:36linuxstb_dan_a: Yes, the first 512KB are identical to the second 512KB...
22:57:45GodEaterdan_a: nor me - I couldn't get it to only grab the rockbox branch - it insisted on grabbing everything
22:57:53linuxstb_But the other ipods are claimed to have 1MB of flash.
22:58:38dan_aI wouldn't expect them to have the same data in the second half
22:59:20dan_aGodEater: It did that for me too, but that's not really a major issue as long as you have the disk space
22:59:44linuxstb_dan_a: Did you see my notes about the mmap registers? Do you have similar info for other PP devices?
23:00:24dan_alinuxstb_: I don't have it to hand, but I can have a look
23:01:49linuxstb_No rush.
23:02:53GodEaterLlorean: resampled down to 16Khz
23:02:56GodEaterstill no sound output
23:03:00GodEatertrying to go to mono now
23:03:24*Llorean is still encoding, heh.
23:03:36dan_alinuxstb_: I've been trying to get Rolo working properly again for the past few days. Even if I make the main CPU wait until the COP has been put to sleep by the new firmware, I still can't get it to boot properly
23:04:19dan_aAny ideas?
23:04:44 Join muesli- [0] (n=muesli_t@
23:04:45linuxstb_Only what's been discussed before - that it may be expecting certain values in registers and/or iram.
23:05:26dan_aNo, because even Rockbox doesn't load properly (the second core is not active)
23:05:45 Quit rp- ("leaving")
23:06:01linuxstb_Ah, that. I get confused about the current status of rolo - it's constantly changing....
23:06:18GodEaterLlorean: yep - works fine as mono
23:06:31dan_aI still don't even understand why the dual core work broke it
23:06:53LloreanGodEater: That's unfortunate.
23:07:02LloreanI thought it was supposed to be fixed.
23:07:20linuxstb_dan_a: Are interrupts disabled on both CPUs?
23:07:22dan_aLlorean: Rolo or Speex?
23:07:45dan_alinuxstb_: Does set_irq_level(HIGHEST_IRQ_LEVEL) do that?
23:07:48Lloreandan_a: Speex stereo.
23:08:26GodEaterLlorean: also, sample rate doesn't matter
23:08:34GodEaterI just reused the original 44.1Khz wav
23:08:41GodEaterknobbled it down to mono
23:08:44GodEaterand that plays fine too
23:08:52***Saving seen data "./dancer.seen"
23:09:45linuxstb_dan_a: See the definition in firmware/export/system.h - I think that only applies to the current CPU.
23:10:12*linuxstb_ isn't sure and looks towards preglow
23:10:44GodEatertoo late to recode them all now
23:10:48GodEaterI'll do them tomorrow
23:10:52GodEaternight all!
23:11:14dan_alinuxstb_: I don't think so - each core will have its own CPSR (I think!)
23:11:22dan_aGood night, GodEater
23:11:41linuxstb_So is set_irq_level() being called on both CPUs?
23:11:46 Join ]MrFWorK[ [0] (n=1@
23:11:51 Part ]MrFWorK[
23:12:25dan_alinuxstb_: Not in SVN, but in my test code it is
23:13:00dan_a(I'd also love to understand why the codec thread doesn't now run twice as fast as it used to)
23:13:58pixelmadoes anybody know how much audio buffer is available on an FM-Recorder (with an SVN build)? I want to know how large a voice file can be in order to work... maybe someone with an FM-Recorder could check (scorche?)
23:15:01 Quit muesli__ (Read error: 110 (Connection timed out))
23:15:10preglowdan_a: lots of mysteries left with dual core support for sure
23:15:34 Join rotator [0] (n=e@rockbox/developer/rotator)
23:16:04dan_apreglow: Hopefully they'll annoy the people who know how to fix them enough so they get fixed, now it's in SVN ;)
23:18:20 Join web-taz [0] (
23:20:46DataGhosteh LinusN
23:20:52DataGhostwill rockbox boot if it can't mount fat32?
23:21:29DataGhostso I just mounted fat32 on my 80GB :)
23:21:35DataGhostit can't find the .rockbox directory, though
23:21:39 Part perl|work
23:21:42preglowdan_a: you're hoping to have pp employees aboard any time soon? ;)
23:21:51LinusNDataGhost: what have you changed?
23:22:01DataGhosteh. same code I changed in the linux kernel
23:22:06DataGhost'superblock' reading
23:22:16DataGhost2 sectors instead of 1.... really simple actually
23:22:21DataGhostbut it has to happen everywhere
23:22:27dan_apreglow: hahaha - I can hope!
23:22:36DataGhosttoo bad rockbox doesn't use multiple reads like linux does
23:22:45LinusNyes, the reading of the BPB is only one of a dozen things
23:22:47preglowman, that would rock. i'm starting to get really bloody tired of all this guesswork
23:22:47DataGhostin linux it was only the reading of the superblock, that's all sorted now
23:22:58DataGhostbut it does indeed require some more work in rb
23:23:19DataGhostI couldn't get the debugging build to work by the way
23:23:26DataGhostso I have written a basic debugging function myself now
23:23:34LinusNthe linux driver is a lot easier to work with, since it's not optimized for low mem usage
23:23:49 Quit ender` (Read error: 110 (Connection timed out))
23:23:59DataGhostand the hardest part isn't even the code
23:24:28DataGhostI write my code on this pc, I save it over samba to my server, then I scp the changed files to my laptop, build it there, scp rockbox.ipod back to my server and then I load that on my iPod
23:24:52DataGhostbecause I don't want to screw up my arm toolchain with yet another version
23:25:15LinusNDataGhost: i don't get it
23:25:28DataGhostnever mind :P
23:25:39linuxstb_You can just install another arm toolchain with a different −−prefix, and modify your path.
23:25:49LloreanOr just use the VMWare image for complete separation.
23:26:00DataGhostyeah I know but crosstool is easier and I'm lazy
23:26:16DataGhostand I can't easily change the prefix in there
23:26:20linuxstb_Easier than running the script?
23:26:27LinusNthen it can't be easier :)
23:26:29linuxstb_(edit the prefix at the top)
23:26:30DataGhostcrossdev even
23:26:46DataGhostoh yeah I actually ignored that script
23:26:50DataGhostI should have looked into it :D
23:27:08dan_aDataGhost: Crossdev is easy to have two tools - you just run gcc-profile to choose between them!
23:27:23DataGhostyeah but the other 2 I have weren't installed with crossdev
23:27:25pixelmaLinusN: do you have an FM-Recorder? I want to ask the guy who's made the new files if he could also provide a smaller one which could work on the other Archos targets as well... I'd like to give him a number of the maximum size
23:27:28DataGhostI'm already using a PATH= for the kernel
23:27:51LinusNpixelma: i have an fm rec
23:28:02linuxstb_DataGhost: Has work been abandoned on getting 2.6 working on the ipod?
23:28:18 Quit JdGordon ("Konversation terminated!")
23:28:19DataGhostI only heard that it existed, never cared much about it
23:28:27pixelmaLinusN: could you test with an SVN version how much buffer is available atm?
23:29:22pixelmaehm... should have mentioned voice files before ;)
23:29:34LinusNpixelma: hang on
23:29:37linuxstb_DataGhost: I noticed some CVS changes to the kernel - does that mean IPL fully supports the 80GB now? i.e. as well as the other 5gs?
23:29:52DataGhosti submitted my patches :)
23:29:55 Join Thundercloud_ [0] (n=thunderc@
23:30:40DataGhostthat's why I'm looking into rockbox now
23:30:43linuxstb_The first IPL kernel commits since last August...
23:31:17DataGhostfunny thing is, after that crappy kernel I first 'released' everyone started using it on the 30GB winpod, which was then the only one that would fully work with that specific release
23:31:39DataGhostafter some time and peer-review we decided to take out some 'redundant' code, actually breaking 30GB support, heh.
23:32:03XavierGrpixelma: with which engine are these voice files generated?
23:32:26pixelmaXavierGr: it's in the mail archive
23:32:46dan_aI'd love to get this bug where if you boot up an iPod with the USB connected it hangs at the USB screen fixed. Can give me a quick overview of what happens when the USB is plugged in?
23:32:46XavierGris it a commercial one?
23:33:21 Join petur [0] (n=petur@rockbox/developer/petur)
23:33:25linuxstb_dan_a: It also freezes for me if I plug it in soon after startup, if the disk is still spinning (probably the dircache thread).
23:33:32pixelmadon't know - but there are now up in the wiki
23:33:38XavierGrbecause I have some nice Loquendo voice files that I could share, but iirc they are commercial and thus can't distribute them
23:33:44XavierGrlet me check then
23:34:10DataGhostdid you get to that already LinusN? or didn't you spend that much time on this issue?
23:35:58LinusNas a test, i patched the fat/ata driver to be able to mount the file system
23:35:58linuxstb_DataGhost: I guess you're loading Rockbox with the IPL bootloader?
23:36:11 Quit web-taz ("( :: NoNameScript 4.02 :: )")
23:36:15LinusNa long time ago
23:36:26DataGhostyes linuxstb
23:36:46DataGhostoh wait. the rb loader needs fixing too, right?
23:36:57linuxstb_Because it's displaying the hw info correctly. I never bothered implementing that on the 5gs, as it's not used anywhere...
23:37:00DataGhostah ok LinusN... too bad :P
23:37:10linuxstb_hw revision
23:37:13DataGhosthw info is set by the bootloader? :o
23:37:32linuxstb_Yes, but that could be changed now. It's read from a location in flash IIRC.
23:37:34DataGhostI know that it reads it but this is new to me :)
23:38:01linuxstb_And normally flash is hidden "underneath" sdram. But we've recently made it available at a different address in Rockbox.
23:38:37preglowlinuxstb_: i don't use dircache on my nano and it still hangs like a madman
23:38:46linuxstb_The only place it's actually used in Rockbox is in the Color/Photo LCD driver, to distinguish between the two LCD revisions.
23:39:03 Join Nico_P [0] (
23:39:40linuxstb_preglow: All I've noticed is that it reliably freezes if I connect whilst the disk is still active after boot. If I wait until the disk stops, and then connect, it reliably works...
23:40:35dan_aI'm guessing that something is waiting on the disk before it acks the USB insert, and that the USB insert disables the disk
23:41:03dan_alogf to the rescue, I hope!
23:41:38linuxstb_What about the ata callback? Would Rockbox be attempting to write the settings?
23:41:57jhMikeSoy...the gigabeat is gonna make optimizing difficult in general for arm... need to think of something
23:42:09 Join surrealisticpill [0] (n=surreali@
23:42:20dan_aOh, it's one of those annoying bugs. A logf build fixes it
23:42:37dan_aThe ATA callback sounds like a good suspect, though
23:42:54 Part surrealisticpill
23:43:07markunjhMikeS: you can try to clock it down to 17MHz :)
23:43:26jhMikeShehe...says it's already running that :D
23:43:26linuxstb_Great idea - get libmad working then...
23:43:48linuxstb_jhMikeS: You have a gigabeat?
23:43:59jhMikeSlinuxstb_: yes...just came in
23:44:00 Quit Thundercloud (Read error: 110 (Connection timed out))
23:44:02markunjhMikeS: there is a 'slow mode' which does that
23:44:09markunjhMikeS: congrats!
23:44:20linuxstb_Which model?
23:44:26jhMikeSnice player with a really great display
23:44:48*linuxstb_ can only see F10s and F20s on ebay at the moment (available to the UK)
23:44:49markunI wonder how much more F40's there are out there
23:45:08markunlinuxstb_: toffe bought another 2 broken F40s :)
23:45:49LinusNpixelma: Buffer: 1.292MB
23:46:06jhMikeSI got it from toffe and it works like a charm. Didn't have time myself to assemble a good one from broken ones.
23:46:16pixelmaLinusN: thank you!
23:46:33dan_ajhMikeS: Treat yourself to a 3G iPod!
23:46:49Kasperleand pwn random boxen with firewire:D
23:46:59 Quit funky ("leaving")
23:47:06jhMikeSFor the ultimate challenge with a broken cache implementation? :)
23:47:32toffe82I don' know I should receive one fridy but it is for XavierGr and I think I will recieve the 2 others next week
23:47:48toffe82sorry wrong windows
23:48:06peturLinusN: just replaced hdd and battery on my h340 and now the backlight of my lcd no longer comes on. lcd works. any idea?
23:49:13jhMikeSThe F40 does chow down on the SPC playback without even flinching :)
23:49:18 Part Llorean
23:49:59LinusNpetur: hmmm, hard to tell
23:50:43 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
23:51:12LinusNpetur: have you tried to run the OF?
23:52:37peturnot yet
23:52:40LinusNdo so
23:53:17LinusNi had a similar issue when i started the h300 port ages ago, and running the OF fixed it
23:53:34jhMikeSdan_a: You've got an inexpensive one? Like I said I'll take a working one that someone wants to unload.
23:53:44peturhmmmm are we missing some init code then?
23:54:11DataGhosthm. so after it's read the partitiontable (and presumably mounted the fat partition) it'll only read 1 sector from sector -1 LinusN :o
23:54:20LinusNi thought i fixed it, but let's hope i didn't :-)
23:54:46LinusNpetur: because then running the OF will help you in this case :-)
23:55:08dan_ajhMikeS: I want to keep mine, but I plan to get some dead ones off ebay and do some voodoo on them
23:55:15LinusNDataGhost: it reads single sectors from all over the place
23:55:29DataGhostat least not through ata_read_sectors
23:55:33 Quit juxtap (Read error: 110 (Connection timed out))
23:55:37DataGhosthold on
23:56:04jhMikeSdan_a: keep me posted
23:56:06peturLinusN: BINGO!
23:56:38LinusNso we still have that backlight init bug... :-(
23:57:03peturI'm not running a recent bootloader, only the latest from the wiki
23:57:12jhMikeSany of this ata stuff relevant to recording sticking while building the database? I'm not following too well here.
23:57:19petur(version 5)
23:57:31LinusNthat bug is much older than bootloader 5
23:57:39LinusNit's been there forever
23:57:51peturwell it's not fixed... and it is only the backlight
23:57:58DataGhost LinusN
23:58:02LinusNbetter look into it then
23:58:15DataGhostand from there on I haven't seen another read but 1 -1
23:59:00dan_ajhMikeS: I'll keep you posted, and I don't think the ATA stuff I'm looking at will affect recording, but I can't be sure

Previous day | Next day