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

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

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

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

#rockbox log for 2009-05-23

00:02:21 Join robin0800 [0] (
00:02:36 Join mak_tn [0] (n=23HA1901@
00:02:48 Part mak_tn
00:06:46 Quit amiconn (Nick collision from services.)
00:06:49 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
00:07:08 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
00:07:28 Quit pixelma (Nick collision from services.)
00:07:30 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
00:07:47 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
00:18:05CIA-38New commit by 03peter (r21045): Update Dutch language file. Not overly happy with the translation of this, though...
00:18:32*petur thinks the performance of cia-38 is impressive
00:20:34JdGordon|it get it before svn ci exited?
00:21:28AlexPyeah, it is rather speedy
00:21:34saratogayeah i think its hooked in as soon as the SVN connection is initiated
00:22:00peturtime between commit and message was < 1s
00:22:11saratogabefore the changes are transmitted, at least i've seen it ping the channel before a commit finishes
00:32:33 Quit matsl (Read error: 110 (Connection timed out))
00:33:54 Quit merbanan (Read error: 60 (Operation timed out))
00:39:34 Quit perrikwp (" ajax IRC Client")
00:44:38CIA-38New commit by 03mcuelenaere (r21046): Lua: ...
00:45:34 Join matsl [0] (
00:46:24 Quit n1s ("Lšmnar")
00:46:57*Llorean doesn't like commit messages that don't say anything
00:47:32AlexPBad CIA-38!
00:47:46AlexPmcuelenaere confused it by including too much information :)
00:48:09mcuelenaerehèhè, /me guesses CIA doesn't like newlines :)
00:48:27Unhelpfulthere's more in the log, but yes, CIA seems to trim at the first newline :/
00:48:46AlexPThat's why I never use them
00:49:03AlexPOr spaces, or in fact punctuation of any kind, just to be sure
00:49:05 Join kachna [0] (
00:49:27 Quit itcheg (" ajax IRC Client")
00:49:40Unhelpfulyou should contract words in non-standard ways as well, beacues i'm pretty sure it trims on character count, too.
00:52:33AlexPUnhelpful: Good plan
00:53:29 Quit robin0800 (Remote closed the connection)
00:53:36 Quit petur ("Zzzzz")
00:54:26Unhelpfulthat way you can make sure that even if you radically alter nearly every part of rockbox, wcnstllrdthcmmtmsgnhr.
00:54:46 Quit JdGordon| (" ajax IRC Client")
00:56:25 Quit ender` (" /[,\(\s]?#1(s)(\s+-?\d+(?{}\.\d+)?\s+-?\d+(?{}\.\d+)?\s+-?\d+(?{}\.\d+)?\s+-?\d+(?{}\.\d+)?\s+-?\d+(?{}\.\d+)?\s+-?\d+(?{}\")
01:02:54linuxstbpetur: (Sorry I'm a bit late to ask this, but it wasn't mentioned in the commit message or visible on the front page) Why commit dbestfit as part of the PDa plugin? If it's in Rockbox at all, wouldn't it be better in the plugin lib?
01:08:30 Quit kachna (Remote closed the connection)
01:08:36 Join kachna [0] (
01:08:46 Join faemir [0] (
01:10:54 Join hd [0] (
01:11:13 Quit HellDragon (Client Quit)
01:11:54domonokylinuxstb: as far as i know, he needs a full blown memory manager, (ie buflib isnt enough). So dbestfit is better as something selfwritten :-) Dont know if moving it into the plugin lib is a good idea (we dont want to encurage use of this, when not really needed).
01:12:10 Quit hd (Read error: 104 (Connection reset by peer))
01:14:01 Join Tyr [0] (
01:14:23Tyrwow lots people in here
01:14:41domonokyand maybe he can modifiy the code, so we could remove dbestfit later again.. :-) But thats something to decide for wincent_balin itself.
01:14:47CIA-38New commit by 03mcuelenaere (r21047): Set svn properties
01:15:09AlexPdomonoky: himself surely - we dont rank students quite that low (yet) :)
01:15:39domonokyups.. :-)
01:16:02 Quit Tyr (Client Quit)
01:16:41 Join HellDragon [0] (
01:18:16saratogadomonoky: we've actually got a patch sitting on the tracker for adding dbestfit to the codeclib
01:18:39saratogaFS #9885 - Add malloc library to codeclib
01:20:23 Join codemonkeyandrew [0] (
01:20:33 Nick codemonkeyandrew is now known as andrewbeveridge (
01:20:39domonokyah, didnt know about that. Then it might be better to finish this task, and remove dbestfit from PDa.
01:20:47 Quit andrewbeveridge (Client Quit)
01:20:59saratogacan codec and plugin code be shared?
01:21:02Unhelpfulbuflib does as much to manage memory as you can do without virtual memory. the major barrier to using it in imported code is that it does *not* have an API anything like malloc.
01:21:30 Join andrewbeveridge [0] (
01:21:45Unhelpfulsaratoga: surely the same methods used for pluginlib scaler/bmp/jpeg would work for importing items from the codec lib...
01:21:53 Part andrewbeveridge
01:23:09Unhelpfulbuflib can compact free space, while dbestfit can't possibly do that - *but* dbestfit gives you a pointer to your allocation that will never change until you free it.
01:24:32saratogadoes pda have a way to flush the malloc buffer? thats the only thing I couldn't figure out in the codeclib
01:25:34Unhelpfuli'm sure if you free al allocations you can get back to one contiguous block, if that's what you mean...
01:26:19saratogai needed someway to reset the malloc on codec changes so that a leak in one codec coudln't crash a different codec later on
01:26:40Unhelpfulactually, buflib can now make *one* allocation with a fixed address, via the buffer_in/buffer_out method
01:27:32domonokysaratoga: woulnt it be better to check for leftover buffers and warn after a codec exits ? Then those leaks could get fixed :-)
01:27:33Unhelpfulsaratoga: you might need to look at how that particular malloc initializes, and add a function to reset its state, if it doesn't have one
01:28:13saratogai tried adding such a function but never got it right
01:28:38saratogai don't see any sense in looking for leaks in the codecs, if I'm going to do that I might as well start rewriting them to not use malloc in the first place
01:29:06saratogaleaks are harmless with the current malloc anyway
01:29:27saratoga(since we don't even have a free!)
01:29:41Unhelpfulthat's not always practical, if they have a certain maximum total allocation, but it's split different ways depending on the file...
01:32:17domonokyand couldnt those leaks build-up if you play many files with the same codec ?
01:33:20 Join andrewbeveridge [0] (
01:33:28 Part andrewbeveridge
01:33:32Unhelpfuli think a malloc reset between files, even with the same codec in use, might be a very good idea...
01:33:48saratogadomonoky: no codecs reset their mallocs on track change
01:33:56saratogathey have to since we don't actually have a free()
01:34:16***Saving seen data "./dancer.seen"
01:34:28domonokyah, now i understand..
01:34:48*Unhelpful thought you just said you were never able to work out a malloc reset...
01:35:39saratoganot for a real malloc
01:36:04saratogafor the current one its really easy, you just reset the malloc pointer
01:36:25saratogaand since this is what free looks like: #define _ogg_free(x) do { } while(0)
01:36:27saratogait works fine
01:37:16saratogabut for dbestfit trying to reset the thing typically crashed on track changes
01:37:39 Join andrewbeveridge [0] (i=586d65ae@gateway/web/ajax/
01:38:13Unhelpfuli seem to recall that the "init" function for dbestfit just adds a new chunk to the buffer?
01:39:15saratogaUnhelpful: yes it does, but readding the same chunk does not work
01:40:03Unhelpfulright, because the chunk is already in whatever list dbestfit keeps... where does it track that? perhaps we just need to be able to reset that list?
01:40:57 Part andrewbeveridge
01:41:39saratogait uses some sort of tree structure that I didn't understand
01:43:02*domonoky would think that to reinit dbestfit it would be enought to reinit top[].
01:43:15*mcuelenaere wonders how hard it would be to get splashf to learn \\n
01:46:06CIA-38New commit by 03mcuelenaere (r21048): Lua: add stack traceback when erroring out
01:51:54domonokywincent_balin: we are thinking about moving dbestfit from your plugin into pluginlib/codeclib as others could benefit from it too (see FS #9885) what do you think ?
01:52:37wincent_balinI say yes. Less for me to take care of :-)
01:52:56domonokyor, more... someone has todo the neccessary work :-)
01:53:13 Join |unjustice| [0] (
01:53:28 Quit HellDragon (Success)
01:54:13|unjustice|I am trying to format my 4G ipod mini with two partitions, but I am not sure how big to make them it says (1-1032) when I try fdisk
01:56:13gevaertsyou could download the mbr from and see what's there
01:56:57domonokywincent_balin: did you already use dbestfit successfully in PDa ? i only see it with #if 0 in your patch.
01:58:00wincent_balinThe next step is to run the messaging system in the allocated memory.
01:59:16|unjustice|I already tried to autodetect my ipod and it did not work becaus of HFS+, but in order to make it FAT32, I have to partition the hd for a filesystem...not sure how big to make the partitions, I erased the pre-existing ones
01:59:47Llorean|unjustice|: Follow the link gevaerts pasted
01:59:53 Join andrewbeveridge [0] (i=586d65ae@gateway/web/ajax/
02:00:28Unhelpfulwould it perhaps be possible, just for the targets that *have* an MMU, to make use of virtual memory for the buffer a malloc implementation uses? that would all a real/useful free
02:00:55 Join HellDragon [0] (
02:01:57 Quit froggyman (" ajax IRC Client")
02:01:59andrewbeveridgehi, anybody here willing to guess why I can't complete my wiki registration? clicking the verify link takes me to a 403 Forbidden page
02:07:11andrewbeveridgeactually, could anybody please click this link for me, just to save my sanity and confirm there is somthing wrong with the rockbox TWiki?;code=AndrewBeveridge.660054076
02:07:16|unjustice|Llorean: that link is not helpful if I cannot write to the ipod
02:07:58AlexPandrewbeveridge: Can you sign in and edit pages?
02:07:59 Join Dauron [0] (
02:08:00|unjustice|Llorean: I need to know how to partition a filesystem prior to adding the mbr thing, I cannot seem to mount the device
02:08:51linuxstbThe mbr _is_ the partitioning. It's simply written to the first sector of the disk. You don't need to (or want to) mount it before doing that.
02:09:35andrewbeveridgeAlexP: nope
02:10:14|unjustice|linuxstb: but I need a filesystem so that I can write a partition table
02:10:33|unjustice|linuxstb: FAT32, this just says plug it into a windows box with itunes
02:10:37|unjustice|but I do not have that
02:10:37AlexPandrewbeveridge: You'll have to ask the wiki admins - that'd be Bagder probably. Unfortunately it is currently 02:10 in the morning, so they are probably asleep
02:11:32andrewbeveridgeAlexP: ah, ok. no problem. Can you confirm that it is a problem with TWiki, and not just me?
02:11:49gevaerts|unjustice|: I think you're seriously confused about what filesystems and partition tables are
02:11:51AlexPandrewbeveridge: I can't confirm it but it seems likely
02:12:16|unjustice|gevaerts: doubtful
02:12:36andrewbeveridgeAlexP: ok. I already emailed but it's been a week so I guess I should just come back here at a less antisocial hour and hope Badger or somebody else can help me
02:12:36AlexP|unjustice|: You haven't siplayed any understanding yet
02:12:50 Join hd [0] (
02:12:53Llorean|unjustice|: Since a filesystem is unnecessary to write an MBR directly to the disk, I'm going to have to agree with gevaerts and AlexP
02:12:58|unjustice|gevaerts: the problem I have is this: there is no /dev/sdN, so I cannot move the mbr to that device
02:13:04AlexPandrewbeveridge: I'll try to remember to prod him if I see him :)
02:13:13|unjustice|the link you suggested recommeds I use FAT32
02:13:14domonoky|unjustice|: trust us, a mbr and partitioning is different to filesystems :-)
02:13:18 Quit d3v14710n ()
02:13:20|unjustice|no doubt
02:13:26|unjustice|never thought they were the same
02:13:32|unjustice|these are 2 separate problems
02:13:54Unhelpful|unjustice|: there are no disk devices *at all* for it?
02:13:57Llorean|unjustice|: A lack of /dev/sdN is completely different than a lack of filesystem, and unrelated.
02:14:15LloreanWell, the lack of filesystem might be a symptom, but not the cause
02:14:22AlexPYou do not need a filesystem to write a partition - in fact, without a partition you cannot have a filesystem
02:14:37andrewbeveridgeAlexP: ok. I already emailed but it's been a week so I guess I should just come back here at a less antisocial hour and hope Badger or somebody else can help me
02:14:37andrewbeveridgeAlexP: thanks =)
02:14:47AlexPIt not appearing in /dev is completely unrelated
02:14:55AlexPandrewbeveridge: no worries :)
02:15:22 Quit HellDragon (Read error: 60 (Operation timed out))
02:15:24|unjustice|last night I mounted it as /dev/sdc to moutn point /mnt/ipod
02:15:24AlexP|unjustice|: What does dmesg say when you plug it in?
02:15:37andrewbeveridgei was about to mention dmesg
02:15:55AlexP|unjustice|: You don't want to mount it to write a partition table, you just need to know the device
02:16:04AlexPWhich dmesg should tell you
02:16:48|unjustice|it is sdb\
02:16:56*andrewbeveridge wishes he could type faster
02:16:59|unjustice|but I cannot mount it, it prompts me to give it a filesystem
02:17:09LloreanYou don't need to mount it.
02:17:10gevaertsof course you can't mount it
02:17:16 Join _Auro [0] (
02:17:32AlexP|unjustice|: At least four different people have now told you that you don't want to mount it
02:17:46AlexP|unjustice|: I suggest you read that page you were linked a good while ago
02:17:50|unjustice|I did
02:17:55|unjustice|4 times
02:17:58|unjustice|I will read it again
02:18:05AlexPclearly not very thoroughly
02:18:13gevaerts|unjustice|: in that case, seriously, find a windows machine with itunes and let that do the job
02:18:33Llorean|unjustice|: I thought you said the device is /dev/sdb
02:18:41LloreanScrolled up on accident, ignore me
02:19:10Llorean|unjustice|: You'll notice the first step on that page for the manual instructions is to *unmount* your device. your inability to mount it means you can skip that step
02:19:32AlexP|unjustice|: It nowhere says to mount it - the very first step is to *unmount* it
02:19:43 Quit domonoky (Read error: 104 (Connection reset by peer))
02:20:00andrewbeveridgeunjustice: i would suggest you unplug your device, plug it back in, and do dd if=/dev/zero of=/dev/sdb bs=512 count=1
02:20:10andrewbeveridgeto start afresh
02:20:24AlexPandrewbeveridge: As long as it reappears on /dev/sdb :)
02:20:42andrewbeveridgeAlexP: aw who cares whats the worst that could happen? xP
02:21:07AlexPandrewbeveridge: Well we shouldn't give people activly bad advice :)
02:21:22AlexPIf he just reads that page and does what it actually says he should be fine
02:21:44andrewbeveridgeAlexP: aw shucks. ok now im gonna go check the irc log for the page url, i'm intrigued
02:21:52 Quit mcuelenaere (" - Chat comfortably. Anywhere.")
02:23:58andrewbeveridgejust thought of something i meant to ask a while ago; any idea who chose the distro for the VMware devel platform image?
02:24:14andrewbeveridgeor, what thought process was used to choose it?
02:24:30AlexPlots of the devs use debian
02:24:46Lloreanandrewbeveridge: It was what the person working on it decided to pick.
02:24:51gevaertsMost probably the person who made the image just chose what he prefers
02:25:03andrewbeveridgeyeah, thats what i assumed.
02:25:05LloreanAs far as I know, I don't think there was really much debate or anything
02:25:10AlexPnone at all
02:25:35AlexPIt just happens that many here favour debian (or debian derivatives), and the person that made it choose that
02:25:54|unjustice|I find this URL vague and not helpful, since the first step recommends unmounting
02:25:57|unjustice|which I cannot do
02:26:05AlexPSo don't
02:26:11andrewbeveridgeits just that for new linux users, that image is probably very scary.
02:26:12AlexPIt isn't rocket science
02:26:33AlexP|unjustice|: First step is unmount, you are already not mounted, job done
02:26:34 Quit _Auron_ (Read error: 110 (Connection timed out))
02:26:34 Quit matsl (Read error: 110 (Connection timed out))
02:26:37 Nick _Auro is now known as _Auron_ (
02:27:03Llorean|unjustice|: I already told you that you can skip that step earlier, anyway.
02:27:09AlexP|unjustice|: Seriously, if you are struggling with this find a Windows machine and itunes restore. It will be easier for all of us
02:27:32andrewbeveridgeAlexP: was that addressed to me? if so, I know it isnt. but i also know if i was somebody who had no experience of linux it would be a lot harder to handle than, say, an opensuse image or possibly ubuntu
02:27:48AlexPandrewbeveridge: No, not at you
02:27:59|unjustice|it was at me
02:28:10 Quit hd (Client Quit)
02:28:11|unjustice|"xxxx" means what?
02:28:14Lloreanandrewbeveridge: Someone who has no experience of Linux should probably be able to manage to learn the minimal knowledge needed if they're planning on developing Rockbox
02:28:18gevaertsandrewbeveridge: this is a bit off-topic, but it isn't any harder to use than those things you mention
02:28:31|unjustice|linux is easy to use
02:28:45AlexPandrewbeveridge: They are given instructions though - it isn't meant as a "proper" VM, just as a linux environment that they can follow instructions in
02:28:54AlexP|unjustice|: It is the mbr that you downloaded
02:29:09|unjustice|okay, so I put that filename into the command?
02:29:16AlexP|unjustice|: yes
02:30:02AlexP|unjustice|: You saw the line that says "Replace mbr-xxxx.bin with the name of the MBR file you downloaded in step b). " I assume?
02:30:10|unjustice|I did
02:30:22AlexPSo what is unclear about that?
02:30:34andrewbeveridgeok, i guess i understand that logic. and to be honest, if somebody wants to develop rockbox, its gonna help if they are happy with linux anyway.
02:31:23AlexPandrewbeveridge: I don't see why this choice is any harder than any other given its function (or just anyway)
02:31:24andrewbeveridge"mbr-xxxx.bin" means "a file starting with "mbr-" and ending with ".bin"
02:31:47AlexPandrewbeveridge: Even if he didn't get that I think "Replace mbr-xxxx.bin with the name of the MBR file you downloaded in step b)." is fairly clear
02:32:20andrewbeveridgeAlexP: oops i didn't read that. sorry ><
02:32:32|unjustice|yes, I understand that now
02:32:46 Quit KBH (Read error: 110 (Connection timed out))
02:34:26andrewbeveridgeAlexP: I guess i measure "difficulty" of distro on a "similarity to MS windows" scale, for most new users. and Debian is pretty low on the scale compared with a couple others. mainly because of the lack of KDE (or even Gnome)
02:34:50gevaertsplease, this is getting far off topic
02:34:54Lloreanandrewbeveridge: But Rockbox development *requires* terminal use.
02:34:59AlexPthat is a choice for the VM to keep the download size down, nothing to do with Debian
02:35:09|unjustice|so, how do I figure out the name of the ipod?
02:35:15LloreanIn that context, "difficulty" is basically equal for the compiling part, and everything else can be done outside of the VM
02:35:31AlexPandrewbeveridge: but yes, #rockbox-community is good for stuff like this
02:35:44Llorean|unjustice|: For which step?
02:36:05andrewbeveridgeokies, topic dropped (it was reallt just a whimsical thought anyway)
02:36:07|unjustice|um...several anytime it says diskN after unplugging
02:36:56Llorean|unjustice|: As it says at the top of the page, those are Mac instructions.
02:37:09|unjustice|I am not getting any output from the first command that contains the name of the ipod
02:37:14gevaertsHave a look at the IpodManualRestore page. That will probably help
02:37:16Llorean/dev/diskN is the Mac equivalent for /dev/sdN
02:38:08|unjustice|I guessed that, but the sdb command does not work, perhaps it was a typo, I will double check
02:38:26 Join SirFunk__ [0] (
02:39:00 Join kuber [0] (
02:39:06andrewbeveridgeunjustice: pastebin the output from "dmesg | tail" after plugging it in?
02:39:37kuberhm. i have an ipod video black 80gb... I can't seem to play m4a files, though the manual lists them in 'Supported audio formats'.. when i go to files -> and try to play it just skips to the next file
02:40:08|unjustice|newfs_msdos is not a command
02:40:17Lloreankuber: What SVN revision are you running?
02:40:35Llorean|unjustice|: That's not the first step you should be doing
02:40:43kuberLlorean: i just grabbed the current build link , r21048
02:40:44|unjustice|oh? what should I be doing?
02:41:04Lloreankuber: Are you sure they're not encrypted? Did you encode them yourself?
02:41:10kuberyeah, encoded with faac
02:41:12 Quit Dauron (Connection timed out)
02:41:15Llorean|unjustice|: You should be doing them in order, not skipping to the filesystem one
02:41:29Llorean|unjustice|: Or you could try the link at the top of the page, the one with the red text, that says it's for people using Linux...
02:41:44Lloreankuber: It may be a metadata issue then. Do they have embedded album art?
02:42:04kuberno, this was just straight from a flac ripped to from cd then to faac
02:42:12|unjustice|ahhhh MUCH better thank you
02:42:26Llorean|unjustice|: You still need to do other things *before* making the filesystem
02:42:32LloreanPlease don't try to do things out of order
02:42:53kuberwhen i try to play i get a now playing screen momentarily, with "TB\n(root)\n(root)"
02:43:05Lloreankuber: Unfortunately, you seem to have some sort of problem file. You'll need to figure out what's different about it, probably
02:44:17kuberhm. i wonder where i can find a reference file that should play
02:44:38JdGordonso, I've just come up with the wierd idea to write a function to put touchable buttons on screen (for touchscreens...) 1) do I bother making it multi-screen aware? I mean its probably not going to be wanted to be shown on remote screens, and 2) how programatically customizable do I make it? right now im only thinking text for the buttons...
02:45:24 Join HellDragon [0] (
02:45:24 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
02:45:33gevaertsfor what context is this?
02:45:39Lloreankuber: Files encoded with iTunes should play. Usually it's a metadata issue, so you could find a tool for manipulating MP4 containers and play with it
02:46:53LloreanJdGordon: I'd say touch buttons kinda need to be able to use images as well.
02:47:22LloreanSome things work much better as icons than text, since icons don't need to resize to fit translated strings
02:47:23JdGordonyeah, but not just yet I think... eventually icons would be nice
02:47:54LloreanWell, you were asking whether text was enough. :-P
02:48:46LloreanI'd also say try to make it multi-screen aware. Who knows when we'll get a DAP with the second screen being the touchscreen, or with multiple touchscreens. If possible, I'd say try to plan for this possibility (within reason if it really makes it far too complex, of course)
02:49:08LloreanI can pretty easily see a DAP with a touchscreen remote happening.
02:49:39JdGordonyeah, um... what I meant was do I let things specify where on the screen each button goes? or just the viewport to put them in and make a nice row of buttons?
02:50:03JdGordonand by multi-screen aware I meant draw it on both screen... the mr500 already has a remote lcd
02:50:15JdGordonbut I doubt we'd ever get a remote lcd touchscreen
02:50:37LloreanI think a non touchscreen main unit with a touchscreen remote isn't too unlikely.
02:51:18JdGordonactually... hmm.. this could work anyway
02:51:41JdGordonseen as each screen would need to have the area set differently anyway
02:53:16 Quit SirFunk_ (Read error: 110 (Connection timed out))
02:54:26 Quit barrywardell (Remote closed the connection)
02:56:56JdGordonhmm... I could easily use the icon themeing system to do this... but that would eat up RAM (I'd do it from a seperate file so it wouldnt touch non TS targets)
02:57:41*JdGordon remembers his old icon patch still in the tracker which was going to be used to clean icons up if/when this sort of thing was ever needed
02:57:52JdGordononly it was done to make the statusbar themeable :D
02:58:32JdGordonyike! the patch is from 11-11-07!
02:59:34dfktis there any simple album art downloader that saves a folder.jpg for rockbox? i know about AlbumArtAggretagtor and Album Art Downloadr XUI - both are pretty bad for my needs
03:01:34|unjustice|mkfs.vfat -F 32 /deokay, so I do not have sdb2, I am trying to do this: mkfs.vfat -F 32 /dev/sdX2
03:01:46|unjustice|not sure how to go about it
03:02:29 Quit krazykit ("Connection reset by beer")
03:03:02 Join Thundercloud [0] (
03:04:08linuxstb|unjustice|: Did you write the mbr successfully?
03:04:32linuxstbAnd then "hdparm -z /dev/sdb" ?
03:05:27linuxstbTry unplugging your ipod, then reattaching it. If you don't get a /dev/sdb2 then, then the mbr wasn't written.
03:06:56|unjustice|still not there
03:07:04|unjustice|so what? I just do it again?
03:07:28linuxstbYes, try the "dd" command again.
03:07:36linuxstb(with the mbr)
03:07:39 Part kuber
03:07:55andrewbeveridgedfkt: mediamonkey? or, mediamonkey in wine?
03:08:46 Join krazykit [0] (
03:13:01|unjustice|umm...should rbutilqt-v1.2.1 autodetect the ipod now?
03:16:14linuxstb|unjustice|: Have you succeeded in restoring it?
03:19:31 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
03:19:58|unjustice|yes, all the steps were completed and sdb 1 and 2 are both in /dev
03:20:21linuxstbDoes the Apple firmware now work?
03:21:37|unjustice|it already worked
03:21:55gevaertsthat doesn't answer the question
03:22:03|unjustice|should I unplug and check?
03:22:20|unjustice|doesn't work
03:22:23|unjustice|wrong mbr?
03:23:58linuxstbDid you dd the "Firmware-X.Y.Z" file?
03:24:31 Join krazykit` [0] (
03:24:40gevaertsdid you dd the file you downloaded, or did you unzip it first?
03:24:41|unjustice|it just keeps showing a folder and an exclamation point
03:24:41 Quit Thundercloud (Read error: 104 (Connection reset by peer))
03:24:48|unjustice|unzip it first
03:27:04 Quit HellDragon (Read error: 54 (Connection reset by peer))
03:27:17 Join HellDragon [0] (
03:30:23linuxstb|unjustice|: Either you did something wrong (try doing it all again), or there's a hardware problem with your ipod...
03:31:43 Join lee321987 [0] (
03:32:27lee321987LUA scripting language on RB: What does the file extension have to be, and can it compile .rock files?
03:33:20lee321987*file extension of an LUA script.
03:34:17***Saving seen data "./dancer.seen"
03:34:19Unhelpfullee321987: "compile .rock files" is not what a scripting language does, surely?
03:34:34linuxstblee321987: "lua"
03:36:19 Quit krazykit (Read error: 110 (Connection timed out))
03:36:48 Quit Lss (Read error: 104 (Connection reset by peer))
03:38:01 Quit |unjustice| (Read error: 113 (No route to host))
03:38:31 Join _unjustice_ [0] (
03:39:02_unjustice_BLKRRPART failed: Device or resource busy I get this error now that I am retrying this command "hdparm -z /dev/sdb"
03:44:51lee321987RB VMware image: How do I _cleanly_ shut down?
03:46:05linuxstb_unjustice_: Is the ipod mounted?
03:46:24_unjustice_lol no
03:46:54 Quit bertrik (Read error: 113 (No route to host))
03:52:59 Quit HellDragon (Client Quit)
03:53:36 Join KBH [0] (
03:53:43 Nick JdGordon is now known as JdGordon|afk (n=jonno@rockbox/developer/JdGordon)
03:54:35 Join _lifeless [0] (n=lifeless@
03:56:34lee321987Whenever I try to shutdown the RB VMware image, it gets to the line "Power down.", and just sits there like that.
03:57:15 Join HellDragon [0] (
03:57:28_unjustice_linuxstb: nope
04:03:12 Quit moos ("Rockbox rules the DAP world")
04:11:42 Quit __lifeless (Read error: 110 (Connection timed out))
04:22:31 Part wincent_balin ("Kopete 0.12.7 :")
04:24:41 Join ipodvideo5g [0] (n=639dd867@gateway/web/cgi-irc/
04:24:55 Quit lee321987 ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
04:25:49ipodvideo5gi cant connect to my ipod
04:26:04ipodvideo5git just goes directly to ok to disconnect
04:27:07andrewbeveridgeno need to shout
04:27:07andrewbeveridgehave you mounted it?
04:27:18andrewbeveridgewait, is this in windows or linux?
04:28:40ipodvideo5git is windows it was working fine ive been using rockbox for about a year now...
04:29:09krazykit`what revision are you using?
04:29:22ipodvideo5gi beleive it might be related to the md5sum...cause stupid me i was wondering what the plugin i ran it
04:29:34_unjustice_BLKRRPART failed: Device or resource busy I get this error now that I am retrying this command "hdparm -z /dev/sdb"
04:29:56andrewbeveridgegoogle it?
04:30:20ipodvideo5gi am using version 3.2-090323
04:30:25krazykit`_unjustice_, it sounds like it's still mounted or something.
04:30:36krazykit`ipodvideo5g, then you're using apple's firmware for usb, which rockbox doesn't affec
04:31:07_unjustice_when I try to unmount, it says device not mounted, however it shows up when I list mounted devices
04:31:32krazykit`so run "sync", wait, and just unplug it
04:31:46ipodvideo5gkrazykit, i dont understand your response
04:32:23krazykit`ipodvideo5g, if you're using the 3.2 release, when you're in a usb transfer mode, you're not running any rockbox code was sdb2
04:33:45ipodvideo5gok so what should i do?
04:34:14andrewbeveridgehave a look in device manager?
04:34:14andrewbeveridgesee if it is listed
04:34:38andrewbeveridgeif it isn't, and other USB devices work fine, then there is something faulty with your iPod, I would guess.
04:35:36ipodvideo5gno it doesnt list at all..i have tried it on both my laptop and my desktop and it reboots when i connect the usb then shows the do not disconnect...then immediately ok to disconnect
04:36:13andrewbeveridgewell, that it nothing to do with rockbox. try it in safe mode or whatever its called with iPods
04:36:27 Nick fxb is now known as fxb__ (
04:37:07 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
04:37:08andrewbeveridgei think you have to hold down two of the buttons on the ipod while you switch it on, or plug it in, or something
04:38:35ipodvideo5gyes thats what i have always doen but when i do that it freezes in ok to disconnect
04:38:49ipodvideo5gits the select and play
04:39:18ipodvideo5gbefore this it didnt require the 2dle button hold
04:39:34ipodvideo5g*double button button hold
04:41:00andrewbeveridgedoes the ipod work fine otherwise? try plugging it into a linux system and check if dmesg reports anything
04:41:55ipodvideo5gyes the ipod works fine
04:41:59_unjustice_ I followed these instructions all the way through and I cannot get my ipod to start it only gets to a folder with an exclamation point, I still want to install rockbox, is this a sign that the ipod is broken?
04:42:23ipodvideo5gwhat is the everything.md5sum file for can i delete it
04:44:21krazykit`it's what happened when you ran the md5sum plugin.
04:44:30krazykit`you can safely delete it
04:49:29ipodvideo5gi dont understand how... i have googled like crazy and found that many have gone thru this yet no one has posted how to fix it
04:50:28_unjustice_does anyone know how to get vfat onto the ipod in linux, so that I can install rockbox?
04:50:44krazykit`_unjustice_, yes, you follow the wiki page you linked. that's the method.
04:51:25_unjustice_okay, but it doesn't work
04:51:47_unjustice_the ipod works with hfs+, which means it probably isn't hardware
04:52:21_unjustice_krazykit`: every command on that page works fine, but the ipod still does not start up
04:52:26krazykit`that suggests that you're doing something wrong
04:52:56_unjustice_I am copying and pasting the instructions\
04:53:02_unjustice_into terminal
04:53:16andrewbeveridgeunjustice: would you be willing to run a vnc server so i can do it for you?
04:54:02krazykit`_unjustice_, you're substituting the proper filenames and device names in the commands, right?
04:54:17_unjustice_sdb and
04:54:49andrewbeveridgeunjustice: is that a yes or a no?
04:57:22_unjustice_andrewbeveridge: I would prefer to understand it first hand\
04:59:01andrewbeveridgeok, well you would if you saw me do it. i think too many people have tried and failed to explain what to do. and if there is something unusual going on which is causing the instructions to fail, it would be a lot easier to figure out if I could *see* what was happening
04:59:16_unjustice_you want pastebin?
05:00:04andrewbeveridgewhat are you on about?
05:00:27_unjustice_none of the commands give errors, so I am not sure what is going on. I tried this manual restore page, and I cannot get vfat onto the ipod
05:00:36_unjustice_the partition table worked fine
05:01:11_unjustice_sdb, sdb1, sdb2 are all there...but I cannot seem to get the ipod to start up
05:01:27andrewbeveridgeright, ok, so from this point as far as i'm concerned the only way i can help you is if i can access the terminal
05:01:38 Quit miepchen^schla (Read error: 101 (Network is unreachable))
05:01:38_unjustice_what would you do?
05:03:29andrewbeveridgefirst i would ask you to unplug and re-plug in the ipod. then i would do "dmesg | tail", process the output to see if anything unusual was happening. if not, i would then do a 'dd if="/dev/zero" of="/dev/whatever" bs=512 count=1'. then i would start fdisk and make some partitions.
05:04:17_unjustice_that is funny
05:04:24_unjustice_I tried each of those in that order
05:05:27andrewbeveridgewell, if it is funny, and you are unwilling to let me try and help you by accessing the terminal myself, there is no point me being here
05:06:25_unjustice_basically, I do not see how granting you access to my terminal will help me, if you plan on doing things that I have already attempted
05:07:09_unjustice_that, and I am not sure I can trust that you have my best interests in mind, as was suggested to me earlier by someone in this room
05:07:32andrewbeveridgeclearly, if you had done the above and succeeded, you would have what you wanted right now. so the fact that you are still here means something must have gone wrong. i have no way of knowing what exactly went wrong unless i can see the commands, and outputs, myself.
05:07:59_unjustice_but you can see them without terminal access
05:08:22andrewbeveridgeyou want to copy and paste outputs from commands into pastebin for me to read? it would be so much easier for you to run a vnc client.
05:09:44_unjustice_I would rather do it myself because if someone else fixes my problem then I learn very little about the nature of the error and how to fix it
05:10:01_unjustice_*poof* it works
05:10:20 Quit HellDragon (Connection timed out)
05:10:35andrewbeveridgeuh, have you learned anything in the past hour or two? many very intelligent people have tried to help you. you would see anything i did, and i would happily explain anything
05:11:03_unjustice_everyone directed me to the link
05:11:06_unjustice_it was helpful
05:11:22_unjustice_however, I am left with the same problem I entered with, an ipod with partitions but no vfat
05:11:27andrewbeveridgeok... so what are you saying? you have got it working now?
05:11:44andrewbeveridgewell, what did you create the partitions with?
05:11:56_unjustice_then, hdparm
05:12:00andrewbeveridgei mean right now
05:12:18krazykit`_unjustice_, you didn't dd the partition table over?
05:13:14_unjustice_krazykit`: if that is on the link, then yes...
05:13:16krazykit`that wiki page doesn't mention fdisk at all
05:13:20_unjustice_I kow
05:13:24_unjustice_*I know that
05:13:58andrewbeveridgewhat is your aim? i must have missed your initial request. i thought it was simply to partition an ipod
05:15:41 Quit efyx_ (Remote closed the connection)
05:15:48_unjustice_I was partitioning with fdisk, I wanted to know the sector size, I was redirected to the wikipage, which had an alternative method for partitioning. fine. I used the mbr-mini2g-4g to partition and I get 3 (sdb, 1 and 2)
05:16:01_unjustice_BUT, I cannot get vfat onto the ipod
05:16:18_unjustice_which is necessary for rockbox installation
05:17:20krazykit`so you dd'd the firmware over as per the instructions, making absolutely sure you have the right firmware file?
05:17:47andrewbeveridgelook, fdisk cannot create filesystems. use parted for that.
05:18:08krazykit`andrewbeveridge, have you even read what he's said? you're entirely off base here.
05:18:31 Join taylor_ [0] (
05:18:39 Part taylor_ ("Leaving")
05:18:46andrewbeveridgekrazykit`: clearly, the instructions haven't worked for him. i thought he was simply trying to create a vfat filesystem on a partition on an ipod
05:19:34krazykit`_unjustice_, can you please paste your exact dd command for copying the firmware?
05:20:11ipodvideo5g have updated my itunes thinking that maybe somwhow the drivers would help me enter diskmode...but nope no cigar
05:20:50andrewbeveridgekrazykit`: his initial message was "I am trying to format my 4G ipod mini with two partitions, but I am not sure how big to make them". how was i off base?
05:21:16krazykit`andrewbeveridge, it sounds like confusion over how to turn a macpod into a winpod at this point.
05:21:56andrewbeveridgekrazykit`: would a dd if=zero then fdisk + parted not solve that?
05:22:04_unjustice_sudo dd if=Firmware- of=/dev/sdb1
05:22:05_unjustice_8999+0 records in
05:22:05_unjustice_8999+0 records out
05:22:05DBUGEnqueued KICK _unjustice_
05:22:05_unjustice_4607488 bytes (4.6 MB) copied, 5.78446 s, 797 kB/s
05:22:10_unjustice_sorry I didn't pastebin
05:22:12krazykit`andrewbeveridge, not if _unjustice_ wants to install rockbox.
05:23:20krazykit`_unjustice_, you downloaded the wrong firmware file. that looks like the "4g iPod", rather than the "mini 2g"
05:23:20andrewbeveridgekrazykit`: uh... why? surely then he would just follow the usual instructions for installing rockbox on it? what am i missing?
05:23:55krazykit`andrewbeveridge, you're missing the fact that the device currently doesn't work?
05:24:48krazykit`_unjustice_, on the firmware download page, look in the dropdown box for "Mini 2g (7/1.4.1)", since you said you have a 2nd gen ipod mini
05:25:07andrewbeveridgekrazykit`: ugh. i'm confused. i shall drop it as i am cluttering up this channel
05:25:29ipodvideo5gandrew: could you help me
05:26:21ipodvideo5gi am still unable to connect to my ipod
05:26:51_unjustice_krazykit`: okay, I already did that
05:27:30andrewbeveridgeipodvideo5g: please? =P sure. though it isn't rockbox related so we should switch to #rockbox-community. if device manager isn't recognizing your ipod at all, no matter what buttons you press or hold or whatever, then all i can recommend is plugging it into a computer running linux and telling me what dmesg says.
05:27:53ipodvideo5gok how do i switch
05:27:58krazykit`_unjustice_, now, once you dd the correct firmware over (this new one), you should be able to follow the rest of the instructions successfully
05:28:18andrewbeveridgeipodvideo5g: well, you joined this channel ok ;-) (this is #rockbox )
05:28:30_unjustice_can you send me a link? I am selecting the same firmware
05:28:36_unjustice_it is just replacing the old file
05:28:37andrewbeveridgeipodvideo5g: type /join #rockbox-community
05:29:44ipodvideo5gok here i go
05:29:50ipodvideo5g /join #rockbox-community
05:30:14krazykit`ipodvideo5g, without the space at the beginning of the line.
05:30:18andrewbeveridge... no spaces
05:30:18andrewbeveridgei mean at the beginning
05:34:18***Saving seen data "./dancer.seen"
05:39:04 Join AndyI [0] (i=AndyI@
05:43:50 Quit bubsy (Read error: 60 (Operation timed out))
05:45:17andrewbeveridgehow does the rockbox database read meta info from tracks managed by iTunes?
05:45:34krazykit`by id3 tags and whatever metadata aac files use
05:45:57Unhelpfulthere are mp4 atoms for metadata
05:46:18andrewbeveridgedoes iTunes store tags in the actual files then?
05:46:33andrewbeveridgei thought it stored them in its own database
05:46:52andrewbeveridgei've never actually used it, just seen friends and clients ipods.
05:46:57UnhelpfuliTunes has a database, but i'm pretty sure tracks from iTunes should already have proper metadata. tracks from elswhere, that becomes your problem.
05:48:45andrewbeveridgeah, ok. i assumed if it had a database of its own on the iPod, and it moves all tracks to funky locations and filenames, then it would wipe the tags in the files.
05:49:23 Quit tchan (Remote closed the connection)
05:50:01 Join tchan [0] (n=tchan@lunar-linux/developer/tchan)
05:50:19andrewbeveridgeok, second question. if i got hold of an ipod, plugged it into a linux box, zeroed the whole thing with dd, then created a single partition + vfat filesystem with parted, would it work?
05:50:30Unhelpfulwhy do that? it's complicated, and the space savings would be miniscule, and actually almost zero in terms of actual filesystem space saved (since files are allocated in whole block)
05:50:42Unhelpful...i wouldn't do that if i were you.
05:52:02 Join safetydan [0] (n=Dan@rockbox/developer/safetydan)
05:53:37andrewbeveridgeUnhelpful: huh? who mentioned space saving? I was just clarifying that what i was trying to get _unjustice_ to do was correct. Am I right in saying that would work for any iPod, no matter what setup was on it before?
05:54:15safetydanHow does dbestfit compare to dlmalloc.c that's part of the LUA plugin?
05:55:36krazykit`andrewbeveridge, no, you're not right. the ipod would not function as a music player and likely would reboot into emergency disk mode
05:55:44Unhelpfulandrewbeveridge: "i assumed if it had a database of its own on the iPod, and it moves all tracks to funky locations and filenames, then it would wipe the tags in the files" <- and i was pointing out that no purpose could be served by this
06:18:21 Join ipodvideo5g2 [0] (n=639dd867@gateway/web/cgi-irc/
06:41:55 Quit ipodvideo5g2 (Excess Flood)
06:42:08 Join ipodvideo5g2 [0] (n=639dd867@gateway/web/cgi-irc/
06:42:24 Quit ipodvideo5g2 (Excess Flood)
06:42:52 Join ipodvideo5g2 [0] (n=639dd867@gateway/web/cgi-irc/
06:47:54ipodvideo5g2ok its there
07:22:19 Join AndyIL [0] (n=pasha_in@
07:33:14 Quit ipodvideo5g2 ("CGI:IRC (EOF)")
07:33:16 Part andrewbeveridge
07:33:17 Quit AndyI (Read error: 110 (Connection timed out))
07:34:21***Saving seen data "./dancer.seen"
07:34:26 Quit ipodvideo5g ("CGI:IRC (EOF)")
07:39:50 Join Horschti [0] (n=Horscht@xbmc/user/horscht)
07:57:25 Quit Horscht (Read error: 110 (Connection timed out))
08:01:40 Quit _unjustice_ ("KVIrc 3.2.4 Anomalies")
08:10:01 Join petur [50] (n=petur@rockbox/developer/petur)
08:24:22CIA-38New commit by 03peter (r21049): Patch from Wincent Balin: Remove some duplicate code from the PDa plugin
08:26:40CIA-38New commit by 03peter (r21050): Patch by Wincent Balin of PDa plugin: update readme and disable shared libraries
08:32:30 Join robin0800 [0] (
08:39:53 Join Rob2222 [0] (
08:44:04 Quit Foxx ()
08:53:57 Quit bzed (Read error: 60 (Operation timed out))
08:54:00 Join bzed [0] (
08:58:05 Quit Rob2223 (Read error: 110 (Connection timed out))
09:01:24 Quit robin0800 ("No Ping reply in 90 seconds.")
09:02:00 Join robin0800 [0] (
09:14:36 Quit petur ("*real life*")
09:20:27 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:34:25***Saving seen data "./dancer.seen"
09:37:38 Quit spyder (
09:37:38 Quit Galois (
09:37:38 Quit yosafbridge (
09:37:38 Quit planetbeing (
09:38:34NJoinspyder [0] (
09:38:34NJoinGalois [0] (
09:38:34NJoinyosafbridge [0] (
09:38:34NJoinplanetbeing [0] (
09:51:11 Join bertrik [0] (
09:55:16CIA-38New commit by 03unhelpful (r21051): Load PictureFlow logo from disk, discarding it after splash screen is done and freeing the used space for the slide cache.
10:00:34 Join flydutch [0] (
10:01:58CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
10:01:58*amiconn was thinking about adding the overlay loader capability (actually just handling a non-standard load address) to the core plugin loader
10:02:49amiconnThis would allow to get rid of the separate overlay loader plugins, and make better use of available ram for "jumbo plugins" (they won't be "overlays" anymore)
10:03:03 Quit Zarggg (Read error: 60 (Operation timed out))
10:07:04 Join FlynDice_ [0] (
10:08:11amiconnThose plugins would no longer need to be loaded before the standard plugin buffer, but they could overlap that buffer
10:09:24FlynDice_what's the best way to pass an argument to a short assembly routine in a .c file?
10:09:59amiconnIt would also allow to ask the user whether he wants to stop playback (localised and with voice support) when such a plugin is about to be loaded
10:10:27amiconn(Voice will only work on swcodec, obviously)
10:11:20Unhelpfuli've summed up >8KB so fare of stuff i'm fairly sure is not needed during cache build :)
10:12:32Unhelpful... plus the ~4.5KB stack for the PF loader thread
10:14:30Unhelpfuli like the chaining idea. it might even be of value on non-tiny-buffer targets, like the color targets with huge screens (and huge cache items)
10:17:25amiconnThe jumbo plugin idea would also save a bit of diskspace (less files - saves 1 cluster per .ovl), and reduce confusion for developers (no duplicate plugin names in apps/plugins/)
10:18:13Unhelpfulwhere are these buffers laid out? is the audio buffer immediately before the plugin buffer?
10:19:08amiconnChaining might be useful for PF, but the jumbo plugin idea is independent of that (there is more than just PF)
10:19:43Unhelpfulright, the other plugins that use overlays on devices with smaller memory, for example.
10:20:00amiconnIt could even be automated in the build system, meaning that making a jumbo plugin wouldn't need special care anymore
10:20:05Unhelpfulwhat about the idea of moving the boundary between audio and plugin buffers?
10:21:12Unhelpfulin particular, that might be *very* nice if we also add the ability to unbuffer the last track and compact the audio buffer, since jumbo plugins could load without forcing a stop
10:21:30 Quit FlynDice (Read error: 110 (Connection timed out))
10:21:55amiconnActually, if plugins are linked in a way that their *end* aligns to a certain bound (that's how overlay linking works now), the plugin buffer size could even be a user setting
10:22:15amiconnI.e. the user would decide how much ram to set aside for plugins to run in parallel with playback
10:23:05Unhelpfulbut, if we make the audio/plugin division something that can change dynamically, that gets even better - there is no space reserved when a plugin is not running, for example. :)
10:23:26amiconnUnhelpful: The buffering mechanism doesn't allow this atm, also, there are two different implementations to take into account (hwcodec and swcodec)...
10:24:21Unhelpfulhow hard would it be to add such a capability? or, we could force stop/restart to move the buffer end, as we already do on AA size change...
10:24:57amiconnCurrently mpeg.c has no mechanism to change buffer allocation once it's initialised
10:25:38*amiconn likes that idea though
10:25:53amiconnExactly zero plugin buffer when no plugin is running...
10:26:43*Unhelpful would think that the extra code to do the added management work couldn't possibly cost more than 32KiB ;)
10:27:19amiconnNo, but it will increase the binsize. There are some limits related to that...
10:28:50 Quit fyrestorm (Read error: 60 (Operation timed out))
10:29:00amiconn(1) The OF loader will only load .mod/.ajz up to a certain size: 200KB on Player and Recorder v1, 400KB on the others. Not all devices are flashable, especially Player and Recorder v1.
10:29:35Unhelpfulwell, *all* of buflib, including the compactor, and the new buffer_in/out functions, comes to 1.2KiB on ARM. less than 800B on SH-1. i can't imagine adding compaction to our existing buffer implementations would approach even that much.
10:30:27amiconnWe have a workaround for that which pushes the limit a bit, and is already in use for recv1: When the binary exceeds 200KB, it will be .ucl compressed (which is done anyway for flashing) and prepended with a small self-extractor
10:31:41amiconn(2) For flashing, the binary needs to fit in the flash ROM. Flash ROM is 256KB; 28KB are taken by the flash loader and bootbox (the rescue firmware in case main rb fails) - that leaves 228KB
10:33:17amiconnThis is what ucl compressed rockbox must not exceed. If rockbox fits uncompressed, it can be run directly from flash rom ("rombox"), leaving ~200KB additional ram for buffering
10:33:43amiconnCurrently this is only possible for Player and Ondio SP - all others have grown too large
10:35:54Unhelpfulso, our major concerns would be 1) making core too large to fit compressed any more 2) making core too large to fit uncompressed on Player and Ondio SP.
10:39:27 Join ender` [0] (
10:40:13Unhelpfulalso, on multicore targets, making sure that compaction doesn't introduce concurrency issues. remember, jhmikes and i talked a bit about this when i was writing buflib, and the conclusion we came to was pretty much that each core could run its own buflib allocator for non-shared data, and that shared data would have to be allocated from a common buffer and kept at static addresses.
10:47:53 Part Llorean
10:48:38 Join {phoenix} [0] (
10:54:49 Join Zarggg [0] (
11:00:20Unhelpfulamiconn: as i understand it, the sh-1 cmp/pz instruction amounts to "true if signed value is non-negative", correct? i think i've got the final version of mul_s16_u32. the math tests out in C:
11:08:45 Quit Zarggg ()
11:10:12 Join intrados [0] (
11:14:51 Nick Horschti is now known as Horscht (n=Horscht@xbmc/user/horscht)
11:16:39*Unhelpful just realized he can make that branchless...
11:16:48 Nick FlynDice_ is now known as FlynDice (
11:23:01 Quit intrados ()
11:25:52Unhelpfulsince op16_hi is 0xffff in the case where we'd need to subtract op32_lo, we can AND the word-swapped op32 with A, SHLR16 the result, and subtract that from the hi part of the final result.
11:27:38 Join Zarggg [0] (
11:34:26***Saving seen data "./dancer.seen"
11:40:28 Join barrywardell [0] (n=barrywar@
11:51:26Unhelpfulthis seems good, i think...
12:08:36 Join SETKEHs [0] (n=setkehs@
12:14:35 Quit SETKEHs (Remote closed the connection)
12:17:31 Join webguest43 [0] (n=57aec1aa@gateway/web/cgi-irc/
12:18:20 Quit webguest43 (Client Quit)
12:24:03 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
12:38:26 Join Peace- [0] (
12:38:31 Part Peace- (" - Chat comfortably. Anywhere.")
12:46:49 Join wincent_balin [0] (
12:53:06 Join kugel [0] (n=kugel@rockbox/developer/kugel)
12:53:56kugelUnhelpful: hmm, mul_s16_u32 returns an unsigned value?
12:54:09kugelCan't the result be negativeß
12:57:20 Quit barrywardell ()
12:58:26 Join renke [0] (
12:59:43 Join barrywardell [0] (n=barrywar@
13:12:04 Join miepchen^schla [0] (
13:25:55 Quit BXCracer ("Quit")
13:28:50 Join keetah [0] (
13:30:13 Part keetah
13:34:30***Saving seen data "./dancer.seen"
13:37:23 Join Thundercloud [0] (
13:38:08 Quit faemir ("Leaving")
13:43:22 Join Makuseru [0] (
13:44:10 Quit Makuseru (Remote closed the connection)
13:44:45 Quit flydutch ("/* empty */")
13:47:51 Quit Thundercloud (Read error: 104 (Connection reset by peer))
13:49:54 Nick krazykit` is now known as krazykit (
13:51:10 Quit kugel (Read error: 60 (Operation timed out))
13:52:36 Quit SirFunk__ (Connection timed out)
13:53:55 Quit FOAD ("I'll be back")
13:59:39 Join FOAD [0] (
14:02:12 Quit robin0800 (Remote closed the connection)
14:09:59 Quit evilnick_home1 (Read error: 104 (Connection reset by peer))
14:11:07 Join evilnick_home [0] (
14:14:13 Join efyx_ [0] (
14:23:38 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
14:30:17 Join robin0800 [0] (
14:33:50 Quit Horscht (Read error: 104 (Connection reset by peer))
14:34:55 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
14:41:49CIA-38New commit by 03gevaerts (r21052): Fix #ifdef nesting to allow USB_HID to be built without USB_CHARGING_ONLY (FS #10236 by Tomer Shalev)
14:47:06 Join Lss [0] (
14:51:40 Join robin0800_ [0] (
14:52:39 Quit safetydan (Read error: 113 (No route to host))
14:54:57 Join BdN3504 [0] (n=55b234ad@gateway/web/cgi-irc/
14:58:05BdN3504I'm using the VMWare Development platform and been recently getting the error "this client is too old to work with working copy of [...]manual, pleae update your svn client" but i always check out to the newest svn version using the svn trunk rockbox command. what can i do?
14:59:28gevaertsdo you use svn both in the VM and from windows on the same checkout?
15:00:17 Join stoffel [0] (
15:05:55BdN3504errm? i use the vm with the terminal provided in the image. i recently installed tourtoisesvn but i don't use it.
15:06:52gevaertsthat error means that you checked out with a newer version of svn than the one you're using now
15:07:26Unhelpfulkugel: it can indeed be negative. it hardly matters, adding a "positive" value with the sign bit set has the same result as adding a negative. the functions that call it store the value as unsigned, too.
15:09:27BdN3504will removeing the whole folder and checking out again resolve the issue?
15:09:52gevaertsit should, yes
15:09:57BdN3504lemme try
15:13:05 Quit robin0800 (Read error: 110 (Connection timed out))
15:14:27 Join n1s [0] (n=n1s@rockbox/developer/n1s)
15:16:01 Join __lifeless [0] (n=lifeless@
15:20:37BdN3504thanks, it worked.
15:28:05 Nick fxb__ is now known as fxb (
15:32:11 Quit _lifeless (Read error: 113 (No route to host))
15:32:27 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:32:43kugelUnhelpful: still, it seems weird to me
15:34:33***Saving seen data "./dancer.seen"
15:35:43 Join shodanX_ [0] (
15:44:57 Quit BdN3504 ("CGI:IRC (EOF)")
15:46:14 Quit alexbobp (Read error: 110 (Connection timed out))
15:55:33 Quit gevaerts (Nick collision from services.)
15:55:42 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
15:57:31 Join moos [0] (i=mustapha@rockbox/staff/moos)
15:58:42 Join krazykit` [0] (
16:05:36gevaertsDoes a recent svn beast bootloader work for anyone?
16:08:48 Quit __lifeless (Remote closed the connection)
16:10:03 Quit krazykit (Read error: 110 (Connection timed out))
16:13:34moosgevaerts: hi, not here either
16:13:58gevaertsThanks. That means I can commit my new code without fear that it's all my fault :)
16:14:15mooshaha :D
16:16:08moosI wanted to put a rb only bootloader, but I fair to have the 120GB drive formated...
16:16:37gevaertsI've had it reformatted twice during the last half hour :)
16:18:58kugellow capacity flash daps ftw :>
16:19:31mooshehe, I also encountered lot of *random* problems in the past with the beast installation, /me blames MS here. Invasive OF. :)
16:20:11kugelwell, you can only fail if you expect to tame the beast easily
16:20:44*gevaerts now is stuck at File not found
16:20:59moosgevaerts: Invasive even at boot when runing dual one. Once you tasted single boot, it's hard to come back to this OF silyness
16:21:09moosgevaerts: the famous one :(
16:22:15moosgevaerts: does the 2nd partition is empty?
16:22:26gevaertsApart from .rockbox, yes
16:23:09moosBecause here, I found that at each time that the 2nd partition have missing some OF files/dirs, 100% of cases I had the file not found
16:23:50AlexPFirst thing I did was delete all the OF rubbish
16:24:02gevaertsah yes. IIRC i goes away by booting the OF once
16:24:03AlexPNever had a problem with file not found with either single or dual boot
16:25:22moosAlexP: hi, I spoke about the files that the OF generated after installation
16:25:54AlexPmoos: I know, and first time I had access to the filesystem I deleted them
16:26:05mooseach time I encoutered this "file not found" bug, I noticed that the 2nd partition was empty
16:26:17moosok after then :)
16:26:36moosnote that I'm a windows user ;)
16:27:01AlexPI'm just saying that it is odd though that these problems seems so hard to pin down
16:27:28gevaertsit's now helpfully removing all media items again
16:27:30moosindeed, and seems "random" :(
16:28:14moossilly OF, I wish we could completly get rid of it
16:30:23CIA-38New commit by 03gevaerts (r21053): Reorganise USB stack defines. Now config.h decides which class drivers get enabled instead of usb_core.h ...
16:30:48 Join {-phoenix-} [0] (
16:33:52 Join BXCracer [0] (
16:34:13moosgevaerts: note that to workaround this "File not found bug", I found that even if the partitions seems empty in Windows explorer, in the basket, there was the 3 .inf files needed for the windows installation. Then I restored them, and rockbox was working again here.
16:35:15gevaertsmoos: I've solved it every time by just booting the OF once. I don't know what it does
16:35:43mooscan you compare the 2nd partitions in the 2 cases?
16:36:05moosi.e once refuse to boot rb and once yes
16:37:14 Quit {-phoenix-} (Remote closed the connection)
16:37:52moosMarcGay had problems recently too IIRC
16:38:03moosoops MarcGuay even
16:40:02 Join {-phoenix-} [0] (
16:40:16moosOne mystery yet for me is: When the OF decide to format or not. Since sometimes it did and other times not.
16:42:57CIA-38New commit by 03gevaerts (r21054): Add HID keymap handling in usb screen. This allows to use the DAP as a keypad to control the PC. The current keymaps are audio-oriented ...
16:43:41 Quit {phoenix} (Read error: 110 (Connection timed out))
16:45:26 Join shodanX__ [0] (
16:47:54 Quit barrywardell ()
16:50:32CIA-38New commit by 03gevaerts (r21055): Clarify comment and clean up a bit (FS #10227 by Tomer Shalev)
16:54:57 Join tomers [0] (
16:57:12 Quit shodanX_ (Read error: 110 (Connection timed out))
16:57:18 Nick fxb is now known as fxb__ (
16:57:53gevaertstomers: is there still anything in FS #10166 or FS #10110 that really needs to be applied, or should I close them?
17:00:24tomersgevaerts: I'm working on it. You did a great job :-) The code is neatier, more readable. I wanted to make similar change myself but didn't want to get into that #ifdef hell, and also I felt I didn't see the full picture yet.
17:02:22*gevaerts spots red :(
17:02:55 Quit robin0800_ (Read error: 104 (Connection reset by peer))
17:05:03CIA-38New commit by 03gevaerts (r21056): Fix wrong #ifdefs that sneaked in
17:05:20 Join Mikachu [0] (
17:06:53tomersgevaerts: FS #10227 can be closed
17:07:27tomersOh, it is!
17:07:39 Quit {-phoenix-} (Read error: 104 (Connection reset by peer))
17:09:53 Join {phoenix} [0] (
17:09:56tomersgevaerts: AFAIC, You can close both FS #10166 and FS #10110
17:10:13CIA-38New commit by 03gevaerts (r21057): Fix wrong #ifdefs
17:10:15 Join daurn [0] (n=daurnima@unaffiliated/daurnimator)
17:11:21gevaertsAlso, your comment about #ifdef hell was spot on. It caught me twice apparently :)
17:13:36Mikachui just installed 21055 on my nano, but dmesg still says it only finds ums, i think
17:13:40Mikachushould it work? :)
17:14:33tomersgevaerts: Can you replace #ifndef BOOTLOADER ... #else .. #endif with #ifdef XXX #else #endif? It is a source of problems, as confusing ifndef with ifdef is very common and hard to spot
17:16:31gevaertsMikachu: does "lsusb -v -d 05ac:" mention hid?
17:17:04kugeltomers: nice work, how can I play with it?
17:17:13 Join Braintix [0] (n=Braintix@
17:17:36Mikachuit does look like it, but should it really say mouse?
17:18:22tomersMikachu: It should work: #define CONFIG_USBOTG USBOTG_ARC
17:18:36 Quit Braintix (Client Quit)
17:19:11 Quit perrikwp (" ajax IRC Client")
17:19:11tomerskugel: Thanks! Just plug it in, and use the keys to control volume and playback. What player do you use?
17:19:14***Alert Mode level 1
17:19:14CIA-38New commit by 03gevaerts (r21058): swap bootloader and non-bootloader cases around to avoid an ugly #ifndef
17:19:29kugeltomers: e200
17:20:20*kugel doesn't quite get how to use simplelist correctly :(
17:21:16tomerskugel: I develop on e200v1, so I am sure it works. Just to verify: Did you restarted the player? What OS do you use?
17:21:19Mikachushould i insert it holding menu so it doesn't go to the disk screen?
17:21:31kugeltomers: I haven't done anything yet :)
17:22:01kugelI'm right now booting my pc with windows 7, in the hope to control media monkey with my e200
17:22:08tomersMikachu: It work when the USB plug is shown on the screen, and the DAP can be accessed by the OS as a mass-storage device
17:22:32 Join robin0800 [0] (
17:22:48Mikachu[2429506.090229] usb-storage: probe of 1-2.3:1.1 failed with error -5
17:22:48Mikachu[2429506.090707] usb 1-2.3: device_add(1-2.3:1.1) −−> -5
17:22:50Mikachuis that related?
17:22:51gevaertsMikachu: if you don't have anything set up to handle these "multimedia" keys, you can use xev to look at incoming raw events
17:23:12 Join mcuelenaere [0] (
17:23:15Mikachugevaerts: it should create a node in /dev/input first ;)
17:23:24tomerskugel: I can control volume on Windows XP, no matter what window is in focus. Foobar2000 also changes tracks and playback mode even when it is in the background, so it is quite cool to be able to control music while working with other applications in foreground
17:23:41kugeltomers: yea, that's how I'd expect it to work
17:23:50kugelthat's what my keyboard is doing too
17:24:00*Mikachu uses win-alt-zxcvb to control his music player
17:24:27kugeland honestly, I don't think I'm going to replace my multimedia keyboard with my e200, but it's already incredible cool that I could do it :P
17:24:55Mikachumy previous rockbox build is from 2007 :P
17:25:13gevaertsThis will mainly be important once we get host support on some players, so you can use one rockbox player as a remote for another :)
17:25:44tomerskugel: Some people would find it nice to use their DAP as a wired alternative to Logitech's PowerPoint presenter, when it will be possible to have different sets of keymaps ;-)
17:26:52 Quit daurnimator (Read error: 101 (Network is unreachable))
17:27:42kugelthat will for sure be very useful
17:27:50gevaertstomers: can you have a look at the remaining build errors on ?
17:28:12tomersgevaerts: Looking...
17:29:15***Alert Mode OFF
17:30:08 Quit robin0800 ("No Ping reply in 90 seconds.")
17:30:21tomersgevaerts: I'll fix it and post a patch now. I tested all platforms using the dist tools - compiled it over night. That's weird. These tools halt on error.
17:30:32 Join robin0800 [0] (
17:31:41 Quit daurn (Read error: 110 (Connection timed out))
17:32:07 Join daurn [0] (n=daurnima@unaffiliated/daurnimator)
17:32:21 Quit robin0800 (Client Quit)
17:32:47 Join robin0800 [0] (
17:34:36***Saving seen data "./dancer.seen"
17:41:11gevaertstomers: when you're working out how to build usb class drivers as plugins, try to make it possible to have some specific ones built-in. I'm mainly thinking about bootloader use here
17:41:36tomersgevaerts: Posting patch to fix red...
17:43:37 Join merbanan [0] (
17:43:55 Join kugel_ [0] (
17:44:01 Quit kugel (Nick collision from services.)
17:44:11 Nick kugel_ is now known as kugel (
17:44:43kugeltomers: \o/ very cool! and impressive
17:45:19kugelgevaerts: ", and eliminates the need for a custom class for charging-only connections. " is too technical for MajorChanges IMO :)
17:45:35gevaertskugel: feel free to fix it :)
17:45:49gevaerts"Windows will now no longer ask for a driver"?
17:46:01AlexP"Charging on Windows now works" :)
17:46:28gevaertsAlexP: careful there. There are other issues as well
17:46:48tomerskugel: Thanks!!! ;-)
17:48:14CIA-38New commit by 03gevaerts (r21059): fix some keymaps to actually compile (FS #10237, by Tomer Shalev)
17:48:48gevaertsuhm, those deltas look very wrong...
17:49:39kugelhm, windows asked for a driver, but found one itself
17:49:51AlexPgevaerts: OK, allow charging on windows to work "at all"
17:50:27CIA-38New commit by 03gevaerts (r21060): don't disable usb storage by accident
17:50:52kugelI guess my change isn't quite complete then
17:50:53gevaertsAnyway, we now know the exact binsize cost of usb storage :)
17:51:30gevaertskugel: it worked all by itself in xp for me
17:51:34kugelthe mobe is still red
17:51:53kugelgevaerts: yes, but didn't you get the pop up about installing a driver?
17:52:41tomerskugel: The driver installed by Windows is the HID driver. It is the same driver that would be loaded if you attach a USB Multimedia Keyboard
17:52:42kugelit didn't ask me for a driver, but it installed one (installing device driver..[some seconds]..the device is now ready to use)
17:53:13tomerskugel: I'm testing the mrobe
17:54:01tomerskugel: Mrobe 500 or 100?
17:54:14*gevaerts finds out that testing mr500 builds to see if mr100 works correctly is not entirely the right way
17:55:07kugelso, now we just need to be able to open the virtual keyboard and type texts using the dap
17:55:37AlexPkugel: masochist
17:55:55gevaertskugel: I think you just need a single extra pluginlib export to be able to make a plugin that does that :)
17:56:48kugelalso, I want to open firefox with the rec button :D
17:56:59*kugel demands too much :(
17:57:09Mikachuyou want to scroll web pages with the wheel too?
17:58:59*domonoky wants full configurability for this, so you can set which key sends which keycode :-)
17:59:52AlexPThis is quite weird, selecting volume up and down on the player and seeing the volume bar changing on my laptop :)
18:00:17gevaertswell, once it's in a plugin, full configurability shouldn't be a big problem :)
18:03:44kugelgevaerts: so what's now in the core won't be there for long?
18:04:40gevaertskugel: I don't know. This "usb as plugins" thing isn't actually designed yet
18:06:34tomersHow do you set your IRc status (i.e. 'tomers is fixing red')? Using chatzilla
18:06:50*AlexP does stuff
18:06:53AlexPLike that?
18:07:05 Join _lifeless [0] (n=lifeless@
18:07:19AlexPIf so, it isn't a status - just use "/me"
18:07:48gevaertskugel: using the current in-core HID, you'd only need to add usb_hid_send_consumer_usage() to the plugin API to have custom keymap plugins
18:09:40 Join Ubuntuxer [0] (
18:10:22tomersgevaerts: Posting yet another patch to fix red
18:10:40*tomers learning to set status
18:11:07Horschtit's not a "status"
18:11:12Horschtit's an action
18:11:23CIA-38New commit by 03lenzone10 (r21061): Updated italian language file.
18:11:37Horschtsomething like... twittering
18:13:05CIA-38New commit by 03gevaerts (r21062): fix another keymap (FS #10237 by Tomer Shalev)
18:14:57tomersA noobs question: The mass storage takes ~5K... When USB classes will be implemented as plugins, and the DAP detached from USB, this space will be spared. What benefit we would get from it?
18:17:54gevaertsFrom that 5K, not really much. It's not the entire picture though, there's also a bit in screens.c (the USB connect screen). As I see it, the real benefit is that it's a lot easier to add class drivers, and to make them more interesting. Some of those "other ideas" on your wiki page aren't really doable in core
18:20:59tomersgevaerts: Thanks for this info. I really couldn't understand how it helps from other aspects. I do agree that it will ease classes development. I'll update my Wiki page. Going now...
18:21:15gevaertsthe usb logo on the connect screen is more than 16K on ipod video, and it's *only* used for mass storage. That means that with careful cutting, the actual benefit of plugins is nearer 20k
18:21:46 Quit _lifeless (Read error: 113 (No route to host))
18:22:12 Join _lifeless [0] (n=lifeless@
18:22:13gevaertsand that's the sort of number that does matter
18:23:21tomersgevaerts: Interensting...
18:23:39*gevaerts wonders if it would be possible to build plugins as totally position independent code, so you could load more than one tsr plugin at the same time
18:23:39*tomers Going to walk
18:25:00 Quit _lifeless (Read error: 54 (Connection reset by peer))
18:25:03 Join _lifeless [0] (n=lifeless@
18:28:25gevaertsAlso, if usb storage is a plugin, it either can use the entire plugin buffer and doesn't need the audio buffer anymore, or (if we go for a separate small plugin buffer for usb) it's actually apps/ code so it's allowed to grab the audio buffer :)
18:30:22 Join BryanJacobs [0] (
18:38:32 Quit Ubuntuxer ("Leaving.")
18:40:10kugelhm, now I managed to get a nice simplelist, but it doesn't scroll anymore :(
18:41:44 Quit MarcGuay (Read error: 110 (Connection timed out))
18:42:00kugelmcuelenaere: ping
18:43:48 Join MarcGuay [0] (
18:56:30kugelhm, the sansa fuze's framerate only varies with changing the dbop clock when the cpu is boosted :/
18:58:49kugelthat makes sense, actually
19:04:54 Quit {phoenix} (Remote closed the connection)
19:08:09 Join petur [50] (n=petur@rockbox/developer/petur)
19:08:27 Join Xerion_ [0] (
19:10:50 Quit Xerion (Read error: 60 (Operation timed out))
19:10:50 Nick Xerion_ is now known as Xerion (
19:19:48 Join krazykit [0] (
19:21:01 Quit krazykit` ("Connection reset by beer")
19:21:37mcuelenaerekugel: pong
19:21:42 Quit robin0800 ("No Ping reply in 90 seconds.")
19:22:11kugelmcuelenaere: I was wondering about the exit() for plugins. What if plugins have to do some cleanup before exiting?
19:22:21 Join robin0800 [0] (
19:23:26 Quit robin0800 (Remote closed the connection)
19:24:55mcuelenaerekugel: hmm yes, they should do that theirselves atm
19:25:03mcuelenaerebut there could be some callback function if wanted
19:25:15kugelbut that isn't possible when doing exit() atm?
19:25:37mcuelenaerekugel: have you looked at ?
19:25:56kugelI thought that a) a cleanup() would be called in exit() which is to be provided by the plugin, or b) the plugin loader handles all possible cleanup
19:25:59mcuelenaereyou can add a new PLUGINLIB_EXIT_INIT_W_CALLBACK(callback) define though
19:26:21mcuelenaerehow could the plugin loader handle all possible cleanup? (e.g. open file handles)
19:26:57kugelhm, right
19:27:15*mcuelenaere thinks adding a PLUGINLIB_EXIT_INIT_W_CALLBACK will be the easiest to do
19:27:16kugelI was more thinking of thinks like reverting to backlight settings if the plugin forced it on and the like
19:27:30kugelsounds reasonable, yes
19:27:36mcuelenaererevertk g
19:27:53 Quit krazykit (Read error: 104 (Connection reset by peer))
19:27:58mcuelenaerereverting to the original backlight settings sounds something for the plugin loader indeed
19:29:25 Join krazykit [0] (
19:34:40***Saving seen data "./dancer.seen"
19:40:26 Join HellDragon [0] (
19:41:35 Join MarcGuay_ [0] (
19:42:47 Join bubsy [0] (i=Bubsy@
19:48:13AlexPmcuelenaere: Incidently, is there any differences between lua in the wild and lua in Rockbox? I'm thinking for the manual. It'd be good to be able to say something like "This is lua, see link. The Rockbox implementation is different/doesn't support..."
19:48:32mcuelenaereAlexP: it doesn't support floating point
19:48:39AlexPmcuelenaere: Is that it?
19:48:45mcuelenaereit also doesn't have some of the default lua libs (io, math, ..)
19:49:10AlexPmcuelenaere: Any chance you could note these down for me (or whoever) to do the manual entry with?
19:49:36AlexPat some point :)
19:49:38mcuelenaereAlexP: I'm thinking of picking up my old plugin API documentation work and rewrite to get it working with Lua
19:49:46mcuelenaereAlexP: sure
19:50:09mcuelenaerealso, I'm not the one who did the initial Lua porting (safetydan did); so I don't know every little detail about it :)
19:50:19AlexPmcuelenaere: Cheers - it'd be good to get something in the manual, but I don't want to get the technical aspects wrong
19:50:36AlexPmcuelenaere: Well we can always fix, it'd just be good to not have any glaring errors straight off :)
19:51:47mcuelenaereAlexP: perhaps you should just prod me with a TeX patch and then I'll see if there are any mistakes in them or so :)
19:52:13AlexPmcuelenaere: OK :)
19:52:29AlexPOr I'll leave you a section to fill in with the missing details :)
19:54:58 Join tomers_ [0] (
19:57:20 Quit MarcGuay (Read error: 110 (Connection timed out))
19:59:55 Quit renke (Read error: 110 (Connection timed out))
20:00:25 Join DrMoos [0] (
20:00:51 Quit moos (Read error: 104 (Connection reset by peer))
20:00:53 Nick DrMoos is now known as moos (
20:03:06AlexPSo, any got any bright ideas on how I translate a check for plugin buffer size into a manual opt?
20:05:09gevaertsAlexP: if all else fails, you could add a feature in features.txt with the exact same logic as in SOURCES
20:06:11kugelgevaerts: He hesitates to add something to features.txt just for the manual
20:06:19AlexPI do?
20:06:30kugeldidn't you? Someone said that
20:06:40kugelsomeone said, that this someone hesitates I mean
20:06:45AlexPkugel: It may well be true, but it wasn't me that said it :)
20:07:14gevaertsI'd actually like a feature for each plugin
20:07:19kugelI think there's no reason to not do it, especially as there's some plugin localization work ongoing
20:07:39kugelso those features are quite likely to be used for langs later too
20:07:43AlexPI could just do it with \nopt for those that don't have it, but that isn't very nice as then people have to check it for every new target, and adjust it if a target's support changes
20:07:58gevaertskugel: well, features are listed in rockbox-info.txt, they aren't pure internal stuff
20:08:33kugelI'd just do #if PLUGIN_BUFFER_SIZE < 512K;low_plugin_buffer;#endif
20:08:48kugelgevaerts: it still doesn't hurt, does it?
20:09:42gevaertsit doesn't hurt *much*, probably.
20:09:44kugelor a features.txt just for plugins (plugins.txt?). that could then be in a seperate (if at all) section in rockbox-info.txt too
20:10:11 Quit tomers (Read error: 110 (Connection timed out))
20:10:29 Quit tessarakt (Read error: 104 (Connection reset by peer))
20:11:18*gevaerts thinks that this needs the opinion of the manual regulars
20:11:51*domonoky reminds everyone to be carefull when messing with rockbox-info.txt, its used by rbutil :-)
20:11:59AlexPI'd prefer an option like that I can check, but I don't know how it affects the rest of Rockbox
20:12:10AlexPdomonoky proves my point :)
20:12:15domonokyespecially the features, as they are needed for voice generation..
20:13:56 Join renke [0] (
20:17:57 Join kugel_ [0] (
20:24:27 Quit kugel (Nick collision from services.)
20:24:38 Quit kugel_ ("exit(0);")
20:26:48 Join tessarakt [0] (
20:33:25 Join __lifeless [0] (n=lifeless@
20:35:15 Quit __lifeless (Remote closed the connection)
20:38:48 Join otro_viajero7 [0] (n=chatzill@
20:39:13 Join __lifeless [0] (n=lifeless@
20:39:24 Quit __lifeless (Read error: 104 (Connection reset by peer))
20:39:31 Part otro_viajero7
20:40:16 Quit daurn (Remote closed the connection)
20:41:48Unhelpfulkugel: if it makes you happier, i can change the type before commit. it's not like it would matter.
20:41:57BryanJacobssomebody really needs to generate a dependency graph of the files in the rockbox tree and rearrange things so that the graph isn't so connected
20:42:30BryanJacobsfor example, successfully compiling apps/metadata/wavpack.c requires apps/metadata/mp3.c
20:44:05BryanJacobsand EVERYTHING requires misc.c
20:44:11BryanJacobsand misc.c requires everything
20:45:57 Quit HellDragon (Client Quit)
20:47:35 Join kugel [0] (i=kugel@rockbox/developer/kugel)
20:50:23 Quit _lifeless (Read error: 113 (No route to host))
20:51:21gevaertsBryanJacobs: I once tried. I killed dot when it was using 110GB of memory
20:51:59domonoky:-) the question is, is it worth the effort ? (grep or "search in files" is very helpfull when navigating rockboxs source) :-)
20:52:24BryanJacobsdomonoky: yes, if Rockbox were properly modular my life trying to build a minimal set of it would be much easier
20:52:50 Join flydutch [0] (
20:53:30BryanJacobsI eventually just took metadata_common.c and stripped out all the stuff I didn't need
20:53:50BryanJacobsall I really needed was the read_uintfoo functions
20:53:55kugelthe mp3.c is probably something historical
20:54:27BryanJacobskugel: graph is metadata/wavpack -> metadata_common -> id3 funcs -> metadata/mp3
20:54:46BryanJacobsthe problem is metadata_common having too much stuff in it
20:55:12BryanJacobsit actually keeps going with mp3 -> codepages + unicode -> misc -> everything
20:55:32kugelwell, guess why it's id3 funcs and struct mp3entry. historical reasons. And apparently nobody bothered to clear it up yet
20:55:52 Join HellDragon [0] (
20:56:19kugelpatches are welcome, you know :)
20:56:20BryanJacobswhy don't codecs implement their own metadata parser?
20:56:48domonokyBryanJacobs: thats not possible.. at the time metadat is read, the codec isnt loaded.
20:56:53kugelbecause lots of codecs share the metadata with other codecs
20:56:53BryanJacobskugel: I'm coding stuff now
20:57:07domonokyand the sharing of course too...
20:57:15BryanJacobsah, I see
20:57:40kugelalso, the metadata reading isn't a codec thing, mostly (IIUC)
20:57:50domonokymetadata code is also the second swicth from file extension to codec (in case two codecs can have the same file ending)
20:58:16BryanJacobsI didn't realize file extensions were one-many mapped to codecs
20:58:31BryanJacobswhich extension(s) would that be?
20:58:40BryanJacobswe don't do container formats like AVI, right?
20:58:49domonokythats a little bit hidden, but metadata can switch the codec type..
20:58:57domonokyfor example .ogg i think
20:59:08kugelwe have many codecs. but many codecs share the metadata type (vorbis comment, id3, etc) and many codecs share its container
20:59:25BryanJacobsI didn't know rockbox did ogg-contained flac
20:59:37kugeland AFAIK it doesn't
20:59:48BryanJacobsI think I understand now, though - just having the shared id3 code would be reason enough
21:00:04BryanJacobsbut still, the dependencies are out of hand
21:00:17domonokywe have some shared containers, but i dont know exactly which ones..
21:00:33BryanJacobsmaybe WAV
21:02:02BryanJacobswhy do I get a backtrace like "#0 0x00000000 in ?? ()" when I compiled everything with -ggdb?
21:02:41domonoky.mp4 does it (line 653)
21:03:07domonoky(switches between ALAC and AAC)
21:03:21mtBryanJacobs : Instead of taking functions from rb's source tree, you could write your own functions that "look" like what you'd use from rb's source. (That's how it's done in RM)
21:03:23BryanJacobsoh yeah, I forgot about .m4a
21:03:31 Join Llorean [0] (
21:03:55BryanJacobsmt: I was trying to use the code in-place because then it automatically stays in sync
21:04:03mtThat's assuming you're trying to get the codec working outside rockbox ?
21:04:28BryanJacobsI've been pulling things out only when absolutely necessary
21:04:53BryanJacobsbut maintenance gets harder when you increase code duplication, right?
21:05:06domonokyBryanJacobs: while its good that you try to clean up the mess, i would suggest to focus on your project... :-) and maybe take a look at the other codec test programs...
21:05:24BryanJacobsdomonoky: I'm using the test_codec plugin code
21:05:36domonokythats probably a good start :-)
21:05:56BryanJacobsand not actually doing cleanup right now... I was just getting annoyed at how the dependencies made it hard to get things working
21:05:57mtBryanJacobs : The RM parser isn't that big so I had no trouble porting it to rockbox .
21:07:35BryanJacobsanyhow, I've got a whole codec testing program compiling but have to fix a segfault
21:08:15BryanJacobscode at for the interested
21:08:32BryanJacobsI'm signing off for today, see you all later
21:09:29domonokythat sounds good.. cu later... :-)
21:09:46 Join n00b81 [0] (
21:09:48 Part BryanJacobs
21:09:54 Part n00b81 ("Leaving")
21:12:00 Quit renke ("leaving")
21:19:21 Quit HellDragon (Client Quit)
21:19:27 Quit flydutch ("/* empty */")
21:31:03 Quit kugel ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042316]")
21:34:42***Saving seen data "./dancer.seen"
21:39:47 Quit tomers_ ("ChatZilla 0.9.84 [Firefox 3.0.10/2009042523]")
21:49:08 Quit stoffel (Remote closed the connection)
21:51:24 Join SirFunk [0] (
21:52:53 Join alexbobp [0] (
21:57:48 Join robin0800 [0] (
22:15:44 Quit spyder (Read error: 110 (Connection timed out))
22:17:01 Quit SirFunk (Read error: 60 (Operation timed out))
22:17:51 Quit robin0800 (Remote closed the connection)
22:19:56 Join SirFunk [0] (
22:23:01 Join andrewbeveridge [0] (i=586d65ae@gateway/web/ajax/
22:26:48 Join archivator [0] (n=archivat@
22:33:42 Join faemir [0] (
22:38:39andrewbeveridgeBadger: oo, i only just noticed you are online! a little birdie told me you are the man in charge regarding TWiki? I tried to register a couple of days ago, but cannot complete because of a 403 Forbidden error =(
22:40:29Bagder403 at what point?
22:41:05andrewbeveridgeupon clicking the link in the confirmation email:;code=AndrewBeveridge.660054076
22:41:08Mikachuandrewbeveridge: bagder, not badger
22:41:26andrewbeveridgeMikachu: man, i don't take enough time to heed usernames huh
22:41:48andrewbeveridgeMikachu: and i thought i got enough of a lecture from gevaerts ;-)
22:44:29Bagderandrewbeveridge: and you tried simply doing it again?
22:45:03andrewbeveridgeseveral times
22:45:09andrewbeveridgebefore and after the outage
22:45:38andrewbeveridge(in the vague hope it was a temporary (perl? php?) error.
22:45:42Mikachufwiw i get the 403 with that link too
22:45:51Bagderyeah, me too
22:46:04Bagderit's just weird since we get N people registered fine every day
22:46:41andrewbeveridgewell, could you maybe cancel my registration so i can try again with the same details? (and i wrote such a thoughtful message too =P)
22:46:52Bagderwell I see no traces of it so I can't cancel anything
22:47:17andrewbeveridgewhenever i try to register again it says that username is taken
22:47:31andrewbeveridgethats just annoying.
22:47:31BagderI see it now
22:47:39andrewbeveridgeah good =)
22:48:15andrewbeveridgeanything unusual? are you looking at a database here, or TWiki admin page, or what? (curious)
22:48:57 Join spyder [0] (
22:53:20*andrewbeveridge shall be back (gotta switch batteries)
22:53:43 Quit andrewbeveridge (" ajax IRC Client")
23:01:21 Join andrewbeveridge [0] (i=586d65ae@gateway/web/ajax/
23:01:37andrewbeveridgehello again
23:02:45andrewbeveridgeanybody know if/where rockboy saves key bindings?
23:06:23 Join fyrestorm [0] (
23:08:43 Quit bertrik (Read error: 113 (No route to host))
23:12:02 Quit archivator ()
23:13:18andrewbeveridgeBagder: could you perhaps tell me where i stand now regarding the wiki please? should i try and create a new account using a different name, or same, or what?
23:13:48BagderI removed the single file I found with your name in it
23:14:00BagderI know nothing further unfortunately
23:14:27andrewbeveridgeok, thanks. i shall try registering again and report back =)
23:16:29 Join archivator [0] (n=archivat@
23:19:00archivatorQuick question - is there any way a plugin can detect buffer underruns and take measures to reduce its cpu usage?
23:20:14saratogabut we have a scheduler, so setting the priority low enough should be good enough to accomplish that I think
23:23:43archivatorI'm not sure I follow - the plugin should lower its thread's priority, is that what you're saying?
23:26:06saratogaif you're using so much CPU time that you compete with the codec for CPU time, you can request a lower priority for your thread I believe
23:27:02Unhelpfulis this about the spectrum visualizer?
23:27:15archivatorUnhelpful: it is.
23:28:12Unhelpfulmy suggestion would be to calculate the FFT in a thread that is of lower priority than the various threads doing decode, buffering, and playback, and display it in the main thread, since displaying it doesn't really cost anything.
23:29:23 Quit SirFunk (Read error: 60 (Operation timed out))
23:29:38saratogahow much CPU time does your plugin really need?
23:29:41archivatorUnhelpful: There's scaling to do before displaying it but I can see where you're going. I've never worked with threads in rockbox, is the overhead significant?
23:30:19saratogai've become interested in getting a fast fft into rockbox once again
23:30:48Mikachuheh, fast fft huh?
23:30:53archivatorsaratoga: hard to say - how would you propose to measure that? I mean, I can only give you updates per second and that's hardly a meaningful number..
23:31:43archivatorMikachu: there's fft and then there's kiss fft (which I'm regretfully using) :)
23:31:48saratogaI would time the execution of one fft using a timer
23:32:17saratogaor else try to estimate how much time you're spending in teh fft by running your plugin with and without fft and observing the difference in clock speed
23:33:04 Join xnyhps [0] (n=xnyhps@2001:470:1f14:da:219:e3ff:fed7:c57c)
23:33:52 Quit bmbl ("Woah!")
23:34:41archivatorsaratoga: I can probably get you those numbers in a day or two (I don't have much spare time right now ) .. what are you expecting to see though? KISS FFT is hardly the fastest implementation out there and all the function calls in there must add quite the overhead...
23:34:47***Saving seen data "./dancer.seen"
23:35:54saratogaknowing how much time you're using is helpful for optimization
23:36:12saratogai guess a better question might be how often you call the fft and of what size
23:36:25Mikachuwhat about fftw?
23:36:35archivatorRight, I thought it had to do with your incentive to add an fft lib in core rockbox..
23:36:55saratogai don't think fftw does fixed point
23:37:11andrewbeveridgeMikachu: seti requires networking ;-)
23:37:23archivatorI think much of my troubles will be solved once I figure out a better way to access PCM data and trade speed for precision
23:37:41saratogathe reason to add fft to core would be as part of the codec library's MDCT functions, if in fact a fast fft is faster then alternative ways of computing the imdct
23:38:29saratogaof course that would have the added advantage of making it available to plugins
23:39:27archivatorSure. I don't see any other plugin using that but that's fine..
23:40:00archivatorI'll give the threaded idea a try next week. Have to go now.
23:40:21 Quit archivator (""break the glass!"")
23:41:23 Quit petur ("Zzzzzz")
23:41:35andrewbeveridgewhy is rockbox a rather than a plugin?
23:41:57 Join SirFunk [0] (
23:42:18*Unhelpful rather doubts imdct-via-fft can beat an optimized imdct
23:42:43Unhelpfulespecially if the optimized imdct has had the windowing function it uses rolled into the same operation
23:43:17*andrewbeveridge wishes he had more experience and understood more jargon
23:44:57saratogaUnhelpful: on the ffmpeg list people seem to hold the opposite view, and dislike the idea of mixing the two operations
23:45:05saratogabut i've seen people claim both are faster
23:45:35saratogai emailed the ffmpeg developer behind their new IMDCT to ask him about it
23:47:00 Quit merbanan (Read error: 110 (Connection timed out))
23:47:25Unhelpfulsaratoga: it probably depends on the windowing function used, but it might be possible to multiply the IMDCT matrix by the windowing function, and then reduce algebraically to produce an optimized imdct+window
23:54:07 Part wincent_balin ("Kopete 0.12.7 :")

Previous day | Next day