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 2009-02-26

00:08:07 Join avis [0] (n=ident@pdpc/supporter/student/avis)
00:08:16 Quit jgarvey ("Leaving")
00:08:50 Quit bluebrother (Nick collision from services.)
00:08:55 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
00:11:55kugelGood job!!!!!!!!
00:12:04*kugel is a bit late to the party
00:13:00gevaertskugel: the last yellow fix isn't built yet, so you're in time :)
00:15:15 Quit pixelma (Nick collision from services.)
00:15:16 Join pixelma_ [50] (n=pixelma@rockbox/staff/pixelma)
00:15:18 Quit amiconn (Nick collision from services.)
00:15:19 Join amiconn_ [50] (n=jens@rockbox/developer/amiconn)
00:15:30 Nick pixelma_ is now known as pixelma (n=pixelma@rockbox/staff/pixelma)
00:15:37 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
00:16:33kugelgevaerts: all targets have usb no, except 1-4gen ipods?
00:16:47gevaertsexcept 1-3gen ipods
00:16:52gevaerts4th gen has it
00:16:57amiconn1st and 2nd gen ipods have no usb anyway
00:17:03amiconn(only firewire)
00:17:28Bagder"the person from nvidia i know wont give me the data sheet" now that came as a complete surprise! ;-)
00:17:29*gevaerts thinks that there's a bit of wiki fixup to do
00:17:29rasherSo 3G is the odd one out?
00:17:37Bagder(quote from the forum)
00:17:43rashergevaerts: and manual..
00:17:47 Quit ender` (" Remember: A secretary isn't permanent until she's been screwed on the desk...")
00:18:11*gevaerts should never have started this...
00:18:34kugelgevaerts: first page is fixed :)
00:19:36kugelwhat a great day :)
00:20:47n1sBagder: I think this usb stuff warrants a news item on the front page
00:21:12Bagderlet's poke Zagor!
00:21:32gevaertsMaybe it also needs a
00:21:32 Quit billenium ("Lost terminal")
00:21:36gevaertsMajorChanges line
00:21:55rasherI'll say
00:21:56kugelgevaerts: yea, someone is already editing it....
00:22:05rasherBagder: can't we all add to the news?
00:22:20Bagderbut it's not updated automatically
00:22:30Bagderso once the page is changed in svn, we need to svn up there
00:22:32kugel"USB stack on..." isn't quite right is it?
00:22:37kugelthe stack was there before
00:23:38*rasher gets lost in the manual
00:23:49amiconnrasher: USB on the 3rd gen is special in several aspects. The 3rd gen can't charge via usb, and iirc it only has usb1.1
00:23:53kugelit reads USB enabled now
00:24:15amiconnThe controller is different - not sure whether it's built into the PP5002 or whether it's separate
00:24:31amiconnI guess the former, judging from the controller addresses
00:24:36*rasher weeps for the 3G owners
00:25:22amiconnThe controller base address follows a rather typical relation between PP5002 and PP502x: 0xc0005000 on PP5002, 0xc5000000 on PP502x
00:25:56amiconnSeveral modules have this relation: 0xX000Y000 <-> 0xXY000000
00:27:28 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
00:28:53gevaertsIpodFAQ, SansaFAQ and IriverH10Port still mention lack of USB support
00:29:25*rasher is on it
00:30:00rasherWow, SansaFAQ has it all over the place
00:30:36*rasher removes information about the bootloader booting OF on usb insert, since it'll be false soon anyway
00:30:40rasherAnd it's not vital
00:31:01rasherIt still explains how to boot it using the keypress
00:34:14***Saving seen data "./dancer.seen"
00:35:05 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
00:36:14 Join jordoex [0] (
00:36:56 Quit rocko ("Leaving")
00:37:11kugelrasher: well, it's false right now. Until they update nobody has USB
00:38:16rasherI guess
00:38:55*kugel wonders if it's safe to wipe the OF entirely now
00:39:09rasherThe wiki has historically documented current builds, but I suppose there's a case for documenting the current release, but I'm not sure how that'd work
00:39:19kugelI still have it in the SYSTEM folder (I replaced the bootloader a long time ago already)
00:39:56kugelI think I keep it, until there's some usb mode in the bootloader
00:40:39rasherThere's always rescue mode
00:42:54kugele200tool you mean?
00:43:23LloreanRecovery mode, unless you want to nuke that too.
00:43:27LloreanWhich would be pretty pointless.
00:43:49kugelpointless? it saves several seconds bootup time to replace to OF bootloader
00:44:10rasherThen there's e200tool, yes.
00:45:28 Quit itcheg (" ajax IRC Client")
00:45:32Lloreankugel: You've verified that removing all of the OF except recovery mode takes significantly longer than also removing recovery mode?
00:45:57kugelLlorean: You can remove the OF without recovery mode?
00:46:22LloreanSince Recovery Mode is used for recovering a damaged OF, I'd imagine they're separate...
00:46:30kugelI have replaced the OF bootloader with the Rockbox one (and if I want to the OF I load the one from the SYSTEM folder)
00:46:49LloreanYou still have recovery mode if you put a rockbox bootloader as PP5022.MI4 on a recovery partition, right?
00:46:57kugelafaik that's the only other way to install the rockbox bootloader
00:47:22kugelno, I don't
00:47:48kugelhm, I didn't know of that way, I haven't tried
00:48:00 Quit bluebrother ("leaving")
00:48:04kugelI used the -b option of sansapatcher, which gets rid of the OF bootloader at all
00:48:06LloreanWhat was the "no, I don't" in response to then, if you haven't tried?
00:48:24kugelI don't have the recovery mode
00:48:40LloreanBut you didn't do it the way I described...
00:49:03kugelNo, I didn't
00:49:10LloreanI didn't ask "Do you have recovery mode", I asked if it was available if the bootloader was installed the way I described.
00:49:21LloreanWhether you have it or not is irrelevant to the question...
00:49:44kugelI didn't even think of doing it the way you described
00:49:58rasherNow, who will start working on iTunes support?
00:50:16Lloreankugel: It should net you the much faster bootup time without losing recovery mode, if my understanding is correct.
00:50:30kugelLlorean: but that doesn't remove the OF bootloader then, and thus don't save much bootup time, I assume
00:51:00LloreanMost of the bootup time is in what it loads, not the fact that it runs, I think. Similar to iPods.
00:51:15LloreanSince it'll be loading a much smaller image (just the RB bootloader, not the combined RB+OF) it should run much faster.
00:51:44 Join rocko [0] (
00:52:03kugelbut that recovery mode doesn't help me much anyway. I'm not searching for a way to recover the OF, but rockbox, in case something goes wrong
00:53:16LloreanYou can put a rockbox.mi4 in recovery mode, recover Rockbox AS the "main" firmware, then use its USB mode to do whatever you want.
00:54:57rasherkugel: well, once you have the OF, you can get Rockbox.
00:55:02kugelrockbox is then installed on this firmware partition, right? the bootloader doesn't look there after the reboot (and IIRC it reboots after recovery mode)
00:55:33kugelrasher: well, of course, that's why I have it in the system folder. that's one step less in case I need to recover
00:56:03Lloreankugel: What bootloader doesn't look there after the reboot?
00:56:23kugelah, the OF bootloader will load rockbox directly then, correct?
00:56:28LloreanIt should, yes.
00:57:33kugelhm, I think I'll just keep things like they are. The of.bin in the system folder doesn't really disturb me
00:58:01rasherLlorean: if I forget, delete the usb test build whenever you feel like it
00:59:01LloreanI'll try to remember to do that after dinner later.
00:59:42LloreanOr, since it's already locked, may as well go ahead
00:59:50LloreanI was gonna leave it a little bit to see if that one guy responded with more info
00:59:57rasherFigured I'd leave it around for a bit so people don't wonder what happened
01:00:04rasher(in case they miss the news)
01:00:29LloreanI think if it's missing they're more likely to look for the news, honestly.
01:00:30rasherToo slow...
01:00:41LloreanIf we leave it, they'll just read the first post, and dowload it instead of the official build.
01:00:41rasherMight be
01:00:59rasherI edited that
01:01:16 Quit Zagor ("Leaving")
01:01:36LloreanAh well, it could be good or bad either way.
01:02:02 Quit T0paz ("Leaving")
01:02:11rasherLet's not waste more time on that, then
01:03:25 Join fdinel [0] (
01:03:50*Llorean posts something to the announcement forum.
01:04:27rasherProbably should include "will be in 3.2, due to be released ____"
01:04:50LambdaCalculus37Here's to PortalPlayer USB!
01:04:53*LambdaCalculus37 hoists a beer
01:05:42*Llorean hopes he got the list of players right
01:06:07gevaertsLlorean: mrobe 100
01:06:15kugelLlorean: nope :)
01:06:35LambdaCalculus37I was just about to post that Llorean missed the m:robe 100.
01:06:51*Llorean always forgets that darn thing.
01:07:16LambdaCalculus37Llorean: It's the unloved target. ;)
01:07:50*Llorean is glad that he only missed one, the topic was at its max length without M:Robe already
01:09:53 Join stripwax [0] (
01:15:42*rasher compiles bootloaders
01:16:05*Llorean cheers
01:19:32 Quit MethoS-- (Remote closed the connection)
01:23:45LambdaCalculus37rasher: \o/
01:25:50 Quit LambdaCalculus37 ("dinner time!")
01:26:11 Quit homielowe ()
01:26:27*stripwax reads logs
01:26:42rasherWhich one is the bootloader file for ipods - bootloader.bin or bootloader-ipod4g.ipod?
01:26:47stripwaxgevaerts - congrats on USB enabled!
01:27:15gevaertsstripwax: Thanks! Don't forget all other people involved though
01:27:29gevaertsrasher: yes
01:27:40rashergevaerts: *stab*
01:27:49stripwaxs/gevaerts/gevaerts+all the other people too!/
01:27:55gevaertsrasher: as in, ipodpatcher can install both
01:29:15gevaertsrasher: the builtin ones seem to be .ipod
01:29:28midgeyn1s was posting the .ipod variants of the bootloaders for 3.0 flyspray
01:29:57Nico_Prasher: I think you got mixed up on the IpodFAQ page
01:29:58Lloreanrasher: I think .ipod just has the small target prefix at the beginning to identify what player it's supposed to be for, and the .bin doesn't, but I could be wrong
01:30:20rasherLlorean: Sounds likely
01:30:23n1smidgey: yes, since those were the ones that got built into ipodpatcher i felt it made the most sense
01:31:11n1smidgey: anyways, great to hear that you intend to take another look at that plugin localization thing :)
01:31:44midgeyi better get something posted soon, my free time is beginning to wear down again...
01:31:47 Join homielowe [0] (n=homielow@unaffiliated/homielowe)
01:32:24rasherFS #9955
01:32:41stripwax(apologies in advance if this sort of stuff was already discussed to death, but that said..) - So looking at the usb speeds I'm wondering how much of the pp rockbox usb code originally came from reverse engineering ipod EDM. Some, much, very little, ..?
01:33:02stripwax(by usb speeds I mean the pastebin posts earlier)
01:33:32gevaertsstripwax: speeds are actually mostly limited by the various storage drivers
01:34:23stripwaxinteresting - was the video 5g a 30g or 60g/80g model?
01:36:36kugelrasher: I suppose we also release new sansa/ipodpatcher (etc) when we release new bootloader?
01:36:42gevaertsAs for reverse engineering, the USB controller setup (and related things) are a bit magic, but the actual driver was implemented using the imx31 datasheet
01:37:52 Join __lifeless [0] (n=lifeless@
01:38:22stripwaxYeah, I remember some of the USB controller setup investigation stuff discussion and the magic pp bit-setting init sequence, wonder if full OF (rather than EDM) uses something different, but if bottleneck is storage driver it's a moot point
01:39:25stripwaxI also know practically nothing about USB (do the incoming commands look like individual requests for sector / multisector reads/writes , or something higher level)
01:39:41 Join midijunkie [0] (
01:39:43gevaertsI once got 13MB/s (IIRC) on my c250 when doing dummy writes (or maybe ramdisk, can't remember). That's a while ago though, I should retest this
01:40:10gevaertsstripwax: it's basically multisector reads/writes, using SCSI commands
01:42:34stripwaxdo we do cached reads from ram or is ram not really used for much during usb? wondering how much to-ing and fro-ing there is between reading fat/filesystem and writing file sectors etc
01:42:45midgeyrasher: ipod 4g bootloader seems good
01:43:00rashermidgey: Probably best to report in the tracker
01:43:02stripwaxI think I saw a patch regarding fat caching / fat read-ahead somewhere in the tracker
01:43:05midgeyalready done
01:43:43gevaertsstripwax: it's at a lower level than that. the OS is supposed to do caching if/when needed
01:43:52 Quit n1s ("Lämnar")
01:44:20stripwaxgevaerts - as in, the host OS? just thinking we could do device OS caching too
01:44:49gevaertsstripwax: host, yes. Device caching doesn't make much sense I think, unless we do readahead
01:45:11 Quit homielowe ()
01:45:14midgeyDoes FS #9708 (PP ATA DMA) improve USB speeds at all?
01:45:24 Quit Thundercloud (Remote closed the connection)
01:46:11gevaertsAnd even then, the cases where we could do readahead are probably the ones where the host OS is already doing it, so we're busy serving USB requests anyway
01:46:13stripwaxfinal random thought (for now), any likelihood OF is using DMA to transfer data from ATA to USB registers?
01:46:25stripwaxgevaerts - ah, good point. scratch that then
01:47:11gevaertsstripwax: DMA straight from ATA to USB can't be done with this USB controller, unless I'm missing something
01:47:48gevaertsthe PP USB controller does DMA from RAM though, so we're not busy-waiting or things like that
01:48:17gevaertsmidgey: probably worth looking into
01:48:47stripwaxgevaerts - 3 for 3 then! :) thanks. I'll stop making random guesses now - what are the known storage driver bottlenecks at the moment
01:49:04 Join JdGordon [0] (
01:49:09rashergevaerts: I just got errors...
01:49:34gevaertsI don't really know. We're slower on write in just about any target, but I don't know those drivers
01:49:38gevaertsrasher: what sort?
01:50:23gevaertsrasher: nothing else?
01:50:26rasherThe cable is quite definitely inserted fully. And it only seemed to happen after I tried to write to it
01:50:52rashergevaerts: Sorry, those messages, repeated to fill the buffer
01:51:18gevaertsrasher: anything in /var/log/kern.log?
01:51:58rasherNope, it starts out like that
01:52:10kugelata-sd-pp has some assembly fast-write function, I don't know how they are used though
01:52:19kugelthat doesn't help hdd targets though
01:52:24gevaertsrasher: does lsusb still see the device?
01:52:39*rasher removed the cable :\
01:53:48 Quit midgey ()
01:54:07 Quit _lifeless (Read error: 110 (Connection timed out))
01:55:27 Quit intrados_ (Read error: 60 (Operation timed out))
02:00:14 Join intrados_ [0] (
02:00:45 Part toffe82
02:04:38 Quit stripwax ("")
02:05:08 Join sarixe [0] (
02:05:24 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
02:10:09 Quit rocko (Remote closed the connection)
02:10:44 Join mib_xrvqm1 [0] (i=47cd1fcf@gateway/web/ajax/
02:10:58 Join rocko [0] (
02:12:25 Quit Nico_P (Remote closed the connection)
02:13:28 Join MTee [0] (n=MTee@
02:18:56 Quit cool_walking_ (" ajax IRC Client")
02:19:22 Join cool_walking_ [0] (i=cb3b81c3@gateway/web/ajax/
02:21:34 Quit bmbl ("Woah!")
02:22:35 Nick fxb is now known as fxb__ (
02:25:30 Quit jordoex (Read error: 110 (Connection timed out))
02:26:02 Join _lifeless [0] (n=lifeless@
02:26:14 Quit __lifeless (Remote closed the connection)
02:29:29 Quit jeffronius (Remote closed the connection)
02:34:18***Saving seen data "./dancer.seen"
02:37:03 Quit mib_xrvqm1 (" ajax IRC Client")
02:48:27 Quit Llorean (Read error: 104 (Connection reset by peer))
02:48:46 Join Llorean [0] (
02:50:00 Quit midijunkie (Read error: 104 (Connection reset by peer))
02:50:24 Quit Aurix_Lexico (Remote closed the connection)
02:50:49 Quit sarixe ("Leaving")
02:54:23 Quit jhMikeS (Read error: 54 (Connection reset by peer))
02:54:35 Quit MTee (Read error: 110 (Connection timed out))
02:55:30 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
03:03:13 Quit miepchen^schla (Read error: 60 (Operation timed out))
03:06:19 Quit Xerion (Read error: 104 (Connection reset by peer))
03:15:52UnhelpfulJdGordon: a fairly big deal for a plugin that's become obviously very memory-hungry as i attempt to remove its use of rb->bufalloc. my first try at that basically needs 1MB plugin buffer, which is a bit ridiculous.
03:17:12Unhelpfulhrm, can tagcache be queried for "Nth unique album name"? that's basically what the album index that PF builds is storing, and it would be nice if the big pile of strings that it needs to store for that could go away...
03:18:38 Join CaptainKewl [0] (
03:18:39rasherUnhelpful: If not, it shouldn't be hard to add. The DB has a file with exactly that
03:18:53JdGordonUnhelpful: yeah, I was being... ram saving are always good...
03:19:26JdGordonI think you need to do a search and then go through the results untill you get the Nth one
03:19:32UnhelpfulJdGordon: well, as it stood before, pictureflow had about 150-200KB left in the plugin buffer when it started trying to load covers.
03:19:34JdGordoni.e no automatic way to do it
03:19:53Unhelpfulheh, that's how it builds the index now. maybe i should get some stats on the size of that index before worrying much.
03:19:54 Quit tvelocity (Remote closed the connection)
03:20:42rasherYou could go through the file manually. It's ideally suited for fast lookup. But that obviously means maintaining code twice
03:20:50Unhelpful"freeing" the uniqbuf after the index is built is the largest single savings i've gotten so for
03:21:06 Quit rocko ("Leaving")
03:24:14 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
03:24:30 Join zman977 [0] (n=483043c8@gateway/web/cgi-irc/
03:24:32saratogaregarding the slow USB speed on the e200, there is a patch for doing DMA writes to SD memory
03:24:36saratogaperhaps that would help
03:25:21saratogakugel: did you ever battery bench using the rockbox bootloader without the sandisk one? I wonder if power consumption is the same without the OF's inits
03:25:21 Quit zman977 (Client Quit)
03:26:13 Join homielowe_ [0] (
03:27:22kugelsaratoga: uhm, not recently. But I remember I did that once, and haven't noticed a noticeable difference to previous benches
03:28:27kugelI could make one, though
03:28:52 Join jordoex [0] (
03:29:24saratogakugel: thats good news
03:29:26 Join zman977 [0] (
03:29:40saratogai'd been wondering if we really did all the inits right in rockbox to run optimally without the OF
03:29:59saratogaalso, did you ever benchmark gcc4.3.2?
03:30:45kugelUnhelpful: you could add such a search, I suppose
03:31:30Unhelpfuli'm not sure it'll be needed now. the uniqbuf frees much of the total space used.
03:31:42kugelsaratoga: I'll do a bench with svn bootloader soon, then we'll see
03:32:47kugelUnhelpful: I always had the suspicion that the static hell eats memory more than it needs to
03:32:54 Quit homielowe_ ()
03:33:30JdGordonwhy is the uniqbuf actually needed anyway? shouldnt all entries on disk be forced to be unique?
03:33:35Unhelpfulin many plugins, it won't matter. but, a really *huge* static buffer needed only while initializing plugin data? that's horrid.
03:34:03UnhelpfulJdGordon: i was wondering about that, as well... aren't *album* entries supposed to be unique already?
03:34:07kugelUnhelpful: yes, but the other plugins don't have some hundreds of static vars
03:34:27kugelIt always looked to me that pf has more static vars that normal ones
03:34:54JdGordonUnhelpful: all string values i would thing should be stored as unique
03:35:09JdGordonSlasheri is the man to talk to about this...
03:35:21kugelis he alive?
03:42:24Unhelpfulhrm, might the uniqbuf be needed here if there are two albums with the same name, by different artists?
03:43:42 Quit kugel (Nick collision from services.)
03:43:47 Join kugel [0] (n=kugel@rockbox/developer/kugel)
03:49:48zman977hey i got a quick question. where would be a good place to start to make a bootloader for the sansa view? the view carrys the mi4 firmware
03:50:37kugelUnhelpful: would it matter? pf doesn't really handle anyway
03:51:29kugelzman977: you first need to find out how to run code, then you can think about a bootloader
03:52:27saratogawe can actually run code, its the same as the Sansa, just theres no drivers
03:53:16zman977so we need to firgure out the drivers then?
03:53:18kugelcan we run code without bricking?
03:53:32kugelif yes I'd try to re-use some e200 drivers
03:53:34saratogainterestingly enough, the exploit they fixed on the Rhapsody players isn't fixed on the View
03:53:59saratogai think the only common hardware is the DAC, so reusing drivers is mostly out
03:54:00kugelparticulary the back/buttonlight ones, those appear to be the same for virtually any sansa
03:54:07saratogaah thats true
03:54:23kugelthe dac? I.e. the as3514 too?
03:54:34saratogaits some AMS chip, maybe not the same one
03:54:34zman977i think some one did run the e200 bootloader
03:54:35kugelthen button and backlight will probably work
03:54:47zman977just no lcd and idk about buttons
03:54:54saratogayou can load the e200 bootloader, but without LCD its rather pointless
03:54:58 Quit perrikwp (" ajax IRC Client")
03:55:07 Join perrikwp [0] (i=18ac0c41@gateway/web/ajax/
03:55:11zman977so first stem is to get the lcd working?
03:55:22LambdaCalculus37kugel: It's an AS3517 in the View.
03:55:27kugelzman977: you should try to get some visual feedback by letting the buttonlight or backlight blink
03:55:44 Quit JdGordon (Read error: 110 (Connection timed out))
03:55:59kugelLambdaCalculus37: ah, ok. Well, the backlight and buttonlight is the same for as3514 and as3525, so it's probably the same on as3517 too
03:56:16LambdaCalculus37kugel: Only one way to find out, right?
03:56:31kugeltwo, actually :)
03:56:37saratogai assume the pins aren't the same though, and PP derivied chips have a lot of GPIOs
03:57:02kugelyea, unlike the ams sansas
03:57:36kugelsaratoga: one could think they reintroduced the exploit :)
03:57:52saratogamaybe they don't hate us afterall
03:58:16 Quit saratoga ("CGI:IRC (EOF)")
03:58:19kugelthey should love us
03:58:21 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
03:58:36saratogahonestly, given the difficulty of the port, I'm not sure what the point is
03:58:40saratogajust get a Fuze
03:59:06kugelwell, I haven't seen a view yet, but I always found that a very interesting player
03:59:28kugelmore interesting than the fuze, to be honest
03:59:45zman977i like it because it carrys the goforce 6100
04:00:24kugelbecause it's called goforce, or what's to like about the goforce 6100?
04:00:56saratogathey renamed PP to goforce I think
04:00:57zman977its from nvidia idk lol
04:01:20zman977oh well like i said im now to the protable player stuff
04:01:22kugelsaratoga: goforce existed before the acquisition
04:01:36saratogayeah but not for ARM i think
04:02:01kugelwell, I thought the goforce 6100 was an extra chip in the view anyway (for video)
04:02:16LambdaCalculus37Nope. It's the main SoC.
04:02:34saratogakugel: no its the pp6110
04:02:50*kugel hides then
04:02:52saratogabasically PP but with an ARM11 core
04:03:07kugelarm11, not bad. could be a second beast
04:03:08LambdaCalculus37I wonder what kind of potential power you can get from that. :)
04:03:46zman977idk but isnt the arm11 close to whats in the ipod touch?
04:03:48*kugel needs to find a funny typo for that before anyone else does
04:04:15LambdaCalculus37zman977: Please stop saying "idk", and use full English here.
04:04:36LambdaCalculus37And the Gigabeast already has an ARM11 core.
04:04:41zman977Oh I'm sorry.
04:04:46saratogaif the fuzev2 is actually an AMS353x, its probably pretty comparable
04:04:54saratogathough the view is 32GB
04:05:11kugelams353x comparable ?
04:05:20saratogahonestly, the sooner the view is canceled the sooner we'll have 32GB AMS targets, so I'm not that big a fan of it
04:06:55saratogai guess the whole "flash based portable media player thats not an ipod" market ended up being a lot smaller then sandisk expected, and so theres little reason to replace the view
04:07:45kugelI think they rather bring a view v2, maybe ams based one; before they get rid of it at all
04:08:48kugelbut it's better touch based, there's little point in bringing a upper budget player without
04:09:06 Join rocko [0] (
04:10:19saratogai assume the ViewV2 would just be a fuzev2 with a bigger screen and more video formats
04:10:55kugeland touch based :)
04:11:36LambdaCalculus37I would like to explore more with the View, but I haven't got one, and I haven't much money to get one. :)
04:11:40saratogadon't the fuze and view already look nearly the same?
04:11:48LambdaCalculus37saratoga: Nope.
04:11:58 Join JdGordon [0] (
04:12:10LambdaCalculus37The View is taller and thinner, with a 240x320 LCD.
04:12:24Unhelpfulkugel: if it doesn't handle that already, it certainly ought to ;)
04:12:26LambdaCalculus37The Fuze resembles the 3rd gen nano, and has a 220x176 LCD.
04:13:14kugelUnhelpful: there's a feature request about it I think
04:13:48saratogathe pictures look just like my fuze, but with a bigger screen
04:13:53saratogaeven the wheel and home button look the same
04:13:54kugelLambdaCalculus37: but that's all, isn't it?
04:13:59 Quit avis (Remote closed the connection)
04:14:12kugelso it's pretty much like the fuze
04:15:46 Join blkhawk- [0] (
04:18:06LambdaCalculus37kugel: But certainly not internally.
04:20:51LambdaCalculus37Well, heading off now... be back tomorrow.
04:21:07 Quit LambdaCalculus37 ("Fwump")
04:21:47Unhelpfulthis is kind of disgusting, removing the static uniqbuf has turned a 300KB+ plugin into a 49KB one :/
04:22:22 Join avis [0] (n=ident@pdpc/supporter/student/avis)
04:22:28kugeldoes it still work at least?
04:22:31Unhelpfulon the up side, i think it means that any target with PLUGIN_BUF_SIZE >= 0x60000 can work decently without bufalloc
04:23:43Unhelpfulthe ipod video will have it worst, with room for only about 14 slides in cache
04:25:00 Join bs66_1 [0] (
04:25:44zman977Well thanks for the help, i will mess around with the view tomarrow.
04:26:39 Join CaptainKwel [0] (
04:27:37Unhelpfuli think that implementing cache-to-core-buffer-first is still worthwhile, and it might be good to let the cache grow larger than the current limit, in that case
04:29:17 Quit CaptainKewl (Read error: 60 (Operation timed out))
04:30:15 Quit zman977 ("Trillian (")
04:30:33FlynDicekugel: Do you know if mp3 playback works from the micro sd slot on the fuze? There was a report that it was working that way on an e280v2.
04:31:45kugelNot for me
04:31:52 Quit blkhawk (Read error: 110 (Connection timed out))
04:32:44 Nick blkhawk- is now known as blkhawk (
04:32:53kugelsaratoga: does mp3 use iram?
04:33:41kugelI was wondering if the reboot is some kind of wrong iram access
04:34:16kugelsaratoga: also, have you tried to build gcc with what I posted to the ml?
04:34:21***Saving seen data "./dancer.seen"
04:39:15saratogakugel: it uses IRAM extensively
04:40:00saratogakugel: i haven't had time lately for programming, I will try to take a look at it as soon as I can
04:40:17saratogai'd guess i'm just not properly flushing some cache or something similar
04:40:36 Quit bs66_ (Read error: 110 (Connection timed out))
04:40:53Unhelpfulok, now that it works decently without requiring 1MB plugin buffer, anybody want to give FS #9919 a try? it's working for me on beast, with the entire contents of the player loaded to playlist :D
04:43:38saratogakugel: are you sure it was in mad_synth_thread? that function is just 5 lines long, all it does is wait on a sempahor then call something else
04:44:02 Join lymeca [0] (
04:44:17kugelsaratoga: I was surprised too, but I'm fairly sure
04:44:57kugelI think I have the map and and codecs file somewhere if you want to take a look at it
04:45:07saratogai will look into it once I build the compiler
04:45:11saratoganow i have to get back to work
04:45:33 Quit saratoga ("CGI:IRC (EOF)")
04:48:50 Join sarixe [0] (
04:53:49 Quit kugel ("ChatZilla 0.9.84-rdmsoft [XULRunner]")
04:56:14 Join BHSPitMonkey [0] (n=stephen@unaffiliated/bhspitmonkey)
05:05:31 Quit rocko ("Leaving")
05:17:11 Quit Lss (Read error: 110 (Connection timed out))
05:17:51*JdGordon seems to have a very stupid but probably simple bug somewhere... and cant find it :(
05:18:41Dhraakelliandude, I owe people certain people beers now, should I ever run into them
05:19:25*JdGordon is happy to collect on their behalf
05:21:40 Quit HellDragon (Read error: 104 (Connection reset by peer))
05:22:36 Join HellDragon [0] (
05:23:41 Quit _lifeless (Read error: 113 (No route to host))
05:31:52 Quit Horschti ("Verlassend")
05:34:06 Join insanepotato [0] (
05:35:26insanepotatodoes the "dual ARM" described in the portal player data sheets mean 2 CPUs? does RB use both CPUs?
05:35:58 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
05:37:18saratogatheres two CPUs, and some things are able to use them [mpegplayer, mp3 decode]
05:45:57 Join rocko [0] (
05:50:18 Quit saratoga ("CGI:IRC (EOF)")
05:51:37 Quit insanepotato ("ChatZilla 0.9.84 [Firefox 3.0.6/2009011913]")
05:53:54 Quit lymeca (Read error: 110 (Connection timed out))
05:56:24 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
06:01:05 Join _lifeless [0] (n=lifeless@
06:04:59 Join homielowe [0] (n=homielow@unaffiliated/homielowe)
06:16:51 Quit tarbo (Connection timed out)
06:17:47 Join tarbo [0] (n=me@unaffiliated/tarbo)
06:26:54 Join bertrik [0] (
06:29:48 Quit avis (Remote closed the connection)
06:34:25***Saving seen data "./dancer.seen"
06:36:44 Join avis [0] (n=ident@pdpc/supporter/student/avis)
06:41:11 Join midgey [0] (
07:06:43 Quit bertrik (Remote closed the connection)
07:15:30midgeyUnhelpful: I'm trying out your patches for pictureflow right now. is there anything i should be looking for?
07:16:44 Quit FlynDice ("Konversation terminated!")
07:17:38 Join Kanzar [0] (
07:17:53Kanzarok, this is only very vaguely related - but does anyone have the old firmware files for the 2nd gen nano? i need to downgrade it :(
07:19:22Unhelpfulmidgey: not really. if you turn the number of slides up too high, you'll likely see some flickering back and forth between the proper cover image and the no-cover image. i'm working on a more stable method (in terms of not repeatedly changing its mind about what to keep) of managing the cache, which should solve that.
07:19:50Unhelpfulthe main thing is that it should still work as well during playback, even with the audio buffer full, as when not playing.
07:20:27 Join FlynDice [0] (
07:22:34midgeyerr, whats up with the status bar in the settings menu?
07:22:51midgey(i havent used pictureflow so it's probably not new)
07:25:11 Quit FlynDice (Remote closed the connection)
07:25:16midgeyseems to work pretty well while playing music (on my gigabeat), nice work
07:30:08 Join FlynDice [0] (
07:33:14amiconnUnhelpful: My opinion on this is that PF should have its own buffering system, but still be able to *optionally* use the audio buffer (also using its own buffering system, not core bufalloc)
07:33:43amiconnThen it would restrict itself to the plugin buffer if music is playing, but use the audiobuffer if music is stopped
07:34:02amiconnOn lowmem targets it will have to use the audiobuffer anyway
07:34:57amiconnDo you know how much of those 49KB are bitmaps, i.e. wil become smaller on monochrome?
07:36:10Unhelpfulthe splash bitmap is about 13KiB on my color targets, i believe. i can also remove the static allocation for the album index, it was just easier to tackle either the titles cache or that by itself, than to do both.
07:36:30Unhelpfulthe titles cache *probably* has to stay for now, as tagcache doesn't
07:36:55midgeysounds like that splash screen is a good place to save size
07:36:57Unhelpfulsilly enter key in the way. tagcache doesn't (i think) allow one to query for "Nth unique album"
07:37:48Unhelpfulmidgey: it will be much, much smaller on greyscale targets - on the order of about 1/8 the size, i believe, for the same resolution?
07:38:44midgeyi meant just size to save in general (mostly for color). it would have basically no effect on greyscale
07:39:39*amiconn is curious whether PF will become small enough to fit into the archos plugin ram
07:39:40Unhelpfulwell, even on e200, that's still only a little over half the size of a single cached slide
07:40:10Unhelpfulamiconn: well, it should be possible, almost, to build it on archos now...
07:41:21Unhelpfulwhy would you say we shouldn't use bufalloc? i'd rather not grab the whole audio buffer, in the event that we can eventually start playback from PF without leaving it.
07:42:01Unhelpfulalso, if audio is playing, but the buffer is not full, core bufalloc can still get us a larger slide cache than just the plugin buffer
07:46:39amiconnYou would need to support two APIs: bufalloc and the internal one
07:46:56amiconnAlso, I don't really trust bufalloc to work properly in a mixed-use case
07:49:40Unhelpfuli don't see a big problem with the two API case, it's not much harder than "try to alloc from one buffer, then the other"
07:50:15Unhelpfulideally, a "discontiguous mode" for buffer_alloc would be best, if it has two buffers to work from
07:50:52cool_walking_Which file (bootloader.bin or bootloader-ipodvideo.ipod) and switch for ipodpatcher do I want to use for testing the 3.2 bootloader?
07:53:08Unhelpfulan easy way to test core bufalloc would be to use the "old" PF with enough music loaded to partially fill the buffer
07:53:39 Quit CaptainKwel (Read error: 110 (Connection timed out))
07:59:43midgeycool_walking_: either one works supposedly (I haven't tried)
07:59:58midgeyI used the .ipod file with the -a flag on my 4G
08:02:07cool_walking_Well that's confusing. Why are there two files then? and 4 switches (-wf, -wfb, -a, -ab) that all sound like they might do the same thing?
08:04:40amiconn-wf and -a aren't doing the same thing
08:05:11cool_walking_but the names "add" vs "write" aren't very enlightening.
08:05:20cool_walking_what's the difference?
08:05:25amiconn-wf *replaces* the OF with the bootloader, meaning faster boot at the cost of losing dual boot
08:05:35amiconn-a *adds* the bootloader to the OF
08:05:35cool_walking_is that safe?
08:06:02midgeyyou'll still have disk mode
08:07:07cool_walking_okay so what's the difference between -wf and -wfb?
08:07:48cool_walking_oh, are the .bin and .ipod just a different file format, but you get the same result?
08:07:55amiconnThe 'b' variants take a .bin file, the ones without 'b' take an .ipod file
08:08:41cool_walking_okay, thanks.
08:09:21amiconnThe .ipod is the .bin with a little header prepended that tells which model it is intended for, the file size, and a checksum, so corruption can be detected and installation on a "wrong" ipod can be prevented
08:09:40amiconnOnce installed, the result is exactly the same
08:17:25 Quit BHSPitMonkey (Remote closed the connection)
08:21:13*Kanzar sighs :(
08:25:35cool_walking_Is Rockbox supposed to stop snap out of the USB screen when you do a "safely remove" or "eject"? It used to on my Gigabeat S, but since a while ago it started just staying at the USB screen. Now after updating my iPod video to a build with Rockbox USB, I notice it's doing it too.
08:29:50 Join Zagor [242] (n=bjorn@rockbox/developer/Zagor)
08:31:26 Join Rob2222 [0] (
08:34:22 Join ender` [0] (
08:34:28***Saving seen data "./dancer.seen"
08:43:04 Quit scorche (Read error: 54 (Connection reset by peer))
08:43:59 Join scorche [50] (n=scorche@rockbox/administrator/scorche)
08:44:06 Quit cool_walking_ (" ajax IRC Client")
08:44:44 Join petur [50] (n=petur@rockbox/developer/petur)
08:49:14 Join Bagderr [241] (n=daniel@rockbox/developer/bagder)
08:49:39 Quit Rob2223 (Read error: 110 (Connection timed out))
08:49:41 Nick Bagderr is now known as B4gder (n=daniel@rockbox/developer/bagder)
08:49:54 Join jordoex_ [0] (
08:56:43 Quit jordoex (Read error: 60 (Operation timed out))
08:57:59 Join rvvs89_ [0] (
08:58:08 Quit rvvs89 (Read error: 60 (Operation timed out))
09:08:30 Quit perrikwp (" ajax IRC Client")
09:21:51 Quit Seed ("cu, Andre")
09:22:56 Join einhirn [0] (
09:27:26 Join gartral [0] (
09:27:45 Quit z35 ("Leaving")
09:28:23 Quit rocko (Remote closed the connection)
09:33:14gartralmy E250 boots into the OF on usb insert still..
09:34:29kadobangartral: i believe you need a new bootloader to change that. there are ones being tested for 3.2 in flyspray
09:35:38linuxstbgartral: are you inserting usb before turning it on, or when Rockbox has started? if the first one, then what kadoban said.
09:39:59 Join rocko [0] (
09:45:27 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
09:46:00gartralnah, found the new bootloader, thanks!
09:46:22*kadoban is glad that his completely unfounded assumption magically turned out correct
09:48:19*LinusN performs the late but happy USB mass storage dance
09:50:24 Quit _lifeless (Remote closed the connection)
09:50:41 Join _lifeless [0] (n=lifeless@
09:55:10 Nick rvvs89_ is now known as rvvs89 (
09:58:26 Join Thundercloud [0] (
10:02:26*gartral joins LinusN
10:04:34Kanzarok, this is only very vaguely related - but does anyone have the old firmware files for the 2nd gen nano? i need to downgrade it :(
10:05:35 Join n1s [0] (n=n1s@rockbox/developer/n1s)
10:10:51midgeyKanzar: your best bet is probably asking the linux4nano guys
10:11:07Kanzaryeah, i'm asking in there atm :P
10:29:47 Join gromit`` [0] (
10:34:30***Saving seen data "./dancer.seen"
10:34:36 Quit Thundercloud (Remote closed the connection)
10:37:56 Quit gromit` (Read error: 110 (Connection timed out))
10:38:13 Quit nuonguy (Read error: 110 (Connection timed out))
10:50:26 Join nibbler [0] (
10:51:12 Quit midgey ()
10:53:12advcomp2019the new bootloader will not boot my sansa e280R anyway i try
10:58:23linuxstbadvcomp2019: Did you build an e200r bootloader? rasher's zip doesn't have one.
10:59:16advcomp2019o i did not know that.. that was useful info now heh
11:01:21advcomp2019i posted in the FS #9955 without knowing that
11:02:12linuxstbI guess rasher forgot about that difference as well - the main builds are the same, but the bootloader .mi4 files need to be signed differently.
11:02:47linuxstb(e200 vs e200r)
11:04:28gartralall running happily on my e250
11:04:53advcomp2019yea and i have not reinstalled vmware to this computer since i have not needed it on my laptop
11:08:12 Join SliMM [0] (n=SliMM@
11:08:54SliMMis the __PCTOOL__ part of tagcahe working?
11:08:54advcomp2019SliMM, hello
11:12:53 Quit kachna (Read error: 113 (No route to host))
11:15:05B4gderSliMM: make in tools/database seems to fail. what exactly are you asking?
11:16:06advcomp2019linuxstb, i reinstalled the older bootloader via recovery and it is working.. plus i posted on that FS#
11:16:30SliMMB4gder: if I can ignore the parts that are in #ifndef __PCTOOL__ when using the code for a pc project
11:16:37 Join MTee [0] (n=MTee@
11:17:10linuxstbSliMM: Yes, that's the idea of that define.
11:17:23linuxstbBut it may not work...
11:17:55linuxstbI think there's also a flyspray task which tries to make independent building of the database code better. I'm not sure what state that is in though.
11:22:45B4gdernow 'database' builds at least
11:24:35 Quit rocko ("Leaving")
11:26:05MTeelinuxstb : after parsing a rm file, with the decoder now known, what should happen ? it seems to me like we enter a DATA chunk, send its packets to be decoded, till we
11:26:27linuxstbMTee: Oops, I've just remembered I forgot to reply to that email... ;)
11:26:34MTee*till we're done with this chunk, and then go to the next DATA chunk and so on
11:26:47MTeelinuxstb : never mind :)
11:27:26linuxstbI'm not sure of the answer though... Did my parser get as far as printing out the details (i.e. offset in file and length) of the data chunk(s) ?
11:28:24linuxstbI think what I would do next is find an example file with an "easy" codec (like mp3 or ac3) and make the parser write the raw audio data to a file. You should then be able to play that file.
11:29:42linuxstbI seem to recall that realaudio files have some strange psuedo-encryption involving XORing the data with something. Looking at the source code of the RealMedia parser in ffmpeg (or I think, vlc) should help a lot.
11:31:04 Part gartral
11:32:28MTeethis is sort of my question ; should I loop through the chunks sending packets to the decoder, or should the decoder handle the rest of the process ? (sorry if it sounds stupid :) )
11:33:27linuxstbDo you know how big these data chunks are?
11:35:46linuxstbBut generally, the actual decoder will contain a function called something like "decode_frame()" which will take a pointer to a small amount of compressed audio, and a pointer to an output buffer, and decode that audio into the output buffer.
11:36:42linuxstbSo the actual codec will first call the parser to get a pointer to the next data chunk, and then (probably repeatedly) call that decode_frame() function to decode the data.
11:37:47MTeeI see, thanks a lot.
11:39:01linuxstbThe codec then calls the Rockbox "pcmbuf_insert()" function to send the output data from decode_frame to the Rockbox DSP layer (resampling, equalizer, crossfeed etc), which then sends it to the DAC...
11:42:08 Quit ChanServ (shutting down)
11:42:40 Join ChanServ [0] (ChanServ@services.)
11:42:40Mode"#rockbox +o ChanServ " by
11:42:40DBUGsent MODE #rockbox -o ChanServ
11:42:40***Server message 485: 'logbot ChanServ #rockbox :User is immune from kick/deop'
11:43:44 Join PaulJam [0] (
11:46:08MTeeis it possible to implement a lyrics viewer for the WPS ?
11:46:29rasheradvcomp2019: Added an e200r bootloader
11:46:54advcomp2019rasher, thanks
11:47:27 Nick flx_ is now known as flux (
11:48:59 Join petur2 [50] (n=petur@rockbox/developer/petur)
11:51:03linuxstbMTee: Only if you add that code to Rockbox itself - it can't be done with a plugin (unless of course, you add plugin capability to the WPS...). However, there's a lyrics viewer patch on flyspray which implements a plugin which shows lyrics (i.e. it replaces the WPS)
11:51:25linuxstb(I hope that makes sense)
11:52:05 Join pyro_maniac [0] (
11:54:41MTeelinuxstb : yes it's clear. I was talking about adding the code not writing a plugin. Thanks
11:57:27 Join robin0800 [0] (
11:58:53linuxstbYou'll have two problems doing that: 1) Managing memory (i.e. where do you store the lyrics?) and 2) other Rockbox developers may not want lyrics support taking up code/RAM in the Rockbox core...
11:59:34Unhelpfullinuxstb: what about doing it both ways - allowing a plugin to load and be displayed in a viewport in the WPS?
11:59:41*rasher sees there's a h10_5gb bootloader issue
12:00:06linuxstbUnhelpful: Yes, that idea has been mentioned a few times. No-one has implemented it though...
12:00:16Unhelpfulsolves the memory management problem, and the code bloat problem, as the bulk of the "new code" ends up in the plugin buffer.
12:00:42Unhelpfulit seems like the way to do that, and lots of other "fancy things some few users think are needed"
12:00:43linuxstbThis discussion always goes down the path of suggesting that the WPS itself becomes a plugin...
12:01:07 Quit petur (Nick collision from services.)
12:01:07linuxstbHow does it solve the memory management problem? Where do wps plugins live?
12:01:11 Nick petur2 is now known as petur (n=petur@rockbox/developer/petur)
12:01:51Unhelpfullinuxstb: they live in the plugin buffer. aside from very few special TSR plugins, you can't actually run a "regular" plugin while using wps anyway, right?
12:02:12 Quit GodEater_ (" ajax IRC Client")
12:02:32linuxstbNo, but you can run plugins when playback is active. And wps plugins probably want to be running all the time.
12:05:46 Quit robin0800 (Read error: 104 (Connection reset by peer))
12:05:53MTeewhat about a plugin which just detects the file being played and displays the lyrics accordingly ? not in a viewport in the WPS.
12:07:00Unhelpfulwhy, to save load delay when you go to the WPS? i prefer the idea of displaying a plugin in a WPS viewport, anyway, to making the WPS itself a plugin - i think it's a nice balance, and that a limit of one "custom item" is reasonable
12:07:36Unhelpfulalso, a million "alternate WPS" plugins turns into a maintainence nightmare, since most are probably based on "normal" WPS
12:08:09 Join GodEater_ [0] (i=c2cbc962@rockbox/staff/GodEater)
12:08:32advcomp2019rasher, i found a bug on the e200r bootloader too
12:08:37LloreanUnhelpful: Displaying some plugins in a WPS viewport would mean that some WPSes could be a single fullscreen plugin viewport.
12:08:50LloreanBut some plugins would have to better handle different screen sizes (or have a flag where they require fullscreen)
12:09:35UnhelpfulLlorean: i don't really think we'd expect absolutely any plugin to work in a WPS viewport - clearly the ones that require input are no-go
12:09:59GodEater_why isn't the USB stuff in the project news bit on the front page? Isn't it big enough?
12:10:04Unhelpfulmaybe they should use the plugin loader, and the plugin buffer, but have an entirely different magic in the header, or something?
12:10:30LloreanUnhelpful: I actually don't think input is the problem. The plugin's input would override the WPSes until it's closed.
12:10:47advcomp2019if you plug in the sansa when it is off, it will turn on to rockbox but it will not mount it as a usb drive
12:10:52*Llorean imagines playing Gameboy games in a viewport while still being able to see playback info, for example
12:11:33Unhelpfulah, that's not really what i was thinking of at all, i was thinking more "a way for people to put visualizaton, lyrics, etc into WPS"
12:11:49LloreanMy thought *started* with visualization, but went on to "why not everything else?"
12:12:00LloreanYou could open the text editor, scroll through lyrics, then close it, etc.
12:12:03Unhelpfulthough, i suppose the same code could support both - a header flag could mark ones that shouldn't steal input
12:12:18LloreanA question is how to close ones that don't steal input.
12:12:31LloreanI'd suggest that "visualization" plugins still steal input, they just implement *most* of the playback controls.
12:13:54Unhelpfulthat might be better, but that means putting keymaps for the playback controls in a bunch of plugins. i was rather thinking that the WPS would specify the plugin to load, and you don't interact with it, which would all work fine for visualizations...
12:14:44LloreanUnhelpful: Yeah, but you could just have a standard visualization plugin header or something.
12:15:34Unhelpfulyou mean to handle the keymaps? yes, that might work
12:15:34gevaertsGodEater_: Zagor wasn't awake :)
12:15:55*Zagor rubs eyes
12:16:22LloreanUnhelpful: Yeah, to handle the keymaps. Of course it's also possible just to have the "Menu" button cancel plugins on first press so you're back to the normal WPS, then bring up the menu on second press when plugins are/were running
12:21:29Zagordo we have a useful wiki page that lists which targets have pp502x ?
12:23:08Zagorthat's the one I was thinking of. it's not terribly newbie friendly though :-)
12:23:35gevaertsZagor: look at the files touched by r20105
12:24:04Zagorgevaerts: yeah I know, but I need something that explains to users which targets got this. for the front page.
12:24:09linuxstbWe have a PortalPlayer wiki page I think - they could be added there.
12:24:25LloreanZagor: I posted a list of targets in the forum
12:24:27Zagormaybe I should just list them all
12:24:32ZagorLlorean: ah, link?
12:25:06ZagorI'll link to that
12:29:06 Join gromit` [0] (
12:34:34***Saving seen data "./dancer.seen"
12:37:04 Quit gromit`` (Read error: 110 (Connection timed out))
12:41:45*linuxstb reads
12:42:34LloreanThey can resell PIDs?
12:42:46 Join gartral [0] (
12:42:54linuxstbI don't think so, but they seem to...
12:43:00LloreanI was just gonna suggest we disable USB again, and tell people we need a VID if we really want to do it "right" and run a funding drive. =P
12:43:25linuxstbThat's not such a bad idea... $2000 isn't a huge amount of money for a community of our size.
12:43:43 Quit SliMM (Remote closed the connection)
12:43:47linuxstbI would expect 200 people could donate $10
12:43:48*GodEater_ wonders what's in the rockbox fund currently
12:44:13LloreanThe donation part was serious, the "hold USB for ransom" not. :)
12:44:32LloreanThat's why I was asking if the $2000 was a one-time fee (if I read it right, basically) as it doesn't seem like an unattainable goal
12:45:07linuxstbHmm, it seems to be $2000 for two years -
12:45:28GodEater_boo hiss
12:45:33Lloreanlinuxstb: Read the last paragraph
12:45:44LloreanIf you wish to purchase a VID without licensing the logo rights, that's a $2000 _purchase_
12:46:01linuxstbAh yes, the license appears to just be for the logo...
12:46:14LloreanThe question is "do they mean by purchase what we'd think they mean" since it mentions no term.
12:47:13linuxstbBut this does seem the "correct" thing to do, especially if we implement usb modes the OF doesn't.
12:47:32*linuxstb wonders what gevaerts thinks
12:47:40LloreanI wouldn't want to bother if we just have to ask for money every year / two years to maintain it.
12:48:04LloreanBut if we really can purchase a VID permanently (which honestly, makes sense since revoking them would cause infinite chaos in the long term) we probably should try to raise the money
12:48:12 Quit _lifeless (Remote closed the connection)
12:48:14 Join moos [0] (i=Mustapha@rockbox/staff/moos)
12:48:58linuxstbThis is the application form - it certainly looks like a one-time payment -
12:49:20LloreanI don't know how you *would* revoke VIDs
12:49:24B4gder"A one time processing fee" it says
12:49:28linuxstbI guess there's the technical issue that there is no "Rockbox" company.
12:49:52B4gderwe could use a proxy company we trust
12:50:03LloreanWe could probably just contact and ask if a "Company" is necessary, or if that's just a field for the ID
12:50:05B4gderlike that, yes
12:50:34linuxstbB4gder: Do you think this is worthwhile?
12:51:11B4gderI'm not sure
12:51:57B4gderI mean I'd be fun and nice, but is it 2000USD nice?
12:52:27linuxstbI think it should be "extra" money - i.e. as Llorean suggests, a fund-raising effort just for this purpose.
12:52:30LloreanWe'd need to come up with a list of things it could/would allow us (and other projects interacting with us in the future) to more easily do.
12:52:41gartralwell, just out of pure wonder, what would buying a VID offer that we dont already have?
12:52:42LloreanWhat hurdles does using the original VIDs present for us?
12:53:52 Join jaykay [0] (
12:54:05Lloreangartral: In one sense, it's just an issue of "doing it right". Arguably there's nothing we can do with one that we can't do without one.
12:54:41*linuxstb tries to find the previous discussion about this in the IRC logs
12:54:42GodEater_well one issue is that iPods don't work 100% properly with iTunes right ? Since we don't support the extra SCSI commands.
12:54:50LloreanBut having Rockboxed players identify as Rockboxed players immediately can allow software interacting with it an easier job of knowing if Rockbox is installed (and currently running / providing the USB connection) so if we have special USB functionality it's easier to know when it can be used.
12:56:28linuxstb starting at 15.11.06
12:56:29gartralinteresting side note, the "no sansa detected" glitch with rbutil does not show at all after 20105 (in rockbox only)
13:00:49LloreanGodEater_: So having our own ID would break iTunes support completely, but we could add an "iTunes compatibility mode" option which is described as 'incomplete' until/unless we add the other magic.
13:03:16 Join Casainho [0] (
13:03:20Casainhohello :-)
13:03:27Casainhocan someone explain me the "pcm_play_data_start(unsigned char *start, size_t size)" ?
13:04:08CasainhoWhen does Rockbox call it? everytime audio fifo buffer is full?
13:04:11linuxstbI was mainly thinking about when we add new USB features - e.g. usb audio or HID. Would new IDs be needed for that, or can the same IDs be safely used for lots of different modes?
13:04:40LloreanThe same IDs can surely be used, but it'd probably make everything easier all around if we presented different ones.
13:05:22linuxstbCasainho: What fifo buffer?
13:06:05Casainholinuxstb: well, the *start
13:06:43LloreanAnd it would be an optional donation drive. We could present, realistically, that buying a USB VID is non-essential but something we'd "like" to do (in part because it's kinda the 'right' thing to do, and in part because it does make the project more professional in outward appearance and more flexible) and let people decide for themselves whether to donate or not. Then if we don't have $2000 within, say, a year we just roll it into the R
13:09:10 Quit Casainho ("ChatZilla 0.9.84 [Firefox 3.0.6/2009011913]")
13:10:04gartralLlorean: you got $10 from me!
13:11:19linuxstbHas anyone tested Rockbox USB with itunes yet?
13:11:23LloreanI guess the question is more, do people object to the idea of us actually "requesting" donations for something like this? It's nonessential, but nice to have.
13:11:38Lloreanlinuxstb: We had one report of iTunes not showing the iPod, but MacOS showing it specifically as an iPod still.
13:11:53linuxstbLlorean: I think it all depends on what the IDs will bring to Rockbox - I'm waiting for gevaerts...
13:12:36Lloreanlinuxstb: I think the only thing they "give" us that we cannot do in some other way is forcing applications not to mistakenly identify us as their player.
13:13:07LloreanEverything else I'd imagine could be done despite original VID/PID by way of some other channel of interrogation.
13:14:34LloreanI think if it was really "worth" $2000, we'd be discussing whether or not to use Rockbox fund funds on it (if they were available in that quantity)
13:18:52B4gdergoogle supposedly will pay that sum for each gsoc year we've participated...
13:19:05 Join _lifeless [0] (n=lifeless@
13:23:50gartraland what about UIDs?
13:25:20linuxstbWe would buy a "vendor ID", which then gives us 65536 (or maybe slight less...) product IDs to choose from.
13:25:20B4gderwhat about them?
13:25:24gartralbut its it possible to reuse those and use a rockbox VID?
13:25:36gartralohh, nvm
13:25:53gartrali didnt understand how it worked
13:26:46B4gderwell, we could also re-sell PIDs like that other company ;-)
13:27:20LloreanB4gder: Well, there might be Rockbox hardware in the future...
13:27:49B4gderyes, but 64K pids are a lot, I'm sure we can manage with a subset of them
13:28:09gartraland i'm just curious how accessories would be handled....
13:28:16linuxstbThe application form says we can't resell though...
13:28:31B4gderisn't that for the vid?
13:30:00LloreanFrom the form, it actually sounds like PIDs can't be resold, rather than VIDs
13:33:13 Join nazir [0] (
13:33:27gevaertsB4gder: I wouldn't recommend reselling PIDs. The USB IF tends to frown on that, and I suspect that it's one of the ways you *can* lose your VID
13:33:32LloreanI think the idea is that the VID/PID combo is "one item", if you transfer the VID it comes with the PIDs and vice versa.
13:33:56fluxgevaerts, but how do you really lose a VID? who would want to reuse a VID that's been used by someone else?
13:34:14fluxjust keep using it, extremely unlikely anything will collide
13:34:16B4gderflux: want? I'm sure you get one assigned
13:34:32fluxb4gder, and perhaps they'd be extremely pissed at that
13:34:34LloreanB4gder: Yeah, but you would be upset if you were assigned a VID in use on dozens of legacy devices.
13:34:37B4gder"here's your VID sir, enjoy"
13:34:49B4gderyou would certainly, yes
13:34:57flux"well, that's great, I'll pick my own VID thankyouverymuch!" and the VID system would collapse :)
13:35:23Lloreangevaerts: Since you're here - any thoughts on pros/cons of having our own VID?
13:35:34gevaertsSome people can choose their own VID. I'm pretty sure that the one Intel has was not just the next in the sequence
13:35:36fluxI don't see why they would have a reason to get upset by someone selling PIDs. in a business, 2k USD is not that much, it is doubtful it is that great business for them anyway?
13:36:00gevaertsflux: whether they have a reason or not doesn't matter. Historically they *have* been upset
13:36:27bombardierhas anybody noticed that fuze OF crashes if you try to delete file while there is .rockbox directory present? shall we report this bug to sansa? :)
13:37:07B4gderbombardier: "sorry, no support with rockbox installed" ;-)
13:37:56gevaertsLlorean: pro: it's cleaner, and we have no risk of conflicts with host-side software (not just itunes. Drivers tend to have lists with devices that need special treatment due to bugs). con: the price
13:40:28LloreanWhat would happen if we just picked an unused VID and went with it?
13:40:57gevaertswe risk future conflicts
13:41:14gartrali would think all would go well until someone gets that VID assigned too theem, and we his conflicts
13:41:22gevaertsI really prefer to use the OF PID/VID to any "shady" tricks
13:41:23BigBambiand pissing off people that matter?
13:41:30gartrals// theem/them
13:42:11BigBambigevaerts: And forgetting the cost for a moment, you would prefer buying a VID to using the OF ones?
13:43:08gevaertsBigBambi: if we forget the cost, clearly yes.
13:43:35BigBambigevaerts: Cool, I wasn't sure if there were any other downsides
13:43:49Lloreangevaerts: I'm not sure just picking one *is* shady, really. I mean, I'd prefer to buy one, but if there's an empty one I'm not really sure I agree that it's particularly wrong to say "we're going to identify ourselves as this, at least until a problem arises".
13:44:22BigBambino licensing or other issues?
13:45:15LloreanWhat license?
13:45:30BigBambiI don't know, that's why I asked
13:46:03BigBambiI didn't know if using them without 'permission' would get some USB licensing body down on you or something
13:46:23gevaertsLlorean: *now* it's probably not a problem. If we ever want to deal with the USB IF directly, I can't imagine this being happy about it
13:46:23 Join kachna [0] (
13:46:37Lloreangevaerts: This is true.
13:46:51LloreanThen again, arguably they really ought to let us buy a number from them in the future to clear up any problems.
13:47:01LloreanIf they don't, they're just encouraging people to ignore them and maintain their own community VID list.
13:47:11LloreanSince if you ever do it "wrong" you're not allowed to fix it.
13:47:11gevaertsMaybe we're better off finding a cheap PID reseller. That's against *their* contract, but we can't help that
13:47:32LloreanI think if we're going to spend money, we at least ought to *try* to get a VID.
13:47:41gevaertsDepends on the price
13:47:58 Part nazir
13:48:06gevaertsi.e. if we could get say 50 PIDs from such a reseller for $100, I'd seriously consider that
13:48:34LloreanThe one linked by linuxstb was a bit more than $10 / PID
13:48:52gevaertsBut anyway, we haven't seen any practical problems with reusing the OFs ids yet, so I'd just do nothing for now
13:49:08LloreanI think we'd probably want $5/PID or less, from someone we expect to stay around in case we need more, if we really wanted to buy just PIDs
13:49:18gevaertsLlorean: $10 per PID is way too much for us. If we use our own, we need a separate one per device
13:52:27Lloreangevaerts: I'd want to see evidence that the PID reseller had the necessary written permission too
13:52:34LloreanI'd hate to buy PIDs just to find out their VID got revoked.
13:52:37linuxstbgevaerts: Do you know how WIndows for example deals with usb drivers? i.e. is there one driver per vid/pid pair, or one per ID/class combination or ...
13:52:56gevaertslinuxstb: yes :) (i.e. all of them)
13:54:05linuxstbSo we could simply add new USB class drivers to Rockbox, and Windows should be happy? (i.e. not interfere with existing drivers for that device)?
13:54:58gevaertsIt *should*, but I imaging that VID/PID specific drivers get priority
13:56:50linuxstbI'm wondering if this is causing problems with the Beast - using the same IDs for both MTP and UMS. Some people have reported that Windows doesn't detect their Beast in MTP mode.
13:57:37gevaertscould be...
13:57:52 Quit Kanzar ("I'm a traitor to the cause, a hero without applause~")
13:59:28linuxstbMy view is that it just seems the "right thing" to use our own IDs, and I think $2000 could be raised by special donations.
14:01:16*B4gder agrees
14:01:34 Join Aurix_Lexico [0] (
14:02:26LloreanMy view is *exactly* linuxstb's stated view just there.
14:03:14LloreanAlso - first draft letter to USB-IF
14:04:24LloreanIt definitely needs some polishing in the wording, but it's the points I'm bringing up that's most important right now.
14:05:28B4gderLlorean: I think the first point needs to point out slightly clearer that we've _replaced_ the original usb software that has a VID/PID pair we reuse
14:05:39gevaertsLlorean: I would cut the " probably least likely to be answered positively" bit
14:06:14gevaertsAlso, maybe explain what rockbox is a bit more (which would help clarifying B4gder's point)
14:06:39 Quit fyrestorm (Read error: 104 (Connection reset by peer))
14:08:42gartralpossibly add the What Is Rockbox block of the FAQ?
14:08:48*Llorean posts an update.
14:08:55Lloreangartral: Don't want to make it too long.
14:09:31linuxstbAre there many other open source projects which would use USB ids? I can imagine hobbyists building their own usb devices wanting one-off IDs, but do we want to go down that road?
14:09:41linuxstbs/many/any/ ?
14:09:44LloreanI've also changed "allow them to avoid having to..." to "allow them to avoid choosing to..." since the idea of them having to makes USB-IF sound like the bad guys, while OSS projects choosing to makes it sound much more neutral
14:13:56 Join midijunkie [0] (
14:14:08linuxstbLlorean: I would say something about Rockbox not changing the USB behaviour on devices with hardware bridges, but completely replacing the implementation, and adding new features on the devices with software USB. This is the reason we want to use our own IDs - because we are completely replacing the USB implementation on those devices.
14:15:08B4gderpossibly usb-if would also consider it to be a new vendor when we have a new stack, thus it would fit how these things are normally done
14:15:36B4gderbut then replacing usb stacks are probably not that that common
14:15:45linuxstbMaybe that's simply the first question to ask them - do they think we should use our own IDs in this situation?
14:18:09 Quit jaykay (Read error: 110 (Connection timed out))
14:19:54B4gderJ-23: on the sansas it was found out that the usb core was identical to the one in imx31 which has docs, so a similar thing _could_ be done for the 3525
14:20:06B4gderbut for this we already know who made the usb block
14:21:58*Llorean updates again, with a new ending to the first paragraph and new beginning to the second -
14:24:28B4gderlooks fine to me
14:26:04LloreanThe only disadvantage to this is that we're running up to them waving a flag saying "We're using other companies' VIDs." I don't really imagine that'll be much a problem beyond a probably worst case of them saying "you shouldn't do that, tsk tsk" at us.
14:26:35 Nick fxb__ is now known as fxb (
14:27:25B4gder"want are you gonna do? bleed on me?!"
14:27:38B4gderargh, typing!
14:27:42*B4gder hides
14:28:25B4gderyou think we can get 0xdead ? ;-)
14:28:45LloreanAnyone have any objections then, or should I send it off into the great aether?
14:29:12B4gder0xdead/0xbeef is gonna be a great pair
14:29:14gartralthey should thank us for coming forth and trying to do this honestly and right!
14:30:17gevaertsLlorean: go for it!
14:31:22*Llorean will give about five minutes in case someone's not looking at the channel, then if no holds, it's off.
14:32:20 Join jgarvey [0] (
14:32:25*GodEater_ thinks we should go for 0xc0ed/0xbabe
14:33:04*gartral thinks they *may* see past the perverse side-intention there...
14:33:05*Llorean wonders if 0xc0d3 or 0xc0de is taken.
14:33:18LloreanWell, we get *ALL* the PIDs that come with the VID
14:33:49B4gderbut I don't know how (in)complete that is
14:34:09LloreanAt least it doesn't have c0ed, c03d, c0de or c0d3 VIDs.
14:34:18LloreanOr dead for that matter
14:34:36B4gder0xc0ca/0xc01a ?
14:34:37***Saving seen data "./dancer.seen"
14:34:47linuxstbLlorean: I'm changing your letter slightly...
14:34:51*Llorean doesn't want to *invite* suits.
14:34:54Lloreanlinuxstb: Oh?
14:36:11gartralnor d3ad
14:44:46 Part B4gder
14:45:07 Join CaptainKewl [0] (
14:47:06 Join LambdaCalculus37 [0] (i=44a04303@rockbox/staff/LambdaCalculus37)
14:50:36Lloreanlinuxstb: What change?
14:52:33linuxstbLlorean: I'm basicallyl rewriting it in a slightly different way. Feel free to ignore when I'm done though if you prefer your draft...
14:53:34LloreanAh, since you said "slightly" I was just trying to figure out if it was something you'd pastebinned and forgotten to link (maybe busy?) or something. If you're still writing that's fine, sorry.
15:00:25linuxstbi'm not sure about the whole "open source community" bit of the letter. Thinking about it, that may just confuse the issue - how do you define "open source"? Also, there are many devices running open source software, but the hardware vendor assigns the ID. I think Rockbox is quite unique in replacing the software in usb devices.
15:00:59 Quit CaptainKewl ("( :: NoNameScript 4.02 :: )")
15:01:16linuxstbAnd then it raises the question of who is responsible for assigning product IDs for this "vendor" ? I doubt the USB-IF would want that job...
15:01:39LloreanWell, PIDs are always assigned "internally" anyway, right?
15:01:48LloreanSo whoever the "owner" of the VID would be would handle that.
15:02:25linuxstbI know - but who would that owner be? You're suggesting Rockbox becomes that organisation?
15:03:15LloreanWell, the idea that came up in -community was that if it were possible for them to be made widely available, instead of just used for us, maybe HAXX would be interested in purchasing a VID for the purpose of making available to OSS software projects.
15:03:31LloreanThat'd still be up to HAXX of course, but it'd be another possible alternative.
15:04:20linuxstbBut thinking from USB-IF's point of view, I don't think vendor/product IDs relate to software, but to specific pieces of hardware.
15:04:48linuxstbAnd even in Rockbox's case, do we want unofficial builds using our IDs?
15:05:03LloreanIf that's the case, we shouldn't really be applying for a VID anyway, and they'll probably tell us that.
15:05:22LloreanWell, unofficial builds could already be doing whatever they want with IDs now.
15:06:57linuxstbI think that's what we need to ask - my letter (I've got stuck at the end...) basically just tries to explain our situtation, and ask them what the "right thing" is to do. So I'm thinking that maybe we don't want to offer too many solutions, but rather just see what they say.
15:07:15 Part gartral
15:09:33gevaertsIf people feel bored, finding a datasheet for the "AMD 5536 UDC" USB device controller would be helpful
15:11:46linuxstbBut this is what I've written so far -
15:14:20 Part LinusN
15:14:34LloreanA few minor changes (typos, change "plan" possibly to "expect to" instead) but other than that, I definitely like yours better than mine.
15:14:52GodEater_and "offer" to "offers"
15:14:55LambdaCalculus37I think it gets the message across very well.
15:14:59GodEater_me too
15:15:04*GodEater_ votes for linuxstb's
15:15:12*LambdaCalculus37 also votes for linuxstb's
15:20:17 Quit jordoex_ (Read error: 110 (Connection timed out))
15:24:05 Quit dfkt (Read error: 104 (Connection reset by peer))
15:26:05 Join evilnick [0] (i=0c140464@gateway/web/ajax/
15:27:50advcomp2019gevaerts, i think i found some info for that but maybe not but i am not sure
15:28:26 Quit maddler (Remote closed the connection)
15:28:31 Join maddler [0] (
15:29:07 Nick evilnick is now known as evilnick_7 (i=0c140464@gateway/web/ajax/
15:29:24 Join tvelocity [0] (
15:29:34*freqmod_gq sees that the linuxstb letter don't touch the issue of using a rocbox vid for multiple oss projects
15:29:50linuxstbfreqmod_gq: Can you name another project that would need them?
15:30:11freqmod_gqmaybe ipodlinux
15:30:28linuxstbUnlikely IMO - no kernel hacking has happened there for years...
15:31:02linuxstb(afaik - I may be wrong...)
15:31:50linuxstbI'm just sure if we want to confuse the issue by suggesting that, at least not in the first contact.
15:31:56linuxstb^not sure
15:32:28 Join gregorovius [0] (n=diego@
15:34:20LloreanProbably best not to, really. It's something that can come up later, once we know what our options etc are.
15:34:23 Quit midijunkie (Read error: 104 (Connection reset by peer))
15:37:58 Join midijunkie [0] (
15:39:43 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
15:40:06advcomp2019gevaerts, i do not know if this helps but it talks about that i think and it is a porting guide:
15:45:27 Join tyfoo [0] (
15:53:04 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
15:53:56 Quit PaulJam (".")
15:54:30 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
15:55:08 Join Sedgewick [0] (n=Sedgewic@
15:55:20Lloreanlinuxstb: Should I patch up the small typos and go ahead and send it, or would you rather send it being the actual author?
15:55:54linuxstbNo, I'm happy for you to send it.
15:56:57 Join kugel [0] (n=kugel@rockbox/developer/kugel)
15:57:34linuxstbI'm not sure why you don't like the word "plan" though...
15:58:42LloreanWell, I didn't think we really "planned" stuff, and I definitely didn't know we had plans for specific features to be added.
15:58:55LloreanSo I just wanted to change the tone slightly to sorta indicate "we don't expect it's unlikely at all not to eventually see these"
15:58:59LloreanRather than "we're working on them now"
15:59:14LloreanBut I'm more or less indifferent about it. I was gonna leave "plan" in since nobody'd commented on my proposed change to it. :)
15:59:27 Join miepchen^schlaf [0] (
15:59:59linuxstbI just think you can "plan" (i.e. intend) to do things without having any fixed timescales...
16:00:48LloreanI just didn't know we intended to.
16:02:00linuxstbgevaerts has mentioned it, and I think it's inevitable that if he doesn't do it, others will. i.e. it's one of those "yes, we would like it if someone does the work" kind of things.
16:06:32LloreanAlright, I've tweaked some of the wording a little (mostly just because I thought it sounded better), fixed a few typos (offer/offers, etc) and did remove "plans" just because I like the way it sounds better this new way.
16:07:37gevaertsLlorean: do we actually need to mention those older devices?
16:08:07 Quit kachna (Read error: 60 (Operation timed out))
16:08:39 Join woyciesjes [0] (n=8024b83a@gateway/web/cgi-irc/
16:08:50Lloreangevaerts: My original thought with mentioning them was "well there are some devices where we absolutely cannot change the VID/PID" but I don't think it's essential to mention them.
16:09:06gevaertsI think it could actually be confusing
16:09:28linuxstbI would say "In some cases" at the start of the sentence, not the end.
16:10:22LloreanShould we drop the USB<->ATA mention then?
16:10:25 Quit woyciesjes (Client Quit)
16:10:39linuxstbYes, I agree it doesn't add anything.
16:10:51LloreanBoth of those are fixed, then
16:15:26 Join woyciesjes [0] (n=8024b83a@gateway/web/cgi-irc/
16:15:48linuxstbBTW, does anyone know the state of USB on telechips?
16:16:34gevaertsIt "works", but it's not very well tested I think
16:18:22 Join {phoenix} [0] (
16:20:52*Llorean is ready to send the email.
16:20:54LloreanAny last-minute objections?
16:21:08 Join nuonguy [0] (
16:21:30 Quit kugel (Nick collision from services.)
16:21:35 Join kugel [0] (n=kugel@rockbox/developer/kugel)
16:23:16 Join woyciesjes_ [0] (
16:23:56 Quit woyciesjes ("CGI:IRC (EOF)")
16:27:49*Llorean clicks send.
16:34:39***Saving seen data "./dancer.seen"
16:40:02 Join Seed [0] (
16:40:42 Quit SirFunk (Read error: 110 (Connection timed out))
16:42:02 Join z35 [0] (
16:44:44 Join SirFunk [0] (
16:45:41rasherAny idea why the E200R wouldn't connect USB if booting with USB connected? I thought the E200R used the same main Rockbox image as the E200?
16:48:15 Join evilnick [0] (i=0c140464@gateway/web/ajax/
16:48:27 Join Lss [0] (
16:48:38 Nick evilnick is now known as evilnick_6 (i=0c140464@gateway/web/ajax/
16:53:57 Join Xerion [0] (
16:54:38 Join toffe82 [0] (
16:56:12linuxstbrasher: What happens?
16:56:18 Join phill [0] (
16:56:36rasherlinuxstb: The bootloader boots Rockbox, but a connection isn't established
16:56:54rasherSeems to me like the bootloader did its job
16:56:55phillrasher: on my E200 (non-R) USB doesn't connect on boot if Start Screen is set to Resume Playback.
16:57:17rasherAh, that might be the key here - advcomp2019?
16:57:17phillrasher: if set to something else it works fine
16:57:34rasherNice catch
16:59:19rasheradvcomp2019: What is your start screen set to?
16:59:37 Join flydutch [0] (
16:59:38advcomp2019rasher, have to look hold on
16:59:40rashergevaerts: Shouldn't this be fixed?
17:00:01gevaertsrasher: probably. What's the FS number? ;)
17:00:17rasherNone, yet
17:00:22advcomp2019rasher, the now playing screen
17:00:50rasheradvcomp2019: Try changing that, then retrying turning on by inserting usb cable
17:01:01advcomp2019ok will do
17:01:40rasherSame happens for my e280
17:01:52rashergevaerts: Do you want me to file a bug?
17:02:15 Quit Zagor ("Client exiting")
17:02:26gevaertsrasher: I think so, yes. We don't want to forget things
17:02:45rasherWill do
17:03:09advcomp2019rasher, i changed it to main menu and it works fine
17:03:30rasherSo this isn't a bootloader issue - please add that to FS #9955 :)
17:04:23linuxstbrasher: Are your bootloaders the current svn versions, or have you changed the usb detection behaviour? (or did someone change that in svn without me noticing)?
17:04:37 Quit gregorovius ()
17:04:52rasherlinuxstb: SVN always had this behavior when USE_ROCKBOX_USB was enabled
17:05:11rasherSo they're plain SVN
17:05:21linuxstbAh, OK. So enabling that define changed the bootloaders...
17:05:28 Quit kugel (Nick collision from services.)
17:05:33 Join kugel [0] (n=kugel@rockbox/developer/kugel)
17:05:35linuxstbBootloader changes by the back door...
17:06:43gevaertsThat's been there since r16279
17:06:54*gevaerts blames Zagor ;)
17:07:06advcomp2019rasher, i added that.. i hope it sounds good ;)
17:07:16evilnick_6How can I change the bootloader on my Sansa E280 to test the new ones?
17:07:53linuxstb"sansapatcher -a bootloader-xxx.sansa"
17:08:15evilnick_6linuxstb: Does that need admin rights on Windows?
17:08:24linuxstbrasher: New sansapatcher/ipodpatcher builds will be needed to...
17:08:30linuxstbevilnick_6: Yes, I think so.
17:08:49rasherlinuxstb: aw crud, there are bootloader-*.sansa files? Didn't include those :\
17:09:04LloreanAren't they .mi4?
17:09:08rasherlinuxstb: Yeah, let's hold off on that until we've tested the bootloaders though
17:09:10LloreanOr did we change that at some point?
17:10:07linuxstbOops, yes, they are probably .mi4
17:10:10Lloreangevaerts: If we presented multiple USB modes, do you think that'd best be done directly through the core (settings somewhere) or possibly by way of some Advanced USB plugin that let you configure behaviour that replaced the standard USB when the plugin was running?
17:10:20*linuxstb is confused with mkamsboot
17:11:05linuxstbSo "sansapatcher -a pp5022.mi4" or whatever they are called.
17:11:10rashergevaerts: FS #9957
17:13:25rasherDo we release Sansa/ipodpatcher with the release, or just once we're ready?
17:13:37gevaertsLlorean: I expect that we'll want plugins once we have a lot of class drivers
17:14:29rasheradvcomp2019: So the E200R bootloader works?
17:14:59advcomp2019rasher, yep
17:15:11linuxstbrasher: I think we can release at any time - they're independent of specific releases. It would be useful to have them no later than the release though.
17:15:35linuxstbOr maybe it would confuse things to have them before 3.2...
17:16:08rasherThat's what I was thinking
17:16:16linuxstbYes, I've just caught up with you.
17:16:33linuxstbBut has anything changed with ipodpatcher or the ipod bootloaders?
17:16:54rasherPossibly not. I wasn't aware they didn't have USB detection
17:17:01rasherHow about the H10?
17:17:03linuxstbSo maybe we can just forget those.
17:17:28linuxstbI'm not sure...
17:18:40rasherIt seems to use the main-pp.c bootloader - that would put it alongside the e200, right?
17:18:44 Part domonoky
17:18:55LloreanI'd rather have the bootloader "release" before even the feature freeze for 3.2, that way all people testing the 3.2 build can be doing it on current bootloaders.
17:19:15linuxstbLlorean: But then there's the problem of new installs of 3.1
17:19:19LloreanI don't think it'd be *too* confusing to have them released early as long as we say "In anticipation of the 3.2 firmware release, we've released a new round of bootloaders for these targets: "
17:19:32Lloreanlinuxstb: Would they harm 3.1 installs? They shouldn't.
17:19:40rasherLet's do it at the time of the freeze, then?
17:20:04rasherIt'd just take away "insert usb cable to turn on and connect" functionality from 3.1 installs
17:20:36rasherNot the end of the world
17:21:07linuxstbNo, but the other way around causes less damage - i.e. releasing 3.2 first, then the new bootloaders.
17:21:15Lloreanrasher: That's only one a couple players anyway, the Sansas I think
17:21:52linuxstbI just think that we need to try and keep the stable builds stable... I agree it's not a big deal though.
17:22:19rasherI'm increasingly of the opinion that we shouldn't have rbutil/*patcher releases other than at release time
17:22:50rasherBut have an area for pre-release versions of them, if development makes it necessary
17:23:07linuxstbWe could even change the version numbers, so everything has the version number of the main Rockbox build.
17:23:37 Quit kugel (Nick collision from services.)
17:23:41 Join kugel [0] (n=kugel@rockbox/developer/kugel)
17:25:20n1srasher: i think the tracker is fine as a place to keep test versions
17:26:10 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
17:26:15rasherWe need bootloader testers for Iriver H10 (20GB), M:Robe 100 and C200 - FS #9955
17:27:06 Join MethoS- [0] (
17:27:13Lloreanlinuxstb: I'd like it very much if everything was version numbered in sync with the main build even if it meant rebuilding old bootloaders with new version numbers each time just so there was a "synced" version.
17:27:31 Quit avis (Remote closed the connection)
17:28:35n1sLlorean: i don't think releasing new bootloaders with every release if noting has changed is good
17:28:41linuxstbMaybe we don't need to change the version numbers of the bootloader itself, just sansapatcher/ipodpatcher/rbutil. Doing a full range of bootloader testing every 3 months could get tiresome.
17:33:45LloreanWhat we need, at least, is a single page people can go to and see if their bootloader is up to date. A list of current bootloader versions (especially for players where rbutil can't check)
17:33:47rasherRebuilding the patchers seems good enough
17:38:17 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
17:39:03kugelso, I have this sorting patch done in Llorean way now
17:39:10 Quit gevaerts (Nick collision from services.)
17:39:19 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
17:39:37kugellooks a bit weird when you choose between "1 2 03 10" and "03 1 10 2", but it works
17:46:01rasherI think there should be some text to make clear that those are example sortings
17:49:06 Join SirFunk_ [0] (
17:49:14 Join avis [0] (n=ident@pdpc/supporter/student/avis)
17:50:24kugelI think it's obvious that these are examples.
17:50:43kugeland if it confuses people, there's always the manual
17:52:12kugelthe menu is called "sort leading numbers", and the options are a series of numbers; really, that's pretty obvious to me
17:54:32 Join MethoS-- [0] (
17:55:43*kugel likes to commit . objections?
17:57:36 Quit sarixe ("Leaving")
17:59:38 Quit MethoS- (Read error: 110 (Connection timed out))
18:01:21 Join SirFunk__ [0] (
18:03:02 Quit Sedgewick ("off")
18:03:28 Quit SirFunk (Connection timed out)
18:14:35 Quit SirFunk_ (Read error: 110 (Connection timed out))
18:14:57linuxstbkugel: Doesn't chessbox stop playback on Archos devices?
18:15:57*linuxstb wonders why kugel asked for comments, then committed anyway...
18:16:40kadobanoh, right...should probably not include that menu on archos for chessbox
18:25:24linuxstbkugel: Regarding the natural sorting patch, the original comment says "it is not limited to leading numbers in a string", but you've called the setting "Sort leading numbers" ?
18:25:38 Quit kugel (Nick collision from services.)
18:25:43 Join kugel [0] (n=kugel@rockbox/developer/kugel)
18:26:08kugellinuxstb: I asked if there are objections, not for comments ;)
18:26:48*linuxstb wonders why kugel asked for objects, but didn't give anyone chance to make any
18:26:57linuxstbBah, I can't type...
18:27:12kugelI waited some minutes, in that time I tested the patch
18:27:34kugelbut it should've been clear to me that archos messes it all up again
18:28:25kadobani have a patch for that, not the prettiest thing, but should fix it. let me put it up somewhere
18:28:42linuxstbkugel: Did you see my other comment, just before you quit and rejoined?
18:29:58kugelyes, now
18:31:18kugelit's the most obvious feature. And it only sorts different to ascii, when there leading numbers, or a "constant" prefix before the numbers (so numbers are still involved)
18:31:32kugeland to make the menu name not kilometers long
18:32:20kugelbut if you have a better idea, I'm still open to suggestions
18:32:28 Join kubiix [0] (n=511bc91c@gateway/web/cgi-irc/
18:33:14kugelis there a way to determine whether a plugin stops playback?
18:33:26kadobani added a patch to FS #9953 to take care of chessbox on archos
18:33:30LloreanI thought at one point it was proposed to affect trailing numbers too, though I never tested to find out if it did.
18:34:05 Quit petur ("work->home")
18:34:08kadobankugel: i think the only way is to look for audio buffer steals or being an overlay
18:34:15kugelyea, if the leading characters are the same (="constant" prefix)
18:34:34linuxstbLlorean: This is the page describing the function we're using -
18:34:42***Saving seen data "./dancer.seen"
18:34:54kugelkadoban: I see a there
18:35:22linuxstbWhich also gives a nice example -
18:35:24Lloreankugel: "The Colour of Magic - 1" isn't something people are going to think of as "a filename where 'The Colour of Magic' is the prefix"
18:36:45*Llorean notes that pic02a comes between pic2 and pic3, suggesting that it's not just prefix either.
18:37:00kadobanfrik...that patch i added doesn't work probably
18:37:04LloreanIf we're using that algorithm, we could just say "Sort numbers" for it.
18:37:10rasherLlorean: it just sees numbers as a unit
18:37:22rasherAs far as I can tell
18:37:32kugelSo just "Sort numbers", I'm fine with it
18:37:40Lloreanrasher: As long as ours matches that example-out.txt it should be "Sort Numbers" then.
18:38:10kugelI think the name of the menu should be as short as possible, further explanations are for the manual
18:38:11rasherYeah, I think that's a far better name
18:38:14rasheror "Number sorting"
18:38:19LloreanOr that
18:38:20LloreanEither one.
18:39:10*rasher slightly favours "Number sorting", since "Sort numbers" sounds like an action, not a method of doing something. To me, non-native speaker, anyway
18:39:37Llorean"Number Sorting" sounds a bit nicer, I think too.
18:39:54Lloreankugel: So it's a modifier now, rather than its own sort method?
18:39:56*rasher mumbles something about title casing
18:40:15woyciesjes_I'd vote for "Number Sorting" too.
18:40:41Lloreanrasher: I thought our menus were Title case.
18:40:43*gevaerts plays with FS #9708
18:40:52rasherLlorean: Yeah, they are. For now!
18:40:54 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
18:41:13Lloreanrasher: Well, it's pretty common in English.
18:41:18kugellinuxstb: Ok, I'm going to use the #ifdef that's in plugin/SOURCES (#if (MEMORYSIZE <= 8) && !defined(SIMULATOR))
18:41:20LloreanEven UK English, if one can judge by say,
18:41:27kugelLlorean: yes
18:41:37 Quit jgarvey ("Leaving")
18:42:21pondlifeReally silly question - should splash(HZ, ...) pause for a second or until a keypress before moving on?
18:42:53kadobankugel: hopefully correct patch now up at the same FS if it helps
18:43:40pondlifeMy patch on FS #8894 has an unexpected bit of screen update where I use a splash in the menu callback to prompt for a reboot. The menu returns first, then draws the splash for some reason..
18:43:51pondlifeJdGordon: You're Mr Menu callback, no?
18:44:12 Join obo [0] (
18:44:14kugelkadoban: seems good, almost the same which I did
18:44:25kugelI forgot to #ifdef the #include though
18:45:41kugelbut this isn't only true for archos, right?
18:46:29kadobani believe it's only archos
18:46:47kugelrasher: Sort XYZ is consistent with the rest of the menu though
18:47:05kadobanclip maybe? but that's not set up yet
18:47:51kugeliirc this should be true for the entire AMS sansas, and I believe there's a iriver with 8MB ram too
18:47:57*pondlife would be happy to see the end of Incorrect Title Case...
18:48:22 Quit phill ()
18:48:24pondlife"Set as Right Quickscreen Item".. :/
18:48:32rasherkugel: not sure that's important
18:48:56kadobankugel: it's using the same logic as the overlay building, so if it's flagging incorrect targets then i think it's already broken. or am i missing something?
18:49:29kugelrasher: Consistency matters more to me than "I slightly like this more"
18:49:54kugelkadoban: depends on how much ram chessbox really needs
18:49:56rasherBut why does it need to be consistent with other options that aren't very similar?
18:50:11kugelit appears to stop playback on all <= 8MB, which may be correct
18:50:25*rasher would rather rename "Sort Case Sensitive" to "Case sensitive sort(ing)"
18:50:35kugelbecause it looks ugly, and something that looks ugly can't be good for the user experience
18:50:48kugelthat could be done, yes
18:50:50linuxstbLlorean: Your BBC example is the perfect reason why attempting to use title case is a bad idea - some links on use it, but many don't... e.g. at the top - "Display Options" and "Take the tour", "Add more to this page"...
18:51:37*rasher doesn't understand how it would look ugly at all
18:51:58kugelpondlife: I'd have a look at the other options which require reboot (cue sheet support, database, disk cache)
18:52:16pondlifekugel: I did - I stole the code from dircache, but it behaves differently.
18:52:26pondlifeProbably something really stupid I've done.
18:52:51gevaertsamiconn: I added some notes to FS #9708 about my tests with the ATA dma code. It looks like there's a lot of room for improvement
18:53:24Lloreanpondlife: "Incorrect" by what definition? Setting names are titles.
18:53:39*gevaerts managed to get 9.6 MB/s over rockbox USB with a real device (i.e. not a ramdisk)
18:53:53pondlifeTitles are proper nouns, surely. Quickscreen maybe, but never verbs..
18:54:14pondlifeI see very little Title Case on the BBC News site.. only in titles and "News Feeds" (a mistake, I guess)
18:54:16LloreanA title can include verbs, can't it?
18:54:58pondlifeI think this is more a UK/US English thing.
18:55:08Llorean"Display Options" "Accessibility Help" "Online Courses" "Multifaith Calender"
18:55:23LloreanOddly, the titles of articles aren't capitalized, but the titles of all sorts of options/pages you can go to are.
18:55:27rashergevaerts: Aw, still not on par with the OF :\
18:55:41linuxstbApparently, there's an ISO standard for this kind of thing...
18:55:51Llorean"Advertise With Us" "Contact Us" "Reception Advice"
18:55:53pondlife has "News feeds" and "News Feeds"
18:56:15kugelLlorean: on, if you click on news, every news is sentence case
18:56:24Lloreankugel: We're not discussing news articles.
18:56:28LloreanWe're discussing menu options.
18:56:31LloreanSettings, etc.
18:56:38Lloreanlinuxstb: That's for form labels.
18:56:39pondlifeSorry, I thought you were using as a ref..
18:56:55kugelI tend to assume the website was done by US (title case), while the content (which is sentence case) was done by UKs
18:56:57rasher> One crucial issue in legibility (possibly the most crucial) is the question of familiarity. The most familiar presentation of sentences is in sentence case. Most reading is reading of sentences. Therefore, sentence case is easiest to read.
18:57:10linuxstbLlorean: Yes, but still relevant IMO...
18:57:32Lloreankugel: That's a silly assumption to make.
18:57:37pondlifeSome of them...!
18:57:44LloreanYou have no evidence of where the website was made other than it uses title case.
18:58:18pondlifeAll I know is that Title Case Looks Odd to me.
18:58:39LloreanIt looks odd when you're providing sentences, sure.
18:58:44LloreanWe don't put periods on the ends of settings names.
18:58:49LloreanThey aren't prose.
18:58:55rasherI really think this lady hit home when she said "Most reading is of sentences. Therefore, sentence case is easiest to read."
18:59:00rasherEven if it's just a fragment
18:59:35pondlifeI just want consistency, so we can go for all Title Case if you prefer
18:59:50kugelthat's my opinion too
18:59:58LloreanI think "Sound Settings" looks MUCH better than "Sound settings"
18:59:59pondlifeBut should it then not be "Set As Right Quickscreen Item"
19:00:03linuxstbpondlife: That's my point - some strings in english.lang are definitely wrong in title case - things like "Are you sure?"
19:00:13Lloreanpondlife: Title Case doesn't mean "Capitalize every word"
19:00:29linuxstbLlorean: And that's the other problem - some strings use "Of", some use "of", etc etc...
19:00:31pondlifeI know, but I don't understand why pick on Item and not As, for example
19:00:35Lloreanlinuxstb: That's not an item in a menu list...
19:00:58linuxstbLlorean: I know. But you try going through the strings in english.lang and maintain consistent rules of that sort...
19:00:58Lloreanlinuxstb: It's very definitely not a title, so I don't see why title case is even a question for "Are you sure?"
19:01:17 Join R3d2Dawn [0] (n=R3d2Dawn@
19:01:21linuxstbWhat about lists of options for settings? They are not setting titles.
19:01:25LloreanSo the argument is "we should use sentence case, because although we have both sentences and titles, people shouldn't be expected to sort it out"?
19:01:31linuxstbe.g. "Mono left", "Mono right".
19:01:40Lloreanlinuxstb: And I don't think those should be title case.
19:01:52pondlifeI just don't know enough about title case, I guess.
19:01:53R3d2DawnHow do I get Coverflow?
19:02:11pondlifeLlorean: In "Set as Right Quickscreen Item" should Item be a capital?
19:02:23rasherLlorean: My argument is "because it's silly to capitalise titles in the first place"
19:02:27pondlife(have a capital I, I meant)
19:02:31rasherDon't know about everyone else.
19:02:39kugelWe agreed on using British English in Rockbox, so the decision on title vs sentence case is up to the British , imo
19:02:49linuxstbLlorean: My argument is that english.lang has _never_ been consistent, and looks a mess. Lots of people (mainly British people) prefer sentence case, others prefer title case, but switching to sentence case would solve the consistency issue forever.
19:02:50Lloreanpondlife: Either "right Quickscreen item" or "Right Quickscreen Item" depending on whether the right quickscreen item is a title, or not.
19:02:57kugelor, the native British English speakers, rather
19:03:20pondlifeOK, so "Set as Right Quickscreen Item" would be correct...
19:03:34LloreanI think probably so
19:03:48linuxstbAnd on the BBC website, "Make this my homepage"...
19:04:14Lloreanlinuxstb: That's a sentence. Subject, predicate.
19:04:42linuxstbAnd "Set as right quickscreen item" is different?
19:04:46 Part pyro_maniac ("Leaving.")
19:04:51Lloreankugel: the problem is mainly that there's not absolutely clear rules for when to use title case (or what, exactly, should be capitalized when using it)
19:04:54*kugel saw many news titles that appeared to be a sentence and had title case
19:04:58Lloreanlinuxstb: Look at what I said in regard to it.
19:05:14linuxstbLlorean: I did, but I don't understand the difference.
19:05:19Lloreanlinuxstb: If "Right Quickscreen Item" is a title, then you capitalize all the words in it. "as" is lowercase, because it's a sentence. Set is uppercase because it's the first word.
19:05:33LloreanIt's like saying "Set as Dave Chapman"
19:05:44LloreanLooks like title case for the whole thing, but it's sentence case with a title case object in it.
19:05:44linuxstbOr Make this My Homepage
19:05:49 Nick fxb is now known as fxb__ (
19:05:54linuxstbIn browsers, you have the text "My Homepage"
19:05:58LloreanIs "My Homepage" a title?
19:06:04 Join gregzx [0] (
19:06:16gevaertsrasher: no, but that's probably at least partly due to this 16k read and write size the UMS code uses with the disk
19:06:16LloreanI have a homepage, but I don't have a Homepage.
19:07:14LloreanI mean, we could go entirely sentence case, but I think it'll end up making our software generally look a lot less 'professional' than trying to maintain the mix we have now.
19:07:15pondlifeSeems title case isn't used much...,, were the first 3 sites I checked.
19:07:35pondlifeCapitals for proper nouns only,.
19:07:36linuxstbWhere do you have "Right Quickscreen Item" ? Proper nouns are capitised...
19:08:11Lloreanpondlife: "Terms and Conditions", "Personal Finance" and others on the telegraph one.
19:08:26kugelthe software looks less profressional with sentence case?
19:08:45pondlifekugel: I think it looks less professional with title case
19:08:47Llorean"Terms and Conditions" and "Privacy Policy" seem to be globally capitalized for osme reason.
19:08:53kubiixhi, I am testing the USB transfer from rockbox on Ipod Color and it is much slower than transfer in OF
19:09:00rasherkubiix: We know.
19:09:02Lloreanpondlife, kugel: How much software in English do you see in sentence case, then?
19:09:24rasherkubiix: Well, define "much"? It's about 30% of the speed, iirc
19:09:24kugelWe look more profession if we follow what other software does?
19:09:24kubiixan is is same on 5.5G ?
19:09:25linuxstbLlorean: That doesn't make it unprofessional.
19:09:37kugelI can'T follow that reasoning
19:09:45pondlifeMicrosoft use title case too much - it makes me cringe a little
19:09:54pondlife"My Computer" indeed ;)
19:10:12rasherMaybe it's time for english-us.lang?
19:10:14Lloreankugel: We look more professional if we use the language in the generally accepted way, rather than making up our own rules just to satisfy a few loud people.
19:10:29linuxstbLlorean: Like "Ipod", "Iriver", ...
19:10:51Lloreanlinuxstb: Personally, I think that's pretty stupid too. Those are 'terms' and should be treated as such.
19:11:34woyciesjes_Well, according to The Steve and Apple, it's "iPod" :)
19:11:36kugela) We replace firmwares on highly embedded devices, that's professional enough, b) I don't care how professional it looks, I prefer what looks less odd to my eyes, c) Maximum consistency always looks more prefessional than some mixed stuff (which looks more like a loose bunch)
19:11:40pondliferasher: Yes, I think so.. :)
19:11:44rasherSeriously, why not create a US English version that uses US spellings and Title Case?
19:11:57*linuxstb is in favour (and favor) of that
19:12:31rasherIt shouldn't be a particularly hard change
19:13:07pondlifeI do think it's a UK/US thing....
19:13:20Lloreankugel: It's not "just mixed." How on Earth is "using it the wrong way but we do it everywhere" more professional looking than following the mix that means using each case where appropriate?
19:13:30LloreanShould we ignore title case even when things are _definitely_ titles, just for consistency Kugel?
19:13:44linuxstbLlorean: The point is that titles do not have to be written in title case.
19:13:52BigBambiLlorean: I don't think you can look at software in general - the vast mojority of software forces US English on us.
19:13:53linuxstbThat's simply a style issue.
19:13:58Lloreanpondlife: Still, you haven't shown me a single instance where someone avoids title case _entirely_ which is what you guys have proposed.
19:14:13pondlifeI could always tell that The Onion was a parody of US news sites just by looking at the layout - probably use of Title Case is why :)
19:14:27pondlifeLlorean: It's never going to be 100% with human input..
19:14:42 Quit kubiix ("CGI:IRC (Ping timeout)")
19:14:56pondlifeI look at Misplaced Title Case and see an error. That's what was schooled into me.
19:15:13LloreanBut if it's a title, how is it misplaced?
19:15:19rasherLlorean: It's a style issue. I don't see why it's hard to get?
19:15:20linuxstbThis is one UK government site where I know their policy is for sentence case -
19:15:29rasherNot a question of Wrong or Right
19:15:34 Join jaykay [0] (
19:15:39pondlifeCapitaliation is for proper nouns and the start of a sentence.
19:15:45linuxstbAlthough even they don't get it right...
19:15:48Lloreanrasher: Well, there's still the question of the language that's built in.
19:15:49pondlifeTitle is neither here nor there
19:16:13rasherLlorean: What?
19:16:15Lloreanpondlife: So you're saying Title case isn't valid, period? You think it should be "Harry Potter and the half blood prince" on the book?
19:16:20R3d2DawnHow do I get CoverFlow on the Ipod
19:16:29BigBambiR3d2Dawn: You don't
19:16:29pondlifeHalf-blood Prince
19:16:33Lloreanrasher: Do we build in the one that looks like most English software, or the UK English one?
19:16:46Lloreanpondlife: But surely that's not a proper noun. He's a prince, and he's half-blooded.
19:16:47rasherLlorean: Most US English software, you mean
19:16:51BigBambiR3d2Dawn: Rockbox has a visualisation called Pictureflow, that is detailed in the manual
19:16:56pondlifeThat's a proper noun, it refers to a specific person (even if not by name)
19:17:01Lloreanrasher: I haven't yet seen an example of UK English software brought up before me.
19:17:05rasherLlorean: And we use UK English, just as we do everywhere
19:17:20Lloreanpondlife: That doesn't make it a proper noun. Otherwise "That Guy Who Does the Dishes" or such.
19:17:21BigBambiLlorean: I don't see why we should put up with having US English forced on us, but you can never have UK English?
19:17:28rasherLlorean: Because everyone uses US English in software for some gosh darn reason
19:17:48R3d2DawnI watched a video via utube that showed someone with rockbox on there ipod and coverflow well it was called something else though
19:17:56pondlifeLlorean: That Guy Who Does The Dishes (or similar) would be correct
19:17:58kugelwhen we choosed UK english in Rockbox, saying that "Other software forces us to US english" isn't valid
19:18:01BigBambiR3d2Dawn: Yes, Pictureflow as I said
19:18:03LloreanBigBambi, rasher: My point is, we go through an effort to try to present to the general public what they _expect_ in cases where it doesn't hurt functionality. What they expect is title case, because that's what other software does.
19:18:19Lloreankugel: We chose UK _spellings_
19:18:22rasherLlorean: So we should switch to US english?
19:18:31rasherUsers _expect_ to see "color"
19:18:46pondlifeSeriously, just make an American.lang
19:18:49rasherRight now, english.lang is a UK/US mutant
19:18:50BigBambirasher: Even though I dislike it
19:18:53linuxstbpondlife: "American" ? ;)
19:18:54R3d2Dawnit was called pictureflow
19:19:06BigBambiR3d2Dawn: Yes, I have told you that *twice*
19:19:12R3d2DawnOk thanks
19:19:14BigBambiR3d2Dawn: It is also in the manual
19:19:39kugelLlorean: So UK spelling and US English?
19:19:45kugeleven worse
19:19:50Lloreanrasher: Spelling is a language thing. Capitalization is an aesthetic thing.
19:19:58 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:20:00pondlifeIt's a cultural thing
19:20:10Lloreankugel: As we've established, there's no clear grammatical rule on the capitalization issue. Otherwise we wouldn't be having this issue.
19:20:16rasherLlorean: It's as much part of the language as the spelling
19:20:24Lloreanrasher: Then find me a rule regarding it.
19:20:35rasherWho said it was a rule?
19:20:53LloreanI thought it was part of the language.
19:20:57rasherIT IS
19:21:02 Join robin0800 [0] (
19:21:02kugelI would've been slapped by my German teacher if I didn't capitalise nounds, it *is* part of the language
19:21:03rasherdoesn't mean it's a rule set in stone
19:21:06BigBambiI'm not actually too bothered as long as the spelling is correct
19:21:17Lloreanpondlife: The page you linked to supports title case.
19:21:22linuxstbI started this discussion last year, because someone changed a large number of strings from sentence case to title case in english.lang, without asking if that was wanted...
19:21:29kugelby my English teacher too
19:21:41linuxstbAnd people still add new sentence case strings, meaning lots of people want them...
19:21:44pondlife"Use capital letters at the start of only the first and important words in a title"
19:22:00*linuxstb learnt that at school...
19:22:04pondlifeI think the point is whether a menu option is a title. I'd say not.
19:22:12Lloreanpondlife: The page you linked says "use them for the start of the first and key words, and not for small words within the title. "
19:22:13pondlifeA title is of a book or a film etc.
19:22:29Llorean"Playback Settings" is not the title of that set of menus?
19:22:49*Llorean thinks anything that will show at the top of the list on the next screen is a "title"
19:23:03LearBut also the name of a menu option.
19:23:31LloreanLear: Titles exist to be the name of *something*
19:24:21rasherLlorean: You still haven't given any reason not to create an english-us.lang
19:24:24LearI just meant that the same string in the lang file can be used in different places, like a menu title and a menu option.
19:24:35pondlifeI don't think you can argue this one either way logically. It's a "what the user expects" thing which (probably) maps to UK/US.
19:24:39oboUK software, with sentence case:
19:25:06 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
19:25:07linuxstbobo: Thank God for that...
19:25:24 Part R3d2Dawn ("Leaving.")
19:25:24kugelobo: haha, nice find
19:25:27Lloreanrasher: I never said we shouldn't.
19:25:35rasherSo why continue this discussion?
19:25:38LloreanLear: But as a menu option, it's still the title.
19:25:42rasherIt's a rather obvious solution...
19:25:56Lloreanrasher: Except that then the question comes up of default... as I said last time you asked...
19:26:11rasherLet people who expect US English pick english-us, and let Rockbox continue being UK English
19:26:28Lloreanobo: I don't see any menu titles there...
19:26:31rasherWe've *already* decided on UK English
19:26:31kugelso, commit , and the first one who bothers enough makes a english-us.lang or Title-Case.lang :)
19:26:55rasherSomeone commit title-case.lang and I'll cut him
19:27:16Lloreanobo: Or at least none with more than one word.
19:27:27 Join pixelma2 [0] (i=d5494c04@rockbox/staff/pixelma)
19:27:44rasherLlorean: Don't tell me "what a user expects" (what other programs use) isn't "Set Work Directory"
19:28:06pondliferasher: That's Microsoft talk, that is ;)
19:28:08Lloreanrasher: I don't understand that sentence at all.
19:28:33rasherLlorean: I'm saying that if this program was using US English and title case, that menu would've been "Set Work Directory"
19:28:34LloreanI'd expect "Set work directory"
19:28:49LloreanIt's not a menu title, it's a command.
19:28:50rasherThen you wouldn't get what you expected
19:28:59 Join fml [0] (n=4fd3c969@gateway/web/cgi-irc/
19:29:22rasherYou're arguing we should do what other Title Case Users do - they would put "Set Work Directory"
19:29:29fmlHello. Does the "release/current build" option in RBUtil apply to both the RB and the bootloader?
19:29:59Lloreanrasher: I'm arguing that we shouldn't remove title case _for titles_. On argument in favour of that is that generally people expect to see title case in software.
19:30:24Lloreanrasher: But the one point in favour isn't what my goal is, it's just an argument in favour to weigh against "we shouldn't do it because I think it looks dumb"
19:30:41bombardierhow about making two releases with different cases :)
19:31:01rasherEither you want us to do what's "familiar" or you don't. If you want to use that argument, there's no sense in not using Title Case like everyone else
19:31:11Lloreanrasher: That's a false argument.
19:31:15fmlkugel: were you speaking to me?
19:31:23Lloreanrasher: You're telling me I'm not allowed to be in favour of using title case for titles, unless I want to use it for everything?
19:32:04rasherLlorean: no, I'm saying using the "familiarity" argument falls down if you're not actually going to do what everyone else does
19:32:20Lloreanrasher: familiarity is not a binary situation. You can be more familiar without doing *everything* the other guy does.
19:32:47rasherThen you'd be sitting between two chairs, which is the *worst* for user-familiarity
19:32:55LloreanThat's silly.
19:32:57LloreanIt's definitely not in this case.
19:32:59fmlkugel: how can I install the "USB enabled" bootloader then? (Could you please prepend the nick to (at least very general and hard-to-relate) phrases?)
19:33:21pondlifeThe logical argument really comes down to whether "Set as Right Quickscreen Item" is a title or not.
19:33:23rasher"Oh, this is clearly not Title Case as I'm used " or "This is all Title Case as I'm used to" are both better than "This is Title Case some places, but Why Not Everywhere?"
19:33:25LloreanIf people are used to seeing a lot of title case, seeing a little less is still more familiar than seeing none.
19:33:30pondlifeI say not, Llorean says it is.
19:33:31rasherLlorean: false
19:33:40LloreanSaying "seeing a little less is going to be more alien than seeing none at all" is hard to back up.
19:33:59Lloreanpondlife: That's not a title. That's a sentence, in sentence case.
19:34:23pondlifeSo it should be "Set as right quickscreen item"
19:34:36rasherI'm not saying "more alien", I'm saying more confusing.
19:34:38pondlifeNo proper nouns there (well, perhaps Quickscreen... (tm))
19:34:50rasherStop putting words in my mouth
19:34:53Lloreanpondlife: I think it's already *in* sentence case, but it's a sentence with a title in it, that title being "Right Quickscreen Item"
19:34:59LloreanThe question is whether "Right Quickscreen Item" is a title.
19:35:27kugelfml: grab sansapatcher, or compile your own
19:35:30Lloreanrasher: "more alien" and "more confusing" are synonymous unless you think people can be confused by familiar things. It has to be less familiar to be more confusing, and familiar is the opposite of alien...
19:36:01LloreanSorry, not "synonymous" but "equivalent"
19:36:03 Join Zoxc [0] (
19:36:21kugelfml: if you use a pre-compiled sansapatcher, make sure that you don't install the build-in bootloader
19:36:29rasherLlorean: I explained already.
19:36:42LloreanUsing title case for the names of menu titles, and sentence case for individual options in lists and actual sentences isn't realistically going to confuse anyone.
19:37:13pondlifeIt's a matter of appearing consistent, not confusion.
19:37:16Lloreanrasher: You didn't explain. You stated a claim with no explanation as to why that claim might be true.
19:37:23rasherBut but consistently using UK English and associated rules are?
19:37:26linuxstbLlorean: So what when the same string is used in both places?
19:37:28kugelor we just use sentence case for everything, and don't have a tiresome discussion on every menu item on whether it's a title or not
19:37:31Lloreanpondlife: And there's no consistency in "items used as actual menu titles are titles, others are not"?
19:37:35rasherLlorean: if you can't follow logic, that's not my problem
19:37:55Lloreanrasher: There's no LOGIC to what you said. You didn't _explain_. You said "the sky is blue" and didn't follow up with any evidence that it is.
19:37:58kugelif title case implies that we have to decide on title or sentence for every single string, that, no thanks
19:38:02LloreanYour opinion on its own isn't "logic" by magic.
19:38:19Lloreanlinuxstb: Do you have examples of items that are used twice?
19:38:19rasherLlorean: Seriously, why haven't you created english-us.lang yet?
19:38:20pondlifeLlorean: So, title case if it has sub-options (and thus could appear as a heading), or sentence case if you're at the bottom of the tree?
19:38:26fmlkugel: ok, will check. Another question (for the others as well): is there a wiki page explaining the differences between the USB charging and USB connection for data transfer? I think I've read (here in IRC) this is not the same but what is active when?
19:38:39 Join webguest84 [0] (n=a87a02e3@gateway/web/cgi-irc/
19:38:43Lloreanrasher: If you ask me that again, I'm just leaving. I've explained that this is about defaults now twice, I'm getting tired of you circling back to pointless things like that.
19:38:54Lloreanpondlife: basically, yes. Title case *for titles*
19:38:54rasherIt's not pointless - it's a SOLUTION
19:39:06Lloreanrasher: it doesn't solve the default issue... at all.
19:39:16rasherLlorean: that's not an issue at all
19:39:21rasherRockbox is UK English.
19:39:34rasherAnyone who wants another language should pick that
19:39:49kugelfml: huh, the difference between charging and connected as ums drive is pretty clear
19:39:52Lloreanrasher: And you've yet to show me that this is an issue of language, rather than aesthetic. If UK people like blue text more than white text, do we change the color too?
19:40:14rasherIf this was an issue, I don't see how we haven't had a huge shitstorm over using UK spellings
19:40:30rasherLlorean: UK English uses sentence case. US English uses Title Case. Why is this hard?
19:40:42Lloreanrasher: UK English also uses Title Case, they just use it in less cases than US English.
19:40:59rasherLlorean: Irrelevant to this discussion though.
19:41:04LloreanNo, it's not.
19:41:17rasherIt damn well is, because UK English wouldn't use title case in Rockbox.
19:41:23LloreanSince this discussion is entirely about the question "do we have things in Rockbox that qualify as titles for the purpose of casing?"
19:41:26amiconngevaerts: Why would the 16K blocksize be the limiting factor on the Video?
19:41:51Lloreanrasher: Based on your own opinion of how the language works, or what?
19:41:58pondlifeUK English users use title case incorrectly sometimes, just like they use the Oxford comma sometimes. It's human to make mistakes.
19:42:01rasherAnd pixies and rainbows.
19:42:09fmlkugel: I mean, when I plug in the USB, does it charge only? Or can I trasfer data as well? Or only transfer? If the latter is the case, what should I do to charge? I think there's been a discussion about how to let the user to decide what should happen on USB plug.
19:42:20kugelit connects
19:42:31kugelyou hold select upon inserting if you only want to charge
19:42:50kugelfml: yes, there was a (very short) discussion
19:43:03kugelletting the user decide is considered as configurable buttons, and thus rejected
19:43:07 Join Thundercloud [0] (
19:43:12Lloreanpondlife: So show me some book titles, and other actual titles of things that use sentence case?
19:43:22LloreanTitles of _things_ as we're discussing.
19:43:36rasherSetting title != Book title
19:43:47Lloreanrasher: They're both titles.
19:43:52oboSo menu items are automatically titles, and not sentences?
19:43:55pondlifeNo, the question is not which case to use for titles, it is "what is a title?"
19:44:04*pondlife types too slowly
19:44:11fmlkugel: aha! So if I hold SELECT then it only charges but RB continues to run (not entering the USB mode)? And if I plug it in without holding anything it enters the USB mode so that I can transfer data but at the same time the battery is charged?
19:44:15Lloreanpondlife: Well, you haven't explained to me why titles at the tops of lists aren't titles, then.
19:44:17rasherLlorean: but different "levels" of titles
19:44:24kugelfml: exactly
19:44:25pondlifeHeading != title
19:44:25Lloreanrasher: Ah, so it's down to interpretation and opinion again?
19:44:36rasherNot really.
19:45:04Lloreanpondlife: According to, that's incorrect.
19:45:14Lloreanrasher: Well, then, what exactly is the level at which the line is drawn?
19:45:15fmlkugel: ok, now I undertand!
19:45:29Lloreanrasher: If it's not up to opinion, I assume you can provide a reference?
19:46:16 Quit pixelma2 (" ajax IRC Client")
19:47:08fmlLlorean, rasher, pondlife: my 2 cents: in an English-"speaking" app, I'd expect all menu entries in Title Case. I'd also expect them in this form (i.e. Title Case) if they are mentioned in a running text, e.g. the manual.
19:47:41pondlifefml: Absolutely, the manual etc. should be consistent with whatever we have.
19:48:12*Llorean goes to lunch, and eagerly awaits some meaningful definitions of "title" that clarify what are are are not titles.
19:48:15pondlifeI don't see why "Right Quickscreen Item" is so important to merit being a title.
19:48:35amiconnWell, it isn't a sentence.
19:48:41Lloreanpondlife: Well that one may not be. That's why I only said "probably." I don't know if it's used as a title anywhere.
19:48:42pondlifeIt's a noun
19:48:44*Llorean really is gone now.
19:49:05fmlpondlife: i.e. if you could put something in "" in the running text it should appear exactly as it does where it really appears in the app (I hope the sentence is not a nonsense)
19:49:13kugelamiconn: but "Set as Right Quickscreen Item" is, isn't it?
19:49:45amiconnYes, almost (missing the period).
19:49:51*kugel finds that looks shity
19:49:59pondlifeWhat about "Touchpad Sensitivity"
19:50:14pondlifeOr "Stop Recording And Shutdown"
19:50:35webguest842nd one is a sentence
19:50:43kugelpondlife: I can agree on title case with the former, not with the latter
19:51:00pondlifeThey both look wrong to me
19:51:06kugelbut I'd prefer "Touchpad sensitivity" too
19:51:09fmlI'd expect those to be in title case since they all apper as menu entries
19:51:34pondlifeThis is the problem with entertaining Title Case.
19:51:56*kugel expects nothing
19:52:05pondlifeIf you exclude it then there's reduced chance of inconsistency
19:52:20linuxstbPlus it looks right ;)
19:52:20kugelI haven't ever put extra attention on title or sentence case in any software
19:52:25pondlifeIn German you always capitalise nouns, right?
19:53:02pondlifeJust checking.. it's been some years since I wrote much.
19:53:16fmlThis is simple: if it's a menu item do it title case. For example, "Set As Right Quickscreen Item". In the manual, we'd have "To set this item as the right quickscreen item, select Set As Right Quickscreen Item"
19:53:40kugelfml: It's not about the manual
19:53:59kugelSure the manual is consitent with Rockbox, that's not the point of this discussion
19:54:19kugels/is/should be/
19:54:21fmlkugel: not *only* about the manual. But my definition covers all cases (so I hope)
19:54:28pondlifeIt might not be 100% language, but I think it's quite a close match.
19:55:47 Quit webguest84 ("CGI:IRC (Ping timeout)")
19:55:56 Join gromit`` [0] (
19:56:04kugelany further opinions about the sorting patch?
19:56:16*linuxstb looks up the definition of "title" -
19:56:16 Quit jaykay (Read error: 110 (Connection timed out))
19:56:21*kugel thinks a -dev ml mail would be appropriate
19:56:32pondlifelinuxstb: Yes, not heading
19:56:44linuxstbThe key being "the name of a book, musical composition, or other artistic work."
19:56:59kugel"Law a right or claim to the ownership of property or to a rank or throne. " certainly applies :D
19:57:08pondlife"a caption or credit in a film or broadcast" is closest to a Rockbox menu option...
19:57:08linuxstbSo menus don't have "titles", they simply have a name.
19:57:17fmlkugel: in the last commit, I think it'd be better to declare a symbol with a meaningful name (e.g. HAVE_PLAYBACK_CONTROL) and use it instead of repeating rather long condition several times.
19:57:40kugelfml: feel free to work that out
19:58:12kugelI simply copied from SOURCES, which decides whether to compile the overlay
19:58:33linuxstbpondlife: What would they be (caption or credit) ?
19:58:39pondlifeNeither ;)
19:59:10pondlifeBut that's exactly what I meant by "heading != title"
19:59:12linuxstbNo, I meant what are examples of captions or credits in a broadcast/film?
19:59:47pondlife"Executive Producer: Max Badhands"
20:00:00pondlifeThe titles are rolling
20:00:16linuxstbSo, "job title" ?
20:00:23kugelrasher: another idea: "Sort numbered filenames" ?
20:00:46pondlifekugel: I look forward to the option whatever it's called :)
20:00:50linuxstbBut I think that makes it clear that the rockbox menu items are not considered titles...
20:00:51kugelhm, but that may imply "only leading" too, or not?
20:01:41kugelI like that definition. We souldn't accept another one, in case Llorean comes up with some
20:02:07linuxstbThat's from the Oxford English Dictionary (or at least, the Compact version...)
20:02:08pondlifeThe OED is the ultimate English language reference, I'd say...
20:02:23kugelyes, just saying
20:02:37Lssits missing some ink
20:02:44LssQED > OED
20:03:11fmlkugel: the patch is at
20:03:54kugelfml: I don't think this is needed to be local to chessbox only
20:04:19 Quit gromit` (Read error: 110 (Connection timed out))
20:04:33fmlkugel: that wouldbe another step. But I get allergic if I see something like this repeated.
20:04:38kadobanbut i don't think it's possible to have a global define for that. it depends on the size of the plugin
20:05:04kadobanfor example, goban is also an overlay, but with slightly different logic
20:05:35kugelthen the logic should be a parameter
20:07:50kugelany opinion on "Sort numbered filename"?
20:08:02fmlkugel: do it as you want, but without repetition of code :-)
20:08:47fmlkugel: that means that the name itself is sorted, not quite what it is
20:09:31 Quit kugel (Nick collision from services.)
20:09:36 Join kugel [0] (n=kugel@rockbox/developer/kugel)
20:09:43fmlRespect numbers in the beginning of file names?
20:09:53kugelfml: I don't quite understand
20:10:04kugelwell, not only in the beginning
20:10:36gevaertsamiconn: Speed stops going up with "higher" UDMA modes. That doesn't have to mean anything as such, but (a) USB *can* be faster (cf. the mini), and (b) it tends to go up with modes on the CF-mini. I think it's at least worth investigating if this is indeed the bottleneck
20:10:38 Part pondlife
20:11:25rashergevaerts: Is the USB code locked at 16kb r/w?
20:11:48fmlkugel: I thought that was only for number in the beginning, i.e. not in "Give me 1000 things"
20:12:32 Join midgey [0] (
20:13:30gevaertsrasher: the way it's currently structured, yes (although it can do lower)
20:14:54amiconngevaerts: Yeah, but why are you saying 16KB blocksize is the limiting factor on the Video, but not on the Mini?
20:16:13gevaertsamiconn: hm, it probably is as well. I was mostly looking at the read numbers when I wrote that bit
20:16:17 Join jaykay [0] (
20:16:30*gevaerts should think a bit more before writing things :)
20:16:56*amiconn wonders how the OF manages the 15MB/s
20:17:12kugelfml: no, it also sorts "Give me 999 things" before "Give me 1000 things"
20:18:52*gevaerts votes to call this "intuitive sorting (but we won't tell you who's intuition)"
20:19:42*kugel points gevaerts to the -dev ml :)
20:20:50fmlkugel: oh, I didn't know. How about "9_give me 1000" and "10_give me 999"? The first number wins? How about "give me 1000" and "2_give me 999"? I think if it applies to other numbers then it can get too far
20:22:07kadobanfml: of course it does the sane thing in those cases, and stops after the first number/difference is resolved
20:22:36gevaertsamiconn: I've been thinking about caching a lot more data and only writing if we either have a lot or a SCSI command comes in that doesn't write to the next consecutive sector. Since any modern OS will send SCSI_ALLOW_MEDIUM_REMOVAL commands to removable devices we know when the OS is done using it, so this *should* be safe unless the DAP crashes
20:23:36 Quit kugel (Nick collision from services.)
20:23:41 Join kugel [0] (n=kugel@rockbox/developer/kugel)
20:24:14linuxstbgevaerts: Do you know when Linux sends that? On umount or eject?
20:24:40amiconngevaerts: Do you know how to detect on the device side that the device was safely ejected?
20:25:02amiconnThe various OF's manage it so it is possible
20:25:03gevaertslinuxstb: SCSI_ALLOW_MEDIUM_REMOVAL is on unmount. Eject is SCSI_START_STOP_UNIT (with various bits set)
20:25:18gevaertsamiconn: I have to look at the details again, but yes.
20:25:26linuxstbgevaerts: What about when multiple partitions are mounted?
20:25:56fmlkadoban: I mean, what happens if you have two titles, one with a number at the start and the other with a number in the middle? E.g. "20_xyz" and "Give me 5"? How would you sort these two?
20:26:00gevaertslinuxstb: I should try, but I would expect SCSI_ALLOW_MEDIUM_REMOVAL when the last one is unmounted
20:26:07kugelfml: numbers are considered as 1 unit
20:26:28kugeland still stops after the first difference, but with complete numbers being 1 unit
20:27:00kugelit sorts 20 before g
20:27:37kadobanfml: the only time that it's different from currently is when two numbers are compared. this doesn't happen unless the strings are identical up to that point
20:28:04linuxstbgevaerts: Have you thought about switching from UMS to charging mode on eject?
20:28:18*linuxstb seems to recall having this discussion before...
20:28:18fmlkugel: aha, now I see. With rule "a number is 'less' than any non-number character?
20:28:27gevaertslinuxstb: it should do that. I know it doesn't though...
20:28:27kugelfml: yes
20:29:29amiconngevaerts: Actually even some hw-bridge targets can do that, but it's not implemented in rockbox yet
20:29:35fmlkugel: that makes sense, yes.
20:29:52gevaertsamiconn: I'm actually not sure if it's legal SCSI behaviour though
20:30:29gevaertsBut as long as people aren't running their mission-critical ACID databases on a sansa I think we're fine
20:31:04 Quit nuonguy ("Leaving")
20:32:32 Quit midijunkie ("?(???~•~)?")
20:33:43 Join rocko [0] (
20:33:52 Quit tmzt (Read error: 60 (Operation timed out))
20:34:47***Saving seen data "./dancer.seen"
20:35:27 Join bluefoxx [0] (
20:35:34rasherWhat's the status of the theme site?
20:37:06kugelthe status didn't change recently, I guess
20:37:08bluefoxxHey. I've got several questions, but I'm not sure which to start with here. I suppose asking how i might rip the factory firmware from a stock e260 to flash onto a rockboxed one with newer firmware might be a start. Anyone care to help?
20:37:22 Join tmzt [0] (
20:37:43kugelinstall the new firmware, then install rockbox?
20:38:45bluefoxxI ask this, because on my rockboxed sansa, I had to redo the firmware after screwing around too much and it locked up on USB connectivity, but now it does the obnoxious database rebuild upon connecting to my PC
20:38:59 Join kachna [0] (
20:39:33 Join zimba42 [0] (
20:39:46linuxstbbluefoxx: Do you use the original firmware, or just to connect to your PC?
20:40:12bluefoxxI've attempted what the page in the FAQ says, but the firmware file seems no good. whenever that stock firmware loads, the sansa crashes and gives green/blue bars and waves
20:40:12 Join Strife89 [0] (n=michael@
20:40:27bluefoxxI use the origional firmware simply for connecting to the PC
20:40:34 Quit robin0800 (Read error: 104 (Connection reset by peer))
20:40:38linuxstbSince yesterday, Rockbox has its own USB mode, so you don't need the original firmware
20:40:45bluefoxxI saw.
20:41:06 Join robin0800 [0] (
20:41:20 Join petur [50] (n=petur@rockbox/developer/petur)
20:41:41zimba42is there anything new about the sansa e200v2/fuze playback problems? is it a storage driver problem?
20:42:17bluefoxxI don't plan to upgrade, until I can figure out how to customize some things. The menu layout is radically different from my liking, as is the typing navigation. Theres other things too, like the revision I'm using I can tap the <rec> button in the WPS screen and it brings up my current playlist.
20:42:45n1szimba42: it is likely a storage problem but i dont' think anyone knows exactly what the problem is
20:45:00bluefoxxWhich would have been my second question. How could I customize a build of rockbox with those features? I'm also looking to find a stripped down version that does recording only, I have a sansa that works, but has a broken screen(cable inside is bad). I don't want to throw it away, so the plan is to turn it into a voice recorder or sorts...
20:45:21zimba42n1s: how to debug it then?
20:45:21 Join pyro_maniac [0] (
20:45:35kadobanbluefoxx: there'sa blind interface...use that? :)
20:45:38n1sbluefoxx: you need to patch the source and compile your own version
20:46:22zimba42yea.. patch it! make the world a better place!
20:47:18bluefoxxI do recall (rather painfully) using the voice menus when that was my main player (screen broke, hunted down another v1, swapped the memory chips)
20:47:35n1szimba42: i'm not sure, probably you should start by asking people who have investigated it
20:47:37bluefoxxSo can I get a link to the page that tells my how to patch this than?
20:48:36bluefoxxI rather have to leave soon, but a link to the page for this would be great, when I get back home tonight I plan to take a look and give it a shot.
20:49:22BigBambibluefoxx: There are many links in the wiki telling you how to compile
20:49:37BigBambibluefoxx: However, what changes to make to the source is down to you
20:49:41n1sbluefoxx: our wiki has plenty of info on how to set up a dev environment and compile, start lookign under the "for developers" section in the index
20:51:11bluefoxxWell I don't have time to dig for a link right now, but if I could get a couple than I would have them waiting for me when I get back, as a nice reminder to me saying "Oh yea, I'm that other fiddly project you have to work on" otherwise I'll end up spending my time updating backups and by the time I remember I'll be going to bed...
20:52:50bluefoxxThank you. Anyways, I have to go now. Outside life and all that.
21:08:17gevaertsI get 8.9 MB/s when reading using udma4, 8.2 MB/s using the USB_USE_RAMDISK code in usb_storage.c, and 8.8 MB/s using the USB_USE_RAMDISK with the memcpy()s commented out (so using whatever data happens to be there)
21:08:48 Quit obo ("Bye")
21:14:00 Join akur [0] (
21:16:18gevaertsThat basically means that whatever we do on the ATA side, it won't go faster. We need to work on the USB side
21:17:57 Quit tvelocity (Remote closed the connection)
21:20:51 Quit Lear ("ChatZilla 0.9.84 [Firefox 3.1b3pre/20090226034605]")
21:23:24 Join stripwax [0] (
21:23:33 Quit stripwax (Client Quit)
21:24:31*gevaerts suspects that we'll have to implement proper chained transfers in the ARC driver to go faster
21:27:01 Part akur
21:27:45 Join zimba421 [0] (
21:28:21 Quit zimba421 (Client Quit)
21:30:12 Quit robin0800 (Remote closed the connection)
21:30:26 Quit zimba42 (Read error: 113 (No route to host))
21:38:55 Quit parafin (Read error: 110 (Connection timed out))
21:39:04 Quit Strife89 ("Gotta go.")
21:39:09n17ikhafter deleting a file on my microsd card (e200) using the new usb mode, it unmounted without warning
21:39:28n17ikhnot sure if it was deleting or reading, I was moving files off using windows
21:39:37 Join bluebrother [0] (n=dom@rockbox/developer/bluebrother)
21:40:48n17ikhwas in the middle of the move and windows complained that the rest of the files no longer existed
21:41:02n17ikhactually, both the microsd card and the sansa itself unmounted
21:41:21gevaertsWhat exactly do you mean by "unmounted" ?
21:41:35n17ikhdisappeared from my computer
21:41:42n17ikhand from safely remove devices, and from diskmgmt
21:42:03n17ikhthe sansa didn't lock up, I unplugged the usb cable and it went back into regular mode
21:42:14n17ikhI'm on vista x64, if that helps
21:42:26gevaertsDoes the event log say anything? (I don't know if it's supposed to...)
21:42:33n17ikhthe windows event log?
21:43:08n17ikhone sec
21:43:42*kugel understands this SCROLL_WHEEL thing in rockblox
21:43:47n17ikhyeah, I've got one
21:43:47*rasher wonders why the theme site is using KiB
21:43:54n17ikhfrom the "VDS dynamic provider"
21:44:00n17ikhThe provider failed while storing notifications from the driver. The Virtual Disk Service should be restarted. hr=80042505
21:44:17gevaertshm, no idea what that means...
21:44:19n17ikhthat's the only event I have from the past week, and it happened five minutes ago
21:44:22n17ikhyeah, me either :/
21:44:44 Join perrikwp [0] (i=98214c9d@gateway/web/ajax/
21:45:03bluebrotherrasher: because it's correct when calculating 2-based? ;-)
21:45:19n17ikh*now* rockbox appears to have locked up, after I plugged it back into usb
21:45:26rasherbluebrother: It's correct, but there's no reason not to use 10-based
21:45:28n17ikhwindows is giving me the old device not recognized thing
21:45:54bluebrotherthere's also no reason to not use 2-based :)
21:46:12n17ikhunplug it from usb, it has the writing icon on the screen but it's been sitting like that for 30-odd seconds
21:46:16n17ikhon the usb screen, still
21:46:23rasherbluebrother: Some might be confused by the i
21:46:36gevaertsYou'd better reset it. It won't come back by itself
21:46:45 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
21:46:55rashermcuelenaere: Hello - I'm prodding the theme site
21:47:10bluebrothertrue, but if its not used people will never recognize the unit. Plus, I assume more people assume Byte-size to be 2-based
21:47:20mcuelenaererasher: haven't done much with it lately..
21:47:22 Quit fml ("CGI:IRC (EOF)")
21:47:25mcuelenaereeverything is in SVN
21:47:45rashermy first mission is to remove the email-validation and replace it with email-confirmation
21:48:12*bluebrother is strongly in favor of using correct prefixes
21:48:17mcuelenaereadmin also needs cleanups
21:48:55n17ikhok, after resetting it and plugging it back in Vista asks me if I want to "scan and fix" the sdhc card, which I guess means chkdsk
21:49:06rasherbluebrother: Me too! I just think the binary prefixes are very rarely useful.
21:49:08gevaertsyes. You'd better let it do that I think
21:49:12n17ikhyeah, did
21:49:26bluebrotherrasher: true, so spread the word!
21:49:26n17ikhactually, I did it through my card reader
21:49:40 Join parafin [0] (
21:50:16rashermcuelenaere: Also, you (or someone else) used a PHP 5.3.0 feature. That's probably not going to work on the server. It didn't even work on my Debian Unstable!
21:50:30rasher(third argument to strstr)
21:50:31mcuelenaererasher: which?
21:50:39mcuelenaereprobably me then
21:51:32*mcuelenaere thinks you could replace it with a self-made function
21:52:02mcuelenaererasher: see first comment at the documentation page about that function
21:52:20n17ikhgevaerts: did it again
21:52:43n17ikhcut a directory from the microsd card, paste it onto my computer, it gets through one file and usb mode dies
21:52:49rashermcuelenaere: I just did the obvoius way and used strpos and substr. Not exactly speed critical piece of code
21:53:27mcuelenaereis there any description/documentation about the target-specific code regarding PCM? (pcm-<target>.c)
21:55:18*gevaerts tries to think of a way to debug this
21:55:32n17ikhlet me see what else, if anything, I can get to trigger it
21:55:38n17ikhlike if it's writing or just deleting
21:55:55 Join mirak [0] (
21:55:55n17ikhbecause I wrote files to the internal flash through rockbox just fine
21:55:58n17ikhdidn't try deleting though
21:56:12 Join jordoex_ [0] (
21:56:22gevaertsIt's unlikely to be anything specific at that level. It just reads and writes ranges of sectors anyway
21:57:11gevaertsCould you test reading and writing to the internal flash a bit more? I'd like to know if that difference is real or just a coincidence
21:57:29kugelhow about we just call this "Recognise numbers in filenames", with an on/off setting?
21:57:42kugelgevaerts, rasher, others: any opinion to that?
21:57:46 Quit LambdaCalculus37 (" ajax IRC Client")
21:58:04rasherThat doesn't even suggest what it does, if you ask me
21:58:07linuxstbkugel: Doesn't it also apply to directory names?
21:58:23kugelyes, right
21:58:37kugel"Regognise numbers while sorting"?
21:59:27gevaertskugel: I'm actually leaning towards this "intuitive sorting". It implies that it does something fuzzy that you probably won't entirely understand, which seems to be precisely what the setting does :)
21:59:52kugelhm, iirc we already had "number-aware" as a suggestion, which was just as controversial as the others
22:00:18kugelgevaerts: I'd go with that too
22:00:24 Quit killan ("( :: NoNameScript 4.22 :: )")
22:00:26kugelI'm open to many names
22:00:46*linuxstb suspects kugel would be happy with any name, just to finally commit it ;)
22:01:59kadobani thought that the whole point of the example lists was to _not_ have to come up with a name for the sorting?
22:02:59evilnick_6"windows-esque fuzzy sorting"
22:03:30gevaertsIs increasing HZ on PP safe?
22:03:38n17ikhgevaerts: it's maddenly inconsistent
22:03:41 Join Matze88 [0] (
22:03:48kugellinuxstb: well, to be honest....
22:03:51gevaertsn17ikh: in what way?
22:03:58n17ikhI copied, pasted, deleted, and moved dozens of files without anything happening and then it happens during one move
22:04:01 Join killan [0] (
22:04:10kugelsurely not any, but quite a few more before all these tiresome discussion
22:04:22n17ikhthe original folder I was trying to move off the sd card I finished moving without incident
22:04:49n17ikhpicked a bigger folder, cut and pasted it elsewhere, and it gets through 24 of 44 mp3 files and then does it
22:04:51kugelkadoban: right, but now bluebrother complains on the task
22:04:57kadobanoh, i see
22:04:58Matze88hi, anyone here has wiki write access? i just wanted to share my experience with CFModGuide - there is a sentence with the cheaper dealextreme IDE-CF Adapter ("This even cheaper one is available as of 12/03/08 and is slightly smaller than the first example given (1.97 in x 1.85 in x 0.2 in vs. 2.2 in x 1.89 in x 0.26 in). One has been ordered for testing. −− EricJorgensen - 03 Dec 2008 "). I ordered that one and successfully used it. m
22:05:08gevaertsn17ikh: internal or sd?
22:05:10kugelhe apparently didn't notice all the discussion about using examples
22:05:29n17ikhI couldn't get it to do it with internal, but I can keep trying
22:06:05gevaertsn17ikh: is the card inserted securely? Could there be a bad contact there?
22:06:08midgeyMatze88: if you register for the wiki you can add that yourself
22:06:20Matze88midgey: but i wont get write access when i register
22:06:25n17ikhit's inserted as securely as I can get it
22:06:33 Quit kugel (Nick collision from services.)
22:06:34midgeyjust ask ;)
22:06:36n17ikhI'm not moving the player around while I'm doing this stuff
22:06:38 Join kugel [0] (n=kugel@rockbox/developer/kugel)
22:06:59n17ikhI can read files off the card without incident, as far as I can tell
22:07:14n17ikhit's just right after deleting a file, very occasionally, it makes usb mode die
22:08:10n17ikhhah, it died during the chkdsk that time
22:08:18n17ikhmust be writing that triggers it
22:08:28n17ikhwindows refuses to believe it's gone
22:08:35Matze88okay so i'm asking for write access now :-) My Wikiname is MatthiasLarisch. any more infos needed? :-)
22:08:55kugelyes, PIN of your credit card
22:10:09gevaertsn17ikh: do you have a build environment set up?
22:10:10kugelMatze88: permission granted! Promise to not spam :)
22:10:20n17ikhgevaerts: no :/
22:10:25Matze88kugel: okay, i promise. thank you
22:10:54gevaertsn17ikh: ok. That just means that I'll have to provide test builds if I see something I guess
22:11:01linuxstbHow about "Sorting algorithm: Alphabetical | ASCIIbetical" (inspired by )
22:11:12n17ikhlet me see if I can get it to trigger on the internal flash
22:11:15*linuxstb suspects that might be hard to translate though...
22:11:19gevaertslinuxstb: that's totally misleading
22:12:02gevaertsNeither Alphabetical nor ASCIIbetical talks about numbers at all, and ASCIIbetical is likely to mean nothing to non-programmers
22:12:09kugelgevaerts: if we picked intuitive, people will probably wonder why it even is an option?
22:12:36kugelgevaerts: and non-programmers don't want that anyway -> problem solved :)
22:13:04midgeywhat was the argument against a name similar to Numerical Sorting?
22:13:20gevaertsn17ikh: is this e200 or c200?
22:13:26linuxstbgevaerts: The problem is that anything that mentions numbers makes people think it's just dealing with numbers, not strings with numbers in the middle.
22:13:34n17ikhI have a c200 I can try, if you'd like
22:14:01gevaertslinuxstb: I'd interpret Alphabetical vs ASCIIbetical as "how does it treat uppercase?"
22:14:11gevaertsn17ikh: that may be helpful
22:14:34 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
22:16:34kugelgevaerts: uppercase is a seperate setting
22:16:52kugelor rather, case-insensitivity is
22:16:59gevaertskugel: which is why ASCIIbetical is confusing
22:17:59linuxstbkugel: BTW, are you sure your fix to chessbox is right? The condition looks reversed...
22:18:24kugellol, indeed
22:18:38*kugel slaps himself once more
22:20:20kugel"#if (MEMORYSIZE > 8) || !defined(SIMULATOR)" would be correct then?
22:20:47shotofaddsmcuelenaere: I need to add an extra column to drivers/nand_id.c to tell the TCC nand driver how many blocks make up one logical unit - currently this is #ifdef'd depending on model (D2/iAudio7 vs. smaller targets), but someone on the forum has a D2 which needs runtime detection.
22:20:55shotofaddsDo you have any objection to me adding this extra field, even if it ends up being TCC specific? I can't really see the point of #ifdef'ing it
22:21:02n17ikhgevaerts: I got it to trigger on the internal flash! but in a different way!
22:21:21n17ikhI moved about 1.1GB of files off the internal flash without a problem
22:21:34n17ikhwent to move them back on, it gets a third of the way through and dies
22:21:44mcuelenaereshotofadds: sure, go ahead. Are you sure you can't generalise this? (there are OF'es who use a similar idea like this and have a lot of more data, like the TCC SDK for example..)
22:21:51n17ikhvista is saying the files don't exist anymore in the folder on my desktop, but I don't think that's the case
22:22:13n17ikhit thinks the sd card and the internal flash are still mounted, but there is nothing in them
22:22:15jaykaybluebrother: there are some // comments in your latest commit, i dont know whether this was wanted
22:22:20mcuelenaereshotofadds: oh, is this specific to the TCC FTL?
22:23:45pyro_maniacgevaerts: I get warnings of redefining USE_ROCKBOX_USB in the h10 5GB build. is there anything left to rework?
22:24:01gevaertspyro_maniac: do you still define it yourself?
22:24:04kugellinuxstb: the simulator part should be removed too right? in SOURCES it was only because of the overlay, not because of the playback menu
22:25:29shotofaddsmcuelenaere: yes it's probably specific to the FTL, but I wasn't sure about polluting the "common" nand_id.c with potentially target-specific stuff... btw. the existing nand_id table is is already similar to what the Cowon OF uses - I'm talking about extending it slightly..
22:25:41pyro_maniacgevaerts: I got EXTRA_DEFINES=-DUSE_ROCKBOX_USB -DLOGF_ENABLE in the makefile. is this obsolete?
22:25:52gevaertspyro_maniac: the -DUSE_ROCKBOX_USB is, yes
22:26:58pyro_maniacgevaerts: ok thanks
22:27:13n17ikhgevaerts: now I got it to trigger on a copy, instead of a move, from my computer to the internal flash
22:27:17linuxstbkugel: I think it should simplly be #if MEMORYSIZE > 8. But I also like fml's suggestion for a single define (FS #9958)
22:27:31mcuelenaereshotofadds: yes, I understand. well I don't really mind, one extra column isn't going to waste that much binsize :)
22:28:19gevaertsn17ikh: I don't really have any ideas. I'd like to see if you can reproduce it on c200, to try to isolate it a bit
22:29:58shotofaddsmcuelenaere: a couple of the fields could be changed to 'char' anyway which more than makes up for the difference
22:31:26kugellinuxstb: yes, I did that now (both) :)
22:32:27 Part Matze88
22:32:42kugelsomething is wrong with kkurbjun's commit messages
22:34:51***Saving seen data "./dancer.seen"
22:35:09 Join Zagor [242] (
22:37:50 Join barrywardell [0] (n=barrywar@
22:39:40n17ikhgevaerts: I do have all kinds of warnings in my event logs
22:40:01n17ikh"An error was detected on device \Device\Harddisk6\DR11 during a paging operation."
22:41:59bluebrotherjaykay: Rockbox Utility is written in C++, and comments are // in c++. So yes, this is correct and wanted.
22:42:43Bagderfor the record, I don't really care about the sorting option name
22:42:53BigBambime neither
22:43:07*kugel neither
22:43:11kugelat least not anymore....
22:43:21jaykaybluebrother: ok, sorry
22:44:06gevaertsn17ikh: are those from the e200 tests?
22:44:22n17ikhgevaerts: yes
22:44:24gevaertsAlso, can you get details for those errors?
22:44:27n17ikhI've got the c200 hooked up now
22:45:03 Part woyciesjes_ ("Ex-Chat")
22:45:37 Join MethoS- [0] (
22:45:53n17ikhgevaerts: I can save them to an event file and send it to you, that might be the best thing
22:45:58 Quit MethoS- (Remote closed the connection)
22:46:08n17ikhif you can open up a vista event file
22:46:11 Join SirFunk_ [0] (
22:46:13gevaertsn17ikh: probably not
22:46:31n17ikhactually, I can save them as csv or as tab-delimited
22:46:35n17ikhlet me see what that looks like
22:47:08n17ikhoh, it's fairly useless, doesn't save details
22:47:11gevaertsIf it has more information, I can use it, otherwise it just says that there was an error, and we know that already...
22:47:32n17ikhwell, in "details", there is what looks like a small hex dump
22:47:39n17ikhI'm not sure what good that will do you
22:47:51gevaertsnot much I think
22:48:01linuxstbHow about "Sort numbers as text: Yes (03, 1, 10, 2) | No (1, 2, 03, 10)" ?
22:48:40n17ikhyeah, that's all the details there are
22:49:00n17ikhI actually have 13 of these, and they are "warnings"
22:49:04n17ikhI only have one error
22:49:26n17ikhbut all the warnings say the same thing, more or less
22:49:27 Quit jordoex_ ("bye")
22:49:37gevaertsLet's see what the c200 does
22:49:41n17ikhyeah, testing that now
22:50:54kugellinuxstb: not sure..
22:51:05 Part pyro_maniac
22:51:23 Quit MethoS-- (Read error: 60 (Operation timed out))
22:51:29 Quit kugel (Read error: 104 (Connection reset by peer))
22:51:39 Join kugel [0] (n=kugel@rockbox/developer/kugel)
22:52:19kugellinuxstb: is "Recognize numbers while sorting" so bad?
22:52:21kugelbecause that's exactly what it does
22:54:22kugelnatsort is number-aware, ascii-sort only digit aware
22:54:46linuxstbI can live with that. So the options would be yes/no with the examples?
22:55:09linuxstbbluebrother: ?
22:56:36linuxstbRead back a few lines - what do you think of that suggestion? Or do you just care about using something in the .cfg?
22:56:53linuxstb^using something sensible
22:57:13 Join florinp3 [0] (n=4e619cc0@gateway/web/cgi-irc/
22:57:43bluebrothersounds reasonable to me (and much less prone to misunderstanding)
22:58:21 Quit florinp3 (Client Quit)
22:58:28 Join florinp3 [0] (n=4e619cc0@gateway/web/cgi-irc/
23:00:27 Quit petur (Remote closed the connection)
23:00:44 Quit SirFunk__ (Read error: 110 (Connection timed out))
23:03:52 Quit florinp3 ("CGI:IRC (Ping timeout)")
23:04:28 Join florinp3 [0] (n=t@
23:05:03florinp3hello everyone
23:06:03florinp3I think I might have a problem with my fuze that's sort of related to Rockbox
23:06:36florinp3I had my both my 1G Nano and my Fuze connected to my PC
23:06:45gevaertsrasher: I assume that WPS in "Start Screen is set to WPS" means "Resume Playback"?
23:06:54rashergevaerts: yes
23:07:12florinp3and used Rbutil to update Rockbox and bootloader
23:07:37rashergevaerts: I wouldn't be surprised if setting it to recording screen or radio might have the same effect (but haven't checked)
23:07:48gevaertsrasher: it seems to work properly here
23:07:48florinp3the problem is that I noticed afterwards that it had my Fuze selected instead of my Nano
23:07:55 Quit bmbl ("Woah!")
23:08:08bluebrotherflorinp3: what's the exact problem now?
23:08:08florinp3what does Install bootloader actually do?
23:08:49bluebrotherit installs the bootloader. But the Ipod bootloader will only install on an Ipod, so no problem here.
23:08:54rashergevaerts: Weird. Which target?
23:09:06florinp3Rbutil updating the Nano bootloader to the Fuze
23:09:10bluebrotherYou'll have the build installed on the wrong player, but that isn't an issue either −− just remove the .rockbox folder
23:09:40bluebrotherflorinp3: again: the Ipod bootloader will only install on an Ipod. The mountpoint / drive letter doesn't have an effect on bootloader installation
23:09:53 Quit {phoenix} (Remote closed the connection)
23:09:59bluebrotheryou *cannot* install an Ipod bootloader on a Sansa, even unintentionally.
23:10:11florinp3ok, thanks
23:10:27bluebrotherjust remove the .rockbox folder on the fuze and reinstall the build on the ipod
23:10:31florinp3I just wanted to double check before I remove the Fuze :)
23:10:59gevaertsrasher: e200. It's still running your test build though.
23:12:03rashergevaerts: Weird. Err, try with a current build?
23:12:58gevaertstrying that now. Isn't the "try with a current build" line meant for the bug *reporter*? :P
23:13:17 Quit evilnick_6 (" ajax IRC Client")
23:15:19gevaertsrasher: still working fine. Official r20120 and bootloader from FS #9955
23:15:59 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
23:16:13rashergevaerts: Okay, so. 1) Play music. 2) Shut down from within the WPS. 3) Plug USB cable. 4) I don't get a USB connection
23:17:15gevaertsrasher: I do exactly the same, except that in step 1 I play a radio show and not music. It works fine
23:17:16mcuelenaereany touchscreen user here willing to help me testing a patch?
23:17:28rashergevaerts: This is absolutely mad.
23:17:45gevaertsrasher: what sort of music do you play? Big files, high bitrates?
23:17:55rasherLow bitrate vorbis
23:18:24rasherHuh, this time it actually crashed right when it was about to play
23:18:32gevaertsThis is low bitrate mp3 (32kbit)
23:19:49n17ikhgevaerts: I am unable to trigger the bug on the c200
23:19:52gevaertsIt does switch to the wps, but then it immediately switches to the USB screen
23:20:13n17ikhboth with internal flash and with sdhc
23:20:15rasherSeems the screen is frozen while the cable is pluggined in. Unplugging makes it start playing
23:20:33gevaertsrasher: same here, except that it's in proper usb mode
23:20:59gevaertsn17ikh: are you using the same cable with both?
23:21:24gevaertsI assume the OF never had this sort of issue?
23:21:27n17ikhsame cable, same sd card
23:21:31 Join MethoS- [0] (
23:21:34n17ikhnot that I ever ran across
23:21:56gevaertsCould you try on the e200 without the SD card inserted?
23:22:07n17ikhshame I don't have a second e200
23:22:08*gevaerts still doesn't have much ideas
23:22:37n17ikhcould be it's something windows is doing
23:22:56n17ikhI'm not sure what kind of mangling it would be doing though that would cause that kind of a bug
23:23:05*rasher isn't running Windows
23:23:12gevaertswell possibly, but why isn't it doing it on the c200?
23:23:52rashergevaerts: It really doesn't leave the wps
23:23:58gevaertsrasher: (switching to full-blown support mode :)) have you tried clearing your settings and checking the filesystem?
23:27:14gevaertsrasher: I can't think of anything else...
23:27:33rasher"differences between boot sector and its backup."
23:28:02*rasher flips a coin and copies backup to original
23:30:11 Join itcheg [0] (i=62db4767@gateway/web/ajax/
23:30:22 Quit tyfoo (Connection reset by peer)
23:30:58rashergevaerts: now it's connecting.. :\
23:31:17rasherActually, no.. but it's at the USB screen
23:31:31gevaertsrasher: you did set the starting screen back to Resume playback, right?
23:31:33rasherSome timing thing? I had a rather large playlist last time
23:31:51rashergevaerts: Yes
23:32:12gevaertsI guess it could be timing
23:32:27rasherThis time it skipped the WPS and went to the USB screen, but no connection has been established ([271590.998770] hub 6-0:1.0: unable to enumerate USB device on port 1
23:32:34rasherThat message appeared in dmesg
23:33:37 Quit florinp3 ()
23:33:50n17ikhgevaerts: triggered almost immediately on the e200
23:33:56n17ikhthis time on reading
23:34:00n17ikhnot writing
23:34:15n17ikhcopied a bunch of files from the e200's internal flash without the sd card in
23:34:16rashergevaerts: Rebuilt a large playlist, now it's stuck at the WPS
23:34:27rasherAnd I get "unable to enumerate" again.
23:34:43n17ikhand soon after I get "could not find this item", as windows thinks the file no longer exists
23:35:28rashergevaerts: Err, with a playlist of one file, I still get unable to enumerate :\
23:35:47n17ikhthe e200, however, is not locked up
23:35:51gevaertsrasher: unable to enumerate should be totally unrelated to playback
23:35:57rashergevaerts: But now it went to the usb screen
23:36:49gevaertsn17ikh: can you submit a bug report? I can't look at it any longer tonoght
23:37:07rashergevaerts: Running short on ideas then
23:37:09n17ikhactually, let me try something else
23:37:24gevaertsrasher: a bit, yes
23:37:27n17ikhthe e200 has an old crufty rockbox install and config file, maybe something is screwing it up
23:37:39n17ikhlet me wipe the .rockbox folder and config
23:38:30rashergevaerts: Plugging while playing works fine
23:39:09gevaertsrasher: I'm wondering if calling audio_get_buffer() while playback is starting could cause trouble
23:39:17moosgevaerts: Pandorabox :) /me missed the dance
23:39:48rashergevaerts: Doesn't sound unlikely. Doesn't really explain why it's working for you though
23:40:28gevaertsrasher: if it's timing related things like filesystem fragmentation might come into play
23:40:41 Quit itcheg (" ajax IRC Client")
23:44:20*freqmod_gq updated the wikipedia article of rockbox mentioning that rockbox has usb support in the daily builds in the microSDHC paragraph
23:45:15kugelfreqmod_gq: don't forget USB
23:45:43freqmod_gqit wrote USB
23:46:02freqmod_gq.... An external reader is required to put new files on the card as Rockbox does not yet have a software USB implementation in a stable release. An USB implementation is available in the daily builds and are scheduled for version 3.2 if no problems are discovered. ...
23:47:03gevaertsrasher: on re-reading usb.c and friends, I think that "unable to enumerate" could be caused by this sort of issue after all
23:47:20gevaertsfreqmod_gq: "are scheduled"?
23:47:29freqmod_gqmaybe planned?
23:47:38gevaerts"is scheduled"
23:47:47freqmod_gqyes, thank you
23:48:18*gevaerts still isn't as familiar with jhMikeS's new usb init sequence as he'd like to be :(
23:48:49rashergevaerts: wouldn't it make more sense to detect usb even before the startup screen stuff is done?
23:48:59 Quit jaykay (Read error: 104 (Connection reset by peer))
23:49:46n17ikhso once USB is deemed stable enough, will the sansa bootloaders be rewritten to load rockbox instead of the OF when usb is plugged in when the player is turned off?
23:49:48 Quit perrikwp (" ajax IRC Client")
23:49:48 Quit mirak (Remote closed the connection)
23:50:17kadobann17ikh: they already are, there are test bootloaders in FS (i think it's 9955)
23:50:46n17ikhoh, ok
23:51:08gevaertsrasher: maybe, but while that may make your issue go away it won't solve the bug. From what I understand you could trigger this by pressing play and plugging in at exactly the right time
23:51:40gevaertsso I'd actually like to keep it as it is to help reproduce it :)
23:51:44freqmod_gqbtw. is the usb support for sansa able to write to the sd card?
23:52:33kugelgevaerts: maybe this is the comment Nico_P left
23:52:41gevaertskugel: ?
23:52:44n17ikhgevaerts: I think it may have been my config file or something old in .rockbox
23:52:56n17ikhthus far I have been unable to re-trigger the bug with a fresh .rockbox directory
23:53:17kugelgevaerts: "The bigger issue is that I'm not sure whether the audio_stop_playback call that happens on USB connect actually manages to close all the handles: if the buffering thread has acknowledged the USB connect before the audio thread, I guess the answer is no and I would even expect a deadlock."
23:53:25gevaertsn17ikh: which revision were you running previously?
23:53:28rashergevaerts: Can't really say much more than "happy to test, and good luck!"
23:53:51n17ikhwell, I installed 20112 today
23:53:53gevaertskugel: ah yes. Yes, I think it's not unlikely to be related to that
23:53:55n17ikhnot sure what I had before that
23:54:12n17ikhbut I didn't wipe .rockbox, I just overwrote it
23:54:23n17ikhbut wiping it and replacing it with 20120 seems to have fixed the problem, so far
23:54:45n17ikhI'm most of the way through writing 800 megs of files to the sd card, which would have triggered the bug previously
23:56:15gevaertsn17ikh: I don't really see how any remains of older installs could cause this, but OK. If you still see it, please file a bug report. Otherwise we'll assume that it was indeed caused by some weird remnants of the past
23:56:32n17ikhI'll continue testing
23:57:36n17ikhI just figured the first step in making it reproducable would be to have a completely fresh rockbox install, and now I can't reproduce it
23:58:22amiconnkugel: This is one more reason why the acknowledging stuff needs to be serialized
23:58:49 Join T0paz [0] (

Previous day | Next day