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 2008-02-10

00:00:26 Quit fyrestorm (Read error: 104 (Connection reset by peer))
00:01:00 Join fyrestorm [0] (
00:01:29rasherDoes anyone see why we're still shipping profont? The license (which we include) says "You may redistribute unmodified copies of ProFont Distriubtion"
00:01:48rasherI don't believe shipping it as .bdf counts as unmodified.
00:03:06scorche|shrasher: hrm...i am trying to remember exactly how we counted it, i think lostlogic might remember...
00:03:08linuxstbAnd we're also not distributing "ProFont Distribution" - which is defined as 7 different things...
00:03:51rasherFWIW Bagder seems to agree
00:04:03LloreanThe original font is "Freeware"
00:04:20LloreanThey do state their intent is to have profont distributed "as widely as possible", but apparently only as that specific package.
00:07:18amiconnsaratoga: That patch is nasty as-is
00:07:52amiconnIn pcf50605.c, it disables voltage setup for all targets, and only implements it for ipod video
00:08:23amiconn-> all other PCF PP targets no longer get those voltages set up
00:09:11 Join austriancoder__ [0] (n=austrian@
00:09:25saratogaamiconn: could you post that in the FS thread so taht buschel sees it?
00:09:30amiconnAnd why it switches off the dock connector voltage even on G5 is beyond me. I'd expect reports about non-working accessories from this
00:09:45Lloreanrasher: Sadly, the maintainer email in the license doesn't work. =/
00:11:02 Join countrymonkey [0] (n=4b05639a@gateway/web/cgi-irc/
00:11:17 Quit ompaul (Client Quit)
00:13:16countrymonkeyrasher: Any news on the arabic translation?
00:17:27rashercountrymonkey: would you mind not asking the same question every 2 days? It's not like I'm going to keep the translation to myself unless you ask me for it.
00:18:34 Join aneka [0] (n=kvirc@
00:22:51karl_rasher: is your nick referring to the irish bacon, or something else?
00:24:29 Quit austriancoder_ (Read error: 110 (Connection timed out))
00:27:08countrymonkeyrasher: speaking of translations, your translate page has a mistake. It reads tradtional chinese. I assume you mean traditional chinese. Could this please be fixed?
00:28:31saratogaamiconn: should i make those corrections now and comitt it or do you want me to kick this back to Buschel for further cleanup?
00:29:46rashercountrymonkey: fixed
00:29:59rasherkarl_: That's off topic, and no, it doesn't refer to anything.
00:30:03 Quit desowin ()
00:30:28gevaertsWrite support still has some problems. I copied a linux kernel source tree to my sansa and back, and there is a problem with 21 files (out of 22390)
00:32:18 Join grndslm [0] (
00:33:21gevaertsActually it might just as well be read support. I'll boot to OF to check
00:35:37 Quit petur ("Zzzzz")
00:45:55 Quit ender` (" Kids. You gotta love them. I adore children. A little salt, a squeeze of lemon--perfect. -- Harry Dresden")
00:46:31 Join crzyboyster [0] (n=6210f026@gateway/web/cgi-irc/
00:46:36 Quit karl_ (Remote closed the connection)
00:46:47crzyboysterJust finished up the update to the Hindi translation >
00:48:52 Join webguest15 [0] (n=5bc02253@gateway/web/cgi-irc/
00:49:17 Quit Domonoky (Read error: 104 (Connection reset by peer))
00:49:36 Quit ctaylorr (Read error: 110 (Connection timed out))
00:49:47gevaertsFalse alarm : the files were not actually different, but there were some files missing due to having names only differing in case, so some of them disapeared when copying them to a FAT filesystem.
00:50:33rashercrzyboyster: Please fix your text editor to not insert a BOM at the beginning of files
00:50:34 Quit webguest15 (Client Quit)
00:50:55crzyboysterSorry, but what is a BOM?
00:52:41Lloreangevaerts: So it seems reliable?
00:52:53crzyboysterI use notepad. Is the "$Id$" what you guys are talking about?
00:53:26 Join [SiC][DeZmOnD] [0] (i=werj@
00:53:27rasher"sed -i -e 's/\xef\xbb\xbf//' filename" to remove a BOM
00:53:32rashercrzyboyster: no, read linuxstb's link
00:54:18crzyboysterI really don't understand any of that...
00:54:54rashercrzyboyster: Just don't use notepad
00:55:21 Quit [SiC][DeZmOnD] (Client Quit)
00:55:33Horschtnotepad++ or TED notepad are good notepad replacements for windows
00:55:56soapnotepad2 as well
00:56:38crzyboysterPoint taken, I won't use notepad anymore (down with Windows!) and can you just remove the BOM yourself for me?
00:57:16rashercrzyboyster: Already committed.
00:59:03 Quit mf0102 ("Verlassend")
00:59:05 Quit crzyboyster ("CGI:IRC (EOF)")
01:00:08 Join Cruleworld [0] (n=Crulewor@
01:03:41 Join ol_schoola [0] (
01:04:37 Quit einhirn (Read error: 104 (Connection reset by peer))
01:07:44***Saving seen data "./dancer.seen"
01:08:28gevaertsLlorean: For me and rasher anyway. There's a report on the tracker about some problems, but I'm still looking into it.
01:14:12 Join star_jasmine [0] (n=arwyneve@
01:14:56gevaertsrasher: do you see the sd card slot appearing in your dmesg output if there is a sdhc card in it with the OF ?
01:15:56gevaertsrasher: I'm asking because I find it strange that in there is no trace of the sd card at all
01:16:36 Join linuxstb____ [0] (
01:17:26 Join ctaylorr [0] (
01:17:30rashergevaerts: So you want the dmesg output when plugging with an sdhc card while in OF?
01:18:13gevaertsrasher: yes. I don't need the entire output, but I just want to know if the OF presents two devices
01:18:30rasherI think it does
01:19:02gevaertsIt should, but maybe it just ignores it entirely if there is an unsupported card in it.
01:19:26rasherActually that seems to be the case
01:20:31 Quit ol_schoola (Read error: 110 (Connection timed out))
01:21:29star_jasminewow... congratulations and thanks to the people responsible for the rbuilt utility. I just took a look at the right one for use with screenreaders. very accessible
01:21:51 Quit Rob2222 (Read error: 104 (Connection reset by peer))
01:23:32gevaertsrasher: it prints '[sdc] Attached SCSI removable disk' for it. The tracker post doesn't contain that (neither for OF nor for rockbox), so I'm a bit doubtful about his hardware
01:23:55rasherAh, so it does. I missed that
01:25:22rashergevaerts: It even does that when no card is present
01:26:08gevaertsrasher: that's needed for hotplug support.
01:26:40rasherSure, I'm just saying there's really no reason why he should not get a second disk
01:27:04gevaertsExcept if he has broken hardware...
01:28:37 Quit linuxstb (Read error: 113 (No route to host))
01:28:37 Join crzyboyster [0] (n=6210f026@gateway/web/cgi-irc/
01:28:48 Quit Cruleworld ("Leaving")
01:29:03rashergevaerts: Where does the OF live? Is it possible that it's become slightly corrupted?
01:29:19crzyboysterWhich font gallery format is better at > ? The top one or the bottom one?
01:29:24 Quit asdrubal (Connection timed out)
01:29:35 Quit Axio ()
01:29:53BigBambigevaerts: I've just been trying the patch with an e260 + 1GB SD in linux - it mounts and I can browse both the internal memory and the sd card fine, but on the unit tiself I don't get a USB screen and I can still browse the menus etc. I havne't tried to do anything like play music as I don't want to corrupt anything :)
01:29:58gevaertsrasher: from what I understand the OF lives on the second partition.
01:30:32rasherBigBambi: I don't think the USB screen is working yet
01:30:40gevaertsBigBambi: I'm seeing the same thing.
01:30:44BigBambiOK cool
01:31:02rashercrzyboyster: The bottom one is full of weird info.. it's hard to make a decision based on an incomplete example
01:31:15BigBambiJolly nice work by the way :)
01:31:25gevaertsActually, I did try high speed again earlier today (it still doesn't work), and then I got the USB screen when I plugged in.
01:31:38BigBambiOne or the other :)
01:32:44rashercrzyboyster: Also, there's little need for such a big screenshot.. and using standard wiki syntax for the tables would be preferable
01:33:41crzyboysterrasher: What about now, is it easier to decide? And the problem is that the regular wiki syntax doesn't allow me to make it look like that...
01:33:46 Join asdrubal [0] (
01:34:30rashercrzyboyster: The top one is certainly possible
01:35:41crzyboysterHold on, can someone fix it to make it wiki syntax compatible? And how do you "log off" of twiki?
01:36:14gevaertsIf I rmmod ehci_hcd, I get the USB screen as well. There's probably something weird effect from this force-full-speed bit.
01:36:28 Join faeries26 [0] (n=arwyneve@
01:36:34crzyboysterI am leaning toward the top one, though...
01:37:10crzyboysterI will be back shortly...
01:37:11 Quit crzyboyster ("CGI:IRC")
01:37:24faeries26I must be doing something wrong. when I ask the rbuilt tool to configure/ create a voice, it downloads the current one from rockbos'x website, and claims not to be able o find an mp3 file in temp
01:37:42BigBambifaeries26: Wich target?
01:38:06BigBambiDo you not want to select speex for voice now?
01:38:20faeries26sansa e200... an e280 to be exact
01:38:33faeries26I selected speex
01:38:54BigBambiHmmm, not sure then
01:39:34BigBambiDomonoky and bluebrother are your rbutil chaps, but it is 01:40 over here at the moment so I suspect they are asleep
01:39:36 Join FOAD_ [0] (
01:39:42 Quit K4rP4D ("Leaving")
01:39:57simonrvncrashd_: ?logout=true IIRC
01:39:58faeries26oh ok. well, I can always use cygwin. very nice utility
01:40:02faeries26rbuilt that is.
01:40:21faeries26just have to figure out what I've done wrong
01:40:29simonrvncrashd_: sorry
01:40:38simonrvnperson it was meant for is gone...
01:40:50countrymonkeyWould somebody please build me a copy of svn rbutil and post it somewhere? I cannot figure out qt.
01:41:26BigBambiI'm not going to, but you might get more takers if you at least specified OS
01:41:39countrymonkeySory, windows
01:41:55BigBambiThen why not use the binary bluebrother posted the link to in the forum?
01:42:00countrymonkeyI don't know why I always forget.
01:42:12countrymonkeywhat link?
01:42:19BigBambiThere is a search button
01:44:25 Quit conando (Read error: 110 (Connection timed out))
01:45:15 Quit faeries26 ()
01:46:04gevaertsI'm going to get some sleep
01:46:39 Quit gevaerts ("ZZZZZZZZZZZZZZzzzzzzzzzzzzzzzzzzz...............")
01:46:58 Quit obo ("bye")
01:47:32 Quit star_jasmine (Read error: 110 (Connection timed out))
01:49:48 Quit lee-qid (Read error: 110 (Connection timed out))
01:51:55 Quit FOAD (Read error: 110 (Connection timed out))
01:51:55 Nick FOAD_ is now known as FOAD (
01:52:30midgeyso, it seems safari doesn't like the live irc log
01:53:04midgeyit works fine on the archived logs, but the current log gets in a loop where it keeps downloading the file over and over and never displays
01:55:12 Quit Nico_P (Remote closed the connection)
02:04:22 Join tvelocity [0] (
02:07:57 Join caysee [0] (
02:08:20DerPapstmidgey: yupp.. iirc firefox is the only browser which likes the live log. not sure about opera though
02:09:13midgeyat least on other non-supported browsers (like IE) you get a log that doesnt update
02:10:43midgeyaccording to wikipedia ( http server push should work
02:14:49countrymonkeyI wish 8456 would be committed. I know less than 3% is translated but it will let other people use rasher's page to finish the work. I don't think alex will keep the translation going.
02:22:50 Join Geeky123 [0] (
02:22:56 Quit Geeky123 (Read error: 104 (Connection reset by peer))
02:24:24DerPapstGood night everyone :-)
02:24:29 Quit countrymonkey ("CGI:IRC (Ping timeout)")
02:24:40 Quit DerPapst ("So Long And Thanks For All The Fish!")
02:30:57 Quit asdrubal ("Leaving")
02:33:50 Quit bertrik ("bye")
02:38:11 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
02:40:19 Join star_jasmine [0] (n=arwyneve@
02:40:55 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
02:41:21 Join MethoS-- [0] (
02:42:49 Join jurrie [0] (
02:56:36 Join FOAD_ [0] (
02:59:32 Quit star_jasmine ()
02:59:58 Quit MethoS- (Read error: 110 (Connection timed out))
03:02:15 Quit DaCapn (Read error: 104 (Connection reset by peer))
03:07:45***Saving seen data "./dancer.seen"
03:10:21 Quit tvelocity ("Αποχώρησε")
03:12:05 Quit FOAD (Read error: 110 (Connection timed out))
03:12:06 Nick FOAD_ is now known as FOAD (
03:32:57 Quit Thundercloud (Remote closed the connection)
03:33:59 Quit goffa (Remote closed the connection)
03:36:05 Quit DerDome (Nick collision from services.)
03:36:08 Join DerDom1 [0] (
03:36:18 Nick DerDom1 is now known as DerDome (
03:36:53 Join DrDnar [0] (
03:39:43 Join goffa [0] (n=goffa@
03:42:38 Quit homielowe (Read error: 110 (Connection timed out))
03:44:25 Join crzyboyster [0] (n=6210f026@gateway/web/cgi-irc/
03:45:56 Quit kslater (Read error: 113 (No route to host))
03:46:32crzyboysterI'm having trouble generating a Hindi voice file in rbutilqt (can it even generate Hindi voice files from espeak HI?)
03:46:57crzyboysterHow exactly should I set it up properly because right now it spits out a 3 kb file.
03:48:28 Quit DerDome ("Leaving.")
03:51:21 Quit crzyboyster ("CGI:IRC")
03:58:38 Join JdGordon [0] (
04:00:47 Join ctaylorr_ [0] (
04:02:20 Quit BitTorment ("SEGFAULT: shit!")
04:05:16 Quit saratoga ("CGI:IRC (EOF)")
04:08:46 Quit csc` (Read error: 110 (Connection timed out))
04:11:38 Quit ctaylorr (Read error: 110 (Connection timed out))
04:17:22 Join goffa_ [0] (n=goffa@
04:26:01 Quit goffa (Read error: 110 (Connection timed out))
04:27:42 Join csc` [0] (n=csc@archlinux/user/csc)
04:31:28 Quit tessarakt ("Client exiting")
04:53:17 Join keanu [0] (n=none@unaffiliated/keanu)
04:53:19 Quit MethoS-- (Remote closed the connection)
05:00:08 Quit csc` (Success)
05:07:45kfazzi get the usb logo every time, unless i hold the center button, but i'm on a fullspeed port
05:07:46***Saving seen data "./dancer.seen"
05:08:04*kfazz talks to the logs
05:20:14 Quit grndslm (Read error: 113 (No route to host))
05:20:55*keanu was hoping JdGordon was on
05:22:49 Join Buschel [0] (
05:31:42 Join webguest76 [0] (n=4644967b@gateway/web/cgi-irc/
05:32:58 Quit webguest76 (Client Quit)
05:32:59 Join webguest82 [0] (n=4644967b@gateway/web/cgi-irc/
05:33:11 Quit webguest82 (Client Quit)
05:33:21 Join webguest82 [0] (n=4644967b@gateway/web/cgi-irc/
05:34:11 Quit webguest82 (Client Quit)
05:36:02keanu...or not
05:36:49 Join homielowe [0] (
05:40:09 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
05:41:24saratogaBuschel: i saw your new patch
05:41:34saratogadoes it still use 24MHz as the default clock rate?
05:43:06 Join FOAD_ [0] (
05:45:51Buschelah, hi :)
05:45:57Buschelyes, it still uses 24MHz
05:46:10Buscheljust reading your reply
05:47:04saratogayou'll have to go through amiconn and the other ipod developers to get the 24MHz core clock change committed
05:47:13saratogabut i can commit the device disable changes sooner
05:47:52saratogatheres actually a long history of people arguing over what the core clock speed should be
05:48:14Buscheli had this discussion with them a few weeks ago :)
05:48:20saratogafor a while, people wanted to increase the boosted speed to improve performance, and there were some bitter arguments on filespray about it
05:48:47saratogaso people will be cautious about it i think
05:48:48Buschelwell, it's a pitty that my GUI-boost didn't work fine...
05:49:00saratogawhat was the issue with that anyway?
05:49:47Buschelit did not boost on scrolling lines. this caused some interference...
05:50:18saratogai always liked the idea of just tying boost to the back light
05:50:21Lloreansaratoga: No, they wanted to raise the base speed to improve performance. Raising the boosted speed really doesn't do much, in almost all cases, I'd think.
05:50:39saratogaLlorean: that makes sense
05:50:55Buschelwe really need to find a solution to scale the CPU for low consuming codecs. FLAC may run at <20MHz, but we're wasting mA's with letting CPU run at 30MHz
05:51:33saratogai really do like the idea of adjusting the base clock speed depending on usage
05:52:00saratogawould it be feasible to lower the base clock speed whenever the back light is off?
05:52:08saratogaand keep it at 30MHz when on?
05:52:20saratogawould sidestep the GUI responsiveness and plugin issues neatly
05:53:10Buschelthis is really something for a seperate patch ;o)
05:53:16Buschelbut it would be feasible
05:53:16saratogayes i know
05:54:00 Join star_jasmine [0] (n=arwyneve@
05:54:04saratogacan you update your patch to do 30MHz now?
05:54:07saratogaI'd like to commit it
05:54:54BuschelI am doing a test build right now. there will be something quick-and-dirty like #if 0/#else/endif, wherease the #if0 contains the 24MHz variant
05:55:18saratogai think thats acceptable
05:57:28 Quit FOAD (Read error: 110 (Connection timed out))
05:57:28 Nick FOAD_ is now known as FOAD (
05:58:22Buschelmy PC is building sloow
05:58:42saratogadidn't people get you a new PC a while back?
05:58:53Buschel? when?
05:59:12saratogai thought there was some donations for an mpc developer's PC
05:59:18saratogaor was the Klemm
05:59:29Buschelthat was Klemm, I never did such things
05:59:53 Quit aneka ("KVIrc 3.2.4 Anomalies")
06:00:35saratogaah got you two confused
06:01:23BuschelKlemm took over the development and then asked the users to donate a PC
06:01:39JdGordonkeanu: sup?
06:01:48 Quit goffa_ ("Lost terminal")
06:01:54Buschelin 2000 or 2001
06:02:37keanuJdGordon, sorry, I gotta go now =/
06:04:03Buschelsaratoga: only a few minutes to go
06:06:11 Quit star_jasmine ()
06:07:33saratogaBuschel: what are you building on?
06:07:54Buschel30MHz, build ready, just running. now creating patch
06:13:18Lloreansaratoga: I do agree that I'd rather not see the UI responsiveness lowered for just two or three codecs, especially when MPC and FLAC aren't exactly the most widely used things ever in the first place.
06:13:51LloreanIf a solution can be come up with where the UI isn't hindered, then I'm all for savings there, of course.
06:14:12Buschelsaratoga: to verify
06:14:58BuschelLlorean: did you use this patch (@24MHz) and found the GUI less responsive?
06:17:54LloreanBuschel: I don't have a 5th generation iPod, the one where it will most be a problem
06:18:06LloreanBut I know there are already people who think it's not responsive enough during music playback at 30mhz
06:18:37Buschelsaratoga: btw, my PC is some x686 700MHz
06:19:33saratogaBuschel: no wonder it takes you a while to build
06:19:49saratogai could probably get you SSH access to a faster system if you'd like
06:20:06saratogathough hopefulyl you have fast internet if you're going to pull builds between them
06:20:48Buschelsaratoga: I'll buy a new notebook soon as I otherwise would have to go into the basement in future. my son will have my home-office as his room.
06:21:14saratogaalright, let me know if you change you mind
06:21:28Buscheli will and thanks for your offer
06:22:26soapmaybe a second build server farm - round-robin assignment of single builds for developers who lack access to a decently fast machine?
06:23:06soapbut that's probably overkill.
06:23:50saratogai could probably just stick an old P4 in a lab somewhere and leave SSH on if anyone really needed it
06:24:13saratogaprobably be good for ~2 minutes a build, maybe a bit better
06:24:45Buschelmine needs 26-28 minutes
06:27:07saratogaheh, i do about ~ 35 seconds
06:27:22Buschelclean build?
06:27:47LloreanWith ccache, sure.
06:27:52LloreanAbout 40 seconds here.
06:28:08BuschelI need a new PC
06:28:27LloreanAs much as 4 minutes in a completely clean environment with ccache emptied, iirc.
06:28:36saratogai don't have ccache
06:28:39Buschelfunny thing is: my wife also wants a new one, so I should take the chance ;-)
06:29:14saratogadoes it make a big difference
06:29:31Lloreansaratoga: It can under a lot of circumstances.
06:30:00saratogaBuschel: for the commit message, how much should I say this improves battery life?
06:30:04soapdepends on your build habits. If the same files keep getting compiled over and over again w/o revision, it will save you some serious CPU.
06:30:39Buschelsaratoga: let me check my Excel file
06:30:47Lloreansaratoga: you could always just say "significant battery life improvement"
06:31:11LloreanIf you quote a number, someone's always going to say "I don't get that much" and turn out to either have wrong math, or just unique circumstances, but they'll file a bug report nonetheless. :-P
06:33:16Buschelabout this (when not taking into account the MPC/FLAC measurements): iPod 4G +~50%, 5G +15-20%, C200 +15%, E200 +20%, H10 +25%, nano +30%
06:33:24 Quit z35 ("Leaving")
06:33:31Buschelbut for H10/nano there were only few measurements
06:33:35 Join ol_schoola [0] (
06:34:45 Quit JdGordon (Remote closed the connection)
06:35:01Buschelbut you should better say "significant improvement" like Llorean suggests
06:35:06 Join JdGordon [0] (
06:39:37saratogaok commited
06:39:51Buschellet's see
06:40:55saratogathis reminds me, i still need to go through the Sansa's power management chip and squeeze some more battery life out of it
06:40:56 Quit homielowe (Read error: 110 (Connection timed out))
06:41:32LloreanAnd get charging working properly?
06:41:42saratogaBuschel: do you mind if I close your FS entry now?
06:41:45Buschelit's quite a pain to play around with power management. you'll need lots of time to find out if it's stable...
06:41:47saratogaor do you want to do the honors
06:42:05saratogaLlorean: yeah that too, though its pretty simple to do
06:42:30Buschelsaratoga: just close it. I might open up a new one for the 24MHz and PCF-fiddling ;)
06:43:37Buschelit's a good feeling to see the work of several weeks finally committed :)
06:46:19saratogai'm just amazed to see this
06:46:27saratogai didn't htink we'd ever get anywhere with the ipod runtime
06:47:38saratogadamn it broke some sims
06:48:02Lloreansaratoga: Improved runtime on iPods, USB support on the horizon... What next, someone cracking the Zune of the iPod Classic?
06:48:56*jhMikeS has a probable answer for shorter run times on PP - the tick runs too slow (not a joke)
06:48:59 Quit ol_schoola (Read error: 104 (Connection reset by peer))
06:49:02Buscheldamn, the debug menu...
06:49:02 Join ol_schoola [0] (
06:51:10jhMikeSe200 fell ~4s behind in 3 minutes which is similar for H10.
06:51:52saratogaBuschel: what does that check for DEBUG_CANCEL even do?
06:52:02saratogajhMikeS: fell behind what?
06:52:26Buschelsaratoga: there's a #ifndef SIMULATOR missing around dbg_pcf()
06:53:26jhMikeSsaratoga: compared with the clock on my desktop PC
06:53:41saratogaBuschel: line 1366 in debug_menu.c?
06:54:05saratogajhMikeS: thats with this latest commit?
06:54:44Buschelyes: #if (CONFIG_RTC == RTC_PCF50605) && !defined(SIMULATOR)
06:54:53jhMikeSno. I was just seeing how accurate ticks are and they're pretty badly off (comparing others to some adjustments on S)
06:55:23*jhMikeS does some triple checking on something
06:57:32 Quit alienbiker99 (Read error: 110 (Connection timed out))
07:03:02saratogajhMikeS: the 5024 can adjust the clock if its running too fast or too slow
07:04:44 Join z35 [0] (
07:07:32LloreanjhMikeS: That'd account for another 2% or so runtime error?
07:07:47***Saving seen data "./dancer.seen"
07:08:48*Buschel hides away
07:09:02Buschelagain build error for PCSIm
07:10:06saratogayeah i'm fixing it
07:10:33saratogathis time i'm even going to compile a sim before i decide its fixed!
07:10:48jhMikeShmmm...did a different check and the theory seems to be disproven. Woulda been neat if it were that simple. :)
07:10:55Buschelthe line should be added to 2383... I am doing the same right now
07:11:55saratogaBuschel: already commited
07:12:12Buschelwell, you're building faster :)
07:12:19*jhMikeS hates jumping the gun so badly like that...argh
07:12:33Buschelmy fault, don't bother
07:21:53saratogaok all green
07:21:56saratogaand i'm going to bed
07:23:47 Quit XavierGr (Nick collision from services.)
07:24:00 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
07:27:10 Quit saratoga ("CGI:IRC (Ping timeout)")
07:33:57 Quit XavierGr (Nick collision from services.)
07:34:10 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
07:36:58 Join mewshi [0] (
07:38:11 Join [potato] [0] (
07:42:19mewshiSo... I got my girlfriend to try rockbox ^_^
07:43:09mewshiand... um...
07:43:14mewshishe has a few complaints
07:47:52 Quit daurnimator (Connection timed out)
07:48:19 Join simonrvn_ [0] (i=simon@unaffiliated/simonrvn)
07:49:23 Quit simonrvn (Nick collision from services.)
07:49:28 Nick simonrvn_ is now known as simonrvn (i=simon@unaffiliated/simonrvn)
07:51:06[potato]me thinks theres a feedback thingy for complaints on the rockbox site.
07:53:58 Quit gtkspert (
07:53:58 Quit mokkurkalve (
07:53:58 Quit maddler (
07:54:02 Quit DrDnar (Read error: 110 (Connection timed out))
07:55:17NJoinmokkurkalve [0] (
07:55:25mewshiJust wondering, how hard would it be to put rockbox into standby?
07:55:55NJoingtkspert [0] (
08:00:49[potato]wouldn't that be a waste of batteries?
08:04:30soapDepends on your point of view. If the point of batteries is to power your device - and the point of your device is to give a satisfying user experience - and the reboot cuts into your satisfaction - I'd say standby mode would be a good usage of battery.
08:05:31soapAlso, depending on your usage pattern - in and out of standby might actually prolong battery life if you are frequently powering off and on a HDD player.
08:08:12[potato]i didn't think of it that way.
08:08:22mewshiYeah :)
08:08:32mewshimy girlfriend would like that, like the original firmware on her player has
08:08:49soapThat being said, and no offense to mewshi intended - but I think it is rather silly to think standby mode would be A-Easy and B-Valuable and C-Not done.
08:09:15[potato]where can i go to check on the status of the USB stack?
08:12:38 Join linuxstb_____ [0] (
08:12:42 Nick linuxstb_____ is now known as linuxstb (
08:12:45scorchebrowse the logs :)
08:12:54[potato]thnx =]
08:12:56kfazzso... trying to compile an interrupt enabled bootloader for e200. i keep getting undefined references to current_tick in thread.c
08:13:06kfazzwas looking at the new gigabeat s one for reference
08:13:40 Quit linuxstb____ (Read error: 113 (No route to host))
08:14:53 Quit spr0k3t (Read error: 104 (Connection reset by peer))
08:16:06 Join spr0k3t [0] (
08:18:29 Quit nicktastic (Connection timed out)
08:18:48[potato]sweeeett. thnx soap ^^
08:19:39 Part [potato]
08:21:34Buschelsoap: so, finally we did it :) Thanks a lot for all the time and effort you've put in supporting the battery patch!
08:22:51midgeythis probably warrants a MajorChanges addition
08:25:06 Join n1s [0] (
08:42:18 Join Buschel_ [0] (
08:42:18 Quit Buschel (Read error: 104 (Connection reset by peer))
08:44:56 Join xushi [0] (n=xushi@unaffiliated/xushi)
09:03:12 Join Buschel [0] (
09:03:13 Quit Buschel_ (Read error: 104 (Connection reset by peer))
09:03:51 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
09:07:31mewshiWell, I mean, maybe like an option to enable! ^_^
09:07:50***Saving seen data "./dancer.seen"
09:08:58 Quit perrikwp (" ajax IRC Client")
09:09:53 Join Llorea1 [0] (
09:10:14 Quit Llorean (Nick collision from services.)
09:10:18 Nick Llorea1 is now known as Llorean (
09:13:01mewshiMew :\
09:14:30 Quit midgey ()
09:21:29 Quit morrijr (Remote closed the connection)
09:35:07 Join miepchen^schlaf [0] (
09:38:21 Join J [0] (
09:55:33amiconnjhMikeS: (re the PP timer) I wonder what effect you're seeing. I just checked both Mini G2 and H10 small. The timer perfectly matches "wall clock" there
09:56:45amiconn(difference < 0.1s over 5 minutes, i.e. precision limit of hand-operated stop watches)
09:59:31amiconnBuschel: around?
10:05:12 Join fehmicans [0] (n=canavar@
10:09:41 Join mewshi__ [0] (
10:11:11amiconnHow did you determine the per-target minimum values for voltages and DEV_ENx?
10:11:58 Quit mewshi (Read error: 104 (Connection reset by peer))
10:12:05BuschelDEV_ENx via massive testing. voltages via different steps:
10:12:25Buschela) switch off and see whats happening (e.g. no audio playback)
10:12:44Buschelb) reducing step by step until same effect occurs
10:13:13Buschelc) reading several specs (e.g. different WM-specs) to find out what typical min/max values are
10:13:47amiconnWell, determining the voltage values via trial & error is somewhat nasty. It might work on *your* G5.5, but that doesn't mean it will work on *all* G5.5s
10:13:49Buscheld) of course reading out the unchanged default values when starting up
10:14:12amiconnI might try to add some settings for Mini G2
10:14:20Buschelthere are 3 values changed: a) the 2nd core voltage set to 0. it is not used
10:14:54Buschelb) WM-codec: 2.5V is in normal range for all WM-specs I've read. It even runs at 1.8V on my one
10:15:00amiconnYes, and lcd voltage plus wmcodec voltage lowered
10:15:31amiconnThe lcd voltage change is somewhat questionable, although it might have an advantage if it's actually the bcm voltage
10:15:44Buschelc) 2.6V for LCD also did not show any negative effect. also not for other users
10:16:29amiconnMy point here is that the actual panel voltages are significantly higher, so that the (any) lcd controller has to do upward conversion anyway
10:16:30 Join ender` [0] (
10:16:38Buschel-> it also works at 2.5V, beyond it there are artifacts visible (tested down to 1.8V)
10:17:16amiconnSo lowering the input voltage might have no effect, or even a bad one, if the lowered input voltage causes the internal converters to run at lower efficiency
10:19:13amiconnAnd regarding the (not committed) change to 24 MHz for unboosted - I'm worried that it would cause complaints about sluggish ui operation. And I don't think we should boost on ui activity - that'd be an extreme measure
10:20:18amiconnThe 'mainstream' codecs boost a little even at 30MHz, so keeping that frequency shouldn't have a negative impact on runtime for those
10:20:41amiconnAnd finally, regarding your extremely slow builds: are you using cygwin?
10:21:57 Join Buschel_ [0] (
10:22:02 Quit amiconn (Nick collision from services.)
10:22:08 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
10:22:15*Buschel_ will kick his USB-WLAN adapter out of the window!
10:22:34amiconnGot my latest remarks/questions?
10:23:30Buschel_yes, just read the logs.
10:23:34Buschel_cygwin, btw
10:24:30 Quit Buschel (Read error: 104 (Connection reset by peer))
10:24:32amiconnYeah, cygwin is slow. If your PC has enough RAM, using a linux VM would allow for way faster builds
10:24:58Buschel_you are right about the mainstream codecs, but wouldn't it be perfect advertisement to support high quality codecs like FLAC and have better runtime with it on top?
10:25:06 Join conando [0] (
10:25:49amiconnMaybe, but not if the price is sluggish ui.
10:26:14Buschel_maybe we really could implement something like 30Mhz/80MHz during GUI acitivty and 20MHz/80MHz when GUI is idle?
10:26:43Buschel_we could even choose 40/80 and 12/80
10:26:48LloreanI'm gonna hafta agree that UI performance shouldn't be sacrificed for runtime. The OF has a pretty snappy UI, and we should be able to do similar.
10:26:50amiconnFor one, I wouldn't go below 24MHz at all
10:27:14amiconnThat seems a 'natural' clock frequency on PP
10:27:50jhMikeSamiconn: I was measuring from the run time screen (a bad idea because it uses deltas) so i measured from a reference tick and all was fine.
10:28:03amiconnjhMikeS: Ah.
10:28:24amiconnA good way to check timer correctness is using the stopwatch plugin, btw
10:28:30amiconn(since it uses the timer tick)
10:28:41 Join robin0800 [0] (
10:28:49Buschel_ok, then 40/80 and 24/80: 40 to have better responsiveness, 24 to save current for efficient codecs
10:29:20amiconnWell, it could use the idle/normal switching mechanism
10:29:27Buschel_the overall runtime should not be influences by the clock needed for GUI interaction
10:29:32amiconnSwitch to idle in wps
10:30:11jhMikeSBuschel_: 2nd core isn't used? ??
10:30:23amiconn2nd core _voltage_
10:30:29Buschel_jhMikeS: no, the second voltage ist not used.
10:30:34amiconnThat's an output of the pcf iiuc
10:30:48Buschel_jhMikeS: you can configure the PCF to switch between low and high supply voltage
10:31:32 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
10:31:36Buschel_but lowering the 1.2V even is instable @24MHz
10:31:50*jhMikeS just thinks "COP" when hearing "2nd core" :)
10:31:55Buschel_so, there will be no savings when playing around with it
10:32:34*amiconn wonders what runtime he will get on his small H10
10:32:49amiconnThe beast seems to have a somewhat worn-out battery
10:34:04jhMikeSisn't that an easy replacement?
10:35:21amiconnTechnically easy, yes, even dead simple
10:35:26Buschel_jhMikeS: seen my last comment in fs?
10:36:01jhMikeSnot yet
10:36:49Buschel_maybe this spinlock-5G-HDD-thrash stuff is connected to ata.c and the special handling which was built in to support large HDDs?
10:37:35HorschtBuschel, does PCFv06 and 7 still need testing for issues? I assume they do, since they seem to do stuff that's not inlcuded in DDvXX
10:38:04Buschel_Horscht: yes, feel free to test them. I am also doing further tests here
10:38:20jhMikeSI'm thinking it may have to do with that. What I'm not clear on is does this happen if no other disk activity is present like some bg scan?
10:38:42Buschel_ask Horscht, he tested the patch
10:39:20HorschtjhMikeS, my test results from that FS entry had the following setup/procedure
10:39:27jhMikeSOne major thing is that spinlock maintains thread position in the list. Right now, a woken thread is placed farthest in the list from the current one (insert last). The patch places it closest to the current one.
10:40:14Buschel_btw, Horscht: you said the HDD in the good case is switched off after ~70s, disk activity (reading) is finished after ~20s. Why is there a delay of 50s before HDD is switched off? my 5G 30GB switches off the HDD ~1s after reading is finished.
10:40:48 Quit linuxstb ("ChatZilla 0.9.80 [Firefox]")
10:40:56Horschti started my ipod, waited for the HD to shut off again. then navigated to the artist in the database. Again, I waited for the Hd to shut off, then qeued all the songs and measured the time
10:40:58 Join pixelma [0] (n=Marianne@rockbox/staff/pixelma)
10:42:00HorschtBuschel_, the first 20 seconds it is continously reading from the disk (you can hear the disk activity noise quite good). after 20 seconds it's basicaly just spinning, with minimal disk activity all 4 or 5 seonds or so
10:42:29Buschel_sounds strange...
10:42:50 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
10:42:53jhMikeSHorscht: care to check my small patch in FS #8568? Only the bin needs rebuilding.
10:42:53HorschtI wish i had some way to record the sound
10:43:25Horschtin a few minutes. I am having breakfast at the moment :)
10:43:40 Quit jhulst (Remote closed the connection)
10:45:21amiconnBuschel_: Now I really wonder how you tested all this. E.g. GPO32_EN has a rather weird value on my mini
10:49:33 Nick fxb__ is now known as fxb (
10:50:08HorschtjhMikeS, I am now compiling SVN with your diff applied. I will post the results to FS, once I have them
10:51:03jhMikeSHorscht: thanks
10:52:19 Join gtkspert_ [0] (
10:55:30Buschel_amiconn: that's why I dropped it and only fiddled out GPO32_VAL
10:55:45 Quit xushi ()
10:58:42Buschel_when unsetting "#define MAX_PHYS_SECTOR_SIZE 1024" in config-ipodvideo.h for my 30GB 5G it seems to work fine. it even does not one strange thing i always had before: HDD is reading (WPS shows HDD on), then finishes reading (WPS shows HDD off). Then it was switched on shortly and off again (WPS shows this + you can hear this).
10:59:35Buschel_now it simply switches off after reading. no more additional on->off anymore
10:59:48 Join Lear [0] (i=chatzill@rockbox/developer/lear)
11:03:46amiconnThe apparent short disk activity shortly after the actual activity ended is the ata sleep routine
11:04:01amiconnIt applies to all hdd based targets
11:04:36jhMikeSamiconn: will it work to use the alternate sector size on another player for testing purposes?
11:04:38amiconnIt's just not always visible on targets with no physical hdd led due to status bar update frequency
11:04:58amiconnjhMikeS: Not by default, but you can force it to
11:05:20amiconnThe actual sector size on which the amount of transferred data is based isn't different
11:05:31jhMikeSso it won't corrupt the FAT or anything?
11:05:47amiconn(i.e. the ata protocol always calculates based on 512-byte sectors)
11:06:28amiconnIt's just that the MK8011 *requires* transfers to happen in multiples of 2 sectors, with an even start sector number
11:06:44 Quit gtkspert (Read error: 101 (Network is unreachable))
11:07:13amiconnYou can easily force that instead of letting the ata driver determine it from the identify info
11:07:51***Saving seen data "./dancer.seen"
11:08:07jhMikeSguess I can try on H10 which is about as close as I can get to 5g in that way
11:08:26HorschtjhMikeS, FS comment added.
11:09:45Horschtah... wait. crap. it's not verified against same stock SVN but 16243
11:09:49amiconnIt should be sufficient to set the MAX_PHYS_SECTOR_SIZE macro, and set phys_sector_mult = 2; in ata_init() instead of evaluating identify_info[106]
11:10:03 Join gtkspert [0] (
11:11:30jhMikeSHorscht: it should be applied on r16197 or later
11:12:15HorschtI applied it to the latest, but the measurements of the stock SVN build were made with 16243
11:12:17jhMikeSI suppose to imitate 5.5G it should be 1024
11:12:53Horschti'll just compile latest SVN unpatched to verify
11:13:27 Join mf0102 [0] (n=michi@
11:14:06jhMikeSHorscht: It very strange since very little if anything in the way of waiting should be happening...but...the ATA thread itself could be starving the disk access by not running enough since it does occasionally lock the mutex.
11:14:35Horschtjust to verify the small decrease in buffering time was actualy caused by your patch and not a change in SVN, namely Buschel_'s commit
11:16:00 Nick fxb is now known as fxb__ (
11:16:27jhMikeSI wonder what would happen if the ATA thread were changed to PRIORITY_REALTIME
11:16:53Horschtsorry, I am not able to help coding wise. I have no coding experience at all. All i can help with, is testing
11:17:30jhMikeSI can give a patch really quickly
11:17:54Buschel_ok, I'll be off for a while (moving my PC and all stuff to the basement)
11:17:58 Quit Buschel_ ()
11:17:59HorschtI will be happy to test it
11:21:55jhMikeSHorscht: just paste this into a blank text file:
11:22:34 Quit gtkspert_ (Read error: 101 (Network is unreachable))
11:26:42Horschtwill try that
11:33:11 Quit amiconn (" HydraIRC -> <- \o/")
11:35:11HorschtjhMikeS, is that patch to be compiled toghether with the one from FS?
11:35:20 Quit mewshi__ (Remote closed the connection)
11:35:22jhMikeSone thing I can safely say is running it like that make buffering _Very_ slow
11:35:36jhMikeSHorscht: yeah, why not. shouldn't matter.
11:35:59*jhMikeS tried MAX_PHYS_SECTOR_SIZE on H10
11:41:33 Join mewshi__ [0] (
11:44:12 Join GodEater [50] (n=bryan@rockbox/staff/GodEater)
11:45:59 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
11:49:54 Join tvelocity [0] (
11:50:02 Quit GodEater (Read error: 113 (No route to host))
11:50:14 Join GodEater [50] (n=bryan@rockbox/staff/GodEater)
11:50:19jhMikeSone piece of nastiness in ata.c is sleeping while holding a mutex :\
11:54:31HorschtI am now rerunning all my test against same SVN base
11:56:04 Quit GodEater (Read error: 104 (Connection reset by peer))
11:56:06 Join GodEater_ [50] (n=bryan@rockbox/staff/GodEater)
11:57:17 Quit GodEater_ (Read error: 104 (Connection reset by peer))
11:57:19 Join GodEater__ [50] (n=bryan@rockbox/staff/GodEater)
11:57:51 Join desowin [0] (n=desowin@atheme/developer/desowin)
11:59:16 Join GodEater [50] (n=bryan@rockbox/staff/GodEater)
11:59:57 Quit GodEater__ (Read error: 113 (No route to host))
12:00:53 Quit GodEater (Client Quit)
12:04:12 Quit mewshi__ (Read error: 104 (Connection reset by peer))
12:04:15 Join mewshi__ [0] (
12:04:34 Quit DataGhost (Nick collision from services.)
12:04:42 Join DataGhost [0] (i=dataghos@unaffiliated/dataghost)
12:09:12 Join Domonoky [0] (
12:09:55 Join Thundercloud [0] (
12:11:29 Join K4rP4D [0] (
12:12:25 Join tessarakt [0] (
12:13:45 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
12:17:09linuxstbHas anyone else seen a kernel message like "scsi scan: INQUIRY result too short (5), using 36" when using the Rockbox UMS mode? (this is on my ipod Color)
12:21:38 Quit mewshi__ (Read error: 104 (Connection reset by peer))
12:21:49BigBambilinuxstb: Yes, I see the same thing with an E260
12:23:08 Join mewshi__ [0] (
12:32:03 Join Rob2222 [0] (
12:32:54 Quit mewshi__ (Connection reset by peer)
12:34:28linuxstbHmm, seems my disk is now corrupt in some way - in both the OF and Rockbox UMS modes I get errors, but Rockbox seems to be able to read the filesystem fine...
12:35:39 Quit mf0102 ("Verlassend")
12:36:02 Join Buschel [0] (
12:36:39HorschtjhMikeS, results are up
12:37:00 Join roxfan2 [0] (
12:37:53 Join mewshi__ [0] (
12:40:14 Join waldo [0] (
12:40:50 Quit mewshi__ (Remote closed the connection)
12:43:01 Join mewshi__ [0] (
12:43:05jhMikeSHorscht: thanks. those results look very odd.
12:46:01desowinoh pure god, #8562 appears to be working :-)
12:46:21 Quit waldo (Read error: 104 (Connection reset by peer))
12:46:30HorschtjhMikeS, i thought so too
12:47:47 Join Imaginativeone [0] (
12:48:39Horschtbut still, spinlock_revert gives quite an improvement
12:51:20 Join austriancoder_ [0] (n=austrian@
12:52:04jhMikeSindeed and for no good reason I can possibly imagine at the moment
12:52:23 Quit roxfan (Read error: 110 (Connection timed out))
12:52:52 Quit mewshi__ (Connection reset by peer)
12:53:02 Join waldo [0] (
12:53:29desowinlinuxstb: the INQUIRY result too short happens on sansa e280 as well (on init)
12:55:52linuxstbdesowin, BigBambi: OK, thanks - I'll add a note to flyspray
13:00:29 Join mewshi__ [0] (
13:00:52 Join homielowe [0] (
13:02:40 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
13:05:54 Quit austriancoder__ (Read error: 110 (Connection timed out))
13:07:18 Join roxfan [0] (
13:07:56***Saving seen data "./dancer.seen"
13:09:41 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
13:10:31amiconnjhMikeS: Oww, where does it do *that*?
13:13:51amiconnIf you mean during spinup, then that's okay
13:14:26 Quit Buschel (Read error: 104 (Connection reset by peer))
13:15:18 Join Buschel [0] (
13:16:21 Quit roxfan2 (Read error: 110 (Connection timed out))
13:17:48rasherWhat do people think about leaving out a couple of the very old translations? Personally I find a 20% translated interface worse than nothing at all.
13:17:49 Quit Buschel (Read error: 104 (Connection reset by peer))
13:18:27 Quit markun (Read error: 104 (Connection reset by peer))
13:19:35 Join Buschel [0] (
13:21:16 Join karashata [0] (
13:22:21 Quit mewshi__ (Connection timed out)
13:22:26 Quit Rob2222 (Read error: 104 (Connection reset by peer))
13:23:17 Join markun [50] (n=markun@rockbox/developer/markun)
13:24:55linuxstbrasher: Until we make a release, it might be nice to leave them there, so users choose them, find out they're not complete, and (hopefully) offer to translate...
13:26:01rasherlinuxstb: This would make more sense if we actually did releases..
13:26:24linuxstbMaybe a prominent link on the homepage (in the news?) seeking translators could help.
13:29:04rasherWorth the try
13:29:54rasherMy online translator seems to have helped a good deal
13:30:00linuxstbSomething like "Rockbox is seeking translators - see [name of page] for details".
13:30:38linuxstbIs there already a page we could link to?
13:31:24linuxstbLooks good to me.
13:32:05rasherIt actually seems to work quite well
13:32:40 Part Imaginativeone
13:34:03rasherPity the Swedes are in the habit of changing the website without committing
13:36:56linuxstbLooking at this page -"> - does it make sense for "none" to be editable in the first item?
13:38:41 Nick fxb__ is now known as fxb (
13:38:46rasherlinuxstb: it isn't editable
13:39:07linuxstbTrue... ;)
13:40:13 Join miepchen^schlaf [0] (
13:40:31linuxstbAlso, what happens if there is no voice string - does it default to the displayed text?
13:41:12linuxstb(I'm talking in general, not your interface)
13:41:46rasherIt doesn't. Some strings don't have voice set at all. It's hard to make an automated decision
13:42:10rasherExcept I guess you could do something along the lines of "does english.lang set a voice for this string? If so, copy the one from the displayed string"
13:42:18 Quit karashata (" HydraIRC -> <- In tests, 0x09 out of 0x0A l33t h4x0rz prefer it :)")
13:44:37linuxstbOK. Next question - should the voice strings be target-specific? Looking at catalan.lang for example, some voice strings are "*:", whereas the displayed text is only defined for some targets.
13:44:57rasherThat does sound odd - which ones?
13:45:07linuxstbe.g. LANG_PM_RELEASE
13:45:15linuxstbBut there are quite a few
13:45:35 Join merbanan [0] (n=banan@
13:45:59rasherlinuxstb: that looks like a mistake
13:46:30rasherAnd a bug in genlang probably, because english.lang has a target-specific voice for that
13:48:07linuxstbSo to fix, would you need to mirror the <dest> structure, or go back to english.lang?
13:48:27*linuxstb wonders if there's a Catalan TTS engine
13:48:35pixelmayes, and as been said earlier... (when discussing the order)
13:48:38rasherI'd say genlang should verify the structure of the <voice> against that of the English one
13:48:54rasherObviously not the contents as it does with <source>
13:49:29n1scan I copy a rockbox.gigabeat to my 'beast using MTP?
13:50:20 Nick austriancoder_ is now known as austriancoder (n=austrian@
13:50:29linuxstbn1s: I guess so, as MTP is the only way to transfer files...
13:50:54n1sOk, I'll try it out
13:51:30 Join roxfan2 [0] (
13:51:59rasherSo basically the problem is that genlang fails to check the structure (targets) of the <voice> part?
13:52:45pixelmaAFAIK yes
13:52:50rasherpixelma: have you seen">
13:54:03pixelmano, not yet but I know that some of the new strings aren't translated yet
13:55:22 Join mewshi__ [0] (
13:55:27 Join Axio [0] (
13:55:37rasherIt also hilights other mistakes that genlang doesn't/can't catch
13:55:54rasherSuch as missing voice strings where english.lang defines one
13:58:15 Quit K4rP4D ("Leaving")
13:59:03linuxstbrasher: Should deprecated strings be shown in your web interface?
13:59:35rasherlinuxstb: I just silently handle that for the user (hopefully!)
13:59:40rasherThat's the idea, at least
14:00:22linuxstbWhat do you mean? They are shown now, and I'm not sure why...
14:00:47*n1s boots rockbox on the beast! yay :)
14:01:46rasherlinuxstb: Maybe I remember wrong
14:01:58 Quit roxfan (Read error: 110 (Connection timed out))
14:03:21rasherYeah, that's a mistake..
14:03:35linuxstbAre they also being included in the stats?
14:03:44linuxstb(i.e. Current Translation Status) ?
14:04:16rasherThey're included in "Changed desc" and "Changed source", but not in the percentage
14:04:29 Join Febs [0] (
14:04:30rasherThat only counts missing strings
14:06:24pixelmarasher: hmm... ok - these did not show up when using genlang. Nice list, helps me to find them.
14:07:22rasherpixelma: Indeed - there are also some things which would be silly to flag with genlang (such as English and translation being the same, which in many cases is not a problem, but might be a missed string)
14:08:29 Join gtkspert_ [0] (
14:09:16 Join runit [0] (n=79293ab6@gateway/web/cgi-irc/
14:10:28rasherlinuxstb: I was for some reason only checking for upper-case "DEPRECATED", should be fixed now
14:10:37 Join gevaerts [0] (
14:10:41pixelmarasher: that's true, almost all "same as in english" are either coincedence (like "pause") or on purpose (not translating button "names") or not translatable ("equalizer" is used here too)
14:10:43 Quit mewshi__ (Read error: 104 (Connection reset by peer))
14:10:49 Join mewshi__ [0] (
14:11:22 Join Horschti [0] (
14:11:41 Join axionix [0] (
14:12:04 Quit Horscht (Nick collision from services.)
14:13:24 Quit runit (Client Quit)
14:15:22 Join asdrubal [0] (
14:20:32 Quit gtkspert (Read error: 101 (Network is unreachable))
14:22:47 Nick ctaylorr_ is now known as ctaylorr (
14:23:28 Join BitTorment [0] (
14:27:49 Quit axionix_ (Read error: 110 (Connection timed out))
14:28:20 Quit mewshi__ (Connection timed out)
14:29:53 Nick Horschti is now known as Horscht (n=Horscht@xbmc/user/horscht)
14:39:36 Quit Horscht ("electromagnetic radiation from satellite debris")
14:40:11 Join Honkboy [0] (
14:42:34amiconnDomonoky: Check out my new commit...
14:43:00Domonokyi have just seen it, will try it.. :-)
14:46:23*gevaerts is updating his svn tree to check if USB still works with the power-saving changes
14:49:46 Join FOAD_ [0] (
14:51:59gevaertsdesowin: can you access the sd card slot properly over usb (i.e. does it even appear in the dmesg output) ?
14:54:25*gevaerts noticed he doesn't even check dmesg output anymure before mounting his sansa...
14:55:31gevaertslinuxstb: did you do anything on the rockbox side when you got disk corruption ? Did you get the USB screen ?
14:56:37 Quit Nico_P (Remote closed the connection)
14:56:43gevaertsThe USB screen not showing is most probably not a purely cosmetic issue. I'm pretty sure it also means stuff like dircache and database updates are not stopped
14:57:16linuxstbgevaerts: I occasionally get the USB screen, but I don't think I did in that test. But I also didn't get it in the subsequent test when things worked fine...
14:59:07gevaertslinuxstb: if it was caused by background disk access by some other rockbox thread, it's bound to be somewhat unpredicable
15:01:35 Part Honkboy ("Leaving.")
15:02:38 Join Geeky123 [0] (
15:02:43 Quit Geeky123 (Read error: 104 (Connection reset by peer))
15:04:32 Join GodEater [50] (n=bryan@rockbox/staff/GodEater)
15:04:38gevaertslinuxstb: I uploaded a new version that should fix the connection screen issue
15:07:10 Quit FOAD (Read error: 110 (Connection timed out))
15:07:11 Nick FOAD_ is now known as FOAD (
15:07:59***Saving seen data "./dancer.seen"
15:08:36 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
15:09:03gevaertsrasher: it waited for 500ms after connect detection for any usb activity. That was apparently near the normal delay, which is why some people didn't have the issue. Now it waits for 1 second.
15:11:18bluebrotherDomonoky: why is rbutil a "householding" utility while it's been "housekeeping" in the about box for ages? ;-)
15:11:36bluebrotheroh, and I managed to build rbutilcli without QtGui :)
15:11:47gevaertsWe could try to do some clever tricks like using SCSI_ALLOW_MEDIUM_REMOVAL to see which devices are actually in use by the host, but that's probably being too clever
15:13:26Domonokybluebrother: nice.. householding / keeping.. just a mistake... i coded too much yesterday :-)
15:15:32bluebrotherhehe ;-)
15:15:39bluebrotherhow's progress on the cli?
15:16:20bluebrotherI made some changes to rbutil for building without QtGui.
15:17:12Domonokyat the moment i try to add support for sapi4 :-)
15:17:23bluebrotherhopefully I'll commit them later today.
15:17:35bluebrotherhow good I'm on linux ;-)
15:18:57desowingevaerts: /dev/sdb2 appears no matter if sd card is in or not, can't access it at all
15:19:25rasherdesowin: which distro and which kernel?
15:19:33desowingentoo 2.6.24
15:19:40rasherI'm getting no sd card in Windows
15:20:04rasherI do get it in Linux though - Debian 2.6.24
15:20:16 Join mewshi__ [0] (
15:20:43gevaertsdesowin: I'm guessing /dev/sdb1 is your internal flash. If so, the sd card should be /dev/sdc, not /dev/sdb2 (that would be your firmware partition). Do you see anything about /dev/sdc ? It should print out one or two lines about it.
15:20:49desowinfdisk says exactly the same, no matter if card is in or not
15:20:54amiconndesowin: sdb2 ? Iiuc the card slot should be another scsi device
15:21:07desowinahh, another, let me check
15:21:16gevaertsrasher: I know. That's my next task.
15:21:30desowinno other device at all
15:21:49desowinsdb1 is internal flash, yes
15:22:22amiconnThen sdb2 is the hidden partition iiuc
15:22:53rasherThere should be a second one, most likely sdc
15:23:00desowinthere isn't
15:23:13rasherDmesg output on pastebin?
15:23:20gevaertsdesowin: OK. That matches what infirit sees on his e280. So why does this work on c250 and e260, but not on e280 ?
15:23:36desowinI'll compile the v3 before pastebin
15:24:10*bluebrother leaves for some more stuff to try for QtGui splitting
15:24:14 Quit bluebrother ("bbl")
15:24:17 Quit mewshi__ (Remote closed the connection)
15:24:28 Quit fehmicans (Remote closed the connection)
15:24:32 Join mewshi__ [0] (
15:24:34Buschelhi, just saw that my device disables change has one usecase where USB devices are still enabled but not used: attaching USB when MENU is pressed (e.g. for charging), then detach USB. In this case the devices are still enabled.
15:25:46BuschelAs I cannot code on my notebokk right now. Can anyone try to add to usb-fw-pp502x.c (the else-patch)?
15:26:02pixelman1s: with the "parsing features.txt" patch, the resulting options are then called like the feature? E.g. "multivolume" and not "HAVE_MULTIVOLUME", I wanted to introduce such a macro now and want to get the naming right.
15:26:59rashergevaerts: Apropos what Buschel just said, there's a problem - if I hold select when plugging I don't get the USB screen, yet the sansa presents the drives over usb
15:27:26gevaertsrasher : I just noticed that as well
15:28:00desowinoh, needed to "hand-reboot" sansa after updating rockbox
15:28:34 Join DerDome [0] (
15:28:52gevaertsdesowin: that should be solved with v3
15:30:39*rasher installs a modified bootloader for good measure
15:31:31n1spixelma: correct
15:32:03gevaertsdesowin: can you also do that for the OF ?
15:32:46pixelman1s: nice, thanks. I always have to keep an eye on upper/lower case?
15:34:22rasherecho foo |sed 's/\(.*\)/HAVE_\U\1/' <−− converts foo to HAVE_FOO. \U might be a GNU extension though
15:34:58desowinhmm, it might be my fault due to kernel, I'll reboot into ubuntu and check
15:35:12 Quit desowin ()
15:36:13n1spixelma: Think so, afaik those macros are case sensitive
15:37:00n1srasher: it's not as simple as that HAVE_FOO is not always "foo" in features.txt
15:37:03pixelmaok... rasher: also thanks but why not getting it right the first time? :)
15:38:13 Join desowin [0] (n=desowin@atheme/developer/desowin)
15:39:03desowinindeed kernel fault
15:39:05desowinworks fine on ubuntu 2.6.24 pre-packaged kernel
15:39:05rasherpixelma: it was more a suggestion for adding into the conversion somewhere
15:39:27rasherdesowin: What fancy stuff do you have in your kernel? Might be useful to know what threw it off
15:40:09desowinI'll try hunt that down
15:40:09 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
15:40:16rasherThat looks a lot nicer
15:40:40 Quit Lear ("ChatZilla 0.9.80 [Firefox 3.0b4pre/2008020708]")
15:40:48gevaertsdesowin: I have some vague memory of there being a setting in the kernel to never scan for more than one LUN on scsi devices.
15:41:20rasherThat soundsCONFIG_SCSI_MULTI_LUN=y
15:41:57 Quit Domonoky (Read error: 104 (Connection reset by peer))
15:42:17gevaertsrasher: that would be the one
15:42:22desowinI think it's it
15:42:32 Quit moos (Read error: 104 (Connection reset by peer))
15:42:43desowinI'll check for sure
15:42:45 Join moos [0] (
15:42:56rasherGoogling for that gives a bunch of people begging their distro to turn it on
15:44:51 Join roolku [0] (
15:46:33gevaertsDo all distros put their kernel config in /boot ?
15:46:58rasherNo idea
15:47:14desowinI know that gentoo doesn't, but gentoo is retarted in many aspects ;)
15:47:47*rasher asks #fedora
15:48:02rasherNot that Fedora is "all distros"
15:48:07gevaertsWhy is it that CONFIG_SCSI_MULTI_LUN=n is only true for people who own a sansa e280 ?
15:48:31 Join Domonoky [0] (
15:48:47desowinI know why it's not set for me - because I'm ignorant enough; but dunno for others ;)
15:48:47rashergevaerts: More likely to be gentoo users?
15:49:10rasherFedora does store config in /boot
15:49:14GodEatergentoo users that are too pansy to configure their own kernels too
15:49:25GodEatergenkernel is for wusses
15:49:34desowinyou shouldn't rely on what gentoo user says
15:49:40desowin(I'm using gentoo as well)
15:49:49rasherDoes Gentoo set CONFIG_SCSI_MULTI_LUN=n by default?
15:49:56desowinit's why I have ubuntu here too
15:50:02gevaertsCONFIG_SCSI_MULTI_LUN=y probably slows down the kernel by some infinitesimal amount
15:50:42desowinwe all agree that gentoo sucks then :D
15:51:14*GodEater doesn't agree that per se, he think it's mostly gentoo users that don't really know what their doing that suck :)
15:51:33desowinuserbase is even worse
15:52:38 Join obo [0] (n=obo@rockbox/developer/obo)
15:52:40gevaertsFor some reason Windows sends no request at all to LUN 1 when in rockbox. Maybe it's related to this INQUIRY too short issue
15:52:51 Join bumbl [0] (
15:52:58gevaertsOr maybe windows and gentoo are related ?
15:53:14Domonokyamiconn: have testet the sapi4 option of sapi_voice.vbs ? for me it always says: "Error - could not get ActiveVoice object. SAPI 4 not installed?". But i installed sapi4 sdk, and other sapi4 programms work..
15:53:20desowinlike I said - don't rely on what gentooers say unless they can reproduce it on other OS
15:54:04GodEaterhave we established this config option is not set by default in gentoo ?
15:54:49 Join einhirn [0] (
15:54:50amiconnDomonoky: You need to register the activevoice control
15:54:53*GodEater seems to recall /proc/config.gz *is* present by default, and so ought to say
15:55:01amiconnThe installer "forgets" to do this
15:55:25gevaertsGodEater: not yet. But we also have no indication that infirit (who posted the issue on FS) uses gentoo
15:55:29 Join karashata [0] (
15:55:33gevaertsGodEater: On debian it isn't
15:55:46Domonokyamiconn: how do i do this ?
15:55:49GodEaterit's not turned on ?
15:55:50desowinGodEater: yes, but you still need to run default kernel
15:56:12GodEaterdesowin: what percentage of gentoo users do you think don't just use "genkernel" and leave it to it ?
15:56:24amiconnDon't remember right now. Will look it up
15:56:41GodEaterreally? I always got the impression most didn't mess with the gentoo defaults
15:56:51desowinI've seen a lot of useless bug reports from gentoo users
15:56:57gevaertsGodEater: not on the installations I've seen, and I usually run default kernels these days
15:57:32GodEaterI must admit, I never did :)
15:57:45desowinit's because when someone choses gentoo, it's to "fiddle with it", not just "use it"
15:57:54*gevaerts goes to find his card reader to see what that one sends back on INQUIRY
15:57:55desowinthat's the majority of gentoo userbase
15:59:59rashergevaerts: The default in pristine kernel sources seems to be to not set it
16:01:33 Join tvelocity_ [0] (
16:02:20 Quit mewshi__ (Connection timed out)
16:02:35gevaertsrasher: nice. We _could_ "solve" this by presenting two mass storage interfaces with 1 LUN each, but that's really dirty. Also, this is not a regression from OF anyway, so I'd say it's not our problem.
16:03:11rashergevaerts: If we behave the same as OF, I'd say it's fine to blame the kernel
16:03:38rasherOnly s390, m32r and sparc64 set it =y
16:03:46gevaertsrasher: exactly
16:03:52rasherProbably not many Rockbox users on those platforms
16:03:54desowinI agree with you guys
16:04:18rasherPlus, we know with 99% certainty what to ask for when people come with this problem, so it's no biggie anyway
16:05:32roolkuBuschel: could you have a look at FS #8581 please
16:09:29rashergevaerts: another place to look for the running kernel config is /proc/config.gz
16:10:04Domonokyamiconn: got it... thx..
16:12:42 Quit BitTorment (Remote closed the connection)
16:12:55 Join mewshi__ [0] (
16:13:21 Join BitTorment [0] (
16:14:57Buschelroolku: just did... does the mrobe need the resets same like the H10?
16:15:51 Quit BitTorment (Remote closed the connection)
16:15:57roolkuI don't know if it needs them, but judging from the diff, this is how it used to be?
16:17:18Buschelroolku: all the #else-paths do nothing (like before). from the bugtracker i understood that you already tried to use the same config than for H10, right?
16:18:27 Quit tvelocity (Read error: 110 (Connection timed out))
16:19:22 Join BitTorment [0] (
16:19:36keanuJdGordon, when you get back - when I was testing the lists using viewports patch (8457) it seems that the viewport created is behaving as if it's the entire screen size - everything outside of it doesn't refresh. however, when a screendump is taken, it's fine
16:19:53roolkuBuschel: yes, but that made it worse = black screen (I just added mrobe to the h10 if branch, so not just the resets were effected but also the dev_en etc)
16:20:36Buschelroolku: what is the exact error you have when using current svn?
16:20:42roolkuI find it really hard to determine from the patch, was has changed for the mr100
16:21:18Buschelroolku: mrobe is pp5020, not 5022?
16:21:27keanuJdGordon, screenshot:
16:21:30roolkuit hangs on the screen with the logo
16:21:41 Join lee-qid [0] (
16:21:46Buschelbacklight switching on when hitting buttons?
16:23:40roolkuBuschel: backlight is on and stays on (set to permanently on in cfg...let me disable)
16:23:50asdrubalI just ordered a 1GB sansa C240 from newegg... I'll let you guys know if it's V1 or V2....
16:23:57 Quit waldo ("Konversation terminated!")
16:24:03asdrubalI think they're clearing out stock on it because it's only 1GB
16:24:08 Quit sbhsu ("leaving")
16:28:38 Quit mewshi__ (Read error: 104 (Connection reset by peer))
16:28:59 Join mewshi__ [0] (
16:29:04linuxstbgevaerts: Your last patch includes +test_disk.c in apps/plugins/SOURCES
16:29:22roolkuif you look at line 294 it shows that the reset is same as h10
16:29:36roolkuBuschel: what happened to outl(inl(0x70000024) | 0xc0, 0x70000024); ?
16:30:10Buschelwasn't needed for H10's −− tried to use it for mrobe again?
16:30:48Buscheldid you make the changes to use the reset like H10? was this needed?
16:32:23roolkuBuschel: btw - backlight stays on regardless of setting
16:32:29gevaertslinuxstb: oops. I'll remove it again
16:32:48roolkuBuschel: not sure what you mean with your last question?
16:33:04linuxstbgevaerts: Also, a cosmetic comment on your patch - IIUC (I'm sure others will correct me), the preference in Rockbox is for #ifdef lines to be left-aligned, not indented with code.
16:33:11 Join seablue [0] (
16:33:50Buschelroolku: did you try to use the former svn code in the #elsif MROBE path?
16:33:59gevaertslinuxstb: ok. I'll fix them
16:34:35roolkuBuschel: I added | defined(MROBE_100) to the first if statement, which should achive the same?
16:34:53roolku|| defined(MROBE_100)
16:35:10Buschelwait, I'll paste it
16:38:06linuxstbgevaerts: Also, would be be possible (and desirable?) for Rockbox to present the same IDs as the original firmware? I mount all my removable devices using the /dev/disk/by-id/ devices, and these are now different with Rockbox...
16:39:25*linuxstb can't remember why Zagor chose "" and "Rockbox media player" instead of the same names presented by the OF
16:41:12roolkuBuschel: I think I tried that, but will try again. hang on
16:42:28 Quit ctaylorr (Read error: 110 (Connection timed out))
16:42:44 Join midgey [0] (
16:43:11Buschelroolku: even it that not works. leave it and we'll check for next thing...
16:43:18roolkuBuschel: doesn't make a difference
16:44:36Buschelroolku: small change,
16:46:34roolkuthat doesn't help either :( was this there before your change?
16:47:22Buschelroolku: not in that place, but below −− normally these registers are set by default.
16:47:29 Join ctaylorr [0] (
16:47:51 Join kugel [0] (n=kugel@unaffiliated/kugel)
16:48:12Buschelnext thing: add || defined(MROBE_100) to "#if defined (IRIVER_H10) || defined(IRIVER_H10_5GB)" in firmware/target/arm/adc-pp5020.c
16:50:39roolkuBuschel: that seems to have done the trick
16:50:40 Quit Buschel (Read error: 104 (Connection reset by peer))
16:50:41Domonokyamiconn: i still have problems with sapi4 and the sapi_voices.vbs script: with sapi it speaks the commands out, instead of writing to the file, and it hangs after the first command.. :-/
16:50:52 Join Buschel [0] (
16:50:58roolkuBuschel: that seems to have done the trick
16:51:08 Quit markun (Remote closed the connection)
16:51:19Buschelroolku: perfect :o)
16:51:40Buschelroolku: could you now try to delete the changes in system-pp502x.c?
16:51:42roolkuBuschel: shall I remove the reset from system
16:52:19gevaertslinuxstd: I'm not sure if that is possible. using the same id as the OF means finding out the exact model (c240 vs c250 vs ...), and the serial number
16:52:45Buschelroolku: are both ADC-channels needed or only of them?
16:52:59Buschelroolku would be interesting for the PP-registers page
16:53:01 Quit MattAndrew (Read error: 113 (No route to host))
16:53:57linuxstbgevaerts: Where are those strings defined? grep is failing me...
16:53:59Horschtrasher, question about the translations. Your Page lists "English voice" and "voice"
16:54:23Horschtshould "voice" simply be the according translation as well?
16:54:38rasherEh.. "voice" should be a translation of "English voice"
16:54:57roolkubuschel: testing - i know channel 3 gives a constant readout - we don't know what 1 and 2 do
16:56:34 Quit ol_schoola ()
16:57:03roolkuBuschel: I disabled channel 3 and it hangs again, now 2
16:58:06gevaertslinuxstb: I don;t know exactly, but as far as I can see the /dev/disk/by-id names come from vendor and device ids from identify2inquiry() in usb_storage.c, and the device serial number
16:59:05roolkuBuschel: nope needs to do both - wonder what for
16:59:53Buschelhmm, just adding these to the PP-register page. Can you create a patch? I am not on my development PC
17:00:52Buschelyou only need to add the || defined(MROBE_100) to the ADC-stuff −− no change in system-pp502x.c, correct?
17:01:46roolkuBuschel: I think I will commit that simple change - yes, just adding || defined(MROBE_100) works
17:01:54rasherHorscht: Would it make more sense if the input fields were empty rather than containing the English string?
17:02:00 Join freenod__ [0] (
17:02:00 Join maddler [0] (
17:02:14linuxstbgevaerts: What's puzzling me is that "lsusb -v" shows me "iManufacturer:" and "iProduct: Rockbox media player", but I can't find those strings anywhere in the Rockbox source...
17:02:19Buschelroolku: perfect! sorry for any inconvenience
17:02:55HorschtI think it just should say "<language> voice" instead
17:02:58roolkuBuschel: that's okay - I am glad it is something simple
17:03:21Buschelroolku: btw, interested in optimizing the mrobe registers?
17:03:42 Join FOAD_ [0] (
17:04:13rasherHorscht: Done
17:04:18soapis anyone else seeing overlapping text and " Notice: Undefined index: 0 in /usr/share/flyspray/htdocs/includes/class.flyspray.php on line 344" errors in the tracker?
17:04:34roolkuBuschel: maybe not yet - we still don't know how the button controller is attached and I don't want to accidently disable it
17:04:51Horschtsoap: no. tracker is displayed fine here
17:04:57Buschelroolku: that's a pretty good reason :)
17:04:59gevaertslinuxstb: Those are in usb_core.c, about line 420 in the patched version
17:05:11rashersoap: it does that once in a while, I've not figured out when
17:05:21rasherSloppy coding
17:05:36 Quit bumbl (Read error: 110 (Connection timed out))
17:05:41linuxstbgevaerts: Ah, no wonder grep didn't find them...
17:07:41pixelmasoap: seen that too, IIRC only in one specific tracker entry (can't remember which)
17:08:02***Saving seen data "./dancer.seen"
17:08:18rasherFlyspray also has a bit of a spasm once in a while when uploading patches
17:09:31gevaertsI can't find what I'm doing to make windows not find the SD card slot. I'll use a real tracer at work tomorrow.
17:12:13 Join axionix_ [0] (
17:16:50rashergevaerts: now to solve the resets
17:18:41Ice``Hammer it.
17:18:58 Quit Axio ()
17:19:06rashergevaerts: You have two warnings now: usbstack/usb_core.c:722: warning: unused parameter ‘in’ and usbstack/usb_storage.c:155: warning: comparison between signed and unsigned
17:19:10gevaertsrasher: I wish I knew how. They usually happen when the rockbox ums driver sees nothing wrong (i.e. the transfer went to the controller properly), but linux doesn't get the data. Maybe the tracer will help there as well
17:19:57 Quit FOAD (Read error: 110 (Connection timed out))
17:19:57 Nick FOAD_ is now known as FOAD (
17:22:23gevaertsrasher: the usb_storage warning will be gone in the next patch, but I'm not sure what to do about the one in usb_core
17:22:32Buschelhorscht: any news about the HDD-issue?
17:23:18Horschtno, latest tests and news are in the tracker.
17:23:18 Quit austriancoder (Remote closed the connection)
17:23:35Horschtso far, only reverting seems to have helped
17:23:49linuxstbgevaerts: I have a question about the usb_strings - I see that strings 0 and 1 are iManufacturer and iProduct. Strings 2 and 3 appear to "iInterface" for the USB_CHARGING_ONLY (4) and USB_BENCHMARK (3). Could I change that so that iSerial is 3, and move the others to 4 and 5?
17:23:51Buschelso, jhMikeS is also confused about the behaviour...
17:24:49 Quit axionix (Read error: 110 (Connection timed out))
17:26:28linuxstbgevaerts: I meant to say strings 1 and 2 are iManufacturer and iProduct, with the other two being 3 and 4...
17:26:46gevaertslinuxstb: that should work.
17:27:12Buschelhmm, amiconn made the basic change for supporting the 80GB HDD. maybe he has an idea...
17:27:28 Join BlackChaos [0] (
17:27:48 Quit BlackChaos (Client Quit)
17:28:34gevaertslinuxstb: usbmassbulk_10.pdf specifies what the serial number should look like on pages 9 and 10
17:30:09linuxstbgevaerts: Thanks.
17:31:00gevaertslinuxstb: actually, if I read that correctly it's a bug not to have a serial number in a mass storage device.
17:37:16amiconnAn usb device with no serial number is a violation of the specs
17:37:49amiconnBut there are many usb devices out there without such a number
17:38:18gevaertsamiconn: The serial number is optional for usb, but some device classes require one.
17:38:49amiconnYeah, I mean ums
17:39:13amiconnOn windows you can detect such devices quite easily without checking the internals
17:40:00gevaertsQuestion is, how do we get a serial number ?
17:40:02amiconnIf you attach a device first time, it gets the next free drive letter drive letter. Change that assignment to another letter
17:40:22pixelman1s: maybe I should have asked this before, but does anything speak against introducing a "multivolume" macro now, without that patch? Just making sure... :)
17:40:48n1snothing that I can think of
17:40:48amiconnThen disconnect the device, and plug it into a different port. If it gets assigned the letter you just chose, it does have a serial number
17:41:10amiconnBut if it gets the next free drive letter instead of the chosen one, it has no serial number
17:41:33 Join csc` [0] (n=csc@archlinux/user/csc)
17:41:38 Nick fxb is now known as fxb__ (
17:43:29linuxstbOn the ipods at least, the serial number appears in the flash ROM - offset 0x2034 on my Ipod COlor.
17:43:38 Join bumbl [0] (
17:44:50linuxstb(as a 64-bit little-endian int)
17:45:55linuxstbgevaerts: BTW, the usb screen is appearing 100% reliably for me with your last patch (v3)
17:48:05gevaertslinuxstb: good news. Did you get any more disk corrutption ?
17:48:16linuxstbNot that I've noticed...
17:48:45gevaertsThat's the dangerous kind
17:49:57 Quit GodEater ("Nice Scotty, now beam my clothes up too!")
17:50:04 Join mf0102 [0] (n=michi@
17:51:52 Quit tessarakt ("Client exiting")
17:56:30Domonokyamiconn: does sapi4 with the sapi_voices script work for you ? ( for me it speaks the commands, instead of writing it to file when i choose sapi4 )
17:56:47 Join MethoS- [0] (
17:56:57 Join tvelocity__ [0] (
17:59:05 Join tessarakt [0] (
17:59:15 Join Nico_P [50] (n=nicolas@rockbox/developer/NicoP)
18:04:16 Quit tessarakt ("Client exiting")
18:08:15gevaertsIt looks like sansa users are most interested in the UMS work. Does that mean that the OF disk modes for other players are better ?
18:08:56 Join petur [50] (n=petur@rockbox/developer/petur)
18:10:33krazykitother players being other portalplayer devices, or overall?
18:11:02gevaertsother portalplayer devices mainly
18:11:11scorche|shwell, the issue with the sansas is that you need to boot into the orginal firmware for USB and part of that has a long "refresh database"
18:11:47rasherPlus the issue of sdhc cards, though that's probably a minority
18:12:55 Quit axionix_ (Read error: 104 (Connection reset by peer))
18:12:59rasherDoes anyone have automatic voice-building script using espeak? Because I'm thinking of adding an additional step for espeak that lets you pick which voice to use
18:13:16gevaertsI just noticed that I've heard of at least 6 sansa users and only 2 ipod users testing it.
18:14:08linuxstbgevaerts: I think there are probably more just Sansa users around... Ipods aren't that popular.
18:14:19linuxstbs/more just/just more/
18:14:28rasherEspecially in the dev-ish crowd
18:14:41krazykitespecially since sansas have been sold for practically nothing recently.
18:14:47scorche|shlostlogic: can you remember our rationale for why our profont fonts were "unmodified" (or however we declared it was alright to commit with that licence)
18:15:13lostlogic... because we downloaded them and converted them and nothing else?
18:15:25lostlogicwe didn't in any way modify the bitmaps, only changed their format
18:15:39lostlogichas an owner stepped forward and questioned it?
18:15:52 Quit tvelocity_ (Read error: 110 (Connection timed out))
18:16:39scorche| was just brought up yesterday and we couldnt remember the rationale..
18:16:42linuxstbThe license only talks about allowing unmodified distribution of the "profont distribution" - which we don't do, we distribute a converted version of one of the items in that distribution.
18:16:42rasherI think that counts as modifying - especially as the license text doesn't mention the bitmaps, but rather the entire package "Profont Distribution"
18:18:07lostlogicMeh, the other side of the coin is that the creator has accepted and posted modified distributions in the past, but now seems to have dropped off of the planet in terms of being contactable about another new distribution
18:18:43rasherThat's hardly terribly explicit permission though
18:22:28linuxstbgevaerts: I've fixed the iProduct/iManufacturer/iSerial strings to be the same as the Apple firmware, but my ipod is still not appearing with the same id as the OF - it's /dev/disk/by-id/usb-Apple_H_000A2700xxxxxxxx-0:0 The "H" is what appears in the kernel logs when linux detects the device - "scsi 43:0:0:0: Direct-Access H PQ: 0 ANSI: 0". The OF gives "scsi 27:0:0:0: Direct-Access Apple iPod
18:22:28linuxstb 1.62 PQ: 0 ANSI: 0"
18:23:55lostlogicFair enough −− I'll try to contact the author for explicit permission or a modified license
18:26:24 Nick mewshi__ is now known as Mewshi (
18:27:24 Quit csc` (Read error: 110 (Connection timed out))
18:27:46 Quit ompaul (Client Quit)
18:27:52Mewshiaww ^_^
18:27:55Mewshithat's sweet ^_^
18:28:11 Quit bumbl (Read error: 110 (Connection timed out))
18:30:22gevaertslinuxstb: Those come from identify2inquiry() in usb_storage.c. You have to make sure inquiry->VendorId and inquiry->ProductId are set properly
18:30:35 Nick Addicted_ is now known as Addicted (
18:30:37*linuxstb wonders why the rockbox.ipod change detection doesn't work correctly on ipods
18:31:40lostlogicprofont author emailed, rockbox-dev CC'd
18:32:09rasherlinuxstb: I noticed it worked on sansa
18:33:06linuxstbgevaerts: I've just tried v4 of your patch - I don't get the INQUIRY result too short error, but the info seems byte-swapped - "OTHSBI A H YB00"
18:33:16rasherEh, how do I commit an older version of a file that I got by using svn up -r ?
18:33:40Nico_Prasher: you need to use svn merge
18:37:10gevaertslinuxstb: I'll fix that.
18:39:53 Join bumbl [0] (
18:42:28n1sNico_P: did you have any luck with usb on the gigabeast?
18:42:38 Quit DerDome ("Leaving.")
18:42:48Nico_Pn1s: no, I didn't have much time and there is quite a lot to do
18:43:01Nico_Pmost of which is far above my knowledge, I fear
18:43:24Nico_Pthe USB driver is completely empty
18:44:07n1sAh, I dunno much about it either...
18:44:57 Join DerDome [0] (
18:45:05 Join axionix [0] (
18:45:46Nico_PI don't think we're going to be able to do much real testing without USB either...
18:45:46gevaertsI'm having some strange issue here. The INQUIRY result too short error was caused by some bytes being zeroed between the time the structure was filled and the time it was used. I "solved" that in v4 by moving the initialisation to right before usage. Now I'm trying to fix it for real, and I find that the issue goes away if I make the inquiry struct non-static
18:45:56 Quit axionix (Read error: 104 (Connection reset by peer))
18:46:23 Join Axio [0] (
18:47:21 Join t0pGuN [0] (
18:49:44 Join d3r3k [0] (
18:49:49d3r3kis there a way to watch .avi or .wmv files? I just put rockbox on my f10 giga and i can only play .mpg video
18:50:07 Join axionix [0] (
18:51:15krazykitd3r3k, no, only mpeg1/2 video. you'd have to convert them anyway to fit the size of the screen. for more information, read the PluginMpegPlayer wiki page
18:51:29krazykitsorry, PluginMpegplayer
18:51:38d3r3kactually, i just copy pasted my .mpg and bang it played it
18:51:41d3r3kno conversion needed
18:52:00d3r3kand sound was nice too
18:52:27krazykitno, i mean if the video is larger than 320x240, you will have to convert it to make it fit, assuming you want to see everything.
18:53:09n1shmm, weird, ld on rashers server segfaulted and caused red in the previous round...
18:53:54rasherOh dear..
18:54:54d3r3kokok one more thing, i had my original gigabeat software and firmware, but then it started creating 3 gig empty files, so i copied my data (all .sat) files to my computer, then formatted using windows format. everything worked for a day and then everything died i got the "no system found on hdd" error and everything died, gigabeat wasnt even recognized by windows anymore
18:55:23d3r3kso i followed the instructions and tried for many hours to get the firmware upgraded and nothing worked, so then i put the rockbox and bang everything works
18:55:42d3r3khow can i put all my 7 gigs of music back on the gigabeat since theyre all .sat files?
18:56:06 Join redwings0921 [0] (n=44249fa6@gateway/web/cgi-irc/
18:56:08krazykityou have to convert them back to mp3 as per the instructions on the wiki... which require the original firmware.
18:56:09rashern1s: can't reproduce it
18:56:25n1sseems kind of random
18:56:35d3r3ki saw a page on how to do it at but these steps are the steps to follow if u had not put the rockbox yet
18:56:42redwings0921Hello, can anyone tell me when the last Rock box update for the Iriver H10 was?
18:56:47rashern1s: First time I've noticed it too
18:56:53krazykitredwings0921, the last SVN commit.
18:57:04d3r3kit says "It is suggested that you do this before you load Rockbox, but there are also steps on what to do if you've already taken the jump"
18:57:10redwings0921What is the SVN?
18:57:15d3r3ki cant find those steps on what to do if i've already taken the hump
18:57:32n1sredwings0921: the version control system we use
18:57:34d3r3koh shit i found it brb
18:58:26Horschtredwings0921, basicaly, there's an update every day, called daily builds
18:58:26 Quit redwings0921 (Client Quit)
18:59:10 Quit keanu (Read error: 104 (Connection reset by peer))
19:00:25 Join keanu [0] (n=none@unaffiliated/keanu)
19:00:39 Join Lear [0] (i=chatzill@rockbox/developer/lear)
19:01:26Nico_Pthere is even a build for each source code change
19:01:56 Join star_jasmine [0] (n=arwyneve@
19:02:43linuxstbgevaerts: Are you going to attempt to fill in the iSerial string correctly? I can help for the ipods, but I've no idea about where to get that number from on the Sansas...
19:03:21gevaertslinuxstb: I'd like to. Maybe we can use the internal flash serial number
19:03:45linuxstbThere's also the H10 and m:robe 100...
19:04:12d3r3kwhere can i find the original FWIMG01.DAT file for my f10 ?
19:05:15krazykitd3r3k, the gigabeat room software should have it
19:05:18 Join thgz [0] (
19:05:49gevaertslinuxstb: are the flash or HD based ? Hard disks should have serial numbers as well that we might easily get at
19:06:46thgzIn r16268 a modified sapi_voice.vbs was committed. It added some kind of support for sapi4. Does it mean that I can use sapi4 for generating a .voice on cygwin?
19:07:39d3r3kgreat thanks
19:08:05***Saving seen data "./dancer.seen"
19:08:46 Quit t0pGuN (Read error: 110 (Connection timed out))
19:08:57thgzOops, I meant r16269
19:09:01 Quit einhirn (Read error: 104 (Connection reset by peer))
19:09:02d3r3kits not there =/
19:09:19d3r3kis it inside another file
19:09:23d3r3kor just FWIMG01.DAT ?
19:09:32 Join davina [0] (
19:09:40rasherthgz: that's been possible a long time
19:09:43 Quit DerDome (Read error: 110 (Connection timed out))
19:10:11thgzRasher: Yes, with a patch which is in flyspray
19:10:19krazykitd3r3k, really, i'm not sure. the forums should have more information about the recovery procedure and where you can find it, exactly
19:10:32linuxstbgevaerts: The m:robe 100 has a 5GB microdrive. The H10 comes in two versions - a 20GB 1.8" hard disk and a 5GB/6GB 1" hard disk.
19:10:38rasherI was under the impression that sapi4 voice generation has been working and fine for a while... amiconn?
19:10:41Horschtquestion about colors. When I try to set the line colors from within rockbox, R and B values are maxed at 31, G goes up to 63
19:10:46Horschtis that intenional?
19:11:04 Join pondlife [50] (n=Steve@rockbox/developer/pondlife)
19:11:08rasherHorscht: That's a property of the display
19:11:11Nico_PHorscht: yes, it's because the colors are 16 bit
19:11:34n1s5:6:5 rgb
19:11:53Horschtah, ok
19:12:36gevaertslinuxstb: let's try going for drive or flash serial numbers as default, and "correct" ones for devices where we know how to get them
19:12:38 Join z35_1 [0] (
19:13:05 Quit z35 (Read error: 113 (No route to host))
19:13:08thgzRasher: Well at least I haven't found a way to do that without fs#7578
19:13:27 Join DerDome [0] (
19:13:57linuxstbgevaerts: Sounds good to me. I'm not sure where this code can go though - it seems messy to put a load of #ifdefs in the usb driver code.
19:14:03 Join advlaptop2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
19:14:17rasheramiconn: I was under the impression that could generate SAPI4 voices - is this not true?
19:14:44 Quit advlaptop2019 (Read error: 104 (Connection reset by peer))
19:15:03n1sgevaerts: doesn't the pp chip itself have a unique id?
19:15:51linuxstbn1s: I think we need to try and work out how the original firmwares do it. I've done that for the ipods, but don't own any other PP devices...
19:16:19n1syeah, that is probably the best way
19:17:06thgzOH no problem now
19:17:41thgzFound from 24.12.2007 irc log Amiconn covering the same question
19:17:52rasherthgz: so that fs entry could be closed?
19:18:29thgzI think so, because that patch existed for a long time.
19:19:24 Part thgz
19:19:33lostlogichere's the original text that told me we are fine in distributing profont as we do: <−− didn't remember where I'd found it, but it's in the file.
19:19:56lostlogictheir express intent is to be credited and to have profont be free and to have it reach as wide of an audience as possible
19:20:31linuxstbBut also "You may redistribute unmodified copies of ProFont Distriubtion" - i.e. the whole thing...
19:20:49 Nick z35_1 is now known as z35 (
19:21:04rasherWhat they meant means nothing if they wrote something else.
19:21:10linuxstbIt's odd though that the "distribution" and the fonts themselves have different (C) holders...
19:21:21 Join sitwon [0] (
19:21:52 Quit bumbl (Remote closed the connection)
19:22:11lostlogicI remember now that I tried to dig up and email the original original authors but couldn't the first time, so this time I've emailed tobias who is not one of the original authors and redistributes non-original versions of the font from his site :-\
19:22:36linuxstbSo he's just doing what we're doing?
19:22:53*rasher sighs at the whole situation
19:23:26lostlogicpossibly :-\ the license is pretty contradictory IMO −− given that the distribution clause says that it may be included in other applications and collections but must be in it's original distribution form
19:23:41sitwonI'm having problems with my rockbox'd e280. It was working fine for a long time, but recently it started freezing up. I tried to connect it to my computer to check the filesystem, but I can't. I use recovery mode to put the original firmware back on it, but it still wont connect properly.
19:24:40sitwondmesg reports: usb 5-2: new high speed USB device using ehci_hcd and address 18
19:24:40sitwonusb 5-2: config 128 interface 0 altsetting 0 endpoint 0x81 has an invalid bInterval 100, changing to 10
19:24:40sitwonusb 5-2: configuration #128 chosen from 1 choice
19:25:01sitwonno mountable devices...
19:25:20rashersitwon: would it by any chance be in MTP mode?
19:26:21 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
19:26:52sitwonoh shi... Sorry, I feel like an idiot now. (although I wonder how it got switched back to MTP in the first place...)
19:27:42rashersitwon: recovery mode does that
19:28:01rasherHad me fooled the first time too
19:28:16sitwonrasher: yes, but I only did recovery mode because it already wan't working... but I know it was mounting fine a week ago
19:28:34lostlogicoh and actually the .pcf version that I converted to .bdf doesn't come with any license file at all *laugh*
19:30:11 Part pondlife
19:31:30sitwonewww. font licensing? have fun w/ that
19:31:34linuxstbAny Sansa users with Linux around? What does "lsusb -v" display for iProduct, iManufacturer and iSerial? (using the OF USB mode)
19:33:08 Join replix [0] (
19:33:15gevaerts"SanDisk", "Sansa c250" and "4453033f-32305453-b9108047-6b008c55-00000000"
19:33:31linuxstbLinux is inserting those hyphens?
19:33:46linuxstbI just get a 16 (hex) digit string...
19:33:52rasherlinuxstb: "SanDisk" "Sansa e260" and "4f027154-4056b477-96d622bf-006b4ca5-00000000"
19:34:00replixi have a sansa e250 with rockbox
19:34:15replixand i want to play pokémon with rockboy
19:34:22replixi got the ROMs
19:34:26sitwoniProduct: Sansa e280 iManufacturer: SanDisk iSerial: 00000000-00000000-8867b393-fd1aff17-00000000
19:34:43replixbut when i play the game, the sound is _crappy_
19:34:52krazykitreplix, so turn off sound.
19:34:57desowin"SanDisk", "Sansa e280", "00000000-00000000-5603b487-d025f902-00000000"
19:35:10gevaertslinuxstb: no. they are sent like that over the wire (except that it's in 16-bit characters there)
19:35:15sitwonreplix: the processor in the Sansa isn't fast enough for gameboy emulation
19:35:34rashersitwon: huh?
19:35:59linuxstbgevaerts: Those serial numbers look invalid to me - IIUC, that PDF you pointed me to says that only hex digits are allowed...
19:36:43sitwonI'm pretty sure I read on the Forums the reason the rockboy was so bad with gbc games was because the cpu wasn't fast enough for full emulation in real-time
19:37:04linuxstbIt should be fast enough, it's just that no-one has worked hard to optimise it.
19:37:22linuxstb(especially as Rockboy is just using one of the two cores)
19:37:31gevaertslinuxstb: let me check what the usb core spec says.
19:37:34sitwonoh, sorry for the misinformation then
19:38:07sitwonI know mine is original cuz I just loaded the OF... and I probably didn't even have to :(
19:38:59kugelpetur: thanks
19:39:39 Quit d3r3k (Read error: 110 (Connection timed out))
19:39:42peturkugel: I was quite surprised about that. I can remember that it worked
19:40:02kugelyea, I was very surprised as well
19:40:48replixand why doesn't it use two cores?
19:41:41linuxstbreplix: Because no-one has implemented it...
19:41:46desowinbecause nobody made it using two
19:41:48gevaertslinuxstb: the core spec doesn't seem to specify anything about serial numbers except their existence
19:44:13linuxstbgevaerts: I'm wondering if the USB_DT_STRING request should be handled in usb_storage.c, rather than usb_core.c ?
19:45:00linuxstbOr am I misunderstanding the hierarchy of things?
19:45:50gevaertslinuxstb: USB_DT_STRING is device-level, not interface-level, so they should be in usb_core.c.
19:47:20gevaertslinuxstb: you could have a device that handles both mass-storage and something else like audio. Those functions will share the same device descriptor (since there is only one of those), and therefore the same vendor,product and serial number strings
19:47:28 Quit karashata (" HydraIRC -> <- Organize your IRC")
19:48:04linuxstbSo a device that offers two disks (like the Sansa, with the SD card) will use the same serial number for both disks?
19:48:06gevaertsMaybe we should actually try to build that combination. Imagine being able to play music through your DAP instead of on your DAP...
19:48:21amiconnrasher: SAPI4 voice building worked for quite some time. What I added today is sapi4 support for listing voices (a helper function for rbutil, nothing to do with voice creation in the build system)
19:48:53rasheramiconn: that's what I thought
19:49:05 Join mirak [0] (
19:49:08amiconnIt's just that it takes ages compared to sapi5
19:49:10linuxstbgevaerts: There have been a few feature requests for that...
19:49:21amiconnANd not all voices work if you're not admin
19:49:55pixelmaNico_P: I get the "skipping track" effect very consistently for some folders (on my c200). Is there something I could do to record what's going on there (LOGF, ...)?
19:49:56amiconnThat's actually a bug in those engines - not the script. Those voices don't even work from the control panel as non-admin
19:50:21kugelpetur: the line I added wasn't present in the commit of user configurable rec path (
19:50:28linuxstbgevaerts: So usb_core_init() will be the place set the serial number string?
19:50:35 Quit mirak (SendQ exceeded)
19:51:19*amiconn will try to generate a .voice file with sapi4
19:51:41 Join mirak [0] (
19:52:00gevaertslinuxstb: yes and no. That serial number is part of the entire device. You could support scsi-level serial numbers as well (although I'm not sure if ums-devices typically support that)
19:52:42peturkugel: weird. Maybe settings were always saved at shutdown in the past? I set my rec dir once and never used it again.
19:53:07*petur was reading and got a bit scared...
19:55:16star_jasminewhen I use the rbuilt tool with a sansa e200, the voice clips get cut off near the ends of words, and sometimes, the phrases don't mach up to the prompt. buffer instead of battery level ec. what am I doing wrong?
19:56:25Nico_Ppixelma: yeah, a logf dump could help
19:56:39Nico_Penabling logf in playback.c should be enough
19:58:00bluebrotherDomonoky: hope I didn't break your work in progress :o
19:58:01linuxstbgevaerts: OK, but for the "iSerial" serial number, that just needs to be unique to the entire device, and so we can initialise it at boot?
19:58:12 Join MattAndrew [0] (
19:58:33gevaertslinuxstb: yes
20:00:05bluebrotherpetur: me too ... :(
20:00:38linuxstbgevaerts: Also, I'm guessing the [32] size for the strings in usb_strings is just arbitrary? i.e. the spec allows the full 8-bits to be used for length? (given that the Sansas have a 44 character iSerial).
20:00:41peturbluebrother: they just took her laptop and never gave it back - wtf?
20:00:49rasherpetur, bluebrother: Nothing to worry about unless you're dark-skinned or have a huge beard
20:01:27bluebrotherwell, it gives entering the US a really bad taste. Despite of the concerns I already have ...
20:01:29peturso I'll have to shave off my beard if going to the us :(
20:03:51 Join dave_ [0] (
20:03:58kugelSo, if we make the meet in the US, we basically risk to travel without computers, or even begin the trip with our laptops and get them stolen at the airport?
20:04:27gevaertslinuxstb: I assume so yes.
20:04:41bluebrotherkugel: not stolen. Confiscated. Or more positive, get them screened intensively.
20:05:36rasherbluebrother: if someone takes something from me without reason and without giving it back, it was stolen.
20:05:46kugelbluebrother: but this Maria Udy never saw her laptop again after this, did she?
20:07:51 Quit replix ("Verlassend")
20:10:02 Join matsl [0] (
20:13:55 Join markun [50] (n=markun@rockbox/developer/markun)
20:13:58 Join karashata [0] (
20:15:18 Quit star_jasmine ()
20:15:39dave_just curious, but is anyone else having trouble as of late with Rockbox skipping seemingly random tracks in the playlist?
20:17:28 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:17:38kugelMe not, but pixelma is having such problem IIRC
20:17:46gevaertsI have a data structure in usb_storage that is OK until usb_core_init() starts the usb_core_thread. Immediately after create_thread() returns my datastructure is corrupted. Any hints on how to debug this ?
20:18:27linuxstbWill it fit in iram?
20:18:56gevaertslinuxstb: it's 36 bytes
20:19:15gevaertsbut I don't know if data around it are also getting corrupted
20:19:38dave_kugel: I am running a patched build (viewports for WPS), but I doubt viewports touches the playlist
20:20:41 Quit Mewshi ("Konversation terminated!")
20:23:21 Quit rocko (Read error: 110 (Connection timed out))
20:25:41pixelmakugel, dave_: for me it's not very random (always the 3rd track in a playlist made of one folder - the random aspect being that some folders are affected some not.
20:26:22pixelmaI don't use a viewport build
20:26:38 Quit bluebrother ("leaving")
20:27:49dave_pixelma: funny you should mention the 3rd track, I just realized the last time it happened it was the 3rd track that was supposed to play that got skipped, but it wasn't the 3rd track in the directory. I started further down in the album.
20:28:28Nico_Pn1s: it seems your commit for the S makes the build fail
20:28:46Nico_Pthe linker, specifically
20:29:02n1sNico_P: weird, works perfectly here...
20:29:43n1sdo you have the multilib patch for gcc applied?
20:29:44Nico_Pwhat version of arm-elf-gcc do you use?
20:29:55n1s4.0.3 with the multilib patch
20:30:00Nico_Pno idea. I used
20:30:03pixelmadave_: yes, same here. As I said it's always the 3rd track in the playlist - if less than 3 tracks are remaining, the last file won't be played and I can't just play the last track at all (or play a single track from this folder by using insert from the context menu). Btw. what player do you have?
20:30:38dave_pixelma: Gigabeat F40
20:30:52dave_I also have trouble playing single-track directories
20:32:14kugelIs this bug limited to playing music via filebrowser?
20:32:25Nico_Pn1s: how can I know?
20:32:31Nico_PI have 4.0.3 btw
20:32:57n1sif you still have the gcc sources you can check, other than that I don't know...
20:33:14Nico_Pdoes include the patch?
20:33:27n1ssince ~november iirc
20:33:41Nico_PI'll try rebuilding the compilers then
20:33:41gevaertsIt goes wrong somewhere inside create_thread. If I check the data at the start of the thread it is already wrong
20:33:58n1sjust the arm gcc should be needed
20:34:58pixelmakugel: will try (somehow I think I read a report that it happens for some using database too) but I got to go now, bbl
20:35:11 Join t0pGuN [0] (
20:35:19 Part pixelma
20:35:29n1sNico_P: see if arm-elf-gcc -print-multi-lib give any output
20:35:38t0pGuNhow can i do things like create a bookmark or select "random play" so that it plays all of the songs in all of the folders
20:35:59Nico_Pn1s: there is output
20:36:07Nico_Pdon't konw what to look for though
20:36:41n1sdoes it list amr926e or arm1136jf-s ?
20:37:22n1sok, then you need to rebuild
20:38:15Nico_Pdoing that
20:39:59 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
20:40:29 Quit amiconn (" HydraIRC -> <-")
20:40:43 Join stripwax [0] (
20:40:47 Quit matsl (Remote closed the connection)
20:41:34 Join Mathiasdm [0] (
20:43:47 Quit advlaptop2019_ (Read error: 113 (No route to host))
20:51:02 Quit Buschel ()
20:52:14 Join Casainho [0] (
20:53:08 Join csc` [0] (n=csc@archlinux/user/csc)
20:53:26Casainhohello RB people :-)
20:55:21Casainhocan someonte tell me If sansa E200have morew than one IC of flash memory?
20:55:48CasainhoI am reading the TWiki page about Sansa E200 but I can't have sure about how many ics of Flash are used...
20:56:30stripwaxCasainho - (I don't know the answer, but) I'm interested in why the number of physical ICs matters
20:56:50 Quit t0pGuN ()
20:56:57Casainhobecause I would like to understand how Sansa works...
20:57:25Casainhowhere are the bootloader of the rockbox and the rockbox application..
20:58:08gevaertslinuxstb: I uploaded a new version which should fix your disk-name byteswapping issue.
21:01:01Nico_Pis mcuelenaere able to run code on the ZVM?
21:01:03Casainhoanyone can answer this, where reside bootloader on Sansa?
21:01:03 Join shivramk [0] (n=shivram@
21:01:14Nico_Pthe latest wiki edits seem to indicate he is
21:03:41 Quit shivramk ("Leaving.")
21:04:38n1sCasainho: which bootloader? the rb one or the of one?
21:05:00 Quit perrikwp (" ajax IRC Client")
21:05:04Casainhon1s: booth, If you can tell me... :-)
21:05:34Casainhon1s: I would like to learn understanding Sansa E200, wich is one player I own
21:06:45n1sthe of one is in a special flash which the cpu starts executiong from, it loads the firmware from a hidden partition of the main storage flash, or if rockbox is installed it loads the rb bootloader from said hidden partition
21:07:19n1sthe rb bootloader then loads rockbox from the fat partition of the main flash
21:08:09***Saving seen data "./dancer.seen"
21:09:13Casainhoso, the main storage flash can be one flash IC with 2 partitions? 1st for bootloader (maybe without file system, just ARM code??) and 2nd with FAT32 file system (with RB and audio files??)??
21:09:55 Join Domonoky_ [0] (n=Domonoky@
21:11:35gevaertsCasainho: It's actually the reverse. The first partition contains audio files (and RB), and the second partition has the OF (and RB bootloader). This second partition is indeed not a normal filesystem
21:11:43n1sthe startup code and of bootloader is not in the main flash it's in a separate flash and the of itself lives in the hidden partition
21:12:09n1s^what he said
21:12:41Casainhookok, I understand now :-)
21:12:59Casainho * Anyone with knowledge of the inner workings of Rockbox know if this is feasable? We can always slightly modify rockbox so that it saves configuration changes (or possibly a database, but that should probably be stored on the SD card in question instead) to the serial flash after a config change. The point is that there's no file system on the spi flash, so will Rockbox work? We could...
21:13:01CasainhoRockbox will be stored in SPI flash and be copied to ram upon bootup.
21:13:02Casainho...have the Rockbox bootloader at the start of the serial flash and another binary chunk that's an image of a FAT filesystem that contains the rest of Rockbox (besides RoLo) and then just pretend that the block of RAM that's a copy of this image is equivalent to the mass storage device that the Rockbox files would normally be served from. Easy / difficult to implement this way?
21:14:38peturbah... saratoga added a line to the H10 debug screen and now the last line is lost :(
21:14:46CasainhoSo I will answer to that question, suggesting a system working like Sansa E200, 1 main flash IC with 2 partitions, one with RB bootloader and other for FAT32...
21:15:45Casainhothe ARM MCU we will use have a bootloader to start loading from NAN flash ic...
21:16:12 Quit gevaerts ("back later tonight or tomorrow")
21:16:17scorche|shwhy do you want 2 partitions?
21:18:03Casainhoscorche|sh: since I understand, one partition will have RB bootloader which will make the other partition with FAT32 available, working
21:19:22SlasheriCasainho: that could be possible if there would be plenty of internal flash available (in the range of several megabytes)
21:19:54CasainhoSlasheri: could be possible, what?
21:20:08Slasherito make a FS in flash memory
21:20:32Slasheribut i don't think any target has that much of flash storage available
21:21:06Casainhobut, I am thinking on that, like 16MB, for RB bootloader + FAT32 with RB code
21:21:43Slasheriah, you are designing a player?
21:21:52CasainhoSlasheri: I don't understand... does Sansa work like that? with a FS FAT32 on main flash IC?
21:21:58 Join LambdaCalculus37 [0] (
21:22:22 Join GodEater [50] (n=bryan@rockbox/staff/GodEater)
21:22:39CasainhoSlasheri: please read more here:
21:22:45Slasherii have no idea about sansa, but a flash based target would be entirely different thing than hd based in anyway
21:23:35karashataSlasheri: the Sansa's are flash-based, as is the iPod Nano
21:24:15 Quit kugel (Remote closed the connection)
21:24:21Slasheriyeah, never mind. i just thought about a different thing to save configuration to flash instead of a hdd
21:24:34 Join kugel [0] (n=kugel@unaffiliated/kugel)
21:26:13 Quit Mathiasdm ("Yuuw!")
21:27:16 Quit Domonoky (Read error: 110 (Connection timed out))
21:27:32Nico_Pn1s: things are better now that I've rebuilt arm-elf-gcc
21:27:42n1sgreat :)
21:33:48 Quit kugel ("Benutzer ist abwesend.")
21:34:57 Join kugel [0] (n=kugel@unaffiliated/kugel)
21:38:49 Quit LambdaCalculus37 (Read error: 110 (Connection timed out))
21:45:28 Join crzyboyster [0] (n=4b596e13@gateway/web/cgi-irc/
21:47:23 Quit davina ("GNU/Linux the free alternative to Windows")
21:49:27MattAndrewCasainho: howdy!
21:49:29linuxstbpetur: Is there a #define to enable for Rockbox to check that rockbox.ext has changed after USB connection? My ipod doesn't seem to do it (with the UMS patch).
21:49:59CasainhoMattAndrew: heelo :-)
21:50:16Casainhohow are things going? - I let a message recently...
21:51:39crzyboysterI am having trouble setting up a hindi voice file. I know I am doing it wrong because I am getting a 3 kb voice file. What is the proper way?
21:52:27rashercrzyboyster: How are you doing it?
21:54:45linuxstbpetur: I've found it - it's explicitly disabled for ipods...
21:54:58MattAndrewCasainho: alright. I just read your conversation above about the flash thing. I had assumed that rockbox uncompressed with all the plugins and such would fit in RAM, but found out yesterday that it's ~35MB, which means it can't be stored in spi flash compressed and still be usable like it is on a hard drive player where the files are just available without significant modification.
21:55:34crzyboysterI get the latest version from extract and run the exe. Then I go to the Accessibility tab, click "Create Voice File" and then select "Hindi" as the language. Then I make the TTS engine be espeak and point it towards C:\Program Files\eSpeak. I didn't input anything for TTS options. I make it rockbox speex encoder and leave the wavtrim as is. I run it. What am
21:56:02MattAndrewcasainho: something like, have rockbox load into ram the main program and whatever plugins are necessary at that moment and then if a new file is needed, to uncompress it from the spi flash on the fly and load it in to ram.
21:56:02rashercrzyboyster: you need to tell espeak to use the hindi voice
21:56:06rashercrzyboyster: -v hi
21:56:28 Join ian_hawdon [0] (
21:57:01crzyboysterAnd that goes into the "Configure TTS Engine" field?
21:57:18crzyboysterAnd what does this recent commit (and decommit) mean? > Voice-building changes: Ask for which eSpeak voice to use for some languages. Please test and report if you have any
21:57:46rashercrzyboyster: it was for the build system thing, not related to rbutil
21:58:46crzyboysterI put "-v hi" into the Configure TTS Engine box and I still get a 3 kb file?
21:59:21ian_hawdonhi, was wondering, has anyone been having problems with playback? e.g. showing it's on track 2 on a playlist when it's only on tack 1? or is this an iPod build problem? currently, I've reverted back to a build sometime in January, where everything seems to work fine
21:59:47rashercrzyboyster: Then you need to see Domonoky_ or bluebrother - it works if you use the Build system, and I've no idea how to tell what goes wrong in rbutil
22:00:38CasainhoMattAndrew: but why need to use SPI flash?
22:00:46MattAndrewcasainho: but I think I underestimated the size of the flash that we need even for a compressed rockbox. The zip file for the uisimulator is 9.4MB, so maybe we need that much plus some room for the future, and that much capacity is just not available in spi flash right now.
22:00:58crzyboysterI will post in the forums and thanks.
22:01:07 Quit crzyboyster ("CGI:IRC")
22:01:16MattAndrewCasainho, to save board space, power and to make the circuit board layout easier
22:01:22CasainhoMattAndrew: 16MB flash can be ok, I think
22:01:43rasherMattAndrew: the simulator isn't stripped
22:01:49rasherMattAndrew: regular rockbox builds are far smaller
22:01:54MattAndrewCasainho, the only spi flash that digikey sells that's bigger than 8MiB is a package that's difficult to solder
22:01:57CasainhoMattAndrew: ah, okok, I understand now.
22:02:21CasainhoSo, let's leave that option and use a larger flash IC
22:02:33MattAndrewCasainho, as I mentioned, I don't have a working ipod anymore, so I can't see what rockbox is like on dedicated hardware
22:03:11Casainho"dedicated hardfware"? - i don't understand what you pretended to say...
22:03:24linuxstbJust download a file from the website to see how big Rockbox currently is.
22:03:32MattAndrewI mean on a portable music player (not a pc)
22:03:56Casainhoeheh - so just do the hardware prototype :-)
22:04:34 Quit z35 ("Leaving")
22:04:43CasainhoMattAndrew: so, will you use a NAN flash? if so, what size? 16MB? more?
22:05:11BigBambiMattAndrew: The current size of extracted for gigabeat F without fonts but with all themes etc is 6.4MB
22:05:25 Join webguest59 [0] (n=503c3057@gateway/web/cgi-irc/
22:05:43BigBambiClearly not all of which is in RAM when running
22:05:48MattAndrewok. So, the ~1.6MB zip file size includes all of rockbox (with all plugins and theme media)? linuxstb: I'm downloading the ipod first gen file, but it's going slowly...
22:06:00BigBambiYes, except fonts
22:06:11linuxstbBigBambi: I would guess LCD size would make a noticable difference to size (the embedded bitmaps). So best to choose a zip with the same size/depth LCD.
22:06:28BigBambiWhat is the proposed zip
22:06:33BigBambisorry, screen size
22:06:57n1sh10small would be closest
22:07:14webguest59i have a question: ive installed rockbox succesfully on my 80gb 5.5g, but windows says it cant install its drivers when i connect it to my computer with USB
22:07:36n1swebguest59: reboot it to the original firmware to connect
22:07:39BigBambiok, h10 small is 4.5 MB
22:08:00linuxstbBigBambi: And the fonts? I don't think you can ignore them...
22:08:01*GodEater prays for working UMS to rid us of these repetitive "I can't find the driver" questions
22:08:05MattAndrewcompressed or uncompressed?
22:08:10 Quit desowin ()
22:08:20ian_hawdonwebguest59 just ignore it, it will go to diskmode and thus, not need the drivers anyway :D
22:08:39*n1s prays for ums to be bale to copy stuff in dirs to the beast
22:08:51webguest59@ian, it doesnt list as a hard drive
22:08:53MattAndrewBigBambi, okay, then it does seem doable with just spi flash then.
22:08:57BigBambiMost of our targets have 32 MB or RAM, of which ~32 MB (someone correct me here) is available for the audio buffer
22:08:59GodEatern1s: that too :)
22:09:26ian_hawdonif it's gone to disk mode... make sure disk mode is enabled on iTunes
22:09:44 Join Genesis [0] (n=genesis@
22:09:52BigBambiSorry, I meant ~30 MB free for buffer
22:10:27BigBambiAnd, as I said before that 4.5 MB is without fonts
22:10:29webguest59ok, i booted the apple firmware and it lists as HD now, thanks
22:10:40webguest59please put this in the FAQ
22:10:52GodEaterwebguest59: it's on the front page of
22:12:02webguest59no it isnt? maybe im not looking good enough?
22:12:03MattAndrewBigBambi, okay, so assuming the fonts aren't huge, can I get away with either having the compressed rockbox zip always accessable on spi flash OR loading rockbox uncompressed into ram upon bootup (both options are WITHOUT any other mass storage device available, since this is an SD card only platform)?
22:12:31GodEaterwebguest59: yes it is. it says "Ignore driver popups from Windows."
22:12:46BigBambiI don't really know, I'd have a look at how other targets do it
22:13:30 Quit JdGordon (Remote closed the connection)
22:13:47webguest59it should list the ipod anyway, even if i didnt ignore the dialog
22:13:52 Join JdGordon [0] (
22:14:05GodEaterwebguest59: it doesn't apply to only the ipod
22:15:41webguest59ok, thanks for your help anyway, sorry for the inconvenience...
22:15:42 Join robin0800_ [0] (
22:16:13GodEaterwebguest59: hopefully it won't be an issue for too much longer :)
22:16:43 Quit robin0800_ (Client Quit)
22:16:43 Quit webguest59 ("CGI:IRC (EOF)")
22:16:57 Join robin0800_ [0] (
22:17:09 Join azdrubal [0] (
22:18:09 Quit robin0800_ (Client Quit)
22:19:42 Join ashes [0] (n=ashes@2001:5c0:8fff:ffff:0:0:0:3f)
22:20:09 Join ol_schoola [0] (
22:21:13 Quit ol_schoola (Client Quit)
22:21:13asheshello. i normally name music files with '01 -' and '02 -' so they play in order. will the rockbox database probe every file for its track number so it can display the correct order, and i can leave the numbers out of the filenames?
22:21:16 Join einhirn [0] (
22:21:37krazykitashes, assuming you're using the database and everything is tagged properly, yes.
22:21:49linuxstbashes: Yes, Rockbox's database will use the track numbers from the metadata (i.e. tags) in the files.
22:23:27 Quit roolku ()
22:25:14 Quit Lear ("ChatZilla 0.9.80 [Firefox 3.0b4pre/2008020708]")
22:25:39Nico_PGodEater: seen ?
22:25:58Nico_Pthe patch *should* fix the audio glitching issues
22:26:49GodEaterNico_P: I'll give it a try first thing in the morning
22:26:57 Quit tvelocity__ (Remote closed the connection)
22:27:22GodEaterdid you arrange with amiconn to get the 5.5G sent yet ?
22:28:00 Quit asdrubal (Read error: 110 (Connection timed out))
22:29:00 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
22:29:20Nico_Pyes, he sent it yesterday
22:30:29linuxstbNico_P: So you're now responsible for large-sector support as well ;)
22:30:47Nico_Phehe :)
22:32:16stripwaxDomonoky_- if you're interested I have a new, better-looking Talkfile button for rbutil: . I tried to match the style and colouring of the existing button icons this time
22:32:17 Join BHSPitLappy [0] (
22:33:33MattAndrewlinuxstb, do you have instant messaging? I'm chatting with Casainho about your forum post if you want in on the discussion
22:33:46stripwaxNico_P - do you know about this bug: if you play a track from an album (which creates the auto playlist for the album), and then edit the playlist, and insert and entire new album in the middle, it doesn't actually buffer or play the tracks you inserted when it gets to them (it continues to play the first album)
22:33:50linuxstbMattAndrew: Why not just talk here? I'm sure others have opinoins.
22:34:28CasainhoI also prefer chating here :-) - because of the logs and others can give ideas also :-)
22:35:51Nico_Pstripwax: ah, no I didn't... actually I thought it worked OK
22:36:12CasainhoMattAndrew: so, SPI flash or not? :-)
22:36:17Nico_PI seem to remember testing that
22:36:56MattAndrewlinuxstb, Casainho : I think there's a subtlety about what's being copied. I'm talking about the circuit board design files (schematic, board layout and gerbers), not the general list of parts we've all been discussing that's currently on the wiki. I *am* planning to be the sole designer of the circuit board, at least for this first version.
22:37:18 Join tessarakt [0] (
22:37:59 Join Kitt [0] (
22:38:34linuxstbMattAndrew: Then that doesn't fall into what I would call an "open/free" design - the whole point of open source software (IMO) is that it's a collaborative process, and everyone has equal rights over the final product. I would like to see the same principles with an open hardware design for Rockbox.
22:38:40CasainhoMattAndrew: can you please add that information about you be the only designer of the circuit biard on TWiki? - because I think Spark would like to work with you for this project, and since that, we will not be able. He should knows that.
22:39:51KittHello, I just put rockbox onto my ipod video (30g) and now explorer doesn't detect the device anymore. I searched in the forums and read that the only way for it to recognize is to switch it back to apple mode whenever you want to transfer songs. Is this the only way to do this?
22:40:04GodEaterKitt: currently yes
22:40:17CasainhoMattAndrew: for me this is no problem, however, understand that are here more people, like Spark and me, that wants to make a totaly free/open hardware, even using KiCad, we will do that!
22:40:22KittOk, thanks GodEater
22:40:47MattAndrewspi flash: I still don't know. If given a bunch of time, I could modify rockbox to work that way, but I can't do it before the summer at any rate if it's really a lot of effort, and the other problem with major changes to rockbox is that maybe everybody else who's a core developer won't like them and won't accept the patches and all the effort will be for nothing or it'll be effort that ends up in a fork that's a dead end or
22:40:51MattAndrewa patch that's not maintained.
22:41:46linuxstbCasainho: This is why I don't think you can ignore the license issue until later - you should try and decide on a license before any design work is done, so everyone knows where they stand.
22:44:07MattAndrewtotally free: I guess we just differ on our definitions of these things. One could argue that public domain is *more* free than open source software / open hardware designs, since you are, in the case of public domain stuff, "free" to change one thing and call it your own copyrighted work. (I'm not arguing for using a public domain "license", I'm just pointing out that "free" has different definitions)
22:44:21GodEaterCasainho: I agree with linuxstb. If you want your design to be properly free/open you cannot share ideas with MattAndrew as long as he wants to license his work under the CC-by-nc-sa
22:44:38BigBambiYou cannot change something under someone elses copyright and claim that copyright yoursdelf
22:44:40 Nick fxb__ is now known as fxb (
22:44:40 Part Kitt
22:45:04MattAndrewclaim copyright: as far as I understand it, public domain means nobody claims copyright ownership on it
22:46:45CasainhoMattAndrew: have a short winodw time to make the hardware, It will be a prototype, not a final version, we after must work on another design, so I think there will be no problem, we can use some parts as Matt is doing from looking at others hardwares :-)
22:47:09BigBambiCasainho: Prototype or not is irrelevent
22:47:41MattAndrewGodEater, "you cannot share ideas": that seems extreme. The design, when finished, will still be freely available for anyone to have a copy of and to use in making a player, just not commercially. This point is really worth saying "don't talk to him at all about it" over?
22:48:25BigBambiI think Casainho (and others) need to be properly aware of what this licence means to them
22:48:37BigBambi(or indeed any other licence)
22:48:41 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
22:48:49linuxstbMattAndrew: The point is not just your final design, it's derivative works as well. So any player based on your design can never be manufacturered commercially, and as hardware costs money (unlike software), that's a big limitation IMO.
22:48:53pixelmaNico_P, kugel: about the skipping song problem earlier, it does happen with the database too. It is also predictable if album = folder, the same thing happens and e.g. all tracks out of my "broken" various tracks folder which end as a single track in a respective sub"folder" in "Album" or "Artist" can't be played
22:49:02MattAndrewand anyway, the list of parts and the reasons behind using them is valuable enough by itself.
22:49:42kugelpixelma: I haven't tested it with a current build, but I don't have this problem with the dircache for multivolume patch applied
22:49:52GodEaterMattAndrew: I think the point is we'd *like* to see a commerical version. It would maker it both cheaper to obtain, and more likely to be better quality.
22:50:09MattAndrewGodEater, ah, I think I'm starting to understand
22:50:17Casainholinuxstb: but in hardware is very dificult to prove that you reuse parts from others... I think there will be no problem in using some parts, as Matt is doing right now! - or I am wrong?
22:50:29BigBambiCasainho: The parts are not the issue
22:50:30MattAndrewGodEater, I can't agree about the quality point though. :)
22:50:43GodEaterMattAndrew: heh - fair enough :)
22:50:58BigBambiManufacturing quality I'd say pretty certainly
22:51:18GodEaterI have only my own soldering skills to look at, and if I saw them in a shop - I wouldn't buy it ;)
22:51:19MattAndrewI can't claim copyright on a non-verbatim copy of a list of parts
22:52:03MattAndrewGodEater, :) I meant more the quality of the pcb layout than the soldering job
22:52:32CasainhoMatt come to make is own version of hardware for his school project
22:52:54linuxstbMattAndrew: I assume you're intending to run Rockbox on your hardware? If so, are you intending to port it yourself, or are you hoping others will help?
22:52:56CasainhoI think we can win something with is work and also we can win :-)
22:52:58BigBambiThing is, I would like an open hardware player, but am never going to contruct it myself, so a commercial version of it is the only option
22:53:01GodEaterin any case, this is really the point we're making. If Casainho end's up using part of your design, it become derivative (iiuc), and therefore will ALSO be unable to be adopted commercially
22:53:03 Quit SirFunk_ (Connection timed out)
22:53:17MattAndrewokay, so who'd like to see a commercial version of a derivative of this player?
22:53:31*GodEater raises his hand
22:53:32rasherGodEater: MattAndrew only wants to license the *schematics*, not the list of parts under by-nc-sa
22:53:37*BigBambi puts hand up
22:53:46CasainhoBigBambi: please read more:
22:53:51BigBambiCasainho: i have
22:53:54pixelmakugel: it seems like it doesn't happen for everyone and as I said, it's only happening in certain folders for me. I don't think that dircache for multivolume patch changes anything there (might have a small, even negative influence as I could "break" some more folder by chosing settings that take up more RAM)
22:53:59BigBambiCasainho: You are not making any sense
22:54:13GodEaterrasher: I understand so little about the hardware manufacturing process, I have no idea how limiting a factor that is
22:54:19linuxstbrasher: What does the list of parts have anything to do with it? Aren't the schemetics the crucial part?
22:54:21MattAndrewlinuxstb, I am *hoping* that I can build me one and a few others one and send it to them and then they will help port it. I am willing to port it myself if nobody else will, and that was what I was planning before I met Casainho
22:55:10 Join LambdaCalculus37 [0] (n=1800dac6@gateway/web/cgi-irc/
22:55:22CasainhoBigBambi: I said that because of the " contruct it myself, so a commercial version of it is the only option"
22:55:38BigBambiAnd I stand by that
22:56:09BigBambiCasainho: I am not going to build a player myself, so if a commercial version of an open player exists, that is the only way I will have one.
22:56:13rasherlinuxstb: I may have misread GodEater. It seemed to me that GodEater was telling Casainho that he couldn't even use the same parts as MattAndrew
22:56:25GodEaterrasher: no I wasn't saying that
22:56:40MattAndrewno, he was saying that he couldn't copy parts of the schematic or pcb layout
22:56:53MattAndrewwhich I don't think he intended to in the first place
22:57:08BigBambiMattAndrew: Who knows :)
22:57:14GodEaterthen I must have misread what Casainho was saying then...
22:57:17rasherStill, until you get to the "draw a schematic" phase, everything can be shared..
22:57:17linuxstbMattAndrew: The point I wanted to make was that a Rockbox port is a very hard task, even if you have full documentation, and I would expect less people to be interested in porting to a player only you have the rights to manufacture and sell.
22:57:23 Join saratoga [0] (n=9803c6dd@gateway/web/cgi-irc/
22:57:57Casainhosorry, but I can copy parts of the schematic, since Matt is also copyning from application notes and things like that! also there is the DSPdap from Spark, there is a lot of Open hardwares htat we can use! :-)
22:58:05GodEaterI would be thoroughly discouraged to buy something from someone that I knew would never be open to improvement by anyone else
22:58:12BigBambiCasainho: No, you can't copy the schematics
22:58:15BigBambiThat is the point
22:59:09CasainhoBigBambi: but can I copy parts of the schametics??
22:59:12 Nick fxb is now known as fxb__ (
22:59:22GodEaterCasainho: no - because then you're making a derivative work
22:59:39*amiconn wonders about the flash rom discussion earlier
22:59:46MattAndrewI still think it's possible to have a commercial player build from this design or its derivative, it's just that it'll be under our (meaning my if I'm the sole designer of the schematic and pcb layout or truely "our" if someone takes my design and creates a derivative work). I think the control for doing so should remain with "us" (me+fellow schematic+layout designers) instead of "them" (any random company that might try to ti
22:59:47MattAndrewvo-ize the design).
22:59:49Casainhowell, I think I can copy from the same original sources that Matt are copying!!
23:00:14*GodEater decides it is time for bed
23:00:17 Quit GodEater ("There are two major products that come out of Berkeley: LSD and UNIX. We don't believe this to be a coincidence.")
23:00:29BigBambiCasainho: There are no sources for the schematic and PCB layout, that is Matt's design
23:00:41*LambdaCalculus37 decides that maybe he'll wait till tomorrow to discuss fixing disktidy.c
23:00:44 Quit LambdaCalculus37 ("CGI:IRC 0.5.9 (2006/06/06)")
23:00:58linuxstbCasainho: The DSPDap design is also under the non-commercial CC license, so you can't copy those sources if you want people to be able to manufacture the design commercially.
23:01:04 Part karashata
23:01:10CasainhoBigBambi: there are sources of parts of the schematics :-)
23:01:11MattAndrewCasainho, you can copy the parts of the schematic that I copied from app notes. Or, to say it another way, you can just copy 'em direct from the app notes
23:01:19 Join z35 [0] (
23:02:12Domonoky_maybe it would be enogh to have someone to hold the complete copyright, then he can re-licence it for a manufacutrer we choose..
23:02:19BigBambiCasainho: Look, any parts of Matt's design that are not directly out of the docs (in which case why bother looking at his design) YOU CANNOT COPY
23:02:27MattAndrewDomonoky_, that's what I'm saying
23:02:50Domonoky_so like give the copyright to some sort of rockbox organisation or such..
23:03:04MattAndrewis there such a legal entity at this time?
23:03:26Casainhothe question here is that this kind of hardware is very simple and we can design quickly taking ideas from here and there! - also sources from Matt don't are important because he will not use a Free/Open source software to design it... we will must do a new design from start, with a free/open osftware like Kicad and not hte Egale
23:03:30BigBambiDomonoky_: If one person holds the copyright, 'we' cannot chose anything
23:03:32MattAndrewit's spread out over the globe, so it might be difficult to get laws to mesh together
23:03:47saratogai don't understand how this argument is happening
23:04:08saratogais there some disagreement on using some sort of open license for the hardware
23:04:16saratogaif not, what is the problem
23:04:21rasherBigBambi: depends how trustworthy that person is
23:04:25MattAndrewdesigning something with eagle doesn't limit how it can be used in any way, as long as the designer paid for the eagle license once.
23:04:48BigBambirasher: Of course, but it is still out of your hands
23:04:53linuxstbsaratoga: Basically, MattAndrew is designing a player and is using the non-commercial CC license for it. Others feel this isn't a license we want for an open Rockbox hardware project.
23:05:35saratoganon-commercial CC would appear to be a poor choice of license
23:05:39Lloreansaratoga: One piece of the hardware is already licensed solely noncommercial to one person, meaning only he can make commercial players using that part of the design (meaning any other contributors to the design as a whole would need to replace that first)
23:05:41CasainhoMattAndrew: about eagle, I want to do cooperative working, we need to easy for everydeveloper, like GCC-ARM, for example...
23:05:47BigBambisaratoga: Hence the argument :)
23:06:21MattAndrewCasainho, I tried using geda first. I crashed every 5 minutes. I havn't tried kicad yet
23:06:26scorche|shCasainho: if he is putting it under NC, what cooperatve working would you be doing?
23:06:34BigBambiCasainho: You need to sort out the licence before you can do any co-operative working regardless of the design software
23:06:42 Join webguest009 [0] (n=d55988b9@gateway/web/cgi-irc/
23:06:46CasainhoMattAndrew: is your hardware, design it with the program you want! :-)
23:06:59*BigBambi bangs head against wall
23:07:02linuxstbDoes anyone have any suggestions for a license for hardware designs? (I don't...)
23:07:44kugelmarkun: I just noticed jott gave you the anti-aliased font patch, what do you think about this?
23:07:47Domonoky_MattAndrew could also licence it as NC to the world, and with another licence to the rockbox community..
23:07:54Casainhoscorche|sh: cooperative in other version, from start!
23:08:07saratogalinuxstb: GPL or BSD would make a lot of sense I think
23:08:10***Saving seen data "./dancer.seen"
23:08:19MattAndrewgpl is for software
23:08:38 Join gromit` [0] (
23:08:38saratogayour design is essentially software until its fabbed
23:08:59Casainhofor me GPL but as this something that everyone can copy and use and we will not be able to "force" nothing, I would go for the BSD
23:09:17MattAndrewcreative commons was started to come up with a gpl-like license for other kinds of works. A schematic is a picture. A pcb layout is a picture. creative commons is the license of choice for these kinds of designs
23:09:30LloreanGPL is for software and "similar works" or something like that. It basically applies to any design that could be described by software too, I should think.
23:09:42saratogayour design is actually not a picture most likely
23:09:49CasainhoI would like to see some chinese company make this harwadre without ask permission, as they do the S1MP3 players :-)
23:09:51Domonoky_Casainho: if you really make another shematic than mad, make sure to use nearly the same parts and the same connections to the cpu, the you can reuse code.. :-)
23:09:53saratogaits probably a list of nodes and a description for how they interconnect
23:10:24MattAndrewbut also the specific place for where I routed those nets
23:10:39saratogasure, thats part of the node description
23:10:43webguest009hehe.. I think all the discussions about incompabilities between open source licenses are quite academic. The copy-shop industry that exists in some countries are not very likely to care. I mean they even copy stuff that are very explicitly forbidden :)
23:11:03BigBambiThat doesn't mean the licence is not worth bothering about
23:11:08saratogawebguest009: I doubt anyone cares about that
23:11:19saratogaits important so that people working agree on what they're going to do
23:11:30saratogaif someone else wants to copy it later, its not our problem or concern most likely
23:11:38BigBambiOr *understand* what they are doing
23:11:47 Quit Domonoky_ ("Trillian (")
23:12:09MattAndrewyeah, but you can argue that about non-open source stuff too. If a company in china manufactures something that's an exact replica of an apple ipod, then there's not much anyone in this country can do, short of talking politicians into an embargo. worldwide copyright protection is a dubious claim
23:12:23LloreanMattAndrew: For me, the only real sticking point is that, were I contributing to the overall player design as a "Rockbox Open Player" I would feel not only uncomfortable with someone unilaterally designing a part of it, but also with the fact that if we decide we wanted a modified version of that part for the final model because some of us saw improvements, despite the fact that we improved it our derivative would still be forcibly noncom
23:12:42CasainhoI still thnik there is no problem since Matt hardware have a few things that will be changed in future, things that were not discussed by everyone, that he choosed by himself
23:13:02BigBambiCasainho: What?
23:13:44Nico_Ppixelma: re - is this r15444?
23:14:23Casainhothere are parts in hardware that were choosed by Matt, wic I thnik are not realy important for the main objectives of the project
23:14:42pixelmaNico_P: no, r16104
23:14:54pixelmaeh... 16105
23:15:09Casainhoso, Matt is working a bit alone
23:15:09BigBambiCasainho: I still don't know what relevance that has to the discussion
23:15:18Nico_Pthese two commits seem to have higlighted many issues
23:15:29MattAndrewso how 'bout I put a big disclaimer at the top of the RockboxPlayerPrototypeB wiki page giving a summary of the two sides of this argument, as a disclaimer of how contributing to it might be worthwhile or not, depending on your interpretation of the license and how it can/will be used? Then people can decide for themselves whether they want to help or discuss ideas on a case-by-case basis. Sound good?
23:15:32 Quit webguest009 ("CGI:IRC")
23:15:40BigBambiMattAndrew: A good idea
23:15:43 Join gevaerts [0] (
23:15:43 Quit seablue ("life, death, life, death")
23:16:11 Quit merbanan (Remote closed the connection)
23:16:29rasherAnd still - the initial phase can work both ways at any rate as long as it's the schematics that will be put under by-nc-sa
23:16:36MattAndrewI will try to write something up and make it fair and somebody else (from the "other" side of the argument) can read it and edit it if he likes. Also good?
23:16:41CasainhoBigBambi: If Matt continues working alone, desiding for himself, he will be alone....
23:16:50 Quit ch4os_ ("Lost terminal")
23:17:01BigBambiMattAndrew: Yup :)
23:17:12*gevaerts has just been using his player to actually listen to stuff
23:17:25BigBambigevaerts: Stop it and get back to work! :)
23:18:21 Quit gromit`` (Read error: 110 (Connection timed out))
23:18:43*gevaerts unplugs the earphones and plugs in the usb cable
23:18:43 Quit mf0102 ("Verlassend")
23:18:50MattAndrewokay, then. I need to go make a sandwich...
23:19:06CasainhoI will go sleep :-)
23:19:14rashergevaerts: that
23:19:18rashergevaerts: that's more like it
23:19:20Casainhobye bye :-)
23:19:26 Quit Casainho ("ChatZilla 0.9.80 [Firefox]")
23:19:48*scorche|sh sighs
23:20:13rashergevaerts: nice to get confirmation that it was in fact the MULTI_LUN thing causing infirit problems
23:20:34rasherI wonder why this hasn't been more of a problem
23:20:44linuxstbMattAndrew: The main issue to me is that I don't want to see Rockbox split into different sub-projects with different licenses.... So all hardware projects should share the same license, so they can share schemetics etc without worry of legal issues.
23:21:04gevaertsrasher: yes. I'm also pretty sure that the IO errors he saw came from simultaneous access by other stuff
23:21:10linuxstb(but I'm just one person, others may not care...)
23:22:00Lloreanlinuxstb: I agree with that wholeheartedly, though I'd add that, to me, it's also nearly as important that all contributors have equivalent/equal rights over the final product. IE, nobody can use the whole product commercially (just their individual contributed bits), or everyone can.
23:23:38LloreanWhich most licenses grant anyway, so it's more a case of making sure contributors know what they're getting into.
23:24:48LloreanNico_P: Regarding the default theme, they picked a larger font because they felt the physical size of the screen was small enough that the larger font was needed for readability (I don't agree, but that's the way the discussion worked out)
23:25:15pixelmaNico_P: not sure if it means something but I noticed when testing the skipping problem I can press "stop" during the first buffer and it only takes effect when it's finished buffering, whereas it immediately stops in a folder without this problem.
23:25:48pixelmathe former in a folder with the problem, obviously
23:25:55Nico_PLlorean: then I don't understand why the ipod video version doesn't use a larger font too (although I find it fine as it is)
23:26:19LloreanThe screen is less tall, so you have less lines you can fit maybe?
23:27:00pixelmasmaller font? I still think the font(s) chosen are too small, at least to "light"
23:28:16Nico_PLlorean: maybe, but the end result on the gigabeat is a really cluttered screen. at least they could take out the clock and make AA a bit smaller
23:28:31LloreanNico_P: I agree, just letting you know why it was the way it was.
23:28:39LloreanPeople seem to think AA needs to be "as large as you can fit" =/
23:29:07LloreanMaybe there could be a "View Album Art" option in the context menu for those sillies. =P
23:29:44Nico_Pmy goal in posting the pics on the forum thread was to spark a discussion and hopefully end up with an improved version
23:29:58LloreanGood hope
23:30:58pixelmaNico_P: re. 15444, isn't that the "easy way out"? Maybe the same applies to 16105 too...
23:31:10jhMikeSis it me or does the mc13783 "datasheet" tell you next to nothing?
23:31:15 Part ian_hawdon
23:31:33 Join spiorf [0] (
23:31:50n1sjhMikeS: I got that feeling too after first reading the imx31 datasheet with details aplenty
23:31:54Nico_Ppixelma: what do you mean by "easy way out"?
23:32:43jhMikeSIt has no descriptions of registers themselves. I even read a mention about that being NDA.
23:34:00n1sjhMikeS: there are drivers in the source of the linux image for the reference design for imx31 that might give a clue
23:34:34jhMikeSwhat does the mc13783 have to do with the imx31?
23:34:48n1sit is used on the reference board too :)
23:35:24Nico_PjhMikeS: what is the mc13783 used in?
23:35:28Nico_Pand for?
23:35:35jhMikeSNico_P: It's the S's PMU
23:35:54Nico_Phave you looked at USB btw?
23:37:01jhMikeShaven't gotten into that yet
23:37:36 Quit petur ("gonne")
23:37:37jhMikeSI did get the backlight to turn on and off though :)
23:37:42Nico_Pnice :)
23:38:19pixelmaNico_P: I mean that reverting those changes maybe don't fix the problem just hiding them again
23:39:16n1sjhMikeS: I have a ~2mb patchfile with the drivers in it her if you want
23:39:27 Join asdrubal [0] (
23:39:28Nico_PI managed to get the USB stack to compile for the S, but when I saw how usb-imx31.c was empty I gave up
23:39:42Nico_Ppixelma: I do agree
23:39:45jhMikeSn1s: thanks. link me up.
23:40:09jhMikeSNico_P: I'm not sure but this mc13783 may come into play somehow with that
23:40:32Nico_Ppixelma: I don't intend to revert those commits. knowing they highlighted the issues can be helpful though
23:40:49Nico_Pand if people can't live with the issues, they can apply the band aid
23:44:13 Quit pixelma (" reboot")
23:44:49jhMikeSI'm just thinking about doing whatever in order to get the info out of them
23:47:35jhMikeSof course perhaps everything is in there too and it's just a matter of integrating it all
23:49:39stripwaxDomonoky seems to have left again - would anyone else be interested in committing a new button .png for rbutil?
23:49:42 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
23:50:55XavierGrSlasheri: ping
23:50:57 Quit spiorf (Remote closed the connection)
23:52:42 Quit J3TC- (".UPP.")
23:53:09 Quit markun (Remote closed the connection)
23:57:46 Quit conando (Read error: 110 (Connection timed out))
23:58:18pixelmaNico_P: can you remind me what I would have to put into playback.c for the logf build?
23:58:37Nico_Pyou just need to uncomment LOGF_ENABLE toward the top
23:58:49 Quit azdrubal (Connection timed out)
23:58:58pixelmathanks, will try

Previous day | Next day