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 2005-12-13

00:00:23Jungti1234I'm going to require much informations to him.
00:01:23MoosJungti1234: you know there are South Korean racist in the world too, they thought the south korean are isolated from the world, dictatorship...
00:02:23Moossure there is dictatorship, but they don't have right to judge human
00:03:34Jungti1234Isn't so
00:04:06Jungti1234What mean is 'Sure there is dictatorship'?
00:04:14peturQUICK reply if possible: where could I force a setting at startup best?
00:04:34thegeek_I've gotten addicted to the foobar2000 dolby headphone plugin
00:04:38thegeek_I can't stand music without it
00:04:46MoosJungti1234: sorry I mixed South and north
00:04:59thegeek_I'm seriously entertaining the possiblity of transcoding all my music through foobar with it enabled
00:05:16Jungti1234South and north are much different.
00:05:24Moosyes sorry :)
00:05:48MoosI don't know why I thought you was North Korean
00:06:25Jungti1234North Korea can't make internet freely.
00:06:48Moosunfortunatly, not a lot of things freely
00:07:21Jungti1234Do you know 'Nanta'?
00:07:32Moosno, what is it?
00:08:16 Quit Rob2222 ()
00:09:30 Join elinenbe [0] (
00:09:30Moosit's pitty "merging" North and South Korea could make one powerfull countrie
00:09:38*petur is convinced allowing backlight value 0 is a big mistake. Even more if it is default...
00:09:59peturbtw, we have coonfigurable backlight setting on H300!
00:10:03mirak_FireFly_: you fix flicking for witch device ?
00:10:21MoosSalut Karim :)
00:10:48mirakI was reading the H300 lcd datasheet and they say there is something to kind of freeze the refreshing
00:10:51mirakMoos: salut
00:11:17Jungti1234nanta is cookin
00:11:33Moosmirak: still learning how to make video thing?
00:11:38 Join olaf_ [0] (
00:11:45MoosJungti1234: ok
00:11:47Jungti1234I see nanta.
00:11:47mirakyes I was reading jpeg and mpeg stuffs
00:11:52 Quit Jungti1234 ("bye")
00:11:58olaf_Hi there
00:12:01mirakMoos: the theory
00:12:07mirakand watched xvid decoder code
00:12:15mirakI am far from having something, ond't worry
00:12:29Moosthat will be big thing to do
00:12:31olaf_anybody knows if repeat A-B works for iriver rockbox
00:12:48Moosmirak: maybe needed more than one dev in this
00:13:19Moosmirak: c'est du lourd ;)
00:13:22_FireFly_mirak: mainly for the remote from h1xx-devices
00:13:39_FireFly_but that should als reduce the flicking a bit for the h300
00:14:00_FireFly_someone has also mentioned that the flicking is gone on h300
00:14:15mirakI had it when using the boxed interface
00:14:29mirakhaven't tried since
00:14:58olaf_mirak: answering my question ?
00:15:04mirakMoos: yes certainly, but if I know someone is already doing it, I will not try
00:15:05_FireFly_hmm i have just discovered that my flicker-patch isn't needed when my new-patch is used
00:15:19mirakolaf_: I think it doesn't
00:15:19Moosmirak: hehe :)
00:15:27_FireFly_it seams at least to me
00:15:38olaf_mirak: yep.
00:15:42peturnew patch?
00:17:03_FireFly_petur: my image-flicking-patch has increased the ticking-noise for the h1xx devices
00:17:16_FireFly_and my new patch reduce it to the "old-value"
00:17:21Moosohoh, 5 hours remaining before wake up, time to go here
00:17:31_FireFly_Moos: me too ;)
00:17:32Moosgood night or day at all
00:17:39mirakwhat is that ticking noise ?
00:17:42Moosnighty FireFly
00:17:45 Quit Moos ("Glory to Rockbox")
00:18:15_FireFly_mirak: the ticking is afaik only on h1xx-devices and there one noticealbe when using the remote
00:18:29peturcan you read settings via the plugin api?
00:18:42mirakoh I heard that with H300 remote also
00:18:54mirakI don't have a remote
00:19:50_FireFly_petur: but now i have discovered that my flicking patch isn't needed with the new patch because the new patch updated the lcd only on one-place -> so it seams no flickering of images is possible anymore
00:20:04_FireFly_i have the build currently running and the images doesn't flicker
00:20:25peturI still saw flickering when text and images are mixed - solved as well?
00:20:29_FireFly_on the remote
00:20:35 Quit PaulJ (".")
00:20:44_FireFly_the patch is on tracker
00:20:55peturI'm talking main display of course - don't have a remote
00:21:27_FireFly_petur: as i had already said this doesn*t matter because it's one code for all displays :)
00:21:32peturI'll soon post a patch for configurable backlight brightness on H300
00:22:06_FireFly_so if i change the wps-code the change is for both displays on the h1xx/h3xx devices
00:22:18peturjust need to find out how to set a decent value if the setting doesn't exist yet
00:22:35peturMy first try gave 0 and a black screen
00:22:43_FireFly_at leat at that palces which my patch changes
00:22:47peturused a bright light to navigate to the settings
00:22:49mirakpetur: !
00:22:54peturthank god it worked
00:23:02_FireFly_petur: what ??
00:23:21peturbacklight brightness?
00:23:40_FireFly_ah yes i thought the flicking-problem is gone on the h3xx ;)
00:24:02petursorry, still busy with my own patch
00:24:54_FireFly_np but now i have to go to bed ;)
00:25:00_FireFly_good night every-body
00:25:07peturis it on patch tracker?
00:25:46 Quit _FireFly_ ("Leaving")
00:26:20peturwhat was the way to reset RB settings again? hold some key?
00:29:01Bagderbut not too early
00:29:13 Join LinusN [0] (
00:29:31LinusNpetur: if the settings are incompatible, you must bump the config vlock version
00:29:44peturI added it at the back
00:29:56peturand used pwm '9' as default
00:30:17LinusNok, then you should be fine
00:30:17peturyou read what I did first? default 0. lol
00:30:22peturgot a black screen
00:30:24LinusNsaw that :-)
00:30:43peturI'll submit the patch before bedtime
00:30:52peturwhat to do with language files?
00:31:35LinusNyou only need to do the english one
00:32:14peturyou know why it flickered?
00:32:34peturHDD access at low intensities, iRiver FW does it as well
00:32:37 Quit _DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
00:34:28peturcan't get my settings to reset
00:34:43peturI hold REC when the bootloader starts?
00:35:39Bagderno, when rockbox starts
00:36:21peturI held it until it loaded completely, didn't reset my setting
00:41:22 Quit actionshrimp (Read error: 104 (Connection reset by peer))
00:42:01 Join actionshrimp [0] (
00:48:47olaf_anybody knows if repeat A-B works for iriver rockbox (repost) ?
00:49:06Bagderthat is it doesn't work
00:49:20olaf_just saw there seems to be support for Archos
00:49:31olaf_and just posted a feature request
00:49:56 Join dropandho [0] (
00:50:54peturnew entry in LAN_xxx file: also at the back?
00:51:16olaf_Bagder: thanxs for the reply. Hope the new feature request makes it sometime
00:53:53 Quit olaf_ ("goodnight")
00:55:08dropandhohey all!
00:55:14dropandhoany word on the forums?
00:56:17dropandhois jeff mia or wha?
00:58:08 Quit yngwi ("Chatzilla [Firefox 1.5/undefined]")
01:00:42 Quit ender` (" Engineers think that equations approximate the real world. Scientists think that the real world approximates equations. Ma)
01:01:38 Quit Kingstone (Read error: 104 (Connection reset by peer))
01:01:40 Join Kingstone [0] (n=sg1@
01:01:49 Quit Kohlrabi ("Leaving")
01:03:04dropandhopeople try have attempted to contact him...
01:03:17dropandhoso no one has had any luck?
01:03:41BagderI don't know
01:04:49dropandhoLinus - u around?
01:04:54dropandhoany insight?
01:09:40 Join StrathAFK [0] (
01:09:42 Quit Strath (Read error: 104 (Connection reset by peer))
01:13:58***Saving seen data "./dancer.seen"
01:16:13peturhehe.. running my battery flat without playing one song. Just code testing :D
01:18:05peturThat patch of _FireFly_ is perrrrfect!
01:18:10peturno more flickering!
01:18:31dropandhoyou got a wps fliker fix patch?!
01:18:58petur_FireFly_ made one
01:19:14peturthis one really works!
01:19:16RickROFL @ Dance Puff theme
01:19:51dropandhoso guys...what can we do to get some forum love here?
01:19:57peturLinusN: around?
01:20:07Rickforum love?
01:20:13dropandhoget it fixed
01:20:35dropandhopeople have IMed Jeff and emailed well as linus
01:21:01peturall right.. off to bed.. goodnight!
01:21:09 Quit petur ("here today, gone tomorrow")
01:21:22dropandhoany1 have any clue what to do next?
01:21:26Rickodd that changing themes takes so long
01:22:02dropandhomaybe start a thread on misticriver?
01:22:55dropandhomisitic river is jeff's site
01:27:37LinusNi'm around now
01:27:51dropandhohehe- you love to hmmmm
01:27:56Rickhas anyone noticed the problem of accessing the 'Rockbox Info' via remote on H120?
01:28:05Rick(you can't exit the info screen from the remote, nor do you see it on the remote)
01:28:10dropandhohey Linus...just wondering you had any word on the forums
01:28:24Ricksame goes for version
01:28:35Rickactually no, Version is fine
01:28:44LinusNjeff mentioned something about an IP address change
01:30:02dropandhoahh- points to
01:31:40dropandhoguess it is out of my hands...let me know if there is anything i can do to help
01:34:09 Quit mirak ("Ex-Chat")
01:36:07 Nick StrathAFK is now known as Strath (
01:40:35dropandholinus- i was also wondering if you ever got that ticking unit
01:42:51LinusNoh, yes, i got it today,!
01:43:11LinusNforgot to tell you
01:43:17dropandhowhat timming! hehe...sorry to bug you
01:43:21dropandhojust curious!
01:43:34dropandholookin fwd to what you discover
01:43:51dropandhothanks again..and to the sender!
01:50:56LinusNthe analog and digital ground are not properly isolated on the ticking board
01:51:41dropandhothose punks! how do you mess that up?!
01:52:58dropandhoi remember the sender said that he didn't have some things on the board you were looking for
01:53:58LinusNjoystick and hold switch
01:54:57LinusNhis board is in fact newer than mine
01:55:28dropandhohehe- now that is would think they would have corrected the problem further into production
01:56:21LinusNlooks like it's the other way around
01:58:04 Join nathanh [0] (
01:58:06dropandhoi thought he was missing something like r22 as well
01:59:16dropandhoi forget was in the forum..hehe
01:59:38 Join leftright [0] (
01:59:43nathanhi cant express enough how grateful i am to the rockbox dev team for the H300 port... gapless playback, car charger mode, on-the-fly playlists... even in beta it's 10x better than the stock firmware :-)
02:01:14leftrightI have just spent a few hours fiddeling with my gf's new Archos gimini XS202, the unit itself is nice, tiny, great from factor, but the firmware in comparison to R-box sucks
02:02:12leftrighteach time you nav out of the wps it stops playback, not nice at all, plus its slow to react to key presses
02:02:59leftrightthis has made me appreciate rockbox even more
02:04:31leftrightthe OTG playlist feature of the XS202 is not OTF, what are those guys at Archos thinking, each time you want to add something to the playlist playback needs to be stopped ? weird
02:05:11dropandhothey aren't rckbx maniacs...thats the problem!
02:05:49leftrightsheesh, what is it with these OEM firmware designers
02:07:38leftrightanyywayyy, my girl friend is happy, she is totally ignorant of Rockbox's features, nothing to compare it to, ignotance is indeed bliss in this case
02:08:23 Join DJDD_ [0] (
02:12:15 Join webguest91 [0] (
02:12:42LinusNtime to sleep
02:12:43webguest91So about - sounds like the dns entry needs updating?
02:12:50LinusNwebguest91: yes
02:13:01LinusNi wish i knew what to put in my dns though
02:13:15*leftright gives rockbox a 0dB hug
02:13:26 Part leftright
02:13:32webguest91But doing a manual http request to with host: gives a "no site configured at this address" message
02:13:51webguest9172.29.70.123 being the ip of
02:16:15webguest91As does putting "" in /etc/hosts and checking it in firefox
02:16:45linuxstbhttp:// looks to be the forums.
02:18:19webguest91Good catch
02:18:28dropandhoweird, hitting that tries to connect to
02:19:23linuxstbI found it by using as a dns server - i.e. "host"
02:22:09 Quit mikearthur (Read error: 104 (Connection reset by peer))
02:22:41dropandhowe want our forums! hehe
02:24:33webguest91dropandho: trying to connect to is no wonder: <link rel="stylesheet" type="text/css" href="" />
02:25:14dropandhogotcha- so what to do?!
02:25:46webguest91Put "" in your HOSTS file
02:25:50webguest91Or wait.
02:25:53webguest91I guess
02:26:09webguest91It's still hella slow.
02:26:33linuxstbslow or dead? They were working fine when I first tried them, but now I can't access them.
02:26:42webguest91Yeah, seconded in fact.
02:27:43webguest91Strange... works if I telnet and do the http manually
02:28:12dropandhosome1 get that man jeff!
02:28:30webguest91I think it's just Firefox going crazy
02:28:33webguest91Lynx works as well
02:29:16dropandhoso jeff needs to setup the dns entry and we might b there?
02:29:19linuxstbI see the problem - src="" in the header.
02:29:42linuxstbSo every page is still trying to get the javascript file from the old IP address.
02:30:10 Join MisticJeff [0] (
02:30:23webguest91With a hardcoded IP at that
02:30:30dropandhospeak of the devil...and he shall appear- hehe
02:30:32linuxstbSpeak of the devil....
02:30:32Weazel_there he is
02:30:49linuxstbWhat's up?
02:31:20webguest91Hah, firefox finally gave up, it seems.
02:31:24webguest91and loaded the page
02:31:51MisticJeffDon't have a clue how to get the forums back up and running... it's an IP issue I believe
02:31:59MisticJeffbut not sure where to change it
02:32:38LinusNMisticJeff: start with changing the IP references in the forum html
02:32:45LinusNi have updated the dns
02:32:53 Join Paul_The_Nerd [0] (
02:33:00dropandholinus- i thought you were sleepin...get to bed young man!
02:33:16LinusN"young man" :-)
02:33:28dropandhohehe- why not?!
02:34:34LinusNit will probably take a while before the dns update propagates
02:34:55LinusNMisticJeff: the forum html code is full of absolute references to the old IP address
02:35:17linuxstb(which was
02:36:05LinusNlinuxstb: odd, the old dns entry pointed to
02:36:33linuxstbMmm. The html says 137 - maybe that's why the IP address was used - those files were on a different server.
02:36:46LinusNjeff should know
02:36:47webguest91Maybe the javascript files were on or something
02:37:00webguest91But yes
02:37:04webguest91Paging MisticJeff
02:38:23Weazel_http:// asks me a login and password could this be something?
02:38:23 Quit nathanh ("CGI:IRC (EOF)")
02:38:48MisticJeffguys this could be a loooooonnnnnngggggg process since i'm not a web guru believe it or not
02:38:57MisticJeffwill take me a long time to figure this out
02:39:22linuxstb returns a javascript file OK.
02:39:37dropandhojeff, its ok- we are here for support!
02:40:43Weazel_mental support
02:40:44linuxstbMisticJeff: The first (and maybe only) thing will be to do a search and replace in the html files and change all references to "" to ""
02:41:16dropandhowe call this "dynamic"...right boys?
02:47:08LinusNMisticJeff: why did the ip address change?
02:47:48 Join Lynx0 [0] (
02:48:34lostlogicugh, apparently my DNS cache for the forums hasn't expired yet *manually hunts up the IP*
02:48:39MisticJeffthe company that I lease the server from purchased their own IP addresses rather than leasing them as they were before
02:49:21MisticJeffi've looked through a handful of files and find no reference to the old IP... not sure where to look gang
02:50:08webguest91Is there a web-frontend to administering SMF?
02:50:34MisticJeffyes but the problem is that I cannot login to the admin area because of this issue
02:50:46MisticJeffsince the login is java script
02:51:01lostlogicMisticJeff: and I take it your DNS is updated and/or you have put in your hosts file?
02:51:34MisticJeffwhat's a host file?
02:52:41lostlogicI forget where it lives in windows, but there's a file that you can locate named hosts with ip -> name pairs, so you could put: in there to make sure it's not a DNS problem keeping you out.
02:53:27dropandhoit is in- C:\WINDOWS\system32\drivers\etc
02:53:29LinusNfyi: the dns is
02:53:53LinusNand i have updated it with the new ip
02:54:09LinusNi wich i would have known about the ip change beforehand
02:54:31dropandhoall you need to do is open up the hosts file in wordpad....and add that line below everything on that page
02:54:49MisticJeffwish i would have known about it before too.... i've been on vacation... just got back 2 days ago
02:54:49lostlogicLinusN: I'm sure it's just cached in Jeff's ISP −− I had to reboot my local bind to force it to update.
02:55:14dropandhowelcome back!
02:55:20dropandhojeff- does this make sense to you?
02:55:37dropandhoi can send you a screen shot if you like
02:56:00MisticJeffthat's why for the past couple of years i've surrounded myself with knowlegdeable people since i don't know dick about this stuff
02:56:08linuxstbMisticJeff: How do you edit the files on the server? Do you have ssh access?
02:56:14dropandhoits all good
02:56:22linuxstbSomehow you need to search for the IP address in all the files.
02:56:45MisticJeffdon't know how to use SSH, i simply download them through FTP, edit and then reupload but luckily don't have to do it very often
02:57:17lostlogiclinuxstb: do you really think the IP is directly referenced somewhere in the files?
02:57:34MisticJeffthere's no way i'm going to search through 100 + files for the IP address..... simply don't have the time or patience
02:57:35dropandhoso maybe you should send one of us the files (if you are comfortable with that)
02:57:36linuxstblostlogic: Yes - do "wget" and look at the html
02:57:59linuxstbMisticJeff: Of course - we need to find an easy way to search them.
02:58:22MisticJeffguys, unless you know what files you're looking for i'd have to send you the whole thing, ain't happenin'
02:58:30linuxstbDo you have copies of every file on your computer, or do you just transfer them individually?
02:58:55MisticJeffyou mean the Rockbox Forums files....?? they are on the server
02:59:23dropandhojeff i didnt follow....would you be willing to send one of us the files- zipped up?
02:59:27linuxstbYes, I meant those files. If you could copy them all to your computer, you can search them easily using the Windows find program.
02:59:38MisticJeffhold on guys
02:59:38dropandhothere u go
03:00:21linuxstbMisticJeff: Do you know the name of the forum software you are using?
03:00:40webguest91there's a link at the bottom of eacj page
03:00:47linuxstbLet's search for SMF and see how it organises its files.
03:00:51lostlogicit'll be probably in the index.php file?
03:01:28MisticJeffit's not there
03:01:42MisticJeffi've given server access to LinusN
03:01:53MisticJeffi'm sure he'll be able to find out something
03:01:55dropandhothat would do it- nice
03:02:00webguest91I'm guessing it's in some database
03:02:02dropandhou trust that german?!
03:02:45dropandhooops! that was my bad
03:02:49dropandhoi was tryna make a joke
03:02:53dropandhoand totally ruined it
03:02:55dropandhomy bad!
03:03:50*webguest91 downloads SMF 1.0-rc1
03:07:22dropandhomaybe it lives in something called- CPanel
03:07:44webguest91Why would you say that?
03:07:57dropandholooking at their forums...just guessing tho
03:11:59dropandhothey also made a near repair tool
03:12:03 Join tvelocity [0] (n=tony@
03:12:22dropandhothe old domain might show up all over the files
03:12:36dropandhoand i believe this tool repairs such things
03:13:50LinusNhmmm, i can't find any absolute ip addresses when i read the html from
03:13:54 Quit Lynx_ (Connection timed out)
03:14:02***Saving seen data "./dancer.seen"
03:14:39 Nick Lynx0 is now known as Lynx_ (
03:15:50 Join ehntoo [0] (
03:16:14dropandhoare u able to run the forum software linus?
03:16:38LinusNin fact, the forum works fine from here, when i use as my primary dns appears you need to run repair_settings.php
03:17:22dropandhothats what they seem to say on their forum
03:17:38LinusNdropandho: i don't see why, it works fine without it
03:17:40webguest91Which version is that for though
03:17:50webguest91doesn't seem to exist in 1.0-rc1
03:17:59dropandhou need to d/l the tool
03:18:23webguest91Ah, thought it was a part of the thing
03:18:24 Quit actionshrimp ("a bird in the bush is worth two in your house")
03:18:33dropandhobut the folks seem to say thats what u run when you do a domain change
03:18:53LinusNbut we haven't done a domain change
03:19:13LinusNonly changed the ip of the existing domain
03:19:31lostlogicyeah, it's fine as is −− the view source from the page shows the correct IP to me now that my DNS server is updated.
03:19:40 Join Cassandra- [0] (
03:19:41lostlogicjust need to wait for caches to reset
03:19:46LinusNlostlogic: same here
03:20:01lostlogiceveryone should run their own DNS server... or something.
03:20:14LinusNthat usually takes a day or so, i believe
03:20:19MisticJeffdo you want me to reboot the server???
03:20:25LinusNno need
03:20:25lostlogicMisticJeff: won't change anything
03:20:34dropandhotime is on our side
03:21:05MisticJeffok then... i need to run and once again A BIG THANKS TO ALL THE ROCKBOX FOLKS!!!!!!!!!! YOU ALL ARE GREAT!!!!!
03:21:10LinusNit'll probably be fine tomorrow
03:21:14lostlogicMisticJeff: it's DNS servers all over the 'net that won't update the cache for however long the refresh time is set in the SOA record for
03:21:22LinusNMisticJeff: thx
03:21:30dropandhothanks jeff!
03:22:03MisticJeffLinusN: need anything for the H300 project??
03:22:42LinusNMisticJeff: a shitload of money .-)
03:22:50LinusNheh, i need time
03:22:54 Quit Cassandra (Read error: 113 (No route to host))
03:22:54 Nick Cassandra- is now known as Cassandra (
03:22:56MisticJeffhmmm... ok
03:23:20 Part MisticJeff
03:23:39dropandhohehe- you confused him!
03:23:42lostlogichehe, rockbox has already singlehandedly made me start using my H340 again after it sat around donig nothing for about a month because I was bored of it :)
03:23:45dropandholike i was confused where you are from
03:25:53LinusNtime to sleep
03:25:55LinusNnite all
03:26:04 Part LinusN
03:29:43dropandhoso does anyone know what is up with the new tick fix?
03:29:57*Weazel_ has ihp120
03:31:54g33i still think songs sound sharper and not as good in rockbox
03:32:49dropandhowhat ver do u have
03:35:27Paul_The_NerdHasn't analysis shown that the outputs are very nearly identical?
03:36:54g33i havent done any serious A/B but its just a general feeling
03:37:09Paul_The_NerdI remember someone showing a comparison of the waveforms or something, a while ago.
03:37:13*Paul_The_Nerd shrugs
03:37:15g33except for this one really low quality song, which sounds worse for sure
03:37:21Paul_The_NerdAre you listening to MP3s, Oggs, or WAVs?
03:37:35g33but that one was a 34khz 56kbps mp3
03:37:35Paul_The_NerdAnd, are the songs being resampled?
03:37:43 Quit ehntoo (Read error: 104 (Connection reset by peer))
03:37:44g33err 24khz
03:38:14g33ok lemme do a test right now
03:38:17webguest91The resampler is not too good. But 44,1khz files are pretty much identical
03:38:39g33do you know if the volume affects the optical output?
03:38:43webguest91At least the frequency response etc.
03:38:48webguest91Don't think it does
03:41:04*g33 is waiting for his new dac
03:42:05g33hmm nope this song sounds pretty much exactly the same
03:42:28 Join Timothy [0] (
03:43:07g33and this one is 44khz 192k mp3
03:43:30 Join amiconn_ [0] (
03:43:41TimothyHi, has anyone had any problems with the simulator not being able to open any .rock files? It looks like it's trying to run them as a windows executable.
03:43:47g33cant wait till i get my dac though... aaawhh
03:44:51Weazel_do you concert tape also with your ihp?
03:44:51TimothyWhat is going on with it, I thought it wa supposed to allow me to test my plugins
03:46:22webguest91Timothy: which sim?
03:46:47linuxstbTimothy: You need to compile your plugins for the simulator - you can't just copy them from your player.
03:46:56Timothyoh ok
03:46:59g33sometimes i record myself playing harmonica using the mic that comes with the iriver :)
03:47:10TimothyHow do I go about doing that?
03:47:21webguest91Just like you compiled the sim in the first place
03:47:22linuxstbDid you compile the simulator yourself?
03:47:39linuxstbWhich plugin do you want to test?
03:48:07linuxstbAfter typing "make" to compile the simulator, did you type "make install" ?
03:48:26g33what is rockbox written in?
03:48:28Timothyi dont think so, ill try that
03:48:48*g33 is trying to learn c++
03:48:53linuxstb"make install" installs the plugins, fonts and all the other files in the archos/.rockbox directory for your sim.
03:49:04Timothyoh alright
03:49:27TimothyI just copied the .rockbox folder from the zip file
03:49:46linuxstbOut of interest, which instructions did you follow for compiling the sim? This seems to be a common mistake.
03:49:56 Join ehntoo [0] (
03:50:05Timothyhold on
03:51:13 Join Lynx0 [0] (
03:51:37TimothyIT actually does say to type make install, but i must have missed that. unless it was just added
03:51:59linuxstbNope - you can see the "last modified" date at the bottom of the page...
03:52:34linuxstbBut that text about enabling sounds make it's less obvious.
03:52:39Timothyyup, well i guess i'm just blind then. heh
03:52:54linuxstbOr just too keen to start the sim.
03:53:08Timothyyeah, I didnt bother with that part, so i must have missed it
03:54:16Timothycool, the plugins seem to work noe. thanks
03:57:39 Join Lynx [0] (
04:00:37 Quit amiconn (Read error: 110 (Connection timed out))
04:00:41 Nick amiconn_ is now known as amiconn (
04:03:37 Quit Timothy ("CGI:IRC (Ping timeout)")
04:05:44 Join webguest12 [0] (
04:09:41 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
04:19:36 Quit Lynx_ (Connection timed out)
04:21:45 Nick Lynx is now known as Lynx_ (
04:22:07 Quit Lynx0 (Read error: 110 (Connection timed out))
04:26:48 Quit webguest91 ("CGI:IRC (EOF)")
04:51:34 Quit dpassen1 ()
05:00:05 Part webguest12
05:05:39 Quit ehntoo ("Leaving")
05:05:52 Quit tvelocity ("Leaving")
05:14:03***Saving seen data "./dancer.seen"
05:18:50Weazel_ah forum works now
05:20:57dropandhonot here yet
05:21:03 Join ehntoo [0] (
05:21:03dropandhodamn dns!
05:23:45lostlogicrunyerown dns ;)
05:25:52dropandhoam i missing anything exciting on the forums?!
05:27:02 Join Timothy [0] (
05:31:12lostlogicyes. it's turned from a pretty slow and technical forum into a wild party with naked women since the DNS changed.
05:31:46dropandhothis is what i was thinkin!
05:31:54dropandhoany illegal substances?
05:35:18 Quit saa[b_r]ider (Read error: 110 (Connection timed out))
05:46:37 Quit ehntoo ("Leaving")
05:51:17Timothydoes anyone know what files i need to edit in order to have my plugin compiled when i do a make install?
06:01:17 Join webguest42 [0] (
06:01:25webguest42Timothy: apps/plugins/SOURCES
06:01:27 Quit webguest42 (Client Quit)
06:02:33Timothyyeah, I found that, im not sure which block to put it in for the h3xx though
06:04:16Timothyalso, when I ran make install, it said CC myplug.c but there is no myplug.rock in the .rockboc/rocks directory
06:04:20 Join Jungti1234 [0] (n=jungti12@
06:05:51Timothywhoops, im retarted. nevermind
06:09:58 Quit dropandho ()
06:13:21Jungti1234 : To iRiver in these page why is no there 'New'?
06:15:42Timothyyay, I wrote my first plugin!
06:17:50 Join NibbIer [0] (
06:18:10 Quit Nibbler (Read error: 110 (Connection timed out))
06:34:37 Join Paul_The_Nerd [0] (
06:49:10 Join YouCeyE [0] (n=YouCeyE@unaffiliated/youceye)
07:05:44 Join perplexity [0] (
07:07:31 Quit DreamTactix291 (Read error: 104 (Connection reset by peer))
07:10:31 Join saa[b_r]ider [0] (n=saab_rid@
07:14:07***Saving seen data "./dancer.seen"
07:15:49 Join DreamTactix291 [0] (
07:40:19Bgermorning :)
07:51:14Jungti1234morning, Bger
07:52:51saa[b_r]iderhello Bger
07:55:05Bgersaa[b_r]ider did u test the patch ?
08:19:27 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
08:35:11 Join ender` [0] (i=ychat@
08:35:54 Join Paul_The_Nerd [0] (
08:54:14 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
08:57:28preglowlinuxstb: the backlight timeout was dead accurate for me yesterday
08:57:36preglowlinuxstb: running at 75mhz, at least
08:57:44 Join _FireFly_ [0] (
08:58:15_FireFly_wow LinusN that was fast
08:58:28 Join einhirn [0] (
09:02:44_FireFly_it seams that my last patch makes the patch before from me obsolete
09:03:30_FireFly_it seams that due the change to only update the lcd on one place eleminates the bmp-flickering as well
09:13:00saa[b_r]iderBger: when I apply the patch, it doesn't confirm the application, or create a backup of the original button.c, am I using the wrong p?
09:14:11***Saving seen data "./dancer.seen"
09:14:52Bgerhm, what .patch do u use ?
09:16:22Bgerthis is old
09:16:42saa[b_r]iderwhat's the ID of the new one?
09:17:16Bgerit's the same, but i've changed the file
09:17:19 Quit perplexity (Read error: 110 (Connection timed out))
09:17:32 Join perplexity [0] (
09:18:06saa[b_r]iderok, I'll redownload it now
09:18:17saa[b_r]ider1367824, right?
09:19:07Bgercd to the root dir where the folders apps/ firmware/ tools/ etc are situated
09:19:45Bgerand patch -p0 < h300_lcd_remote_buttons_v2.patch
09:20:23saa[b_r]idersourceforge takes ages to load here for somereason :)
09:20:36Bagdersourceforge is like that
09:21:11saa[b_r]iderupdated yesterday, nice
09:23:38Slasheriah, "tcs.type = tag_artist; tcs.search_by_type = tag_genre; tcs.search_by_seek = extra; .." now even relative searches from tag cache are really simple :)
09:23:58Bger"relative" ?
09:24:05Bgerhm, slasheri
09:24:10Slasherithat search would lisk all artists by the given genre
09:24:44Bgerwhat about making dircache not to cashe one directory tree (for data files)?
09:24:46Timothywhere are all of the BUTTON_WHATEVER macros defined?
09:25:08Bgerconfigurable, of course
09:25:28BgerTimothy firmware/export/button.h
09:25:53preglowBger: i think that would be a bit cumbersome
09:26:30Bgerpreglow maybe, but i have one big directory D, in which i keep all non-music, rockbox etc data
09:26:39Bgerand there are many little files ...
09:27:18Bgeri suspect that about 1/3 of the cache size is from this dir
09:39:42Timothywhen you call yield(), does it save the stack? I'm getting wierd results when I use it in a for loop, and thats the only thing I can think of because it's using local variables
09:40:10BgerTimothy : of course it saves the stack
09:40:40Bgerthis is one of the ways to give the cpu to other thread (other one is sleep)
09:40:52 Join Kohlrabi [0] (
09:40:59Timothyyeah i know
09:41:09TimothyI"m writing a mandelbrot set explorer
09:41:36Timothyand im putting a yield call before it calculates each pixel, but when I do that all I see is black. If i take out that yield, it's fine
09:42:07Timothyand I'm drawing the screen all at once, so its not that if forgot the color state
09:42:37Bgermove the yield after lcd_update ?
09:43:09preglowTimothy: you know there already is a mandelbrot set explorer in rockbox, right? :>
09:43:29Timothyyeah, it doesnt seem to do anything on my simulator though
09:43:42Timothyplus its a nice way to get used to the API
09:43:48preglowbut yes
09:43:57preglowyield() works like a normal function call
09:44:04preglowyou should not have to care that anything special happens inside it
09:44:14preglowplus, it switches stacks
09:46:14Timothyman this is so strange...
09:46:31preglowoh, strange things do happen
09:47:45Timothyso im claculating each pixels color inside a for loop, storing it into an array, and then calling my display function. but if i put a yield() inside the for loop before it calculates the pixel, which is the most taxing, it just doesnt display anything
09:47:48preglowcoffee time!
09:48:20*Bagder offers preglow a cup
09:48:33preglowone's enough for now
09:48:40preglowreached new levels of tiredness today
09:49:13Timothyit also stops responding to any input
09:49:29preglowso, looks like we'll be having an ifp port soon
09:49:47Timothymaybe it's just switching tasks too often, and can't get anything done?
09:49:57preglowTimothy: well, no, not possible
09:50:11preglowTimothy: a yield() call does one cycle through the tasks, then end up in your task again
09:50:19 Join Zagor [0] (n=bjst@pdpc/supporter/sustaining/Zagor)
09:50:25preglowit always does get back to you again
09:52:16Timothyactually, that seems to be what's happening
09:52:56Bgerpreglow not *always* :)
09:53:02Timothyi put in a if(i%20)==0) rb->yield(); and now it works, albeit really slowly...
09:53:18Timothyi guess i should just take that yield out. lol
09:53:45Timothyso much for sharing the cpu
09:53:51preglowyes, you need to do a balance there, heh
09:54:14Timothyit will probably halt music playback on the real thing, hehe
09:54:39preglowthe mandelbrot set explorer we already have does halt music playback from time to time
09:55:13Timothyi wonder why it deosnt do anything for me...when I try to run mandelbrot the screen flashes and is back to the plugins directory
09:58:03BgerTimothy that's because mandelbrot uses gfxapi, which is not available in the sim
09:58:32Bgerdo u own a h1x0 ?
09:59:08Bgerx11 or win32 sim ?
09:59:40Bgeri guess u want to add color to mandelbrot or ?
09:59:42 Quit Vlad0man (Read error: 110 (Connection timed out))
09:59:42SlasheriBger: That idea not to cache certain directories could be good. But user would need to create a list of those directories and put the list file to the player
10:00:00BgerSlasheri i think one dir is enough
10:00:02Timothyyeah mine is using color
10:00:05Bgerso it could be setting
10:00:30 Join Vlad0man [0] (
10:00:58SlasheriAh, true. Well, then it would be user-friendly to configure also
10:01:43Bgeras in my case ... i would just put in my /D dir everything i wouldn't use on the player itself
10:01:58SlasheriBut that would eat a lot of space from the configuration structure (not possible with current implementation)
10:01:59Timothyhowever, i'm not using any inline assembly, so it will be much slower
10:02:08Slasheriof course, we could use separate file for that
10:02:20Bgerhm, why lot of space ?
10:02:37Bger20 bytes or so ?
10:03:01Slasherithat is quite much still
10:03:13Slasheri120 bits..
10:03:17Lynx_I assume the 10 fps only video on the h300 series is a hardware limitation?
10:03:30Bgerhow much is the "free" space in the config sector ?
10:03:43BgerLynx_ : mix between hardware and software
10:04:01Bgeri mean, the mpeg4 codec is heavy for the cpu of h300
10:04:11Lynx_Bger: how is it hardware? Is the display too slow?
10:04:26Lynx_Bger: ah, ok
10:04:36saa[b_r]iderhas anyone changed the RGB values in lcd.h? I changed the colors, and it worked on my H300, but in file tree the highlight bar color was wrong. do I need to change the color contrast values too?
10:05:50Lynx_So it would be possible to have video at more than 10 fps with an easier to decode format, theoretically?
10:06:23 Join LinusN [0] (
10:07:01Bgersaa[b_r]ider i suspect u'll have to look at apps/tree.[ch]
10:07:06BgerLynx_ yep
10:11:00saa[b_r]iderBger: couldn't it be this value
10:15:14 Join hshah [0] (
10:16:53Timothyhaha, holy crap my mandelbrot set viewer is so slow on the actual player
10:22:53Bgersaa[b_r]ider i haven't digged into this code enough to help you
10:23:48Timothyit takes like 20 seconds to render at 1/10th the native resolution...
10:24:06Timothybut at least it runs
10:24:27Timothyand now im off to bed. thanks for the help everyone
10:26:04Jungti1234good night
10:26:43saa[b_r]iderBger: it's ok, thanks
10:29:40 Join mirak [0] (
10:33:30preglowLinusN: it's pretty uncommon for a cache not to have address selectivity, no?
10:35:16LinusNhave what?
10:35:56Bagderyou mean to map memory regions and not cached by the data cache?
10:36:15*Bagder can't type properly
10:36:23preglowi mean like telling the cache to cache that and that memory area, not those
10:36:29preglowby specifying masks or whatever
10:37:06preglowif the pp cache has anything like it, they sure haven't found it yet
10:37:24preglowthis will probably end with me having to do full scale reverse engineering
10:37:40LinusNthat is usuallu not a function of the cache
10:37:48LinusNbut the mmu
10:39:27preglowoh well, i'll just enable it naively and ignore it for me
10:39:38preglowthen find out why the flaming hell the button driver is acting so cool on me
10:51:16XavierGrso are rockbox forums up and running or it just me?
10:51:25Bagderthey are up
10:51:40Bagderjust that the new IP has not propagated through all over
10:51:42preglowdepends on your dns
10:52:09XavierGrwhen I go with my dns it will not load. Without one it works but slow.
10:55:56 Join Polo_o [0] (
10:56:57XavierGrok I have a question about memory, but please be gentle I don't that stuff: Can I imagine the whole memory on a target like a box. Which is sliced up for different purposes? (E.g: Plugin memory, Rockbox memory, Codec memory etc)
10:57:12Bagderyes you can
10:57:45linuxstbpreglow: Are you using the same algorithm as IPL for the scrollwheel events?
10:57:55preglowlinuxstb: indeed i am
10:58:00preglowit's almost unmodified
10:58:38XavierGrSo when rockbox runs how many areas does it have? I know that there is the plugin buffer and the codec one, but where rockbox code runs? All these stuff run in the 32MB SDRAM?
10:58:58BagderXavierGr: every single buffer in the code uses a little piece of memory
10:59:17Bagderthen there's the separate "areas" like codec mem, plugin mem, audio buffer mem
10:59:25Bagderand font buffer, language buffer
10:59:40Bagderand possibly more I don't remember atm
10:59:59XavierGrare these all mapped and standart loactions on memory?
11:00:15Bagdernone of them are
11:00:41XavierGrso every time the places are random?
11:00:41Bagderhm, wrong, the plugin and codec areas are
11:01:00Bagdernot random, but may differ yes
11:01:03 Join nathanh [0] (
11:01:11BgerXavierGr : the same code with the same compiler will have the same allocations
11:01:44Bgerand the same binutils
11:02:47XavierGrso is there a dir buffer which holds info about current folder? (I imagine this will be bigger when dircache is enabled)
11:02:58linuxstbXavierGr: The file apps/ (in your build directory) is a map generated by the linker that tells you where it has put everything in memory.
11:03:04Bgernot, it won't
11:03:08BagderXavierGr: yes there's a dir buffer
11:03:19linuxstbThe dir buffer is taken from the audio buffer when Rockbox starts up.
11:03:29Bgerthe also the dircache buffer
11:03:47Bgerand also
11:03:48linuxstbI meant dircache buffer.
11:04:03Bgerbut there is a dir buffer too
11:04:24XavierGrwihtout dircache enabled dir buffer holds what? Current folder?
11:04:36linuxstbThat's also taken from the audio buffer at startup - because its size is configurable.
11:04:49preglowSlasheri: are you far away from having a testable tagcache prototype?
11:04:53Bgerthere is also playlist buffer
11:05:02Bagdermany buffers there are
11:05:30XavierGrInstead of opening the dir myself and sorting the jpeg list, I could read the data from there.
11:06:00XavierGrbut it will be very difficult to know were that data is.
11:06:03linuxstbpreglow: I wonder if it's the application dealing with scrollwheel events differently in IPL.
11:06:11preglowSlasheri: it's a bit annoying that just entering the dircache setting, even without changing it, just plain disables it until the next reboot
11:06:13Bgeri mean the audio buffer, the dir buffer, the playlist buffer and the dircache buffer are allocated at startup and are with pseudodynamic sizes
11:06:18BagderXavierGr: if you want the current dir and you have access to it from a plugin, then yes
11:06:56BagderXavierGr: plugins uses the API and thus they never need to know where things are
11:07:02preglowlinuxstb: hmm, i don't think so, the driver just seems to insert events into a button queue
11:07:07preglowlinuxstb: i think the app just deals with those
11:07:27linuxstbYes, but the app may ignore multiple events received within X ms
11:07:48preglowno, i don't think so
11:08:02XavierGrBagder: I guess I would need a memory finder for that :). I think this would be too difficult.
11:08:03preglowthe clickwheel driver works by inserting several keypresses at once if you move it fast enough
11:08:12preglowthat functionality woudln't do anything if the app worked like that
11:08:22BagderXavierGr: no, you need a function added to the plugin API
11:08:36Bagderthat would return a pointer to the buffer, or a function that can return contents from the cache
11:09:12Bagderbut again, that is only if you really are interested in the current dir
11:09:34XavierGryes that's good. only current dir.
11:09:46XavierGrthe jpeg viewer scrolls only in the current dir.
11:09:56Bagderok, then it could work
11:10:15Bgerhm, but this is viewer ...
11:10:31Bgerso it gets a filename through its parameter or i'm wrong ?
11:10:43Bagderyes it does
11:10:45XavierGryes you are right
11:11:04BgerXavierGr then get the dir from this param
11:11:05XavierGrin my patch the viewer gets the filename and then opens that dir.
11:11:06Bgerit's easy
11:11:16XavierGrit will scan the dir and make a list with all jpeg files in that list
11:11:34XavierGrBger that's what I've done already.
11:11:44Bgerhm, ok
11:11:55XavierGrBger: My patch works pretty good with many files.
11:12:24XavierGrThe point is that if you exceed (1000 files) opening that dir and sorting that list will take 3-4 seconds.
11:12:37XavierGrIf I had the data ready there is no need for this.
11:12:44Bgerpreglow any plans of making the jpeg viewer to work in color ?
11:12:49preglowBger: hell no
11:12:55preglowBger: that is, perhaps, but not yet
11:13:05preglowi'm pretty sure someone will beat me to it
11:13:06XavierGrThough I will excuse my plugin because even the filetree in rockbox will get a significant lag past the 1000 files.
11:13:13Bgeryeah, i suppose u have much work on ipod port
11:13:16Bagderthere was an embryo for a color bmp loader posted
11:13:27Bagdersomeone could grab it and run for it
11:13:35saa[b_r]iderBger: i I just finished compiling, I'll test it out now
11:13:44Bgersaa[b_r]ider oki ;)
11:14:01BgerBagder the problem with jpeg is more serious ...
11:14:03Bagderpersonally I don't do any color stuff since I can't use the sim with colors
11:14:13***Saving seen data "./dancer.seen"
11:14:16preglowBger: not very much, luma is already decoded, you only need to decode chroma as well
11:14:25preglowand the decoding proc is similar
11:14:27BagderBger: I disagree, since bmp in colors will introduce bg colors and wps in colors etc
11:14:31preglowyou just need to do some basic resampling as well
11:14:44Bgerok, but this is dsp stuff :)
11:15:21BagderI fear I will have to do the SDL stuff
11:15:31preglowBagder: and we'll love you for it
11:15:32Bgerno fear :)
11:15:43Bagderwell, I guess it could be cool to get to know it
11:15:47preglowBagder: i actually kind of feel like doing it myself, but i guess i should stick to ipod
11:15:52 Nick Lost-ash is now known as ashridah (n=ashridah@
11:16:10BagderI'm not really up to it just yet anyway
11:16:12preglowanywho, having a single sim target will be very nice
11:16:22preglowsdl is a pretty easy learn
11:17:24Bagderbtw, 7zip sure is a compression monster
11:17:49preglowyes, yes it is
11:18:08BagderI have some pending patches to introduce a make 7zip target
11:20:53Bagderbut compression is all it does nicely
11:21:00Bagderthe rest is a mess
11:21:45Bagderis 7za available in cygwin?
11:21:55nathanhmany devs use this channel?
11:22:05Bagdernathanh: yes, on and off
11:22:34saa[b_r]iderBger: USB doesn't want to connect well... I'll restart
11:23:41linuxstbBagder: What difference in size does 7zip give for
11:23:56Bagdermy test went from 1500KB to 1100KB, roughly
11:24:04linuxstbHave you tried tar.bz2 ?
11:24:21Bagderyes, the same test package bz2 was 1400KB
11:25:15ashridahbz2 is highly dependent on the data type
11:25:17preglowcompression speed?
11:25:26BagderI didn't compare speeds
11:25:32preglowbzip2 is slow as a dog
11:25:39nathanhbz2 is burrows wheeler, which uses heaps of memory
11:25:44linuxstbBagder: Are you against splitting the downloads into separate zip files for things like fonts, themes etc?
11:26:02nathanh7zip uses ppmd, which is better than burrows wheeler
11:26:12BagderI'm not against it, as long as we still can build one version that include the lot
11:27:25Slasheripreglow: not far away. A test prototype should be ready probably on this week
11:32:15XavierGrBagder: On the dir buffer thingy: All data for current directory I think are stored in the static struct tree_context tc; in tree.c
11:32:36XavierGrbut it ism't an extern variable so how could i get the address of it?
11:33:27 Quit mirak (Connection timed out)
11:34:30Bagdertree_get_context() perhaps?
11:35:51XavierGrnice indeed :)
11:36:43XavierGrthis can make the jpeg viewer a lot more faster.... (situations with more than 1000 files)
11:37:14 Quit saa[b_r]ider (Read error: 104 (Connection reset by peer))
11:38:11 Join mirak [0] (
11:38:35Bagder7zip takes roughly three times longer than zip
11:38:56Bagder1595509 vs 1143132
11:39:02Bagderh120 recent cvs size
11:39:19nathanhis this for snapshot distribution?
11:39:36BagderI'm considering it
11:39:56Bgertools under linux ?
11:40:01nathanhim not sure 7zip is available everywhere, like linux
11:40:10BagderI use nothing but linux ;-)
11:40:34*nathanh hang head
11:40:46 Join saa[b_r]ider [0] (n=saab_rid@
11:41:09*Bger emerges
11:44:03nathanhany info on the remote control buttons for the h300?
11:44:23Bgerthere is a patch in the patchtracker (made by me)
11:44:24Bagderbmp with space
11:44:28Bagderbmp name
11:44:37Bgeralso, there is info in the wiki
11:45:11saa[b_r]iderBger: it's a charm!!
11:45:19nathanhim amazed at how quickly this has swung into action... 6 months ago, nothing, 4 weeks ago, it booted, today, i prefer it to the official iriver firmware
11:45:32Bgerhehe :)
11:46:05saa[b_r]iderthis patch should be comitted
11:46:17 Quit perplexity (Read error: 110 (Connection timed out))
11:46:19Bgerbtw i think that rb on h300 has shorted battery life than rb on h100
11:46:29nathanhyeah, i noticed that on the wiki
11:46:30Bgersaa[b_r]ider no, it shouldn't ...
11:46:35nathanhsomething about the backlight turning off but not the lcd
11:46:42saa[b_r]iderBger: any bugs?
11:46:56Bgerno, but there are many people with h300 and h100 remotes
11:47:17XavierGrMy battery_bencmark plugin could help you make realtime tests.
11:47:28Bgerhehe XavierGr :)
11:47:29XavierGrBut still it waits for commitment
11:47:33Bgeri'll test it
11:47:49nathanhthe remote buttons patch wont be committed? why not?
11:47:51saa[b_r]iderhmm... you're right. I also noticed the battery life is not very high now
11:47:59Bagderhm, I wonder if we should trigger new builds on changes to the tools/
11:48:04XavierGrYou can make some very good graphs with it, if you use the log.
11:48:25BgerXavierGr i'll give it a go, but not now ...
11:48:55XavierGrjust mentioning, my patch will go nowhere? At most it may be commited :D
11:49:02Bger(regarding batt. life) i think that a part of the problem is the slow lcd_update ...
11:49:22BagderXavierGr: I'm +1 on a commit
11:49:44 Quit Kohlrabi ("Leaving")
11:50:09XavierGrYou mean you think it would be good to be commited?
11:50:20XavierGrOk thanks :D
11:50:41Bgerhm, maybe i should change the mappings of h100 remote also... and let someone with h100's remote to try them ...
11:50:44XavierGrStill we would need on archos tester to be sure that it wont brake anythin.
11:50:58XavierGranyone with an archos?
11:51:05BgerBagder :)
11:51:55saa[b_r]iderit seems that the highlight bar is defined as the inverse of the background color, and the highlighted text is the inverse of the foreground color (if I'm not wrong)
11:53:07saa[b_r]idercan anyone help me find out which strings make these definitions? I'd like to change the highlight to the foreground color, and the highlight text to background
11:53:45Bgercould someone give me a pic of h100 remote ?
11:54:05linuxstbsaa[b_r]ider: It's not that simple. You will have to look at the code in lcd-16bit.c
11:54:07Bgerlink to such pic i mean
11:54:17saa[b_r]iderBger: just to know what it looks like?
11:54:33linuxstbBger: Just search ebay - I'm sure you will find some.
11:54:46Bgerto see where the buttons are situated
11:55:08nathanhi see lcd_update on h300 just does a full framebuffer copy
11:55:12linuxstbThe buttons are on the top, the bottom, and the front.
11:55:12nathanhno wonder it flickers :-)
11:55:32saa[b_r]iderBger: try this
11:57:29saa[b_r]iderany suggestions for a Unix-friendly text editor for windows? I hate going back and forth between this machine and my linux machine to view the codes properly
11:57:35Bgercan u give me the image locations
11:57:53nathanhsaa: vim, emacs, take your pick
11:58:09Bgersaa[b_r]ider silly, but i use winrar's internal viewer for unix formated text (only for view of course)
11:58:19Cassandrapfe is OK, if it's still around.
11:58:51saa[b_r]iderthanks guys, I'll google them and see
11:59:07saa[b_r]iderbgerL did that link work?
11:59:25preglowsaa[b_r]ider: gvim is nice
11:59:52Bgersaa[b_r]ider yes it worked
11:59:56Bgerbut i need more :)
12:00:06linuxstbsaa[b_r]ider: I know Windows users that use JEdit happilly
12:00:26saa[b_r]iderif you check the MR gallery, I'm sure there are loads...
12:00:47 Join aliask [0] (
12:01:13saa[b_r]iderthanks nath cass linux :)
12:03:59saa[b_r]iderBger: I think you'll like this :)
12:04:11amiconnTimothy: mandelbrot.rock doesn't work in the simulator yet because it uses the grayscale library, which isn't supported in the sims
12:04:36Bgersaa[b_r]ider bingo :)
12:05:36amiconnTwo things need to be changed: (1) The sims need to support the grayscale lib for b&w and low-depth grayscale targets. (2) For colour units, mandelbrot.rock needs to use the standard lcd api
12:05:50amiconnAll this is already on my todo list :/
12:06:24Bgeranyone here who knows how the original fw treats "vol+", "vol-", "<<" and ">>" on h100 remote?
12:06:42nathanhi know how it works on the h300 remote, will that help you?
12:07:01amiconnsaa[b_r]ider: I can recommend ConTEXT. It keeps the line end format whatever it was, can convert between formats, supports syntax highlighting, ...
12:07:02Bgernathanh no, i know how it works on h300, i have one
12:07:15amiconnOnly drawback is that it doesn't support UTF-8 yet
12:07:35amiconnBger: I wouldn't care
12:07:46saa[b_r]ideramiconn: great suggestion. I can live without Unicode
12:07:53amiconnOh, and ConTEXT is free (but not open source)
12:07:55*Bger expected such answer from amiconn :)
12:08:09*preglow plays the advocate: gvim!gvim!gvim!
12:08:10nathanhoh, is amiconn = linas?
12:08:16Bgerbtw, i'm up with amiconn about conTEXT, tried it
12:08:22Bgernathanh no
12:08:25Bgeramiconn = Jens
12:08:30BgerLinusN = Linus
12:08:32nathanhwoo, jens
12:09:44Bgerotoh, saa[b_r]ider, if u want to try a "classic" editor, give gvim a go
12:10:11markunsaa[b_r]ider: You can try
12:10:56XavierGrI use notepad2 very good program
12:11:03markunalso supports UTF-8
12:11:35Jungti1234thanks markun
12:11:37XavierGrit has the tab to spaces option and column marking.
12:12:09XavierGrso no more whitespaces or more than 80 columns in patches.
12:12:10Bgerhighlighting ?
12:12:22XavierGrdiff, c c++ and many others.
12:12:33XavierGrnot so configurable like Context though.
12:15:24markunsaa[b_r]ider: no support for arabic in notepad2.. :)
12:16:24TimothyI like notepad++
12:17:24pregloware we going to want to use the piezo speaker on the ipod for anything? :>
12:17:45Timothyanti theft alarm?
12:18:20 Join saab_rider [0] (n=saab_rid@
12:18:28XavierGrpiezo speaker?
12:18:38XavierGra tweeter in an iPod why?
12:18:45preglowask steve jobs
12:18:50linuxstbpreglow: A key click?
12:18:52preglowi just wonder if anyone can think of any uses for it
12:19:00preglowlinuxstb: yeah, we might as well keep that, i guess
12:19:47XavierGrput it in the plugin API, many users may come up with something....
12:22:13aliaskCan anyone think of a reason that rockboy would run 20-30% slower on the H300 to the H100's?
12:22:13nathanhits for the click noises when you spin the "wheel"
12:22:29ashridahwhat he said
12:22:31preglowaliask: screen update?
12:22:55ashridahaliask: LCD requires more data per pixel
12:22:59 Quit Timothy ("CGI:IRC (Ping timeout)")
12:23:03ashridahprobably not overly efficient yet either
12:23:08aliaskHrm, maybe... could we limit the FPS to boost it?
12:23:18nathanhbetter off fixing the lcd_update routine
12:23:23nathanhim taking a look at it, and its ripe for improvement
12:23:33preglowfind someone to optimie rockboy for us, there's tons to be done
12:23:46preglowi guess someone could also make a frame skipper for it
12:23:48preglowthat'd rock
12:24:01aliaskCan be done in ASM or does it need dynrec?
12:24:14XavierGrthen it would be like playing Doom 3 with an 64MB gfx card!!!!!
12:24:24preglowaliask: asm
12:24:36preglowaliask: i don't think dynrec is very viable for out platforms
12:24:51preglowXavierGr: do you prefer it to be slow as hell instead?
12:24:57preglowplaying in slow motion isn't exactly a blast
12:25:36nathanhgameboy = $20 on ebay
12:25:56Bgernathanh u don't get the whole idea behind this ... :)
12:25:58aliaskYeah but you'd have to carry it around.
12:26:07*nathanh is just teasing
12:27:16*saab_rider just realizes that the forums are up again!
12:28:42 Quit saa[b_r]ider (Read error: 110 (Connection timed out))
12:28:53aliaskIs there any method for text input in rockbox?
12:29:05nathanhtheres a mini virtual keyboard
12:29:14Bgerwith morse input also :)
12:29:18aliaskBut to edit files?
12:31:55Bgernot yet
12:31:57linuxstbSomeone was working on "rockword" to edit text files - that was for the h300 only.
12:32:36aliaskI have been thinking about making it, but seeing as I started learning C about a week ago, I think it would take me too long.
12:32:50nathanhaliask: linus torvalds learnt C while writing linux
12:32:53aliaskDid the guy who was making it show any source?
12:33:12aliasknathanh: That's why I said slow, not that I wouldn't do it :)
12:33:31linuxstbYes, I think so. There is a thread on the h300 mistic river forum about colour plugins for the h300 - he links to his plugins from there.
12:35:07aliaskExcellent, I'll take a look.
12:36:39aliaskWow, it's far from basic. Even has text formatting.
12:37:11XavierGrpreglow: I was just kidding on the rockboy subject ;p
12:37:26 Join mind_less [0] (n=Tim@
12:37:55nathanhive read the wiki about rockbox-ipod... is anybody here using it or developing it?
12:39:02*ashridah looks at preglow
12:39:20preglowyes, that'd be me and linuxstb
12:39:35nathanhthe wiki doesnt make it clear... is it in a useable state?
12:39:41ashridahaliask: iirc, what he posted were mockups tho
12:40:12aliaskWell, I'll see soon, I'm installing the plugins.
12:40:36ashridahi haven't been watching too closely tho
12:41:13preglownathanh: i implemented the button driver just yesterday, so you can navigate (after a fashion) now
12:41:18preglownathanh: no audio, though
12:42:11 Join drumrboy [0] (
12:42:32linuxstbnathanh: Do you own an iPod?
12:42:41nathanhlinuxstb: no
12:42:53aliaskI would if they got rockbox though.
12:42:58nathanhbut i find it interesting how quickly rockbox is speading :-)
12:43:01linuxstbFor some reason, ipod owners seem rare in this channel...
12:43:09linuxstbBut we are now three.
12:43:21nathanhi had an ipod, i wasnt overly impressed with it
12:43:23XavierGrCassandra got one too?
12:43:32nathanhi thought the iriver was better... and rockbox is better again
12:43:53linuxstbXavierGr: Yes - a Nano
12:45:26Bgerjust to ask who is the 3rd
12:45:54nathanhgot a question about the bootloader on the h300
12:46:22nathanhwhen you hold play, it boots the bootloader from flash, which is the rockbox patched bootloader, and based on the record button state you can boot the original firmware or the rockbox firmware
12:47:01nathanhbut if you plug in the car adapter the iriver jumps straight to the charging screen, then if you hit play it boots straight into the iriver firmware
12:47:24nathanhis that second path also in the flash firmware?
12:48:12aliaskWhen you either plug AC or USB in while it's off, the iRiver firmware catches it and boots. This will change in v4
12:48:31nathanhso the iriver firmware has multiple entry points
12:48:38nathanhand the v3 bootloader only catches one of them so far
12:48:44aliaskThe charging screen is actually the iRiver firmware.
12:49:07aliaskI thought that v3 didn't catch the USB... let me check.
12:49:35aliaskYeah, plug in USB from off, and it boots the iRiver firmware
12:51:06nathanhoh i think i get it, so in mkboot the bootloader/main.c is appended to the original iriver firmware
12:51:34nathanhthen the vector interrupts at the start of the firmware are rewritten to point at bootloader/main.c
12:52:11nathanhbut there are multiple vector interrupts, at least play, usb and power
12:52:31aliaskThat's my understanding anyway.
12:52:32saab_riderlinuxstb: so RB on an iPod will allow filetree structure, and playing MP3s directly off the HD (as oposed to transfering them through iTunes)?
12:52:56linuxstbsaab_rider: Yes, Rockbox on iPod will make the iPod behave like every other Rockbox target.
12:53:07nathanhideally if the power interrupt vector could be caught as well
12:53:16nathanhthen rockbox would be the ideal car player
12:53:19aliaskOh well, I'm off, gotta do christmas shopping tomorrow. Joy. :|
12:53:24nathanhturn the car on, starts playing, turn the car off, turns off automatically
12:53:42aliaskComing in bootloader v4
12:53:45aliaskG'night all.
12:53:49*nathanh does a happy jig
12:53:58 Quit aliask ("Chatzilla 0.9.69 [Firefox 1.5/2005111116]")
12:53:59miraklinuxstb: lol fuck ipod.
12:54:12mirakrockbox is betraying us
12:54:22Bagderrockbox is not an iriver project you know
12:54:38ashridahif you're going to be all eliteist, rockbox could just take its ball and go back to archos only :)
12:54:48Zagornathanh: rockbox already does that
12:55:00nathanhzagor: just the iriver bootloader v3 doesnt, i get it
12:55:28mirakBagder: ipod support can only be good for rockbox, considering the large userbase and potential developers isn't it ?
12:55:42nathanhsome ipod users will clamour for it
12:55:48nathanhogg support, gapless playback
12:55:49Bagderthe more the merrier
12:55:59nathanhive got an ipod friend who curses the gaps in his dj albums
12:56:05mind_lesspeople who prefer iPod in general probably got it for it's simplicity, not to hack it.
12:56:11miraknathanh: there was a rumour saying ipod hardware could not support ogg
12:56:19miraknathanh: have you heard about that ?
12:56:30Bgermirak it's actual only for old ipods
12:56:32nathanhi heard it, i reckon newer hardware must be able to, if it has enough cpu grunt to play video
12:56:32ashridahmirak: they claimed iriver's hardware couldn't do gapless when asked too :)
12:56:51Bagdernathanh: it doesn't use the cpu for the video, it has a separate dsp
12:57:04nathanhfair enough
12:57:14Bgerthere is a problem with the cpu cache in older portalplayer chipsets iirc...
12:57:38preglowyes, it's slow
12:57:39amiconnnathanh: Rockbox already has 'car adapter mode' on platforms which power on when plugged. Currently this is archos only, but it sounds like the H3x0 is the next target which will support it
12:57:49nathanhamiconn: woo
12:58:04nathanhamiconn: the power off bit works fine, i can stop the car and it pauses, start the car and it resumes
12:58:09mirakamiconn: what is this car adapter mode ? I have read that on the wiki, I don't get the difference
12:58:15preglownathanh: the video is decoded on a more powerful chip
12:58:16ashridahnathanh: just need to wire the remote from a H1xx into your steering column :)
12:58:21amiconnThe H1x0 can't, because it doesn't power on automatically
12:58:23preglowi reckon we'll be able to do ogg nicely on ipod hardware
12:58:28preglowat least the pp5020 based ones
12:58:32nathanhamiconn: its the power-on from off state that doesnt work, but as already said, the v4 will fix that
12:58:44saab_riderlinuxstb: which gen iPod do you have?
12:58:48mirakpreglow: that's a dsp or a dedicated ship ?
12:59:04 Quit DrumRBoy|Away (Read error: 110 (Connection timed out))
13:00:06 Nick ashridah is now known as Lost-ash (n=ashridah@
13:00:22amiconnmirak: If you have a car with a power outlet that is switched together with the ignition, just connect the jukebox to it via a car charging adapter. Then starting the engine will unpause playback, and stopping it will pause (and poweroff after timeout)
13:00:24Bagdermirak: both
13:00:54preglowmirak: dsp-ish, it's hard to say
13:01:05mirakamiconn: my car suck
13:01:06 Quit nathanh ("CGI:IRC (EOF)")
13:01:12mirakone speaker is dead
13:01:20mirakI must put everything on left channel
13:01:24amiconnAll this does work on all platforms, just the power-on does only work for platforms which power on when plugged
13:01:53mirakis there a menu for radio on H300 os it's not implemetned yet ?
13:02:43amiconnNot implemented yet
13:03:43saab_riderI wish i understand code... can't figure out where the colors of the highlight bar, and higlighted text are
13:04:15amiconnThey aren't defined separately, they are just the inverse of the standard colour
13:06:15saab_rideramiconn: exactly, I want to disable the inverse, and switch the colors, background=highlight text color, and foreground=highlight
13:06:32saab_rideris this possible?
13:07:36 Join nathanh [0] (
13:09:29XavierGrstruct tree_context->name_buffer holds all the names of current folder, but which variable in there holds the positions of each string on the buffer?
13:10:17 Quit linuxstb (Read error: 110 (Connection timed out))
13:10:25 Join Febs [0] (
13:10:29BagderXavierGr: read the tree.c code and see what that does
13:10:39 Join linuxstb [0] (
13:10:46XavierGrokay thanks
13:11:49 Quit edx (Read error: 110 (Connection timed out))
13:13:21linuxstbsaab_rider: I have a 4G color (60GB).
13:14:11saab_riderhmm.. so the current port won't work on 3G ipods I assume
13:14:17***Saving seen data "./dancer.seen"
13:14:50linuxstbNo. It also won't work on greyscale 4G ipods until someone write an LCD driver.
13:15:07linuxstb(that includes the mini)
13:15:13preglowmaking an pp5002 port shouldn't be that hard
13:15:16preglowbut we need someone to do it
13:15:28preglowwe could use an ipod developer or two extra, actually
13:15:35saab_riderI wonder if my friend is capable... but other than the LCD driver, the buttons are different as well
13:15:42preglowoh yes
13:15:49saab_rider(at least for 3G)
13:15:58preglowthe hardware on 3g and below is pretty different than 4g and up
13:16:45saab_riderI already sent him a message (but he's not around now). he loves a challange, if he's interested ;)
13:16:47linuxstbBut the infrastructure is there - for the hardware drivers, it will mainly be a case of copy and paste from the IPL code.
13:17:40linuxstbThe LCD driver isn't trivial - the ipods use a different framebuffer format to the H1x0, so the high-level LCD code will need to be written as well.
13:17:55linuxstb(I'm talking about the greyscale LCD)
13:18:46saab_riderI sent him RB's link so he can see what's going on... there's also the iPod linux port, but I'm not sure what that's capable of
13:20:25saab_riderlinuxstb: any idea about the highlight color question?
13:21:41linuxstbI think amiconn answered it - that there is no concept of a "highlight color" in Rockbox at the moment - only an "inverse"
13:22:47SlasheriHmm, do i have a bug in my code or am i correct that rockbox can only append to a file opened with O_RDWR and not write anywhere there i go with lseek?
13:23:38saab_riderlinuxstb: yeah I know that it's only an inverse. hmm.. but doesn't it mention which previously defined colors to inverse?
13:26:11nathanhthe extent of this source code is bloody amazing
13:26:28nathanhtheres a mini libc, a kernel, threads, ata drivers
13:26:48Bagderyeps, we have it all ;-)
13:27:02Bagdernow, where's our tcp/ip stack? ;-P
13:27:19linuxstbsaab_rider: When drawing text, it will inverse the FG and BG colours.
13:27:37Slasherihehe, tcp/ip will come in future along with players equipped with wlan etc ;)
13:27:50nathanhrumor mill says that next gen ipod will have wifi
13:27:50Bagderthat'll be fun
13:27:58nathanhwhether that means bluetooth or wimax or 802, who knows
13:28:16Bagderseem inappropriate
13:28:20Bagderbut who knows
13:28:47Bagderand btw, wimax is 802 too ;-)
13:28:48XavierGrSlasheri: no you can write
13:28:54nathanhhrm, bluetooth hardware hack onto h300 w/ bluetooth headphones
13:28:57*nathanh dreams
13:28:58Slasheribluetooth would be too easy because i have already coded a stack for it :)
13:29:18SlasheriXavierGr: ah, interesting. Then i need to add more debugging
13:29:51XavierGrSlasheri: Or so I know from my test....
13:30:20XavierGrhow do you write to the file?
13:30:32saab_riderI get it now, it just inverses. I thought maybe it was actually specific by telling it to assign the inverse of the BG to the highlight, and an inverse of the FG to the text. but it's simpler than that, it just asks the current entry's color to be an inverse of the default colors
13:30:50 Quit mirak ("Ex-Chat")
13:30:59SlasheriXavierGr: i open a file created earlier, then seek to a correct location and write a chunk of data
13:31:12Slasheriah, in fact i read a chunk from that file first
13:31:27 Join mirak [0] (
13:31:28Slasherithen i edit that chunk a little, lseek back and write the new one
13:32:00XavierGrI think that you can't overwrite
13:32:02Bgersaab_rider btw i forgot to add a comment regarding the fact that the h300 remote lcd button patch adds the CONFIG_VIRTUAL led define too
13:32:05XavierGronly add
13:32:33Slasherioh, so that's the problem then.. :/
13:33:02Slasheriwhy overwriting is not possible? I think it shouldn't be hard to implement..
13:33:08XavierGrI had the same problem once
13:33:16nathanhim guessing nobody needed to overwrite before :-)
13:33:17BagderSlasheri: I would guess it just hasn't been done
13:33:18Slasherii think we need to fix that :)
13:33:18saab_riderthe inverse of black and blue is white and yellow (yellow text with white highlight doesn't look great), so I changed the line select to the pointer :)
13:33:33saab_riderBger: what does that do?
13:33:34SlasheriBagder: ah :) probably i try to figure that out
13:33:39XavierGryou can make a trick with read_line
13:34:26nathanhare the gpio regs memory mapped, or are they shadows?
13:34:43Slasheribut overwriting in the new tagcache code is quite essential for performance reasons.. so i try to fix the write routine then
13:34:47XavierGrBut yeah Slasheri if you knwo how to implement such function please do. I had the same problem (though I solved with another approach)
13:34:59*nathanh asking about h300
13:35:16Bagdernathanh: memory mapped, iirc
13:35:36SlasheriXavierGr: i will try :) but doing it safe takes some time to learn the essential fat structure first
13:36:05BagderSlasheri: there are some test cases for the fat code that might come handy
13:36:21SlasheriBagder: ah, that sounds good :)
13:36:26Bagderthe linux mount-a-file is a dream for that
13:37:00Bagderso you can make a fs-dump from a live fs, mount that and run your test code
13:37:37Slasherihmm, that sounds like a great idea to test the code
13:37:55Bgerbtw can someone with cvs access add the following to firmware/export/config-h300.h
13:38:07Bger#define CONFIG_LED LED_VIRTUAL
13:38:36BagderI wonder why MR spell Rockbox so oddly
13:38:47Bagder"ROCKbox for H1xx Series"
13:39:21saab_riderin the forum titles?
13:39:25crwlRrrOcKBOX sup1337 h4x0r f1rmw4r3x
13:39:41amiconnSlasheri: rockbox should be able to write anywhere in a file opened with O_RDWR
13:39:51amiconn(unless you open with O_APPEND of course)
13:40:15saab_riderwell the logo says ROCKbox
13:40:51saab_riderI personally prefer H100 rather than H1xx or H1x0
13:40:51Bagderbut that's the graphical logo
13:41:08Slasheriamiconn: Hmm, interesting. I will write a simpler code to test that.. probably it's just a bug in my code then
13:41:22saab_riderI guess they took it as the "official" way of writing it
13:41:25XavierGramiconn: He wants to overwrite previous data, can that be done?
13:41:25Bagdersaab_rider: the problem with h100 is that there is a model h100 that differs slightly from h120 and h140
13:41:48Bagdersaab_rider: so h1x0 is more "covering" while H100 could feel like it isn't h120 and h140
13:42:03Slasheriamiconn: So i would need to read a data first, and write it back (in same location) with modified
13:42:06saab_riderthat makes sense...
13:42:36Bagderat least I try to make that distinction
13:43:13Jungti1234 ~
13:43:27Bagderbless you
13:43:46saab_riderI didn't know that there existed an H100 that was different, what's the difference?
13:44:03Bagdersaab_rider: the ihp100 has 16MB ram as the biggest noticable diff
13:44:20Jungti1234Snort snort
13:44:21Bagderbut there are some more minor differences
13:45:13saab_riderbut they look the same, right?
13:46:03 Join Membrillo [0] (
13:46:18saab_riderinteresting... I thought they just renamed them because of their use of the HP prefix
13:46:32Bagderthe renaming was done on the 120 and 140 as well
13:46:45Bagderthere are ihp-120 and h120 models
13:46:56saab_ridervery confusing :)
13:47:51XavierGrand as one nice guy said on MR forums once: "I bought an iHP and I was left with an H. :("
13:47:53Bagderand the ihp-100 is also known as iHP-110 sometimes
13:48:08Bagderand even an iHP-115 is claimed to exist
13:48:49Jungti1234I have H120.
13:49:36saab_riderI wonder what they iRiver have up their sleve as their new HD-based player
13:50:03XavierGrI have an iR-160-e :))
13:50:18XavierGrR for rockbox
13:50:48Jungti1234A new player? :)
13:50:58BagderI don't believe in iriver's future players
13:50:58amiconnSlasheri, XavierGr: It should work. Of course you need to seek back after reading, and not open the file with O_TRUNC specified
13:51:35saab_riderwell I don't want it to be like the H10
13:51:37Jungti1234iriver will make H20.
13:51:44XavierGrStrange, it din't work out on my test. Maybe I was doing something wrong.
13:51:53XavierGrSlasheri please tell me if you get this working.
13:52:17nathanhwhats wrong with the h10?
13:52:22saab_riderjungti1234: are any of the specs released/rumored?
13:52:56saab_riderdoes it function like the H1x0/H300 with filetree, and drag-n-drop playback?
13:53:07Jungti1234Don't detail yet.
13:53:49Jungti1234It can change.
13:54:02 Quit Febs (Read error: 110 (Connection timed out))
13:54:13SlasheriXavierGr: yes, i will. I just tried a very simple test case and it seems to work there but it doesn't with my real program :/ i will still try to find the bug..
13:54:34Bagdernathanh: it seems it isn't mass storage anymore for example
13:55:56 Join Rob2222 [0] (
13:56:08Jungti1234nathanh: People say that H10 has many bugs.
13:56:19Bagderit seems there are different versions of H10
13:56:24Bagdersome are mass storage
13:56:27linuxstbBagder: Anything interesting happening with Neuros? Or is that idea dead?
13:56:41XavierGrSlasheri: Did you just write a message with white letters on awhite background?
13:56:43Bagderlinuxstb: I consider it dead for a good while ahead
13:57:25Bagderlinuxstb: Rockbox would only be considered for their N3 model which isn't even discussed and the only plans mentioned said something like Q3 next year
13:57:39Bagderand there is no source
13:57:49Bagderand they use a TI chip again
13:57:55linuxstbThey would seem better off collaborating with the ipodlinux people on an open source userland application.
13:58:26linuxstbThey could use the Rockbox simulator...
13:58:27Bagderthere first model on this arch is their 442 model which is gonna be linux based
13:58:50saab_riderBagder: I thought H10 was non-UMS :D
13:59:00Bagder"Related question. How do I know whether my player is UMS or MTP?"
13:59:00SlasheriXavierGr: i created a file and wrote a structure and closed the file. Then i opened the file again with O_RDWR, read that structure, changed one value and wrote it back. That seems to work fine..
13:59:15Bagdersaab_rider: "If your player comes with iriver plus1 or 2, then you have to use iriver plus software for organizing your music and this can be regarded as UMS version."
13:59:21Bagderindicates UMS ;-)
13:59:32nathanhwtf is ums?
13:59:39Bagderusb mass storage
14:00:18linuxstbi.e. you plug your player into any computer running any operating system, and it appears as another hard disk.
14:00:46nathanhthe primary reason why i got the iriver
14:01:05nathanhthe idea of having to use a windows/mac software package to move audio files onto the player was boggling
14:01:10saab_riderbagderL hang on... so H1x0 and H300 is just "mass storage" not UMS? or what would they be called?
14:01:24BagderUMS is mass storage
14:01:46Bagderthe H10 is also available as "MTP only"
14:01:50Bagderwithout UMS support
14:02:11linuxstbSo what's MTP?
14:02:36Bagderits a cloned named from PTP (picture transfer protocol), meaning music transfer protocol
14:02:48XavierGrSlasheri: OK thanks. I was doing something completely wrong. And when I asked here I thought that it wasn't supported.
14:03:00linuxstbIs that some kind of standard?
14:03:08Bagderfrom MS, yes
14:03:17Bagder"Why Partners Need to Better Understand MTP"
14:03:23nathanhhar har har
14:03:31linuxstbAh, so WMP can communicate with MTP players.
14:03:34nathanh"You Will Like MTP, damnit"
14:03:38Bagderlinuxstb: yes
14:03:47BagderI'm guessing there is a lot of DRM all over
14:03:49*linuxstb shudders
14:03:51Jungti1234H10 20GB?
14:04:00nathanhits the way of the future, rockbox is the only way out :-)
14:04:19saab_riderBagder: so H1x0s, H300s, and H10 are all UMS, but the first 2 support playback of drag-n-dropped files, and UMS H10 requires iriverplus for music transfer?
14:04:48Bagdersaab_rider: no, if H10 supports UMS it supports the same drag-and-drop
14:05:04Bagderbut it doesn't have to mean that it can play music transfered that way
14:05:37markunBagder: just like the Gigabeat, who refuses to play music if it's not transfered with their special software
14:05:42Bagderand iriver's site is not likely to tell
14:05:43linuxstbSounds like a move back in time 5 years when all DAPs had their own proprietory protocol. Good to see the industry innovating and moving forward.
14:06:07Bagdermarkun: :-(
14:06:31Bagder"There is simply no way to understate the importance of MTP to Microsoft's digital entertainment strategy"
14:06:36nathanhthe media companies will be pushing for proprietary protocols
14:06:38linuxstbmarkun: Have you tried to run code on your gigabeat yet?
14:06:52nathanhexpect it to happen, no matter the suckage
14:07:09markunno, couldn't get the an arm-linux crosscompiler to build so far.
14:07:27markunlibc refused to be build on my FreeBSD machine
14:07:39Bagdermarkun: tried it with crosstools?
14:07:45markunno, I didn't
14:07:53saab_riderso the main difference between the UMS H10 and the MTP H10 is that the UMS version uses iriverplus, and MTP uses WMP and doesn't act as an external drive?
14:08:07Bagdersaab_rider: as far as I can tell, yes
14:08:24saab_riderthat's why I don't like the H10 :
14:08:27Bagderthey mention DRM differences too
14:08:58saab_riderI read about that MS compatability...
14:09:12saab_riderthey had a funny name, play for sure?
14:09:13SlasheriXavierGr: i would say that there is nothing wrong with the fs code.. Now my hexdumps are beginning to look too weird.. :) Probably i need some extensive testing now
14:10:25linuxstbAnyone have any idea what kind of market share DRM'd WMP has compared with the iTunes Store?
14:10:48saab_riderprobably 0.0000000001
14:11:47preglowbloody button driver
14:13:37linuxstbI love Sunncomm's FAQ about the DRM they provide on Sony CDs - they provide a set of DRM'ed WMA files and then blame Apple for the fact that users can't transfer the CD to an iPod because "Apple uses a proprietory system".
14:15:54SlasheriXavierGr: i am now quite sure overwriting doesn't work as it should.. it just corrupts the file
14:16:37_FireFly_Slasheri how did you open the file for overwriting ??
14:17:02Slasheri_FireFly_: O_RDWR
14:17:25_FireFly_and how does the file get corrupted ??
14:17:28Slasheriand that problem doesn't seem to happen with a very small files..
14:17:36Slasherii will implement a test function soon
14:18:05Slasheri_FireFly_: it doesn't update from the beginning, it will lost the alignment and data are being inserted to the end or something like that.. :)
14:18:13Slasherii don't know well just yet
14:18:34_FireFly_Slasheri try to explicitly seek to the start of the file
14:18:46_FireFly_before writing
14:19:01linuxstbpreglow: No luck with your buttons then?
14:19:04Slasheri_FireFly_: i have done that
14:19:08preglowhaven't had much time, but no
14:19:11Slasheribut it doesn't have effect
14:19:21Slasherithe old data just remains..
14:19:25_FireFly_Slasheri or some other fns which do you use modify the seek-pos
14:19:25preglowgot cpu clocked at 75mhz, still no dice, it lags like mad
14:20:27linuxstbpreglow: Do you send release events as well for each scroll event?
14:20:38preglowlinuxstb: scroll, yes, the other buttons, no
14:20:50preglowthat's a bug i just thinked of
14:21:41linuxstbMaybe the list viewer code is expecting something different - have you looked at the application side of things?
14:21:59 Join Kohlrabi [0] (
14:22:28preglowqueue_post(&button_queue, wheel_keycode, NULL);
14:22:29preglowqueue_post(&button_queue, BUTTON_REL | wheel_keycode, NULL);
14:22:43preglowthat's what i do for each scroll event
14:22:53preglowno, i haven't
14:23:31linuxstbSo presumably wheel_keycode is either BUTTON_SCROLL_FWD or BUTTON_SCROLL_BACK ?
14:23:57preglowaye <-
14:25:04linuxstbSeems that the button definitions in tree.h are not being used any more.
14:25:40linuxstbGuess it must be in gui/list.[ch] now
14:25:51LinusNpreglow: why the release event?
14:26:16preglowLinusN: because i thought they were needed :]
14:26:37LinusNafaik they aren't
14:26:43preglowokies, then i'll do a quick mod
14:27:26LinusNcan you detect that you are holding still on the wheel?
14:27:48linuxstbMy thoughts about the scrollwheel were that it would fire successive button press/release events, and would never send a repeat. This way, the button driver would control how fast the events are sent to the application.
14:28:13linuxstbBut maybe that was crazy.
14:28:25LinusNbetter send repeated press events
14:28:43LinusNwithout the REPEAT flag
14:28:44 Quit nathanh ("CGI:IRC (EOF)")
14:29:00LinusNcan you detect that you are holding still on the wheel?
14:29:10preglowi don't know
14:29:18linuxstbI think you can.
14:29:22Jungti1234Good night
14:29:22LinusNcan you detect that you push it?
14:29:39 Quit Jungti1234 ("bye")
14:29:40preglowagain, i don't know, buttons send a press interrupt, and depress
14:29:43preglowdon't know how the wheel does it
14:29:58LinusNbut aren't you writing the driver? :-)
14:30:16preglowthe driver was a quick cowboy coding session i did yesterday
14:30:25preglowit's more or less a direct rip of the ipl code with some #if 0 in it
14:30:40preglowman, cache really makes a difference
14:30:57LinusNhow does the apple firmware operate? does it detect presses as well as sliding?
14:31:27preglowpressing the wheel and sliding are two different things
14:31:32linuxstbLinusN: There are four "real" buttons located underneath the scroll wheel.
14:31:35preglowthe presses are are actual buttons
14:32:06linuxstbIIUC, the driver detects where on the wheel you are currently touching.
14:32:17preglowwhy's that lame?
14:32:31linuxstbThere are two actions - touching and pressing
14:32:49preglowscroll wheel is usable now
14:33:39*LinusN does the scrollwheel dance
14:33:47*linuxstb wants a webcam URL
14:33:48preglowclick wheel, i mean!
14:33:55preglowyou can have a build
14:34:14preglowit's oversensitive still, though
14:34:15preglowi need to fix that
14:35:08linuxstbIs it possible to move one position now? With your previous version, the best I could do was two.
14:36:10linuxstbSo have you enabled the cache now as well?
14:36:24preglowyes, it flies
14:36:52linuxstbIs the ATA driver still causing you problems?
14:36:56preglowat least it can handle really heavy interrupt loads now
14:37:11preglowa couple of quick circuits around the wheel with the finger and it's going for a good two seconds
14:37:53preglowyes it is
14:38:19preglowmy nano is making sound
14:38:23preglowhigh pitched whine
14:38:25preglowpretty weak
14:38:33 Quit XavierGr (Read error: 110 (Connection timed out))
14:39:09LinusNpreglow: not exactly a "gentlemen" mail, is it? :-)
14:39:42 Quit markun (Read error: 113 (No route to host))
14:39:43preglowGentlemen, we have unintended high-pitched whine!
14:40:04LinusNpreglow: do you need to handle the wheel with an interrupt?
14:40:23preglowall buttons are handled with interrupts
14:40:35LinusNcan't you poll the wheel interrupt state within the button tick interrupt?
14:40:52preglowi guess i can
14:41:11LinusNi don't think you need to handle each and every wheel event
14:41:17LinusNthe same goes for the buttons
14:41:39preglowi would want the wheel event speed to follow the speed of my finger
14:41:49preglowthat'd work still, of course
14:41:54preglowdepends if i can get polling to work properly
14:42:12preglowi have no idea what part of this jumble of status registers writes does what
14:42:14LinusNthat way you would get a much more even load
14:42:44preglowcomments are in the style of /* Interrupt ACK? */
14:43:10LinusNif you don't find out, you can do a pseudo-poll:
14:43:38LinusNin the wheel interrupt only set a flag indicating what the event was, then poll that flag in the button tick driver
14:44:10 Join Hooligan [0] (
14:44:22preglowbut argh
14:44:30preglowthe flash accesses are really sluggish
14:44:45LinusNstill, heavu usage of dbg_ports() should reveal some stuff
14:45:01preglowthat's true
14:46:29preglowi press "browse fonts", and it just hangs for a couple of seconds, then pretends nothing happenede
14:46:36preglowwith the file listing being absent
14:47:07preglowman, how deep is the event queue?
14:47:28preglowi had it going for like twenty seconds, here, with five or six quick circuits on the click wheel
14:47:30 Join markun [0] (
14:47:40preglowi mask the i2c interrupt while dealing with it
14:47:45LinusNpreglow: 16 events, i believe
14:47:57preglowok, so the clickwheel controller itself really needs to have quite a queue going
14:48:29LinusNif you really need to capture that many events, yes
14:48:43preglowno, i mean the hardware controller
14:48:54preglowit can't be a software issue if you've only got sixteen levels of queue
14:48:59preglowthat'd be spent in a blink
14:49:13preglowbut i obviously continue to receive events for ages
14:49:28LinusNany data sheets for that controller?
14:49:35preglowno idea
14:49:39LinusNor is that in the PP chipset too?
14:49:43amiconnEvent queues are 64 events afaik
14:49:43preglowdoubt it
14:49:50preglowit's a custom apple controller, i bet
14:49:55preglowthey get these click wheels made especially for them
14:50:10preglowbut it might also be a pp thing
14:50:13LinusNamiconn: #define QUEUE_LENGTH 16
14:51:10LinusNnasty FAT32 buffer overflow alert
14:51:44LinusNthe filenames are nowadays stored in utf8, but the direntry buffer is only 256 bytes
14:51:54LinusN(as pointed out in the rockbox forums)
14:52:06linuxstbpreglow: From the ipodlinux wiki - "Cypress CY8C21434 - Click wheel controller"
14:52:19linuxstb(this is under the entry for the Nano)
14:53:36 Join edx [0] (
14:54:38linuxstbAnd a datasheet:
14:55:10LinusNlinuxstb: put it in the rockbox datasheet wiki
14:56:42preglowit's programmable
14:56:45preglowso really not much help, i guess
14:57:11linuxstbI'll upload it anyway.
14:57:52Slasheriamiconn: hehe, it seems to be a caching bug in the file system code because if i call fsync after every write, the problem is gone
14:58:02Slasherii will look fixing that later, now i need to go
14:59:55 Join Moos [0] (
15:00:02MoosHello here!
15:00:34 Quit Maxime (Read error: 104 (Connection reset by peer))
15:02:02amiconnLinusN: Ah ok, so the default queues are 16 entries deep. I confused it with the talk queue
15:03:08Bgerhm, what about adding a "HAVE_ADJUSTABLE_BRIGHTNESS" define ?
15:03:17preglowlinuxstb: it's still hypersensitive
15:03:29Bgerregardinh all colour targets ?
15:03:30preglowjust touching the wheel makes the cursor move
15:05:06 Join Maxime` [0] (
15:05:17Bgeranyone ?
15:05:33BagderBger: seems reasonable
15:06:07Bagderor only HAVE_LCD_BRIGHTNESS
15:07:36LinusNbetter separate main and remote lcd brightness
15:07:54Bagderbut most defines already use LCD for main and REMOTE_LCD for remote
15:08:22LinusNah, of course
15:08:33Bgeri'm editing the h300 brightness patch atm, which is awfully made
15:10:17preglowlinuxstb: any idea why the disk read would just fail like that?
15:11:17linuxstbNo. The only thing I can think of is the interrupts, but then I would expect it to fail earlier.
15:11:32preglowi have disabled all interrupts but timer1 and i2c
15:11:38preglowbut you didn't use them anyway, no?
15:11:44preglowi mean, you couldn't :)
15:11:50preglowbut anywho
15:11:52preglowit does read
15:11:56linuxstbBut who knows what the IPL init code does...
15:11:57preglowat least the root dir
15:12:24linuxstbYep. And it seems to save the config sector when shutting down.
15:12:43saab_riderBger: what does the CONFIG_VIRTUAL led define you mention earlier do?
15:12:54preglowlinuxstb: hmm, no, actually, it doesn't do that
15:13:06preglowor maybe it does
15:13:11preglowi should try
15:13:24linuxstbI changed the font, and now that font is the default when I start Rockbox.
15:13:33preglowwell, yeah, but you can change font as well
15:13:35preglowi need to try it
15:13:56linuxstbYou may also want to decrease the idle timeout
15:14:18***Saving seen data "./dancer.seen"
15:14:48linuxstbOr not - forget that. It should be written during the next disk access.
15:15:21preglowi did
15:15:41linuxstbI did manage to set the date and time from within Rockbox - that's one thing IPL can't do for some reason.
15:16:12preglowyes, you did
15:16:15preglowthat works fine here as well
15:17:02linuxstbPlugins seem to work - the databox plugin loads and runs.
15:17:06 Quit Bger (Read error: 104 (Connection reset by peer))
15:17:12preglowyes, it does save the config sector
15:17:21preglowi can't access plugins either :/
15:17:46 Join Bger [0] (n=Bager@
15:18:14linuxstbAnyway, I'll do some more tests and look at it this evening. Hopefully you'll have committed your button driver by then.
15:18:50Bgersaab_rider this enables the "disk activity" "led" on the statusbar
15:19:09 Join mikearthur [0] (
15:19:40preglowlinuxstb: don't count on it, but we'll see
15:19:55preglowi don't want to commit code looking like this
15:20:14 Quit Bger (Client Quit)
15:23:41 Join perplexity [0] (n=joust@
15:23:57 Join Kaggen [0] (
15:24:04linuxstbpreglow: OK, no problem.
15:26:37 Join Bger [0] (n=Bager@
15:28:51 Quit Membrillo ()
15:34:03 Join petur [0] (
15:36:13peturBger, can you tell me what's wrong with that H300 brightness patch? Excuse my first steps into this project...
15:37:50BgerBger it's not wrong, but u've done it verry... how to say it
15:37:55Bgermany little files
15:38:02Bgerdos format
15:38:25peturSorry, used tortoisecvs to create them one by one
15:38:35Bgerthere will be other targets with brightness setting
15:38:40Bgerso i'm adapting it
15:38:43peturthis cvs stuff is all new to me
15:38:56Bgerheh it is nearly the same to me
15:39:04peturI agree with the other targets, was a bit short-sighted of me
15:39:27Bgeru wanted to do it asap, yep ? :)
15:39:40peturmy first real patch :D
15:39:54Bgeri understand u
15:40:11petursorry for the inconvenience
15:40:30Bgerbtw, u've also included another patch in this one
15:40:33petur(at least the code works and is commented)
15:40:42HooliganWhy is it that under Rockbox the H300 LCD doesn't turn off, just the backlight?
15:40:43Bgeryep, u're right
15:41:05Bgercongrats for this
15:41:07BagderHooligan: because no one wrote anything that turns it off
15:41:25Bagderthings usually don't happen by themselves
15:41:25HooliganBadger: Well, I suppose that would do it!
15:41:28Bagdernot even in Rockbox
15:41:49peturLinus, is it not turned off?
15:42:11Bagderno its not
15:42:19Bagder(even though I'm not Linus)
15:42:46peturhehe... more work, but not now and not tonight...
15:43:23peturshouldn't be doing this from at work - gotta go
15:43:36 Quit petur ("CGI:IRC 0.5.7 (2005/06/19)")
15:46:09HooliganAll great things happen on borrowed time. Especially that which you are being paid for.
15:47:12HooliganOh, one more question if you don't mind. Why can't the JPEG viewer be used on the H300 currently?
15:47:30Bagderbecause it isn't adjusted
15:48:29 Join arkascha [0] (
15:49:06HooliganCan Rockbox see why kids love Cinnamon Toast Crunch?
15:53:07amiconnLinusN: Regarding MAX_PATH - utf-8 will triple the size in the worst case.
15:53:50amiconnWhile UTF-8 chars can be 1..4 bytes, it won't be more than 3 bytes when converting from UCS-2 (that's what VFAT uses)
15:55:08LinusNamiconn: ok
15:55:34LinusN768 bytes per entry then...ouch!
15:56:07amiconnEither that, or we need to handle possible overflows
15:56:26LinusNthat is not very nice
15:56:47LinusNsince the user won't be able to rename such a file
15:57:08 Join nathanh [0] (
15:57:12amiconnIt wouldn't be possible to open such a file either
15:57:33LinusNthat may be fixable, but yes
15:57:46amiconnWe could keep the path in UCS-2 internally, that would mean 512 bytes
15:58:00LinusNof course
15:58:08amiconn...but complicate the handling, since it can contain 0 bytes
15:58:42 Join Febs [0] (
15:59:13LinusNamiconn: which character is UCS-2 code 0x0000?
15:59:13 Quit nathanh (Client Quit)
15:59:14markunamiconn: how is the \0 displayed in a filename in windows?
16:00:34amiconnUCS-2 is essentially the same as UTF-16, but covering the 'basic multilingual plane' only
16:01:15amiconnLinusN, markun: I mean that an UCS-2 or UTF-16 string can contail \0 _bytes_ because an UCS-2 char is 16bits
16:01:26markunAh, of couese.
16:01:29amiconnThat means we can't use standard string functions to handle them
16:01:35LinusNof course we can't
16:03:57LinusNi'm mostly worried about the file browser buffer size
16:05:08 Quit Moos (Read error: 104 (Connection reset by peer))
16:05:10amiconnI'm more worried about buffers on the stack
16:05:39amiconnThe file browser buffer size needs to be increased of course
16:06:03LinusNamiconn: yes, stack buffers will be a problem
16:06:58markunI hope the unicode transition will not cause more problems than it solved.
16:07:51Slasheriamiconn: yes, now the tagcache works very well with the fsync call added. Without it, the index files got really corrupted :/
16:08:25LinusNmarkun: initially, it will be shaky, of course
16:08:26Slasheribut fsyncing after each write is of course a way too slow, so the real problem should be fixed
16:08:39LinusNi'm mostly worried about the font chaos we have today
16:09:07LinusNwe have a lot of fonts with so-so quality
16:10:09markunI've been trying to convert the crox* fonts, but it's not a trivial tast because the encoding used is not standard and not consistend between the fonts.
16:10:11LinusNbtw, perhaps rockbox should try to warn the user when selecting a language for which the font has no support. i wonder how on earth we could pull that off, though...
16:10:13 Join Moos [0] (
16:11:05amiconnmarkun: Afaik, most of the crox* fonts use Win1251
16:11:37markunLinusN: Bagder and I were talking about a list of most used chars at the beginning of the language file so they can be preloaded. It could also be used to detect if the language file is supported by the current font
16:12:30markunamiconn: I know, but still they are not right. Some have a E umlaut in the same place where others have an upside down '?'
16:12:41markunamiconn: and some have smiley icons at the beginning
16:12:53amiconn"Warning - 85% of the characters in the new .lng file are not covered by the current font. Do you want to switch anyway?"
16:13:08LinusNamiconn: nice
16:13:21markunyes, that would be great
16:13:39 Join Hooligan| [0] (
16:13:59LinusNand that has to be shown when you switch font as well
16:14:16markunamiconn: maybe iconv's CP1251 table is not correct, or else the fonts look more like a mix of the lower part of latin1 + higher part of CP1251
16:14:23 Quit hshah (Read error: 104 (Connection reset by peer))
16:14:52LinusNsometimes i envy the coders of the original firmware, they only need to care about one or two fonts :-)
16:15:22markunAnd as phaedrus961 noticed: the defauld chars in some fonts should be replaced by something visible.
16:15:36markundefault glyph
16:17:45 Join leftright [0] (
16:19:03saab_riderBger: so that's what it does! (late reply)
16:19:03leftright*quietly* asks amiconn how the wakeup feature is coming along
16:19:34preglowcan't we just delete the fonts that won't work? :-)
16:19:54 Quit mikearthur (Connection timed out)
16:20:07 Join mikearthur [0] (
16:21:14 Quit _FireFly_ ("Easy as 3.14159265358979323846...")
16:22:15leftrightRockbox is just streaks ahead of the Archos gmini's XS202's firmware, my gf got a gminilast night.
16:24:56saab_riderjust wondering, Rockboy works without sound right?
16:25:15saab_riderthanks :)
16:26:35Bgersaab_rider yes, that's what it does
16:28:41saab_ridercheers Bger, cool feature indeed
16:32:21Hooligan|saab_rider: I doubt Rockboy could run in it's present state trying to do sound as well
16:32:33Hooligan|GB music wasn't very good anyway :P
16:33:01Bgersaab_rider heh it's just one added line to enable a feature that already exists
16:33:28saab_riderHooligan: yeah, I just wanted to check to know if I was doing something wrong. and what are you talking about GB music being bad, it rocked :D
16:34:01 Quit Hooligan (Read error: 110 (Connection timed out))
16:34:10Hooligan|saab_rider, I'm still a sucker for NES music myself.
16:34:29saab_riderBger: I always liked the red LED on the H100s, now we have something similar, albeit not exactly the same
16:34:29Hooligan|Speaking of which, it'd be cool if RB could play NSF and SPC files eventually.
16:35:02saab_riderHooligan: you want real old school stuff, MSX music.... if you know what MSX is
16:35:42Hooligan|Now that's old school.
16:36:54Hooligan|But I'd settle for NSF and SPC just the same :)
16:38:59Hooligan|Isn't the SPC decoder just an SNES emulator that only emulates a few bits of the hardware?
16:39:10preglowit's a cpu, in its own right
16:39:12Bgerok, gotta go, nite
16:39:32Hooligan| So it is.
16:39:39saab_ridergood night Bger
16:43:08Hooligan|So how hard would it be to port the SNES APU emulation code to Rockbox?
16:45:03preglowdepends on the source code we can get
16:45:12preglowthe best engine is written in x86 asm, unfortuntely
16:45:36SlasheriHmm, i might just fixed the fs bug.. testing now :)
16:45:55Hooligan|I see. Too bad.
16:46:21Hooligan|Most PC SPC plugins seem to be based on this fellow's work
16:46:21preglowHooligan|: it's far from impossible, just need someone to work on it
16:46:27preglowi'd like to, but i've got a zillion other things to do first
16:46:37preglowHooligan|: yes, and that's the engine i was talking about
16:46:41preglowit's the best i know about
16:46:43Hooligan|Exactly, it's possible but nobody has time to do the work.
16:46:48Slasherioops.. didn't work very well :) couldn't even read/write settings
16:47:50Hooligan|Ah well, there are certainly plenty of more important issues to work on.
16:48:40linuxstbDoes Rockbox understand the System and/or Hidden bits in a FAT directory entry and not show the directory?
16:49:37preglowHooligan|: well, it's just about 13000 lines of assembler that needs translating...
16:50:14 Quit DjDeaf (Read error: 104 (Connection reset by peer))
16:50:38markunDoes anyone actually use the starmap font?
16:51:11preglowgod, i forgot how ugly x86 asm isd
16:51:43linuxstbpreglow: I'm playing with the build you gave me yesterday, and the ATA driver seems fine.
16:52:01preglowlinuxstb: well, it's pretty much got to be a nano issue
16:52:39linuxstbI wonder if it's something to do with the code to spin the disk up and down.
16:52:49preglowmakes sense, i guess
16:52:54linuxstbThat code is working fine for me - but maybe we shouldn't do it for the Nano.
16:52:56preglowbut then i wonder why the data sector write worked
16:53:26 Quit Febs ()
16:54:15 Join XavierGr [0] (
16:54:22preglowHooligan|: looks like it would require a subtantial amount of work to port that
16:54:28amiconnlinuxstb: Rockbox understands the hidden bit and doesn't show the file/folder unless 'Show files' is set to 'all'
16:54:37preglowHooligan|: i wish someone would port it to c, though, it's a great engine
16:54:40Hooligan|preglow: It's never easy, is it?
16:54:57linuxstbamiconn: Thanks. I thought that was the case. The directory on the iPod containing all the music files is hidden.
16:55:17Hooligan|preglow: Sure is. I suppose the SNES freaks will just have to wait it out :P
16:56:11preglowHooligan|: i'm one of them :/
16:58:09Hooligan|preglow: Misery loves company
16:58:20Hooligan|and working codecs
16:58:37preglowright now i want a working ipod button driver
17:00:25leftrighta remote control for dap's would be usefull, for my ass is da laziest ting
17:02:08leftrightbut guess you guys have enough on your plates :)
17:11:11SlasheriFound, fixed and committed the fs bug fix :)
17:11:17SlasheriXavierGr: that problem should be gone now
17:12:16 Quit YouCeyE ("Leaving")
17:14:22***Saving seen data "./dancer.seen"
17:20:57 Quit DJDD_ ("Trillian (")
17:24:39 Nick saab_rider is now known as saa[b_r]ider (n=saab_rid@
17:26:51edxCan anyone tell me where tetris.c has gone, please? −− Im not up to date, it used to be there like a year ago :P
17:27:19edxah "rockblox" got it..
17:35:11linuxstbpreglow: Have you noticed whether the ipod charges when connected to USB in Rockbox?
17:37:07 Part leftright
17:38:13preglowhow can i notice that?
17:38:24CassandraSomething's been bothering me. Is there any conceivable reason to need to be able to pass references to variables in C. I mean is there ever a circumstance where passing a pointer to the variable wouldn't do exactly the same thing?
17:38:36preglowyou know what
17:38:38preglowit looks like it
17:38:51preglowi forgot it plugged in at 'shutting down' for over an hour now
17:38:57preglowwhen i went into disk mode, it wont charge
17:39:32preglowCassandra: references are pointers, just a bit prettier
17:40:12CassandraThat's exactly what I thought.
17:40:22CassandraPointers are more transparent though.
17:41:54preglowcoming from someone used to c, sure
17:41:55*Cassandra is annoyed with libraries that want pointers passed half the time, and references the other half with no predictable pattern.
17:42:04preglowc++ is kind of a hybrid
17:42:11preglowordinary oo languages have just references
17:42:12preglownot pointers
17:42:19preglowand all objects are reference variables
17:42:25preglowso it depends on your perspective
17:42:57 Join Sanitarium [0] (n=sanitari@
17:43:27CassandraExcept C predates OO languages, so that can't be the reason they bothered with references.
17:43:45CassandraI admit pointers take some getting your head around, but it you're going to have them anyway ....
17:44:57preglowyou can shoot yourself in the head with pointers
17:44:59preglownot with references
17:45:02preglowso they're safer
17:45:27CassandraC should not have trainer wheels.
17:45:46CassandraC is all about giving you the freedom to shoot yourself in the foot.
17:46:00preglowi thought we were talking about languages which used references
17:46:11preglowc should be what it is, sure
17:46:27preglowbut i can perfectly well see why someone would embed references into a language and not pointers
17:46:34CassandraC does use references. funct(int& a), for example.
17:46:45preglowis that valid c?
17:46:52CassandraI'm afraid so.
17:47:00pregloware you absolutely sure?
17:47:07CassandraAnd it doesn't mean what you'd expect it to either.
17:47:31CassandraIt means that a is passed as a reference rather than copied to the function.
17:47:43preglowwell, that's what i would have thought it meant
17:48:05preglowbut i thought that was c++
17:48:06preglownot sc
17:48:08preglowc <-
17:48:22CassandraOh, it might be. I thought I remembered it in C though.
17:48:32preglowwell, you might, but it would surprise me a great deal
17:48:55linuxstbpreglow: Have you done anything in the (bootloader) startup code that might prevent Linux and the Apple firmware from running? I can't get either of them to work any more.
17:49:39preglownothing i can think of, i remap memory, but never in the bootloader
17:49:47linuxstbThat's what I thought.
17:49:48CassandraI seem to remeber it being in K&R. We're talking 15 years ago, mind.
17:50:05 Join hshah [0] (
17:50:26*Cassandra suspects that ipod_fw is corrupting the images in some way.
17:50:32preglowwell, i can't possibly imagine what references would be doing there
17:51:04preglowpointers cover all that's needed
17:53:43preglowlinuxstb: not linux either?
17:53:51linuxstbpreglow: Nope.
17:53:56preglowi can't get either of them to work either
17:53:59preglowand my bootloader is old
17:53:59linuxstbI'm just trying an old version of crt0.S now.
17:54:23 Part Kaggen
17:54:24Cassandrapreglow: Exactly. And if they do it in C, they do it in C++.
17:54:48CassandraI suppose it enables one to write pointer free code, but why, oh god, why?
17:54:50preglowCassandra: yup, but like i said, c++ is kind of an hybrid
17:54:59linuxstbpreglow: Do we enable interrupts (e.g. the timer) in the bootloader now?
17:55:05preglowlinuxstb: no
17:55:17preglowoh oh oh
17:55:22preglowwait a bit now
17:55:28preglowsystem.c is included in the bootloader?
17:55:43linuxstbYes - the bootloader calls a lot of the init functions.
17:55:48preglowthen that needs fixing
17:55:53preglowi init timers and all that stuff there now
17:56:02linuxstbI just tried an old crt0.S and it doesn't work either - so it's not that part of the code.
17:56:09preglowbut that's not related, like i said, my bootloader is ancient, and it can't load retailos either
17:56:41preglowi think we should continue to live without interrupts in the bootloader, btw
17:57:10preglowit would be nice for unity, but i don't know what linux and retailos says to being called with bunches and interrupts and remapped memory
17:57:11linuxstbpreglow: Yes. But the problem is detecting a keypress reliably.
17:57:22preglowlinuxstb: why do we need that? we're rockbox!
17:57:36linuxstbSure :)
17:57:40preglowi haven't used retailos yet, and i'm not going to start now
17:58:12preglowCassandra: but yeah, c++ took a lot of baggage from c that it shouldn't, imho
17:58:25preglowCassandra: but of course, if they hadn't done that, we probably wouldn't been using c++ today
18:00:51 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
18:01:19CassandraWouldn't that be nice.
18:01:26CassandraC++ sucks.
18:02:14preglowyes, it isn't very nice at all
18:02:19preglowunfortunately it still manages to be quite usable
18:02:33preglowand pretty fast
18:03:51CassandraThe latter surprises me somewhat. It ought to run like a dog as far as I can tell.
18:04:05CassandraStill, it's probably the best OO language available.
18:04:14Cassandra(I hear good things about python though)
18:04:34Cassandraand for a windowed application, OO is kind of useful.
18:04:36preglowi rather prefer most other oo languages to c++
18:04:43preglowc#'s nice enough
18:05:55preglowi never could like python
18:07:40Cassandrac# always struck me as MS' plan to close C++.
18:08:14CassandraNever used it though.
18:09:49preglowwell, of course
18:09:54preglowit's a nicer language
18:09:56preglowthat's for sure
18:10:06preglowc++ is too complex
18:10:10preglowand the complexity doesn't help it
18:16:45preglowcan anyone explain to me why iriver can't play parts of vorbis files with low bitrate?
18:17:12preglowi thought this was an issue of using a too long window size, but that can't be true when it also can't play local parts of a track that's using a small bitrate
18:28:46linuxstbhehe. Cube is slow on the ipod at whatever the default clock speed is.
18:30:26preglowprobably does some faulty frame skip
18:37:35 Join Lear [0] (
18:39:57 Quit Hooligan| ()
18:44:42preglowlinuxstb: any new ideas on what's wrong with bootloader?
18:49:36 Join Febs [0] (
18:50:38linuxstbpreglow: No.
18:51:00linuxstbI may go back through some old CVS versions and see when loading the original firmware broke.
18:52:21 Quit ender` (Read error: 104 (Connection reset by peer))
18:53:50 Join Mmmm [0] (
18:55:07 Join ender` [0] (n=ender@
18:58:14preglowwhat's up with it just suddenly stops responding for large amounts of time
18:58:24linuxstbbootloader from 20 November works....
18:58:38preglowok, what's different?
18:58:52MmmmThe forum is not quite up to speed yet either
19:00:20 Join Acksaw [0] (
19:02:31preglowsomething's funny
19:02:38preglowsuddenly it comes back for about one second
19:02:40preglowthen goes again
19:05:16linuxstbpreglow: Your made your first set of startup code changes on the 20th - the bootloader after that works fine...
19:05:32preglowwhat did i change then?
19:05:36preglowright, remember
19:06:47preglowbut yeah, i'd be surprised if it was anything in crt0.S
19:08:54preglowam i the only one having troubles with the site?
19:09:22linuxstbIt seems fine for me.
19:09:46preglowdoesn't answer pings here
19:09:48preglowit does now
19:09:53preglowno, stopped
19:10:32linuxstbwebsite still fine for me. ping is fine.
19:11:16linuxstbMmm. 7 December bootloader is fine...
19:11:52preglowthat can't be right
19:11:57preglowi flashed mine waaaaay before that
19:12:29linuxstbLet me double-check.
19:13:40 Join _FireFly_ [0] ( is completely broken for me
19:13:46preglowoh well
19:14:24***Saving seen data "./dancer.seen"
19:14:30_FireFly_preglow: for me only the forum
19:14:59linuxstbYep, 7 December bootloader still loads the retailos and Linux for me.
19:15:17preglowthen it's pretty much got to be my crt0.S changes fine for me but forum is sloooooooow!
19:17:22 Join hshah_ [0] (
19:18:08preglowthere's got to be something wrong with the connection here
19:18:11MmmmFirefly was that you with the flicker fix?
19:18:43_FireFly_Mmmm: yepp
19:19:16MmmmWell done...its perfect...You made me a happy man :D
19:19:55preglowthis is the weirdest connection problem i've ever seen
19:20:17preglowi wonder if i should even bother trying to restart this piece of shit
19:21:15linuxstbpreglow: 12 December (00:00) bootloader doesn't work. Getting closer...
19:22:16 Part Polo_o
19:24:28preglowreboot, brb
19:25:01linuxstbOK, looks like I broke it on the 11 December...
19:27:53_FireFly_Mmmm: could you test this patch please ??
19:28:13_FireFly_this is an combination of my already commited patches
19:28:46 Join webguest58 [0] (
19:29:08_FireFly_if this also works then is this a more simple solution as the two other together
19:29:12preglowlinuxstb: how?
19:29:33 Quit webguest58 (Client Quit)
19:29:58linuxstbI've no idea - it was my i2c and rtc code which isn't used in the bootloader afaics
19:30:35linuxstbBut 2005-12-11 00:00 works, and 2005-12-12 00:00 doesn't.
19:31:19MmmmFirefly: I would but what is there to patch? CVS doesn't flicker now! Or shall I test on an old build?
19:31:54_FireFly_Mmmm: this patch is a more simple one which also reduce the codesize a bit
19:32:28_FireFly_Mmmm: this patch is against the latest cvs
19:33:36MmmmOkey Dokey... :)
19:34:47 Quit hshah (Read error: 113 (No route to host))
19:35:39 Quit Sanitarium (Read error: 110 (Connection timed out))
19:36:35 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:38:20 Quit _FireFly_ (Read error: 113 (No route to host))
19:38:42 Join _FireFly_ [0] (
19:40:03MmmmNot a flicker in sight!
19:40:19_FireFly_:) still acts up
19:41:03 Join tvelocity [0] (n=tony@
19:41:22MmmmHow can this patch reduce the codesize? it has more additions than subtractions! or does it make some code redundant?
19:42:17_FireFly_Mmmm: some code have a simply commented
19:42:26_FireFly_for testing
19:42:31MmmmOh yeah ! :D
19:43:34MmmmI really should take a closer look before i start talking rubbish eh?
19:43:57_FireFly_i do just an test if this really reduce the code-size
19:44:00_FireFly_on iriver
19:44:24linuxstbpreglow: I don't know what yet, but it was definitely something done on the 11 December.
19:46:36_FireFly_yepp it reduce code-size abit :) 250020(org) to 249996
19:46:54_FireFly_24 byte decrease
19:47:44_FireFly_this is/was a three-step way to get a good result :)
19:49:17preglowmarkun: yo, you use freebsd, yea?
19:49:56preglowhard to get a dev env working?
19:50:07MmmmGood one firefly... Next you've got to fix that darn ticking once and for all!! :D
19:50:33_FireFly_the ticking it self is hardware related
19:50:40Learhm... can you do a jump table in an asm statement? :)
19:50:54MmmmOk but improve the fix then
19:51:36_FireFly_but with the reduce-ticking option the ticking is at least on my device not noticeable
19:52:00preglowLear: yes?
19:52:32MmmmIn a quiet room you can only just hear it, but the main prob is the fact that oggs dont play when it's switched on! too processor intensive i think!
19:52:36Learbit unsure about syntax in general for asm() stuff, so how would the labels be?
19:52:51preglowLear: have a look at coldfire.S in ffmpegFLAC
19:53:07preglowLear: featuring the most efficient jump table implementation i could think of
19:53:31_FireFly_Mmmm: sometimes i have the feeling that the ticking is also not noticeable although the reduce-ticking option isn't enabled
19:54:09LearHm.. The leading "2" in the jump instruction itself (or rather, the address word)?
19:54:55MmmmFirefly: Really? you are lucky! It seems that there are all sorts of units with different ticking levels out there.
19:55:22preglowLear: ip always points to the extension word of the instruction currently executing
19:55:31 Join tucoz [0] (
19:55:45preglowLear: that's just for correcting
19:56:02tucozpreglow, just to let you know. The rockbox site is weird for me as well.
19:56:03wubbla_FireFly_: reduce-ticking reduce-ticking
19:56:08wubbla_FireFly_: reduce-ticking option
19:56:14wubbla_FireFly_: sorry :-)
19:56:20preglowtucoz: perhaps it's a norwegian pipe problem
19:56:34_FireFly_wubbla: ?? what do you want to try to tell me ??
19:57:17wubbla_FireFly_: i just wanted to ask you about your reduce-ticking option... (e.g. where to find it in the settings menu) ;-)
19:57:52_FireFly_the reduce-ticking option isn't from me it's from Slasheri
19:58:07preglowLear: the jmp instruction is coded as two words for that particular addressing mode, the first word is the instruction itself, the second is the addressing data, and the instruction pointer always points at the extension word (even if the instruction hasn't got one), so i need to add 2 to make it point at the first jump table entry
19:58:08_FireFly_wubbla: it is located in the remote-lcd-settings menu
19:58:39MmmmFirefly: My problem is that with a complicated WPS (lots of scrolling) and ticking fix on, high bitrate oggs skip! The only way to play them is to go into the filebrowser.
19:58:50wubbla_FireFly_: wasn't the flickering related to the h300's main display?
19:59:17Learyes, I suspected that much; I did do quite a bit 68000 programming a number of years ago...
19:59:19 Quit hshah_ (Read error: 113 (No route to host))
19:59:58MmmmWubbla: mode->General settings->Display->Remote-lcd settings->Reduce ticking it's for h1xx remote ticking
19:59:59_FireFly_wubbla: on h3x0 that's true i had noticed only on the h1x0 remote but the code which i had changed is used for both displays
20:00:31_FireFly_wubbla: i think you mix something :)
20:01:15Mmmmright...I'm hungry..Someone's got to cook round here!
20:01:27wubbla_FireFly_: indeed... ;-)
20:01:27_FireFly_the image-flicker patch has no option :) why want someone that all images flickers :)
20:01:40 Quit Mmmm ("Goes to heat up the stove")
20:03:15 Part tucoz ("Leaving")
20:03:33wubbla_FireFly_: that's why i was that astonished
20:06:48 Join lamed [0] (
20:07:19_FireFly_LinusN: now i have a more simple solution for the image-flicker-problem :)
20:08:20_FireFly_LinusN: with which i could also reduce the code-size a bit
20:08:27_FireFly_for lcd-bitmap targets
20:08:54LearHm... gcc 3.4.5 is out...
20:09:38 Quit lamed (Client Quit)
20:09:44_FireFly_Lear: and ??
20:10:12LearJust as a note to those who do builds... :)
20:12:31LinusN_FireFly_: nice
20:13:11_FireFly_the new code size for rockbox.iriver is 29952
20:13:34_FireFly_ups i mean 249952
20:13:36preglowLear: yeah, noticed, i'll be trying it out now, since i'm just about to build a cross compiler anyway
20:13:36LinusN29k! that's what i call reduced
20:13:59_FireFly_ok next try to type it correct :)
20:14:38_FireFly_the new size is 249952 to 250020
20:15:34_FireFly_it isn't much but any code-size reducement is a good one :)
20:17:43 Join webguest99 [0] (
20:17:44_FireFly_LinusN: but how did you come to a 29k reduce ??
20:18:08_FireFly_was it a bit math magic ?? :)
20:18:17webguest99preglow: do your vbr files skip on your nano ?,
20:18:54preglowwebguest99: i don't use the apple os
20:19:08webguest99wonder if apple purposely sabotaged mp3 ?, its a long shot
20:19:41webguest99after all they do know how to program codecs,
20:20:19preglowi don't really care
20:20:29preglowi can see how that can be a legit bug
20:20:33preglowbut they should of course fix it
20:20:42webguest99Long Live Rockbox
20:20:44preglowthen again, apple isn't known for listening too much to user requests
20:20:56webguest99they dont really give a toss
20:21:23webguest99glad creative is having a go at them, a bit of their own medicine for a change
20:21:26Learin an asm(), do I need to tell that I clobber an input operand, even if I'm not interested in that operand afterwards (I return soon after)?
20:22:08preglowLear: yes
20:22:20preglowLear: all input operands that are clobbered, need to be specified as output operands
20:22:20 Part webguest99
20:22:43preglowif i remember correctly
20:23:03Learwith earlyclobber ("&") set, I suspect...
20:24:06preglowwill be fun to see how the ifp port fares
20:26:33Bagderthat emulator approach sure is effective
20:26:37Bagderbut a lot of work
20:27:09preglowhell yes, i can't see how he did it by himself
20:28:07LearOuch... I want to inline some asm - with labels - , but it can be used in multiple places in one function
20:28:20Lear, so I get label redefinitions. How avoid that?
20:28:29preglowuse local labels
20:28:44preglowlook at crt0.S
20:28:56LearI do that already - I think - as in ".4:", no go... I'll check though.
20:29:08preglowi mean pure numerical labels, like 1: all the way
20:29:20preglowdepends how far you need to branch
20:29:56Learyep, that worked, thanks.
20:35:05_FireFly_i have another idea to reduce the code-size in the wps-code and also remove an loop
20:35:21preglowi tried to test rockbox compiled with -mthumb
20:35:49preglowit crashed with some of the libgcc functions, however :/
20:37:43preglowlibrockbox.a only differs by 24kb
20:38:20_FireFly_when compiling with -mthumb ??
20:38:47LearYay, it actually seems to work, once I got past the assembler. :)
20:38:50linuxstbpreglow: Did you use -mthumb-interwork ?
20:39:04preglowlinuxstb: that's what crashed
20:39:09preglowlinuxstb: libgcc didn't support that
20:39:17Leardang, forgot make zip, so I don't know yet... :)
20:39:19preglowLear: what're you assembling?
20:39:42Learsome assembler optimizations (attempts to, at least) for Tremor...
20:40:08 Join DangerousDan [0] (
20:41:27preglowdid you look at lostlogics patch?
20:42:15 Join zeero [0] (
20:43:09zeerodoes anyone know if the x11 version of the sim ever worked properly? looking at the code i don't see how it would ever have draw the image for the player you were using, for example
20:43:49LinusNzeero: of course it has ever worked, and it still works
20:44:01LinusNbut not for the h300
20:44:31LinusNand it doesn't draw the player image in the x11 version
20:44:48zeeroahh, okay
20:45:07zeeroi assumed it worked for drawing the lcd, but i wondered about the player image
20:45:20_FireFly_yeah another 16 bytes reducement
20:45:46LinusNzeero: the x11 sim developers didn't find the player image very important
20:46:21zeeroLinusN: fair enough
20:47:03LinusNzeero: but you are most welcome to port the x11 sim to SDL if you want
20:48:03zeeroLinusN: i imagine i am, and if i manage to find the time, i just might
20:48:11preglowbagder also said he might do this
20:48:47zeeroit doesn't seem like it would be too hard, given that the lcd drawing code is already there, but i've only been looking at the code for a day, so i've probably missed something
20:48:53linuxstbpreglow: I've discovered one thing - the Apple OS seems to want the COP enabled.
20:49:08preglowlinuxstb: lucky it, i'll enable it soon
20:49:13linuxstbLinux doesn't mind either way.
20:56:25_FireFly_ok the new patch makes a code-size-reduction about 84-bytes :)
20:58:13wubbla_FireFly_: just to make that clear: you're not trying to get rid of the h300 flickering, are you? ;-)
20:58:34_FireFly_the image-flickering is gone and already in cvs
20:58:43_FireFly_i only improve the patch :)
20:58:53_FireFly_as some h3xx users has already said
20:59:02 Quit Acksaw ()
20:59:17wubblaPatch #1379076 by Stephan Wezel - reduce the ticking problem a bit
20:59:37wubbla_FireFly_: that is yours right?
21:00:01_FireFly_ Patch #1377894 by Stephan Wezel - less image flickering in WPS
21:00:58_FireFly_wubbla: yepp the second patch reduced the ticking-noise to the "old" value because the first patch which had solved the image-flicker-problem has increased the ticking-noise
21:04:08wubbla_FireFly_: i'm using the latest CVS source and the flickering is still there...
21:04:49wubbla_FireFly_: i might add that i changed the background/foreground colours in lcd.h (white fg on darkblue bg)
21:04:54 Join solexx_ [0] (
21:05:14wubblamaybe the flickering is less visible on dark bg and light fg...
21:05:15wubblano idea
21:06:14wubblai'll check that...
21:06:58_FireFly_wubbla: then the only reason which i due this see is that the lcd-codes for the h3x0 isn't yet optimised
21:07:08_FireFly_s/due/see due
21:08:00wubblahmm... LinusN said something about DMA access for the LCD/HDD some days ago...
21:14:28***Saving seen data "./dancer.seen"
21:14:39preglowfor a second there, i actually considered writing an ipod emulator
21:16:44 Join Sanitarium [0] (
21:20:01XavierGrnow I am completely confused. I made a plugin and included tree.h while it perfectly included the tree_flush() fn it will not do the same for get_tree_context(). why is that?
21:20:48 Quit solexx__ (Read error: 110 (Connection timed out))
21:24:49LearBecause it is called tree_get_context()? Not that it is something that's accessible from a plugin anyway...
21:25:14 Quit Maxime` (Read error: 104 (Connection reset by peer))
21:25:20 Join Maxime [0] (
21:25:27LinusNwubbla: i have made experiments with dma for the h300 lcd, and it seems that it won't work
21:25:52LinusNprobably due to the lcd bus timing constraints
21:26:02wubblaLinusN: hum. :-(
21:26:28wubblaLinusN: this would have made the access to LCD lightning fast, right? ;-)
21:27:06LinusNwell, a lot faster anyway
21:27:35wubblaLinusN: it didn't work at all?
21:27:42LinusNi haven't given up yet, but i'll have to hoox up the logic analyzer to the lcd bus
21:28:03LinusNit worked so-so, the lcd didn't always recognize the writes
21:28:24_FireFly_what a "stupid" lcd :)
21:30:34LinusNmy guess is that the time-between-accesses constraint isn't met by the coldfire bus controller
21:30:52LinusNi'll know more when i do the bus measurements
21:31:05preglowcome to the other side, linus, get an ipod
21:31:13preglowthe water's nice over here!
21:31:15LinusNpreglow: i just might
21:31:25LinusNbut only after the h300 and the x5
21:31:27preglowleaving the porting to the n00bs is irresponsible!
21:32:02dwihnoYou are great! The ipod port will be really sugarcake:ish! :)
21:32:03LinusNpreglow: exactly, that's why i don't leave the h300 ;-)
21:32:06dwihnommm, sugarcake
21:32:23dwihnox5 = iaudio disk thing?
21:32:35wubblaLinusN: you have to be strong and simply ignore these persuasions ;-)
21:33:06LinusNit's hard...the dark side is tempting...easier...more elusive...
21:33:27preglowwe have beer!
21:33:34LinusNdone deal!
21:34:05*LinusN walks away for a while
21:34:10*preglow brands linus with an apple logo
21:36:10 Join Midgey34 [0] (
21:36:27wubblaLinusN: don't you want to be alternative and "oldschool"? remember that rockbox is an opensource project, that's why you simply have to use alternative hardware instead of supporting monopolies :-)
21:37:39preglowpft, it's a fringe case, they merely have 90% of the market
21:37:50preglowthere's a "mono" in "monopoly" you know!!!!111one
21:39:02wubblafringe case... pffff...
21:40:33preglowbesides, it's clear he's in it for the fame and glory, imagine what the chicks'll be like when they find out they're running his code on their nice shiny ipods
21:40:44linuxstbpreglow: Congratulations on the assimilation. amiconn is next...
21:41:19preglowlinuxstb: he's a tough nut, but he'll crack under some skilled brainwashing
21:41:26_FireFly_ipods are nice but the to must use an db is dumb
21:41:34*wubbla simply knows that LinusN will never ever convert to the dark side...
21:42:27preglowlinuxstb: i know a guy at apple who's known for creating reality distortion fields. i'll have him talk to amiconn, and we'll see
21:42:28_FireFly_i like to simply through some files on the drive
21:42:44preglow_FireFly_: well lucky you, rockbox on ipod will be like that
21:42:51preglowas a matter of fact, it already is!
21:43:06_FireFly_:) then i only need the money to buy an ipod :)
21:43:33_FireFly_preglow: or do you want to give an ipod as a gift to me ;) *g*
21:43:37linuxstbTwo converts in 5 minutes. Who's next?
21:43:44preglow_FireFly_: "yes"
21:44:13XavierGr"NIce" struct entry is redifined in dbinterface!
21:44:30preglowlinuxstb: starting to look dim for the button driver today
21:44:40XavierGrIf I inlcude tree.h in the plugin api there is a clash between them
21:45:11linuxstbpreglow: I know how you feel. This bootloader's driving me mad.
21:46:28 Join PaulJ [0] (
21:46:44linuxstbXavierGr: Which struct are you talking about?
21:46:56preglowwoo, i just shaved off twenty bytes or so in crt0.S for arm
21:47:11_FireFly_preglow: :)
21:47:23XavierGrstruct entry in tree.h is redefined in dbinterface.h
21:47:31XavierGrthere is no colision yet because tree.h isn't included in the plugin api.
21:47:32preglowhahha, dancepuffduo wps rocks
21:48:25preglowlinuxstb: we've gotta start giving thought to what to do with the cop as well
21:48:40XavierGrI changed the struct entry to sturct dbentry in "dbinterface.c and dbinterface.h" and now it will compile fine.
21:49:45 Join petur [0] (
21:50:14linuxstbpreglow: I think that depends on how the CPU scaling will work. IIUC, (from the PP propaganda I think), the main CPU can be clocked at any speed, and the COP runs at either 100% or 50% or 0% of the speed of the main CPU.
21:50:28preglowcop has clock scaling, yes
21:50:39linuxstbBut the COP can't run faster than the main CPU.
21:50:41preglowso you're saying we might want to use the cpu for codecs
21:50:48preglowand cop for everything else
21:51:24linuxstbThat's one option.
21:51:31preglowwell, it does make sense
21:51:45preglowwe wont need a very high clock rate for the cpu that just does gui handling and plugins
21:52:09_FireFly_yeah decoding needs more power
21:52:10preglowmost of the time, at least
21:52:30 Part mind_less
21:53:05XavierGrSlasheri: Remember that you told me that it would be easy to acces the directory data from the dir buffer? Well in fact I can access them if I inlude another function in the plugin api. This way the jpeg viewer will have no lag for sorting or large buffer arrays for filenames.
21:56:30linuxstbpreglow: What have you learnt about the cache? Does that affect how we can use the COP?
21:56:56 Quit tvelocity ("Leaving")
21:57:23preglowwell, yes
21:57:45preglowthe bus is not snooped, so we can't count on memory being coherent between processors
21:58:12 Quit Cassandra (Read error: 110 (Connection timed out))
21:58:16preglowso if one processor needs data from another, that always has to go through iram
21:58:35linuxstbSo basically that means that they can't use the same data structures?
21:58:44linuxstb(unless they are in iram...)
21:58:45preglowunless they're in iram: correct
21:59:01 Join Noel_sad_song [0] (
21:59:02linuxstbSo that's not too bad then.
22:00:36preglowperhaps some of the gurus can elaborate on how much data is actually shared between the playback thread and the rest
22:00:41preglowor guru: slasheri
22:02:42preglowlinuxstb: i've put some of the consts that differ between 5002 and 5020 and moved them to an ifdef at the top, so we can minimise ifdefs in the code itself, ok?
22:02:56preglowi dislike ifdefs in code
22:03:46linuxstbThey must just come from the fact that IPL uses real if statements.
22:06:17linuxstbCan you give me a build running at 75MHz? I want to test the speed of wav2wv
22:06:28preglowgimme a sec
22:07:41*preglow crosses fingers
22:07:57preglowlet's just hope i haven't broken crt0.S in the meantime
22:08:00preglowanywho, you'll find out :>
22:08:32XavierGrplease tell us the results, it would be fun to know how a 75Mhz processor did comparing to an 120 one.
22:09:02LearWell, in this case it would really be mips...
22:09:20Lear(As in meaningless information about processor speed :) )
22:09:23XavierGrof course
22:09:35 Join hshah [0] (
22:09:38 Join linuxstb_ [0] (
22:09:50 Quit linuxstb (Nick collision from services.)
22:09:54preglowlinuxstb: oh, and the button driver is probably on crack
22:09:55 Nick linuxstb_ is now known as linuxstb (
22:09:58preglowshould work, though
22:10:18markunmight still give a nice indication how the 300MHz of the Gigabeat will perform
22:10:39preglowif it doesn't just crash, that is
22:10:48preglowit's a miracle if this build works at all
22:11:17 Join San [0] (
22:12:12 Join Gibbed [0] (
22:12:14 Quit Rick (Nick collision from services.)
22:12:23 Nick Gibbed is now known as Rick (
22:12:35markunpreglow: you didn't test it on your nano?
22:12:53preglowmarkun: it didn't even compile here
22:12:59preglowi think
22:13:06*preglow finds a wav
22:13:13 Join Cassandra [0] (
22:13:24linuxstbTesting now...
22:13:33linuxstbAbout 113% realtime
22:13:42preglowdamns fucking amazing
22:13:46preglowthat's <-
22:13:49XavierGririver one?
22:13:50linuxstbNot bad :)
22:14:05preglowXavierGr: ipod
22:14:09linuxstbAnyone remember the % on the iriver?
22:14:09preglowat 75mhz
22:14:15preglowaround 140, i believe
22:14:32linuxstbWho needs a cop?
22:14:39XavierGryes I meant how the iriver can manage with the same file?
22:14:39preglowno one, that's who
22:14:46linuxstbXavierGr: I'll test now.
22:14:47 Join Gibbed [0] (
22:14:50 Quit Rick (Nick collision from services.)
22:14:59preglowdoing a test myself
22:14:59 Nick Gibbed is now known as Rick (
22:15:04XavierGrI am shaking....
22:15:16preglowlinuxstb: riiight, open with menu doesn't work
22:15:45XavierGrif the iriver cant get way better than this.... it will be disastrous fir irver fans...
22:15:47preglowit does
22:15:56preglowbut it only lists two things: searchengine and searchengine
22:16:01linuxstbpreglow: You need to compile wav2wv - it's not compiled by default.
22:16:41preglowwe don't have SWCODEC defined yet?
22:16:54preglowahh, right
22:17:09preglowcan i just remove the line at the top of SOURCES?
22:17:20 Join jlo [0] (
22:17:28jlohi all
22:17:43linuxstbpreglow: No - you will get button errors in all the plugins. Just move the plugins you want outside SOURCES.
22:17:50XavierGrhi jlo
22:17:57linuxstbI mean outside that #ifdef in SOURCES
22:18:25jlopreglow : I'm just uploading some lsd for you to test
22:18:35LinusNpreglow: can't you share the main memory by using semaphores and mailboxes in iram?
22:18:57_FireFly_LinusN: mailboxes ??
22:19:18linuxstbThe same file encodes at 200% realtime on the h140...
22:19:20preglowLinusN: we seem to have dedicated mailboxes with interrupts, btw
22:19:21LinusNor message queues
22:19:30LinusNpreglow: goodie
22:19:34LinusNthere we have it
22:19:36preglowjlo: wont have time today, gotta go in ten minutes
22:19:42mirakpreglow: hi, I a reading xvid code ...
22:19:49XavierGrthat's good for the iriver tight?
22:19:50LearInteresting, having a static const array in a static inline function works; wonders how much space it needs... :)
22:19:58 Quit arkascha (Read error: 110 (Connection timed out))
22:20:17mirakpreglow: I wonder how it will be able to ever compute that
22:20:29jlopreglow : just try it when you have time
22:21:27LearPhew, only got one extra copy...
22:21:54 Join linuxstb_ [0] (
22:22:06 Quit linuxstb (Nick collision from services.)
22:22:35 Nick linuxstb_ is now known as linuxstb (
22:22:54preglowjlo: sure, just msg me an url or something
22:23:45preglowlinuxstb: is there some spinup wait in the ide code or something?
22:24:49jlopreglow : you get in a subdirectory lsd at same place than yesterday
22:25:18preglowjlo: then please msg me that url, i had to reboot so lost it
22:26:45jlopreglow :
22:27:12linuxstbpreglow: I don't know the ATA code that well - it's all generic. I basically just added the register addresses for the ipod and fixed some endian issues.
22:27:40preglowlinuxstb: no matter what i do, i get a "cannot open /.rockbox/viewers/wav2wv.rock" message
22:27:50 Quit Sanitarium (Read error: 110 (Connection timed out))
22:28:25linuxstbWhat happens if you try and play the WAV file? I get the WPS appearing.
22:28:30preglowme too
22:28:59mirakhem, in C if you have a function called myfunction you can assign it to a variable ?
22:29:06miraktransfer_8to16copy = transfer_8to16copy_altivec_c
22:29:11mirakI see that in the code
22:29:26preglowhere it goes
22:29:29preglow140% realtime
22:29:39mirakthe function itself if used without () is pointer ?
22:29:53preglowdamn, this is impressive
22:30:01^BeN^h300 users: big images soliter by me =]
22:30:21preglowlinuxstb: i say we just let the cop sleep if this is any indication of what stuff is going to be like
22:30:37mirakpreglow: hey are you ignoring me or something ?
22:30:45linuxstbYep - I think we can get a fair distance without using it.
22:30:58preglowmirak: i'm a bit stressed here, i should have been running towards the bus now
22:31:07mirakoh go ahead
22:31:10 Quit hshah ("Leaving")
22:31:17preglowlinuxstb: but okie, there's a bug in the ata driver somewhere
22:31:26peturBger, are you around?
22:32:24LinusNmirak: yes, the function name without the () is treated like a pointer
22:34:10preglowok, on the same file, wav2wv on h120 yields: 198% realtime, peak
22:34:12peturLinusN: what's the difference between GPIO1 pin 49 (backlight on) and switching the PWM on? is it connected to an enable or something?
22:34:15preglow150 for arm
22:34:19petur(just curious)
22:34:21preglowi say we're probably going to be fine...
22:34:42LinusNpetur: the GPIO1 pin is used as an enable signal
22:35:16peturdo you know how they do contrast?
22:35:35preglowbut ok, gotta go, later
22:35:47linuxstbpreglow: The whole playback code seems to work just fine. It seems that the writing of the output data is blocking. So it should be possible to write a dummy audio driver that just discards all the data. We can then see how fast it is.
22:36:21linuxstbBut goodnight - don't miss your bus.
22:38:15XavierGrhow he manages to be online when he have to catch a bus?
22:41:23 Join webguest02 [0] (
22:41:47 Join RotAtoR [0] (
22:42:32 Part Noel_sad_song
22:42:59LinusNpetur: maybe they fiddle with the grayscale voltages
22:43:12 Join qwisp11 [0] (
22:46:53peturwould we care to do contrast? would be nice once we do colour... just wait until that is ready.
22:46:53qwisp11does anyone know the extension for saving sudoku games?
22:46:57 Join DrMoos [0] (
22:47:21_FireFly_qwisp11: afaik .ss
22:47:37mirakso I want to adapt a pogramm that uses malloc and free. What method did you use for other codecs that use that ?
22:47:41_FireFly_these are simple-sudoku-files
22:48:16markunmirak: are you porting the ffmpeg mpeg-4 decoder?
22:48:37mirakmarkun: I am reading the code at least ...
22:49:17qwisp11FireFly: I'm getting save failed. Doesn't matter. May well not be the extension I'm using. Thanks anyway
22:49:32mirakmaybe LinusN knows
22:49:56 Quit Moos (Read error: 104 (Connection reset by peer))
22:50:06 Nick DrMoos is now known as Moos (
22:50:17peturBger wasn't quite happy with my brightness patch. I can understand I need to get those line endings right, but he also said 'too many files'
22:50:38peturshould I create a patch that contains multiple changes?
22:50:57markunpetur: can I see you patch?
22:51:39LinusNpetur: cvs can generate a single patch file for you
22:51:56peturbut I see Bger has re-submitted it
22:52:15webguest02markun: did you see that scrolling in the dir browser using a unicode font is too slow at 45mhz?
22:52:39markunwebguest02: H1x0?
22:53:10markunno, didn't notice
22:53:17webguest02At least it is on large dirs.
22:54:13markunwebguest02: which font did you try?
22:54:22webguest026x13 and unifont at least
22:54:30markun6x13 doesn't seem very slow here
22:54:51webguest02Do you have music playing?
22:55:21markunI'll do some more tests later
22:56:02webguest02It will keep scrolling for a few seconds after I release the button
22:56:07LinusNmaybe we should put some of the unicode functions in iram
22:56:55webguest02I should remember that setting to default settings will turn on the voice ui. Scared the crap out of me.
22:57:03Learmaybe the font cache isn't large enough?
22:57:20webguest02Strange that you can't replicate this
22:57:30mirakLinusN: hem, how did you resolved the malloc and free problem for the codecs ?
22:57:56Learmirak: ugly malloc hack without a working free. :)
22:58:33webguest02I can also make the the music skip by scrolling if playing music.
22:58:44mirakLear: in fact you do a big free when all mallocs are supposed to not be used anymore
22:58:48Learany disk access during scroll?
22:59:19webguest02Lear: hard to tell.. once in a while
22:59:24webguest02Might as well be buffering though
23:00:09Learshouldn't be hard to tell, I'd think... If it is, it could mean rockbox needs to load font characters from disk; depends on what kind of filenames you have.
23:00:35webguest02Only stuff that's in latin-1.. how big is the font-cache?
23:00:54 Quit jlo ()
23:01:08webguest02The skips aren't connected with harddisk access.
23:01:23webguest02that much's for sure
23:01:35webguest02I get skips without hdd access.
23:02:21webguest02I suppose I'll try to create a directory of "easy" filenames
23:02:50peturLinusN: those I2C H300 pin assignments for FM tuner, are they correct? (GPIO1 pins 56/57)? So I should just update the defines in fmradio_i2c.c and try?
23:03:31LinusNpetur: hang on
23:03:46*petur swings to Linus
23:04:29 Quit perplexity (No route to host)
23:06:08_FireFly_good night everybody
23:06:11 Join linuxstb_ [0] (
23:06:32 Quit _FireFly_ ("Leaving")
23:08:00webguest02Aha. So. Even with the very few non-latin1 characters, that still matters. A dir with only a-z filenames will not result in sluggishness
23:09:11LinusNpetur: i just double-checked, they seem to be correct
23:09:32LinusNmake sure you change the audio mux select pin as well
23:09:32peturgreat, will try to get fm radio working then
23:09:47peturgood you say that
23:10:11mirakLinusN: have you heard of someone creating is own malloc function ?
23:10:21LinusNmirak: yes, why?
23:10:38mirakLinusN: in rockbox ?
23:10:45mirakfor rockbox ?
23:10:55markunthere is a fake malloc
23:11:16mirakmarkun: and a fake free ?
23:11:30markunmirak: yes, but the fake free doesn't free any memory :)
23:11:39linuxstb_mirak: You should really make every effort to replace mallocs with static buffers. It will also help when optimising - you can put the most-accessed buffers in fast IRAM.
23:11:55 Quit Midgey34 (Read error: 110 (Connection timed out))
23:12:23miraklinuxstb: maybe having something that run with as few modification as possible could be a good start
23:12:28zeerois the lcd api that needs to be implimented by the sim specified anywhere?
23:12:45webguest02zeero: graphicsapi in the wiki
23:12:52linuxstb_mirak: Yes, so just use the fake malloc/free approach that we already use for the other codecs.
23:13:11zeerowebguest02: thanks :)
23:14:03LinusNzeero: basically, the only functions you should need to port are lcd_update() and lcd_update_rect()
23:14:27zeeroLinusN: okay cool, thanks
23:14:30***Saving seen data "./dancer.seen"
23:14:52LinusNthose are the functions that copy the frame buffer contents to the lcd
23:15:02webguest02Hrm. So this directory where I'm getting sluggish scrolling has 17 files (of 400) containing non-ascii characters. 7 different ones.
23:15:06miraklinuxstb: to summarize you increase the stack for each malloc calls, then at one moment there should be no use of all what was allocated so you restart the stack from scratch ?
23:15:25zeeroLinusN: nod
23:15:32markunmirak: look at apps/codecs/lib/codeclib.c
23:15:58linuxstb_mirak: Basically, yes.
23:16:09zeeroLinusN: also remote_update stuff right?
23:16:22LinusNzeero: yes
23:16:51linuxstb_If your code uses lots of free/realloc then you may have problems - you should try and remove the use of realloc.
23:17:26^BeN^bmp2rb -f 4 *.bmp
23:17:28peturLinusN: I would think #if (CONFIG_I2C == I2C_COLDFIRE) meant it's using the coldfire I2C, yet the code looks more like bit-banging.
23:17:30LinusNrealloc is evil in its purest form
23:17:34^BeN^what i need to write
23:17:41^BeN^in the source
23:17:42peturTime for that generic I2C bitbanging thing?
23:17:54LinusNpetur: yes, that is not a good #if condition
23:18:04LinusNpetur: would be nice
23:18:05^BeN^for its work
23:18:44LinusN^BeN^: you need to do it once per file
23:18:46linuxstb_^BeN^: Just do "for a in *.bmp ; do bmp2rb -f 4 $a ; done"
23:18:47LinusNyou can't do *.bmp
23:18:55^BeN^i know
23:19:06^BeN^its only example
23:19:38^BeN^what i need to wirte in the ".c" for its works?
23:19:48LinusN^BeN^: call lcd_bitmap()
23:19:59^BeN^i 10x
23:20:13^BeN^i use mono lcd bitmap
23:23:04^BeN^solitaire.c:898: warning: passing arg 1 of pointer to function makes pointer fro
23:23:04^BeN^m integer without a cast
23:23:04^BeN^solitaire.c:967: warning: passing arg 1 of pointer to function makes pointer fro
23:23:04DBUGEnqueued KICK ^BeN^
23:23:04^BeN^m integer without a cast
23:23:08^BeN^what wrong?
23:25:39^BeN^working now
23:31:00mirakwhat's memory allignement ?
23:31:58 Quit linuxstb_ ("CGI:IRC (Ping timeout)")
23:32:31 Quit qwisp11 ()
23:32:54CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:32:54*webguest02 struggles to recreate the sluggish dir browser
23:33:00 Join linuxstb_ [0] (
23:36:42linuxstb_mirak: It means that values are stored in memory locations aligned to the size of that value - i.e. 32-bit words are stored on 4-byte boundaries, 16-bit words are stored on 2-byte boundaries.
23:37:26linuxstb_Some CPUs will give errors when doing unaligned reads, others will just perform slower.
23:38:35mirakyou mean relative to the adress that needs to be pair for exemple ?
23:38:42mirakor a multiple of 4 or 2 ?
23:38:46mirakor bit wise ?
23:39:05miraklinuxstb: ?
23:39:37linuxstb_The address needs to be a multiple of 4 or 2.
23:39:48linuxstb_e.g. storing a 32-bit value at location 0x1000005 is unaligned
23:40:00linuxstb_But storing it at 0x1000004 is aligned
23:40:11 Join Midgey34 [0] (
23:40:22mirakok in xvid code there is a xvid_malloc, that have in plus an allignement parameter
23:40:30mirakmaybe that's different
23:40:50linuxstb_No, that's probably the same thing.
23:40:57mirakbecause I don't see how we can know the pointer is not aligned before allocation
23:42:17linuxstb_Just look at the malloc implementation in apps/codecs/lib/codeclib - that always returns an address 32-bit aligned.
23:42:46mirakwell I believe you, I just wondered what exactly was doing the malloc in xvid
23:42:59peturLinusN: I don't understand the port pin numbering in the wiki:the ColdFire doc says pin 56 on the 144QFP packages is PADD-VDD. Not really I/O :(
23:43:01mirakthey kind of wrap the original malloc
23:43:24LinusNpetur: it's not qfp
23:43:33LinusNit's bga
23:43:49peturdamn - need to find another datasheet
23:43:56LinusNor turn the page
23:44:17 Quit Lear ("Chatzilla [Firefox 1.5/undefined]")
23:44:21LinusNwhy do you need the pin number?
23:44:34saa[b_r]ider^Ben^: so does the solitare plug-in work now?
23:44:46peturto find out what bits to toggle?
23:45:03^BeN^with the big images?
23:45:32saa[b_r]idergreat, is there another place for me to get it? i can't ipen
23:45:52peturThe wiki only says GPIO1 and a pin number...
23:46:21saa[b_r]ider^Ben^ did you upload it to ?
23:46:37^BeN^not yet
23:46:52^BeN^i can send you
23:47:14saa[b_r]iderplease :)
23:47:53LinusNpetur: the wiki says which *port pin* it is
23:48:14LinusNGPIO is one port, GPIO1 is another
23:48:27saa[b_r]iderare you sending it through DCC?
23:48:35LinusNthey are two halves of a 64-bit I/O port
23:48:50peturoh - ok
23:48:55LinusNGPIO32 is bit 0 in GPIO1
23:49:13petursorry, haven't read the coldfire manual yest :(
23:49:13LinusNget it?
23:49:14saa[b_r]idersorry, please try again
23:49:20peturyep - go it now!
23:49:56^BeN^linusn: when i use this function rb->lcd_bitmap(carddeck[1], LCD_WIDTH2 - CARD_WIDTH+1, LCD_HEIGHT-CARD_HEIGHT+1, 19, 24);
23:50:04saa[b_r]ider^Ben^ I just fixed the settings, so it should work fine.... underline SHOULD :D
23:50:05^BeN^the color goen crazy
23:51:00^BeN^no spesific color
23:51:06linuxstb_Do you set a drawmode?
23:51:12^BeN^all the colors mixed
23:51:17^BeN^LOL no
23:51:30^BeN^i dont know that i need to set it
23:52:05saa[b_r]ider^Ben^ did you resend it?
23:52:16^BeN^look in private
23:52:23peturLinusN: it all makes sense now - now I understand the H1xx defines....
23:54:53^BeN^12linuxstb_12 -1 which draw mode i need to use?
23:55:27 Join tucoz [0] (
23:55:58*tucoz is listening to a review for blind users of rockbox on the h140 :)
23:57:10tucozIt is impressive what computers make possible for the blind. Never heard of a braille device before I started using rockbox
23:57:15LinusN^BeN^: yes?
23:57:23^BeN^which draw mode i need to use?
23:58:38linuxstb_DRMODE_SOLID is the default

Previous day | Next day