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 2005-02-09

00:00:07amiconnBleeding edge builds are incomplete anyway (no rocks for either Ondio, no gmini, no iriver)
00:00:18Bagdera zip file would fix that
00:00:51linuxstbDamn, found another unbounded malloc - the seek table.
00:00:52Bagdera make zip for every build
00:01:58Bagderhehe, and the player debug bleeding edge file is from april 25 2003 ;-)
00:02:33amiconnrofl
00:03:15lImbushehe
00:04:32HClhehe
00:05:35 Quit webguest69 ("CGI:IRC")
00:06:20rashermosaic looks groovy
00:07:19TrevmarI GOT A PROBLEM - the oldest daily build in the archive at rockbox.org that I can access is 9 Jan and the software is already broken by then. I tried to construct a URL for older files and failed. Where do I get the older daily builds from, please???
00:07:43Bagdernowhere
00:07:46Bagderthey're removed
00:08:04Bagderyou build older yourself by checking out older sources
00:08:28Bagdercvs co -D "date" rockbox
00:08:29Bagderiirc
00:10:23TrevmarOK, I didn't change over to the new CVS when you left Sourceforge, so I am way behind on source builds - lazy, I guess.. let me plug away at this because I may be wrong with Jan 9 working - I am looking at it in more detail. I shouldn't jump to conclusions :) Been out of software testing too long, I am afraid...
00:10:42Bagderno worries
00:11:16rasherwhen do the daily builds happen?
00:11:44amiconn6 a.m. CET iirc
00:11:53Trevmarconfirmed - it was broken before Jan 9th - sheez...
00:12:12Bagderok, I modified the bleeding edge builds
00:12:20Bagderand the links to them
00:12:30Bagdernow they're all rockbox.zip
00:12:44amiconnreboot, brb
00:12:47 Part amiconn
00:13:06Bagderso now there's an iRiver h1x0 link as well
00:14:07 Quit thegeek (Read error: 104 (Connection reset by peer))
00:14:18 Quit ripnetuk (Read error: 104 (Connection reset by peer))
00:15:27 Join thegeek [0] (na@ti521110a080-3006.bb.online.no)
00:16:23rasheroh, and it actually works?
00:16:33BagderI assume it does
00:16:38rasherrocking
00:17:44Bagderoops, the description was wrong, fixed now
00:18:57TrevmarOK, I have found that the 23 December OndioFM build is OK with 03 08 units but the Jan 9th build is not seeing the MMC card. It is going to be serious work for me to set up CVS so I can build some more daily builds between that period. Doesn't anybody have a backup or an archive?
00:20:08*HCl goes back to sleep..
00:20:15BagderTrevmar: during that period, there was only two changes in ata.c
00:20:16 Quit izzy_ (Read error: 60 (Operation timed out))
00:20:34rasherwow, cube looks great in full screen
00:20:48rasherbut the perspective thing does look a bit odd at some moments
00:20:48Bagderand three in ata_mmc.c
00:21:04[IDC]Dragonata.c is not involved
00:21:09Bagderok
00:21:28 Join amiconn [0] (~jens@pD95D135F.dip.t-dialin.net)
00:22:22Trevmarif anybody has any build ZIPs maybe they could email them to me? I can test them pretty quick... (and without risk, 'cos I can always go back to uart_bot LOL)
00:23:08rasherWhat in the world does "xsanke: " mean in the "bounce" plugin?
00:23:21Bagderits a parameter
00:23:43Bagderits a speed for a second sine table, whose data is added to the first
00:24:31Bagdernamed like that as a tribute to the weird person who taught me that trick, many eons ago
00:25:07rasherheh okay
00:26:59rasherit'd be fun (although possibly more draining to the batteries) to be able to use these as actual screen savers
00:27:54BagderI doubt very many would actually want that
00:28:11amiconnHmm, strange. Both recorder and iRiver sim scroll when compiled for X11, however, when compiling Win32 sims, only the recorder sim does scroll...
00:28:13rasherprobably
00:28:49Bagderamiconn: that is mighty strange
00:29:41linuxstbAt what stage does Rockbox read the ID3 tags from a file? Is it when playing starts, or earlier?
00:29:46 Quit midk ("Leaving")
00:29:57Bagderlinuxstb: when it loads the file to ram
00:30:00 Quit Sucka ("( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de )")
00:30:36linuxstbBagder: Does it need to read it that early?
00:31:01Bagderyes, in order not to waste battery it reads the data when the disk is already spinning
00:31:15Bagderand then it also has info about next song etc
00:31:46linuxstbIs the next song info displayed (sorry, I haven't used my Archos for a while)
00:31:50 Join midk [0] (~midk@c66-235-14-120.sea2.cablespeed.com)
00:32:11 Quit jyp ("poof!")
00:32:18Bagderyes, if you use such a WPS-tag
00:33:10 Join izzy [0] (laitinei@huippu.net)
00:33:22amiconn..but the screen update of these is broken, depending on the line type you insert them they may not get updated :(
00:33:33Bagderyes, it still needs improvement
00:33:45Bagderbut that's not limited to next-song
00:33:49linuxstbWith FLAC, the procedure is: init decoder, process file up to end of meta-data, process rest of file, reset decoder ready for next file
00:34:24Bagderamiconn: try a fixed string with "time remaining" appended, long enough to scroll
00:34:27linuxstbSo it looks like I need an efficient way to read the metadata from a FLAC file, without creating a full decoder instance (needing 150K RAM).
00:35:03Bagderthat would be the best, yes
00:41:53Bagdertime to hit bed
00:44:13[IDC]Dragonhere too
00:45:39 Part [IDC]Dragon
00:46:40rasherHere as well
00:47:09lImbuswell then. good night on all sides
00:48:42TrevmarGoonight. I guess I will do more testing if anybody sends me some OndioFM daily build Zips from 23 Dec to 9 Jan to test..
00:51:42 Join webguest43 [0] (~d99b54b6@labb.contactor.se)
00:51:52***Saving seen data "./dancer.seen"
00:51:59webguest43hi ! has anyone managed to run the iriver without battery ?
00:52:11webguest43i.e. direct hookup to a power supply ?
00:52:26 Quit webguest43 (Client Quit)
00:55:09 Join ashridah [0] (ashridah@220-253-119-189.VIC.netspace.net.au)
00:55:35 Quit Aison ("( www.nnscript.de :: NoNameScript 3.72 :: www.XLhost.de )")
01:00
01:01:11 Join xen` [0] (~xen@ADijon-151-1-40-153.w83-196.abo.wanadoo.fr)
01:14:03 Quit lImbus ("must...stay...awake...can't..anymore..hold")
01:16:27 Quit lolo-laptop ("Client exiting")
01:37:44 Part amiconn
01:43:19 Quit cYmen_ ("leaving")
01:55:07 Quit Trevmar (Read error: 110 (Connection timed out))
01:55:45 Quit jpburton5150 (Read error: 110 (Connection timed out))
01:56:00 Quit edx ()
02:00
02:05:11 Join edx [0] (edx@p54879303.dip.t-dialin.net)
02:12:57 Join edx_ [0] (edx@pD9EAB343.dip.t-dialin.net)
02:13:43 Quit gromit` (Read error: 60 (Operation timed out))
02:25:21 Quit Patr3ck ("User pushed the X - because it's Xtra, baby")
02:27:29 Quit edx (Read error: 110 (Connection timed out))
02:32:53 Quit ashridah ("out")
02:35:10 Nick edx_ is now known as edx (edx@pD9EAB343.dip.t-dialin.net)
02:48:43 Join gromit` [0] (~gromit`@ras75-5-82-234-244-69.fbx.proxad.net)
02:51:55***Saving seen data "./dancer.seen"
03:00
03:03:04 Quit midk ("Leaving")
03:08:17 Quit mecraw ()
03:22:23 Quit linuxstb (Read error: 110 (Connection timed out))
03:22:44 Join ashridah [0] (ashridah@220-253-118-208.VIC.netspace.net.au)
03:33:41 Join webguest99 [0] (~44e5cff2@labb.contactor.se)
03:46:29 Quit webguest99 ("CGI:IRC (EOF)")
03:57:13XShocKhmm. I realized that SPDIF output is active all the time in Rockbox. IS it supposed to be like that?
03:57:22XShocKiriver ihp120
04:00
04:05:40 Join QT_ [0] (as@area51.users.madwifi)
04:05:49ashridahwell. probably not, but it hasn't been handled yet i imagine
04:08:29XShocKok, just in case only me alone getting it. :)
04:08:44 Join midk [0] (~midk@c66-235-14-120.sea2.cablespeed.com)
04:08:49 Quit Stryke` (Read error: 104 (Connection reset by peer))
04:10:18ashridahnah, mine's on too
04:10:47midkso is mine!
04:12:30ashridahthere's a lot of differences between the archos platform and iriver's that need to be reconciled
04:13:29 Join lostlogic [0] (~lostlogic@node-4024215a.mdw.onnet.us.uu.net)
04:18:53 Quit QT (Read error: 110 (Connection timed out))
04:19:04XShocKbuy, going to sleep.
04:24:01 Quit XShocK (" HydraIRC -> http://www.hydrairc.com <- IRC for those that like to be different")
04:27:26 Join webguest79 [0] (~18e0b44d@labb.contactor.se)
04:28:12 Quit webguest79 (Client Quit)
04:51:58***Saving seen data "./dancer.seen"
04:55:52coobhttp://www.ipodlinux.org/forums/viewtopic.php?p=7200#7200 <−− cube.c ported to ipod linux :D
05:00
05:14:15 Quit xen` (Read error: 110 (Connection timed out))
05:26:46 Quit gromit` (zelazny.freenode.net irc.freenode.net)
05:26:46NSplitzelazny.freenode.net irc.freenode.net
05:26:46 Quit edx (zelazny.freenode.net irc.freenode.net)
05:26:46 Quit izzy (zelazny.freenode.net irc.freenode.net)
05:26:46 Quit mbr (zelazny.freenode.net irc.freenode.net)
05:27:16NHealzelazny.freenode.net irc.freenode.net
05:27:16NJoingromit` [0] (~gromit`@ras75-5-82-234-244-69.fbx.proxad.net)
05:27:16NJoinedx [0] (edx@pD9EAB343.dip.t-dialin.net)
05:27:16NJoinizzy [0] (laitinei@huippu.net)
05:27:16NJoinmbr [0] (~mb@stz-softwaretechnik.de)
05:39:49 Join jpburton5150 [0] (knoppix@cpe-24-94-54-216.stny.res.rr.com)
05:39:55jpburton5150hey HCL
05:40:21jpburton5150how did you end up fixing that "unrecognized architecture specification '5249'" problem?
05:40:30jpburton5150(or anyone else who got it)
06:00
06:20:57 Quit midk (Read error: 104 (Connection reset by peer))
06:22:17HCljpburton5150: cvs binutils
06:29:30 Quit jpburton5150 (Read error: 60 (Operation timed out))
06:52:01***Saving seen data "./dancer.seen"
07:00
07:12:20 Join midk [0] (~midk@c66-235-14-120.sea2.cablespeed.com)
07:30:27 Quit MooMaunder (Read error: 110 (Connection timed out))
07:32:32 Join webguest96 [0] (~18221819@labb.contactor.se)
07:34:12 Part webguest96
08:00
08:14:51 Join Stryke` [0] (~Derek@resnet-241-86.resnet.UMBC.EDU)
08:15:11 Join LinusN [0] (~linus@labb.contactor.se)
08:15:22 Quit Stryke` (Read error: 54 (Connection reset by peer))
08:42:06 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
08:50:39 Join Trevmar [0] (~trevor@ca-agoura-cuda2h-53.ventca.adelphia.net)
08:51:40TrevmarI am trying to get my source tree to the date of 29 Dec daily CVS build so that I can generate old versions to track a bug but Cygwin CVS says it doesn't accept the -D "20051229" option. What to do??
08:52:02***Saving seen data "./dancer.seen"
08:52:22LinusNshould work
08:53:02midk20051229?
08:53:07midkthat's about ten and a half months away.
08:53:11LinusNhaha
08:53:24midk:)
08:54:42TrevmarOk, I checked, I put -D "20041229" and got 'invalid option −−D'
08:54:53LinusNwhere did you put the -D?
08:54:54midkthat's happened to me before..
08:55:01midkyes, it has to be in a certain place.
08:55:04LinusNit should be after the "co"
08:55:11TrevmarOk, thanks, let me try
08:55:14LinusNor "update" if you did that
08:57:07Trevmarbrilliant, files are being updated. Now my memory is foggy cos last time I used Rock CVS it was stil on sourceforge :( Is there a script to automatically generate the daily build ZIP? or do I have to make and then pick files?
08:57:27LinusN"make zip"
08:57:31Trevmarthanks
08:57:51LinusNyou might need to check out "fonts" as well
08:57:53Trevmarfrom the home/Adminstrator/rockbox directory, I assume?
08:58:15LinusNfrom the dir where you built the firmware
08:58:39TrevmarOk, so I have to build formware first? doesn't make do that?
08:58:57LinusNif you type make without arguments, yes
08:59:21TrevmarOK, so first I will run make w/o args and then make zip OK?
08:59:36LinusNyou need to run the configure script as usual, then build the firmware, then "make zip"
09:00
09:00:31Trevmaroops, forgot configure..let me do that.. make without args builds the firmeare?
09:00:43LinusNyes
09:00:51LinusNmkdir ondio
09:00:54LinusNcd ondio
09:00:59Trevmarthanks, sorry for being thick... it has been a long time...
09:01:02LinusN../tools/configure
09:01:06LinusNmake
09:01:07LinusNmake zip
09:01:20Trevmar:)
09:01:26LinusNbut first you should check out the "fonts" module
09:01:38LinusNfrom cvs
09:02:35TrevmarI already have fonts on the boxes I am flashing, will it be OK to ignore updates to them in the last 2 months?
09:02:40LinusN"/home/Administrator"...hehe i wouldn't dare to work logged in as root on my linux box... :-)
09:02:56LinusNthe make zip script needs the fonts
09:03:05Trevmaryes, LOL, I am not a developer, and I won't be using the simulators -LOL
09:03:20TrevmarOK, will get fontd
09:04:41TrevmarOK, CVS gave me fints :)
09:04:49Trevmartypos sheez...
09:05:00Trevmargetting late..
09:05:04LinusN:-)
09:09:43Trevmaryou were correct, 'make' was OK but 'make zip' told me "cant open fonts at /home/Administrator/rockbox/tools/buildzip.pl" I downloaded the CVS, did I have to make the fonts seperately?
09:10:06TrevmarI did get the font pacakge from CVS
09:10:16LinusNi think i misled yoiu a little
09:10:30LinusNthe fonts package should be in the rockbox dir
09:10:38LinusNrockbox/fonts
09:10:44TrevmarOK, I founf that fonts are not in the rockbox directory. Let me get them in the right place
09:11:18Trevmarso I should issue that CVS command in the rockbox directory, I guess??
09:11:23LinusNyes
09:11:27Trevmar:)
09:12:10dwihnoLinusN: How does the current fonts look on the ihp?
09:12:25LinusNwonderful
09:12:32LinusN16 lines
09:12:40LinusNwith the default font
09:12:48LinusNand the larger ones look really good
09:12:56LinusNpure pleasure
09:13:10LinusNgo get an iriver
09:13:19dwihno:)
09:13:32*LinusN got one off ebay yesterday
09:13:33dwihnoI'm buying a new computer this month. The iriver has to wait.
09:16:31TrevmarHmm. After I downloaded and Rolo ut t he "info" screen says version 20050209-0008 yet it works, and I told CVS to checkout -D "2004 1229"
09:17:20LinusNthat info is the build date, not the checkout date
09:17:39LinusN"built from CVS on 20050209"
09:17:53TrevmarOk, thanks, so is there anyway I can doublecheck I am flashing the correct build date?
09:18:01LinusNnot really
09:18:28TrevmarOk, I will go ahead and try a few dates. I know that Jan 9 was dead. Let me try that one again...
09:18:36LinusNgood luck
09:18:40Trevmarthanks
09:18:51LinusNyour efforts are apprciated
09:19:04Trevmaryou rockbox guys make it soooo easy.. thanks a bunch
09:38:24 Join einhirn [0] (Miranda@bsod.rz.tu-clausthal.de)
09:40:15 Join kurzhaarrocker [0] (~Phil@p5487C51D.dip0.t-ipconnect.de)
09:44:48kurzhaarrockerAre there any iriver models that run on AA sized batteries (mignon) and are about to be supported by rockbox?
09:45:36sneakumsnot that i know of.
09:46:33dwihnokurzhaarrocker: the only iriver models using AA type batteries are the flash based...
09:46:58kurzhaarrocker.. which are not very interesting to me either :(
09:49:41dwihno:(
09:49:49dwihnoYou are concerned about the battery?
09:49:55dwihno(in the hard drive models)?
09:50:48 Join DMJC [0] (~James@220-245-162-47-sa-nt.tpgi.com.au)
09:50:57kurzhaarrockerI just don't want to carry around a bunch of special chargers for each toy I have. I won't buy portable electronic that doesn't run on AA sized batteries.
09:51:12DMJCsucks to be you
09:51:39kurzhaarrocker:)
09:52:03DMJCmyself, i think everything should use the USB voltage..
09:52:14DMJCyou can get car to usb adapters..
09:52:29sneakumsthere's a usb to iriver adapter
09:53:16sneakumsiriver-h120_354.htm">http://www.boxwave.com/products/minisync/minisync-retractable-cable-iriver-h120_354.htm
09:53:34sneakumsrasher has one, it appears to work as advertised
09:54:37kurzhaarrockerBut that would mean I had to buy a car to generate the energy for my portatronix -> to big for my pocket.
09:54:40rasherYup, got it yesterday. It's pretty sweet.
09:54:57sneakumsor plug it into a handy computer.
09:55:18DMJCyou apid for that?
09:55:21DMJCpaid
09:55:45DMJCI just got a $10.00 usb camera cable, and soldered a power end on the power wires
09:55:53DMJCrecharges and synchs
09:55:54 Quit Trevmar (Read error: 110 (Connection timed out))
09:56:14 Join Trevmar [0] (~trevor@ca-agoura-cuda2h-53.ventca.adelphia.net)
09:56:15rasherI'm not too handy with the soldering, plus I like the retractable
09:56:17rasherthing
09:56:51DMJCmy soldering is disgustingly poor heh
09:56:59TrevmarLinus, when I tried to make for a new date I got LD giving me the error "region FLASH is full". I used CVS -D and then "make clean" and "make"
09:57:02DMJChad a mate take a look at it when i'd finished
09:57:53DMJCis anyone working on porting tremor atm?
09:59:09Trevmarcan anybody help me with my compilation problem? I am trying to compile cersions of CVS from early january to track an Ondio S/W bug
09:59:46sneakumsDMJC: linuxstb or whoever has been doing work on the codecs, not sure if he got around to tremor yet
10:00
10:00:04DMJCk
10:07:26 Join amiconn [0] (~jens@pD95D135F.dip.t-dialin.net)
10:07:32Trevmarcan anybody help me with my compilation problem? I am trying to compile cersions of CVS from early january to track an Ondio S/W bug
10:07:34TrevmarLinus, when I tried to make for a new date I got LD giving me the error "region FLASH is full". I used CVS -D and then "make clean" and "make"
10:07:54Trevmaramiconn (linus is gone..)
10:08:25TrevmarI made the 20041229 OK, and it worked, but having trouble making newer dates
10:08:32amiconnTrevmar: That's no problem, it only indicates that your can't build rombox, because it grew too big
10:09:07Trevmarbut the one that couldn'y link was the CVS -D "20050109"
10:09:30Trevmarsurely that release could be flashed?
10:09:52TrevmarI couldn't link 20050110 either :(
10:10:14 Quit kurzhaarrocker (Remote closed the connection)
10:10:16amiconnyup; only that there is no rombox.ucl. rockbox.ucl will work just fine.
10:10:41TrevmarOh, I get it.. let me try again
10:13:59 Join Zagor [242] (~bjst@labb.contactor.se)
10:15:53Trevmarhmm. the 10 Jan build I did worked. Let me go back to the online repository and double-check, maybe I was too tire this afternoon and identified Jan 9 incorrectly...
10:16:04Trevmartired
10:19:31Trevmaramiconn, the Rolo version of 10 Jan works to recognize the MMC card, but when I flash the UCL it doesn't see them. OK, let me make some more dates...
10:26:42 Join Aison [0] (~hans@zux166-181.adsl.green.ch)
10:52:05***Saving seen data "./dancer.seen"
10:58:56 Join Patr3ck [0] (~patr3ck@pD9ECFF0B.dip.t-dialin.net)
11:00
11:04:15 Join ripnetuk [0] (~george@82-70-100-230.dsl.in-addr.zen.co.uk)
11:06:36 Join linuxstb [0] (~linuxstb@dsl-212-23-31-215.zen.co.uk)
11:07:40ripnetukare we sure that iRiver rockbox doesnt eat file systems? ive had 3 instances of corrupted fat after running Rockbox, on an iRiver thats never corrupted before. Each time, I have formatted the drive using Windows, then synced my ~36 gigs of music on, and unzipped the rockbox zip. This works. If I then allow Rockbox to boot, and then boot back into the original firmware, and unzip the rockbox.zip again, the fat driver panics saying its
11:08:15rashersounds like a bug
11:08:25ripnetukdelete beyond eof, and remounts the jukebox read only. The only other thing I did recently was upgrade my kernel. Could it be the infamous usb scsi kernel bug? can anyone remember the patch for the kernel to get irivers working reliably?
11:09:00LinusNripnetuk: the vfat driver?
11:09:01rasherHm, I haven't had any problems with that
11:09:24LinusNi haven't had any file system problems
11:10:13ripnetukit was something to do with not reporting back to the scsi layer.
11:10:21LinusNi know that
11:10:44ripnetukim gonna try again on my old kernel tongiht. Just checking no one else had problems with it
11:10:52LinusNi just meant that i haven't experienced any fat problems at all since i started developing the iriver stuff
11:10:53rasherripnetuk: which kernel version is this?
11:11:10ripnetuk 2.6.8.1-24mdksmp
11:11:33LinusNi have 2.6.9
11:11:39rasherI have 2.6.10
11:12:03ripnetukhmmm... i guess the next sensible thing to try is either installing from windows, or trying my old kernel again.
11:12:11rasherthink I skipped over 2.6.8.1
11:12:17ripnetuksorry to have bl;amed Rockbox :) im sure the ata code is well tested on the Archos :)
11:12:44LinusNripnetuk: yes, the rockbox ata driver is well-tested, but it may still contain bugs
11:13:04ripnetukbut im sure someone else would have noticed if it destroyed fs's
11:13:07LinusNand the iriver ATA driver is quite fresh
11:13:24ripnetuki will do some more testing tonight and report back. thanks
11:14:16LinusNgood luck
11:14:18 Join xen` [0] (~xen@ADijon-151-1-51-175.w83-196.abo.wanadoo.fr)
11:14:19ripnetukta
11:15:01rasherwhat're people working on right now?
11:15:17LinusNstuff
11:15:18LinusN:-)
11:15:25rasher:)
11:15:56LinusNi have some changes to the bootloader coming up, and then i'll work on the usb
11:16:13xen`means we'll have to reflash ? ;)
11:16:21 Quit Trevmar (Read error: 110 (Connection timed out))
11:16:25LinusNif you want the new functionality, yes
11:16:26xen`heh
11:17:34rasherwhat are the changes?
11:17:47rasheranything specific?
11:18:13LinusN1) it starts the original if rockbox.iriver isn't found
11:18:25LinusN2) it respects the hold switches
11:20:02xen`I checked libmodplug
11:20:10xen`it got some little part in float
11:20:17xen`gonna check if its possible to convert them
11:20:21xen`but also, its written in C++
11:20:37xen`but that should be possible to convert it more or less easily
11:20:43rasherThat's great news
11:21:36 Quit Christi-S (Read error: 60 (Operation timed out))
11:22:32xen`and with it, we would get s3m, mod, xm, it, far, 669, umx, etc...
11:22:33xen`even wav =)
11:22:55rashermodplug does wav?
11:23:06xen`yep
11:23:08rashersome formats using wav samples
11:23:10rasherI guess?
11:24:00xen`config.h.in load_dmf.cpp load_mod.cpp load_stm.cpp mmcmp.cpp sndmix.cpp
11:24:00xen`fastmix.cpp load_dsm.cpp load_mt2.cpp load_ult.cpp modplug.cpp tables.cpp
11:24:00xen`libmodplug load_far.cpp load_mtm.cpp load_umx.cpp modplug.h
11:24:00DBUGEnqueued KICK xen`
11:24:00xen`load_669.cpp load_it.cpp load_okt.cpp load_wav.cpp snd_dsp.cpp
11:24:00xen`load_amf.cpp load_j2b.cpp load_psm.cpp load_xm.cpp sndfile.cpp
11:24:01***Alert Mode level 1
11:24:01xen`load_ams.cpp load_mdl.cpp load_ptm.cpp Makefile.am snd_flt.cpp
11:24:03xen`load_dbm.cpp load_med.cpp load_s3m.cpp Makefile.in snd_fx.cpp
11:24:14xen`there are plenty of format =)
11:24:23linuxstbxen`: I'm working on some audio codecs for Rockbox, and the main problem so far is memory usage - Rockbox doesn't have malloc, and we want to try and avoid using it.
11:24:33xen`ho ok
11:24:42xen`what other limitations there is ?
11:25:09xen`and how many memory I'm restricted to ?
11:25:31linuxstbNo other big limitations. I suppose just the speed of the processor.
11:25:53linuxstbRe: memory - I think the rule is just as little as possible.
11:25:58xen`ok
11:26:24xen`I'll have to check rockbox others codec to learn how to adapt them
11:26:38LinusNthere are none
11:26:55linuxstbAlso, you will need to restrict yourself you a pre-defined maximum memory size. For FLAC, this means not supporting files that have been encoded with large block sizes (which I don't think happens in the real world anyway).
11:27:13LinusNhow large is "large"?
11:27:45linuxstbAll my FLAC files are encoded with a blocksize of 4608 bytes. But the standard allows block sizes up to 32K
11:28:40LinusNbut how does that affect the overall mem usage?
11:28:42xen`oki
11:29:21DMJChow's the tremor port going?
11:29:37xen`right now everything is ready to start working on them right ? I just wont get sound output cause of the sound part not done ?
11:29:44xen`or is sound part the codec themself only ?
11:29:47linuxstblibFLAC allocates 4 buffers of about 4xblocksize - So 4608*16 is about 73K, but 32K*16 is 524K - a big difference in "typical" and maximum usage
11:30:20linuxstbI'll just make "maximum supported blocksize" a #define, and set it initially to 4608
11:30:57 Join Christi-S [0] (~christi@213.78.116.104)
11:31:52xen`hmm
11:31:54xen`things is
11:31:59xen`mod/xm are tracker format
11:32:03linuxstbxen': the audio "framework" hasn't been designed yet. I'm first looking in detail at how the codecs work, and then I'll propose something on the Wiki for others to comment on.
11:32:13LinusNlinuxstb: i don't see a problem with 524K
11:32:21xen`samples has to be loaded on the fly
11:32:32xen`it gonna be painfull with no malloc =)
11:32:43xen`otherwise I can check if its possible to read them as they are played
11:32:45linuxstbLinusN: in which case, I'll just change the #define to 32K
11:32:49LinusNdo so
11:32:50 Join Trevmar [0] (LUMOS@206-72-67-118.dsl.dock.net)
11:33:02TrevmarLinusN or Amiconn, are you tehre??
11:33:12LinusNi am
11:33:47TrevmarLinus, I found that the Jan 02 build doesn't see the MMC card in the 03 08 mask ondios and that Jan 01 is OK
11:34:02***Alert Mode OFF
11:34:10linuxstbThe other "unbounded" memory usage is the FLAC seektable. One seekpoint takes 20 bytes of storage, and even a 1 hour FLAC file only has about 400 seek points, so I'll just limit that to about 1000 seek points.
11:34:17TrevmarThis coincides with when the LD comes up with "FLASh is full" error message on ROMbox
11:34:24xen`hmm, finally I think it will need some complete rewrite :/
11:34:30linuxstb*The only "unbounded"...
11:34:50Trevmarand the files changed are dir.c config-ondiofm.h and config-ondisp.h
11:35:12linuxstbDMJC: I haven't looked at Tremor yet.
11:35:48TrevmarLinus, shoudl I post that on the Forum?
11:36:07LinusNTrevmar: i'm afraid that it is a little more complicated than that
11:36:14amiconnTrevmar: Did you flash "rombox.ucl" or "rockbox.ucl" for your Jan 01 build?
11:36:14linuxstbLinusN: Apart from those 4 big buffers, libFLAC uses about another 80K for various other buffers. - so mem usage would be between 150K and 600K depending on how we set the maximum blocksize.
11:36:35LinusNlinuxstb: ok, we'll have to live with that, i guess
11:36:42Trevmarlet me check.. rockbox.ucl
11:37:03LinusNjan 01 was when the MULTIVOLUME was activated
11:37:13Trevmararrrgh
11:37:18linuxstbI suppose it depends on how many codecs we want to be running at the same time. If it's just two, then I guess 600K is OK.
11:37:37xen`loading whole file into memory would be possible you think ?
11:37:55LinusNxen`: that depends on the file size, of course
11:38:04xen`ok
11:38:24rashermost tracker files are less than a few mb
11:38:27LinusNwe will have 25-30Mbytes for buffering
11:38:35xen`because since samples are stored all over the file and several can be readen at once
11:38:41xen`ho fine
11:38:54xen`ok, if the whole file is in buffering, it should require no malloc at all =)
11:38:55linuxstbxen': The codec itself shouldn't read the file from disk - the basic concept is to have a "compressed data buffer" and a "PCM data buffer". The audio subsystem moves data from one buffer to the other with the help of the codecs.
11:39:12amiconnTrevmar: There was a bug with the first multivolume implementation, that was fixed on Jan-03. Could you try building with Jan-04 cvs?
11:39:12xen`oki
11:39:25Trevmaramiconn, I guess i wasted time on a binary search - if I had known about Jan 1... Sigh...
11:39:32xen`yes 30 mb is way enough for nearly all tracker files =)
11:39:43Trevmaramiconn- yes, I will try a jan 4 for you, stand by
11:40:17xen`finally it should be fine with modplug, I'll just wait the audio framework is out and i'll work on it
11:40:42Trevmardir.c and fat.h were changed between Jan 2 and Jan 4... compiling...
11:41:54linuxstbxen': we may need to do something special with tracker files (and SID files) - "normal" codecs are streams - i.e. you read a frame of compressed data from the input, and write a frame of PCM data to the output. It sounds like tracker files aren't like that.
11:42:17xen`yes, right, the whole file is needed from the start
11:42:34rasherluckily there's no size problem with either
11:42:45rashersid files are 100kb at most
11:42:49xen`yep =)
11:43:10DMJCis there a script to copy the completed iriver firmware build onto the iriver?
11:43:33linuxstbDMJC: use "make zip" and then unzip the rockbox.zip on the irver.
11:43:34LinusNyou can have half the hvsc collection in memory at once :-)
11:43:40DMJCcool
11:44:08rasherheh
11:44:35TrevmarAmiconn: it doesn't see the MMC on Jan 4th either
11:44:36rasherwell 100kb was overdoing it
11:44:44rasher63kb is the largest in hvsc
11:45:09DMJCcan't open dir fonts at /home/james/development/iriver/rockbox/tools/buildzip.pl line 76.
11:45:17linuxstbrasher: someone said yesterday that SID files were 65?? programs. So I'm guessing that limits them to 64K.
11:45:24LinusNDMJC: cvs co fonts
11:45:39DMJCk
11:46:06rasherlinuxstb: they are, and apparentlly it does indeed limit them to 64k
11:46:18rasherI was thinking maybe there were multi-disk sids or something
11:46:33rasherapparently not so
11:46:58linuxstbBut I'm assuming there's no limits on the sizes of various tracker files.
11:47:11rasherprobably not
11:47:24amiconnTrevmar: Could you try one more thing: Compile from current cvs, but disable multivolume support and hotswap support in config-ondiofm.h
11:47:26amiconn?
11:47:42rasherbut there'd probalby be no harm done in putting a sufficiently high artificial limit on them
11:48:21Trevmaramiconn- OK, I am doing that. It will take 5 mins or so
11:52:11DMJCanyone writing a tetris port?
11:52:18rasherThere is already one
11:52:23rasherrockblox
11:52:24DMJCcool
11:52:34rasheralthough it's horizontal
11:52:41rasherwhich is quite odd on the iriver screen
11:52:57DMJCyeah
11:53:14DMJCthat's something I noticed...
11:53:21DMJCthe play area for snake is hell tiny
11:53:29LinusNthe idea on the archos is to hold your archos rotated
11:54:07LinusNthe iriver screen is large enough to play it without rotating it
11:54:33LinusNlunch time
11:54:58DMJCyeah, need to start exploiting that screen
11:56:17rasheraw, snow in fullscreen
11:56:37 Quit ripnetuk (Read error: 104 (Connection reset by peer))
11:56:55amiconnrasher: :)
11:57:30sneakumsyou could fit a nice pac man on the iriver screen, i'll wager
11:57:50 Join iriver [0] (~50a04882@labb.contactor.se)
11:58:52iriverdo your plugins work on the H1x0? I get a adress error for all plugins - except helloworld.
11:59:04LinusNiriver: use the latest daily build
11:59:14iriverI do
11:59:21LinusNi doubt it
12:00
12:00:01irivercvs checked out and compiled this morning
12:00:02LinusNdid you unzip the entire zip to the iriver?
12:00:06iriveryes
12:00:50LinusNweird, i fixed that but 2 days ago
12:00:54LinusNbug
12:01:53iriverI don't know what is wrong with my iriver builds... the power off/shutdown function does not wok neither :(
12:02:35LinusNi changed that, you are now supposed to hols the stop button
12:02:41LinusNhold
12:03:22iriverI implemented my own shutdown if HOLD is ON... STOP hold does not work - if have to reset to poweroff Rockbox
12:04:19iriverwhen holding STOP the file browser starts
12:04:41TrevmarAmiconn: as you expected, that seems OK. I have my advanced audio tweaks now too :) Shal I just try disabling hotswap or multicolume one at a time? Any dependencies?
12:04:55LinusNiriver: then what is on screen when you hold stop?
12:05:35iriverIt does not react - it stays in filebrowser
12:06:02rasherwhy are recording options not under options?
12:06:05rasherin the menu
12:06:15rasheririver: do you have a song playing?
12:06:16LinusNyou said it started the browser, so where was it when you pushed stop in the first place?
12:06:29rasheririver: well, trying to play
12:08:03iriverwhen Rockbox starts, the main screen is clear. only shown battery, volume, playing status, shuffle status
12:08:40rasheririver: what is the playing status?
12:08:45iriverrasher: yes, i'm "playing" - at least the play icon is on, but stop does not stop it
12:08:52rasherthat's the problem
12:09:01LinusNiriver: i guess you enabled the "resume" option
12:09:15iriverbut what is the solution :)
12:09:28rasheririver: go into the settings, set resume to "No"
12:09:46rasherthen try and generate disk activity (I think.. I browsed some folders..)
12:10:14rasherthen reset
12:10:27rasherthat's exactly the behaviour I got
12:10:44iriverso I can't poweroff, if Resume in ON. Weird :)
12:10:58rasherwell you can't poweroff if the iriver's trying to play a song
12:11:10rashersetting resume to off is just making sure that you're not doing that
12:11:19iriverbut why can't I stop a song by pressing STOP
12:11:51LinusNiriver: because the mpeg thread doesn't respond to the SYS_POWEROFF event
12:12:10TrevmarAmiconn: now I am trying just disabling multivolume, and not hotswap
12:12:11LinusNand it doesn't respond to MPEG_PLAY either
12:12:55iriverTurning off RESUME solved the problem... Thanks :)
12:13:11rasherMaybe I should add this to the wiki somewhere
12:13:42LinusNi wonder why people enable resume when you can't play any music?
12:13:57TrevmarAmiconn: nope, there are unsatisfied dependencies with hotplug and no multivolume.
12:14:09rasherLinusN: Had to try all the options!
12:14:11Bagderwe should add an option "remove all my music" and people will select that too ;-)
12:14:14iriveryeah, it is not obvious that RESUME should be left alone until the MPEG_PLAY and SYS_POWEROFF is handled correctly
12:14:53iriverpeople play around to learn Rockbox
12:16:24TrevmarIriver: but I played around with my Jukebox recorder - I expect it solid now for the Ondio -LOL
12:17:53rasherit seems weird that "Playback" is under general settings, but "recording options" is under recording
12:17:57rasherany specific reason for this?
12:20:36Bagderthe menu structure is always subject for debate
12:23:41rasherI guess
12:25:42 Join methangas [0] (methangas@0x50a43276.virnxx10.adsl-dhcp.tele.dk)
12:32:55LinusNfunny, i can stop the playback, and i can shut down after playing too
12:33:16LinusNiriver: are you really sure that you use the latest cvs?
12:33:43linuxstbLinusN: Are you confident of being able to control the rest of the hardware on the iRiver, or are there still some unknowns?
12:33:47LinusNi bet you don't, since you have plugin problems as well
12:34:05HCloh, is the stop playback thing fixed?
12:34:10HClgmorning
12:34:11HClby the way
12:34:13LinusNi don't see any problems with the hardware
12:34:37linuxstbSo there is no reason to think the iRiver port won't be 100% successful?
12:35:03LinusNlinuxstb: no, i'm confident that we will succeed
12:35:41LinusNthe stop playback thing was fixed feb 6
12:35:46HClokay
12:36:00HClmy checkout is from just after the plugin fix
12:36:06HClso its prolly from before that
12:36:51LinusNplugins were fixed on feb 7
12:37:11linuxstbSo I can now recommend that people by iRivers because it will run Rockbox :-)
12:37:20LinusNsure :-)
12:37:32Bagderwhile there still are any around to buy ;-)
12:37:42linuxstbBagder: exactly.
12:37:56linuxstbIs the H3xx still in production?
12:38:09Bagderpeople it isn't
12:38:11LinusNrumours say the are out of production
12:38:13Bagder+say
12:38:40 Quit courtc (Read error: 60 (Operation timed out))
12:39:53rasherhrm
12:39:56rasherI can't stop
12:40:02rasherplayback, that is
12:40:07LinusNiriver: i suggest you try "make clean"
12:40:29rasherand I just unpzcked the latest bleeding edge build
12:40:35 Join walnut [0] (~garryh@frink.nuigalway.ie)
12:41:03LinusNrasher: yes you can. it's just the status bar that isn't updated
12:41:13rasherah
12:41:15rasheralright
12:41:29HClLinusN: thats kinda odd, cause i still get the unable to shutdown bug.. i think..
12:41:32rasherI see the "can't turn off while charging" is still there
12:41:42LinusNis the charger attached?
12:41:55CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
12:41:55*HCl goes to fetch his iriver to see
12:42:53HCland i need to update my bootloader.. it gives the wrong size for rockbox..
12:43:12Bagderit is in hex
12:43:19HCloh.
12:43:20HClokay
12:43:26*Bagder did the same mistake
12:43:29HCland nope, i can shut down now
12:43:30HCl :)
12:43:46HClhas someone with cvs access added strchr to the plugin api yet? :x
12:43:55Bagdernot yet
12:44:12HClk..
12:44:31Bagderyou can easily add that in your local version for now
12:44:48HCli guess
12:45:04 Quit walnut ("[BX] Reserve your copy of BitchX-1.0c19 for the Nintendo 64 today!")
12:45:22HClwhere do i fill in the api structure before its passed to plugins?
12:45:22Trevmar[amiconn] Are you still here? I have some more info
12:45:39BagderHCl: apps/plugin.c
12:45:42HClok
12:45:55 Join courtc [0] (~court@adsl-154-40-167.asm.bellsouth.net)
12:47:41BagderI don't get an ld when I intall the calmrisc binutils
12:47:48Zagorhas anyone used the "flyspray" bug tracker tool? i noticed the ipodlinux gang uses it and from my first glances it looks pretty nice.
12:47:55Zagorhttp://flyspray.rocks.cc/bts/
12:48:09 Join ripnetuk [0] (~george@82-70-100-230.dsl.in-addr.zen.co.uk)
12:49:02Zagori really want to get off the SF trackers
12:49:21Bagderhttp://wget-bugs.ferrara.linux.it/ <= roundup in use
12:50:12linuxstbHCl: are you going to be ablle to manage without the malloc family of functions?
12:50:22LinusNHCl: i just added strchr() in cvs for you
12:50:26ZagorBagder: looks good too
12:50:34Bagderthey look similar
12:51:06LinusN"flyspray", now that's cute...
12:51:32Zagor"roundup" is the same bad joke
12:51:36LinusNyeah
12:51:55Bagderflyspray seems the more advanced of these two
12:52:02dwihnoflyspray looks nice
12:52:04HClLinusN: thanks :)
12:52:07***Saving seen data "./dancer.seen"
12:52:26HCllinuxstb: probably, i'm just gonna write my own replacement that'll use the mp3 buffer.. a very crude one that doesn't free or anything.
12:52:28LinusN"roundup" - an obvious trademark violation :-)
12:52:32HCland just assume that 32mb will be enough.
12:52:41BagderLinusN: indeed
12:53:13HClcan rockbox play mp3 and run plugins at the same time?
12:53:14Bagderpossibly we could offer malloc for the plugins, in a separate plugin pool
12:53:21LinusNHCl: yes
12:53:23HClBagder: that would be nice..
12:53:30HClLinusN: that would seriously mess up with gnuboy then :X
12:53:36HClas it is now, anyways..
12:53:46LinusNgnuboy should just stop the music then
12:53:59LinusNother plugins do
12:53:59HClno um.. >.> badger suggested i'd use the mp3 buffer for memory
12:54:06LinusNmetronome for example
12:54:09HClsince i can't make everything static
12:54:11linuxstbMy understanding is that plugins have a fixed memory area to work in, so it would make sense to allow malloc within that memory area.
12:54:34Bagderlinuxstb: yes, that's my thinking too
12:54:36LinusNi can live with that
12:54:44HClsounds ok.
12:54:48LinusNBagder: add malloc() to the plugin lib
12:54:51Bagderat least when we extend the plugin area for irivers
12:55:08HClyea, gnuboy is gonna need... about... 6 mb to run?
12:55:13LinusN6!
12:55:17HClwell
12:55:18HClat max
12:55:19Bagderhah
12:55:19HClanyways
12:55:19linuxstbBagder: no reason why it couldn't work (on a smaller scale) for all targets.
12:55:19Zagoryikes
12:55:27HClthere are some gameboy roms that are 4mb
12:55:31LinusNah
12:55:40linuxstbMaybe "big_malloc" as well...
12:55:51LinusNthen the mp3 buffer is a better choice
12:55:52Zagorwell, since gnuboy will halt the music it can have all the ram it wants
12:55:56HCland there's something like 64k savegame ram and 256k work ram..
12:55:59Bagderthe plugin-malloc would not allow that huge memory anyway
12:56:03HCland then it needs ram for the code
12:56:21HClBagder: well, i somewhat need a spot to load the gameboy rom into ram...
12:56:30HCland since they can get up to 4mb.. :/
12:56:38Bagderthe mp3 buffer is for you my son ;-)
12:56:40LinusNBagder: how about a malloc() in the plugin lib and also be able to assign a pool to it?
12:56:51Bagderyes, I guess that's what we want
12:56:56HClgnuboy will use reallocs to limit memory usage as much as possible
12:56:58LinusNso you can grab the mp3 buffer and give it to malloc()
12:57:02ZagorHCl: why not use static sizes for those things?
12:57:11linuxstbLinusN: That would be useful for the codecs as well - same functions, different pools.
12:57:14BagderLinusN: in fact, my code already supports that kind of operation
12:57:20LinusNZagor: for the same reasons as we don't change all mallocs in the codecs?
12:57:28 Join LUMOS1 [0] (~trevor@ca-agoura-cuda2h-53.ventca.adelphia.net)
12:57:36HClZagor: well, it'd almost always use more ram than it'd actually need..
12:57:43LinusNlinuxstb: it is a lot more complicated for the codecs
12:57:46HCland it like
12:57:55HClreads the amount of memory its gonna need, from the rom header
12:58:03HClso at the moment i just have to assume maximum
12:58:36ZagorHCl: well the point with static allocation is that you always know you can handle the worst case. with dynamic, that requires testing which is sometimes non-trivial.
12:59:04linuxstbLinusN: Yes, at the moment, my plan is to try and remove mallocs (in a clean way) from the codecs. e.g. for FLAC, I would like to submit a patch to the libFLAC people to implement a "FLAC__USE_STATIC_MEMORY" or similar compile directive.
12:59:20LinusNlinuxstb: way to go
12:59:23Zagorand there's no point in using 6 instead of 7 megs of ram, since we'll be having 24+ allocated for it anyway
12:59:47HClokay.
12:59:52Zagorlinuxstb: sounds excellent!
12:59:58HClwell, i don't really see a problem with making gnuboy use static buffers...
13:00
13:00:16HClit'll just use a huge memory space though
13:00:31LinusNit will anyway
13:00:39HClmhm
13:01:00LinusNsince you have to reserve a malloc memory pool
13:01:31LinusNthat can accomodate such huge malloc'c
13:01:32Zagori'd say the only good reason to use malloc is if you want to minimize the changes done to the original code
13:01:58LinusNamen
13:02:04HClmhm.
13:02:19linuxstbZagor. In theory :-). In practice, there are lots of mallocs which aren't executed on my test files. So there would be a big difference between a patch that just works for Rockbox, and a general-pupose patch that removes every single malloc from libFLAC
13:02:20HClwell, that doesn't matter much for gnuboy
13:02:24HClit hasn't been updated in 4 years
13:02:43 Join webguest43 [0] (~8f416314@labb.contactor.se)
13:02:58HCloh, um.
13:03:00HCli get this too >.>
13:03:01HCl/home/hcl/rockbox/cross/lib/gcc/m68k-elf/3.4.3/../../../../m68k-elf/bin/ld: region PLUGIN_RAM is full (/home/hcl/rockbox/rockbox-devel/iriver/rockboy.elf section .text)
13:03:04HCl/home/hcl/rockbox/cross/lib/gcc/m68k-elf/3.4.3/../../../../m68k-elf/bin/ld: region PLUGIN_RAM is full (/home/hcl/rockbox/rockbox-devel/iriver/rockboy.elf section .bss)
13:03:16HClthats not good, is it >.>
13:03:21ZagorHCl: yeah, the plugin ram area is very small for the sh units. you need to expand it.
13:03:23LinusNyes, it is only 32k at the moment
13:03:25HClmhm
13:03:29HClyea, 32k won't fit xD
13:03:31webguest43Hi ! Does anyone know if it is possible to power the iriver without the battery ?
13:03:42LinusNi'm sure it is possible
13:03:47LinusNi haven't tried though
13:03:49HClwebguest43: just yesterday or so i heard someone talk about having a charger usb cable for the iriver
13:03:52HCland said it was working
13:04:01 Join webguest86 [0] (~50fca7f2@labb.contactor.se)
13:04:07HClah.
13:04:10HCl13:07 * rasher quietly notes that the Boxwave usb+charger cable for iRiver
13:04:10HCl appears to work
13:04:22rasherYeah, but that's still with battery
13:04:36HClohh.
13:04:37HClokay
13:04:53ashridahrasher: is that usb 1.1 or 2.0/
13:04:54ashridah?
13:05:04rasherusb2
13:05:11rasherit's pretty new I think
13:05:20rasheruntil recently they only had a usb1.1 version
13:05:25 Join cYmen [0] (~cymen@nat-ph3-wh.rz.uni-karlsruhe.de)
13:05:27*HCl continues writing wrapper functions for the plugin api
13:05:29rasherwww.boxwave.com/products/minisync/minisync-retractable-cable-iriver-h120_354.htm
13:05:52 Quit Trevmar (Read error: 60 (Operation timed out))
13:06:52LinusNHCl: to expand the plugin buffer area, edit firmware/app.lds and apps/plugins/plugin.lds
13:07:09HClokay, thanks
13:07:28webguest43the charger usb cable is simple, jou just splice the five volt from the usb cable into a power plug
13:07:43HCli'm not there just yet, still some basic libc functions to fill in/replace, and i haven't even started on the iriver <-> gnuboy interface
13:07:59 Join lImbus [0] (~manuel@kernel.cycos.net)
13:08:00webguest43but powering the player via an external power supply would be great, so you could attach it to a stereo and use it all day without draining the battery
13:08:03lImbushoi
13:08:44webguest43afaik the built in socket is only for the charger, and will not power the unit ? (dont own one yet)
13:08:49HCldo we have strcat in rockbox?
13:09:59LinusNyes
13:10:28LinusNyou need it?
13:10:58linuxstbIs anyone interested in WMA playback? There's a nice GPL'd decoder in ffmpeg, but it needs optimising and changing to fixed-point arithmetic. Could be fun for the right sort of person.
13:11:58HClLinusN: yup
13:12:08HClLinusN: i am, sortof.
13:12:14HClthe more rockbox can support, the better
13:12:22HClnot that i own any wma's, if any
13:12:37linuxstbIt could be important to users coming from the iRiver firmware.
13:12:39HCllinuxstb: *
13:12:43HClmhm
13:14:32 Join Patr3ck_ [0] (~patr3ck@pD9ECF876.dip.t-dialin.net)
13:15:08HClLinusN: tolower? do we have that?
13:15:29HClsprintf?
13:15:34 Quit Bagder (Read error: 110 (Connection timed out))
13:15:45 Join Bagder [0] (~daniel@1-1-5-26a.hud.sth.bostream.se)
13:16:13linuxstbSpeaking of "tolower", are there any plans to move Rockbox to use UTF-8 internally?
13:16:38Bagderyeeeeeellow build
13:16:46Bagderand the gmini one died hard
13:21:52 Join [IDC]Dragon [0] (~d90a3255@labb.contactor.se)
13:23:42LinusNtolower exists
13:23:49LinusNsnprintf() exists
13:25:43HClokay
13:26:05LinusNdo you get compiler errors on tolower?
13:26:16HClyea, its not in the plugin api
13:26:17HClso
13:26:29LinusN#include "ctype.h"
13:26:30Zagorlinuxstb: utf-8 would probably be a good idea (not the least to invite all asian users) but it's a rather big change
13:26:38HClok
13:27:51LinusNstrcat() added
13:29:02 Quit Patr3ck (Read error: 110 (Connection timed out))
13:29:13webguest43LinusN: its not possible to power the iriver from the charger and play music without draining the battery right ?
13:30:41linuxstbZagor: The meta-data in FLAC (and I'm guessing OGG) files is always UTF-8. Should I map this to iso-8859-1 ?
13:31:07Zagoryes, i think that's the best solution for now
13:31:22linuxstbThat's fine - I already have a function to do that.
13:35:11HClLinusN: thanks
13:35:46HClLinusN: is tolower in the api yet?
13:35:51HClcan't find it..
13:36:55HClhrm.
13:37:25HClokay. i have a problem with writing a wrapper function for snprintf :X how do i relay the ... arguments ?
13:37:50[IDC]Dragonphew, an iriver 140 is expensive
13:38:22Bagder[IDC]Dragon: I think the project could sponsor you
13:38:51[IDC]DragonI can't promise to seriously work on it
13:38:55BagderHCl: check the export/debug.h header file for an example
13:39:12Bagdernobody promises anything here ;-)
13:39:17ZagorHCl: if you are using a macro, the magic define is __VA_ARGS__
13:39:44 Quit dwihno (Remote closed the connection)
13:39:47[IDC]DragonI could morally only accept by taking it for AJB work
13:39:47LinusNHCl: tolower() is a macro, defined in ctype.h
13:39:48HCli'm not using a macro...
13:39:53HClLinusN: ah...
13:40:06ZagorHCl: then do that :-)
13:40:14[IDC]Dragonwhich I bought on my own, of course
13:40:16 Join dwihno [0] (~dw@81.8.224.89)
13:40:18HCli guess.. gah :x
13:40:20Bagder[IDC]Dragon: that would work for me ;-)
13:40:29HCli'll go rewrite my wrapper functions into macros for gnuboy..
13:40:41LinusNwebguest43: i think you can play music with the charger attached without draining the batteries
13:40:53HClit'll make them faster anyways
13:41:06HClyea, you can.
13:41:10*[IDC]Dragon is tempted
13:41:35lImbus:-)
13:42:10webguest43Great ! I did not know that !!! Could someone confirm please ?
13:42:35*lImbus nods his head
13:42:36LinusNdwihno: i got your recorder
13:43:13[IDC]DragonBagder: tell me the datails
13:43:17HClwebguest43: yes, you can
13:43:23ripnetukDoes anyone have a reference to the kernel problem that causes corruption of iRiver? from what I can remember the SCSI layer reports 0 bytes read, and should signal a retry, but signals a failure instead which causes read corruption. There is a patch, but I cannot find it right now. I want to see if my kernel is fixed or not
13:43:26webguest86webguest43: Yes, you can.
13:45:42 Quit lostlogic ("Going to the moon")
13:49:27 Quit iriver ("CGI:IRC")
13:50:13 Join jyp [0] (~jp@64.198-200-80.adsl.skynet.be)
13:50:27Bagderhi jyp
13:50:37Bagdermy calmrisc binutils are a mess ;-)
13:51:16jypaha
13:51:22jypin what way ?
13:51:40Bagder1 - I didn't have any 'ld' so I updated from cvs and installed
13:51:52Bagder2 - now the build breaks terribly
13:52:05jypokok
13:52:23jypI must have overlooked something
13:52:23Bagdersee the cvs build output
13:53:07jyp2 thouthands warnings ;)
13:53:19 Join quelsaruk [0] (~kvirc@85.48.33.190)
13:54:40quelsarukhi
13:55:10lImbushi
13:55:59quelsaruki received your mail
13:56:08quelsarukbut i saw amiconn commited it already
13:56:09quelsaruk:)
13:56:45Bagderand now the builds are even more yellow
13:57:13Bagderjyp: let me know if there's anything I can do or check or something
13:57:35Bagderthose regex outputs look like a debug output
13:59:02LinusN2117 warnings, is that a record?
13:59:09 Quit xen` (Read error: 60 (Operation timed out))
13:59:11HCllol.
13:59:26BagderI don't think so
13:59:40LinusNdamn
13:59:57DMJClol
14:00
14:00:02DMJCchicago font looks funny
14:00:08DMJCas if I owned an ipod
14:01:09lImbusthose 2117 are errors, isn't it ?
14:01:39 Join preglow [0] (thomj@s183a.studby.ntnu.no)
14:02:11Bagdertrue
14:02:49lImbusquelsaruk: and jyp even corrected a few typos i forgot
14:03:06quelsarukhehe
14:03:08quelsaruknice
14:04:16jypYeah those are errors
14:04:34lImbus_that_ might by a record :-)
14:04:46jypA messed up gas got in
14:04:51jypno wonder ;)
14:04:58lImbusgas ?
14:05:05jypGNU assembler
14:05:14lImbusah
14:05:54jypAlright, I got the "mess" fixed (locally)
14:06:05*[IDC]Dragon reboots
14:06:05jypnot I gotta check why ld doesn't install
14:06:09 Quit [IDC]Dragon ("CGI:IRC")
14:07:58jypWell, it installs fine here ...
14:13:29 Join lolo-laptop [0] (~lostlogic@68.251.84.226)
14:19:49 Quit webguest43 ("CGI:IRC (Ping timeout)")
14:20:06jypI'm now able to reproduce the problem
14:20:39 Quit webguest86 ("CGI:IRC")
14:21:28jypomfg...
14:21:51jypIt looks like that's due to a wrong automake version
14:23:02 Join webguest03 [0] (~507e1b43@labb.contactor.se)
14:23:11webguest03hello
14:23:26*jyp tips his hat.
14:23:32HClhmm..
14:23:49HCli'm having a slight problem with including plugin.h from a different directory..
14:24:02HCl#ifdef HAVE_LCD_BITMAP
14:24:02HCl#include "widgets.h"
14:24:02HCl#endif
14:24:07webguest03dudes, have you anything news about open source and the xclef hd-500?
14:24:10HClis that a bug and should it be recorder/widgets.h?
14:24:11HClor what?
14:24:17Bagderwebguest03: no
14:25:16webguest03fuck it... openjukebox.free.fr are 'developping' a new firmware, but they have restricted the source area... silly Frenches...
14:25:56webguest03I think your project is great by the way, but now I have to go to the Spongebob Squarepants movie
14:26:37 Join Etienne [0] (~51fb4e38@labb.contactor.se)
14:27:04 Quit webguest03 (Client Quit)
14:27:48EtienneIs there somebody here?
14:27:59quelsarukyes
14:28:01HClplenty.
14:28:03lolo-laptopnope, the 44 people in the list are not here.
14:28:04lolo-laptop:)
14:28:23HClBagder: do you know why that include fails..?
14:28:23Etienne: ) You guys are active ! COOL
14:29:21jypXD
14:29:46EtienneI'm french and I speak english too I am proposing my help for translation if you need any (I also owned both h120 and now h320)
14:29:48BagderHCl: the plugins can't include that file like that I guess
14:30:36HClBagder: i'm trying to include plugin.h in the gnuboy sources..
14:30:46quelsarukEtienne: that's cool, actually, francais.lang is up to date, but you can speak with lImbus and help him I think bobthc also helps with that)
14:31:03HClBagder: is HAVE_LCD_BITMAP defined for iriver?
14:31:07Bagderyes
14:31:09HClk..
14:31:22HCli'll just make a softlink for now..
14:32:25HClrockmacros.h:3: error: multiple storage classes in declaration of `rb'
14:32:30HClany idea what thats about?
14:32:38HCloh, nm :)
14:33:35Etienne@quelsaruk −− ok no problem!
14:34:08quelsaruktalk with them, i think help is always wanted :)
14:34:31quelsarukthe more people working in a translation, the best result you get :)
14:34:37EtienneIyes!
14:34:38 Join [IDC]Dragon [0] (~d90a3255@labb.contactor.se)
14:34:57jypAs opposed to a software project ;)
14:35:25EtienneAnd how could i help you to advance on the iriver port?
14:35:40quelsarukwhy not? :)
14:35:48jypBagder: the problems with binutils should be fixed
14:35:56Bagderok, updating...
14:36:12 Quit methangas (" HydraIRC -> http://www.hydrairc.com <- The dawn of a new IRC era")
14:36:25lolo-laptopactually I was going to ask that too −− I don't have the time/skills to really read the code and dig in for the iRiver port, would donating money in some way help advance it, or something else?
14:36:43Bagdermoney will help us buy more hw
14:38:07lolo-laptopis that pretty much what most of the donation money is used for?
14:38:14Bagderyes
14:38:25lolo-laptopcool, that's a worthy cause :)
14:39:23Bagdersponsoring more developers, buying 3x0s to dissect etc
14:39:29Lynx_how much donation money is there anyway?
14:39:43BagderI don't know
14:39:52Bagderzagor is the money man
14:40:45lolo-laptopconsider yourselves donated to :)
14:42:35 Quit HCl ("Lost terminal")
14:42:38Lynx_does paypal work for europe?
14:42:39Bagderjyp: still no ld installed though
14:42:50 Join HCl [0] (hcl@titania.student.utwente.nl)
14:42:53jypmmmm
14:42:55HClhm.
14:42:57BagderLynx_: yes it does
14:43:08*HCl slaps irssi
14:43:13jypmaybe, delete the ld dir & retry ?
14:43:25EtienneI can have h320 for cheaper if you want?
14:43:36EtienneWith suntec...
14:43:41Lynx_Bagder: ah, and with credit card as i see. hehe, the 'amount' box does not show a currency ;)
14:43:50Bagderhaha
14:44:12BagderI never paid to paypal, I only extract from ;-)
14:44:25Bagderand between paypal accounts of course
14:44:26Lynx_so i donate "10", hmm
14:44:40HClum......
14:44:53HCl/home/hcl/rockbox/rockbox-devel/apps/plugins/gnuboy/inflate.c:132: undefined reference to `memcpy'
14:45:04HClline 132 =
14:45:05HCl const int clen_weird_tangle[CLEN_TSIZE]
14:45:05HCl = { 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15 };
14:45:07HClo.o;
14:45:17HClthat calls memcpy?
14:45:39Bagderunder the hood it apparently does
14:45:45HClokay..
14:45:54HClguess i'll need wrappers for memcpy and set after all
14:46:15EtienneGuys another little question,you have probably heard of the drm problem when changing firmware with the h320 anybody know how to solve it?
14:47:00dwihnoLinusN: Was the packaging OK? :)
14:47:18dwihnoLinusN: you should see the way my christmas gifts look! ;D
14:47:27lolo-laptopEtienne: there is a thread on misticriver.net about it −− I don't know if they solved it or just talked about it though.
14:47:42quelsarukamiconn: are you here?
14:47:47lolo-laptopLynx_: it's USD afaik
14:48:07EtienneThey only talked about it...
14:48:21lolo-laptopEtienne: tragic.
14:48:26Lynx_lolo-laptop: yeah, i would have guessed that. kinda hard to find out what this will cost, with exchange rates and so on
14:48:37EtienneSome guy said he could remove the message that came up but not put the function back....
14:49:02lolo-laptopLynx_: yeah, isn't it like USD1.40=1EURO?
14:49:17Etienneno less about i.28
14:49:24Etiennego to xe.com to compare
14:49:48LinusNdwihno: the package is tightly sealed, according to sara :-)
14:49:49Lynx_lolo-laptop: yes, but i meant the stuff you have to pay paypal and the credit card company.
14:49:54Bagderhttp://daniel.haxx.se/currency/
14:50:01jypBagder, did 'ld' install yet ?
14:50:06Bagderno
14:50:15BagderI'm checking details now
14:50:25lolo-laptopahh
14:50:43Bagderoops
14:50:48Bagderjyp: yes it does, now it works
14:50:57jypah, good
14:51:31Bagderbut I still get regex mania
14:51:37jypDo you ?
14:52:06jypThere's something I don't understand
14:52:09EtienneOk i need to go but keep up the good work , you guys are great and I hope you suceed (and do the h320)!!
14:52:10***Saving seen data "./dancer.seen"
14:52:15dwihnoLinusN: If you need someone to package stuff - I'm your man! :)
14:52:19Bagderperhaps I should remove more files and rebuild
14:52:23BagderI'll do that
14:52:52jypthe offending files are in opcodes
14:53:05jypbut under CVS
14:53:12 Quit Etienne ("CGI:IRC")
14:54:12BagderI removed all files in my build tree and restart
14:54:13jypThe thing is, there are generated files in cvs; maybe this confuses 'make' or another tool
14:56:12Bagdermight be the reason, yes
14:58:00Bagderopenjukebox.free.fr is fun ;-)
14:58:01 Join xen` [0] (~xen@ADijon-151-1-57-9.w83-196.abo.wanadoo.fr)
14:59:20Bagderjyp: nope, didn't take
14:59:29lImbusBagder: yea, it's french :-)
14:59:38jypargh
14:59:58BagderlImbus: well, I thought more about the fact that there's no source
15:00
15:00:25*HCl grins.
15:00:38HClis there even a build of openjukebox?
15:00:42HClall i can find is hacked firmware
15:00:42Bagderno
15:00:46Bagderthere's nothing
15:00:48HClxD
15:00:58HClah well
15:00:59HClwhats to expect
15:01:02HClthey're french >.o
15:01:14HCl :P
15:01:15lImbusI just saw the "Currently active users counter" be bumped from 3 to 73 :D
15:01:17 Join webguest70 [0] (~d4963cc9@labb.contactor.se)
15:02:26lImbusXClef seem to be happy, except for some icons, rofl
15:05:03jypBagder, sorry that was my fault ...
15:05:38BagderI'm only happy if we proceed, whoever's fault it is! ;-)
15:05:45jypSourceforge cvs is so slow that I didn't let the commit finish ... closed the window too soon
15:05:52Bagderhaha
15:06:08jypGood that makes you laugh ;)
15:07:59ripnetukback
15:08:05ripnetukquit
15:08:06 Quit ripnetuk ("Leaving")
15:08:10 Join ripnetuk [0] (~george@82-70-100-230.dsl.in-addr.zen.co.uk)
15:08:19jypalright, it should be in this time
15:08:32jyp(opcodes/calmrisc16-asm.c)
15:08:35 Join R3nTiL [0] (~zorroz@217.30.249.98)
15:09:10 Quit ripnetuk (Client Quit)
15:09:11*Bagder does the rare cvs update dance while waiting for it to complete
15:10:45Zagorhow about moving the calmrisc cvs to rockbox.org?
15:11:48 Quit webguest70 ("CGI:IRC (EOF)")
15:13:21 Join bobTHC [0] (~foo@l07v-3-160.d1.club-internet.fr)
15:13:36bobTHChi all
15:14:09jypZagor, thanks for the offer; but changing the cvs host is a pain by itself
15:15:02jypso I guess we'll wait some more... :P
15:17:23jypOpenjukebox head dev. says "We don't publish our results coz we don't want problems with MCL (Jukebox' vendor)" −− May 2004
15:18:22Bagderyes, they've said that for a very long time
15:18:53Zagorthey only get in trouble if they distribute the vendors code, in which case it's not open any way
15:19:09sneakumswhat's French for "idiot"?
15:19:21Bagdercalling it open source is clearly a joke
15:19:37bobTHCtotally
15:19:37jypfrench(idiot) = idiot
15:19:43Shulberrysneakums: idiot, or imbécile perhaps
15:19:53bobTHCdont b silly
15:20:01Zagorsneakums: please behave nicely
15:20:09sneakumspardon?
15:20:12bobTHCevery country have this idiot
15:20:20Bagderindeed
15:20:40jypWonder what wiktionnary says about it
15:20:42*Bagder installs calmrisc binutils again
15:20:45mrmagsours is called 'gw bush'
15:20:47HCljyp: did he literally use "coz" ?
15:20:57Zagorsneakums: don't post flames on their forums or anything
15:21:02sneakumsi wasn't going to
15:21:12Zagorok, good. just making sure :-)
15:21:20BagderI asked them for the source once
15:21:27jypHCl, It's a free translation by me
15:21:36HClok. gnuboy compiles. all thats left is the iriver lcd/button <-> gnuboy interface
15:21:39HClok :p
15:21:43Bagderjyp: hey, it works! ;-)
15:21:50*jyp bounces
15:21:51*HCl stretches
15:22:02Bagderapart from the link
15:22:06Bagder"target elf32-sh not found"
15:22:07jyponto fix app.lds i suppose
15:22:10Bagderno surprise there
15:22:11bobTHClook@this and it's not frenchies ;) : http://uk.news.yahoo.com/050208/80/fbzpn.html
15:22:32bobTHCtotaly offtopic but so funny, sorry :)
15:23:27jypI guess *that* is idiocy
15:23:58jypSo, who's to win the tournament ? England ? ;)
15:24:01bobTHCtotally
15:24:14quelsarukthatis weird
15:24:17bobTHCthe best
15:24:27bobTHCwin
15:24:46jypWhile we're on the OT thread ... http://en.wiktionary.org/wiki/Idiot
15:26:05Bagderthe xclef has a hw-based codec
15:26:43Bagderor perhaps I misunderstand
15:26:46Bagdernever mind
15:26:58HClum...
15:27:09HClat least he was good for his word o.o;
15:28:39Bagdergmini down to "only" 386 warnings now ;-)
15:29:14HCl :P
15:30:05HCldoes rockbox have usleep? or how much is a kernel tick in sleep() ?
15:30:42BagderHZ is the number of ticks per second
15:31:06Zagoralways use fractions of HZ when calling sleep()
15:31:33HClhrm.
15:31:55HClso milleseconds would be..... HZ/1000*ms ?
15:31:59HClmilli*
15:32:16Bagderyes
15:32:27Bagderand hope that HZ is larger than 1000 ;-)
15:32:36HClhmhm :x
15:32:49Bagder>= actually
15:32:56HClyea.
15:33:02bobTHCall battlefront progress, gmini , iriver..... it's so great
15:33:33Bagder#define HZ 100
15:33:44Bagder(from export/kernel.h)
15:33:54ZagorHCl: you simply can't sleep for such short durations. use yield() if you just want to give up some time.
15:34:27Bagderpossibly we can that when we increase the clock speed
15:34:35HClerm.
15:35:20HCli need the fairly exact delay.. for sound..
15:35:41Zagorwhat sound?
15:35:49HClgameboy.
15:36:00Zagoryeah, but why do you need millisecond sleep for sound?
15:36:15HClbecause it needs to slightly pause emulation till the sound has been played, i think.
15:36:30Bagderyou play a millisecond sound?
15:36:32Zagorsurely the sound is longer than 1ms?
15:36:33HClwe'll just see how it goes.
15:36:39HClZagor: yea.
15:37:02HCli can alter it.. i guess.
15:37:25jypIn talk.h, how about changing ..
15:37:30jyp#define UNIT_SHIFT (32-4) /* this many bits left from UNIT_xx enum */
15:37:33HCl rb->sleep(HZ*us/1000);
15:37:36jypto (16-4) ?
15:37:42HClnow it'll just go to 0 if us is lower than 1000
15:38:04Bagderjyp: [IDC]Dragon and amiconn are the main talk guys
15:38:24jypok,thanks.
15:38:45jypcoz that would fix like a hundred warnings ;)
15:39:38Zagorjyp: (sizeof(int)*8 - 4) ?
15:40:27jypmaybe; the thing is I don't know if 12 bits would be enough for whatever should fit in there...
15:41:01Zagorok
15:42:11jypamiconn? [IDC]Dragon ?
15:43:19HCli got a question about get_time...
15:43:37HClsay you do b=get_time();
15:43:41HClthen do a=get_time();
15:43:50HClwill b be overwritten with a?
15:44:30HCl(does the result have a static memspace?)
15:45:16Bagderyes, it points to static data
15:45:18HClok.
15:46:40 Nick quelsaruk is now known as Quel|away (~kvirc@85.48.33.190)
15:49:13HClwhats tm_isdst in the time structure?
15:49:26LinusNdaylight savings time
15:49:45Bagderas if Rockbox knew ;-)
15:49:50LinusN:-)
15:50:08HCllol.
15:51:45linuxstbiRiver rockbox doesn't even know what day it is...
15:51:56HClheh... i need to return microseconds.. but the only time resolution rockbox provides is seconds at most...
15:52:04[IDC]Dragonjyp: I'm here
15:52:23jypHave you seen my question ?
15:52:34*[IDC]Dragon scrolls back...
15:52:34HCli'm not sure if this is gonna work with just a resolution of seconds.. actually, the unix implementation returns 1 second + microseconds if more than a second have passed..
15:53:19LinusNHCl: what do you need the resolution for?
15:53:25HClLinusN: timing in the emulator..
15:53:29[IDC]Dragonthe number talking shouldn't get restricted to 16 bit, imho
15:53:53ZagorHCl: is current_tick enough?
15:53:55LinusNHCl: how high resolution do you need, and what for?
15:54:00HClZagor: probably
15:54:13HClif you can guarantee that current_tick will keep increasing
15:54:25Zagoronly until it wraps :)
15:54:25jyp[IDC]Dragon: Really ? Because it would lead to quite some changes
15:54:25LinusNit does
15:54:44HCli'm not sure how high the resolution should be, but less than a second, i know that.
15:55:07LinusNthe rockbox tick is 10ms
15:55:22HClshould be more than enough resolution
15:55:57[IDC]Dragonjyp: we need 4 bits for the unit one for the sign, so this would leave only 11 bits
15:56:00HCli don't think unix system provide a much better resolution than 10ms anyways
15:56:11HClsystems*
15:56:12sneakumsdepends on the unix
15:56:15HClmhm
15:56:33[IDC]Dragonwe wouldn't be able to say numbers beyond 2047
15:56:46HClok, current_tick is just in the api, right?
15:57:02LinusNHCl: ???
15:57:15HClhm? o.o
15:57:29HCli'll use current tick
15:57:33[IDC]Dragonjyp: what are the consequences of having this long?
15:57:33LinusNcurrent_tick is a global variable
15:57:39HClits mostly to check how much time has been passed between a certain time
15:57:41HCland another time
15:57:43LinusNand the plugin api has a pointer to it
15:57:43HClyea
15:57:47HCl*nods*
15:58:10jyp[IDC]Dragon, mainly changing it all along the call chain; qutite tedious
15:58:14LinusNthere are two useful macros, TIME_AFTER() and TIME_BEFORE()
15:58:23HClso i'll just do current_tick-old_tick*HZ*1000000
15:58:27HCloh
15:58:33HClwhat do they do exactly?
15:58:53LinusNthey tell you if a tick value is before or after another
15:58:56jypNothing too difficult nor costly I guess
15:59:04HClokay..
15:59:06LinusNand it handles the wrapping
15:59:06HClthanks
15:59:07HCl :)
15:59:09HCl*nods*
15:59:14HClthat'll do nicely
15:59:31LinusNjust subtracting them will probably be fine
16:00
16:00:24linuxstbLooking in firmware/export/kernel.h, TIME_AFTER just subtracts them... How does that handle wrapping?
16:00:33HClneeds to be divide, by the way
16:01:19HClis the size of a pointer the same size as a long on iriver?
16:01:28HCli have a bit of a bad x86 coding habit where i assume they are..
16:02:34jyp[IDC]Dragon, looks like TALK_ID is used only for small numbers
16:02:39 Join webguest73 [0] (~d91ef962@labb.contactor.se)
16:03:29Zagorthey are, but please avoid depending on it
16:04:44HCl*nods*
16:05:32 Quit webguest73 (Client Quit)
16:05:32LinusNlinuxstb: in fact, subtracting them works just fine
16:06:12LinusNas they are unsigned
16:08:18HCloi, i don't like how C uses * for both pointer dereference and multiply
16:08:38preglowyou rarely end up mixing them together
16:08:56HClint sys_elapsed(long *oldtick) {
16:08:56HCl return ((rb->current_tick-(*oldtick))*1000000)/HZ;
16:08:56HCl}
16:08:59preglowthough stuff like *a * *b looks nasty
16:09:00HClwhats wrong with that?
16:09:08HClgives an compile error
16:09:14HClsys/iriver/iriver.c:54: error: invalid operands to binary *
16:09:57HCli'm confused to whats wrong with it.. :/
16:10:12LinusNrb->current tick is a pointer
16:10:18HClok
16:10:25LinusNuse *(rb->current_tick)
16:10:43 Quit Patr3ck_ ("User pushed the X - because it's Xtra, baby")
16:10:56HCl*nods*
16:11:02LinusNgotta go, cu later
16:11:06HClbye..
16:11:21linuxstbLinusN: OK, I will believe you. I guess it takes a very long time for them to wrap anyway.
16:11:40LinusNindeed
16:11:43LinusNbye
16:11:45 Part LinusN
16:12:34linuxstbAbout 8 days... (for current_tick to wrap)
16:13:22Bagder100 ticks per sec surely takes longer to wrap
16:14:38Zagori get 497 days
16:14:55Bagderit might be signed, making 248 days
16:15:17HCllol.
16:15:30HClno risk there
16:16:28linuxstbYes, sorry - (2^32)/100 seconds is about 11930 hours = 497 days
16:16:42Bagderhttp://uptime.netcraft.com/up/today/top.last.html :-)
16:18:29Zagorinteresting how 5 of the top-10 are swedish servers and 8 of 10 are scandinavian
16:18:34HClhm..
16:18:46linuxstbSurely that is easy to fake?
16:18:50sneakumsit's all the clean living
16:19:40Bagderalso, there's not a single Linux in the top-50
16:19:42Zagorlinuxstb: probably, but those are major state-owned institutions that I doubt would do that just for fun.
16:20:03bobTHCZagor> swedish wizards once again hit ;)
16:20:33HClhrm.
16:20:54HClrockbox doesn't update its lcd buffer automatically, does it?
16:21:03Zagorno, you need to call lcd_update()
16:21:06HClmhm..
16:21:18HCland i'm guessing it doesn't support 4bit pixel data either..
16:21:30Bagder1bit only so far
16:21:33HClmhm...
16:21:52jyphmhmhhn ;)
16:21:56HClgnuboy is designed to directly write onto a framebuffer
16:22:07HClwith either 8,16,24, or 32bit modes
16:22:47Bagderenters 1bit mode... :-)
16:23:02jyphttp://www.namesuppressed.com/kenny/
16:23:46Bagderwell, hopefully you'll get 2 bits "soon"
16:24:13preglowHCl: afaik, gameboy uses four shades of gray, just like the h1x0, so you COULD just hack in a 'write directly' mode
16:25:33HClpreglow: yea, i was thinking about that.
16:25:41HClBagder: thats sort of what i'm waiting for :P
16:25:51HCli think i'll take a break
16:26:17HClgnuboy pretty much compiles, all thats left is to point it at the correct rom to load, make the lcd interface and make the joystick interface
16:36:00linuxstbHCl: Maybe your emulator be implemented as a "viewer" - i.e. like the text file or jpeg viewer. You would then just "play" a ROM in the normal file browser.
16:37:04preglowhahaha
16:37:05preglowthat would rock
16:37:09bobTHCit's a good way to simply browse
16:37:17preglowindeed
16:37:45Bagdermaking a plugin a viewer is fairly easy
16:39:48lImbusit's already done for those chip8-emus, isn't it ?
16:40:04Bagderyes
16:40:10HCllinuxstb: yea.
16:40:21HClthat was sort of what i had in mind
16:40:56lImbusit's mainly like argv and argc except that argc is always 1
16:41:59lImbusdefine a file type (extension, that is) add a line to the viewers.cfg and you're done. no need to modify anything else in rockbox
16:48:06 Quit R3nTiL ()
16:48:55HClheh, i'll be happy if it runs first
16:52:14***Saving seen data "./dancer.seen"
16:58:05 Part Zagor
17:00
17:05:35HClhm
17:06:08HClhow do buttons work?
17:06:25HCli read something about a queue? i just repeat button_get till i get BUTTON_NONE ?
17:06:28 Join webguest61 [0] (~d570e850@labb.contactor.se)
17:06:58preglowit seems to be queue based, yes
17:07:01preglowso i guess so
17:09:03HClokay.
17:11:15ashridahhrm. wouldn't you want raw access to the buttons to get smoother gameplay? (and i still have no idea how you intend to deal with the fact that the joystick looks like it can't press two directions to get a diagonal...
17:12:32Bagderyou can check for buttons, it doesn't block if you don't want to
17:13:29lolo-laptopis it the button processing hardware that doesn't send multiple button signals simultaneously? That is the l4m3.
17:13:30ashridahfair enough then
17:14:18preglowlolo-laptop: that's the way it usually is, you can get away with one wire instead of several then
17:15:38ashridahi don't see the problem with that. it's not a gameboy after all, it's an mp3 player. :)
17:16:22lolo-laptopserial >= parallel in most cases −− if the bus is fast enough couldn't the game controller simulate it similar to how X-Windows similates a 3rd button with chord middle?
17:16:23bobTHC:)
17:16:53ashridahhell, it's how i designed my data input systems last time i built something using a 6811 uc.
17:17:09HClwhere's a list of all the button values etc?
17:17:17Bagderbutton.h
17:17:18 Join mecraw [0] (~mecraw@69.2.235.2)
17:17:22HClashridah: eh.. i think it should be fast enough o.o;
17:17:35HClas badger said, i just make it non-blocking
17:17:52ashridahtrue.
17:18:09lolo-laptopchuckle he probably has the most often typo'd name around. −− it'll take some getting used to in games to do a rolling press to get diagonal, but it should work, eh?
17:18:43preglowlolo-laptop: it's a hardware issue, one buttons physically blocks the other
17:19:01ashridahlolo-laptop: pff. that's what tab complete is for
17:19:30lolo-laptoppreglow: so a rolling press doesn't send both button codes :-(!
17:20:22ashridahpreglow: i was more thinking of the physical issues, as in the joystick seems to have no way to press both at once, at a quick glance
17:23:22HClhrm.
17:23:41HClashridah: correct
17:24:20*HCl isn't sure how to detect releasing buttons o.o;
17:24:46HCli.. guess.. i'll.. just add a last-state var and check whether it was pressed last time, and if it wasn't, release it..
17:24:47preglowyou could maintain an internal map
17:24:50HClwas*
17:24:54BagderBUTTON_REL
17:25:01preglowthat simplifies it :P
17:25:04HClohh
17:25:06HClgreat.
17:25:09HClabsolutely great :p
17:25:36Bagderreading existing code will help you finding out these things
17:26:05HClyea
17:26:08HCli guess
17:26:16lolo-laptopdespite being overall less game good the H340 buttons seem to cooperate with being pressed 2 at once alright...
17:26:19HClsorry
17:29:21LUMOS1Joerg, are you there??
17:29:32[IDC]Dragonyes
17:29:32 Nick LUMOS1 is now known as Trevmar (~trevor@ca-agoura-cuda2h-53.ventca.adelphia.net)
17:29:35TrevmarJoerg, are you there??
17:29:44[IDC]Dragonstill, yet
17:30:27TrevmarI noticed that post about the USB which disappeared from nansen2, and it fits the remaining Ondio sitting on my workbench to a T
17:30:55TrevmarCan you think of any testing which might be useful now I have CVS running?
17:31:22[IDC]Dragonhang on
17:31:29[IDC]Dragonaway for a sec
17:31:32Trevmarok
17:32:14 Quit Christi-S (Read error: 110 (Connection timed out))
17:32:20 Join Christi-S [0] (~christi@213.78.97.157)
17:38:34 Nick Trevmar is now known as LUMOS (~trevor@ca-agoura-cuda2h-53.ventca.adelphia.net)
17:38:36 Quit jyp ("poof!")
17:38:40 Nick LUMOS is now known as Trevmar (~trevor@ca-agoura-cuda2h-53.ventca.adelphia.net)
17:46:13 Nick Trevmar is now known as LUMOS (~trevor@ca-agoura-cuda2h-53.ventca.adelphia.net)
17:46:18 Nick LUMOS is now known as Trevmar (~trevor@ca-agoura-cuda2h-53.ventca.adelphia.net)
17:49:29HClok, that covers iriver buttons <-> gnuboy
17:50:32HClwho wants to extend rockbox to support 4 bits grayscale? :x
17:50:46Bagder2bits
17:50:54Bagder4 colors
17:50:57*Bagder runs away
17:52:01*preglow hides.
17:52:17lolo-laptopwho wants to extend it to support full color on the H3x0?
17:52:22*lolo-laptop runs farther
17:57:28HClsorry
17:57:29HCl2 bits
17:58:13[IDC]DragonTrovor: back again
17:58:23[IDC]DragonTrevor even
18:00
18:00:00 Join R3nTiL [0] (~zorroz@217.30.249.40)
18:00:26[IDC]Dragonor Trevmar
18:01:15Trevmarok i am here
18:01:59[IDC]Dragonhave you tried just multivolume, no hotswap?
18:02:25TrevmarI tried hotswap and no multivolume, but there were dependencies
18:02:36HCligh..
18:02:38[IDC]Dragonyes, that doesn't work
18:02:47[IDC]Dragonandmakes no sense either
18:03:13TrevmarOK, but I thought that the Jan 2 breaking date would indicate it was multivolume
18:03:24Trevmarthat was broken
18:03:26[IDC]Dragontry wit current cvs
18:03:42[IDC]Dragonto make sure you'r not missing any later fix
18:03:50TrevmarOk, will do.
18:06:18HCllinus isn't around...
18:06:28HCli'm slightly annoyed with the way i have to do video at the moment..
18:06:45HClcause it goes from.. 8bit framebuffer in memory -> 1 bit framebuffer in memory -> lcd framebuffer
18:10:49[IDC]Dragonvideo?
18:11:02HClgameboy video output.
18:11:06HCldisplay is generally called video
18:11:18[IDC]Dragonah, graphics
18:11:26[IDC]Dragon;-)
18:11:52HCli'm trying to get at least the 8bit -> 1bit conversion out.. while at the same time telling rockbox to only update one scanline..
18:12:53thegeekhave you tested a binary?
18:12:57thegeekjust to see if it actually works
18:13:11thegeekI know there is no gfx/controls/sound or anything
18:13:33[IDC]DragonHC1: minimum update is 8 scanlines
18:13:47[IDC]Dragonthe line may be short
18:14:16[IDC]Dragonbut a byte in frame buffer is 8 bits *below*
18:14:35[IDC]Dragonnot across, as intuition tells
18:15:52HClthegeek: no.
18:16:01HClthegeek: whats the point of testing it without gfx/controls/sound?
18:16:18HClthegeek: i haven't even linked the plugin entry point to the gnuboy emu
18:16:31HCli wouldn't even be able to tell whether it worked
18:17:13HCl[IDC]Dragon: one byte is 8 scanlines, right?
18:17:26thegeeksurely you could just add some debug messages in some critical places
18:17:29thegeekto see if it runs
18:17:43HCli guess.
18:17:44thegeekI don't know, I really don't know rockbox code;)
18:17:48thegeekdo it your way
18:17:52thegeekdon't listen to me;)
18:18:32HCl :P
18:18:33HClok
18:18:48*HCl scratches his head
18:19:53HClthis is so frustrating, cause gnuboy has 0 comments/code documentation
18:22:22 Quit bobTHC ("( www.nnscript.de :: NoNameScript 3.81 :: www.XLhost.de )")
18:22:29HClah.. palette thing..
18:24:23 Join webguest62 [0] (~513ec04a@labb.contactor.se)
18:25:21 Join Cap_d_Stuzz [0] (~SHEET@host162-222.pool80183.interbusiness.it)
18:25:51Cap_d_Stuzzso?
18:26:23TrevmarJoerg: It says *Panic* Reading past end of card" on the 03 08 unit with multivolume enabled
18:27:28Trevmarluckily my USB still works
18:27:51HCligh.
18:28:00TrevmarJoerg: please confirm you have seen this :)
18:28:06*HCl doesn't like math.
18:28:15 Part Cap_d_Stuzz
18:29:18preglowi love math
18:29:42*ashridah flees
18:30:50 Quit webguest62 ("CGI:IRC")
18:30:56[IDC]DragonTrevmar: the panic is checking if an access beyond the card's capacity is tried
18:31:16[IDC]Dragonno, I haven't seen it happening
18:31:47[IDC]DragonJens introduced that check not too long ago
18:32:03 Join Sucka [0] (~NNSCRIPT@host81-156-215-25.range81-156.btcentralplus.com)
18:32:37[IDC]Dragonif it triggers, something is ill before
18:33:37[IDC]Dragonwe should extend it with some output: which card, which address
18:33:53[IDC]Dragonmaybe MV is confusing cards
18:36:19 Quit lImbus (Remote closed the connection)
18:36:32 Quit R3nTiL ()
18:38:45[IDC]DragonTrevmar: do you feel capable to do so?
18:41:39HCl[IDC]Dragon: to map a scanline onto an rockbox scanline i need to divide the scanline by eight and the modulo 8 of the scanline will be the individual scanline within the byte, right?
18:43:16HClcan i just shift a one by that much to set it?
18:44:28TrevmarSorry Joerg, I went to get coffee :) Yes, I used to embedded program a lot (about 20 years ago :( ) Tell me what you need doing, and try and make it easy for me to handle, as this is what I have been doing for th elast 5 years or so: http://autoimmunityresearch.org/chicago2005.htm
18:47:36 Join lImbus [0] (~manuel@kernel.cycos.net)
18:48:57 Quit preglow ("and i'm away")
18:52:15***Saving seen data "./dancer.seen"
18:52:23*HCl prods people who know about the lcd framebuffer :/
18:59:36 Join webguest78 [0] (~50fca7f2@labb.contactor.se)
19:00
19:00:38HClthats just evil...
19:00:44HClgcc shouldn't even accept this line of code..
19:00:45HCl *(dest+++scanline_mapped)&=~(1<<scanline_index);
19:02:24[IDC]DragonTrevmar: are you around for a while? We can cantinue this in an hour or so
19:02:37HCl[IDC]Dragon: do you know anything about the lcd buffer format?
19:02:46[IDC]Dragonyes
19:03:45TrevmarJoerg,
19:03:48[IDC]Dragonas I said, a byte are the piyels ontop of each other for 8 scanlines
19:04:10[IDC]Dragonthe next byte wil be the 8 pixels to the right of that column
19:04:11amiconnHCl: It's somewhat described in firmware/drivers/lcd-h100.c. Search for "About Rockbox' internal bitmap format"
19:04:16Trevmaryes I will check back from time to time. Best to send me an email in case I lose my IRC connection, though
19:04:46TrevmarIRC keeps disconnecting every few hours :(
19:05:12[IDC]DragonTrevmar: I'm still confused about which disease I should contact you ;-)
19:05:14HClamiconn: okay
19:05:33Trevmarno problems, we study them lol -LOL
19:05:35 Nick Lynx_ is now known as Lynx_awy (HydraIRC@134.95.189.59)
19:06:39TrevmarJoerg, I did do a transition from pure Elecronic Engr to Biomed Engr in 1985 - then I got caught up in RISc and Intrenet and whatever until about 2000 -LOL
19:07:10HClokay.
19:08:13 Quit webguest78 ("CGI:IRC (EOF)")
19:08:15HClamiconn: ok.. i *think* i got it right then..
19:09:04Suckahows it going guys?
19:12:14HClhmm.
19:12:16HClso-so
19:13:30Suckaooh
19:13:39Suckaare you the guy trying to get a gameboy emulater to work on it?
19:13:45Suckai was reading irc logs ;)
19:13:57HClthats me, yea.
19:14:00Suckaawesome :D
19:14:08HClthe button interface was trivial...
19:14:14HCli *think* i just finished the lcd interface
19:14:29HClso i just need to connect gnuboy's main() to the plugin interface
19:14:33HClbut i'm gonna eat dinner, ta
19:14:38Suckabye
19:19:01 Join Stryke` [0] (~Chairman8@resnet-241-86.resnet.umbc.edu)
19:23:18*HCl returns
19:23:38lImbusdinner in 9 minutes. wow
19:27:25HCl :P
19:27:38HClno interesting people to talk with.
19:28:46linuxstbHCl: I've just had a quick look at the gnuboy sources - have you included inflate.c ?
19:29:18HClno.
19:29:36HClcompression needed yet another realloc buffer.
19:29:49HCland i didn't think it was worth the memory sacrifice.
19:30:01linuxstbNo - I think 40GB is enough for GB roms...
19:30:05HClmhm.
19:30:21HCldo you know the lcd framebuffer some?
19:30:34HCli'm kind of wanting/needing someone to tell me whether this code i have is correctish
19:30:42linuxstbMe? No, not at all.
19:30:57HClk :/
19:31:36linuxstbHave you tried building for the iRiver simulator?
19:33:28HClno.
19:33:39HClthats a good idea, by the way.
19:34:03HClaside from not having a windows iriver simulator, nor a build environment to make one
19:34:29linuxstbWhat OS do you use? There's an X11 simulator for Unix as well.
19:35:09linuxstbOr are we talking about different things?
19:35:30linuxstbI'm talking about the Rockbox simulator.
19:35:32HClwe are.. my workstations are windows, but my server runs linux
19:35:38HClhence i prefer a windows simulator
19:35:46HClbut i guess i can use vnc
19:35:50HCland use the x11 one
19:36:06HCllet me hook up its main() thing first..
19:36:06 Join amx [0] (~amx@Ottawa-HSE-ppp262227.sympatico.ca)
19:40:15ashridahhm. the simulator doesn't work with cygwin?
19:42:29HCli don't know
19:42:31HCli don't have cygwin
19:43:05 Quit thegeek (Read error: 101 (Network is unreachable))
19:43:24 Join thegeek [0] (na@ti521110a080-3006.bb.online.no)
19:47:13 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
19:47:51 Join einhirn [0] (Miranda@carlsberg.heim2.tu-clausthal.de)
19:48:22HClwhere could i alter the memory sizes assigned to plugins again?
19:49:10 Join ripnetUK [0] (~mirc@82-70-100-230.dsl.in-addr.zen.co.uk)
19:49:16lImbusuh. that must be somewhere in the startup of rockbox. isn't there a define for that in plugin.h ?
19:49:27HCllinus told me where it was earlier
19:49:31HClbut my irc client crashed
19:49:33HClso i lost it
19:49:37[IDC]Dragonit's in the linker definition file
19:49:39ashridahHCl: so use the online logs
19:49:46[IDC]Dragon*.lds
19:49:47HClohyea.
19:49:52HCl[IDC]Dragon: ok..
19:50:31 Quit ripnetUK (Client Quit)
19:54:33[IDC]Dragonbbl
19:54:37 Quit [IDC]Dragon ("CGI:IRC")
19:54:51HCltitania:/home/hcl/rockbox/rockbox-devel/iriver# ls -al rockboy.rock
19:54:51HCl-rwxr-xr-x 1 root root 4873512 Feb 9 19:53 rockboy.rock
19:54:57HCl4.8 mb :x
19:55:03ashridahrofl
19:55:11ashridahdid you link in libc by accident? :)
19:55:14HClno xD
19:55:24 Join webguest99 [0] (~3e234217@labb.contactor.se)
19:55:40HCli'm assuming it allocates space for the buffers right in the .rock?
19:55:55webguest99has linus work on iriver port since yesterday ?
19:56:02webguest99has linus work on iriver port since yesterday ?
19:56:06HCllook at the cvs weblog.
19:56:29webguest99i spoke about the bootloader and the hold button ?
19:56:50HCli don't think he added support for that
19:56:58thegeekhe's working on it
19:57:06webguest99cool
19:57:27webguest99my battery suffered today , my iriver stand powered on all the day
19:58:44*HCl stares at gnuboy.
19:58:54thegeekgo!
19:58:55HClnow i'm starting to get scared to finish it and find out it doesn't work at all
19:58:56HClxD
19:59:19HClall thats left is linking the plugin entry point to the main() and adding proper procedures to shut the emulator down
19:59:40 Quit Aison (Read error: 104 (Connection reset by peer))
19:59:59ashridahplease, you've done 90% of the work. now you've just got 90% to go
20:00
20:00:05ashridahyou're not finished :)
20:00:08linuxstbHCl: Congratulations on getting it to build. Now the hard work begins...
20:00:16 Join Aison [0] (~hans@zux166-181.adsl.green.ch)
20:01:07HClwell. hm.
20:01:15HClsay i wanted to access a file in the root of the drive
20:01:22HClwould i just open "\\blah" ?
20:02:05linuxstbI think you use "/blah"
20:02:10HClok
20:02:30HCli'm gonna try and make it just.. load /rom.gb by default
20:02:41linuxstbLooking in bootloader/main.c, the bootloader looks for "/rockbox.iriver"
20:03:53HClok
20:03:59amiconnHCl: /blah is correct; rockbox uses unix path syntax (and only supports absolute paths)
20:04:02HClokay
20:05:24HClis there a function that like, terminates a plugin immediately?
20:05:40HClor am i just gonna have to let my entry point return nicely..
20:05:41 Join ripnetUK [0] (~mirc@82-70-100-230.dsl.in-addr.zen.co.uk)
20:05:44HClwhich is gonna be a bit of a pain..
20:06:05ripnetUKfor the record, its looking a lot like its the new Mandrake kernel thats been eating my filesystem, not Rockbox :)
20:06:10thegeekhumhum
20:08:36HCl :P
20:13:57HClokay.
20:14:01HClnow i'm scared to test it :x
20:15:02ashridahwimp :)
20:15:42HCl :p
20:15:50HClwell.
20:15:52HClits just that.
20:15:54HClif it doesn't work
20:15:57HClit means more work
20:16:03ashridahbah
20:16:05ashridahwimp.
20:16:05HCland otherwise, i'll just not know :p
20:16:07HCl :P
20:16:30HClit'll most likely crash horribly.
20:16:37HCladding a stop button might be a good idea too, heh
20:17:56 Join hubble [0] (hubble@h13n1fls302o1033.telia.com)
20:18:43 Join [IDC]Dragon [0] (~idc-drago@p3EE2D52B.dip.t-dialin.net)
20:19:32HClhow do i check the hold button?
20:21:47hubbleAnyone know how UDA1380 i2s interface is connected to the CPU?
20:22:01[IDC]DragonTrevmar: do you read?
20:23:20Trevmaryes, I am here
20:24:00[IDC]Dragonso, where were we? MV works, but no hotswap?
20:24:19Trevmarno, MV enabled gives PANIC
20:24:26[IDC]Dragonok
20:24:29 Quit webguest99 ("CGI:IRC")
20:24:31Trevmarusing the latest CVS
20:24:38[IDC]Dragoncan you go into ata_mmc.c ?
20:24:45Trevmarjust a minute
20:24:56*HCl does a make zip...
20:25:11[IDC]Dragonline 646
20:25:51Trevmararrgh this editor doesnt have line count, wait a bit...
20:26:06Trevmarcopy and paste the line pse
20:26:12[IDC]Dragonsearch for panicf
20:26:31TrevmarOK, got the message
20:26:53Trevmaryou want to porint some variables there?
20:26:54[IDC]Dragonextend this to print the offending values
20:27:06Trevmarer, which values?
20:27:20[IDC]Dragonjust a sec
20:29:24HClhrm.
20:29:35[IDC]Dragonpanicf("Reading %d past end %d of card %d\n", start+incount-card->numsectors, card->numsectors, drive);
20:30:02HClhrm.
20:30:10[IDC]Dragonthis only works for MV enabled, because of the drive param
20:30:11HClrockboy isn't getting zipped when i do a make zip... why?
20:30:36[IDC]DragonHCl: you need to add it to SOURCES
20:30:59HClit is
20:31:11TrevmarOK, will compile and test...
20:31:12HClotherwise it wouldn't get built
20:31:19HCli guess i can copy it manually...
20:33:07amiconnTrevmar: I have one more thing you could check (with your test build or plain daily, and with a build without multivolume & hotswap for comparison)
20:33:18TrevmarGA
20:34:22amiconnWith _both builds_, go to Info->Debug->View MMC info, and note all 4 pages _both with & without_ MMC inserted
20:35:31amiconnBe sure you don't mix the headlines; there are [MMC0 p1], [MMC0 p2], [MMC1 p1], [MMC1 p2]
20:36:03[IDC]Dragonlooking for a wrong enable?
20:36:21TrevmarJoerg, I got *PANIC* Reading 1 past end 0 of card 1
20:36:35[IDC]Dragonend 0 is no good
20:36:53[IDC]Dragonthanks
20:36:57amiconnLooks like the card doesn't get initialised properly
20:37:18Trevmarsame card worked fine with MV disabled :)
20:37:27amiconnTrevmar: Do you get this even when no card is inserted?
20:37:50TrevmarNope, no panic, machine works fine until I insert card
20:37:59HClhrm.
20:38:12amiconnTrevmar: What if the card is already inserted at boot?
20:38:16*HCl uploaded rockboy, put a rom on his iriver..
20:38:17[IDC]DragonHCl is stuck ina loop
20:38:20HClyea
20:38:22HClxd
20:38:26thegeekoh the excitement!
20:38:38HClit'll just crash horribly *nods*
20:38:39HClxD
20:39:12Trevmarthe card has to be inserted at boot because I have hotplug turned off still
20:39:21Trevmarjust MV enabled
20:39:26amiconnAH.
20:39:28HClincompatible version.
20:39:29HClo.o
20:39:40HClhey! how's that possible o.o;
20:39:47HClhm.
20:39:51HClnm o.o
20:40:21amiconnHrrrrmmm. I suspect problems with logic level when switching the select. I'd really prefer to have such a 0x0308 unit here for testing...
20:40:37HClthat explains why it didn't pack rockboy >.o old .zip
20:41:56[IDC]Dragonamiconn: I admit that I don't understand why the circuit works with both AND and OR
20:42:10amiconn???
20:42:12[IDC]Dragonand just flipping a polarity
20:42:21Trevmarthe hotplug fails with my AC08 units as well, remember
20:42:44Trevmareven though the MV works with them
20:42:45amiconnTrevmar: Yes, but multivolume does work there
20:42:51Trevmarand the ones reported on the forum were 07 08
20:43:09amiconnI think there are 2 separate problems
20:43:27 Join thegeek_ [0] (na@ti521110a080-0636.bb.online.no)
20:43:36 Quit webguest61 ("CGI:IRC")
20:43:56amiconn[IDC]Dragon: I think the circuit is a bit more different, not only AND replaced by OR
20:45:28amiconnUsing AND is the better choice imho (and I think this is why they changed it after all). With AND, the "idle" level is high, which is also idle level for MMC.
20:45:52*[IDC]Dragon stares at the pics
20:46:08[IDC]Dragonyes, the traces are somwhat different
20:46:39amiconnTrevmar: Does USB access work for both internal & external with MV-enabled rockbox?
20:47:02Trevmarwait - let me test it
20:47:40amiconnlImbus: r u there?
20:49:55 Quit thegeek_ (Read error: 54 (Connection reset by peer))
20:50:01Trevmaramiconn, the 0708 says "please remove inseted MMC" when I plug in SUB. daily build 20050204
20:50:12Trevmar..plug in USB..
20:50:23amiconnOf course.
20:50:24 Join webguest00 [0] (~0c977851@labb.contactor.se)
20:50:28 Quit thegeek (Read error: 60 (Operation timed out))
20:50:39 Join XShocK [0] (~XShocK@pcp09492659pcs.nrockv01.md.comcast.net)
20:50:46amiconnUSB connect *can* only work without the card inserted.
20:51:06XShocKhi all
20:51:22amiconnIf you want to access the card via USB, you have to pull it before and reinsert it after connecting
20:51:36amiconnThat's a hardware limitation, nothing we can do about
20:52:19***Saving seen data "./dancer.seen"
20:52:19amiconnAnd: MMC hotswap in USB has been working long before (and independent from) MMC hotswap for rockbox access
20:52:47 Part webguest00
20:53:49amiconnTrevmar: Does the 0308 unit say the same under your test condition? It should...
20:54:02TrevmarHotplug doesn't work in either of my 07 08 boxes. When I pull the card out the unit switches off with horizontal and vertical lines on the display
20:54:54amiconnYeah, I mean, test with MV but *without* hotswap, and primarily on the 0308 units
20:54:57TrevmarYes, the 03 08 and 07 08 both require card to be pulled out when USB plugged in
20:55:33amiconnDoes the actual access work both to the internal flash (when card is pulled) and the card (when plugged)?
20:56:05Trevmaryes, I get the <MMC1> on the 07 08 boxes
20:56:26amiconnNo, I mean USB access.
20:57:03TrevmarWhen I plug in the USB while the card is installed it tells me I have to pull the card
20:57:13amiconnYeah, of course.
20:57:15*[IDC]Dragon draws circuit diagrams
20:57:49amiconnTrevmar: I said: connect USB *without* card inserted. (1) Can you access the internal flash from the PC?
20:58:27amiconn(2) Plug the card now. Does the drive content on the PC now change to the card's content, i.e. access to card works too?
20:59:37Trevmaryes, the unit works fine without a card plugged in
21:00
21:00:20HClthat took a while to load, and as i said, it exited immediately XD
21:00:25HCli wonder why though
21:00:48TrevmarAmiconn, when I plug in USB
21:01:08amiconnTrevmar: Okay. What about (2)?
21:01:26Trevmarand then plug in the MMC the Ondio did not switch off, and on my PC I can not see any <MMC1> designator
21:01:37Trevmarjust the internal ondio memory
21:01:51amiconnOf course; this trick cannot work for USB access.
21:02:36amiconnThe drive (if you are on Windows) should now show the contents of the card *instead* of the internal flash content
21:02:44TrevmarThe moment I unlugged the USB cable I got the message "PANIC reading 1 past end 0 of card 1
21:03:16Trevmarlet me try again and see what I get in more detail..
21:03:24 Quit xen` (Read error: 104 (Connection reset by peer))
21:03:41amiconnTrevmar: Btw, that was on a 0308 unit, correct?
21:04:21TrevmarAmiconn, you are correct, I am looking at the card directory, not the Ondio's internal with my PC
21:04:38Trevmarthis is an 07 08 unit with 20050204
21:04:48amiconnOkay, nice.
21:05:06amiconnCould you please do the same test with a 0308 unit?
21:05:47TrevmarThat is interesting
21:07:35TrevmarI have gotten the units mixed up. With the 03 08 unit compiled without MV then I can see the card directory if I plug in the card while the USB is enabled
21:08:02Trevmaron the 07 08 unit with 20050204 I enable USB and then the unit switches off as I plug in the MMC card
21:09:14Trevmarthis is consistent with my 07 08 units switching off as the card is plugged in without being in the USB transfer mode as well
21:09:49TrevmarI think a variation of this is what is acusing those other huys to lose their USB maybe? CPU going off into the weeds?
21:10:23Trevmarcausing these other guys
21:10:25amiconnHmmm. For the 0708 problem, if you have a rockbox _without_ hotswap on the unit, and insert a card into the running box (what obviously wouldn't cause it to be detected), does the unit switch off too?
21:10:49Trevmarlet me try that
21:12:14TrevmarI am putting the latest CVS without MV or hotswap into the 0708 unit. Then later I will recompile with only hotswap off. one step at a time...
21:14:48HClcannot open romfile
21:14:48HClhrm.
21:15:07TrevmarNo, the 0708 unit resets when I insert either the MMC or the Sd even with MV and Hotswap disabled
21:16:28TrevmarAll ofthese units came with lots of brass shielding soldered into place...
21:16:51TrevmarI took that all out
21:17:05Trevmarno apparent change in performace
21:17:23Trevmareven the radio was just as cruddy reception with and without, IMO
21:17:31 Quit DMJC (zelazny.freenode.net irc.freenode.net)
21:17:31NSplitzelazny.freenode.net irc.freenode.net
21:17:39 Quit XShocK (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit [IDC]Dragon (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit ripnetUK (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit amx (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit Stryke` (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit lImbus (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit Sucka (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit Christi-S (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit HCl (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit Quel|away (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit amiconn (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit QT_ (zelazny.freenode.net irc.freenode.net)
21:21:52 Quit Shulberry (zelazny.freenode.net irc.freenode.net)
21:21:53 Quit crash__ (zelazny.freenode.net irc.freenode.net)
21:21:53 Quit rasher (zelazny.freenode.net irc.freenode.net)
21:21:53 Quit sneakums (zelazny.freenode.net irc.freenode.net)
21:22:29NHealzelazny.freenode.net irc.freenode.net
21:22:29NJoinXShocK [0] (~XShocK@pcp09492659pcs.nrockv01.md.comcast.net)
21:22:35TrevmarI reloaded the latest CVS with MV enabled into the 07 08 unit and it still switches off when you hotplug a card, same as before, but the <MMC1> directory is accessible and music plays OK
21:23:44NJoin[IDC]Dragon [0] (~idc-drago@p3EE2D52B.dip.t-dialin.net)
21:23:44NJoinamx [0] (~amx@Ottawa-HSE-ppp262227.sympatico.ca)
21:23:44NJoinStryke` [0] (~Chairman8@resnet-241-86.resnet.umbc.edu)
21:23:44NJoinlImbus [0] (~manuel@kernel.cycos.net)
21:23:44NJoinSucka [0] (~NNSCRIPT@host81-156-215-25.range81-156.btcentralplus.com)
21:23:44NJoinChristi-S [0] (~christi@213.78.97.157)
21:23:44NJoinHCl [0] (hcl@titania.student.utwente.nl)
21:23:44NJoinQuel|away [0] (~kvirc@85.48.33.190)
21:23:44NJoinamiconn [0] (~jens@pD95D135F.dip.t-dialin.net)
21:23:44NJoinShulberry [0] (Taxi@oslo-dhcp-248-180.bluecom.no)
21:23:44NJoincrash__ [0] (~crash@a15167580.alturo-server.de)
21:23:44NJoinsneakums [0] (~sneakums@doublethink.psax.org)
21:23:44NJoinrasher [0] (~rasher@62.79.64.148.adsl.hs.tiscali.dk)
21:24:07*amiconn hates sudden mega-splits
21:24:11 Join QT [0] (as@area51.users.madwifi)
21:24:13NJoinDMJC [0] (~James@220-245-162-47-sa-nt.tpgi.com.au)
21:25:26NJoinripnetUK [0] (~mirc@82-70-100-230.dsl.in-addr.zen.co.uk)
21:25:27amiconnTrevmar: I had to gather your latest statements from the log because of the split...
21:26:14amiconnIf your 0708 unit shuts down regardless of rockbox features, I'm pretty sure this unit has got a hardware problem
21:26:42amiconnSomehow the MMC insertion force cuts the power
21:29:37lImbushö ? what's up with chanserv ? it just asked me to be sure the question was not on faq...
21:30:18amiconnI dunno....
21:30:23 Join webguest93 [0] (~513ec04a@labb.contactor.se)
21:30:28lImbusah, i c. netsplit, and the window was not yet scrollen down, so I did not notice tzhe split
21:30:34 Quit webguest93 (Client Quit)
21:31:22amiconnlImbus: As you are obviously here now - I remember you having an Ondio with 0308 mask, and that it is broken, but I don't remember *what* broke.
21:31:26lImbusbtw, amiconn: I am there. and now my window is scrolling again
21:31:30lImbushehe
21:32:05amiconnCould you tell me again?
21:32:17lImbuscurrently typing :-)
21:32:39lImbusmy ondio: it was not yet flashed, so archos firmware should boot it up, then load the ajb-file from "disk".
21:32:43lImbusbut it doesn't.
21:32:58lImbusit takes several minutes to boot (if it does not fail quitely)
21:33:05lImbusthen is in the archos-menue
21:33:08HClopen("/rom.gb", "rb");
21:33:13HClanyone any idea why that would fail?
21:33:21HClrom.gb exists on my player
21:33:41 Join Zagor [242] (foobar@h254n2fls31o265.telia.com)
21:33:54lImbusamiconn: (that one menue that can bee seens on joergs screenshot for OndioBacklight
21:34:31lImbusif I try then to do ANYTHING if aborts with several error messages. "not formatted" or "incorrectly formattet"
21:34:38HClanyone? :/
21:34:43HClshould it just be "r" ?
21:35:33lImbusHCl: what fails? to compile, or run ?
21:35:37amiconnHCl: The second parameter is flags, *not* a string.
21:35:37 Nick Quel|away is now known as quelsaruk (~kvirc@85.48.33.190)
21:35:40HClneither
21:35:42quelsarukhi
21:35:43HClamiconn: ahhhhhh
21:35:44HCldoh.
21:35:46HClokay
21:36:01HClmy bad
21:36:08HCllittle fopen -> open bug
21:36:11amiconnHCl: I *really* recommend reading exisiting code...
21:36:29HClsorry o.o
21:36:38lImbusamiconn: if I insert an external mmc-card, it does the same, but tells me to extract the card
21:36:40HCli don't use the open() stuff much.
21:36:40amiconnlImbus: Did you once try to reformat the internal memory? If yes, maybe you accidentally formatted anything else than FAT16?
21:36:47lImbusno
21:36:50lImbusdidn't try
21:36:58lImbusworked until it died one bad day
21:37:12lImbuscan't access it with usb neother
21:37:18lImbuss/neother/neither
21:37:37amiconnHmm, that sounds like the internal flash failed. Too bad it's a bga...
21:37:53lImbusuart-boot
21:37:54lImbus?
21:38:20lImbusmhmm wait. why would external mmc fail, if it's the internal flash ?
21:38:39amiconn???
21:39:24lImbussometimes I mess it up with the different memory types in archjos devices...
21:39:31lImbusmust be a misunderstanding
21:42:19Trevmaramiconn, I am sorry, I was distracted. here is my last message before we got diconnected −−>
21:42:22TrevmarI reloaded the latest CVS with MV enabled into the 07 08 unit and it still switches off when you hotplug a card, same as before, but the <MMC1> directory is accessible and music plays OK
21:43:55TrevmarAmiconn: if the shutdown is a hrdware problem (and I don't necessarily disagree) I have 2 07 08 Ondios with identical hardware problems, then.
21:44:28Trevmarbut it has nothing to do with MMC insertion force
21:44:35amiconnThat may still be the case. To be sure, you could check with archos firmware
21:44:47*[IDC]Dragon is back from dinner
21:45:58TrevmarYes, let me delete the ajbrec and see what I get
21:46:01[IDC]Dragonand done with circuit diagrams and truth tables
21:46:40lImbusamiconn: would the ondio have been working (with internal memory) after having formatted it with something else than fat16 ? i gave it once to a mate...
21:46:42[IDC]Dragonamiconn: I think the reason for the AND vs. OR is different
21:47:08[IDC]Dragonin rockbox, 2 things are changed:
21:47:33amiconnlImbus: If it's not flashed, no. Archos fw can only boot from FAT16
21:47:34TrevmarYes, with the archos software the insertion of an MMC card into those units causes them to shut down.
21:47:36[IDC]Dragon1) the SCk line is idle high or low, depending on the chip
21:48:00lImbusamiconn: ok, then I had never something else than fat16.
21:48:12[IDC]Dragon2) the clock gate polarity is reversed
21:48:24[IDC]Dragonnow comes the interesting:
21:48:30amiconn[IDC]Dragon: Yes, I know. Thats catered for.
21:49:00[IDC]Dragonthe clock line from the bridge must also be idle high or idle low
21:49:28[IDC]Dragonwith the OR, it must be 0 to allow us to clock the cards
21:49:41[IDC]Dragonwith the AND, itmust be 1
21:49:54amiconnAh, yes.
21:50:09amiconnHmm, how is this achieved?
21:50:22HClit still doesn't work x.x
21:50:36HClis it possible i'm getting assigned fd 0?
21:50:40[IDC]Dragonwe can't control it, so I guess the driving reason for the change was a different bridge behaviour
21:51:00amiconnHmm.
21:51:21amiconnThat should be obvious from the datasheets, shouldn't it?
21:51:39*HCl checks the other sources and guesses it is..
21:51:39amiconnMaybe it was the AU9330->AU9331 transition?
21:53:10*[IDC]Dragon tries to remember
21:54:52amiconnlImbus: I remember having read on the archos site that earlier Ondio fw versions could even get stuck when the internal flash is not formatted with FAT16.
21:55:28amiconnThe only chance to revive such a unit would be to uart-boot it, and flash a newer firmware (or preferably rockbox ;-) )
21:55:58 Quit einhirn (zelazny.freenode.net irc.freenode.net)
21:55:58 Quit dwihno (zelazny.freenode.net irc.freenode.net)
21:55:58 Quit Bagder (zelazny.freenode.net irc.freenode.net)
21:55:58 Quit linuxstb (zelazny.freenode.net irc.freenode.net)
21:55:58 Quit Hadaka (zelazny.freenode.net irc.freenode.net)
21:55:58 Quit hile (zelazny.freenode.net irc.freenode.net)
21:55:58 Quit coob (zelazny.freenode.net irc.freenode.net)
21:56:21amiconnlImbus: What archos fw version does this Ondio have?
21:56:51lImbusuh.
21:56:54lImbusdunno by hard
21:57:05lImbushave to look when I'm at home
21:57:36NJoineinhirn [0] (Miranda@carlsberg.heim2.tu-clausthal.de)
21:57:36NJoindwihno [0] (~dw@81.8.224.89)
21:57:36NJoinBagder [0] (~daniel@1-1-5-26a.hud.sth.bostream.se)
21:57:36NJoinlinuxstb [0] (~linuxstb@dsl-212-23-31-215.zen.co.uk)
21:57:36NJoincoob [0] (pen0r@82-44-227-205.cable.ubr11.haye.blueyonder.co.uk)
21:57:36NJoinHadaka [0] (naked@naked.iki.fi)
21:57:36NJoinhile [0] (hile@hack.fi)
21:57:42amiconnSince this is an older hw version, chances are that fw version is rather old too...
21:57:53lImbusof course
21:58:12amiconnMaybe it can be revived with uart boot.
21:58:43[IDC]DragonI was offering that
21:58:46lImbusit's planned that I send it to jörg anyways.
21:59:00lImbusas soon as I got it back from my co-worker
21:59:30lImbusif uart-boot and flash works-> nice, if not I donate the dead device. at least you can beep the tracks
21:59:55HClhrmf.
22:00
22:00:08HClwatashi wa baka desu yo.
22:00:16amiconnHCl: Yes, you can get assigned fd 0
22:00:25lImbuslol
22:01:57HClamiconn: ok
22:02:12*HCl scratches his head, at least got a little further
22:03:03HCli hope linus manages usb soon.. its annoying to having to keep switching back and forth between original and rockbox firmware
22:03:24 Quit Christi-S (Read error: 60 (Operation timed out))
22:03:44lImbushi cristi
22:04:05lImbusoops :-/
22:04:11[IDC]Dragonwrong direction
22:04:34lImbusI'm sure she's reading the logs :-)
22:05:44 Join lImbus_ [0] (~c24d9e4b@labb.contactor.se)
22:05:56 Part lImbus
22:06:52 Nick lImbus_ is now known as lImbus (~c24d9e4b@labb.contactor.se)
22:06:56HClwelp. thats somewhat what i expected....
22:06:59amiconnHCl: You get the first unused fd. If no other file is open, this is 0. Error is indicated by <0
22:07:02HClyea
22:07:12HClits videodriver isn't updating..
22:07:44HClat least it shuts down semi-properly
22:09:32HClwhich means its input driver is working
22:10:04HCland the code is running in general
22:12:31HClit might just be extremely slow
22:12:37amiconnlImbus: http://www.archos.com/download/firmware/README_ONDIO_FM_history.txt Read the notes regarding v 1.31f ...
22:12:44 Join Christi-S [0] (~christi@213.78.108.199)
22:13:04 Join jpburton5150 [0] (knoppix@cpe-24-94-54-216.stny.res.rr.com)
22:13:59[IDC]Dragonamiconn: could the order of things on idle level and clock select be important?
22:14:18[IDC]Dragonmaybe we produce a spike for the OR models
22:14:32jpburton5150hey... so, im trying to compile rockbox... but i get "Error: unrecognized arcgitecture specification
22:14:38jpburton5150'5249'
22:14:53jpburton5150which HCl, i read you had...
22:15:03jpburton5150so, how did you fix it?
22:15:13amiconn[IDC]Dragon: Yes of course, that's what I suspect.However, it is very difficult to test without such a unit.
22:15:24HCljpburton5150: cvs binutils
22:15:32[IDC]Dragontry to think...
22:15:39jpburton5150k thanks a lot, ill try that
22:15:44amiconnOR means idle low, and low isn't idle for MMC SPI protocol...
22:16:47lImbusamiconn: I really can't recall the version.
22:17:15amiconnYou'll have to look what the screen shows at startup
22:17:57lImbussilly) co-worker
22:18:04lImbusargl.
22:18:16lImbusthat cgi::irc is buggy
22:18:45lImbusI meant: yes, as soon as I get back my ondio from that BRACKET-OPEN silly BRACKET-CLOSE co-worker
22:19:33amiconnTrevmar: Btw, what version do your 0308 Ondios show at startup? If it's lower than 1.31f, you should consider either updating archos fw in flash, or flashing rockbox....
22:19:55amiconn...if it's not yet flashed of course
22:21:25TrevmarRockbox is flashed
22:22:05Trevmarleft-button boot shows 1.32b
22:23:01amiconnYes, that's expected then. The archos image in the flash packages is the latest (flash) version.
22:24:29amiconnAh, I just remember that you uart-booted and flashed a number of Ondios.
22:24:41amiconnWere these 0308 units among them?
22:24:50HClwill lcd_putsxy print to the right of the x,y you feed it or to the left?
22:25:10amiconnHCl: Right & bottom
22:25:23HClokay, and the top left pixel is 0,0, right?
22:25:32amiconnyup
22:25:38HClokay..
22:26:52TrevmarYes, these 03 08 units were flashed, after chnaing the mask at FC and FD in the binary image :)
22:27:16amiconnWhat was the problem with them before?
22:27:18Trevmarchnaing=changing sheez...
22:27:28amiconnDid they get stuck at the archos boot screen?
22:28:51Trevmarmost of them had no life. Looked like they had been badly flashed by their previous owners, I guess. One of the 03 08 units was fully operational until I flashed Rockbix, and it took me several days to figure out that the guy hadn't sold me a lemon, that it was actually rockbox at fault :(
22:29:42amiconnOkay, thanks.
22:31:06Trevmarno screen mainly, as I remember. I have one still on the workbench which I can't get at the main memory to make that visible to a PC over USB. I can display the ocntents of an inserted MMc with ajbrec on it. Remember that one? I keep figuring that when we get these probelsm understood that it will magically come to life again. Or that somebody wil write "low_level_format.rock" for me :)
22:32:31Trevmarbecause it will access and exceute off the MMC OK, only the main memory is dead in some way. It says "no partitions found, please insert USB and fix it" but the PC sees nothing across the USB
22:33:08amiconnYou could try flashing archos firmware (via uart). That should definitely work if it's not a hardware problem
22:33:15TrevmarIf I boot with an MMC card in place it takes that as the main partition. I described the screens in a thread on the forum
22:34:56TrevmarWhen I left-boot that one it comes up with the "Ondio FM" screen but never drops through to that 4-way selecetor display of Archos
22:37:44Trevmarwith Archos software I dont get to that 4-way selector even when I boot with an MMC card in place. With Rockbox it sees the MMC card
22:37:58Trevmarbut still no USB with rockbox
22:38:19amiconnWell, you should be able to access the MMC via USB
22:38:30 Join thegeek [0] (na@ti521110a080-0636.bb.online.no)
22:38:59amiconnSo to come back to the 0380 vs. MV+hotswap problem, I ask again if you could do the tests I proposed 2 hours ago.
22:39:07amiconn*0308
22:39:58amiconnFirst, put on a build without MV+hotswap, go to Info->Debug->View MMC info, and check all 4 screens both with & without MMC inserted
22:40:35HClwell...... that goes... slow..... :X
22:42:21TrevmarAmiconn, I tried your trick of plugging in the card after USB is active. It didn't switch off this one (which is 07 08) but Win2K prompts me to "insert a disk in drive h:" repeatedly
22:42:55Trevmaramicon, let me read those screens for you
22:43:26amiconnIt's not a trick... it's the required way of handling MMC in USB, even with archos firmware
22:44:41Trevmarwithout MMC: MMC0 p1 STM128 rev 6.6 prod:2/2006 M=02 O=0000 size 122MB
22:45:41TrevmarMMC0 p2: speed 20 Mbit/s Tsac 10ms Nsac 0clk R2W *4 IR max 35..35ma IW 35..45ms
22:47:37TrevmarMMC1 p1 not found and MMC1 p2 not found. OK lets' plug in the 1 gig MMC
22:48:34amiconnMMC0 p1: It really display 02/2006? Then you have a chip from the future....
22:48:55TrevmarMMC1 p1 and p2 not found. Let's reboot with the card in place
22:49:41Trevmaroops - displays PANIC (maybe cos there is an ajbrec on the card, let me fix that)
22:50:07amiconnquelsaruk: Do you have some goats to spare? ;-)
22:51:12amiconn[IDC]Dragon: Do you have the schematics in an accessible place?
22:51:26Trevmar8I keep getting panics. Let me triple check the build I have in there :)
22:51:39quelsarukhmmm
22:51:41quelsaruklet me check
22:51:45quelsaruki have some here
22:51:48quelsarukhow many do you want?
22:51:50quelsaruk;)
22:52:21***Saving seen data "./dancer.seen"
22:52:39[IDC]Dragonamiconn: no, I hand-doodled them
22:52:53 Join jyp [0] (~jp@202-181.240.81.adsl.skynet.be)
22:53:32crash__can someone ell me if the iriver port already runs on the h100 ? linus recently told me it only runs on h120 and h140
22:53:54Bagderno h100 user has tried it
22:54:06Bagderthere may be reasons why it won't
22:54:15amiconn[IDC]Dragon: Hmm. I try to understand if there would be a glitch, this would only affect MV-enabled rockbox. The clock select function did not change at all...
22:54:16Bagderlike the lesser amount of ram
22:54:47crash__Bagder: linus told me that there are also different chips or layout on it
22:55:09Bagdercrash__: you can't flash the 120/140 iRiver firmware to a 100, can you?
22:55:10[IDC]Dragonamiconn: but it's selected at rockbox time, not while USB
22:55:46amiconnIt's also selected without MV... and that's done immediately before the first access....
22:56:04crash__unfortunatly i havent a h100 myself
22:56:15crash__and would like to assure it can be done
22:56:15TrevmarMMC1 p1: MMC01G rev 0.0 Prod 12/2004 M=37 O=ffff Size 979MB
22:56:21HClh100...
22:56:26crash__so there are no hardware issues why it shouldn runt
22:56:26HClwhat kind of disk does that have?
22:56:29Bagdercrash__: it is not sure it can be done yet
22:56:45Bagderwe're trying to get our hands on a h100
22:56:53crash__Bagder: i ask cause linus seemd to be quite sure it wont
22:56:59crash__ah
22:57:12crash__i only could tell you someone who has it
22:57:18amiconnTrevmar: MMC0 p1/p2 still displaying correctly?
22:57:30HCldoes the iriver simulator miss the mode button?
22:57:42crash__but we wouldnt have the hardware to get it back to life if it fails ;)
22:57:50TrevmarMMC1 p2 speed 20Mbit/sTsac 80ms Nsac 0clk R2W *32 IRmax 60..80ma IWmax 60..80ma
22:58:52TrevmarMMC0 still the same, just MMC1 added
22:59:41BagderHCl: what sim? the x11 one?
22:59:45HClyen
22:59:45HClyes
23:00
23:00:23Bagderor rather *which* sim
23:00:37HCliriver x11
23:01:41Bagdertry the insert button
23:01:50HClokay
23:02:05BagderI usually read button-x11.c to figure out the keys
23:02:06HClah.
23:02:10HClok
23:02:22Bagderwe should write a wiki about it
23:02:45HClwell, it prints keys when it starts
23:02:47HClbut obviously
23:02:49HClthey're not correct
23:02:56amiconnBagder: That's why the Win32 sim bg pic is useful: It shows the button assigments...
23:03:10amiconn...and will soon be even clickable, I hope
23:04:32*HCl stares at his iriver and has a really memexpensive counter program at the moment :X
23:04:42amiconn[IDC]Dragon: I just had an idea. Maybe on these old units, the MMC presence detection switch reacts earlier, when the card is not yet fully inserted?
23:05:15[IDC]Dragondunno, perhaps
23:06:03[IDC]Dragonbut the idea is (while pulling out) to have the switch trigger before the contacts go loose
23:06:12amiconnTrevmar: Do you currently have a not MV and hotswap enabled .ajz on the internal?
23:06:49TrevmarI couldn't compile that version because of unsatisfied dependencies
23:07:03HCligh, the simulator isn't building the gnuboy sourcetree for some reason..
23:07:03amiconnNo I mean both features disabled
23:07:36Trevmaryes
23:07:38amiconnIf so, please do the following test: (1) Insert the card, while the Ondio is switched off. (2) Switch it on while holding LEFT, thereby starting archos firmware and having it loading rockbox from the internal
23:07:58amiconnDoes the access to the card work then?
23:08:00HCldoes the sim build have a different makefile or what? :/
23:08:11BagderHCl: yes
23:08:30HClBagder: where? for the plugins?
23:08:45Bagderwe changed the build process a while back and we haven't yet made the sim builds follow that style
23:08:52HClmhm..
23:09:07TrevmarPlease calrify what access: "Does the access to the card work then?"
23:09:12Trevmarclarify
23:09:34amiconnI mean, if booted that way, do you get the card's contents displayed within rockbox?
23:09:55Trevmaryes, but that works with the flashed version too, as long as I diable MV and Hotplug
23:10:11amiconnYes of course.
23:10:23amiconnBooting that way is still different,
23:10:45amiconnbecause the archos fw in flash does *always* boot from the internal flash.
23:11:30amiconnThat means, the internal is selected when rockbox kicks in, and rockbox properly manages to switch selection to the external.
23:11:38Trevmarright, but I noticed no difference between the flash boot and coming up via archos over the last few days of testing
23:12:00HClBagder: so where can i find the makefile that controls the build of the plugins for the simulator?
23:12:04TrevmarI tried the archos left arrow boot on just about every software build I did
23:13:05BagderHCl: uisimulator/x11/Makefile builds all plugins in the plugin dir
23:13:26HClok
23:13:30 Join bagawk [0] (~Lee@bagawk.user)
23:13:44amiconnTrevmar: Hmm, I'm a bit confused. Do you also get the panic with multivolume if the card is already inserted at boot?
23:14:38 Join preglow [0] (thomj@s183a.studby.ntnu.no)
23:14:43preglowwassyp
23:14:50*HCl needs to read up on make files :/
23:14:54HClwell.
23:14:57HClgnuboy runs.
23:14:58HClbut
23:15:01Trevmarthat's when I get the panic, with the 07 08 units with MMC card at boot. The latest daily builds of the 03 08 dont show me any card, if I recall correctly
23:15:03preglowyes, i saw
23:15:08HClok :/
23:15:17preglowread the logs
23:15:19HCl*nods*
23:15:22preglowbut that's not bad by itself
23:15:29ripnetUKgnuboy RUNS?
23:15:30preglowand i see linus has done some speed improvements
23:15:42 Join Patr3ck [0] (~patr3ck@pD9E5C239.dip.t-dialin.net)
23:15:54preglowenabling the instruction cache is pretty clever
23:16:22HClripnetUK: pretty much?
23:16:27HClanyways
23:16:31HCli hadn't finished my but
23:16:31ripnetUKnice work
23:16:48ripnetUKit would be funny to get a gb emu before mp3 playback :)
23:16:51HClthe but is that either my routine to use the lcd buffer messed up
23:16:57HClor its just toooooooo slow
23:17:02HClwhich i wouldn't be suprised of
23:17:04HClsince at the moment
23:17:05preglowHCl: does it manage to draw anything?
23:17:18HClits doing about 2-3 seconds for a single emulation loop cycle
23:17:23ripnetUKlol
23:17:27preglowwell, it IS slow at the moment
23:17:31HClso yea.
23:17:32preglowis this with linus latest updates, btw?
23:17:36HClhm?
23:17:41HCli haven't really checked out today
23:17:42HCli think
23:17:47preglowinstruction cache enable, lcd wait state removal
23:17:51HClno.
23:17:54HCllet me update
23:17:56preglowthen hell, do so
23:18:32Bagderand moving the frame buf to iram
23:18:58HClBagder: i'm not having a framebuffer in gnuboy at the moment
23:19:11Bagderok
23:19:12HCli wrote a function thats mapping them onto the lcd_framebuffer
23:19:16HClor at least, it should.
23:19:27HCli'm still not sure about the palette stuff.
23:19:32Bagderlcd_framebuffer is the framebuffer
23:20:09Bagderdon't you use lcd_update() ?
23:21:16preglowwell, he wouldn't have gotten anything at all if he didn't use that, or?
23:21:34hubblebtw, anyone else looking at i2c and i2s interfaces for audio codec?
23:21:34HCli use the lcd_updaterect
23:21:39HClsince i thought it'd be more efficient
23:21:44Bagderit is
23:21:45HCli can revert to lcd_update
23:21:50Bagderbut that uses the frame buffer
23:21:58Bagderand moving the frame buffer to iram boosts performance
23:21:59HCli can paste you the function i wrote.. and explain..
23:22:00ashridahHCl: uh. isn't the iriver currently runnign at the lowest clock rate in rockbox atm?
23:22:04HClashridah: yup :P
23:22:15preglowhubble: linus, perhaps
23:22:31preglowhubble: which of them is the dac part connected with?
23:22:44hubblepreglow: i2c for control, i2s for audio data (PCM)
23:22:46*amiconn is rather confused now :-(
23:23:02Trevmaramiconn: sorry
23:23:05preglowhubble: ahhh, that explains it, i wondered how they could transfer audio data fast enough through i2c, heh
23:23:44hubblepreglow: yes, i was wondering too because the UDA1380 manual almost only talked about i2c and it is max 400 kbps
23:24:43amiconncard inserted/removed, 0308/0708, panic, card invisible, MV, hotswap...
23:24:45preglowyes, exactly
23:24:50preglowhubble: that's what made me wonder as well
23:24:55hubblepreglow: anyways, i'm reading about how to program these interfaces from the coldfire =) pretty interesting
23:25:08preglowhubble: i looked at that yesterday myself, but i won't have time to code anything
23:25:44hubblepreglow: ok.. i've been compiling notes.. probably should do a wiki writeup
23:26:35preglowhubble: that would be great
23:26:55preglowbut yes, does the coldfire have i2s support?
23:27:08quelsarukrebooting
23:27:10 Quit quelsaruk ("KVIrc 3.0.1.99 'Realia'")
23:27:29hubblepreglow: sure, two of them
23:28:37preglowgreat, 'cause i looked for that in the coldfire manual yesterday, and somehow couldn't find anything on that
23:28:40preglowheh
23:28:48preglowbut i could swear i'd read it somewhere
23:28:50hubblepreglow: hm.. actually 4 i2s interfaces
23:28:56preglowdamn, not bad
23:29:03preglowaudio from adc is i2s as well, i presume?
23:29:08hubblepreglow: they are called Philips IIS (Sony AIAJ)
23:29:25hubblepreglow: yes
23:29:29preglowahhh
23:29:31preglowi saw that
23:30:55hubbleanother interesting thing is that the coldfire have like 3 S/PDIF interfaces (2 in, 1 out)
23:31:11preglowyes, that comes in handy for the optical ports
23:31:53HClspdif can do 5.1.. right?
23:31:55hubblenot sure if they are connected to the external port.. it could also be the audio codec..
23:32:26hubbleI think someone said that the external ports connected to the codec which only handles 2 channels
23:32:34hubbleso no 5.1 :(
23:32:43HCl :/
23:33:04hubblethat would be cool HW mod to get 5.1 output =)
23:33:24Bagderwould go nice with the video plugin ;-)
23:33:30HClheh
23:33:54*HCl refueled himself with some food.
23:33:55preglowhubble: you think the spdif ports are connected to the codec?
23:34:06preglowhubble: that doesn't make much sense if you've got on-chip support for it
23:34:21hubblepreglow: dont know, cant tell looking at the board scans
23:34:27hubblepreglow: they are connected somewhere
23:34:35preglowno, i couldn't find out either
23:34:35HCli thought linus said two channel only
23:34:51hubblepreglow: hm.. no.. havn't actually checked spdif pins
23:34:56preglowtwo channel only is no problem if the spdif ports are connected directly to the cpu
23:35:13preglowwe can even do ac-3 passthrough :P
23:37:27*amiconn stares at the mmc code
23:38:17bagawkamiconn, blink once in awhile :)
23:38:18*[IDC]Dragon has drawn a nicer diagram
23:39:06XShocKso, if I now use doubles and floats in rockbox, gcc will emulate them?
23:39:15BagderXShocK: yes
23:39:32Bagderand they are slow
23:39:45Bagderand we have no sprintf() for them
23:39:54XShocKok
23:39:57preglowXShocK: never use floating point in rockbox
23:40:04preglowXShocK: read up on fixed point math and start again
23:40:19amiconnI don't get wth a single select at startup works, but more frequent selects while running don't. At least not why the **** this only happens on some units...
23:40:42ashridahXShocK: you have to emulate them by hand, building them into integers of various lengths
23:40:49XShocKi haven't done anything in floating-points yet. :) I always ask before doing. :)
23:41:01 Quit jpburton5150 (Read error: 54 (Connection reset by peer))
23:41:23Bagderwise move ;-)
23:41:51 Join quelsaruk [0] (~kvirc@85.48.33.190)
23:41:52quelsarukre
23:41:55HClhm.
23:42:05HClis it common for .rock plugins to get smaller for the simulator?
23:42:26ashridahXShocK: take a look at: http://www.bookofhook.com/Article/GameDevelopment/AnIntroductiontoFixedPoin.html
23:42:36ashridahthat'll give you enough knowledge to at least find detailed stuff
23:42:53[IDC]Dragonhttp://joerg.hohensohn.bei.t-online.de/archos/OndioClockGate.jpg
23:43:07[IDC]Dragonamiconn: see above
23:43:22HClwoot. segmentation fault.
23:43:32HClthe simulator kind of crashes roughly on rockboy
23:43:40XShocKthanks
23:44:23hubblepreglow: wierd.. looking at EBUOUT2 (spdif #2) which is pin P9, looks like it goes to the DRAM.. EBUOUT2 is also GPIO37 so it can be used for anything
23:44:34TrevmarJoerg - you call that nicer -LOL (just kidding - its excellent)
23:45:06BagderHCl: that's usually a good indication of badness in the code ;-)
23:45:25Bagderit'll be a lot harder to track down when it crashes on target
23:45:36amiconn[IDC]Dragon: Your idea about the changed bridge behaviour coincides with the changed polarity of the bridge activity indication. I really think those old units use AU9330 instead of AU9331
23:46:33TrevmarIs that a CPU code or something I can look for on them?
23:46:47amiconnThat's the USB-MMC bridge
23:47:02amiconnAU9330 is older, and iirc handles MMC only.
23:47:06HClBagder: i'm sort of getting the feeling its its static buffer
23:47:07HCls
23:47:26amiconnAU9331 is newer, and handles both MMC and SD (if you manage to fit an SD card)
23:47:35BagderHCl: gdb is your friend now ;-)
23:47:41*HCl sighs.
23:47:43HClok ok.
23:47:45Trevmarmy scrapped board says AU9331
23:47:56amiconnIs that a 0308 unit?
23:48:00Trevmarcant see the others easily. DO you want me to look
23:48:30 Part lImbus
23:48:35Trevmarno, that is one ofthe 07 08. Let me open up an 03 08
23:48:44amiconnI'd expect 0708 having AU9331, 0308 having AU9330
23:49:09 Join lImbus_ [0] (~manuel@kernel.cycos.net)
23:49:16HClwhy the hell does it crash on this ? o.o...
23:49:24HClplugin_start (api=0x8072400, parameter=0x0)
23:49:24HCl at /home/hcl/rockbox/rockbox-devel/apps/plugins/rockboy.c:77
23:49:24HCl77 shutdown=0;
23:49:29HClshutdown is a variable o.o;
23:49:36HCldo all global variables need to be static?
23:49:40lImbus_mhmm. cgi::irc only /part's me. instead of /quit when closing :-/#
23:49:54[IDC]Dragonamiconn: another thing I noticed while drawing the circuit:
23:49:55TrevmarIt says 9330D4
23:50:21Trevmar9330D04-UXL
23:50:23[IDC]Dragonin essence, the clocks are ORed together
23:50:33amiconnTrevmar: As expected....really interesting!
23:50:37[IDC]Dragonbut the internal MMC can be "muted"
23:50:56[IDC]Dragonthe external doesn't need disabling, it is just absent
23:50:57 Nick lImbus_ is now known as lImbus (~manuel@kernel.cycos.net)
23:51:15amiconn[IDC]Dragon: Yes, that's quite obvious from your schematics
23:51:25*HCl sighs.
23:51:53amiconn[IDC]Dragon: The AU9330/AU9331 should be added to the wiki, along with the datasheet for the 9330
23:52:05HClfor anyone who wants to try/play/look; there's a zip on my ftp with the rockboy source.. ftp://titania.student.utwente.nl/
23:52:07Trevmarbut it works fine with my Kingston SD cards and the Kingmax MMC
23:52:27[IDC]Dragondo we have the datasheet?
23:52:30preglowbattery.dummy?
23:52:44amiconnOkay, so I remembered wrong, and the 9330 also supports SD
23:53:02TrevmarJoerg and amiconn - sorry that chip is even a different package - i should have noticed :(
23:53:55preglowhah
23:54:00 Quit lImbus (Read error: 104 (Connection reset by peer))
23:54:03preglowrockbox on iriver is a hell of a lot more responsive now
23:54:13Bagderneat
23:54:20 Join lImbus [0] (~manuel@kernel.cycos.net)
23:54:21pregloworders of magnitude better
23:54:40preglowif this can be done with just instruction cache and no lcd wait states, clocking correctly should work wonders
23:55:01preglowbut yes, enabling the instruction was probably very decisive
23:55:08 Join amiconn_ [0] (~jens@pD95D10E9.dip.t-dialin.net)
23:55:27 Quit amiconn (Nick collision from services.)
23:55:27 Nick amiconn_ is now known as amiconn (~jens@pD95D10E9.dip.t-dialin.net)
23:55:38amiconnTrevmar: 44-pin LQFP vs. 28-pin SSOP
23:56:10bagawkif Compact Flash is using an IDE interface...
23:56:39bagawkTHen thereticly, you could make a cf-ide cable and hook a hd into a cf slot right?
23:56:50bagawk(IKNow know it works vice-versa)
23:57:24TrevmarJoerg, I just opened up the unit with the dead USb - and it is a 9331
23:57:46quelsarukBagder: 2 quick things, first of all... is the firmware smaller than the one 1 week earlier? at least i just built one 1KB smaller, and second thing, the initial "resume question" use default font instead of custom one, can that be changed?
23:57:56*HCl goes to build the new rockbox...
23:58:34Bagderquelsaruk: it could be smaller, I don't know. No, I believe it hasn't loaded the custom font at that point
23:58:46quelsarukok
23:58:47quelsarukthanks
23:58:48quelsaruk:)

Previous day | Next day