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).

#rockbox log for 2005-01-28

00:04:30 Join Stryke` [0] (
00:05:03preglowcharging is actually software controlled for some players?
00:06:21preglowwow, i'd have guessed that was a hardware thing, mostly
00:09:17preglowdoes anyone know how the h1x0 players manage that yet?
00:11:25preglowi'm also thinking of getting mine a new and higher capacity battery, anyone heard something that suggests that won't be a good idea?
00:11:51LinusNthe H1x0?
00:12:31LinusNi heard that some people have problems fitting the new batteries in the casing
00:12:43preglowyes, that would of course be an issue, hehe
00:14:30***Saving seen data "./dancer.seen"
00:16:01 Join QT_ [0] (as@area51.users.madwifi)
00:22:55amiconnZagor: Yellow build for win32 sims...
00:23:40Zagoryeah, apparently win32 doesn't have strcasestr().
00:24:41amiconnIt must have it, since otherwise linking would fail...
00:25:08amiconnA propos win32 sim: I looked into the LITTLE_ENDIAN issue with cygwin. This is really strange, but the LITTLE_ENDIAN thing is relatively easy to work around.
00:26:05amiconnHowever, if compiling the Win32 sim with cygwin, netinet/in.h doesn't seem to exist....
00:26:20amiconnIt does work if building the X11 sim with cygwin though...
00:26:45Zagorright, it should use something other than htonl. that was just a quick hack.
00:27:19Zagori'm off to bed now, though. we'll look at it tomorrow.
00:27:28amiconnAhahhh! When compiling the sim now with cygwin, the build *fails*
00:27:33 Quit QT (Read error: 110 (Connection timed out))
00:27:43 Quit Zagor ("Client exiting")
00:39:07quelsarukleaving, cu
00:39:57 Quit quelsaruk ("KVIrc 'Realia'")
00:42:15 Join einhirn [0] (
00:54:03 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
01:04:41 Join bagawk [0] (Lee@bagawk.user)
01:05:26bagawkhey amiconn, could you do a quick favor for me?
01:08:20bagawkGo to and tell me if you get an apche confimation page
01:08:37preglowi certainly do
01:08:46amiconnbagawk: yup
01:09:01amiconnThe page is even localized :)
01:09:21 Join johntb [0] (~Baumans@
01:09:26johntbjoin #freeciv-dev
01:09:29 Quit johntb (Client Quit)
01:10:03bagawklocalized? In what way?
01:10:53preglowapaches default page comes in several languages
01:11:19preglowand your browser usually requests one in your language, if you've set it up that way
01:11:33bagawkI was making a little ftp server on my machine so I could keep my files in one place
01:12:06bagawknow to get dyndns going :)
01:12:23 Join elinenbe_ [0] (
01:20:45bagawkdoes work?
01:23:36amiconnErm, yup
01:23:55bagawkz is a little far away for u or e
01:24:03bagawk(on a QWERTY atleast(
01:24:25amiconnNo, it's not. The german keyboard has y and z swapped compared to english keyboard
01:25:27amiconnIt seems I should get some sleep now.
01:25:53 Part amiconn
01:26:08 Quit bagawk ("Leaving")
01:32:27 Part LinusN
02:05:39 Join elinenbe1 [0] (~elinenbe_@
02:07:47 Quit elinenbe (Read error: 54 (Connection reset by peer))
02:07:48 Nick elinenbe1 is now known as elinenbe (~elinenbe_@
02:13:16 Quit preglow ("fisle")
02:14:34***Saving seen data "./dancer.seen"
02:14:43 Quit mecraw ()
02:25:24 Quit edx (Read error: 110 (Connection timed out))
02:28:23 Join jipi [0] (
02:32:03 Quit jipi (Client Quit)
02:42:26 Join DarthWufei [0] (
02:42:40DarthWufeiI think my Archos died. :/
02:44:17DarthWufeiIt was charging and then I noticed it had some error on it, didn't quite catch what it was, but the unit wouldn't shut off. So I had to remove the AC adapter, it shut off, and then wouldn't turn on. It turns on witht he adapter, and charges, but when I try to load Rockbox.. it stops about 3 steps in and just sits there.
03:03:33DarthWufeiOh whoa it works now, I guess the battery got super low. :X
04:05:46 Join edx [0] (
04:14:18 Quit MooMaunder (Read error: 60 (Operation timed out))
04:14:35***Saving seen data "./dancer.seen"
04:14:48 Quit Shulberry (Read error: 60 (Operation timed out))
04:29:09 Quit gromit` (Remote closed the connection)
05:03:51 Join Shulberry [0] (
05:09:32 Join lostlogic [0] (
05:35:28 Quit DarthWufei ()
06:14:37***Saving seen data "./dancer.seen"
07:12:25 Quit Shulberry (Read error: 60 (Operation timed out))
07:23:09 Join methangas [0] (
07:24:04dwihnoGood morning everyone
07:33:12 Quit midk (Read error: 104 (Connection reset by peer))
07:35:37 Join Shulberry [0] (
07:43:04 Join midk [0] (
07:43:24 Join gromit` [0] (~gromit`
07:57:36 Join dwihno_ [0] (~dw@
07:59:37 Quit dwihno_ (Client Quit)
08:00:15 Join dwihno_ [0] (~dw@
08:07:20 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
08:14:39***Saving seen data "./dancer.seen"
08:15:42 Quit dwihno (Read error: 110 (Connection timed out))
08:44:26 Nick QT_ is now known as QT (as@area51.users.madwifi)
09:05:48 Join Zagor [242] (
09:47:56 Join lImbus [0] (
09:48:19 Join Schnueff [0] (~mah@
09:48:40lImbusHi, I just finished very successfully my car power supply. somebody interested in pictures/shematics ?
09:49:25lImbusit's mainly a question to the core devs if it's worth to get on the wiki.
09:54:42ZagorlImbus: what exactly is the mod?
09:57:19lImbusit's just a L4940 very low drop regulator with two capacitors
09:58:13lImbusI ask because I knew it couldn't be complicated, but I didn't dare to do it, until this morning
09:58:36Zagorno, i mean what does it do?
09:59:00lImbusuh. it gives power to my archos recorder, so I can charge it while driving
10:00:21Zagorrecorder v2?
10:00:31lImbusnope, v1
10:02:35Zagorok. well sure, throw it up in the wiki.
10:03:39lImbusI'll lik it within docs->hardware and a note in the battery and charging FAQ, if I manage it :-)
10:04:25Zagorsounds good
10:14:44***Saving seen data "./dancer.seen"
10:31:12 Join LinusN [0] (
10:33:34 Join MooMaunder [0] (~me@
10:34:11lImbusuhhh. took my a while to understand that the left side link "Doc" on the main page is NOT the same than the DocIndex-Page in the wiki...
10:34:53lImbusHÄ ? now it is ?!?
10:35:12lImbusok. my fault
10:35:14Zagorit is
10:35:31lImbusI picked the old "" from my browsers proposal...
10:35:49lImbusit better should redirect into wiki then.
10:36:03lImbusI got that old url in a bookmark as well.
10:40:43 Part LinusN
10:47:39 Join LinusN [0] (
10:48:14*LinusN is installing the bootloader via iriver's own firmware upgrade function
10:48:39*lImbus is clapping and bowing.
10:48:59dwihno_Heja Linus, friskt humör! Skjortan hänger utanför! :D
10:49:15LinusNwell, the boot loader isn't finished
10:49:28LinusNthe loading itself isn't implemented
10:50:01dwihno_Well, the skjorta still is inside the pants then :)
10:50:06LinusNbut i made the tool that inserts the boot loader into the original firmware
10:50:07dwihno_But still, friskt humör!
10:50:37LinusNso now i can select which code to run by holding the REC button when turning it on
10:51:27LinusNonly thing left is to make ROLO understand the new unscrambled file format
11:00:08izzy_sounds really good :)
11:06:04 Join einhirn [0] (
11:15:08 Join amiconn [0] (
11:32:49amiconnZagor: r u there?
12:08:51 Part LinusN
12:14:48***Saving seen data "./dancer.seen"
12:22:12 Join LinusN [0] (
12:24:03 Join ripnetuk [0] (
12:26:26Zagoramiconn: here now
12:27:23amiconnZagor: I checked my cygwin installation - there is no strcasestr()
12:28:42lImbusZagor: btw, I'm just done with that wiki article about my circuit
12:29:49ripnetukiRiver lookin' good :)
12:31:52 Join kurzhaarrocker [0] (
12:32:26LinusNlImbus: woo, long paragraphs!
12:32:59lImbusI just wanted to explain to the people what I did not understand before. Shorten as you like :-)
12:33:21Zagori propose renaming the subject to CarPowerSupply aswell
12:34:03lImbusthe thing is I have to hurry a bit now. it took longer than expected. do you want me to do it this evening ? or is it so important that you want to fix that now ? :-)
12:34:14lImbusI dont even know how to rename topics
12:34:25Zagorit's not important, don't worry
12:35:04lImbuskk (->goes to my todo list, right after, lemme see, ummm, well, right after coming home from office this evening :-)
12:35:05Schnueff`More' -> `Rename / move topic'
12:35:16lImbusahh, danke
12:36:14kurzhaarrockerot: Does anybody here know / use a freeware tool for making use case diagrams?
12:36:15LinusNlImbus: i wish i have such a short todo list :-)
12:37:36lImbusLinusN. Mine is very long, but I assume it's not hard to render this topic nice, so...
12:39:28lImbusuhh. Is Christi in here ? She locked my topic, already squashing out the bugs, I suppose.
12:39:37lImbusbug == typos, of course
12:40:53 Part kurzhaarrocker
12:41:05 Quit lostlogic (
13:05:56 Join jyp [0] (
13:05:56NJoinlostlogic [0] (
13:10:30 Quit lostlogic (
13:10:30 Quit jyp (
13:10:52NJoinlostlogic [0] (
13:20:45*ripnetuk claps
13:20:55LinusNi just booted rockbox with the rockbox boot loader
13:20:58ripnetukhave you rolo'ed a exe then :)
13:21:04ripnetuknice one
13:22:40LinusNi created a tool for inserting our loader into an iriver firmware
13:22:47LinusNin tools/
13:23:05LinusNZagor: i called it "mkboot", is that ok?
13:25:46dwihno_really cool linus!
13:26:13*lImbus gives out a beer or two to LinusN... or three
13:26:41lImbusgotta run
13:26:42lImbuscz later
13:26:42ZagorLinusN: sounds good to me
13:26:48 Quit lImbus (" HydraIRC -> <- irc client ownage!")
13:33:24ripnetukAre you going to provide a pre-built binary of your boot loader so we can use your mkboot to inject it into a iRiver firmware? im not sure I want my first test of my m68k cross compiler to be the boot loader :)
13:34:02ripnetukalso, if it goes wrong, how many points do you have to solder to get a wiggler working to repair ?
13:34:41 Quit midk (Ping timeout: 14400 seconds)
13:35:14LinusNripnetuk: 13 points
13:35:24LinusNplus a resistor
13:35:48LinusNyes, i can provide a boot loader image
13:36:10LinusNbut, i strongly suggest you wait a little until zagor and bagder have tried it
13:36:50LinusNi can help them restore their flash if anything goes wrong
13:37:06 Quit lostlogic (
13:37:17LinusNi want to include MiniMon in the loader before releasing it to the public
13:37:41LinusNmuch less soldering if something goes wrong
13:37:48NJoinlostlogic [0] (
13:39:18ripnetuki assume minimon is a little debugger that works on the serial port@?
13:39:34ripnetukso if it went wrong we could use a max232 or similar to connect it to a pc?
13:41:30ripnetukwell im going to build myself a cross compiler using your latest instructions then (i built one before, but that was AGES ago)
13:41:42ripnetukare the instructions on the rockbox site up to date?
13:42:21Bagderyou rock LinusN
13:47:30BagderLinusN: FIRMWARE_OFFSET_FILE_DATA should be 8, not 6 right?
13:47:58Bagderor am I missing something?
13:48:45LinusNi am a silly person, yes it should
13:48:57Bagderand the crc offset should then be 0
13:49:16LinusNbrain damage on my part
13:49:27Bagdercommit mails are great for review
13:55:16Zagori intend to start working on the new sound system today
13:55:38amiconnSound system?
13:55:53Zagorthe new architecture, to support software codecs
13:56:09amiconnAh ok.
13:56:13 Join [IDC]Dragon [0] (
13:56:41 Join kurzhaarrocker [0] (
13:57:07kurzhaarrockerWill that new sound system take peak-meter like info into account, Zagor?
13:57:26*[IDC]Dragon saw some success in the log
13:57:34amiconnZagor: Btw, I think the comressed file buffering will buffer the files straight, i.e. without bitswap. Imho it would be a good idea to do the same for the hw codec boxes, doing the bitswap shortly before the data is needed, in small portions.
13:57:36Zagorkurzhaarrocker: no :-)
13:57:40[IDC]Dragoncongrats, LinusN
13:58:18[IDC]Dragonamiconn: I think so, too
13:58:23Zagoramiconn: the problem with that is it requires double buffers for hw codecs too
13:58:37*kurzhaarrocker cries and throws the peak meter & split editor into the trashcan
13:58:40[IDC]Dragonthe bitswap is a problem of the transmission
13:58:55[IDC]Dragonshoudn't affect the whole system
13:59:00amiconnThis would be more consistent, and also allow to leave the id3 tag info within the buffer, because it won't get swapped.
13:59:20amiconn[IDC]Dragon: Doing this would however require a rework of video and voice ui.
13:59:26[IDC]Dragonand parse the data, to do tricks like ff/fr
13:59:27Zagorumm, why would we want the id3 tag in the audio buffer?
13:59:58amiconnZagor: In order to get rid of the hard limit on the number of tags
14:00:17Zagorwe still need to keep track of the songs in ram
14:01:20amiconnYes, but this could be done dynamically, as a linked list
14:01:39Zagor...which needs to be a fixed size, which means we have a fixed limit.
14:01:56Zagorwhere do you get limitless ram on the archos?
14:02:07[IDC]Dragonthe list is interleaved in the mp3 buffer
14:02:30amiconnWe need exactly *one* buffer for the tag of the currently playing track, because this one can get overwritten while the track is still playing.
14:02:32[IDC]Dragonor rather could/will be
14:03:10Zagorthe primary question remains: why?
14:03:17amiconnThen we have a pointer pointing to the start of the next track. The first longword is a pointer to the next track. Then comes the tag, then the mp3 data.
14:03:50kurzhaarrockeramiconn: so you still need a (limited) list of pointers to id3 tags?
14:03:59LinusNkurzhaarrocker: of course we will have a peak meter of some kind
14:04:11BagderZagor: the point of such a system would be to get rid of some fixed id3 limits
14:04:42Bagderbut it would make it slightly tricker to get next-song info
14:04:56amiconnkurzhaarrocker: No. We'd need exactly *one* pointer, pointing to the tag info of the next track. This one has a poiter to the next etc. (or NULL if there is not (yet) a next track)
14:05:06Bagderpossibly it would also be tricky when the id3 data wraps in the buffer
14:05:10 Quit mbr (
14:05:32NJoinmbr [0] (
14:05:57amiconnBagder: I already thought about this. We could simply set a rule that id3 tag info never wraps, i.e. if it doesn't fit at the end, wrap *before* the tag
14:06:15kurzhaarrockerI understand, amiconn, something like a linked list only the links are generated on the fly
14:06:22amiconnid3 tag info is fixed size, so no problem.
14:06:49Bagderamiconn: if the id3 tags contain album arts etc, that could waste a good bunch of valuable ram
14:06:56Zagorso, we add complexity and decrease runtime. all to be able to fit >16 tracks in 2MB ram?
14:07:14LinusNZagor: or 32
14:07:41amiconnBagder: I didn't think about storing the tags as-is (this would really be a problem), but the tag info as currently used in rockbox. This is fixed size.
14:07:44Zagoririver can have 128 id3 slots
14:07:56Bagderamiconn: ah, ok. I misunderstood
14:08:35amiconnStoring as-is would also be problematic because id3v1 is located at the end of the file, so preprocessing is necessary anyway
14:11:05amiconnkurzhaarrocker: yes. When a new track gets loaded the pointer to the next is NULL at first. If then another track is loaded, the pointer of the former one gets updated to point to the newly loaded track etc.
14:12:04amiconnThis requires walking the list once each time a new track gets loaded. This list walking is actually fairly simple
14:14:50***Saving seen data "./dancer.seen"
14:14:58dwihno_Just a question... how many tracks does the current rockbox code keep in buffer? as much as needed or max 2?
14:15:55LinusNmax 16
14:16:26LinusNthen it becomes crazy
14:16:35BagderI like amiconn's suggestion
14:16:41LinusN(see the bug report about the norwegian language course)
14:16:59Bagderbut I'm not mr mpeg.c :-)
14:16:59LinusNcrazy == it repeats every 16th song
14:17:09dwihno_Well, 16 tracks should be enough for everyone :)
14:17:14ZagorBagder: i think it costs much more than it gains
14:17:46kurzhaarrocker"640kb should be enough for everyone" :)
14:17:51LinusNwe still need to handle stuff like album art etc
14:18:22Bagderand that makes me a bit scared ;-)
14:19:12ripnetukis there a lag between booking stuff into cvs and being able to book out? i wanted to have a look at the new bootloader committed at 12:51 by Linus, but is doesnt seem to be in the cvs. Ive found a bootloader dir, but (flash/bootloader) but i think thats the Archos one
14:19:20*[IDC]Dragon whispers "malloc"
14:19:27[IDC]Dragonand runs off
14:19:32LinusNripnetuk: cvs co bootloader
14:19:49ripnetukim doing rockbox-all does that not cover bootloader yet?
14:19:52ripnetukthanks btw
14:19:57Bagderyou guys seen the latest openneo commits btw?
14:20:05LinusNno, rockbox-all doesn't contain the boot loader
14:20:09LinusNBagder: yes
14:20:15[IDC]DragonI don't monitor it
14:20:26[IDC]Dragonwhat's it about?
14:20:30Bagder"max number of random songs buffer"
14:20:45ripnetukok... booked it out... thanks
14:20:45Bagderthey have removed our shuffle
14:20:55Bagderand keeps a buffer of last-played songs
14:21:08Bagderfunny choice
14:21:12Zagorthat's amazingly silly!
14:21:19amiconnLinusN: The album art etc should be handled the same way as now. I didn't propose to buffer the *tags* as-is, but to simply buffer the id3tag structure within the mp3 buffer instead of a separate array.
14:21:55LinusNbut i guess they have implemented whole-disk random without playlists
14:22:07dwihno_the rockbox shuffle is really, really smart! me loves it
14:22:18BagderI think our shuffle rocks
14:22:35amiconnThat's why it's called rockbox? ;-)
14:22:37LinusNthe only thing it can't handle is whole-disk random without playlist
14:22:48kurzhaarrocker(since Linus fixed the random)
14:23:12[IDC]Dragonwe could use a more simple PRNG, though
14:23:37[IDC]Dragongives you memory for your song buffer ;-)
14:24:47 Part kurzhaarrocker
14:25:03LinusNa faster random will be needed though
14:25:27LinusNfor dithering
14:25:51Zagornot necessarily the same function though
14:25:57Bagderit can probably also be less good
14:26:03[IDC]Dragonamiconn has one i the grayscale bib
14:26:12LinusNwe could take that one
14:26:33LinusNi believe it's the same one bluechip have been proposing all along
14:27:20[IDC]DragonI think it's even more simple
14:27:52amiconnLinusN, [IDC]Dragon: While the PRNG within the grayscale lib is very fast, it is certainly not good enough for shuffling.
14:31:36[IDC]DragonI know
14:32:26LinusNthat's why we still use the twister
14:32:28LinusNgotta go
14:32:31LinusNcu later
14:32:37 Part LinusN
14:34:04 Join ashridah [0] (
14:36:18Bagderthe iriver ifdef in firmware/SOURCES is not that good, imho
14:38:06Zagorawfully much duplication
14:38:29ashridahhm. iriver boot loader checked into cvs eh? complete? shell? highly explosive?
14:38:31Bagderbut I guess this is only for now
14:40:16ripnetukWhen the cross compiler instructions say book binutils out of cvs, do I need the one listed for calmRISC16? i cvs -z3 co binutils-2.15 ? or do i need a more recent version?
14:40:25ashridahmy eyes!
14:40:48 Join LinusN [0] (
14:40:56*ashridah tosses mad props to LinusN
14:40:59LinusNthe SOURCES is temporary
14:41:08Bagderripnetuk: that would work
14:41:21LinusNi just didn't like having everything checked out all the time
14:41:33ripnetukthanks Bagder
14:41:42Bagderripnetuk: when you install the binutils, use 'make -k install'
14:41:55Bagderthe -k being the important part
14:42:09ripnetukok, thanks... what does -k do?
14:42:25Bagderat least for me, the install failed on some non-important subdir and -k makes it continue with the other install things
14:42:27LinusN-k tells make not to stop on errors
14:42:29Bagderwithout -k, it stops
14:42:29ripnetukand do i need it both for the original non cvs install and the cvs one?
14:43:00Bagderthis was for the calmrisc16 binutils from CVS the other day
14:43:15Bagderit may have been fixed since, I don't know
14:44:28ripnetukok, thanks for the heads up
14:45:50ripnetukdammit the cvs is asking for a password when I try to check out binutils. Im doing cvs login
14:46:01ripnetukbtw what is gemoss?
14:46:07Bagderpress return only
14:46:17Bagdergemoss is the gmini development tools project
14:46:45Bagderjyp and the guys run it
14:47:08ripnetukok... so the binutils for that is OK for iriver then?... im pressing enter only, and getting cvs [login aborted]: end of file from server (consult above messages if any). Is this just a sourceforge crappy cvs issue?
14:47:36Bagderno that is for Gmini dev
14:48:32Bagderfor iRiver, you need a coldfire setup
14:51:45Bagderwhich needs bintutils from the official binutils cvs ;-)
14:52:10BagderI added a description ot the CrossCompiler wiki
14:52:33ripnetuki will get on with building the non cvs one for now...
14:55:42ripnetukthanks Bagder - thats checkout out (from Redhat) ok now
14:56:37Bagderour development environments are not the easiest accessible ones ;-)
14:56:46ripnetukthats part of the fun
14:58:22ripnetukbinutils seems to require flex - might be worth noting on the wiki
14:58:28ripnetukthank god for urpmi :)
14:58:49Bagderfeel free to update
15:00:02*Bagder runs off
15:03:41ripnetukDoes Bagder== Daniel Stenberg?
15:08:26ripnetukok... just checking as it said the wiki was locked by him... i will go ahead anyway as he just invited me to alter it
15:08:39Zagordo so
15:09:26ashridahso has anyone else tried linusn's iriver bootloader? :)
15:10:03Zagorno, not yet
15:10:13ripnetuknot yet... he advised we wait until Badger tried it first, as (i assume) they live close enough to use the wiggler to fix it is it goes wrong
15:10:19ripnetukim well tempted to just try it tho...
15:10:53ashridahheh. i'm not that desparate myself. don't really have the disposable income to test yet :)
15:11:49ripnetukyeah, me neither...
15:11:55*ripnetuk remembers PaulS's situation
15:12:25ripnetukhe tried to hack the original iRiver firmware to load code from disk
15:12:47ashridahaah, that's right
15:12:55ripnetukby replacing the re-flashing code in the stock firmware... it went wrong, and now he cant flash anything, (but the player works!)
15:12:57ashridahand managed to render it unable to flash itself again or something?
15:13:05ripnetukthat must suck :(
15:23:01ripnetukalmost there... im getting a command no not found... i assume no just returns a non-success error code... anyone help me lash up an alias that implements this?
15:23:26ripnetukthis is building the cvs binutils. The original one built and installed no probs
15:29:32 Join thegeek [0] (
15:32:54thegeekhmm, has badger tested the bootloader yet?
15:38:08 Quit Schnueff ("leaving")
15:38:22 Part lostlogic ("Kernels are for popping")
15:47:00 Join thegeek_ [0] (
15:47:05 Quit thegeek (Read error: 104 (Connection reset by peer))
15:53:37 Quit thegeek_ ()
15:58:45 Quit ashridah ("sleep")
16:02:14amiconnZagor: Could you please fix the simulator issue? Otherwise I can't use the sim. The official build produces a warning, however I get a linker error.
16:03:21Zagori'm not sure what the proper fix is. does win32 use stristr instead of strcasestr, or do we need to include strcasestr.c in the win32 build?
16:04:14Zagor(unrelated: check out this player with user-swappable 2.5" harddrive.
16:04:34Zagorlooks fragile
16:09:20 Join thegeek [0] (
16:09:43Zagorthegeek: no, nobody other than linus has tested the bootloader yet
16:10:06thegeekI'll probably wait a little then;)
16:10:31thegeekgreat work btw LinusN:)
16:11:15amiconnZagor: Btw, same problem with building the X11 sim on cygwin (linker error).
16:11:38Zagorcygwin is weird :)
16:13:35Zagorneuros pocketplayer will support WMA+DRM, *and* be open source. wanna bet not all source will be open...?
16:14:21Zagordapreview makes your head spin. so many new player models every week!
16:14:53Zagoramiconn: since I don't run cygwin, could you help me fix this?
16:14:54***Saving seen data "./dancer.seen"
16:16:05amiconnNo stristr() either. I searched all .h files included with cygwin
16:19:15amiconnGotta go, cu l8er
16:19:17Zagori guess the win32 simulator will have to include the strcasestr.c file then
16:19:38amiconnI think so too.
16:19:54amiconnNot only the Win32 sim, but all simulator builds on cygwin
16:20:14Zagorannoying :-(
16:21:41Bagderperhaps all sims could use our own strcasestr then
16:22:19Bagderit isn't a very standard function
16:22:42Zagoryeah, that's the easiest solution
16:26:24[IDC]DragonI'm looking into hotswap again
16:26:31*ripnetuk has my cross compiler for m68k built with cvs and all :)
16:26:45[IDC]Dragonseems the file code needs only little additions
16:26:48 Quit thegeek ()
16:27:19[IDC]Dragonmost of the api functions are bailing out if the file is marked as not busy
16:28:10 Join El_Barto2 [0] (
16:28:26El_Barto2what do you think of the gimini 120?
16:28:28[IDC]Dragonso, I'd only need a function to set all file/dir handles on a geven volume to not busy
16:28:44El_Barto2do you think its better than a jukebox recorder 20?
16:29:06El_Barto2cause there is no alternative soft..
16:29:24[IDC]Dragonhow would you call such a function?
16:29:58Zagor[IDC]Dragon: release_files() perhaps?
16:30:08ZagorEl_Barto2: gmini support is being added to rockbox.
16:30:24El_Barto2so its a better choice right?
16:30:25[IDC]Dragonand release_dirs()
16:30:34El_Barto2an cheeper..
16:31:32[IDC]Dragonthe busy flag may have a reenty problem
16:31:47ZagorEl_Barto2: better is a matter of opinion. they are different.
16:31:48[IDC]Dragonbut I'll save that for later ;-)
16:32:27El_Barto2can i charge it through usb?
16:33:08ripnetukLinus - I just got the iRiver rockbox built, but it built very quick... is it just a shell at the moment? i ended up with a rockbox.bin of size 6718 and rockbox.elf of size 42320
16:33:12[IDC]Dragonactually, the whole file/open closy handle code must not be reentered, as of now
16:34:10*[IDC]Dragon types weird
16:35:13Zagoryes, but we have only very few short windows of potential reentrance. although of course it should be fixed anyway.
16:36:26[IDC]DragonI don't know if there is a window
16:36:41[IDC]Dragononly if we sleep inbetween
16:36:57 Quit Shulberry ()
16:37:07El_Barto2do you know some practival advantage of the gimini 120?
16:37:28El_Barto2vs jukebox recorder 20 ;)
16:37:42[IDC]DragonEl_Barto2: this is rockbox *developer* IRC
16:37:43El_Barto2or disadvantages..
16:37:48El_Barto2ok :)
16:38:09El_Barto2so where can i find the rockbox soft for the gimini?
16:38:16Zagor[IDC]Dragon: you're right, since the fat driver is now mutexed
16:38:34ZagorEl_Barto2: it's not available yet. it's work in progress.
16:38:41[IDC]Dragononly the caching, no more
16:38:59Zagor[IDC]Dragon: ok, then there's probably a window left open.
16:40:13El_Barto2ok..i'll wait ;)
16:40:26El_Barto2but what mp3 player would you prefer?
16:48:21ripnetukbagder - i didnt need the -k when installing the cvs binutils - i can now build m68 binaries on my Linux box :) cant get the bootloader compiled (maybe just as well as the temptation is so strong to try it!)
16:49:36 Join Shulberry [0] (
16:50:11El_Barto2ok...understand...i leave you alone ;)
16:50:22El_Barto2succes for the firmware! ;)
16:50:31 Quit El_Barto2 ()
16:53:10Zagorgotta go
16:53:12 Part Zagor
16:58:31 Join amiconn_ [0] (
17:09:54 Join mecraw [0] (~mecraw@
17:11:30ripnetukcan any c compentent person give me a clue please? im trying to build LinusN's new bootloader - the directory I checked out has a main.c, Makefile and sources in it. I cant figure out how to get it building - make gives me an error, and lookign at the makefile, it seems to rely on $FIRMDIR and soon being set, so i think its supposed to be built as part of a 'bigger' build job. Where should I put the bootloader dir to get it compiling?
17:15:54LinusNripnetuk: it is not ready yet
17:15:59[IDC]DragonI think LinusN didn't commit the full thing yet
17:16:02 Quit amiconn (Read error: 110 (Connection timed out))
17:16:02 Nick amiconn_ is now known as amiconn (
17:16:06LinusNthe configure script is not done
17:16:43LinusNand i suggest that you stay away from the boot loader unless you have a bdm emulator to save you
17:16:45 Join ripnet [0] (
17:17:12ripnetdammit! my linux box has crashed @ home, and i am @work :( thats the end of playijng for today :(
17:17:59[IDC]DragonLinusN: you see how important it is not to commit the last part until fully tested
17:18:19LinusNgotta go, cu around
17:19:12 Part LinusN
17:22:30 Join ripnetuk_ [0] (
17:25:24 Quit ripnetuk_ (Client Quit)
17:26:17 Quit edx ()
17:32:17 Quit ripnet ("CGI:IRC (Ping timeout)")
17:37:12 Join edx [0] (
17:42:57 Join Spida [0] (
17:46:23 Quit Spida_ (Read error: 60 (Operation timed out))
17:47:48 Quit Shulberry ()
17:59:46 Join Atur [0] (
18:14:55***Saving seen data "./dancer.seen"
18:16:43 Join jyp [0] (
18:37:58[IDC]Dragonamiconn: r u there?
18:40:27 Quit mecraw (Read error: 60 (Operation timed out))
18:45:40[IDC]Dragonamiconn: in case you will: where should the mounting/unmounting go, into the MMC thread or into the default event handler?
19:01:43 Join Stryke` [0] (
19:01:59 Quit [IDC]Dragon ("CGI:IRC")
19:03:28 Quit jyp ("poof!")
19:05:14 Quit ripnetuk (Read error: 104 (Connection reset by peer))
19:29:34 Join R3nTiL [0] (~zorroz@
19:33:23 Quit Stryke` (Read error: 60 (Operation timed out))
19:50:44 Join preglow [0] (
20:02:49 Quit elinenbe (" HydraIRC -> <- :P")
20:04:12preglowso, this bootloader is more or less finished and ready for use?
20:10:39 Quit R3nTiL ()
20:14:57***Saving seen data "./dancer.seen"
20:21:11 Join Bluechip [0] (
20:23:48 Join Stryke` [0] (
20:25:04 Join webguest25 [0] (
20:25:35 Quit webguest25 (Client Quit)
20:31:23 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
20:33:18 Join webguest13 [0] (
20:36:28 Join jyp [0] (
20:46:03 Join jyp_ [0] (
20:47:27 Quit jyp ("poof!")
20:47:57 Nick jyp_ is now known as jyp (
20:58:19 Join [IDC]Dragon [0] (
20:59:27pregloware there any never versions of the m68k tools out somewhere? mine seems to puke on crt0.S
21:00:49 Part Bluechip
21:23:00amiconnhi again [IDC]Dragon
21:23:09[IDC]Dragonhi there
21:23:37[IDC]Dragondid you see my msg?
21:23:42amiconnGot your question - I think the mount/unmount should not rely on the foreground thread calling the event handler.
21:24:28[IDC]Dragontrue, it may "eat" it
21:24:49[IDC]Dragonbut I need the event hdl. to return nonzero
21:24:57amiconn..or be called way later, when another MMC is already inserted..
21:25:30[IDC]Dragonperhaps not...
21:25:47[IDC]Dragonto have the file tree react on it
21:25:55amiconnBtw., there is no MMC thread. The MMC driver only uses a tick task.
21:26:25[IDC]Dragonoh, so this is irq context?
21:26:41[IDC]Dragonthen I can't unmount/mount there
21:27:04amiconnYeps. In addition, only short actions should be done there.
21:27:22amiconnWe may have to resurrect the thread...
21:27:47[IDC]DragonI think so, or even both
21:29:31amiconnPerhaps the resurrected mmc thread should react on SYS_MMC_INSERTED/SYS_USB_EXTRACTED, do the mount/ unmount, and broadcast a message itself after doing that.
21:30:03[IDC]Dragona separate one?
21:30:19[IDC]Dragonto be sure there's no race?
21:30:34[IDC]Dragonof mount vs. redraw
21:30:34amiconnSomething like SYS_FS_CHANGED, to trigger a tree reload (and a dir change to the root if necessary)
21:31:44[IDC]Dragonsounds reasonable
21:32:11amiconnMaybe the SYS_MMC_* messages shouldn't be global. This could be done if the queue functions allow posting to a specific queue which is not ours (the USB thread needs them)
21:33:24amiconnHmm, our own thread will also need these messages, so we'd have to post to 2 specific queues.
21:35:38[IDC]Dragonwhat's bad about global?
21:37:09amiconnWell, it's not that bad. We inform many threads about MMC changes which simply don't need the info.
21:37:24amiconn"many" < 10 here
21:42:00amiconnI'm currently doing runtime tests on Ondio with a modified battery_test plugin. The first test (900 mAh NiMHs) ran for 12 hours.
21:42:22amiconnCurrently I'm sacrificing a fresh set of alkalines...
21:44:14[IDC]Dragonmy Ondio has a terrible runtime now
21:44:21amiconnWell, cheap ones from Plus (€ 1.89 for 8 batteries)
21:44:25[IDC]DragonI've messed with the backlight
21:44:39[IDC]Dragonwhich takes ~160 mA now
21:45:00amiconnWhat did you do?
21:45:06[IDC]Dragonno improvement over the ~40 mA
21:45:22[IDC]Dragondifferent coil different capacitor
21:45:40[IDC]Dragonand it's noisy!
21:46:12[IDC]Dragonlesson learned: no ceramic caps for such, they work as piezos
21:46:33amiconnCouldn't you undo the change?
21:46:53[IDC]Dragonit wasn't perfect before, either
21:47:05[IDC]DragonI have some foil caps on order
21:47:18[IDC]Dragonwill change again when I get them
21:47:31amiconnOk, I understand.
21:47:50amiconn900 mAh NiMHs are nice :)
21:47:56[IDC]Dragonaway for some minutes
22:08:29 Join [IDC]Dragon2 [0] (
22:14:59***Saving seen data "./dancer.seen"
22:17:02 Quit [IDC]Dragon (Read error: 60 (Operation timed out))
22:17:54 Nick [IDC]Dragon2 is now known as [IDC]Dragon (
22:34:45 Quit webguest13 ("CGI:IRC (EOF)")
22:43:18 Quit [IDC]Dragon (Read error: 60 (Operation timed out))
22:46:46 Join mecraw [0] (~mecraw@
22:48:21Bagderpreglow: you built the binutils from cvs?
22:48:32preglowno, i did not, that explains it
22:48:53BagderI added the cvs info to the wiki page earlier today
22:51:22preglowreally? what page?
22:52:24Bagderthat is, the info about building the CVS version was already there, I just added the exact cvs command lines needed
22:53:46preglowso i really need to build my own tools from cvs?
22:54:12preglowait, i'll do that then
22:54:20preglowwhat will be simpler building under, cygwin or linux?
22:54:37BagderI think the process is identical
22:54:54preglowyes, sure, i'm just not familiar with home completely unix like cygwin is, i just installed it
22:54:56BagderI personally would use linux
22:54:59preglowhave been using mingw up until now
22:55:11preglowyes, i think i'll use linux as well
22:57:25 Quit gromit` (Read error: 54 (Connection reset by peer))
22:57:40 Join gromit` [0] (~gromit`
22:58:25preglowhas anyone apart from linus himself tried the bootloader, btw?
22:58:38BagderI don't think so
22:58:47BagderI was expecting some kind of docs first
22:59:01preglowi'm way too eager for my own good
22:59:16Bagdersetting up a dev env is good anyway
23:01:13 Join [IDC]Dragon [0] (
23:01:21preglowapart from cvs the build process as described is accurate?
23:01:45Bagderits the same for all the cross compilers
23:01:55Bagderjust different targets
23:02:54amiconnHere :)
23:03:07[IDC]DragonI guess I have to avoid trying mout/unmount in USB mode ;-/
23:03:40amiconnOf course. The MMC thread must do the same as the ata thread for the jukeboxes
23:04:16amiconn...acknowledge USB and then wait until USB is extracted again, ignoring any SYS_MMC* events meanwhile
23:04:26*[IDC]Dragon looks how it does
23:05:11amiconnI plotted my first NiMH discharge curve now. There is a strange "jump" in it.
23:06:11amiconnPerhaps this is caused by the LTC3440 changing its switching mode
23:06:30[IDC]Dragonah, maybe
23:09:28preglowi like this, configure actually writes a makefile, but it contains syntax errors
23:10:58 Quit Atur (Remote closed the connection)
23:11:30[IDC]Dragonlooks nice
23:11:43[IDC]Dragonif you smoothen it a bit
23:12:40[IDC]Dragonlow level code for hotswap is ready
23:13:00[IDC]Dragonnow I need to teach tree.c what to do
23:13:10[IDC]Dragonor. filetree.c
23:14:03amiconnWell, basically filetree.c needs to react on a fs_change message in 2 cases:
23:14:31amiconn(1) If currently in the root, reload the root to reflect the change.
23:15:11amiconn(2) If currently in a sub-dir that was on the removed MMC, jump back to the root
23:15:13[IDC]Dragonyes, that's clear
23:15:41amiconnOtherwise no action is required
23:16:53amiconnThe default event handler should return nonzero for the fs change message. It doesn't need to do anything
23:17:16[IDC]Dragonhow about the callback?
23:17:17 Join lImbus [0] (
23:17:23lImbusmoin moin
23:19:04[IDC]DragonSYS_FS_CHANGED could be handled like a button, why bother the default handler?
23:19:21amiconnI just wanted to suggest the same :)
23:19:47amiconnI think there is no default action for fs change, it depends on the context
23:20:09amiconnfiletree.c might be the only one actually interested in this event.
23:20:20[IDC]Dragonthe UI part ii in tree.c
23:20:38[IDC]Dragonwhich makes it hairy
23:20:45amiconnPerhaps an upcoming norton commander plugin could use it too
23:20:52[IDC]Dragonit's shared with the db browsing
23:21:04[IDC]DragonZagor, help!
23:22:58preglowBagder: yes, seems linux will be the way to go for me
23:29:20amiconn[IDC]Dragon: There is a "bool id3db" in dirbrowse()
23:30:26preglowall i need to pull from cvs is binutils, yes?
23:30:41[IDC]Dragonamiconn: thanks
23:30:42 Join Zagor [242] (
23:30:46Zagoryou called?
23:30:59[IDC]DragonJini out of the bottle
23:31:06Bagderpreglow: but that's quite a lot ;-)
23:31:25lImbusZagor: I just need that script ^^
23:31:34amiconn[IDC]Dragon: Either our alkaline discharge is totally off, or my Ondio has a different battery scale factor than yours. It's now running for >4 hours, but still showing 91% battery
23:31:42[IDC]DragonZagor: as you may have seen, I "only" have to make the browser react on a new sys message
23:31:46 Join Atur [0] (
23:32:02 Join mecraw_ [0] (~mecraw@
23:32:07[IDC]Dragonamiconn: that curve was from a textbook
23:32:12Zagor[IDC]Dragon: right. that should be tree.c. file- and dbtree don't read sys messages
23:32:14[IDC]Dragonno real measurement
23:32:35[IDC]Dragonbtw, hotswap works!
23:33:07amiconn[IDC]Dragon: Yes, I read that comment. I'll compare Ondio voltage display with multimeter after the alkaline test finishes.
23:34:05preglowBagder: i completely missed the step about first installing binutils 2.15, is it important?
23:34:13BagderI don't know
23:34:20BagderI don't see why that would be needed
23:34:21amiconnZagor: So tree.c needs to react on that event under the conditions mentioned, plus only if it's in file browse mode.
23:34:23Zagor[IDC]Dragon: isn't it just a question of adding an elseif after tree.c:1087
23:34:28preglownor do i, so i ignore it
23:34:40Bagderpreglow: try without it, and let's see what happens! ;-)
23:35:22preglowand both gcc 3.3.4 and 3.4.2 is mentioned, which should i use? i think i read that gcc 3.3.x produces more compact code somewhere
23:35:41amiconnZagor: I think another case: before line 1078 would be more appropriate (#ifdef'ed MMC)
23:36:14Zagoramiconn: yeah, i agree
23:36:31[IDC]Dragonthere's a HAVE_HOTSWAP now
23:36:58Bagderpreglow: get the latest, I'd say. that 3.3.x note is most important for the SH
23:37:15amiconn[IDC]Dragon: Prepared for a future SATA hotswap player ;-)
23:37:30[IDC]Dragonthe gmini can do that
23:37:47[IDC]Dragonit has a CF slot
23:38:19amiconnAh, yes
23:38:38preglowgcc 4.0 it is :P
23:38:57 Quit mecraw (Read error: 60 (Operation timed out))
23:39:36amiconnZagor: There is still a warning about implicit declaration for the win32 sims.
23:40:02[IDC]Dragona button code has the disadvantage that it can't carry data
23:40:18Zagoryes, i didn't have time to make the complete fix this afternoon.
23:40:24amiconnI didn't have a good idea how to solve this. Including our own string.h instead of the system's string.h wouldn't be a good idea imho.
23:40:48[IDC]Dragonin general, it would be good to carry the information which part of the f/s has changed
23:41:39Zagoramiconn: no, we need to declare it separately somewhere else. i couldn't make up my mind about what a good place was before I had to go
23:41:47amiconnZagor: Then the endianess problem with cygwin is also still there
23:43:58Zagori need help with cygwin-specific problems. i only rarely run win32.
23:44:09 Join amiconn_ [0] (
23:44:45amiconn_Ooops, disconnected by accidentally switching off the UPS
23:45:00 Quit amiconn (Nick collision from services.)
23:45:01 Nick amiconn_ is now known as amiconn (
23:45:33amiconnZagor: I could do some tests.
23:45:58[IDC]Dragontree.c insiders: how do I force a redraw from within?
23:46:52amiconnZagor: Cygwin defines neither LITTLE_ENDIAN nor BIG_ENDIAN (of course), and also no __LITTLE_ENDIAN__ or BYTE_ORDER.
23:47:01[IDC]Dragonok, set reload_dir = true;
23:47:19amiconnThere are only symbols like _X86_
23:47:46amiconnI tested a simple kludge in dbtree.c:
23:47:51Zagor[IDC]Dragon: you might want to make that reload_root instead, otherwise it may fail (if the new card doesn't have the current directory)
23:48:16amiconn#if !dfefined(LITTLE_ENDIAN) && defned(_X86_)
23:48:21 Quit methangas (" HydraIRC -> <- :P")
23:48:22amiconn#define LITTLE_ENDIAN
23:48:28[IDC]Dragonyeah, finding that out is the advanced version
23:49:30amiconnHowever, this does only help for building the X11 sim under cygwin, because for building the Win32 sim, cygwin has no netinet/in.h in that mode
23:50:04*amiconn can't type
23:55:19[IDC]Dragonhehe, this is cool
23:55:36[IDC]Dragonto see <MMC1> come and go
23:55:57amiconnAh :)
23:56:19amiconnCan't test atm, I don't want to have wasted a set of alkalines for no reason.
23:56:38[IDC]Dragonyou said so, yes
23:56:48[IDC]Dragon"no reason" ;-)

Previous day | Next day