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 2005-10-10

00:24:51 Join DrMoos [0] (
00:25:32 Quit Moos (Read error: 104 (Connection reset by peer))
01:08:16 Join aegray [0] (
01:08:32 Nick DrMoos is now known as Moos (
01:14:27 Join amiconn_ [0] (
01:18:17 Quit Moos ("Glory to Rockbox")
01:21:39 Join webguest24 [0] (
01:22:46webguest24ok, if i disable dircaching, isnt it supposed to not scan for anything?
01:24:30preglowamiconn: sure
01:24:59webguest24is dir caching here to stay?
01:25:20preglowwhy, does it scan for something when you turn it off?
01:25:25webguest24i mean, for me, that means i will not be able to be up-to-date for quite a long time
01:25:29preglowit doesn't scan for anything here if i disable it
01:25:31webguest24no the thing is
01:25:39webguest24i use korean patch
01:25:50webguest24and that seems to be interfering with the dircache
01:26:14preglowwell, maintaining other people's patches isn't really our job
01:26:17webguest24i got a dircache-less source from timid and thats what im using for now
01:26:25webguest24i know, but im just curious
01:26:36preglowbut yeah, i can't imagine why dircache would vanish
01:26:41preglowi love it, and lots of other people do as well
01:26:50webguest24if there's a way to not have take out dircache by myself
01:27:05preglowhow does the patch fail?
01:27:28***Saving seen data "./dancer.seen"
01:27:34webguest24when i boot it up, it freezes at the startup screen (Rockbox pic)
01:27:48webguest24actually, it doesnt freeze; the hdd keeps on spinning
01:28:16webguest24from my shallow background knowledge, dircache is probably having a problem recognizing korean dirs
01:28:39preglowcan't imagine why, though
01:28:44preglowis the korean patch maintained?
01:29:01webguest24i thought it was but it seems really uh... "quiet" as of now
01:29:13webguest24someone seems to be working on the unicode patch though
01:29:25preglowyou COULD ask slasheri why he thinks the patch fails
01:29:46preglowbut like i said, it's not our responsibility to make sure other peoples patches work
01:29:55preglowi'd love unicode rockbox myself
01:29:58webguest24right, i perfectly understand that
01:30:19webguest24i jus wanna know if i can get it "removed" manually
01:30:33markunphaedrus961 and I are working on unicode
01:30:41preglowmarkun: actively?
01:31:00markunwebguest24: Are you the one who has a korean language file?
01:31:10webguest24no it was el.. something but
01:31:12webguest24he uploaded it
01:31:24webguest24if that's not good enough, i can do one myself
01:31:34webguest24but i saw it posted
01:31:35markunpreglow: Well, not very active. It works fine so far.
01:31:44preglowmarkun: still, that's great news
01:32:13webguest24markun: is that completely diff than that chinese/japanese patch updated to support korean?
01:32:18preglowmarkun: how does it work? can you compile rockbox both with and without unicode support? or doesn't it alter rockbox size too much, so it can be enabled for all builds?
01:32:54 Quit amiconn (Read error: 110 (Connection timed out))
01:32:55 Nick amiconn_ is now known as amiconn (
01:32:58markunThe current unicode patch is my original unicode patch + the font caching stolen from the chinese/japanese patch.
01:33:42webguest24if korean works, i can be the beta tester :p
01:33:47markunpreglow: There are no ifdefs so it always uses unicode if that's what you mean.
01:34:03webguest24this is the uploaded lang file
01:34:25markunlatin1 fonts work, but for other fonts you need a unicode font. X11 ships with some.
01:34:27webguest24if u need any other help, ill do what i can (except programming...)
01:35:06webguest24is the font format the same? fnt file cs file?
01:35:46markunYes, the font format is the same.
01:36:27webguest24i see
01:37:07 Join Marcus [0] (
01:37:12webguest24so... anything i can do to help?
01:37:50Marcusquick question please: what the memacces debug feature is it for?
01:38:11Marcusamiconn ported it
01:38:20markunwebguest24: Hm. I will try to get korean to work properly and then you can play with it and tell me if there's a problem.
01:38:38markunwebguest24: What's your normal nick on irc?
01:38:44webguest24markun: ok :) tell me if u need the font
01:38:49Marcusamiconn: are you here?
01:38:56webguest24markun: ... random...
01:39:13webguest24markun: it's probably better for u to email me, g2tL84 at
01:39:18markunwebguest24: I have a font, but I could use some more. A bdf unicode font with arabic would also be very nice.
01:39:21webguest24or the username "user" on forums
01:39:39markunok, I'll mail you then if I have any updates.
01:39:43webguest24arabic as in.. numbers? or the language?
01:40:05Marcusthe language
01:40:53Marcusanyone here know what is it the new memory guard debug option please?
01:43:10preglowwell, it probably lets you trap memory accesses in some way
01:43:17preglowlike null pointer dereferences
01:43:23preglowexactly how i don't know
01:44:46Marcushave you examples for I understand well please?
01:45:37 Quit actionshrimp ("a bird in the bush is worth two in your house")
01:45:48preglowi'm not quite certain how it works myself
01:46:40Marcusok thanks anyway
01:47:17Marcusamiconn could explain me, but apperently isn't here
01:47:35preglowplace is usually quite dead sundays
01:48:54MarcusI'll ask an other day if I'll much of luck
01:49:13Marcusgood night @ all
01:49:22 Quit Marcus ("CGI:IRC")
01:55:59markunpreglow: I put "@$(MAKE) -C $(TOOLSDIR)" in the makefile, so the tools will always be up to date. Do you know why this is not done?
01:59:27markunwebguest24: ok, the korean language file now works with the unicode patch, but it has to load from the HD all the time because of the font caching.
02:09:56webguest24ok ok im back from dinner
02:10:06webguest24what do u mean it has to load from the hd?
02:10:10webguest24doesnt it always?
02:11:55 Join _user_ [0] (
02:11:55 Quit webguest24 ("CGI:IRC (EOF)")
02:12:10_user_gay web irc...
02:12:24_user_this is "webguest24"
02:12:51_user_markun: so what's going on exactly?
02:15:14_user_haha i think i might wanna re-translate some of them later
02:15:45markunThe problem is that the font caching only loads a small piece of the font. Every time it sees a new char it has to load it from disk.
02:15:59_user_oh i see
02:16:07_user_why is that?
02:16:17markunA solution would be to save all the chars that are in the cache when you turn of the player and load them again when you turn it on..
02:16:39markunWell, the font's are so big that you don't want to load the whole font into memory.
02:16:49_user_i see
02:16:57_user_well, that seems like a valid choice
02:17:30preglowmarkun: i have no idea, it's a known issue, but nothing has been done about it
02:17:36XavierGrmarkun how big is a font file with unicode?
02:17:45_user_or maybe cache the "korean part" first and cache the really big "chinese part" whenever needed?
02:17:51_user_fonts are usually around 1-2MB
02:17:58_user_oh wait thats bdf
02:18:03markunpreglow: shall I just commit it?
02:18:05XavierGrvery big
02:18:28_user_fnt i think my fnt file is 514KB
02:19:00XavierGrit wii be a problem for archos models
02:19:22markun_user_: greek people want to preload the greek part, russian people want the cyrilic part.. if you just remember what was in the cache it will work for everybody.
02:19:43_user_i see
02:20:09XavierGrbiggest font now is like 7,20 kb
02:20:46markunThe unicode font I used for the korean language file is 736K.
02:20:52_user_well the font's big size is due to the chinses chars, which arent used as frequently ( i think i have like less than 10 songs that have a chinese char in it)
02:21:00_user_wow even larger
02:22:19XavierGris there any relation between chines, japanese and korean languages?
02:22:43preglowmarkun: i say yes, if it seems to work
02:22:55_user_well, for one thing, korean fonts usually have all three
02:22:59preglowi have no idea why this hasn't been fixed yet, but i can't imagine anything more fancy than that is needed
02:23:39XavierGrI got a little confused, are you going to commit unicode support?
02:25:41preglowno, just a makefile change to compile the tools dir automatically if it's not already done
02:25:58preglowsince that just stopped working at some point
02:26:54markunpreglow: It has worked before?
02:27:16preglowmarkun: yes, i can distinctly remember make in build/ built the tools for you if they weren't built
02:27:21preglowbut it stopped working at some point
02:27:49XavierGrgreat idea.
02:27:55preglowwell, yes
02:28:00preglowthat's the way it's supposed to be
02:28:04preglowanywho, i need to go to bed
02:28:04preglowlater, all
02:28:12 Quit preglow ("leaving")
02:28:15markun_user_: Are your fonts unicode (ISO-10646) ?
02:28:46XavierGrnow for the unicode, how is it working? I mean regular fonts have 256 thesis to place english chars numerical, others and another 1 language.
02:29:14XavierGrSo if I load Greek on rockbox I cant read french and if I load french I can't read greek
02:29:22XavierGrso how a unicode font looks like?
02:29:33XavierGrand how it recognises which language to show?
02:30:20_user_its uhm.. ksx1001.1998
02:30:44_user_which is what the chinese/japanese patch uses
02:31:08markunXavierGr: Well, a char in unicode is 16-bit (or even 32-bit), so japanese, greek and french can be in the same file.
02:31:46markun_user_: I don't know if there is a converter, but it would be nice.
02:32:09_user_ill try to look for that type of fonts
02:32:17 Quit XavierGr ()
02:33:08 Join XavierGr [0] (
02:33:10markun_user_: I hope it will work. I remember a lof of people were angry because they felt unicode didn't handle CJK very well
02:33:37markunXavierGr: The unicode patch uses UTF-8 as encoding for the strings. Do you know what it is?
02:33:59XavierGrI know only that it is a coding for languages.
02:35:02XavierGras Western (ISO -xxx) e.t.c
02:35:05markunXavierGr: It a ASCII compatible way to store unicode strings. Included in the patch is a small script that will convert all the language files to UTF-8 so you can use all the languages together.
02:35:47XavierGrnice so that's why it makes the font file so big.
02:35:56XavierGrand how it can decide which language is shown?
02:36:19XavierGrI mean I may have tags in Greek and French or German language.
02:37:19markunXavierGr: If the tags are using unicode than it should all work. Ogg Vorbis and APE tags use UTF-8 for the tags, but I don't know about ID3 and ID3v2
02:37:40 Join lostlogic_ [0] (
02:38:00_user_markun: kinda seems wrong for someone doing nothing to get angry with someone who's working
02:39:08XavierGrI think that id3 tags support Unicode too.
02:39:30markun_user_: It was not related to rockbox. People were angry with the unicode standard.
02:39:43XavierGrso what is still missing on the unicde patch to be commited?
02:39:48markunXavierGr: Then maybe you will have to convert your tags, but I'm not sure.
02:39:57_user_markun: oh, ok
02:41:49markunHebrew is left-to-right with the patch, so that should be fixed. All the rockbox fonts that are not latin1 don't work anymore. And the font cache doesn't remember what was in the cache when you reboot or load a different font.
02:42:06markunAnd I have not found a proper unicode font with arabic.
02:42:08phaedrus961id3v2 tags do support unicode, but most are written in the local codepage
02:42:20markunphaedrus961: Ah, happy you are here!
02:42:27markunI have to go to bed so you can take over :)
02:42:38phaedrus961hebrew is right to left, by the way
02:42:51phaedrus961sure :)
02:43:12markunphaedrus961: great that hebrew works. I thought I tried it last week and it didn't.
02:43:23XavierGrID3v2 supports Unicode
02:44:01phaedrus961XavierGr: I just said that :)
02:44:17XavierGropps yeah sorry I just saw!
02:44:31phaedrus961I'm also working on conversions for cjk tags that aren't unicode
02:45:41phaedrus961see here for what needs to be done:
02:45:47markunphaedrus961: korean lang with the unicode patch:
02:49:06_user_markun: ok i downloaded this fontpack thingy.. and i see a bunch of .pcf files.. can you use this?
02:50:05phaedrus961we need bdf fonts for rockbox, but maybe they can be converted
02:51:05markunwell, good night everyone.
02:53:45 Quit lostlogic (Read error: 110 (Connection timed out))
02:55:48phaedrus961_user_: are those fonts unicode?
02:57:13_user_seems like it... i got it from this korean website and the file name is iso10646-1.exe
02:57:35_user_im trying to look for something like pcftobdf
02:58:00_user_bdftopcf seems like a common command though.. (i don't know since i don't use linux)
02:58:02phaedrus961can you give me the site? I have a util to convert them
03:00:53_user_if d/l speed is unbearably slow, ill upload it some place else or something
03:03:13phaedrus961is 110kb/s unbearably slow?
03:03:55_user_meh, seems alright for a 20MB file
03:07:08phaedrus961hmm, I have the file but I can't seem to extract it. do you know what kind of archive it is?
03:07:23_user_i had to install it
03:07:37_user_then it extracts the fonts in the destination folder
03:07:52_user_or are u on linux?
03:08:35phaedrus961I'm on linux, I'll have to move it to a windows box
03:08:54_user_oh isee
03:09:03_user_lemme send it to you by email then
03:11:20_user_actually, i have better idea, gimme a sec
03:18:44phaedrus961thanks, downloading it now
03:19:35_user_np, hope that helps
03:23:23 Join GaNgSta69 [0] (
03:23:27phaedrus961these fonts don't look big enough to contain cjk characters
03:25:06GaNgSta69i was currently using Rockbox on my h140 and did not notice how low the battery was.. it got down to basically no battery which i usually never do and now the dam thing wont turn on at all and does not recognise that its connected to the charger cable ... is there any hints to help me out or forums i can go 2
03:25:08phaedrus961nevermind, they're compressed
03:25:41_user_GaNgSta69: is the batt life that bad??
03:25:58_user_GaNgSta69: oh u mean it's broken?
03:26:15_user_GaNgSta69: i remeber reading some posts about low battery bugs...
03:26:28GaNgSta69its not turning on or doesnt recognise that cable is in to charge it..
03:27:03_user_did u try to reset it?
03:27:18_user_i meant, did u try resetting it?
03:27:22GaNgSta69i pressed it in .. yeh but nuthin happens
03:27:32***Saving seen data "./dancer.seen"
03:28:13_user_hmm so when u connect it to the charger, the light doesnt turn on?
03:28:25_user_did u try leaving it for a long time?
03:28:30GaNgSta69no light at all
03:28:38GaNgSta69ive had it in for around 10mins
03:29:24GaNgSta69im starting to worry.. ive had h140 for over yr and nuthin ever happened to it.. its my life ;) jkin.. but my prob is serious
03:29:54_user_hmm... i think something like that happend to me too
03:30:16_user_what i did was, i let it be for a long time w/o the charger
03:30:26_user_then tried to turn it on, which it did for a split sec
03:30:51GaNgSta69its been left alone for around 12hrs.. i went to sleep ... now im connecting to charger ..
03:31:21_user_if it ever turns on, even for a split sec, try plugging in ur usb cable
03:31:25_user_and back up everything if u can
03:31:32_user_my h120 failed 2 times
03:31:42_user_one time, i was not able to recover anything
03:31:49GaNgSta69shit i dont give a crap about the files.. i want my thing to work
03:31:55_user_lol well
03:32:18_user_iriver is usually generous about that kinda stuff even if ur warranty is gone
03:32:24_user_try calling them
03:32:29GaNgSta69im in AUS
03:32:33_user_you will probably have to wait a looonnnng
03:32:50_user_well in that case... i cant really think of anything.. sorry
03:33:02GaNgSta69dont they help with ppl in AUS
03:33:23_user_i have no idea
03:33:27_user_they probably do...
03:33:39_user_did u try
03:34:52GaNgSta69site isnt workin atm
03:35:11GaNgSta69or is that just me
03:35:49phaedrus961_user_: I don't see any cjk chars in those fonts
03:36:53_user_phaedrus961: aw man.. i guess ill look for more...
03:37:48GaNgSta69ive got my brothers h140 as well.. u think it would be a good idea to open mine up and swap with battery or somethin like that.. ?? or shouldnt do that ?
03:37:56GaNgSta69how much is a battery ?
03:38:12GaNgSta69is there a model i can look at or somethin
03:38:37phaedrus961maybe there's a way to convert the ksx1001 fonts to unicode
03:38:43phaedrus961I'll look into that
03:38:50_user_GaNgSta69: hmm doesnt work for me either
03:38:59_user_phaedrus961: yeah that'd be sooo nice
03:39:39_user_GaNgSta69: with my virtually non-existent knowledge, i suggest u wait for to work...
03:39:58_user_even rockbox forum doesnt work for me :(
03:40:19GaNgSta69major breakthrough
03:40:29_user_so it was ur charger then?
03:40:38GaNgSta69maybe.. ill connect mine to his
03:40:55 Join Shani}{Coder [0] (i=shani@Ariel.Atlantica.US)
03:41:13_user_elome.. or something like that claimed that he has the korean iso font in the forum
03:41:22_user_but the forum seems to be down or something.. i cant get in
03:42:26phaedrus961yeah, it's down for me too
03:46:28_user_btw, are you the guy who did the chinese/japanese patch?
03:47:33 Join random_man [0] (
03:47:46phaedrus961I updated it and added a few things
03:47:56phaedrus961but I'm not the original author
03:48:11random_mananybody else have the forums link time out on there?
03:49:05_user_phaedrus961: hmm so do you think there's a way to get around that dircache? so that i can get my rockbox uptodate?
03:49:26_user_random_man: forum doesnt work for me either
03:49:49random_manok just making sure its not my connection. you try too?
03:50:02_user_is misticriver on the same server? (i'm guessing tha's the case since jeff mods it) it doesnt work either...
03:50:36random_manyes now i remember it is
03:50:50phaedrus961I don't know anything about the dircache code, sorry
03:51:46phaedrus961but markun says that it works with the unicode patch, so you can use that when I get korean working
03:52:14random_manwhy do you guys think the remote patch commited to cvs?
03:54:06_user_phaedrus961: ok :)
03:54:19_user_i think timid and firefly are working on that
03:54:33_user_oh xaviergr too probably
04:02:13XavierGr no I am not on the remote anymore, unfortunately.
04:02:37XavierGrFirefly is doing something like I did, but more complete and bugfee.
04:02:51_user_oh ok
04:02:58XavierGrTiMiD has another approach which probably be commitable
04:03:15_user_so firefly's is still temporary?
04:03:38XavierGryeah, it wil not be commited and thats for sure.
04:04:54_user_i see
04:05:12 Join QT [0] (i=as@madwifi/users/area51)
04:16:06 Quit RotAtoR ()
04:17:18 Quit QT_ (Read error: 110 (Connection timed out))
04:24:27random_manwhy wont fireflys be commited?
04:35:22 Quit _user_ ("CGI:IRC (EOF)")
04:46:11 Quit Shani}{Coder (Read error: 113 (No route to host))
04:49:09 Part aegray
04:51:19XavierGrrandom_man: Because his way is not good for commitment, it will add clutter to the code, rockbox isn't made for the iriver specifically, so the devs are right that they don't want this way. After all they denied my remote patch too, because it was this way(same as FireFly's). But have patience, TiMiD is working on another remote approach which probably be accepted for commitment.
04:56:06 Quit GaNgSta69 ()
05:16:33random_manhmmm they seem kinda picky...
05:19:23random_manwell i hope when they do commited i hope its as good as as the latest remote patch from firefly
05:20:11XavierGrit will be even better because this way rockbox can handle any number of screens indpendently
05:20:34XavierGrand will be good for further ports too
05:24:02random_manthe forums are working again if you didnt know i just found out....
05:24:03TiMiDstill awake guys .?
05:27:35***Saving seen data "./dancer.seen"
05:30:14TiMiDahthey flew away :(
05:31:09random_manwat is link for updated remote implamented in rockbox build?
05:32:16TiMiDthe firefly's one ?
05:32:35TiMiDit may be here
05:32:37random_manyes. isnt it ?
05:32:40TiMiDor not ...
05:32:55random_mani looked its only updated to the 8th but check this out.....
05:33:32random_manI have uploaded my cvs-sources with the latest cvs changes(9 Oct 13:37) or does this mean something different
05:34:28TiMiDthis one seems to be from yesterday
05:36:54TiMiD(else you have my patch, but I'm not sure it's what you want :p)
05:37:03random_manwho was it who had the other page with other sorts of versions of rockbox with plugins and stuff in them. is site kinda looked like rockboxes. with the backround colors
05:37:23random_manyes i heard you are making one. how is that comming?
05:37:36random_mani havnt head much bout yours TiMiD
05:39:40TiMiDit was rasher's page I thing
05:40:04TiMiDyou can find it on the forum (if it wants to work :p )
05:41:05random_manfourms are working... for me
05:41:05TiMiDabout my patch : I'm waiting for core dev's feedback, but they seems very busy right now
05:41:21TiMiDfor me it's working but very slow">
05:42:25TiMiDhere it is
05:42:27random_mani have slow internet so its the same for me. \
05:43:00TiMiDI have no internet :D
05:43:24TiMiDit's my neighbour wifi ... and it depends on the wether, etc ...
05:44:27random_manlol nice the nearest neighbor for dsl is bout 10 miles away so im stuck with less then 56k
05:44:41TiMiDuse a parabol :)
05:45:37random_mani dont know what that is
05:46:27TiMiDmaybe my english taht sucks ;p
05:46:57TiMiDparabola ?
05:48:50random_manexplain what it does and i might know
05:49:11TiMiDit's used to receive radio waves
05:49:18TiMiDfor example from satellites
05:49:51random_manoh ok. i dont really know the kid anymore so i dont think i could get it anyways
05:50:28TiMiDthat :)
05:51:02TiMiD(I want one in my garden XD)
05:51:17random_manuhh thats a little big
05:51:49random_manwe use to have an old sat dish that was the old school ones (very huge and white) mabey i could of riged it up to do that
05:52:59TiMiDthere are guys who managed to get wifi signal from 10 miles
05:53:20random_manholy crap
05:53:49TiMiDwith a .. I don't know how to say this in english, but it's very cheap
05:53:58TiMiDlook at the pictures :)
05:55:44random_manwow different language. ill use babel fish
05:55:57TiMiDhmmm :)
05:56:09TiMiDI remember
05:56:11random_manwhat language is that?
05:56:24TiMiDthe word is "cans"
05:56:30TiMiDfrench :)
05:58:02random_mancool. mabey i can get ffaster internet then. my town is bout 4 miles away...
05:58:15TiMiDha :)
05:58:40TiMiDprepare your drill !
05:58:54random_manthis isnt illegal is it?
05:59:13TiMiDthe antenna is legal
05:59:19random_mando you know how much this costs?
05:59:22TiMiDstealing the connexion isn't
06:00:00TiMiDthe cost should'nt exceed 10USD
06:00:13random_manwow really. sweet
06:00:33 Join linuxstb_ [0] (
06:00:36random_mannow if only i could read french cuz trasnlation dosnt make sense
06:00:54TiMiDthis seems to be english :)
06:01:17TiMiDall you need is a can, a drill and some wires
06:01:30TiMiDbut I can't guaranty that it will work ...
06:01:40TiMiDI didn't tested myself
06:01:48random_mansweet. that one is under 5 bucks
06:02:43TiMiDyou can do amazing things with a can :)
06:04:09random_manhaha yes you can
06:05:22TiMiDbtw, are you a native english speaker ?
06:05:43TiMiDcool :)
06:05:47TiMiDI have a question
06:05:50TiMiDshould I ?
06:06:37XavierGrTiMiD:still awake guys .?-> I didn't sleep at all. I have college lesson in 8.00 am. If i sleep now I will surely miss it!!!
06:06:43TiMiDI have to say a sentence but i don't know how to say il properly
06:07:02TiMiDit's for my CV
06:07:09TiMiD(I must do in in english ...)
06:07:10random_manuse translator haha
06:07:31TiMiDI tried but it's far too complicated :p
06:07:46TiMiDwhat I want to say is that sometimes I hold the student bar
06:08:10 Join brunofone [0] (
06:08:12TiMiD(I receive custommers)
06:08:54TiMiDXavierGr: no school tomorrow XD
06:09:08TiMiD(but i must do this fckù$ù CV)
06:09:17brunofoneyo guys, any word on the progress of rockbox for the gmini 220?? as in, is the project still going on at all??
06:09:54TiMiDit's paused :)
06:10:13brunofonehaha with a smiley face?
06:10:27brunofonei just bought the 220, man the firmware is bad
06:10:29TiMiDI don't own a gmini :p
06:10:30brunofoneit hangs like crazy
06:11:26brunofonewhy is it paused? the guy doing it doesnt have time?
06:11:43TiMiDno he stopped working on it
06:11:54TiMiD(but a lot of work has been done
06:12:08brunofoneno plans to pick it back up?
06:12:16TiMiDyou could do it yourself :p
06:12:33random_manhow far is 200 meters ?
06:12:50brunofonehaha i go to georgia tech, i could probably find some cs kid that would do it
06:14:03TiMiDrandom_man: 1/5 km
06:14:10TiMiD1/10 miles
06:14:24random_man1/10 of a mile?
06:14:43XavierGrI hate metric inconsistencies!!!
06:14:55TiMiDa little more but that's around that
06:15:24 Quit linuxstb (Read error: 113 (No route to host))
06:15:27 Join webguest64 [0] (
06:15:58 Quit webguest64 (Client Quit)
06:16:48random_manhow long is 5000 meters?
06:16:52XavierGr200 metres = 7874.02 inches or 656.17 feet or 218,72 yards or 0,12 miles or 0.2 km
06:17:10XavierGrin what you want the answer?
06:17:13 Join Paul_The_Nerd [0] (
06:17:37XavierGr1640.42 feet
06:17:55random_man5 000 meters = 3.10685596 miles
06:18:03TiMiDrandom_man: for the 10 miles, it was with parabolas
06:18:19XavierGroh I thought he said 500 metres
06:18:45random_manoh. well i found one that was tested wtih 5000 meters for a tin can
06:18:46TiMiD3 miles with a can antenna
06:19:10TiMiD79km with a parabola :)
06:19:29TiMiD(it's in french but look the numbers :p)
06:19:35XavierGrparabola, you mean the big satelite dish you posted?
06:20:37XavierGror a little tv-like antenna?
06:22:05XavierGrcan we build that ourselves?
06:22:17random_mani would like one of those
06:23:00random_mancouldl i jsut use an old small dish instead of havint that one ?
06:23:11TiMiDI don't know
06:23:57TiMiDantennas are not random whapes
06:24:10TiMiDit must be calculated
06:24:11random_manwell i have to go. nice talking to you. later
06:24:15 Quit brunofone ()
06:24:24 Part random_man
06:24:43TiMiDand my question :(
06:28:53TiMiDany native english speaker awake ?
06:29:45Paul_The_NerdHow can I help you?'
06:29:56TiMiDI have to do a CV in english
06:30:14TiMiDI've done it but I'm not sure it's "english"
06:30:25TiMiDsome sentences sounds weird to me
06:31:22TiMiDCurriculum Vitae
06:31:39TiMiDit's a document you provide a company when ytou want them to employ you
06:32:02TiMiDto postulate for a job
06:33:26Paul_The_NerdWell, I wouldn't mind looking at it. I'm a native english speaker, though I'll admit my grammar is imperfect.
06:33:26TiMiDI think it's how you say it
06:33:30Paul_The_NerdYes, it is.
06:33:35TiMiDthx :):)
06:33:43TiMiDdo you have openoffice ?
06:33:57TiMiDa .odt is finethen ;)
06:34:05Paul_The_NerdIn fact, it is preferred.
06:35:10TiMiDa .doc would look crappy with the conversion
06:35:19TiMiD Kevin FERRARE anglais.odt
06:35:25TiMiDit's here
06:35:55TiMiDthank you very much !
06:36:05TiMiD(it's not a very big document)
06:37:58Paul_The_NerdHrm... your use of the word "achievement" may not be the best choice. I think it works, but it's a little awkward.
06:38:40TiMiDwhat word would you use instead ?
06:38:44Paul_The_NerdI'm thinking about that.
06:38:48TiMiDok :)
06:39:16Paul_The_NerdThese are things you made, or helped make?
06:39:29TiMiDwhat I want to say is just that it wasa part of my work
06:39:43TiMiDthings I made
06:40:54Paul_The_NerdI think 'creation' might sound better then.
06:41:30TiMiDok I change that :p
06:42:18Paul_The_NerdBeside "Clanlib" you spelled multiplatform wrong. There's an E in it that doesn't belong.
06:43:03TiMiDI wrote it the french way :)
06:44:22Paul_The_NerdI'm just trying to figure out how to reword this sentence: "handling and reception of the public in a show room and in a swimming pool." It sounds very strange, but I'm not sure I can think of a better way of wording it.
06:44:57Paul_The_NerdIt provides the information clearly, is just seems strange reading it.
06:45:12TiMiDbecause the words are in french order maybe :)
06:45:32Paul_The_NerdI think it actually works. I don't know if there's a better way to say it.
06:45:54TiMiDshow room is correct ?
06:46:35TiMiD(I meant a place where you listen concerts and so on)
06:46:42TiMiDI wasn't sure
06:47:06Paul_The_NerdWell then, perhaps "Theatre" "Auditorium" or "Concert hall"
06:47:25TiMiDit was a multi purpose place
06:47:36Paul_The_NerdWell, show room works.
06:48:11TiMiDwhere there was concerts, art, dance, associations, political meetings ...
06:49:24Paul_The_NerdI think that's probably the best way of saying it then.
06:49:55TiMiDthe sentence still looks strange ?
06:49:58Paul_The_NerdYou may want to change "in a swimming pool" to something like "around a swimming pool." I don't know exactly what you intend, but it sounds like everyone was in the water the whole time to me.
06:50:46TiMiDI was in the cloakroom and at the counter
06:50:55TiMiDto help people
06:51:12Paul_The_NerdSo maybe change it to "at a swimming pool"
06:51:20Paul_The_NerdI think it's the "in" that makes it seem strange to me.
06:51:22TiMiD(and to punish them if they did silly things XD)
06:51:45TiMiD"at" si better ?
06:52:11Paul_The_NerdI believe so, yes.
06:52:40TiMiD(yes, tricking girls with sprinkler pipe with cold water was the best ^^)
06:53:30TiMiDat doesn't sounds good I think
06:53:41TiMiD"at a swimming pool." is correct ?
06:54:22XavierGrI think you must remove the "the" before public
06:54:59Paul_The_NerdWell, "a swimming pool" generally refers to the water itself, and not the facilities around it such as the counter. So I think "at" would be best.
06:55:10TiMiDok :)
06:56:00Paul_The_NerdXavierGr: The "the" before public is fine, assuming "the public" refers to people in general.
06:56:45TiMiDoops I found a fatal error : my age
06:56:57TiMiD22 years old is more correct :)
06:57:24Paul_The_NerdI didn't even look at the date.
06:57:25XavierGrYes that why I said it, tough no native speaker of English, I remember someone saying that "the" seems to has a better use when naming something unique and not general.
06:58:10TiMiDthat's whyI put the "the" here ;)
06:58:10Paul_The_NerdYes, but Public is an adjective usually. You can have a public item, meaning that it belongs to everyone. When you say "the public" you make public into a noun.
06:58:34TiMiDsame in french :)
06:58:56XavierGrHey TiMiD: We have the same age!
06:59:02Paul_The_NerdI'm 22 as well.
06:59:11TiMiDow ^^
06:59:18TiMiDwhen were you born ?
06:59:30Paul_The_NerdJanuary 12, 1983 here.
06:59:41XavierGroh then I am 21 and heading to 22
07:00:08TiMiDI'm not the only one mistaken about my age :p
07:00:23XavierGrwhen TiMiD?
07:00:34TiMiDwritten in my corrent :p
07:00:52TiMiDSeptember 19, 1983
07:01:22XavierGrI am the newbie here then...
07:02:04TiMiDI'm not far from you then :)
07:02:42TiMiDPaul_The_Nerd: I wanted to add smth to my current but I don't know how to say it in english
07:02:51TiMiDit's about the studies
07:03:14TiMiDI'm in computer engineering
07:03:34TiMiDbut there is a specialisation in systems and networks
07:04:15TiMiD(but I don't want to say I'm specialized because it's not that)
07:04:25Paul_The_NerdYou could just say "Computer engineering with a focus on systems and networks"
07:04:31TiMiDok :)
07:04:41TiMiDthat's perfect :)
07:04:47Paul_The_NerdGlad I could help
07:05:46TiMiDthank you verrryyyyyyyyyyyyyyyyyy much :)
07:06:08TiMiDI was afraid my current looked like chinese to a native speaker
07:07:14TiMiDI need it because I want to do my last training period abroad
07:07:39Paul_The_NerdNo, it's very readable
07:08:09Paul_The_NerdThough "LANGUES ETRANGERES" should probably be translated to English. ;-)
07:08:35TiMiDI didn't even noticed :)
07:08:45Paul_The_NerdI read right over it the first time.
07:09:05TiMiDme 3 :)
07:09:20TiMiD(I don't read the titles ^^)
07:09:44TiMiD'FOREIGN LANGUAGES' should be fine :)
07:10:03Paul_The_NerdOr simply "LANGUAGES"
07:10:34TiMiDok :)
07:11:43TiMiDanother expression that sounded strange was "board of directors"
07:12:29TiMiDwhat I just wanted to say was that I was in the decision group of the assiciation
07:12:39TiMiDbut nothing to do with buisness
07:12:53Paul_The_Nerd"Directing committee"
07:12:56TiMiD(in french, director is associated to a company)
07:13:09Paul_The_NerdBoard of Directors does sound like a more business related position
07:13:15TiMiDok :)
07:14:30TiMiDI'm reassured now that you read it :p
07:16:05TiMiDmaybe I should add that I get the first place at the inter-university bier drinking contest :D
07:16:24TiMiD(or how to break the whole thing)
07:17:31TiMiDmaybe just the 'Holding the student bar.' is inappropriate
07:18:13Paul_The_NerdI'm actually not certain what you mean by that.
07:18:45TiMiDI couldn't find an exact translation
07:19:06TiMiDwe have a student bar holded by students
07:19:43TiMiD(they prepare cofee, biers, it's a free work (they are not paid for that))
07:19:52TiMiDand sometimes I do this too
07:19:56TiMiDthat what I meant
07:21:02Paul_The_Nerd"Volunteering at the student bar"
07:21:15TiMiDsounds great :)
07:23:24 Join matsl [0] (
07:23:48TiMiDthanks for having granted me your time :)
07:24:00TiMiDit's very kind
07:24:22Paul_The_NerdIt's no problem
07:25:38TiMiDI think I'm gonna go to bed
07:25:46TiMiD(it's 7am here)
07:26:07TiMiD(tonight, I couldn't sleep)
07:26:22TiMiDa party until 5 am
07:26:31Paul_The_NerdSee you later then
07:27:31TiMiDand then this current to do (I was anxious because there are few works abroad and alot of students, so I must do it very quickly)
07:27:37***Saving seen data "./dancer.seen"
07:27:57TiMiDwell good night then!
07:28:12TiMiD(it must be night too where you are)
07:28:21Paul_The_Nerd12:28 AM
07:28:47TiMiDyou are near to the atlantic side then
07:28:54Paul_The_NerdTexas. :(
07:29:13TiMiDforget it :)
07:29:49TiMiDI hope some core devs give me their opinions about the remote patch I did
07:29:58TiMiDbut they seems very busy
07:30:07Paul_The_NerdYeah, it can take a while
07:30:13TiMiD(I havn't seen one since Isubmutted it)
07:30:58Bgermorning :)
07:31:04Paul_The_NerdMornin' Bger
07:31:09TiMiD'night :)
07:31:12 Quit XavierGr (Read error: 110 (Connection timed out))
07:31:14Bgeryep :)
07:31:17TiMiD(even if it's morning here)
07:31:40BgerTiMiD u haven't slept ?
07:31:42TiMiDtoday will be a zombie day
07:31:46TiMiDnot yet
07:31:51Bgerbad :(
07:31:59TiMiDI had lot of things to do
07:32:09TiMiDand I don't have school today :)
07:32:11 Quit matsl (Remote closed the connection)
07:32:18*Bger haven't awaken yet
07:32:34TiMiDwhat time is it 4 you ?
07:33:05Maxime`hum, ctcp time don't work :p lol
07:33:22Bger Private messages from unregistered users are currently blocked due to spam
07:33:22Bger problems, but you can always message a staffer. Please register! (
07:33:22Bger )
07:34:09Maxime`ah yup
07:34:11TiMiDwell cu guys
07:34:26Maxime`it's strange that ctcp-replys are filtered too
07:34:29TiMiD(and thanks again Paul_The_Nerd :) )
07:34:54Paul_The_NerdNo problem TiMiD
07:35:30 Nick TiMiD is now known as TiMiD[sleepingAw (n=TiMiD[
07:35:30DBUGEnqueued KICK TiMiD[sleepingAw
07:35:36Bgernite, TiMiD[sleepingAw
07:36:01TiMiD[sleepingAwnight .. it's almost day here :)
07:36:20Bgeryea, i know, i'm in east europe
07:36:32Bger8:36 here
07:36:57TiMiD[sleepingAw7:36 :)
07:41:59Bgerhm, Maxime` it should work now :)
07:42:40Maxime`Bger: i've turned "unfiltered" on, then it worked ^^
07:42:52Maxime`it's just that CTCP-REPLY uses "privmsg"
07:42:57Maxime`so the server filter it
07:43:18 Nick Vlad0man is now known as Vladoman (
07:43:23CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
07:43:23*Bger wasn't identified
07:43:33Bgerthat was the problem
07:43:36Maxime`ah :p
07:43:55Maxime`yup ^^
07:44:53Maxime`time to school
07:48:59 Join ender` [0] (i=ychat@
08:03:09 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
08:14:41 Join ashridah [0] (
08:42:16 Join _FireFly_ [0] (
08:45:18 Quit Bger ("bbl")
08:46:07 Join Vlad0man [0] (
08:49:25 Join einhirn [0] (
08:58:40 Quit linuxstb_ (Read error: 104 (Connection reset by peer))
08:59:14 Join linuxstb_ [0] (
09:02:51 Join linuxstb__ [0] (
09:03:55 Quit Vladoman (Read error: 110 (Connection timed out))
09:07:32 Join Bger [0] (n=Bager@
09:19:36 Quit linuxstb_ (Read error: 110 (Connection timed out))
09:26:20 Quit linuxstb__ (Read error: 104 (Connection reset by peer))
09:26:54 Join linuxstb__ [0] (
09:27:38***Saving seen data "./dancer.seen"
09:31:44 Quit Maxime` (Read error: 104 (Connection reset by peer))
09:31:46 Join linuxstb_ [0] (
09:33:00 Join Maxime [0] (
09:38:54 Join ender1 [0] (i=ychat@
09:39:27 Quit leachbj ("Client exiting")
09:48:29 Quit linuxstb__ (Read error: 110 (Connection timed out))
09:51:51 Quit ender` (Read error: 110 (Connection timed out))
09:56:38 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
09:57:40 Join |Lupin| [0] (
09:57:46|Lupin|Hello, everybody.
09:58:41|Lupin|Just a quick question. Can I copy rockbox on an Iriver before booting on the new firmware, or should I first copy the firmware, then upgrade it, then reboot, then copy rockbox, and then finally runrockbox ?
09:59:36 Nick paugh is now known as AliasCoffee (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
10:01:29ashridahit doesn't matter. having rockbox on the iriver will do nothing until a patched firmware is on
10:02:02linuxstb_|Lupin|: It doesn't matter. If the bootloader hasn't been installed, rockbox files will just be ignored on the hard disk. If you've installed the bootloader, but rockbox doesn't exist, then the bootloader will just load the original firmware.
10:03:47|Lupin|So I can copy everything at once
10:03:58 Join XavierGr [0] (
10:04:13XavierGrGood Morning!
10:04:13|Lupin|and then oncethe firmware is upgraded, the next start will run rockboxAutomatically.
10:04:18|Lupin|Am I correct ?
10:05:47Bgermorning, XavierGr
10:06:11Bgerlinuxstb_ 10x for looking for "password wallets"
10:06:20|Lupin|Thanks all
10:06:22|Lupin|C U later
10:06:29 Part |Lupin|
10:07:59 Join Musicmad [0] (
10:09:32 Join B4gder [0] (
10:36:54B4gderbusy morning today :-)
10:38:10Musicmadwould it be considered rude to push this one to the front :)
10:39:13B4gderyou can push all you want, I think the problem with this is that no dev can repeat this
10:40:07Musicmadyes - I fear that too... :)
10:41:28Musicmadwhere do I need to send my player then!
10:41:39 Join linuxstb__ [0] (
10:41:40Musicmadthere is a danish dev right? (rasher?)
10:54:58 Quit linuxstb_ (Read error: 110 (Connection timed out))
11:09:00 Quit Slasheri (
11:09:58NJoinSlasheri [0] (
11:13:26XavierGrhi Slasheri!
11:13:33SlasheriXavierGr: hi :)
11:22:17 Join Moos [0] (
11:22:26MoosMorning guys
11:22:53MoosSlsheri: hi, you there?
11:23:02Moos*Slasheri :-)
11:24:22markunMorning everyone
11:24:39MoosHi markun
11:25:21markunMoos: I now have rockbox in korean ;)
11:25:48 Join LinusN [0] (
11:25:54Mooscongratulations :-)
11:25:59LinusNthanks! :-)
11:26:08amiconnooh, hi LinusN
11:26:12MoosBonjour Linus
11:26:43LinusNi asked varitronix directly about the H300 LCD, showing them the JPG of the markings
11:27:02LinusN"The mentioned display is our custom tooled
11:27:02LinusNmodule, not our standard module, hence we are reluntant to provide its
11:27:02LinusNspecification and details to others."
11:27:42***Saving seen data "./dancer.seen"
11:27:52MoosSlasheri: still doesn't know why there are shuting down problem with dircache? I read guys with my problem too in the forum :-(
11:28:07BgerLinusN is this very bad :(
11:28:39LinusNBger: yes, but let's hope that it still has the HD66789R controller
11:29:00LinusNit might just be a form factor thing
11:29:15Bgerwhat's the chance of getting info from iriver firmware otherwise ?
11:29:49Slasherihi Moos, LinusN and everyone :)
11:30:11MoosHi :)
11:30:34SlasheriMoos: i don't know because i have never experienced that myself..
11:30:55Moosany idea what can cause this?
11:31:43Moosit systematicly freeze when shuting down
11:31:54Moosserious bug, :-(
11:31:58Slasheriinteresting.. what it shows on the display when it freezes?
11:32:19Slasheritry the debug console and paste all lines you have on the remote lcd before it freezes
11:32:24LinusNBger: we can of course reverse engineer the firmware, or use a logic analyzer to snoop the LCD communication
11:32:25Moosit freeze in the browser screen
11:32:48LinusNMoos: but not with the cache disabled?
11:32:51Bgeri suppose it'll slow down the bootloader...
11:33:03LinusNBger: somewhat
11:33:05MoosLinus: yes just with the cache
11:33:19LinusNtodays daily?
11:33:25Bgerwhat about Moos saving the config to .cfg file ...
11:33:46Moosyes since 1 day or 2
11:34:05SlasheriMoos: that is weird.. in fact there might be one possible problem that cache doesn't list the . and .. files currently but i don't think that should cause it
11:35:22SlasheriMoos: try saving the cfg file as Bger just mentioned and then try resetting configuration. And send the config to me too
11:35:23MoosI thought I'm alone with this bug, but it was reported in Rockbox forums too
11:35:35MoosI trying
11:36:28LinusNMoos: so you just start it, and shut down directly?
11:36:35LinusNwithout resume
11:37:12LinusNMoos: database or file mode?
11:37:25Moosstart player, play new music, stop and long press stop for turned off, but freeze in the browser screen
11:37:25 Join |Lupin| [0] (
11:37:31|Lupin|Hi again
11:37:39Moosfile mode
11:37:53|Lupin|This is a question for Ljoux users: which USB controller (modules) do you use tomount an Iriver H1xx, please ?
11:37:56Moosdircahe don't work for database
11:38:09SlasheriMoos: so first you press stop shortly and the file browser still works? then you long press it and freeze?
11:40:18LinusN|Lupin|: what is Ljoux?
11:40:32ashridah|Lupin|: generic usb-storage + scsi disk support works fine here
11:40:38Bgerlinux maybe :)
11:40:39MoosSlasheri: long press freeze the player
11:41:02Moosfor shut down
11:41:11amiconnMoos: It's a dircache, not a database cache
11:41:14ashridahso usb-storage and sd need to be loaded (plus pre-requisites)
11:41:22SlasheriMoos: ok, and before the long press you have a short press?
11:41:30|Lupin|yes Linux...
11:41:33Moosamiconn: indeed :)
11:41:35LinusNMoos: playlist or regular dir play?
11:41:40|Lupin|but nothing happens when I plug in the Iriver.
11:41:45ashridah|Lupin|: there's nothing special fortunately, the filesystem will be vfat (fat32) and you just mount it and copy files across like any hard disk
11:41:54LinusN|Lupin|: /var/log/messages?
11:41:55MoosSlasheri: yes short press for quit WPS
11:41:58|Lupin|I started it, though. Is there something else to do before pluging it to the computer ?
11:42:02SlasheriMoos: ok, good
11:42:30ashridah|Lupin|: did the device go into usb mode?
11:42:48|Lupin|I don't know.
11:42:52|Lupin|How can I check this ?
11:43:01Bgeris there something on the screen
11:43:02SlasheriMoos: do you have any patches or just a fresh daily build?
11:43:14Bgersaying "CONNECTED"
11:43:23Bgeror something similar
11:43:30MoosSlasheri: fresh daily build of yesterday
11:45:02|Lupin|for the moment the Iriver has the normal Iriver firmware.
11:45:11|Lupin|I would like to mount it to copy Rockbox on it.
11:45:26Bger|Lupin| so it runs iriver fw ?
11:46:02Bgerok, when u start it and plug in it to the USB ... is there something on the screen? like "CONNECTED" "USB" etc ?
11:46:03Musicmadquestion regarding the no sound bug. To make a clean install do I need to do anything else than deleting all rockbox related and copy the newest build over?
11:46:33BgerMusicmad overwriting should be sufficient
11:46:54Slasheri|Lupin|: check you have at least the following modules loaded: usb_storage and usbhid (kernel 2.6)
11:47:04MusicmadBger: thanks - what about RoLo? - I have only used the fwpatcher once a while back.
11:47:06ashridahnot usb_storage
11:47:06Slasherithen tail -f /var/log/kern.log and plug the player
11:47:19 Quit _FireFly_ (Read error: 110 (Connection timed out))
11:47:24Slasheriashridah: hmm, ok :)
11:47:28markun|Lupin|: How do you transfer music to your iriver if you can't get USB to work?
11:47:33Bgerashridah it's usb_storage
11:47:43ashridahi stand corrected
11:48:12ashridahit shows up as usb_storage in lsmod. it loads as usb-storage.
11:48:13Slasheritry loading also uhci_hcd and ehci_hcd
11:48:15*ashridah gives himself half marks
11:48:17|Lupin|Mounted. Thanks !
11:48:23Slasherigreat :)
11:48:37|Lupin|ehci_hcd seems the right choice.
11:48:47|Lupin|Btw: Is Iriver also USB 1.1 compatible ?
11:48:56Slasheriyes it is
11:49:01|Lupin|ok, thx.
11:49:30MusicmadBger: what do you think?
11:49:46Musicmaddo I need to use fwpatcher again or?
11:50:14Bgerit's for the bootloader
11:50:24Bgerbtw, what's your bootloader version?
11:51:24SlasheriMoos: i think i will provide you a few test builds you can try when i get home
11:51:56Moosok I will go too, I could test this night
11:52:08Slasheriok, good :)
11:52:18Slasherii add some debug numbers to the remote lcd
11:52:24Moosin waiting can't use your cache here :-)
11:52:34MusicmadBger: let me check
11:52:37Slasheriyou then just need to tell me the last number it displays when it freezes
11:53:53Musicmadhm where do you check bootloader version?
11:54:34BgerSlasheri does the last version of the bootloader display build date
11:54:42MoosSlasheri: logs in remote isn't in the CVS version
11:54:54SlasheriMoos: no
11:55:07SlasheriBger: at least it displays the version
11:55:18Moosok, I will test with your builds
11:55:27BgerMusicmad restart your player
11:55:32Slasheribut not the date
11:55:46Bgerit should show it in the very beginning
11:55:50Slasherirockbox displays build date but not the boot loader
11:56:18SlasheriMoos: in fact i can do one build you immediately if you have little time to test :)
11:56:23MusicmadBger: ok - what am I looking for?
11:56:32Slasheri(i can't test right now but i can remotely create the build still)
11:56:46BgerMusicmad, Slasheri could describe it better than me ...
11:57:03 Join preglow [0] (
11:57:07Musicmadok thanks - Slasheri how do I check the bootloader version please? :)
11:57:20Bgerhm, easier q: when did you put the rockbox's bootloader ?
11:57:35SlasheriMoos: what is the first line the bootloader says when you power on the unit?
11:57:51XavierGrit says the version of the bootloader right at the start.
11:58:00MusicmadBger: maybe 3 months ago when I first used fwpatcher. Have only done daily build overwrites since that.
11:58:01Slasheriyes, at least version 5 says
11:58:07XavierGris says Bootloader vx
11:58:16MusicmadI might have an old version then?
11:58:26MoosSlasheri: I'll go to university right now, I'll back in few hours
11:58:30XavierGrcheck it out.
11:58:34BgerMusicmad then i definitely suggest u to upgrade ...
11:58:40markunMusicmad: or menu -> info -> Rockbox info
11:58:42Bgerthere are some bugs
11:58:49Bgerin older bootloaders
11:59:02markunEhh.. Version :)
11:59:14MusicmadBger: excellent news. Let me check the site on info on how to upgrade.
11:59:24SlasheriMoos: ah, ok :) see you later :)
11:59:43Moosc you later guys
11:59:55BgerMusicmad i'm not telling you that this will get out your problems...
12:00:05 Quit Moos ("Glory to Rockbox")
12:00:28MusicmadBger: I know, but it's something to try at least.
12:00:48BgerMusicmad did you find the page ?
12:01:09XavierGra slight detail may cause disastrous results, you have nothing to loose if you try
12:02:19Musicmadok thanks.
12:02:40Musicmaddoes fwpatcher contain the newest bootloader?
12:02:59Musicmadhow cool is that
12:03:19preglowwould be pretty useless otherwise...
12:04:12Bgerbtw, LinusN, while on bootloader dev, what about really moving the rockbox.iriver in the .rockbox dir ?
12:04:47XavierGryeah I agree on that too.
12:05:30Bgerone less entry in the root dir :)
12:05:49preglowwell, it's more or less the same to me
12:05:53XavierGrand safety reasons.
12:05:58preglowsafety? how?
12:05:58LinusNand less "my .rockbox dir is missing" questions
12:06:05Bgeryep :)
12:06:08XavierGrsomeone could delete it by accident
12:06:20XavierGrand dont say that applies too .rockbox
12:06:28LinusNbesides, it's time to rename ".rockbox" to "Rockbox"
12:06:48XavierGrbecause .rockbox is filtered off and not showing on the Supported filter.
12:06:52Bgerhm :)
12:07:02XavierGrWhile rockbox.iriver is a rolo file and can be clearly seen
12:07:02LinusNbecause of the poor os x lusers
12:07:35XavierGralso if you want to keep it to root you could at least make it read-only proof.
12:07:44preglowhmm, i like my .rockbox dir hidden :)
12:07:57*Bger likes hidden .rb too
12:07:57markunLinusN: OS X cannot see hidden files?
12:08:18XavierGrme too but rolo files aren't hidden on "Supported" mode
12:08:26*Bger even likes the exact ".rb" or ".RB"
12:08:30LinusNmarkun: it hides dot files
12:08:49MusicmadBger: got the newest bootloader. Will have to wait until tomorrow to see if it's fixed. Will report back in the forum. Thanks a lot for your help.
12:08:59LinusNmarkun: and the gui unzipping tools skip hidden data
12:09:01Bgerfor nothing:)
12:09:28XavierGrLinusN: By the way, I upadted both patches (FM presets, JPEG scroller) with brackets fixed and solved the bug in FM preset.
12:09:36Bgerthese .DS... dirs are awful
12:10:10XavierGrIf you find anything else please let me know.
12:10:40LinusNwill do when i find the time
12:13:49BgerLinusN any other news on the H3x0 front ?
12:14:16LinusNnot at all, i haven't had time too even look at the h300 for a long time
12:14:24XavierGrstrange, when I try to make rockbox.iriver a hidden file, Rockbox can't find the file and original firmware starts. Is that intended?
12:16:49LinusNdon't think so
12:17:16XavierGryou mean I am wrong or maybe a bug?
12:17:28preglowLinusN: can you think of any reason why the h3x0 lcd should have a custom controller as well?
12:17:46preglowno, me neither
12:19:12LinusNXavierGr: maybe a bug
12:19:41LinusNreboot, brb
12:19:44 Part LinusN
12:19:44*XavierGr imagines the beautifull magenda Rockbox logo on a H300!!
12:19:59BgerXavierGr do u have h3x0 ?
12:20:08markunLinusN: Do you think unicode support in rockbox is a good idea?
12:20:30Bgeri don't think anyone objects this ...
12:20:35XavierGrBger: No, markun: He just quit.
12:20:45markunAh :)
12:23:02 Join LinusN [0] (
12:23:17XavierGrSo, who actually made and drawed the Rockbox logo?
12:23:21markunLinusN: Ah, there you are..
12:23:41markunLinusN: Would you object to adding unicode support to rockbox soon?
12:24:03preglowmarkun: does it add much to the binary?
12:24:46LinusNmarkun: unicode is welcome, if it works well
12:24:52markunThe size of yesterday's rockbox.iriver is 253268
12:24:59markunWith unicode..
12:25:02preglowwell, uyeah
12:25:03preglowbut archoses
12:25:11preglowwe're very tight there right now
12:25:34markunHm, didn't think of that. We can have a HAVE_UNICODE define of course..
12:25:56preglowwell, not having it would be nicer, heh
12:26:12LinusNmarkun: which patch is it?
12:26:16markunphaedrus961 is working on all kinds of codepage conversions.
12:26:38markunLinusN: The unicode patch in the tracker, but I don't think it will apply cleanly.
12:27:56markunI don't like that I have to "cvs add" files before they can be included in the patch with "cvs diff -N"
12:28:21LinusNme neither
12:28:24markunLinusN: If you like you can have my to test.
12:28:44LinusNi'm only interested in the source
12:29:51markunShould the archos player also use UTF-8 for it's strings?
12:30:21amiconnOtherwise we would have to maintain the .lang file in different encodings per target
12:31:02amiconnPlus, some day I will add cyrillic and greek support if at all possible
12:31:12markunGood luck :)
12:31:40markunYou only have 4 free chars on some players, right?
12:31:53 Quit Musicmad ("Trillian (")
12:31:54XavierGradd where?
12:32:11LinusNmarkun: but we remap them dynamically
12:32:12XavierGrisn't there already a greek lang?
12:32:41markunXavierGr: There is, but it doesn't work on the old archos players.
12:32:52LinusNXavierGr: yes, but it doesn't display correctly on the character LCD models
12:33:13XavierGrI got it LCD_CHARCELLS
12:33:26amiconnmarkun: I think it's possible. The old LCD has less definable chars, but more that are re-usable
12:34:06XavierGramiconn was talking about making a greyscale lib for the old players and porting jpeg.c for it too in a resolution of 14*20 I think!!! XD
12:34:38markunToo bad I can't get the simulator to compile on FreeBSD to work on archos unicode.
12:35:42amiconnXavierGr: Yes, I was saying that the same technique used for grayscale on archos recorders, Ondios and H1x0 should work on the player lcd too
12:35:58amiconnI wasn't saying that I will actually do it
12:39:16markunI wonder howmany names in the credit file are spelled wrong because it uses latin1.
12:40:32Bgeri don't think there are many
12:40:39XavierGramiconn: of course just mentioning...
12:44:12|Lupin|Ok, now Rockbox is installed on my new Iriver.
12:44:24|Lupin|How can I go to the menu to check that RCkbox speaks ?
12:44:52|Lupin|(I'm blind)
12:45:17preglowit should be speaking be default if you remembered to copy a voice file to your iriver as well
12:45:35amiconnXavierGr: Btw, the colour rockbox logo is black & grey on orange
12:46:01XavierGrwell yes magenda orange whatever.
12:46:08XavierGrand who made it?
12:46:20|Lupin|I did that: copied an english.voice file to the .rockbox directory...
12:47:17|Lupin|When One starts rockbox, is itSupposed to say something ?
12:47:22preglowahh, i needs to be in the langs directory which is inside the .rockbox directory
12:47:53preglowno, it doesn't say anything when you start it, you need to press the menu button or do some other action
12:49:13preglowamiconn: do you know if iriver voice gui supports talking dirs and filenames?
12:49:35|Lupin|It doesn't
12:49:43|Lupin|Well it can spell them out ?
12:50:10preglow|Lupin|: yes, but that requires further preparations
12:50:57amiconnpreglow: It can either say the number and file type, spell them, or play the associated .talk clip the same way as on archos
12:51:05markun|Lupin|: Would text to speech be a useful feature for you on rockbox?
12:51:31preglow|Lupin|: but yes, if you have copied the voice file to the .rockbox/langs directory and not changed any vital settings, rockbox should be speaking after you start it and do some action
12:51:41amiconn|Lupin|: The englsih.voice file must be in /.rockbox/lang
12:51:51markunthe ipodlinux guys got flite (festival light) working.
12:52:08preglowmarkun: yes, and we should most definitely play around with that as well
12:52:50|Lupin|markun: Definitely.
12:54:12amiconnRecently I get the impression that work on many things is started, but most of them never really finished, instead work starts on yet another thing...
12:54:23markunIt sounds quite ok. I have listened to a lot of texts for fun with festival and you get used to it :)
12:54:29preglowyes, that is the way of interest driver development
12:54:31amiconnThis way rockbox will never become stable enough for a release
12:54:36amiconn(on iriver)
12:54:42amiconnJust my 0.02
12:55:27preglowamiconn: btw, your new catch mem accesses commit, how does it work?
12:55:52amiconnIt works similar as on archos; I should update the wiki article about it...
12:57:21amiconnBasically, the user break controller (SH1) and the debug module's breakpoint logic (coldfire) allow to fire an exeception on certain memory accesses
12:57:38amiconnYou can try it live:
12:57:50preglowyeah, i kind of figured that, just wondered how the access is trapped
12:58:20amiconnWrite a primitive plugin that tries to read from address 0 or close to that
12:58:46amiconnSet 'Catch mem accesses' to 'Zero area (all)', then run the plugin
12:59:02amiconn-> I0C (Debug) at <address>
12:59:44preglowyep, that's what i wanted to know
12:59:47preglowindeed a useful feature
12:59:54amiconnAs mentioned in the commit msg, using Play to reboot doesn't work, you need the paper clip
13:00:24amiconnAlso, catching writes doesn't do anything, because the bus just hangs on write
13:01:21preglownow somone do a rombox for iriver ;)
13:01:29amiconnLinusN: What do you think? Should we remove the ROM write protection from crt0.S ?
13:01:43amiconnOr should the memguard do it when enabled?
13:02:05preglowwell, you can't really write to the rom that way anyway, can you?
13:02:08amiconnpreglow: Perhaps not rombox, but I certainly want rockbox in flash
13:02:19LinusNamiconn: do that
13:02:24preglowamiconn: which is exactly what i meant, excuse my ignorance
13:02:44amiconnpreglow: No, but if writes are allowed, you can switch the ROM to programming mode
13:02:45preglowhavent dealt too much with it
13:02:55amiconnLinusN: Which of the 2 options?
13:03:20amiconnBtw, if we want to flash some day, we'll need to remove wp anyway
13:03:25Slasheriamiconn: Hmm, i think it can be good to have the protection as default on and removed only when needed
13:03:38amiconn...even for some debug functions, like reading the flash id
13:05:19preglowSlasheri: so, how's recording coming?
13:06:45Slasheripreglow: well, there is not much left before i could commit it but currently i don't have almost time at all to work with it :/
13:07:01LinusNi think it should be write enabled
13:07:46amiconnLinusN: Okay, will do it. Just a single bit change :)
13:08:30amiconnpreglow: Regarding rockbox in flash: On archos we have two methods:
13:08:50SlasheriHmm, is there then any increased risk of accidential flash write if the system crashes? I hope the flash still requires the initialization commands before it accepts write
13:09:11preglowSlasheri: it can't possibly imagine it being any other way
13:09:16amiconn(1) Rockbox compressed in flash, decompressed into RAM and executed from there. Needs less ROM space than (2)
13:09:46Slasheripreglow: ok, good
13:09:49amiconn(2) Rombox, meaning the code is uncompressed, and directly running from ROM. Only .data and .bss are in RAM
13:09:55preglowSlasheri: programming flash is always a widely different procedure than writing to ram, in my experience
13:10:11amiconnNeeds less RAM, and slightly less boot time (no decompression), but more ROM space
13:10:16Slasheriyep, it should be except if the cpu does some for you
13:10:31preglowamiconn: i wouldn't mind 2, but how fast are rom accesses on iriver?
13:10:45amiconnYes, that's the interesting question.
13:11:14amiconnEven on archos there is a speed penalty, because the RAM has 16 bit bus and the ROM has 8 bit bus
13:11:28amiconn...but it pays off, because of the bigger RAM buffer
13:11:29preglowwell, iriver has 16 bit buses for both, at least
13:11:51amiconn...and the most time critical code is in IRAM
13:12:00preglowyes, iriver does it like that as well
13:12:37amiconnIt might be that it is actually faster to run from ROM on iriver, but that needs to be tested
13:13:37amiconnThe DRAM is sooo slow when burst mode isn't used... but then icache line reads are using burst...
13:14:15preglowi distinctly remember linus saying rom was faster than ram on iriver
13:14:24preglowbut i also remember him saying that that's not true :)
13:15:06amiconnOn archos, we also have dual-boot for safety
13:15:40amiconnThere are 2 firmware images in flash, selectable at boot time by a button press
13:16:10amiconnThe first one is the backup. It used to be the original archos firmware (compressed), but is now replaced with bootbox
13:16:15markunamiconn: Is the iriver firmware compressed?
13:16:52amiconnOnly the second image gets overwritten by day-to-day flashing, so if that fails, the first image can be selected on boot and would then read rockbox from disk
13:17:28amiconnThis .ucl compression is quite effective
13:17:52amiconnMaybe we could use the same on iriver to compress the original firmware and make room for rockbox
13:17:53preglowyes, and small decompressor if i remember correctly
13:18:32preglowsure, but how do we execute the original firmware when it expects to be executed from flash? modify the memory map slightly?
13:18:51amiconnAfaik the original firmware copies itself to ram before execution
13:19:02preglowin its entirety?
13:19:26amiconnIn several pieces I think, but I didn't check myself
13:19:30markunJust noticed the unscrambled iriver firmware compresses a lot better with gzip than the scrambled one.
13:19:32preglowme neither
13:19:43preglowof course
13:19:47preglowit's not scrambled in rom anyway
13:19:53amiconnWe would have to replicate this functionality
13:20:01amiconn(the copying)
13:20:13preglowand then just use a different entrypoint to the flash?
13:20:17preglowiriver firmware, i mean
13:20:42preglowsounds decent
13:20:52amiconnI'm not sure whether it's worth it though
13:21:11amiconnWe could just have bootbox on iriver too, and drop the original firmware
13:21:30preglowindeed we could
13:21:36preglowi'll never use it again, that's for sure
13:21:49markunBut we will need our own flash util then
13:22:00 Quit AliasCoffee ("Leaving")
13:22:35markunhow do you flash on archos?
13:23:16preglowamiconn: but we've already got a very stable bootloader, we could just include that anyway and include a couple of safety images on the disk, the iriver firmware being one
13:23:37preglowamiconn: of course being generated by fwpatcher or some such
13:23:38amiconnmarkun: The flash procedure of the flash rom is documented
13:24:09amiconnpreglow: Basically the iriver rockbox bootloader already does what bootbox does on archos
13:24:23amiconn(allowing usb access and starting rockbox from disk)
13:24:54preglowbut anyway
13:24:59amiconnBootbox on archos recorder v1 does one additional thing - charging, as the v1 has software controlled charging
13:25:06preglowtwo megs of flash should be more than enough for both the compressed firmwares
13:25:18preglowor just uncompressed rockbox and compressed iriver fw
13:26:03amiconnYes, but the question is whether it's worth to do the reverse engineering needed for having the iriver fw compressed
13:26:17preglowi think it's already been done
13:26:29preglowthere was a page up somewhere that described how the iriver firmware booted
13:26:40preglowif someone got that far, they can give us another entry point as well
13:26:56preglowassuming it's as simple as it copying itself to ram, of course
13:27:11preglowif it does a lot of chunk copying it'll be worse
13:27:34preglowas for me, i don't care about iriver fw, but it is another layer of security for when things will one day go wrong
13:27:45***Saving seen data "./dancer.seen"
13:27:57 Join pengo [0] (i=xtofu@
13:28:25 Join Zagor [0] (
13:30:06amiconnBagder, Zagor, LinusN: Could one of you pls delete the wiki spammer...
13:30:56Bgerwhy don't use e-mail validation for twiki registration ?
13:31:07LinusNafk for a while
13:33:10|Lupin|Ok, guys, I've to leave, since I'm at theoffice.
13:33:21|Lupin|But I'll certainly come back either tonight or in a few days
13:33:47LinusNamiconn: which spammer?
13:33:49|Lupin|For instance to talk about text-to-speech
13:34:12 Part |Lupin|
13:35:36 Nick TiMiD[sleepingAw is now known as TiMiD (n=TiMiD[
13:35:36DBUGEnqueued KICK TiMiD
13:35:38Bgermorning :)
13:36:49amiconnLinusN: SinerAram
13:38:09TiMiDamiconn: this guy is on many wikis
13:38:16TiMiDjust look on google ...
13:41:58XavierGrTiMiD: You are awake, I am getting sleepy now.... I can hardly keep my eyes on, ZZZzzzzzz....
13:43:06TiMiDyou should have slept
13:43:30TiMiDjust sleep 4 or 5 hours
13:44:26XavierGrI would have miss my lesson, but yeah I shoudln't go anyway, there were no sits to sit down, and the proffessor was saying things I already new, boooriiing...
13:45:04XavierGrF*cked up Greek education!
13:45:29TiMiDsame here :)
13:45:39XavierGrI guess I will leave you now for a little noon nap.
13:46:04XavierGrOh and I got lesson again in 17.00
13:46:23TiMiD1h of nap can rest you like full night
13:47:28XavierGrI miss the summer days where I could sleep 12 hours a day :)
13:56:25 Join {Mike} [0] (
13:56:35{Mike}any devs who are here:
13:56:42{Mike}You've done a bloody good job, nice one
13:59:21preglowoh, they know ;)
14:04:36TiMiDbtw, has any core dev looked at my remote patch ?
14:14:29preglownot that i know of
14:17:20preglowi don't know enough about those parts of rockbox to do it myself
14:19:52TiMiDthey are not interrested by remote ?
14:20:12preglowthey don't have too much time on their hands
14:20:25preglowand possibly that as well, i don't use the remote myself
14:22:06TiMiDI just want to know if what I did was good or bad :p
14:23:44 Quit pengo ("Curious? Open binary with Qubero")
14:41:51 Quit Febs (Read error: 110 (Connection timed out))
14:46:33 Quit yosemite (
14:46:49NJoinyosemite [0] (
14:53:08 Quit ashridah ("Leaving")
14:57:48 Join |Lupin| [0] (
14:57:52|Lupin|Hello again, guys.
14:58:38|Lupin|May I ask what's this fc*p utility which is used to build the .html doc from the .t files, plS ?
14:59:00 Quit |Lupin| (Client Quit)
14:59:49LinusNit's a modified C preprocessor
15:00:04CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
15:00:04*preglow prods slasheri
15:02:24 Join |Lupin| [0] (
15:02:46|Lupin|Hmm disconnected due to a wrong command. Did someone reply to my question in the meantime ?
15:02:53|Lupin|(s)rry for the trouble, all)
15:02:59preglowyes, linus said:
15:03:00preglowit's a modified C preprocessor
15:03:20|Lupin|Is it possible to download it from somewhere ?
15:05:22|Lupin|Thanks :)
15:08:48 Quit linuxstb__ (Remote closed the connection)
15:12:36|Lupin|And what's about themanual ? Is the version from the CVS newer than the one inthe wiki, or the contrary ?
15:13:50Bgerin the wiki is the most recent version
15:14:09Bgerbecause it's the easiest to edit
15:14:16LinusNthe cvs manual is old as hell
15:16:07|Lupin|ok. Thanks Bger and LinusN. I was surprised by finding nothing about Iriver, there, hence the question.
15:19:00Bgernp, |Lupin|
15:21:36TiMiDLinusN: did you looked at the remote patch I submitted ?
15:25:56LinusNTiMiD: haven't had time, but i intend to
15:26:22preglowargh, more mixing of spaces and tabs :/
15:27:23preglowthe codecs are such a frigging mess of styles
15:27:28TiMiDno problems :p
15:27:46***Saving seen data "./dancer.seen"
15:28:35preglowLinusN: i take it USE_IRAM is a preferable define to check for than SIMULATOR when doing the iram init stuff?
15:30:11preglowwhy perhaps? why is it there if it's not
15:30:49LinusNi honestly don't know the purpose of it, but i guess it might be for this situation
15:31:27preglowcan't say i care much either, i just know USE_IRAM exists but is almost never used for the one purpose it has
15:32:06Bgermaybe somewhere should be a list with such defines ...
15:32:13Bgersomewhere == wiki ?
15:32:28amiconnpreglow: apps/plugins/rockboy/rockboy.c line 117
15:32:52TiMiDLinusN: btw, I have a question about the bootloader
15:33:01preglowamiconn: 'never' was an overstatement, i've used it myself when i have the chance
15:33:11TiMiDis the ata check still enabled in it ?
15:33:20TiMiD(cf this thread :
15:34:21preglowand btw, i think codecs should never include anything but codec.h and codeclib.h, anyone agree?
15:34:58preglowthat is, apart from the libraries they need, that is
15:35:00preglowlike libmad, etc
15:35:11preglowright now they include playback.h and dsp.h alot as well
15:35:35amiconnpreglow: you're right
15:36:20preglowthink i'll just spend some time cleaning up the codecs
15:36:51preglowok, i'll stuff whatever else is needed in codec.h instead, the only thing codecs currently need from playback.h and dsp.h, are some defines
15:37:25amiconnIt's the same with plugins
15:37:55amiconnPlugins should only include plugin.h, and the .h files from apps/plugins/lib they need
15:39:42preglowcodec.h is just a couple of support routine declaratiopns
15:39:49preglowcodecs.h is what they should include
15:39:57preglowbut that's currently only included by codeclib.h ...
15:40:17TiMiDI think about one thing
15:40:56TiMiDcould sine and cos functions be integrated into a lib
15:41:11TiMiDI'm talking about the ones in plasma and cube
15:41:12preglowi don't get the current codec header system
15:41:22preglowTiMiD: yes, i've thought about that
15:41:43preglowTiMiD: i can code a nice interpolated sine routine in a jiffy
15:41:51TiMiDsince I was working on another plugin which also uses them
15:42:17preglowproblem is just what form the argument and return value should be in
15:42:30preglowi'd like the argument to be from 0 to 0xffffffff
15:42:35preglowwhich represents 0 and 2*pi
15:42:40TiMiDisn't there also some trigonometrics fn in the codecs code ?
15:42:47preglowand gives lots of precision
15:43:15preglowreturn value should be value from -2^31 to 2^31
15:43:21preglowthis is just the way i'd like to use it
15:43:23TiMiDthat is not a problem, you just have to generate the table :)
15:43:30preglownah, you precalc that
15:43:33amiconnThe current implementation for plugins uses way less precision, but is sufficient for gfx stuff
15:43:55amiconnargument in degrees, result in 18.14 fixed point
15:44:06preglowyeah, argument in degrees is a bad thing, if you ask me
15:44:20TiMiDI agree too
15:44:23preglowbecause it doesn't offer enough granularity
15:44:30preglowi'd like this sine routine to be used in dsp as well
15:44:41preglowwhere i'll need lots more precision than 0..360 integer
15:45:50amiconnHigher precision will be slower, and the degree->whatever conversion will add even more
15:46:00amiconn-> Not recommended for gfx, imho
15:46:04preglowno, you just need to recalc the deltas you use somewhat
15:46:28preglowthe sine routine in itself will be very high quality if it's just a linearly interpolated lookup with 128 samples in the table
15:46:53amiconnThe current plugin sine doesn't interpolate at all
15:46:58preglowi know
15:47:15preglowbut it does a low of branching which makes it slower than a plain interpolated lookup
15:47:35preglowforget it, that's just the precalc
15:47:39amiconnIt does exactly 2 conditional branches, and a table lookup
15:48:22preglowthree branches, one mod and a table looki
15:48:24preglowlookup <-
15:48:34preglowthat is, two branches and one conditiional
15:49:05preglowthe mod in itself isn't cheap, which is exactly why i want to use the full range of the argument, so we don't need to handle that
15:49:18preglowplus it gives lots of nice precision for audio work as well
15:51:02 Quit MasteR` ("I'm a quit message virus. Please replace your old line with this line and help me take over the world.")
15:51:05|Lupin|The Quick Menu can be accessed only while playing ?
15:51:41 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
15:51:43 Quit thegeek (Read error: 104 (Connection reset by peer))
15:53:00preglowbut whatever, a lower precision one will do just fine for gfx, but we'll also need it for audio sooner or later, that's almost certain
15:54:04preglowi did a nice full precision 32 bit sine function for coldfire using 512 bytes of ram, six simple bit arithmetic instructions and a 64 bit mul
15:54:05amiconnI already thought about a generic table & interpolation function.
15:54:17preglowthe mul doesn't have to be 64 bit either, but that loses some precision of course
15:54:21amiconnI want to improve the log() and exp() for the peakmeter
15:54:55preglowehh, i forgot two memory accesses of course ;)
15:55:15amiconnThe 64bit mult takes quite some instructions... you could reuse my routine from mandelbrot.c
15:55:27amiconnIt's somewhat faster than gcc's muldi3
15:55:41preglowyeah, i can imagine, i've just done it with the emac unit thus far
15:55:57amiconnI tried emac as well, but it wasn't any faster
15:56:07preglownot if you want the full 64 bit result, no
15:56:19preglowbut for sine() the 32 bit high bit result will be more than enough
15:56:54preglowbut i could check and see if a 32 bit mul is enough
15:58:13preglowdoing a quick test now
16:06:40preglowlooks completely decent
16:08:03*amiconn just found a superfluous instruction in his 64 bit mul...
16:13:08preglowwith a 32 bit mul, the sine has a first harmonic at -80db
16:15:02 Join Paul_The_Nerd [0] (
16:18:38 Quit B4gder ("CGI:IRC")
16:19:55 Join webguest80 [0] (
16:24:00|Lupin|(How) Is it possible to play radio with RockBox on an Iriver H120, please ?
16:24:11|Lupin|I can't findthe eoc for this in the Rockbox manual.
16:24:47Slasheri|Lupin|: you can find the radio from the settings menu
16:25:14 Join thegeek [0] (
16:25:31|Lupin|oh !
16:25:33|Lupin|Ok, thanks.
16:28:19 Join tvelocity [0] (
16:29:20preglowSlasheri: you got time to look at the mpc.c source code to see if i'm doing something wrong? if not, there's most definitely something wrong with the dsp interleaved setero handling
16:29:38preglowif so, it's here:
16:29:51preglowhad a look at dsp.c and can't say i see anything obviously wrong
16:36:35preglowbut something is wrong, the right channel is full amplitude noise, and the left channel is correct speed, but half pitch and too low amplitude
16:36:45preglowthe disk writing i do there gives a perfect result
16:41:22Slasheripreglow: ah, checking that soon :)
16:41:31 Join Paul_The_Nerd_ [0] (
16:41:59Slasheripreglow: hehe, that sounds like a common problem.. i had always that kind of things while coding the dsp long time ago ;)
16:42:29preglowyes, i've had tons of them myself when doing similar work
16:47:30webguest80w00t so how's mpc coming along?
16:47:54Slasheripreglow: just tried and the pcm output sounds pretty fun :D
16:48:03preglowSlasheri: yes, yes it does, heh
16:48:12Bger dj scratch ? :)
16:48:25preglowwebguest80: badly, i'm just refurbishing the codec plugin, i might add seeking to see how bad that is, though
16:53:28 Quit Paul_The_Nerd (Read error: 110 (Connection timed out))
17:06:40 Nick Paul_The_Nerd_ is now known as Paul_The_Nerd (
17:13:49 Join _FireFly_ [0] (
17:15:37 Join solexx___ [0] (
17:17:26 Join dpassen1 [0] (
17:23:03Slasheripreglow: fixed
17:23:10Slasherithere was two problems in mpc.c:
17:23:35Slasherici->configure(DSP_SET_SAMPLE_DEPTH, (long *)(MPC_FIXED_POINT_SCALE_SHIFT - 1)); and while (!ci->pcmbuf_insert((char *)sample_buffer, status*sizeof(MPC_SAMPLE_FORMAT)*2))
17:23:41Slasherijust do those changes and it will work :)
17:25:28preglowthe first one is too stupid to be true
17:25:32Slasherihehe :D
17:26:06preglowbut why the last *2 ?
17:26:26Slasheriif interleaved, it wants the real byte length
17:26:35preglownot exactly intuitive...
17:26:37Slasheriand with non-interleaved mode the byte length / 2
17:27:09preglowcomment that somewhere
17:27:26Slasherihmm, good idea :)
17:27:47***Saving seen data "./dancer.seen"
17:28:16preglowi'm going to include dsp.h and playback.h in codecs.h
17:28:32preglowsince codecs really shouldn't need to include them explicitely
17:29:51 Quit solexx_ (Read error: 110 (Connection timed out))
17:32:23preglowSlasheri: thanks for finding my bugs anyway
17:32:33Slasherinp :)
17:34:21preglowHWCODEC == SWCODEC on h1x0, yes?
17:34:49_FireFly_afaik is h1x0 SWCODEC
17:35:29preglowi added dsp.h and playback.h to the #if (HWCODEC == SWCODEC) part in codecs.h, still some codecs refuse to compile
17:35:50amiconnpreglow: CONFIG_CODEC, not HWCODEC
17:35:53preglowwell, it's quite certainly not true
17:36:03preglowthen what the hell does HWCODEC == SWCODEC mean?
17:36:40amiconnHWCODEC did never exist
17:36:57preglowok, so it's a typo and i should make it #if (CONFIG_CODEC == SWCODEC) instead ?
17:37:48preglowworks great if i do so, at least
17:38:16 Quit {Mike} (Remote closed the connection)
17:39:24 Join winjer [0] (
17:39:40winjeri've got an iriver iHP-140, and i want to run rockbox on it
17:39:45winjerdo i just use the h120 stuff?
17:40:10preglowamiconn: you're the one who added that line, why'd you do that if it'll never trigger? :)
17:40:24amiconnHmm? which line?
17:40:39preglowcodecs.h line 46
17:41:41preglownah, you just touched it last, according to cvs
17:41:58amiconnI didn't add it, I just changed it (overlooking the mistake) :/
17:42:03_FireFly_winjer: yepp
17:42:05preglowit's been there from the start
17:42:20winjer_FireFly_: cool, just copy it on and do the firmware thing?
17:42:22preglowamiconn: so we agree it's supposed to be CONFIG_CODEC == SWCODEC, then?
17:42:25amiconnI changed the right part of the comparison
17:42:34preglowaight, changed and it works just fine
17:42:41_FireFly_winjer: h120 and h140 are the same exept that the h140 has an bigger hd than h120
17:42:52amiconnHowever, I don't see the point of that #if
17:43:06_FireFly_winjer: look at
17:43:14amiconncodecs.h is for swcodec platforms only anyway
17:43:16winjer_FireFly_: wicked thanks
17:43:50preglowamiconn: well, that depends on your plans
17:44:02 Nick TiMiD is now known as TiMiD[starcraft] (n=TiMiD[
17:44:02DBUGEnqueued KICK TiMiD[starcraft]
17:44:13preglowamiconn: or do you want to use a very different codec subsystem for archos codecs when that day comes?
17:44:45amiconnI think it has to be quite different...
17:45:42preglowyes, indeed
17:46:02preglowbut ok, i'll just rip out the entire CONFIG_CODEC, then?
17:46:12preglowthe ifdef, i mean
17:49:42preglowwe can almost play some low quality musepack files now
17:50:16preglowwill anyone shoot me if i kill the goto in the codecs?
17:50:19preglowit just feels... nasty...
17:51:06Paul_The_NerdThere's a goto?
17:51:11preglowoh yes
17:54:05 Quit LinusN ("disconnecting from stoned server.")
17:54:21Paul_The_NerdI have never actually seen one used outside of BASIC
17:54:49preglowi've seen tons of gotos in c
17:54:52preglowsometimes they're nice to have
17:54:59preglowbut i don't think this is one of those places :)
17:55:03 Quit Slasheri (
17:58:25winjer_FireFly_: thanks, it works perfectly
17:58:25winjervery nice
17:58:30 Join yosemite_ [0] (
17:58:56winjeris there any way of getting the current status displayed on the remote, or is that on the wishlist? :)
17:59:21NJoinSlasheri [0] (
17:59:23 Quit yosemite (
17:59:40 Join DangerousDan [0] (
17:59:57_FireFly_i have made a patch wich adds remote-supprt and on,wezel can you find a compiled version
18:00:45winjer_FireFly_: is empty
18:00:53_FireFly_ups s.wezel
18:02:01winjerah yes ;)
18:04:54winjer\o/ works
18:05:04winjer_FireFly_: great
18:05:12 Join LinusN [0] (
18:10:55preglowi just encountered a bug that resulted in the backlight never turning off
18:12:51TiMiD[starcraft]preglow: a goto can sometimes be useful and the best way to solve a problem
18:12:57preglowTiMiD[starcraft]: yes, i know
18:13:05 Nick TiMiD[starcraft] is now known as TiMiD (n=TiMiD[
18:13:05DBUGEnqueued KICK TiMiD
18:13:08 Part LinusN
18:13:22TiMiDhehe :)
18:14:03TiMiDI have a suggestion
18:14:15TiMiDabout the display code in the driver
18:14:40TiMiDa lot of high level fn use exacly the same code for remote and main display
18:15:04TiMiDwhy not to move them to "user space" and make them use the same code
18:15:37TiMiDfor example lcd_remote_vline and lcd_vline
18:15:43TiMiDit's the same code
18:15:55TiMiDso it eats memory
18:15:55_FireFly_not really :)
18:16:12_FireFly_they call different lcd_ functions :)
18:16:32TiMiDit could be a fn ptr :p
18:18:15TiMiDexcepted the low level code that handles registers registers and so on, these are very similar
18:22:59preglowSlasheri: does one have to set seek_time to 0 after a seek is complete? shouldn't seek_complete() do that for you?
18:23:52Slasheripreglow: ah, good point. It should but currently it doesn't..
18:24:10Slasheriso you have manually set it to 0 or fix the code :)
18:24:22preglowi'll fix the code and correct the codecs to use seek_complete
18:24:25preglowsome of them don't
18:24:30Slasherigood :)
18:24:36 Join Luthion [0] (
18:25:51preglowSlasheri: seek_complete doesn't know where it's called from, shall i just assume ci.seek_time is what i need to clear? voice codecs should not need to seek...
18:26:13 Quit Luthion (Client Quit)
18:26:32 Join Lear [0] (
18:26:42Slasheriyes, i think you can assume that. And current_codec will tell you if it's voice or audio codec
18:27:06preglowi just assume ci and add a comment, voice codec should never have to seek, if you ask me
18:28:22 Nick ender1 is now known as ender` (i=ychat@
18:29:09preglowmusepack seeking doesn't look too good
18:29:09Slasheriyep, true
18:34:57amiconnTiMiD: In fact this idea (unifying higher-level code of the lcd driver) was already discussed a while ago
18:35:58amiconnI agree that it can save space on iriver, but I wouldn't want such a unification to make the code bigger for units with just one display
18:36:42amiconnPlus, we should not pull on too many ends at once, the results might be unsatisfying..
18:37:33TiMiDI was thinking about that because with my remote patch, it would be very simple to do this for these high level fn
18:37:57amiconnThere's not that much reusable code
18:38:33TiMiDatleast the code that draws lines and strings
18:38:35amiconnxxx_drawpixel() (tiny fn), xxx_drawline(), xxx_drawrect()
18:38:37TiMiDthe scrolling thread
18:39:07amiconnYes, and the string drawing
18:39:19amiconnScrolling thread can't be unified
18:40:43TiMiDoh why ?
18:40:52TiMiD(the code is also almost the same
18:41:26amiconnYes, but the contexts are different, and the settings too
18:42:08preglowis there some way to pass track length info to the playback system from a codec?
18:42:16preglowi don't want to alter metadata.c to understand mpc files :/
18:42:39amiconn...and when thinking about an improved concept (scrolling boxes, and retaining attributes like colour, font etc), even the context format will be different
18:44:37TiMiDit's sure that the scroll system is not very flexible for pixel displays
19:05:00 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:11:10 Quit |Lupin| ("leaving")
19:27:10 Join Moos [0] (
19:27:48***Saving seen data "./dancer.seen"
19:30:16_FireFly_i have a little patch so that comments are really ignored in wps-files
19:31:29_FireFly_currently they aren't e.g. when the bitmap loading lines are on the end of the file and between this lines comments are used
19:31:32MoosSlasheri: aware?
19:31:40Mooshi Firefly btw
19:33:37Slasheriah, hi Moos :)
19:34:06MoosRe Miika
19:34:08SlasheriMoos: would you like to try now?
19:34:17Moosyes if you want
19:34:31_FireFly_what to try ??
19:34:34Slasheriok, just a moment.. i will try to register with nickserv so i can dcc you
19:34:44Slasheri_FireFly_: little dircache debugging
19:35:24SlasheriMoos: ups
19:35:35Slasheridon't use that file
19:36:00Slasherithat's weird.. dcc sent only 7 kB
19:36:08Slasheritry deleting previous file if you have one
19:36:24Slasheri(in the dcc directory)
19:36:28Mooslet me check where incomed
19:37:37MoosI find nothing
19:37:40Moostry again
19:37:44Slasheriok, trying
19:37:57Slasherino, it doesn't work :/
19:38:33Mooshow did you do for send me files in the past
19:38:46Slasheriwith dcc :) but i will put it to the web
19:38:52Moosa ok
19:39:08Moosor mail if you want
19:41:24LearWhy is -fno-builtin set when compiling the simulator?
19:42:33 Quit tvelocity ("Leaving")
19:45:07 Join linuxstb [0] (
19:47:59linuxstbpreglow: You should just be able to set id3->length (an integer number of milliseconds) from mpc.c
19:48:31linuxstbI mean ci->id3->length
19:48:46preglowlinuxstb: will try that
19:49:29linuxstbBut how much work would it be to add it to metadata.c ? What does an mpc file look like?
19:50:10preglowyou can have a look at streaminfo.c in libmusepack
19:50:15preglowthat's what's used to extract the information
19:50:24preglowi've already added ape tag reading for mpc files to metadata.c
19:50:42 Join tvelocity [0] (
19:51:01_FireFly_it would be nice if ape tags can be also used in flag files :)
19:51:22linuxstbWhy? What's wrong with vorbis comments?
19:52:08_FireFly_oh foget it i have found the cmd-line tool to show vorbis comments
19:52:12preglowlinuxstb: that's 32 bit int, yes?
19:53:16linuxstbpreglow: It's defined in firmware/export/id3.h - "unsigned long"
19:53:28linuxstbSo yes - 32-bit int.
19:53:55amiconnLear: (-fno-builtin) We want our own implementations for standard functions to be used in the simulator as well
19:54:27LearHm.. I see. But it makes it harder to use a proper alloca there then...
19:54:46preglowlinuxstb: worked just fine, thanks
19:55:00preglowLear: why? it's called __builtin_alloca there as well
19:55:33LearWhen -fno-builtin is set, __builtin functions aren't available...
19:56:00LearIf I read the manual right... :)
19:57:09LearPerhaps not, but still the problem is that __builtin_alloca does not work in the simulator...
19:57:38linuxstballoca manpage: "The alloca function is machine and compiler dependent. On many systems its implementation is buggy. Its use is discouraged."
19:58:08preglowLear: the __builtin functions are still there, it's just that they're prefixed with just that... __builtin
19:58:09Learlinuxstb: Which implementation is that?
19:58:55linuxstbI'm not sure, but it's a GNU manpage.
19:59:06linuxstb"man 3 alloca" in Debian.
19:59:24preglow__builtin_alloca works just fine for gcc
19:59:27preglowso this doesn't affect us
19:59:53Learpreglow: Then explain why "#define alloca __builtin_alloca" works fine on target but not in the simulator.
20:01:37linuxstbLear: Is it just Tremor that uses alloca ?
20:01:38Learpreglow: think I found it, a stray alloca declaration in Tremor/misc.h...
20:01:53Learlinuxstb: as far as I know, yes...
20:02:07preglowDon't recognize built-in functions that do not begin with __builtin_ as prefix.
20:02:10preglowthat's what the manual says
20:03:02linuxstbWouldn't it be better to try and get rid of the alloca calls in the first place?
20:03:24linuxstbIt seems an easy way to get a stack overflow.
20:03:29preglowok, i've got musepack seeking working
20:03:31preglowbut it's _SLOW_
20:03:38preglowshould i even bother commiting it?
20:03:46preglowlinuxstb: using the stack is good
20:03:54preglowlinuxstb: and alloca use is pretty rare as it is
20:03:55Learlinuxstb: in the Tremor case, it isn't that easily done... It was written like that for a reason...
20:04:15Learlinuxstb: and it is mainly only used for smallish allocations
20:04:21linuxstbBut shouldn't they be statically allocated on the stack?
20:04:29Learsorry, "mainly used for smallish allocations".
20:04:32preglowyou can't statically allocate something on the stack
20:04:36preglowthat's not possible
20:04:58linuxstbBad choice of word - but I think you know what I mean?
20:05:04preglowhmm, not sure
20:05:06Learstatic size allocation, I guess he means, but for tremor, that would often waste stack space, I guess, if it was easy to know the max size...
20:05:13preglowahh, right
20:05:14preglowi get it
20:05:29linuxstbLear: Why is the stack wasted? We have to make the stack large enough for the worse case.
20:05:40preglowlinuxstb: and as of now, tremor isn't even close
20:05:44preglowbut yeah
20:05:54preglowvorbis is written to allow massive encoder changes without a decoder change
20:06:02preglowthat's why tremor is so vague with its max limits
20:06:09Learit would often allocate more stuff than necessary...
20:06:11preglowthere are none
20:06:30Learpreglow: btw, did you know that TREMOROPTS isn't used in the builds? Not that it makes much of a difference... :/
20:06:36linuxstbBut we should enforce some limits based on current encoders - otherwise we can get stack overflows in the future.
20:07:14linuxstbEven if they are just #defines. The FLAC decoder has some of those - something like MAX_SUPPORTED_BLOCKSIZE
20:08:24preglowLear: didn't know that, no
20:08:33 Join matsl [0] (
20:08:45preglowbut anyone, should i commit mpc seeking even though it's dog slow?
20:08:59LearEasy to tell; build output should include (Tremor), but doesn't... The rule is wrong.
20:09:12LearIs all of MPC slow, or just the seeking?
20:09:38linuxstbpreglow: Yes, I think you should. Unless you think your approach is fundamentally wrong.
20:10:30preglowno, no, perfectly logical and simple
20:10:34preglowit's libmpcdec that is at fault
20:12:59linuxstbSo what's the problem with libmpcdec?
20:15:11preglowthe format plain and simple doesn't support seeking
20:15:25preglowi knew this from before, just had to try it to believe it
20:15:56preglowno seektable, no nothing
20:16:13preglowit uses an aeon to search thirty seconds into the future
20:16:18preglowgimme a sec, and i'll commit it
20:18:42CtcpIgnored 2 channel CTCP requests in 29 minutes and 3 seconds at the last flood
20:18:42*amiconn successfully read his H140's flash chip maufacturer and device id :)
20:18:56amiconnHad to correct the flash wait states for this to work
20:19:18preglowthen hooray!
20:19:34Slasheriamiconn: hehe, and the next step would be to erase and write some unused (or original firmware) flash areas? :)
20:19:46amiconnPerhaps, but not now
20:20:16amiconnI just wanted to complement the removal of the flash write protection with something minimal useful
20:20:46amiconnIf I only knew whether the H1x0 flash chips are 70ns or 90ns...
20:21:39SlasheriHmm, isn't there a rdy pin or something like that you could probe if the chip is ready?
20:27:20 Quit ghode|afk (Read error: 104 (Connection reset by peer))
20:28:32MoosSlasheri: just back
20:33:10 Join ghode|afk [0] (
20:33:13 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
20:35:42MoosSlasheri: still here?
20:36:09Moosit's maybe one runtime db bug for your dircache :-(
20:36:49Moosstill freezed
20:38:39 Nick paugh is now known as AliasCoffee (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
20:42:47_FireFly_argh bad ripping/tagging tool to rip/tag from cd to flac :)
20:43:08Paul_The_NerdWhat are you using _FireFly_?
20:43:48_FireFly_the tool had set the vorbis comment tags lowercase so metaflag didn't see the comments bad after retagging the files with easytag it works with metaflag :)
20:45:43_FireFly_a similar problem was with grip and ogg-files
20:46:01_FireFly_these ogg-files produced codec-failure on my h120
20:46:38_FireFly_ogginfo said that there where a hole in the file but with easytag i could "repair" these files
20:49:00Paul_The_NerdSo, easytag would be a good first program for me to remember to acquire.
20:49:35_FireFly_Paul_The_Nerd: the tool which i had ripped the flac files is rip(
20:49:53_FireFly_a simple retagging with easytag :)
20:50:04_FireFly_easytag had no problem to read the tags
20:50:30Paul_The_NerdRight now I'm using windows, because I'm waiting to be able to use my laptop's wireless properly in linux before I switch over.
20:51:22_FireFly_which chip is used in the wireless ??
20:51:50Paul_The_NerdI honestly don't know.
20:52:04Paul_The_NerdI didn't bother to check... I'd already read the only known way to make it work was ndiswrapper
20:52:37Paul_The_NerdWhich doesn't work in linux compiled to run with an AMD64 under 64-bit mode, apparently.
20:56:08_FireFly_oh i was wron rip set id3tags in the flag files
20:59:54Paul_The_Nerdid3 in the flacs?
21:01:04 Quit Moos (Read error: 110 (Connection timed out))
21:02:36preglowSlasheri: what is the set_offset call?
21:05:35 Join Coldtoast [0] (i=tor@tor/session/x-54364c3fea0da5fb)
21:06:34Coldtoastinteresting! I though tyou pretty much got klined on EFnet for using TOR
21:12:24_FireFly_is someone interesting about my patch for really ignore comment-lines in wps-files ??
21:13:09Paul_The_NerdDoes that make them not count against size limits?
21:13:57_FireFly_currently not because i check the comments in wps_format()
21:15:32_FireFly_to ignore comments when loading the wps_loaf function must be changed to read each line insteed to read the file complete
21:16:52amiconnThe flash rom does indeed look unscrambled
21:17:05*amiconn added ROM dump for H1x0 as well
21:22:33Paul_The_NerdAah, so what benefit does the true ignore add _FireFly_?
21:24:57_FireFly_if you have your bitmap loading lines(%x %xl) are on the end of file, you have some comments bewteen this lines and you don't use all viewable lines to display something, then some of these comments are shown while in wps
21:25:05preglowSlasheri: really, i think the number passed to pcmbuf_insert should always be samples per channel :>
21:25:08_FireFly_i hope it is understandable
21:26:06 Join actionshrimp [0] (
21:26:11Paul_The_NerdSo it is beneficial. Though it's still not a good idea to put comments in a WPS anyway, though.
21:26:14Slasheripreglow: hmm, you might be correct :) i don't know why but long time ago i think it could be better to use bytes instead of samples
21:26:23Slasheripreglow: set_offset updates the id3 elapsed information
21:26:57_FireFly_i try to change wps_load so that comments won't be included in the temp format-buffer for the wps
21:27:10 Join Moos [0] (
21:27:49***Saving seen data "./dancer.seen"
21:28:09Moosconnexion problem at home :(
21:28:24MoosSlasheri: scuse me, still here
21:28:57Slasherihi Moos :) gotta go sleep soon
21:29:40Moosirc client sucks here :(
21:29:51Slasherihmm :/
21:30:02Moosdid you identified more precisely the problem?
21:30:21Moosthe runtime problem I mean
21:30:21Slasherino, because i don't use the runtime db.. you really should consult HCl
21:30:30preglowSlasheri: and why should i call that if i've also called set_elapsed?
21:31:16Slasherioh, weird.. there were two functions
21:31:24Learinteresting detail: in filetree there's a big if statement with lots of "(dptr->attr & TREE_ATTR_MASK)" cases that aren't optimized away (on m68k at least)... :)
21:31:27MoosSlasheri: ah, but your code need to be compatible with runtime db, no? :-)
21:31:53Slasheripreglow: i think i haven't even added that offset thing.. but still looking what it does ;)
21:32:13SlasheriMoos: yes, or rundb less buggy. I hope hcl finds out the problem
21:32:17 Join banan_ [0] (
21:32:26Mooswhen I tuned runtime off, the bug didn't apear, it's better than yesterday :D
21:32:31 Join thegeek_ [0] (
21:32:38 Join muesli- [0] (
21:32:41Moosnow I can use the dircache again
21:33:02Moosbut not the runtime, HCl ? :-)
21:33:06HClwho what?
21:33:15MoosHello HCl
21:33:20Moosand muesli
21:33:23HClwhat bug in the runtime db? O.o
21:33:28HClruntime db is pretty clean as far as i know.
21:33:34Mooswith the new dircach code
21:33:58 Join banan__ [0] (
21:33:58muesli-hiho _FireFly_ Moos, hcl and all i forgot...
21:34:03Moosno just with last Slasheri changes, your runtime db sucks :D
21:34:09Slasheripreglow: hmm, i think that offset thing has something to do with resume.. but now i have to go :/
21:34:20HClwhats the dircache and what does it do?
21:34:41muesli-hcl..your beyond the times ;)
21:34:42 Quit thegeek_ (Read error: 104 (Connection reset by peer))
21:34:51HClno, just been busy with other stuff
21:34:52HClmostly gf
21:34:57 Join thegeek_ [0] (
21:35:03Mooshehe apear you don't look Rb news those days ;-)
21:35:06muesli-you have only one gf and thats your iriver ;)
21:35:57HClwhats the dircache and what does it do?
21:35:59MoosHCl: your new java design will never commited?
21:36:14muesli-hcl remember the "spin bug" ?
21:36:15Moosvery good feature for the crazy browserers
21:36:26HClMoos: it hasn't yet till i'm more sure to what people want.
21:36:31HClmuesli-: not really..?
21:36:32muesli-rbx cashes now the whole hdd
21:37:02HClin the filewrite /reading layer or?
21:37:50MoosSlasheri is your man for details ;-)
21:38:01HCli don't understand how such a thing would mess with the runtime database
21:38:10HClunless its closing and re-opening files
21:38:50Mooshehe, here if I tuned both the new dircach and your runtime db, isn't possible to shutting down
21:38:51muesli-_FireFly_ hasb been those lockups sorted out?
21:40:01MoosHCl: if you didn't test yet your "old" .jar, there fews little bugs ;-)
21:40:33HClits hard for me to understand the consequences for my runtime db if i don't fully understand what this dir cache does
21:40:56Mooslike rating thinks, the runing time debug features, databox....
21:41:26Moosnot dramatical
21:41:36HClthe rating and such is stored in the mp3entry structure
21:42:08_FireFly_muesli-: i don't know but Slasheri and moos are working on it
21:42:27muesli-sounds good :D
21:45:24preglowi don't even think i'll bother adding resume support to musepack
21:45:29preglowpeople will just think it's locked up
21:45:49Paul_The_NerdIt's that slow?
21:46:00linuxstbpreglow: hehe. You don't sound like a fan of libmusepack.
21:46:25linuxstbIs poor seeking a limitation of the musepack format?
21:46:26preglowin principle i like the musepack format
21:46:32preglowbut its execution, no
21:46:42preglowand libmusepack is more or less tailored to arm math
21:46:50preglowlinuxstb: yes it is
21:46:59preglowaccording to the libmusepack people, at least
21:47:03preglowand they should know
21:47:58linuxstbIs it another case of variable sized frames and no seek table?
21:48:36 Quit thegeek (Read error: 113 (No route to host))
21:48:57 Quit merbanan (Read error: 110 (Connection timed out))
21:51:10preglowyes, i guess so
21:51:15 Quit banan_ (Read error: 110 (Connection timed out))
21:51:21preglowand some kind of inter-frame dependecies, i think
21:52:46muesli-HCl started a runtime contest yesterday. already 22h continious playback with irivers fw with the ionity. pray the new one is even better :)
21:58:39HCli'm sure it is
21:58:39HCl :p
21:58:49muesli-1 bar left
21:59:22muesli-hope its done before midnight
21:59:52preglowmp3s? oggs?
22:00:12muesli-toggled shuffle on
22:00:17muesli-some vbr's, cbr
22:01:29amiconnFor a reproducable runtime comparison, the very same album should be selected, no shuffle, repeat = all, and of course irivers using the same hd
22:01:37muesli-unfortunately somebody switched shuffle on. i put some files in a dir and set it to endless repeat
22:01:53muesli-yeah amiconn...
22:01:59muesli-dunno who put shuffle on
22:02:52amiconnWould be interesting to compare battery runtime iriver fw <-> rockbox on the very same unit,
22:03:11amiconntesting each format supported by both firmwares
22:03:32linuxstbAre the ata_init() error codes documented anywhere (apart from reading through ata.c) ?
22:03:58muesli-amiconn that was my ideas
22:04:06amiconnlinuxstb: no
22:04:10muesli-at least testing on the some mp3s
22:05:18linuxstbamiconn: OK, thanks.
22:05:31 Nick banan__ is now known as merbanan (
22:07:57 Quit matsl (Read error: 104 (Connection reset by peer))
22:08:59 Join matsl [0] (
22:11:14*preglow cleans up codec plugin sources
22:11:22 Nick solexx___ is now known as solexx_ (
22:16:32Moosgood night all
22:16:36 Quit Moos ("CGI:IRC")
22:20:38 Join Shani}{Coder [0] (i=shani@Ariel.Atlantica.US)
22:22:06 Quit Shani}{Coder (Remote closed the connection)
22:22:54CoCoLUSare 40 euro fair for an ionity 1900 mah replacement battery?
22:23:17 Quit Paul_The_Nerd ("Chatzilla 0.9.68a [Firefox 1.0.7/20050915]")
22:25:33muesli-CoCoLUS yes, but wait
22:26:05muesli-thats the price on
22:26:40muesli-i've just shopped this one and will test it within the next days
22:26:55*preglow gasps as he sees a float in a52.c
22:27:02 Join Shani}{Coder [0] (i=shani@Ariel.Atlantica.US)
22:27:51 Quit AliasCoffee ("Leaving")
22:28:45CoCoLUSyeah i looked at dapstore...
22:28:57CoCoLUSthey sure ship to austria and they look trustworthy... so...
22:31:18_FireFly_yehaa, i got it ;)
22:32:06_FireFly_i could change wps_load in that way that now only non-comment lines are included in the format-buffer
22:32:44_FireFly_and i have test the code with a 2.6kb byte wps-file where the last xl-line was on the end of the file
22:33:07_FireFly_and this image is displayed in wps :)
22:34:11preglowlinuxstb: do you know what the output format of liba52 is? from here it looks like it uses the full signed 32 bits
22:34:48 Quit ender` (Read error: 104 (Connection reset by peer))
22:35:55linuxstbpreglow: I've never really understood liba52's output. But I think it's a 32-bit unsigned output.
22:36:16preglow_unsigned_ ?
22:36:59linuxstbNo, I'm mistaken. It's signed.
22:37:04muesli-CoCoLUS query
22:37:35preglowthe way it's handled in a52, it's first scaled fifteen bits down, then clipped
22:37:44preglowso seems like it's signed and uses 30 bits or something
22:38:53preglowi was thinking of modifying a52.c to not do its own sixteen bit conversion as well
22:39:00preglowi'd very much like that to be handled by dsp in all cases
22:39:55linuxstbPlease do. But one thing we need to get working properly is the multi-channel downmixing to stereo - I'm not sure that's working right.
22:40:19preglowi think you gave me a five channel oasis file that worked fine
22:41:06linuxstbAre you sure it was 5-channel, and not just plain 2.0 ?
22:41:29preglownot by any means
22:41:44preglowit's 448khbps
22:41:47preglowsounds like more than stereo to me
22:43:06preglownah, you're right, foobar says it's stereo
22:43:23amiconnGetting a 5.1 ac3 file should be simple
22:43:36amiconnJust rip one from a dvd
22:43:43linuxstbI've got a sample already - just a moment....
22:44:16preglowlinuxstb: but man, that's a hefty bitrate for a stereo file anyway
22:45:01linuxstbI know - it's how the German TV station Pro-7 broadcasts.
22:45:17linuxstbIt's a direct recording of their broadcast bitstream.
22:45:48 Join ]RowaN[ [0] (
22:46:56]RowaN[#rockbox, where everybady knows your name
22:46:57]RowaN[la la la
22:47:03]RowaN[and you're always glad you came
22:47:09muesli-li lo la
22:47:16dpassen1you wanna bee where you can see that troubles are all the same
22:47:43]RowaN[you wanan b where everybady knows yo' name
22:47:47]RowaN[trala lala la la
22:48:11muesli-rbx should have an own anthem
22:48:30]RowaN[i cant decide what hd to buy for my pc
22:48:33]RowaN[Toshiba 5400 16mb 68 80gb MK8026GAX
22:48:33]RowaN[Toshiba 4200 8mb 63 80gb MK8025GAS
22:48:36 Quit webguest80 ("CGI:IRC")
22:48:45]RowaN[gonna get me a 2.5" drive, coz less noisey
22:48:58]RowaN[im guessing the 5400 speed will be higher pitched noise =/
22:50:10]RowaN[i found a great comparison table but it doesnt have drives bigger than 40gig (
22:51:08 Join webguest80 [0] (
22:52:34linuxstbI've uploaded a test 5.1 AC3 file here: (22MB)
22:53:17linuxstbIt doesn't sound right to me in Rockbox - as if the front audio is coming out of the right channel, and the rear audio out of the left.
22:53:30 Quit _FireFly_ ("Leaving")
22:54:36]RowaN[u want me to try it?
22:54:43linuxstbI want someone to fix a52.c :)
22:56:09 Join Moos [0] (
22:56:16muesli-CoCoLUS ?
22:56:20]RowaN[slightly offtopic, but my pc's soundcard is linked to my amp with a digital connection, but when i play 5.1 content on my pc (e.g. Prison.Break.S01E07.HR.HDTV.AC3.5.1.XviD-CTU.avi using media player), my amp doesnt recognise it as 5.1.. any ideas guys?
22:57:05crwlyour player software must be misconfigured
22:57:06HCldo you have your digital out turned on and set for 5.1..?
22:57:09linuxstbNo idea, but Prison Break is a great series.
22:57:21crwlplayer or mixer/something :)
22:57:28HCli have no problem with 5.1, though i have an analog connection to my amp
22:57:40]RowaN[digital out is "on" yes, ive never found an option to tweak any settings related to it at all
22:57:50HClspeaker settings set to 5.1?
22:58:09]RowaN[ive got treble or bass, heh
22:58:19]RowaN[found it!
22:58:31]RowaN[mayben not
22:58:39HClits just in sound properties..
22:58:41]RowaN[i can choose "2 speakers" or "4 speakers"
22:58:50HClyou should have a whole range of options
22:58:57]RowaN[Creative has installed its own EAX sound mixer crap with a ton of options
22:59:16HCljust use the windows control panel
23:00:01muesli-gotta go
23:01:12]RowaN[ive now set my speaker mode to 5.1 in control panel, but it makes no difference to the digital out
23:01:37]RowaN[my amp just knows its "pcm", thats all, and when i enable amp effects, it just says prologic, as if its recieving a normal stereo signal
23:03:10amiconnpreglow: liba52 seems to be rather efficient... :)
23:05:38*]RowaN[ emails Creative to moan about their shite software
23:08:46 Join Vladoman [0] (
23:10:11amiconnpreglow, linuxstb: The a52 downmix is definitely broken
23:11:11linuxstbI know :(
23:12:30linuxstbI tried, but could never work it out. The best thing would be to go back to the original "a52dec" source (the player that comes with liba52).
23:13:43preglowamiconn: yes, it was efficient before i did my thing on it, and very efficient after
23:14:43amiconn5.1 mode isn't as efficient, but 2.0 had ~3 % boost ratio - with resampling
23:15:10preglowit seems audio is always stereo before it reaches the resampler
23:15:20preglowso i think i'm going to make it do both channels in one pass
23:21:07Learpreglow: remember that mono might look like two channels in the resampler, but really isn't...
23:21:12*amiconn wonders whether the current a52.codec will handle mono correctly
23:21:31preglowLear: riiight, that's how it is...
23:22:03preglowi WAS going to code a resampler for doing both mono and stereo in one pass, but i bloody hate duplicating code
23:22:28preglowa branch isn't THAT dear on coldfire, i could just do a compare per sample
23:22:37preglowto see if i've actually got two channels
23:22:57amiconn_per sample_ ??
23:22:58LearHm... When I rewrote dsp, I think I tried to avoid just that...
23:23:13amiconnI'd rather avoid that...
23:23:48 Quit muesli- (Read error: 110 (Connection timed out))
23:24:10 Quit Vlad0man (Read error: 110 (Connection timed out))
23:25:26preglowamiconn: yes, me too, but the alternative is writing two perfectly identical resamplers, one doing stereo, the other mono
23:25:58preglowthe way we do now, with processing the channels separately, is very wasteful
23:26:50Learmore waste than an if per sample?
23:26:54preglowoh yes
23:27:14preglowthere's already a compare per sample (the while loop) that can be avoided if we do both channels per pass
23:27:19preglowand additional delta calculation, etc
23:27:37amiconnI don't see the problem with having 2 resample loops
23:27:52***Saving seen data "./dancer.seen"
23:28:09preglowamiconn: no problem per se, but there could be just one loop, it goes without saying one loop instead of two identical loops is more efficient
23:28:09amiconnIt's just a bit more code... and if it helps to speed things up...
23:28:20preglowahh, like that
23:28:27preglowcould make it a macro, i suppose
23:28:40preglowsometimes i actually miss templates from c++ :)
23:30:59Learfor the stereo case, some delta calculations would be avoided, yes, but you'd do completely unecessary processing in the mono case...
23:31:17Learso neither case is perfect. I chose this as the least bad, iirc...
23:33:17 Join Vlad0man [0] (
23:33:44preglowLear: not if make a resampler for just mono and one for stereo, no ;)
23:34:07Lear...because memory access is fairly expensive (not so bad with both buffers in iram though...)
23:34:57amiconnI did just that (splitting one loop into 2 almost identical ones) some days ago for speeding up fire.rock
23:35:01]RowaN[guys whats the url of that german shop i bought my replacment iriver hd from
23:35:48Learpreglow: so, would that mean 4 resamplers then (2 up and 2 down)? :)
23:36:22preglowLear: yes, yes it would
23:36:41preglowamiconn: i don't want to do this until i can measure things properly, boost factor doesn't cut it
23:37:17amiconnYou could simply time unsynchronised decoding, w/o any output
23:37:48amiconnThat's how I did the performance measurements for the various memory functions
23:38:13Leartricky to make it generic for any codec though...
23:38:21Lear(I did a quick one for Vorbis...)
23:38:31 Quit Vladoman (Read error: 110 (Connection timed out))
23:41:50linuxstbI've just updated the table at the end of SoundCodecs page a little to make it more critical of the codecs. Does anyone have anything to add there about the state of Vorbis?
23:41:58 Join webguest53 [0] (
23:43:31 Quit webguest53 (Client Quit)
23:46:39Learlinuxstd: files encoded by old versions (pre 1.0) might not play properly...
23:47:31amiconnvorbis seems to be the most mysterious codec in rockbox
23:47:41Learhow so?
23:48:01amiconnFiles crashing it or not playing properly etc
23:48:36Learcrashes are typically the codec running out of memory; not so mysterious...
23:49:19Learthat's the only problem I've seen anyway (including the recent one in the forums, which Linus got a copy of, that he forwarded to me, and which played fine.)
23:50:22Learoh well, really gotta go to bed now. bye!
23:50:27 Quit Lear ("Chatzilla [Firefox 1.4.1/undefined]")
23:53:42preglowbut yeah, back to a52.c
23:55:06preglowbut the conversion from 5.1 to stereo, i thought liba52 itself took care of that

Previous day | Next day