Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2005-12-12

00:03:22 Join Massa [0] (
00:04:47 Join JazzBone [0] (
00:05:11MassaHi everybody!
00:06:41idanmgood night everyone
00:09:40Massadoes someone know what happened to
00:10:23idanmthey are dead for now...
00:10:34 Quit idanm ()
00:10:44Massawhat happened?
00:10:46 Quit JazzBone ("Leaving")
00:11:04Massa(and when will i be available again?
00:11:12Moosthe administarator isn't here
00:11:25Moosyou need to wait from MisticJeff
00:11:45Mooswe can't tell when the site will relive again
00:11:47 Join einhirn [0] (
00:11:50 Join ts-x [0] (
00:12:07Massawas it discussed already this day?
00:12:21BagderMassa: MisticJeff is not here very often
00:12:23 Quit mirak (Remote closed the connection)
00:12:28Bagderso mentioning it here doesn't do much good
00:12:41MassaBTW, is MisticJeff also the admin of
00:13:18ts-xHey for anyone working on the H3xx wps flickering issue, I think I found something that might help (and maybe you are already aware of?)
00:13:35MassaI thought he's "only" responsible for
00:13:44Moosboth of them
00:13:54Massats-x: Isn't the flickering already fixed in the latest CVS version?
00:14:09ts-xNot entirely
00:14:24ts-xBut what I've found is that it's only the %x images that are flickering
00:14:51ts-xPreloaded and conditionaly displayed bitmaps are not flickering
00:15:05ts-xI tried the exact same image both ways and found this to be true
00:15:29Moosgood night @ all
00:15:37 Quit Moos ("Glory to Rockbox")
00:15:56MassaI'm not very experienced in WPS, but isn't %x for preloaded images?
00:16:25ts-x%x loads and displays immediately
00:16:25MassaMoos: good night!
00:18:33Massaso only %xl preloads images?
00:19:17ts-xI just loaded Rockbox for the first time on Friday and kind of stumbled upon this during the wps creation process
00:19:47MassaHmm, when you look e.g. in engineer2 wps, all images will be preloaded with %xl
00:19:49 Join tarn [0] (n=irc@tor/session/x-3d6ef1891163788f)
00:20:08Massa(at least in the version I use ;) - and it still flickers...
00:20:24ts-xThe background is %x
00:20:37ts-xImage 'a'
00:21:24MassaYou're right - and you changed it so it's preloaded?
00:21:32MassaAnd it did not flicker any more?
00:21:40 Quit tarn ()
00:22:16ts-xI just tested engineer as well, just the background is flickering. Everything else is ok.
00:22:58ts-xLet me change it and test...
00:23:17dropandhothat would be dope to fix that bug!
00:25:30MassaI didn't test the latest CVS changes, is it still flickering?
00:26:54 Quit Kohlriba ("Leaving")
00:27:12 Join chiller_ [0] (
00:27:19 Quit chiller (Read error: 104 (Connection reset by peer))
00:27:28 Nick chiller_ is now known as chiller (
00:30:40 Quit chiller (Read error: 104 (Connection reset by peer))
00:32:11 Join chiller [0] (
00:39:47ts-xNot having much luck with engineer since not preloading the background seems to cause it to overlap everything else
00:40:16MassaAnd removing it would help?
00:44:21ts-xI just did that - it significantly reduces flickering but it still flickers when pausing and changing tracks
00:45:01ts-xWhat's strange is that the wps I'm creating doesn't flicker on pausing or track changes
00:45:37MassaYeah, that's really strange...
00:46:09MassaI'll leave now and go to bed...
00:46:12MassaGood night!
00:46:22ts-xSeems to indicate that multiple factors are contributing the flickering
00:46:25ts-xGood night
00:46:37ts-xWell if nothing else maybe this information will help one of the devs...
00:47:30 Quit PaulJ (".")
00:52:55 Quit Massa ("ChatZilla 0.9.61 [Mozilla rv:1.7.12/20050915]")
00:53:54 Join Jungti1234 [0] (n=jungti12@
01:07:03 Quit San||Away (Read error: 110 (Connection timed out))
01:10:25 Quit saa[b_r]ider (Read error: 104 (Connection reset by peer))
01:10:55 Quit ts-x ("CGI:IRC (EOF)")
01:11:17 Quit ratpack91 ("Azureus")
01:13:23***Saving seen data "./dancer.seen"
01:15:32 Quit Bger ()
01:26:32 Quit gromit` (Remote closed the connection)
01:29:15 Quit Dima202 ("Leaving")
01:30:54 Quit ender` (Read error: 113 (No route to host))
02:21:29 Quit einhirn (Read error: 110 (Connection timed out))
02:26:53 Join actionshrimp [0] (
02:46:50 Join San [0] (
02:57:11 Quit tvelocity ("Leaving")
03:09:18 Quit Jungti1234 ("bye")
03:13:26***Saving seen data "./dancer.seen"
03:37:53 Quit San (Read error: 110 (Connection timed out))
03:43:08 Join amiconn_ [0] (
04:01:16 Quit amiconn (Read error: 110 (Connection timed out))
04:01:17 Nick amiconn_ is now known as amiconn (
04:02:22 Quit actionshrimp ("a bird in the bush is worth two in your house")
04:06:26 Quit ehntoo ("Leaving")
04:06:47 Join ehntoo [0] (
04:19:38 Quit TCK (Read error: 104 (Connection reset by peer))
04:31:15 Join Midgey34 [0] (
04:42:17 Join Jungti1234 [0] (n=jungti12@
04:48:49dropandhohey man...what up?!
04:52:25Jungti1234what up?
04:53:45 Quit Jungti1234 ("bye")
05:04:01 Join Rob2222_ [0] (
05:10:35 Join Paul_The_Nerd [0] (
05:13:14 Part Midgey34
05:13:31***Saving seen data "./dancer.seen"
05:14:45 Join San [0] (
05:19:39 Quit Rob2222 (Read error: 110 (Connection timed out))
05:28:00 Quit ehntoo ("Leaving")
05:28:01 Quit San (Read error: 104 (Connection reset by peer))
05:36:59 Join gursikh [0] (
05:40:25 Quit ReKleSS (
05:40:30NJoinReKleSS [0] (
05:45:01gursikhThis forums have been down for some time, Is this intentional?
05:45:13dropandhonot that i am aware of
05:45:34dropandhoapparently Jeff, from misticriver, hosts the forums
05:45:54dropandhoi have reported it to him on IM today, but haven't received any word back from him
05:46:16dropandhoi hope someone else has some word on it
05:46:31 Join webguest24 [0] (
05:46:48gursikhThis morning (14 hours ago) before work it was down, so it's been quite some time
05:47:10dropandhoyeah, it has been a while
05:47:15 Quit webguest24 (Client Quit)
05:47:20dropandhohaven't caught Linus on here- to report it to him
05:47:37dropandhoany ideas what to do next?
05:47:59gursikhIs there any other contact for Jeff?
05:48:44 Join Midgey34 [0] (
05:49:21dropandhou wanna hit him up?
05:49:35gursikhhe has not responded via AIM ?
05:49:51dropandhocorrect...he has not written back
05:50:17gursikhvia AIM or Y! messengers?
05:50:17dropandhoi cant ping the forum address either
05:50:37dropandhoit said he was "active" also
05:51:20gursikhhe has both, Let me attempt yahoo.
05:53:23gursikhIs there additional contact Infomation for Linus?
05:54:33dropandhooff of the site
05:54:40gursikhyeah, i got that one.
06:01:33 Quit Paul_The_Nerd ("Chatzilla [Firefox 1.5/undefined]")
06:03:59dropandhoany luck?
06:05:00gursikhhe just put on his away message on aim after i IMMED him.
06:05:00dropandhoi need to head out now for bed
06:05:32dropandhowelp, im off now
06:05:38dropandhohope something comes up
06:05:54dropandhou email linus and jeff?
06:06:29dropandhogreat- we have it covered
06:06:31dropandhothanks man
06:06:34dropandhohave a good night
06:07:28 Quit dropandho ()
06:17:43 Quit Nibbler (Read error: 110 (Connection timed out))
06:21:39 Quit mikearthur (Connection timed out)
06:21:51 Join mikearthur [0] (
06:24:21 Join Nibbler [0] (
06:36:48 Quit mikearthur (Read error: 104 (Connection reset by peer))
06:40:44 Quit gursikh (Remote closed the connection)
07:13:35***Saving seen data "./dancer.seen"
07:17:22 Part Midgey34
07:19:30 Quit RotAtoR ()
07:27:21 Join San [0] (
07:47:57 Join Jungti1234 [0] (n=jungti12@
07:59:53 Join saa[b_r]ider [0] (n=saab_rid@
08:17:48 Join Bger [0] (n=Bager@
08:19:20Jungti1234hi Bger
08:19:26Bgerhi, Jungti1234 :)
08:34:47 Quit San (Read error: 104 (Connection reset by peer))
08:34:47 Quit saa[b_r]ider (Read error: 104 (Connection reset by peer))
08:36:16 Join ender` [0] (i=ychat@
08:50:07 Join Acksaw [0] (
09:03:38 Join einhirn [0] (
09:04:38 Quit Acksaw ()
09:04:58Bgerwow, we have 3 independend buttons on h300 (main unit) : ON (aka play/pause), MODE (aka A-B) and REC
09:09:34 Join LinusN [0] (
09:10:01Jungti1234hi Linus
09:11:11 Quit Rob2222_ ()
09:12:37LinusNhey ho
09:13:37***Saving seen data "./dancer.seen"
09:15:43XavierGrHi LinusN!
09:17:54XavierGrLinusN: Theoritically, can I start a new thread within a plugin?
09:18:17XavierGr(but current plugin continue to run)
09:18:37XavierGrlike having a sub thread at the time the current plugin runs.
09:20:00XavierGrI know I can start a thread in a tsr plugin but then I exit the plugin (it runs in the background) so I am not sure if I can.
09:22:51LinusNXavierGr: yes you can
09:23:20XavierGrI quite finished the new version of the jpeg filescroller.
09:24:24XavierGrIn order to have as more entries as you can, I decided that filenames must be written in a file and a pointer array would hold their seek positions in the file.
09:24:49XavierGrThat is nice until sorting comes into play.
09:25:06LinusNhow do you create the file?
09:25:15XavierGrTo sort ~850 entries with qsort an interval of 6-7 seconds is needed.
09:25:54XavierGryou mean how I write the entries in the file?
09:26:21XavierGr(I create the file with rb->open) then I scan the folder and filter jpeg files
09:26:38XavierGrfor each entry I write down to the file the name of the jpg.
09:26:47LinusNah ok
09:26:57 Join B4gder [0] (
09:27:36XavierGrthing is that when I try to sort the compare function must open read 2 position get the names close the file and then compare the 2 entries.
09:27:58XavierGrfor a small number of files this is quite easy and fast but if you have a big list then....
09:30:36XavierGrso I had the idea of letting the user see the jpeg of his choice and then run a different transpaerent thread for sorting.
09:31:56pregloww00t, my player locked up
09:31:57XavierGralso for the iriver targets I made a small interface to check if playback occurs to try to fetch plugin memory instead of the audio buffer.
09:32:10XavierGryour iPod nano?
09:32:35preglowof course not
09:32:49preglowthat never locks up!!
09:32:52preglowyeah, ihp
09:32:58preglowwhen trying to start playback right after using usb
09:33:00preglowplain locked
09:33:05preglowwouldn't turn off no how
09:33:12XavierGrreset :)
09:33:16preglowyes, of course
09:33:16 Join Zagor [0] (n=bjst@pdpc/supporter/sustaining/Zagor)
09:34:18XavierGrDoes anyone know if the grayscale needs the same memory on all targets? Eg. iriver needs ~86kB what about archos models?
09:34:40XavierGrmake that 8kB I think
09:34:58B4gderit most definately is not the same amount
09:36:11XavierGrwe can easily determine that with a splash in th jpeg viewer. (at least the method I used) anyone willing to help me find it?
09:37:01amiconnYou can't splash() while grayscale is running
09:37:20XavierGrwell I splash right before.
09:37:26amiconnHowever, the amount of mem is easy to calculate; the formula is given in the grayscale lib sources
09:37:41XavierGrgray_init outputs the size needed with an argument
09:38:43amiconnapps/plugins/lib/gray_core.c, line 106ff
09:41:02XavierGrwhat's bitpatterns?
09:44:38XavierGranyway I am asking this to see if Archos targets are capable of viewing while playing (at least a very small pic < 5kB) but as I see it with current plugin memory on archos this is not possible, am I right on this?
09:45:30XavierGrgrayscale cannot fit in the archos plugin memory, I think.
09:45:33Bgercan someone help me in understanding the following
09:45:35Bger@@ -166,9 +166,12 @@
09:45:43Bgerin a diff file ?
09:46:01XavierGrlines and offsets to apply the patch
09:46:11Bgeryes, but...
09:46:34Bgerwhat does the ",9" and ",12" mean ?
09:46:51XavierGrdunno :(
09:46:59B4gderwhy does it matter?
09:47:09zeBger: count the lines
09:47:45BgerB4gder: because i'm trying to edit a diff file
09:47:58B4gderthen that line won't matter much
09:48:02amiconnXavierGr: Yes you're right. On Archos it's possible to fit unbuffered grayscale with <33 shades in plugin memory with small plugins (e.g. mandelbrot), but not with full depth and the jpeg plugin takling almost all plugin memory for itself
09:48:08B4gderpatch only uses that as a hint
09:48:33zeBger: i think the 1st is the number of lines that the specified chunk has before the patch, and the 2nd is the number of lines in that chunk after patching
09:48:46Bgerthen what does the "unexpected end of patch" or simillar mean ?
09:49:03B4gderit means patch didn't like the format of the patch
09:49:08zeand i think patch does expect it to be right
09:49:35Bgerheh, B4gder ...
09:49:45XavierGramiconn: Thanks then I will remove this option for archos with low plugin memory.
09:51:24Bgerso if i edit a diff file and remove a number of different lines, should i change these values as well ?
09:52:09B4gderI would advice against hand editing a diff file
09:52:53XavierGrexcept if it is rather small change (E.g on the same line)
09:54:05BgerB4gder then what u do when u've changed many files but u want to make a diff of only 5-6 ?
09:54:27 Join b0br [0] (
09:54:34zejust diff those files?
09:54:36B4gderthen I diff only those
09:54:37Bgeralso, there are some very little diffs (like whitespace) ?
09:54:47B4gderthen use an option that ignores whitespace
09:55:07Bgerwhich is ?
09:55:11B4gder-b and -B
09:55:25B4gderor possibly -w ;-)
09:55:43Bgerdoes the cvs diff support these ?
09:55:44B4gderand -E
09:55:46*B4gder grins
09:55:57B4gderit supports -B and -b at least
09:56:07*Bger goes to RTFM
09:56:32B4gderand to make a single patch, you can just concatenate any number of single-file diffs
09:57:02B4gderdiff files built this planet
09:57:06Bgerbut cvs diff does it afaics ... or i'm wrong
09:58:06B4gderyes it does, but I mean in case you use multiple command line invokes
09:58:18Bgerahap... 10x :)
10:00:18XavierGrI hate when I am diffing from CVS and I have a new file.
10:00:28XavierGrI just can't make it to add it.
10:00:36 Join Vlad0man [0] (
10:00:40Bgersomeone told there is a command for doing it
10:00:48Bgeriirc markun
10:00:54XavierGrInstead i do single diff of 2 identical cvs'es (the other has the changes)
10:01:43 Quit _Vladoman (Read error: 110 (Connection timed out))
10:02:19markunyes, install cvs tools. Then: cvsdo add 'file'
10:02:34markunand make the patch with cvs diff -uN 'files'
10:03:52XavierGrGod I am an as*hole!!!!
10:04:37XavierGrmarkun I suppose that I don't need cvs acces for this right?
10:04:48markunnot with cvsdo
10:05:14XavierGrok thanks.
10:06:22Bgerheh i don't see such ebuild ... (cvs tools or sth similar)
10:06:35 Part b0br
10:07:20BgerB4gder cvs diff supports all three -b -B -w
10:08:53linuxstbpreglow: When did you last compile and install the bootloader on your Nano? The current CVS doesn't load the original firmware any more for me.
10:09:30preglowlinuxstb: aeons ago
10:09:48preglowlinuxstb: it doesn't here either
10:09:51preglowlinuxstb: it just hangs
10:10:23linuxstbOK. I'm thinking of changing it anyway, so instead of having the apple OS in the boot partition, it loads it from the disk. This should speed up booting a little.
10:10:47preglowoh, you mean dumping it to a file first?
10:11:03linuxstbYes - you have to do that as part of building the bootloader anyway.
10:12:35XavierGrOMFG Sorting the list from the file will need 18 seconds.
10:12:59XavierGr(for 850 files)
10:13:14B4gderXavierGr: 18 seconds to sort 850 names?
10:13:19linuxstbWow, I've just tested it with a 512KB "dummy" firmware file, and the boot time on my iPod decreased by about 4 seconds.
10:13:31XavierGryes. but the names are sorted in a file
10:13:55B4gderXavierGr: you mean you don't read the file, only move them around within the file?
10:15:02XavierGrNo, the file remains unchanged, I just move the pointers that hold the seek position of each name. The strings are never loaded into memory (except 2 strings each time in the comapre function)
10:15:22XavierGrMaybe I can do that a little faster, but not much...
10:15:32B4gderthat's what the playlist code does
10:15:36B4gderit is _waaaaay_ faster
10:15:43XavierGrA different thread will make it run at least transparent.
10:16:02amiconnXavierGr: Directly sorting a file will probably be way slower onm Ondio
10:16:19B4gderoh well, it doesn't always sort them but anyway
10:16:38XavierGris it that I open and close the file inside the compare? I think i can just leave open the file and close it in the end of the sorting process.
10:16:39B4gderI guess that's the reason
10:17:03B4gderXavierGr: how do you do the sort in the first place? qsort() ?
10:17:05linuxstbXavierGr: Why don't you store the list of files in memory, then sort, then write to disk - at least if it is possible to do so, which on the iriver I expect it would be 99% of the time.
10:17:09XavierGramiconn: we need to do some tests after I finish this.
10:17:36XavierGrB4gder: yes I use qsort.
10:18:14XavierGrlinuxstb: Storing in memory 1000 * 20 names will eat all my plugin memory buffer, which I need to hold for the jpeg in order not to stop playback.
10:18:36linuxstbYes, but you can write them to disk _after_ sorting them.
10:18:58XavierGrhmm and then flash that space eh?
10:19:01linuxstbI'm assuming you load the file list before you load any images.
10:19:28linuxstbYes, and then you can free that space and use it for the images.
10:19:59XavierGrsounds neat, I could try that.
10:20:12 Join _FireFly_ [0] (
10:20:28XavierGrHi _FireFly_.
10:20:44XavierGr_FireFly_: and we were talking about sorting :)
10:24:52preglowthink i'll need to upgrade my bootloader soon
10:24:58preglownow for finding a linux pc where i can do it
10:26:30linuxstbpreglow: If I remove the 5.5MB Apple firmware from my boot partition, then my Ipod will boot to the file browser in about 4 seconds.
10:26:42linuxstbIt was about 7-8 seconds with it.
10:26:48XavierGrI will be damned. Moving the open, close file functions outside of compare functins dropped the sorting time from 18,5 seconds down to 3,5 seconds!!! Isn't this great?
10:27:59_FireFly_XavierGr ,9
10:28:17preglowlinuxstb: how the hell can it be so much faster
10:28:54linuxstbThe flash bootloader is obviously very slow at loading the boot partition. It will also have to calculate and verify a checksum before executing it.
10:29:47linuxstbBut the good thing is that it doesn't do any check on the minimum size - I'm writing a 84KB boot partition and it's working perfectly.
10:31:08linuxstbSo I'm definitely in favour of putting the apple_os.bin file on the fat32 partition. What do you think?
10:31:32linuxstbIt probably has less of an effect on your Nano, but I'm sure you will still notice a speed-up.
10:31:58 Join San||Away [0] (
10:32:59linuxstbYou can test it by creating a dummy 512 byte file (e.g. dummy.bin), and then create a boot partition using "ipod_fw -g nano -o rockboot.bin -l dummy.bin bootloader.bin"
10:33:22 Join saa[b_r]ider [0] (n=saab_rid@
10:44:14preglowi'm going to start working on a linux box soon
10:48:01linuxstbDo you think we should try and fix loading the apple firmware via the boot partition, or just abandon that method and always load it from a file?
10:48:19preglowloading from a file sounds just dandy to me
10:48:23preglowif you need to make that file anyway
10:48:24preglowwhich you do
10:48:39linuxstbOK then. I'm happy to to shave 4 seconds off my boot time to Rockbox.
10:49:34preglowbut okiedokie
10:49:42preglowwhy the hell doesn't this work on your ipod
10:50:17linuxstbI've tried the new button driver, and it's not as good as I would have hoped. I added a loop in the bootloader that reads the key state and displays it, but it still doesn't always recognise a key that is pressed before the bootloader starts.
10:50:41linuxstbBut I think it could be used to poll the button status if we wanted to write a polling button driver.
10:51:18linuxstbThe code that's there only checks the 5 real buttons - it doesn't give a status for the scrollwheel. But hopefully that can be added.
10:51:30preglowdo we really want a polling driver?
10:51:39preglowif you ask me, an interrupt driven one is better
10:51:48preglowyou only poll when you need to, and polling interval issues are gone
10:52:40linuxstbOf course - an interrupt driven driver will be better.
10:53:03preglowhmm, i think i'll introduce the sleep loop again
10:53:10preglowto see if the tick timer is really working
10:53:40linuxstbOK. Let me know if you want me to do any more tests.
10:54:22linuxstbI'll try and make the bootloader changes later today.
10:55:58linuxstbAnyone got any news about the forums? I hope it's not a serious problem.
10:56:06XavierGrwhat is working currently on rockbox for ipods?
10:56:22preglowwe'll see
10:56:26preglowXavierGr: i am
10:56:31linuxstbXavierGr: It boots up into the file browser, but there is currently no button driver.
10:56:34preglowa lot of forums have been hacket lately
10:56:45preglowahh, what, not who :-)
10:56:50 Join tucoz [0] (
10:56:58XavierGrlinuxstb: And how do you turn it off? :D
10:57:17linuxstbYou never turn the ipod off...
10:57:32linuxstbBut you press a key combination for a hardware-controlled reboot.
10:57:45linuxstbThis works even if Rockbox has crashed.
10:57:46preglowwoot, works even with that loop now
10:57:47XavierGralso what do you think about audio on iPod. Will it be easy now that we have iriver targets working on that aspect?
10:58:05preglowaudio on ipod is going to need a lot of work
10:58:13preglowto get it running on two cores, at least
10:58:32preglowlinuxstb: but okay, what's the status on ipod button mappings?
10:58:42XavierGrit has a different DAC too, right?
11:00:02linuxstbThose are pretty much finalised - see button.h. In addition to the five buttons specific to the ipod, there is the generic BUTTON_LEFT and BUTTON_RIGHT
11:00:36tucozsorry for the off-topic, anyone familiar with dvd regions? I can play region 1 dvd's in linux, but when I try to play it in windows it tells me I only have 4 more region changes left. Is this shifting done in sw or the dvd-player hw?
11:00:51linuxstbBoth IIRC.
11:01:07linuxstbThe DVD-ROM can have region settings, and so can the DVD playing software.
11:01:11tucozbut linux doesn't care?
11:02:03linuxstbI think it cares about the DVD-ROM settings (it can't affect them), so it sounds like it is just a software issue with your Windows player.
11:02:26linuxstbYou can try a free Windows DVD player such as VLC -
11:03:00tucozlinuxstb, ok. good. thanks. Then the dvd-rom is probably region free, but the player-sw is not.
11:03:06linuxstbYes, sounds that way.
11:03:40preglowbut ok, first, find out why it doesn't work with you
11:03:57preglowwhat approach should we take to that? debug stuff before commit, or commit then debug? :)
11:04:24linuxstbI would say commit then debug. But commit in small stages if that's possible.
11:05:09tucozlinuxstb, are you in london btw?
11:05:20linuxstbtucoz: Yes.
11:05:23 Quit San||Away (Read error: 110 (Connection timed out))
11:06:00tucozlinuxstb, looks like the end of the world over there.
11:06:12SlasheriHmm, i think now the tagcache building should work on archos too :) And simple searches works now also
11:07:42tucozSlasheri, cool.
11:08:22SlasheriIt creates one temporary file when building the cache and then parses that file into 5 separate files with one master lookup table
11:08:48Slasherithen it should be possible on some platforms load that structure into ram also
11:08:49tucozSlasheri, when adding or removing files to the player, will the tagcache rebuild the enitre db or just the add/remove the changes?
11:09:08Slasheritucoz: no, it can add files without completely rebuilding it
11:09:24tucozoh, that sounds good
11:09:34linuxstbIs that part working yet?
11:09:53Slasheriwith dircache that rebuilding is automatic on boot, and without it manual start of database check is required
11:10:23Slasherilinuxstb: building works and update worked too some time ago (but i just changed the structure a little), so it should be work quite well
11:11:24tucozI like the morse code input. Never thought that I would be able to learn morse, but it is actually a really good input method.
11:11:30preglowtucoz: nice pic
11:11:41Slasheritucoz: hehe =)
11:12:01tucozpreglow, it looks manipulated, but according to the news-article it is not.
11:12:08preglowjust what london needs, more smog
11:12:10amiconnXavierGr: Building and sorting the list beforehand sounds good, however, the drawback is that you can't do it in the background (not a major concern imho)
11:13:40***Saving seen data "./dancer.seen"
11:15:20tucozI think, if I would ever use the database is that it is built just the way Slasheri has done. I never get used to running a host-program to set up the database.
11:15:46tucozas I tend to remove and add stuff all the time, and do not have java on my computer :)
11:15:57B4gderwell, the java version is overbloated
11:16:19XavierGramiconn: There are many things that I have to plan if I do that. 1) I must split the remaining plugin buffer into 2 parts. (fortunately I can predict where to split) 1 buffer for the pointer array and one for the filenames.
11:16:23tucozB4gder, well the perl script is probably nicer.
11:16:29 Join Membrillo [0] (
11:16:36 Join Polo_o [0] (
11:16:39B4gderI think this not only because I wrote the perl version
11:16:42preglowthe target way is nice in that it supports everything rockbox does
11:16:52XavierGramiconn: Then when the list is sorted into memory, I will have to change the pointer array values with the seek points of the file list.
11:17:00B4gder"the proper way" would be to offer host host and target to do it
11:17:00preglowand automatically
11:17:10XavierGrthen I can discard the data for the jpeg buffer.
11:17:12preglowbut in this case i'd really like to use the same code
11:17:17B4gderpreglow: indeed
11:17:21tucozB4gder, no. But the fact that I need a full JRE to build the database is not that nice.
11:17:24preglowi wonder if it's possible to make get_metadata work for host as well
11:17:26B4gderand that's why there's a start for db code in C in CVS
11:17:28preglowthat would be just dandy
11:17:38linuxstbIIUC, Slasheri mentioned that it was faster to build the database on the iriver than on a PC.
11:17:40B4gderpreglow: that would be the best approach
11:17:54tucozreally, hehe that is amazing
11:17:54B4gderlinuxstb: his code is faster than the current tools on PC, sure
11:18:02B4gderblame the crc mania
11:18:08B4gderwhich I doubt Slasheri uses
11:18:17B4gderso his code can't do the runtimdb
11:18:17linuxstbBut I would expect the bottleneck to be the disk accesses, not the CPU.
11:18:28linuxstb(apart from CRC of course)
11:19:09B4gderwell, I can't see how the target can be even close to host, speedwise, if they'd do the same thing
11:19:30B4gdersince both CPU and disk are waaay faster on host
11:19:33tucozI see, so title, album, artist is not good enough for the runtimedb?
11:19:35B4gderand amounts of ram
11:19:41B4gdertucoz: nopes
11:19:48tucozWhy is that?
11:19:53B4gderread up
11:20:15B4gderbecause HCl wanted it to survive changed tags
11:20:20B4gderfor example
11:20:45tucozI get it
11:21:02B4gderpersonally I think it is overkill
11:21:10tucozIt is the actual track that is the key here, not the meta-data.
11:21:29B4gderand that's why it does CRC on a chunk of the audio data
11:21:38B4gderto use as key
11:21:45tucozI understand
11:22:24tucozHmm, of course that has it's benefits but might be a little overkill yes.
11:22:56tucozI think the audioscrobbler way is good enough for me.
11:23:17Slasheriin fact that crc code is not a solution for the runtimedb..
11:23:17XavierGrwhat to choose for my big char array buffer, unsigned char, or simple char?
11:23:18preglowwhich is?
11:23:28Slasheriit only calculates the first 32 KiB of data or something like that
11:23:28B4gderSlasheri: why not?
11:23:50B4gder"only" 32kb
11:23:54Slasheriso it's more than likely there will be more than one identical song the db sees
11:24:02B4gderI doubt that
11:24:05Bgerok, new version of the h300 lcd remote buttons patch is on the patchtracker
11:24:14B4gderbut I don't like that CRC approach anyway
11:24:38linuxstbXavierGr: It depends what the functions using that buffer are expecting. What are you storing in that buffer?
11:24:40tucozpreglow, I believe audioscrobbler just records the artist-title, and adds to the db if the track has been played more than 50% or 2.5 minutes (whichever comes first)
11:25:25SlasheriB4gder: for example wav files without any tag entries and similar mixes at the beginning..
11:25:31 Join mirak [0] (
11:25:34linuxstbThen you should use the same type as the string functions (strncmp, strlen etc) - i.e. char
11:25:34Slasherithen db would see those as the same song
11:25:40B4gderSlasheri: the db tools don't even work with wav files ;-)
11:25:46Slasheriah, hehe :D
11:25:54B4gdertucoz: audioscrobbler is more naive than we can be
11:25:57XavierGrlinuxstb: ok thanks
11:27:09B4gderSlasheri: still, using that lame CRC is more likely to find the song than you are without it
11:27:31tucozB4gder, ok
11:27:36B4gdergiven the premises
11:27:42B4gderbut I think we should lower the goals
11:29:00 Join yngwi [0] (
11:29:24linuxstbB4gder: Are you suggesting we don't use a CRC at all?
11:29:25 Quit yngwi (Client Quit)
11:29:55B4gderI haven't fully thought about it properly to have a final suggestion, but I think I lean towards dropping the CRC yes
11:30:50linuxstbThe only alternatives I can think of is to use the path and filename (which can change) or the tags (which can be missing).
11:31:15B4gderI know
11:31:22SlasheriB4gder: Hmm, what will the current db do, if i move the song to another folder. Will it still keep the runtime data and find the song in future if the tagdb is rebuilt?
11:31:23linuxstbIt's not uncommon for someone to re-organise the files on their player. So at least the paths will change.
11:31:39B4gderSlasheri: yes, I think that's the plan
11:31:48SlasheriIt could be possible to add the crc or something similar identification to a separate tag index file
11:37:20tucozB4gder, did you have a look at the LaTeX stuff?
11:39:39 Join DJDD_ [0] (
11:41:10preglowlinuxstb: btw, i don't remember if i disable the cop in my bootloader
11:41:24preglowlinuxstb: if that is so, it's a small wonder it doesn't work with you
11:43:05 Join Delazon [0] (
11:44:54 Quit Membrillo ()
11:45:50DelazonThanx boys for the release of rockbox for H3xx!!!
11:46:12_FireFly_Delazon: there isn't yet an release for h1xx and h3xx yet
11:46:14BgerDelazon there's no *release*
11:46:27_FireFly_these are only testing builds :)
11:46:47_FireFly_because not all functions are implemented or working correctly
11:48:02linuxstbpreglow: I'm running the CVS bootloader at the moment - I think it does wake up the cop
11:48:17preglowtry wrwapping the wakeup with an if 0
11:48:20preglowthat's what i have here
11:48:53preglowi don't really know what we're waking it up to
11:48:56preglowso i just let it sleep
11:49:10*XavierGr *sighs* and imagines the day that he will see an iHP Icon on the Download Rockbox page....
11:49:12mirakis it possible to put mp3 in an ogg container ?
11:49:31linuxstbpreglow: OK, trying now.
11:49:32amiconnXavierGr: If you sort the file list in-memory, you can drop the pointer array too after writing the file. Just write the file sorted, then you can calculate the seek position
11:50:56Delazonok, i mean releasing a testbuild... i tested it and im elightend. It will be great when it?s finsished....
11:51:05DelazonOne question?
11:51:09linuxstbpreglow: Success :)
11:51:15BgerDelazon just ask
11:51:18linuxstbBacklight faded after about 3 seconds.
11:51:43DelazonWill the user interface come in color or will it be in black and white all the time?
11:51:58mirakanyone could send me a H300 V3 firmware ? I can't patch it on linux, I don't know what I am doing wrong
11:52:03linuxstbDelazon: It is currently black and light-blue.
11:52:20LinusNmirak: what happens?
11:52:30linuxstbDelazon: But yes, Rockbox can be made in as many colours as you want - it just needs a developer to work on it.
11:52:38DelazonYes, you guy?s marks my words ;)
11:52:46DelazonI mean monochrome then ;)
11:53:35tucozDelazon, it will be in color eventually
11:54:31Delazoni switched back to iRiver fw for now... im a big radio listener, but when it?s released with full features i will swtich again. :D It was fun to test it.
11:54:59DelazonAnd this function WPS... i like it alot...
11:55:05BgerLinusN what has to be done before having working radio on h300 ?
11:55:22LinusNthe i2c communication with the radio
11:55:25DelazonBig chans to do your on stuff fairly easy.
11:55:34LinusNi believe it uses different pins
11:56:03Delazonhow manydevelopers are tou on this project?
11:57:38XavierGrReminds me that my radio patch stills sits idle that enables mutlipreset list handling...
11:58:24 Quit DJDD__ (Read error: 110 (Connection timed out))
11:58:35tucozDelazon, I don't know how many active developers there are, but the credits list more than a hundred. (all of them are of course not active)
11:58:47XavierGramiconn: No, because if I don't hold the array with the seek possitions I will have to loop when I use a read_line(). (if I have the seek position I will just seek there first and read only 1 line)
11:59:09Delazoncool many devs then :) many to thnx later on then i guess ;)
11:59:18amiconnI thought you will use fixed field leghts...
11:59:47XavierGramiconn: I think it is better to avoid that for various reasons.
11:59:50preglowlinuxstb: i thought as much
11:59:50amiconnThen seeking becomes icredibly simple.
12:00:01preglowlinuxstb: what's the default backlight time?
12:00:35Delazoncan i come with a suggestion?
12:00:42preglowlinuxstb: i'll let the if 0 stay in crt0.S when i commit
12:00:42XavierGramiconn: If I hold the seek positions it becomes simple too. I will just loose entries * sizeof(int) bytes from memory
12:00:50tucozDelazon, sure
12:01:25LinusNthe h300 lcd doesn't like dma...
12:01:32preglowlinuxstb: hooray
12:01:42amiconnpreglow: 4 seconds iirc
12:01:54preglowok, so the tick timer is probably off
12:02:01preglowno surprise
12:02:13XavierGroh I have another problem: the range o an unsigned int is 0-65535 what if I need to go beyond that? Is it right to have longs?
12:02:40preglowwhy shouldn't that be ok?
12:02:41Delazonwhen i use the player and the display shuts down or going in to sleep. To wake it upp you have to push a button. I would be glad if this push would result in the display woke up without the button push would for an example paus the musik or go in to the browser mode or change the volume... as it is in the iRiver FW.
12:02:51LinusNXavierGr: int is 32-bit on all current rockbox platforms
12:03:05XavierGrwhich means up to?
12:03:18preglow4 billion
12:03:19preglowmore than
12:03:31XavierGr~4 millions?
12:03:40amiconnint *can* be 16bit, but we don't need to care
12:03:50XavierGroh isn't this great!
12:03:51LinusNDelazon: you want the first push to only enable the lcd?
12:03:55amiconnThere can't be more than 65536 files in a FAT directory
12:04:00tucozDelazon, I like that idea as well. I normally use the record button for that, but I think it would be better if the first key press would wake up the lcd.
12:04:02DelazonYes exatly :D
12:04:02merbanan2raised by 32 for unsigned
12:04:18preglowLinusN: can you imagine any reason why i wouldn not want to use interrupts when implementing button driver?
12:04:28LinusNpreglow: no
12:04:40amiconn...and even 65536 files are only possible if none of them has long names
12:05:47XavierGramiconn: seek position can be larger than 65535.
12:05:56amiconnAh, yes
12:06:06amiconnShould be a long then
12:06:34XavierGrbut Linus said that an int is considered 32 bit
12:07:18preglowint is sometimes considered 32 bits
12:07:20preglowmost often, perhaps
12:07:22preglowbut it can be 16
12:07:26preglowif you need more than 16, use long
12:07:38XavierGrso how would I know?
12:07:42preglowyou don't
12:07:45preglowthere are no guarantees
12:07:58preglowyou also have int32_t, of course
12:08:03preglowwhich is always guaranteed to be 32 bits
12:08:12preglowbut those aren't recommended for use in rockbox, for some reason
12:08:14 Join muesli- [0] (n=muesli_t@
12:08:31XavierGrhmm long is double the bytes of an int, yes?
12:08:42preglowlong == int for most pplatforms
12:08:54 Quit muesli- (Client Quit)
12:09:16XavierGrso I don't need to change sizeof(int) to sizeof (long int) right?
12:11:42 Join Bger_ [0] (n=Bager@
12:12:13 Quit Bger (Nick collision from services.)
12:12:20 Nick Bger_ is now known as Bger (n=Bager@
12:12:59 Join webguest54 [0] (
12:13:13Bgeruse Rec button
12:13:21Bgerfor only lighting the LCD
12:13:54webguest54is the db scale (which I like), realtive to line out, and if so is line out taken before amplification or after ?
12:14:18Bgerwhile on this topic, is there any reason for the rec button doesn't have any function ?
12:14:39preglowwhat's the declaration specs for a naked function, again
12:14:57LinusNBger: no
12:15:23preglowwebguest54: relative to line out? it is relative to max level, regardless of output
12:15:47 Join webguest58 [0] (
12:15:48 Quit webguest54 (Client Quit)
12:15:55BgerXavierGr u should use sizeof(long int) if u use long int var ...
12:16:17linuxstbpreglow: What's a "naked function" ?
12:16:43preglowlinuxstb: a function without any wrapping code added by the compiler, afaik
12:16:55preglowlinuxstb: like implicit return instruction, i want to use my own return instruction
12:17:05Bgerhm, one more question
12:17:16preglowarm gets pretty fancy with its exception handler return methods
12:17:36Bgerwhat about making some button (like ON/play) to light the screen even when the hold button is turned on ?
12:17:55webguest58preglow:, thanks for the db info, now, is line out pre or post amplifier ?
12:18:01preglowBger: i wouldn't want to waste my battery for the backlight when my hold is on
12:18:07preglowwebguest58: post
12:18:17preglowwebguest58: only outputs that are pre-amp is optical out
12:18:17Bgerpreglow even on nano ?
12:18:28preglowBger: why even on nano?
12:18:41 Nick Lynx_awy is now known as Lynx_ (
12:18:42Bgerhm, i don't know whether nano has hold button
12:18:54preglowit does
12:18:55Bgerbut it has color LCD
12:19:01preglowit does
12:19:06webguest58so to achieve best line out output the volime should be set to max ?
12:19:15Bgerand u cant see the display with backlight off ...
12:19:16preglowbut if i enable hold, it is for a reason: i keep it in my pocket, and buttons might be pressed all the time
12:19:20preglowwhich means the backlight might be constantly on
12:19:24preglowi don't want that
12:20:11Bgerhm, my GSM brings the backlight on when i press the "call" button and i like this ...
12:21:16Bgeri mean even when the keypad is locked
12:21:59*Bger whispers "new option" and goes to the corner
12:22:45Jungti1234hey guys
12:23:31Bgeryes, Jungti1234
12:23:35Jungti1234Where is Rockbox option saved?
12:23:51Bgeron the hard disk
12:24:03Bgerbut in one sector out of the file system
12:24:30 Part webguest58
12:24:49LinusNsector 62
12:25:30Jungti1234How do I do to delete it?
12:25:42Delazonok thanx for your answers on the lcd matter...
12:25:49BgerJungti1234 : if you have problems with your settings, you can reset them with pressing "rec" key while booting rockbox
12:25:54Bgeror from the menu
12:26:22BgerManage settings->reset settings
12:26:33Jungti1234ah thanks
12:28:42*Bger remembers one of the first days of rb on h300, when he enabled the dircache and the unit self powered off right after booting because of the button driver problems...
12:29:13 Part LinusN
12:32:35preglowwhy would i get an interrupt from the ide controller?
12:34:29linuxstbMmm. I guess they haven't been disabled. I have no idea what the four cryptic lines in ata_init do.
12:35:15 Part tucoz ("Leaving")
12:37:25 Join webguest97 [0] (
12:37:38preglowno, but anyway
12:37:42preglowwhy would i get it?
12:37:52preglowif i'm expecting data, i guess i should know about it
12:37:54preglowhaving requested it and all
12:38:04preglowthen again, there's the latency, of course
12:38:33webguest97would it be possible to attach a different screen to the h-1x0 player via the remote output?
12:38:42linuxstbpreglow: So when are you getting the interrupt?
12:39:01webguest97i mean is the resolution given by the player or by the remote itself?
12:40:16preglowlinuxstb: i don't know if i'm getting it
12:40:25preglowlinuxstb: i'm just wondering to what degree i'm going to bother with it it all
12:40:28preglowit at all
12:42:17 Quit DJDD_ ("Trillian (")
12:43:15linuxstbDo you know what IPL does?
12:43:23preglownot in the least
12:44:24 Join San||Away [0] (
12:45:33 Quit webguest97 ("CGI:IRC (EOF)")
12:46:50linuxstbpreglow: If we wanted to cheat with the timing of the ticker interrupt, we could update the current_tick variable by using the value in the RTC variable we are currently using to simulate the current_tick.
12:46:59linuxstbI don't know how accurate the timer needs to be for other purposes.
12:48:04preglowwell, yeah, we could
12:48:10preglowbut i'd rather do it properly first
12:48:13preglowthen we'll see
12:48:16mirakhem, how do I make the tools now ?
12:48:21miraksomething have changed ?
12:48:32preglowyou type 'make' in your build dir
12:48:48mirakpreglow: after running a full configure ?
12:49:24mirakwhat devel option change ?
12:49:27 Part Polo_o
12:49:28mirakdebuging ?
12:50:21 Join Polo_o [0] (
12:51:51linuxstbpreglow: So you don't think it will be a problem to get the timer tick accurate?
12:52:41preglowlinuxstb: the answer to that question depends a lot on how tricky cpu freq change is
12:53:07preglowbut no, let's just try this approach first
12:53:11preglowi'll see about doing some commiting now
12:54:59linuxstbDo you update the current_tick variable yet, or have you left that the way it is?
12:59:42preglowi update it
13:00:37 Join Rob2222 [0] (
13:02:35 Quit _FireFly_ ("IceChat - Its whips the llama's butt")
13:03:56XavierGrin rockbox: char = 2bytes, long int = 4bytes. Please correct me if I am wrong.
13:04:32B4gderchar = 1 byte
13:04:38mirakhow does the lcd screen is represented in memory ? is it just an offset that you set and wich is located into ram, or does it have it's own ram ?
13:04:57B4gdermirak: we have a framebuffer that represents what the LCD looks like
13:05:06B4gdermirak: lcd_update() writes that framebuffer to the actual hw
13:05:26mirakB4gder: the buffer size depends of the device it's been compiled on
13:05:53mirakwich file is it defined ?
13:05:57B4gderand even how the frame buffer is layed out
13:06:07B4gdercheck firmware/drivers
13:06:11B4gderand the lcd*.c files
13:09:28XavierGrB4gder: and what about char* and long int*?
13:09:36B4gder32 bits
13:09:59XavierGrso 4bytes both.
13:10:21B4gderbut you should use sizeof() of course, if your code relies on the size
13:10:32XavierGryes of course
13:11:50XavierGrI have a plugin memory orgie right now. I have to set 3 buffers, at least in the end only 1 will be left.
13:13:41***Saving seen data "./dancer.seen"
13:14:52 Quit Delazon ("Chatzilla [Firefox 1.5/undefined]")
13:15:59*B4gder debugs code booting from flash without any kind of hw attached... tedious and painful are words that pop up
13:16:11B4gderdebug-hw I mean
13:20:11mirakwhat does it mean when there is only a line with (void)data;
13:20:24B4gderthat the variable isn't used
13:20:27linuxstbIt's used to stop the compiler complaining.
13:20:41mirakso that's a dummy method
13:21:00mirak /* TODO: Implement lcd_blit() */
13:21:05mirakI gues it is :)
13:21:18B4gderits not a "method"
13:21:22B4gderit is a variable
13:21:34linuxstbThat's probably misleading - I don't think that function is on the to-do list. It probably won't be needed for colour LCDs.
13:23:43BgerB4gder could you see patch No 1378064 ? i think it's ok ...
13:23:53Bger(sokoban h300 updates)
13:24:08B4gderno time atm
13:24:26Bgerokay :(
13:25:13 Join webguest37 [0] (
13:25:28webguest37Febs: are you around ?
13:25:57Jungti1234bad australian
13:25:59webguest37uggh, should open my eyes and look right −−->>
13:26:44Jungti1234Do they do racism yet?
13:26:45mirakB4gder: color is limited to 16bits ?
13:26:50*webguest37 does magical dance to summon Febs
13:26:50B4gderyes, atm
13:29:29 Part webguest37
13:31:22saa[b_r]idercustom H300 LCD colors :)
13:31:49saa[b_r]ider(this is an actual screendump, not from a simulator)
13:32:57B4gderI still wait for a background image and alfa blended text on top ;-)
13:33:42preglowcrt0.S is becoming a royal mess
13:34:12 Join mikearthur [0] (
13:34:36B4gderhow come?
13:34:45mirakB4gder: what is the reason to use 16bits instead of the 18 ?
13:34:50preglowsome general arm code
13:34:53preglowsome PP specific code
13:34:57preglowsome code that only goes in the bootloader
13:35:12B4gdermirak: why use 18bits when most of the UI only uses two colors?
13:35:42mirakB4gder: why do you answer that ?
13:35:43B4gderapart from the apparent reason that of course 16bits is much faster than 18
13:35:49mirakI want to to know what are the technical reasons
13:35:51B4gdermirak: ?
13:35:57B4gderno there aren't
13:36:16 Quit Bger ("Don't squeeze the BitchX")
13:36:23mirakso just changing the lcd depth would make it ?
13:36:41B4gderyou need to fix the code too
13:37:04 Join Bger [0] (n=Bager@
13:38:01B4gderlike most things in Rockbox it can be extended to become better if you just feel for it
13:38:12B4gderand have the skill
13:44:14preglowlinuxstb: but ok, so far, i won't enable interrupts in bootloader
13:45:28amiconnI don't think 18 bit support for H300 would be wise, as 16bit is kinda slow already...
13:45:51B4gderonly for some special mode if so
13:45:59B4gderlike jpeg viewer mode or similar
13:46:19amiconnI'd rather want to avoid LCD mode switching
13:46:24B4gderif you happen to have a jpeg with only blue shades or something ;-)
13:47:03B4gderwe have different "modes"already with the greyscale lib
13:47:45B4gderbut I don't think of 18bit as particularly important
13:48:07amiconnThe grayscale lib doesn't switch the LCD mode, it works on top of that
13:48:25B4gderyes, but it is a different graphic mode
13:49:12mirak#define _RGBPACK(r, g, b) ( ((((r) * 31 + 127) / 255) << 11) \
13:49:23mirakwouldn't it be faster to just use shifts ?
13:49:50mirakplus mask and a or to mix all colors
13:49:53miraka mask
13:50:40amiconnmirak: Shifts are imprecise
13:50:49mirakamiconn: what do you mean ?
13:50:54amiconnPlus, this #define is evaluated at build time
13:50:58preglowwhat the hell
13:51:00preglowhow can this work
13:51:08preglowahh, right, forget me
13:51:41amiconnmirak: I mean that the conversion between bit depths can't be done with shifts only. It wouldn't be precise
13:52:02mirakamiconn: explain that
13:52:10mirakplease :)
13:52:13preglowlinuxstb: for consistency, we might want to have the bootloader remap memory as well, but i don't know how the apple firmware, or even linux, will take to suddenly run at address 0x0
13:52:49mirakamiconn: the color passed to the argument are between 0 and 255 ?
13:53:25mirakok te method try to rescale to a 5bits ?
13:53:27mirakfor exemple
13:53:59mirakamiconn: ?
13:54:14miraka 5 bits intervals for red and blue
13:55:17mirakchunking just the low bits should do it for me
13:55:35linuxstbpreglow: Are you saying you map the DRAM to address 0x0 ?
13:56:07linuxstbMmm. That will make rolo-ing the Apple firmware or Linux tricky.
13:56:15linuxstbUnless it's easy to map it back.
13:56:18preglowit is
13:56:22preglowthat is
13:56:24preglowi have no idea how
13:56:33preglowor rather, how, but not what to write
13:56:45linuxstbI assume the IRAM stays where it is?
13:56:48preglowit's just a couple of registers
13:56:50preglowiram stays
13:56:52pregloweverything stays
13:56:58preglowbut sdram, and flash, obviously
13:57:01preglowwhich we can't access after this
13:58:09linuxstbBut I assume that the rolo function (which is in iram already) can change it?
13:58:39linuxstbOr the bootloader could change it back before executing Linux or the Apple firmware?
13:59:48preglowsure, just need to find out how
14:00:09preglowthis is the same way of doing things ipl does
14:00:17preglowthere is another mecanism for pp5020, it seems
14:00:22preglowbut not for pp5002
14:00:31preglowand this way is consistent with how other arms do it
14:00:36preglowso we can share some code
14:00:59linuxstbI am happy to forget about the pp5002 for now - a lucky owner of such a device can solve those problems.
14:01:11mirakamiconn: wouldn't this be enough ? (((r) >> 3 ) << 11 ) & (((g) >> 2 ) << 5 ) & ((b) >> 3 )
14:01:38preglowlinuxstb: well, now at least this part is working
14:01:46preglowlinuxstb: and besides, i never could get the other method to work
14:01:57preglowlinuxstb: ipl people haven't researched it too much, since they don't use it
14:02:17mirakI mean if the color is from 0 to 255 on 8 bits, if you want to reduce to 6 or 5 bits, you are forced to lose precision in the color.
14:02:22linuxstbmirak: As amiconn said, the shifts are inprecise - they always round down
14:02:55miraklinuxstb: ok I see
14:03:04preglownow it doesn't work anymore :/(
14:03:09linuxstbAnd it's a macro that's designed to be use at compile-time, so the speed doesn't matter.
14:03:17miraklinuxstb: I wasn't figuring the rounding down
14:03:39Bgerhm, guys, anything other than unix2dos and dos2unix ?
14:03:51linuxstbiconv or recode
14:04:03linuxstb(I'm not sure about iconv....)
14:04:16linuxstb(in fact, forget i mentioned it...)
14:05:07Bger10x :)
14:05:07B4gderBger: to do what?
14:05:15Bgerjust to ask you about iconv ...
14:05:25preglowi just built a rockbox that rebooted on start
14:05:38BgerB4gder i want to convert one patch from dos to unix eol format
14:07:15*B4gder uses tr for that
14:09:00Bgerhm, good idea
14:09:41B4gderI actually have a perl script too, but that's mostly for windows-versions of diff that produces the drive letter and backward slashes in the diff -u output
14:10:10 Join linuxstb_ [0] (
14:10:13 Join muesli- [0] (n=muesli_t@
14:10:20 Quit linuxstb (Nick collision from services.)
14:10:29 Nick linuxstb_ is now known as linuxstb (
14:10:34XavierGrGOD!!!!! I am gonna rip my hairs out!!!!
14:10:56Bgerthis snow.c patch is awful ...
14:11:02XavierGrI can't make the pointers work
14:11:44 Join dropandho [0] (
14:12:01dropandhohey all
14:12:08dropandhoany news regarding the forums?
14:12:20B4gderJeff is not here
14:12:37dropandhogot it
14:12:46dropandhothere have been some attempts to reach out to him
14:12:53dropandhoso i was hoping something came of that
14:13:02XavierGrI have 2 char* pointers how can the one pointer the other? Or one of 2 must be **pt?
14:13:04B4gdernot to my knowledge
14:13:32B4gderXavierGr: I don't understand the question
14:14:39BgerXavierGr if you have char *pt1,*pt2;
14:14:53Bgerthen u just make pt2=pt1; and it's done
14:16:39preglownow, why did i disable backup files in vim...
14:16:41XavierGrI will describe it better. I have 2 buffers. 1) The first holds all the strings (char *names) 2) (Let's name that str_pt) The second has the places where each string starts in "names". I want every time names get another element to be pointed by str_pt.
14:17:10XavierGrpreglow: dont tell me?
14:17:30XavierGryou lost your work?
14:17:39preglowjust a tiny bit
14:17:44preglowwill be back in a jiffy
14:17:47BgerXavierGr so you need an array of pointers to char iiuc
14:18:23Bgerlike char *str_pt[MAX_NAMES];
14:18:40Bgerstr_pt[i]=names[j] etc ...
14:19:01XavierGrno because I can't be sure how many I will have
14:19:16XavierGrso I use a buffer for that too
14:19:52B4gderyou need a max value anyway
14:19:52 Join DjDeaf [0] (
14:20:30BgerXavierGr : as u already know, rb doesn't have dynamic memory allocation
14:20:40XavierGrBagder: Yes but at least I will get that value during runtime.
14:20:44Bgerso, whether u want or not, u must assume some limit
14:21:04XavierGrThe limit comes when the code is running
14:21:13DjDeafhi i have a problem can abyone help me?
14:21:22BgerDjDeaf just ask
14:21:30DjDeafive installed rockbox
14:21:36DjDeafon h340
14:21:41Bgernoone can help you if (s)he doesn't know what's the problem
14:21:42XavierGrbut let me think it over again.
14:21:55DjDeafand when im accessing the menu
14:22:08DjDeafi cant see it , the text is blank
14:22:09BgerXavierGr if the limit comes runtime, then u want dynamic memory alloc. ..
14:22:24BgerDjDeaf any language ?
14:22:31Bgerdifferent from english ?
14:22:36Bgeror any font set ?
14:22:40DjDeafmaybe i didnt notice
14:22:45DjDeafits in the originalk theme
14:23:14preglowok, commited exceptions
14:23:16Bgeri suggest u to stop and start the unit and hold the "rec" button while it's booting
14:23:35Bgerbut not too early, otherwise u'll start the iriver fw
14:23:40 Quit dropandho ()
14:23:51DjDeafok i did it
14:24:00BgerB4gder am i right in using tr -d "\r" < dos_formatted_file > unix_formatted_file
14:24:07DjDeafoh here, it works
14:24:12DjDeafthanks :)
14:24:26BgerDjDeaf np ... this resets all settings, it was added recently
14:24:36DjDeafand another tiny question, it there video on images option on rockbox?
14:24:50Bgervideo : not
14:25:19Bgerimages also, but they will come in a more short period
14:25:49XavierGrBger: Excuse me but when I run a plugin I have a big chunk of memory available to do anything I like. Maybe it is called dynamic memory allocation, but anyway rockbox lets me to do anything I like with that left memory in the end.
14:26:16BgerXavierGr it has statical limit, so it's statical
14:26:40 Join linuxstb_ [0] (
14:26:45 Quit linuxstb (Nick collision from services.)
14:26:48 Nick linuxstb_ is now known as linuxstb (
14:26:59preglowlinuxstb: care to test what's in cvs now?
14:27:05linuxstbXavierGr: you could store your strings starting from the start of the buffer, and the array can work back from the end.
14:27:11Bgeru can use all this memory to make an array and after that use this array in the way u want
14:27:22preglowhere filebrowser starts, but then just stands there looking silly
14:27:36Bgeryep, something like the linuxstb's note
14:27:43linuxstbpreglow: Hasn't it always done that? :)
14:27:44preglowbut then again, i've got some further changes
14:27:58preglowlinuxstb: perhaps, i've forgotten how it behaves in cvs, heh
14:28:32XavierGrguys I think I fixed what I wanted, one of the char pointers had to be char **str_pt
14:28:35preglowlinuxstb: but anywho, as long as it doesn't lock up
14:28:48linuxstbpreglow: I'm testing now.
14:28:48BgerXavierGr ... i don't think so
14:28:50XavierGrIt is just annoying that I haven;t clearly understand how pointers work.
14:28:55Bgerdo u initialize the pointers ?
14:29:10Bgersee, before you use any pointer, u must initialize it
14:29:32Bgerso, how do u initialize it ?
14:29:54XavierGrthese pointers get always a start address from the end memory of plugin buffer.
14:30:06Bgerchar **str_pt; means pointer to pointer to char
14:30:12XavierGr(different according to my use)
14:30:26Bgerso u must assign something to str_pt
14:30:32XavierGrthat's what I wanted after all, it is just that I am not sure about it.
14:30:52miraklinuxstb: ok, I will insist a bit ^^. For me the precision doesn't matter. What the actual thing is doing is considering that the colors below and above in fixed interval can be assimilated to that color. But in reality the colors truncated with a right shift are considered to be a subset of a base color. For exemple with what you are doing with 6 bits for a taint, the color 0 have a subset of colors of 2 colors {000 001
14:30:52mirak} while the color 100 have a subset of {010,011,100,101}
14:30:56Bgeri mean, str_pt must point to free memory
14:31:00XavierGryes I assign a position from the other buffer
14:31:53XavierGrI must read a very good tutorial about pointers which covers all strange situations.
14:31:54miraktruncation is the way to go for colors
14:33:16mirakimho. It was this approach wich was used on Atari ste when they extended the color range
14:33:17linuxstbpreglow: Your CVS changes make no difference to how Rockbox behaves :). Which is a success I think.
14:33:36BgerXavierGr : really, it's simple
14:33:37preglowlinuxstb: indeed, then i shall queue forth a new set of changes
14:33:48preglowlinuxstb: btw, what to do about the yield issue? :/
14:33:55miraklinuxstb: very intersting ;D
14:34:05Bgeru just must know that the pointer is a var that points to other var
14:34:14 Join amar [0] (
14:34:31Bgerand u must give the pointer a valid address
14:35:41XavierGrBger: I know that, but sometimes I get confused by the indirection.
14:36:28Bgerthen u just need more practice
14:36:41mirak0,8 should be considered as 0 and not 1 for the same reason that the color 255 will not be rounded up, because it's just impossible to go above 255.
14:36:43linuxstbmirak: I don't really have any views about that macro - amiconn wrote it. But I think it may have something to do with making the reverse operation possible (i.e. rgb565 to rgb888). If you do the reverse operation by simply shifting, then you don't get pure white any more.
14:37:29linuxstbBut as I say, I'm not the person you should be talking to.
14:39:24miraklinuxstb: that's the inevitable boundary effect
14:39:33 Join Kohlrabi [0] (
14:40:02 Join Mmmm [0] (
14:40:05mirakdo you know how the screen manage colros internaly ?
14:40:20mirakdoes it use one byte per colors, or is packed ?
14:40:38mirakto save bits
14:41:07 Quit Mmmm (Client Quit)
14:41:11B4gderit is packed
14:41:19B4gderthe framebuffer uses lcd native format
14:41:48 Join Mmmm [0] (
14:42:02mirakB4gder: ok so you can control the depth of the lcd screen ?
14:42:25 Quit Mmmm (Client Quit)
14:42:43mirakthe screen uses 18 bits colors, rockbox 16 bits, something doesn't match :)
14:42:55B4gderthe LCD is programmable
14:43:00 Part Sando
14:43:02B4gderrockbox sets it in 16bit mode
14:43:09 Join Sando [0] (
14:43:12mirakok thank you
14:43:22mirakB4gder: is there some docs about that ?
14:43:31B4gderthe source code and the lcd datasheet
14:44:52mirakthis reminds me my atari days, where programming was very close from the hardware
14:45:26*B4gder never left that :-)
14:45:29linuxstbmirak: I believe the LCD controller uses 18bpp internally, but it can work in a 16bpp where the missing two bits are just set to (I think) 1.
14:45:45linuxstb(i.e. the least significant red and blue bits)
14:45:58linuxstbAs B4gder said, the lcd datasheet describes it.
14:46:22 Quit San||Away (Read error: 104 (Connection reset by peer))
14:47:09XavierGrAre we sure that iriver uses 18 bit colors?
14:47:17XavierGrmaybe they use 16 too. :D
14:47:19mirakhow was it possible to find wich system adress were coresponding to this or that device ?
14:47:37XavierGrWho can tell?
14:47:44B4gderXavierGr: we're not sure no, I doubt anyone cares enough to figure it out
14:47:56B4gdermirak: hard work
14:48:26XavierGrThen they say that their product shows 18bit colours so.... it could get them into trouble if not.
14:48:28mirakXavierGr: yep, anyway since the screen resolution is around 5 times less than traditional screens resolution with a 24 bits depth, I don't think that any image converted to that resolution can display that much color nuances
14:48:59mirakso yes 16bits colors is not much of loss probably
14:49:12B4gderI think you should be able to tell
14:49:17linuxstbAnyone heard anything more from the person working on the iFP790 port? He doesn't seem to have updated the wiki or his website since his initial announcement.
14:49:20B4gderif you make a pic with a gradient blue only
14:50:37XavierGrI will make the same discussion then: Does human eye has a measurable colour range?
14:51:09XavierGr(The range is known, I think)
14:51:17XavierGrbut how manu increments?
14:51:26g33i have a suggestion for filebrowsing... dunno how easy it is to implement though..
14:51:38mirakXavierGr: that's subjective and depend of persons
14:52:02mirakXavierGr: there is probably a physiological limit
14:52:05preglowB4gder: just doing MY_ADDR; on a volatile address will read it, yes?
14:52:26XavierGrmirak: yes that's what I am talking about.
14:52:50linuxstbpreglow: Yes - look at how current_tick is defined and used.
14:53:10mirakXavierGr: I think 24 bits is below that limit, but enough
14:53:10XavierGrIs there any plan to use 64bits for displays?
14:53:26preglowok, here comes timer
14:53:30g33you know sometimes the idiot who ripped an album put the artist name and album name in the file name, so when you list the files in a folder all you see is 15 "eric clapton - from the cra.." and have to wait and scroll every file to find the right song
14:53:30preglowlinuxstb: please test
14:54:02mirakg33: I have seen idiots that put album and artist into the artist id3 tag
14:54:45g33man i should spend one day and just sort out all my mp3s
14:54:54mirakg33: I have done that
14:55:16g33maybe i can pay some polish kid to do it for me! :D
14:55:29mirakat the begining I was puting artist - album - track - title in the name, because it was easier for P2P
14:55:38linuxstbg33: There are tools to have retag and rename MP3s
14:55:53g33yeah but i dont always trust em automatic tools
14:55:54mirakI use easytag on linux
14:56:00g33im not on linux
14:56:04mirakg33: that's semy manual
14:56:09g33im l33t so i use xp home :)
14:56:30mirakI have seen that some uses musical signatures to identify songs
14:57:00mirakit means that two mp3 encoded with different bitrates by different persons can have the same signature
14:57:05linuxstbpreglow: That's perfect - the backlight now fades.
14:57:15mirakeven if they don't match bit for bit
14:57:20preglowok, button driver next!
14:58:51markung33: with 'mmv' you can rename files very easy under linux.
14:59:06g33[13:56] <g33> im l33t so i use xp home :)
14:59:08markunAh, NOT on linux :)
15:00:02g33im learning c++.. maybe one day ill write my own tool..
15:00:11g33eta 2009 :P
15:00:42 Join webguest11 [0] (
15:06:10 Part webguest11
15:06:28CassandraI had a look at the button code in ipodlinux. It looked kind of scary.
15:06:35CassandraAnd interupt driven.
15:06:53preglowyeah, it did
15:07:06preglowbut i'm going to have a crack at it now, or so help me god!
15:07:12preglowi like being melodramatic
15:07:15CassandraGood luck with that.
15:07:28preglowat least we have interrupts now
15:07:29CassandraI kind of got distracted by the Secret Project.
15:07:31preglowwhich isn't a bad start
15:07:47preglowthe wha?
15:07:54CassandraThe bear bones of which is now implemented. It's looking good. I may have a first release in a few days.
15:08:20CassandraIt's a wxWidgets program.
15:08:26CassandraErm, "bare" bones.
15:08:39preglowit'd be cooler if it contained bear bones
15:09:02preglowwasn't it you who was going to implement the sims with wxwidgets?
15:09:42CassandraSomeone suggested it was, that, but no.
15:10:04CassandraAlthough that might be an interesting project at some point.
15:10:17CassandraHowever, I regard the sim as low priority.
15:10:27g33hey how about making the iriver to a digital soundcard? you connect the iriver to computer with usb, and then it works as a usb soundcard and you can use the digital output from the computer? :D
15:10:38preglowyup, but i think using something like sdl would be wiser for that
15:10:43preglowwe don't exactly need widgets for sims
15:11:22preglowbut that would be a cool project
15:11:22Cassandrag33: Can't be done, I'm afraid. All the computer sees is a disk.
15:11:34preglowwould eliminate the need for multiple targets, and would probably result in a better simulator
15:11:50Cassandrapreglow, on the other hand it would mean polluting the sim with C++.
15:11:55preglowCassandra: oh?
15:12:01preglowsdl is c
15:12:10CassandrawxW is C++, sadly.
15:12:18preglowahh, yes
15:12:19preglowthat it is
15:12:21preglowbut again
15:12:24preglowsdl would be better
15:12:25g33ohwell.. usb sound card is only $20 anways
15:12:27preglowsince it's got sound drivers as well
15:12:32CassandraPart of whats been taking time on this project has been learning C++.
15:13:05CassandraSounds like a more sensible choice, yes.
15:13:16 Join PaulJ [0] (
15:13:45***Saving seen data "./dancer.seen"
15:14:11mirakamiconn: are you around ?
15:14:16 Quit Jungti1234 ("bye")
15:16:12 Join tucoz [0] (
15:16:31tucozCassandra, are you here?
15:17:04 Join LinusN [0] (
15:17:57preglowlinuxstb: btw, did you try the new ipl bootloader?
15:18:08XavierGrI will be damned now sorting 850 files droped to 68 ticks
15:18:10 Join Febs [0] (
15:18:20XavierGrPrevious was 618 ticks for the same list!
15:18:56XavierGrNow I have a killer jpeg viewer, optimised as hell.
15:19:24XavierGrI think 1 second for 850 files is a good deal.
15:19:35*preglow points to the dct functions, which are surprisingly asm free, just to spoil the fun
15:19:49B4gderXavierGr: and it does color on h3x0 ? B-]
15:19:50mirakXavierGr: can we see that ?
15:20:31XavierGrBagder: Unfortunately I am too idiot to understand the jpeg code to do that.
15:20:44*LinusN drops the dct functions in preglow's lap
15:20:51linuxstbpreglow: No, it needs yet another version of gcc installing.
15:21:01XavierGrmirak: Of course I will put some more memory failsafe code and I will upload it to the tracker.
15:21:23preglowlinuxstb: eh?
15:21:24Cassandratucoz, hello, yes.
15:21:32*preglow looks at the viper in his lap with disgust
15:21:52tucozCassandra, if you find some time. Would you like to have a look at the latex docs? It is organized like the 2.4 docs atm.
15:22:47CassandraNot sure when I'll have time though.
15:23:45tucozCassandra, not that this is looking that good. It is only the writer2latex output, that I split into chapter files and made compilable :)
15:24:35preglowlinuxstb: do you have any idea why the ipl button code for 4g and up seems to use both the mini and 4g interrupts?
15:25:09*Bger spots an old code in playlist_viewer.c
15:25:14linuxstbI think the "mini" interrupt is needed for the hold switch
15:25:58linuxstbAnd forget what I said about needing a different compiler for the new ipl bootloader - I was thinking about podzilla2
15:26:24preglowwhy the hell would one need a different compiler for compiling a program?
15:26:24linuxstbThat needs an arm-uclinux- toolchain and libraries
15:27:10tucozCassandra, this zip includes the LaTex build system. It should be quite easy to make it look prettier from here. Make the chapters (parts) more organized. The file structure is what is important right now I think.
15:27:11 Quit VikraMarkA_ ("Leaving")
15:27:34preglowtime for food
15:27:34Cassandratucoz: I've had a quick look, and it looks like a very good start. You seem to be missing the appendices and introductory blurb.
15:27:58tucozCassandra, yes, but the files are there
15:28:29CassandraCool. Just having a peek inside the archive now.
15:28:37tucozI didn't bother to make the appendeces compile, but the text is there in appendix.tex
15:29:14tucozCassandra, I think some of the text needs some attention as well, as some features have changed since 2.4.
15:29:23CassandraHow clean is the generated code?
15:29:35tucozCassandra, and of course the tables. The generated code looks like ****
15:29:48Cassandratucoz: Yeah, it needs a lot of attention in fact.
15:29:57tucozAnd I generated it with the -clean flag
15:30:08Cassandra*nod* very sensible.
15:30:09tucozthe generated docs is here
15:30:28CassandraFor a first shot, I think it looks better than I'd have expected.
15:30:59tucozYes, me too. But as you see, e.g. the tables need some major work.
15:31:28tucozBut, it should serve as a starter at least.
15:32:52miraklinuxstb: just a question, what would be the point of converting from 16 to 24 ?
15:32:52preglowi think it should do nicely for a starter
15:33:22CassandraCode is nice and clean too.
15:33:30 Join perplexity [0] (n=joust@
15:33:30tucoz(and of course fix the references, as I didn't fix all of them)
15:33:31linuxstbmirak: There probably isn't any point any more - but the first colour version of the Simulator worked in 24-bit
15:33:31B4gdermirak: for screendumps for example
15:33:35CassandraI think that's a very good starting point, tucoz.
15:33:41tucozCassandra, :) nice
15:33:44linuxstbB4gder: The screendump also uses rgb565
15:33:55B4gderbmps can do 565?
15:34:06linuxstbWe write a 16-bit bitmap.
15:34:09*B4gder learned something
15:34:31miraklinuxstb: I was reading the datasheets and it seems you can pass directly 18bits and it does the conversion in hardware on the lcd side
15:34:46mirakfrom 18 to 16 bits
15:34:53tucozCassandra, ok, now you know where they are. I don't think I work on that before we could send changes as patches.
15:35:02Febstucoz, FYI, everything in the WikiManual (at least everything that I wrote) started from version 2.4 of the manual if it existed there in the first place.
15:35:10linuxstbmirak: I thought it was the other way around - you write 16-bits and the hardware converts it to 18-bits.
15:35:12B4gdermirak: why would you want to do that?
15:35:55miraklinuxstb: it can do that too, wait a minute :)
15:35:58tucozFebs: good. I think the wiki will serve as a good source of copy and paste for now :)
15:36:26mirakB4gder: to not to do the conversion in software mode
15:36:27FebsGreat. That was always my intention as I updated the 2.4 sections and put them in the WikiManual.
15:36:42linuxstbmirak: What are you trying to achieve? Do you want to make the h300 lcd work in 18-bit mode?
15:36:44B4gdermirak: you'd have to write a lot more data, which is bad
15:36:49FebsI need to get LaTex installed on my computer so that I can see what you've done.
15:37:06tucozFebs, you could look at the pdf?
15:38:06mirakB4gder: yep
15:38:09tucozFebs, that is mostly generated code from the writer2latex command, which I have cleaned up and made to compile.
15:38:34miraklinuxstb: no not really, I am figure out how it works globally, so sometime I fix on some details
15:39:18mirakmaybe there is some docs about rockbox interfaces ?
15:39:42B4gdermirak: nope, none that are any recent
15:39:53linuxstbThere's a good "graphics API" wiki page that amiconn keeps up to date.
15:39:58tucozFebs, you can find the other sources in that directory. That is, the writer2latex output (which won't compile) and the latex sources.
15:40:04mirakLinusN: ok I will read that
15:40:36Febstucoz, looks like a great start.
15:41:15tucozFebs, at least under the hood :)
15:41:58FebsHey, you must start somewhere!
15:42:47tucozyes, that is true. Now to agree on what to put in cvs, and start merging the docs from the wiki with this.
15:42:49*Febs is now installing XEmTeX.
15:43:02tucozFebs, are you using windows?
15:43:24tucozOk, is that a good LaTeX system?
15:43:29 Quit muesli- ("ich will Kühe!!!")
15:43:36mirakI guess it's possible to call the file browser from an application ?
15:43:50Lynx_just got my h340!
15:43:56mirakdman that's not easy
15:44:11tucozI have only used MiKTex under windows
15:44:15Febstucoz, having never used any LaTeX system, I have no basis for comparison.
15:44:27tucozOk, probably good enough
15:47:15 Join webguest27 [0] (
15:47:19Lynx_argh, why in the world would they put a paper sticker on the display that tears on removal
15:47:47saa[b_r]iderLynx: wow, the one with the pixeled dancers on it?
15:48:34*webguest27 thanks every develloper here for all the Rockbox works !!!
15:48:36mirakextensions are usable on the simulator ?
15:48:39Lynx_saa[b_r]ider: yes
15:49:09saa[b_r]ideryou're a bit late to the party :D
15:49:29mirakinvalid ELF header
15:49:35mirakwhen I try to run an extension
15:49:40webguest27anyone here now if we can replace the doble plater 40go HD of h140 by a 80 go one?
15:49:42Lynx_hmm, it seems i didn't get any remote with it :(
15:49:54saa[b_r]iderlynx: is this your first DAP, or did you have another one before it?
15:50:05webguest27seems Toshiba made 80 doble plater HD now
15:50:10Lynx_saa[b_r]ider: had an archos until last week
15:50:27saa[b_r]ider(the LCD remote is optional, only comes with the non-LCD one)
15:50:41webguest27iHP- 180, that would be the best music device :
15:50:43Lynx_saa[b_r]ider: it didn't come with any remote
15:50:56Lynx_saa[b_r]ider: unless i haven't found i yet. but it's not listed on the box
15:51:13webguest27anyone know?, please
15:51:14saa[b_r]iderwebguest: they announced it a while ago, but afaik it's still not available
15:51:48webguest27but when it will be possible to by, will it be compatible with h140 one?
15:52:15saa[b_r]iderlynx: to be honest, I never used that remote. I only connected it cause the the standard headphones are short
15:52:15webguest27I love my hp140 and don't want to change it
15:52:22linuxstbmirak: Are you trying to use a plugin compiled for the target on the simulator? You need to use the simulator versions - they are installed when you do a "make install" in your simulator build directory.
15:53:13Lynx_saa[b_r]ider: a remote was the only thing i was missing on the archos. i thought there would be one with this iriver because you told me some time ago, but they seemed to have changed this.
15:53:22 Join tvelocity [0] (n=tony@
15:53:27saa[b_r]iderwebguest: check the MR forums, or check the specs on toshiba's site. techincaly, it should fit if the dimentions in the specs are correct
15:53:56webguest27but I heard 2 differents size for it
15:54:33webguest27and don't know wich is a good one :(
15:55:16saa[b_r]iderlynx: well my H340 came with it... but changes did happen. my H340 came with the non-windowed case. newer H340s came with the windowed case (which I had to buy my self, together with the LCD remote)
15:55:34Lynx_saa[b_r]ider: ok, i have a newer one then
15:55:49linuxstbwebguest27: The h140 has a Toshiba MK4004GAH - I am sure you can find the dimensions of that drive on the web and compare it with the one you want to buy.
15:55:53saa[b_r]iderwebguest: beleive me, a lot of people are waiting for it... including me
15:56:31webguest27linuxstb: afaike isn't the iriver size problem, but the 80go HD one
15:56:43miraklinuxstb: yes I am dumb
15:56:50saa[b_r]iderlynx: you got the windowed case?
15:57:00Lynx_saa[b_r]ider: yes
15:57:30Lynx_saa[b_r]ider: maybe i can buy the non-lcd remote off someone who has it and does not need it like you
15:57:31Bgersaa[b_r]ider btw if you're using the h300 lcd remote, i've updated the button patch
15:57:36webguest27saa[b_r]ider: you want it too :), did you see if we can upgrade or no yet?
15:58:18saa[b_r]iderI would suggest that you get a t-dimention skin, or rock.. I forget the name, always confused me with rockbox
15:58:37saa[b_r]iderBger: the button.diff patch?
15:58:42Bgeryes, the same
15:58:43miraksaa[b_r]ider: I have the windowed case, it's a bit large
15:59:18Bgernow the remote behaves much more like the buttons on the main unit
15:59:30webguest27saa[b_r]ider: but apparently Toshiba said 2 differents size for the same product, and I don't know which one is good :(
15:59:47saa[b_r]iderwebguest: it's not out yet, so I still don't know. we'll have to wait :)
16:00:36webguest27ok thanks for help me a bit
16:00:42webguest27I'll wait
16:00:49webguest27in hoping it will fit
16:00:57webguest2780go O.O
16:01:06saa[b_r]iderbger: yesterday I used the remote_type.patch, to scroll through menus you have to use right and left.
16:01:23Bgersaa[b_r]ider yes, i know
16:01:29saa[b_r]iderwhen I tried to patch using button.diff I think nothing happened
16:01:47Bgerhm, see, u can't use the 2 patches at the same time
16:01:48saa[b_r]iderso how does the new patch work?
16:01:51Bgerdo u have 2 remotes ?
16:01:55webguest27ciao everyone
16:02:02 Quit webguest27 ("CGI:IRC")
16:02:07saa[b_r]iderI didn't use them at the same time
16:02:11saa[b_r]idercompiled twice
16:02:16Bgeri mean, h100's one and h300's one
16:02:24saa[b_r]idernope, only the H300
16:02:39Bgerthen u don't need to change settings
16:03:30saa[b_r]iderso if I patched a new build with button.diff, how are the buttons now?
16:03:34Bgerwith the new version u scroll with up/down
16:03:45Bgerand enter with left/right
16:03:52Bgerin menus, in browser ...
16:03:55saa[b_r]iderwhat about the +10 and -10?
16:04:02Bgerpage up/page down
16:04:04mirakis there a way to slow down the simulator ?
16:04:10mirakplugins display nothing
16:04:13mirakor to fast
16:04:26Bgermirak put a sleep :)
16:04:43linuxstbmirak: Which sim are you using?
16:04:56linuxstb(i.e. x11 or win32)
16:05:06Bger*exit/enter with left/right*
16:05:07linuxstbFor the h300?
16:05:18linuxstbThat doesn't work...
16:05:25linuxstbNo-one has added colour support to the x11 sim yet.
16:05:46saa[b_r]iderBger: I think I'll have a go at it today. last night I changed the colors in lcd.h, but they still need tweaking
16:05:48mirakthe cube displays something
16:05:49mirakbut bad
16:05:56linuxstbBut if you have Wine installed and a windows cross-compiler, you can use the win32 sim.
16:06:18 Part tucoz ("Leaving")
16:06:18linuxstbWhich Linux distribution are you using?
16:06:19mirakI don't have win32 cross compiler
16:06:23mirakI use ubuntu
16:06:29B4gderthen you can have it
16:06:30mirakon x86
16:06:42mirakwell I must built it
16:06:50mirakI don't want to do that no
16:06:52linuxstbNope - there is a Debian package.
16:06:56B4gderwe use the debian package fine
16:07:02mirakok great
16:07:17linuxstbI use it as well to build the windows sim (and other things)
16:07:35B4gderbut of course we'd also appreciate a fix for the x11 sim!
16:07:44linuxstbMaybe the fix is to port it to SDL
16:07:53B4gderyes, I would agree to that
16:08:14linuxstbDoes the windows sim use cygwin or is it a native windows app?
16:08:48linuxstbObviously it doesn't use cygwin....
16:09:03linuxstbOtherwise mingw32 wouldn't work.
16:09:05saa[b_r]iderlynx: btw, the windowed case scratches your H300 after a while, that's why I suggested the skin
16:09:24saa[b_r]iderplus it's bulky
16:09:54Lynx_saa[b_r]ider: yes, i usually don't use those things
16:10:14Lynx_saa[b_r]ider: how much is the lcd remote?
16:10:39BgerLynx_ i suggest u to contact someone in korea...
16:10:51Bgeror buy the h100 lcd remote
16:11:07Bgeri bought my this way: through Kylera @ MR
16:11:39saa[b_r]iderI'm in china, and I asked a korean classmate to get it for me during our spring break :)
16:12:07amiconnmirak: Now I'm here (sort of)
16:12:20Lynx_Bger: is it cheaper in korea? its 20 euro at amazon
16:12:26saa[b_r]iderI've gotten so used to it, I barely take out my H340 out of my bag. I just use the remote
16:12:40BgerLynx_ ? 20 eu for *h300* lcd remote ?
16:12:45Lynx_ah, not that's the no lcd one
16:12:57saa[b_r]iderno one wants that :)
16:13:09Lynx_saa[b_r]ider: i don't really care, i just want any remote
16:13:22Bgerthis is a robbery for non-lcd :)
16:13:24mirakamiconn: it was about the precision thing
16:13:29amiconnB4gder: Standard 16 bit BMP is actually 15 bit RGB555, but you can do all sorts of 16 bit RGB, i.e. RGB565, RGB664 etc. by using the bitfield feature of BMP
16:13:33saa[b_r]iderlynx: did you check if they still have the LCD remote in MR's on-line shop?
16:13:44Lynx_saa[b_r]ider: MR?
16:14:11saa[b_r]ideryour new friend for everything iRiver :D
16:14:14mirakamiconn: what I was saying is the truncation was the way to go, because, the extra bits are just a subset of the base color
16:14:23amiconnNo, that's wrong
16:14:59amiconnBy having less bits per colour you're getting fewer levels, so you have to round.
16:15:06Lynx_saa[b_r]ider: didn't know they have a shop... There's one for 50 euro currently at ebay. that's a lot of money for a remote
16:15:23Lynx_saa[b_r]ider: i'll buy your non-lcd remote for 5 eur ;)
16:15:37amiconnThe rounding should happen towards the nearest destination level, and in addition, black always stays black and white always stays white
16:15:38mirakamiconn: For me the precision doesn't matter. What the actual thing is doing is considering that the colors below and above in fixed interval can be assimilated to that color. But in reality the colors truncated with a right shift are considered to be a subset of a base color. For exemple with what you are doing with 6 bits for a taint, the color 0 have a subset of colors of 2 colors {000 001} while the color 100 have a
16:15:38miraksubset of {010,011,100,101}
16:15:46 Join Hooligan [0] (
16:16:10saa[b_r]iderlynx: sure, and i'll charge you 45 euros for shipping :)
16:16:32Lynx_saa[b_r]ider: heh, do you live on the moon? ;)
16:16:45amiconnmirak: Yes, and you can't achieve that with right shifting
16:17:06saa[b_r]iderclose, china :D
16:17:24amiconnThe point is that the lowest and the highest interval have half the width of all others
16:17:27mirakamiconn: what I am saying is that 0 should also have a 4 colors subset
16:17:39amiconnThat would be plain wrong
16:18:20saa[b_r]iderlynx: US$ 61 at misticaudio, and backordered! don't know how much that is in euro
16:18:40 Join muesli- [0] (n=muesli_t@
16:18:46mirakamiconn: this doesn't matter, because no matter what, 4 taint of green in 8 bit per taint mode will be assimilated to the same color in 6 bit mode.
16:18:50Lynx_saa[b_r]ider: yes, i saw that. i don't think i will buy one for that price
16:19:14 Join b0br [0] (
16:19:33mirakwith what you are doing the color 0 will have 2 and the highest will have 6. Well whatever, but that's not even
16:20:03*saa[b_r]ider checking iriver china online shop
16:20:09amiconnI spread evenly, the bottom and top intervals both have half the width of the rest
16:20:21mirakwhat I am saying is that if you go from 4 nuances −−-> the same nuance, truncating the lowest bytes is enough and not less accurate than doing a rounding
16:20:24amiconnThat's why I calculate /255, not /256
16:20:41amiconnIt is less accurate
16:21:36miraksince colors are linear it's not
16:21:53amiconnThe error is bigger
16:22:15amiconnIf you always round down, the change in brightness can be 0, -1, -2 or -3
16:22:22saa[b_r]iderlynx: not available... I'm hoping one day I can use my non-LCD remote as a modded car remote or something (over ambitios)
16:22:42mirakamiconn: explain the numbers
16:22:59linuxstbmirak: You are dividing by four and trucating.
16:23:23amiconnHowever, if you *round* the change is +1, 0, -1, -2 (or +2, +1 ,0, -1 depending how you treat the "border values")
16:23:52 Quit B4gder ("time to say moo")
16:24:05amiconnThe number mean the error when representing the old values (full scale) with the new ones (reduced scale)
16:24:09mirakfor me the reason it doesnt matter is that the difference between 0,5 and 1,5 is still one
16:24:34mirakthe error is still 4 no matter what you do
16:24:40amiconnIt's not only the difference that matters, but also the absolute value
16:24:45mirakhem not 4 , 1
16:25:10mirakwell I don't think that we can perceive a visual difference
16:25:44amiconnIf you just truncate, the pictures will end up too dark in the dark parts, and too bright in the bright parts
16:26:22amiconnIf you make a simple drawing on a paper (with few levels) or a spreadsheet, you'll see what I mean
16:26:22preglowlinuxstb: just tested with the yield uncommented again, works just dandy
16:26:41 Quit muesli- ("ich will Kühe!!!")
16:26:45linuxstbSo what changed?
16:27:11preglowthe fun thing about not understanding the error in the first place, is that i don't care :)
16:27:16amiconnIt took me quite a while to figure this out, we implementing the conversion functions in the grayscale lib´rary
16:27:27mirakamiconn: I agree with darkness thing, but not with brightness :)
16:27:32linuxstbpreglow: hehe. Let's just hope it doesn't return.
16:27:32preglowi'd appreciate it if you tried it too
16:28:34amiconnmirak: The difference is that with truncation, you map 256->64, with rounding down. The correct mapping is 255->63, with proper rounding and half-sized upper and lower interval
16:29:04mirakwhy 256 ?
16:29:11linuxstbpreglow: Yes, it seems happy.
16:29:38amiconnIf you put together a spreadsheet, you'll also see that *not* all intermediate intervals contain 4 source values, some will contain 5
16:29:50 Join adiamas [0] (n=adiamas@
16:30:09amiconnmirak: Because that's how truncation works. Just that you'll hever see the topmost value
16:30:36mirakamiconn: you can't have a value of 256
16:30:39mirakwhat do you mean
16:30:47mirakit's between 0 and 255
16:31:06amiconn[16:30:09] <amiconn> mirak: Because that's how truncation works. Just that you'll hever see the topmost value <==
16:31:53mirakwith truncation 0, 1, 2, 3, gives 0 and 251 253 254 255 gives 63 with 63 that should be the same than the 251
16:32:25mirak251 color, though there is some weird thing in 16bit mode in the datasheet of the lcd screen
16:32:29linuxstbpreglow: I don't know if you've tested, but the line scrolling works fine in the file browser - I have a test directory starting with A with a very long name that scrolls nicely.
16:32:41saa[b_r]iderbger: any known bugs with the H300 remote patch?
16:32:47mirakremove "251 color,"
16:33:09preglowlinuxstb: cool, will test
16:33:22mirakamiconn: from what I understand, colors 253 254 and 255 are lost and can't be obtained in 16bit mode
16:33:43mirakwell not exactly but well ..
16:33:56amarsilly question, is there anywhare specidic you have to go to decaler a global variable/setting
16:34:01linuxstbpreglow: I wish you hadn't made the backlight timeout work...
16:34:17preglowlinuxstb: i find the lcd pretty readable even without it
16:35:14linuxstbMine needs a light shining directly onto it, and I haven't got one.
16:35:19mirakamiconn: well nevermind, that's not that much of a problem anyway exept it needs a bit more computation
16:35:25linuxstbI've just changed the default setting in settings.c though.
16:35:36preglowmine is completely readable without light
16:35:56preglowscrolling doesn't work here
16:36:00preglowahh, right
16:36:03preglowit needs to be selected...
16:36:10preglowbetter make it a dir, then
16:36:50preglowcolour lcds where you can't see without backlight seems really annoying...
16:37:15linuxstbI have trouble reading the h140's screen without a light as well. Maybe it's just me.
16:37:23preglowno, that's not just you
16:37:25linuxstbOr my dark house.
16:37:25preglowit sucks
16:37:47preglowworks just dandy
16:37:48mirakpreglow: I can't see anything without backlight on the H300
16:38:00preglowmirak: the h300 has a very sucky lcd in that regard
16:38:19mirakdon't know why it's sucky
16:38:27preglowbut okies
16:38:30preglownow for buttons
16:38:34mirakpreglow: that's the only way to have black black
16:38:35amiconnMost colour LCDs are like that
16:39:48saa[b_r]iderthe H300 LCD shows pretty well without backlight in direct sunlight
16:40:22saa[b_r]iderso if you're outdoors, you could save some battery power if you kept the LCD off...
16:40:57mirakyes it would be better to put the lcd off
16:41:12mirakthat's not what you meant I think
16:41:16saa[b_r]iderI think the reason the H300 LCD doesn't show well without back light because there's a thick piece of glass
16:41:32saa[b_r]iderI ment the back light
16:42:01mirakactually, for me it would be the same to have the screen turned off than having just the backlight off
16:42:14 Join San||Away [0] (
16:42:19mirakI wouldn't have noticed it was on if you didn't told it to me
16:42:52saa[b_r]iderunless you were under sulight :)
16:43:21mirakiriver firmware turn it off isn't it ?
16:43:37saa[b_r]iderI've always wanted the LCD to stay on with the backlight off, thinking that it would be as clear as with color-screen mobile phones
16:45:04mirakI still have a nokia 3310
16:45:40mirakthe battery is dead
16:45:43saa[b_r]iderdid the model numbers go that low :D
16:46:50saa[b_r]iderI have my faithful SE T68i as my backup phone... had to use it a few days ago when i dropped my current phone in a glass of water by accident :|
16:47:35mirakadmit it was a beer
16:47:38mirakjust admit it
16:48:37 Quit adiamas ("Chatzilla 0.9.69 [Firefox 1.5/2005111116]")
16:48:43saa[b_r]iderhahaha.... actually, it was the iron refiller, with water and toothpaste inside it (from when I opened my H340 to insert an inSkin)
16:49:27saa[b_r]iderI bet you could never have guessed ;)
16:49:49preglowlinuxstb: does the ipod control pads differ between generations? i thought they all had the same basic functionalty
16:50:35preglowhah, rockbox.ipod is 300k, rockbox.iriver is 250k
16:50:42preglowwe really should play around with using thumb code
16:51:08linuxstbYes - especially for code in iram.
16:51:25preglowit's usually the other way around
16:51:27preglowespecially for code in ra
16:51:35linuxstbI know - but I mean to save space in iram.
16:51:48preglowwell, we wont use too much iram for code, hopefully
16:52:01preglowthough the ipod code cache is just as bad as motorolas
16:52:27preglowi wonder if we can configure it to never cache data
16:52:37preglowwe'll need to do some tests, obviously
16:52:38mirakwhat exactly is ffmpeg ? Is it an interface or codecs etcetera ?
16:52:43preglowmirak: it's a codec library
16:52:55mirakyou used that for ogg mp3 etcetera ?
16:52:57linuxstbThe ipod control pads are (I think) the same for all the PP5020 models - apart from the mini.
16:53:20linuxstbmirak: No. But we use the ffmpeg and shorten decoders from it.
16:53:22preglowwhat's different for the mini?
16:53:29preglowflac and shorten decoders
16:53:44linuxstbYes, that.
16:54:30linuxstbpreglow: All I know is what's in the kernel's keyboard.c file
16:54:46miraklinuxstb: hem ffmpeg is a codec in itself ?
16:55:10linuxstbmirak: ffmpeg is a high level library for decoding and encoding audio and video. It consists of libavcodec (the codecs themselves) and libavformat (for the container formats)
16:55:11 Join hshah [0] (
16:55:31preglowlinuxstb: i was just thinking about the need for several IPOD_*_PAD defines
16:55:37preglowlinuxstb: when they're all basically the same
16:55:38linuxstbmirak: It does far too much for it to be of general use in Rockbox.
16:55:54linuxstbpreglow: Yes, I've thought about that. An IPOD_4G_PAD is probably enough.
16:56:06preglowlinuxstb: what about IPOD_PAD ?
16:56:36linuxstbThe earlier PP5002-based models are possibly different - I haven't looked at the code for those.
16:56:39preglowthey alle have the same buttons, don't they?
16:56:52linuxstbYes, but at least the drivers will be different.
16:56:57preglowyes they will
16:57:15preglowbut this way all plugins and such don't have to do the full string off checks for IPOD_*_PAD
16:57:33preglowbut okies, we'll see
16:57:36linuxstbThe H100/H300 is the same - they look the same to the application, but have different drivers.
16:58:03preglowhmm, true
16:58:09linuxstbSo it could be worth having something like IPOD_PAD and CONFIG_IPOD_PAD
16:58:27linuxstbOr in fact, just have IPOD_PAD contain different (non-zero) values
16:58:45preglowand btw, we should probably drop all the PP5020 prefixes we use now
16:58:53preglowit's kind of redundant in rockbox
16:59:14preglowat least for the ones in pp5020.h
16:59:49linuxstbYep, that would make sense.
16:59:53preglowshould i stuff opto_i2c_init in i2c-pp5020.c as well?
17:00:27preglowi don't really know what it does, so i have no idea, i just see 'i2c' in its name
17:00:27linuxstbNo, I think that code belongs in button_init()
17:00:54linuxstbBut I guess it doesn't really matter.
17:01:03 Quit Zagor ("Client exiting")
17:01:41 Join muesli- [0] (n=muesli_t@
17:05:23mirakhem what does it mean when it's said that rockbox doesn't do dynamic memory allocation ?
17:05:44linuxstbThere is no malloc()
17:06:03linuxstb(apart from a very restricted implementation used by some of the codecs - but that's an exception)
17:06:17 Quit muesli- ("ich will Kühe!!!")
17:06:55mirakso how can you know what memory you can use ?
17:07:11preglowman, how i hate ifdefs
17:07:20preglowyou take whatever you need
17:07:21preglowthen use it
17:07:26 Join muesli- [0] (n=muesli_t@
17:07:34preglowif you need 256 bytes, you just do char buf[256]; then use that
17:08:27mirakwhat's the difference with the malloc then ?
17:08:34XavierGrQuestion: What #if would you recommend for the file scroller plugin? It is based on plugin memory needs.
17:08:34preglowlinuxstb: btw, you changed the backlight timer, yes? does it seem accurate on longer times?
17:08:44mirak(I am mostly do java actually, so that's a bit far for me now)
17:09:00preglowmirak: the difference is that we don't need to keep a lot of memory free so you can use malloc
17:09:36mirakok, don't know what malloc is doing exactly
17:09:36 Quit edx (Read error: 104 (Connection reset by peer))
17:09:49preglowmirak: malloc finds a large enough space of free memory, then lets you use that
17:10:03preglowmirak: that memory can't be used any place else in rockbox, so we need a large free pool of memory
17:10:23preglowmirak: wasting space like that is bad in rockbox, since we need all the memory we can get for the file buffer
17:10:25linuxstbpreglow: No, changing the default in settings.c doesn't work unless I reset the settings...
17:10:26mirakbut malloc doesn't take what you say him to take only ?
17:10:43mirakok nevermind
17:10:46 Quit hshah ("Leaving")
17:11:05preglowmirak: yes, but since we never know how much memory people are going to need, we have to reserve rather a lot of space
17:11:17linuxstbmirak: malloc() also has a corresponding free() function - it's the free() function that's the main troublemaker.
17:11:47preglowbah, i should get keys working
17:11:47linuxstbSo it's probably more informative to say that Rockbox doesn't have free().
17:12:00 Quit San||Away (Read error: 110 (Connection timed out))
17:12:05preglowwe should write a faq entry on this
17:12:23miraklinuxstb: so how the space allocated is given back ?
17:12:33linuxstbIt never is. That's the point.
17:13:07amiconnmalloc() on a platform like rockbox has some drawbacks: (1) it needs a memory pool for allocation, which can't be used otherwise. We want to use most RAM for buffering, so that would be a waste.
17:13:47***Saving seen data "./dancer.seen"
17:14:08PaulJcan someone explain to me what exactly the crossfadeoptions "fade in delay" and "fade out delay" mean? (this is missing in the wikimanual)
17:14:09amiconn(2) error handling would become more complex, because a malloc() can fail. It would also be more confusing for the user, why a feature works one time, but not the other, even with proper error messages
17:14:29mirakwhen you declare a new variable, a char buf[256]; how can you know it will not give the memory adress of an existing programm ?
17:14:30XavierGrplugin_get_buffer is a nice example of "like" memory allocation in rockbox.
17:14:43XavierGras linuxstb taught me.
17:14:53amiconn(3) On platforms without a MMU, we would also run into memory fragmentation problems with a true malloc() implementation
17:14:58linuxstbmirak: Rockbox is just a single program - the compiler takes care of allocating the static buffers.
17:15:05Slasheriin fact it could be possible to do malloc quite working but then we would need to implement a function like p = refresh(p); that would memmove and relocate the allocated memory
17:15:35mirakok so that's still less confusing than assembler then ^^
17:15:40linuxstbThat sounds like memory handles which need to be de-referenced before accessing. OSes like Symbian use those IIRC
17:15:51*amiconn really likes assembler
17:16:31amiconnmirak: Have a look at memcpy_a.S ;)
17:16:34preglowassembler is nice
17:16:53linuxstbamiconn: I'm sure you would love the ARM...
17:16:57mirakok, so when you exit a plugin, what happens ?
17:17:20preglowamiconn: nice, shiny risc assembler
17:17:34amiconnlinuxstb: I'm not so sure...
17:17:37amiconnI'm biased
17:17:43preglowarm is lovely for assembler
17:17:47preglowthere can be no doubt about it
17:17:54preglowone of the nicer asm variants i have seen
17:18:09linuxstbmirak: There is a fixed block of memory (32KB for Archos, 768KB for iriver) assigned for plugins - all the code and data go in there.
17:19:01 Join hshah [0] (
17:19:03amiconnSlasheri: Hmm, the memory handle idea could actually work - if the overhead isn't too big
17:19:14preglowbtw, can't we decrease the 768kb a bit?
17:19:20preglow512kb sounds like it should be enough
17:19:20amiconnI think so
17:19:25Slasheriamiconn: true, especially if the allocated memory chunks are small
17:19:35linuxstbThat would ruin XavierGr's changes to the JPEG viewer though...
17:19:52preglowdoes that use plugin space?
17:20:02amiconnpreglow: 512KB/32MB == 32KB/2MB, just by coincidence...
17:20:05preglowjpeg has heaps of memory left even then
17:20:22preglowjpeg is only 26kb big, tons and tons of memory left
17:20:42linuxstbThe jpeg viewer currently uses the audio buffer for the actual image buffer - XavierGr is changing it to use the plugin buffer if there is enough room.
17:21:04linuxstbPlus adding a slideshow type feature IIUC.
17:21:15amiconnYou can get quite far with 512KB (minus 26KB for the plugin)
17:21:18preglowbut i think reserving plugin space just because of features like that is a bit... wrong
17:21:44*amiconn would like to get down codec ram size as well
17:21:56preglowi think 512kb sounds nice
17:22:01amiconnBad aac ;)
17:22:09preglowaac needs work
17:22:22linuxstbYes, the codecs also have a huge 512KB malloc buffer which should be reduced or even removed.
17:22:28XavierGrEven if you change the plugin memory to 512 my change still work
17:22:48XavierGrIt is just that it would fit smaller pictures.
17:22:48linuxstbIs it still as useful?
17:23:03amiconnThe biggest plugin is rockboy, and even that needs a little less than 400 KB
17:23:16preglowit would still be useful, but yes
17:23:31preglowi'd probably love the feature, but i still don't think the needs of just one plugin merits keeping the plugin buffer huge
17:23:45XavierGrwhat feature?
17:23:52linuxstbYour feature.
17:23:54preglowyour jpeg plugin extensions
17:24:08XavierGrit will be still usable if set to 512
17:25:00XavierGr29 kb for plugin + 86kb (I think) for grayscale + (number of entries) * 1 byte
17:25:05XavierGrthe reset for the picture
17:26:35mirakso if I declare many times a variable in a block ( a loop for exemple) will it take all the space in ram ?
17:27:09preglowaac.codec is actually smaller for ipod
17:27:24preglowmirak: no
17:27:31preglowmirak: it will just stay one variable
17:28:01 Quit b0br ("CGI:IRC")
17:28:27preglowlinuxstb: what to do about the udelay() calls? use something else or just include that as well?
17:28:36preglowwhat does other driver code use for delays?
17:29:26linuxstbI don't think anything else has delays yet.
17:29:36Slasherimirak: you can think that everything inside { } makes the stack one step deeper and variables inside that block are allocated from the stack. However, at the end of }, one level of stack is freed. So for(..) { .. } would increase and decrease the stack on every loop
17:30:10linuxstbpreglow: There are loops for things like "wait_not_busy" which use current_tick to implement a timeout whilst waiting for a hardware register to change.
17:30:22preglowi need 5 musecs of delay, hehe
17:30:52preglowthin i'll just stuff udelay in for now
17:30:56mirakSlasheri: ok
17:31:43linuxstbpreglow: Yes, put udelay somewhere central - maybe an inline function in system.h
17:32:59preglowi wont yield inside it
17:33:07preglowso it's just for driver use
17:36:31 Quit Febs ()
17:39:13 Join Moos [0] (
17:39:23Lynx_so, does the iaudio x5 play all xvid files, or do they have to be of a special size and such?
17:40:22mirakdamn i am reading xvid decoder
17:40:23preglowyou think it'll full size 640x480 xvids at 25fps? :P
17:40:45preglowinsert 'play'
17:40:54mirakthe H300 ?
17:41:17mirak220*174 25fps would be fine yes
17:41:34mirakthough after seeing the cube running, I am wondering if the screen doesn't have to much latency
17:41:41linuxstbmirak: Which xvid decoder are you looking at?
17:41:45mirakto use above 10 fps
17:41:53XavierGrmirak: no kidding. you have an xvid video in rockbox playing?
17:42:04miraklinuxstb: I apt-get source xvidcore
17:42:10mirakXavierGr: rofl
17:42:13linuxstbDoes it use floats?
17:42:21mirakit seems it doesn't
17:42:26mirakI have seen only int
17:42:55mirakfor interpolation I don't know yet
17:43:03mirakit's in other files I haven't checked
17:43:10linuxstbDo you know what license it is released under?
17:43:28mirakthat's on debian so it should be ok
17:43:43 Join lush [0] (
17:43:47lushhola guys
17:43:53mirakThis program is free software ; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by
17:44:29mirakthere is assembly optimisations
17:44:35miraknot for coldfire ;)
17:44:40 Quit lush (Client Quit)
17:44:55linuxstbWhat about ARM?
17:46:54mirakanyway I don't see why they would use float, most of the time everything is converted to discret stuffs, even for fourier computations
17:47:02linuxstbGet porting then :)
17:47:22mirakthat's my main motivation
17:47:23mirakbut well
17:47:28mirakyou know :)
17:48:09mirakI am interested in knowing how mpeg compression works.
17:48:25mirakthat's not from that code that I will learn that I think
17:48:37preglowi wouldn't want to learn that from the code
17:48:41preglowmpeg4 isn't that complex
17:48:50preglowstill builds on good old mpeg1/2 methods
17:48:59mirakI don't know them
17:49:12preglowa good first step is finding out how jpeg works
17:49:17mirakI have studied some things related to that in univ though
17:49:21preglowthen after that learning about motion estimation
17:49:29preglowand entropy coding
17:49:44mirakthat's what I learned. motion estimation for tracking etecetera
17:49:56mirakc'etait imbaisable
17:50:01miraklike we say here
17:50:15mirak(traduction : it was unfuckable)
17:50:25preglowwould most definitely be hard
17:51:00mirakI got 10 at this exam. it was about fourier transformation and stuffs
17:51:15mirakmy god
17:51:47preglowten being how good?
17:51:55 Quit hshah ("Leaving")
17:53:24miraknot less not more
17:54:15mirakit could have been worse, I didn't do pure mathematics for 2 years before doing that.
17:58:29mirakpreglow: ! ok I have studied that but we didn't seen how to use that for compression
17:58:37preglowuse what?
17:58:50mirakhem discret fourier transformation
17:59:11preglowit's just a matter of removing coefficients that aren't imptant
17:59:21preglowand coding the result
17:59:24mirakright !
17:59:29mirakI remember
17:59:35mirakwe cut the frequencies
18:00:00mirakhowever for me it would have been better to study that for sound
18:00:20mirakbecause in 2 dimension that's not very clear what happens
18:00:47miraka spectral representation of a sound wave is interpretable
18:01:09mirakand understandable, while for an image that's really not the case
18:01:28mirakso you studied that ?
18:02:10preglowstudied and studied
18:02:14preglowit's more a hobby thing
18:03:12miraklol, fourier transform as a hobbi
18:03:14mirakwhy not :)
18:04:36preglowa part of it
18:04:43preglowfourier transform isn't exactly my main area of expertise
18:07:27mirakwhat's your dada ?
18:08:16mirakyour passion
18:08:28mirakyour main domain of expertise
18:09:21preglowgod known
18:09:25preglowsound synthesis, i'd guess
18:09:31preglowwhich includes quite a few areas again
18:09:48preglowmore or less anything that has to do with sound
18:09:57mirakyou did the vocal stuff for archos ?
18:10:07miraklike fart sound ?
18:10:07preglowthat was [idc]dragon
18:10:26miraksorry ... :)
18:13:30mirakI had a module of vocal recognition
18:13:43mirakI got 17/20 but it was a bit sucky
18:17:50 Join lush [0] (
18:18:04lushanybody home/
18:18:27lushhey just started using the rockbox fw on my iriver h120 and i frekking love it
18:18:32lushbut it wont play back my wma files
18:18:37 Join NicoFR [0] (
18:18:44lushit just skips em, and then says codec error
18:18:56lushany idea as to why?
18:19:00muesli-it wont in near future..mayb sometime
18:19:32linuxstbRockbox doesn't have a WMA decoder.
18:19:45linuxstbSo it can't play wma decoders until someone decides to write one.
18:20:00linuxstbI mean it can't play wma files...
18:20:06lushok i got ya
18:20:30lushhey one other question,
18:20:40lushhow do you add plugins to the player?
18:20:58lushdo i dl em and then just, copy them over to the rockbox dir?
18:21:29 Quit XavierGr (Read error: 110 (Connection timed out))
18:21:41muesli-copy them where you want
18:21:47muesli-it doesnt matter
18:22:02lushsame for those wps things that are the skins?
18:22:15muesli-see .rockbox/wps
18:22:27lushthanks for the help
18:22:34muesli-no worries
18:22:49lushthis stuff is soo much better than default crap
18:23:07muesli-100% agree ;)
18:23:08 Quit lush ("CGI:IRC (EOF)")
18:23:23 Join webguest34 [0] (
18:23:36 Quit webguest34 (Client Quit)
18:23:51 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
18:29:19mirakexept for video ...
18:29:37muesli-time will tell ;)
18:29:41 Join DangerousDan [0] (
18:31:41preglowman, button drivers are not my favourite
18:34:48 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
18:37:12preglowlinuxstb: are there any ipod special cases hanging around that you know about?
18:39:12linuxstbNo, I think the current_tick was the last one.
18:39:29linuxstbWhat's the problem?
18:39:36preglowjust wondering
18:39:43preglowi've implemented a small keyboard driver now
18:39:45preglowand it hangs rockbox
18:40:54linuxstbDoes it crash immediately, or when you press a button?
18:41:18preglowhangs at logo
18:41:19 Join DrMoos [0] (
18:41:35preglowgot it
18:41:40preglowwas a dead giveaway
18:41:43preglowi didn't ack the interrupt
18:42:10preglownow it hangs when i press a button :-)
18:42:41linuxstbThat's progress...
18:42:55 Quit Moos (Read error: 104 (Connection reset by peer))
18:44:11 Nick DrMoos is now known as Moos (
18:44:14 Join Mmmm [0] (
18:44:27preglowhmm, no
18:44:32preglowit seems i'm still not acking the interrupt
18:44:41preglowi must have touched the clickwheel the first time
18:44:44preglowit behaves just the same
18:46:12amarwhare its the global_settings structure diffined?
18:46:26 Join jlo [0] (
18:48:11 Quit muesli- ("ich will Kühe!!!")
18:50:15jlopreglow : I did some test with Reaktor (I used soft this weekend because I have only harware at the office !), simulating crossfeed, now can you and markun tell me which music you'd like to test so I can do some recordings with/without
18:53:39*preglow kicks markun
18:53:42 Join San [0] (
18:55:17jloI'll come back later
18:55:20 Quit jlo ()
18:57:25 Quit Rick (Read error: 104 (Connection reset by peer))
19:00:23 Join Mmmm_ [0] (
19:00:38mirakpreglow: a keyboard driver on usb host ?
19:00:38 Quit Mmmm ("CGI:IRC (EOF)")
19:00:59 Quit NicoFR ()
19:02:03preglowmirak: eh?
19:02:17 Join akaidiot [0] (
19:02:44 Join Mmmm [0] (
19:04:10preglowi would kill for a data sheet
19:04:27preglowsomeone try that :-)
19:07:01mirak<preglow> i've implemented a small keyboard driver now
19:07:03 Join XavierGr [0] (
19:07:32 Join ender1 [0] (i=ychat@
19:07:47mirakrockbox is better than ipod linux ?
19:07:55XavierGrwhich is faster: memmove or memcpy?
19:08:20preglowmirak: i'm talking about ipod
19:08:27preglowmirak: depends what you want
19:08:32preglowXavierGr: memcpy for now
19:08:41preglowif you don't need memmoves functionality, use memcpy
19:08:50XavierGrokay then it should be my preffered choice.
19:09:11 Join edx [0] (
19:09:23XavierGralso on the jpeg viewer I saw that: #define MEMCPY(d,s,c) rb->memcpy(d,s,c)
19:09:41XavierGris that an optimization, should I use it that way or call with the rb api pointer?
19:10:02XavierGror it is just an alias?
19:10:07mirakpreglow: don't probably ipodlinux is less cryptic
19:10:17miraksince we are maybe more used to linux
19:10:19mirakdon't know
19:10:46mirakbut well if rockbox is faster anyway
19:11:00preglownothing says it is
19:11:08preglowrockbox and linux are two different approaches
19:11:11preglowit depends what you want
19:11:18mirakRockbox's version is a lot smaller and faster, while IpodLinux's is bigger and slower but very well known, very well tested and offers a multitude of more services and features.
19:12:57mirakwell actually for me the main difficulty with rockbox is that it's not clear what are syscalls usable from user programs, what is not etecetera
19:13:26mirakbut that just me, and I am starting
19:13:50***Saving seen data "./dancer.seen"
19:15:04 Quit ender` (Read error: 110 (Connection timed out))
19:16:38 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
19:17:08 Join Rick [0] (
19:17:25preglowmost things are usuable from user programs in rockbox?
19:20:02preglowit no longer hangs!
19:20:31preglowi can activate backlight by keypress
19:20:36 Quit akaidiota (Read error: 110 (Connection timed out))
19:20:59amiconnXavierGr: (1) We don't have memmove() in rockbox yet. (2) The MEMCPY() macro is just for portability
19:21:47 Nick San is now known as San||Halo (
19:22:00 Quit Mmmm_ ("CGI:IRC (EOF)")
19:22:06 Quit Mmmm ("CGI:IRC (EOF)")
19:22:41 Part Polo_o
19:22:55 Join Mmmm [0] (
19:26:04preglowlinuxstb: you there?
19:27:20 Join petur [0] (
19:28:18preglowlinuxstb: if you are, check out, please
19:28:27preglowseems the hold switch functionality is done in hardware on the ipod
19:29:45 Join yngwi [0] (
19:34:14XavierGrf*cking iriver fw.
19:35:00XavierGrI try to debug, when my unti hungs and reset it then if I press play iriver fw starts instead of Rockbox, why?
19:35:21XavierGrIt takes ages to reboot when iriver fw starts... grrrr
19:35:24amiconnpreglow: detectable?
19:36:14preglowamiconn: yes
19:36:20preglowamiconn: that is, i think so
19:36:31preglowamiconn: it's a bit strange, switching hold off seems to generate an interrupt
19:36:36preglowamiconn: but switching on doesn't seem to
19:37:04preglowi was wrong
19:37:05preglowboth do
19:38:24preglowipl installs two keyboard related interrupts, i only use one, and i seem to get all the interrupts i need
19:38:29mirakpreglow: that's what I don't know, I guess documentation will come with time
19:41:21 Join jlo [0] (
19:42:22 Join webguest34 [0] (
19:42:48 Join xmixahlx [0] (n=xmixahlx@
19:44:24preglowjlo: so, have you modelled your new crossfeed proposal yet?
19:45:29 Quit edx (Read error: 110 (Connection timed out))
19:45:50jlopreglow : yes, if you have reaktor, I can send you the modelling file
19:47:04 Join _FireFly_ [0] (
19:47:11jloI'll put wav files with/without correction on ftp
19:48:35preglowhaven't got reaktor here, i'm afraid
19:51:07jlopreglow : so I put the schematics (in png) on in the reaktor directory, have a look
19:53:13 Quit webguest34 ("CGI:IRC")
19:53:37preglowthose filters are second order
19:53:42preglowi wish you wouldn't use them ;)
19:54:41jlopreglow : the pb for you is that the filters are shelving , no?
19:55:13preglowwell, yeah, but i was hoping we'd be using first order filters
19:55:24preglowand using second order filters instead of first order might give different results
19:55:36preglowi think reaktor has got first order shelving filters as well
19:55:39preglowin the core modules somewhere
19:55:47preglowstatic eq filters or something
19:55:52jlopreglow : it's first order (6dB/oct) but shelving
19:55:58preglowoh, they are?
19:56:01preglowthen it's fine
19:56:07 Quit Hooligan ()
19:59:47jlopreglow : I first tested with correlated pink noise to check that no major tonal unbalance occurs, and after with different music tracks, but now I'd like to check also with your files
20:02:50preglowgimme a sec and i'll se if i've got what markun gave me
20:02:51 Join edx [0] (
20:04:27 Join _DangerousDan [0] (
20:05:10jlopreglow : see, he's giving the source of his winamp crossfeed plugin, have you allready looked that one ?
20:06:04 Quit ender1 (Read error: 104 (Connection reset by peer))
20:07:14preglowjlo: no i havenĉt
20:07:15 Join JazzBone [0] (
20:07:27 Join Lear [0] (
20:07:52 Join zeero [0] (
20:08:13XavierGrOK The new JPEG plugin viewer is ready. Any testers? (Especially arhos users) I will upload it to the patch tracker soon.
20:08:23 Join ender` [0] (n=ender@
20:09:13XavierGrI manged to do quite well to my tests. 1074 files 75 characters each. (path doesn't matter, only filename)
20:09:25zeeroso, maybe i'm a idiot to not be able to work this out, but i'd like to start helping out with the h300 effort now that i can get it on my player, but in building the sim, when i run ./rockboxui i get something that looks not at all like what i see on my player. also, i can't work out what the keys for a-b and so on are, only up/down/left/right seem to work. am i missing something obvious?
20:10:04XavierGrMaybe you biuld a simulator for another target.
20:10:22XavierGrDid you choose your model in coniguration?
20:10:26zeeroXavierGr: i selected the h300 in config
20:10:51XavierGrwhat do you use: cygwin, devkit, linux?
20:14:38XavierGrI can't help you. I use the devkit
20:15:13peturzeero: do you use w32 (wine) or x11?
20:18:33jlopreglow : I'll put on the web three anonymous files (original, rb crossfeed, my version, winamp plugin) so you can try with abc/hr soft ( the one you prefer. if you and others can use it, I can do it for other tracks
20:19:48 Quit tvelocity ("Leaving")
20:19:58jlopreglow : I think In
20:20:00peturSlasheri, are you around?
20:20:12jloI think I need an hour or so
20:20:14 Quit DangerousDan (Read error: 110 (Connection timed out))
20:20:21preglowjlo: that's ok, i'll still be here then
20:22:00Slasheripetur: yes, hi
20:22:19peturgot a quick Q about the H1xx backlight dimming
20:22:34Slasherihmm, original code yes
20:23:01peturis that dimming through a timer routine switching on/off?
20:24:04Slasheriyes, with cpu timer interrupt. So we have a fixed cycle time and then the on and off pulse times are being modified
20:24:31petursoftware PWM then... the H1xx has only on/oof, no PWM?
20:24:44peturzeero: x11 sim seems broken
20:24:46Slasheriyep, pure software pwm :)
20:24:53peturok thanks.
20:25:06peturwanted to be sure I understood what I saw
20:25:06Slasheriit might have pwm on some outputs, but not on that backlight output at least
20:25:11zeeropetur: hrmm, does the wine one work?
20:25:30zeeropetur: k, thanks, i'll try building that
20:25:37linuxstbpreglow: I've just tried your build - I can turn the backlight back on...
20:25:51peturcan't help with wine, using W2K here ;)
20:26:02petur(that's windows2000)
20:26:12peturbut it should work
20:26:24zeeropetur: nod, dang, so i need a cross compiler to do that though huh?
20:26:46linuxstbYes - which linux distro are you using?
20:27:03Slasherilinuxstb: btw, now the tag cache has a stable building routine that should most definately work on archos too and simple artist listing. Seems to work pretty well :) But still many things left to be done..
20:27:04zeerolinuxstb: redhat el4
20:27:21zeerolinuxstb: i have a debian box i can use too, but i'd prefer to do it on this one
20:27:47linuxstbzeero: On Debian, you can just do "apt-get install mingw32" to install the Windows cross-compiler. I've no idea about RH.
20:28:04peturLinusN: have you tried different backlight PWM values on the pcf50606? Is there a safe range or do they all work?
20:28:11linuxstbSlasheri: Nice. Are you far away from having something others can try?
20:28:37LinusNpetur: most values flicker
20:28:39 Quit jlo (Read error: 104 (Connection reset by peer))
20:28:44Slasherilinuxstb: not very far, but still that code is not very usable. Maybe in one or two weeks :)
20:28:55zeerolinuxstb: k, thanks, i'll poke around a bit and see if there's an rpm somewhere (how i wish apt worked as well for redhat as it did for debian)
20:28:55peturI wonder how iRiver does it then
20:29:24peturno harm to try them all? I mean: over-volting the backlight or something?
20:30:20preglowlinuxstb: backlight turns on when you press something, yes=
20:30:25Slasherilinuxstb: at least the api to the tag cache is now quite simple. For example listing all artists: struct tagcache_search tcs; tcs.type = tag_artist; tagcache_search(&tcs); while (tagcache_get_next(&tcs)) { do_something(tcs.result, tcs.result_len) } tagcache_search_finish();
20:30:32peturLinusN: did you play with the frequency as well?
20:30:39linuxstbpreglow: Yes. For the hold switch as well.
20:30:44preglowlinuxstb: excellent
20:31:01peturreason: want to have a go at configurable backlight
20:31:07linuxstbSo it looks like the hold switch is hardware?
20:31:12preglowlinuxstb: indeed
20:31:20preglowlinuxstb: all i do currently is register and handle the interrupt
20:31:32preglowlinuxstb: no other handling is done by me, and the keys are ignored when hold is on
20:31:45LinusNpetur: yes
20:32:08peturis 512Hz what iRiver uses or your best pick?
20:32:18LinusNpetur: feel free to experiment
20:32:18linuxstbpreglow: Have you looked at this code?
20:32:26preglowlinuxstb: somewhat
20:32:29LinusNpetur: that's whet the original uses
20:34:18peturLinusN: you didn't find out how they set the different levels...
20:34:54XavierGrok here is the patch for the viewer. Please test or give me your opinion. Thanks.
20:35:08*preglow looks forward to getting multicore support going......
20:35:12LinusNpetur: no
20:35:22peturok, thanks
20:35:23preglowi'm going to need divine intervention or a lot of help
20:35:58Slasherilinuxstb: i think that in future even an integrated tag editor to the tag cache browser would be really nice :)
20:36:06Cassandrapreglow, scary stuff.
20:36:19preglowCassandra: got a bootloader stuffed in your ipod yet, btw?
20:36:45CassandraYep. For some reason it won't boot the original firmware when the Rockbox loader is on.
20:36:57preglowmine can't boot it any longer either
20:37:05Cassandra(Oh, and as I minor point, I think we should remove the penguin.)
20:37:26preglowdoesn't seem much point in keeping it anymore, no, unless we want it as some absurd "thank you" to the ipl people
20:37:51 Quit Mmmm ()
20:37:55CassandraWell, I think that it says "Rockbox/iPod Linux bootloader is enough of a nod.
20:38:06amiconnpreglow: I think multicore support is the most interesting part of the ipod port...
20:38:49zeeropetur: so, any idea if/when the x11 version will be fixed, and what's wrong with it?
20:38:52preglowamiconn: it'll need some thinking
20:39:07peturzeero: no idea, never used it...
20:39:13preglowamiconn: i think we will want to dedicate the second core to codecs, but again, there will be nuances that will need resolving
20:39:20preglowamiconn: i think you should get one to help me along :)
20:39:24CassandraMultiprocessor Rockbox. Scary.
20:39:24zeeropetur: okay, thanks
20:39:34peturyou could always fix it ;)
20:39:42CassandraI suspect Rockboy is going to need both processors.
20:39:50CassandraBut we can always not bother with that.
20:39:53amiconnNah, no ipod for me, that's for sure
20:40:35preglowthink of all the glorious asm!
20:40:40CassandraWhy not amiconn?
20:41:11preglowleaving the n00bs to it is irresponsible :P
20:41:19mirakpreglow: I shoudln't have erased Matlab
20:41:29mirak(to test compression)
20:41:34preglowmatlab's nice
20:41:52mirakyes to have a math approach
20:42:00mirakmatrix computing is easy and handy
20:42:01Moosamiconn: apple allergy like here? :D
20:42:29*Cassandra thinks the Nano is very pretty.
20:42:33peturLinusN: sorry to irq again: do you think there would be a problem setting the prescaler to something like 7kHz?
20:42:40preglowthe nano is the prettiest one
20:42:42Cassandra(Although obviously way too small.)
20:42:43preglowthe rest look worse
20:42:46amiconnI don't like the design, and many details distract me as well the docking connector instead of a standard USB socket, and the worst thing - the touch wheel
20:43:31CassandraI quite like the touch wheel.
20:43:34preglowme too
20:45:35LinusNpetur: try it
20:45:38LinusNi gotta go
20:45:42 Part LinusN
20:45:43amiconnThe only pro of the Nano is that it's really thin, but that's about it
20:46:35CassandraThe Nano gives me tech lust.
20:46:46CassandraApple make a 60gb version and I'm so there. ;)
20:47:46preglowfour gig is enough for anyone!
20:48:09preglowCassandra: how many gigs is yours?
20:48:18*amiconn prefers the expandable storage of the Ondio
20:48:27preglowmine too
20:48:32preglowi can't be bothered with memory cards
20:48:43Moosamiconn: me too, it's why I bought one
20:48:53 Join webguest25 [0] (
20:49:13CassandraI never know what kind of mood I'm going to be in. I like to have all my music with me.
20:49:23preglowgood point
20:49:41preglowbut i plan on using my h120 for my main player anyway
20:49:47preglowand the nano more for small trips and the like
20:49:50zeeroso, what's the right way to create the .rockbox dir for the sim? (make zip and then unzip in the archos dir?)
20:49:52preglowwhere the size is nice
20:49:56CassandraAlso random shuffle is better with your whole collection. Means you listen to stuff you've forgotten about.
20:50:06*Moos waiting for 80 GO Toshiba HD
20:50:07linuxstbzerro: "make install" - which I think is just a zip and unzip
20:50:14linuxstb(sorry, zeero)
20:50:15amiconnMoos: One day I'll write a file management plugin if no-one does it before. Then there are many new possibilities, like updating rockbox from a memory card, without the computer around...
20:50:26amiconn...or copying files from one Ondio to another
20:50:38zeerolinuxstb: so that puts everything in the archos dir?
20:50:49CassandraWhat I want is something that will store about 500 albums lossless.
20:50:50zeerocool, thanks
20:50:52Moosamiconn: that sounds good
20:51:01Cassandra(I only own 250 or so, but room to expand is good.)
20:51:05amiconnAn unzip plugin would do for a start
20:51:21zeeroahh, the wine version seems to work quite nicely :)
20:51:37preglowsomeone start porting sims to sdl!
20:52:00peturdoes wine support sound as well?
20:52:17Moosamiconn: be sure anyone will work on it until you, not a lot of ondio dev :(
20:52:22zeeropetur: checking :)
20:53:07zeeropetur: hrmm, seems like no :( i get a unhandled div by zero
20:53:14Moosamiconn: is you and [IDC]Dragon, ported Rockbox on Ondios, right?
20:53:14peturthe current w32 sim isn't OK, I wrote a patch (1375787) that fixes this
20:53:20amiconnpetur: It does, at least for me
20:53:24linuxstbzeero: I think you have to manually enable it by setting ROCKBOX_HAS_SIMSOUND in the autoconf.h
20:53:28amiconnMoos: yes
20:53:29peturAnd that is fixed as well in another patch
20:53:43zeerolinuxstb: ahh, okay, let's try that
20:54:12amarcan anyone tell me where is the global_settings structure is deffined?
20:54:38zeerolinuxstb: do i define it as 1?
20:54:39amaror is that a very stupid question
20:54:53linuxstbzeero: Yes, define it as anything.
20:55:00zeerookay, cool
20:55:05linuxstbIt's just checked using #ifdef
20:55:08amiconnMoos: [IDC]Dragon figured out most hardware tidbits and added FAT16 support, I wrote the MMC driver and helped debugging FAT16
20:55:12zeeroahh, nod
20:55:43linuxstbamar: apps/settings.h
20:55:51peturamiconn: the cvs implementation either doesn't play or gives 100% cpu load (Sleep(0) or Sleep(1))
20:55:51XavierGrdon't try to delete a folder with 2000+ files using rockbox.....
20:55:53amiconnMany small contributions from other authors of course, like the soft button definitions etc
20:56:52amiconnOh, and since it was me who bought the Ondio SP, I had to adjust rockbox to the differences between MAS3587 and MAS3539
20:57:06amiconnNow I have both an FM and the SP ...
20:57:07zeeropetur/linuxstd: sound works fine with that def :)
20:57:50preglowsell one and get a nano!
20:58:15HClhows rockbox with the nano?
20:58:23preglowHCl: it responds to keypresses now...
20:59:51 Join mojo [0] (n=mojoski@
21:00:14mojohey everyone.. Anyone alive?
21:00:15amiconnpreglow: The nano flash access is boooring, as it just uses the standard ata driver ;)
21:00:28mojooops, sorry
21:01:09preglowamiconn: lucky for me, i don't think i enjoy that kind of programming
21:01:23amiconnOh, forgot that [IDC]Dragon added philips tuner support, which was then helpful on iriver too
21:01:42preglowmy, i'd love some datasheets about now
21:02:20Moosamiconn: is there one hardware limitation of MMC capacity? I mean in the future is there one size limit max?
21:03:24Moos4go, no?
21:03:46amiconnThe current MMC protocol uses byte addressing with a 32 bit address
21:03:58 Quit solexx (Read error: 104 (Connection reset by peer))
21:04:25Moos10 little cards equivalent of my iriver :)
21:04:33amiconnAfaik, even MMC4.0 doesn't raise the limit, it only defines faster transfer modes (with additional contacts)
21:04:38 Join solexx__ [0] (
21:05:18Moosthanks for he infos
21:05:31*Moos going to do some food
21:06:06amiconnIn case some future standard allows more than 4GB, and this standard would still support SPI mode, rockbox could be adapted to support it
21:06:20MoosCool !
21:06:47amiconnThe USB->MMC brigde wouldn't understand it, but that's not a big problem on the Ondio. You would just need a card reader to fill the card
21:07:10 Quit San||Halo (Read error: 110 (Connection timed out))
21:07:36_FireFly_amiconn: about unzip i have some functions which can read uncompressed zip-file
21:08:06linuxstbWe definitely need to speed up wps loading somehow.
21:08:19_FireFly_linuxstb: ??
21:08:47_FireFly_amiconn: with crc
21:08:50preglowi thought bmp loading was what took time
21:08:53linuxstbLoading the 500 tiny bitmaps...
21:08:58linuxstbThat's what I mean.
21:09:03_FireFly_linuxstb: ah yes
21:09:15_FireFly_mine combined-bitmap support could help
21:09:17_FireFly_a bit
21:09:32_FireFly_but it isn't very userfriendly :)
21:09:59_FireFly_because they need to merge the images together yourself
21:10:26preglowi think that approach is definitely best
21:10:34SlasheriHmm, just combine the bitmaps runtime and create a fast binary version of the current wps?
21:10:49preglowcombine bitmaps runtime?
21:10:52preglowthen you need to load them anyway
21:10:58Slasheriwhen loading the wps first time
21:10:59linuxstbThe build script could combine the bitmaps into a single file.
21:11:07Slasherithat's true
21:11:29_FireFly_linuxstb: but how to discover the position of the images in the image
21:11:43linuxstbThe build script knows them (it put them there)
21:11:53linuxstbSo the build script would have to modify the wps.
21:12:06amiconnI don't see a problem with all gfx in one image
21:12:08linuxstbMaybe not a 5 minute job though...
21:12:37linuxstbamiconn: I agree, but it's more awkward for users to produce them like that.
21:12:46amiconnIn fact, I think editing wps graphics is easier this way, instead of dealing with dozens of tiny bitmaps
21:13:51***Saving seen data "./dancer.seen"
21:14:37_FireFly_my wps for example has fit all used images in an the image which was the background-image
21:16:07_FireFly_the size is 160x25 pixel
21:16:50preglowlinuxstb: i wonder how to handle repeat with an interrupt based driver
21:16:57preglowlinuxstb: i hope the hardware deals with it...
21:17:14 Quit perplexity (Read error: 113 (No route to host))
21:17:43preglowmight not be that hard anyway, of course, we'll see
21:19:11amiconnDepending on whether you get an interrupt both for press and release, it might be simple
21:20:13 Quit joshn_454 ("KVIrc 3.2.0 'Realia'")
21:20:20preglowseems like i only get interrupts for press
21:20:32preglowlemme check
21:20:54preglowfor release as well
21:21:10preglowso i could just schedule a tick task to deal with it for me while a key is pressed
21:22:01amiconnThe button handling wouldn't be all that different from the other platforms, just the button_status variable won't be updated by polling, but by the button interrupt(s) instead
21:22:21amiconnThe button tick could work the same way, just without the button_read() call
21:22:35_FireFly_amiconn: the patch for combined bmp support is already on tracker but i must update it do the cvs-change which my wps-image-flicker fix had produced :)
21:25:11_FireFly_amiconn: do you want my fns for reading uncompressed zip-files ??
21:25:51amiconnNo time for zip atm :/
21:26:52_FireFly_hmm ok the only thing which would be hard to implement due the small amount on resources is an uncompression algo
21:27:17preglowi don't believe that'd take much time
21:27:23preglowdepends on the files we're talking about
21:27:34preglowwell, yeah
21:27:36preglowbut contents
21:27:40pregloware we talking about zipping wpses here?
21:27:56preglowthat would be a pretty nice application, actually
21:27:59preglowhaving wps contained in one file
21:28:20amiconnI would like to be able to unzip a rockbox build on target
21:28:23_FireFly_i have already an function set which work with uncompressed zip-files
21:28:38preglowthere is an unzip library hanging around
21:28:39amiconnNothing that needs to be especially fast
21:28:42preglowwhich uses zlib
21:28:47preglowshould be very portable
21:29:02amiconnIt should just fit in the archos plugin ram
21:29:05_FireFly_preglow: afaik malloc might be needed
21:29:10preglow_FireFly_: good point
21:29:29_FireFly_because you need an bigger buffer for the uncompressed data
21:29:34_FireFly_as the compressed one
21:29:46amiconnI think it should work as a viewer for .zip files, asking for the destination dir to unzip it to
21:29:56_FireFly_and my short overlook on zlib it uses malloc
21:30:22_FireFly_so only the file-list could be in ram
21:31:01_FireFly_and then a spezified amount of compressed data is read from file ,uncompressed and written to the destionation file
21:31:10preglowamiconn: so, use of mp3 buffer is banished?
21:31:27_FireFly_my fn's it reads into an array all files in the zip-file whith the seek-points to the start of the data
21:31:28amiconnpreglow: No, it could be used for data
21:31:34linuxstbWhat's "human68k" ? (it's a target for unzip)
21:31:34preglowthen you'll be alright
21:31:48mirak_FireFly_: what do you do to convert the malloc ?
21:31:54_FireFly_only an uncompressen algo is missing
21:32:07_FireFly_mirak: i*m doing nothing
21:32:29preglowlinuxstb: i think it's the same as x68000, and old kind of computer from asian country
21:32:29mirak_FireFly_: ok so if an app uses malloc, how can I adapt it ?
21:32:42preglow68k based
21:32:50mirakatari and amiga
21:33:12preglowyes, but this was something different
21:33:21mirakmy god there is was only 1 mega of ram
21:33:26preglowi can't quite remember, it's been yonks since i've even heard the name mentioned
21:33:31 Quit webguest25 ("CGI:IRC")
21:33:48 Join Mmmm [0] (
21:34:41linuxstbunzip has m68k assembler optimisations courtesy of that machine.
21:34:59preglowlinuxstb: too bad we can't use them
21:35:06linuxstbNo use at all?
21:35:08preglowcoldfire is a subset of m68k
21:35:12preglowdepends on how it's written
21:35:58 Quit xmixahlx ("blah blah blah")
21:36:52_FireFly_if someone have interrest about my read-zip-fns (with crc-check) then the sources can be found here:
21:37:21linuxstbThe unzip source code has even more #ifdefs than Rockbox
21:37:22_FireFly_it was written to work in rockbox -> only functions used, which are present in rockbox
21:37:30_FireFly_linuxstb: yeag
21:37:34_FireFly_-g -h
21:37:57_FireFly_but i havn't yet written an plugin or such
21:38:32 Part Mmmm
21:39:20 Join Mmmm [0] (
21:40:17linuxstbmirak: You have two options for porting an app that uses malloc - either replace all the mallocs by static buffers, or implement your own malloc that allocates from a large static buffer. Either way, remember that you don't have unlimited RAM available - unlike a PC.
21:40:59_FireFly_RAM on PC isn't unlimited either but much more then on the targets
21:41:23zepeople oughtta use ram on a PC like they're forced to on small devices :p
21:41:57preglowgod, i'm tired of bloat
21:42:01 Part mojo
21:42:20_FireFly_i think you should only use as much as ram as really needed
21:43:08preglownow, how to set last_btn properly for an interrupt driven button driver?
21:43:17amiconn'small' devices are dependent on the point of view. The ZX81 had 1KB of RAM - including the framebuffer (!)
21:43:33linuxstbAnd I remember a chess game being available for it...
21:44:17preglowyou probably couldn't castle in it, due to memory constraints
21:46:46amiconnChess was available for the ZX Spectrum, which had either 16KB or 48KB of RAM
21:46:56amiconnThe even was VoiceChess...
21:48:08linuxstbRockChess - where is it?
21:48:16MmmmHas anyone noticed an increased ticking noise with the remote possibly since the flicker improvements?
21:48:56preglowlinuxstb: ipl has chess...
21:49:20preglowit's called (surprise) tuxchess
21:49:28linuxstbNot iChess?
21:50:04preglowhow do they choose?
21:50:15preglowmust be half of the development process alone
21:50:52 Quit JazzBone ("Leaving")
21:51:26linuxstbMmm. Seems tuxchess is not GPL - but the author's own "do whatever, as long as you don't make money from it" kind of license.
21:51:57preglowlinuxstb: btw, slowcoder says the new button driver isn't perfect either
21:52:29preglowhow the hell do they survive all these numerical addresses?
21:53:38 Join ender1 [0] (i=ychat@
21:54:40linuxstbAny ideas about how games like Sudoku and Bejewelled could work with the ipod's scrollwheel?
21:54:42miraklinuxstb: ok. I have read that . I am not sure about quantization
21:54:55mirakI don't understand the equation
21:55:15preglowlinuxstb: plain horizontal scroll with wraparound?
21:55:21preglowgod known
21:55:27preglowi think i'd like to use button action for bejeweled
21:58:05*preglow gets bitten by the strange musepack buzzing
21:58:12linuxstbYep, I was thinking of just horizontal scroll with wraparound. It's going to take a lot of work to get all the plugins working.
21:58:25preglowwell, perhaps
21:58:31preglowwith the scroll wheel, yes
21:58:40preglowbut we'll have ordinary up/down/left/right as well
22:00:55 Join [1]ender [0] (i=ychat@
22:02:06 Join qwisp11 [0] (
22:02:48preglowlinuxstb: so, are we to call it the MENU or UP button?
22:03:23linuxstbI'm happy with what they are called now - MENU and PLAY
22:03:27*preglow has a sneaking feeling we'll be struggling with the amount of buttons...
22:03:45linuxstbIt's no worse than the ondio I think.
22:04:08preglowlinuxstb: so you're thinking that the only way of generating UP and DOWN would be with the click wheel?
22:04:36 Join webguest02 [0] (
22:04:42linuxstbI'm thinking we don't have UP and DOWN - we have SCROLL_FWD and SCROLL_BACK
22:04:56linuxstbWhich is left/right or up/down depending on the context
22:05:17CassandraMakes life interesting, doesn't it?
22:05:23 Quit ender` (Read error: 110 (Connection timed out))
22:06:01 Join muesli_- [0] (
22:06:07qwisp11Could anyone tell me the extension rockboy reads? or .gbc
22:06:21qwisp11ok thanks
22:06:24 Quit webguest02 (Client Quit)
22:06:34muesli_-what does rockbox.iriver btw?
22:06:49 Join webguest86 [0] (
22:06:57_FireFly_it's the firmware file for the iriver ;)
22:06:59_FireFly_of rockbox
22:07:28muesli_-i mean..does the bootloader access on it to start rockbox?
22:08:03linuxstbYes, the bootloader reads that file into memory and then runs it.
22:08:20muesli_-cheers linuxstb
22:11:04 Quit webguest86 (Client Quit)
22:11:12mirakin the codecs, is there a distinction from rockbox of the container and the codec itself or is it all handled by the codec ?
22:11:32mirakthe codec programm in rockbox I mean
22:11:37linuxstbIt's all handled by the codec.
22:11:38mirakI can't be less clear
22:11:44linuxstbI knew what you meant :)
22:12:07preglowmy, it'll be fun to see if this works
22:12:11preglowthe odds are slim, slim, slim
22:12:12mirakso in the case there is two channels in a container
22:12:26mirakpreglow: xvid ?
22:12:33preglowno, ipod button driver
22:12:37preglowi don't care about video on portables
22:12:46mirakdon't be negative
22:12:55preglowtowards portable video i am negative
22:12:57preglowit's a waste of time
22:13:10preglowimho, of course
22:13:22mirakI used it on various ocasion
22:13:30preglowbut the scary part is i'll probably end up helping if you get some code going
22:13:31mirakI used it in the train
22:13:50mirakI am not there yet
22:14:05linuxstbI'm sure lots of people will help - it's the first step that no-one's taken yet.
22:14:24mirakfor me that's less useless than playing PSP in the metro
22:14:29mirakI have seen guys doing that
22:14:38mirakI would puke my breakfast for sure
22:14:38MmmmSo then...about this ticking... could the flickering improvements have worsened the ticking noise? Does that sound likely?
22:14:39preglowi'm more keen on reading on the train
22:14:43preglowand listening to music, of course
22:14:58miraksleeping also is nice
22:14:59preglowbut yes, i would of course help optimising it
22:15:12preglowit's always fun to see what one can get out of limited hardware
22:15:47mirakI am sure somebody wil have done it before I even start to code
22:16:30mirakbut well anyway I am understanding mpeg compression, I wanted to study that since a long time
22:17:04*preglow tries out his cowboy style button driver
22:17:08linuxstbI wonder if Apple will launch a video Nano before we get Rockbox working.
22:17:20 Quit ender1 (Read error: 110 (Connection timed out))
22:17:43preglowIT BLOODY WORKS
22:18:02linuxstbHas it crashed yet?
22:18:18preglowbut this was way, way more than i expected
22:18:38*petur crosses his fingers as he tries his backlight pwm code
22:18:45linuxstbSo did the file browser actually browse?
22:18:45preglowrockbox keeps assuring me there's nothing to resume
22:18:58mirakpreglow: you try to run rockbox on ipod ?
22:19:01preglowi have no up and down
22:19:03preglowi'll try it now
22:19:10peturLinusN is right - it flickers :( but it works!
22:19:35mirakhave though of giving max priority to the music thread ?
22:19:45mirakI have ogg pausing sometime when scrolling
22:19:48preglowlinuxstb: well, i need to redefine some code to make it work first, there is no UP and DOWN for ipod
22:19:53mirakor pushing a button too long
22:19:58preglowor have you fixed that in the file browser already?
22:20:18preglowi have not attempted to make
22:20:23preglowcliock wheel work
22:20:32 Quit qwisp11 ()
22:20:50linuxstbup/down are mapped to the scroll wheel in the file browser - see apps/tree.h
22:21:07preglowi haven't got scrool wheel working
22:21:36preglowit's a bit flakey as of yet
22:22:19preglowmenu works
22:22:46peturis it OK to add a setting for LCD backlight for the H300 target?
22:22:53CassandraDoes sprintf(string, "a%s", string) work OK?
22:23:08linuxstbYes - but use snprintf
22:23:10preglowbookmarks empty
22:23:22 Quit [1]ender (Read error: 104 (Connection reset by peer))
22:23:51 Quit Mmmm ()
22:24:01preglowcontext menu works
22:24:18 Join ender` [0] (i=ychat@
22:25:51linuxstbpetur: Is that a setting to adjust the brightness?
22:27:08linuxstbDoes the contrast setting do anything?
22:27:15peturnot that I know
22:27:22petur9TO DO) :)
22:27:33petur9 = (
22:28:02peturiRiver FW can set the contrast
22:28:16peturmaybe they play with the LCD signal timings
22:30:06peturlinuxstb: contrast for h300 is dummy
22:30:09 Quit akaidiot ("( :: NoNameScript 3.81 :: )")
22:30:40 Quit ze (Read error: 110 (Connection timed out))
22:31:16Cassandrapreglow, doesn't work in wxWidgets. Bleh to them.
22:31:47_FireFly_Cassandra: is string an wxString ??
22:32:15_FireFly_why not use the .Format instance-methode of the class
22:32:19_FireFly_or similar
22:32:34CassandraAnd I'm using wxString->Printf, of course (which is equivalent to a vsnprintf, whatever that might be.)
22:32:39peturwould we need backlight fade in/out on the H300 as well?
22:32:51petur(while I'm at the settings stuff)
22:33:22Learcassandra: sounds like you want wxString::Format...
22:33:50muesli_-petur sure! its fantastic on h100
22:34:11CassandraI just copied to a temp buffer. It's all good.
22:34:25peturbut the PWM only has 16 levels :( unless I also do the Slasheri trick
22:34:39linuxstbI think Linus said that backlight fading didn't work very well on the h300 when he played with it.
22:34:43_FireFly_Cassandra: you should also do following string = "a"+string
22:35:00peturso hardware PWM for continuous setting and software PWM for fading
22:35:58peturI think I know what Linus saw: with lower PWM values, you get a bit of flickering if the drive is being accessed. The original FW has this as well
22:36:15peturif the drive spins down - steady light ;)
22:36:43linuxstbA nice way to indicate disk activity...
22:36:57peturI would also make this setting available to plugins - flashlight anyone?
22:37:34peturbtw, I already used my H340 as a flashlight once!
22:37:41linuxstbThat's the main use for my ipod at the moment - I find my way to bed with it instead of turning the light on.
22:38:06CassandraSome bugger has been stealing my brackets.
22:38:07preglowthis is just too funny
22:38:28preglowi've got a feeling we're not clocked too high as of yet
22:38:39linuxstbNo, I don't think so.
22:38:53preglowi've got click wheel working
22:38:57preglowbut damn, it handles like a dog
22:39:06 Quit paugh ("Leaving")
22:39:33preglowand a small, small, teensy stroke just keeps it going for seconds
22:41:21preglowit's a wee bit flakey as of yet too
22:41:23preglowwanna try it out?
22:42:20preglowif so:
22:42:52 Quit mikearthur (Read error: 104 (Connection reset by peer))
22:43:03preglowthat's for 4g
22:43:48 Join mikearthur [0] (
22:44:56 Join Jungti1234 [0] (n=jungti12@
22:45:44 Join actionshrimp [0] (
22:48:44 Join San [0] (
22:49:08linuxstbpreglow: It works, but the scrollwheel doesn't seem quite right....
22:49:44preglowlinuxstb: no, it's completely off the bat
22:49:50preglowlinuxstb: setting time works just nice
22:50:00preglowso kudos on that driver
22:50:16preglowscroll wheel works here, but it's hypersensitive
22:50:18preglowand lagged
22:50:50linuxstbAt last, I've managed to change the font...
22:51:03preglowto what? :P
22:51:05 Join guillaumh [0] (
22:51:13preglowthe default is fine like wine here
22:51:28linuxstbTo chicago12
22:51:29 Part guillaumh ("bye")
22:52:00preglowmy, this things is lagged
22:52:17linuxstbIs it right that the backlight only comes on when you press action?
22:52:26preglowit should come on for all buttons
22:52:28preglownot click wheel
22:52:31preglowi just hacked that in right now
22:52:50preglowi can't browse fonts
22:52:52preglowjust doesn't work
22:53:26preglowbut ok
22:53:32preglowthis things seems really, really slow at the moment
22:53:39preglowperhaps i should try clocking it up a notch
22:54:08linuxstbIt's gone crazy now - moving up and down randomly...
22:54:20linuxstbAh, finally it's stopped...
22:54:30preglowyeah, it can queue a bunch of event
22:54:35preglowand they take aeons to complete
22:55:13peturmust new settings be added at the back? Or keep them where it's logical
22:55:34 Join Mmmm [0] (
22:56:49preglowclocking the cpu up gives results that are completely nuts
22:56:59preglowit behaves just like before, just faster
22:58:17peturechoing... must new settings be added at the back? Or keep them where it's logical
22:58:38preglowlinuxstb: it might seem your fat driver is somewhat off here
22:58:48preglowit uses lagea mounts of time to find out bookmarks don't exist
23:00:12 Part Mmmm
23:01:46 Join akaidiot [0] (
23:03:38 Join jlo [0] (
23:04:09preglowyou got some tests for me now?
23:04:11preglowi've gotta go soon
23:06:47 Quit jlo (Read error: 104 (Connection reset by peer))
23:08:18 Join jlo [0] (
23:08:51 Quit amar ("CGI:IRC")
23:09:04jlopreglow : I'm uploading now, in some minutes it's ok
23:12:42Learpetur: preferably add at the back, that way you don't need to bump the config version.
23:12:54linuxstbpreglow: I hope it isn't a problem with the fat driver...
23:13:08preglowwell, i'm way out of my league no matter what it is
23:13:12linuxstbAnyway, it's nice to see that most of Rockbox is working nicely.
23:13:23preglowsoon only playback remains...
23:13:32 Join muesli- [0] (
23:13:54***Saving seen data "./dancer.seen"
23:13:55linuxstbYep, once we get a reliable button driver, we can look at playback.
23:14:03preglowi don't want to :////
23:14:05peturif I enable backlight fading for H300 as well, I do have a problem (config files of H300 no longer compatible)
23:14:16peturbest to kick out fading at the moment?
23:14:46jlopreglow : so it's 4 times "starwberry fields" (original, winamp, rb, reaktor anonymously numbered). it's not the best song to do it (quality is not so good) but it's just a trial
23:15:10preglowdon't i need a txt file for the abchr prog?
23:15:33preglowno idea have i make it
23:16:59jlopreglow : no, do it yourself, just choose in setup the 4 files, the first will be reference (easy to recognize the original !)
23:18:08jlopreglow : and after you can compare two by two to gives quality notes
23:18:19 Quit Lear ("Chatzilla [Firefox 1.5/undefined]")
23:18:22 Quit San (Read error: 110 (Connection timed out))
23:19:39preglowi didn't get this
23:19:42preglowi only get three slider groups
23:20:28jlopreglow : after your trials, you save the results and then it will create the text file
23:21:18jloyou have 3 sliders : 1 compared to 2, 1 to 3, 1 to 4
23:22:02jlo1 is the one called orig wav
23:22:44preglowbtw, is this really the best example track?
23:22:46preglowsounds pretty mono to me
23:24:35jlopreglow : the beginning is very unbalanced to left so it's very happy with crossfeed (60's songs !)
23:25:30preglowjlo: there's something wrong withg my rig here, i get mono from everything...
23:25:33MoosJungti1234: we fucks him ;)
23:25:44preglowpreglow's special mixer setup
23:26:29 Quit zeero ("Leaving")
23:26:42jlopreglow : it should be very un-mono !
23:27:11MoosJungti1234: since 11 september lot of arabic and muslim racism
23:27:28preglowjlo: yes, i fucked up in the mixer matrix here, i've got it correct now
23:27:38preglow48 channels get confusing at times
23:27:56 Quit muesli_- (Read error: 110 (Connection timed out))
23:29:16Jungti1234Moos: Isn't the good event.
23:31:03jloso preglow, can you hear something now
23:31:15preglowyes, it's fixed now, but i gotta catch a bus now
23:31:24preglowso i've got to do the test tomorrow
23:31:32preglowlater, all
23:31:53Mooswe are trust to be arabic and/or muslims, and we fucked all those stupid people, the violence didn't solve any problem
23:31:57Moosnight preglow
23:32:21jlook, take your time, and if you have some ideas about other songs...
23:32:33Jungti1234Does Australian hate Korean?
23:32:40preglowi'd love lucy in the sky with diamonds
23:32:49preglowjlo: just post links here or sokmething, i'll read the log tomorrow
23:32:51*preglow runs awayt
23:32:59jlook ,
23:33:24MoosJungti1234: fortunatly not all australian people are racist ;)
23:34:06jlofine, I'do "lucy" tomorrow, goodbye
23:34:07Jungti1234Don't look like so to me.
23:34:12 Part jlo
23:34:21MoosJungti1234: ?
23:34:32Jungti1234They look like 'A mad dog'.
23:35:25Moosthis kind of people is international, you can found they around the worls, fortunatly just a minority of people
23:36:00Jungti1234They spoke 'Disappear in this land if is not Australian'.
23:36:26 Join ze [0] (
23:36:50Moosyes for exemple we had those kind of vocabulary in France too, around the world, but they are minoritary
23:37:34 Quit muesli- (Read error: 110 (Connection timed out))
23:37:44Jungti1234I can't understand it.
23:37:50Moosthe stupidity don't have color or nationality
23:38:14Moossounds like you are'nt stupid so ;)
23:38:28Jungti1234All humans are equal.
23:38:36Moosyes they are
23:38:48Bagderbut some are more equal!
23:38:58*Bagder animal farms
23:39:19Jungti1234I go to have breakfast.
23:39:47Mooshave a good breakfast, and don't hold your heart about this
23:39:58 Join webguest15 [0] (
23:40:01Moosthose behaviours are antiguos
23:40:23Moosand they will be again for long time unfortunatly
23:40:51 Quit webguest15 (Client Quit)
23:41:29_FireFly_someone on this channel as mentioned that the image-flicker fix had the ticking getting louder
23:41:37_FireFly_he was right
23:42:01Moosoops :)
23:42:02_FireFly_i have a little patch which will reduce the ticking again
23:42:11Moosit was your fix, no?
23:42:34_FireFly_and with the reduce-ticking fix the ticking is gone (at least on my device)
23:42:38_FireFly_Moos: yepp :)
23:43:40Mooshere I noticed since I use rockbox the remote tick is gone without the Slasheri's option
23:43:42_FireFly_my new patch uses only one place where the framebuffer gets putted to the lcd
23:44:40_FireFly_Moos: sometimes i didn't hear it also but if a play a bit with the remote-buttons the ticking went back
23:44:48*Moos didn't use his remote, he is use it since TiMiD works and no tick :)
23:45:18_FireFly_do you use the remote or not ??
23:45:37MoosI use Rockbox since few month now, but I use remote just in those cold times
23:46:04_FireFly_maybe you are a lucky guy with an device without the problem
23:46:07Moosbut 1 year ago I remenber I heared the tick with remote
23:46:17Moosyes maybe :)
23:47:36MoosAt start I thought it was one intentionel behaviour from iriver for mimick the beep mode, but...
23:47:38_FireFly_LinuxN: btw, have you any success about this ticking through an check of an "infected" device
23:48:01MoosLinus isn't here
23:48:05_FireFly_LinusN: which you got send by someone
23:48:21_FireFly_Moos: yes i know but i gess he will read the logs ;)
23:48:41Moosyes true :)
23:49:43 Quit t0mas (Read error: 104 (Connection reset by peer))
23:51:25MoosBagder: I just checked the Webstatistics, hehe more pages read, Rockbox will become one big "communauty" :)
23:51:39Mooswith a lot of devices
23:51:58Bagderthe more the merrier!
23:52:36*Bagder discovered that his 'virus' mailbox is now almost 1GB
23:52:47Jungti1234I'm full. :)
23:53:05MoosBagder: ugh :(
23:53:37MoosJungti1234: you are fast :)
23:53:51Bagderspams and virus like me a lot
23:53:52Jungti1234Virus mail is very many.
23:54:33Jungti1234Gmail isolates all it.
23:55:17*Moos is looking in one debate on TV about the colonization
23:55:18Jungti1234G-Mail is safe. :)
23:58:12Jungti1234Is he master?
23:58:54Mooslook like one animal

Previous day | Next day