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 2008-12-31

00:00:19mcuelenaereusb_core_enable_driver()*
00:01:25gevaertsusb_drv_recv() tells the controller to start expecting data, so nothing should be received yet (i.e. the device should be NAKing packets coming from the host until then)
00:01:40mcuelenaereok good
00:02:18mcuelenaerewhat should it do when it's called like this? usb_drv_recv(0, NULL, 0)
00:02:31mcuelenaereso a NULL pointer with size 0 at EP_CONTROL
00:04:07gevaertsI'd say "ack a control transaction", but I guess you want more detail...
00:05:17mcuelenaere:)
00:05:20bertrikZagor, maybe the code on the bottom of the clip is a clue regarding v1/v2 too. My v1 starts with BB, the v2 in the abi thread starts with BH
00:05:33mcuelenaerebut ack a control transaction by using a usb_drv_recv() call, that sounds confusing to me
00:05:42mcuelenaereshouldn't you be using a usb_drv_send() call then?
00:05:47mcuelenaere(which gets used in the code)
00:05:49Zagorbertrik: mine starts with BE
00:06:48 Quit solexx (Read error: 110 (Connection timed out))
00:07:37 Quit fredddy (Remote closed the connection)
00:07:59mcuelenaerehmm I'm not sure if the controller is able to NACK a control packet :/
00:08:06mcuelenaereeh a packet in general
00:08:37gevaertsmcuelenaere: what actually happens is that the host sends a SETUP packet (which contains information about what the control transfer is about), followed by the device sending an IN packet. If that IN packet is empty, (i.e. you don't send data back), you use usb_drv_send(0,0,0).
00:08:42mcuelenaereah no I misread what you said, ignore that last sentence
00:09:12linuxstbTorne: That seems odd - I'm sure I remember much more than a 400ms improvement (i.e. 4 or 5 seconds) when replacing "retailos+RB bootloader" with just the RB bootloader.
00:09:21gevaertsIf on the other hand you do send back data, the host then sends an empty OUT packet to finish the control transaction. You need to receive that with usb_drv_recv(0,0,0)
00:10:06mcuelenaereyou mean the device needs to acknowledge it (as there's nothing to receive)?
00:10:16mcuelenaereif so, that's already done in the EP0_handler()
00:10:40gevaertsAre you sure? This is not the low-level ACK
00:12:22mcuelenaerepackets at EP0 consists of setup and data stages, right? I meant it acknowledged the DATA stage
00:12:24Tornelinuxstb: i'm doing it over now with multiple timings, trying to account for reactions
00:12:39 Quit itcheg ("http://www.mibbit.com ajax IRC Client")
00:12:47mcuelenaerethe datasheet describes it as: ServicedOutPktRdy: The CPU writes a 1 to this bit to clear the OutPktRdy bit. It is cleared automatically.
00:12:57linuxstbTorne: How long is your total boot time?
00:13:20Tornelinuxstb: from hitting menu to rockbox bringing up the menu is 10.3 seconds with retail+bootloader
00:13:30*gevaerts decides to clear up some terminology first :)
00:13:33 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
00:13:52*mcuelenaere hates the USB spec IN/OUT terminology
00:13:59mcuelenaereI always get confused between the two
00:14:11Tornewith just the loader in osos, it's 9 seconds
00:14:33 Join itcheg [0] (i=41d59de2@gateway/web/ajax/mibbit.com/x-2e1936183ca5b1e9)
00:14:34gevaertsAt the lowest level, you have "packets", which come in several kinds, "IN", "OUT", "DATA0
00:14:46Tornewith my patch it's 8.9
00:14:50mcuelenaereyou mean in general, or just for EP0?
00:14:53gevaerts", "DATA1", "ACK, "NAK", "SETUP" and "PING" IIRC
00:14:54linuxstbTorne: Which ipod do you have?
00:14:54Tornebased on five times for each averaged
00:14:57gevaertsin general
00:15:01mcuelenaereok
00:15:01Torneipod video 80gb
00:15:03Torneso, 64mb ram
00:15:15Torneand a 7.4mb retailos, which i think is the largest
00:15:23mcuelenaereI presume the usb controller handles most of them automatically
00:15:31gevaertsA series of these packets is a transaction. You have IN, OUT and SETUP transactions (SETUP only for control)
00:15:33Tornealso the large sectors on the disk..
00:15:50Torneanyway, yeah. i suspect my patchis not worth it after all
00:15:57Torneand i was suffering from poor reactions and confirmation bias
00:16:01gevaertsI would expect controllers to work at transaction or higher level
00:16:09Tornebut neither way is as much of an improvement as 5 seconds
00:16:12gevaertsA group of transactions is a transfer
00:16:18Torne1.5 at most
00:16:25Zagorgevaerts: what do I need to do to make my target run try_reboot() on usb connect? I thought that was the default in SVN but that doesn't seem to be the case?
00:17:29 Join peter__b [0] (n=peter_b@p54978D46.dip0.t-ipconnect.de)
00:17:40 Quit itcheg (Client Quit)
00:17:43peter__bhi all
00:17:45linuxstbTorne: It must depend on the type of ipod - as I said, using "-ef bootloader.ipod" gave a very noticable improvement on my Photo. I'm sure others have reported the same on other ipods.
00:17:51linuxstbI mean -wf
00:18:26gevaertsZagor: actually, the way the code works now, it first enumerates and then reboots when it knows that this isn't just a charger, so you need a USB driver...
00:18:27Tornelinuxstb: hm. someone else has posted on the patch and said it works
00:18:44Tornelinuxstb: it would be interesting to see if anyone else is willing to time it carefully :)
00:18:52Zagorgevaerts: ugh. ok.
00:18:57Tornethey only said it works, though, not how much of an improvement it was
00:19:16gevaertsmcuelenaere: how did you handle these empty packets yesterday?
00:19:19peter__bcan someone tell me if the lcd_* mehtods in the plugin api are threadsafe ?
00:19:25linuxstbTorne: That's dreamlayers - the person who started the forum thread
00:19:27Tornelinuxstb: i'm sure it was very easily noticable the first time i try it
00:19:38Tornebut i can't reproduce that
00:19:43mcuelenaeregevaerts: you mean the usb_drv_recv(0,0,0) ones?
00:19:46gevaertsyes
00:19:55mcuelenaereI ignored them (still do)
00:20:01mcuelenaerenot the usb_drv_send(0,0,0)
00:20:35gevaertsOK. That means they are handled by the controller, or you wouldn't have gotten anywhere
00:21:03gevaertsSo ignoring them is probably the right thing to do (although probably also the right place to document why :)
00:22:03mcuelenaerewill do :)
00:22:11mcuelenaere@documenting
00:23:28mcuelenaereI still wonder though how come when I enable the USB serial driver I don't get a USB driver request popup but an unrecognized USB device
00:23:39mcuelenaerewhich doesn't appear when using the USB charger only driver
00:24:08mcuelenaereso I think it's a related issue like yesterday, meaning something in the USB EP_CONTROL transfers is set wrong
00:24:17 Quit gevaerts (Nick collision from services.)
00:24:26 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
00:24:47mcuelenaeregevaerts: did you get those last lines?
00:25:15gevaertsNot realtime, but I do have the website log open :)
00:25:29Tornelinuxstb: have put a comment on the flyspray entry with my timing results and asked if anyone else is interested in trying it :)
00:25:40*gevaerts doesn't always understand these controllers. The ARC controller works at transfer level in general but you need to handle these control pseudo-ACKs yourself, while other controllers that work at a lower transaction level do that for you...
00:25:43mcuelenaereok :) another question, does usb_drv_send() need to be a blocking call?
00:26:16mcuelenaeregevaerts: I could send you the USB controller datasheet if you want it..
00:26:27mcuelenaerenot that it's THAT helpfull..
00:26:27 Quit ender` (" Who is this General Failure and why is he reading my hard drive?")
00:27:09gevaertsSome code is going to rely on usb_drv_send() being blocking, yes.
00:27:54gevaertsAlthough, when I look at it, on non-0 endpoints all callers actually use usb_drv_send_nonblocking()...
00:28:31*gevaerts suddenly knows what's wrong with that ptp code on the forums...
00:28:45mcuelenaerethe current codeflow makes it kinda asynchronous, so I'll need to manually check and yield() till the transfer gets done
00:29:56Tornelinuxstb: i wonder if portalplayer has any kind of performance counter registers.. that would be a better way to time it :)
00:30:03RockRabbitGartral - Sorry - had to go answer the door. I have converted from flac because so few of the command line compression apps support it.
00:30:09Tornelinuxstb: i doubt it but later ARMs do
00:30:12linuxstbTorne: Yes, there's a microsecond timer register.
00:30:22Torneinitialised at poweron?
00:30:37gevaertsmcuelenaere: if it's for non-control endpoints, don't worry too much. They
00:30:41gevaerts're all asynchronous
00:30:44Torne(or by the rom, at least)
00:30:52linuxstbTorne: Yes. It's referred to by the USEC_TIMER macro in Rockbox - i.e. "int now = USEC_TIMER;"
00:31:31Tornehmm
00:31:50mcuelenaeregevaerts: is usb_drv_request_endpoint() get used heavily? (currently it's not implemented)
00:32:28gevaertsnot heavily, but once per needed endpoint. Not having it will cause trouble
00:32:57mcuelenaerethat could be it (I couldn't see the call for it as there were too many other logf()'s)
00:33:14mcuelenaereit's only there for requesting BULK endpoints, right?
00:33:18Tornelinuxstb: if i hacked rockbox's main to store that right at the beginning and display it in the debug menu, or similar, that would be a much better way to time it
00:33:38gevaertsBULK only for now, until we actually have code that wants interrupt or isochronous :)
00:33:57linuxstbTorne: Or maybe just display it in the bootloader - before the branch to 0x0
00:34:24mcuelenaeregevaerts: this controller doesn't support isochronous, only interrupt and bulk :)
00:34:33mcuelenaereso do I just return eg (1|USB_DIR_OUT) ?
00:34:36Tornelinuxstb: then have it fatal?
00:35:02linuxstbTorne: e.g. something like: "verbose = true; printf("%d",USEC_TIMER); sleep(5*HZ);" <- to sleep 5 seconds, then continue
00:35:13Torneok, will try that :)
00:35:21gevaertsThat would work. It all depends on the controller
00:35:30mcuelenaereok that did the trick :)
00:35:30Tornethat should be far better than 'time read' and me hitting enter :)
00:35:59mcuelenaereso is there any setup guide/tutorial to do usb serial on Windows?
00:36:27mcuelenaerethat way I can test non-EP_CONTROL transfers
00:36:35 Quit RockRabbit ("CGI:IRC (EOF)")
00:36:38mcuelenaereor never mind, I'll use a Linux VM
00:37:30gevaertsJust use storage. You can tell it to use a ramdisk for test purposes by setting a define in usb_storage.c
00:38:39mcuelenaerethat's weird, for some reason it identifies as a Creative Zen Removable Disk storage device
00:38:44mcuelenaereprobably something weird with VMWare
00:39:51gevaertsWhat USB IDs did you use?
00:41:00 Join MethoS- [0] (n=clemens@host-091-096-213-245.ewe-ip-backbone.de)
00:41:00mcuelenaerehmm I probably did a copy-paste job from the config-creativezvm.h
00:43:18gartraljust got off the phone with my brother, he says his v1 clips have serials starting with HA and HH
00:44:01Tornelinuxstb: ahh, i think i have the source of the problem :)
00:44:24Tornelinuxstb: the time is *massively* different depending if you reboot with menu+select, or shut down in rockbox then boot with menu, or reboot from emergency disk mode
00:44:35bertrikgartral, so it doesn't look like a useful indicator
00:44:37gevaertsmcuelenaere: I'd really recommend testing with #define USB_USE_RAMDISK in usb_storage.c. That code is tested a lot more than serial
00:44:47Tornelinuxstb: shutting down then booting with menu takes 6.2 secs, but rebooting with menu+select takes 8.1
00:44:58mcuelenaereok, but last time I enabled that I got some compile errors
00:45:07mcuelenaereit'll probably need some customisation
00:45:21mcuelenaere(and I really want to test logf over USB sometime ;) )
00:46:11linuxstbTorne: Yes, I would expect that. I always tested from power off.
00:46:18gevaertsYou may have to adapt usb_storage_init_connection() a bit
00:46:32 Join blkhawk [0] (n=blkhawk@e176226199.adsl.alicedsl.de)
00:46:35Tornelinuxstb: i only noticed because the apple rom did that thing it does where it doesn't boot and you have to reboot :)
00:46:42linuxstbTorne: Or rather, I wouldn't have expected menu+select to take longer...
00:46:43mcuelenaerebtw has usb_storage been tested with BOOTLOADER defined?
00:46:51***Saving seen data "./dancer.seen"
00:46:54gevaertsgigabeat S does that, yes
00:46:56Tornelinuxstb: that's rebooting from the apple rom having crashed, rather than from insdie rockbox
00:47:26Tornelinuxstb: so, er, yeah. i'll ignore all those and just do shutdown -> boot
00:48:57mcuelenaereI get some 'warning: suggest parentheses around arithmetic in operand of |' warnings in usb_storage.c (MIPSel uses a more recent GCC version)
00:49:37gartralwhy does *menu* go back in pictureflow, and the "rewind" key goes back everywhere else?
00:53:30 Join lilmackie [0] (n=483546c5@gateway/web/cgi-irc/labb.contactor.se/x-ec0a4724ac9567d9)
00:53:39lilmackiehey all, i'm back with another problem
00:54:03lilmackieI installed rockbox with no problems at all, new themes and stuff, and i tested it by puttin on some music
00:54:27lilmackieeverything went fine, and i proceeded to put more music on.... but now that i have it's not showing the second batch
00:54:28lilmackiehelp?
00:54:42LloreanDid you update the database?
00:54:58lilmackiehow do i update it?
00:55:04lilmackiethe first time it asked me to update it and i said ok
00:55:10lilmackienow it's not giving me the option
00:55:29 Quit Zagor ("Leaving")
00:56:16Lloreanlilmackie: How to update manually, and how to turn on auto-update, should both be covered in the manual.
00:57:14lilmackiesorry, of course i should have checked.
00:57:15lilmackielet me go look now
01:00
01:00:43lilmackieLlorean do you recommend that i select the auto update feature?
01:00:49lilmackieor should i just update now, as i see fit?
01:01:16 Join massiveH [0] (n=massiveH@ool-44c48a1e.dyn.optonline.net)
01:03:15Lloreanlilmackie: It's more or less up to you. Auto update is *supposed* to work pretty well, but it has been glitchy for some. I don't use database personally
01:04:48Torneaha, that's what i forgot: i was going to write a patch to have update done on boot if a certain file exists, so my auto syncing script can create it :)
01:04:57Tornetotally forgot bout it till now
01:08:56 Quit massiveH ("Leaving")
01:09:15lilmackieyou don't use database? then what do you use rockbox for?
01:09:19*lilmackie 's eye twitches
01:09:32 Quit nplus (Remote closed the connection)
01:09:36*linuxstb guesses Llorean only plays Doom
01:09:55Lloreanlilmackie: I use it to play music.
01:10:02LloreanI just have a nice, tidy file tree.
01:10:03JdGordonlinuxstb: hey, have you got your wps-statusbar patch somewhere I can have a look at?
01:12:35 Quit tessarakt ("Client exiting")
01:13:32linuxstbJdGordon: This is the only version I can find - linuxstb.cream.org/rockbox/wps_statusbar-v0.1.diff">http://linuxstb.cream.org/rockbox/wps_statusbar-v0.1.diff
01:15:12 Nick tyfoo2 is now known as tyfoo (n=tyfoo@dyndsl-095-033-064-207.ewe-ip-backbone.de)
01:15:19lilmackiefile tree?
01:15:37*gartral syncs all his music in folder tree style, following a [genre]/[artist]/[album]/[file].[ext]
01:15:44gartralformat
01:15:56*lilmackie is uber confused
01:15:58linuxstblilmackie: What you see in the "Files" menu in Rockbox....
01:16:10lilmackieoooooooooooh
01:16:19lilmackiesee you have to explain things to me :P
01:16:32lilmackiei'm a nerdy girl but i have to pick things up first before i master them.
01:16:45kugellucent, keanu: new patch at FS #9645
01:17:18keanuk
01:17:27lilmackieon that note
01:17:29lilmackiethanks guys ^_^
01:18:03Tornelinuxstb: ok, that's much less variable timings: retail+loader takes 6.1s, loader takes 4.3s, rockbox+loader (my patch) takes 4.0s
01:18:57gartral14.4 seconds?
01:19:12 Quit bertrik ("PC LOAD LETTER")
01:19:12Tornegartral: no, average for each different boot method
01:19:38Tornei.e. you can save about two seconds by not loading retailos into ram at boot
01:19:43gartralohh, ok, sorry, im trying o keep up with more than i can handle
01:20:01Torneand my patched loader method might be a little faster than just using the loader on its own
01:20:03 Quit lilmackie ("CGI:IRC (EOF)")
01:20:04Tornebut not by a lot :)
01:21:55gartralany speedup is an improvment in my eyes, no matter where, or what it is
01:22:23Tornegartral: yah, but the point is you don't need to change the code at all to get most of the improvement
01:22:39linuxstbOr do anything special when you update Rockbox...
01:22:47Tornethe tiny extra benefit from my patch, even if it's consistent on different models, is probably not worth the effort
01:23:21Tornei guess it's the time it takes rockbox to go look for the file on disk
01:23:33Torneand maybe the apple rom is a little faster at accessing the drive?
01:24:14linuxstbThe apple rom has it easy - it just needs to read consecutive sectors from a known offset on the disk. Rockbox deals with the FAT filesystem.
01:24:21Torneindeed
01:26:38JdGordonlinuxstb: thanks
01:27:28 Quit faemir (Remote closed the connection)
01:27:34 Join CaptainKewl [0] (n=jason@cpe-68-173-40-122.nyc.res.rr.com)
01:28:00Tornelinuxstb: i guess wait and see what anyone else reports :)
01:28:06 Quit culture (Read error: 60 (Operation timed out))
01:28:21gartrali think this is a one-off bug, but ill bring too the table, anyway, i was fumbling with my e250, and i accidentally held the menu key and pressed the play key from main menu, too get too the WPS, but it rendered the MM over the WPS and the weird thing is, im unable to reproduce it....
01:31:09linuxstbTorne: Yes. There's one other test which could be interesting - loading apple_os.ipod from a file (with only loader in the firmware partition).
01:32:20Tornelinuxstb: compared to booting it from an osos with retail+loader?
01:32:29linuxstbTorne: Yes.
01:32:40 Join mcuelenaere_ [0] (n=mcuelena@rockbox/developer/mcuelenaere)
01:32:49Tornethat might show up the difference better as it's much bigger
01:33:19 Quit mcuelenaere (Read error: 113 (No route to host))
01:33:23 Nick mcuelenaere_ is now known as mcuelenaere (n=mcuelena@rockbox/developer/mcuelenaere)
01:33:41Tornebut i think my ipod is gettign tired of being rebooted for now
01:33:42Torne:)
01:33:50Torneand my other half is looking bored :)
01:33:53gartrallol
01:34:01Torneso i'll try that another time, and leave it here for now
01:34:08 Join swimmer [0] (i=swimmer@xs7.xs4all.nl)
01:34:13Tornethanks for the sounding board and assistance here!
01:34:19Tornescience is clearly the approach
01:35:07*gartral plays clutch - subtle hustle
01:37:58Torneseeya folks :)
01:38:07Tornerockbox is really fun to hack on
01:38:22Torneso i expect i shall be around more :)
01:38:24 Part Torne
01:43:20 Quit tvelocity (Remote closed the connection)
01:54:55 Join Thundercloud [0] (n=thunderc@84-51-130-71.judith186.adsl.metronet.co.uk)
01:55:51 Quit robin0800 (Remote closed the connection)
01:58:32 Join tessarakt [0] (i=nobody@vpn-cl-164-35.rz.uni-karlsruhe.de)
02:00
02:04:00mcuelenaeregevaerts (logs): I get this in a Linux VM over Windows: http://pastebin.com/m6724bcb0
02:06:13mcuelenaereI think it's due to a bad EP>0 implementation in my code
02:07:47kugelkeanu: did you test it or still busy?
02:07:58keanustill busy =/
02:12:14 Quit Bensawesome (Read error: 54 (Connection reset by peer))
02:14:59 Join Bensawsome [0] (n=Bensawso@c-24-63-80-124.hsd1.ct.comcast.net)
02:16:47 Quit tessarakt (Connection timed out)
02:26:05 Quit moos ("Rockbox rules the DAP world")
02:27:18 Quit einhirn (Read error: 104 (Connection reset by peer))
02:33:04JdGordonkugel: did you play with that patch?
02:39:23 Quit mcuelenaere (Read error: 54 (Connection reset by peer))
02:44:23 Quit Thundercloud (Read error: 60 (Operation timed out))
02:45:04 Join Thundercloud [0] (n=thunderc@84-51-130-71.judith186.adsl.metronet.co.uk)
02:46:54***Saving seen data "./dancer.seen"
02:51:21kugelJdGordon: yours?
02:51:57kugelI gave it a try and noticed some glitches here and there (late redrawing etc) which are probably due to the 250ms timer
02:52:56kugelbut there was some talk about keeping manual redraws in some cases, so that we can cope better with information changes in the statusbar or screen changes
02:54:43kugelLlorean: I wondered if you could have a look at http://www.rockbox.org/tracker/task/9662 again and maybe give it a test
02:55:22LloreanNot until after the holidays, probably. Remind me again on the 5th or later and I'll have a look. :)
02:55:37kugelsure!
02:59:07kugelJdGordon: with "some glitches here and there" I acutally meant "it apparently works as supposed to be", as the late redraw issue was basically the only one I noticed
03:00
03:04:52 Join casainho [0] (n=chatzill@87.196.150.31)
03:05:14casainhoI am getting this error: ./arm-elf/bin/ld: cannot find target/arm/at91sam/rockboxplayerlittle/crt0.o −− what shoulkd I do?
03:05:55 Quit agaffney (Read error: 104 (Connection reset by peer))
03:07:54casainhoI have that file crt0.S on that path, and If I put it on SOURCES and build it is ok If I build again, after remove it again from the SOURCES....
03:08:58casainhoamiconn: are you there?
03:09:25casainhokugel: are you there?
03:10:57 Join agaffney [0] (n=agaffney@gentoo/developer/agaffney)
03:12:09Unhelpfulkugel: broke how, exactly? my ability to debug hardware at all, especially when i don't have it, is a tiny bit limited.
03:12:44kugelcasainho: why do you remove it from sources?
03:13:29casainhokugel: If I have it on SOURCES, I get the error telling that I have duplicated the reset_handler label!!!
03:13:43 Quit Thundercloud (Remote closed the connection)
03:14:07casainhokugel: but If I put it there, build and remove again, I don't have the error of it is missing :-)
03:14:15kugelcasainho: what does your system-target.h look like now?
03:15:02casainhokugel: just this line: #include "system-arm.h"
03:15:05kugelUnhelpful: uhm, some wrong pixels in the main menu at the "Rockbox" lettering, and the yellow in the wps is more looks a bit weird too
03:15:38Unhelpfulbizarre. can you reproduce on the clip sim? then at least i can try to debug it myself... :/
03:16:07kugelcasainho: haven't you put the usual #ifdef _FILENALE_H_ around the stuff in it?
03:16:16kugelUnhelpful: fuze
03:16:31kugelI'm talking about cabbiev2 btw
03:16:37Unhelpful...sorry, yes, fuze. does fuze *have* a working sim?
03:16:58casainhokugel: yes, it's there also that defines...
03:17:56kugelcasainho: I have no idea, maybe you care to actually show the error message?
03:18:00kugelUnhelpful: yes
03:18:37casainhokugel: here is the error: /usr/local/arm-elf/lib/gcc/arm-elf/4.0.3/../../../../arm-elf/bin/ld: cannot find target/arm/at91sam/rockboxplayerlittle/crt0.o
03:19:22kugelnot this one
03:19:39kugelit's obvious that you have to fix the other one
03:19:52kugelUnhelpful: looks correct in the sim
03:20:11kugelUnhelpful: I've nailed it down to that revision, it's really that one that introduced it
03:20:37Unhelpfuli'm not sure how best to debug, on HW i don't have. does logf work on fuze?
03:21:16casainhokugel: did you talked to me?
03:21:24kugelyes
03:21:39kugelUnhelpful: I think so
03:21:55casainhokugel: but I just have that error... what I have to fix? I don't understand...
03:22:18kugellol? "<casainho>kugel: If I have it on SOURCES, I get the error telling that I have duplicated the reset_handler label!!!"
03:23:28casainhokugel: If I remove it from SOURCES. then I got that error about it missing... and other targets don't have that file on SOURCES!
03:23:57casainhokugel: it's like if it is not building automatically...
03:24:03kugelcome one
03:24:09kugelon*
03:24:10kugellisten to me
03:24:17kugelcrt0.S needs to be in sources
03:24:30kugeland you need to fix that error you get when it is in sources
03:25:06kugelsure you get a missing crt0.o error if you didn't compile crt0.S
03:25:46kugelUnhelpful: it's a bit weird that apps/ changes break something
03:27:11kugelUnhelpful: it actually seems like some very few colors are inverted
03:28:15Unhelpfulkugel: sounds like overflow, to me.
03:28:32casainhokugel: but on bootloader/SOURCES I see for other targets like this: #elif defined(CREATIVE_ZVx)
03:28:34casainhocreativezvm.c
03:28:36casainho#elif CONFIG_CPU==AS3525
03:28:59Unhelpfuland in WPS? that uses unscaled loads, exclusively.
03:29:19casainhokugel: no others targets puts there the crt0.S file! Or I am not seeing it? :-)
03:29:19 Quit tyfoo ("Carpe diem")
03:30:14kugelUnhelpful: I haven't said it's related to scaling
03:30:49kugelcasainho: I see many
03:30:54creamycenter2003im sooo excited, i just put rockbox on my ipod nano 1g!
03:31:00kugeljust search for crt0.S
03:31:08creamycenter2003so what should i put on my nano that now plays videos???
03:31:16kugelor crt0 alone, that you even find more
03:31:52 Quit _Auron_ ("Infinity repeatedly denies rumours of plotting with zero to bring down the Universe.")
03:32:11kugelUnhelpful: overflow?
03:32:27casainhokugel: but that crt0.S file is not on SOURCES file, under bootloader direcotry, or not? −− yes, for sure that I need the crt0.S!! I had to costumize my file and the linker script, mas last months were working on it ;-)
03:32:50kugelcasainho: firmware/ of course
03:33:21kugelyou're supposed to put your crt0.S into firmware/target/../rockboxplayerlittle/
03:33:29kugeland add it to firmware/SOURCES
03:34:00casainhokugel: AH :-) on SOURCES of firmware and not on SOURCES bootloader ;-)
03:34:18kugelbootloader/ contains very little, most stuff is in firmware, all the drivers and stuff
03:34:36 Quit tchan ("WeeChat 0.2.7-dev")
03:34:40kugelyou only add your bootloader.c in bootloader/SOURCES
03:35:38 Quit DerDome (Read error: 60 (Operation timed out))
03:35:48 Join DerDome [0] (n=DerDome@dslb-082-083-234-119.pools.arcor-ip.net)
03:37:19 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
03:37:22 Quit MethoS- (Remote closed the connection)
03:37:49kugelUnhelpful: I'll read the logs, in case you have some ideas
03:38:00 Quit kugel (Remote closed the connection)
03:40:25 Quit Actium (Read error: 110 (Connection timed out))
03:41:00 Quit Aurix_Lexico (Read error: 110 (Connection timed out))
03:43:12 Quit Schmogel (Read error: 104 (Connection reset by peer))
03:44:26 Join Aurix_Lexico [0] (n=comrade@c-68-56-205-239.hsd1.fl.comcast.net)
03:52:29 Join _Auron_ [0] (n=DarkAuro@ppp-70-244-161-118.dsl.rcsntx.swbell.net)
04:00
04:00:19Unhelpfulkugel: there can be an error where a bright value in one of the channels "wraps" to a dark one. i'm not sure *how* that could happen in the unscaled case, or how that changeset could be related to such a thing... but that sounds like what you described, i think?
04:06:19offsetIs there some way to change the resolution of X to something higher than 1024x768 when using the vmware dev image? I've been editing the XF86Config-4 file and trying to use xrandr with no luck.
04:06:21 Quit casainho ("ChatZilla 0.9.84 [Firefox 3.0.5/2008121622]")
04:07:45offsetnever mind
04:11:29 Quit dfkt ("-= SysReset 2.53=- Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn.")
04:11:51 Join blkhawk- [0] (n=blkhawk@g228064121.adsl.alicedsl.de)
04:12:50 Quit blkhawk (Read error: 60 (Operation timed out))
04:13:49 Nick blkhawk- is now known as blkhawk (n=blkhawk@g228064121.adsl.alicedsl.de)
04:27:03 Quit Slack_ (Remote closed the connection)
04:46:57***Saving seen data "./dancer.seen"
04:47:48 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:51:23 Quit ChadFM (" HydraIRC -> http://www.hydrairc.com <- The alternative IRC client")
04:52:41 Join Darksair [0] (n=user@58.192.37.107)
04:54:05 Join Barahir_ [0] (n=jonathan@BAAfcf0.baa.pppool.de)
04:55:54 Join gartral1 [0] (n=Gareth@adsl-75-33-78-225.dsl.bcvloh.sbcglobal.net)
05:00
05:00:02 Part gartral1
05:01:52 Quit gartral (Read error: 60 (Operation timed out))
05:04:32 Join tchan [0] (n=tchan@c-67-173-9-133.hsd1.il.comcast.net)
05:10:37 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/mibbit.com/x-d1d49619732b7621)
05:10:59 Quit Barahir (Read error: 110 (Connection timed out))
05:13:44 Join itcheg [0] (i=62db4767@gateway/web/ajax/mibbit.com/x-3e6329cc8580aacc)
05:16:05 Join gartral [0] (n=Gareth@adsl-75-33-78-225.dsl.bcvloh.sbcglobal.net)
05:17:09gartralanyone know why rbutil reports *ALL* versions of rockbox as "3.0/196xx"
05:19:31LloreanWhich version of rbutil, do you mean different 196xx versions, or do you mean it literally says "196xx" as indicated by your use of " marks?
05:20:40gartralin m1.0.8 the read for rockbox version is, without the quotes, "3.0/r19621"
05:20:52 Join goffa_ [0] (n=goffa@216.220.23.105)
05:21:21LloreanAnd what version is actually on the player, and what does the rockbox-info.txt on the player report?
05:24:16gartralwell, i just updated it, so i know the version is r19621, and rockbox-info.txt confirms
05:25:03LloreanSo the problem is just the 3.0/ preceding it?
05:25:39gartralyep, that sums it up, also, i should not, just well its on my mind, that ANY file masquerading as "mingwm10.dll" will quell rbutils anger... ;p
05:25:46gartralnote*
05:26:09LloreanYou should file a bug report about the presence of the 3.0/ before the version number then
05:26:54gartralwell, first, are we shure its a bug, and not someones ill-attempt at a "original installation" feature?
05:27:34LloreanYes.
05:29:15gartralok, the reason i asked is cause there are other things with the same naming conventions
05:30:35gartralnamely the fonts, which reads, again, minus quotations, "20080823/20081230"
05:32:21gartraleverything else, minus the boot-loader section which is appears to be an MD5 hash, reads either 20080823, or 20080404
05:32:30 Quit goffa (Read error: 110 (Connection timed out))
05:32:59gartralreplace "823" with 411" sorry, my arin is crosswired
05:33:03gartralbrain
05:34:33gartralanyway, how should i file that report?
05:37:46LloreanJust describe that some of the version numbers seem to either be hardcoded or report something other than the version currently installed and that it's unclear what it's reporting
05:39:26gartrali get that... under WHAT should i file it?
05:40:03 Quit Darksair ("Everything that has a beginning has an end.")
05:40:09LloreanThere's a category for rbutil...
05:41:04gartralan its not reporting wrong numbers, its just report 2, on older, and the current
05:41:06 Join ment [0] (i=thement@ibawizard.net)
05:41:45menthi
05:41:52menthas anyone tried to port rockbox to nintendo ds?
05:42:58LloreanNot yet.
05:44:19gartralwhy in the world would you want to put Rockbox on the DS?!?!
05:44:57_Auron_um
05:44:58gartrali could almost see the PSP, but the DS has enough hacks for media...
05:45:01 Quit perrikwp ("http://www.mibbit.com ajax IRC Client")
05:45:01_Auron_just use moonshell. o.O
05:45:05 Quit itcheg ("http://www.mibbit.com ajax IRC Client")
05:45:53gartralnot too mention how hard that would be to circumvent there OS, let alone the hashed bootloader
05:47:48mentwell the output quality of most mp3 decoders for nds sucks (don't know why, maybe because of libmad)
05:48:42LloreanWe use libmad as well, but the nds is pretty slow.
05:50:15mentthe arm9 should be able to decode 256kbps cbr mp3 (helix)
05:50:47LloreanHelix is much more optimized than Libmad
05:57:29gartralwhat catagory should i file my report for mixed version numbers in rbutil under?~!
05:58:09LloreanThere's a category explicitly called "rbutil"
05:58:12LloreanI already told you this.
05:58:24gartralahh, looked twich, just saw it, sorry
05:58:31gartraltwice
05:58:56gartralok FS link http://www.rockbox.org/tracker/task/9732
06:00
06:05:56 Join zz_Bensawsome-1 [0] (n=Bensawso@adsl-074-167-039-241.sip.cha.bellsouth.net)
06:06:26 Quit Bensawsome ("The awsome is gone :(")
06:06:56 Nick zz_Bensawsome-1 is now known as Bensawsome (n=Bensawso@adsl-074-167-039-241.sip.cha.bellsouth.net)
06:07:49 Join saratoga [0] (n=62e14d04@gateway/web/cgi-irc/labb.contactor.se/x-dff9e0fd8fdd766d)
06:08:17saratogaLlorean: for what its worth, the NDS is probably on par with PP502x in terms of CPU power
06:08:26saratogaat least ignoring dual core
06:08:40 Quit amiconn (Nick collision from services.)
06:08:43 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
06:09:01 Join pixelma_ [0] (n=pixelma@rockbox/staff/pixelma)
06:09:07JdGordonthe NDS isnt really a good target other than to say we did it...
06:09:36JdGordonnot really enough storage to be worth it (assuming you have plenty of games on the chip)
06:10:07Lloreansaratoga: I thought it was a 33mhz and a 66mhz core.
06:10:26mentJdGordon: in most recent homebrew cards you have sdhc slot
06:10:29saratogaLLorean: it is, but its ARM9 w/ DSP extensions
06:10:45LloreanAh
06:10:48JdGordonment: yeah so? my 6gb chip has 6gb of games on it
06:11:06 Quit pixelma (Read error: 110 (Connection timed out))
06:11:07saratogathe DS would be a pretty easy target to port since theres already GPLed drivers for its hardware
06:11:19saratogaand you can put uSD and CF cards in some of the adapters
06:12:04mentJdGordon: i have about ~3GB music on 4GB card
06:12:36JdGordonyeah, waste... my muisc is on my phone or daps
06:12:48gartralyea, and you have a lot of hardware to spoof and wright to flash, unless you want to really patch there BL, then thats at least a jtag overwright
06:13:15gartralthier*
06:13:27gartraltheir* smegging fingers
06:14:19mentbtw rockbox has it's own implementation of libc-like functions or it uses bsd libc (from devkitarm)?
06:17:44saratogament: i believe we use newlib
06:18:48gartralnot that would prove very usful, but does anyone think that the Sandisk sansa slot player would benefit from rockbox? of course, it would be mainly a deno of rockbox's voice UI
06:25:50saratogaapparently devkitarm uses newlib too
06:32:28 Join geokite [0] (n=geokite@ip68-6-252-29.sd.sd.cox.net)
06:41:52 Join _lifeless [0] (n=lifeless@89.20.104.191)
06:41:57 Quit __lifeless (Read error: 131 (Connection reset by peer))
06:47:02***Saving seen data "./dancer.seen"
06:50:49 Nick geokite is now known as geokite_ (n=geokite@ip68-6-252-29.sd.sd.cox.net)
06:51:13 Nick geokite_ is now known as geokite (n=geokite@ip68-6-252-29.sd.sd.cox.net)
06:58:38 Join jbenjg [0] (n=jbenjg@c-24-91-157-7.hsd1.ma.comcast.net)
07:00
07:02:29JdGordonlet me know if you notice any wierdness with the stausbar now please... I'm hoping I got all of them but dunno
07:06:18 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/mibbit.com/x-cbf274510b301e78)
07:08:06 Quit dew (Read error: 104 (Connection reset by peer))
07:10:00saratogawhats the advantage of decoupling the status bar drawing?
07:11:15 Join peter-b [0] (n=peter_b@p54978E44.dip0.t-ipconnect.de)
07:11:42 Join webguest06 [0] (n=4cbea446@gateway/web/cgi-irc/labb.contactor.se/x-b1563ebc1e187593)
07:11:57 Join Kaervak [0] (n=kaervak@cpe-76-190-164-70.neo.res.rr.com)
07:12:06 Quit webguest06 (Client Quit)
07:12:06JdGordonapart from a nice green... its one less thing to worry about when drawing screens, and sizing viewports
07:12:45 Quit peter__b (Read error: 110 (Connection timed out))
07:12:58Kaervakquick stupid question, is anyone else having problems downloading rockbox?
07:14:07JdGordonflcush your dns.... one of the mirrors might be having problems
07:14:15 Join doomed [0] (n=44c126d4@gateway/web/cgi-irc/labb.contactor.se/x-7dcb80c104dbc515)
07:15:28 Nick doomed is now known as alienbiker99 (n=44c126d4@gateway/web/cgi-irc/labb.contactor.se/x-7dcb80c104dbc515)
07:15:50 Quit alienbiker99 (Client Quit)
07:15:51 Part creamycenter2003
07:15:59 Join alienbiker99 [0] (n=44c126d4@gateway/web/cgi-irc/labb.contactor.se/x-d7ffcf25af370ab0)
07:16:31Kaervakhmm, something must be really screwed up on my end. it's still not working after flushing things.
07:17:22JdGordonare you trying through rbutil? or the website?
07:18:08Kaervakjust the website
07:19:47JdGordontry the http://haxx.rockbox.org mirror (I tihnk thats the master one...)
07:20:01geokitein the wps Widecabbie for the e200, the current track information is yellowish, while the next track information is whitish. I'm trying to modify the wps file to change these font colors. The foreground color setting does not change them (odd to me). What tag specifies the font color in the wps? (I've searched the manual, the forum for "font color", and the wps file for the color values (three values of xx, correct?)
07:20:34Kaervakoooh, that one works. Thanks a lot
07:21:02JdGordongeokite: if you want them as differnt colous you need to fiddle with their viewport lines (the %V tag line)
07:23:07 Quit Aurix_Lexico (Read error: 110 (Connection timed out))
07:25:37geokiteexcelent. Just did a search in the manual, no luck. Did a search in the forum for %v, got the message that search words must be at least two characters long. Any tips on where I can find out what the values after %v stand for?
07:26:00 Quit alienbiker99 ("CGI:IRC (Ping timeout)")
07:26:22JdGordonCustomWPS wiki page
07:29:09geokiteMost excellent, found just what I was looking for. Thanks Jd
07:30:06 Join Tofuik [0] (n=44d19c0c@gateway/web/cgi-irc/labb.contactor.se/x-3b065df3102dbeb2)
07:32:32Tofuikdoes rockbox work with 3rd gen Nanos?
07:34:17 Quit Tofuik (Client Quit)
07:35:43 Part toffe82
07:59:55 Quit saratoga ("CGI:IRC (EOF)")
08:00
08:00:02 Quit BHSPitLappy (Remote closed the connection)
08:03:20geokiteAfter looking at the wiki concerning x and y coordinates; where is (0,0) on the players screen?
08:08:11 Part Kaervak
08:11:22rashergeokite: top-left
08:12:22 Join Rob2223 [0] (n=Miranda@p4FDCE41B.dip.t-dialin.net)
08:14:53 Quit tophe ("Ex-Chat")
08:17:15 Quit abo_ (Read error: 110 (Connection timed out))
08:19:10geokiteThank you!! Cheers everyone
08:19:21 Part geokite
08:24:34 Quit perrikwp ("http://www.mibbit.com ajax IRC Client")
08:27:52 Quit jhMikeS (Nick collision from services.)
08:27:58 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
08:30:45 Quit Rob2222 (Read error: 110 (Connection timed out))
08:34:16 Quit _lifeless (Remote closed the connection)
08:35:36 Join _lifeless [0] (n=lifeless@89.20.104.191)
08:36:41 Quit XavierGr ()
08:40:27 Join Darksair [0] (n=user@58.192.37.107)
08:46:20 Quit anewuser (Client Quit)
08:47:05***Saving seen data "./dancer.seen"
08:49:07 Part gartral
08:57:58 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
09:00
09:00:34 Quit amiconn (Remote closed the connection)
09:00:35 Quit pixelma_ (Remote closed the connection)
09:01:18 Join pixelma [0] (n=pixelma@rockbox/staff/pixelma)
09:01:19 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
09:02:29 Join stoffel_ [0] (n=sfr@p57B4FCA0.dip.t-dialin.net)
09:23:42 Nick Barahir_ is now known as Barahir (n=jonathan@BAAfcf0.baa.pppool.de)
09:24:37 Join dotch [0] (n=dotch@212.49.89.23)
09:28:43 Quit _lifeless (Remote closed the connection)
09:32:06 Join _lifeless [0] (n=lifeless@89.20.104.191)
10:00
10:00:05 Quit JdGordon (Remote closed the connection)
10:00:53 Quit lucent (Read error: 60 (Operation timed out))
10:01:40 Join JdGordon [0] (n=Miranda@123-243-140-31.static.tpgi.com.au)
10:02:08 Join Horschti [0] (n=Horscht@p4FD4F6D9.dip.t-dialin.net)
10:02:22 Quit Horscht (Nick collision from services.)
10:05:37 Join culture [0] (n=none@cpc1-bele3-0-0-cust658.belf.cable.ntl.com)
10:14:18 Join write_erase [0] (n=Olivier@217.108.144.241)
10:15:25write_eraseHi ... I get Undefined instrucion at 01e88078 when I try to play a .mp3 file. I run rockbox - svn on Ipod Mini 2ndGen .... Help please
10:21:19 Quit stoffel_ ("leaving")
10:22:58 Join fredddy [0] (n=freddy@p3E9E3883.dip0.t-ipconnect.de)
10:24:17 Part dotch
10:36:33pixelmawrite_erase: is it always the same mp3 or did you try different ones?
10:38:34pixelmaand stating the exact revision of the build you are using is probably more helpful than "svn"
10:47:06***Saving seen data "./dancer.seen"
10:49:49 Join gartral [0] (n=Gareth@adsl-75-33-78-225.dsl.bcvloh.sbcglobal.net)
10:50:58 Quit write_erase (Read error: 113 (No route to host))
10:51:17gartralok, this is a severe bug, i cannot add anything too the playlist, all it does is ask if i want to delete the current playlist, for literally a nano-second, and then quits to menu, resuming where it paused
10:51:47gartrali cannot purge the playlist by saving, or opening another, ether, same behavour
10:51:57gartraleither*
10:52:35JdGordonwhich version?
10:53:49JdGordonbah, ok yeah I see the problem... i cant fix it right now though
10:53:55gartrali just updated, like 35 seconds ago, too the latest
10:54:15JdGordongo back to the latest daily...
10:54:20gartralstaarted playing my old playlist, decided i wanted to change it, and now, i cant
10:54:31gartrali am running latest daily
10:54:43JdGordonrun yesterdays daily
10:55:07pixelmaJdGordon: something to do with your statusbar commit?
10:55:07gartrali would, but your website is reporting 404
10:55:24 Join bs66_1 [0] (n=sysuser@79.138.146.122.bredband.tre.se)
10:55:28scorche|shgartral: flush your DNS
10:56:15JdGordonpixelma: yep... yesno screen.. easy fix.. i'm in windows now so cant actually test, but ill commit the fix
10:56:17gartralits a direct bridge network, i have to pay AT&T $40 to flush the DNS for my network, ide rather not seeing as im $500 in the hole as it is
10:56:53gartrali think ill just wait till it reupdates
10:56:54scorche|shgartral: then you just got an unlucky mirror...
10:56:58*scorche|sh nods
10:57:28gartralbut its "www.rockbox.com" that, i never had problems with
10:57:29*JdGordon hopes visual studio doesnt mangle the line endings...
10:57:42amiconnWhoa a JdGordon commit caused a green delta. Now if it also works properly... ;)
10:57:43scorche|shoh?...nothing seems wrong with www.r.o
10:57:45gartralJdGordon: it will, more than likely
10:57:59gartralwww.r.o?
10:58:05scorche|shwww.rockbox.org
10:58:23pixelmagartral: rockbox.com?
10:58:39JdGordonthanks amiconn :p
10:58:48gartrallol, www.r.o comes up as a potentially malicious site
10:59:26scorche|shsorry...colloquialism from OLPC stuffs
10:59:27gartralblah, its 5 am, i havent slept since yesterday, leave me alones
11:00
11:00:21gartralof course, it does help when i type ORG not OGR
11:02:52JdGordonok fixed... the next build should work fine
11:03:21gartraland it helps when trying to call a website through an AP, not a printer
11:03:27JdGordonshould even have correct line ending and spaces!
11:04:12gartrallol
11:05:59 Join ender` [0] (i=krneki@foo.eternallybored.org)
11:06:20gartralhow long till its up and avilable?
11:06:57JdGordon7min or so
11:07:06JdGordonunder the current build link
11:07:31gartralill just use rbutil
11:08:14gartralwhich reminds me, the mingwm10.dll dependy can be filled wit any file renamed to mingwm10.dll
11:08:46gartrali used a simple txt that contains "hello world"
11:08:54JdGordonwtf? haha
11:10:38gartrali figured youed get a kick outa thar
11:10:44gartralthat
11:11:16 Quit bs66_ (Read error: 110 (Connection timed out))
11:12:31gartraldoes the iAudio M3 not have a screen?
11:12:47amiconnBagder: The Clip and the Fuze are missing in the delta table
11:13:40 Join abo_ [0] (n=abo@rm-ddb-1ef24.adsl.wanadoo.nl)
11:14:03pixelmagartral: no, the Iaudio M3 does not have a screen on its main unit but can use a lcd remote
11:14:05amiconngartral: Not on the main unit. It does have an lcd remote though
11:14:44*gartral thinks thats just simply bassackwards, but, dosnt have an M3 too worry about it
11:16:43gartralnow that its at r19624, mabey i can play somthing other than Still Alive
11:20:14gartralso, just courios, what caused RB to ignore input like that?
11:21:37JdGordonI idnt fix that screen to handle the new system event which happens very often.. so as soon as it got it it exited
11:22:45gartralhuh
11:23:04gartralwell, im glad it was caught quickly
11:23:26gartralthat was truly anoying
11:30:22 Part jbenjg
11:30:31 Join jbenjg [0] (n=jbenjg@c-24-91-157-7.hsd1.ma.comcast.net)
11:35:48 Join bertrik [0] (n=bertrik@ip117-49-211-87.adsl2.static.versatel.nl)
11:41:16 Quit JdGordon (Read error: 104 (Connection reset by peer))
11:44:16gartralwhy wont the bitrate feild change while playing a VBR file?
11:46:57linuxstbIt just shows the average bitrate.
11:47:53gartralnot really... it shows what the bitrate is at start...
11:48:31gartralmost of the files in playing go from 170-350 kbs, no way average is 93
11:49:57 Join MethoS- [0] (n=clemens@host-091-096-213-245.ewe-ip-backbone.de)
11:51:43linuxstbThen the VBR header in the file is incorrect. I assume these are MP3s?
11:52:06linuxstbDoes Rockbox show the correct length for the track?
11:53:37gartralno, its anywhere from 1 too 4 seconds sort
11:53:44gartralshort
11:54:06gartraland yes, there MP3s and OGAs
11:54:17gartralOGM, rather
11:55:48gartraland vbrfix further breaks them, too the point they report 50min+
12:00
12:01:04gartralim sorry, nvm, i had a corrupt file, vbrfix does nothing at all
12:20:49 Join Thundercloud [0] (n=thunderc@84-51-130-71.judith186.adsl.metronet.co.uk)
12:40:47 Part gartral
12:47:08***Saving seen data "./dancer.seen"
12:58:37 Join kugel [0] (n=chatzill@unaffiliated/kugel)
13:00
13:02:20kugelhalf of JdGordon's greens delta is my job :)
13:02:25 Join Jaykay [0] (n=chatzill@p579E78B7.dip.t-dialin.net)
13:07:57 Join Dauron [0] (n=DarkAuro@ppp-70-249-156-112.dsl.rcsntx.swbell.net)
13:16:12 Quit _Auron_ (Read error: 110 (Connection timed out))
13:16:12 Nick Dauron is now known as _Auron_ (n=DarkAuro@ppp-70-249-156-112.dsl.rcsntx.swbell.net)
13:18:10 Join Domonoky [0] (n=domonoky@rockbox/developer/domonoky)
13:20:21 Join tvelocity [0] (n=tony@athedsl-4399031.home.otenet.gr)
13:20:38bertrikjhMikeS, you increased the volume on the si4700 radio, and I think it's still fine on my clip when compared to music playback
13:25:57 Join AndyIL [0] (i=AndyI@212.14.205.32)
13:28:51 Join moos [0] (i=Mustapha@rockbox/staff/moos)
13:29:23 Join faemir [0] (n=faemir@88-106-244-173.dynamic.dsl.as9105.com)
13:29:58 Quit bertrik (Remote closed the connection)
13:31:25 Join tyfoo [0] (n=tyfoo@dyndsl-095-033-105-007.ewe-ip-backbone.de)
13:35:23kugelJaykay: I'm wondering why you didn't post your benches onto the SansaRuntime page
13:36:15Jaykayyeah im wondering too
13:36:17Jaykayill do it
13:37:11 Quit AndyI (Read error: 110 (Connection timed out))
13:39:25Jaykaykugel: should i upload the file and write the link to the dev mailing list or should i write all of them in the table?
13:39:50kugelI'd write them down in the table
13:40:34 Quit Horschti ("I am root. If you see me laughing, you better have a backup")
13:41:33 Join InDAVIDualism [0] (n=David@c-98-224-111-57.hsd1.ca.comcast.net)
13:41:46InDAVIDualismHello.
13:43:37InDAVIDualismNeed some help...
13:44:12 Quit thegeek_ (Read error: 110 (Connection timed out))
13:49:13 Join splinter- [0] (n=splinter@c-68-46-245-71.hsd1.ar.comcast.net)
13:49:13 Quit splinter (Read error: 104 (Connection reset by peer))
13:49:18 Nick splinter- is now known as splinter (n=splinter@c-68-46-245-71.hsd1.ar.comcast.net)
13:54:05advcomp2019InDAVIDualism, we cant help without knowing what you need help with
13:54:42InDAVIDualismOkay, well, I didn't know if anybody was active or not.
13:55:13InDAVIDualismWhat I'm trying to do is get the rockbox firmware running on my sansa fuze, and I'm completely lost here. Pretty sure what I need at this point is a boot manager or something.
13:56:10advcomp2019are you building the bootloader and firmware because the fuze is not supported yet
13:56:55 Join fdinel [0] (n=Miranda@modemcable204.232-203-24.mc.videotron.ca)
13:57:28InDAVIDualismIndeed.
13:59:28advcomp2019ok.. you might want to wait till someone that knows a bit more about the ams sansas then.. i have not built rockbox yet for one of them
13:59:52linuxstbInDAVIDualism: You need to merge the Rockbox bootloader with a copy of the Fuze original firmware using the rbutil/mkamsboot/ utility.
14:00
14:00:30 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
14:01:41InDAVIDualismIs there a guide or walkthrough or something I could use to do that?
14:02:10linuxstbI don't think so - just type "make" in that directory, and run mkamsboot to get usage information.
14:02:49 Join itcheg [0] (i=62db4767@gateway/web/ajax/mibbit.com/x-a8a95de4179c7cd6)
14:03:29 Join thegeek [0] (n=nnscript@85-89-4.72.3p.ntebredband.no)
14:04:05Jaykaykugel: the tag in the empty wps was %wd?
14:04:27kugelI don't know what you put there?
14:05:12LambdaCalculus37linuxstb: Can you take a look at this forum post? http://forums.rockbox.org/index.php?topic=19981.0
14:05:42LambdaCalculus37I think we should look for a couple more mi4 keys for c200/e200 firmwares if possible.
14:06:29linuxstbYes, you should ;)
14:06:46InDAVIDualismrbutil/mkamsboot/dualboot.S is what I want?
14:07:18linuxstbInDAVIDualism: No. You type "make" in that directory, then run the "mkamsboot" program that's created. Don't touch anything else.
14:07:40LambdaCalculus37linuxstb: I shall. :)
14:08:07InDAVIDualism..So what is it I'll need to download exactly, if you please..
14:08:56linuxstbLambdaCalculus37: I think they also need adding to the bootloader, so it can load encrypted "OF.mi4" files
14:09:35linuxstbInDAVIDualism: Have you already downloaded the source code to Rockbox, and compiled the bootloader and main Rockbox builds?
14:11:07LambdaCalculus37linuxstb: No problem.
14:12:05InDAVIDualismI've no idea what I've done.
14:12:06 Quit splinter (Read error: 104 (Connection reset by peer))
14:12:18InDAVIDualismI've a .rockbox folder on my Fuze already
14:12:27 Join splinter [0] (n=splinter@c-68-46-245-71.hsd1.ar.comcast.net)
14:12:40InDAVIDualismAnd I got the latest firmware for the fuze off of.. Daniel's site or something I think
14:13:15InDAVIDualismI've tried with a few other programs to try and get this working, but I kept getting error messages along the way, because I"m not using a supported model.
14:14:10 Quit tyfoo ("Carpe diem")
14:16:26 Quit Thundercloud (Read error: 60 (Operation timed out))
14:17:09InDAVIDualismThere was a file.. Fuze-in-tree.diff, in the section of stuff that people've been working on regarding the Fuze, but I didn't know what to do with it, and when I opened it, it was opened as text in my brower, so I just copied and pasted as a .diff on the root hdd. That file description said something about the boot manager or something I'm sure.
14:18:42 Join splinter- [0] (n=splinter@c-68-46-245-71.hsd1.ar.comcast.net)
14:18:45 Join miepchen^schlaf [0] (n=miepel@p579ECC82.dip.t-dialin.net)
14:19:50 Quit splinter (Read error: 104 (Connection reset by peer))
14:19:50 Nick splinter- is now known as splinter (n=splinter@c-68-46-245-71.hsd1.ar.comcast.net)
14:19:51kugelInDAVIDualism: the bootloader installation is described on the SansaV2 wiki page
14:20:09 Join JdGordon [0] (n=jonno@rockbox/developer/JdGordon)
14:20:30kugeland "Fuze-in-tree.diff" is a patch which is already in SVN
14:20:56 Join RoC_MasterMind [0] (n=Free@c-76-122-43-188.hsd1.fl.comcast.net)
14:21:10kugelbut really, if you don't know what you're doing, I recommend to do nothing
14:21:14JdGordonkugel: taking my credit are you... :D
14:21:16kugelrockbox isn't usable anyway
14:21:35kugelJdGordon: You took mine! :P
14:22:00InDAVIDualismidk how to use it as it is... w/o having to do the other stuff, 'cause the apps that do it automatically don't run, non-recognized model...
14:23:15kugelInDAVIDualism: there's no "i do it for you" app right now
14:23:26InDAVIDualism;-(
14:24:10InDAVIDualismI just need direction.. where it is I need to go to get it done... I've been wandering around four a couple hours now! D=
14:25:21kugelit's all documented
14:27:41InDAVIDualismSo much of it!
14:29:45 Join splinter- [0] (n=splinter@c-68-46-245-71.hsd1.ar.comcast.net)
14:29:48 Quit itcheg ("http://www.mibbit.com ajax IRC Client")
14:29:55 Quit splinter (Read error: 104 (Connection reset by peer))
14:29:55 Nick splinter- is now known as splinter (n=splinter@c-68-46-245-71.hsd1.ar.comcast.net)
14:30:27InDAVIDualismWhere does the rockbox utility install the downloaded version of rockboxes it retrieved? to my media player?
14:32:04kugelInDAVIDualism: do you understand that the fuze is not supported?
14:32:15InDAVIDualismI understand, yes.
14:32:20kugelthere's no rockbox utility for unsupported players
14:32:34InDAVIDualismHence my need to do it manually. =\
14:34:54LambdaCalculus37InDAVIDualism: Right now, Rockbox on the Fuze is at a point where not a lot of stuff is working. It's only useful for developers looking to test and debug. It's not at a point for end users to pick up and use every day.
14:36:19LambdaCalculus37There's still a lot of missing functionality, and while I can't tell you what to do, I'm wondering why you're so anxious to run it now.
14:36:37 Join Thundercloud [0] (n=thunderc@84-51-130-71.judith186.adsl.metronet.co.uk)
14:36:38InDAVIDualism=\
14:36:50InDAVIDualismI want to play games at work on it?
14:37:38gevaertsGood:) They need work, so we need volunteers for that
14:37:39 Join Blackfrosch [0] (n=Blackfro@p57A75C37.dip.t-dialin.net)
14:38:04Jaykaygevaerts: the games need work?
14:38:18LambdaCalculus37I think the majority of plugins need work.
14:38:53Blackfroschhi
14:38:57kugelhe wants to play games...at work, and not work on games ;)
14:39:26Blackfroschyeah cause playing is much easier than working on them :-)
14:40:09InDAVIDualismHey, I might work on them, I get bored sometimes!
14:40:11Blackfroschdoes anybody know where I can get a rockbox 3.0 release for an ipod mini gen2?
14:40:40LambdaCalculus37Rockbox 3.1 has been released already.
14:41:00LambdaCalculus37Go to the Releases page of the site.
14:41:34Blackfroschthx for helping - but I am not that lazy as you think
14:41:51BlackfroschI know the rb page pretty good
14:42:11Blackfroschthe thing is car adapter mode is no working on my ipod with release 3.1
14:43:02Blackfroschso I am looking für 3.0 - last version it was working with...
14:43:12gevaertshttp://download.rockbox.org/release/3.0/rockbox-ipodmini2g-3.0.zip
14:43:40Blackfroschwhere did you get this link?
14:43:57Blackfroschcompiled on your own?
14:43:59gevaertsI took the 3.1 link, replaced a bit, and tested
14:44:11Blackfroschsorry for that
14:44:59Blackfroschcould have been my idea - but even google cache was up to date
14:45:16InDAVIDualism...
14:45:21InDAVIDualismI'm still entirely lost here...
14:46:48Jaykayindavidualism: the fuze is not supported, many things dont work, you are not supposed to play games on work and i think the plugins are partly not working. let it be.
14:47:06Jaykayit doesnt make much sense.....
14:47:11***Saving seen data "./dancer.seen"
14:47:23InDAVIDualism=\
14:47:25LambdaCalculus37InDAVIDualism: Then I suggest you stop what you're doing, go do your work, and wait until the Fuze is supported (which we don't know when that'll be).
14:48:00LambdaCalculus37If you want to use Rockbox that badly, go get one of the supported players listed on the front page.
14:48:27Jaykaykugel: the bench-results are now on the sansaruntime-page.... do you feel better now? :)
14:49:23kugelyep
14:49:34InDAVIDualismSupposing I did get it to work at some point, how would I know if it were capable of booting in RockBox? Would it default to it?
14:51:09 Quit LambdaCalculus37 ("Ka-chunka")
14:51:44 Join dfkt [0] (i=dfkt@unaffiliated/dfkt)
14:54:16Blackfroschthx again car adapter mode and backlight on while plugged in is working again!
14:56:08 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
14:56:13 Join stoffel_ [0] (n=sfr@p57B4FCA0.dip.t-dialin.net)
14:56:25mcuelenaeregevaerts: have you read the logs?
14:57:56Unhelpfulkugel: get anywhere with the weird graphics corruption issue?
14:58:06kugelnope
14:59:06kugelUnhelpful: I noticed though that it's not limited to cabbit
14:59:11kugelcabbie*
15:00
15:00:18 Join itcheg [0] (i=41d59de2@gateway/web/ajax/mibbit.com/x-581d39242f219901)
15:01:28 Join japc [0] (n=japc@194.65.5.235)
15:02:06 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
15:02:55 Quit jhMikeS (Nick collision from services.)
15:03:01 Join jhMikeS [50] (n=jethead7@rockbox/developer/jhMikeS)
15:04:08kugeljhMikeS: hey
15:04:35kugellol!
15:04:59*kugel just coped ~3,5GB from his microSD to the internal storage within Rockbox!!
15:05:20JdGordonhow long did it take?
15:05:50kugelnot that long actually
15:06:02kugelthe thing is, that you cannot read/write past 1GB
15:06:20kugelless than 10min
15:06:38JdGordonthat would explain the speed then'
15:06:42JdGordonit didnt actually work :p
15:07:11 Nick JdGordon is now known as JdGordon|zzz (n=jonno@rockbox/developer/JdGordon)
15:07:21kugelJdGordon|zzz: it worked
15:07:50kugelJdGordon|zzz: no, it didn't!
15:12:01 Join m0f0x [0] (n=m0f0x@189.107.183.109)
15:24:30InDAVIDualismWhat language are the games written in
15:24:32InDAVIDualism?
15:26:12J-23Rockbox games?
15:27:39 Quit swimmer (Client Quit)
15:30:01BlackfroschI guess C
15:33:07mcuelenaereInDAVIDualism: there's a LUA interpreter patch in the tracker
15:34:50mcuelenaere(FS #9174)
15:35:20 Join MethoS-- [0] (n=clemens@dyndsl-085-016-165-064.ewe-ip-backbone.de)
15:36:04gevaertsmcuelenaere: yes. I agree that it looks like bad EP!=0. The fact that it finds the interface correctly seems to show that the EP0 bit is OK
15:36:22*gevaerts thinks
15:36:32gevaertsWhat endianness are you using?
15:36:43mcuelenaereLittle Endian
15:36:51mcuelenaereso the same as the ARM targets
15:37:03mcuelenaere(except for the meizu)
15:37:28gevaertsok. That can't cause problems then
15:37:52mcuelenaereI think this issue is probably something on my side (something wrong in the USB driver)
15:37:58mcuelenaereso not in the USB stack
15:39:08mcuelenaereI see the mass storage driver doing an usb_drv_recv() on EP1, but I only get one EP1 request of 31bytes big after some seconds
15:39:25mcuelenaereso I'm not sure whether the host is doing something wrong or whether it's the client
15:42:29 Join gregzx [0] (n=chatzill@dse135.neoplus.adsl.tpnet.pl)
15:42:51gevaertsI suspect the client
15:42:58 Quit MethoS- (Read error: 145 (Connection timed out))
15:47:38 Quit bs66_1 ("Leaving.")
15:47:55 Quit itcheg ("http://www.mibbit.com ajax IRC Client")
15:49:35 Join LambdaCalculus37 [0] (i=44a04303@rockbox/staff/LambdaCalculus37)
16:00
16:03:43 Nick Kopfgeldjaeger is now known as Kopfgeldjaeger20 (n=nicolai@monitor-mode-enabled-on-mon0.phy0.de)
16:03:47 Nick Kopfgeldjaeger20 is now known as Kopfgeldjaeger (n=nicolai@monitor-mode-enabled-on-mon0.phy0.de)
16:08:21 Quit RoC_MasterMind ("Leaving")
16:09:49 Quit Domonoky ("Leaving.")
16:18:11 Join n1s [0] (n=nils@rockbox/developer/n1s)
16:22:02 Join einhirn [0] (i=Miranda@p5B033333.dip0.t-ipconnect.de)
16:23:18martian67http://gizmodo.com/5121311/reports-30gb-zunes-failing-everywhere-all-at-once
16:25:29LambdaCalculus37martian67: No, we still won't port to the Zune. :)
16:25:50fluxnice easter egg they have there
16:30:26Blackfroschhelp help OpenSource Community - Microsoft does not understand their own hardware - help them by reverse engineering
16:33:53 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
16:34:00 Quit advcomp2019 (Read error: 104 (Connection reset by peer))
16:34:53 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
16:35:55 Join thegeek_ [0] (n=nnscript@85-89-4.72.3p.ntebredband.no)
16:36:04 Join Aurix_Lexico [0] (n=comrade@c-68-56-205-239.hsd1.fl.comcast.net)
16:39:01 Quit thegeek (Read error: 60 (Operation timed out))
16:46:28 Quit Jaykay (Read error: 110 (Connection timed out))
16:47:15***Saving seen data "./dancer.seen"
16:47:42 Join Schmogel [0] (n=Miranda@essn-4db6c9fb.pool.einsundeins.de)
16:47:57 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
16:51:24Unhelpfulkugel: hrm... screendump working on fuze? it might help... and if you're seeing corruption in theme bitmaps, it might be worth putting a splash or logf at the top of resize_on_load, just to make sure that it really is using the unscaled output like it should.
16:51:56kugelUnhelpful: should work, let me see
16:53:59mcuelenaeregevaerts: I found a mistake in the driver and now I have this: http://pastebin.com/m748f699c
16:57:26mcuelenaeregevaerts: should I forward EP0 packets who don't have a DATA stage to usb_core_control_request()?
16:58:33 Quit miepchen^schlaf ()
17:00
17:00:33kugelUnhelpful: indeed, ss is working and you see the weirdnesses
17:04:02 Join anewuser [0] (i=anewuser@unaffiliated/anewuser)
17:05:51kugelUnhelpful: http://www.alice-dsl.net/simonemartitz/rockbox/mainmenu1.png
17:06:01kugeland http://www.alice-dsl.net/simonemartitz/rockbox/wps.png
17:08:13 Quit Blackfrosch (Read error: 54 (Connection reset by peer))
17:10:10kugelUnhelpful: the grey logos look a bit weird too in the wos
17:10:12kugelwps
17:19:21 Quit stoffel_ ("leaving")
17:23:01 Join mick_laptop [0] (n=skywalke@clamwin/admin/mickhome)
17:26:46 Join __lifeless [0] (n=lifeless@89.20.117.26)
17:29:19mcuelenaeregevaerts: I get http://pastebin.com/m6aee3f9d currently, doesn't that imply a working EP1?
17:33:51 Join write__erase [0] (n=write__e@royale.aixmarseille.com)
17:34:25 Join toffe82 [0] (n=chatzill@adsl-70-137-196-62.dsl.frs2ca.sbcglobal.net)
17:35:26mcuelenaereanyone opposed to committing FS #9733?
17:36:44write__eraseHi, what is the revision corresponding to latest stable ?
17:37:20write__eraseI mean the -r value to pass SVN to get the latest stable code
17:39:26write__eraseFound v3_1 tag . nevermind
17:40:58Unhelpfulwrite__erase: why do you want latest stable? if you're developing, you should probably be using the most current revision
17:41:17 Quit _lifeless (Read error: 110 (Connection timed out))
17:41:43Unhelpfulkugel: i *see*... that's clearly not even any of the type of problems i was thinking of. weird.
17:42:55Unhelpfulif screendump works, i assume i can open a file and write to it, as well? maybe i should try dumping some stuff to a logfile, or savinge the the load results as a bmp before returning... :/
17:44:17kugelUnhelpful: I think it's weird that it seems limited to my the fuze
17:44:37 Quit einhirn (Read error: 104 (Connection reset by peer))
17:45:06write__eraseUnhelpful, because I get :
17:45:26write__erase* Kernel oops with usb-storage on FreeBSD
17:45:57write__erase* Undefined instructions while playing mp3
17:47:42mcuelenaerewrite__erase: I don't think usb mass storage has changed since 3.1
17:47:50linuxstbwrite__erase: When did you install your Rockbox bootloader?
17:48:07write__eraseYesterday
17:48:24jhMikeSkugel: ...hey hey?
17:49:13Unhelpfulis the screen dump of the in-memory framebuffer, or are we getting it "back" from the LCD driver somehow?
17:49:32kugelMe and Zagor wondered about naming the backlight defines with USE_* instead of HAVE_* (what's your opinion about that?), also we wondered about the name you choosed for the beasts fade
17:50:05Unhelpfuland the fact that it's only happening on target, and only on one target, really bugs me. i can't imagine how that would happen. :/
17:50:26write__eraseWhat about "undefined instuction at 01E99078" when trying to play an MP3 ?
17:51:04 Join xovan [0] (n=chatzill@ip70-173-252-122.lv.lv.cox.net)
17:51:37xovanhello
17:52:25 Part xovan
17:52:41jhMikeSkugel: Hmmm. What it the point of HAVE_ vs. USE_ anyway? Also, what would you call the beasts fade in some general sense?
17:54:35kugeljhMikeS: I'm fine with either (and USE makes for a bit more sense for me in such a (eye-candy) feature-only), but Zagor said using HAVE_ is the consensus and should be consistent within the source (I wouldn't disagree).
17:55:20kugelMe and Zagor think the name of the beasts fade define is a bit confusing
17:55:24 Quit n1s ()
17:57:51 Quit Darksair (Read error: 104 (Connection reset by peer))
17:57:53jhMikeSWhat is it? I was trying to avoid proliferation of new #defines when a target may use something that doesn't fit any generic one.
17:57:56 Quit pabs (Read error: 113 (No route to host))
17:58:29kugeljhMikeS: see also here http://www.rockbox.org/tracker/task/9724 for some discussion
17:58:41jhMikeSBoth gigabeats implement a unique algorthm and both could use a variant of that define
17:59:00 Join pabs [0] (n=pabs@xor.pablotron.org)
17:59:47 Join Seed [0] (n=ben@bzq-84-108-232-45.cablep.bezeqint.net)
18:00
18:03:24jhMikeSPerhaps a CONFIG_BACKLIGHT_FADING == .... could be done instead
18:03:50 Join itcheg [0] (i=41d59de2@gateway/web/ajax/mibbit.com/x-71f50f39d6ecbd29)
18:07:32kugelsounds good
18:10:55jhMikeSAnother idea. You could encode the setting type within the type value, if even, then integers, if odd then bool then a simple #if (type & 1) #else
18:13:06 Join miepchen^schlaf [0] (n=miepel@e178226156.adsl.alicedsl.de)
18:13:56kugelhm
18:15:15 Join WarpTI [0] (n=WarpTI@c213-100-134-54.swipnet.se)
18:15:16jhMikeSat least then it would select throughout without editing files other than the target ones
18:15:25WarpTIhello
18:16:19WarpTIarnt the ams devices like ams sansas used in shitty chinese mp4 players?
18:16:34WarpTI*ams hardware
18:16:57kugeljhMikeS: how about CONFIG_BACKLIGHT_FADE with PWM_, SW_TYPE_1_, SW_TYPE_2, SIMPLE_ [I don't really know how to call beasts fading]
18:17:32kugeland then in config.h #if (CONFIG_BACKLIGHT_FADE == PWM) #define BACKLIGHT_SETTING_INT
18:17:41kugel#elif (...) #define BACKLIGHT_SETTING_BOOL
18:18:09kugelthe apps/ would only use the BACKLIGHT_SETTING_* ones
18:18:51mcuelenaeregevaerts: now I have http://pastebin.com/m1daad028
18:20:05jhMikeSI don't know what to call it either and the same basic define could handle F/X as well as S. For charging I used CONFIG_CHARGING == CHARGING_TARGET :\ and Recorder and beast just define their own algorithms for it.
18:21:17kugelso, e.g. BACKLIGHT_FADE_TARGET /* this requires a target-specific implementation */ ?
18:22:01jhMikeSsure. but how to select setting type?
18:22:29kugeluhm, maybe #define BACKLIGHT_SETTING_BOOL in backlight-target.h then
18:22:50kugelor config-<target>.h
18:23:05 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
18:23:09 Join casainho [0] (n=chatzill@bl10-160-22.dsl.telepac.pt)
18:23:15gevaertsmcuelenaere: usb_core_control_request() should get all control transfers. Whether or not they have a DATA stage is irrelevant
18:23:23jhMikeSBACKLIGHT_FADE_SETTING_BOOL? :) BACKLIGHT_SETTING_BOOL sounds ambiguous.
18:23:42mcuelenaeregevaerts: but I can't pass a NULL pointer to it, can I?
18:23:55kugelyea, just examples
18:25:03gevaertsWhy would you?
18:25:06kugelwhat do you think about the USE_ vs HAVE_ now? I'm undecided
18:25:24mcuelenaeregevaerts: if it doesn't have a DATA stage, what data should I give usb_core_control_request() then?
18:25:47gevaertsIt gets the setup data, not the DATA data
18:25:53casainhohello :-) - I am getting data abort interruption on my code, at random points but everytime related to thread, as example: "thread->tmo.prev = NULL;" and doing debug, looks like there is always a tentative of reading/writing to an reserved memory region. Can someone please help me? give me any clue?
18:25:55jhMikeSkugel: I don't know. USE_HISTORICAL_PRECEDENT? :p
18:26:04kugel:D
18:26:11casainhokugel: any idea?
18:26:42kugelno idea, sorry
18:26:52jhMikeSWhere else is USE_* used?
18:27:05mcuelenaeregevaerts: hmm never mind that, I misread something
18:27:20casainhokugel: thanks :-)
18:27:51kugeljhMikeS: grep shows some (USE_ALLOCA) but not so much
18:29:03kugelI think USE_ROCKBOX_USB is the most prominent one
18:30:00gevaertsUSE_ROCKBOX_USB is meant as a temporary thing. Once the stack works properly, it goes.
18:30:11jhMikeSAh, yes. Hmmm. Odd about that since its a switch on top of HAVE_USB_STACK
18:30:30jhMikeSgevaerts: what happened to the USB tester?
18:30:36gevaertsAlso, it's (I think) different from the HAVE_* things, since the stack is there, and this one just decides whether to use it
18:30:38 Quit japc (Read error: 110 (Connection timed out))
18:30:42mcuelenaerecasainho: if you get random behaviour when using interrupts, double check your interrupt handling code whether it puts the CPU in the same state as before the interrupt was triggered
18:31:22kugelgevaerts: I don't see how's that much different from our backlight fading
18:31:38kugelbacklight is there, you just decide to use fading too
18:31:46casainhomcuelenaere: random behaviour as it trying to access at reserved memory areas?
18:31:53gevaertskugel: the backlight may be there, but the fading isn't :)
18:31:58mcuelenaerecasainho: yep
18:32:11mcuelenaerecasainho: I had related issues when porting Rockbox to MIPS
18:32:13kugelgevaerts: well, with the sw fade..
18:32:26kugelthat can be used for pretty much every backlight
18:32:35kugelas long as it has levels
18:32:50casainhomcuelenaere: can you explain me a bit of what are you thinking? or give me a link?
18:32:51gevaertskugel: I mean, with usb, all code is there. USE_ROCKBOX_USB only decides whether to reboot
18:33:51casainhomcuelenaere: like, why some pointers gets an address of reserved memory area with a different CPU state?
18:33:52*jhMikeS just tries to minimize the complexity of the preprocessor selections and the amount of editing required for new types
18:34:01gevaertsjhMikeS: the tracer? It's here. I've used it a few times, but not very intensively yet. I do know now that even if the PP stack seems to work fine from a user point of view, there are still *lots* of issues with it on the wire
18:34:17gevaertsAnd those issues are definitely not there on the beast
18:34:33mcuelenaerecasainho: that's architecture-specific, and as I'm not familiar with yours I can't really answer that..
18:34:58mcuelenaerebut the fact that pointers (in RAM?) change when switching CPU states sounds a bit suspicious
18:34:59casainhomcuelenaere: thank you!! I will investigate ;-)
18:35:25jhMikeSgevaerts: I'm glad the beast works. I suppose that implies the stack itself is ok? Can you get current readings from the tracer?
18:35:37gevaertsNot with this one, no
18:36:16casainhomcuelenaere: I didn't say that they change when CPU states changes, I just know that pointers address seems to go to a reserved memory areas
18:36:37gevaertsI'm not going to claim that the stack itself is bug-free, but I know of no issues anyway (except a wrong logf I just spotted :)
18:38:32jhMikeSgevaerts: But you were saying that basically the PP issue is something is HW setup, not the stack?
18:38:51gevaertsjhMikeS: that's what the data point to, yes
18:39:37 Join Reptile211 [0] (n=chatzill@host-216-66-248-9.static.linkline.com)
18:40:31jhMikeSgevaerts: one thing I'm curious about on the beast is if anything would point to why the controller needs to be started and stopped before the first connect (if you recall that little glitch).
18:42:36gevaertsmcuelenaere: are you sure the direction bit you pass to usb_core_transfer_complete() is correct?
18:42:49mcuelenaeregevaerts: I fixed that in my local tree
18:43:02mcuelenaereyou mean the endpoint|USB_DIR_IN, right?
18:43:25gevaertsmcuelenaere: I'm just looking at the last pastebin you linked to, not at your actual code :)
18:43:33mcuelenaereah ok :)
18:44:22mcuelenaeregevaerts: is this correct usb_core_transfer_complete(endpoint, USB_DIR_IN, 0, ep->sent); ?
18:47:18***Saving seen data "./dancer.seen"
18:47:47gevaertsIf ep->sent is the size of the transfered data, then yes. From reading the code I think endpoint needs to be with the direction bit stripped though (maybe usb_core_handle_transfer_completion() should mask it with 0x7f itself)
18:48:34mcuelenaereyes, I removed the |USB_DIR_IN & |USB_DIR_OUT in my local tree
18:49:12gevaertsthe USB spec is not always clear on what exactly an endpoint number is, so my code inherits a bit of that confusion :)
18:49:34mcuelenaerebut I'm not sure when I should send an usb_core_transfer_complete() upon receiving a packet
18:49:52mcuelenaereEP1 is set with length=1024 by the mass storage driver
18:50:27gevaertshm, ok. This was easy with the ARC...
18:51:44mcuelenaerecurrently I just forward all packets with usb_core_transfer_complete()
18:51:44gevaertsA transfer is complete when (a) you have all data the caller asked for, or (b) you receive a non-maxpacketsize packet.
18:51:55 Quit MethoS-- (Remote closed the connection)
18:52:00mcuelenaerehmm ok
18:52:48gevaertse.g. if you have a maxpacketsize of 64, and you get three 64-byte packets and one 23 byte packet, you have a 215-byte transfer
18:53:39gevaertsIf you get three 64-byte packets and one zero-byte packet you have a 192-byte transfer.
18:53:59gevaertsIf you just get three 64-byte packets (and the caller asked more), the transfer isn't complete yet
18:54:31mcuelenaereok, I implemented it like that now
18:55:39mcuelenaerehmm I only get scsi: inquiry 0 & ums: set config now
18:55:44mcuelenaerethat's less than I used to get
18:56:08*mcuelenaere hates returning to previous code revisions and figuring out what he did wrong
18:56:29*gevaerts has to go now...
19:00
19:00:11 Quit Reptile211 ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
19:12:30 Quit mcuelenaere (Read error: 113 (No route to host))
19:12:33 Join mcuelenaere_ [0] (n=mcuelena@rockbox/developer/mcuelenaere)
19:13:34 Nick mcuelenaere_ is now known as mcuelenaere (n=mcuelena@rockbox/developer/mcuelenaere)
19:16:23 Quit bmbl ("Woah!")
19:16:39 Quit Schmogel (Read error: 104 (Connection reset by peer))
19:19:28 Quit miepchen^schlaf (Read error: 54 (Connection reset by peer))
19:28:21 Join CaptainKwel [0] (i=2669ecc2@gateway/web/ajax/mibbit.com/x-d639d10ec4cab557)
19:35:59 Part WarpTI
19:38:44 Join vwbusguy [0] (n=scott@fedora/vwbusguy)
19:38:45 Quit casainho ("ChatZilla 0.9.84 [Firefox 3.0.5/2008121622]")
19:39:01vwbusguyWill Rockbox work with any Creative Zen ?
19:40:02LambdaCalculus37The Zen Vision:M is the only Zen with any work done on it, and even that is still not meant for any serious use.
19:40:34LambdaCalculus37It's still very early in development and is missing a lot of functionality.
19:41:37 Quit mick_laptop ("leaving")
19:42:31mcuelenaerevwbusguy: LambdaCalculus37 is right, only the Zen Vision:M, Zen Vision:M 60GB and Zen Vision have _some_ functionality
19:50:37vwbusguyok, thanks
19:50:56vwbusguystill using my blackberry foran mp3 player for now, but thinnking of replacing it
19:51:10vwbusguyor at least getting a designated mp3 player
19:56:17LambdaCalculus37Then you may want to read http://www.rockbox.org/twiki/bin/view/Main/BuyersGuide to get an idea of what works with Rockbox.
19:58:30 Join Davide-NYC [0] (n=Davide-N@user-12ld9li.cable.mindspring.com)
19:59:16Davide-NYCLambdaCalculus37: I cannot find my GigabeatS (LOL) what's the word on the charging code?
20:00
20:01:00 Part vwbusguy ("Leaving")
20:02:32 Join gregzx [0] (n=chatzill@dse135.neoplus.adsl.tpnet.pl)
20:02:49LambdaCalculus37Davide-NYC: jhMikeS is the man to ask.
20:03:32Davide-NYCBet.
20:04:16LambdaCalculus37Davide-NYC: On my end, charging from USB works well. Haven't tried charging from AC within Rockbox yet (force of habit makes me do it from the OF).
20:04:34Davide-NYCjhMikeS: cannot find the gigabeast. Out of curiosity how does it look form a charging code perspective?
20:04:59Davide-NYCLambdaCalculus37: that is stellar news. I just moved and have some boxes still to sift through. Hopefully the beast is not lost.
20:07:50 Quit __lifeless (Read error: 60 (Operation timed out))
20:07:51 Join theredone [0] (n=18837ab0@gateway/web/cgi-irc/labb.contactor.se/x-96e40f5643b33d2d)
20:09:03theredonehello does any one know how i can get my g5 to work with the dlo video docking station
20:10:30krazykittheredone, what is a g5?
20:10:39 Quit Thundercloud (Remote closed the connection)
20:10:44theredoneipod video
20:12:08theredoneive spent three days reserching this with no luck
20:13:12Unhelpfultheredone: use apple's firmware? seriously, the video isn't too great for video playback without its hardware codec.
20:13:35 Join Thundercloud [0] (n=thunderc@84-51-130-71.judith186.adsl.metronet.co.uk)
20:14:12 Quit Davide-NYC ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
20:16:00theredoneya the ipod firmware sucks and id have to use ipods database in order to do it wich means my music will be on ther in two locations
20:16:39theredonethe root and database
20:16:50Unhelpfulso, only put videos in there for ipod firmware? really, 18fps or so is not worth it for video :P
20:18:21theredoneno no im not trying to watch a video im trying to use the docking station witch has an video out so you can see your ipod on a tv
20:19:19LambdaCalculus37Rockbox doesn't make use of video out, nor the Broadcom video chip in the iPod video.
20:19:22Unhelpfuloh. are you sure that the thing is truly an "accessory" at all? there are more than a few devices that treat the ipod as a drive with music files and a database on it.
20:20:06LambdaCalculus37So that will answer the question as "this accessory will not work with Rockbox, at all".
20:20:12theredoneya its an accessory
20:20:23theredoneoh ok
20:20:43theredonewell thats 100 bucks wasted
20:20:57*Unhelpful wonders why people don't check
20:21:39LambdaCalculus37Unhelpful: Because some people are too hasty and don't bother reading the manual before putting Rockbox onto their shiny $300 toy?
20:21:41theredonewell theres no real info about this ive been researching this for three days
20:22:20theredoneand the manual says nothing about a docking station
20:22:25Unhelpfulyou don't *have* to put files on twice to use either firmware, assuming they're itunes-supported formats
20:22:39Unhelpfultheredone: the rockbox manual? that should probably have been a clue.
20:24:06theredoneactualy if i use ipods system rockbox cant read the file names correctly
20:24:10 Quit Thundercloud (Remote closed the connection)
20:24:46theredonecomes up as mmdb or som other four letter name
20:24:52Unhelpfultheredone: no, because you don't get to choose filenames on the device if you use itunes. but you can load songs with itunes and use the rockbox db.
20:25:07theredoneya
20:25:09 Join _lifeless [0] (n=lifeless@89.20.110.205)
20:25:19 Quit gregzx ("ChatZilla 0.9.84 [Firefox 3.0.5/2008120122]")
20:25:22Unhelpfuland then you can start it up in appleOS when you want to use your video dock thingy
20:25:23theredonethats why i hate itunes
20:25:56 Quit m0f0x ()
20:26:01LambdaCalculus37theredone: Haven't thought about using the Database function in Rockbox to view your music?
20:26:58*Unhelpful is sure there must be *some* users doing that to be able to share between appleOS and rockbox
20:27:14theredonei did at first but it changed everything interms of my organization and most of the music was recorded from albums so i didnt tag them
20:27:33theredoneits alot of work
20:27:44 Join japc [0] (n=japc@81.193.11.216)
20:28:20Unhelpfulpop them in picard, and they'll be tagged. ;D
20:28:56theredoneive tyed that alot of themusic i have is pretty obscure
20:29:56theredonealot of drum and bass (techno)
20:30:37theredoneand there stuff comes from mix cds so the tags are from those
20:31:06 Quit japc (Client Quit)
20:33:58_lifelesshappy new year guys, 2009 here ^^
20:34:17theredonesame for you cheers
20:38:37Unhelpfulkugel: bah, screen_dump uses the in-memory framebuffer, so i can't readily blame your LCD driver :)
20:39:09*kugel isn't really surprised
20:40:03Unhelpfulnor am i, really... i've been given the impression that on pretty much all targets, data is essentially "gone" once it goes to the LCD driver, aside from the copy in lcd_framebuffer.
20:42:11Unhelpfuli'll have to kludge up a patch to dump some data to a file while loading, and see if we can't figure out where it gets corrupted.
20:43:21theredonewell hay guys im out thanks for the info i hope this will be looked at for future updates of rockbox
20:44:16Unhelpfulonly if somebody who has one works on it.
20:47:21***Saving seen data "./dancer.seen"
20:48:29 Quit theredone ("CGI:IRC (EOF)")
20:51:17 Quit CaptainKwel ("http://www.mibbit.com ajax IRC Client")
20:58:25 Quit LambdaCalculus37 ("http://www.mibbit.com ajax IRC Client")
21:00
21:02:37 Join Strife89 [0] (n=michael@204.116.245.152)
21:08:59 Join Zagor [242] (n=bjst@rockbox/developer/Zagor)
21:09:34kugelZagor: hi
21:10:03Zagorhi
21:10:14ZagorjhMikeS: still here?
21:10:18 Quit itcheg ("http://www.mibbit.com ajax IRC Client")
21:10:38 Join stoffel_ [0] (n=sfr@p57B4FCA0.dip.t-dialin.net)
21:10:43kugelI put the button driver online which could go in. Due to the still missing Home button, I copied the e200 keymap, so that everything should be usable (Rec is hardly mapped on e200)
21:11:32kugeland I watched the buffering screen a bit on my fuze, interesting things happen before playback crashes
21:12:32Zagorkugel: was that with svn buffering?
21:12:37Zagoror my patch?
21:12:37kugelyep
21:12:41kugelsvn
21:13:12kugelI can't tell why your patch fails on my fuze (and I haven't watched the buffering screen when I tried it)
21:13:29Zagorit fails? how?
21:13:46kugeluhm, I explained what happened on the tracker
21:13:52Zagorok, checking
21:14:27Zagorodd
21:14:46Zagorit works perfectly on my clip, c200 and h140
21:15:01 Quit Seed ("cu, Andre")
21:15:25ZagorI would like more testing of it, especially on ipods. FS #9703
21:16:31Zagorthe pcm buffer back and forth is correct. I changed what the pcm buffer displays and it is now more correct.
21:16:44kugelquite odd, I thought so too
21:16:48 Quit mcuelenaere ()
21:16:58kugelyea, I don't think your patch is wrong
21:17:34kugelit's something on the fuze, just revealed now. I mean, playback isn't working as it should anyway
21:18:00ZagorI am going to order a fuze to be able to test myself. Also I'd like to find someplace to buy a clip v2.
21:18:24kugelah, the port got you :)
21:18:41Zagoryeah, new ports are inspiring.
21:18:47kugelindeed
21:19:59*kugel is glad to get some other people helping him
21:20:03 Join B4gder [241] (n=daniel@rockbox/developer/bagder)
21:21:06kugelZagor: so what I observed in the buffering screen when trying to play ogg/aac
21:21:55kugelthe audio buffer is filled very slowly. the pcm buffer decreases and doesn't quite catch up again (so it's effectively decreasing slowly)
21:22:27kugelonce the pcm buffer hits a certain level, both audio buffer and pcm buffer refill quite quickly
21:22:33Zagordoes cpu boost work on fuze?
21:22:34kugelbut the playback doesn't advance anymore
21:23:00kugelwell, from what I can tell, yes. The buffering screen shows that it's boosted
21:23:12Zagorthe pcm buffer has a critical mark where it starts boosting. that could be related.
21:23:34kugelthe screen shows that it's boosted before and after
21:23:38kugelbut yea, I found that weird too
21:23:40Zagorodd
21:24:02kugelbut when the pcm buffer refills, playback stopped already, so decoding doesn't happen anymore
21:24:36kugeland when the pcm buffer is refilled, the screen looks frozen (neither buffer changes)
21:25:33 Quit faemir (Remote closed the connection)
21:25:58Zagorthe pcm, buffering and playback code contains rather a lot of fragile areas
21:26:19 Join faemir [0] (n=faemir@88-106-244-173.dynamic.dsl.as9105.com)
21:27:02kugelZagor: what I also noticed is, that even though flac plays fine, skipping and seeking is bugged
21:27:47Zagorafter the holidays I plan to start looking into simplification and robustness enhancements
21:29:17kugelsounds good, but I don't think the code is wrong. Something on the fuze is wrong sounds more likely to me
21:29:55Zagoryes, I agree
21:31:58Zagordoes the clip and fuze have the same sd controller?
21:32:49kugelyes
21:33:48kugelpretty much everything is the same code, except buttons, lcd and backlight
21:33:48ZagorI assume playback works in the sim?
21:33:52Zagorok
21:39:20 Join gregzx [0] (n=chatzill@dse135.neoplus.adsl.tpnet.pl)
21:44:14 Quit stoffel_ ("leaving")
21:45:26kugelZagor: is 1MB codec buffer size normal?
21:46:43Zagorkugel: yes
21:47:53Zagorbut clip of course uses a much smaller buffer. 288 KB
21:49:15 Join telliott [0] (n=Tim@68-179-182-168.bsr-c3-d1.evv.dhcp.sigecom.net)
21:49:58*Zagor notes the #if MEMORYSIZE line in apps/codecs/SOURCES should rather be #if CODEC_SIZE > ..
21:50:41telliottI need help with the Archos bootloader. I want to upgrade but I don't see a bootloader for the recorder V1 here http://download.rockbox.org/bootloader/archos/
21:51:38telliottEs the v1 recorder bootloader the same as V2?
21:51:41telliottIs
21:51:42Zagortelliott: it's simply unfortunate file names. flash-recorder-v2.zip is for the v1
21:51:53Zagorflash-recorderv2-v2.zip is for recorder v2
21:52:56telliottIs there anything significant in the updated bootloader?
21:53:13ZagorI don't know
21:53:41telliottok. I'll update just to get rid of the warning anyway. Thankd
21:53:44Zagorask amiconn, he's our main Archos guy
21:53:49B4gderamiconn is our main archos man
21:53:59*B4gder was too late
21:54:40telliottI want to try the Sansa Clip V1 build when it's fully baked. I have a 2 gig player
21:55:54jhMikeSZagor: Still here indeed
21:58:25ZagorjhMikeS: my idea was to rename USE_BACKLIGHT_CUSTOM_FADING_BOOL to smply HAVE_BACKLIGHT_HW_FADING. however I just now noticed we also have HAVE_BACKLIGHT_PWM_FADING. what is special about the gigabeat fading?
21:58:55Zagorkugel: I assume your new button code is the last patch in FS #9645?
22:00
22:00:13kugelthat's right
22:01:24Zagorkugel: ok, committing it now
22:03:51telliottDoes backlight fading work on Archos V1?
22:04:19jhMikeSZagor: What was on my mind: The S isn't generic and uses the PMIC hardware to fade. If the F/X is given options, it's also a unique method but not hardware. For higher-level code, why should it care of the difference?
22:06:17ZagorjhMikeS: I don't follow you. Wasn't it you who added the new define for gigabeat fading?
22:09:37Zagortelliott: no
22:09:38jhMikeSZagor: I did but I was thinking that there shouldn't be a different define for each implementation-specific way of doing fading.
22:10:27ZagorjhMikeS: I agree completely. that's why I want to get rid of USE_BACKLIGHT_CUSTOM_FADING_BOOL
22:10:29 Join saratoga [0] (n=62e14d04@gateway/web/cgi-irc/labb.contactor.se/x-cc351f542bb63c5e)
22:10:51Zagorand possibly merge some of the other fading options too
22:10:59telliottOh well. I love my Archos, upgraded to 120 gigs. I've alwayse said I hope this lasts long enough to large SSD's to become affordable.
22:11:10Zagorahh, fuze build is green at last
22:11:11telliottfor
22:11:12jhMikeSZagor: But it wasn't intended to be used by S alone. F/X could also define that. Each implements it's own algorthm in the target tree instead.
22:11:40B4gdera green line is nice line!
22:11:59ZagorjhMikeS: do they really need a define at all then?
22:12:43jhMikeSPossibly not except for including the proper bool/int options
22:12:48kugelall yellows gone :)
22:14:13kugelso let's do what we had in mind today
22:14:41B4gderdrink champagne and do fireworks?
22:14:54kugelthat too :)
22:15:12B4gder:-)
22:15:35ZagorjhMikeS: I don't understand why we have different settings for different fading implementations. in my view HAVE_BACKLIGHT_FADING is enough and then each implementation solves it best it can.
22:15:50kugeleven though it doesn't look like I'd do fireworks today/tomorrow
22:16:04*jhMikeS has to cure the headache he currently has so he can be sure only the alcohol is causing it later.
22:17:27kugelZagor: my idea was to have #define BACKLIGHT_FADE_SETTING_INT/_BOOL depending on the algorithm used, as in PWM is _INT, all others _BOOL
22:17:42jhMikeSZagor: The beast is constant time but F/X and thread fading depend on the starting level so displaying a time as used in the int version isn't really accurate.
22:18:09kugeland having CONFIG_BACKLIGHT_FADE == ...
22:18:27saratogaJdGordon: Could I ask you about alternating sublines in gwps-common.c?
22:18:49ZagorI rather want to remove the time settings for PWM. that level of configuration is just silly. next we'll be getting a time setting for fade_on_stop
22:19:19*kugel didn't dare to say that, but agrees ;)
22:19:25jhMikeSI guess they can all be bool then :)
22:19:33Zagorexactly
22:20:01jhMikeSPersonally, I think one menu item would nicer with Off/In/Out/Both as the options
22:20:03kugelZagor: I guess there's some people who disagree though. amiconn for example said he has a long fade out, but a short fade in
22:20:25Zagoryeah well guess who added that code...
22:20:32LloreanThe default setting is a short fade in and a long fade out, and I rather like it.
22:20:41kugeljhMikeS: that's what I had in mind too at first, but then we decided to reuse existing LANG strings
22:20:57ZagorLlorean: that sounds very reasonable
22:21:06LloreanIt makes sense that the screen should be instantly (or nearly instantly) readable upon button press, but doesn't necessarily need a near-instant out.
22:21:30jhMikeSkugel: yeah, I think I was part of that discussion and just gave in to keeping them mostly alike
22:21:41Zagordefault is 0.3s fade-in and 2s fade-out
22:23:14ZagorI vote for hardcoding those values and removing the options, and then unite the fading defines.
22:24:17Zagorkugel: is the last patch of FS #9617 commit ready too?
22:24:23*kugel wonders who'd choose a long fade in anyway, and who'd feel strong about the length of the fade out
22:24:49LloreanSome people might want short/short
22:24:57kugelZagor: no, I changed the keymap, so it needs sync
22:25:02Zagorkugel: ok
22:25:04LloreanAlmost like having fade-in/out off, but a little bit softer than that.
22:25:34kugelshould be easy though, mostly use e200 keymap for fuze too
22:26:01ZagorLlorean: but do we really want to cater to every possible opinion? these are really minute details, and they're messing up the code.
22:26:45kugelLlorean: yea, I can imagine people having a short fade out, but I'm wondering who of them will really miss it if they get a hard-coded long one
22:26:49LloreanZagor: Would it be possible to have "Short in, Long out" and "Short in, Short out" and "Off" as options?
22:27:01jhMikeSLlorean: If they want short, Off is the shortest fade possible. :)
22:27:05LloreanOr is the idea to take away an entire option?
22:27:44*Llorean missed what the exact goal is here, to simplify or to remove.
22:27:45ZagorI want a single bool "fading on/off", and the leave the details to the implementation
22:28:10Zagorfight setting bloat! :-)
22:28:26kugel:)
22:28:31LloreanI would imagine then that the best overall option is short in, long out.
22:28:38*jhMikeS still votes for killing crossfade then
22:29:33ZagorjhMikeS: entirely? that's rather different isn't it? I can agree we could probably shave a few settings off of it though.
22:30:37kugelimho, crossfade is one of those things, where you can safe lots of problems with choosing sane defaults and optimizing the code for it
22:30:38jhMikeSZagor: It's just something that makes life harder when thinking about improving the playback end.
22:30:52ZagorjhMikeS: ah, indeed
22:31:54jhMikeSRidiculously long crossfades require a huge pcm buffer just for that.
22:32:02saratogai've wanted to remove crossfade for ages
22:32:13saratogaand maybe readd it someday when sw playback is more stable
22:32:48Zagor30 seconds is really a silly long crossfade duration
22:33:06LloreanCrossfade is probably something we should've removed right before 3.0
22:33:10Zagorand requires a 5.4MB pcm buffer...
22:33:17saratogalets drop archos support and hwcodec at the same time
22:33:55 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/mibbit.com/x-1193079987eab676)
22:33:56Zagoris crossfade buggy, or just messing up the pcm code?
22:34:14kugelback when I used it it wasn't very reliable
22:34:26kugelnever turned it on again since then
22:34:59LloreanZagor: It's been buggy on and off. I think it works right now except maybe across directory changes
22:35:17LloreanLike auto-change-directory directory changes, that is
22:36:29Zagorit sure would streamline pcm a lot... otoh I a can imagine it is a bit of a crowd pleaser, isn't it?
22:36:38 Join ZincAlloy [0] (n=d9eec0ed@gateway/web/cgi-irc/labb.contactor.se/x-a3465a7540a16709)
22:36:38saratogais HWCODEC gapless for mp3?
22:37:02Zagorsaratoga: for lame-gapless files, yes
22:37:31LloreanI thought HWCODEC required properly cut files, not just lame tags.
22:38:14Zagoras I remember, lame has an option to not use the bit reservoir at the end of the track
22:38:22saratogai'm actually kind of surprised that crossfading worked at all on PP until recently
22:38:35saratogamp3 wasn't even 2x real time until the COP patch went in
22:38:39LloreanZagor: I thought it was just that it still need to be cut on the end of a frame.
22:39:02LloreanI wouldn't be against cutting crossfade for the time being, though.
22:39:18saratogathe nogap thing in lame just cuts tracks so that each is an integer number of frames long, thus no gaps
22:39:29saratogai assume we support proper gapless in lame though for hwcodec
22:39:33*kugel thinks removing it altogether is a bit too much
22:40:06jhMikeSsaratoga: int number of frames? I thought it cut the filter delay away which isn't always 1152 samples.
22:40:06saratogawell i guess now that we have releases if anyone really wants it, they could stick with an older release
22:40:25Llorean"Crossfade as a feature never worked satisfactorily and complicates playback code to the point where improvements are far more difficult until they have to be. Until/unless someone comes forward and re-implements it in a better way it will be removed from future builds" or something like that sounds fine to me.
22:40:25saratogajhMikeS: not sure what you mean
22:40:36Llorean*than they have to be.
22:41:18saratogai think no gap just borrows samples from adjoining tracks so that the number of samples in all but the first and last track are an integer multiple of the subband decomposition size
22:41:38saratogawhich is why it sometimes adds bits of the next track to the present track
22:41:56jhMikeSsaratoga: The encoded mp3 with the padding is always an int number of frames.
22:44:01Zagordoes lame have another setting than −−nores? that one disables the bit reservoir, which means each frame is atomic and can be cut/joined gaplessly
22:44:19saratogaZagor: i was referring to −−nogap
22:45:01saratogajhMikeS: yeah i know that the decoded is always an int number of frames, but the incoming PCM isn't, which is all I meant
22:45:03jhMikeSZagor: I think that is in regard to the editable option
22:45:20*kugel is a bit surprised that Llorean actually votes for removing an existing feature
22:45:45jhMikeSsaratoga: I misread what you meant then.
22:45:56saratogai think playback is sort of a special case since no one has been very satisfied with the state of it for many years now
22:46:15Zagorsaratoga: ah right. −−nogap works when encoding a whole album. −−nores works for individual tracks (with some size penalty)
22:46:18Lloreankugel: I vote for removing it because it's apparently done badly. I'd like it very much if it comes back if someone can rewrite it to work better and cause less problems for code maintenance.
22:46:19 Quit XavierGr (Nick collision from services.)
22:46:30 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
22:46:35LloreanBut I'd rather have it gone and everything else work better if that's what's necessary
22:46:47kugelyea, I understand
22:47:04saratogafrankly it always seemed a sort of silly feature anyway given that we're fully gapless
22:47:25***Saving seen data "./dancer.seen"
22:47:27kugelI think removing some/most/all of the configurability will already make life easier with crossfade
22:47:31Lloreansaratoga: People like it for parties and such.
22:47:47ZagorI'm not sure it is necessary for improvement. at least my impression from reading the pcm/buffering/playback code have not made me want to rip out crossfade.
22:47:48LloreanWithout crossfade you still have distinct ends/beginnings.
22:47:53saratogaif you really want crossfading you could just encode the tracks with a crossfade built in and our playback engine would give the same effect thanks to gapless
22:48:01saratogaLLorean: ah good point
22:48:25saratogaZagor: well lets keep the idea on the table
22:48:38saratogaif someone steps up and wants to work on it, we consider it, otherwise, leave it in
22:48:43ZagorI'm not hard to convince about simplifying code :-)
22:49:02saratogano sense pulling a feature if no one is working on playback anyway
22:50:31Zagorwell I'm doing a bit of work on it, and might want to do more.
22:50:38jhMikeSThe worst thing is the huge buffer which if late DSP is ever done, doubles the size since 32-bit samples are backlogged instead of 16-bit.
22:51:33Zagorand I would welcome testing of FS #9703 on targets I don't have
22:53:01*jhMikeS also would love to get rid of that ATA mutex hack for 60/80GB video and the real problem identified.
22:53:13saratogai just want to see stable playback on my sansa without a 2 second delay between when i click a new track and when the old one stops playing
22:55:07kugelhuh? 2s? it's more like 0.2s to me
22:55:33kugela bit more if the next is unbuffered/skipping backwards
22:55:48ZagorI would actually like to commit FS #9703 to get broader testing. I don't know of any problems with it.
22:56:12saratogakugel: for me at least it seems quite random, sometimes its less then a second, other times it seems to take forever
22:56:20saratogaZagor: go for it!
22:56:22kugelZagor: except 1, but that's unsupported :)
22:56:42Zagorkugel: yeah well our fuze owner base isn't likely to be very annoyed :-)
22:57:31kugelI think the patch is good, no matter of the problems my fuze has with it, and should go in soon
22:58:03kugelit should also give a bit more battery life time (if even measurable), right?
22:58:25Zagoryes
22:59:20kugelperfect
23:00
23:04:55 Join LambdaCalculus37 [0] (n=rmenes@rockbox/staff/LambdaCalculus37)
23:05:19*jhMikeS notices "(spinup / HZ) + 1.5" in the patch :)
23:05:46kugeloh is that still in? I already told Zagor about that
23:07:02jhMikeSIn the last one posted there, yes
23:08:23Zagorah, heh
23:08:36kugeluhm
23:09:07kugellistening to a flac, my sansa is really unresponsive
23:09:23kugelbacklight fade takes ages
23:10:12kugelfine without playback
23:10:24kugelfine with mp3
23:10:48saratogais anyone here familar with the code for doing alternating lines in the WPS?
23:11:49kugelnot so fine with ogg
23:12:08Lloreansaratoga: Don't you just separate lines with a ;?
23:12:35LloreanPossibly with the %t (I think) tag for display time?
23:12:49*kugel tries a current build
23:15:36kugelcurrent build is fine
23:15:55saratogaLlorean: i'm trying to figure out how the code for updating those works so i can fix fs#8523
23:16:08kugelZagor: there's apparently some problems with your patch
23:16:31saratogait sounds like the only bug known with that patch is that the counters for the subline updates don't get reset after switching the display off
23:17:07Zagorkugel: those flac+fade issues are with my patch?
23:17:10Lloreansaratoga: Oh, sorry, I thought you meant the in-wps code.
23:17:13*Llorean wasn't thinking
23:17:38kugelZagor: yes, but it's also with mp3 (not so bad, but noticeable) and ogg (even worse than flac)
23:18:03Zagorkugel: you might want to say that earlier :-)
23:18:25ZagorI'll test it too
23:18:25kugelsorry, I just tested it on my sansa
23:22:06kugeljhMikeS, saratoga: I think FS #8523 should be dependent on lcd_enable and/or lcd_sleep, those indicate well if the screen is readable without backlight
23:22:51saratogakugel: verses just a define in the config files?
23:23:39kugelsaratoga: well, given that lcd is pretty much always disabled without backlight once HAVE_LCD_ENABLE is defined, an extra define seems a bit redundant to me
23:24:17saratogakugel: so HAVE_LCD_ENABLE tells you if the LCD will be turned off?
23:24:54kugelif implemented, yes. The backlight driver I looked at do lcd_enable(false) in backlight_off()
23:25:49saratogabut we probably want to have this dependent on the preprocessor not runtime checks
23:26:05 Quit gregzx (Read error: 110 (Connection timed out))
23:26:18kugelHAVE_LCD_ENABLE is a preprocessor define
23:26:36kugelwhich just says if a lcd_enable function is implemented, which in turns may disable the lcd
23:26:46saratogaah ok misunderstood you
23:27:55kugeland I'd think that only on targets being unreadable without backlight such a function is used with backlight_off
23:32:10saratogadoes cabbiev2 use alternating WPS lines?
23:33:02jhMikeSkugel: I'll give a look
23:34:45*jhMikeS wonders how the screen is refreshed properly when the lcd comes back on
23:37:56kugeljhMikeS: lcd_enable does lcd_update
23:39:12 Join miepchen^schlaf [0] (n=miepel@e178226156.adsl.alicedsl.de)
23:40:07jhMikeSkugel: it depends on the contents of the framebuffer being current
23:40:27jhMikeSI mentioned something there that mpegplayer uses to update the YUV
23:40:29Zagorkugel: which target do you have slow fading with? my c200 has no problems.
23:40:44kugele200
23:41:00kugelnot only slow fading, scrolling through lists is a pain too
23:41:23Zagorodd
23:41:49ZagorI don't see any difference from svn
23:45:12 Quit gromit` ("Coyote finally caught me")
23:50:32Zagortesting on e200 now and still don't see anything slow or painful. can you try my build?
23:51:31jhMikeScheck the OS Stacks screen and see if the codec thread priority is elevated for some reason?
23:52:16kugeldata abort now :/
23:52:35kugelat 0x90d0
23:53:20Zagorwow
23:54:38kugelcodec is at 18%, audio at 56%
23:54:53kugelit also sleems to get slower overtime
23:55:53jhMikeSPriority is the "16 16" bit. The % is maximum stack usage so far.
23:56:01 Join ender1 [0] (i=krneki@foo.eternallybored.org)
23:56:31 Join solexx [0] (n=jrschulz@e176113137.adsl.alicedsl.de)
23:57:33Zagormy stack view shows less elevated for flac than for mp3...
23:58:08Zagorkugel: 18% sounds low
23:58:20kugelcodec is "S 9 9 18%", but sometimes jump to "*R 10 10 %18"
23:58:45Zagor*R just means running. +R means elevated.

Previous day | Next day