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

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

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

#rockbox log for 2004-10-09

00:08:39*[IDC]Dragon says goodnight
00:08:55 Quit [IDC]Dragon ()
00:11:49 Quit gromit` (Read error: 110 (Connection timed out))
00:18:29 Quit methangas (" Try HydraIRC -> <-")
00:41:33JoeBorngoodnight all, thanks for letting me hang out and listen
00:41:59 Part JoeBorn
00:54:07***Saving seen data "./dancer.seen"
01:06:37 Join iRiverMan [0] (
01:08:08iRiverManhow is iriverbox coming on?
01:09:18midk'coming on'?
01:16:34iRiverManwell has someone made it boot on the iriver yet?
01:16:45iRiverManmidk u have a karma righ?
01:16:58midkit's not *quite* that simple. the backlight flashes, that's about it, as far as i know
01:17:01midkand no, of course not
01:17:13iRiverMani thought u said u had a rio karma
01:17:30midku throught wrong
01:21:30iRiverManok so i thought wrong
01:21:33iRiverManso what DO you have?
01:22:25iRiverMan u have an ipod
01:25:57 Quit _aLF ("Leaving")
01:27:29iRiverManperhaps not then
01:31:15amiconnmidk: How is breakout coming?
01:34:39iRiverManhe's asleep
01:36:09midkamiconn, it's mostly done, i haven't messed with it recently
01:36:25midkthe collision detection (written by strath) is good but not perfect
01:36:39amiconnThe button repeat rate should no longer be a problem
01:37:06midkit's not
01:42:08midkif you'd like to give collision detection a try, i can send it over..
01:42:32iRiverManmiss my archos
01:42:35iRiverManstill got the hd for it
01:43:51iRiverManis the old hd worth keeping?
01:43:57iRiverManits a hitachi 10gb laptop drive
01:44:10amiconnmidk: I would like to, but unfortunately I will be busy for some more time with Ondio hacking.
01:44:41midki understand... if/when you want it, just let me know
02:01:17 Part amiconn
02:05:52Sebulba02how are you doing the collision?
02:06:27midkit's pretty simple, you just check the position of the ball in comparison to the center of the nearest block
02:07:56Sebulba02sounds easy enough
02:50:00 Quit klaxon (Read error: 110 (Connection timed out))
02:54:11***Saving seen data "./dancer.seen"
03:00:30 Quit AciD (Read error: 232 (Connection reset by peer))
03:09:53 Quit mecraw ("Trillian (")
03:22:51 Quit iRiverMan ("CGI:IRC (Ping timeout)")
04:54:12***Saving seen data "./dancer.seen"
04:56:41 Part scott666_
05:28:33 Join scott666_ [0] (
05:28:42 Part scott666_
06:34:13 Join kramerica [0] (
06:41:06 Join ashridah [0] (ashridah@
06:54:13***Saving seen data "./dancer.seen"
07:57:26 Join scott666_ [0] (
07:57:30 Part scott666_
08:54:14***Saving seen data "./dancer.seen"
09:17:52 Join methangas [0] (
09:28:33 Join [IDC]Dragon [0] (
09:37:05 Join amiconn [0] (
09:38:05amiconnMorning all
09:38:27[IDC]DragonMorning logpeeker!
09:38:53amiconnNah, not logpeeking
09:38:57[IDC]Dragoncongrats to your driver!
09:39:07[IDC]Dragonmission accomplished, I'd say
09:39:07amiconnAlready tried it?
09:39:23[IDC]Dragon(briefly, I admit)
09:39:48[IDC]Dragonbut I just had to flash a ROMbox build of it
09:40:00[IDC]Dragonto see how it starts up
09:40:05amiconnThe only thing left to do is consistent error handling. As it is now it may return the same error code for different failures
09:40:38[IDC]DragonI still get the 50 sec pause with the high bitrate file I tried
09:40:54[IDC]Dragonbut that may be an issue in mpeg.c
09:41:05amiconnI did a total of 3 runs of my fs test plugin, testing with 100 MB.
09:42:22amiconnOne test (with the internal flash) produced a bad fat sector, but the 2 other tests (external MMC with fat16 and fat32) ran without problems
09:43:29amiconnHave to test again with the internal and dump an image for analysis in case it happens again
09:44:26amiconnI can't imagine that it was caused by the driver though
09:45:34[IDC]Dragonhow was the bad sector produced? Rockbox does not verify nor mark bad sectors.
09:46:34amiconnNo, I checked with window's chkdsk afterwards. It complained about a number of cross-linked cluster chains
09:46:58[IDC]Dragonthat sounds like a file system bug
09:47:49amiconnThere were ~20 of them, cluster range 14080..14196. These are all handled within one fat sector (no. 55 if I calculated right)
09:49:26amiconnMaybe the bad sector was caused by the flash chip itself, and it would have told the driver if the driver would check the status register
09:50:06[IDC]Dragonthere's a status for each sector?
09:51:26amiconnNo. You can check status after a write operation by explicitly requesting it (CMD_SEND_STATUS)
09:52:04[IDC]Dragonhmm, might be a good idea
09:52:42[IDC]Dragonbut requires an architecture change in the file system if we want to really mark it as bad
09:53:48amiconnIiuc MMC handles bad sectors internally, but maybe some cards need the data to be re-send if they find such a sector
09:58:01 Join amiconn_ [0] (
09:58:56amiconn_msg nickserv ghost amiconn x14zpt23!
09:59:07 Quit amiconn (Nick collision from services.)
09:59:07 Nick amiconn_ is now known as amiconn (
10:01:30amiconnpassword changed
10:03:48amiconn[IDC]Dragon: Worst case for playback would be 384 kbps MP2, or even video with such an MP2 as audio part if you adapt the video plugin
10:04:43[IDC]Dragonhave you tried high bitrate?
10:05:13[IDC]DragonI played a 224 kBit/s mp3
10:05:21amiconnI just started another test run on the internal flash
10:06:30[IDC]Dragonthis 50 sec pause is a mytery to me, you can't possibly take so long to load <2 MB
10:11:49amiconnNo. I think might caused by the mmc driver starving the bitswap when loading the first big chunk (mpeg.c loads small chunk as long as the buffer level is low, then switches to big chunks)
10:12:10amiconnThe mas is then fed with unswapped data, so it produces silence
10:13:43amiconnThe beast just switched off, and lost the settings again, grr
10:14:13amiconnMay be due to battery level though
10:17:47[IDC]Dragonit can feed *unswapped* data? sounds like a bad design, missing check, or such
10:24:38amiconnI dunno if it may do that, we should ask Linus
10:40:58amiconn[IDC]Dragon: Interesting finding - my last test (that one that stopped and switched off the player) produced a bad fat sector again
10:41:24amiconnWhat is really weird about that - it's *exactly* the same sector!
10:46:34 Quit kramerica ()
10:53:32 Join marc77 [0] (
10:54:18***Saving seen data "./dancer.seen"
10:58:53amiconnAh, I just found why rockbox doesn't keep settings with the internal flash. I accidentally reformatted it in large-floppy mode...
11:02:41 Join Schoki2 [0] (
11:10:51 Join edx [0] (
11:14:00amiconnArgh! I found what happened to the file system. There seems to be a bug in the config sector handling. In super-floppy mode, it wrote the config sector right into the fat area!
11:14:42amiconnGuess what! This is sector 61
11:17:07[IDC]Dragonuh, ha
11:17:51amiconnAt least it's neither the driver nor the file system
11:18:13*[IDC]Dragon says phew!
11:18:57amiconnA nice little "Roc..." within the fat area
11:19:29[IDC]Dragonyes, that's striking
11:19:43[IDC]Dragonthis is how you found it?
11:21:19amiconnYes, although it was by accident that I reformatted my internal flash in superfloppy mode. Still looking how to revert this (WinXP doesn't allow me to do that)
11:22:35 Join [1]marc77 [0] (
11:22:35 Quit marc77 (Read error: 104 (Connection reset by peer))
11:22:36amiconnIt seems there is a bug in the MBR detection in Rockbox. Although the Partiton boot sector is not an MBR, it seems to get detected as such
11:22:38 Nick [1]marc77 is now known as marc77 (
11:23:28amiconnSo rockbox thinks there is a partition table, but does not find a partition. This sets the config sector number to 61.
11:24:13amiconnAfterwards it tries to mount in superfloppy mode anyway, and succeeds. Now the config sector is located within the fat...
11:24:27amiconn[IDC]Dragon: Do you know about MBR docs?
11:24:40*[IDC]Dragon looks
11:26:33[IDC]Dragondo you read the c't?
11:26:58amiconnSometimes... I have no c't issues at home :(
11:29:41[IDC]Dragonissue 6/2000 had background articles
11:30:16amiconnWe need a proper method to distinguish an MBR from a partition boot sector
11:32:56[IDC]DragonI can email you the article
11:33:07[IDC]Dragonhave it on CD
11:34:26amiconnPlease do so, this could be interesting
11:35:27[IDC]Dragonor DCC?
11:35:29amiconnHowever, from what I found on google, distinguishing a boot sector from the mbr is not easily possible, because the only fixed marking of an mbr seems to be the 0x55aa at the end
11:36:01amiconnThe fat boot sector of my image does have the same marking...
11:36:06amiconnDCC is fine
11:37:12[IDC]Dragonlet's try, it rarely worked for me
11:37:23[IDC]Dragonmaybe firewall,router, etc.
11:37:40[IDC]Dragoncoming in?
11:37:46amiconnConnect failed...
11:38:30amiconnMy router is configured for dcc, and it already worked with some people
11:41:34 Join scott666_ [0] (
11:42:41[IDC]Dragonemail then (upload takes a while, 1.5 MB)
11:42:42amiconn[IDC]Dragon: The mbr check is not necessary. I found the bug in my config sector calculation... really stupid one
11:43:05[IDC]Dragontoo late, on the way...
11:43:16[IDC]Dragonbut perhaps worth keeping
11:43:39amiconnI wonder why the compiler did not complain... I wrote if(fat_startsector != 0) instead of if(fat_startsector() != 0) Argh!
11:44:04[IDC]Dragonyou compared the function pointer
11:49:22amiconnTesting again...
11:53:10 Quit marc77 (" WOW! This IRC Client ownz! HydraIRC -> <-")
12:14:04 Quit Schoki2 ("Client Exiting")
12:14:44 Join webguest99 [0] (
12:16:35webguest99I connected my Jukebox to the computer and played musik from it. Suddenly it turned off. Now it doesn`t react, no matter which buttons I try.
12:17:12webguest99I connected it to the wallcharger because I thought of a Batterie problem but that doesn`t help.
12:17:19webguest99Who can help me???
12:22:23 Quit webguest99 ("CGI:IRC (EOF)")
12:24:17[IDC]Dragonoh, he left
12:30:08 Join webguest91 [0] (
12:30:55 Quit webguest91 (Client Quit)
12:42:03 Join AciD [0] (
12:45:31 Join Toni [0] (
12:45:42ToniHi there.
12:46:03ToniBig progress in Ondio development. :-)
12:47:08ToniToday I got a new SD card 1GB which shows at startup "No partition found Insert USB cable to fix it"
12:47:32[IDC]DragonHi Tony
12:47:38ToniAfter doing it, playing works fine
12:47:54[IDC]Dragonyou got an Ondio, obviously?
12:48:07ToniSeems that this disk has no 55aa at the end of boot sector
12:48:38ToniYes the one with 1GB MMC> problems
12:49:13[IDC]Dragonour 1st user
12:49:20ToniNow MMC works fine except USB connection on WinXP
12:49:41ToniIs there any timeout?
12:50:10Tonibecause Win tries to read the card, but gives up after e few second
12:51:15ToniAfter that I can select the new Drive, but the Message says "Please insert disc"
12:52:51 Quit AciD (Read error: 104 (Connection reset by peer))
12:53:50 Join R3nTiL [0] (
12:54:15ToniThe original archos firmware refuses the USB connection with inserted MMC by saying: "Please remove MMC card first"
12:54:20***Saving seen data "./dancer.seen"
12:54:50ToniCurrently the only way to load the MMC is to use a card reader
12:55:55[IDC]Dragonthere's nothing we can do for USB mode
12:56:22Toniooh :-(
12:58:37 Part Toni
13:07:03 Join tron_ [0] (
13:10:16 Join AciD [0] (
13:19:50 Quit R3nTiL ()
13:21:48[IDC]Dragonamiconn: recording works
13:22:05[IDC]Dragonno changes, just adapting the screen
13:26:03amiconnArgh, missed Toni
13:26:22[IDC]Dragonany news for him?
13:26:37amiconnMy test ran ok, no corrupted fat
13:26:56[IDC]Dragonas expected
13:27:55amiconnYes, I know what causes the "no partition", because now that I fixed the config sector calc I get this too. His new MMC is obviously formatted in superfloppy mode
13:29:38amiconnThere seems to be no way with WinXP to create a partition table on such a card.
13:31:25[IDC]Dragonnbo void space on superfloppy, right?
13:32:06amiconnYes, superfloppy means no partition table, file system starts at sector 0
13:39:36amiconnI know what causes the non-working usb access to an inserted mmc from Windows. I'm going to implement mmc change handling while on usb, and the "Please remove..." message thing
13:40:16[IDC]Dragonthe transition from SPI to MMC mode, I guess
13:40:24[IDC]Dragonor, the lack of
13:41:09amiconnYes, partially. You have to extract the inserted mmc first, because that transition is not possible without power cycling.
13:41:38[IDC]DragonI wish they've spent a transistor for that...
13:42:12amiconnThen USB connects to the internal flash of course. If you insert a card while on USB, the bridge gets deactivated and reactivated by the archos fw, causing the PC to detect the change
13:42:25amiconnWe have to implement this too
13:42:59amiconnYes a transistor would have solved the problem, the same way as the reset pin does for the internal flash
13:43:13[IDC]Dragonother topic: recording from mic sounds crappy
13:43:30[IDC]Dragona buzzing noise
13:47:29amiconnTry if it gets better when you change mp3_playback.c, line 556 to val = 0x2d; This should disable spdif out from the mas
13:54:05[IDC]Dragonhmm, still noisy, I'm not sure if this is better
13:55:08amiconnSo it's likely caused by something else you have to find. I presume with the original fw recording from mic is ok?
13:55:34[IDC]Dragonguess not
13:57:34[IDC]Dragonproblem: without RTC, all recordings get the same filename
13:58:08[IDC]DragonI'll have to make a counter or such
13:58:56amiconnThey should not. Isn't there a pseudo-time calculation for the fs that gradually increments for each file written?
14:00:01[IDC]Dragonno, get_time() returns a constant
14:00:35 Join _aLF [0] (
14:00:39amiconnAh, then it's used for the file system only, unavailable from get_time()
14:01:34[IDC]DragonI need to do something like the default name for .cfg saving
14:02:37amiconnTested high bitrate, it goes silent for me too, and returns after ~50 s
14:03:42[IDC]Dragonyou could use your scope to see if there's any MMC activity meanwhile
14:04:07amiconnOtherwise playback works fine even for the highest possible itrate (384 kbps mp2)
14:04:32[IDC]Dragonwit 3 MBit we're way ahead of this
14:08:57amiconnIt also happens when you skip to a high-bitrate track
14:10:17 Join Sebulba03 [0] (
14:10:49 Quit Sebulba02 ("++")
14:14:29[IDC]Dragongotta leave now
14:14:36 Quit [IDC]Dragon ()
14:37:43 Join Toni [0] (
14:37:56Toniamiconn, r u there?
14:38:42ToniI donīt think, my SD-card is super-floppy formatted.
14:39:43ToniWhen I do system_reboot() right after first time disk_init() the next time disk_init() is ok.
14:39:49ToniStrange, or?
14:40:31amiconnHmm. What does the debug info tell about that problematic MMC?
14:40:50Toni* just checking
14:42:36 Join _Schoki [0] (
14:42:56Toni25MBit/s, 1.5ms, 0clk, *16
14:43:33amiconnIt's really 25 MBit/s ?? That value
14:43:47ToniItīs a SD-card. not specified for MMC, at least for 3.x standard
14:44:21amiconnHow do you fit an SD card into the slot?
14:44:45Tonihard, but its fit somewhat
14:45:30amiconnUrgs. SD is thicker than MMC, I tried it, doesn't fit without forcing it
14:46:08amiconnI wonder if the SD init is sligtly different, and if I could adapt the driver to that
14:46:29amiconnOT: I just fixed the high-bitrate playback probs in cvs
14:46:58ToniYes, thatīs true. My sister (another Ondio user) got that SD-card by accident and is happy with herīs.
14:47:30amiconnDid you try rockbox on her Ondio too?
14:47:37ToniI just wanted to know, wether archos firmware prefers SD-card rather than MMCs, but it doesnīt seem so.
14:48:19ToniNot yet, because she is no technician, but I can ask her to try.
14:48:24amiconnArchos says only MMC is supported, but the SPI protocol is almost identical for SD and MMC, so it works if you can make the card fit
14:49:02amiconnI would be interested in the debug data of her Ondio (Type, hw mask etc)
14:49:23ToniOK, I will ask.
14:50:42 Part Toni
14:54:24***Saving seen data "./dancer.seen"
14:56:25 Quit AciD (Read error: 104 (Connection reset by peer))
14:59:29 Join AciD [0] (
15:20:45 Quit Sebulba03 ("brb")
15:24:05 Join Sebulba03 [0] (
15:35:39 Quit _Schoki ("Client Exiting")
15:37:34 Quit AciD (Read error: 104 (Connection reset by peer))
15:47:49 Join Sebulba02 [0] (
15:50:17 Part Sebulba03
15:58:01 Join webguest32 [0] (
16:00:12 Join AciD [0] (
16:08:46 Quit webguest32 ("CGI:IRC")
16:54:25***Saving seen data "./dancer.seen"
17:28:10 Join thegeek [0] (no@
17:33:20 Quit AciD (Read error: 232 (Connection reset by peer))
17:36:53 Join AciD [0] (
17:38:28 Part thegeek
18:19:59 Quit ashridah (Read error: 110 (Connection timed out))
18:54:28***Saving seen data "./dancer.seen"
19:14:14 Join midk_ [0] (
19:14:44 Quit midk (Nick collision from services.)
19:14:45 Nick midk_ is now known as midk (
19:21:33 Join webguest73 [0] (
19:31:14 Quit webguest73 ("CGI:IRC")
19:34:51 Quit midk ("just STOP it arspy")
20:54:30***Saving seen data "./dancer.seen"
21:37:37 Quit tron_ (Remote closed the connection)
21:41:33 Join [IDC]Dragon [0] (
21:42:12amiconnGood evening Jörg
21:42:22[IDC]Dragonhello again
21:42:38[IDC]Dragonback from shopping, etc.
21:42:56amiconnI just extracted the remainders of backlight (menu entries, settings bits). Commit pending
21:43:18[IDC]Dragonoh, I was about to revive that
21:43:25amiconnUh? Why?
21:43:40[IDC]Dragonbecause the Ondio has a preparation for backlight
21:43:54[IDC]Dragonmaybe the earlier model have it
21:44:40amiconnHmm. It's a config option anyway, so it's simply changing a #define to reenable it
21:44:40[IDC]Dragonthe Archos s/w uses that pin like such
21:45:00[IDC]Dragonafter keypress, it stays on for a while
21:45:45amiconnCould the backlight presence be runtime detected? Perhaps switching the pin to input and check level?
21:45:51[IDC]Dragonwe just have a different kind of backlight (different port bit)
21:46:18[IDC]Dragonit doesn't hurt to just drive it
21:46:48[IDC]Dragondunno if we could detect, perhaps not
21:46:57amiconnOf course driving it doesn't hurt, but I think it'd be a bit confusing to a user to find settings for a non-existing backlight
21:47:49amiconnWe should add that to our "help needed" question: "Does your Ondio have a backlight?"
21:48:25[IDC]Dragonbacklight should be a multi-value config entry, too
21:48:32amiconnSo what should I do now with my changes?
21:49:13[IDC]Dragonthe player has it on a port, Recorders on I2C, Ondio on another port, iriver again different
21:49:43[IDC]Dragoncommit them, I'd say, as you mentioned it's a switch
21:50:48amiconnThe switch was already there, I only #ifdef'ed out some sections in wps-display.c, settings_menu.c, settings.c
21:51:19[IDC]Dragonyes, currently it's HAVE_BACKLIGHT
21:51:46amiconnOk, so I should commit before you change it to a multi-value option
21:52:04[IDC]Dragonyes, please.
21:52:07amiconnDon't forget to account for no beacklight in your choices
21:52:31amiconnJust checking if it works for recorder (i.e. does remove nothing)
21:52:33[IDC]Dragonlike some multi-values, not defined means absent
21:52:51[IDC]DragonI need to reboot
21:52:53amiconnAh ok. The config options should be documented somewhere
21:53:15 Quit [IDC]Dragon ()
22:00:48 Join [IDC]Dragon [0] (
22:00:54amiconnHi again
22:01:08amiconnBacklight changes committed.
22:01:08[IDC]Dragonthat was 2 reboots
22:01:53amiconnI plan to do the same with the disk settings (only setting here is disk spindown, which has no meaning on Ondio)
22:02:38[IDC]Dragonmakes sense, yes
22:02:58amiconnThe keyboard needs to be adapted. I wanted to save a .cfg in order to avoid re-setting all my settings every time the config block changes...
22:03:16[IDC]Dragonyou can, just have to wait
22:03:31[IDC]Dragonit creates one
22:03:50amiconnAh. Still the keyboard would be handy.
22:04:00[IDC]Dragonof course.
22:04:21[IDC]Dragonif only we had keys... ;-)
22:04:29amiconnBefore you are going to implement your file count for recording, please note that this could be handy in other places too
22:04:43[IDC]Dragonlike what?
22:05:16[IDC]Dragonhow is it done now?
22:05:28[IDC]Dragonlike .cfg?
22:05:40[IDC]Dragonthe name generation, I mean
22:05:52amiconnScreenshot works the same way as recording now (prefix + time stamp)
22:06:48amiconnThere is already some code that does the check-and-increment for .cfg, so I think you could just try to make this generic and globally reusable
22:07:15[IDC]Dragonthe .cfg counts and tries to open to check for existence
22:07:28amiconnYes, that's what I thought
22:07:38[IDC]Dragonif it finds a nonexisting, this will be the name to use
22:07:59amiconnScreenshot is a debug menu function anyway, but it would be handy to be able to take several screenshot without overwriting the old ones
22:08:06[IDC]Dragonso it "fills holes", if there should be any
22:08:19[IDC]Dragonfor recording, I plan to do different
22:08:32amiconnWhy/ how?
22:08:41[IDC]DragonI'd like to iterate the directory and determine the highest number
22:08:55[IDC]Dragonthen increment and create that
22:09:15amiconnAh ok. Could be used for .cfg and screenshots too
22:09:27[IDC]Dragonthis way, later recordings will always have higher numbers, even if you erased some inbetween
22:09:49[IDC]Dragonmakes it easier to find it
22:10:12 Quit scott666_ ("i'll be back...eventually...")
22:10:58amiconnI'd suggest to create 2 functions: One that creates these incremental-numbered files with a prefix, and one that creates these prefix + timestamp variant (for platforms with an RTC)
22:11:38amiconnFor recording and screenshot, decide via #define which one to use. Config should always use the incremented version
22:11:40[IDC]Dragonif the latter is worth the outsourcing, yes
22:13:09[IDC]DragonOT: bought a new toy today, a DVD writer
22:13:52 Join scott666_ [0] (
22:15:36amiconn[IDC]Dragon: More settings to remove: Car adapter mode, Button bar
22:15:58[IDC]Dragonyes, how to define that?
22:16:21[IDC]Dragoncar adapter mode should depend on HAVE_BATTERIES
22:16:28scott666_does the ondio have an RTC?
22:16:33amiconnscott666_: no
22:16:33 Nick scott666_ is now known as scott666 (
22:16:45[IDC]Dragonwhich should better be renamed to HAVE_CHARGING
22:17:03[IDC]Dragonbutton bar can depend on the recorder keypad
22:17:13amiconnButton bar is obvious, yes
22:18:32amiconnThe HAVE_BATTERIES had a slightly different meaning when there was neo support. We have to check if undefing it has no unwanted side effects.
22:18:49amiconnThen it could be renamed to HAVE_CHARGING
22:20:20amiconnThe defaults for max. files per dir and max. playlist size should be changed for Ondio too.
22:20:35amiconn(Ok, at least the latter)
22:22:24[IDC]Dragonwho knows what people will do with their GB cards
22:22:58[IDC]Dragonno real need to save RAM, we're not afraid of spinups
22:23:13amiconnI'm merely talking about defaults here.
22:23:52amiconnI noticed that the config struct members are mostly unconditionally defined, so they are there for non-existing features. Dunno if it's worth to make them conditional
22:24:39[IDC]DragonI chatted with LinusN or Zagor about it, they said they intentionally didn't bother
22:25:03amiconnAh ok.
22:26:21amiconnThere is still no #define for ordinary ATA, so I have to use that strange #ifndef HAVE_MMC to get rid of the harddsik menu
22:27:49[IDC]Dragonanother mutivalue item: CONFIG_STORAGE or such
22:28:10amiconn[IDC]Dragon the multivalue guy ;)
22:28:40[IDC]Dragon#ifndef HAVE_MMC is proably fine for now, I don't expect a large variety soon
22:29:01*[IDC]Dragon feels so multivalent
22:31:40amiconnOT: Do you still have some spare flash chips?
22:36:45amiconnMay I send you my Studio for some soldering?
22:37:17[IDC]Dragonplease do the uart boot first
22:37:32amiconnOkay, I'll try that first.
22:37:34[IDC]Dragonbut yes, you may ;-)
22:38:06[IDC]Dragonmobile phone cables are good for level converters
22:42:48amiconnWhat do you mean?
22:43:24[IDC]Dragonthey contain the +/- 12V to 3V level converter which we need
22:43:53[IDC]DragonRS232 adapters
22:44:08amiconnAh. Maybe a self-made converter is cheaper
22:44:31amiconnFirst I need a serial interface anyway
22:45:01[IDC]Dragonsometimes those are found for cheap on a "Grabbeltisch"
22:45:48[IDC]DragonI have a USB to serial adapter, should check itf that works
22:46:54amiconnYes, that would be nice to know before I buy one. I am restricted to use such a beast
22:47:37[IDC]Dragonhang on...
22:54:34***Saving seen data "./dancer.seen"
22:56:18[IDC]Dragonhmm, not exactly working right away
22:58:37amiconnWhat does it do (or not)?
23:00:48[IDC]Dragonit doesn't move at all
23:03:55amiconnHmm, I wonder why. If the serial port is available in windows, everything that doesn't directly bang the hardware should work. Gdb (in turn cygwin) certainly uses system calls, so I wonder what prevents it from working
23:05:05[IDC]Dragonthis is not gdb
23:05:26[IDC]DragonI tried my uart boot, but meanwhile a terminal
23:05:37[IDC]Dragonno go even with that
23:05:53amiconnHmm, Hyperterminal should definitely work
23:06:09[IDC]Dragonso I'd say there's something wrong with the USB -> RS232 cable
23:07:01amiconnI can get one for € 19,-. Do you think the price is okay?
23:07:22[IDC]DragonI think so
23:07:38amiconnButtonbar is gone...
23:10:18[IDC]Dragonwould I bug you if I ask for flashing again?
23:10:47[IDC]Dragonyou can write the Archos s/w in the second flash half
23:11:12[IDC]Dragongiving you back the classic behaviour, even without holding a button
23:11:13amiconnWhat's the first (alternate) image then? I thought it is the archos fw?
23:11:19amiconnAh ok
23:11:23[IDC]Dragonas well, yes
23:11:34[IDC]Dragonyou decide on the second
23:11:41amiconnWhat Rockbox image is in the second image? Does it contain my new driver?
23:12:00[IDC]Dragonnot yet, but I canprepare that
23:12:11[IDC]Dragonor, you flash it afterwards
23:12:19amiconnIs it empty then , or does it contain an older rockbox?
23:12:35[IDC]Dragonfrom 2 days ago, working filesystem
23:13:24amiconnI'd suggest you prepare a new image (although I'll use the current one now), e.g. because of the bugs fixed
23:13:50amiconnLet's hope that I don't mis-flash, with no serial port...
23:13:56[IDC]Dragonfor "real" distribution, I will leave the 2nd empty
23:14:17[IDC]Dragonso it's the user's choice
23:14:44[IDC]Dragonfor now, I have the second in to improve chances that at least one works
23:15:23amiconnWhat are the keys for the bootloader to select the image?
23:15:50[IDC]Dragonback is F1, up F2, right F3
23:16:14[IDC]Dragonso hold back to force image #1
23:16:22amiconnSo you can go "back" to the archos fw - lol
23:16:40[IDC]Dragonyea, that came to me, too
23:17:51[IDC]Dragonwhen you did firmware_flash, you can play your favourite .ucl right after that
23:18:26amiconnSo, now let's hope for the best...
23:19:34amiconnPhew, all went well
23:20:04[IDC]Dragonplayed your latest ucl?
23:20:48[IDC]Dragondo both images start as desired?
23:20:51amiconnArchos fw boots ok as well
23:21:34[IDC]Dragonok, so I can shrink the image to just Archos
23:21:45amiconnRombox hangs...
23:22:00[IDC]Dragonoh, works for me
23:22:29[IDC]Dragonordinary rockbox ucl works?
23:23:46amiconnRombox is not completely hung, it is stuck on the boot logo, but the button poweroff is captured
23:23:52[IDC]Dragonmaybe ROMbox got a wrong start address?
23:24:18amiconnI think you figured the start address from the image I sent you?
23:24:31[IDC]Dragonyes, but I'm human
23:25:06[IDC]Dragonplease compare what the flash plugin tells with the config value
23:25:34amiconnWhere does the flash plugin tell that?
23:25:48[IDC]Dragonwhen you play a ucl
23:26:07[IDC]Dragonit says found image at xx KB or so
23:26:10amiconnAh. It says at 72 KB
23:26:52amiconnso 0x12010 should be correct
23:27:56[IDC]Dragonthen I don't know
23:28:39[IDC]Dragontry flashing the Archos ucl, to get back "classic boot"
23:29:04amiconnI have to create one first
23:29:15[IDC]Dragonno, it's in my webspace
23:30:26 Quit methangas (" HydraIRC -> <- :P")
23:31:26amiconnWorks, and loads rockbox from disk.
23:31:39amiconnI wonder what is wrong with rombox...
23:32:26[IDC]Dragonthis new "classic" boot should be a bit faster then it used to
23:32:57[IDC]Dragonbecause my decompressing bootloader is faster than the Archos' descrambling one
23:33:00amiconnYes it is. The progress bar is no longer there
23:33:33amiconnThe archos bootloader takes *that* long for just descrambling?
23:34:19[IDC]Dragonprobably it's that monitor again, they had a 2 stage boot on recorders
23:35:02[IDC]Dragonthe boot rom descrambles not the image, but a monitor
23:35:38[IDC]Dragonthe monitor waits very briefly for a RS232 command, then descrambles the real image
23:36:24[IDC]Dragonths monitor can in fact be used for download, flashing, and a few other tricks
23:36:39amiconnI think you approach (selecting the monitor via a button) is way better. I hate devices that take long to initialize. Best is instant-on
23:37:13 Join methangas [0] (
23:37:15 Quit methangas (Client Quit)
23:43:24amiconnWhat do you think which thread should monitor the MMC button and generate the SYS_ events? I vote for the mmc thread...
23:43:47[IDC]Dragonor USB?
23:44:30amiconnImho this has nothing to do with USB, although the USB needs those events. Later the filesystem will need them too
23:44:34[IDC]DragonUSB already monitors a switch
23:44:59[IDC]DragonMMC can only do this if it doesn't block anywhere
23:45:17[IDC]Dragondunno i that's the case
23:46:21amiconnThe mmc thread currently blocks on usb, but that could be changed
23:46:50amiconnThe mmc thread currently does nothing useful
23:48:24amiconnFor real ata, the thread blocks on usb in order to generate no unwanted accesses while usb is connected. This is a non-issue with mmc (no idle timeout)
23:50:26[IDC]Dragonif it fits, better place it in the MMC thread
23:53:25 Join gromit``` [0] (
23:56:33amiconn[IDC]Dragon: Did you already do something to the backlight config options? I found a place that is missing an #ifdef - plugin.[ch]
23:57:18amiconnAh no, you provided dummy functions

Previous day | Next day