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 2008-10-12

00:02:20 Quit gp_g (Remote closed the connection)
00:08:03 Quit n1s ()
00:09:58 Join dabujo [0] (
00:10:58 Quit nuonguy ("This computer has gone to sleep")
00:17:32 Quit bluebrother ("leaving")
00:19:58denes_gevaerts: I see there has been some action regarding the crt0.s for s5l8700
00:20:40denes_gevaerts: isn't crt0.s where setting up the memory, the memory protection and copying stuff from rom to the runtime address should happen?
00:21:06denes_gevaerts: because at least setting up the memory is target dependent and not generic on arm
00:21:26denes_gevaerts: afaik
00:22:25 Quit avis ()
00:22:55Lloreanamiconn: Any idea why a RecorderV1 wouldn't boot from an firmware file then?
00:23:16amiconnThat's a bit... vague
00:23:29denes_gevaerts: perhaps I am wrong, and setting up the memory etc. should be done in the bootloader
00:23:59Lloreanamiconn: A rockbox ajbrec.ajz is present, but it's simply booting into the OF.
00:24:15amiconnI assume this is a non-flashed archos?
00:24:40Lloreanamiconn: It's a used one he picked up at a store, no knowledge of previous history.
00:25:09amiconnCheck the versions displayed at boot and in the OF
00:25:36amiconnI suspect that it's non-flashed, and the in-flash OF is rather old
00:26:16LloreanI've asked him for the firmware version number (in the forums). I'm not sure if he's ever coming back, though.
00:26:26amiconnOlder versions sometimes don't pick up an ajbrec.ajz from disk if the rootdir is large/ fragmented or similar
00:27:16amiconnWhat might help is defragmenting the disk. If that doesn't help, reformat and put back ajbrec.ajz as the first thing
00:27:50amiconnAnother reason for an ajbrec.ajz not picked up would be a wrong one (i.e. one made for a different archos)
00:32:38 Join Waldo000000 [0] (
00:32:51 Part Waldo000000
00:35:42 Join blkhawk- [0] (
00:39:29 Quit Tetracomm (Read error: 110 (Connection timed out))
00:40:01 Quit dabujo ("( :: NoNameScript 4.2 :: )")
00:40:05 Join Tetracomm [0] (n=nicholas@
00:41:22 Join reacocard [0] (n=reacocar@WL-358.CINE.HMC.Edu)
00:42:37denes_linuxstb: hi, are you here?
00:47:19 Join MethoS- [0] (
00:50:07 Quit blkhawk (Read error: 110 (Connection timed out))
00:54:03 Part Llorean
00:54:22 Join Llorean [0] (
00:57:26denes_anybody: I don't get it. is the crtS0.s in ipod executed from rom? or is it loaded by some bootloader into ram first?
00:57:55***Saving seen data "./dancer.seen"
00:57:57 Join Waldo000000 [0] (
00:58:13denes_because it can't be in rom, if it allocates the FIQ/IRQ stacks in crt0s itself
00:58:17 Quit robin0800 (Remote closed the connection)
01:00:35 Part Waldo000000
01:04:01 Quit ajonat (Read error: 110 (Connection timed out))
01:04:52 Join ajonat [0] (n=ajonat@
01:05:23amiconndenes_: On ipod none of our code is executed from rom
01:06:59 Quit tvelocity (Read error: 110 (Connection timed out))
01:09:49denes_amiconn: ok
01:10:27denes_amiconn: thanks. and how does the code get into (i)ram then? there is a bootrom that does that?
01:10:32amiconnWe don't touch the rom at all
01:10:53amiconnThe apple loader (in rom) loads the OF from the firmware partition
01:10:59denes_amiconn: i see thanks
01:11:14denes_amiconn: because the thing is, meizu m3 will be executing from rom
01:11:15amiconnWe patch that on-disk firmware and insert our bootloader
01:11:31denes_amiconn: ok thanks, that clears things up
01:11:45amiconn(well, patch by default, we can also replace it, making it boot faster)
01:12:09denes_amiconn: is there a target that runs from rom?
01:12:35amiconnThat disables dual-boot, obviously. ipods are brick-safe due to the fact that the rom also contains a disk mode
01:12:44denes_amiconn: right
01:13:13amiconnIirc no arm target executes our bootloader from rom
01:13:31amiconnThe irivers do, but they're coldfire
01:13:40amiconn(H1x0 and H300 I mean)
01:13:41denes_amiconn: i see
01:14:19 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
01:14:20denes_amiconn: and another question: what is the bootloader supposed to do when everything is running from rom?
01:14:54denes_amiconn: it just loads stuff to (i)ram and hand execution to the firmware main function?
01:15:22denes_amiconn: i mean crt0.s loads stuff into ram and iram, and zeros out bss
01:15:40denes_amiconn: before that of course the ram controller has to be set up
01:15:58denes_amiconn: and then we jump to the firmware entry point
01:16:07denes_amiconn: does that sound right?
01:16:15 Quit bertrik ("Leaving")
01:16:25amiconnThat depends a lot on the actual target. Targets need quite different handling depending on ROM, size, partitioning, method/ possibility to load the OF for dual boot etc
01:17:04aarcanewhat options do I need to compile gcc with to build rockbox successfully?
01:17:17 Quit reacocard (Read error: 110 (Connection timed out))
01:18:28BigBambiaarcane: I'd use the script in tools - it will apply the required patches then build for you
01:18:48amiconnOn all swcodec targets the bootloader loads the main rockbox binary from disk (with the exception of the still unofficial H1x0 flashing)
01:18:58denes_amiconn: i see
01:19:18aarcaneBigBambi, o,.,0 will it install it in a private prefix, or in my main path somewhere?
01:20:10BigBambiaarcane: it puts it in /usr/local by default, but you can of course edit the script if you like
01:20:10denes_amiconn: is there a target, where the whole firmware executes from rom (that's how I am planning to do it on the meizu m3)
01:20:28 Quit domonoky (Read error: 104 (Connection reset by peer))
01:20:52 Join HBK- [0] (
01:20:53BigBambiaarcane: Also, we recommend/need specific versions of gcc depending on the architecture - will get the correct versions for you too
01:21:39amiconndenes_: The archoses can be flashed to load the firmware from ROM. It *could* also be executed directly from ron if it hadn't grown too big meanwhile
01:21:49 Quit Tetracomm ("Leaving")
01:22:08 Join Tetracomm [0] (n=nicholas@
01:22:16amiconnBut this isn't the standard method, and imho it shouldn't be the standard method on any target
01:22:26BigBambidenes_: Why not load from disk?
01:22:46amiconnUpdating rockbox is much easier if you load the rockbox binary from disk.
01:22:54BigBambiIt means to update you don't constantly need to flash
01:23:00BigBambiamiconn beat me to it :)
01:23:13amiconnThe bootloader could of course run directly from rom if possible
01:23:24amiconn(saves boot time)
01:23:53amiconnAnother reason to not execute from ROM would be that the ROM is almost always slower than RAM
01:24:13amiconnThat's okay for a bootloader, but nasty for main rockbox
01:24:28TetracommRockbo does not work on the Ipod Nano, right?
01:24:49BigBambiIt works on the first gen Nano just fine
01:24:55BigBambiAs says
01:25:21denes_amiconn: and on other targets how does the bootloader load the firmware? just copies it to ram and jump to the start?
01:25:23TetracommHow do I tell which generation my Nano is?
01:25:38aarcaneokay, anything I need to know for between running and tools/configure in my build dir ?
01:25:47amiconndenes_: Yes, basically
01:25:53 Quit ajonat (Read error: 110 (Connection timed out))
01:25:59BigBambiTetracomm: The very first link on google for "What ipod do I have" is a hint
01:26:12amiconnThe firmware should depend as little as possible on the bootloader, and set up things itself
01:26:36BigBambiaarcane: At the end will tell you to add things to your path, that's it
01:27:12BigBambiyou will need to run as root incidently
01:27:38 Join druidu [0] (n=root@
01:27:46 Quit gregzx ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
01:28:24amiconnThe main firmware always has a crt0.S as well, which does things like zeroing bss/ibss, copying data (for optional rom execution - right now only applies to archos and H1x0), setting up the main stacks etc
01:28:52druiduany tips for finding which GPIO pins are for each button on a device (iRiver X20)?
01:29:46druiduI know from the datasheet the address of all the registers and also the special function each can have (like external interrupts −− I am guessing they use one of those pins for the power button so it can wake up the device from sleep state)
01:30:01 Quit bmbl (Client Quit)
01:31:18 Quit bughunter2 (Read error: 104 (Connection reset by peer))
01:32:15denes_amiconn: yes, I am looking at the HXX0 linker scripts
01:33:38aarcanetime for me to poof
01:34:14amiconndruidu: Most targets don't use sleep state, but true poweroff. One of the buttons typically doubles as a power switch, and then the CPU needs to enable a "hold power" signal before the users releases the button
01:34:22denes_amiconn: so you are suggesting that the meizu m3's 8 Mbit flash should be used for a 4KB bootloader?
01:34:38amiconnIn the simplest case this "hold power" is a switching transistor that's enabled via a GPIO output
01:35:04amiconndenes_: The bootloader is typically much larger than that, around 40..50KB
01:35:31BigBambidenes_: Given the speed at which Rockbox changes, we really don't want people flashing multiple times per day
01:35:37amiconnIt needs the ata (or sd, or similar) driver, file system driver etc
01:36:17amiconnOf course you can offer flashing main rockbox, but this shouldn't touch our bootloader, and it should be optional
01:36:18denes_BigBambi: which one wears of faster, the nor flash or the mlc nand flash?
01:36:31 Quit HBK (Read error: 110 (Connection timed out))
01:36:39BigBambidenes_: It isn't just that, it is the danger/ease of use too
01:36:53amiconnNor flash can take less cycles, because you cannot do wear leveling on it
01:36:54BigBambidenes_: And I wouldn't have thought eaither wear quickly enough for that to matter
01:37:11amiconnPlus flashing a nor flash is quite slow
01:37:14denes_BigBambi: even if the firmware is not flashed to rom, if you break it on the nand flash, you can't do anything (usb is done from software)
01:37:37amiconnThen the bootloader also needs the usb stack
01:37:45BigBambibootloader usb....
01:37:56BigBambidammit - amiconn beat me again :)
01:39:21denes_just a side note, there is no usb stack yet on the m3
01:39:44amiconnThe typical rockbox boot sequence is that the OF loader loads our bootloader from disk (usually a firmware partition, but may also be a file), and then that bootloader loads main rockbox (from the data partition) or the OF (if the ttarget supports dual boot)
01:39:46BigBambidenes_: If you have a usb enabled bootloader in flash then rockbox is updated on disk, then any temporary glitch in Rockbox can be repaired by the stable tested bootloader in flash. If we are relying on having only main Rockbox in flash, then if that glitches it it causes major problems
01:39:56druiduok, so regardless of that... I can have some visual output (lcd backlight on/off and a led on/off) and also disassembled the original firmware
01:40:08druiduwhat's the best way to detect which gpio is for which button?
01:41:13druiduI tried some kind of bruteforce (setting all the pins to input & enabling pull-up resistors then checking if some gpio pins change state), but didn't get anywhere... some gpio pins kind of fuck everything up when I mess with them
01:41:24denes_BigBambi: i don't understand. you can always use dfu upload - it is stored in the bootrom baked into the chip.
01:41:27amiconnThere are some targets with a different sequence though. The mentioned irivers have no separate OF loader, so we inject our bootloader to an OF, and that patched OF is then flashed using the OF. SO on these, our bootloader is actually the first thing that gets executed. It then either loads rockbox from disk, or jumps to the of
01:41:37BigBambidenes_: That isn't very user friendly
01:41:47amiconnIt can also load rockbox from ROM on the H1x0 (unofficially)
01:41:50druiduI got usb boot to work
01:41:55BigBambiOr rather less userfriendly
01:42:35BigBambiYou have to put it in the mode, use a tool to send a firmware etc. If it is on disk you just plug it into usb, use bootloader usb and copy a new build
01:43:47denes_BigBambi: sounds great. one problem though: nothing of all this has been written. and if the bootlader has to be perfected to 100% utility first, all of this will be written even later.
01:43:52druiduacutally, I'm not at that point yet −− I still have a lot of parameters to figure out before the device could do something useful (make the lcd to work, figure out controls etc)
01:44:01amiconnArchos and the coldfire based iaudios are again different
01:44:08BigBambidenes_: Everything you need for the bootloader, you need for main rockbox
01:44:13BigBambiI don't see the issue.
01:44:31BigBambiThe only difference is that you try it, say it works, and there you go, that is a bootloader release
01:44:50denes_BigBambi: you could already use a half working (but working) rockbox, without usb support for example
01:45:03BigBambiSure, for development
01:45:03amiconndenes_: Tae a look at how our swcodec bootloaders are built. All of the driver code is taken from firmware/, with just a single file representing the apps/ code
01:45:14denes_amiconn: yes, I have seen all that.
01:45:53amiconnIt basically *is* a cut-down rockbox, which can just load rockbox from disk, charge, provide usb (the latter 2 depending on whether the hardware requires it)
01:45:57BigBambidenes_: The only reason we get away with no usb for e.g. ipod is because Apple disk mode still exists
01:46:06druidu<amiconn> In the simplest case this "hold power" is a switching transistor that's enabled via a GPIO output
01:46:12denes_BigBambi: yes, for development. but final release quality is many months away.
01:46:27BigBambiOf course, but that is what you want to work for
01:46:28druiduthis means that if I set all pins to output and write FF everywhere, it should turn off?
01:46:41BigBambidenes_: Why do it one way now, then have to change it all before release?
01:47:04BigBambiWhy not work in the final way now, it'll save time in the long run
01:47:24TetracommDo you have a bot that answers FAQ?
01:47:33BigBambiTetracomm: No, not really
01:47:33 Quit culture (Connection timed out)
01:47:42BigBambiTetracomm: We have a fine manual :)
01:48:04TetracommOk. What can your current bot (if any) do?
01:48:09amiconndruidu: No. It means you need to research, either by reverse engineering the OF, or by trying things and measuring what happens(could be potentially dangerous), or by getting a defective device, stripping all parts, and then use a multimeter to trace the board
01:48:36BigBambiTetracomm: kick people if they flood. Certain people can prompt it to explain stuff I think. It also logs the channel
01:49:03amiconndenes_: Nothing stops you from working on the bootloader and rockbox in parallel
01:49:03BigBambiAlso, certain people can get it to kick people
01:49:15druiduamiconn: well, this is what I want to ask, tips for doing that... already tried a lot of stuff and I'm not really getting anywhere, maybe you guys have experience with other devices
01:49:29denes_BigBambi: why do it one way when it will have to be changed later? because getting rockbox itsself to work first, is much more entertaing - and remember, i am not paid for this, i am planning to do this in my free time.
01:49:37druidufor example, patterns to look for in the OF... I tried searching for some register addresses, but they don't help too much
01:49:40BigBambidenes_: As is everyone else
01:49:55amiconnAlso, as I already mentioned, rockbox should be self contained, meaning you should be able to flash rockbox *instead* of a rockbox bootloader (provided there's enough space in ROM) and it should work
01:49:59BigBambidenes_: But why do all the work if it won't get accepted because of the way you did it?
01:50:21BigBambiSurely the idea is that it becomes a full target and therefore gets the value of everyone else changes too?
01:50:53BigBambiIf you do a port that isn't accepted because of the approach it takes, then it won't become a part of Rockbox as a whole (extreme case)
01:51:20denes_BigBambi: you don't have to threaten me that my work will not be included. all that you can gain by that is me loosing interest (not a tragedy, but still... ;) )
01:51:25amiconnNote that I used 'bootloader' in the sense we use it on the swcodec targets. Archos is a bit different as our system is a bit more modular there
01:51:37BigBambiI'm not threatening at all, I have zero say in this
01:52:03BigBambiI'm just trying to persuade you that doing it in the 'correct' or preferred way would be a good idea
01:52:10denes_amiconn: yes, i understand all you have said. it's true that the bootloader is mostly a subset of rockbox.
01:52:26denes_amiconn: and with a little linker script tinkering, rockbox can run standalone
01:52:35BigBambiAnything you do would be hugely appreciated
01:52:43BigBambiI just don't understand the opposition
01:52:54denes_amiconn: now that's why I wanted to have rockbox run standalone from rom - less work and faster results
01:53:12 Quit faemir (Read error: 54 (Connection reset by peer))
01:53:47denes_amiconn: which of course doesn't mean it can not - or should not - be fixed up later correctly (that again is not much work as we might agrree)
01:54:09amiconnWell, you can develop this way if you want, but it should not become the standard way of running rockbox on your target
01:54:28BigBambiExactly my point
01:54:36amiconnAlthough, I really think that it's more difficult to do it that way
01:55:14denes_amiconn: the question is will it get accepted to svn like that - as a first try at the things.
01:55:19amiconnHaving the bootloader load rockbox from disk makes it so much easier to work on (non-bootloader) rockbox stuff...
01:55:29 Quit Nico_P (Remote closed the connection)
01:55:53denes_amiconn: because if it doesn't then I just work on my own, nothing gets back to the other developers, work gets duplicated and lost, etc.
01:56:10BigBambidenes_: Surely you will still post patches?
01:56:27BigBambiThen at least others could fix it up if they had the desire?
01:56:54denes_BigBambi: well yes.
01:57:30denes_ok guys thanks for clearing some questions up
01:57:30BigBambidenes_: As amiconn says, you are free to develop as you like. However, it would make it much easier all round if you did it in a rockbox way
01:58:05BigBambidenes_: Don't get me wrong, I'm very much looking forward to any and all progress :)
01:58:28denes_BigBambi: look, at the current stage, when nothing is working (except for lcd) it doesn't matter if one works on the bootloader or rockbox itself - same thing.
01:58:35BigBambivery true
01:58:54BigBambibut we were talking about a future stage
01:59:10BigBambicertainly drivers etc is applicable to both
01:59:22denes_BigBambi: but when things already work, but there is no usb stack for example, then things get stalled - as progress could continue on sound etc., but it cannot because a perfect bootloader is the first priority
01:59:34BigBambino, that isn't true
01:59:58BigBambiAnd I din't think anyone said that
02:00:07denes_BigBambi: that's how I understood it.
02:01:26BigBambiWell, to an extent I guess - all I was saying was is that the bootloader should load Rockbox from disk
02:02:05BigBambiSo what we don't want is a full Rockbox as the bootloader
02:02:26BigBambiBut if you want to do that to work on the sound output, then brilliant, let's get sound!
02:02:45BigBambiI'm talking as in eventual design for release etc
02:03:01denes_BigBambi: ok, that is understood.
02:03:28BigBambiIn general, as amiconn said, having Rockbox on disk makes development easier - I guess this all mainly depends on how hard USB is :)
02:04:31denes_so thanks again for answering my questions. so I will start working on the bootloader first - especiall as it doesn't really matter at this stage at all :)
02:04:33BigBambidenes_: I'm very sorry if I came across as negative at all - I've contributed pretty much sod all and I want to encourage, not the other way round :)
02:05:06amiconnAn usb stack already exists, so usb support shouldn't be too difficult. It just needs the driver for the chip
02:05:13 Quit soap ()
02:05:45denes_s/especiall as it doesn't really matter/especiall as it doesn't really matter what I am working on
02:05:50amiconnI think the actual disk support is the biggest problem on many of the new flash-based targets, due to the lack of a hardware flash controller
02:06:03denes_amiconn: possibly
02:06:59denes_BigBambi: ok, at least I understand the idea behind the bootloader and main firmware stuff
02:07:06amiconnWithout disk support there isn't much point in working on sound, so disk support is needed anyway. And if you have that, you basically have the bootloader
02:07:18denes_amiconn: indeed
02:08:12*amiconn *knows* how difficult it is to do development without disk access
02:09:36denes_ok so thanks for the help again - i will hopefully get back to you with some patches soon - good night all
02:12:03 Nick JdGordon|zzz is now known as JdGordon (n=jonno@rockbox/developer/JdGordon)
02:15:35 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
02:20:40 Quit MethoS- (Read error: 104 (Connection reset by peer))
02:22:42 Join Waldo000000 [0] (
02:23:14 Part Waldo000000
02:23:23 Join MethoS- [0] (
02:23:36 Quit ender` (" Sex is dirty - if you do it right. -- Isaac Asimov")
02:31:38TetracommCan I run Rockbox in Qemu?
02:32:43 Quit ompaul (Client Quit)
02:33:11JdGordonTetracomm: no, but it can be run as a simulator so you dont need to
02:34:13TetracommOh, how?
02:34:48JdGordonin the configure script choose simulator instead of normal build
02:35:30 Quit BlakeJohnson86 ("Leaving.")
02:36:56TetracommWill it ask me? Or would I have to edit configure?
02:38:51TetracommI want to try and get Rockbox to work on the Zen, what can I start by doing?
02:38:56TetracommCReative Zen*
02:39:48 Join BlakeJohnson86 [0] (
02:42:09JdGordonTetracomm: a sim build makes a sdl app simulates the UI... if you want to actually get code working on target its not very sueful
02:44:21JdGordonare any of the sansa v2 hackers still around?
02:45:01TetracommAre any Creative player hackers around to help me get started?
02:50:25TetracommOr at least tell me where you left off.
02:51:58 Join funman [0] (
02:51:59 Join Soap [50] (n=Soap@rockbox/staff/soap)
02:52:11funmanJdGordon: you called?
02:52:40JdGordondo any of the v2 hackers have a e200 yet?
02:53:19funmanthe e200 owners are not very active
02:53:49funmanfor testing they are present, but not working on any driver
02:53:58JdGordonok, ill try swapping my v1 for my bro-in-laws v2 so i can help out
02:54:21funmancool :)
02:57:04 Quit Schmogel (Read error: 110 (Connection timed out))
02:57:58***Saving seen data "./dancer.seen"
02:59:11Soapif I buy a v2 by accident when collecting v1s, JdGordon, would you want it?
02:59:42JdGordonI think one is enough :p
02:59:58JdGordonI'm starting to get wierd looks by my family over the amount of mp3 players :)
03:00:05lytewhat is a v2
03:00:38JdGordonfunman: last week of uni now, so ill get into it soonish
03:00:52 Join reacocard [0] (n=reacocar@WL-358.CINE.HMC.Edu)
03:01:54funmannice, I get mad trying to use the SD ;)
03:02:34funmanlyte: something like
03:15:35funmanoh oh it seems the SD answers my commands, but with a wrong crc
03:16:33 Quit fdinel ("Miranda IM! Smaller, Faster, Easier.")
03:20:17 Quit Tetracomm ("Leaving")
03:23:10 Join homielowe [0] (
03:23:20 Quit homielowe (Remote closed the connection)
03:41:14funmanbut I never got the busy bit set in acmd41 response :(
03:45:26funmanit means the card power up procedure never finishes
03:48:13funmanif it answers the commands at all, that must mean it is powered; so maybe not enough ?
03:55:09 Join Thundercloud [0] (
03:55:26funmansaratoga: any news about your jtag wire ?
03:55:30 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
03:58:34 Join Tetracomm [0] (n=nicholas@
04:00:54 Join homielowe [0] (
04:01:56 Quit funman ("leaving")
04:08:11 Quit Thundercloud (Remote closed the connection)
04:14:19 Quit wpyh ("Leaving.")
04:16:01 Nick blkhawk- is now known as blkhawk (
04:16:20 Quit kugel ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092510]")
04:24:58 Join blkhawk- [0] (
04:28:37 Quit thegeek (Read error: 54 (Connection reset by peer))
04:29:58 Quit agaffney (Read error: 104 (Connection reset by peer))
04:30:23 Join agaffney [0] (n=agaffney@gentoo/developer/
04:41:50 Join DaCapn [0] (
04:42:10 Quit blkhawk (Read error: 110 (Connection timed out))
04:42:43 Quit reacocard (Read error: 110 (Connection timed out))
04:42:56 Nick blkhawk- is now known as blkhawk (
04:43:37 Join ddalton [0] (
04:44:08ddaltonwhat is the maximum size sd the sansa e260 can take?
04:44:13ddaltonsd card
04:45:48ddaltonmy other question: can rockbox play video on the e200? if so do I need to convert the video to suit the smaller screen?
04:45:50 Join reacocard [0] (n=reacocar@
04:48:26ddaltonhello? am i getting through?
04:49:00ddaltonso wats da answer to me question?
04:49:07ddaltonsorry the im used to msn chat
04:49:31scorche|shhave a bit of patience...and as for your answers: the maximum rockbox can take i assume? - theoretically much more than exists
04:49:32 Join thegeek [0] (
04:49:51scorche|shas for the second question, yes and you will need to convert it, yes
04:50:18ddaltonok, i thought it said somewhere (not regarding rb the of) that the sansa could only take up to 2 gb?
04:50:34scorche|shthe original firmware can
04:51:05ddaltonso the of can only take up to 2 gb and rockbox can take the highest that is available?
04:51:40ddaltonis that correct?
04:52:27 Quit thegeek (Read error: 104 (Connection reset by peer))
04:52:39 Quit MethoS- (Read error: 104 (Connection reset by peer))
04:52:51scorche|shthe OF cannot read SDHC cards...rockbox can...theoretically, Rockbox supports the same size as defined in the SDHC standard
04:55:28ddaltonwow that sounds complicated ok
04:55:43ddaltonso how much could rockbox take would u estimate at max?
04:56:25scorche|shthe SDHC specifies 2TB as the theoretical limit
04:56:59 Join evilnick [0] (i=60e81b07@gateway/web/ajax/
04:57:47ddaltonand with converting videos wat do I do?
04:57:59***Saving seen data "./dancer.seen"
04:59:49 Join miepchen^schlaf_ [0] (
05:00:10scorche|shthere are instructions on the PluginMpegplayer wiki page
05:00:36ddaltongot a link? or where do I look?
05:01:33ddaltoni think i got it
05:01:55ddalton write?
05:03:52 Join thegeek [0] (
05:07:07 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
05:08:19scorche|shthere isnt any other page named that..
05:18:06ameyerscorche|sh: I actually thought 32 GB was the theoretical max of the microsdhc specification
05:19:35scorche|shameyer: "SDHC uses a different memory addressing method (sector addressing vs byte addressing), thus theoretically reaching a maximum capacity of up to 2 TB (2048 GB). However the SD Card association have artificially defined the maximum limit of SDHC capacity to 32 GB"
05:20:41ameyersimilar to how they fudged sd cards from 1 gb to 4 gb
05:23:10ameyeralthough on a larger scale, apparently
05:38:36 Join mokkurkalve [0] (
05:40:41 Join webguest79 [0] (n=d4fbbf38@gateway/web/cgi-irc/
05:40:49 Quit webguest79 (Client Quit)
05:42:09 Join webguest20 [0] (n=d4fbbf38@gateway/web/cgi-irc/
05:43:01 Part mokkurkalve ("Not the spoon! Oh, no! Not the spooon!!! AAAAAAAAARRRHHH......")
05:46:36 Join avis [0] (n=ident@pdpc/supporter/student/avis)
05:49:12 Join kkurbjun [0] (
05:49:15 Quit Tetracomm ("Leaving")
05:53:00JdGordonhey kkurbjun
06:03:47 Quit XavierGr ()
06:03:59kkurbjunHey JdGordon
06:04:06kkurbjunhow's it going?
06:05:19JdGordonnot too bad, u?
06:05:27JdGordonfeel like trying to get sound working on the mr500 again ? :D
06:07:02kkurbjunpretty well, but got back from a local hockey game. :), I do want to work on it, I've been slowly coding some more. I am about to finalize the gigabeat flash boot code and the mr500 was the next on my list to tackle
06:07:27kkurbjunHave you talked to Cat recently? I wonder if he's still doing anything with it
06:11:23 Quit agaffney (Read error: 54 (Connection reset by peer))
06:12:38 Join agaffney [0] (n=agaffney@gentoo/developer/
06:18:35 Quit evilnick (" ajax IRC Client")
06:21:18 Join Tetracomm [0] (n=nicholas@
06:23:08 Join AndyI [0] (i=AndyI@
06:25:51 Join nicholas__ [0] (n=nicholas@
06:26:07 Quit Tetracomm (Nick collision from services.)
06:26:25 Nick nicholas__ is now known as Tetracomm (n=nicholas@
06:27:58 Quit sarixe ("Ex-Chat")
06:33:43 Join Makuseru [0] (
06:43:53 Quit Makuseru (Remote closed the connection)
06:46:51 Join Makuseru [0] (
06:47:34 Join someone972 [0] (n=97760872@gateway/web/cgi-irc/
06:47:37 Quit agaffney (Read error: 104 (Connection reset by peer))
06:47:41 Join agaffney_ [0] (n=agaffney@gentoo/developer/
06:48:49someone972Anyone here?
06:48:57 Nick agaffney_ is now known as agaffney (n=agaffney@gentoo/developer/
06:49:37 Quit someone972 (Client Quit)
06:57:49 Join evilnick [0] (i=60e81b07@gateway/web/ajax/
06:58:03***Saving seen data "./dancer.seen"
07:06:53 Join Darksair [0] (n=user@
07:11:39 Quit Makuseru ("Konversation terminated!")
07:18:51 Quit Darksair (Read error: 60 (Operation timed out))
07:19:10 Nick HBK- is now known as HBK (
07:37:39 Part toffe82
07:41:13JdGordonkkurbjun: nup, havnt heard from him in ages.... id like to get stuck into it again after exams
07:45:23 Join pierre- [0] (n=pierre@
07:54:58 Quit Tetracomm ("Leaving")
08:18:12 Quit PaulJam (".")
08:26:34 Quit Seed ("cu, Andre")
08:36:47 Quit lyte ("bye")
08:38:04 Join Rob2222 [0] (
08:43:40*JdGordon almost has a new abomination ready for the wps!
08:45:09 Join nplus [0] (n=nplus@141.25.Globcom.Net)
08:48:17 Join Schmogel [0] (
08:51:03 Join J-23 [0] (
08:52:34 Quit Schmogel (Read error: 104 (Connection reset by peer))
08:56:13 Quit Rob2223 (Read error: 110 (Connection timed out))
08:58:00 Quit midkay_ ("Leaving")
08:58:06***Saving seen data "./dancer.seen"
09:00:16 Quit jhulst (Read error: 113 (No route to host))
09:01:15 Quit reacocard (Read error: 110 (Connection timed out))
09:04:49 Join nuonguy [0] (
09:09:18 Quit gevaerts (Nick collision from services.)
09:09:30 Join gevaerts [0] (n=fg@rockbox/developer/gevaerts)
09:10:29*JdGordon thinks we need to talk about upping the wps image buffer size
09:11:42linuxstbJdGordon: What's up your sleeve?
09:12:12 Join midkay [0] (n=midkay@rockbox/developer/midkay)
09:12:31JdGordonbassically I've added a tag so the entire wps can change modes with a button press
09:12:43JdGordonso it could very easily go over the buffer limit
09:14:21linuxstbI don't understand it... If you replace the "go to file browser" button, then how do you go to the file browser?
09:14:36JdGordonyou'd have to go through the menu
09:15:07linuxstbAnd from the description (I haven't looked at the patch), I don't understand how the tag is used in the WPS.
09:15:09JdGordonI doubt the people that will use the tag will complain about that
09:15:40JdGordonreplace one|two|three in the example with the tag to enable viewports
09:15:41linuxstbIt means that a usual button press won't work, depending on your theme.
09:16:28JdGordonif there is a better button to steal then I'll swap, but I dont think there is
09:16:52linuxstbDon't forget "char" is unsigned on arm...
09:18:09JdGordonbah, OK, ill force unsigned and use 255 for "not being used" then
09:18:38JdGordonactually, no I can use 0 for that seen as the conditionals apparently start at 1
09:18:51*amiconn wonders what that should be good for
09:19:15amiconnJdGordon: Btw, don't forget that not all targets even have a "go to browser" button in the wps
09:19:27JdGordonso you dont have to try cramming everything into 1 tiny screen
09:19:30JdGordonoh? which dont
09:19:39 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
09:20:12amiconnAh no, I confused something
09:20:27amiconnThey can browse from the wps, but you can't go directly to the menu
09:21:02amiconnYou need to go through the context menu (not a separate item, but a button press there)
09:21:34JdGordonhmm.. ondios would benefit from this, maybe we change the button for it then?
09:21:59amiconnGuess why the Ondio can't go directly to the menu?
09:22:06JdGordonnot enough buttons :D
09:22:11amiconnThere simply are no buttons left.
09:23:54JdGordonwell, this would only affect people who actually use a theme with the tag, but maybe the last mode can be browse like normal then?
09:23:56amiconnI think that making button behaviour depending on the chosen WPS will be rather confusing (on all targets)
09:24:14JdGordonI dont think so
09:24:39amiconnBesides, did you actually use any of our small-screen targets yet?
09:24:52amiconn(Archos, c200 or similar)
09:25:40JdGordonno, e200 only
09:25:48JdGordondinner.. back in 30
09:25:58amiconnI nnever really felt the need to cramp more into the wps than what fits on those screens
09:26:36amiconnIt might be that someone else would appreciate modes, but I think that going wps->browser is more important functionality than switching those modes
09:27:23*amiconn would like to see some of the viewport and list bugs fixed :\
09:42:15JdGordonwhen they show up on the tracker they can be fixed....
09:46:22 Join dabujo [0] (
09:52:01amiconn9360 would be one..
09:53:15amiconnThere are at least 2 others which I didn't put on the tracker yet, but mentioned several times
09:58:49 Join bmbl [0] (n=Miranda@unaffiliated/bmbl)
10:04:30 Quit nuonguy ("This computer has gone to sleep")
10:07:14*JdGordon thought 9360 was fixed ages ago... probably forgot to commit the fix
10:09:45JdGordonamiconn: you sure thats still a bug? the code is all there to handle it...
10:13:19amiconnHmm, then someone seems to have forgotten to close the task
10:15:06*amiconn just entered fs #9478 on your special request
10:16:27 Quit Horscht ("I got raided by the FBI and all i got is this lousy quit message")
10:17:00amiconnThe simplelist bug seems to be gone now. Weird, since I didn't see the fix for it entering svn.
10:17:19JdGordonwhich was that?
10:17:27amiconnAh, no
10:17:41amiconnNo fs# yet
10:17:48 Quit evilnick (" ajax IRC Client")
10:18:14amiconnI was about to enter it, and first checked whether it's still there. It is, I just remembered wrong where it happens
10:19:20 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
10:23:04 Join gregzx [0] (
10:23:27JdGordonanother minor thing... we need a single place to define the different app/ level events...
10:23:54JdGordonatm there are 3 different users in 3 places and I wouldnt be surprised if they actually overlap already
10:24:01amiconnFS #9479
10:24:35 Quit Dementio (Read error: 104 (Connection reset by peer))
10:25:00JdGordonok, they dont, but new ones may easily accidently overlap
10:25:18JdGordonadding apps/ level defines to firmware/export/events.h feels wrong
10:26:09 Quit jeffdameth (Read error: 110 (Connection timed out))
10:26:32 Join reacocard [0] (n=reacocar@WL-358.CINE.HMC.Edu)
10:26:57 Join jeffdameth [0] (
10:28:56 Join bluebrother [0] (n=dom@rockbox/staff/bluebrother)
10:37:43 Join cyantific [0] (i=c@
10:41:46 Join bertrik [0] (
10:42:21 Join Dementio [0] (n=dementio@
10:58:07***Saving seen data "./dancer.seen"
11:00:51 Quit cyantific ()
11:09:10 Join cyantific [0] (i=c@
11:13:40 Join n1s [0] (n=nils@rockbox/developer/n1s)
11:18:04 Join pixelma [0] (n=Marianne@rockbox/staff/pixelma)
11:18:57 Quit pixelma2 ("-")
11:19:23 Quit n1s ()
11:26:59 Join culture [0] (
11:36:03 Quit amiconn (" l8er")
11:39:27 Join bughunter2 [0] (n=Jelle@
11:42:24 Quit J-23 (Read error: 60 (Operation timed out))
11:44:48 Join J-23 [0] (
11:45:23 Part WarpTI
11:45:52 Join tvelocity [0] (n=tony@
11:50:02 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
11:51:42 Join ender` [0] (
11:52:40 Join fml [0] (n=4fd3f684@gateway/web/cgi-irc/
11:57:28fmlHello developers. Any comments on FS #9457 (Replace multiple usage of the same condition with a locally defined symbol)?
11:59:28*JdGordon saw it and cant think of a reason to not commit it
12:00:23pixelmaJdGordon: I don't understand "* Make %t stricter by aborting if a value is not given. ". I thought that plain %t will also fall back to a default value if no value is given...
12:00:41pixelmabut I'm off to lunch now, later
12:00:50JdGordonpreviously it defaulted to 0 which is meaningless?
12:01:37pixelmahuh, I thought it was 2 seconds or something? At least that's what I remember reading in the description (wiki or so)
12:01:55JdGordonI was going by the code...
12:02:30pixelmamaybe that was lost at some point in time (WPS tokenizer)?
12:02:53JdGordoncould be... just had another look... there is no defalt value handling for it
12:03:05pixelmaI'll try to find out where I read that
12:04:26fmlpixelma: %t without a value should not be used. Either don't use %t at all (which then results in the default 2s) or provide a parameter.
12:05:20 Join Nico_P [0] (
12:06:05 Quit BHSPitLappy (Remote closed the connection)
12:06:17fmlSo I think the commit was ok. %mv is a different thing. It specifies a behaviour and hence should allow to omit the param. In the case of %t the behaviour is specified by the ';', and %t is used to specify a param.
12:19:41gevaertsdenes_: USB shouldn't take too long. It's the same controller as in the tcc chips, and support for that is well underway
12:22:29JdGordongevaerts: you wanna look after fml's patch there? iirc its fiddling with your usb code
12:25:05ddaltonjdgordon: so i see ur not a fan of my idea re accssibility patches in svn?
12:25:47gevaertsJdGordon: it's not really touching usb, it's ata.c :) Anyway, I think it's technically obviously correct, but I still feel too much of a n00b to decide whether that sort of preprocessor-symbol-consolidation should go in
12:26:05 Join lasser [0] (
12:27:06gevaertsddalton: the way I read that thread you're basically asking to have a branch with code that the authors explicitelly didn't feel like making as good as possible
12:29:12ddaltonno, not exactly, blind users in a perfect world would want a lot of stuff that would interfear with sighted people's use of enjoyment of the player eg. wasting a keymap on voice or modifying the quickscreen just for blind people so then the user can choose whether they want full accessibility all they shall do is apply all patches or just some form the accessibility dir otherwise they dont have too...
12:29:39ddaltonalso other people may come across patches that are out of sync and can help fix them eg people interested on working on accessibility
12:29:56ddaltonlike the archoes they are holding voice back i believe
12:30:31gevaerts"I just did my own quick way since it was just for me and I didn't care about coding styles or anything". Sorry, but that sort of thing really isn't helping
12:31:01bluebrotherddalton: why should the archos hold anything back? It can always get ifdef-ed
12:31:44ddaltonyeah i guess so y are massive voice files an issue...
12:32:09*gevaerts tries to translate again
12:32:11gevaertsy == why ?
12:32:49gevaertsPlease use real words. Those shortcuts make your text really hard to read
12:33:08ddaltonand well i just code because i want things and what ever I want wI'll do I'll share it but if you dont like it that is ok, but I can have a big voice and dont have time to push for it to be comitted so I just code and use one big voice file since y waste my time making it do something i dont need only the sighted people need
12:33:30ddaltonsorry im used to msn chat
12:33:53*gevaerts isn't on msn, but he would also complain there
12:34:00denes_gevaerts: sounds good, I didn't know that it was the same controller
12:35:43gevaertsddalton: as long as patches are obviously not fit for commit from a technical point of view, a lot of people are not even going to think about whether the actual feature would be good to have
12:36:48gevaertsAnd people are never going to want to endorse such code in any way, such as by giving it an svn branch
12:37:26ddaltoni gtg but ill discuss this later when i can
12:37:49*bluebrother suggests grouping a17y diffs using a category in the tracker
12:37:55gevaertsdenes_: as soon as vitja commits his code, we can get busy.
12:39:28fmlgevaerts: I'm also an absolut newbe in the ata.c code, but I just saw the repeating (and rather complicated) #if's and thought that it might be worth to do it only once.
12:40:00gevaertsfml: I fully see your point. I just don't know how this is usually handled
12:40:09 Join crazy_bus [0] (
12:40:49crazy_busI'm converting videos using winff and the medibuntu 8.04 package of ffmpeg. But the videos while playing fine in xine are sometimes going out of sync in rockbox. I.e the audio gets ahead so things are out of sync and you end up with a few seconds of just video. Is this a rockbox or encoding problem?
12:41:25 Quit ddalton ("leaving")
12:42:19gevaertscrazy_bus: maybe you're using too high bitrates for rockbox to handle properly. What player is this?
12:42:36crazy_busit's a sansa e200
12:43:30 Quit fml ("CGI:IRC (EOF)")
12:43:59crazy_busthe ffmpeg code it uses is this -acodec mp3 -ab 128 -ar 44100 -vcodec mpeg2video -s 224x176 -b 320kb -strict -1
12:44:48 Join jon-kha [0] (
12:45:50Lloreancrazy_bus: What do you have mpegplayer's frame skip and frame rate limit settings set to?
12:45:58gevaertsYou could try various framskipping options as described on the wiki
12:48:54 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
12:49:25crazy_buswhoops I turned frame skipping off :)
12:51:37denes_gevaerts: I try to get busy with the setup code (setting up the memory controller, etc.)
12:52:23gevaertsdenes_: great! For usb we'll also need working interrupts I think, but we can handle that when we get there
12:58:11***Saving seen data "./dancer.seen"
13:00:05 Quit dabujo ("( :: NoNameScript 4.2 :: )")
13:02:47 Part jon-kha ("[IRSSI] the choice of a GNU generation")
13:03:04 Join jon-kha [0] (
13:04:21denes_gevaerts: I think we have to write the testcode/bootloader into the flash rom, because at least on the m3, it doesn't really fit in the available half of iram
13:04:31denes_gevaerts: and then having interrupts is not a problem
13:05:04denes_gevaerts: and the nor flash should endure many thousand writes, so even the mp3 player will not break
13:05:16gevaertsYes. It shouldn't be too hard to do
13:06:23 Quit crazy_bus (Remote closed the connection)
13:07:09 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
13:14:02 Join faemir [0] (
13:17:06pixelmaI don't get fml's response about the %t completely - does he mean using ";" alone? And I don't see the difference to %mv at the moment, well except that sublines need the "%t" and the ";" and %mv is a single tag
13:19:26gevaertsJdGordon: I haven't looked very hard at FS #9477 yes, but I can imagine that some people would like automatic cycling as well
13:21:26JdGordonok, ill do it if it means being able to get commited :p
13:22:42*gevaerts thinks that this patch is very useful. It made him find out that there is a button to go from the wps straight to the file browser :)
13:22:55*Llorean thinks that WPS tags absolutely should not be able to remap buttons.
13:23:11JdGordonpixelma: I tihnk if you split it with just ; you get the 2s gap, the only reason for the %t is to change that time... it could/(should?) be changed so one of those tags be removed.. but thats a relic from the early wps system
13:23:27JdGordongevaerts: haha, happy to help then
13:23:48JdGordonLlorean: in principle i agree... but i also like this patch and think its an ok exception
13:24:59LloreanJdGordon: Maybe add an option in the settings somewhere that defaults to "No" that lets you pick whether the WPS can take control of that key.
13:25:07LloreanBasically, the tag only works if you say "Yes, it's okay for this tag to work."
13:25:13JdGordonyeah, i tihnk i said that in one comment
13:25:24JdGordonif everyone is happy with that then so am I
13:25:35LloreanI'm not sure I am.
13:25:46LloreanBut at least it's better, since by default WPSes will never change the controls.
13:26:16pixelmaJdGordon: ok, that's the meaning I slowly worked out myself. I don't see how you could remove one of the two tags while still keeping the same functions (being able to set the time)
13:26:18*linuxstb would like to see a useful use of this feature before it's committed (yes, I know it's chicken and egg...)
13:26:23JdGordonalthough, that can easily be defeated by a them .cfg anyway so is probably pointless
13:26:42JdGordonlinuxstb: yeah, I was hoping the themers in the forum would get onto it quickly
13:27:10*gevaerts finds a bug in the manual
13:27:12JdGordonpixelma: the same way %mv is done... i.e ;<time> or just ;
13:27:20LloreanJdGordon: It's not pointless.
13:27:33LloreanWhen we start the new theme site with scripting, we reject themes that change the setting.
13:27:40LloreanSo it's not pointless in official builds with "official" themes.
13:27:42pixelmaJdGordon: and what if you want to start a line with a number?
13:27:45*domonoky just notices, that we forget to distribute the qt accessiility plugin for rbutil mac builds :-/ thats why screenreaders on mac dont work with rbutil :-)
13:27:50gevaerts"Return to the FILE BROWSER" is not really correct, as it goes to the database if you started from there
13:27:55pixelmagevaerts: just one?
13:28:05JdGordonpixelma: umm... good point, which is another reason why it shuoldnt be changed
13:28:28gevaertspixelma: Only one at a time ;)
13:29:04gevaertsAny suggestions for better wording? "Return to the FILE BROWSER to the DATABASE, depending on how you started this playlist" is a bit wordy
13:29:19JdGordonLlorean: well, yes, that would work, but that requires the theme site...
13:29:21Lloreangevaerts: Actually, it's "Return to the screen you launched the WPS from"
13:29:38pixelmagevaerts: return to previous screen?
13:29:43LloreanJdGordon: Yes, but it does mean it's not pointless, so we shouldn't ignore adding that functionality just because the theme site isn't here yet.
13:30:02 Quit tvelocity (Read error: 110 (Connection timed out))
13:30:18JdGordonuntil that happens though we will have the same "issue"
13:30:36JdGordonjust saying
13:30:43LloreanWe tell users not to include that setting in their themes, and hope they do the right thing
13:30:58LloreanRight now users can change *any* setting in their theme, and we don't stop them.
13:31:17JdGordonalso, another reason why its pointless.... themes which use this will probably be very empty if that tag is ignored
13:31:36gevaertsLlorean: not entirely. I came from the main menu, where I chose "Resume playback"
13:31:47Lloreangevaerts: And select didn't send you back to it?
13:32:00LloreanJdGordon: Depends on how they're designed. Doesn't make it pointless.
13:32:08LloreanIt makes people who design themes assuming users will make use of the function stupid.
13:32:48LloreanUsually alternate screens show information you *don't* need to see all the time.
13:33:00LloreanSo you'd still get a basic, useful WPS on-screen all the time.
13:33:15LloreanIf people made the WPS so you had to cycle with select just for it to be useful, it'd be damned annoying.
13:33:28LloreanSo I don't see how it's any argument *against* having the menu option.
13:33:48JdGordonyes, but you would also ONLY use the theme if you wanted that, which means you dont need the browse button
13:33:55LloreanJdGordon: That's a stupid assumption to make.
13:34:06LloreanThere's plenty of themes I'd use without ever using the album art.
13:34:11LloreanOr the "hold-switch to see this information"
13:34:50LloreanI'd avoid a theme *because* I can't make it not take over my controls sooner than I'd ignore a theme simply because it has extra screens I don't need, if I can easily ignore them.
13:35:08JdGordonyes, but we already know most of our user base isnt you...
13:35:20LloreanDoesn't make adding the menu option a bad idea though.
13:35:35LloreanJdGordon: You're basically saying "I wouldn't use the menu option, so it doesn't need to be done."
13:35:50 Quit lasser ("ChatZilla 0.9.83 [Firefox]")
13:35:59JdGordonno... i said themers would change that setting in the theme cfg so there is little point adding it
13:36:00bluebrotherdomonoky: I noticed that rbutil's make clean also tries to delete a core file. Remembered me that a core file is called "core" usually, which wouldn't go nicely with the folder "core".
13:36:06LloreanJdGordon: Well, your patch currently lets the WPS take over the controls and redefine one of them. THAT shouldn't be done. I'm proposing a way you can at least mitigate that extremely negative aspect of it, and you're arguing against it why?
13:36:09JdGordonI've already said im happy to add the option
13:36:25LloreanThe option isn't pointless on the assumption that we'll have the theme site one day.
13:36:43LloreanAnd on the assumption that themers are intelligent people too, and will just tell their users "turn this on for extra features"
13:36:56JdGordonso if the option is added I wont cop a mouthful if its commited?
13:37:11domonokybluebrother: then change the name of the new folder:mke it "base" instead of "core" :-)
13:37:18LloreanIf the option defaults to not allowing the tag to take over control, *I* won't complain. Can't say nobody else won't.
13:37:19bluebrotherI was wondering if there's an alternative.
13:37:26bluebrotherhehe, sounds good. ;-)
13:37:31JdGordonand no.. its more likely themers will get annoyed at users saying "your theme is boring or not working" so they will just disable the setting
13:37:51LloreanJdGordon: Maybe? Who cares.
13:38:02LloreanOur theme site will go up eventually, and all those themes will get rejected until fixed.
13:38:12LloreanDon't forget to document, in the manual, the new options in your patch before committing.
13:38:24JdGordonna, fml will add it :D
13:38:36LloreanThat's being lazy and an ass.
13:38:43JdGordonand sarcatic
13:38:44*bluebrother wonders why we need that much of wps tags
13:38:58JdGordonbluebrother: why do we need any?
13:39:03JdGordonscrew customizability!
13:39:12bluebrothertrue! Do it like apple!
13:39:24JdGordonworse than apple.. they let you disable menu items
13:39:27bluebrothersomeone needs to write rTunes :P
13:39:40bluebrotheror rockTunes?
13:39:48*JdGordon points out that adding the option would make it more than a unnoticable delta
13:40:15bluebrotherbut seriously, there's always the line between customizability and over-customizability.
13:40:17amiconnbluebrother: Make it delete the folder, but not the file 'core'?
13:40:19domonokyrTunes, which renames all your songs to rock-1234.mp3 :-)
13:40:39*pixelma wonders why you would want more info than you can get on one screen?
13:40:40amiconnrm -r core -> rm -r core/*; rmdir core
13:40:42bluebrotherusers repeatedly complain that our menus are overloaded and there are too many options ...
13:40:49Lloreanpixelma: Very small screens.
13:41:04pixelmaLlorean: well, I have an Ondio and a c200...
13:41:17bluebrotheramiconn: that file is generated by qmake, and one problem is when the system tries to write a core file −− that filename is already taken
13:41:22Lloreanpixelma: Some people want large, readable fonts *and* the option to see what's next
13:42:50pixelmaLlorean: how much smaller screens do you mean? I use a slightly larger font on my Ondio (9-Nedore) and think, have quite a lot of info in my WPS, even textual info not much graphical
13:42:53bluebrotherLlorean: have them buy a player with a larger screen?
13:43:20 Quit domonoky (Read error: 104 (Connection reset by peer))
13:43:27pixelmaand with sublines and conditionals you already have some options
13:43:50Lloreanpixelma: Sublines mean it's not within the user's control when they get to see the information.
13:43:54 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
13:43:59LloreanOtherwise sublines would basically do what JdGordon's patch proposed.
13:44:08LloreanIt's basically just manual sublines
13:44:25JdGordonwell, yes
13:45:33bluebrotherwhy do we need this?
13:45:40*bluebrother doesn't see the point
13:46:15 Join globi [0] (
13:46:30pixelmaI am very sceptical about different infos - I remember having a look at the beast's OF (short try with amiconn's) and there was a button that made the UI cycle through different screen with info, it was very confusing - I still don't know what that actually was
13:48:41*JdGordon thinks the e200 OF does this nicely
13:50:48pixelmaI don't know what the e200's OF does but I can't remember anything like this in the c200 OF
13:51:45bluebrotherI played around with a fuze at the local store the other day, and it does something like that. Very confusing
13:52:40 Quit domonoky (Read error: 104 (Connection reset by peer))
13:53:39 Join Schmogel [0] (
13:53:48denes_are there any supported bigendian targets?
13:54:24bluebrotheraren't the coldfire targets big endian?
13:55:23 Join domonoky [0] (n=Domonoky@rockbox/developer/domonoky)
14:01:04denes_ok, thanks
14:02:09 Join moos [0] (
14:02:52bluebrothernot 100% sure though −− can't find a reference right now
14:03:21Lloreanbluebrother: I think it is.
14:03:44*bluebrother found something on wikipedia
14:03:57JdGordonthere is at least a few places where be and le code is #defined in the tree.. so i assume we do
14:04:31bluebrotherm68k is usually be, arm usually le
14:07:11scorche|shthough arm can be both...i dont think m68k can be, so "usually" = "always"
14:12:47amiconnColdfire and SH1 are big endian
14:17:12JdGordonis there a theoritical binsize limit on any targets? (rockbox not rombox)
14:17:52gevaertsJdGordon: ram size?
14:18:14*gevaerts investigated the WPS browse button behaviour
14:19:11 Quit reacocard (".")
14:19:30gevaertsIf you start from the database, it goes back to the database at the place where you left it. If you start from the filebrowser, it goes back there. If you reboot and "resume playback", it goes to the filebrowser, at the root
14:20:37domonokyJdGordon: i think the old archos ones have binsize limits (dont know the actual numbers), on newer targets only ram limits the binsize
14:22:31 Quit JdGordon (Remote closed the connection)
14:22:47 Join JdGordon [0] (
14:25:22JdGordongevaerts: iirc it gets more compliocated if you have follow playlist en/disabled
14:26:03gevaertsJdGordon: so it basically means "Go somewhere, but don't even try to guess where you'll end up"
14:26:18JdGordonsomething like that
14:26:21gevaertsDo people actually use this button?
14:26:43 Join einhirn [0] (
14:27:06*JdGordon likes its randomness :p
14:27:23gevaertsWe have shuffle for that :)
14:29:54*pixelma uses the WPS browse button with follow playlist enabled
14:30:35Lloreangevaerts: It's used by people with Follow Playlist enabled, as pixelma suggested
14:30:43LloreanPeople have even requested that follow playlist work with Database
14:30:46gevaertsThat makes sense, yes
14:31:17gevaertsWithout follow playlist I think the browser button is more or less useless
14:31:41JdGordonand you wouldnt be sad to have it stop working completly on you... :D
14:32:33Lloreangevaerts: I also use it all the time after starting playback, and wanting to the add some more songs in nearby directories.
14:33:15*bluebrother uses the browser button from the wps without having follow playlist enabled
14:33:45pixelmagevaerts: why? I'd still think it's useful to browse music (to add to the playlist etc.). If there wasn't a button for this, you would always have to go via the menu
14:34:04*JdGordon lies... he does actually use it but has no idea what follow playlist is set to
14:34:08*bluebrother would hate it to always have to go through the menu
14:34:23JdGordonthen you wouldnt use the tag
14:34:26bluebrotherusing the menu is way slower
14:34:32JdGordonyes it is
14:34:42bluebrotherwhat if some theme I'm using uses the tag?
14:34:45gevaertspixelma: "more or less useless" is a bit strong, but it's only three button presses
14:35:00Lloreanbluebrother: That's why I like the idea of having an option "Ignore %mo tag" (with some better name) that defaults to "true"
14:35:08 Quit Horscht (Read error: 110 (Connection timed out))
14:35:11JdGordonbluebrother: then you would hope that the themes .cfg doesnt do a nasty and change the setting which lets you disable it
14:35:15LloreanThen the %mo tag does nothing, by default, but you can enable it in the settings and then you get this other functionality.
14:35:28gevaertsJdGordon: the theme site would check that
14:35:37bluebrotherLlorean: I still don't see a reason why we should have such a tag in the first place ... with a similar reasoning we could argue about customizable buttons
14:35:41LloreanSince, once we're in possession of a working Themes site (/me glares in random directions) we can deny themes which change that setting.
14:35:42*JdGordon decides to learn php and get this bloody site up!
14:35:51bertrikJ-23, the FM chip in the sansa clip is the Si4702/03. Probably the same in the e200v2 too.
14:36:03 Join Horscht [0] (
14:36:11pixelmagevaerts: 3 button presses are 3 times of 1 button press... ;)
14:36:17bertrik(but I can't read the exact type number in the photos)
14:36:22Lloreanbluebrother: In the end, all it does is save you a few seconds of waiting for alternating sublines to alternate.
14:36:27LloreanBut it MIGHT also improve battery life.
14:36:34JdGordonbertrik: you want me to open up a e200v2 and check for sure?
14:36:38LloreanSince alternating sublines are much more active than static lines that you can manually switch.
14:36:58JdGordonLlorean: dont kid yourself.... the difference would be bassically nothing
14:37:11 Quit einhirn (Read error: 104 (Connection reset by peer))
14:37:13gevaertspixelma: true, but if you end up in a useless place in the file browser, but need many more button presses anyway
14:37:14bertrikJdGordon, I you offer, yes please. Can you also measure another thing?
14:37:21LloreanJdGordon: Displaying the WPS vs showing the static menu makes a pretty huge difference in battery life.
14:37:35pixelmagevaerts: and e.g. on Ondio you need to go through the WPS context menu to get to the main menu
14:37:35LloreanI bet if you have the whole WPS constantly alternating on a 2s timer, it'd make a measurable hit.
14:37:41JdGordonbertrik: neasure? probably not.. look yes
14:38:05JdGordonLlorean: yes, but comparing a dynamic diplsay with a slightly less dynamic one is not going to show any diff
14:38:11bertrikJdGordon, do you have a DMM to measure if two pads are connected?
14:38:46LloreanJdGordon: If your alternating redraws 75% of the screen (say if you're alternating between large AA and lots of info) I bet it costs something.
14:39:10gevaertspixelma: but there is a button to go to the file browser? I would have made a different choice...
14:39:20bluebrotherLlorean: you really think it would make that much of a difference?
14:39:36Lloreanbluebrother: I think it'd be worth checking.
14:39:58*gevaerts points out that the "mode" button is used for this, so it would make sense to drop that function and use it for JdGordon's mode instead ;)
14:40:11LloreanIf simply displaying even mostly inactive WPSes costs a measurable amount more power than the menu, making subline alternations (especially ones that draw new graphics) manual instead of constant probably makes *some* difference.
14:40:16JdGordongevaerts: on which target?
14:40:20bluebrotherwell, that should be rather easy ... just make one of the MHz guys measure the power consumption of a wps with sublines and one without :)
14:40:26gevaertsJdGordon: ondio
14:40:36pixelmagevaerts: yes, I get along very well with it but got used to for a while...
14:40:46*JdGordon is open to changing the buttons per target
14:40:50pixelmasounds weird
14:41:23LloreanJdGordon: Some targets even have an extra button still (Short-REC, right?)
14:42:10gevaertsAnyway, I don't really like the idea of making this %mo button optional. Either make buttons configurable, or don't
14:42:38Lloreangevaerts: I think WPSes shouldn't change what a button does without warning.
14:43:01LloreanSo either a setting the user is expected to enable that "Allow WPS to change buttons" or similar, or a warning when the WPS is loaded that it will do so.
14:43:05*JdGordon is *really* looking forward to implementing his toushscreen button idea for the wps
14:43:07gevaertsLlorean: I fully agree, but I don't like the configuration workaround either
14:43:30Lloreangevaerts: It's better than nothing, I think.
14:43:35LloreanMaybe just a warning though
14:43:57Llorean"Warning: This WPS changes the function of the SELECT button" when you load it, with a "Select to continue, anything else to cancel"
14:44:07gevaertsI'm not sure. Is this feature useful enough to accept that sort of compromise?
14:44:27LloreanI'd say "no it's not" in my opinion.
14:44:50gevaertsSo it needs more thinking. Maybe there's a better way
14:45:01bertrikJdGordon, don't bother to open your e200v2 now, I think I need to make a list first of stuff to measure / verify (and pictures of what pins to measure exactly)
14:45:06LloreanIt's a feature users have requested in the past (inspired by the Apple and Sandisk OFs) but I think it's mostly a very, very cosmetic feature that doesn't actually allow the user to do anything truly "new" with WPSes
14:45:16JdGordonbertrik: just got the front off :p
14:45:16LloreanYou can already make them more than one screen, it's just automated.
14:45:26JdGordone200+ ver1.5
14:46:25JdGordonjust to rub salt in the wound.... adding this would almost definatly make calls for bigger wps buffers increase
14:46:28bertrikok, can you have a look at the chip if of the FM chip (it's under the display and there's a couple of traces running away from it saying FN_L and FM_R
14:46:34bluebrotherLlorean: uses also want to have video playback completely integrated ...
14:46:35bertrik*chip id
14:46:38LloreanSpeaking of the v2s, has any of the work been tested on (or is expected to work on) the m200s?
14:47:18JdGordonbertrik: under the display? or on the underside of the pcb?
14:47:20Lloreanbluebrother: I still think that could be done with two things: The option to have mpegplayer go straight into the video, and the option for playlists that have had "viewer" associated files manually inserted to call the viewer. :-P
14:48:30JdGordonhmm, no SD daughter board on the v2s
14:48:34*gevaerts wants the playlist/viewer one
14:48:42bertrikJdGordon, the chip in the bottom left of this picture:
14:49:11bluebrotherLlorean: well, even if it's doable (it definitely is) the main question is if such a request is feasible in the current state
14:49:31Lloreanbluebrother: Well, if done like that, I'd *imagine* it'd have minimal to no negative impact.
14:50:52bluebrothertrue. My point is just that people request all weird kind of things regardless if it makes sense or not. And we never cared about the OFs ...
14:50:54JdGordonbertrik: ok the chip says "0216 b4d2 741"
14:51:36Lloreanbluebrother: True.
14:52:00LloreanIt *might* be nice with "Multifont" in the future. You could have one "page" be small and another "large" so you could very, very easily switch to the "large" page in the car or something.
14:52:04JdGordonbertrik: I can try getting highish res scans tomorow if needed?
14:52:13LloreanBut that can also be done with just loading a .cfg as you get into the car.
14:52:30LloreanSo again, no real new functionality, just a different way of approaching the same stuff
14:52:45JdGordonnot sure why thats a negative
14:52:46gevaertsFor that case the context menu is also more than sufficient
14:52:56JdGordonsure if the delat was huge then id say its not worth it.. but its tiny
14:53:47bertrikJdGordon, ah the chip id looks vaguely similar to the one in the clip
14:54:06bluebrotherthe delta is not the only thing to consider −− don't forget what expectations for future development this will trigger at users
14:54:33JdGordonfuture users arnt something to worry about... current ones are
14:54:46*JdGordon misread
14:55:34JdGordonbertrik: are the 8 solder pads under the lcd the jtag connection?
14:55:36bertrikJdGordon, photos of higher resolution than the ones currently linked at the sansae200v2 wiki page are always welcome I think
14:57:27bertrikprobably pinout of jtag is at I don't know if anyone actually connected successfully to JTAG yet
14:58:13***Saving seen data "./dancer.seen"
14:59:15JdGordonany ideas why they bothered adding a backup battery to the v2s?
14:59:23bertrikthe sansav2 thread is getting quite long now and it's a lot of work to comb through it :)
14:59:25*JdGordon doesnt remember one in the v1
15:00:15*bertrik vaguely remembers a tiny button cell in either his c200 or e200
15:00:25JdGordonscorche: is the theme site code accesable anywhere?
15:00:43bertrikI suppose it runs the RTC even when the main battery is disconnected
15:01:46JdGordonscorche|sh: svn where?
15:02:16 Join fdinel [0] (
15:02:21JdGordonah, not under trunk
15:02:27*JdGordon co's
15:04:24*JdGordon sees alot less code than he was expteing :p
15:10:36JdGordonscorche|sh: is the README current?
15:10:44 Join noaftschoarer [0] (
15:11:41 Quit Horscht ("Snak 5.3.3 Unregistered copy. Evaluation period is over. Program will now quit. Thanks for using Snak.")
15:12:08 Part noaftschoarer
15:12:59scorche|shJdGordon: no clue...i havent really touched the svn code
15:13:47JdGordonsays it needs a <yourname>.ini file?
15:13:56JdGordonfiddling with ini.php isnt enough?
15:14:04 Join PaulJam [0] (
15:15:38JdGordonthe email regex is wrong... isnt accepted
15:17:06 Quit webguest20 ("CGI:IRC (EOF)")
15:23:54 Nick fxb__ is now known as fxb (
15:27:08*gevaerts is very surprised when he sees actual correct email regexes
15:28:58*bertrik has been unsuccessful so far in finding out more about the as3525 i2c master controller
15:29:49Nico_Pthe "real" email regex is awefully long and complex
15:29:51bertrikif SDA and SCL were GPIOs we could just bitbang them, but that's not the case unfortunately...
15:33:51 Join gregzx_ [0] (
15:36:45J-23How can I update bootloader on c240? I runned configure and selected bootloader build, now I have firmware.mi4 file. What should I do with it?
15:37:03 Quit gregzx (Nick collision from services.)
15:37:10 Nick gregzx_ is now known as gregzx (
15:37:27domonokyuse sansapatcher with the bootloader file.. (with the correct commandline options)
15:39:39bertrik-a IIRC
15:40:00J-23I runned it with −−install, -of and -a :/
15:40:56bertriklas time I did it, I first uninstalled the current bootloader, then installed it with -a (as root)
15:41:30 Join DerDome [0] (
15:47:03 Join Seed [0] (
15:47:38*BigBambi doesn't think letting themes change buttons is a good idea, even with a menu option. We already have people that change settings without realising (e.g. file vew) - imagine if they changed this by mistake and suddenly select didn't do what they thought
15:49:05 Join barrywardell [0] (n=barrywar@
15:49:07J-23What should I do now?
15:49:09JdGordonselect changing really isnt a massive deal though
15:49:24BigBambiIt is customisable buttons under a different banner
15:49:57BigBambiIf themes can customise a button, why can't other things, why aren't all buttons customisable, .... etc
15:51:39bertrikJ-23, have you ever restored an original firmware in recovery mode?
15:52:03 Quit kachna|lappy ("Konversation terminated!")
15:52:17gevaertsJ-23: if you run sansapatcher -h, you'll get lots of helpful hints
15:52:33*amiconn isn't sure how to handle a re-release of 3.0 (i.e. 3.0.1) which will be necessary for the player if we want up-to-date flash images
15:53:04J-23bertrik: yes, I wanted to reformat my player, because fdisk was reporting that my player has bad partition table
15:53:18bertrikno, don't reformat
15:53:33JdGordonamiconn: whats the issue? just change the build... all 3.0.x builds can be said to be 3.0 release
15:53:45gevaertsJ-23: reformatting won't fix a partition table anyway
15:53:46 Join kachna [0] (
15:54:00gevaertsAnd what exactly does fdisk dislike about it?
15:54:13amiconnJdGordon: It would need to be placed on the download servers
15:54:30J-23"Partition {1,2} does not end on cylinder boundary."
15:54:51gevaertsJ-23: that's not "bad"
15:54:53 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
15:55:12JdGordonamiconn: so? Bagder has the powerr!
15:55:57J-23ok, so what should I do to reinstall the OF?
15:57:08*domonoky needs blind/visually impaired users with macs to test if this rbutil version (FS #9483 , there is also a test binary available) helps accessibility.
15:57:23bertrikreinstalling the OF through recovery mode, then reinstalling rockbox would be what I would do, but maybe others disagree
15:57:52bertrikJ-23, how did you get the c200 in this situation?
15:58:10 Join Acky [0] (
15:58:33JdGordondomonoky: what actually needs testing? I can try using it with the display turned off if thats whats needed?
15:58:41J-23I don't know. I remember some time ago I installed bootloader with RButil.
15:58:57domonokyJdGordon: i need testing with a macosx screenreader...
15:59:44gevaertsJ-23: as I said five minites ago, run sansapatcher -h
15:59:56domonokyi have now included the qt accessibility plugins, so it *should* work with a screenreader now, but i am not sure if i did it correctly..
16:00:20J-23gevaerts: "-of −−update-original-firmware filename.mi4" - is that information I need?
16:01:27JdGordondomonoky: is the "
16:01:40JdGordon"voiceover" in universal access enough to test?
16:03:07domonokyJdGordon: i dont really know what screenreaders are common on mac, but a test with voiceover would surely be nice, but dont know if it is enough...
16:03:12*JdGordon reaslises vnc doesnt forward sound and decides to leave this for tomorow
16:04:39JdGordondomonoky: because I have such painfully slow upload its best to upload it somewhere else first if you have faster hosting
16:04:48JdGordonseen as its pretty big
16:04:54domonokyJdGordon: will do.. :-)
16:05:32JdGordonill send this one to my uni host if you dont
16:06:05domonokyi am just downloading it, and pushing it to another sever
16:06:23JdGordonok great
16:06:29JdGordonwhat speed you getting?
16:06:36domonoky~ 5k
16:07:02JdGordonmuch worse than it should be
16:09:05J-23what files from 7zip archive (downloaded from should I copy to 16MB disk?
16:09:22J-23pribootLoader.rom and firmware.mi4?
16:11:45 Quit Acksaw (Connection timed out)
16:14:32 Join tvelocity [0] (
16:16:23 Join Tetracomm [0] (n=nicholas@
16:16:40J-23hmm, I reinstalled properly, but I don't see USB mode selector in Options menu.
16:29:31 Quit kkurbjun ("Lost terminal")
16:38:59 Quit ender` (" We live in an age when pizza gets to your home before the police. -- Jeff Marder")
16:41:27 Join funman [0] (
16:44:53 Join ender` [0] (
16:53:10 Join XavierGr [0] (n=xavier@rockbox/staff/XavierGr)
16:58:14***Saving seen data "./dancer.seen"
17:01:52bertrikJ-23, nice to hear you were able to restore it. The 1.01.06 indeed has no option for USB in the menu. There is a 1.01.07 version however that does it (it's not on the zefie site)
17:07:01bertrikhi funman, no luck yet with the i2c controller, I can only see some code to set the clocks and slave address, but not how to actually send/receive data over i2c
17:07:20J-23I downgraded firmware to version recommended in SansaFAQ on wiki, installed Rockbox and patched it. Everything works.
17:09:33bertrikfunman, I wrote a little bit of code for the clip that configures all as3525 GPIO pins as GPIO, switches them as input and read the values, so I can tell the bootloader settings
17:09:44funmanbertrik: are you reverse engineering the OF ? is there some useful information about i2c in the datasheet ?
17:10:20funmanwhich information do you want to get from the values of GPIO pins ?
17:10:47 Join n1s [0] (n=nils@rockbox/developer/n1s)
17:10:50bertrikfunman, the datasheet has very little about i2c, it shows which registers are where but does not explain them any further.
17:11:33bertrikI looked if the register locations / order was similar to the pp i2c registers but that is not the case. Can't find anything in the OF either.
17:12:13*amiconn summons one of the Swedes
17:13:06 Quit nplus (Remote closed the connection)
17:13:15bertrikAs far as I understand from the datasheet, some of the GPIO pins on port A and C determine how the clip boots and I thought this would be interesting to know
17:13:53 Join nplus [0] (n=nplus@141.25.Globcom.Net)
17:23:10denes_gevaerts: could you test FS #9484
17:23:20 Quit druidu ("My damn controlling terminal disappeared!")
17:23:31denes_gevaerts: the lcd clock was 4 times that in the OF
17:23:46denes_gevaerts: so maybe now your lcd will work, even at 200MHz cpu clock
17:24:33 Quit Horscht ("IRC is just multiplayer notepad")
17:37:55funmanbertrik: I think they would be useful if we could control them, but for now we can only make the AS3525 bootloader load the firmware from the NAND
17:38:51n1sJdGordon: do you have an intel mac and an ipod?
17:39:23bertrikok I know, but it never hurts to investigate. I wonder if we can find a way to access the USB promer without having to short-circuit the flash
17:40:07n1s...or anyone else for that matter who can test the new ipodpatcher on an intel mac?
17:43:47 Join Horscht [0] (n=Horscht@xbmc/user/horscht)
17:44:16 Join shotofadds [0] (n=rob@rockbox/developer/shotofadds)
17:46:34domonokybluebrother: new autodetection for rbutil (FS #9485) :-) (not completly ready, but successfully finds 5 devices on my PC :-) )
18:05:28linuxstbfunman: Thanks for posting a new patch - I'm looking at it now....
18:07:43 Join shot0fadds [0] (n=rob@rockbox/developer/shotofadds)
18:08:53bluebrotherdomonoky: will check it later ...
18:09:07 Join MethoS [0] (
18:11:01funmanlinuxstb: cool :) on my side I'll fix a very stupid error in my SD code and try to run it again
18:12:14linuxstbfunman: I noticed you've put "??" as the (C) holder in the dummy .h files (which is fine IMO), but button-target.h looks complete - who should I add as (C) holder there?
18:12:30funmanFrancois Dinel
18:13:04linuxstbHe's atomikpunk I assume?
18:13:38funmanmy stupid error had no effect on the code (at this moment), the ready bit still isn't set in the acmd41 response :/
18:13:47 Quit Twst (Read error: 110 (Connection timed out))
18:14:31 Quit shotofadds (Read error: 60 (Operation timed out))
18:15:07linuxstbfunman: I tried that patch for dualboot.S and I still can't get it to boot with the home button - does it work for you?
18:15:44funmanI didn't try your mkamsboot but the button check is the same
18:15:59funmanand yes the home button is fine here
18:16:03 Join fragilematter [0] (n=fragilem@
18:16:22bertrikthe home button works here (just tried it), but I remember it didn't work the first few times I tried
18:17:26 Quit avis (Read error: 54 (Connection reset by peer))
18:17:52 Join MethoS- [0] (
18:17:53linuxstbbertrik: With the svn mkamsboot?
18:18:24funmanlinuxstb: can you try other buttons ? (reading button-clip.c of the patch)
18:18:40linuxstbfunman: Maybe later - I want to get your patch committed first...
18:19:07funmansince A3 (hold) works, we can delay that bug without fearing bricking devices
18:20:55 Join avis [0] (n=ident@pdpc/supporter/student/avis)
18:21:02gevaertsdenes_: that seems to work. Shall I commit it, or do you still want to do more changes first?
18:22:06bertriklinuxstb, no I only ran code from git so far
18:25:18denes_gevaerts: great. no I don't plan on other changes for the lcd driver for now.
18:25:38denes_gevaerts: the bootloader will change, but that will also be a bit later
18:25:58denes_gevaerts: first I must hack together the and crt0.s
18:26:12gevaertsWell, yes. The bootloader currently is more of a demo playground
18:26:29gevaertsLCD is pretty fast now
18:26:37denes_gevaerts: btw there is one thing: crt0.s will contain s5l8700 specific code. so imho it wasn't such a good idea to have the same crt0.s for several different targets
18:27:22linuxstbfunman: What about bootloader/sansa_as3525.c - there is no (C) header at all. I'm going to add the standard Rockbox header, but with which name?
18:28:09denes_gevaerts: what will be target specific in crt0.s is - as i see it now - setting some clocks, and setting up the iram timings.
18:28:19denes_gevaerts: and some other stuff
18:28:22denes_gevaerts: too
18:28:30funmanlinuxstb: well add mine if you want, it's not a very big piece of work
18:28:41*gevaerts lets denes_ and linuxstb fight this out
18:29:11funmandenes_: can't that be done in the system-xx.c file ?
18:29:41linuxstbdenes_: Where were you last night? ;) If the crt0.S starts to get too many target-specific #ifdefs, then it can be easily split.
18:30:06denes_so I am not sure how to do it best.
18:30:37denes_in the OF they set 2 gpio pins (one is power hold, the other is unknown). then they set the cloks and setup the iram (sram) timings
18:31:06denes_now i am guessing setting the sram timings might be needed before copying stuff to iram
18:31:34denes_and copying stuff to iram is done by crt0.s by definition - so the sram timings and clocks have to be set there too
18:31:49denes_that's the way I think of it
18:33:06denes_also setting/flushing cache and setting up the protection (maybe that's not needed, well in that case it can be removed later) might be target specific
18:35:06denes_these are things which are/might be target specific, ans should perhaps be done in crt0.s imho
18:35:14denes_what do you guys say?
18:36:24linuxstbdenes_: Yes, sounds like you need to copy it back to the s5l8700 directory...
18:37:42gevaertsIt depends. How much code are we talking about? Maybe the generit crt0.S can call a target-specific block that's in the target/cpu directory
18:39:12 Join massiveH [0] (
18:39:13denes_setting the gpio pins, setting the clock, setting the sram timing, setting up/flushing cache and setting up protection
18:39:32 Quit Tetracomm (Read error: 110 (Connection timed out))
18:39:32denes_these might be target specific, and should be done before copying data segments
18:39:36 Quit MethoS (Success)
18:39:56 Join Tetracomm [0] (n=nicholas@
18:40:11gevaertsI'm not convinced about the gpio pins being needed that early
18:40:25gevaertsRAM timings and clock, yes
18:41:45denes_gevaerts: well that's the first thing the OF does - this is a fact.
18:42:08gevaertsYes, but that doesn't mean it's the only way to do it
18:42:22denes_also setting up the protection and cache might be arm940t specific. at least I have no idea what kind of mpu other cores have, if any.
18:42:31denes_gevaerts: yes
18:43:25funmanwhat about another assembly file, with a function called by crt0.S before main ?
18:43:32 Quit perrikwp (" ajax IRC Client")
18:43:33gevaertsAt least the power hold can be done later. I believe it's usually done in the C part in rockbox (but I'll let someone else confirm that)
18:45:46 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
18:45:53denes_funman: the shared part would be like 2 loops, copying data to ram, and zeroing out bss. that much can be duplicated imho
18:46:30denes_because although the protection might be the same on other arm cores - but the memory layout could be different. and if the memory layout is different, setting up the protection is different also.
18:47:04linuxstbfunman: FS #9467 committed.
18:47:11funmancool :)
18:47:38denes_anyhow, imho the two targets should be split, and if it turns out that everything is the same, then they can be consolidated into one file. so i prefer the other way around as it is done now.
18:48:57linuxstbdenes_: I would prefer the other way - start with a unified file, and split if/when needed. Otherwise the consolidation will never happen...
18:50:22 Quit barrywardell ()
18:50:48denes_linuxstb: i won't mess with ifdef's , especially because I have no idea about other targets - so i assume everything will work on other targets.
18:51:14denes_linuxstb: but this is unlikely
18:51:50denes_linuxstb: but what else could I do
18:52:11 Join Nazdaq [0] (
18:53:18NazdaqHow do i know exactly which ipod i have? i think that i have the 3rd gen nano but i am not sure
18:53:57bluebrotherNazdaq: the IpodFAQ wiki page links an apple article that tells how to distinguish the ipod models
18:54:18NazdaqGreat, will have a look
18:54:35bluebrotherplus, Rockbox Utility can detect your Ipod −− the only thing it doesn't do is tell you detailled which it is if your Ipod is unsupported
18:55:05bluebrotherthe nano 3G is the "fat" one
18:55:12Nazdaqshort and fat
18:55:12denes_funman: what kind of arm core do you have in your player?
18:55:28linuxstbdenes_: If you're adding code that you know is specific to either your CPU, or your target, then simply surround it by that #ifdef. But I think it will become clear once you start adding code whether it needs to be split or not.
18:55:31bluebrothersize doesn't matter for this.
18:55:34funmandenes_: arm922tdmi
18:56:49 Join toffe82 [0] (
18:56:55denes_linuxstb: that means I have to add an ifdef around the whole file minus 10 lines. because I think that everything I listed above is cpu/memory layout specific.
18:57:11denes_linuxstb: even if I move the gpio part to the .c file
18:58:16denes_linuxstb: i mean that's what I think. I don't know because I am not familiar with other targets.
18:58:17***Saving seen data "./dancer.seen"
18:58:25NazdaqI think i will have to do some more reading
19:00:06linuxstbdenes_: Just write the patch first, and then we can decide what to do with the file.
19:00:28Nazdaqyep, just says ipod, maybe the 3rd gen nano isn't supported then?
19:01:06linuxstbNazdaq: See (at the very top)
19:01:10bluebrotherjust says Ipod? It should bring a popup telling that an unsupported Ipod was found.
19:01:37Nazdaqdidn't say anything at all
19:01:49bluebrotherwell, did it select any model in the list?
19:02:09bluebrotherreminds me, does someone know the USB PID of the Nano 4G?
19:02:34Nazdaqsays invalid, well thats ashame
19:04:20 Join petur [50] (n=petur@rockbox/developer/petur)
19:04:47denes_linuxstb: ok
19:06:07 Part fragilematter
19:06:23NazdaqThanks anyways bluebrother, guess i'll stop by from time to time to see if there is anything new available
19:07:56linuxstbfunman: In the LCD driver, I think we can just replace the "xoffset" variable with a constant (#define XOFFSET 2) - IIUC it will be the same when the display is flipped because the framebuffer is "centre-aligned"?
19:09:31funmanlinuxstb: yes it's center aligned, but I didn't try flipping though, let me give it a try
19:11:44 Quit ompaul (Client Quit)
19:12:06amiconnlinuxstb: Shouldn't the folder be called firmware/target/arm/as3525/sansa-clip ?
19:12:49funmanmodelname should also be changed in tools/configure if you want to do that
19:17:16linuxstbamiconn: Hmm,. I guess it should. I'll rename.
19:17:59linuxstbfunman: The other sansas just use the last part of the name "e200" etc for modelname
19:18:02bertrikI'm thinking about reorganizing SansaV2HardwareMappings a bit, by putting the GPIO first and indicating what button/led it represents
19:18:31funmanlinuxstb: but $modelname is used to find this directory
19:19:03linuxstbfunman: I think that's t_model
19:19:16linuxstbBut yes - that should change to sansa-clip
19:19:25funmanhm ok
19:19:47linuxstbThere are far too many similar looking variables like that...
19:20:59linuxstbHmm, seems the c100 and m200 ports also omitted the "sansa-" parts.
19:21:12funmanIf I use lcd_drawrect(0,0,128,64); I can see the whole rectangle on the screen, so it's definitely centered
19:21:21funmanI could find vertical flip, but no horizontal flip
19:21:39 Join BHSPitLappy [0] (n=BHSPitLa@unaffiliated/bhspitmonkey)
19:22:33 Join Twst [0] (
19:22:40*linuxstb wonders why "svn mv clip sansa-clip" didn't work as expected...
19:22:46amiconnfunman: The flip flips hor+vert, i.e. it is actually a rotation by 180
19:23:08amiconnAnd since it's centered, xoffset will indeed be fixed
19:23:23funmanah ok
19:24:10funmanI'll look at adding functions for flipping, maybe also scrolling, as well as the button led
19:24:35funmanamiconn: you know a bit SD cards, don't you ?
19:24:57funmanwhat could mean a card not setting the ready bit in the acmd41 response ?
19:25:02gevaertsdenes_: are the colours correct on the new m3?
19:25:14funmanIt answers forever 0xFF8000 (all voltages supported, still in power up phase)
19:25:56amiconnNot sure. I'm far more familiar with MMC than SD
19:26:43n1sfunman: i think preglow had a similar issue with the d2 but i don't know if/how he fixed it...
19:27:32 Join nuonguy [0] (
19:27:35 Quit Nazdaq ()
19:31:50bertrikfunman, you already sent a send_if_cmd at that point right?
19:32:54funmanwhat's _if_ ? I'm at the send_op_cond stade
19:34:37bertrikas far as I understand the spec, you first send a go_idle_state until it indicates readiness, then a send_if_cond and check the response and only then a send_op_cond
19:34:41funmann1s: it seems the d2 only use nand (according to firmware/SOURCES)
19:34:42bertrikif = interface I think
19:34:53funmanbertrik: ah, let me check that then
19:35:44n1sfunman: I'm not sure if he committed anything to svn but though it might be worth it to contact him and see if he found something interesting out
19:37:04bertrikoh, that is for the SPI mode initialisation sequence
19:38:18denes_gevaerts: no
19:39:02shot0faddslinuxstb: I've finally figured out the weird nand_identify problem - tcctool doesn't send the last part-packet of the firmware file!
19:39:08shot0faddsline 137 needs to be <= not <
19:39:08denes_gevaerts: just a guess, but maybe it's a problem with endiannes?
19:39:15 Join mcuelenaere_ [0] (n=mcuelena@rockbox/developer/mcuelenaere)
19:39:17bertrikjust checked: it is mandatory to send cmd8 (send_if_cond) for the sd mode too for physical spec version 2.00
19:39:59shot0faddsunfortunately I can't build a fixed tcctool right now :(
19:40:21 Quit tvelocity (Read error: 110 (Connection timed out))
19:40:24shot0faddsthe VMware image is too outdated, and since my hard disk crash I don't have Visual Studio installed so can't build for win32
19:40:40 Join tvelocity [0] (n=tony@
19:42:35funmanbertrik: it is mandatory for SDHC
19:42:56linuxstbshot0fadds: I'll build one now - nice catch. Have you committed the fix?
19:42:58funmanand this makes the card know that the host supports specification v2.00
19:43:02 Quit pierre- (Read error: 104 (Connection reset by peer))
19:43:17linuxstbshot0fadds: I take it you need a win32 binary?
19:43:31denes_funman: busy high = not busy imho
19:44:13denes_funman: busy low = busy
19:44:28shot0faddslinuxstb: I haven't committed it, I was waiting to confirm it fixed the problem. win32 is best for me, thanks.
19:44:40shot0faddsthat was *NOT* a nice bug to track down :p
19:44:51*linuxstb hangs head in shame
19:45:17linuxstbI'll just make the change locally and build it for you.
19:45:50shot0faddsthanks. I'm downloading Visual Studio atm, but that could take a while...
19:46:44*shot0fadds wonders why that bug hasn't shown up til now
19:47:01linuxstbHmm, are you sure that's the right fix? Shouldn't the data be padded instead?
19:47:59shot0faddspadding's probably safest, but is there really a practical difference?
19:48:13linuxstbIt will be reading outside the malloc'd memory
19:49:02shot0faddshmm. yes, you're right - I should have spotted that too ;-)
19:49:13pixelma...and another change to docs/CREDITS just to annoy me... ;\
19:49:18shot0faddswhat's the best way to pad the binary?
19:49:48linuxstbMy first thought is some kind of "PAD" macro, which I apply to "len" in the malloc, and also in the division in the for loop
19:50:21linuxstbOr maybe "len" should just be modified to round it up to a multiple of 64
19:50:26funmandenes_: yes right, but the bit is still low
19:51:55funman'card returns busy when : - still in initialization process, or - card is SDHC and cmd8 hasn't been issued'
19:51:58linuxstbshot0fadds: I think I'll do the second option - create a new "padded_len" variable.
19:52:25 Join wpyh [0] (n=william@
19:52:41 Quit massiveH (Read error: 104 (Connection reset by peer))
19:53:08shot0faddslinuxstb: yes - malloc(padded_len), read(len), do_patching(padded_len) works for me.
19:53:36linuxstbshot0fadds: Exactly what I just wrote ;)
19:54:12gevaertsdenes_: found the colour problem. LCD_WDATA is right-aligned, so those shifts are not needed
19:55:36funmanpixelma: why are you annoyed?
19:56:27 Quit mcuelenaere (Read error: 113 (No route to host))
19:57:59pixelmawasn't very serious - it's because I have a local change to it and will get a conflict on SVN up, just my bad luck that it changes this frequently lately... :)
19:58:37denes_funman: maybe the endiannes is not what you expect, and the card is in fact not busy. maybe that "8" in ff8000 is bit 31.
19:58:52funmanhu ??
19:59:23funmanI don't think bytes get reordered with weird endianess, and if I use little endian for other answers, the bits match correctly
19:59:28denes_funman: are you reading the buffer in 16 bit quantities or in 32 bit quantities
19:59:58denes_funman: from the controller?
20:00:02funmanthe bits set in 0xFF8000 correspond to supported voltages
20:00:09bertrikcmd8 is also a bit tricky. If it's not supported you can just issue another command for the rest of the initialisation sequence. If it succeeds, there's 4 more bytes worth of response to read first
20:00:57denes_funman: then I was wrong
20:01:12funmandenes_: thanks for the suggestion ;)
20:01:27denes_funman: what frequency do you set the clock to ?
20:01:42 Join pierre- [0] (n=pierre@
20:02:08funmanbertrik: I read response R7 (aka normal response)
20:02:16funmanas opposed to long
20:02:16denes_funman: iirc it can't be more than 400KHz
20:02:19funmandenes_: 240kHz
20:02:26shot0faddslinxstb: yes, that works nicely. cheers!
20:02:41funmanit must also be more than 100kHz
20:02:44shot0faddsat least I can do some nand driver work now ;)
20:03:21denes_funman: and do you send ff8000 as the register window?
20:03:40funman*PANIC*\ncmd8 failed!
20:04:02funmannicely the panic message fits in the yellow part of the lcd ;)
20:04:34*shot0fadds curses his previous inability to find the TAB button and lets linuxstb know that worked fine :)
20:05:01bertrik:P yeah actually quite clever. Same with the spectrum display in the OF: loud frequencies have a yellow top
20:05:37 Join mf0102 [0] (
20:05:49denes_gevaerts: really? I thought i checked that.
20:05:54funmancmd response timeout, I guess cmd8 not supported
20:06:38bertrikweird, I would have expected an error instead of a timeout
20:06:55funmanI send (1 /*2.7-3.6V*/ << 8)|0xAA /* test pattern */
20:07:29gevaertsdenes_: at least that fixes it here. Maybe you or wpyh can try on the new m3
20:07:33funmanwell in itself it is an error
20:07:59funmanit's the only meaningful error for unsupported commands in the status register of pl180
20:08:49linuxstbshot0fadds: OK, I'll commit. Sorry about that...
20:09:41 Join mcuelenaere [0] (n=mcuelena@rockbox/developer/mcuelenaere)
20:10:03denes_gevaerts: you are right, it fixed it here also
20:10:12denes_gevaerts: I wonder what I was checking then ...
20:10:44bertrikfunman, I'm only really familiar with SPI mode, SD mode is a bit different and possibly uses a timeout to indicate an error
20:12:04funmanbertrik: well the pl180 controller doesn't give access to all the bits in the SD/MMC protocol
20:12:31 Join ghen [0] (
20:13:46 Join {phoenix} [0] (
20:14:49ghenthe front page, subversion part, still links to "all changes since 2.5"
20:15:19ghenshould perhaps be since 3.0 now?
20:15:26ghen(the since30.html page does exist)
20:15:57funmanbertrik: according to wikipedia SPI is a subset of the SD protocol (designed for use with microcontrollers)
20:18:16denes_gevaerts: ah, I mixed up rgb_unpack_xxx with rgb_unpack_xxx_lcd I think, that's what happened
20:18:34gevaertsdenes_: ah, could be
20:20:01gevaertsAnyway this all needs to be looked at again when the LCD driver gets a bit more advanced (things like DMA, lcd_update_rect(),...)
20:20:38shot0faddslinuxstb: no worries, just glad I eventually found it. hopefully I'll be able to try some NAND stuff later too.
20:25:03*wpyh is here
20:25:53*bluebrother notices that the detection class uses GUI stuff :/
20:27:30domonokybluebrother: detecteddevices.h/cpp/frm.ui is the gui to present the detected devices, the code in autodetection.cpp/h shouldnt use GUI parts..
20:27:49 Quit mcuelenaere_ (Read error: 113 (No route to host))
20:28:03bluebrotherdomonoky: I was referring to the detect.cpp in current svn
20:29:10bluebrotherhmm, and autodetection.cpp doesn't include QtCore but relies on the header to do it. Not nice IMO
20:29:41 Quit HBK ()
20:30:10bluebrotherany objections to shuffling some files to base/ ?
20:30:37domonokyautodetection.h includes QtCore, so its not needed anymore in the .cpp, but perhaps it can be moved ? (nearly all includes in the cpp files is good)
20:31:35bluebrotherIMO the includes should be in the cpp file unless it's needed in the header too. It should still stay in the cpp to make clear what includes are needed
20:32:34domonokybluebrother: no objections against moving things to "base"
20:32:50bluebrotherone could even argue that a header shouldn't include anything and the including cpp file is responsible for having the needed ones in front ...
20:33:12domonokyif we pull the msgbox out of the Detect::check function, we can also put it into "base" :-)
20:33:28bluebrotheryep, that's what I want to achieve.
20:33:34 Part pixelma
20:34:23 Join pixelma2 [0] (n=marianne@rockbox/staff/pixelma)
20:35:24 Part Llorean
20:35:48 Join Llorean [0] (
20:35:58 Join goffa [0] (n=goffa@
20:36:03domonokydid you take a look at my new autodetection ? :-)
20:44:08 Join HBK [0] (
20:45:34 Quit J-23 (Remote closed the connection)
20:46:34 Quit goffa_ (Read error: 110 (Connection timed out))
20:55:40 Quit globi (Read error: 110 (Connection timed out))
20:56:21 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
20:56:41 Quit malo ("ChatZilla 0.9.83 [Firefox 3.0.3/2008092417]")
20:58:18***Saving seen data "./dancer.seen"
21:03:19funmanfor the clip button led, should I add an ifdef in firmware/drivers/led.c ?
21:03:42 Quit amiconn (" l8er")
21:04:37funmanit seems a hardcoded setting (LED_REAL) for archos player
21:04:54funmanhm it seems it's only used for ata (I guess busy)
21:08:18 Join sarixe [0] (
21:14:50 Join ajonat [0] (n=ajonat@
21:17:03linuxstbfunman: Yes, that's probably the hard disk LED. You probably want to look at the e200 port to see where/how the button light is handled.
21:18:48 Join lasser [0] (
21:22:23bluebrotherdomonoky: not yet, only skimmed it briefly. Just had a phone call ...
21:23:12bluebrotherI have a longer train trip tomorrow, will take the diff with me and try to give it a look
21:23:31 Quit avis (Remote closed the connection)
21:23:47 Join massiveH [0] (
21:24:16domonokygood :-)
21:31:01 Join veryevil12345 [0] (
21:32:48 Quit funman ("leaving")
21:36:29 Quit pierre- (Success)
21:39:13 Quit perrikwp (" ajax IRC Client")
21:40:01 Join perrikwp [0] (i=4aa794a0@gateway/web/ajax/
21:40:03domonokybluebrother: i think we can also move uninstall.cpp/h installzip.h/cpp and rbsettings to "base" without modification..
21:40:57bluebrotheryou sure? uninstall.h includes QtGui
21:41:18domonokyprogressloggerinterface.h should also go to base
21:42:19bluebrotheryeah, some files should get cleaned up and moved. From my point of view, every file in base/ must not include QtGui and not include anything from outside base/
21:42:26domonokyah, yes, uninstall.h includes QtGUI, but should only need QtCore
21:42:45bluebrotheryep, just tried it
21:43:34bluebrotherthat's what I meant with cleaning up −− I just grepped for all cpp files including QtCore in this first attempt
21:43:51bluebrotheranother reason why I want those includes in the cpp file too ;-)
21:44:12domonokyyep, no depencie to QtGui, and selfcontained code is good. (of course linking to parts of rockbox (non-gui) is ok)
21:45:05bluebrothersure. Kinda like the firmware/ and apps/ thing in main Rockbox
21:45:12 Quit moos (Read error: 110 (Connection timed out))
21:47:58 Join EspeonEefi [0] (i=espeonee@CERULEANCITY.MIT.EDU)
21:48:25bluebrotherI started writing a basic CONTRIBUTING document for rbutil. I guess I should finish that and include such rules ;-)
21:49:04domonokyjup, sounds good :-)
21:51:46 Quit Tetracomm ("Leaving")
21:55:38 Join sharp15 [0] (
21:55:39*linuxstb now waits for an e200v2 patch...
21:58:15 Join FOAD_ [0] (
22:03:06 Quit perrikwp (" ajax IRC Client")
22:03:28 Join perrikwp [0] (i=98212f4d@gateway/web/ajax/
22:04:01sharp15are any of the supported players still available for new purchase? i only found the iRiver h3** from 2 vendors and I'm not sure what version of IPod is available.
22:06:31n1ssharp15: no
22:07:14sharp15thanks. i can stop looking now.
22:07:34n1snot generally anyway, there may be a few here and there but used players is the way to go to get a supported rockbox player
22:08:05 Quit n17ikh|Lappy ()
22:08:53 Join jeffdameth1 [0] (
22:09:21 Quit jeffdameth (Read error: 60 (Operation timed out))
22:09:33 Quit sharp15 ("leaving")
22:09:44 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
22:11:00 Join miepchen^schlaf [0] (
22:13:16 Quit FOAD (Read error: 110 (Connection timed out))
22:13:16 Nick FOAD_ is now known as FOAD (
22:15:08 Quit miepchen^schlaf_ (Connection timed out)
22:20:07 Quit sarixe (Read error: 110 (Connection timed out))
22:21:14 Join sarixe [0] (n=sarixe@
22:22:23 Quit ajonat ()
22:22:30 Quit sarixe (Client Quit)
22:27:45 Join n17ikh|Lappy [0] (
22:28:17 Quit domonoky (Read error: 104 (Connection reset by peer))
22:32:53 Join TENK [0] (
22:39:38 Quit pixelma2 ("-")
22:39:53 Join pixelma [50] (i=pixelma@rockbox/staff/pixelma)
22:41:05 Quit ryanakca (Read error: 104 (Connection reset by peer))
22:42:20 Join amiconn [50] (n=jens@rockbox/developer/amiconn)
22:44:01amiconnHmm, no funman around
22:44:20*amiconn summons one of the 3 Swedes
22:45:00 Join mick_home [0] (n=mick@clamwin/admin/mickhome)
22:46:59 Join ryanakca [0] (n=ryan@ubuntu/member/ryanakca)
22:47:46bertrikamiconn, are you working on sansa v2s now too?
22:48:11amiconnI just wanted to give a hint related to buttonlights
22:48:52*amiconn is trying to keep an eye on overall code structure, especially since the LED code needs a good deal of rearranging
22:49:59shot0faddslinuxstb: I just got a 394kB rockbox.m200 to load and pass the checksum test, so things a looking a bit healthier :)
22:50:27*shot0fadds didn't realise the m200 only had 1MB RAM
22:50:37linuxstbI thought it was 2MB?
22:51:04shot0faddsso did I, but reality (and the wiki page) say otherwise..
22:51:53linuxstbWhat tcc chip (exactly) does it have?
22:52:23shot0faddsit's a TCC770 (no builtin RAM, no DAC)
22:52:45linuxstbAh, I thought it had one of the ones with 2MB RAM builtin...
22:53:26shot0faddsyeah, somewhere along the line I picked up that assumption too. peeling the sticker off and having a look helped..
22:54:11 Quit perrikwp (" ajax IRC Client")
22:54:25shot0faddsthere are a few more tweaks I want to do to the NAND driver, but it's looking a whole lot better than the existing version. I'll try to commit something later this week
22:54:33 Quit petur ("*plop*")
22:55:23 Join funman [0] (
22:56:04linuxstbshot0fadds: Sounds good.
22:56:31funmanamiconn: here is what I used (quite simple I think)
22:57:58amiconnfunman: My main hint is that buttonlights (for several targets) asre handled together with the backlight in the backlight thread (with separate timeout)
22:58:08 Quit {phoenix} (Remote closed the connection)
22:58:19***Saving seen data "./dancer.seen"
22:59:22amiconnAnother point is that the clip will need "backlight" handling too. Even though there is no real backlight, an OLED display consumes a considerable amout of power when active, and hence it needs a (configurable) deactivation timeout
23:01:13funmanwell I'm not here yet
23:02:59shot0faddsthe "RAM usage" count in rockbox-info.txt include the binary, right?
23:03:35 Quit jhulst (Connection timed out)
23:03:50amiconnfunman: Imo you should put that code into a separate .c file, backlight-clip.c
23:04:11 Quit bmbl ("Woah!")
23:04:17funmanand write "backlight" as a timer which shuts down the lcd display ?
23:04:20shot0faddsamiconn: how's that numebr calculated? does it include the plugin/codec buffers too?
23:04:38amiconnshot0fadds: No. It's just binary+bss
23:05:09amiconnfunman: No. Just use the backlight thread, which does these things for all other targets with backlight and/or buttonlight
23:05:30amiconnYou shouldn't need any more code than what you linked to if you have a working tick
23:05:37BagdershodanX: size = _end - _loadaddress (map symbols)
23:05:56BagderI meant to direct that to shot0fadds
23:06:02amiconnJust enable buttonlight handling in config-clip.h (or how that's called)
23:06:04shot0faddshehe.. in that case the current m200 build has a 43kB audio buffer ;-)
23:06:25amiconnhi Bagder
23:07:06amiconnBagder: I'd like to know how to proceed for making a 3.0.1 for archos (Player)
23:08:13amiconnI have some (uncommitted) code for an improved (reduced size) flash loader, and want to make new flash images. This requires moving the rombox start address for the Player
23:08:24linuxstbshot0fadds: Perfect ;) I think Nico_P had a patch to make the buffering code use a 32KB buffer.
23:08:44shot0faddswhat more could you need?
23:08:59amiconnRight now I am preparing to test whether minimon actually works, but hopefully that won't take long
23:09:25amiconn(need to open the Player and hook up wires to the serial)
23:10:31Nico_Plinuxstb: I confirm that. IIRC I put it on the tracker
23:10:34 Join Thundercloud [0] (
23:10:38Bagdera release for a single target of course makes things a bit complicated
23:10:56Bagderbut that's just things to fix
23:11:21amiconnIt's only the Player which requires it. Of course we could make it for all targets, but it wouldn't be different for those apart from the version number
23:11:34Bagderthat would feel a bit silly too
23:11:46amiconn(unless we also want a micro-fix in firmware_flash.rock - it says "BootBox mismatch"
23:11:55linuxstbIt would keep things simple though.
23:12:01gevaertsAt least it would keep a simple svn revision vs version mapping
23:12:07amiconn...if the rombox start address doesn't match. It should say "RomBox mismatch"
23:12:13Nico_Plinuxstb, shot0fadds: FS #9332
23:12:14 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
23:12:30linuxstbNico_P: I see it finally getting some testing soon...
23:12:35amiconnIf we do want this micro-fix, the 3.0.1 would change (a bit) for all archoses
23:12:45Bagdergevaerts: we can still do that, just that 3.0.1 is only for player so the next one would be 3.0.2 for whatever player(s) the release is for
23:12:56Nico_Plinuxstb: kugel gave it some testing and apparently it had quite a significant impact on runtime
23:12:57gevaertsAh, yes. Good point
23:13:08LloreanBagder, amiconn: Any reason this can't just wait a couple months for 3.1 then?
23:13:17LloreanI mean it's not like 3.1 is so far away this time.
23:13:27amiconnLlorean: The flashing instructions + files ...
23:13:40amiconnRight now it's a mess, and the ones we have are way outdated
23:13:46 Quit EspeonEefi ("さよなら")
23:13:50linuxstbNico_P: An impact in what way?
23:14:23Nico_Pit shortened runtime quite a bit
23:14:36linuxstbIs the CPU boosted more often?
23:15:01shot0faddsthe m200 doesn't even boot to the main menu yet, so there's not really any runtime to shorten :p
23:15:20Nico_Pprobably, also IIRC the flash memory doesn't get to go to sleep (I think it was amiconn who said that)
23:15:20bluebrotheryou could shorten the boot time ;-)
23:15:58Nico_Pof course with no more memory available there's no comparison to make ;)
23:17:25linuxstbNico_P: But it sounds like it's reliable - if it lasted for a runtime test...
23:17:46 Quit ompaul (Nick collision from services.)
23:18:13 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
23:19:53 Quit mcuelenaere ()
23:20:03shot0faddslinuxstb: in tcc77x\crt0.s isn't the copying (copied_start etc) redundant when TCCBOOT is undefined? I don't do this in the 78x version.
23:20:34*n1s still wants a osx (intel) tester for ipodpatcher
23:20:38shot0faddsin fact "1MB from the end of DRAM" sounds fun ;-)
23:21:02Nico_Plinuxstb: yes, I think it's pretty reliable
23:21:17linuxstbshot0fadds: That's probably not what you want... ;)
23:21:29Nico_Pthere is probably some room left for optimization and some bugs left, but it's a working proof of concept
23:21:41 Quit bluebrother ("leaving")
23:22:04linuxstbNico_P: Is it something that's easily enabled with a #define?
23:22:33Nico_Plinuxstb: it could be. it's just a matter of compiling buffering_flash.c instead of buffering.c
23:22:55Nico_Pthey both define the same things
23:24:52 Quit mf0102 ("Ex-Chat")
23:25:53 Quit funman ("leaving")
23:26:35linuxstbshot0fadds: The copying is needed to move the bootloader out of the way of where the main binary will get loaded by the bootloader. Or do I misunderstand your question?
23:27:10shot0faddslinuxstb: but doesn't it only need to do that if TCCBOOT is defined?
23:27:22shot0faddsor have I misunderstood?
23:28:05linuxstbTCCBOOT just defines whether the bootloader is installed into the OF as dual-boot, or if it's a standalone binary (uploaded with tcctool). In both cases it needs to be relocated.
23:30:52shot0faddslinuxstb: in that case, the current relocation doesn't work. If TCCBOOT is undefined the bootloader is linked with base = 0x20000000
23:31:44linuxstbI would like to get rid of TCCBOOT anyway - it's really just there to help people debug new bootloaders.
23:32:14 Join barrywardell [0] (n=barrywar@
23:32:53linuxstbBut that can still be done by patching an OF and uploading with tcctool anyway....
23:34:50 Part toffe82
23:35:17 Join toffe82 [0] (n=chatzill@
23:37:57 Join avis [0] (n=ident@pdpc/supporter/student/avis)
23:49:56 Join xoxa [0] (n=pear@
23:50:30 Quit massiveH ("Leaving")
23:54:15 Quit bertrik (Remote closed the connection)
23:55:49 Join saratoga [0] (n=41becb3b@gateway/web/cgi-irc/
23:56:47 Join perrikwp [0] (i=d1a8d351@gateway/web/ajax/
23:57:06 Quit xoxa (Remote closed the connection)
23:58:45 Quit avis (Remote closed the connection)

Previous day | Next day