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 2020-10-19

00:00:25Stanley00ah, too bad, I tried sleep timer last night, and it drain the battery again =]]
00:02:46efqwI have almost zero programming expertise so it would take me a long while to figure these things out and be able to write a c program that does what I just described.
00:11:24Stanley00hmm... I guess the official poweroff could handle all of them except for axp192 part
00:14:07efqwwe have a known list of all possible process names
00:14:34efqwso basically everything could be a shell script with the exception of the ioctl() part
00:15:39efqw`kill -9` the list of processes, `sync`, `umount /mnt; umount /data`
00:16:44 Quit Stanley00 ()
00:17:14 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)
00:40:28***Saving seen data "./dancer.seen"
02:37:48 Join cockroach [0] (~blattodea@pdpc/supporter/active/cockroach)
02:38:06 Part cockroach
02:40:29***Saving seen data "./dancer.seen"
04:20:55 Quit simpleOne (Quit: My MacBook has gone to sleep. ZZZzzz…)
04:22:16 Join simpleOne [0] (
04:40:32***Saving seen data "./dancer.seen"
04:48:01efqwI had a look at some of those small stuff, and the Shanling M1 seems to be a much better option than the FiiO M5
04:48:31efqwIt's hyby based and has a rotary encoder with a couple of buttons.
04:48:57efqwThis should be sufficient for navigating through rb's menus.
04:49:05efqwAnd it has Bluetooth
04:50:38efqwThe newer M0 only has one rotary encoder and I'm not a fan of it.
04:51:31Stanley00I heard that hiby R2 also has more buttons, no rotary, and with wifi
04:53:08Stanley00actually, I think we should have support for touchscreen soon, because many modern players come with it
04:57:42efqwwow, that's actually quite nice
04:57:47efqw5 buttons + power
04:59:16Stanley00yup, r2 looks so interesting to me
04:59:33Stanley00bigger screen, and also bigger battery :3
05:00:59efqwIt's an odd horizontal block
05:02:17efqwBut the hardware looks quite good, we should be able to make it work with the 5 buttons.
05:02:55efqwTouchscreen can wait but it's good to have a functional port before tackling with touchscreen
05:04:20efqwBleh, M1 is not available on amazon anymore.
05:04:47efqwPorting to hardware that are no longer for sale is not worth it.
05:05:38Stanley00efqw: if you like r2, you better get it soon, <= early gonna ends very soon
05:05:48Stanley00*early bird
05:13:13efqweither way that looks like a really good candidate
05:14:04efqwX1000E, Bluetooth, enough physical buttons, USB-C, relatively affordable, good screen, and no shitty recovery.
05:15:46efqwWith 5 buttons (excluding power), navigation should be trivial, and later we can add simple gesture controls to the touch screen too.
05:38:48 Quit Stanley00 ()
05:38:55 Quit atsampson (Ping timeout: 272 seconds)
05:40:23 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)
06:00:41 Join pamaury [0] (
06:00:41 Quit pamaury (Changing host)
06:00:41 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
06:02:09 Quit S|h|a|w|n (Read error: Connection reset by peer)
06:03:23simpleOnegood morning pamaury :)
06:03:48simpleOneSo I managed to decompress the different custom firmwares thanks to upgtools. The differences between the firmwares (stock/modded) are relatively small. I tried to find relevant information such as DSPs, or equalizers, but I couldn't find it. Would you know where I should look?
06:06:18simpleOneDseeHxCustom is sadly encrypted, unreadable
06:06:39simpleOneleft>stock ; right>modded
06:07:27 Join ufdm_ [0] (
06:08:04 Quit ufdm_ (Client Quit)
06:08:53 Quit ufdm (Ping timeout: 260 seconds)
06:15:15 Join atsampson [0] (
06:40:33***Saving seen data "./dancer.seen"
06:50:22 Quit sakax (Remote host closed the connection)
06:55:44speachyefqw: The shanling M1, M2S, M3S are all on the Hiby platform. Hizids AP60, AP60ii, AP60Pro, AP80, AP80 Pro, and AP100. the xDupp X10Tii, and finally last night I found the Cayin N3.
06:56:48speachyfunny thing −− the CAyin N3Pro has *tube amps*
07:01:55speachyStanley00: got a link to the updated patch?
07:02:54speachyoh, there's also the EROS TEN (aka Hifiwalker H8)
07:03:04speachy(er, hifiwalker HX)
07:13:26speachytiny form factor, 128x64 OLED screen, 5 hard buttons + rotary wheel.
07:22:16efqwEROS ten would be a cool target but I don't see how that is better than the xDuoo x3 (1st gen)
07:23:12efqwM1 looks like an interesting platform but I'm not really motivated to work on something that is no longer available on the mass market.
07:23:32speachyefqw: advantage is that it's still easily obtainable
07:23:48speachyand small form factor
07:23:49efqwTube amps to me just sounds like something would break if you accidentally drop it :P
07:24:14efqweros ten is x1000 too?
07:24:25speachyyup, hiby-based.
07:26:03efqwGot a fw download?
07:26:10efqwI'm not entirely sure if it has BT.
07:26:25speachyit does
07:26:52speachy(aka Hifiwalker HX. no firmware for that one but it's clearly the same player)
07:31:09efqwThere's an EROS J too which looks similar
07:31:16efqwBut no fw downloads so far
07:33:45speachyyeaH, the J looks like another re-casing of an existing PCB.
07:56:27efqwIs EROS still actively manufacturing the TEN?
07:57:34speachyit was more recently updtated than the Q, for whatever that tells you
07:58:04speachyI think that, generally, all of these hiby-based models are at their EoL.
07:59:21braewoodsi found these all over the place
07:59:38braewoodsbut i can't find it on the iriver website so is it even genuine?
08:00:02efqwEROS TEN was last updated in 2017-11-22 :/
08:00:30speachyeven the K was last updated in 2018.
08:00:38speachyso who knows.
08:01:11braewoodsspeachy: any insights into why the iriver porting stopped with the H10, H100, and H300 models?
08:01:45speachybraewoods: the realities of reverse-engineering hardware collided with "what we already have is Good enough for our needs" ?
08:01:56braewoodsyea maybe so
08:08:13efqwspeachy: well, I found this in the eros ten fw
08:10:21efqwso those two probably use the same fw too
08:11:20 Quit Stanley00 (Remote host closed the connection)
08:16:57 Join MrZeus_ [0] (MrZeus@gateway/vpn/mullvad/mrzeus)
08:23:16ArsenArsenalright got my iflash
08:23:30ArsenArsenthe rockbox bootloader is still there, all I need to do is get rockbox on it
08:23:45ArsenArsendo I just make a single MBR partition spanning the entire drive, mkfs.fat -F32, mount, and install?
08:23:53ArsenArsenipod 6g btw
08:24:41speachyput the iflash in, have itunes restore it, then do the usual rockbox installation.
08:28:05ArsenArsenyeah I ask that because I don't have itunes
08:28:10ArsenArsenso I guess my best route is a windows vm?
08:28:45speachyall I can tell you is to folllow the instructions in the manual and/or wiki.
08:29:19speachyif there's a non-itunes installation method then great! :)
08:30:10ArsenArsenhm okay
08:30:10braewoodsIs this still relevant?
08:31:11speachybraewoods: yes, unfortunately. because there are many "iflash" models that (mis)behave differently.
08:31:59braewoodsspeachy: wow. i'm starting to feel smart for using my simple passive adapters for CF cards.
08:32:00speachyand even today I see conflicting reports, and efforts to confirm the exact software/hardware is in use tends to be quite difficult.
08:32:43braewoodstoo bad real CF cards seem to stop at 128GB
08:32:51speachyso it's safer to leave those warnings up as a "you should have at least done some cursory research before buying this intending to put rockbox on it" warning.
08:33:37speachy(plus even if the main fw works great we might need to respin the bootloaders still..)
08:34:53braewoodsthough i'm stopping at 64GB for real CF cards
08:35:11braewoodsfor RB larger capacities are probaly best done through adapters since SD cards are so cheap
08:37:32braewoodsat least we know the upper limit on these is 2TB :)
08:37:42braewoodssince the disk label won't support anything larger
08:38:09speachythat's not to say there aren't OF limitations that preclude sizes that large.
08:38:22speachy(IIRC the ipod6g is one of those)
08:39:16braewoodsbraewoods: indeed.
08:39:19braewoodsspeachy: indeed.
08:39:37braewoodsthe 2 main barriers seems to be
08:39:40braewoods128GB and 2TB
08:39:44braewoodsdepending on LBA48 support
08:40:37***Saving seen data "./dancer.seen"
08:42:33braewoodsseems that one of the reasons bootloader development ceased was...
08:42:40braewoodsthe principle author bricked their device
08:43:15braewoodsprobably why i want to do this development on expendable devices
08:43:38braewoodsif i brick 'em, oh well. it's the cost of me trying to learn some stuff.
08:43:54braewoodsi will try to avoid mistakes...
08:46:16speachyeh, mistakes happen.
08:49:14braewoodsspeachy: indeed but
08:50:02braewoodsspeachy: anyway i'll see what i find out but my first intended addition to H300 is support for flashing the bootloader through iriver_flash
08:50:42braewoodsbut i will be researching the assumptions to see if i find any incompatibilties
08:50:54braewoodsthe H100 and H300 have different ROM chips
08:50:59braewoodsnot sure how compatible they are
08:51:21braewoodsthey may be similiar enough though
09:07:55braewoodsspeachy: i guess one thing i like about rockbox... the source code isn't so massive as to be a giant hassle to trace stuff.
09:08:18braewoodsit's easier to get into when all you know is C and Linux.
09:09:40 Quit t0mato (Quit: Ping timeout (120 seconds))
09:15:06ArsenArsenalright the restore is done
09:15:08ArsenArsennow for the reinstall
09:16:00ArsenArsenis the website down?
09:16:32ArsenArsenah, it's back
09:16:36ArsenArsenthe wiki was taking a while to load
09:16:45speachythat thing is slow in the best of times
09:17:09braewoodsspeachy: who runs the web server stuff?
09:17:47braewoodsi wonder how slow the host server is...
09:17:53braewoodsi've been running stuff on OVH for awhile
09:18:01braewoodsthe latest VPS are pretty nice
09:18:27braewoodsbut perhaps not cost effective. RB may use a lot of storage.
09:18:33speachyit's an old dual-socket opteron 2345.
09:18:40braewoodsOh, so bare-metal.
09:18:56speachythe wiki is really, really, really awful performance-wise.
09:19:13braewoodsand the underlying host won't help much there?
09:19:23speachynope. it's lightly loaded overall
09:19:35braewoodsso uh, is it PHP based something or other?
09:19:36speachyfoswiki has horrid concurrency/locking problems
09:19:59braewoodswhat version of php are you using?
09:20:02speachy(well, foswiki 1.x anyway. 2.x might be better)
09:20:23ArsenArsendoes rbutil use root or anything like that on linux? because I am not sure I got it to work on linux last time I had to install the bootloader
09:20:30 Join t0mato [0] (~t0mato@
09:20:39braewoodsArsenArsen: no, it shouldn't need it.
09:20:41speachyfoswiki is running under mod_perl right now, as it's the only way it can handle the current (mostly bot) load.
09:20:51braewoodsArsenArsen: at most it needs some way to mount partitions
09:20:54braewoodsor MTP
09:20:54speachymost of the rb stuff is perl cgi or static pages
09:20:56braewoodsor whatever
09:21:01ArsenArsenI mounted it myself
09:21:12braewoodsspeachy: oh it's perl
09:21:20speachythe theme and translation sites (and forum) is php
09:21:27gevaertsroot requirements for rbutil depend on the device
09:21:33braewoodsis there a reason you haven't upgraded to 2.x?
09:22:00speachybraewoods: no in-place upgrade is possible; it's effectively a new installation with manual data migration.
09:22:06speachy1.x->2.x I mean
09:22:16braewoodsso a royal hassle
09:22:43speachyyeah. If' I'm going to put that much work into it, I'll scrap foswiki altogether for something taht sucks less. which is nearly anything.
09:23:09braewoodsi have a suggestion. let me dig it up.
09:23:26ArsenArsengevaerts: it's an ipod 6g
09:23:37braewoodsspeachy: pmwiki might be a good choice to transition to
09:23:49braewoodsPHP based and i know CRUX uses it
09:24:05speachyI already have a dokuwiki instance running for other stuff
09:24:13braewoodsOk. Nevermind.
09:24:38speachybut remember, pages have to have their syntax translated, custom crap has to be ported, etc etc..
09:24:50speachyattachments, images, etc..
09:24:55braewoodsi see.
09:25:09braewoodsin time maybe we will want to freeze the existing wiki to transition to a new one.
09:25:09gevaertsI don't actually know how the 6g installation works... If it requires raw partition access (like you need on earlier ipods), you'll need root
09:25:29speachyit's been effectively frozen for several years. :D
09:27:29ArsenArsenalright no root appears to be needed
09:27:36ArsenArsenjust needing to manually mount the stuff as you'd expect
09:27:52braewoodsArsenArsen: depends on the distribution
09:28:07braewoodsmost distributions have auto-mount or other ways for users to mount removeable volumes
09:28:18braewoodsso no root needed for your part
09:28:20ArsenArsengentoo with no dbus meme that would let you mount disks (no gvfs, no udisks2, nothing, I don't like their design)
09:28:28ArsenArsenso yeah, have to mount it manually
09:28:51braewoodswhy do i always seem to run into gentoo or other niche users here? lol
09:28:58*braewoods feels alone over on Linux Mint.
09:29:15ArsenArsenhey it's only a niche on my main machine because I like playing around with stuff
09:29:18ArsenArsenmy go-to is debian
09:29:25ArsenArsenand yeah, it boots, wonderful!
09:29:33ArsenArsennow to restore a backup I got right before the last drive decided to die
09:30:21braewoodsArsenArsen: it can make decisions? wow. smart hard drive.
09:30:58ArsenArsenheh heh
09:31:05ArsenArsenironic because it would probably not pass a SMART test
09:31:29ArsenArsenright, to restore a backup, what do I do
09:31:51ArsenArsencopy everything except for rockbox.ipod?
09:32:01*gevaerts thinks automounting is weird :)
09:32:10ArsenArsenyou and me both
09:32:30ArsenArsen"[627477.464238] RockboxUtility[18518]: segfault at 5647f8541000 ip 00005647ef6f3fe2 sp 00007fffc8ee82a0 error 4 in RockboxUtility[5647ef60f000+139000]" right, I didn't debug this yet
09:32:55speachyoh! ArsenArsen, make sure you install a dev version, not the 3.15 release.
09:32:55braewoodsi once joked about replacing segfaults with facefaults
09:33:01ArsenArsen#0 0x00005647ef6f3fe2 in Logger::~Logger() ()
09:33:07ArsenArsenspeachy: ah well I used the release
09:33:10ArsenArsenwhat's the reason?
09:33:31braewoodsArsenArsen: probably iFlash workarounds and such
09:33:39ArsenArsenseems to be working for me
09:33:46speachyArsenArsen: there are definitely issues with 3.15 and the iflash SD adapters.
09:34:05ArsenArsenso, I just select rockbox (no bootloader, fonts, themes or game files) and hit install with that version?
09:34:05speachyreads are fine, but writes... you will get data corruption.
09:34:23ArsenArsenthat makes sense then, I didn't try writing data yet
09:34:34speachybecause the ATA->SD convertor in that thing has totally screwed up power management
09:34:45ArsenArsenoh huh
09:34:52ArsenArsenand this was not noticed until recently?
09:35:46ArsenArsenalright, installed the dev version
09:35:54ArsenArsenmight as well take the moment to copy over my old config and stuff
09:37:25ArsenArsenthis seems sufficient "cp -r config.cfg fonts/ wps backdrops/ icons/ /media/rockbox/.rockbox/"
09:37:29ArsenArsenI'll copy over some music to test with
09:37:42ArsenArsenupdating the db file should be enough to see whether it's broken
09:37:55gevaertsWell, we knew about problems for a long time, but only recently have people started working out causes
09:38:21ArsenArsenI mean, wouldn't the iflash team have noticed? They seem to endorse (or at least like) using rockbox on ipods
09:38:42 Join Stanley00 [0] (2a77e787@unaffiliated/stanley00)
09:41:26ArsenArsenLGTM I guess
09:41:35speachymain reason this one got fixed is because I came across someone who was using an ATA->SD adapter on an old PC, and the article included a screenshot of the ATA identity info.
09:41:48ArsenArsenI mean, it looks like my battery died, but otherwise LGTM
09:42:48Stanley00speachy: I updated to the same gist as before, you can check again here Stanley00/6d9b121ccfca7b8f70a75991d0264b31">
09:43:20speachythe FC1306 converer ASIC basically didn't implement a mandatory part of the ATA spec.
09:43:48ArsenArsenis it the same converter?
09:43:48speachyrockbox assumed that since it was mandatory, it would be present and work; fortunattely the ATA folks put in a capability flag for that feature even though it was mandatory.
09:43:50Stanley00efqw: I forgot my m5 at workplace :( but quick test show that the remain button actually work, can't test with real cloner today
09:45:30gevaertsOf course it's entirely possible that the people who built that thing saw that there's a flag for it and deduced that it's not mandatory
09:45:51ArsenArsenthat's a weird set of decisions
09:45:53ArsenArsenwhat feature was it?
09:46:03speachy"power management"
09:46:15ArsenArsendisk spinup and spindown?
09:46:16speachy(as in, all of it)
09:47:30speachyI suspect the actual problem wasn't that the device went haywire when power managment commands were issued, but because the RB code would actually kill power to the ATA interface/device as well.
09:47:51speachykilling power before the SD card had finished commiting any writes
09:48:22ArsenArsenso the solution is to check if power management exists, and if it doesn't, don't cut power?
09:48:22speachy(because the adapter lied about it being safe to kill power)
09:48:38speachyif power management doens't exist, then never put the ATA interface to sleep
09:48:44ArsenArsenyeah, alright
09:49:00ArsenArsenI guess the chip is going to, or at least should, do that instead then
09:49:01speachyit burns a bit of power, but a lot less than a real hard drive would.
09:49:21speachylow hanging fruit but it seemed to work
09:52:28gevaertsSeems likely. I don't know the ATA code, really, but I'd expect a panic if a command fails, so I'd assume it's not impossible that the commands succeed but don't actually do anything
09:53:28ArsenArsenif I got it right the command would do nothing and report success, leading to the actual power supply being cut off, which is fine in most scenarios but not when the card still hasn't written to flash, which I guess isn't even reported over sd
09:53:31speachyStanley00: BATTERY_CAPACITY 5 ?
09:54:03speachyArsenArsen: the native SD protocol actually handles that just fine; it's just that the iflash adapter hides that from rockbox
09:54:12speachyso we don't know that it isn't actually idle.
09:54:17Stanley00speachy: yup it just report 5 level
09:54:24ArsenArsenoh, so it exists in sd
09:54:30Stanley00the actual cap is 550 mAh or so,
09:54:52speachyof course this is all speculation on my part; it might be that the iflash adapter is actually losing its mind.
09:55:11ArsenArsenhey if the fix works, that's what matters
09:55:15ArsenArsentried telling iflash about it?
09:55:20speachyStanley00: BATTERY_CAPACITY_* is used for runtime estimation, needs to be the actual capacity.
09:56:11Stanley00I think I did use VOLTAGE detect or something, because we can't get it anywhere from fiio driver as I can tell for now
09:57:45speachybattery level reporting is different, implemented via _battery_level() amd #define CONFIG_BATTERY_MEASURE PERCENTAGE
09:58:14speachy(if it's exposed as a level versus a raw voltage)
09:58:52speachyPERCENTAGE_MEASURE that is
09:59:16Stanley00speachy: what I mean is on m5, /sys/class/power_supply/battery/capacity <= this report max at 5 when fully charged
09:59:36speachy(See firmware/target/hosted/xduoo/power*-xduoo.c)
10:01:12speachyStanley00: and when it's less-than full, the value is somewhere in the 0-4 range?
10:01:26speachy(so you take that value *= 20, and that's your reported percentage)
10:01:36Stanley00yeah, right
10:02:16Stanley00it looks like fiio like to implement all the things in kernel space on m5 :(
10:02:38Stanley00even the keypress handle of short/long press is in kernel too
10:03:25speachywhen you're happier with the port, please submit it into our gerrit system, it'll be much easier to properly review
10:04:28Stanley00okay, sure, the volume button somehow not working with recent build, it only effect when I pause and playing the song again
10:04:44Stanley00also I still have issue with poweroff
10:12:17 Join Stanley|00 [0] (~stanley00@unaffiliated/stanley00)
10:14:21 Quit Stanley|00 (Disconnected by services)
10:16:10Stanley00I'm reading the gerrit process, we push directly to master branch for review?
10:16:30Stanley00also, is it okay to push large commit, as in my patch, it's really big
10:17:07speachythere really isn't a good way to slice up a new port
10:17:27speachyunless you want to do the plugins separately I suppose
10:17:37Stanley00got it, thanks
10:18:14Stanley00I will push it after I playing around with the poweroff or volume button issue then
10:25:47 Join cockroach [0] (~blattodea@pdpc/supporter/active/cockroach)
10:27:59 Quit Stanley00 (Ping timeout: 245 seconds)
10:31:23 Quit massiveH (Quit: Leaving)
10:35:27 Join ufdm [0] (
10:40:38***Saving seen data "./dancer.seen"
10:52:18 Quit pamaury (Quit: Konversation terminated!)
11:02:47speachyhuh. this is promising, but the license makes it a non-starter:
11:05:45speachy3-clause BSD, but with a fourth clause: "Any redistribution, use, or modification is done solely for personal benefit and not for any commercial purpose or monetary gain."
11:08:41gevaertsSo basically no point in the first three clauses
11:11:15speachyI know BTStack is/was widely used in automotive applications.
11:12:50braewoodsbecause they want to sell it
11:12:58braewoodsrockbox is generally non-commercial but
11:13:04speachyand they dont' want to end up competing with themselves.
11:13:17braewoodsnot really something we want to deal with
11:13:20speachythe fact we accept donations makes us "commercial"
11:13:27braewoodsi see.
11:13:38braewoodsi didn't consider that since you're not selling anything really
11:13:56speachythat term is so loose that one can drive a fleet of supertankers through it.
11:14:18gevaerts"non-commercial only" really means "don't use"
11:14:46speachywhat are you going to do, use one of the competing BTClassic stacks? :)
11:15:02braewoodsi take it bluez is a non-starter
11:15:08braewoodsprobably too much cruft
11:15:45braewoodsthough it's for Linux so if you were to use it
11:15:55braewoodsit'd have to be ported somehow
11:16:00speachyit would require a lot of surgery since it's closely tied with Linux, both in the POSIX sense and the kernel's BT framework.
11:17:28braewoodsfound something under apache
11:17:44efqwBlueZ would make sense for a hosted target.
11:18:24speachyIndeed, all of our existing hosted targets with bt already use bluez.
11:18:48braewoodsdo we have any native targets with BT hardware?
11:18:48efqwBy the way, btstack's documentation is quite nice.
11:19:09braewoodsif not there's not much use in our own BT stack...
11:19:10speachyyes, btstack is quite good!
11:19:12efqwIt has quite a bit of docs about HCI and UART
11:19:55braewoodsi'm not too surprised 99% of RB is asm or C...
11:20:05braewoodsstuff like lua is too slow to be practical for much
11:20:39braewoodsthough i could see it being useful for stuff you don't want to recompile for
11:53:57speachyI'm actually asking BlueKitchen if they'd be willing to add a GPL license option to BTStack.
12:11:09efqwOr even AGPL would work if they are super hostile to commercial use.
12:11:10speachyoddly, btstack isn't on wikipedia's list of bluetooth stacks.
12:12:32speachythis is what I sent:
12:13:12speachyAGPL wouldn't gain anything except perhaps as a slightly-more-poisonous pill.
12:13:17 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)
12:18:08 Quit Stanley00 (Ping timeout: 256 seconds)
12:28:06braewoodsefqw: AGPL wouldn't be of much use for a BT stack...
12:28:37braewoodsAGPL is mainly for the SAAS situation where no actual source distribution occurs.
12:28:45braewoodsIt's just used as a remote service.
12:29:10braewoodserr no actual program distribution
12:29:26braewoodsbut a BT stack is local so it has to be distributed where it is used
12:29:39braewoodsso GPL is enough
12:31:55speachyAGPL could arguably be extended to applying to anyone connecting to the device.
12:32:13speachy(when it's not the same person on both ends, heh)
12:40:39***Saving seen data "./dancer.seen"
12:54:18 Quit Huntereb (Ping timeout: 260 seconds)
13:13:00 Join lebellium [0] (
13:17:17lebelliumbraewoods: the seller told me he just went on holiday and will only ship the iRiver remote early November
13:20:19braewoodslebellium: ok.
13:20:23braewoodslebellium: i can wait.
13:20:31braewoodsi have lots of other stuff to do in the meantime.
13:20:53braewoodsjust wanted it to be tested eventually
13:25:00lebelliummeanwhile I found evidences that they all work with Rockbox
13:26:57braewoodsso old information then
13:28:53lebelliumAnd the H300 LCD remote is so rare because it was only sold in Korea.
13:29:24lebelliumso only a few enthusiasts pay money to import it
13:29:24braewoodsso if the support isn't already present
13:29:32braewoodsit's probably never going to be
13:30:23braewoodsjust removed the h100 remote reference since it is apparently no longer true
13:31:13braewoodslebellium: thanks
13:31:34braewoodsso far it appears the main addition is supporting flash updates
13:31:49braewoodsthe H100 bootloaders support it and so does iriver_flash
13:32:12braewoodssince the H300 has more ROM space it will need to be modified accordingly
13:32:20braewoods2x the amount the H100 has
13:37:55 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
13:43:06pamaurysimpleOne: hi, just saw your message in the log
13:43:42pamauryI am not entirely how to interpret your image containing the diff
13:44:20pamauryIf I understand correctly, the load_sony_driver is different, and that is an important file that you can look at
13:46:01pamauryalso maybe DseeHxCustom was taken from another player and that would explain the different sound, not sure about this
13:54:04simpleOnepamaury cheers, what you would rec to analyze this load_sony_driver? hex editor? IDA? something else?
14:03:22pamaurythis is just a bash file iirc
14:03:28pamauryso text editor
14:13:27ArsenArsenah hell yeah
14:13:35ArsenArsenfinally got the ol ipod 6g reassembled and in good nick
14:13:40ArsenArsenthat was bloody painful honestly
14:14:17ArsenArsenit's also highly bent out of shape but that's how I found it so it's no biggie
14:14:32 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)
14:16:09ArsenArsenI wonder if someone sells less painful cases for this
14:16:50simpleOneok sorry I was thinking of another file, so it looks like there's no significant differences, they just removed a few things, but nothing about DSPs :(
14:17:40 Join petur [0] (~petur@rockbox/developer/petur)
14:17:41pamaurysimpleOne: my guess is that he might have just copied over some DSP stuff from another player
14:19:02 Quit Stanley00 (Ping timeout: 258 seconds)
14:20:28simpleOneMaybe so, but when you look at the audiophile descriptions for each fw, it really sounds like it's just an hidden EQ profile, take a look (in spoilers):
14:20:55simpleOnemaybe It's just 100% bullshit though haha
14:21:16simpleOneI've to admit that I expected better from these modded shit :(
14:23:15simpleOnethe DseeHx profiles haven't been reversed yet, right?
14:34:48pamaurynot that I know
14:39:43braewoodsArsenArsen: check out gomadic. they sell a lot of after market accessories for old electronics.
14:40:03braewoodsArsenArsen: most notably replacement power/data cables.
14:40:40braewoodsi found stuff for the iriver h10
14:40:41***Saving seen data "./dancer.seen"
14:40:52braewoodsand such when i couldn't otherwise find replacement "chargers"
14:41:16simpleOnethere's this vinylcoeff file that differs for sure pamaury: ; do you know what it is?
14:42:51simpleOneI'm not sure if these are the default DSP settings or just another activable DSP in the DAP, I can't see any mentions of a possible vinyl "emulation" on Sony website
15:26:11pamaurysimpleOne: no idea to be honest, but I can send you the email of someone who has tried to figure out a bit how the audio parameters work
15:26:20pamauryto be honest, I am not so interested in how the OF works ^^
16:11:44braewoodsspeachy: interesting. the fastest i've gotten from rockbox so far is around 13 MB/s for sequential block device writes
16:11:59braewoodsthat's on the h120 using CF storage.
16:29:04speachybraewoods: context? not sure what's in response to..
16:29:21braewoodsspeachy: nothing i guess. i'm just collecting data.
16:29:40speachyah, okay. glad I didn't miss something
16:34:07speachyheh, here's another option for BTStack −− the noncommercial clause was added in april 2012. Before that it was just 3-clause BSD.
16:34:27speachyIn other words, we could safely integrate that old version if we so chose
16:40:43***No seen item changed, no save performed.
16:41:39 Join Oksana_ [0] (~Wikiwide@Maemo/community/ex-council/Wikiwide)
16:56:51speachyand looking at the history, the 4th clause was intentionally chosen over a GPL route.
16:58:43speachyaaaaand pre-fork doesn't have classic a2dp or avcrp support. So much to that idea.
16:59:40speachyI guess we'll see what they have to say in response to my inquiry, but it's probably safe to assume this is a dead end.
17:00:12braewoodsspeachy: what about the code I found that is apache licensed?
17:01:10speachybraewoods: I _think_ that's just the hostside glue code for cypress's proprietary stack. if's of no use if you're not using cypress hardware
17:01:20braewoodsi wonder if their BT stack is worth anything
17:01:32speachyand it's LE-only anyway
17:01:47braewoodsOh. So not backward compatible?
17:02:20speachyit's not a so-called "dual-mode" stack, no.
17:02:55speachyand yeah, there's no actual source code to the stack there, just some platform glue code and a pair of binary blobs.
17:03:46 Quit petur (Quit: Leaving)
17:03:58speachy1MB .a file :)
17:04:34speachyit is dual-mode though, but doesn't implement A2DP.
17:05:19speachyoh wait, I was referring to the cypress thing.
17:06:49braewoodsalmost seems like we'd have to roll our own implementation
17:06:49speachythe apache mynewt thing is still LE-only. (well, LoRa too)
17:07:36braewoodsor borrow code from the linux kernel
17:07:38braewoodsor something
17:08:23braewoodsso serious question...
17:08:30speachyhack BlueZ into something more manageable, perhaps
17:09:01braewoodswhat native targets even have bluetooth hardware?
17:09:06speachythere's literally two decades of bugfixes and workarounds for broken hardware and other tribal knowledge embedded into that codebase.
17:09:08braewoodsi mean
17:09:17speachybraewoods: there are none. I want to change that.
17:09:30braewoodsbut isn't this a chicken or the egg problem then?
17:09:37braewoodsyou need a platform before you can really target it.
17:10:30speachyheh, actually a UART is a UART no matter if it's via Linux TTY or DMA'ing a buffer directly. :)
17:10:57speachyso from the stack perspective, there are OS interfacing bits and the rest is self-contained.
17:11:28speachy(that said I'm sure bluez expects to be able to malloc() at will, whereas something starting out for embedded use likely uses mostly static buffers)
17:14:01braewoodsspeachy: so we basically need our own BT stack seeing as there's almost none out there.
17:39:29 Quit lebellium (Quit: Leaving)
17:48:25speachybut it's all moot until we have a native target that requires it.
17:49:31speachyeven on these ingenic platforms it's not clear how much is actually open-source. I'm of the opinion we should rely as little as possible on the "helper" stuff that eg hiby provides.
18:00:53 Quit pixelma (Quit: .)
18:00:54 Quit amiconn (Quit: - Chat comfortably. Anywhere.)
18:03:29 Join pixelma [0] (marianne@rockbox/staff/pixelma)
18:03:30 Join amiconn [0] (jens@rockbox/developer/amiconn)
18:06:51 Quit amiconn (Client Quit)
18:06:51 Quit pixelma (Client Quit)
18:24:36braewoodsspeachy: might be easier to do if we can design our own PCB
18:24:52braewoodsspeachy: maybe pine can help there
18:25:02braewoodsi'm no electrical engineer so that is beyond me
18:27:35 Join S|h|a|w|n [0] (~shawn156@unaffiliated/shawn156)
18:27:49 Quit pamaury (Ping timeout: 260 seconds)
18:32:12 Join pixelma [0] (marianne@rockbox/staff/pixelma)
18:32:13 Join amiconn [0] (jens@rockbox/developer/amiconn)
18:40:44***Saving seen data "./dancer.seen"
19:06:41 Quit cockroach (Quit: leaving)
19:23:53 Quit St3ak (Quit: Stable ZNC by ##bnc4you)
19:24:56 Join St3ak [0] (
19:30:25 Quit JanC (Remote host closed the connection)
19:30:48 Join JanC [0] (~janc@lugwv/member/JanC)
19:40:19speachybraewoods: no.. I'm referring to hiby's "system_server" that handles a lot of the BT manipulation for the player
19:40:49speachyit controls powering on/off the module, and provides a wrapper for stuff like scanning, pairing, etc.
19:41:00speachyoh, and active connecting too
19:41:11 Quit MrZeus_ (Ping timeout: 244 seconds)
19:41:40 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)
19:41:55speachyaudio-wise, I _think_ it's basically limited to switching audio input/output; the actual player basically knows nothing about that stuff beyond being able to call into the system server
19:42:54speachyand that's why there's no AVRCP integration with the player −− ie no passthrough of playback controls or audio track information
19:43:23 Quit bluebrother (Disconnected by services)
19:43:28 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
19:43:59 Join fs-bluebot_ [0] (
19:43:59speachybluez definitely supports AVRCP; but there's nothing hooking it up on these things.
19:45:03speachyand that's not going to be solveable unless we bypass hiby's system_server altogether.
19:46:24 Quit Stanley00 (Ping timeout: 265 seconds)
19:46:25 Quit fs-bluebot (Ping timeout: 264 seconds)
19:58:14braewoodsspeachy: AVRCP?
20:05:49braewoodsidea for a new app. it detects if you're separated from your smartphone and it starts crying like a baby if it feels neglected.
20:13:18 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)
20:18:30 Quit Stanley00 (Ping timeout: 256 seconds)
20:40:45***Saving seen data "./dancer.seen"
22:14:54speachybraewoods: You mean this? :)
22:40:47***No seen item changed, no save performed.
22:42:27 Join Stanley00 [0] (~stanley00@unaffiliated/stanley00)
22:50:42Stanley00lol... I also had the idea to turn m5 into tamagochi =]]
23:13:57Stanley00the capacity reading on m5 is worst than I think
23:15:04Stanley00it's use reference from m3k driver, and it max out as 5 at voltage level great than around 3.8V :(
23:28:45 Quit t0mato (Quit: Ping timeout (120 seconds))
23:38:07 Quit JanC (Ping timeout: 272 seconds)
23:40:22 Join t0mato [0] (~t0mato@
23:42:31 Join JanC [0] (~janc@lugwv/member/JanC)
23:48:12 Quit TheSeven (Ping timeout: 260 seconds)
23:48:22 Join [7] [0] (~quassel@rockbox/developer/TheSeven)

Previous day | Next day