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 2020-07-22

00:20:33***Saving seen data "./dancer.seen"
00:59:18 Join Rower [0] (
01:12:40 Quit saratoga (Remote host closed the connection)
01:33:22 Quit koniu (Remote host closed the connection)
01:33:54 Join koniu [0] (~koniu@gateway/tor-sasl/koniu)
01:38:54 Quit craftyguy (Ping timeout: 260 seconds)
01:43:36 Join craftyguy [0] (~craftyguy@qutebrowser/dev/craftyguy)
01:49:23 Quit Rower (Ping timeout: 240 seconds)
02:20:36***Saving seen data "./dancer.seen"
03:04:02 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:4961:9a78:8dc2:c7a6)
03:51:49 Quit Lonoxmont (Ping timeout: 264 seconds)
04:20:39***Saving seen data "./dancer.seen"
04:36:25 Join sakax [0] (~r0b0t@unaffiliated/r0b0t)
04:43:19 Join Lonoxmont [0] (
06:14:27 Join Rower [0] (
06:20:41***No seen item changed, no save performed.
07:00:48fs-bluebotBuild Server message: New build round started. Revision cb94b3a, 295 builds, 12 clients.
07:01:15 Join ac_laptop [0] (~ac_laptop@
07:20:54fs-bluebotBuild Server message: Build round completed after 1207 seconds.
07:20:58fs-bluebotBuild Server message: Revision cb94b3a result: All green
07:57:52 Quit ac_laptop (Ping timeout: 256 seconds)
08:02:17 Quit blackyus17 (Quit: AFK)
08:02:29 Join blackyus17 [0] (
08:20:43***Saving seen data "./dancer.seen"
08:38:44 Quit danielp3344 (*.net *.split)
08:39:21 Join massiveH [0] (
08:40:40 Quit nihilazo (Write error: Connection reset by peer)
08:40:41 Quit kadoban (Write error: Connection reset by peer)
08:40:45 Quit blbro[m] (Remote host closed the connection)
08:48:56 Join blbro[m] [0] (blbrostrat@gateway/shell/
08:57:37speachybluebrother, the commits to allow the buildclient to specify a 'dummy' arch have landed.
08:58:13speachyalso, the manual targets now require both latex and the native toolchain for that target before they will proceed.
08:59:24speachyalso, saratoga, do you want me to try and combine your two gerrit accounts?
09:11:11 Join MrZeus_ [0] (
09:12:27speachyaaaand anyone care to review the archos removal patches (1/4 especially)
09:15:15 Join nihilazo [0] (nihilazoma@gateway/shell/
09:15:15 Join danielp3344 [0] (danielp334@gateway/shell/
09:15:15 Join kadoban [0] (kadobanmat@gateway/shell/
09:56:27 Quit MrZeus_ (Ping timeout: 244 seconds)
10:16:53speachythere. last of the manual warnings are fixed. I think.
10:17:06 Quit massiveH (Quit: Leaving)
10:20:46***Saving seen data "./dancer.seen"
10:24:19 Join ac_laptop [0] (~ac_laptop@
11:59:02 Quit Rower ()
11:59:14 Join MrZeus_ [0] (
12:01:09 Join Rower [0] (
12:17:07__builtinspeachy: is there any hope in fixing the non-www domain?
12:17:29__builtinseems to be causing lots of issues for new users
12:17:49speachyit's out of my hands. All I can do is ask the Haxx crew again
12:19:02speachyI don't think there are any non-www URLs on the official sites
12:20:47***Saving seen data "./dancer.seen"
12:29:32__builtinbut third-party sites still have the non-www url (and perhaps search engines, too)
12:29:49speachyA little while back I did find some refecences to it in the sitemap
12:31:34 Quit ac_laptop (Quit: WeeChat 2.8)
12:31:53 Join ac_laptop [0] (~ac_laptop@
12:32:46__builtinwhat has the haxx crew said?
12:33:57speachyIMO my taking over the domain is the least-effort option.
12:34:11speachybut obviously that's not something I can just do.
12:35:08efqwI've spent a little bit more time with m3k's firmware, and there are some things that needs to be addressed by rb itself.
12:36:03efqw The firmware on there doesn't have SDL, so we can either try to get SDL1.2 working on it, or just use the raw framebuffer as-is. Which option would be the most convenient?
12:36:28speachyyou're using their toolchain?
12:37:04speachyIIRC we compule libsdl as part of our cross-toolchain for hosted targets
12:38:18efqwdoes this mean we don't need functional libsdl libraries on the host?
12:38:53 Join jdarnley [0] (
12:41:12 Quit J_Darnley (Ping timeout: 256 seconds)
12:42:54efqwdac power management relies on raw ioctl commands in the stock player binary, there will be no audio output if you don't manually enable the headphone amp portion in the dac on startup (and optionally disable it after playback is done to optimize for battery life)
12:44:44speachyefqw, if they don't exist we can always compile our own. less convenient to be sure, but not a showstopper.
12:45:08speachyand we can issue ioctls (or whatever) in the target-specific code
12:45:39efqwwill we run into libc compatibility issues?
12:46:21efqwthe libc on the stock os is eglibc 2.16
12:47:20speachyif you're using their cross-toolchain then no
12:54:00efqwI've been reading commit d55680993df9b6743506814d98b5cc1859828f8a to understand what kind of changes would be required to make that port work on this device.
12:54:59efqwFortunately the PMU is exactly the same
12:55:47speachyyou can probably ignore everything outside of firmware/target/hosted in that commit
12:56:49speachyor not
12:57:55efqw bootloader/rocker_linux.c looks intimidating
12:58:37speachyfor the m3k all you'd need to change is a few lines there
12:59:05 Quit lemon_jesus (Ping timeout: 240 seconds)
12:59:22speachylook at the difference between rocker_linux and xduoo_linux
12:59:44speachyand that's all you'd need to change. (merging the two is on my todo list)
12:59:51efqwfirmware/drivers/audio/rocker_codec.c will have to be rewritten :/
13:04:22 Join __Bilgus_ [0] (41ba23be@
13:04:46efqweither way, I'll probably start with trivial targets like keymaps, backlight controls, etc. I don't know how to write C, but I can document these in plain text.
13:06:13 Join lemon_jesus [0] (
13:10:52efqwby the way, that existing m3k port doesn't seem to be using sdl
13:11:25speachyusing native framebuffer and alsa? that works too..
13:12:07speachywe have exsting fb code that you can re-use, ditto with alsa
13:12:16efqwinitially I thought they got sdl working, but looking at their rootfs, I can't find `` at all
13:12:27speachyany libSDL at all?
13:12:32speachy(it could have SDL2..)
13:20:08efqwThe only changes they made to the rootfs is a new startup entry in /etc/init.d and the new /usr/rockbox folder.
13:21:36efqwNo new libs as far as I'm aware of
13:21:42 Join lebellium [0] (
13:27:46 Quit ac_laptop (Quit: WeeChat 2.8)
13:28:03 Join ac_laptop [0] (~ac_laptop@
13:40:14speachyefqw, I consolidated the bootloaders in g#2575
13:40:16fs-bluebotGerrit review #2575 at : bootloader: Consolidate xDuoo X3ii/X20 and AGPTek Rocker bootloaders by Solomon Peachy
13:43:30fs-bluebotBuild Server message: New build round started. Revision 0c4ae41, 295 builds, 9 clients.
13:43:51efqwcool, thanks
13:44:01speachyit's the commit that's building
13:44:41speachyin theory all you'll need to change is a single ifdef at the top to add the m3k's icon size and firmware file name.
13:45:04efqwhow should we store the rb binaries? since this device has nand flash, I'd prefer to stash everything onto the nand instead of using microsd like some other players seem to rely on
13:48:45speachythat's the approach xvortex took, but experience has shown us that keeping .rockbox in the onboard flash is a bad idea
13:48:53speachyas onboard flash wears out
13:49:47speachy(and will, longer-term, brick your device)
13:50:19speachywe want to modify the OF (and filesystem) as little as possible.
13:51:22efqwyeah, fiio didn't care much about the nand unforturnately
13:52:21efqwbetween the two separate ubi partitions (instead of volumes) and the incredibly verbose logging, the devices will probably die after a couple of years
13:53:34efqwif this is a dualboot fw, I'll probably put in something that symlinks the log file on the data partition to /dev/null
13:56:40speachyprobably less than that.
13:57:10efqwoh, and the database is also written to the nand, lol
14:01:27efqwso if I'm reading this correctly, the only file that needs to be on the rootfs is the multiboot bootloader?
14:01:51efqw(suppose we take the agpteck rocker approach)
14:05:08fs-bluebotBuild Server message: Build round completed after 1298 seconds.
14:05:12fs-bluebotBuild Server message: Revision 0c4ae41 result: All green
14:07:16speachy...and whatever is needed to launch it, yes.
14:07:31__Bilgus_hosted targets don't use multiboot
14:07:34speachypart of that commit I made was to move the bootloader's config file to the sd card too
14:08:22speachyshould probably add in the ability to set the boot timeout too.
14:09:11efqw"multiboot" in this case refers to "choose which player binary to launch"
14:10:11speachy...probably have an hour left on $dayjob build, bah.
14:12:41efqwso basically all we need to do to the stock firmware are: debloat the rootfs (optional), get sdl1.2 working on the device, add a startup script to check & preemptively symlink the log file to /dev/null (to reduce flash wear), and finally add the rb bootloader binary and the startup script for it?
14:14:32speachyno need for sdl; go native with alsa + fbdev + input −− existing hosted players already have most of what you'll need
14:15:42efqwI was hoping to be able to reuse as much of the rocker port as possible tbh
14:20:49***Saving seen data "./dancer.seen"
14:28:02speachyrocker doesn't use sdl.
14:30:03speachyuses bare linux fbdev, alsa, and /dev/input
14:31:04efqwI thought it used sdl but apparently it's just the *simulator*
14:35:12 Quit __Bilgus_ (Remote host closed the connection)
15:21:38 Join reductum [0] (
15:41:12 Join mendel_munkis_ [0] (
15:44:04 Quit mendel_munkis (Ping timeout: 256 seconds)
16:10:47 Join mendelmunkis [0] (
16:13:32 Quit mendel_munkis_ (Ping timeout: 256 seconds)
16:18:30 Quit reductum (Quit: WeeChat 2.9)
16:20:53***Saving seen data "./dancer.seen"
17:02:52 Quit sakax (Remote host closed the connection)
17:24:22 Quit ac_laptop (Ping timeout: 260 seconds)
17:39:42 Quit lebellium (Quit: Leaving)
18:14:59 Quit Rower (Ping timeout: 240 seconds)
18:15:35 Join ac_laptop [0] (~ac_laptop@
18:20:45 Join Rower [0] (
18:20:56***Saving seen data "./dancer.seen"
18:27:13 Quit Rower (Ping timeout: 258 seconds)
18:41:41 Quit michaelni (Quit: Leaving)
18:42:16 Join michaelni [0] (
19:43:25 Quit MrZeus_ (Ping timeout: 240 seconds)
20:20:57***Saving seen data "./dancer.seen"
20:46:30 Join LambdaCalculus37 [0] (
20:49:40LambdaCalculus37Feels weird being back after so many years away.
20:52:13speachyc'est la vie
20:53:03LambdaCalculus37I don't know what's making me come back to messing with the Rockbox code after, what... 8 or so years?... but I'm playing around with trying to get OpenBSD to be a supported build environment.
20:54:03 Quit LambdaCalculus37 (Quit: Leaving)
20:54:40 Join LambdaCalculus37 [0] (
20:54:45 Quit LambdaCalculus37 (Remote host closed the connection)
20:55:12 Join LambdaCalculus37 [0] (
20:56:35LambdaCalculus37I was able to at least get tools/configure to recognize the uname of OpenBSD, tools/ doesn't work under OpenBSD at all, of course.
20:57:51LambdaCalculus37Building even the sim in OpenBSD breaks spectacularly with a boatload of parse errors in the Makefile:
20:58:13 Quit ZincAlloy (Quit: Leaving.)
20:59:33LambdaCalculus37In OpenBSD, just spits out the following when envoked: './[35]: nproc: not found'
20:59:44speachynot terribly surprising given reliance upon quasi-newish GNUisms
21:00:07LambdaCalculus37And OpenBSD not being a very big fan of said quasi-newish GNUisms.
21:00:39LambdaCalculus37Honestly, I'd be surprised if the tools even still work in FreeBSD.
21:01:36speachyif the sim barfs on the makefile then so will the main builds
21:02:47LambdaCalculus37Of course. I still remember that. :)
21:03:44LambdaCalculus37And I know I can't get the toolchains installed because of the script barfing on me.
21:04:13LambdaCalculus37I don't know which is crazier: trying to build Rockbox in OpenBSD, or trying to build Rockbox in Haiku. :P
21:06:19LambdaCalculus37What I found hilarious is Haiku actually kind of started working, but the tools barfed when it came to the uname check.
21:06:29LambdaCalculus37I posted about it on the forum:,53420.0.html
21:07:06mendelmunkisLambdaCalculus37: At one point I considered trying to make a compiler plugin for rockbox.
21:10:35LambdaCalculus37I mean, it would be cool to see additional OSes be able to be used to have functional build environments for Rockbox, but it's going to be a little tough when those OSes are not quite the same as what everyone is used to, or are obscure as all getup.
21:16:30 Nick mendelmunkis is now known as mendel_munkis (
21:19:20 Join __Bilgus_ [0] (41ba23be@
21:21:27__Bilgus_I like the idea of the vm to a controlled enviroment
21:23:08__Bilgus_each time I start setting up a new one I intend to make a build for the masses but 4/4x now I've forgotten to do so before I start bloating it up with tools I prefer
21:24:03__Bilgus_maybe it could be like a very minimal headless install that acts as a build farm for the user
21:24:45LambdaCalculus37Guess for now I'll set working on OpenBSD as a build environment aside for a bit until I can get around those weird GNUisms.
21:25:27__Bilgus_lambda I feel that the dev script has wayyyy too much content to pickout meaningful errors
21:25:51__Bilgus_I have a patch on gerrit that fails early and only prints errors
21:29:04__Bilgus_ON to other topics.. I'm thinking with grabbing playlist information (adding playlist functions to plugin API)
21:29:14LambdaCalculus37__Bilgus_, you mean itself?
21:29:29__Bilgus_Lambda, yes
21:31:56__Bilgus_the functions I need are Playlist_Amount and Playlist_Get_Display_Index
21:33:13__Bilgus_Instead I think I'd rather add playlist_get_current() and parse the playlist struct local
21:34:15__Bilgus_that also will expand the functionality for the other broken out functions
21:35:25speachythe 'nprocs' thing is easy enough to work around; it's just the number of cores available. But I'm sure there will be plenty more issues.
21:36:38LambdaCalculus37speachy, I'm running OpenBSD 6.7 in a VM on my ThinkPad X220, which is only a dual-core i5-2520M, so I've given only one core to the VM.
21:37:35speachythere's no strategic benefit to rockbox to enabling *BSDs as a build host, especially if it comes at the expense of maintainability.
21:38:03LambdaCalculus37In my case, I just wanted to see if it could be done.
21:38:49speachyOh, I'm sure it's doable, just a question of how much effort will it require. IMO trying to avoid gmake is probably a lost cause
21:40:20LambdaCalculus37Hey, be glad I'm not trying this under something like 9front. :P
21:40:47speachyit's Free Software, far be it for me to say what you should do in your spare time. :)
21:41:52LambdaCalculus37True. I'm just tinkering and exploring again, but that tinkering and exploring is what got me to become a Rockbox dev in the first place. :)
21:42:22speachyI would just greatly prefer to see these rare and valuable developer-hours go towards making rockbox (as it runs on the players) better.
21:42:23__Bilgus_welcome back :)
21:43:08speachyindeed, welcome back. feel free to ignore this old grump.
21:44:00__Bilgus_speachy I want to try bumping the API do you have an idea of when you want to complete the PLAYER removal?
21:44:01LambdaCalculus37But moving onto other subjects: the fix added with rev 54629073ae has also made a nice difference in the performance in the SSD in my iriver H300.
21:44:28__Bilgus_try to avoid! jeeez
21:44:29speachy__Bilgus_ as far as I'm concerned, it's ready as soon as someone else reviews my work
21:44:58__Bilgus_ah ok i'll get at least another reviewed this eve
21:45:26speachyLambdaCalculus37, I'm glad that such a simple fix has proven to be more widely useful.
21:45:45speachydo you have a sense of how it's affected power consumption?
21:46:23speachysince it basically disables all power management, including killing power supplies on the host/controller side.
21:46:51LambdaCalculus37I haven't noticed a major difference in power consumption, but I should do a proper battery bench one day soon.
21:47:21speachywhat sort of SSD is it? ATA/CF->SD?
21:48:27LambdaCalculus37It's a proper PATA SSD. 1.8", 50 pin connector and all. It's a KingSpec KSD-CF18. 128GB.
21:48:44speachy__Bilgus_, I have made a few changes since you last looked, basically fixing up the breakages I'd introduced in the manual.
21:49:03LambdaCalculus37Bought it from Amazon about 4 years ago.
21:49:43LambdaCalculus37It basically is like a proper drop-in replacement. No adapters needed.
21:50:26speachyLambaCalculus37... and it doesn't support mandatory ATA power management. I wonder if you pop it open it won't turn out to be an eMMC chip bolted onto the side of an FC1307.
21:50:50speachycan you dump the ata_identity info so I can take a peek?
21:53:12LambdaCalculus37speachy, here ya go:
21:54:24speachygot it, thank you
21:59:37LambdaCalculus37Anything interesting in the identity info?
22:01:27 Quit livvy (Remote host closed the connection)
22:03:53Strife89speachy, have you had a chance to install and use that iPod Mini battery much?
22:04:16speachyI put the original microdrive back in, which meant the new battery didn't fit.
22:04:24Strife89Mine is showing possible signs of being unable to keep a charge while idle
22:04:41speachymy old battery goes dead if left alone for a couple of days
22:07:59Strife89Oh dear
22:08:28Strife89It's going haywire after I unplugged it from my desktop
22:15:00Strife89Starting to wonder if I managed to loosen or strip one of the battery wires when I made the "final" reassembly.
22:16:34Strife89If I *tilt* the device sufficiently, it shuts off and restarts.
22:17:12Strife89Sometimes it goes to the iPod's "low battery" screen for a moment (or internal charging screen if it's still plugged in)
22:18:36Strife89It's not consistent but it's happening on 3.15 (not that I think software has anything to do with it)
22:20:23speachyLambaCalculus37, definitly doesn't advertise power management.. and it also claims to be an SSD but has that word in the wrong endianness.
22:20:48speachyso ... yeah.
22:21:00***Saving seen data "./dancer.seen"
22:22:40speachyquite sloppy.
22:26:55Strife89 I was holding Select + Play
22:34:30Strife89I'm not sure whether it will be helpful, but I recorded some footage of the Mini's odd behavior. It will take all night to upload, though.
22:35:16Strife89I want to wait for an outside opinion before I start disassembling the Mini again, though.
23:05:07 Quit LambdaCalculus37 (Ping timeout: 246 seconds)
23:05:41 Join LambdaCalculus37 [0] (LambdaCalc@gateway/vpn/privateinternetaccess/lambdacalculus37)
23:09:09 Quit LambdaCalculus37 (Client Quit)
23:23:04 Quit ac_laptop (Ping timeout: 256 seconds)
23:33:30 Quit TheSeven (Disconnected by services)
23:33:39 Join [7] [0] (~quassel@rockbox/developer/TheSeven)

Previous day | Next day