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 | 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 2021-11-28

00:09:34 Join annoying_questio [0] (~annoying_@c-76-20-140-205.hsd1.mi.comcast.net)
00:13:08annoying_questioMaybe I should just find a small android type
00:15:43 Quit annoying_questio (Quit: Connection closed)
00:16:04braewoodsup to you. i was just looking at what's rockbox compatible.
00:16:06***Saving seen data "./dancer.seen"
00:16:30braewoodsthough not all android devices have an easy to remove sd card
00:29:46_bilgusone of those androids with the slide out kb would be perfect
00:30:50_bilgussince they are likely not working with carriers probable sub $30 area i'd imagine
02:00
02:16:10***No seen item changed, no save performed.
02:43:37 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:6150:bff7:9ca0:928)
03:00
03:20:07PaulFertserbraewoods: is fiio m3k really more problematic than the other x1000 boards (erosq)?
03:22:37PaulFertserftr, I have sansa e200 and it boots from internal memory and uSD card hotswap is working nicely.
03:25:16PaulFertserI'd say e200 works wonder if not for the varying background hum :/
03:30:34PaulFertserbtw, erosq seems to be available on newegg for little over 100 bucks but I think one shouldn't be buying a device from a company that violates GPL.
04:00
04:16:12***No seen item changed, no save performed.
04:18:13JanCjust buy it and then demand the sources from Newegg :)
04:56:20 Nick mendel_munkis is now known as munkis (~mendel_mu@ool-ae2cb218.dyn.optonline.net)
05:00
05:03:25 Join lebellium [0] (~lebellium@2a01cb04012c0900e825a5871cc31e54.ipv6.abo.wanadoo.fr)
05:04:50 Quit Rondom (Remote host closed the connection)
05:05:03 Join Rondom [0] (~rondom@user/rondom)
06:00
06:16:14***Saving seen data "./dancer.seen"
07:00
07:39:08 Quit amdj (Quit: quit (v.): to give up or resign, let go, relinquish)
07:39:13 Join amdj [0] (amdj@libera/staff/amdj)
07:46:35 Quit ZincAlloy (Quit: Leaving.)
08:00
08:07:21sporkwhat is problematic about the m3k? i use it daily
08:07:47sporkfiio tries to follow the gpl better than most
08:09:20PaulFertserspork: I was kind of scared by the list of issues on https://www.rockbox.org/wiki/FiioM3K compared to https://www.rockbox.org/wiki/Main/AIGOErosQK
08:09:56sporkamachronic provides a lot of detail
08:10:35sporkmost x1000 platforms are similar, some m3k issues might apply to others even though they are not listed
08:11:27spork*most x1000 devices that is
08:14:01sporkfrom what i read, all supported x1000-based players are very usable
08:14:35sporkboth mine (m3k/q1) have no meaningful issues
08:16:17***Saving seen data "./dancer.seen"
08:37:01PaulFertserGood to know, thank you!
08:38:55PaulFertser" The DAC makes popping noises when changing frequency. " this might get annoying I can imagine.
08:39:35PaulFertserAnd erosq has a different DAC.
08:45:21speachyThe hosted x1000-based units work well too FWIW.
08:54:31PaulFertserI am too much allergic to android to consider a hosted port.
08:55:22speachyFWIW, the x1000s use an android kernel but are otherwise "traditional" userspace.
08:55:37speachy(basically built on top of a ingenic-supplied kernel)
08:57:57PaulFertserI've read the sources for some hosted rockbox ports and that didn't fill me with joy at all. Not sure I can enjoy using a device knowing it runs such a downstream kernel without proper sources and with plenty of sysfs and other hacks instead of proper APIs.
09:00
09:12:35yangJust to get a confirmation, the Ipod headphones volume/up/down and Microphone on the cabling, dones't work on m3k?
09:28:24sporkx1000s use a linux kernel
09:28:50sporkhow is it related to android ?
09:33:00speachyPaulFertser: the system you're using to interact with this IRC channel is running on proprietary firmware.
09:33:20speachyperfection is the enemy of good.
09:33:37speachy(plus all the intermediate routers)
09:34:09speachy*every* single peripheral in or attached to the system I'm using is running non-free software. from the keyboard to the displays, and everything in between.
09:35:26speachyfar be it for me to say what you should or should not "enjoy using" but unless you build your own hardware (including the ASICs!) you're going to have non-free software running on some level.
09:36:12speachyin the mean time, all we can do is make do with what we have, and try to maximize the Freedom available to our users.
09:37:20speachyspork: the reference kernel supplied by Ingenic is an Android kernel. because Android was the primary target for these SoCs.
09:38:10speachyYes, Android is Linux, but it is/was a significant fork with a lot of non-mainline stuff in it, even before the hardware-specific stuff was added.
09:39:23sporki get that, but do you know how much android-related stuff survives ?
09:39:47sporki see the android config for the m3k, but not so much in the ingenic sources
09:40:09speachyspork: given they shipped it with a "traditional" Linux userspace (buildroot FWIW) it's rather moot.
09:40:19sporknot that i suppose it matters much
09:40:47sporkmy x1000's do not have hosted rockbox
09:40:55speachystuff we specifically take advantage of is adb and the much saner usb gadget approach of android
09:41:46speachythe m3k kernel is a dumpster fire.
09:41:59sporkyou mentioned that :)
09:42:12sporkespecially the excessive logging
09:42:29speachywhich is hilarious as they're the only ingenic-based folks to actually _attempt_ to comply with the GPL.
09:42:52sporkshanling does on request, but it is equally useless
09:43:23sporksomething outdated and incomplere
09:43:25sporkt
09:43:26speachythe fact that ingenic even publishes english documentation is highly unusual.
09:43:33speachy(or public documentation at all, honestly)
09:45:01speachybut fiio includes a copy of the GPL in the box, includes a reference to it on their "about" page on the player, etc.
09:45:26speachytheir internal processes are ...lacking but they at least tried, so good for them.
09:45:38sporki would buy a spare m3k if i could do at what i paid for my first
09:46:25speachyI suspect most of the other x1000 players don't even have the source code to (re)publish, as the base OS/platform was provided by HiBy
09:46:47speachy(and the player application too)
09:47:06PaulFertserspeachy: I know where there's proprietary code involved and where not. E.g. my wireless card is ath5k, no firmware, the wifi router is ath9k, no firmware (OpenWrt OS). When I have any choice at all I'm trying to avoid closed vendor codes at all costs.
09:47:10sporkhiby should have for the OS part
09:48:47speachypudding aside the jankiness of the m3k hosted kernel, most of the problems with our native port are our own fault.
09:49:18speachys/pudding/putting/ (gaah)
09:49:41sporkfiio m3 pro is also x1000e-based too, perhaps i should get amachronic one
09:51:14PaulFertserEven uSD cards have firmware and that often hurts, I'd prefer raw NAND to be exposed.
09:53:08speachymost of my F/OSS "career" has been spent reverse engineering hardware to make free software drivers (or firmware) feasible.
09:55:07PaulFertserRE is always hard and you never know if you're going to succeed. That's amazing you're doing what you do.
09:55:40speachythe "don't share unless forced to" mentality is pervasive.
09:56:58CtcpIgnored 2 channel CTCP requests in 3 minutes and 26 seconds at the last flood
09:56:58*speachy sighs.
09:58:15speachythe current DAP market is twofold; cheap shovelware based on the rknano/atj2137-class parts with <512K of RAM, or the equivalent of an ipod touch −− essentially an android phone minus the phone part.
09:58:21sporkshort term sharing does not make money
09:58:49speachylong-term sharing doesn't make money either, at least not in any easily-quantifiable way. at best it reduces costs.
09:59:16speachythe big tech companies (and even non-tech) know this, and use industry groups/associations to pool resources
09:59:24sporkit could improve reputation, but that might not show up in sales
10:00
10:00:16sporkso many examples of big companies taking from oss, not giving back
10:01:07speachycommercial use was _always_ based around economics −− go with whatever is cheaper. especially in the short term.
10:02:16speachymost organizations genuinely don't hav anything worth giving back.
10:03:34speachybut if they are thinking beyond a single product (which they rarely do) they might see that treating F/OSS as two-way means they can reduce their longer-term costs.
10:04:05speachybut that sort of thing is a lot harder to justify on the quarterly financial reports.
10:04:35sporkmanagers making decisions not to share probably have little faith their competitors actually will share
10:04:37speachy(my iminent layoff is a testament to that. as was the last one, incidently)
10:05:18speachyit actually has little to do with competitors IMO −− more to do with internal politics (which in turn are largrely driven by the metrics used to hand out financial rewards)
10:06:22speachywhere competitors come into play is more to do with paranoia "they'll steal our stuff, or see that we stole theirs" paranoia.
10:07:38sporkeach time i think there will be no future targets for rockbox. i suppose the x1000's were a bonus.
10:10:49speachythe tide is turning though, as the security realities (==liabilities) of IoT products are beginning to make proper software BOM tracking critically important
10:11:40speachyyeah, the ingenic stuff was an unexpected bonus. I don't know if we can expect another round of non-touchscreen DAPs.
10:13:07speachyI've tried reaching out to the EROS folks −− they released a Q II that uses a different HW/SW platform in the same physical form factor.
10:14:00speachythis goes back to the "having our own player produced" pipe dream
10:14:48speachybut if we could place a sufficiently large order, we could ask for appropriate source code/documentation and expect to get it.
10:15:09speachyit would be a lot cheaper than designing/manufacturing somethign from scratch
10:15:31speachy(especially the case..)
10:16:20***No seen item changed, no save performed.
10:17:35sporkhard to avoid touchscreens
10:17:59speachyIMO physical controls are a hard requirement.
10:19:16sporki prefer buttons, but a combination is ok if there at least 3-4 physical buttons too
10:20:29speachyI'm thinking primarily of blind folks, but that same approach benefits everyone if the device can be operated without looking at the screen.
10:22:16speachyhttps://www-eroshifi-com.translate.goog/news/130.html?_x_tr_sl=zh-CN&_x_tr_tl=en&_x_tr_hl=en-US&_x_tr_pto=nui&_x_tr_sch=http
10:25:08PaulFertserMy first portable flash player didn't have any display at all, just some kind of joystick which was essentially 3 buttons in one. Worked from a single AAA cell.
10:26:04speachyI can try to get one (via aliexpress or something) with the intent of taking it apart
10:30:19sporkdo you know what soc it uses ?
10:30:52PaulFertserI can't even find a photo of one, it was probably 15 years ago. It had about 128 MiB internal flash IIRC.
10:31:46speachyspork: nope, only they claim it's their own and runs an in-house RTOS vs Linux.
10:32:41speachybased on the capabilities it's much more likely to be in the x1000 level of capabilities rather than a microcontroller.
10:33:29spork'eros e2' is not helpful
10:33:32speachyyeah
10:33:58sporkyeah, bluetooth, dac, etc all sounds similar to x1000 platform players
10:34:02speachyPaulFertser: yeah, I recall a few of those things
10:34:11sporkmaybe their rtos is rockbox
10:34:12speachyspork: it's definitely not linux though
10:34:30speachyand highly unlikely to be rockbox too. :)
10:34:52sporkthat would catch several birds at once
10:36:06speachythere's a "Q2 touchscreen" variant, that's probably x1000-based.
10:36:11speachy(hibyos)
10:36:51sporkthey like that form factor
10:43:22speachyyeah
10:46:28speachyANYway.
10:51:16PaulFertserI first heard about Ingenic when they released some electronic ads in Vogue or Playboy magazines.
11:00
11:00:21speachythere are some allwinner parts of roughtly the same capabilities (DRAM-on-package)
11:02:14 Join dconrad [0] (~dconrad@152.117.104.224)
11:02:20 Quit dconrad (Remote host closed the connection)
11:15:54 Join advcomp2019__ [0] (~advcomp20@user/advcomp2019)
11:19:28 Quit advcomp2019_ (Ping timeout: 265 seconds)
11:22:00_bilgusPaulFertser, at this point we are just looking for viable hardware open or not we don't really have a luxury of being picky we just need hardware to stay relevant
11:25:10_bilgusIve been working on this keyremap plugin for the last week, I really need to come up with a better input method than lists
11:27:22PaulFertser_bilgus: I wonder if Rockbox can collaborate with Pine64 on creating a superb device. They seem to have all the necessary connections and are pro-free software.
11:27:55_bilguswe are already trying that but its too low on RAM for us
11:28:17PaulFertserIt being what exactly?
11:28:32speachythey've proposed two things; essentially a phone-less variant of the pinephone, and something based on a BT headphone ASIC that only has 1/2 the RAM we need to be viable.
11:28:36_bilgusit wouldnt be superb itd be a neutered mess
11:28:41 Join dconrad [0] (~dconrad@152.117.104.224)
11:28:45 Quit dconrad (Client Quit)
11:29:03speachy(or was thet 3/4? either way we'd have to make drastic cuts in functionality)
11:33:45PaulFertserSo it might become viable when a newer version of that BT ASIC is released or when probably another suitable SoC is found. They seem to have all the capabilities to design and manufacture proper cases, and they have huge audience.
11:34:19speachybased on earlier discussions the case is the biggest problem due to up-front tooling costs
11:35:40 Join amachronic [0] (~amachroni@user/amachronic)
11:35:56PaulFertserIndeed, and they already have the expertise. I do not think the pinephone keyboards are numerous and yet they managed it somehow.
11:38:43PaulFertserAn alternative would be to design a replacement motherboard for some cheap mass players that could be bought just for their cases.
11:39:36 Part frosteyes (WeeChat 2.8)
11:39:37PaulFertser(for the first run, and then the right connections might allow to secure a deal with the ODM for the cases alone)
11:40:45rb-bluebotBuild Server message: New build round started. Revision 1ad60c3344, 303 builds, 11 clients.
11:50:22rb-bluebotBuild Server message: Build round completed after 578 seconds.
11:50:25rb-bluebotBuild Server message: Revision 1ad60c3344 result: All green
11:56:49amachronicPaulFertser, I wrote those issues on the m3k wiki page over a year ago and they aren't as bad as they sound
11:57:33amachronicpopping on frequency change mostly affects the > 96 KHz range in practice
11:59:02amachronicwonky touchpad is a real but minor annoyance, the rest is more technical problems which don't really affect day to day usage
11:59:44amachronic(that's why it's also "TODO list" :)
12:00
12:01:25amachronicthe main downside to the m3k IMO is that in the long run it might not be that easy to repair
12:05:28rb-bluebotBuild Server message: New build round started. Revision dbb7371065, 303 builds, 11 clients.
12:05:37 Join ZincAlloy [0] (~Adium@ip5f5abcae.dynamic.kabel-deutschland.de)
12:14:15 Join dconrad [0] (~dconrad@152.117.104.224)
12:15:18rb-bluebotBuild Server message: Build round completed after 590 seconds.
12:15:19rb-bluebotBuild Server message: Revision dbb7371065 result: All green
12:16:22***Saving seen data "./dancer.seen"
12:18:03PaulFertseramachronic: hey, thank you for the info. What kind of repair do you mean?
12:32:32 Quit dconrad (Remote host closed the connection)
12:34:04amachronicPaulFertser: I mean after many years it's not likely to be repairable like the ipods are
12:35:15amachronicbut I never took mine apart (I lack the tools to properly reassemble it) so take my advice with a grain of salt
12:35:21amachronic(my m3k I mean)
12:40:00sporkhppefully the battery will not die too fast
12:45:34sporkthe instructions on the wiki are very clear and completely outside my skills
13:00
13:01:19 Nick vup2 is now known as vup (~~~~@46.101.193.235)
13:23:34yangJust to get a confirmation, the Ipod headphones volume/up/down and Microphone on the cable, doesn't work on m3k?
13:23:39PaulFertserHm, disassembly instructions look similar to what one does with modern smartphones.
13:26:17speachyyang: correct, the m3k lacks that functionality.
13:26:58braewoodsspeachy: given everything it wouldn't surprise me if rockbox's future could end up being as a Linux distribution
13:27:12braewoodsit's getting more difficult to keep up with new native ports
13:27:23 Quit amachronic (Ping timeout: 265 seconds)
13:27:38speachybraewoods: quite possibly, yes. but managing a "native" linux distribution also requires more than we traditionally have available too.
13:27:58braewoodsindeed but we'd be more like openwrt if that's where we ended up
13:29:15sporkopenbox
13:29:41yangthanks for clarification speachy
13:31:35PaulFertserOpenWrt devs spend great deal of time on Linux programming, so it's not just a distro.
13:32:16braewoodsi'm well aware. it's just the closest one to what rockbox is.
13:33:16PaulFertserAnd probably it's actually a good project to base development of such new ports on.
13:34:35 Join dconrad [0] (~dconrad@152.117.104.224)
13:34:57PaulFertserIt's nice for constrained systems, and the build system is more advanced than regular buildroot.
13:37:14braewoodsIt's optimized for networking devices so not totally correct.
13:38:02PaulFertserWhat exactly makes it optimised better for networking devices than for a DAP?
13:38:24braewoodseverything in the design is optimized for router or similar use
13:38:34braewoodsit assumes you have wifi, etc, which a DAP may not have
13:39:29PaulFertserThe recent port to realtek-based managed switches works just fine without wifi (it's a wired ethernet switch after all). hostapd and kernel drivers are just not included.
13:39:31braewoodsif anything Alpine would probably make more sense
13:41:14braewoodswell a DAP is pretty much guaranteed to lack ethernet
13:42:11braewoodsin any case
13:42:28braewoodsit's somewhat moot until there's a device worth managing like this
13:42:46PaulFertserOpenWrt uses procd which consumes way fewer resources than systemd. And OpenWrt doesn't require any networking to run, really.
13:43:44 Join amachronic [0] (~amachroni@user/amachronic)
13:47:44amachronicA 'Linux port' to the x1000 would be feasible, but only useful for the bluetooth and wifi-enabled targets.
13:48:33amachronic(a lot of the low level stuff is already upstream and what isn't can be easily ported from Rockbox)
13:48:49braewoodsi've just been thinking our own Linux managed stack would be desireable if that's even practical, not being stuck with whatever the vendor used
13:49:28 Quit dconrad ()
13:49:33braewoodsbut hit or miss with ARM linux
13:49:40amachronicagreed it'd be useful but it significantly shifts the scope of the project
13:49:57amachronicplus it's no good for the older MMU-less devices
13:50:22braewoodsit'd probably deserve its own repo
13:50:37braewoodssince it's far removed from regular rockbox
13:50:50braewoodsrockbox could target this if we managed it somehow
13:51:41amachronicthen there's the problem that the hosted version of RB isn't exactly great
13:52:08amachronicit basically emulates a 100 Hz kernel tick to get the co-operative threading model
13:52:29amachronicand it uses a lot of polling when it could take more advantage of event-based interfaces
13:54:00amachronicand there's less RAM available too
13:54:13braewoodshuh. i found an interesting option.
13:54:24braewoodsthough kinda bulky
14:00
14:00:15 Quit amachronic (Ping timeout: 265 seconds)
14:04:42speachyamachronic: yeah, I have some WIP code to replace the soft threads with native linux threads.
14:04:55speachythat alone will make a huge difference
14:09:49 Join amachronic [0] (~amachroni@user/amachronic)
14:16:25***Saving seen data "./dancer.seen"
14:20:29 Quit amachronic (Ping timeout: 256 seconds)
14:41:47 Join jschwart_ [0] (~quassel@2001:985:2c6e:0:b00b:32ff:fe28:5567)
14:41:50 Quit jschwart (Ping timeout: 268 seconds)
14:45:09tertuis the mmuless branch of linux still around
14:47:58 Nick jschwart_ is now known as jschwart (~quassel@2001:985:2c6e:0:b00b:32ff:fe28:5567)
14:54:22PaulFertsertertu: it's upstream, called "nommu"
15:00
15:26:13tertuthat seems useful
15:33:13 Quit munkis (Remote host closed the connection)
15:33:37 Join munkis [0] (~mendel_mu@ool-ae2cb218.dyn.optonline.net)
16:00
16:16:29***Saving seen data "./dancer.seen"
16:30:53 Quit munkis (Remote host closed the connection)
16:34:45 Join munkis [0] (~mendel_mu@ool-ae2cb218.dyn.optonline.net)
17:00
17:35:50 Quit munkis (Remote host closed the connection)
17:42:15 Join munkis [0] (~mendel_mu@ool-ae2cb218.dyn.optonline.net)
17:45:55 Quit skipwich (Quit: DISCONNECT)
17:46:08 Join skipwich [0] (~skipwich@user/skipwich)
17:52:00 Quit lebellium (Quit: Leaving)
18:00
18:16:33***Saving seen data "./dancer.seen"
19:00
19:58:36 Quit ZincAlloy (Quit: Leaving.)
20:00
20:16:37***Saving seen data "./dancer.seen"
21:00
21:02:07 Quit TorC (Quit: Vanishing into the mist)
21:02:25 Join TorC [0] (~Tor@fsf/member/TorC)
22:00
22:16:40***Saving seen data "./dancer.seen"

Previous day | Next day