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 2024-10-10

01:00
01:45:03***No seen item changed, no save performed.
02:00
02:19:20 Quit spork (Quit: leaving)
02:53:39__builtinspeachy: you can't decode in chunks and scale on the fly?
02:53:47__builtin(also, hello.)
02:53:52 Join spork [0] (topic@77-164-207-8.fixed.kpn.net)
03:00
03:28:10edhelas_speachy That's what progressive JPEG is all about, afaik you can stop decoding when your resolution is "good enough"
03:28:13edhelas_https://stackoverflow.com/questions/33539225/prevent-progressive-jpeg-from-loading-completely
03:29:25edhelas_Most of the Rockbox devices have small resolution screen, so even the first pass could be "good enough". Higher resolution devices also tend to have higher memory and processing power.
03:45:05***Saving seen data "./dancer.seen"
03:51:51 Join jj5 [0] (~jj5@100.80.216.139.dynamic.dsl.dv.iprimus.net.au)
05:00
05:04:18 Quit jacobk (Ping timeout: 248 seconds)
05:04:53 Join jacobk [0] (~quassel@47-186-105-237.dlls.tx.frontiernet.net)
05:45:10***Saving seen data "./dancer.seen"
06:00
06:11:33 Quit PheralSparky (Read error: Connection reset by peer)
07:00
07:18:03 Quit cnx (Ping timeout: 252 seconds)
07:26:14 Join cnx [0] (~cnx@tem.loang.net)
07:45:14***Saving seen data "./dancer.seen"
07:52:38 Join Trzyzet [0] (~Trzyzet@88.97.248.229)
07:53:51speachyedhelas_: It doens't work terribly well in practice. Even Apple's pictureflow stuff relied on pre-scaled stuff.
07:55:03speachy"higher resolution devices" refers to stuff that's not a microcontroller-based DAP.
07:55:55speachyand that patch throws out the existant scaling for non-progressive images, which also makes it a nonstarter.
07:59:23speachy__builtin: you described the status quo; taking advantage of joeg's 8x8 block size.
08:00
08:00:30speachyIIRC progressive decoding has each pass buildong on the other; ie if you scale as you go, latter passes won't work properly.
08:04:55speachyif the first pass is "good enough" then great, but things fall apart otherwise.
08:07:18speachy(as for PNG, we'd have to essentially write our own scanline-oriended decoder. Not impossible, but we're not exactly swimming in developers)
08:09:11speachy(Though IIRC the main problem with PNG is its use of gzip compression that doesn't do so well with strict constraints on memory use)
09:00
09:15:57 Join dconrad [0] (~dconrad@152.117.104.217)
09:18:40dconradspeachy: yeah I saw that erosq_2024, kinda works out to give us some safety for our bootloader update file
09:19:21dconradI hope to find some time this weekend to decompile this new update file... it's been a while since the first time so it'll be good practice :D
09:21:18dconradNot sure if they changed power management chip or just moved some wires around on gpios or what...
09:21:55dconradit's pretty surprising they changed the linux battery reporting as well
09:22:53 Join dys [0] (~dys@user/dys)
09:35:52speachyyeah.. I wonder if they rebvased to a newer kernel release or something? seems like a fairly gratitutious change to rename 'battery' to 'axp_battery'
09:36:17speachystill 3.10.14
09:36:53speachyI'll get that change into the hosted port and we'll see what else breaks
09:37:26speachy(I half expect the ALSA controls to be renamed or something..)
09:39:43dconradaxp_battery is good though, they're still using axpXXX
09:45:17***No seen item changed, no save performed.
09:48:22dconradany specifics (lcd, axp version, dac, x1000[E/NoE]) we can learn from adb?
09:50:05speachyI don't know offhand; it's been a while.
09:50:41speachy(could ask for the usual assortment; dmesg, alsactl dump, etc..)
09:51:11dconradah
09:55:22speachybbiab
09:55:23 Quit speachy (Quit: WeeChat 4.4.2)
10:00
10:04:08 Join speachy [0] (~speachy@rockbox/developer/speachy)
10:04:08Mode"#rockbox +v speachy" by ChanServ (ChanServ@services.libera.chat)
10:14:05rb-bluebotBuild Server message: New build round started. Revision edbcf0b0f4, 345 builds, 10 clients.
10:14:05rb-blueboterosq: Initial PM support for "hw4" variant by Solomon Peachy
10:19:27dconrad... the cynical side of me wonders if this is someone spinning new changes just to seem busy to their boss... I can respect that
10:19:53dconradit's annoying, but I can respect it haha
10:20:18speachyhopefully I didn't break existing players with that. :D
10:20:44dconraddon't worry, you'll know haha
10:23:18speachyneed to un-bury my v1 hardware from the bottom of my rockbox box
10:34:40 Quit dconrad (Remote host closed the connection)
10:34:45rb-bluebotBuild Server message: Build round completed after 1241 seconds.
10:34:47rb-bluebotBuild Server message: Revision edbcf0b0f4 result: All green
10:58:02speachydconrad: I doubt they'd revise the PCB just to appear busy. :D
11:00
11:05:03sporkmoney
11:05:12speachycosts less money do do nothing
11:05:32sporknot if you can source cheaper parts or have run out of stock and cannot source the original
11:07:03speachyof course, but that's not "appear busy" work
11:10:08sporktrue
11:45:20***Saving seen data "./dancer.seen"
11:58:54 Quit braewoods (Read error: Connection reset by peer)
11:59:03 Join braewoods_ [0] (~braewoods@user/braewoods)
12:00
12:02:38 Quit _bilgus (Remote host closed the connection)
12:02:49 Quit Trzyzet (Quit: The Lounge - https://thelounge.chat)
12:03:06 Join Trzyzet [0] (~Trzyzet@88.97.248.229)
12:03:46 Join _bilgus [0] (~bilgus@syn-162-154-213-134.res.spectrum.com)
12:04:56 Quit Trzyzet (Client Quit)
12:05:12 Join Trzyzet [0] (~Trzyzet@88.97.248.229)
12:07:36 Quit Trzyzet (Client Quit)
12:08:48 Join Trzyzet [0] (~Trzyzet@88.97.248.229)
12:39:43 Quit Trzyzet (Quit: The Lounge - https://thelounge.chat)
12:41:34 Join Trzyzet [0] (~Trzyzet@88.97.248.229)
12:46:46 Quit Trzyzet (Quit: The Lounge - https://thelounge.chat)
12:47:34 Join Trzyzet [0] (~Trzyzet@88.97.248.229)
13:00
13:01:13 Join lebellium [0] (~lebellium@2a01cb0405d07f0035f3c8468153dc53.ipv6.abo.wanadoo.fr)
13:02:35speachydconrad: at least with the AXP20x vs AXP192 we should be able to detect the difference at runtime. :)
13:04:23speachymight be able to save some juice too by turning off HP/LO when they're not plugged.
13:05:03speachy(I only have a rev1 hw, I'm curious what rev2 shows)
13:07:08speachyhmm, hosted port should use the HW volume instead of SW volume on v2+.
13:08:19 Join dconrad_ph [0] (~dconrad_p@108.147.197.72)
13:09:21dconrad_phspeachy: interesting, I was under the impression the hosted port always did its own volume scaling rather than calling alsa's volume adjustment
13:10:29speachydepends on the DAC that was used. the other hibyos targets use hw controls, the sony targets are all sw...
13:14:29dconrad_phwell I can get a dmesg output sometime from my player, I'll just have to put the hosted bl on it as well
13:16:40 Quit dconrad_ph (Quit: Client closed)
13:31:23speachynow have a patch that uses HW volume control on the v2+ devices.
13:37:57 Join dconrad_ph [0] (~dconrad_p@205.237.113.8)
13:38:31speachyrev1 doesn't log gpios like rev4 does
13:38:59dconrad_phwell for all the work its been, my one big complaint with the hosted was not having hwvol... it certainly is easier to maintain the hosted...
13:39:56dconrad_phi think that was the whole reason to do the native port haha
13:42:18speachyAXP192/20x are on the same i2c address, and report ID 0x03 and 0x4a, respectively.
13:42:34dconrad_phoh thats nice
13:45:05speachycan't find corrobating info
13:45:23***Saving seen data "./dancer.seen"
13:49:16speachyhmm.
13:58:14 Quit dconrad_ph (Quit: Client closed)
13:58:56 Join dconrad_ph [0] (~dconrad_p@205.237.113.8)
14:00
14:00:13speachyso g#5963 is there for anyone wanting to try it
14:00:16rb-bluebotGerrit review #5963 at https://gerrit.rockbox.org/r/c/rockbox/+/5963 : erosqhosted: Support HW volume control on rev2 hardware by Solomon Peachy
14:01:00speachyit doesn't regress on rev1 hardware
14:06:06dconrad_phI can try that in the next couple days, is there a debug menu readout for which system it detected?
14:06:34speachynope! :D
14:10:12dconrad_phhaha well that should be pretty easy to add I think
14:10:40speachyso with this I know how to distinguish between rev1, rev2/3 and rev4.
14:10:59dconrad_phamazing!
14:11:11speachyonly indirectly, of course.
14:11:37 Join PheralSparky [0] (~S|h|a|w|n@user/shawn/x-4432647)
14:15:56speachy...If I order a new F20 from Amazon now, will I get a v3.4 player? Ugh.
14:16:07 Quit dconrad_ph (Ping timeout: 256 seconds)
14:16:51 Join dconrad_ph [0] (~dconrad_p@205.237.113.8)
14:17:13dconrad_phI was wondering the same thing myself tbh
14:25:27speachywould be nice to be able to tell the difference between v2/v3 as well.
14:25:54speachy(I don't think anyone shared a dmesg dump that showed the display info)
14:48:00 Quit dconrad_ph (Quit: Client closed)
15:00
15:15:13 Join dconrad_ph [0] (~dconrad_p@205.237.113.8)
15:15:14 Quit dconrad_ph (Client Quit)
15:24:06 Quit jacobk (Ping timeout: 276 seconds)
15:24:20 Join jacobk_ [0] (~quassel@47-186-105-237.dlls.tx.frontiernet.net)
15:45:24***Saving seen data "./dancer.seen"
16:00
16:26:30 Quit jacobk_ (Ping timeout: 276 seconds)
17:00
17:00:21 Join Moriar [0] (~moriar@107-200-193-159.lightspeed.stlsmo.sbcglobal.net)
17:45:28***Saving seen data "./dancer.seen"
17:54:16 Quit lebellium (Quit: Leaving)
18:00
18:11:40 Join dconrad [0] (~dconrad@152.117.104.217)
18:13:45dconradwith how easy it was to add support for the hw4 erosq, I am starting to question whether the effort to maintain the native port is worth it...
18:14:22dconradI might take a second look at the hosted port if that hwvol works as intended
18:14:34dconradfighting sunk cost fallacy hard here
18:16:46dconradbut first, I'll get that dmesg etc. output
18:19:49speachythe hosted port is not that great; we only have access to about 1/4 of the RAM, the UI performance is sluggish, boot times are crappy..
18:21:08dconradah, I thought it was on par
18:21:22dconradperformance-wise
18:21:35dconradmaybe it is worth it
18:22:14speachyraw oomph it's roughly equivalent (really, a 1ghz mips core hides a lot of sins) but there's a lot of overhead due to linux itself plus the non-optimized underlying platform drivers.
18:22:28dconradI see
18:22:46speachygranted, we get some nice things (ntfs, exfat, bluetooth in theory, floating point codecs)
18:24:12dconradMaybe these dmesg etc. outputs would be good to attach to the wiki page
18:24:13speachybut the userspace platform is postively ancient, and while there's definitely some paths to better optimization we'd have to write a new threading layer first.
18:25:19speachy(that would give us true multithreading, allow use of a blocking ALSA API that would enable passing audio to the BT layer, not get blocked on the slow UI updates, etc etc.
18:25:39speachy(would also eat into our already-limited RAM headroom)
18:26:28speachyre dmesg and alsa dumps, yes.
18:26:29dconradany other adb stuff you're interested in on the v2 player? I've got it open and running right now
18:26:48dconradI got dmesg and "amixer contents"
18:27:06speachythat's what comes to mind.
18:27:58dconradalright, it's easy to do if any other info is needed
18:28:22dconradI'll put it up on the wiki
18:28:35speachythank you
18:29:33speachytbh I also want to do a native port to the xduoo x3ii, but I don't relish the prospect of REing how it's put together given that I may end up being the only one to use it.
18:31:29speachyI'll take a look later tonight to see what I can glean from those dumps.
18:38:30dconradI'm also thinking about pulling the trigger on a hw4 unit (or, more accurately, pulling the trigger on rolling the dice to get a hw4 unit)
18:48:40dconradwell it looks like if they took the time to remove networking stuff etc. from the linux kernel it could be /so/ much leaner
18:54:17dconradI still dream of having usb-dac mode in rockbox... I was just thinking of that recently. It could be a plugin, but I suspect decoupling a plugin to allow it to do its own usb thing might be a struggle
18:54:56dconrada personal usb dsp box, if you will
19:00
19:17:45 Quit dconrad (Remote host closed the connection)
19:29:08 Join dconrad [0] (~dconrad@152.117.104.217)
19:38:02 Quit dconrad ()
19:45:32***Saving seen data "./dancer.seen"
19:49:01 Join massiveH [0] (~massiveH@2600:4040:a982:dc00:c2c:fc6c:960e:54c5)
21:00
21:39:35 Join jacobk [0] (~quassel@129.110.242.224)
21:44:04 Quit jacobk (Ping timeout: 260 seconds)
21:45:19 Join jacobk [0] (~quassel@129.110.242.173)
21:45:34***Saving seen data "./dancer.seen"
21:51:39 Quit jacobk (Ping timeout: 260 seconds)
21:54:04 Join jacobk [0] (~quassel@utdpat242024.utdallas.edu)
22:00
22:06:27 Quit jacobk (Ping timeout: 276 seconds)
22:11:14 Quit Moriar (Quit: Leaving.)
22:50:03 Join decky_e [0] (~decky_@69.9.139.14)
23:00
23:02:42 Quit massiveH (Quit: Leaving)
23:38:05 Join decky [0] (~decky_@69.9.139.14)
23:39:07 Quit Natch (Read error: Connection reset by peer)
23:39:46 Quit sam_d (Quit: Bye)
23:39:55 Quit wLLm (Remote host closed the connection)
23:39:58 Quit Bobathan_ (Quit: ZNC 1.8.2+deb2+b1 - https://znc.in)
23:40:13 Join wLLm [0] (~wLLm@wllm.connected.by.freedominter.net)
23:40:54 Quit jj5 (Remote host closed the connection)
23:41:19 Join jj5 [0] (~jj5@100.80.216.139.dynamic.dsl.dv.iprimus.net.au)
23:41:47 Join Bobathan_ [0] (~admin@syn-065-029-248-157.res.spectrum.com)
23:41:50 Quit decky_e (Ping timeout: 252 seconds)
23:42:32 Join sam_d [0] (~sam@user/sam-d/x-8933526)
23:44:02 Quit dys (Ping timeout: 252 seconds)
23:44:02 Quit baltazar (Ping timeout: 252 seconds)
23:45:38***Saving seen data "./dancer.seen"
23:49:52 Join Natch [0] (~natch@c-9e07225c.038-60-73746f7.bbcust.telenor.se)

Previous day | Next day