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 2020-07-23

00:15:58 Join advcomp2019__ [0] (
00:15:58 Quit advcomp2019__ (Changing host)
00:15:58 Join advcomp2019__ [0] (~advcomp20@unaffiliated/advcomp2019)
00:18:25 Quit advcomp2019_ (Ping timeout: 240 seconds)
00:21:01***Saving seen data "./dancer.seen"
00:45:38 Quit akaWolf (Ping timeout: 244 seconds)
01:59:58 Quit paulk-leonov (Quit: Leaving)
02:00:13 Join paulk-leonov [0] (
02:21:05***Saving seen data "./dancer.seen"
04:21:07***No seen item changed, no save performed.
04:24:44 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:a051:3387:ec2:ec2a)
04:27:26 Join Rower [0] (
04:58:42 Quit michaelni (Ping timeout: 256 seconds)
05:11:44 Join michaelni [0] (
06:20:28 Join sakax [0] (~r0b0t@unaffiliated/r0b0t)
06:21:09***Saving seen data "./dancer.seen"
06:50:42 Join vmx [0] (
07:25:36 Join ac_laptop [0] (~ac_laptop@
07:26:52 Part edhelas
07:27:30 Join edhelas [0] (
07:28:21 Join krabador [0] (~krabador@unaffiliated/krabador)
07:29:48efqwAbout the controlled build vm, you could mostly do away with some kind of ubuntu, and possibly support everything else on a best-effort basis.
07:31:28efqwUbuntu is not the best distribution for various reasons that I'm not going to get into, but I still trust it the most for building large software projects.
07:31:28 Quit krabador (Remote host closed the connection)
07:36:49speachyonce an "official" build environment is created, then everything else bitrots and effectively all development of build environments ceases.
07:37:40speachyplus, speaking personally, I don't actually use ubuntu for anything. rubs me the wrong way on multiple levels. it's all the quirks of debian (and then some) minus the principles that make it worth putting up with
07:39:32speachyI've put a lot of work into fixing issues with the build environment; arguably all that's needed is a "cut-n-paste this apt-install cmdline to get everything that isn't isntalled by default"
07:40:07speachyand that cmdline/list can be maintained by folks who actually use that distro.
07:40:49speachywe build everything from source, even the toolchains, in no small part because we don't want to have to depend on anything. (Other than GNU tools.. :D)
07:48:35speachy(at $dayjob I'm dealing with a _very_ brittle tool that only functions properly on a single ubuntu release. And even then only bog-default configurations. and also requires passwordless sudo to be available..)
07:59:25 Quit edhelas (Quit: Gateway shutdown)
08:03:34 Join edhelas [0] (9d94237298@2a01:7c8:aab8:6b9:5054:ff:fec9:fd84)
08:11:21 Join krabador [0] (~krabador@unaffiliated/krabador)
08:21:10***Saving seen data "./dancer.seen"
08:30:47efqwFYI the ingenic X1000E SDK still asks for ubuntu trusty.
08:31:17efqwI ignored it and went with xenial, and everything still worked.
08:35:34speachyphew, looks like they rely on a hacked gcc 4.7
08:36:00speachyso the odds are we won't trip over unexpected behaviour from a newer toolchain
08:36:43speachywe'll eventually want to build our own toolchain from source, which shouldn't be a problem.
08:37:13speachy(since we don't use the xburst extensions that were never contributed to gcc upstream)
08:37:33 Join massiveH [0] (
08:38:44speachyhuh, looks like Ingenic just announced the X2000
08:39:41speachydual core, up to 1.5GHz, plus a small realtime "security" core that application writers don't get to touch.
08:40:07speachyaka the Halley5
08:42:13speachy128 or 256MB RAM on-package
08:46:10speachyaaand looks like a bot is trying to create a bunch of spammy accounts in the bugtracker
08:50:41efqwI've read about the X2000 last month, I wonder what kind of devices will they show up in.
08:50:54speachymore IoT shovelware? :)
08:51:44efqwPretty much. And if we were lucky, more DAPs.
08:55:00efqwInterestingly x1830 has been used in one of those retro emulator consoles lately.
08:55:05speachytouchstreen, wifi-connected, app-enabled "DAP"s
08:55:43efqwWell, probably no apps but it'll be touchscreen no doubt
08:56:44speachydepends on the price then. unless the X2K is cheaper than the X1K there's no reason to use it.
08:57:01speachybecause if you're not going to use the fancy capabilities it gives, why incur the price?
08:58:04efqwThe x1000e DAPs will probably be around for a while, just like the old jz4760[b] stuff.
08:58:24efqwSo it would be quite nice to have rb on more x1000e players.
08:58:46 Join J_Darnley [0] (
09:00:49 Quit jdarnley (Ping timeout: 264 seconds)
09:01:16efqwI'll need to whip something up to verify the key maps defined in their kernel source first.
09:45:51 Quit koniu (Remote host closed the connection)
09:46:17 Join koniu [0] (~koniu@gateway/tor-sasl/koniu)
10:16:26 Quit vmx (Ping timeout: 256 seconds)
10:16:45 Join nwz-e384_user [0] (
10:18:32 Quit massiveH (Quit: Leaving)
10:21:12***Saving seen data "./dancer.seen"
10:54:00 Join vmx [0] (
11:23:46 Quit nwz-e384_user (Remote host closed the connection)
11:54:30 Quit __Bilgus_ (Remote host closed the connection)
12:21:13***Saving seen data "./dancer.seen"
12:25:48 Quit ShadowHntr (Ping timeout: 256 seconds)
12:27:57 Join ShadowHntr [0] (
12:28:22 Nick ShadowHntr is now known as Guest81860 (
12:29:17 Quit krabador (Remote host closed the connection)
12:32:43 Quit vmx (Quit: Leaving)
13:35:02 Join MrZeus_ [0] (
14:14:23 Join LambdaCalculus37 [0] (LambdaCalc@gateway/vpn/privateinternetaccess/lambdacalculus37)
14:14:50LambdaCalculus37speachy, ping!
14:16:59LambdaCalculus37I'm not on the mailing list but I saw you're looking for people to help test builds with the newer toolchain components.
14:17:14LambdaCalculus37I can do a test with my H300 so you can have an m68k target tested.
14:18:04speachyI think we're actually okay on the m68k side of things. The problem child turned out to be portalplayer ARMs.
14:19:06LambdaCalculus37Ahh, okay.
14:20:39speachyipodcolor is still broken with -Os. That's the main holdup.
14:21:16***Saving seen data "./dancer.seen"
14:21:17speachyonce archos/sh is gone I'll pick that stuff back up again.
14:22:00speachyactually there is one thing you could do −− take a current build and run the codecbench on an opus file
14:24:07speachy(on the h300 I mean)
14:25:24 Join S|h|a|w|n [0] (~shawn156@unaffiliated/shawn156)
14:45:15LambdaCalculus37I haven't got any opus encoded files.
14:48:05Strife89You will in a minute. :)
15:01:46 Join mendelmunkis [0] (
15:03:51 Quit mendel_munkis (Ping timeout: 258 seconds)
15:10:38LambdaCalculus37Strife89, thanks!
15:19:20 Quit dweeber_ (Read error: Connection reset by peer)
15:21:17LambdaCalculus37Building on the current version with test codecs enabled now.
15:32:33 Nick mendelmunkis is now known as mendel_munkis (
15:33:06 Quit Guest81860 (Quit: Lost terminal)
15:40:17LambdaCalculus37speachy, first Opus file Strife89 gave gave the following results with test_codec: Decode time 57.13s, File duration 106.16s, 185.82% realtime, 66.83 MHz needed for realtime.
15:40:36LambdaCalculus37That's regular speed test.
15:41:10speachythank you
15:42:19LambdaCalculus37With DSP test, same file decoded in 64.88s and was at 163.62% realtime.
15:42:36 Join __Bilgus_ [0] (41ba23be@
15:43:25__Bilgus_speachy g#2551 has a few things that are suspect
15:43:27fs-bluebotGerrit review #2551 at : [4/4] Remove HAVE_LCD_BITMAP, as it's now the only choice. by Solomon Peachy
15:43:33speachyI saw the email
15:43:51speachyI also added g#2557 removing a few more stragglers
15:43:53fs-bluebotGerrit review #2557 at : do_menu pass internal synclist reference to callback by William Wilgus
15:44:03speachyer, g#2577
15:44:05fs-bluebotGerrit review #2577 at : [5/4] Remove the archos-specific flash and uclpack tools by Solomon Peachy
15:45:57__Bilgus_does that affect anything else that uses ucl?
15:46:22speachynothing seemed to use it based on a quick grep.
15:46:25__Bilgus_I was under the impression all the sansas use ucl pack / upack
15:47:04__Bilgus_maybe there is another route or something?
15:49:11 Join livvy [0] (~livvy@gateway/tor-sasl/livvy)
15:49:36speachyuclpack was only listed as a tool
15:50:17speachywhelp, time to redo that patch then. :)
15:50:25Strife89It turns out that .opus test files exist here already:
15:51:11Strife89It also turns out that I have no clue how to use wget :')
16:02:01speachy__Bilgus_, it should be okay to remove it from the default make targets though, right?
16:05:42speachyokay, the non-ucl parts of the 5/4 patch are merged into the 1st, and the problems with 4/4 are fixed, I think.
16:10:01speachy__Bilgus_ it almost seems like we should have some sort of ceremony to mark this occasion
16:21:19***Saving seen data "./dancer.seen"
16:22:54LambdaCalculus37Wait, you guys are tearing HWCODEC completely out of the codebase?
16:24:00LambdaCalculus37Welp, guess it's gonna be the end of an era, then.
16:24:31speachytbh the charcell display was actually more of an impact to the codebase than hwcodec.
16:25:24LambdaCalculus37I was figuring you guys would've probably split the Archos code completely off and had one codebase for just those targets, and the main one for the other targets.
16:25:35LambdaCalculus37But I guess nothing lasts forever.
16:25:43speachymost of the hwcodec code was closely tied with the sh/archos port code, and not terribly portable should any future hwcodec platform come up
16:27:36LambdaCalculus37At this point, I doubt very much another HWCODEC target would ever show up. Any opportunities to add such haven't been done years ago and probably won't be done now.
16:30:36speachymost of the low-end players on the market are hwcodec targets now (eg atj2127 and rknano)
16:32:26 Quit sakax (Remote host closed the connection)
16:33:47speachywith ludicrously tiny amounts of RAM (especially the atj with its 88KB...)
16:35:27LambdaCalculus37Yikes! 88KB is even less than what the Archos targets had!
16:36:13speachythey execute code from flash and presumably directly DMA the audio file data from the SD card
16:36:57speachyeach gen of the rknano has more ram; the latest I think has closer to 1MB.
16:37:10 Join johnb2 [0] (
16:40:03Strife89 Finally got the footage uploaded.
16:41:36johnb2speachy : does Sansa e200 (v1) fall into the portalplayer ARM category?
16:41:58johnb2for testing the new toolchain ...
16:42:32speachyyeah, it's a PP5024
16:42:44 Join jdarnley [0] (
16:42:45speachyI can toss you a build if you'd like
16:44:21speachy(it would be with the existing toolchain bit with -Os optimization; this breaks on the ipodcolor but works with the mini2g)
16:45:48johnb2ok, I can run tests then tomorrow.
16:45:49 Quit J_Darnley (Ping timeout: 264 seconds)
16:50:26speachyif that works (and it will be immediately obvious IMO) I will generate something with the newer toolchain.
16:56:20johnb2it boots and plays mp3
17:04:26 Quit johnb2 (Ping timeout: 260 seconds)
17:05:34 Join johnb2 [0] (
17:07:53 Join kugel_ [0] (
17:07:53 Quit kugel_ (Changing host)
17:07:53 Join kugel_ [0] (~kugel@rockbox/developer/kugel)
17:09:42 Quit kugel (Ping timeout: 256 seconds)
17:10:07 Quit ac_laptop (Ping timeout: 246 seconds)
17:11:19johnb2works also
17:11:46speachyokay, good data point. thank you
17:11:58speachyjust need to figure out WTF is up with the ipodcolor
17:12:47 Quit johnb2 (Quit: Nettalk6 -
17:13:57 Join kugel [0] (
17:13:57 Quit kugel (Changing host)
17:13:57 Join kugel [0] (~kugel@rockbox/developer/kugel)
17:16:59 Quit kugel_ (Ping timeout: 240 seconds)
17:17:00 Join Soap [0] (~Soap@rockbox/staff/soap)
17:19:40 Quit Soap_ (Ping timeout: 256 seconds)
17:45:03 Quit livvy (Ping timeout: 240 seconds)
17:50:51 Join kugel_ [0] (
17:50:51 Quit kugel_ (Changing host)
17:50:51 Join kugel_ [0] (~kugel@rockbox/developer/kugel)
17:51:45 Quit kugel (Ping timeout: 240 seconds)
18:06:49 Join kugel [0] (
18:06:49 Quit kugel (Changing host)
18:06:49 Join kugel [0] (~kugel@rockbox/developer/kugel)
18:08:13 Quit kugel_ (Ping timeout: 246 seconds)
18:21:21***Saving seen data "./dancer.seen"
18:30:03 Quit efqw (Quit: Connection closed for inactivity)
18:43:43 Quit Rower (Ping timeout: 265 seconds)
18:51:26 Join meatloaf [0] (~ambassado@2601:246:4400:b3c0:64f0:c013:f21:ae75)
18:52:27 Join ac_laptop [0] (~ac_laptop@
19:07:23 Quit mendel_munkis (Ping timeout: 240 seconds)
19:12:10 Join mendelmunkis [0] (
19:22:46 Quit meatloaf (Ping timeout: 244 seconds)
19:26:30__Bilgus_speachy sorry I was AFK
19:26:43 Join meatloaf [0] (~ambassado@2601:246:4400:b3c0:64f0:c013:f21:ae75)
19:30:18__Bilgus_LambdaCalculus37 The plan was/is to just leave a branch at 3.15 for those old targets
19:30:18 Join cockroach [0] (~blattodea@pdpc/supporter/active/cockroach)
19:31:12__Bilgus_then someone can pick it up if they really want it but we won't be supporting them anymore or coding around it
19:31:38__Bilgus_speachy default make targets?
19:44:01 Quit MrZeus_ (Ping timeout: 264 seconds)
19:45:51 Quit craftyguy (Quit: WeeChat 2.8)
19:46:35 Join craftyguy [0] (~craftyguy@qutebrowser/dev/craftyguy)
19:52:48 Quit ufdm (Read error: Connection reset by peer)
19:53:08 Join ufdm [0] (
20:21:24***Saving seen data "./dancer.seen"
20:31:01speachy__Bilgus_, I went back and left ucl completely alone.
20:31:43__Bilgus_did you look at any of the code in UCL?
20:38:41speachyonce you pointed out where it was still being used, I nuked that 5/4 patch and folded the non-ucl bits into 1/4.
20:39:29__Bilgus_it'd be nice to have the compression on device but I bet it'd be slowwwww
20:41:44speachyno, I didn't look into it beyond that
20:42:29__Bilgus_pretty neat :)
20:48:10speachyso things are ready to land; I have a commit for the www side to disable the archos builds; once the commits land in master the infra will pick up the new status from and update accordingly.
20:48:20speachyThough I suppose I should set a tag first
20:57:23__Bilgus_Did you bump the plugin API in any of those patches?
21:02:59speachynot yet
21:05:31 Quit ZincAlloy (Quit: Leaving.)
21:12:59 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:a051:3387:ec2:ec2a)
21:13:03 Quit ZincAlloy (Read error: Connection reset by peer)
21:13:22 Quit jdarnley (Ping timeout: 246 seconds)
21:15:33 Quit edhelas (Ping timeout: 260 seconds)
21:15:50 Join J_Darnley [0] (
21:16:22 Quit kadoban (Ping timeout: 244 seconds)
21:16:22 Quit ender| (Ping timeout: 244 seconds)
21:16:43 Quit ats (Ping timeout: 260 seconds)
21:18:04LambdaCalculus37Saw the mail announcment...
21:18:06 Join ender| [0] (~ender1@2a01:260:4094:1:6045:6fff:fedd:cbf3)
21:18:14*LambdaCalculus37 pours one out for the HWCODEC targets
21:18:19 Join ats [0] (~ats@2001:8b0:83b:b53f::a)
21:18:22LambdaCalculus37RIP Archos devices.
21:19:12 Join kadoban [0] (kadobanmat@gateway/shell/
21:20:16speachy__Bilgus_, 4/4 updated with an API bump.
21:20:57__Bilgus_I think I just need to push my patch within the same day and no one should be the wiser
21:21:20LambdaCalculus37__Bilgus_, if you guys branch at 3.15, is that specific branch going to be HWCODEC code only or no?
21:21:23speachyI also have that remove-six-unusued-API-calls patch to followup too.
21:22:04__Bilgus_so push them all together and no need to bump more than once
21:22:20speachyLambdaCalculus37, I figured we'd just set a tag right before the archos stuff goes away; no sense in creating a branch unless someone wants to maintain it. Which I doubt will happen.
21:22:25mendelmunkisLambdaCalculus37: 3.15 has been around for a while
21:22:32__Bilgus_Lambda I'm not sure how that works in github
21:22:48mendelmunkisnvm I got confused with 3.14 sorry
21:23:06__Bilgus_really nothing after 3.13? maybe older was even added to archos
21:23:09speachywell, 3.15 is now what, 8 months old?
21:23:21speachy3.13 was the final recorderv1 release
21:23:35speachysince then there's just been whatever has landed in the core
21:23:51__Bilgus_I suppose there could be some updates to core
21:24:12__Bilgus_but no one has come foward saying they tested it so who knows
21:24:28speachy3.13 was late 2013. :)
21:24:56speachywell, I did have at least one person test one of my gcc4.9.4 sh builds with a modded ondio
21:25:51__Bilgus_well I guess 3.15 being the last official and wherever your mentioned build is to be the last dev?
21:26:25speachydev builds older than 30 days are automatically deleted
21:26:50speachybut once the archos stuff is all marked as retired the www site (and theoretically, rbutil) will stop displaying the option to download dev builds
21:27:08speachyit'll just be the final release
21:27:24speachylike what is shown for recorderv1
21:29:16__Bilgus_who has the permissions for the github account?
21:30:56speachyzagor, scorche, bagder, and myself have admin access.
21:31:15__Bilgus_we don't need the builds just a branch for anyone that wants it
21:31:38__Bilgus_as far as compiled builds I think 3.15 is fine for em
21:32:27 Quit ender| (Ping timeout: 260 seconds)
21:33:02speachythe github project is just a repo mirror and an oauth provider for gerrit.
21:45:56 Join ender| [0] (~ender1@2a01:260:4094:1:6045:6fff:fedd:cbf3)
21:46:10__Bilgus_oh we can't just put a branch there does it have to be mirrored only?
21:46:58speachyit would have to be a new repo
21:47:21speachywhy not on the main repo?
21:48:29speachy(and, for that matter, why bother with a branch? just tag it, and if anyone cares enough to pick development back up, we can create the branch then..)
21:55:47 Quit meatloaf (Ping timeout: 244 seconds)
21:57:00 Join meatloaf [0] (~ambassado@2601:246:4400:b3c0:64f0:c013:f21:ae75)
22:02:16 Quit cockroach (Quit: leaving)
22:10:13 Quit LambdaCalculus37 (Quit: Leaving)
22:21:28***Saving seen data "./dancer.seen"
22:22:14__Bilgus_ah ok, that works
22:24:37speachyI figure tomorrow evening is a good time as any to drop the axe.
22:24:54speachygive the mailing list lurkers a chance to decry my awful judgement
22:42:36__Bilgus_ok although I doubt many compelling arguments exist
23:16:47speachyhuh, I'm getting a "TLB refill handler" panic when I switch languages on the X3. I don't recall it doing that before.
23:17:11speachyit's basically a bogus pointer in a memcpy()
23:31:18fs-bluebotBuild Server message: New build round started. Revision 09f40e2, 295 builds, 8 clients.
23:31:49speachyfixed. kinda amazed this didn't come up before
23:32:28 Quit [7] (Disconnected by services)
23:32:35 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
23:32:44 Quit meatloaf (Ping timeout: 244 seconds)
23:36:08__Bilgus_found it a lot faster than I expected
23:39:46speachyme too
23:40:29speachyI figured starting at where changing language invoked a voice change was warranted
23:47:15speachygoes to show you how many corner cases are probably still lurking in the talk subsystem
23:48:48speachyoh, I can confirm that rbutil can now generate working voice files from up-to-date dev builds
23:49:04speachyfrom both complete and incomplete translations
23:52:28__Bilgus_I think there are lots of corner cases throughout lots of 'little' things like on off errors etc.
23:53:24speachygenlang still needs more de-spaghettifying TLC
23:53:25fs-bluebotBuild Server message: Build round completed after 1328 seconds.
23:53:27fs-bluebotBuild Server message: Revision 09f40e2 result: All green

Previous day | Next day