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 2023-05-23

00:00:32 Quit jacobk (Ping timeout: 248 seconds)
00:01:43 Join jacobk [0] (
00:05:53***Saving seen data "./dancer.seen"
00:15:30 Join tchan1 [0] (
00:16:47 Quit tchan (Ping timeout: 240 seconds)
00:19:49 Join othello8 [0] (
00:21:14 Quit othello7 (Ping timeout: 246 seconds)
00:21:14 Nick othello8 is now known as othello7 (
00:32:32 Quit jacobk (Ping timeout: 248 seconds)
00:37:12 Quit m01 (Quit: Konversation terminated.)
00:39:40 Join m01 [0] (
00:53:59 Join jacobk [0] (
00:55:28 Quit JanC (Ping timeout: 248 seconds)
00:55:31 Join JanC_ [0] (~janc@user/janc)
00:55:45 Nick JanC_ is now known as JanC (~janc@user/janc)
01:11:05 Join JanC_ [0] (~janc@user/janc)
01:11:19 Quit JanC (Killed ( (Nickname regained by services)))
01:11:19 Nick JanC_ is now known as JanC (~janc@user/janc)
01:12:08 Quit jacobk (Ping timeout: 240 seconds)
01:14:12 Join jacobk [0] (
01:21:59 Quit jacobk (Ping timeout: 240 seconds)
01:37:39 Join JanC_ [0] (~janc@user/janc)
01:37:44 Quit JanC (Read error: Connection reset by peer)
01:37:53 Nick JanC_ is now known as JanC (~janc@user/janc)
01:39:41 Quit melmothX (Quit: reboot)
01:51:04 Join melmothX [0] (~marco@amusewiki/marco)
02:05:57***Saving seen data "./dancer.seen"
02:17:36 Quit JanC (Read error: Connection reset by peer)
02:17:56 Join JanC [0] (~janc@user/janc)
02:47:52 Join advcomp2019_ [0] (~advcomp20@user/advcomp2019)
02:51:01 Quit advcomp2019 (Ping timeout: 240 seconds)
03:09:20 Quit lizzie (Ping timeout: 248 seconds)
03:09:35 Join lizzie [0] (
03:24:06 Quit lizzie (Remote host closed the connection)
04:06:00***Saving seen data "./dancer.seen"
04:39:58 Quit Pokey (Quit: Hecc! My server must have died!)
04:41:35 Quit troglodito (Ping timeout: 240 seconds)
04:42:39 Join troglodito [0] (~cave@
04:42:59 Join Pokey [0] (~pokey@spikeyCactus/hoosky)
05:09:44 Quit JanC (Remote host closed the connection)
05:09:57 Join JanC [0] (~janc@user/janc)
05:31:59 Quit troglodito (Ping timeout: 240 seconds)
05:35:55 Quit S|h|a|w|n (Quit: Leaving)
05:36:44 Join troglodito [0] (~cave@
05:52:42 Quit JanC (Read error: Connection reset by peer)
05:52:50 Join JanC [0] (~janc@user/janc)
06:02:28 Quit JanC (Read error: Connection reset by peer)
06:02:53 Join JanC [0] (~janc@user/janc)
06:06:02***Saving seen data "./dancer.seen"
07:26:22 Nick tchan1 is now known as tchan (
08:06:05***No seen item changed, no save performed.
08:24:27 Quit bluebrother (Ping timeout: 256 seconds)
08:25:11 Quit rb-bluebot (Ping timeout: 268 seconds)
08:26:17 Join bluebrother [0] (~dom@user/bluebrother)
08:38:32 Join rb-bluebot [0] (~rb-bluebo@rockbox/bot/utility)
10:05:53 Quit Trzyzet (Ping timeout: 256 seconds)
10:06:07***Saving seen data "./dancer.seen"
10:06:19 Quit Piece_Maker (Ping timeout: 268 seconds)
10:08:55 Join Piece_Maker [0] (
10:10:05 Join Trzyzet [0] (
10:26:37 Join olk [0] (
10:28:57olkhey, is there any info on NWZ-S765 port? there's no nwz-s760 entry in Player list on SonyNW page
10:31:16speachyIf it's not there, it doesn't exist.
10:31:26speachynot the answer you wanted, but that's what it is...
10:32:12olksad. it's also said there that s760 and a860 have same hardware
10:33:29olkmp200 is pretty well documented, isn't it?
10:35:20olkyeah, the SoC
10:35:38olk... the NW-E060/NW-S760/NW-A860/NWZ-S760/NWZ-A860 series all share the same hardware ...
10:35:47braewoodseven if it is, it's moot if there's no one to put in the effort.
10:36:37olkwell, I have the device but lack knowledge
10:37:26braewoodsspeachy, will there come a time that we drop ports for devices that are so rare today that we can no longer guarantee current code is still working on them?
10:37:41speachyheh, that would be most of them?
10:37:46braewoodsYea, I guess so.
10:38:01braewoodsI would define that as anything that is too rare to find second hand within reason.
10:38:36 Join jacobk [0] (
10:38:40braewoodsIt would simplify the codebase, but what would we be left with?
10:38:58speachywe tried a big push a couple of years back to get folks to re-test things after the toolchain bump; I don't think we got rports than about 2/3rds.
10:39:26speachythe big simplification was nuking HWCODEC support, everything else is just incremental.
10:39:38braewoodsFair I guess.
10:39:55braewoodsI was just noting that the MPIO stuff is so rare now that it is next to impossible to find at all.
10:40:28braewoodsCan't really fix issues for stuff we can't test.
10:40:54speachyOther complications: tiny-mem targets (original Clip, and a couple others) and devices that have dedicated USB hardware (eg the old iRiver ports)
10:41:11braewoodsright, no software USB.
10:41:20braewoodsbut the iriver stuff is still used.
10:41:37speachythose are the ones that cause mutiple architectural approaches in the code.
10:42:14speachybeyond that, each additional target adds some device drivers but that's about it.
10:42:59speachyI'd wager that, other than the various iPods and the dwindling numbers of still-operational Sansa players, the rest of the list is largely irrelevant.
10:43:14speachybasically, flash wears out.
10:43:25braewoodsprobably, at least the units which cannot be repaired easily.
10:43:33braewoodsthe HDD units are an exception.
10:43:42braewoodsSocketed storage in a sense.
10:43:48speachythe ErosQ/K clones seem to still be popping up
10:44:05braewoodsOr ones like the x3 where all the storage is external sd cards.
10:44:06olksorry to interrupt, do you need the help_us paste from nwz-s765? or there will be no efforts to port s760?
10:44:08speachybasically if it doesn't have removeable/replaceable storage, it's gonna die.
10:44:39braewoodsolk, if it's not already supported i doubt anyone would put in the effort now.
10:44:48speachyIIRC the linux-based Sonys are all pretty similar, but I don't know what specifics are needed to generate a working build for any given device.
10:45:43olkspeachy: if you confirm what braewoods said, I'm not going to bother you guys any longer
10:45:57speachyour longtime Sony guy hasn't been active in a long time
10:46:25braewoodsI've done some contributions like repairing the H100/H200 bootloaders.
10:46:30speachyolk: what braewoods didn't mention is that that somebody could be you. :D
10:46:57braewoodsYea, but I can't recommend that kind of work to someone who has never written C before.
10:47:58speachygotta start somewhere. And if it's just a matter of adding another NWZ variant, it's mostly a copy-paste job, and nearly none of that is C.
10:48:07braewoodsmy little project of writing a kind of background service for some unix OS is coming along but slowly.
10:48:30olkspeachy: huh, I can do that
10:49:19braewoodsMaybe, but we can't guarantee the end result will work, given the subject matter expert is no longer around.
10:49:48olkwho's the sony expert?
10:50:14braewoodsI don't know, you'd have to ask speachy. Anyway, you should get a rockbox development environment setup if you want to try this.
10:50:40braewoodsIt sounds like you won't have to change anything but some script code I think.
10:51:45speachyapouly was the Sony maintainer, but hasn't been heard from in about two years.
10:52:16speachywell, from rockbox's perspective anyway.
10:53:16braewoodsspeachy, we still on gcc 4.9?
10:53:23braewoodsOr have we upgraded again?
10:54:29speachyyep, 4.9.4 for all targets. I have a 8.5.x-based toolchain in gerrit, but only a handful of targets have been tested with it.
10:55:44olkmp200 uses arm-elf-eabi target?
10:56:02speachyannoyingly due to make 4.4, the hosted toolchains won't build properly; havent' managed to fix it yet.
10:56:15braewoodsmost of the rockbox targets are some flavor of ARM so probably.
10:56:39braewoodsspeachy, what's special with gnu make 4.4?
10:57:08 Quit jacobk (Ping timeout: 250 seconds)
10:57:12speachyshort version is they broke some stuff relating to recursive parallel builds.
10:57:19braewoodsso, upstream bug?
10:57:28speachy(broke other stuff too but that was easily fixed in a backwards-compatible manner)
10:58:17speachyupstream in the sense that the failures are in the ancient glibc versions we need.
10:59:05speachythe symptom is that the build basically hangs/deadlocks, with two CPU cores pegged.
10:59:18speachymight be able to work around it by moving to a non-parallel build though.
10:59:51braewoodsBut why would the glibc version matter to make?
11:00:01braewoodsI could see it matter to the compiler but make?
11:00:31speachyolk: if it's a linux-based target then it would use the arm-rockbox-linux-gnueabi- prefix
11:00:54braewoodsbasically hosted vs native.
11:00:55speachybraewoods: becasuse glibc utilizes some seriously convoluted makefiles?
11:01:24braewoodsOne is where Rockbox runs as the kernel and the other as a Linux application.
11:01:51speachywe have to compile a libc for a non-baremetal toolchain, and since all of our hosted targets use glibc, we're stuck with that.
11:02:06olkspeachy: I chose arm-linux target in if that's what you meant
11:02:23braewoodswhich would be moot if we had the full original linux distribution source but we don't usually.
11:02:32speachybasically I went as new as I could but what we build has to be ABI compatible with the userspace that's already on the targets.
11:02:59braewoodsI mean, it is possible to bypass libc, but not sure that's advisable for rockbox to do.
11:03:10speachythe original source isn't a problem; it's _building_ those sources on modern systems.
11:03:48speachysony/etc didn't do anything special.
11:04:07braewoodsUsing ASM, it is possible to bypass libc but it's not usually advisable.
11:04:48speachyeh, it's not asm that's the problem, but we'd have to implement our own syscall layer.
11:04:48braewoodsI have done it in experiments, making direct system calls, but it requires you to discover the kernel APIs for yourself instead of letting libc sort it out for you.
11:05:03braewoodsYea, I know.
11:05:46braewoodsI prefer to just let the compiler and libc deal with most of the ASM matters though. :P
11:13:48speachyfwiw the current hosted 4.9.4 toolchains won't build with make 4.4 too.
11:33:36braewoodsheh. my ubuntu ships with 4.3
11:46:55 Quit olk (Quit: olk)
11:48:21 Join olk [0] (
12:00:28 Join jacobk [0] (
12:02:25 Join fourHZ [0] (
12:06:08***Saving seen data "./dancer.seen"
12:19:35olkdamn, glibc still compiling
12:21:24 Join S|h|a|w|n [0] (~shawn156@user/shawn/x-4432647)
12:30:13speachywell, crap. even for a non-parallel build, glibc 2.27 with gmake 4.4 still hangs.
12:32:01fourHZi drowned my xduoo xII with rain rain water. the keys are sensitive now. giving multiple impulses after press. and after shutdown the thing boots up again in seconds. would you suggest to open the thing up and clean the contacts?
12:33:42 Quit S|h|a|w|n (Ping timeout: 268 seconds)
12:34:03 Join S|h|a|w|n [0] (~shawn156@user/shawn/x-4432647)
12:40:28 Quit jacobk (Ping timeout: 240 seconds)
12:47:34speachyyeah, open it up, and some rubbing alcohol or (better yet) contact cleaner should do the trick.
13:09:36 Join jacobk [0] (
13:15:17 Quit Natch (Remote host closed the connection)
13:16:27fourHZi expected an X screwdriver and it is torx... got to get it.
13:23:41 Join Natch [0] (
13:59:40 Quit jacobk (Quit: - Chat comfortably. Anywhere.)
14:06:09***Saving seen data "./dancer.seen"
14:28:33 Quit ddevault (Remote host closed the connection)
14:31:06 Join ddevault [0] (e7b4bb6755@sourcehut/staff/ddevault)
14:34:46 Quit fourHZ (Quit: cause gone)
14:36:04 Join lebellium [0] (
14:50:53olkspeachy: hey, you still here? I think it's not ok for glibc-2.19 to compile for ~4 hours
14:53:44speachyif you check 'top' I'll bet you have one or two 'make' processes using 100% of the CPU.
14:53:54olkyeah, two
14:54:05speachyyou have make 4.4?
14:54:11speachy(make -v)
14:54:34speachyyep, that'll do it.
14:54:56speachyI have something working for glibc 2.27, trying to backport it to 2.20 right now.
14:55:25speachywith luck they'll apply to the 2.26/2.19 that we're using for the current toolchains.
14:56:34olkterminate this build and wait for your ping?
14:57:22speachythere are two independent problems, the first can be fixed by not doign a parallel build, but the second is actually a cicrular dependency.
15:01:02olkhow do I tell to use one thread for building toolchain? −−makeflags="-j1"?
15:02:01speachyyou'd have to edit the script to disable that actually
15:02:06speachybut it's not enough
15:03:41braewoodsspeachy, i assume that issue was reported to make's upstream?
15:03:50speachyit's not a make bug per se, it's glibc.
15:04:01braewoodsso why does old make work?
15:04:42speachylook for "Backward-incompatibility"
15:04:55 Part olk (Good Bye)
15:05:13 Join olk [0] (
15:05:53speachyok, successfully built both mips and arm gcc8.5 toolchains.
15:06:05speachynow to see about backporting these patches..
15:06:18speachygranted I don't know if they actually _work_ :D
15:06:33braewoodsspeachy, could a temporary workaround be to create a custom build of make that still works?
15:07:07speachyin theory, yep.
15:07:28braewoodsI don't see what else you can do but patch the glibc source, if you could figure out what needed to change.
15:09:03braewoodsspeachy, is it about the order of makefiles being rebuilt?
15:09:13braewoodsI can't see anything else that might be relevant.
15:10:03speachy2.19 applies cleanly enough...
15:10:10speachybut 2.25 doesn't.
15:10:35braewoodsoh, you already have some kind of fix patches?
15:12:01speachyI've had a partial fix for at least a month, but the other stuff I didn't try to dig into until last night.
15:26:49speachyok, doing rebuilds of the toolchain now. fingers crossed.
15:58:28rb-bluebotBuild Server message: New build round started. Revision 3acd5f86e5, 303 builds, 9 clients.
15:58:28rb-bluebotbuild: Fix problems compiling old glibc with gmake >= 4.4 by Solomon Peachy
15:59:07speachyolk: if you grab the latest git code, it will have the fixes for generating the olde toolchains with gmake 4.4
16:06:11***Saving seen data "./dancer.seen"
16:12:13rb-bluebotBuild Server message: Build round completed after 826 seconds.
16:12:14rb-bluebotBuild Server message: Revision 3acd5f86e5 result: All green
17:08:59 Quit olk (Quit: olk)
17:15:41 Join olk [0] (
17:18:26olkI have an error building alsa-lib, here's my /tmp/rbdev-build/build-alsa-lib-1.0.19.log:
17:19:01olkit's midnight here though, so I'll come back to this in the morning, in 9 hours
17:19:27olkgoodnight o/
17:19:42 Quit olk (Quit: olk)
18:06:14***Saving seen data "./dancer.seen"
18:06:42 Quit lebellium (Quit: Leaving)
18:33:37 Join massiveH [0] (~massiveH@2600:4040:a99f:1f00:d58a:2a33:95ae:bb27)
20:06:18***Saving seen data "./dancer.seen"
20:39:03 Join jacobk [0] (
21:21:34 Quit JanC (Read error: Connection reset by peer)
21:40:28 Quit massiveH (Quit: Leaving)
21:44:47 Quit jacobk (Ping timeout: 240 seconds)
21:45:38 Quit LjL (Read error: Connection reset by peer)
21:46:17 Join LjL [0] (~ljl@user/ljl)
22:06:21***Saving seen data "./dancer.seen"
23:11:04 Quit Nyaa (Read error: Connection reset by peer)
23:12:41 Join Nyaa [0] (
23:25:15 Quit othello7 (Quit: othello7)
23:25:30 Join othello7 [0] (

Previous day | Next day