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 2022-12-01

00:18:38_bilgusdamn still got the wrong resume after a bunch of tries, turns out get_filename alters the seek pos and doesn't restore it
00:18:53_bilgusonly issue is that utf8 check
00:24:29_bilgusits been broken for 18-20 years
00:42:50_bilgusnow the question is do I use the utf8 check to seek back to the beginning or just open a new filedescriptor
00:43:16_bilguswe did just up the number of fds
00:44:12_bilgusif I just open a new one there is no worries about the mutex anymore
00:53:38_bilgus g#4859
00:53:41rb-bluebotGerrit review #4859 at : RFC playlist.c add_indices_to_playlist() seek back to start of file by William Wilgus
00:58:27 Join S|h|a|w|n [0] (~shawn156@user/shawn/x-4432647)
00:59:07 Join SammysHP_ [0] (
00:59:28 Join jacobk_ [0] (~quassel@
00:59:47 Join yosafbridge` [0] (
01:05:52 Quit jacobk (*.net *.split)
01:05:52 Quit amiconn (*.net *.split)
01:05:53 Quit pixelma (*.net *.split)
01:05:53 Quit ats (*.net *.split)
01:05:53 Quit dys (*.net *.split)
01:05:53 Quit saanaito (*.net *.split)
01:05:53 Quit yosafbridge (*.net *.split)
01:05:53 Quit SammysHP (*.net *.split)
01:05:54 Nick SammysHP_ is now known as SammysHP (
01:21:23 Quit SammysHP (Quit: *wuff*)
01:21:43 Join SammysHP [0] (
01:30:53 Join advcomp2019_ [0] (~advcomp20@user/advcomp2019)
01:34:06 Quit advcomp2019 (Ping timeout: 265 seconds)
01:49:32***Saving seen data "./dancer.seen"
02:32:07 Join amiconn [0] (
02:34:53 Join pixelma [0] (
02:34:54 Join ats [0] (
02:34:54 Join saanaito [0] (
02:44:41 Join mink [0] (~mink@2a07:3e00:81:0:7b6:4663:f93b:9fae)
03:49:36***No seen item changed, no save performed.
03:56:56 Quit CH23_M (Read error: Connection reset by peer)
03:57:39 Join CH23_M [0] (~CH23@revspace/participant/ch23)
04:15:29 Quit S|h|a|w|n (Read error: Connection reset by peer)
04:34:19 Quit prg (Read error: Connection reset by peer)
04:36:30 Join prg [0] (
05:49:40***Saving seen data "./dancer.seen"
07:49:41***No seen item changed, no save performed.
08:07:34 Join dys [0] (~dys@user/dys)
08:14:56 Join massiveH [0] (~massiveH@2600:4040:a993:4900:78d9:d42c:f371:5c0c)
08:55:46 Quit mink (Ping timeout: 252 seconds)
09:00:42 Join mink [0] (~mink@
09:49:42***Saving seen data "./dancer.seen"
10:00:09 Quit CH23_M (Ping timeout: 260 seconds)
10:01:17 Join CH23_M [0] (~CH23@revspace/participant/ch23)
10:08:22speachyhmm, I wonder if we should be using -fwrapv and -fwrapv-pointer
10:12:16speachyor -fstrict-overflow which turns them both on
10:13:46speachyaaaand the SSL cert on the website _finally_ covers just straight '' as well as '*' :D
10:18:49_bilgusspeachy probably
10:19:14_bilgusthere are a lot of assumptions hidden in the depths
10:20:12speachywith newer toolchains there's better optimization but also better "optimization"
10:21:53speachyI have a hard time seeing using those flags as introducing any regressions.
10:24:50speachyoh that's right, we turned on -Wstrict-overflow a long time ago so we get warnings when the compiler finds anything that would get affected by that flag.
10:28:22speachyno, we don't. whoops.
10:29:44speachy-Wall turns on -Wstrict-overflow=2
10:30:52speachybumping it to =3 spews a ton of crap
10:42:47 Nick jacobk_ is now known as jacobk (~quassel@
10:45:44speachyso it's probably a no-op if we enable it
10:49:30 Quit Malinux (Ping timeout: 260 seconds)
10:55:50 Quit massiveH (Quit: Leaving)
11:36:02 Join fourHZ [0] (
11:41:26 Quit mink (Remote host closed the connection)
11:49:43***Saving seen data "./dancer.seen"
12:05:15 Join Malinux [0] (~malin@2001:4641:4dfa::12c:c4a7)
12:08:00 Join amachronic [0] (~amachroni@user/amachronic)
12:11:49rb-bluebotBuild Server message: New build round started. Revision 52ca658069, 303 builds, 8 clients.
12:15:28 Quit jacobk (Ping timeout: 252 seconds)
12:17:36speachyamachronic: we're still not really any closer to an automated installer for the x1000-based devices, correct?
12:18:28amachronicthe best bet for Windows is spawning an HTTP server
12:19:01amachronicjztooljs is the only method that doesn't require a driver install
12:26:12 Quit fourHZ (Quit: Client closed)
12:30:34 Join jacobk [0] (
12:31:27amachronicthe install also needs to be automatic on the device side to avoid requiring extra user interaction
12:32:18amachronicnice, it seems qt6 comes with an HTTP server component.
12:33:45 Quit jacobk (Client Quit)
12:35:01rb-bluebotBuild Server message: Build round completed after 1392 seconds.
12:35:03rb-bluebotBuild Server message: Revision 52ca658069 result: All green
12:35:15 Join jacobk [0] (
12:36:30rb-bluebotBuild Server message: New build round started. Revision 7819a06d74, 303 builds, 8 clients.
12:37:30 Join fourHZ [0] (
12:45:59 Join lebellium [0] (
12:48:34 Quit fourHZ (Quit: host reboot)
12:56:18rb-bluebotBuild Server message: Build round completed after 1188 seconds.
12:56:19rb-bluebotBuild Server message: Revision 7819a06d74 result: All green
12:57:35 Join fourHZ [0] (
13:07:26rb-bluebotBuild Server message: New build round started. Revision 5b1dd64f50, 303 builds, 8 clients.
13:20:41 Quit fourHZ (Quit: Client closed)
13:20:54 Join fourHZ [0] (
13:35:44rb-bluebotBuild Server message: Build round completed after 1699 seconds.
13:35:46rb-bluebotBuild Server message: Revision 5b1dd64f50 result: All green
13:37:03rb-bluebotBuild Server message: New build round started. Revision 6346be51a3, 303 builds, 8 clients.
13:49:47***Saving seen data "./dancer.seen"
13:57:48rb-bluebotBuild Server message: Build round completed after 1244 seconds.
13:57:49rb-bluebotBuild Server message: Revision 6346be51a3 result: All green
13:57:52 Quit Malinux (Ping timeout: 256 seconds)
14:00:11rb-bluebotBuild Server message: New build round started. Revision fb0757b913, 303 builds, 8 clients.
14:02:58 Quit jacobk (Ping timeout: 256 seconds)
14:07:30 Join jacobk [0] (~quassel@
14:20:42rb-bluebotBuild Server message: Build round completed after 1230 seconds.
14:20:44rb-bluebotBuild Server message: Revision fb0757b913 result: All green
14:24:53rb-bluebotBuild Server message: New build round started. Revision d7557e8da8, 303 builds, 8 clients.
14:45:14 Quit fourHZ (Quit: Client closed)
14:46:36 Quit amachronic (Ping timeout: 256 seconds)
14:51:30rb-bluebotBuild Server message: Build round completed after 1596 seconds.
14:51:32rb-bluebotBuild Server message: Revision d7557e8da8 result: All green
14:54:34 Join Malinux [0] (~malin@2001:4641:4dfa::12c:c4a7)
15:23:24 Quit CH23_M (Read error: Connection reset by peer)
15:23:43 Join CH23_M [0] (~CH23@revspace/participant/ch23)
15:27:58 Quit CH23_M (Ping timeout: 256 seconds)
15:29:04 Join CH23_M [0] (~CH23@revspace/participant/ch23)
15:30:53 Quit CH23_M (Read error: Connection reset by peer)
15:31:12 Join CH23_M [0] (~CH23@revspace/participant/ch23)
15:49:51***Saving seen data "./dancer.seen"
16:15:53 Quit Malinux (Ping timeout: 252 seconds)
16:34:49 Join Malinux [0] (~malin@2001:4641:4dfa::12c:c4a7)
16:56:22 Quit Malinux (Ping timeout: 256 seconds)
17:49:53***Saving seen data "./dancer.seen"
17:56:31rb-bluebotBuild Server message: New build round started. Revision 773fa7874d, 303 builds, 7 clients.
17:57:16_bilguswith that playlist should be fixed I'll try a rewrite see how it turns out
17:58:33speachythe 256GB chinesium SSD for the iPod arrived already. that was quick.
18:01:34CH23speachy, msata, right?
18:01:45speachysupposedly. :)
18:02:13CH23would be interested to see if it's faster than an iflash quad with good µSD cards
18:02:47speachyI doubt it; I think the ipod's ATA interface is the actual bottleneck.
18:06:21 Quit lebellium (Quit: Leaving)
18:06:23_bilgushurry up and wait
18:06:49speachyI expect the SD-based iFlash devices are much better on a power consumption front.
18:07:07speachythe only reason to go to the dual/quads are for more (cheaper) capacity.
18:08:20speachyIncidently we could support >2TB SDUC cards without too much work (ie from a storage and sd driver perspective) but we'd need an exFAT implementation, which isn't so simple.
18:09:48speachyI don't think the older (<6g) ipods would be bootable on GPT-equipped drives though. (Might be able to do a GPT-MBR hybrid so the ipod system parition is bootable, but... )
18:15:37CH23my 5th gen was too slow to really do much with my 500GB of music (however it's been dead since before RB 3.15 was released)
18:22:23rb-bluebotBuild Server message: Build round completed after 1551 seconds.
18:22:29rb-bluebotBuild Server message: Revision 773fa7874d result: All green
18:23:17_bilgusCH23 I tried a 3.15 Install the other day I could not get over how much less responsive it was over our current dev versions
18:24:02speachywe've made a _lot_ of improvements. speaking of.. the wiki changes page is ridiculously out of date
18:24:07_bilgusI'd attribute most of tht to the dircache rewrite by JhMikeS but its probably more than just that
18:24:44speachya lot of little improvements eventually add up
18:26:53CH23I am using 0ba3392b9fM-221130 now, I did see something about issues with playlists but so far my playlists seems to play correctly
18:27:29CH23(that's on my 6th gen 3rd revision iPod Classic)
18:28:03_bilgusI got most of the bugs with that one but todays should be solid
18:28:59_bilgusI'm going to refactor it over the next week or so and see if I can clean it up a bit
18:29:36_bilgusI think I've finally squashed the USB unplug bug with the playlist file but time will tell
18:29:47 Quit TorC (Ping timeout: 248 seconds)
18:30:06speachyit's been really nice to see lots of green on the build size table.
18:31:06_bilgusI've been trying to get our binsizes down I figure with all these features its a good idea to do some cleanup
18:31:37_bilgusamachronic has been making some headway with that too I saw
18:32:59speachyI need to finish reworking the jz4760 usb driver
18:34:12speachyat least I got the worst of the infra backlog sorted out last weekend.
18:34:14_bilgusI think he is working on the display system with plans to drag it in to the 20th century it'd be nice to use something half way close to windows (the items not the OS)
18:35:21_bilgusthe USB is half way to black magic to me I'm glad you are working on it
18:35:35speachythe original ctrl URB handling was always ... suspect.
18:36:20_bilgusI think we spent a few years with the Clip(s) having super flaky USB
18:36:24speachywhen amachronic rewrote the core, it broke the jz4760.
18:36:48_bilgusyou had to got to the OF to even transfer files
18:37:01speachyI need to determine if the DMA stuff being broken is due to something else or my use of that newer toolchain
18:37:28_bilgusthats a toss up A.B or a mix
18:37:41speachyAnd, heh, we have another toolchain to consider upgrading to. :)
18:37:58_bilguswell when you are coming from circe 2004
18:38:11speachyunfortunately we're still stuck on gcc49 for android-based stuff.
18:38:44speachynah, current stuff is circa-2013 gcc 4.9.4 (which was one hell of a leap, so much breakage)
18:38:48_bilgusbut since goolgle changed android it needs major rework right?
18:39:09_bilgusno I was saying before you brought it to 4.9
18:39:17speachyspeaking of android there's a substantial patch set in the tracker to bring it into a more modern era
18:39:41speachybut the ibasso targets also rely on the android ndk toolchain
18:39:45_bilgusthat would make some really happy people
18:40:16speachypost-494 google moved to using llvm. We could probably get away with that for our hosted builds but I've not been that brave yet.
18:41:05_bilgusI think with a bit more hardening of core it probably wouldn't be so horribl now Coverity, UBSAN and ASAN mainly to thank
18:41:31speachykeeping the sims building warning-free on modern linux distros has shaken out the core, but there's probably a lot of target-specific stuff with issues. surprisingly the handful of builds I made only generated a net total o 3 new warnings.
18:41:50_bilgusI would really like asserts on the device though
18:41:51speachy(with gcc8.5)
18:42:23speachywhat would an assert do, a BSOD?
18:42:27_bilgusIt wouldn't immediately benefit but as time wore on itd get much more stable
18:43:37speachythe most recent generation of chinesium daps using old flip-phone chipsets would make for an interesting port
18:43:42_bilgusyeah true the panic isn't too great a mechanism
18:44:04speachyif only we could guarantee they'd be around (in an identifiable manner) by the time the port is usable...
18:44:06_bilguson the forum we have the initial part of that
18:44:34speachythey're outwardly identical to all the shovelware rknano/atj2127 devices.
18:45:26speachyironically lately for $dayjob I've been writing a bluetooth controller
18:45:32_bilgusI have one of the devices left the first one died a horrible death before I could even get to trying to get the bootloader flashed
18:45:57_bilgusI guess with <$20 device its to be expected
18:46:25_bilgusthe second one I've gotten my own code on it but its not anything like needs done
18:46:48_bilgusyou know a viable bootloader
18:47:40_bilgusits a path forward though audio quality isn't anything like what I we are used to though
18:48:17_bilgusaudio books itd be perfect
18:48:27speachywhoops, the translate 'problems' page is busted.
18:52:31_bilgusI didn't even know it existed
18:53:26speachywow, that was a pretty embarassing bug; the incorrect argument order has been there since approximately forever.
18:53:36speachyphp <8 was a lot more forgiving apparently.
18:54:56speachyclik the little triangle warning logo next to the flag and it'll tell you what's potentially wrong. Also lets you upload your own to check.
18:55:47speachythe edit page is a superset of that functionality. More accurate too as it handles per-target/feature variations better.
18:56:50speachyI partially rewrote the latter when I also rewrote the genlang/updatelang stuff way back when
18:57:14speachyperl is so much nicer for this string manipulation crap. :D
18:58:33speachyI'm personally amazed that the themesite is still trucking along. Granted I've done a bit of work on it too.
19:03:49 Join TorC [0] (~Tor@fsf/member/TorC)
19:04:18_bilgusits nice that we have a way for users to contribute that
19:10:59 Join Malinux [0] (~malin@2001:4641:4dfa::12c:c4a7)
19:15:48 Quit Malinux (Ping timeout: 252 seconds)
19:27:06_bilgusthere is still at least one more race in the playlist code I just saw a playlist access error haven't been able to repro it yet so I think the rewrite with an eye towards thread safety is a good idea
19:27:55_bilgusI did leave debug code in there to make them come out of the woodwork but haven't seen one all day
19:29:53_bilguslike 10 ish hours with using bookmarks didn't play the wrong track once so I think I can remove the code that checks if the wrong track is played but it can stay in till after a rewrite
19:31:13_bilgusthis one was with the playlist catalog I haven't tested it much today but I was checking out the remove static buffers patch
19:49:55***Saving seen data "./dancer.seen"
20:26:21 Quit jacobk (Ping timeout: 255 seconds)
20:37:11 Quit paulk (Ping timeout: 265 seconds)
20:37:16 Join jacobk [0] (~quassel@
20:37:16 Join paulk-bis [0] (
21:46:11 Quit mrkrisprolls (Quit: Tchussss!)
21:49:57***Saving seen data "./dancer.seen"
22:29:40rb-bluebotBuild Server message: New build round started. Revision 177a15b2ed, 303 builds, 7 clients.
22:30:47 Join mrkrisprolls [0] (
22:59:41rb-bluebotBuild Server message: Build round completed after 1801 seconds.
22:59:43rb-bluebotBuild Server message: Revision 177a15b2ed result: All green
23:37:15 Quit m01 (Quit: Konversation terminated.)
23:39:22 Join m01 [0] (
23:49:31 Quit danwellby (Ping timeout: 268 seconds)
23:50:00***Saving seen data "./dancer.seen"
23:56:17 Join danwellby [0] (

Previous day | Next day