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 2022-03-10

00:54:47***No seen item changed, no save performed.
01:18:56 Quit pixelma (Quit: .)
01:18:56 Quit amiconn (Quit: - Chat comfortably. Anywhere.)
01:19:53 Join pixelma [0] (
01:19:59 Join amiconn [0] (
02:54:48***Saving seen data "./dancer.seen"
03:00:28_bilgus_adding PUSH{lr} and POP{pc} seems to have fixed it the functions don't get run often but I just don't understand why now
03:00:55_bilgus_maybe its just a compiler bug
04:00:06 Quit blbro[m] (Quit: You have been kicked for being idle)
04:05:23 Quit asaba (Quit: Relay server offline)
04:05:43 Join asaba [0] (~asabas@
04:28:55 Quit JanC (Remote host closed the connection)
04:29:09 Join JanC [0] (~janc@user/janc)
04:54:52***Saving seen data "./dancer.seen"
05:10:10 Quit olspookishmagus (Ping timeout: 250 seconds)
05:12:05 Join olspookishmagus [0] (
06:50:19 Quit kugel (Ping timeout: 256 seconds)
06:52:03 Join kugel [0] (
06:54:56***Saving seen data "./dancer.seen"
07:56:31 Join massiveH [0] (
08:54:59***No seen item changed, no save performed.
10:11:14 Quit massiveH (Quit: Leaving)
10:55:02***Saving seen data "./dancer.seen"
12:27:20 Join ZincAlloy [0] (
12:46:59 Quit bluebrother (Ping timeout: 256 seconds)
12:47:31 Quit rb-bluebot (Ping timeout: 250 seconds)
12:48:28 Join bluebrother [0] (~dom@user/bluebrother)
12:55:04***Saving seen data "./dancer.seen"
13:01:32 Join rb-bluebot [0] (~rb-bluebo@rockbox/bot/utility)
13:27:05 Join lebellium [0] (
13:51:12 Join blbro[m] [0] (~blbrostra@2001:470:69fc:105::8f7)
14:55:06***No seen item changed, no save performed.
15:44:47 Join dconrad_web [0] (
15:46:16dconrad_webI'm running into a new issue, could someone try to reproduce and make sure it's not something wacky with my particular install?
15:47:30dconrad_webwith the current dev build, start playing in the middle of an album or playlist, let it play for 10 seconds or so, then stop playback. Use the resume playback hotkey, and it should resume at the /beginning/ of the album/playlist
15:47:53dconrad_webvery strange - it doesn't seem to do it if I stop playback within the first 3-5 seconds, but longer than that and it does
15:49:50dconrad_webit also seems to do the same behavior with the "Resume Playback" on the main menu, which makes sense
15:50:48dconrad_weboh, I guess also I'm starting playback from the file browser rather than the database, which i haven't tried
16:07:38dconrad_weboh, also heads up, gerrit's down again
16:09:44_bilgus_dconrad it won't save if <15 seconds IIRC
16:10:33_bilgus_its to keep it from constantly saving position when you are doing track changes
16:11:07dconrad_webthat doesn't seem right, it remembers its position if I stop within 3-5 seconds, but forgets if I go longer than that
16:11:26dconrad_webit remembers the album/folder/playlist, but not the track and position
16:13:33_bilgus_I saw code that said <15 checking duration in one of the playback fns
16:14:16_bilgus_and I similarly had to skip to >15 seconds when testing tagcache
16:14:30_bilgus_but hitting power made it save immediate
16:19:21dconrad_webit even happens when I start playback via a bookmark - for instance, let's say I start playback 30 seconds into track 5 of 10, listen for a bit, shut off the player, start it back up and try to resume, but it starts at the beginning of the first track
16:19:55dconrad_webany chance you could reproduce it?
16:21:14_bilgus_ATM all the builds I make aren't working I'm trying to figure out if its this player or this machine
16:23:05_bilgus_its this machine hmm weird
16:23:14 Quit _bilgus_ (Quit: Leaving)
16:28:25 Join _bilgus [0] (~bilgus@
16:45:15_bilguswell crap lua doesn't respect the root redirect, and corrupts the state on top of it
16:45:52 Quit ufdm (Remote host closed the connection)
16:46:10 Join ufdm [0] (
16:55:10***Saving seen data "./dancer.seen"
16:59:07_bilgusdconrad if you haven't heard back from me in a few days ask me again, I'll see what I see as I'm implementing buffer shrink into the tagcache
17:10:45dconrad_web_bilgus sounds good - I'm going to try to find some time to bisect and see if I can find the root cause (when gerrit's back up), I also have my other player I should check on when I get home
17:11:04_bilgusyou don't need gerrit for that
17:11:06dconrad_webI just figured someone would be able to give a quick did/didn't work for them
17:11:16dconrad_webwell, yeah I guess so
17:12:29_bilgusit works for me
17:13:35dconrad_webyou weren't able to see the bug?
17:14:22dconrad_webprobably something screwy with my install then
17:14:42_bilgusi'll try again
17:16:43_bilguseh this time it took like 20 seconds to resume and then finally did, and still went to where I stopped
17:17:20_bilgusI think thats because amachronic stopped tagcache from giving up its buffer and now th playback buffer has to wait for it
17:17:51_bilgus(I haven't looked into it yet)
17:17:53dconrad_webalright, I'll probably do a clean install minus the config and theme stuff and see if it goes away
17:18:50_bilgusI've been tracking a arm ASM bug for signed division for the last 2 or 3 days
17:19:55dconrad_weboh that's fun
17:21:45_bilgusits fixed as soon as I can upload it, I was thinking I can do that to git directly but I can't rem how
17:25:10dconrad_webwell I did a clean install, it still does it, though I tried it twice just now and it only occurred once
17:26:32_bilgusmight be that the file isn't getting written (i think nvram file has the resume info)
17:28:35dconrad_webwell I think it's getting written, as it remembers the album/folder, but it goes to the beginning... which I think points toward my SD card being ok as well?
17:28:52_bilgusmight be two separate things
17:45:09 Join amachronic [0] (~amachroni@user/amachronic)
18:01:44amachronicdconrad_web: I can reproduce that resume bug on the latest daily build too
18:02:12dconrad_weboh, good to hear it's not just in my head haha
18:03:52dconrad_webI'll probably be back in a bit when I get home
18:03:59 Quit dconrad_web (Quit: Connection closed)
18:13:22amachronici think root redirect is the problem
18:13:39amachronicf88ea12bac is ok
18:15:12amachronicafter that it's ok *if* I use root redirect
18:15:59amachronicif I remove the redir file and run from the real root it stops working. go figure.
18:16:21 Quit lebellium (Quit: Leaving)
18:19:44amachronicguess I'll look into it this weekend...
18:20:29 Quit amachronic (Quit: amachronic)
18:41:12 Join dconrad [0] (~dconrad@
18:45:52dconradthat reminds me, I want to try to get root redirects set up for my devices, it would be cool to share the same card between them
18:45:58dconradtry it out
18:52:52 Quit ZincAlloy (Quit: Leaving.)
18:53:04 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:b810:b519:d281:16ef)
18:53:06 Quit ZincAlloy (Client Quit)
18:55:11***Saving seen data "./dancer.seen"
19:05:11_bilgusamachonic yes something is afoot with root redirect
19:05:28_bilgusyou can cause it to happen real quick by running a lua program
19:05:56_bilgusthen if it makes it through run it again and the dircache is trashed won't even remount cards
19:06:37_bilgusso I'm thinking it probably has a race or it loses the root entry (probably needs guarded against that..
19:07:13_bilgusI have a good test suite :P hopefully that means it won't take long to find
19:07:44_bilgusamachronic the issue in the buffering thread is common to all ARM devices >= ARM5
19:08:16_bilgussorry the bug I found in the buffering thread
19:09:06_bilgusthe signed div ASM specifically
20:33:27 Quit dconrad ()
20:47:10_bilgus! Too many open Files !
20:55:13***Saving seen data "./dancer.seen"
20:59:49braewoodswhat does division have to do with running out of file handles?
21:25:37_bilgusbut I did just find a bug that doesn't close file handles if the path is absolute
21:26:26_bilgusthats why lua is so good at triggering it, its module loader tries like 2 abs paths per module opening
21:29:31_bilgushaha I wonder if root redirect is making all those calls absolute now
21:29:55_bilgusfd is at 1 now after fixing that
21:30:07_bilguswas at 11 which is max
22:55:14***No seen item changed, no save performed.

Previous day | Next day