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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2022-05-07

00:10:22 Quit massiveH (Quit: Leaving)
00:42:58 Join ZincAlloy [0] (
00:57:48***Saving seen data "./dancer.seen"
02:06:52 Join lebellium [0] (
02:09:08 Quit ZincAlloy (Quit: Leaving.)
02:19:47 Quit amk (Ping timeout: 240 seconds)
02:22:49 Join amk [0] (~amk@
02:57:51***Saving seen data "./dancer.seen"
02:59:51 Quit akaWolf (Ping timeout: 252 seconds)
04:03:48 Join lebellium_ [0] (
04:05:48 Quit lebellium (Ping timeout: 240 seconds)
04:57:52***Saving seen data "./dancer.seen"
06:36:29 Join akaWolf [0] (
06:57:53***No seen item changed, no save performed.
07:05:33 Quit jschwart (Quit: - Chat comfortably. Anywhere.)
07:07:47 Join jschwart [0] (
07:08:06 Join shawn196 [0] (
07:09:29 Quit S|h|a|w|n (Ping timeout: 256 seconds)
07:15:25 Quit shawn196 (Quit: Leaving)
08:22:47 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
08:57:56***Saving seen data "./dancer.seen"
10:10:41 Part emacsomancer (WeeChat 3.5)
10:17:32 Join amachronic [0] (~amachroni@user/amachronic)
10:19:31amachronicchris_s, I can't reproduce a crash on pictureflow with the current daily build could you provide more details?
10:28:23 Join chris_s [0] (
10:35:44chris_sIt seems to crash consistently for me (as long as the database contains more than 1 album) as soon as I (re-)launch the plugin after the album index has already been built.
10:35:45chris_s(Clean, daily build with default settings, except “Load to RAM” enabled (and restarted afterwards).
10:35:45chris_sReverting the commit reliably fixes it. I can try to find out if it is somehow only triggered by certain albums, if you continue to be unable to reproduce it.
10:37:09amachronicwhat's the panic message?
10:38:27chris_s*PANIC* Exception occurred: Address Error (Logd or Ifetch) [0x83e1ecb3] at 0x83e0708c (stack at 0x80002a14)
10:38:41amachronicokay awesome, let's see what that is.
10:42:17amachronicsomewhere in pictureflow_main, apparently
10:45:25_bilgusamachronic looks suspicious
10:46:08_bilgusdiv by 4 + 3 or 7??
10:47:03amachronicthat's a div by 4, plus sizeof(long)-1 to cover the ALIGN_BUFFER below
10:47:13_bilgussould be - maybe the macro normally flips the sign?
10:47:51_bilgusah you remeber when you fixed my bug a long time ago in pf
10:48:16amachronicALIGN_DOWN is just rounding pf_idx.buf_size/4 down to a multiple of four.
10:49:02amachronicalthough now that i think about it aa_bufsz is wrong since it's including the alignment padding
10:49:39_bilgussame spot i think
10:51:45 Join Retr0id6 [0] (~Retr0id@user/retr0id)
10:55:35 Join Acou_Bass [0] (
10:56:59 Quit Piece_Maker (*.net *.split)
10:56:59 Quit bleb (*.net *.split)
10:56:59 Quit jackie (*.net *.split)
10:57:00 Quit berber6 (*.net *.split)
10:57:00 Quit rudi_s (*.net *.split)
10:57:00 Quit Arsen (*.net *.split)
10:57:00 Quit Bobathan (*.net *.split)
10:57:00 Quit Retr0id (*.net *.split)
10:57:00 Quit cb (*.net *.split)
10:57:00 Nick Retr0id6 is now known as Retr0id (~Retr0id@user/retr0id)
10:57:00_bilgusamachronic I'd say you found it it later subtracts from the buffer and pulls it out of alignment
10:57:14 Join bleb [0] (~cm@user/bleb)
10:57:14 Join jackie [0] (
10:57:14 Join berber6 [0] (
10:57:14 Join rudi_s [0] (~simon@user/rudi-s/x-7673890)
10:57:14 Join Arsen [0] (arsen@managarm/dev/Arsen)
10:57:14 Join Bobathan [0] (
10:57:14 Join cb [0] (
10:57:58***Saving seen data "./dancer.seen"
10:58:27amachronicit could be but if pf_idx.buf is not aligned it should get fixed up later anyhow
10:59:15amachronici guess we'll see if it fixes the problem
10:59:42rb-bluebotBuild Server message: New build round started. Revision f661dc596e, 303 builds, 8 clients.
11:10:59_bilguswell unless alignof has some weird behavior thats the only thing I see off
11:13:22_bilgusah actually it appears it does
11:13:54_bilgusstruct foo { int x; char y; } foo1;
11:13:54_bilgusthe value of __alignof__ (foo1.y) is 1, even though its actual alignment is probably 2 or 4, the same as __alignof__ (int). It is an error to ask for the alignment of an incomplete type other than void.
11:14:23_bilgusbut you aren't pulling members
11:14:28chris_ssame panic message, it looks like
11:18:12rb-bluebotBuild Server message: Build round completed after 1111 seconds.
11:18:19rb-bluebotBuild Server message: Revision f661dc596e result: 4 errors 0 warnings
11:19:07amachronicok. all the inlining is making it hard to see where it's faulting from the disasm
11:21:05_bilgusthere is also something about some not requiring alignment and then defaulting t the minimum (??1 char presumably)
11:22:28amachronici doubt they actually do that on any sane machine - even on x86 alignment makes a huge performance difference
11:22:37_bilgusmaybe warn_if_not_aligned (alignment)
11:22:59_bilgusthen you might see it at compile time
11:23:33amachronicin any event on riscs, alignment is a hard requirement so i wouldn't worry about that
11:24:11_bilgusidk after chasing a 0x4 vs 0x8 bug through ARM Thumb I don't assume
11:24:38_bilgusidk what processors those two are but..
11:26:57amachronicokay I think I found it
11:27:23amachronicload_album_index is not aligning stuff properly, surprise surprise...
11:28:10amachronic g#4542
11:28:14rb-bluebotGerrit review #4542 at : pictureflow: fix album index alignment in load_album_index by Aidan MacDonald
11:28:46_bilgusoh see I saw it here
11:29:21amachronichmm that looks pretty sketchy too
11:29:34_bilgusits char aligned after a few of those
11:30:43_bilgusprobably a better idea is to give them an allocator
11:31:07_bilgusbut its that way because a lack of resources
11:33:24amachronici'll give it a 2nd pass later and fix up that stuff
11:34:03amachronican allocator sounds like a good idea but hard to see how it'd work cause of how it builds the indexes backward etc.
11:34:19_bilgusI didn't like working on PF whatsoever so I'm glad chris_s is interested in it :)
11:34:41chris_slol, thanks – seems to get rid of the crash  :)
11:34:51amachronicyes it's a nightmare :D
11:35:01rb-bluebotBuild Server message: New build round started. Revision 4b293285ea, 303 builds, 8 clients.
11:39:25_bilgusI'm still working on the keyboard code Ive started over a few times now because I feel its way too large code wise
11:39:44_bilguslike adding 5-10k large
11:40:50_bilgusIve got the NUMPAD though and indexes I'm debating to keep the code to capitalize unicode or go back to just shifting pages
11:41:38_bilgusI'm leaning towards the latter bc its a lot less code
11:42:06_bilguslet the person making the keyboard reson about it
11:44:16_bilgusI've also started thinking about touch screens with my current scheme I think I can just auto size the font to max that will fit in the available space and hopefully get it large enough to touch
11:45:48_bilgusI'm currently using macros to exclude things from small screen vs large screens but I think it might be better to use anouther few PUA unicode chars to allow different targets in the same board
11:47:58_bilguslike FEF0 for large screens FEF1 for small screens FEF2 for touchscreens then the target can parse the def from user kbd too
11:48:28rb-bluebotBuild Server message: Build round completed after 808 seconds.
11:48:30rb-bluebotBuild Server message: Revision 4b293285ea result: 4 errors 0 warnings
11:49:30amachronic_bilgus that all sounds like good ideas but I wonder is the kbd flexible enough for small vs. large screen to make a big difference?
11:49:53amachronicthe last time i tried your patch (weeks ago) it didn't make a lot of use of the available screen space
12:00:39_bilgusthe num pad takes up some of that but the clip zip has one page for just A-Z
12:01:15_bilgusthe bigger screens should be able to show most common symbols too
12:02:25_bilgusbut probably not touchscreens..
12:03:30_bilgusI still dislike the menu I think* want it as a horizontal scroller
12:04:56_bilgusthen you could touch arrows or hit the index to go to a new page
12:06:43_bilgusbut I figure thats something to add for everyone to use so it can just be a later possibility
12:07:21amachronicyeah, it's better not to go overboard just get the basic functionality in and build extras on top
12:08:27amachronicthe current kbd is nearly unusable with the touchscreen anyway so leaving that out at first isn't a big deal imho
12:10:34_bilgusYeah I figure get everything as modular as possible for later expansion in its current form its very rigid
12:11:51_bilgusthe vports help a lot in that regard
12:12:40_bilgusI am liking the long press for caps though its become v ingrained to te point its bothersome to use the old builds
12:17:20 Quit amachronic (Quit: amachronic)
12:58:02***Saving seen data "./dancer.seen"
13:27:10 Join ZincAlloy [0] (
13:51:31 Quit ZincAlloy (Quit: Leaving.)
14:07:46 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:6cf3:540e:ced2:aa6b)
14:24:18 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
14:58:05***Saving seen data "./dancer.seen"
16:37:17 Quit Acou_Bass (Ping timeout: 256 seconds)
16:58:08***Saving seen data "./dancer.seen"
17:29:18 Quit ZincAlloy (Quit: Leaving.)
18:08:12 Quit chris_s (Ping timeout: 260 seconds)
18:29:21 Join massiveH [0] (
18:58:11***Saving seen data "./dancer.seen"
20:32:25 Quit lebellium_ (Quit: Leaving)
20:58:12***Saving seen data "./dancer.seen"
21:33:45 Join jadzia [0] (~jadzia@2604:3d09:1b85:4200::6005)
21:56:52 Quit munkis (Remote host closed the connection)
22:58:16***Saving seen data "./dancer.seen"
23:09:43 Quit dbohdan (Read error: Connection reset by peer)
23:10:14 Join dbohdan [0] (~dbohdan@user/dbohdan)

Previous day | Next day