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 2019-07-30

00:16:35 Quit ZincAlloy (Quit: Leaving.)
00:19:17 Quit paulk-leonov (Ping timeout: 250 seconds)
00:20:08 Join dys [0] (
00:21:45 Join paulk-leonov [0] (
00:28:23 Quit pR0Ps (Ping timeout: 268 seconds)
00:30:07 Join mendelmunkis [0] (
00:30:11 Join amiconn_ [0] (jens@rockbox/developer/amiconn)
00:30:11 Nick amiconn is now known as Guest49930 (~jens@rockbox/developer/amiconn)
00:30:11 Nick amiconn_ is now known as amiconn (jens@rockbox/developer/amiconn)
00:30:40 Join pixelma_ [0] (marianne@rockbox/staff/pixelma)
00:30:40 Nick pixelma is now known as Guest95349 (~marianne@rockbox/staff/pixelma)
00:30:40 Nick pixelma_ is now known as pixelma (marianne@rockbox/staff/pixelma)
00:31:50 Quit Guest49930 (Ping timeout: 264 seconds)
00:31:50 Quit Guest95349 (Ping timeout: 264 seconds)
00:38:07mendelmunkisdoes anyone know why ROLO would give me a bad checksum error?
00:38:23mendelmunkisassuming that I am attempting to use the correct binary.
00:38:55 Quit lebellium (Quit: Leaving)
00:41:37***Saving seen data "./dancer.seen"
00:48:01 Quit Jinx (Ping timeout: 248 seconds)
00:49:59mendelmunkiswhats the main function in a puzzles plugin?
00:52:08_Bilguson ROLO you mean you ARE loading the right binary?
00:55:58__builtinmendelmunkis: plugin_start()
00:57:19__builtineach puzzle is built from a set of common sources (see rockbox.c) and the puzzle code itself in src/
00:57:52 Join Jinx [0] (~Jinx@unaffiliated/jinx)
01:06:58 Quit vmx (Remote host closed the connection)
01:08:55speachywhat target for the ROLO error?
01:08:59mendelmunkis_builtin thanks
01:09:46mendelmunkis_bilgus: speachy: on my fuze+ my tests builds will occasionaly fail with a rolo bad checksum error then work next time I try them.
01:10:23speachycan't comment on that one. just fixed up some incorrect IDs causing ROLO failures on a few less-stable targets..
01:10:35_Bilgusso like are you copying the file to the device and then ROLO picks up the new file or are you calling rolo?
01:11:21_Bilguslike going thru the menus and manually selecting the firmware
01:11:34mendelmunkisI have a separate RBDIR for test builds. I'm running it from there.
01:11:50mendelmunkisIn other words calling ROLO
01:13:03_Bilgushmm thats odd I could see copying it failing but if it's already on disk... thats worse
01:13:53_Bilguscould you try putting your file on the sdcard and running it from there
01:14:39mendelmunkisIt's intermittent a file will error and then run the next time I try with the same file.
01:15:02mendelmunkisWhy would an sd card change anything?
01:15:46mendelmunkisIt's unlikely to be hardware given as I think it followed me across devices.
01:18:27_Bilgusits a pretty simple set of prerequisites for ROLO to load the file so if bad checksum then the checksum it calculated doesn't match
01:18:44_Bilguswhat causes that? bad data
01:19:23_Bilgusif its intermittent then its bad data from the file or something corrupting it
01:20:19_BilgusI've never seen bad checksum during rolo on fuze+ clipzip i've seen it on the clip+ but that was genuine
01:20:28mendelmunkisThats what I thought. but how does bad data get uncorrupted?
01:21:24_Bilgusthe read succeeds, if its at least slightly reproducible I'd put a splashf in the disk handler on re-read
01:21:44_Bilgusi'd say logf build but that is probably a bad idea
01:24:41_Bilgusyeah nm if it fails then you can still dump it
01:24:53_Bilgusand if not then who cares
01:29:37 Quit mendelmunkis (Ping timeout: 248 seconds)
01:29:49 Join mendelmunkis [0] (
01:34:57 Quit mendelmunkis (Ping timeout: 258 seconds)
01:55:01 Join pR0Ps [0] (
02:27:12__builtinhmm... looks like I may have a race condition between the sound and model code
02:27:42__builtinthat would explain why the crash only occurs at high sample rates
02:29:53*__builtin thinks two threads might be allocating the same chunk of memory simultaneously
02:37:12*__builtin confirms that... disabling one of the conflicting threads eliminates the issue
02:37:55__builtinyep... the alloc function isn't reentrant
02:41:40***Saving seen data "./dancer.seen"
02:59:38speachymalloc/free are usually prime candidates for locking
03:00:57__builtinthing is, quake rolls its own allocator
03:01:06__builtinso anything in tlsf is rendered useless
03:02:34__builtinbut adding a couple of locks to zone.c isn't a bad idea, yes
03:03:32__builtinwell, it looks like we have tlsf's locks disabled
03:06:22 Quit pR0Ps (Quit: Quitting)
03:06:34_Bilgusleave them that way lol
03:09:53 Join pR0Ps [0] (
03:17:35*speachy grumbles at his feline helper.
03:26:44fs-bluebotBuild Server message: New build round started. Revision 4f469e9, 280 builds, 11 clients.
03:27:25speachyminor fix for the debug menu's partition dumper
03:32:53*__builtin is curious how the original SDLQuake managed without synchronization
03:33:28 Join mendelmunkis [0] (
03:33:53_Bilgusor segregation...
03:34:04__builtinprobably the former
03:34:21__builtinthe x86 version had a bunch of weird crashes, I bet they're related
03:49:04fs-bluebotBuild Server message: Build round completed after 1340 seconds.
03:49:09fs-bluebotBuild Server message: Revision 4f469e9 result: All green
03:49:10fs-bluebotBuild Server message: New build round started. Revision 94109a9, 280 builds, 11 clients.
03:49:52 Quit mendelmunkis (Ping timeout: 258 seconds)
03:57:41_Bilguswell I think I got G#2167 figured out
03:57:43fs-bluebotGerrit review #2167 at : tagtree add %tagtree_reload to allow hot reloading of tagnavi config by William Wilgus
03:58:31_Bilgushopefully someone that knows more about the tagcache can tell me if its just a BAD IDEA (tm) but it seems to work pretty well
03:58:52__builtinspeachy, _Bilgus: stickers should arrive at my place by this Friday. Will let you know when I ship.
03:59:33_Bilgusbeing able to unload the tagcache could also be used for freeing ram for emergency or in times of need but its static for now
03:59:40_Bilgus__builtin, thanks
04:00:34*__builtin might not be able to ship until next week, though
04:07:49fs-bluebotBuild Server message: Build round completed after 1120 seconds.
04:07:51fs-bluebotBuild Server message: Revision 94109a9 result: All green
04:32:23fs-bluebotBuild Server message: New build round started. Revision a89bf68, 280 builds, 11 clients.
04:33:23speachyOkay, I think g#2151 is finally ready. Sanely handles HAVE_MULTIDRIVE and !HAVE_MULTIDRIVE including hotplug.
04:33:24fs-bluebotGerrit review #2151 at : Make info screen properly multi-volume aware. by Solomon Peachy
04:33:50speachyinterestingly there are no HAVE_MULTIVOLUME targets that lack HAVE_MULTIDRIVE.
04:38:38speachyat worst it's no more broken than the old code.
04:41:43***Saving seen data "./dancer.seen"
04:48:57fs-bluebotBuild Server message: Build round completed after 993 seconds.
04:49:01fs-bluebotBuild Server message: Revision a89bf68 result: All green
04:59:05_Bilgusspeachynot even the soinys?
04:59:30_Bilgusspeachy, not even the sonys?
05:00:19_Bilgusor the creative whatever the brick is called
05:36:37speachyyeah, only the four creative bricks had MULTIVOLUME defined at all, and they all defined MULTIDRIVE too.
05:39:32 Quit TheSeven (Ping timeout: 250 seconds)
05:40:04 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
06:41:47***Saving seen data "./dancer.seen"
07:01:16 Join mendelmunkis [0] (
07:15:51 Quit mendelmunkis (Ping timeout: 245 seconds)
07:34:55 Join ZincAlloy [0] (~Adium@2a02:8108:9440:dfc:a87b:17cb:6e21:3ffb)
07:40:09 Quit ZincAlloy (Ping timeout: 276 seconds)
08:00:30 Join mendelmunkis [0] (
08:07:06 Quit mendelmunkis (Ping timeout: 258 seconds)
08:07:24 Join mendelmunkis [0] (
08:17:02 Quit mendelmunkis (Remote host closed the connection)
08:34:15 Quit TheSeven (Ping timeout: 264 seconds)
08:37:06 Join mendelmunkis [0] (
08:41:08 Quit mendelmunkis (Read error: Connection reset by peer)
08:41:22 Join mendelmunkis [0] (
08:41:51***Saving seen data "./dancer.seen"
08:47:21 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
08:47:56 Join petur [0] (~petur@
08:47:56 Quit petur (Changing host)
08:47:56 Join petur [0] (~petur@rockbox/developer/petur)
08:48:45 Quit mendelmunkis (Ping timeout: 272 seconds)
09:24:17 Quit dys (Ping timeout: 248 seconds)
09:53:48 Join mendelmunkis [0] (
09:54:46 Join ZincAlloy [0] (~Adium@2a02:8108:9440:dfc:ac3a:d114:725b:ffee)
09:59:00 Quit ZincAlloy (Ping timeout: 252 seconds)
10:04:50 Quit yosafbridge (Quit: Leaving)
10:09:23 Quit mendelmunkis (Ping timeout: 258 seconds)
10:10:26 Join yosafbridge [0] (
10:13:58 Join cereal_eater [0] (b945f46a@
10:15:04cereal_eaterI use a partitioned SD card in my Cowon D2 to hide files from the OF. Do the latest code changes relatet to HAVE_MULTIDRIVE etc. afffect this?
10:35:55 Join mendelmunkis [0] (
10:36:12 Join dys [0] (~dys@2003:5b:203b:100:a64c:c8ff:fef4:13a6)
10:40:18mendelmunkisis pamaury around anymore?
10:41:52***Saving seen data "./dancer.seen"
10:46:33 Quit mendelmunkis (Ping timeout: 272 seconds)
10:47:35 Join TheSphinX_ [0] (
10:48:39 Quit TheSphinX^ (Ping timeout: 276 seconds)
10:48:39 Nick TheSphinX_ is now known as TheSphinX^ (
10:51:27*user890104 successfully built __builtin's sdl apps for win32 and win64 simulator
11:17:50 Quit cereal_eater (Ping timeout: 260 seconds)
11:52:36 Quit flabrus (Remote host closed the connection)
11:52:52 Join flabrus [0] (
11:52:52 Quit flabrus (Changing host)
11:52:52 Join flabrus [0] (~beard@unaffiliated/flabrus)
12:01:06 Join ZincAlloy [0] (
12:05:15 Quit ZincAlloy (Ping timeout: 244 seconds)
12:17:44 Join vmx [0] (
12:34:43 Quit speachy (Remote host closed the connection)
12:41:53***Saving seen data "./dancer.seen"
13:02:57 Quit Jinx (Ping timeout: 248 seconds)
13:13:28 Join Jinx [0] (~Jinx@unaffiliated/jinx)
13:19:49 Quit rasher (Ping timeout: 272 seconds)
13:19:58 Join rasher [0] (~rasher@rockbox/developer/rasher)
13:28:23 Join speachy [0] (d102414d@
13:43:36speachyso, it looks like an explicit HAVE_MULTIVOLUME in the presence of HAVE_MULTIDRIVE has a meaning after all −− it enables support for up to 4 volumes per disk.
13:45:31speachyso I have two choices −− revert the changes to the Creative Zen configurations to re-enable that quirk, OR make it so that all targets with MULTIDRIVE get that 4-per-disk thing too.
13:46:40speachyI good argument can be made either way. Personally I think that anything with removable storage should support the full 4 partitions/volumes per drive.
13:48:56speachycode size in the main binary goes up by ~100 bytes on the handful of ARM targets I rebuilt, and ~250 bytes on the MIPS. (Interestingly, the zen x-fi saw a 64 byte bump, vs the 6.5KB bump on the build page..)
13:50:37speachyoh, I see that number incldes RAM. nm.
14:08:51 Join massiveH [0] (
14:26:56 Quit speachy (Remote host closed the connection)
14:41:56***Saving seen data "./dancer.seen"
14:49:45 Join derp [0] (
14:50:45 Quit derp (Client Quit)
15:11:42 Quit massiveH (Quit: Leaving)
15:17:48 Join speachy [0] (40eebded@
16:09:37 Quit michaelni (Ping timeout: 248 seconds)
16:21:52 Join michaelni [0] (
16:41:57***Saving seen data "./dancer.seen"
17:16:55 Quit petur (Quit: Connection reset by beer)
17:19:04__builtinuser890104: nice!
17:37:47 Quit IncognitoMan (Ping timeout: 246 seconds)
17:38:03_BilgusI'm still not sure why cereal_eater thinks code changes in rb would affect his OF or am I missing something about the cowond2?
17:41:01speachyI think what's going on is that he's relying on rockbox only picking up one partition, so the partition with his OF is completely ignored.
17:42:06speachymeanwhile I need to decide what to do about my patch changing the configs on the Creative Zens
17:42:46speachyie restore their old behavior (so only they see all partitions on all drives) or enable full multivolume for all multidrive targets.
17:43:44speachyIMO the latter is the better choice.
17:43:47_BilgusIIRC the zen has an incompatible partition I don't remember the specifics
17:44:40_Bilgusits like the OF is something funky that rb doesn't like or the otherway around
17:45:32 Join IncognitoMan [0] (~Incognito@unaffiliated/incognitoman)
17:45:44speachyWell, the way it was before I committed, rockbox would see all the partitions. That is a unique setup for the Zen, and the X-Fis.
17:45:55_BilgusRockbox and Creative OS do not use the same file system so one cannot use both at the same time. In order to run Rockbox, you will need to reformat the internal storage as FAT. If you want to restore the Creative OS, you will need to reformat it using the recovery mode. It is strongly advised to put music on the SD card since the internal storage handling might change in the future and require formatting once again
17:46:10_Bilgusso I'd say it doesn't really matter
17:47:12_Bilgushonestly how important is multi partitions to the normal user?
17:47:50_BilgusFat32 supports what like ~2Tb
17:48:25speachyprobably not very.
17:48:59speachysince one would have to go out of their way to format an SD card with multiple partitions
17:49:04_Bilguspersonally I'd say chuck the multi partition in favor of code savings and make it easy enough to enable
17:50:00_Bilgusanyone savy enough to make multi partitions is savy enough to make a multi part build too
17:51:06_Bilgusthough it could be argued that you might want to see multiple partitions say from an android card or raspi card but again how likely?
17:51:24speachymost of the random SD cards I have lying around are multi-partitioned due to Linux boots
17:51:40speachythat's how I discovered this um, quirk. :)
17:51:46_Bilgusbut others might feel otherwise thats what sucks about making decisions (in a vacuum too)
17:52:23_Bilgusstill do you want to be mucking with that partition in your DAP?
17:52:25speachypersonally I think it's worth enabling across the board, as the only real argument against it is code/ram.
17:53:10speachyit concerns me that the MULTIDRIVE && !MULTIVOLUME code has somewhat bitrotted too.
17:53:25_BilgusIt doesn't really matter from the code side ram OTOH
17:54:04speachyRAM is more concerning. code size is under 100 bytes.
17:54:36_Bilgusnow 5k of ram from 8-64 mb is nothing but 5k from 1mb might be
17:55:00_Bilgushow many targets are multi volume and 1mb ram?
17:55:11_BilgusI'd say 1
17:55:38_Bilgusand its probably HWcodec which we want to end support for anyways
17:56:43speachyThere are no MULTIVOLUME targets other than the Creative bricks.
17:57:38_Bilgusmaybe if __builtin gets github access he can branch off the HWCodecs ;)
17:57:57_Bilgussorry I meany multidrive
18:00:33 Quit vmx (Ping timeout: 248 seconds)
18:28:45 Join cereal_eater [0] (b945f425@
18:31:19 Join ZincAlloy [0] (~Adium@2a02:8108:9440:dfc:910a:bca4:b792:12f6)
18:31:41cereal_eater_Bilgus: I use mainly Rockbox and sometimes the OF on my D2. All my files would exceed the file limit that the OF has. So I have all my music on the second partition of a SD card, which Rockbox can see, but the OF not.
18:36:30 Quit cereal_eater (Ping timeout: 260 seconds)
18:41:59***Saving seen data "./dancer.seen"
18:45:41 Join cereal_eater [0] (b945f471@
18:47:13cereal_eaterSome music is on the internal storage fpr toying with flash skins, that the OF supports and on the first very small partition of the SD card is only the .rockbox folder..
18:47:42cereal_eaterSo I was curious if my usecase is still possible with future builds.
18:49:13 Part cereal_eater
19:01:32speachyI don't see why it wouldn't work.
19:02:08speachyall of my bellyaching here is because I'm trying to fix some bugs and stumbling down rabbit holes in the process.
19:09:10 Join krabador [0] (~krabador@unaffiliated/krabador)
19:09:40 Quit krabador (Remote host closed the connection)
19:19:41 Join lebellium [0] (
19:31:01 Join krabador [0] (~krabador@unaffiliated/krabador)
20:02:16 Join MrZeus [0] (~MrZeus@2a02:c7f:70d4:cd00:6cd0:882b:9574:e719)
20:10:50 Quit dys (Ping timeout: 264 seconds)
20:42:03***Saving seen data "./dancer.seen"
20:57:09 Join dys [0] (
21:37:17 Quit krabador (Remote host closed the connection)
22:15:09__builtinuser890104: push your changes and I'll review
22:15:34user890104ah right, i forgot... too much $dayjob today
22:21:42user890104__builtin: g#2169
22:21:44fs-bluebotGerrit review #2169 at : sdl: Fixes for the SDL apps, so they support building simulator on Windows by Vencislav Atanasov
22:30:20__builtinI'll take a look later
22:35:46 Join MrZeus_ [0] (~MrZeus@2a02:c7f:70d4:cd00:6cd0:882b:9574:e719)
22:39:05 Quit MrZeus (Ping timeout: 276 seconds)
22:42:06***Saving seen data "./dancer.seen"
22:55:09__builtinuser890104: one question. Why did you add #undef WIN32 in SDL_video.c?
23:40:17 Quit speachy (Ping timeout: 260 seconds)
23:45:11 Join reductum [0] (

Previous day | Next day