--- Log for 12.01.123  Server: calcium.libera.chat  Channel: #rockbox
--- Nick: rb-logbot  Version: Dancer V4.16  Started: 17 days and 10 hours ago

00.12.33 ***     Saving seen data "./dancer.seen"
02.12.34 ***     No seen item changed, no save performed.
04.12.36 ***     No seen item changed, no save performed.
04.16.05 Join    mendel_munkis [0] (~mendel_mu@2600:4041:5ad1:6800:b225:aaff:fe48:a92c)
04.16.35 Quit    munkis (Ping timeout: 255 seconds)
04.56.08 #       <rb-bluebot> Build Server message: 3New build round started. Revision 541960a110, 303 builds, 8 clients.
04.56.09 #       <rb-bluebot> 3rbcodec/dsp: restore configure loop in dsp_init() (FS#13386) by Aidan MacDonald
04.56.12 Join    amachronic [0] (~amachroni@user/amachronic)
05.21.32 #       <rb-bluebot> Build Server message: 3Build round completed after 1524 seconds.
05.21.35 #       <rb-bluebot> Build Server message: 3Revision 541960a110 result: All green
05.29.06 Quit    advcomp2019 (Read error: Connection reset by peer)
06.05.57 #       <rb-bluebot> Build Server message: 3New build round started. Revision 39439f6909, 303 builds, 8 clients.
06.05.57 #       <rb-bluebot> 3Add INIT_ATTR to system_init() by Aidan MacDonald
06.06.39 #       <_bilgus> amachronic, judging by ifdef HAVE_EXTENDED_MESSAGING_AND_NAME guarding queue_send I don't think every device has it
06.06.46 #       <amachronic> it's there everywhere
06.06.54 #       <amachronic> except bootloaders
06.07.09 #       <amachronic> it's used by codec & audio thread
06.07.14 #       <_bilgus> ah ok
06.07.43 #       <amachronic> playlist.c is a horrible mess :)
06.08.08 #       <_bilgus> yes I'm well aware after that slast mutex adventure
06.12.39 ***     Saving seen data "./dancer.seen"
06.32.44 #       <rb-bluebot> Build Server message: 3Build round completed after 1609 seconds.
06.32.45 #       <rb-bluebot> Build Server message: 3Revision 39439f6909 result: All green
06.33.52 #       <rb-bluebot> Build Server message: 3New build round started. Revision 7e5fc4076a, 303 builds, 8 clients.
06.33.53 #       <rb-bluebot> 3Add INIT_ATTR to i2c_init() by Aidan MacDonald
06.54.15 #       <rb-bluebot> Build Server message: 3Build round completed after 1224 seconds.
06.54.17 #       <rb-bluebot> Build Server message: 3Revision 7e5fc4076a result: 130 errors 4 warnings
06.57.26 #       <rb-bluebot> Build Server message: 3New build round started. Revision 16a32c576c, 303 builds, 8 clients.
06.57.26 #       <rb-bluebot> 3Fix red in 7e5fc4076a by Aidan MacDonald
07.16.46 #       <rb-bluebot> Build Server message: 3Build round completed after 1162 seconds.
07.16.48 #       <rb-bluebot> Build Server message: 3Revision 16a32c576c result: All green
07.25.18 #       <_bilgus> ok I think I have the dc_thread_playlist thing worked out with g#5041
07.25.20 #       <rb-bluebot> 3Gerrit review #5041 at https://gerrit.rockbox.org/r/c/rockbox/+/5041 : 3[BugFix] playlist.c DIRCACHE stop scanning when changing indices by William Wilgus
07.27.57 #       <_bilgus> the only really dicey spot is playlist_insert_track() as it has a sync switch
07.29.25 #       <_bilgus> otherwise I think everything else shouldn't be affected much since I kept it out of loops or at least I didn't observe anywhere else that was noticeably slower
07.31.39 #       <_bilgus> esentially everything that calls dc_load_playlist_pointers gets a matching call to dc_discard prior or the functions further up the call stack do at least 
07.36.15 #       <amachronic> _bilgus the thread needs to do a queue_reply() in the switch to unblock the sender
07.36.57 #       <_bilgus> ah crap thanks!
07.37.30 #       <amachronic> np :D
07.42.20 #       <amachronic> I think recreate_control_unlocked() needs to stop the thread too since it's modifying playlist->indices
07.43.47 #       <amachronic> also missing dc_discard_playlist_pointers() in add_indices_to_playlist()
07.45.43 #       <amachronic> the static bool discard in playlist_insert_track() makes me suspicious
08.00.26 #       <_bilgus> all callers to add_indices call discard first
08.01.06 #       <_bilgus> either in empty_playlist or other places
08.02.14 #       <_bilgus> playlist_insert_track gets called in a loop with a switch to indicate sync if caller doesn't call sync it might get funky but hence the dicey comment
08.03.21 #       <_bilgus> the caller of recreate_control discards first
08.04.42 #       <amachronic> ah sorry I missed that for recreate_control
08.05.11 #       <_bilgus> eh its messy I appreciate the extr eyes
08.05.47 #       <_bilgus> i'm not sure what else to do with pl_insert_track though
08.06.27 #       <amachronic> add_indices I don't like since it reloads the pointer but doesn't discard it
08.07.15 #       <amachronic> probably better to keep the discard & the reload in the same function
08.09.06 #       <_bilgus> mmm good point since it gets clled in loops thats probably better anyway 
08.10.40 #       <amachronic> with insert_track() it looks like sync=false means the caller is responsible for calling playlist_sync() afterwards
08.11.18 #       <amachronic> if so then doing an unconditional discard upon entry would work
08.11.52 #       <amachronic> then the problem is you don't want the delays from queue send <-> reply
08.12.34 #       <amachronic> perhaps you could add a boolean flag controlled by load/discard pointers
08.12.40 ***     No seen item changed, no save performed.
08.13.31 #       <amachronic> eg: if (dc_thread_running) { queue_send(...); dc_thread_running = false; }
08.17.24 #       <_bilgus> thats essentially what the dirty pointers flag does maybe just hoist it out of the thread
08.23.25 #       <_bilgus> sounds like some kind of insult dc_has_dirty_pointers
08.23.38 #       <_bilgus> probably british slang or something
08.27.54 #       <_bilgus> i'm amazed it let me rebase in gerrit after those changes I don't have great luck with that there
08.28.19 Quit    amachronic (Ping timeout: 260 seconds)
08.39.29 #       <_bilgus> amachronic i'll run around with these two patches the next day or two and see what pops out thanks for looking it over plus suggestions
08.41.25 #       <_bilgus> it seems to be faster too probably add_indices losing dc_load_playlist_pointers() I'd guess
08.52.29 Join    massiveH [0] (~massiveH@2600:4040:a992:a300:fd9b:f0e5:bbbd:9004)
09.07.42 Quit    massiveH (Ping timeout: 252 seconds)
09.20.26 Join    amachronic [0] (~amachroni@user/amachronic)
09.46.58 Quit    desowin (Changing host)
09.46.58 Join    desowin [0] (~linux@rockbox/developer/desowin)
10.12.41 ***     Saving seen data "./dancer.seen"
10.12.45 Join    paulk [0] (~paulk@vpn-0-22.aquilenet.fr)
10.24.54 Quit    amachronic (Quit: amachronic)
11.42.16 Join    othello7 [0] (~Thunderbi@pool-100-36-133-223.washdc.fios.verizon.net)
11.43.12 Quit    othello7 (Client Quit)
11.43.30 Join    othello7 [0] (~Thunderbi@pool-100-36-133-223.washdc.fios.verizon.net)
12.00.55 Join    advcomp2019 [0] (~advcomp20@user/advcomp2019)
12.12.43 ***     Saving seen data "./dancer.seen"
13.00.24 Join    l0x0l [0] (~l0x0l@097-093-024-013.res.spectrum.com)
13.23.33 Join    lebellium [0] (~lebellium@2a01cb040109a6009168b8f82cfd1bfd.ipv6.abo.wanadoo.fr)
13.41.01 Join    amachronic [0] (~amachroni@user/amachronic)
14.09.28 Quit    l0x0l (Changing host)
14.09.28 Join    l0x0l [0] (~l0x0l@user/l0x0l)
14.12.47 ***     Saving seen data "./dancer.seen"
14.20.24 Quit    l0x0l (Quit: Leaving)
14.24.55 Quit    amachronic (Quit: amachronic)
15.01.03 Join    amachronic [0] (~amachroni@user/amachronic)
15.10.33 #       <amachronic> has anyone used the pdbox plugin successfully on an ipod 1g to 4g or sansa e200?
15.11.21 #       <amachronic> I ask because it uses a COP thread, but doesn't look remotely set up for it and I can find no evidence it was ever tested on those targets.
15.11.47 #       <amachronic> I suspect it's broken, but I don't want to disable it if it does actually work.
15.40.35 #       <speachy> I can fire it up.
15.42.54 #       <speachy> ... no, I can't, as what I have handy is a mini2g.  d'oh.
15.53.29 #       <speachy> huh, it's installed as a viewer plugin, no wonder I iddn't see it.
15.53.40 #       <speachy> it runs but doesn't seem to "work"
15.57.23 #       <amachronic> admittedly I don't know what "works" is supposed to look like :-)
15.57.37 #       <speachy> me neither; it's the first time I've ever fired it up
15.57.59 #       <speachy> if it's a "beatbox" then I'd surmise the different buttons trigger different sounds.
15.58.13 #       <speachy> the display seems to indicate that the buttons are registered but that's it.
15.58.46 #       <amachronic> https://www.rockbox.org/wiki/PluginPdbox
15.59.07 #       <amachronic> seems you need to load a .pd file?
16.00.31 #       <amachronic> but, given it's lacking IBSS or SHAREDBSS attributes basically everywhere, I'd imagine it doesn't work on PP.
16.12.48 ***     Saving seen data "./dancer.seen"
16.15.49 Quit    amachronic (Quit: amachronic)
16.29.50 Quit    jacobk (Ping timeout: 272 seconds)
18.12.50 ***     Saving seen data "./dancer.seen"
18.22.05 Quit    lebellium (Quit: Leaving)
19.36.41 #       <rb-bluebot> Build Server message: 3New build round started. Revision 2d9cb673e5, 303 builds, 8 clients.
19.36.41 #       <rb-bluebot> 3add chunk_alloc to playlist.c #2 by William Wilgus
19.36.55 #       <_bilgus> lets see how try #2 goes
20.00.54 #       <rb-bluebot> Build Server message: 3Build round completed after 1453 seconds.
20.00.56 #       <rb-bluebot> Build Server message: 3Revision 2d9cb673e5 result: All green
20.12.53 ***     Saving seen data "./dancer.seen"
20.51.40 Join    jacobk [0] (~quassel@47-186-81-17.dlls.tx.frontiernet.net)
22.06.49 Quit    JanC (Remote host closed the connection)
22.07.08 Join    JanC [0] (~janc@user/janc)
22.12.56 ***     Saving seen data "./dancer.seen"
22.54.09 Join    l0x0l [0] (~l0x0l@097-093-024-013.res.spectrum.com)
23.09.12 Quit    hactar|ant (Ping timeout: 252 seconds)
23.16.10 Join    hactar|ant [0] (~zem@c-24-21-103-100.hsd1.or.comcast.net)
23.37.16 Quit    m01 (Quit: Konversation terminated.)
23.38.58 Quit    l0x0l (Quit: Leaving)
23.39.28 Join    m01 [0] (~quassel@vps-b172b88b.vps.ovh.net)