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 2020-08-27

00:22:44 Quit Galois (Remote host closed the connection)
00:22:50 Join nkf [0] (
00:22:50 Nick nkf is now known as Galois (
00:37:15***Saving seen data "./dancer.seen"
01:41:15 Quit amiconn (Quit: - Chat comfortably. Anywhere.)
01:41:15 Quit pixelma (Quit: .)
01:44:07 Join pixelma [0] (marianne@rockbox/staff/pixelma)
01:44:08 Join amiconn [0] (jens@rockbox/developer/amiconn)
02:35:49 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
02:37:19***Saving seen data "./dancer.seen"
04:30:14 Join akaWolf [0] (
04:33:11 Quit akaWolf (Remote host closed the connection)
04:37:23***Saving seen data "./dancer.seen"
05:04:06 Quit JanC (Remote host closed the connection)
05:04:28 Join JanC [0] (~janc@lugwv/member/JanC)
06:06:26 Join akaWolf [0] (
06:15:33 Quit akaWolf (Remote host closed the connection)
06:19:46 Join akaWolf [0] (
06:37:27***Saving seen data "./dancer.seen"
06:39:27 Join johnb2 [0] (
07:17:18 Join sakax [0] (~r0b0t@unaffiliated/r0b0t)
07:17:19 Join lebellium [0] (
08:08:25 Quit johnb2 (Ping timeout: 240 seconds)
08:37:30***Saving seen data "./dancer.seen"
09:44:35 Quit mendel_munkis (Ping timeout: 240 seconds)
09:47:40 Join mendelmunkis [0] (
09:54:42speachy __builtin: IMO a rebased g#1780 should go in now. Since
09:54:44fs-bluebot_Gerrit review #1780 at : Enable semaphore objects unconditionally by Franklin Wei
09:55:08speachysince the removal of hwcodec the conovoluted #ifdef ends up turning it on for nearly everything anyway.
09:56:42speachy(I think only the ihp-1xx series and maybe a couple of the PP targets don't already have it turned on now)
10:12:09 Join __Bilgus [0] (41ba23be@
10:12:50__Bilgusspeachy I got the grey lib stuff figured out g#2700
10:12:52fs-bluebot_Gerrit review #2700 at : Xduoo X3 - Grey scale lib update by William Wilgus
10:13:41__Bilgusthe routine is in C someone with mips assembler experience might wanna make it ASM but at least the code flow is easy to figure out :)
10:13:51speachywhat would be a good demonstrator of this?
10:14:14__Bilguscube, pictureflow, fire, test_grey
10:14:34speachyokeydokey, I'll add it to my next build
10:14:56speachydid you try this with g#2697?
10:14:58fs-bluebot_Gerrit review #2697 at : xduoox3: Fix delay funciton in LCD code to be cpufreq-agnostic (WIP) by Solomon Peachy
10:15:26__BilgusI compared to the clip+ to check it
10:15:54__Bilgusuh I don't think I did
10:16:06__BilgusI will real fast copypasta it
10:20:00 Join johnb4 [0] (
10:20:31speachyI can't see any difference, tbh.
10:21:50__Bilguswith the grey lib or the setupdelay?
10:22:09speachyah, cube looks different. plasma and fire, I can't really tell
10:22:28speachybefore/after the grey change. I already had the setupdelay going
10:22:43__Bilgussetupdelay works fine but I don't really notice it feeling faster
10:23:01__Bilgusbut it works
10:24:16__Bilgusso greylib works right now, test_scanrate is still broken so I'll get to it next so I can create a proper grey map calibration
10:24:20johnb4Am I right, that the new bootloader for Sansa E200 somehow needs blessing to be picked up by RB Utility?
10:24:24speachyif anything, the setupdelay should make things _slower_ −− the asm routine was, if anything, too short for the display setup/hold time.
10:24:49__Bilgusah well slower faster it felt the same
10:25:08speachyyeah, with nearly 600mhz in boost mode this thing is quite zippy. :)
10:25:21__Bilgusjohnb4 yes
10:25:49__Bilgusspeachy I haven't even tried it boosted on any of the plugins yet
10:26:10__BilgusI suppose the cube one probably is
10:26:11speachyunboosted it's 192MHz these days
10:26:52speachywas considering dialing it down further but I don't think that will make much of a difference for battery life.
10:27:36__BilgusI still have a few things to look over in the lcd controller I'll see if I can do some write combine like I did in the clip+ (zip?)
10:29:58__Bilgusoh the rolo lockups haven't happened again since one of those commits
10:30:16__Bilgusi'm guessing sd ..
10:31:19speachyI think I had it lock up once since then, but that might have been prior to the final SD changes
10:31:34__Bilgusgonna push 2700 unless someone wants to do it in ASM
10:32:00__BilgusI'm not familar enough with it to do better than the compiler I doubt
10:32:41speachygo for it.
10:33:25fs-bluebot_Build Server message: New build round started. Revision 5fb4c74, 280 builds, 9 clients.
10:33:26speachydo it correct first, we can make it faster later if it ever matters
10:33:59__Bilgusmy thoughts as well
10:37:33***Saving seen data "./dancer.seen"
10:40:51__builtin__Bilgus: some mips would be fun :)
10:40:59__builtinI'll take a look at some point
10:47:21fs-bluebot_Build Server message: Build round completed after 836 seconds.
10:47:23fs-bluebot_Build Server message: Revision 5fb4c74 result: All green
10:47:56speachy__builtin, I was toying with the idea of unifying the various midi players −− get the midiplay plugin to use the much, much newer timidity code in the sdl tree.
10:51:31 Quit johnb4 (Quit: Nettalk6 -
10:57:26 Join livvy [0] (~livvy@gateway/tor-sasl/livvy)
11:00:45__Bilgus__builtin, I'm sure you can make pretty quick gains looking at the currently generated asm
11:06:43speachyon the other hand, it's also easy to introduce subtle bugs doing that. I've fixed quite a few of those recently...
11:13:21__builtinspeachy: yeah, the timidity port was kind of a hack
11:13:46__builtinmy methodology was "dump in the source and pray"
11:13:48speachyso, dusting off g#2191, the response that came back was "set stereo_width to -1" but there doesn't appear to be a way to do that via the menus; only 0-250%.
11:13:50fs-bluebot_Gerrit review #2191 at : Support swapping playback channels by Solomon Peachy
11:37:52 Quit __Bilgus (Remote host closed the connection)
11:38:03 Quit Rower ()
11:40:24 Join Rower [0] (
12:07:38 Join MrZeus [0] (
12:37:36***Saving seen data "./dancer.seen"
13:34:04 Quit akaWolf (Quit: leaving)
13:38:21 Join johnb2 [0] (
13:41:41 Join akaWolf [0] (
13:42:30 Quit akaWolf (Remote host closed the connection)
13:44:14 Join akaWolf [0] (
14:03:25 Quit johnb2 (Ping timeout: 240 seconds)
14:19:25 Quit MrZeus (Ping timeout: 240 seconds)
14:37:39***Saving seen data "./dancer.seen"
14:52:30 Join PimpiN8 [0] (PimpiN8@gateway/vpn/nordvpn/pimpin8)
15:08:50 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…)
15:13:12 Join MrZeus [0] (
16:06:14 Quit MrZeus (Ping timeout: 258 seconds)
16:37:41***Saving seen data "./dancer.seen"
16:45:19speachyfinally firing up the X3ii. First impression −− the UI is _very_ laggy. and I just segfaulted it in the oscilloscope plugin. :P
16:47:58speachycrashed it exiting another plugin. so that's definitely a little wonky.
16:51:41 Join __BILGUS [0] (41ba23be@
16:52:08__BILGUSspeachy did the oscope crash when it got to the end of the screen?
16:53:23speachyno, when I exited.
16:53:54speachyalso crashed the matrix plugin the same way
16:54:05speachyon the rocker and the x3ii
16:54:26speachygranted for all intents and purposes they're identical except for the screen size
16:57:30__BILGUSoh ok then its not a rect size error at least
16:57:48 Join cockroach [0] (~blattodea@pdpc/supporter/active/cockroach)
16:58:20__BILGUSI had fixed an error in the oscope plugin way long ago was hoping it hadnt come back
16:58:25speachythe screen update was pretty bad on the x3ii −− the voice would finish speaking before the menu would finish painting
17:10:03speachythese things could _really_ use a native port
17:14:18__BILGUSI really don't like the as an app but it is quicker for sure
17:14:36__BILGUSto get working not running
17:15:17__BILGUSdo any of those have multiple cores or co processors?
17:24:56 Part edhelas
17:25:29 Join edhelas [0] (9d94237298@2a01:7c8:aab8:6b9:5054:ff:fec9:fd84)
17:31:36 Join ac_laptop [0] (~ac_laptop@
17:41:40speachythe x1000e has coprocessors (that we probably won't ever know how to use effectively) but the CPU core is running at 1ghz so...
17:42:06__BILGUSoh wow
17:47:52speachyyep. it's _fast_. the jz4760 is already the fastest core we have going
17:48:00speachyand this is even faster..
17:48:07speachys/core/native core/
17:50:08speachythe SoC datasheets and programming manuals are all readily available too, as is the source code for the (reference design) kernel and bootloaders that ingenic put out
17:50:25speachyobviously the per-model variations would need figuring out
17:50:55 Quit pamaury (Ping timeout: 240 seconds)
18:03:04 Join MrZeus [0] (
18:19:29 Quit __BILGUS (Remote host closed the connection)
18:24:12 Quit Moarc (Quit: i znowu NADMUCHAŁ BALONA)
18:28:36 Join Moarc [0] (
18:37:45***Saving seen data "./dancer.seen"
19:35:04 Quit lebellium (Quit: Leaving)
19:45:04 Join fs-bluebot [0] (
19:45:05 Quit bluebrother (Disconnected by services)
19:45:10 Join bluebrother^ [0] (~dom@rockbox/developer/bluebrother)
19:46:25 Quit fs-bluebot_ (Ping timeout: 240 seconds)
20:01:08 Quit MrZeus (Ping timeout: 256 seconds)
20:13:23 Quit livvy (Ping timeout: 240 seconds)
20:34:50 Quit sakax (Quit: Leaving)
20:37:48***Saving seen data "./dancer.seen"
20:43:07 Join johnb2 [0] (
20:46:04 Quit ac_laptop (Ping timeout: 265 seconds)
21:04:08 Join ac_laptop [0] (~ac_laptop@
21:10:22 Join mendel_munkis_ [0] (
21:12:25 Quit mendelmunkis (Ping timeout: 240 seconds)
21:36:18 Quit johnb2 (Quit: Nettalk6 -
21:38:09 Join __Bilgus [0] (41ba23be@
22:03:43 Quit cockroach (Quit: leaving)
22:34:07__Bilgusthis is cool looks like we can enable contrast on the xduoo
22:34:39__BilgusI wonder if I can adjust the precharge period and get some brightness..
22:36:12speachyI did order one of those DS screen replacements, a "why not" type of experiment. the dimness is really my main complaint.
22:37:35__BilgusI think i'm going to go the glass route Ill just use a diamond dremel blade and make a little radius jig
22:37:49***Saving seen data "./dancer.seen"
22:38:08__Bilgusor even free hand on the bench grinder maybe if the glass will handle it
22:40:00__BilgusI also caught another error in the lcd setup LCD_SET_VCOM_DESELECT_LEVEL was set to 0x40 an invalid value
22:40:47__Bilgusdidn't seem to make a difference though setting it lower like (0) did make a difference it was all smeared on the background
22:41:46__BilgusI set it at the default (reset) value and it seemed to be slightly sharper
22:42:24*speachy nods.
22:43:08speachyI haven't attempted to audit any of that stuff since it seemed to basically work. versus stuff that clearly had issues
22:44:29speachyoh, been meaning to ask, any reason why you never landed g#2082 ?
22:44:31fs-bluebotGerrit review #2082 at : Fix voice files play at wrong volume after startup prior to playback by William Wilgus
22:45:04__Bilgusthats just a test maybe I don't think I ever got anywhere
22:46:02__Bilgusyeah I couldn't repro and I forget who it was that I was talking to about it but he never came back
22:47:14__BilgusI could abandon it if you like but i'm sure its probably still an issue to (somebody)
22:48:24__Bilgusjust got a rolo crash first in probably 30
23:04:30speachythe issue I was having with the first menu voice getting cut off turned out to be due to the busted udelay(), not voiced volume
23:18:05 Quit mendel_munkis_ (Ping timeout: 240 seconds)
23:19:48 Quit TheSeven (Ping timeout: 244 seconds)
23:20:27 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
23:28:46speachythat dselect_level value was probably in the OF to try and boost the brightness in compensation for the smoked screen window
23:30:31__Bilgusmaybe assuming its valid but looking for it it appears it was in the sources everyone has copied for that chip
23:32:42speachyincidently, the as3525/lcd-ssd1303.c sets it to 0x25, which appears to also be illegal
23:33:18speachythe lower nibble should be 0, according to the datasheet..
23:34:04__Bilgusnice lol and I had already poured over that one at one time or another
23:35:15speachytbh I don't think we should drop it
23:35:22speachyif it makes the screen dimmer
23:36:12__BilgusI don't think it does on the xduoo
23:36:28__Bilgusif anything it made it crisper
23:37:01__Bilguschanging the value to the spec
23:38:51__Bilgusreading the docs from the manufacturer the key is to precharge long enough to get the cap charged high enough before it draws current longer precharge = higher voltage and less time to do the current drive
23:39:46__Bilgusso in theory we could get max brightness with a period of 1/1 (two phases) and a really high drive voltage
23:40:03__BilgusI'll try upping the value and see if it gets brighter
23:40:15__Bilgus(and lowering the time)
23:41:18speachytoo bad the actual screen specs/requirements are completely unknown
23:43:26__Bilguswhich is typical
23:43:57__Bilguslooks like brightness messes with the contrast to set brightness how 1-16 relates to 0-255 no clue
23:45:15__Bilgusoh that did make it brighter
23:45:30__Bilgushmm might be on to something here
23:46:10__Bilgusso lets set this low when you hit 100% contrast up the voltage and lower the time
23:46:29__Bilgusotherwise we will make it behave at saner levels
23:47:03speachythe brightness value is *16, -1 to gield the 0-255 value
23:48:13__Bilgusare you looking at teh src file atm or do you just have info like that on the top of your head?
23:49:17speachythe latter. when I found the SSD1306 datasheet I was curious how it was mapped so I looked it up in the source code.
23:49:28speachyso it's still in the LRU cache
23:49:56__Bilgusi'm thinking maybe I can remap the contrast to do my scheme and leave brightness to dfo the actual contrast or visav
23:50:03 Join mendelmunkis [0] (
23:50:39__Bilgusthis part is so fun like crack or something
23:53:58speachy(meanwhile it took me over half an hour to figure out how/where the stereo width 0-250% was actually defined)
23:58:26 Quit ac_laptop (Ping timeout: 265 seconds)

Previous day | Next day