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 2017-01-02

00:21:16 Quit skapazzo (Quit: Lost terminal)
00:28:53 Join Kofaz [0] (4582f871@gateway/web/freenode/ip.
00:30:52 Quit Kofaz (Client Quit)
00:32:57 Join anonus [0] (
00:42:15 Quit lebellium (Quit: ChatZilla 0.9.93 [Firefox 50.1.0/20161208153507])
00:45:06 Quit edhelas (Ping timeout: 248 seconds)
00:46:47 Quit Senji (Ping timeout: 264 seconds)
00:49:50 Join Senji [0] (~Senji@
01:10:18 Quit tracktheripper (Ping timeout: 260 seconds)
01:16:35 Quit paulk-collins (Quit: Leaving)
01:25:47 Quit Senji (Ping timeout: 264 seconds)
01:36:04 Quit ZincAlloy (Quit: Leaving.)
01:39:56 Join Senji [0] (~Senji@
01:44:50 Quit Senji (Ping timeout: 248 seconds)
01:48:36 Quit ender` (Quit: Washing your car to make it rain doesn't work.)
01:57:21***Saving seen data "./dancer.seen"
02:01:23 Quit Mihail (Quit: - A hand crafted IRC client)
02:22:44 Join rela_ [0] (
02:26:14 Quit rela (Ping timeout: 260 seconds)
03:06:31 Join nlogex [0] (
03:21:57 Join Senji [0] (~Senji@
03:24:15 Quit diox (Ping timeout: 246 seconds)
03:25:04 Join diox [0] (
03:27:24 Quit Senji (Ping timeout: 246 seconds)
03:31:00dongshm so 4/8gib flash is very cheap. ill get a couple of each to test
03:31:02 Join Senji [0] (~Senji@
03:34:46[Saint]what are you attempting to do?
03:36:43 Quit Senji (Ping timeout: 256 seconds)
03:38:38dongs[Saint]: same thing as before that i've been talking about in the logs.
03:39:04*[Saint] hasn't been online in ~3 days
03:39:27dongsya missed all the fun
03:41:04[Saint]care to gimme a rundown?
03:42:22dongsmany dead flash sansa clip+'s. want to swap sandisk flash part for generic micron/etc nand. checked pinout. looks legit. soldered in 512meg part (all i had laying around). got to boot over jtag w/usb bootloader. shows up as 1.3tb drive. now here.
03:42:59dongsbetween then some discussion of where nand>sd translation occurs,what changes I might need to do in USB bootloader to make it open entier flash area to flash firmware on a blank chip (if it works). etc
03:46:33[Saint]in some regards there's not really generic flash, per se.
03:49:00 Join Senji [0] (~Senji@
03:49:15 Quit Senji (Read error: Connection reset by peer)
03:49:43 Join Senji [0] (~Senji@
03:54:05dongs[Saint]: thats what everyone keeps saying but hardware evidence points otherwise
03:54:58 Quit Senji (Ping timeout: 248 seconds)
03:57:24***Saving seen data "./dancer.seen"
04:01:07[Saint]and that doesn't even begin to deal with the fact that the bootloader and OF probably expect to be dealing with a small subset of very specific credentials from expected silicon.
04:01:32dongsyep, i will be looking into that
04:01:39[Saint]you'll assuredly never boot the OF on this thing.
04:02:01dongsit might never boot beyond using jtag, if ROM bootloader sets up nand and only supports sandisk flash IDs
04:03:14[Saint]it'll also be expecting a very specific capacity.
04:03:44[Saint]the 1.3TB (presumably entirely inaccessible at high or low levels) reported smacks hard of an overflow.
04:04:13 Join CrashBash-Kun [0] (~CrashBash@unaffiliated/crashbash-kun)
04:05:42[Saint]But, yeah - the only thing that is generic, per se, about flash any one of the several standards for ordering the physical composition of it.
04:06:46dongswhich is why im gonna try 4 and 8gb parts
04:06:59dongs512 is really old one, its just the only thing i had laying around.
04:07:09dongs8gb is only like $3.6 so i dont mind wasting that to try
04:07:27[Saint]then you'll just get an overflow with a different offset.
04:07:39dongsthen I'll step through bootloader and see why it does that
04:08:18 Quit CrashBash-Kun (Read error: Connection reset by peer)
04:09:29[Saint]errrr...can we even step the OF bootloader?
04:10:26[Saint]like pretty much every other target the rb bootloader is just a tiny subset inserted that says "hit here; unless <foo>" and allmost all of the hardware setup is handled by the OEM bootloader.
04:10:41dongslast part: no
04:10:52dongsi connect to cpu under reset, load rockbox bootloader into ram, and jump to it
04:10:58dongsROM boot never even runs
04:11:28dongsand screen works and it connects in bootloader usb mode.
04:11:40dongsso i can certainly step through that part once I get build going and can do source-level debugging/stepping
04:12:30[Saint]by the time you get to handoff to soemthing you can step through, hardware setup has already happened.
04:12:45dongs... do you know what "connecting under reset" means?
04:12:56[Saint]I do.
04:12:57dongsnothing runs except the code I put in ram
04:13:45dongsand I'm loading bootloader-clipplus.bin which is all original code, no OF bits
04:14:07dongsso i can certainly step through it all
04:14:13[Saint]that's mainly OF bits, in fact.
04:14:25[Saint]about 98% OF bits in fact
04:14:56[Saint]it's literally just the original firmware with a tiny little edit in it giving it the jump point for Rockbox.
04:15:03dongsno, its not
04:17:49dongsthis is the rockbox bootloader that gets compressed and packed into OF boot
04:18:06dongsi am NOT booting the code thats in clppa.bin that combines OF + rockbox
04:19:19[Saint]so, actually mean bootloader-clipplus.sansa?
04:19:39dongs(renamed to .bin cuz i cut off 8byte or whatever header that was in .sansa file
04:25:45dongsso yea no hardware is setup before that runs.
05:11:57 Join Senji [0] (~Senji@
05:17:25 Join treaki__ [0] (
05:18:01 Quit Senji (Ping timeout: 268 seconds)
05:21:34 Quit treaki_ (Ping timeout: 265 seconds)
05:34:06 Quit aptzero (Quit: Leaving.)
05:37:20 Join JanC_ [0] (~janc@lugwv/member/JanC)
05:38:36 Nick JanC is now known as Guest56827 (~janc@lugwv/member/JanC)
05:38:36 Nick JanC_ is now known as JanC (~janc@lugwv/member/JanC)
05:39:09 Quit Guest56827 (Ping timeout: 256 seconds)
05:57:28***Saving seen data "./dancer.seen"
06:21:14 Join Senji [0] (~Senji@
06:27:05 Quit Senji (Ping timeout: 268 seconds)
06:46:34 Quit advcomp2019 (Ping timeout: 246 seconds)
06:51:33 Quit TheSeven (Disconnected by services)
06:51:42 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
07:15:14 Quit idonob_ (Ping timeout: 252 seconds)
07:16:18 Join idonob_ [0] (
07:27:27 Join edhelas [0] (~edhelas@
07:31:52 Quit edhelas (Client Quit)
07:45:30 Join edhelas [0] (~edhelas@
07:45:40 Quit edhelas (Client Quit)
07:57:31***Saving seen data "./dancer.seen"
08:01:03 Quit amiconn (Quit: - Chat comfortably. Anywhere.)
08:01:04 Quit pixelma_ (Quit: .)
08:01:14 Join pixelma [0] (~pixelma@rockbox/staff/pixelma)
08:01:16 Join amiconn [0] (~amiconn@rockbox/developer/amiconn)
08:08:38 Join Senji [0] (~Senji@
08:14:24 Quit Senji (Ping timeout: 264 seconds)
08:44:22 Join advcomp2019 [0] (
08:44:22 Quit advcomp2019 (Changing host)
08:44:22 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
09:14:28 Quit idonob_ (Ping timeout: 256 seconds)
09:17:58 Quit nlogex (Ping timeout: 258 seconds)
09:57:35***Saving seen data "./dancer.seen"
10:04:21 Join n3m9 [0] (
10:19:31 Quit krnlyng (Ping timeout: 248 seconds)
10:20:04 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
10:26:14 Join Senji [0] (~Senji@
10:26:21 Join Strife1989 [0] (
10:29:19 Quit CommunistWitchDr (Quit: No Ping reply in 180 seconds.)
10:29:26 Quit Strife89 (Ping timeout: 268 seconds)
10:30:23 Join Strife89 [0] (
10:31:01 Join SovietShaman_ [0] (
10:31:12 Quit Strife1989 (Ping timeout: 264 seconds)
10:31:36 Join krnlyng [0] (
10:31:47 Quit Senji (Ping timeout: 248 seconds)
10:35:30 Quit Strife89 (Ping timeout: 256 seconds)
10:49:54 Join ender` [0] (
10:59:24 Quit pamaury (Ping timeout: 264 seconds)
11:15:32 Join Senji [0] (~Senji@
11:21:01 Quit Senji (Ping timeout: 258 seconds)
11:32:58 Join lebellium [0] (
11:46:55 Join pamaury [0] (
11:46:55 Quit pamaury (Changing host)
11:46:55 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:51:28 Join pamaury_ [0] (~pamaury@rockbox/developer/pamaury)
11:55:43 Quit pamaury_ (Ping timeout: 252 seconds)
11:57:39***Saving seen data "./dancer.seen"
12:00:29 Join Senji [0] (~Senji@
12:06:00 Quit Senji (Ping timeout: 264 seconds)
12:08:49 Join robertd1 [0] (~root@
12:15:57 Join skapazzo [0] (~skapazzo@
12:24:46 Join paulk-collins [0] (
12:52:00 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:49fc:d35:f21c:c820)
12:54:45 Join cc___ [0] (~ac@2001:910:113f:1:6a05:caff:fe1c:1627)
13:01:33 Join Senji [0] (~Senji@
13:16:23lebelliumpamaury: I found the Sony official website where you can download all service manuals from free. The only issue is that you have to know the service manual reference number to get the proper URL... I'm mass testing many URLs and by chance I found some NWZ service manuals. The E390 one is ridiculous....
13:17:06pamaurylebellium: lol, indeed
13:20:09pamaurylebellium: I plan to resume work on NWZ port this week. I think I'll add an option to boot the service menu from the bootloader
13:20:54lebelliumyou could do that? I'm looking for the E580 and A10 service manual especially to get the key combo to enter the service menu
13:25:46lebelliumoh I got a nice one! NW-WM1
13:26:00lebelliumthe new high-end player
13:26:36pamauryas far as I know, the service menu is just another program. when you hit the magic sequence, it sets a flags and reboots, on reboot it launches a different program
13:27:30lebelliumlook at NW-WM1, it's not a key combo anymore
13:27:50lebelliumyou have to connect the device to PC and run a separated programm
13:31:38pamauryit's very complete! Well I don't plan to buy a WM1 anytime :-p
13:32:41lebelliumbut I know some users already asked me if you planned to port Rockbox to it anyway :P
13:34:15pamauryI'll try to make the port as generic as possible, we'll see how it can go
13:34:20pamaury*how far
13:35:55 Join johnb2 [0] (
13:40:58 Join pamaury_ [0] (~pamaury@rockbox/developer/pamaury)
13:45:24lebelliumpamaury argh A10 so disappointing
13:45:28lebelliumspent so many hours for that :(
13:51:20dongslebellium: whats the pattern for search stuff
13:51:31dongsI could use a service manual for some other junk they make
13:54:40lebelliumlast 2 digits is the service manual revision number. Most of time it's 01 (v1.0), sometimes 02 (v1.1). There is no real pattern for the 7 digits before. I made an excel file with all URLs from 989603401_SM.pdf to 989999901_SM.pdf and test all of them one by one. I selected prefix 989 but NWZ-A840 from 2010 is still in prefix 988
13:55:54lebelliumif you can find a preview of your service manual on a bastard website where they want to charge you $10 or $20 for it, you can read the service manual number on the 1st page of the manual
13:56:11lebelliumbut for most service manuals there is no preview available
13:56:15lebelliumso you have to guess
13:56:16dongsah yeah no this is a japan-only oddball device so I doubt it would be on the site
13:57:30dongshmm i wonder if thats it
13:57:41***Saving seen data "./dancer.seen"
13:58:03dongshm nope 404.
13:59:37lebelliumI just checked on the box of my A15
13:59:52lebelliumthe reference written has unfortunately nothing to do with the associated service manual
14:06:40 Quit sLite (Remote host closed the connection)
14:19:56Bilgusany reason it isn't indexed by google?
14:20:01Bilgus filetype:pdf
14:20:25 Quit Senji (Ping timeout: 258 seconds)
14:21:49pamaury_Bilgus: haven't checked, but it's only indexed if there are links to it and afaik the directory is not readable
14:21:59Bilgus^not to be an ass but figured it was easier :)
14:22:06lebelliumpamaury_: E470 :D
14:22:36pamaury_Bilgus: great :)
14:23:29 Join Senji [0] (~Senji@
14:25:30Bilgusit seems like lebellium has ones that weren't indexed
14:26:12lebelliumI can have anyone, I'm just checking hundreds URLs by hand. That may be stupid but I don't find a better solution
14:26:25pamaury_lebellium: script ?
14:26:45lebelliumI can't write scripts. Could a script read the model number from the 1st page?
14:27:33pamaury_coudln't you write script to find all *valid* URLs ? I guess most urls you find are 404
14:28:23lebelliumI used that but 150 URLs in one go was boring
14:28:29lebelliumbut indeed I get many 404
14:29:03 Quit Senji (Quit: KVIrc 4.9.2 Aria
14:30:21lebelliumpamaury_: I'll put all NWZ links I found here . Sounds good?
14:31:19pamaury_feel free to fix the NW/NWZ clarification
14:31:31pamaury_I really have to rework this page
14:33:41 Join Link8 [0] (~me@
14:36:05 Join dfkt [0] (~dfkt@unaffiliated/dfkt)
14:37:31anonusis lyre project abandoned?
14:38:20gevaertsWas it ever real?
14:38:41anonusi dunno, just asking
14:39:04anonusi'm just thinkging about having fun with i.mx233
14:39:31anonusto make simple board with, mem and audio
14:39:46anonusi wonder how likely to make it run rockbox
14:40:18gevaertsOh, if yoy know about hardware you can certainly do that
14:40:23 Quit diox (Quit: patchigt)
14:40:33pamaury_anonus: i.mx233 is well supported by rockbox, if you can do the hardware side that should be relatively easy
14:41:02gevaertsJust don't have delusions about doing that as a hobby project easily resulting in a marketable commercial device :)
14:41:24anonusyeah, i just think can i do the trick to make it closer to Fuze+ to be able to run rockbox with minimal modifications
14:42:25anonuswell, certainly it will be a hobby project, i don't want even touch BGA chips with 5 meter stick and with TSSOP DRAM it would eat battery even faster than modern smartphones :)
14:42:55anonusi believe there is no mobile dram in diy-solderable case
14:43:18pamaury_there are some imx233 boards at olimex. It's quite easy to create new targets in rockbox, you don't have to worry about making it close to the fuze+. The only real limitations are 1) don't use raw NAND 2) make sure you have at least 4MB of RAM
14:43:22anonusbut i'll take your point ;)
14:43:44anonuspamaury_: already ordered one to check my ideas
14:43:51pamaury_also the imx233 supports others types of memories I tink, probably ddr2
14:44:13 Join Senji [0] (~Senji@
14:44:14anonushow much ram do you need to run rockbock btw?
14:44:20anonusi mean what the main factor?
14:44:57anonusif i will use simplest monochrome 80x25 LCD will it save me some RAM?
14:45:38gevaertsYes, but I'd argue that 80x25 is a bit too low
14:45:52pamaury_some targets can do with 2MB but for the imx233, I think you should aim for at least 4MB. A bigger screen will use more memory but I don't think a 240x320 screen would be a problem
14:45:55anonuswill 4 MB be enough?
14:46:11pamaury_also ram is cheap, if I were you I would put 8 or 16 MB
14:46:12anonusi mean Fuze+ and all Creatives with i.MX233 have at least 32mb
14:46:33anonusi'm not concerned about cost but about power consumption
14:46:35gevaertsI'd recommend going for a screen resolution that's already supported. That will save you a lot of plugin work, and open up more themes
14:47:07pamaury_the system basically needs 2MB to just run. Everything after that is for audio buffer and codecs. If you have more, you can cache more but 32MB is overkill one could say
14:47:31anonusso 8 will be more than enough?
14:47:48gevaertsOn flash audio buffer isn't as critical as on spinning disks
14:48:02gevaerts8 should be plenty unless you go for big colour screens
14:48:13anonusi am definitely not :)
14:49:12gevaertsIf your primary goal is fancy plugins, you might also want more. If your primary goal is audio playback on a flash device, 4 is probably enough but going for 8 gives you some headroom
14:49:34gevaertsFor low-res 2.5 is going to be enough, even. That's what the clipv1 has
14:49:48gevaerts(2MB + a bit of on-SoC RAM)
14:49:50pamaury_also just so you know, if you go for a button interface, I suggest you have at least 6/7 buttons (direction pad, back, enter). We support touchscreen but it's far from optimal
14:50:17anonuswell, my primary goal to have some fun and experience with designing hardware a bit more complicated than basic stuff with STM32 controllers, never tried to route DRAM before
14:50:18pamaury_gevaerts: imx233 has 32KiB of iram so it's not really helping in this case ;)
14:50:46anonusbut as side product i could get myself a working portable player with rockbox, my sansa clip+ is slowly dying
14:51:30anonuss/clip+/clip zip/
14:52:10anonuslyre project tried to use 5510 nokia lcd, is it supported?
14:53:00pamaury_each lcd is different, adding a new one is not hard if you have the datasheet but there is 0% you can buy an already supported one
14:54:00anonusah, ok
14:54:02anonusi see
14:54:06anonusno problems then
14:54:34gevaertspamaury_: yes, true :). IIRC those few hundreds of kilobytes extra were actually fairly critical in getting rockbox to run well though
14:54:38pamaury_if you can, I think it will simpler to have an lcd with a 8080 or 8086 interface, it's much faster than an i2c or spi
14:54:55pamaury_(but more wires obviously)
14:55:17 Quit johnb2 (Ping timeout: 265 seconds)
14:55:21anonuspamaury_: well, there is one more reason: with 8080 i can have two sd cards
14:55:38gevaertsIf your goal is to get design experience, more wires is actually an advantage!
14:55:47anonusas far as i understand i.mx233 has only two ssps
14:56:35anonusyeah, i just don't want to mess with ribbon cables, and 5510 has nice zebra strip interface
14:58:33pamaury_anonus: the imx233 only has two ssp, although you could get away with two cards on the same ssp, for example by using a multipler a pin to switch between the two
14:59:00pamaury_there is no code for it in rockbox but that can be implemented without too much difficulty I think
14:59:12anonusdoes it run microsds as SPI? i'm not sure, should check datasheet
14:59:54pamaury_no rockbox only supports SD/MMC protocol, we never run microsd as spi
15:00:36anonusso AFAIK you could not mulpitlex microsds if they are not working as SPI
15:00:57anonussd bus has no CS pin
15:01:12anonusor am i missing something?
15:01:14pamaury_that's why you need a multiplex and a gpio to act as CS for an external multiplexer
15:01:37anonusah, ok
15:01:40anonusi see
15:02:08pamaury_if you prefer, you wire all lines to both cards, with a bidir buffer for each of them and you wire a gpio to the enable of each, one inverted, something like that
15:03:22*pamaury_ doesn't claim this is electronically the right solution
15:03:33anonusyeah, that could be done, but i think it does not worth it
15:03:57pamaury_I *think* in theory you can have several microsd on a single bus, because you can select the active card using commands, but I have never done that
15:04:08lebelliumpamaury_: E580!!
15:05:16pamaury_I think the imx233 also has an ATA interface
15:06:02pamaury_at least the older stmp37xx series did and I think Freescale did not advertise the imx233 as been ATA capable but in reality it can do it
15:07:43anonusmmm, i see nothing about ATA in datasheet
15:09:15chrisjjpamaury: did you get my ZEN battery_bench.txt?
15:09:32pamaury_it's not in the imx233 datasheet, but the GMPI block could do nand + ata in the previous generation
15:09:46pamaury_chrisjj: I think so, ?
15:09:50pamaury_I didn't look at it yet
15:10:12chrisjjYup, that's it.
15:10:13pamaury_anonus: google "stmp37xx datasheet"
15:11:12pamaury_although only the BGA version could, and it is possible that the imx233 chip really drop support for it, I don't know
15:12:25pamaury_also double-check the datasheet for shared IO lines, not all combinations are possible
15:12:53pamaury_like ata/nand is shared with ssp2
15:14:01 Join johnb2 [0] (
15:19:24lebelliumpamaury_: when you have some time, can you try the key combo on your E580. I can't enter the test mode the way described in the service manual
15:22:39 Quit johnb2 (Quit: Nettalk6 -
15:23:37pamaury_sure, remind me if I forget
15:25:40lebelliumpamaury_: nevermind, I found the mistake in the usermanual
15:25:45lebelliumservice manual
15:25:54lebelliumthey wrote "up" instead of "down"
15:26:03lebelliumand didn't even notice it a updated it
15:33:44 Quit StaticAmbience (Remote host closed the connection)
15:36:32 Join StaticAmbience [0] (
15:57:42***Saving seen data "./dancer.seen"
15:58:00Bilgus__builtin: did you ever figure out a way to save the current screen buffer and restore it after you switch viewports?
16:00:23 Join aptzero [0] (
16:02:31 Join edhelas [0] (~edhelas@
16:11:16 Join amayer [0] (
16:16:00 Quit fs-bluebot_ (Ping timeout: 268 seconds)
16:16:26 Quit Link8 (Remote host closed the connection)
16:16:48 Quit bluebrother (Ping timeout: 264 seconds)
16:18:29 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
16:26:03 Quit rela_ (Quit: Leaving)
16:30:12 Join fs-bluebot [0] (
16:33:23 Join Guest93 [0] (~textual@
16:35:04 Quit edhelas (Read error: Connection reset by peer)
16:35:21 Quit Guest93 (Client Quit)
16:35:54 Join edhelas [0] (~edhelas@
16:49:16dongscool chats above
16:57:49 Join rela [0] (~x@pdpc/supporter/active/rela)
17:41:00 Quit pamaury (Remote host closed the connection)
17:48:42 Join diox [0] (
17:57:43***Saving seen data "./dancer.seen"
18:02:36 Quit diox (Quit: en till reboot..!)
18:08:35 Join diox [0] (
18:11:40 Quit pamaury_ (Ping timeout: 256 seconds)
18:15:20 Join girafe [0] (
18:29:01 Join nlogex [0] (
18:42:20 Quit lebellium (Quit: ChatZilla 0.9.93 [Firefox 50.1.0/20161208153507])
18:42:41 Join girafe2 [0] (
18:44:23 Quit girafe (Ping timeout: 265 seconds)
18:45:35 Quit Senji (Read error: Connection reset by peer)
18:46:27 Join Senji [0] (~Senji@
18:50:00 Join Senji_ [0] (~Senji@
18:50:16 Quit Senji_ (Read error: Connection reset by peer)
18:50:44 Join Senji_ [0] (~Senji@
18:51:51 Quit Senji (Ping timeout: 260 seconds)
18:52:13 Quit marex-cloud (Ping timeout: 258 seconds)
18:57:48 Quit diox (Read error: Connection reset by peer)
18:58:18 Join diox [0] (
18:58:49 Quit StaticAmbience (Read error: Connection reset by peer)
19:00:45 Part robertd1
19:01:47 Join StaticAmbience [0] (
19:13:39 Quit prg318 (Quit: WeeChat 1.6)
19:19:45 Join prg318 [0] (~prg318@deadcodersociety/prg318)
19:28:20 Quit Senji_ (Read error: Connection reset by peer)
19:30:14 Join Senji [0] (~Senji@
19:30:29 Quit Senji (Read error: Connection reset by peer)
19:31:39 Join Senji [0] (~Senji@
19:32:34 Quit Senji (Read error: Connection reset by peer)
19:32:59 Join Senji [0] (~Senji@
19:33:25 Quit Senji (Read error: Connection reset by peer)
19:35:48 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
19:56:03 Quit edhelas (Ping timeout: 252 seconds)
19:57:47***Saving seen data "./dancer.seen"
20:00:32 Join Senji [0] (~Senji@
20:49:42Bilguschrisjj: want to test the play pause fade plugin?
21:04:09 Join marex-cloud [0] (sid137234@gateway/web/
21:15:30 Quit aptzero (Quit: Leaving.)
21:23:13 Join idonob_ [0] (
21:31:29[Saint]Bilgus: errrr...the what?
21:31:37[Saint]Don't all capable targets already do this?
21:32:08[Saint]Settings - Playback Settings - Fade On Stop/Pause
21:32:14Bilgusthsi plugin allows you to set a variable delay, min and max volume
21:32:21[Saint](and it's in-core, too)
21:32:54[Saint]Why do it in a plugin?
21:33:02[Saint]I would also be very wary about trying to get anything like that into the core.
21:33:13Bilgusit will eventually also be a fade to sleep plugin
21:33:15[Saint]We just plain don't fuck with playback from core with plugins.
21:33:27[Saint]Pretty sure we've been through this.
21:33:39Bilguswell it actually just fades the volume out / in
21:33:48[Saint]I'm aware.
21:34:18[Saint]Do It Right(TM) and add this behavior to the appropriate core setting bundle.
21:34:42Bilgusand I already tried a lua script which works but as I'm sure you are aware doesn't work on all targets
21:34:54[Saint]Sorry, but I'll fight playback boundaries crossing between core and plugins with a passion.
21:35:05 Join aptzero [0] (
21:35:17BilgusI can't see adding any of this to the core
21:35:38[Saint]Why? We already have variable fade delays elsewhere.
21:36:41BilgusI admit it would probably be easier to add it in to the core but why add the extra overhead?
21:36:42[Saint]It's basically just crossfade but with a different trigger and no secondary track to bleed.
21:37:12[Saint]Bilgus: the why would be "plugins have no business governing playback they didn't start"
21:38:11Bilguswell technically the plugin does start the playback
21:39:00Bilgusno different than the alarm plugin
21:39:02[Saint]the bulk of this is already sitting around in the core code just waiting for someone to join the dots.
21:39:17[Saint]and, yeah - that's my one reservation of plugin/core crossover.
21:39:26[Saint]I dislike it but I accept it.
21:40:11[Saint]there's nice fade control and at least two different models for time pickers sitting in core already.
21:40:55BilgusIf you think it'd be better served in core I'll check it out, what about fade to sleep?
21:41:22[Saint]It just seems weird to me to disconnect the existing fade on stop/pause behavior in core.
21:41:42[Saint]Literally all that requires to meet parity is a time picker for fade duration.
21:42:05Bilgusthe two basically go together code wise
21:42:15fs-bluebotBuild Server message: New build round started. Revision 782d9c0, 255 builds, 14 clients.
21:42:42[Saint]No part of me would expect to find control to govern this behavior in two entirely different sections of the codebase.
21:42:42[Saint]UX-wise that's plain bad JuJu
21:43:12Bilgusnow the only question is how to let the user know that the fade is going on
21:43:35[Saint]Why do you need to let them know? We never do elsewhere.
21:43:41[Saint]They set it explicitly. They know.
21:43:59Bilgusfor instance if I pick a fade over 10 minutes in the sleep one and the user changes volume
21:44:42[Saint]I think any values over 60s are pretty nuts personally.
21:44:57Bilgusdo we. start fading from the new volume; disable the fade; disable vol change(uh no)
21:45:07Bilgusnot in a sleep to fade plugin
21:45:33Bilgusthey might want to faded for 30 minutes or something crazy
21:45:59[Saint]That's going to get messy real quick.
21:46:17BilgusANND plugin :)
21:46:19[Saint]Like, how to deal with if they already have a dramatic crossfade set.
21:46:27[Saint]Ignore it?
21:46:29[Saint]Stack it?
21:46:55[Saint]I guess now I'm seeing your point, but it would seem VERY weird to me to have this behavior in plugin and not in core.
21:47:43[Saint]This is a niche case that should probably be tied into whichever plugin it is that I forget the name of that governs sleep on lack of user input.
21:48:11[Saint]I thought you just wanted ~10s of fade on stop/pause like almost every other playback setting has.
21:48:18[Saint]Not spanning multiple minutes.
21:49:37[Saint]I guess the "play pause fade" wording confused me.
21:49:53Bilguswell thats why I started with the fade on play/pause as it is essentially the same thing minus the sleep part
21:49:55[Saint]As this use case seems largely disconnected to playback control in and of itself.
21:50:06 Quit aptzero (Quit: Leaving.)
21:50:20[Saint]IIUC you don't want this to actually fire on user interaction.
21:50:29[Saint]You want it to fire on lack of user interaction?
21:51:05[Saint]pressing a button to fade to a stop ~N minutes in the future isn't something that strikes me as a common use case.
21:51:20Bilguswell yes and no it doesn't have to be linked to user interaction atm though it does grab pause and play state
21:51:46 Join aptzero [0] (
21:51:50Bilgusand the current fade on pause uses the pcm buffer to fade
21:52:05[Saint]Yeah, I feel like this should be tied in with the plugin I can't recall the name of but you likely are familiar with.
21:52:11[Saint]The sleep timer dealy.
21:52:23BilgusI have no idea on the plugin I was looking
21:53:23fs-bluebotBuild Server message: Build round completed after 667 seconds.
21:53:24fs-bluebotBuild Server message: Revision 782d9c0 result: All green
21:53:25fs-bluebotBuild Server message: New build round started. Revision e259836, 255 builds, 12 clients.
21:53:40[Saint]we already have a plugin that sets an interactive timer for sleep.
21:53:56[Saint]so if the device hasn't been interacted with in N minutes, it'll stop playback and then idle timeout.
21:54:09[Saint]seems to me a protracted fade would go well with that.
21:54:20Bilgussounds like a great place to inset this
21:54:26*[Saint] nods
21:55:21Bilgusok so I think i'll look to adding play / pause fade to core and then fade to sleep to this nameless plugin
21:55:28[Saint]I thought you meant a more general stop/pause fade, like we alreadyhave, but just adding configurability to it.
21:56:07BilgusI did but only because they share the same code
21:56:13[Saint]This is indeed more generalized. Lack of having a time picker for the fade period was really an oversite there I think.
21:56:24 Join edhelas [0] (~edhelas@
21:56:36Bilguswell again that is because it uses the pcm buffer
21:57:12[Saint]well, as does crossfade.
21:57:13[Saint]that's configurable.
21:57:49***Saving seen data "./dancer.seen"
21:57:53[Saint]Bah - you've reminded me now how broken the time picker values are across the board.
21:57:59BilgusI'll look at it
21:58:24BilgusI've made some pretty nice interfaces the last few days
21:58:54[Saint]we have everything from 0~10s or 0.00s~60.00s in the time pickers. Various other unit pickers have mismatched values too.
21:59:05[Saint]Noticing this again makes me unreasonably angsty.
22:00:29BilgusI don't much like that many values in a menu
22:01:00Bilgusmy solution to that was to have a menu with 1, 5 , 15 seconds and a + and - button
22:01:17TorCanonus: For your dying Clip-x, depending on how it is dying, you might want to install the modified bootloader here:,51304.msg237134.html#msg237134
22:01:30Bilgushit more and it goes to 2, 10 , 30
22:01:48TorCIt will allow the thing to keep working for you even if the internal nand flash dies.
22:02:09TorCDon't count on installing it after flash dies, though.
22:02:15[Saint]Bilgus: that gets fucky quickly.
22:02:23Bilgusif given an absolute range we could use a scrollbar
22:02:43[Saint]Bilgus: Like, if you wanted to select '8' as the unit, the upper value would be several hundred.
22:02:43[Saint](for instance)
22:03:09[Saint]unless you do some weird non-linear scaling.
22:03:12fs-bluebotBuild Server message: Build round completed after 588 seconds.
22:03:12pamaurygevaerts: do you know the build server a bit ? I've created a new build client but the server doesn't like it :(
22:03:13fs-bluebotBuild Server message: Revision e259836 result: All green
22:03:57BilgusI suppose you could do a sliding scale
22:04:05[Saint]Bilgus: and, on some devices - no, we couldn't. Menus need to work on charcell too.
22:04:26[Saint]No nice pretty selector bars there.
22:04:37Bilgusyeah, true
22:04:55anonusTorC: 3.5 mm jack is dying and it has really strange jack i could find nowhere
22:05:27[Saint]anonus: define dying? there's not a lot there that can go wrong. Likely just needs to be reflowed.
22:05:43[Saint]Their connection with the board is far from rigid.
22:05:57gevaertspamaury: what did you do wrong?
22:06:00anonusjack just started falling out
22:06:15[Saint]Ahhh, right.
22:06:19anonusi did reflow once, but it seems that it is mechanical problem
22:06:21pamaurygevaerts: here is a typical part of my log:
22:06:40pamauryFetching origin
22:06:40pamaury2017-01-02 22:01:00 (../tools/configure −−target=ipod3g −−type=b && make) >> wks-50-63-pamaury-ipod3gboot.log 2>&1
22:06:40pamaury2017-01-02 22:01:04 Uploading wks-50-63-pamaury-ipod3gboot.log...
22:06:40DBUGEnqueued KICK pamaury
22:06:40pamaury2017-01-02 22:01:04 curl -s -F upfile=@wks-50-63-pamaury-ipod3gboot.log
22:06:41pamaury2017-01-02 22:01:04 child ipod3gboot (24138) is uploading
22:06:43pamaury2017-01-02 22:01:04 No upload
22:06:45pamaury2017-01-02 22:01:04 child: ipod3gboot (24138) done
22:06:47pamaury2017-01-02 22:01:04 Completed build ipod3gboot
22:07:05pamauryit does that for a while and then I am disabled temporarily
22:07:11[Saint]anonus: realistically you can use any TRS connector that fits as long as you keep track of the pinout.
22:07:35pamauryI checked and the builds are successful, but upload fails for some reason, apparently
22:08:07anonusyeah, i could just solder some wires and wire-mounted socket outside the case
22:08:22gevaertsYes, sounds like it doesn't like the log you're uploading for some reason
22:08:35gevaertsYou could try reading the server code or asking zagor I guess
22:08:40CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
22:08:40*gevaerts doesn't really know
22:08:53Bilgusanonus: while you are at it cover it in hot melt snot to really finish the look :p
22:09:16anonusworth trying!
22:09:29anonushot snort is my friend
22:09:57TorCanonus: Another option would be to get a spare or two and hope that when they go, something else dies and you can steal the jack for this one.
22:10:02[Saint]I saw a guy with a clip zip with a 6mm TRS hanging out the side of it the other day actually.
22:10:35[Saint]it looked hilarious, just hanging there off his headphone cable.
22:11:05[Saint]not actually clipped to anything, presumably because SanDisk's clips on the actual Clips are all shit.
22:11:29pamaurygevaerts: is it possible that the server gives the same builds to several clients and one beat me every time ? I don't know how the server works
22:11:56gevaertsThat's possible, but only near the end, and it won't disable you for that
22:12:12[Saint]Yeah, that's possible.
22:16:48pamaurygevaerts: ok I think this is an SSL error, I tried to manually upload something and it fails
22:16:56pamauryI recall someone mentioning this problem
22:17:39[Saint]Bilgus: you've got me thinking about a single generic picker but it's such a pain in the tits to do across the board.
22:17:39[Saint]Bilgus: Having each function just pass out the mix, max, and unit and having a generic picker function would be neat...
22:18:11gevaertsAh well, if it's curl that malfunctions you want bagder instead :)
22:18:55pamaurycurl is right, the domain doesn't have a certificate
22:19:29pamaurycurl: (51) SSL: no alternative certificate subject name matches target host name ''
22:19:44Bilgus[Saint]: question is how to lay it out
22:19:59[Saint]Bilgus: yeah, right.
22:19:59gevaertsIt's not using https here...
22:20:59gevaertsAh I see....
22:21:05gevaertsIt's all __builtin's fault!
22:21:16chrisjjBilgus: Sure. Sounds right up my street.
22:22:13Bilguschrisjj: saint has put me on a more enlightened path ill let you know when that part is finished :p
22:22:37pamaury__builtin: ping
22:24:21gevaerts__builtin changed the client to use https, but he didn't change the revision so no client auto-updates (I can't actually remember if you need to change the number somewhere else too...), and apparently he didn't test :)
22:24:22Bilgusthe pcmbuf_fade_tick function is not linear
22:25:24[Saint]Why does the build client even need https?
22:25:54Bilgusrather it is linear
22:26:00[Saint]We don't need anonymity and any security gains are very questionable.
22:26:02chrisjj" Like, how to deal with if they already have a dramatic crossfade set." The two should be independent.
22:26:26pamaurygevaerts: yeah I just realized, I'm sending an email to him zagor and badger
22:27:45[Saint]What practical benefit does https between server and client achieve?
22:27:58[Saint]It's not like we publish hashes or have verifiable builds.
22:28:00Bilguschrisjj I can send you the .rock if you want to check it out but be aware that i'm putting it in core
22:28:38Bilgusthe functions and such will be put into the fade to sleep function though
22:28:42gevaertsHmm, I think the revision number is in a config file on the server as well, so if it's bumped I suspect zagor needs to be told
22:28:50chrisjjBilgues: OK.
22:29:11Bilguswhat target are you running?
22:29:41[Saint]gevaerts: is there something I'm missing about this? As far as I see it, aside from being general good practice, there's little to no functional benefit in https for the build clients.
22:29:56gevaerts[Saint]: right now I care about unbreaking the system
22:30:03[Saint]revert it then.
22:30:17gevaertsIt's not *that* simple
22:31:06gevaertsThere might be clients (like pamaury's, but he of course knows there's the issue) that have the current revision, which has a wrong revision number
22:31:18gevaertsSo we need a revision bump either way
22:31:42gevaertsAnd that requires a zagor-intervention as far as I can see
22:31:52pamaurygevaerts: for the moment I have editted my client manually to switch to http
22:32:05gevaertsSo if zagor decides to enable https at the same time, why not?
22:32:24pamauryI think the simplest option is just to enable https and bump version on the server
22:32:27chrisjjI'm happy to wait until it goes in core. Though the only good ZEN build has not yet reached the master.
22:32:47[Saint]Yeah, I guess. I'd be surprised it anyone more than pamaury got caught up in this.
22:32:47[Saint]I mean, __builtin assuredly didn't test it.
22:32:56gevaertsWell, bump version on both sides
22:33:05chrisjjBilgus: I'm happy to wait until it goes in core. Though the only good ZEN build has not yet reached the master.
22:34:21pamauryhopefully we will eventually find out all subdomains of that need a certificate ;)
22:34:41[Saint]mildly hilarious it's not already the case.
22:34:47[Saint]given who operates the server.
22:35:24[Saint]The dude's like, Cpt. HTTPS or Mjr. Opportunistic Encryption.
22:36:24[Saint]I understand the exit from the community but I would've thought you'd test in your own back yard and https is far from new tech.
22:37:35pamauryWell we still have problems with the tools. For example there is the licence problem that prevents Rockbox Utility from using https
22:37:36gevaerts[Saint]: he was going to, but he wanted to relax so he curled up with a book
22:43:02[Saint]pamaury: I'm not really sure that's much of a realistic concern given where theyre served from is it?
22:43:10[Saint]Its more of a meta-concern.
22:44:17[Saint]Do Swedes give a shit about bands on strong encryption export? I thought not.
22:45:28pamauryyou are referring to the fact that the build server is insecure ? yes I agree with that, but it's everyone's fault
22:46:22[Saint]No. I'm referring to your statement about RbUtil et al not using https.
22:46:35[Saint]I don't think that matters at all given where the binaries are served from.
22:46:44[Saint]Good luck telling .se not to do something.
22:46:50gevaerts[Saint]: does Sweden ban exporting strong encryption?
22:47:06[Saint]gevaerts: no, that's my point.
22:47:20gevaertsSo why is it even a question?
22:47:55[Saint]I believe that's the problem pamaury is referring to with including https in rbutil.
22:48:16gevaertsYes, I know *that*...
22:48:25*gevaerts is confused now
22:48:37gevaertsI thought you said something about Swedes and exporting encryption?
22:49:02 Join lebellium [0] (
22:50:12[Saint]Well, I did. But...gah. It's all fucky now. What I was saying was basically a long winded way of saying that AFAIK the concern I think pamaury has about https vs. RbUtil and other distributed Rb tools isn't particularly valid by my understanding.
22:50:42[Saint]If anyone ever did get pissy about it best they could do is say "don't" to which the reply would be "nah, we're OK. Thanks"
22:52:32[Saint]I think the list of projects that actually do take that license aspect seriously are probably few and far between anyway.
22:53:33[Saint]huh what? I have no idea what you're confused about.
22:54:14[Saint]Is there some _other_ other problem with the distributed tools and https I'm unaware of? I know of precisely one, and it's questionable at best.
22:54:33Bilgus__builtin: I think this belongs to you.
22:55:09Bilgusrockbox/apps/plugins/puzzles/rbwrappers.c: In function ‘vsscanf’: /usr/include/i386-linux-gnu/bits/string3.h:90:1: sorry, unimplemented: inlining failed in call to ‘bcopy’: redefined extern inline functions are not considered for inlining /rockbox/apps/plugins/puzzles/rbwrappers.c:1892:38: sorry, unimplemented: called from here
22:56:53[Saint]Bilgus: Well, at least it has the good sense to be apologetic about it.
22:56:54gevaertsThe issue is that the openssl license is not compatible with the gpl. I don't think that's questionable at all
23:01:58Bilgusthat is weird he has bcopy defined as memcopy which I believe to be correct
23:02:14Bilgussorry memmove
23:06:48Bilgushell IDk anyways I commented out your version of bcopy and replaced it with #define bcopy(s1, s2, n) memmove((s2), (s1), (n)) and that fixed the compile no clue if it broke anything
23:20:33 Quit pamaury (Ping timeout: 246 seconds)
23:24:20 Quit edhelas (Quit: Leaving.)
23:27:56 Quit n3m9 (Read error: Connection reset by peer)
23:30:38__builtingevaerts: what did I break?
23:31:14gevaertsThe build client
23:31:27__builtingive me a second to read the backlog
23:33:16__builtinok, looks like I enabled https for too many URLs, including services that didn't support it
23:33:42gevaertsThat, and you didn't bump the revision and tell zagor about it :)
23:34:25__builtinthat also explains why /my/ client isn't uploading anything...
23:35:26__builtinI could just revert the line that makes the buildmaster go over TLS
23:38:12__builtinand increment the versions in the client and master
23:38:53*gevaerts nods
23:39:18gevaertsAlthough waiting for zagor's response might be good. He might want to just emable https, in which case just incrementing versions is enough
23:39:31__builtinhas anyone emailed him yet?
23:40:14gevaertsI think pamaury did?
23:40:18__builtinit actually looks like the buildmaster url technically has HTTPS enabled right now, but the certificate is for the wrong domain
23:40:39gevaerts22:26:28 <pamaury> gevaerts: yeah I just realized, I'm sending an email to him zagor and badger
23:40:57gevaertsWell yes, that's the same server that does and many others
23:41:54 Quit diox (Ping timeout: 246 seconds)
23:45:50__builtinfor now shall I just push a commit to disable HTTPS for the upload url and increment the revision?
23:46:16 Quit JdGordon (Ping timeout: 245 seconds)
23:47:13 Join diox [0] (
23:47:25*gevaerts isn't sure
23:47:54__builtinit would fix any problems with people who download the script now
23:51:15__builtinthough incrementing the revision might be a bad idea because it would cause the clients to have a higher revision than the server
23:53:46gevaertsThat's not actually a problem
23:54:03gevaertsThe server checks for less than, and nothing else
23:54:32gevaertsSo higher is fine, and IIRC we actually have sometimes bumped a few times before forcing an update
23:55:56Bilguschrisjj: does the play and pause buttoin not work in the creative zen sim??
23:56:28__builtinBilgus: in what context did the aforementioned error appear?
23:56:45Bilgusbuilding the creative zen sim
23:57:53***Saving seen data "./dancer.seen"
23:58:08__builtinit's probably just a naming conflict from what I gather

Previous day | Next day