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 2017-10-16

00:03:14melissa_wodz where are the instructions for that?
00:03:35melissa_So why does the homepage list ipod classic 6G under unstable?
00:07:58wodzmelissa_: 'Unstable' in rockbox context means not everything is finished (like missing manual or simple install method). It doesn't correspond to how often player will crash.
00:12:25 Quit pamaury (Ping timeout: 240 seconds)
00:17:08 Quit wodz (Quit: Leaving)
00:22:04 Quit almog1006 (Quit: Page closed)
00:23:03 Quit melissa_ (Quit: Leaving)
00:43:54 Quit ender` (Quit: PCMCIA: People Can't Memorize Computer Industry Acronyms)
00:57:33 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…)
00:59:14 Quit petur (Remote host closed the connection)
01:15:15__builtinI think the only thing that's preventing it from being marked stable is the manual (which has been partially written), and rbutil support (which just has to be merged)
01:28:30 Quit [Saint] (Disconnected by services)
01:28:35 Join [Saint] [0] (~sinner@rockbox/staff/saint)
01:31:49 Join krabador [0] (~krabador@unaffiliated/krabador)
01:50:01***Saving seen data "./dancer.seen"
02:10:11 Quit krabador (Remote host closed the connection)
02:13:34 Join krabador [0] (~krabador@unaffiliated/krabador)
02:36:46 Join krabador_ [0] (~krabador@unaffiliated/krabador)
02:39:20 Quit krabador (Ping timeout: 255 seconds)
02:56:17 Quit bray90820 ()
03:04:06 Quit Moarc (Ping timeout: 240 seconds)
03:04:51 Join Moarc [0] (
03:07:55 Quit krabador_ (Remote host closed the connection)
03:08:14 Join krabador_ [0] (~krabador@unaffiliated/krabador)
03:08:40 Quit krabador_ (Remote host closed the connection)
03:08:44 Quit Huntereb (Quit: See ya!)
03:09:30 Join Huntereb [0] (
03:50:05***Saving seen data "./dancer.seen"
04:10:29 Quit alexweissman ()
04:29:26 Quit paulk-gagarine (Ping timeout: 246 seconds)
04:30:59 Join paulk-gagarine [0] (
05:15:45_Bilgusjohnb i put a new clip+ image up that will give you error messages on read and write failures
05:25:28 Join TheEaterOfSouls [0] (~souls@unaffiliated/theeaterofsouls)
05:26:09TheEaterOfSoulsDoes the as3525 do hardware mp3 decoding? I didn't see it listed in the features.
05:30:51 Join JdGordon_ [0] (~jonno@rockbox/developer/JdGordon)
05:33:58 Quit JdGordon (Ping timeout: 255 seconds)
05:46:44TheEaterOfSoulsHow come playing mp3s seems to take less battery life then? My understanding is that it's more complex to decode than flac.
05:47:02TheEaterOfSoulsThough I'm not very knowledgeable about audio codecs.
05:50:09***Saving seen data "./dancer.seen"
05:53:07_BilgusI'm not overly familar with the audio codecs but my guess would be it has been very optimized
05:55:33_Bilgusthis shows that the flac codecs are way less processor intensive AFAICT
05:56:18_Bilgus**ON MOST PROCESSORS
06:01:56 Join jhMikeS [0] (
06:03:33TheEaterOfSoulsI always thought battery life with mp3 was a bit better, but maybe I just wasn't paying attention.
06:05:13_BilgusIt is hard to say for sure really I mean there are other things to consider as well such as file size
06:05:19 Quit TheSeven (Ping timeout: 246 seconds)
06:05:53 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
06:07:15_Bilgusfigure if the device had to buffer less on your mp3 for a comparable bitrate that would skew those results
06:08:01_Bilgusor if the whole thing fit in the buffer in one shot that would add up in power savings even more
06:10:53 Quit TheSeven (Ping timeout: 258 seconds)
06:11:35TheEaterOfSoulsIdk, they're good sized files usually.
06:11:47TheEaterOfSoulsMy guess is just that my battery isn't holding as much charge anymore.
06:12:09TheEaterOfSoulsStill gets me +8 hours so I'm not complaining :D
06:12:54 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
06:22:23 Quit TheSeven (Ping timeout: 258 seconds)
06:22:46 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
06:32:04 Join johnb4 [0] (
06:41:24 Quit Ruhan (Quit: Connection closed for inactivity)
06:45:07 Quit TheEaterOfSouls (Quit: ChatZilla 0.9.93 [Firefox 56.0/20170926190823])
06:48:41johnb4_Bilgus: image or recovery FW?
06:49:49johnb4I see a new recovery FW.
06:50:33_Bilgusyeah firmware the clip+ thats 32mb is probably lost
06:52:00johnb4I will try the new recovery on the read-only player today after work (in 12h).
06:52:58_Bilgusone thing to be aware of is you have to wait for the init message to dissappear before you plug the USB (10s)
06:54:47 Quit johnb4 (Quit: Nettalk6 -
07:43:40_BilgusPamaury (Logs) No Auto Slow had a decrease of 1hr of runtime versus the normal code which ran for 20 hours so overall 5% less runtime
07:50:11***Saving seen data "./dancer.seen"
08:07:59 Join ender` [0] (
08:24:26 Quit jeandet (Ping timeout: 255 seconds)
08:27:27 Quit utrack (Ping timeout: 240 seconds)
08:29:38 Join utrack [0] (~utrack@unaffiliated/utrack)
08:43:51 Join Bilgus [0] (~Bilgus@unaffiliated/bilgus)
08:47:28 Quit _Bilgus (Ping timeout: 248 seconds)
08:47:34 Quit _meg (Ping timeout: 240 seconds)
08:50:43 Join _meg [0] (~notsure@
08:57:16 Join wodz [0] (
09:09:05 Join petur [0] (~petur@
09:09:05 Quit petur (Changing host)
09:09:05 Join petur [0] (~petur@rockbox/developer/petur)
09:12:51 Join bray90820 [0] (
09:13:17 Quit man_in_shack (Quit: WeeChat 1.9)
09:27:49 Quit dys (Ping timeout: 252 seconds)
09:31:29 Join dys [0] (
09:34:17 Join _Bilgus [0] (~Bilgus@unaffiliated/bilgus)
09:37:26 Quit Bilgus (Ping timeout: 240 seconds)
09:38:41 Quit _Bilgus (Ping timeout: 255 seconds)
09:50:13***Saving seen data "./dancer.seen"
09:56:19 Quit [Saint] (Read error: Connection reset by peer)
09:56:35 Join [Saint] [0] (~sinner@rockbox/staff/saint)
10:02:34 Quit dys (Ping timeout: 240 seconds)
10:12:42 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
10:14:12 Quit [Saint] (Remote host closed the connection)
10:15:04 Join [Saint] [0] (~sinner@rockbox/staff/saint)
10:22:32 Join jeandet [0] (sid239360@gateway/web/
10:35:27 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
10:44:47 Join dys [0] (~dys@2003:5b:203b:100:6af7:28ff:fe06:801)
10:55:16 Join PimpiN8 [0] (~textual@
10:59:07 Quit pamaury (Ping timeout: 252 seconds)
11:40:42 Join _jhMikeS_ [0] (
11:40:42 Quit jhMikeS (Disconnected by services)
11:40:43 Nick _jhMikeS_ is now known as jhMikeS (
11:47:02 Join Bilgus [0] (~Bilgus@unaffiliated/bilgus)
11:50:17***Saving seen data "./dancer.seen"
11:52:08 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
11:55:46 Quit Bilgus (Ping timeout: 252 seconds)
11:56:07 Join lukeoftheaura [0] (5eb95632@gateway/web/cgi-irc/
11:57:21 Join Bilgus [0] (~Bilgus@unaffiliated/bilgus)
11:57:24lukeoftheaurawhat's the best cheap rockbox device available today? Would I have to get an older/second hand player or are some currently-made players supported?
11:59:50 Join _Bilgus [0] (~Bilgus@unaffiliated/bilgus)
12:01:40 Quit Bilgus (Ping timeout: 240 seconds)
12:04:57 Quit lukeoftheaura (Quit: - A hand crafted IRC client)
12:05:31 Quit _Bilgus (Ping timeout: 260 seconds)
12:23:56 Join Bilgus [0] (~Bilgus@unaffiliated/bilgus)
12:27:58 Quit kugel (Ping timeout: 255 seconds)
12:28:27 Quit Bilgus (Ping timeout: 258 seconds)
12:29:25 Join kugel [0] (
12:29:25 Quit kugel (Changing host)
12:29:25 Join kugel [0] (~kugel@rockbox/developer/kugel)
12:42:19 Join Bilgus [0] (~Bilgus@unaffiliated/bilgus)
12:53:53 Join _Bilgus [0] (~Bilgus@unaffiliated/bilgus)
12:56:47 Join _bilgus_ [0] (~Bilgus@unaffiliated/bilgus)
12:57:07 Quit Bilgus (Ping timeout: 240 seconds)
12:58:36 Quit _Bilgus (Ping timeout: 260 seconds)
13:07:02 Nick _bilgus_ is now known as Bilgus (~Bilgus@unaffiliated/bilgus)
13:07:28 Join CH23 [0] (4dfa0218@gateway/web/freenode/ip.
13:08:05 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…)
13:13:27wodzpamaury: If you are interested wodz/qemu-atj">
13:21:20pamaurywodz: thnaks, I'm interested to see that indeed
13:21:56wodzpamaury: It is pretty hacky and vastly incomplete :-)
13:44:01 Join Aldem [0] (~Aldem@unaffiliated/aldem)
13:50:21***Saving seen data "./dancer.seen"
14:01:40 Join krabador [0] (~krabador@unaffiliated/krabador)
14:33:00 Join Ruhan [0] (uid76353@gateway/web/
14:48:21 Join amayer [0] (
15:24:17 Quit shrizza_ (Quit: shrizza_)
15:24:44 Join shrizza [0] (
15:24:57 Quit wodz (Ping timeout: 264 seconds)
15:43:17 Join Aldem_ [0] (~Aldem@unaffiliated/aldem)
15:45:25 Join user890104_ [0] (Venci@unaffiliated/user890104)
15:47:35pamauryjhMikeS: ping
15:48:11 Join TheLemon1an [0] (~lemonboy@irssi/staff/TheLemonMan)
15:48:14 Quit Aldem (Read error: Connection reset by peer)
15:48:14 Quit user890104 (Quit: .)
15:48:14 Quit TheLemonMan (Ping timeout: 246 seconds)
15:48:14 Nick user890104_ is now known as user890104 (Venci@unaffiliated/user890104)
15:50:02 Nick TheLemon1an is now known as TheLemonMan (~lemonboy@irssi/staff/TheLemonMan)
15:50:22***Saving seen data "./dancer.seen"
15:50:41 Quit pystar89 (Ping timeout: 260 seconds)
16:19:59 Join ZincAlloy [0] (~Adium@2a02:8108:8b80:1700:edfb:6c6b:6f08:9dc3)
16:21:30 Quit deevious (Ping timeout: 255 seconds)
16:23:15 Quit krabador (Quit: Leaving)
16:28:31 Join deevious [0] (~Thunderbi@
16:29:37 Quit Acou_Bass (Quit: byeeeeeeeeeeeeeee)
16:31:03 Join Acou_Bass [0] (
16:38:18shrizzaJust for your guys' info: a page of vendors and responses with regard to the krack thing:
16:51:35 Join PimpiN8 [0] (~textual@
17:00:18 Join krabador [0] (~krabador@unaffiliated/krabador)
17:20:47 Quit LjL (Ping timeout: 240 seconds)
17:27:26 Quit krabador (Quit: Leaving)
17:39:56 Quit jhMikeS (Ping timeout: 240 seconds)
17:40:55 Join jhMikeS [0] (
17:45:33 Quit jhMikeS (Ping timeout: 252 seconds)
17:46:21 Quit petur (Quit: Connection reset by beer)
17:50:24***Saving seen data "./dancer.seen"
17:52:29 Join pystar89 [0] (
18:19:55 Join LjL [0] (~ljl@unaffiliated/ljl)
18:20:05 Join johnb2 [0] (
18:24:11 Quit LjL (Read error: Connection reset by peer)
18:26:58pamauryBilgus: re auto slow: I'm surprised by the 20hours runtime, I thought the battery life was ~40h. Or is your batter old?
18:27:20pamauryanyway thanks for test. Now I guess I'd like to benchmark the HBUS clock change
18:27:24BilgusI didn't run it completely dead
18:27:31pamauryah I see
18:27:33Bilgusits running the hbus one now
18:28:04pamauryoh cool :)
18:28:27BilgusI figured 20 hours was a good enough sampling without running the darn thing for nearly a week
18:29:33pamauryyeah probably
18:29:59BilgusHmm looking at the log though it loos more like it would get 27 hours or so what are the 40 hrs based on?
18:31:48BilgusAh nm its starting at 93% battery since I can never catch the damn thing to turn it off when its done charging
18:33:43 Quit dys (Ping timeout: 255 seconds)
18:33:58 Join LjL [0] (~ljl@unaffiliated/ljl)
18:34:36 Quit pamaury (Ping timeout: 260 seconds)
18:43:56 Quit johnb2 (Ping timeout: 240 seconds)
18:44:40 Quit LjL (Read error: Connection reset by peer)
18:51:56 Join dys [0] (
18:52:51 Join LjL [0] (~ljl@unaffiliated/ljl)
18:59:03 Quit LjL (Read error: Connection reset by peer)
19:07:23 Join LjL [0] (~ljl@unaffiliated/ljl)
19:15:06 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
19:20:49Bilgusmy test consisted of 10 files total, 1 opus 67kbs 7 mp3 128 and 2 AAC-HE files nero_hev2_64 and 37kbs
19:21:07Bilgusso yeah it would be a bit lower I imagine
19:22:18Bilguswhen the hbus one gets the same range as the others ill put it all up on mediafire
19:23:04pamaurybe sure to check it's actually the changing hbus (by looking at debug > HW info > clock)
19:25:44Bilgusit flickers hbus idiv 1 fdiv _ freq 64000 AND hbus idiv _ fdiv 14 freq 198947
19:27:49 Quit LjL (Remote host closed the connection)
19:28:18 Join LjL [0] (~ljl@unaffiliated/ljl)
19:28:51 Quit LjL (Remote host closed the connection)
19:31:38 Join LjL [0] (~ljl@unaffiliated/ljl)
19:36:20 Join lebellium [0] (
19:43:17 Join krabador [0] (~krabador@unaffiliated/krabador)
19:49:37 Quit krabador (Quit: Leaving)
19:50:26***Saving seen data "./dancer.seen"
19:50:38 Join krabador [0] (~krabador@unaffiliated/krabador)
20:03:47 Quit krabador (Quit: Leaving)
20:09:08 Join krabador [0] (~krabador@unaffiliated/krabador)
20:12:20 Nick Aldem_ is now known as Aldem (~Aldem@unaffiliated/aldem)
20:20:05 Join johnb2 [0] (
20:40:44 Join petur [0] (~petur@rockbox/developer/petur)
20:43:15 Quit krabador (Quit: Leaving)
20:44:40 Join krabador [0] (~krabador@unaffiliated/krabador)
20:46:07 Quit johnb2 (Quit: Nettalk6 -
20:53:32 Quit krabador (Remote host closed the connection)
20:54:54 Join krabador [0] (~krabador@unaffiliated/krabador)
20:56:35 Quit krabador (Remote host closed the connection)
21:08:46 Join johnb2 [0] (5b3aff9d@gateway/web/freenode/ip.
21:09:12johnb2Bilgus: where should the log of the new recovery fw go?
21:09:33Bilgusit just displays to screen
21:09:42johnb2I am trying to write to the read-only clip+.
21:09:52BilgusI didn't want to risk writing to disk
21:10:02johnb2Nothing is shown. I had one panic with a stkov message.
21:10:35Bilgusthe panic is called if it can't find the drive at all
21:10:43johnb2I mean apart from the init screen after Rolo.
21:11:07Bilgusbut if no messages pop up then as far as the firmware is concerned the write happened
21:11:55Bilgusif you can get the panic back or the stack overflow could you tell me the messages?
21:12:02johnb2After reboot it still show the 170128 fw, i.e. loads from internal.
21:12:57Bilgusok so lets run this down a bit 1st you are writing the clppa_new.bin to the device with dd?
21:13:10johnb2yeah, I will give it a few more trials. Right now I see the Input/output error again while dumping, with just 16mb.
21:13:33Bilgusthe dump part what are you putting as count?
21:13:54johnb2bs=512 count=$((0xF000))
21:14:12johnb2I dded clip+4GB_Recovery_image32.bin
21:14:15Bilgusah you are still ulling 32 mb so let me calculate how much that needs to be
21:14:39Bilgusyou'll want to put the multiboot bl on there directly
21:14:55BilgusIIRC its clppa_new.bin
21:15:20johnb2The one from the forum post or the 16MB no internal?
21:15:35johnb2MultBoot forum post
21:15:36Bilgusfrom the forum under unsupported builds
21:15:48johnb2got it.
21:15:51Bilgusthe one with no internal was for your clip+ with ATA error
21:16:05Bilguslet me calc count brb
21:16:08johnb2I got the panic again. About to pastebin
21:20:22Bilguscount you'll want 0x7800
21:21:20johnb2for the clppa_new.bin one, right?
21:23:08*johnb2 not having focus follows mouse ...
21:23:24Bilgusmake count 0x8000 to be safe
21:23:53Bilguswell not when you put it on the device but when you use dd to pull it off
21:24:07johnb2of course.
21:24:23Bilguswhen if=clppa_new.bin the count will be determined by file size
21:27:18johnb2well, the md5 don't match, but as the file sizes of clppa_new.bin and download.bin are different that is expected.
21:27:22Bilgushuh the stack overflow is happening in cpu_boost not sure what that would mean
21:27:37Bilguswhat are the file sizes
21:27:48johnb215728640 Feb 10 2017 clppa_new.bin
21:27:54johnb216777216 Oct 16 17:25 download.bin
21:28:34Bilgusyeah make count 0x7800 and they should match
21:29:19 Join _Bilgus [0] (~Bilgus@unaffiliated/bilgus)
21:29:43 Join Bilgus_ph [0] (~Bilgus_ph@
21:29:44johnb2ls -rtl
21:31:03johnb2No luck: download.bin.7800 has a different size
21:31:26_Bilguswhats the size now?
21:31:50johnb2just a typo, size is correct, md5 is different
21:32:14_Bilgusis dd working without error now?
21:33:06 Quit Bilgus (Ping timeout: 260 seconds)
21:33:26_Bilgusok lets do a script that runs for 10-20x
21:44:18johnb2back in 10min
21:45:41_Bilgusjohnb2 consider that pseudo code lol
21:50:29***Saving seen data "./dancer.seen"
21:52:30 Quit dys (Ping timeout: 248 seconds)
21:56:06 Join krabador [0] (~krabador@unaffiliated/krabador)
21:58:22johnb2what is download.bin supposed to be?
21:58:47 Quit krabador (Client Quit)
22:03:12_Bilgusitd be your first dd
22:03:19 Quit Bilgus_ph (Read error: Connection reset by peer)
22:03:19_Bilguslike the original state
22:03:41_Bilgusgive me a second i'm writing it for real
22:07:42 Join krabador [0] (~krabador@unaffiliated/krabador)
22:09:43johnb2Bilgus: I had put the one with the ATA error in the fridge for 30min. Now when I connect it to usb, it starts to load RB, but turns off due to low battery :-(
22:10:10_Bilgusnice that worked though lol
22:10:31 Quit lebellium (Quit: ChatZilla 0.9.93 [Firefox 56.0/20170926190823])
22:17:35 Join dovber__ [0] (~dovber@2600:8801:3180:45:beee:7bff:fee3:335f)
22:17:38 Quit dovber (Ping timeout: 258 seconds)
22:18:18*pamaury wonders how many allocations are involved when the clip+ ROM receives a setup packet. The answer is at least 2, and that's even before starting to process it
22:18:27 Quit krabador (Quit: Leaving)
22:18:37*pamaury and not counting allocations done just to handle the irq of course
22:18:56 Quit pixelma (Quit: .)
22:18:56 Quit amiconn (Quit: - Chat comfortably. Anywhere.)
22:19:18[Saint]pamaury: I'm going to be rebuilding my arm build cluster today.
22:19:30[Saint]Is your crosscompiler all done?
22:19:46 Join krabador [0] (~krabador@unaffiliated/krabador)
22:19:54[Saint]Are there any gotchas I need to be aware of before building your toolchain?
22:21:40pamaury[Saint]: yes, not particularly. Most probably it will complain that you don't have the required library (with a nice error message). You want to install lib{mpc,gmp,mpfr}-dev
22:21:41 Quit Huntereb (Read error: Connection reset by peer)
22:21:54 Join Huntereb [0] (
22:22:19 Join amiconn [0] (~amiconn@rockbox/developer/amiconn)
22:22:19 Join pixelma [0] (~pixelma@rockbox/staff/pixelma)
22:22:26_Bilgusthat should work to push the fw onto the device
22:22:52_Bilgusit'll also warn you if the current md5 doesn't match the original file
22:25:50[Saint]pamaury: Ok, thank you.
22:26:01[Saint]pamaury: it's just the one Gerrit task then?
22:26:14[Saint]I don't need to go fishing across several related tasks?
22:26:44[Saint]Sorry for the questions, I had built it a while ago but a lot has changed since then and I'm in entirely different architecture now.
22:27:53pamaury[Saint]: it's all committed
22:28:10pamauryyou don't need any gerrit task
22:28:13 Quit Huntereb (Read error: Connection reset by peer)
22:28:19[Saint]Oh, I see. Excellent.
22:28:31[Saint]That's great news. Sorry I'm sorry behind the times.
22:28:34pamauryjust fire from git and select the linux arm rockbox
22:29:00[Saint]I suppose I'll need to edit the "Linux Simple Guide" then to reflect this.
22:29:04[Saint]Thank you.
22:29:59[Saint]pamaury: Incidentally, is there any trivial way of building the legacy toolchain with the current source?
22:30:26[Saint]Or would I just take the easy route and checkout a commit prior to this?
22:31:08pamaury[Saint]: you mean the ypr0 toolchain?
22:31:44 Join Huntereb [0] (
22:31:44pamauryI haven't deleted any toolchain from, it can still build the ypr0 toolchain
22:31:54johnb2_Bilgus: the /dev disappeared after the first round:
22:31:57pamaurybut I switched the ypr0 to the new toolchain so it's mostly irrelevant now
22:32:06johnb2I will restart the player and try again.
22:32:32johnb2It shows the same panic message again.
22:32:51[Saint]pamaury: right, yes. Good to know.
22:33:15 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
22:34:09_Bilgushmm ok let me try and disable the cpuboost not sure why its throwing up there
22:35:38[Saint]I want to finish something I started a long while back today so I'm just making sure I know what current pathways are.
22:36:41[Saint]I would like to finish my one-stop-shop-to-deploy-a-build-farm-client script.
22:37:24[Saint]Because I feel like there would be more available clients if end users didn't need to understand a singular thing about the build system and underlying mechanics.
22:37:47johnb2_Bilgus: this time it performed 2 rounds, then /dev disappeared.
22:37:57_Bilguslets put a delay in there
22:38:20 Join p3tur [0] (~petur@rockbox/developer/petur)
22:38:21pamauryseriously, the people who wrote the clip+ are morons
22:38:34[Saint]I feel like if they could just run a script that did everything from gathering source, to building the toolchains, and setting up a service to keep the build environment up more end users would be giving us their extra CPU cycles.
22:38:42pamaurythey created a class for usb standard requests, and then subclasses it for EVERY SINGLE request
22:38:50pamauryjust to implement a SINGLE callback
22:39:20[Saint]I would also like to make sure that it if they desire to do so that an end user can set resource caps so they're not negatively impacted.
22:39:24_Bilgusjohnb put sleep 10s after each dd
22:40:03[Saint]Though, it's questionable as to whether or not capping disk, CPU, and RAM usage would be wide here.
22:40:30_Bilgusalso johnb if you plug the usb before that init message disappears it will panic
22:40:59_BilgusI disabled the error messages in this next build to see if it helps its compiling now
22:41:02johnb2I always waited for the init to disappear.
22:42:58[Saint]Am I dreaming or being too idealistic perhaps?
22:43:26 Quit ZincAlloy (Quit: Leaving.)
22:43:35[Saint]If we put a reminder on the downloads page, and end users needed to do nothing but run a single script, do you think they would join our build farm?
22:43:46 Join maffe [0] (
22:44:21[Saint]I'm talking gathering source, building, and making sure there's a service running to keep the build client up. Handling literally everything.
22:44:39[Saint]Am I being too idealistic again? I suppose so.
22:45:41johnb2_Bilgus: the delay seems to help, it's now int the 5th round
22:45:52johnb2*in the
22:46:32_Bilguswell up that delay to like 30s I guess maybe we are just doing it before it can recover
22:46:50johnb2it is still running
22:46:59_Bilgushere is a fw with the error messages disabled
22:47:04_Bilgusah ok cool
22:49:38_BilgusPamaury I'm amazed you can even do anything with that clip+ OF I looked at it for like 3 days and just gave up
22:51:56pamaury_Bilgus: I'm spent countless hours on it, I can tell you it's very hard work
22:52:24pamauryI have found dozens of classes, it's horrible
22:53:09[Saint]It's likely no exaggeration to say that in his time here pamaury has spent hundreds of not thousands of man hours in firmware reversal.
22:53:19_BilgusI was wondering if they had some kind of code obfuscater
22:53:21[Saint]wodz would also be right up there.
22:53:57pamauryWell it's unclear to me if they tried to obfuscate it, it sounds unlikely to me
22:54:11pamaurybecause I'm ready to bet that the original c++ code is essentially unreadable
22:54:44pamaurytypical "lasagna" code where in doubt you throw another object at the problem
22:55:14pamaurythat plus dumb compiler that didn't devirtualize method calls
22:55:30[Saint]In their defense, lasagna is awesome.
22:55:41_Bilgusnot mondays though
22:55:55pamaury[Saint]: only when there are at most 4 layers ;;)
22:56:44_BilgusI feel kinda like c++ just lends its self to spaghetti code though
22:57:28_Bilgusbut maybe i'm just OOP dumb
22:58:40pamauryyeah, you kind of get the worst of everything in OOP
22:59:31__builtin[Saint]: I don't see why that wouldn't be possible
22:59:49johnb2_Bilgus: 20 rounds with no success. I am going to loop 200 over night.
23:00:11__builtinThough dealing with different platforms would be a pain
23:00:29_Bilgusjohnb2 I don't know that it would help
23:00:55johnb2Tomorrow I would try and work on the one with the ATA error. What would be the count to be used to verify the "no internal image"? ?
23:01:06[Saint]__builtin: oh, sure, the *possible* part doesn't concern me.
23:01:11[Saint]I know the answer to that.
23:01:24[Saint]It's the *practical* part that worries me.
23:01:30_Bilgusuh hold on a second and ill open it up in a hex editor and see probably the same
23:02:18[Saint]__builtin: y'know what I'm saying, right?
23:02:33[Saint]It's all very well and good to build the thing, but, would anyone actually use it?
23:03:42__builtinAh yes, I see
23:04:01[Saint]Dealing with different platforms would indeed be an area of concern, but I'm not /too/ worried there.
23:04:25pamaury_Bilgus: from what I see, I think the ROM accepts three vendor defined requests, but I can't tell what they do at the moment
23:04:31[Saint]It's more "will clients actually use this, regardless of how easy anyone makes it for them to do so"?
23:04:42[Saint]I'm likely overthinking it.
23:04:54[Saint]It definitely couldn't *hurt* anyone.
23:05:09__builtinIt really depends on who our "users" are
23:05:46[Saint]Yes. For the better part of a decade we have had really no idea who our userbase actually is.
23:05:55_Bilgusjohnb2 the count is the same for both
23:06:08[Saint]I may do a survey on this and throw it out on the main page and the mailing lists.
23:06:30[Saint]Just to get an idea of the group(s) that comprise our userbase.
23:06:35__builtinmight be some sampling bias there
23:06:41__builtinespecially on the mailing list
23:06:45[Saint]Yes. Indeed.
23:06:45johnb2_Bilgus: thanks for your help today. Going to bed now.
23:06:52[Saint]It's troubling.
23:07:02_Bilgusthanks john good luck
23:07:12__builtinI mean, do we *need* to expand the build system that much?
23:07:12[Saint]It would indeed be somewhat of an echo chamber.
23:07:39[Saint]__builtin: eh...not really, but redundancy is always a good thing.
23:07:39 Quit johnb2 (Quit: Page closed)
23:07:48_Bilguspamaury I find it interesting that they have that functionality in the OF surely one does the badblocks ident
23:08:08[Saint]I don't want to get back into the situation we were in where only one client can build for any given architecture.
23:08:39[Saint]Everyone wants to build for arm...but the other architectures are overlooked.
23:09:08__builtinhmm, I have a second pi3 sitting idle now
23:09:12[Saint]There's never anything wrong with throwing more clients at a distributed build cluster, at least in my opinion.
23:09:32[Saint]It can only help.
23:09:40_Bilgusnow by distributed is it piecemeal or just roundrobin?
23:09:54[Saint]First last the post.
23:10:00__builtincan a windows machine be a client?
23:10:02pamaury_Bilgus: I'd like to find out if the ROM sonehow has a "magic" access to the raw flash or some undocumented sd/flash command
23:10:08[Saint]Yes. It can.
23:10:34__builtinthrough mingw/cygwin or some other way?
23:10:43[Saint]Though it's non-trivial.
23:11:35__builtincould such a script automate the process?
23:11:37[Saint]You could do a number of routes. I imagine mingw would be preferable? I'm not too savvy with the windows layers.
23:11:49[Saint]Yes. It could.
23:12:02__builtinwell, in that case, I think it'd be worthwhile
23:12:14__builtineven a couple more clients would speed the builds greatly
23:12:16[Saint]I couldn't do it personally though. I'm just not experienced with Windows building.
23:12:28[Saint]And I don't have a single MS machine here.
23:12:59__builtinif it's limited to -nix it'd be pretty redundant, though
23:13:04_BilgusI think in the case of windows machines it might be better to have a VM that literally only does builds
23:13:32_Bilguslike to the point of no gui even
23:13:53[Saint]__builtin: eh,'s speculative but I think we still catch a lot of the young crowd who want to replace everything with "Linux"
23:14:34[Saint]You know the ones. They don't know why they want to run Linux or replace the firmware of everything they own, but it doesn't stop them.
23:15:30[Saint]As it is none of them are going to stumble upon our build system organically.
23:15:56[Saint]Hell, ideally I think it would be bloody neat to integrate this into Rockbox Utility.
23:15:59__builtinthat still seems redundant when we already have a linux client script
23:16:17[Saint]There's not a shitload of reasons to actually build the toolchains on the build ckient
23:16:31__builtinwhy not just improve that to automate some of the setup and promote it some more?
23:16:47[Saint]So that could just download the toolchains and then make sure they're in the path.
23:16:57[Saint]That reduces complexity by an order of magnitude.
23:17:27[Saint]The toolchains don't change enough to warrant each intended client building them themselves.
23:17:32[Saint]At least in my opinion.
23:18:24[Saint]You could always add the toolchains to the build targets, as well.
23:18:50 Join dys [0] (
23:19:05[Saint]But that would generate a lot of churn because as it is we build for everything regardless as to whether it has been changed or not.
23:19:37[Saint]Because it's actually really difficult to know in advance if any given change hits or misses any given target in advance.
23:19:52[Saint]But that's the nature of the source and layout. It is what it is.
23:20:02[Saint]Changing that is an absurd amount of work.
23:20:10__builtinyou could just modify the build client to automatically clone the repo and get the toolchains
23:20:24__builtinthen just link to it on the frontpage and from rbutil
23:21:27[Saint]One thing I would really like to investigate at some point in the future is making our builds reproducible.
23:21:47[Saint]But that is also non-trivial with the way our project is laid out.
23:22:26[Saint]I think reproducible builds are awesome and something an open project should arrive for.
23:22:36[Saint]It's really fucking difficult though.
23:22:58[Saint]*strive for (phone autocorrect, sorry).
23:23:01__builtinI don't know what benefit that would provide though
23:23:38[Saint]Peace of mind, largely.
23:24:00__builtinso security
23:24:17[Saint]I think any open project should provide its end users the means to recreate the binaries they ship *exactly*.
23:24:44[Saint]It's moving slowly towards adoption in a few other projects, which is nice.
23:25:37__builtinI just don't see how it would be useful
23:26:03__builtinsure, it'd ensure that the builds aren't being tampered with, but why would someone want to tamper with them in the first place?
23:27:01__builtinthe devices we target are largely isolated from the outside world
23:27:07[Saint]It's not just that, you also know that absolutely everybody is on exactly the same page at any given point for debugging.
23:27:49[Saint]You know that version x of binary y didn't cause something to link weird in their build, yadda yadda...
23:28:03__builtinI see
23:28:06 Quit krabador (Remote host closed the connection)
23:29:43 Join krabador [0] (~krabador@unaffiliated/krabador)
23:31:15__builtinhow difficult would it be to be fully reproducible?
23:31:30__builtinobviously stuff like __FILE__ and __DATE__ would need to go
23:33:29 Quit amayer (Quit: Leaving)
23:33:29__builtinI suppose our version numbers would have to be changed as well
23:39:48 Part robertd1
23:42:34[Saint]The entire process is quite involved.
23:43:48 Quit krabador (Quit: Leaving)
23:47:52 Quit [Saint] (Read error: Connection reset by peer)
23:49:07 Join [Saint] [0] (~sinner@rockbox/staff/saint)
23:50:33***Saving seen data "./dancer.seen"

Previous day | Next day