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 2020-04-16

00:01:53***Saving seen data "./dancer.seen"
00:10:21speachyoaky, the buildmaster is still thinking my latex manual builds are bad; when I build them by hand they're good so I don't know why it's being a poopy-head.
00:10:38speachyit failed my builder, so the android stuff didn't get built, hence the two errors.
00:11:12speachyso, I'm disabling latex on my builders until I can figure out WTF it's being so persnickety..
00:11:28gevaertsThe buildmaster uses regex matching on the log IIRC to see if things are OK
00:11:46speachythe magic string "not found" it's looking for isn't present when I fire off the exact cmdline the builder uses
00:22:13 Quit prof_wolfff (Ping timeout: 250 seconds)
00:24:42speachywait, my log file was stale.. did I forget to restart the builder with the new stuff? yeah.. looks like it. sigh. in www.git is busted. fixing it will have to wait until after dinner.
01:10:59 Quit ZincAlloy (Quit: Leaving.)
01:15:56 Quit sakax (Quit: Leaving)
01:26:35 Quit pamaury (Ping timeout: 264 seconds)
01:36:40speachyokay, fixed I think.
02:01:57***Saving seen data "./dancer.seen"
02:27:11 Quit MrZeus (Ping timeout: 264 seconds)
02:41:39_Bilgusspeachy do you have a todo list up anywhere?
02:43:29speachyfor infra stuff, code, or all the above?
02:46:38speachyfor the former, definitely. the latter, I don't really have anything beyond various half-started git branches and patches already in gerrit.
02:59:03speachythe lion's share of my code changes lately have been driven by things I'm finding while trying to get the infra going
02:59:19speachyI consider the toolchain bumps as part of that
03:25:41 Quit Natch (Remote host closed the connection)
03:31:01 Join Natch [0] (
04:01:59***Saving seen data "./dancer.seen"
04:53:32speachyI've been trying to do a braindump into a proper todo list
05:02:01speachy...did anyone ever explore adding a TTS engine into rockbox? wait, I see FS #7660 that was abandoned over a decade ago.
05:02:02fs-bluebot_ Plugin espeak TTS engine for rockbox (patches, unconfirmed)
05:03:13 Quit mixfix41 (Ping timeout: 264 seconds)
05:03:16speachythe "pipe dreams" section is growing. :D
05:53:48 Quit Aldem (Read error: Connection reset by peer)
06:00:19 Quit [7] (Ping timeout: 252 seconds)
06:00:27 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
06:02:00***Saving seen data "./dancer.seen"
06:30:49 Quit ac_laptop (Ping timeout: 256 seconds)
07:15:47_BilgusSpeachy uhm Mr(s).Somebody I mean
07:46:36blbro[m]speachy have a look at manual/
08:02:02***Saving seen data "./dancer.seen"
08:04:48 Quit yosafbridge (Quit: Leaving)
08:11:34 Join yosafbridge [0] (
08:57:50__builtinx86 emulator, of course
08:58:05__builtinuser890104 had a working prototype
09:02:07user890104uhm, what about the x86 emulator? :)
09:02:32user890104i even got a 6502 one (oricutron), but it runs slow as hell
09:14:34 Join dys [0] (
09:28:42 Quit _Bilgus (Quit: Leaving)
09:31:24 Join Bilgus [0] (
09:33:40 Join ZincAlloy [0] (
09:38:23 Quit ZincAlloy (Ping timeout: 256 seconds)
09:45:39 Quit Bilgus (Read error: Connection reset by peer)
10:02:03***Saving seen data "./dancer.seen"
10:09:41 Join prof_wolfff [0] (prof_wolff@gateway/vpn/nordvpn/profwolfff/x-22956931)
10:14:01 Quit fragal (Ping timeout: 264 seconds)
10:31:35 Join ZincAlloy [0] (
10:59:03pixelmaspeachy: there was also
11:00:51gevaertsIIRC the main TTS stumbling block was memory management
11:01:07gevaertsI don't know the details
11:02:14 Quit prof_wolfff (Read error: Connection reset by peer)
11:17:36 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
11:20:14 Join prof_wolfff [0] (
11:28:43 Join Bilgus [0] (
11:49:18 Quit pamaury (Ping timeout: 260 seconds)
11:55:35 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
11:57:47 Join PimpiN8 [0] (~PimpiN8@
12:02:05***Saving seen data "./dancer.seen"
13:02:44 Quit prof_wolfff (Ping timeout: 250 seconds)
13:15:54 Join prof_wolfff [0] (~prof_wolf@
13:25:03 Quit Moarc (Ping timeout: 256 seconds)
13:26:01 Join Soap [0] (~Soap@rockbox/staff/soap)
13:29:09 Quit ZincAlloy (Quit: Leaving.)
13:37:02 Join Moarc [0] (
13:47:36 Join ac_laptop [0] (~ac_laptop@
14:01:41 Join rbox [0] (
14:02:09***Saving seen data "./dancer.seen"
14:04:02rboxhi, i've used rockbox for years on a sansa clip 2, until the device died. i've looked through the list of supported devices but many seem to no longer be on sale. could you suggest some devices that are still around, looking for a simple mp3 player, where you could still buy them?
14:15:53 Join MrZeus [0] (
14:17:38speachyblbro[m], I fixed the wrapping issue on the command line. But that that filter script wasn't being used for the html manuals
14:20:27speachy_Bilgus, Hmm, I guess I could augment the MrSomeonesTodoList wiki page, but I can't claim that the stuff on my list was "agreed needs to be done" by anyone other than the voices in my head
14:30:20speachy_Bilgus, here's the state of my brain dump todo:
14:31:40speachyI have a separate one for the infra stuff, mostly migration-related things, but the "misc improvements" section is growing theretoo..
14:50:11 Join massiveH [0] (
15:01:15 Quit MrZeus (Ping timeout: 258 seconds)
15:51:03 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
16:02:11***Saving seen data "./dancer.seen"
16:03:50 Join brasello [0] (
16:36:42 Quit massiveH (Quit: Leaving)
16:42:13 Quit prof_wolfff (Ping timeout: 264 seconds)
17:56:41 Join ubervison_ [0] (~ubervison@2a02:aa12:b106:1b80:4978:337a:24bd:4bbc)
17:57:08 Quit ubervison (Ping timeout: 246 seconds)
18:02:14***Saving seen data "./dancer.seen"
18:06:40 Join Aldem [0] (~Aldem@unaffiliated/aldem)
18:18:42 Join sakax [0] (~r0b0t@unaffiliated/r0b0t)
18:24:36speachyhmm. looking at the MrSomeoneTodoList wiki page, I noticed "show transfer times on the developer build page"
18:25:17 Quit rbox (Remote host closed the connection)
18:25:20speachythis is actually really easy to do (the data is already there) but I don't see it as being terribly useful as the detail page already shows the average transfer speed of each builder in that round.
18:26:23 Join lebellium [0] (
18:26:56speachyanyone think otherwise?
19:04:50Bilguslike the time it took to send the data to a build client and return trip?
19:06:19speachythe data captured is the time it takes to upload the built blob.
19:06:47speachysince for many(most?) of the builds the results aren't actually uploaded.
19:07:18Bilgusyeah so that seems kinda pointless except maybe a metric for performance
19:07:56speachyyeah, the existing page already averages the upload speed for the entire batch on that host.
19:08:14BilgusIf it were going on the build page maybe set a threshold and only show outliers
19:08:44Bilgusbut really a clean buildpage is easier to grok\
19:10:12Bilgusyou said the results aren't uploaded do they get CRCd and rejected then?
19:11:34speachyit only uploads the firmware builds; stuff like checkwps, sims, etc are just there to catch regressions
19:12:54Bilgushmm and every build round uploads all the built fw?
19:13:38Bilgusseems like a lot of extra traffic for something that didn't change
19:14:10speachywell, it's done on every commit, and there's no really good way to know if something would have changed or not
19:14:57Bilguscouldn't you do a crc over the fw and check it against the record on the server?
19:15:17speachyno, becuase the commit id changes therefore the checksum would change. :)
19:15:22speachyevery build is genuinely different
19:16:04speachywe keep 30 days' worth of builds, no more than 1 per day.
19:16:26Bilgusah yeah because we show the Build number in the FW
19:18:13Bilgusthere is already a delay on dev build hitting the main page how is that handled in the 1/day thing?
19:19:21speachyevery night the final build for a given day is archived and becomes that day's build.
19:19:41speachyand if no new build happened on that day, the previous day's build is copied.
19:20:41speachythis way we can easily expire anything over 30 days old rather "delete all but one build for a given day but then keep the most recent 30 files"
19:21:12speachyI think the latter would be nice to have but it's a lot more complicated to implement.
19:27:38Bilgusper your TODO I'm going to start purging HWCODEC (again) as soon as you decide your ARCHOS itches are scratched WIKI - group effort GERRIT maybe we can mark all the cannot merge as abandoned −−the rest maybe have everyone vote for what should be removed
19:28:33speachyhmm. source snapshots, daily changelogs and mapfiles are broken. probably disappeared once the builds page started using the output of the farm instead of generating everything itself.
19:28:49 Quit dys (Ping timeout: 258 seconds)
19:29:10BilgusI think we could knock out 25% of the backlog on Cannot Merge
19:29:24speachygiven that Archos appears to work with a modern toolchain, there's no longer any immediate need to nuke it, except for the hwcodec thing
19:29:52Bilgus*OH mapfiles in the releases is important if you are taking requests for your TODO*
19:30:10speachyreleases or daily builds?
19:30:39BilgusI thought the releases already had it
19:30:40speachyok.. they're not captured as part of the normal 'make zip
19:30:44Bilgusif not there too
19:31:33speachyI mean, adding it to the zip would be the simplest way to capture them for everyone, just bloats the zips a little
19:32:32speachyadding in the map increases the xduoox3 zipfile by ~100k, for example
19:32:46speachy860k extracted.
19:32:49Bilgusyeah like 1-2Kb it'd be worth it to have it on device I could even make up a lua script to search it
19:33:17speachyand that won't cover plugins, codecs, etc. just the main fw image.
19:34:12Bilguswell if its that big them maybe not releases but for sure dev builds
19:35:36speachyno.. it would need to be across the board, we dont' want to have someone unzip a new firmware file that doesn't include a mapfile −− we'll end up with a stale map which is beyond worthless.
19:36:04Bilgusah yeah true\
19:36:23speachythe buildmaster doesn't care all that much; it's the client that decides what to upload.
19:36:34Bilgusits definitely a good trouble shooting tool
19:37:04speachyI'll add maps, changelogs, and source snapshots to the TODO.
19:37:44speachysource snapshots in particular.
19:38:30 Join ZincAlloy [0] (
19:39:22speachyany changes to these probably won't go live until the infra moves to my servers.
19:43:01 Quit ZincAlloy (Ping timeout: 256 seconds)
19:43:16*speachy is tempted to tear into the latest recruiter to cold-call him.
19:44:30speachy... what makes him think that someone with two decades of deeply embedded stuff wants to jump to testing webapps with Selenium?
19:45:30speachyand relocate to sillycon valley, for a 6mo temp-to-maybe-hire gig.
19:49:23Bilguspeople still play the lottery even knowing the odds..
19:50:26speachyI'd rather go tromping around barefoot in that patch of poison ivy I just discovered.
19:51:08 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…)
19:52:16speachybut I digress.
19:53:55speachygot android sdk stuffs installed on my other builder and eventual buildmaster, so I can go ahead and enable the nightly voice and manual build crons and see if anything explodes
19:57:40 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:a570:cc9f:46d2:ae73)
19:57:50 Join MrZeus [0] (
20:00:38 Quit sakax (Remote host closed the connection)
20:02:17***Saving seen data "./dancer.seen"
20:10:33speachyhmm. with respect to the map file, bloating the zip by 100k isn't really a big deal, but I'm more concerned with wasting space on the target. we could always just gzip it first, and put that into the zip file instead?
20:22:40 Quit koniu (Remote host closed the connection)
20:24:35 Join koniu [0] (~koniu@gateway/tor-sasl/koniu)
20:34:22 Join prof_wolfff [0] (~prof_wolf@
20:49:19Bilgusi Suppose but even 1meg isn't going to be that big of a deal or strip the junk from it first
20:49:42Bilgusbut that involves processing and room for mistakes
21:05:12 Join petur [0] (~petur@rockbox/developer/petur)
21:18:09 Join krabador [0] (~krabador@unaffiliated/krabador)
21:22:00speachyhuh, mimic takes 2:38 to generate a voice set where festival takes.. 20:43. yow.
21:25:54 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
21:28:43fs-bluebot_Build Server message: New build round started. Revision e0d7c7c, 295 builds, 9 clients.
21:40:34 Quit prof_wolfff (Ping timeout: 250 seconds)
21:54:49fs-bluebot_Build Server message: Build round completed after 1566 seconds.
21:54:50fs-bluebot_Build Server message: Revision e0d7c7c result: All green
21:58:40speachyd'oh.. I wedged the site
22:00:23speachyshould be okay; next round will un-wedge itself.
22:02:18***Saving seen data "./dancer.seen"
22:37:13 Join S|h|a|w|n [0] (~shawn156@unaffiliated/shawn156)
22:43:07 Join prof_wolfff [0] (~prof_wolf@
22:51:16 Quit lebellium (Quit: Leaving)
23:11:23speachyasked badger & co to unwedge things the right way so we don't lose info on this build round
23:12:21 Quit zagor[m] (Changing host)
23:12:21 Join zagor[m] [0] (bjstmatrix@rockbox/developer/Zagor)
23:12:34zagor[m]speachy: I'm looking at it now
23:12:46speachyawesome, thank you
23:15:15speachyI think I've broken eought things for one day
23:29:01zagor[m]There. Double-check I haven't overwritten your changes. I had a ton of uncommitted local changes I committed now. Mostly http->https, but likely more things too...
23:31:23speachyI have all of the local changes from the original site dump.. I think the cron'd 'git update' was keeping things updated okay
23:31:50speachythe build page is unwedged, and time will tell if anything regressed
23:32:56speachythank you for doing this

Previous day | Next day