--- Log for 23.07.124  Server: copper.libera.chat  Channel: #rockbox
--- Nick: rb-logbot  Version: Dancer V4.16  Started: 4 days and 12 hours ago

00.04.02 #       <_bilgus> hmm well maybe you could watch the watermark in the pcm buffer instead
00.05.07 #       <_bilgus> like you know as you end a track it'll start winding down maybe you could look for that?
00.11.24 #       <_bilgus> pcmbuf_unplayed_bytes()
00.24.23 Quit    dconrad (Remote host closed the connection)
00.25.23 Join    dconrad [0] (~dconrad@152.117.104.217)
00.30.36 Quit    dconrad (Ping timeout: 276 seconds)
00.51.50 #       <_bilgus> g#5828
00.51.53 #       <rb-bluebot> 3Gerrit review #5828 at https://gerrit.rockbox.org/r/c/rockbox/+/5828 : 3[Feature] Allow hiding drive names by William Wilgus
00.54.39 Quit    jacobk (Ping timeout: 276 seconds)
01.22.41 Join    davisr [0] (~davisr@fsf/emeritus/davisr)
01.35.55 Quit    davisr (Quit: yeehaw)
01.41.49 Join    davisr [0] (~davisr@fsf/emeritus/davisr)
01.55.52 ***     Saving seen data "./dancer.seen"
02.03.33 Quit    davisr (Quit: yeehaw)
02.24.40 Quit    othello7 (Ping timeout: 260 seconds)
02.39.00 Join    dconrad [0] (~dconrad@152.117.104.217)
02.43.42 Quit    dconrad (Ping timeout: 272 seconds)
03.55.53 ***     Saving seen data "./dancer.seen"
05.02.24 Join    dconrad [0] (~dconrad@152.117.104.217)
05.06.37 Quit    dconrad (Ping timeout: 252 seconds)
05.53.42 Quit    sebagala (Read error: Connection reset by peer)
05.54.07 Join    sebagala [0] (~Burak@185.25.123.34)
05.55.54 ***     Saving seen data "./dancer.seen"
06.06.33 #       <rb-bluebot> Build Server message: 3New build round started. Revision 62b0456a4b, 337 builds, 9 clients.
06.06.34 #       <rb-bluebot> 3puzzles: resync with upstream (adds new Palisade cursor interface). by Franklin Wei
06.18.15 #       <rb-bluebot> Build Server message: 3Build round completed after 702 seconds.
06.18.16 #       <rb-bluebot> Build Server message: 3Revision 62b0456a4b result: All green
07.38.24 Join    dconrad [0] (~dconrad@152.117.104.217)
07.43.30 Quit    dconrad (Ping timeout: 276 seconds)
07.55.55 ***     Saving seen data "./dancer.seen"
07.58.57 Join    dconrad [0] (~dconrad@152.117.104.217)
08.00.30 Quit    dconrad (Remote host closed the connection)
08.29.44 #       <speachy> __builtin: could you also update README.rockbox to reference the latest update and upstream commit it's based off of?
09.01.35 Quit    ursa-major (Remote host closed the connection)
09.01.40 Quit    dook (Remote host closed the connection)
09.01.53 Quit    lane (Remote host closed the connection)
09.02.08 Join    lane [0] (809450f172@2a03:6000:1812:100::1300)
09.02.09 Join    dook [0] (c9db4a6425@2a03:6000:1812:100::1222)
09.02.14 Join    ursa-major [0] (114efe6c39@2a03:6000:1812:100::11f3)
09.21.21 Join    IPG [0] (~InvoxiPla@2.122.6.66)
09.23.13 Quit    speachy (Quit: WeeChat 4.3.4)
09.38.25 Join    jacobk [0] (~quassel@2603:3013:103:ef00:1113:1c75:8d3b:2220)
09.42.24 Join    speachy [0] (~speachy@hurricane.shaftnet.org)
09.42.25 Quit    speachy (Changing host)
09.42.25 Join    speachy [0] (~speachy@rockbox/developer/speachy)
09.42.25 Mode    "#rockbox +v speachy" by ChanServ (ChanServ@services.libera.chat)
09.44.19 #       <speachy> _bilgus: I think we should not bother making that "hide names" thing configurable.   on non-multidrive targets, it's never relevant.  On multi-drive targets with only one drive used/present, it's also not relevant.  It only comes into play when there is more than one drive in active use.
09.46.11 #       <speachy> and we still need to fix the path voice function so that does the right thing with VOL_START/STOP_TOK  :D
09.55.24 Quit    jacobk (Ping timeout: 260 seconds)
09.55.56 ***     Saving seen data "./dancer.seen"
11.14.34 Join    othello7 [0] (~Thunderbi@pool-100-36-176-164.washdc.fios.verizon.net)
11.14.46 Quit    othello7 (Client Quit)
11.19.28 #       <_bilgus> well either way I did get something out of that it turns out building those string every call makes a noticeable difference in the disk test plugin and generating them at boot shouldn't be terrible they are pretty small 12 bytes each I think 
11.27.59 #       <speachy> ew, yeah, let's do it only once
11.30.05 #       <speachy> that will also make it simpler to reverse-lookup to the logical volume so we can voice it properly.
11.49.10 #       <_bilgus> its still not going to be fun figuring out which storage type it is each time maybe it should store the drive type too and let you query that directly 
11.52.17 #       <_bilgus> at init store the proper ID in vollang[v] later volumename_voice(volume) {talk(vollang[volume]) talknum(volume)}
11.55.59 ***     Saving seen data "./dancer.seen"
11.57.39 #       <speachy> pretty sure we already have that info -- volume id maps to drive id, and from there storage_driver_type(driveid)
11.57.41 Join    othello7 [0] (~Thunderbi@pool-100-36-176-164.washdc.fios.verizon.net)
11.58.46 #       <speachy> otoh, drive 0 is always "internal" (the exact string varies by target) and drive 1 is in a similar situation.
11.59.46 #       <user890104> i've updated the Bulgarian translation in g#5830
11.59.49 #       <rb-bluebot> 3Gerrit review #5830 at https://gerrit.rockbox.org/r/c/rockbox/+/5830 : 3Update Bulgarian translation by Vencislav Atanasov
11.59.54 #       <speachy> sweet
12.00.18 #       <user890104> it's a mix of two or three styles, so it needs a general rework in order to be pleasing to read/hear
12.01.25 #       <user890104> i wouldn't use it personally, but check if the most annoying things can be fixed, there are genders on adjectives in Bulgarian, and some strings are reused regardless of it, so it sounds strange
12.03.18 #       <speachy> hmm, that sort of thing is going to bite other languages too
12.03.40 Quit    mrkrisprolls (Killed (NickServ (GHOST command used by mrkrispr`!~user@h-85-24-169-71.A192.priv.bahnhof.se)))
12.05.34 #       <speachy> want to get added to the notification list should this translation need updates later?
12.06.14 #       <rb-bluebot> Build Server message: 3New build round started. Revision ba4a6f2adc, 337 builds, 6 clients.
12.06.14 #       <rb-bluebot> 3Update Bulgarian translation by Vencislav Atanasov
12.06.40 #       <speachy> I'll add it to the nightly voice list as well.  just espeak, since there's no piper model.
12.16.47 #       <speachy> ok, that's also committed, along with a couple of mechanical corrections to the lang file.  31 issues remain, nearly all of which are of the "identical to English" variety.  One in particular (VOICE_LANG_INVALID_VOICE_FILE) definitely needs to corrected.
12.17.46 Join    mskrisprolls [0] (mrkrisprol@2605:6400:20:5cc:e8af:da1c:ee53:f678)
12.23.07 #       <speachy> __builtin: oh, do we need an updated puzzles font to go with that resync?
12.29.10 #       <rb-bluebot> Build Server message: 3Build round completed after 1377 seconds.
12.29.11 #       <rb-bluebot> Build Server message: 3Revision ba4a6f2adc result: All green
12.29.14 #       <rb-bluebot> Build Server message: 3New build round started. Revision a153f6c266, 337 builds, 7 clients.
12.29.15 #       <rb-bluebot> 3voice:  Add defaults for Bulgarian voicing, and add it to the nightly builds. by Solomon Peachy
12.32.13 #       <user890104> speachy: feel free to add me to the list
12.37.59 #       <user890104> also, which email address does this list use? i haven't been subscribed to the rockbox mailing list for a couple of years
12.47.16 #       <rb-bluebot> Build Server message: 3Build round completed after 1083 seconds.
12.47.17 #       <rb-bluebot> Build Server message: 3Revision a153f6c266 result: All green
13.00.56 #       <user890104> speachy: is this message on the translation site still relevant?
13.01.03 #       <user890104> Note: This is updated at most every 15 minutes, but only if the build farm is idle.
13.01.20 #       <user890104> my changes are still not shown there
13.03.48 #       <speachy> user890104: it'll be whatever you want it to be.  I could just use the address in your gerrit commits?
13.04.57 #       <speachy> and yes, that message is still relevant.  the site updates on the quarter-hour after the build farm goes idle.
13.05.06 #       <user890104> that should also be fine, but I'll PM you my preferred address
13.08.56 #       <user890104> also, there was some problem with the POST max size of https://translate.rockbox.org/submit.php, because the last few translated strings were lost when I submitted the page
13.13.06 #       <speachy> oh, interesting, thanks for that. 
13.14.00 #       <speachy> ...ran into the max_input_vars limit (defaults to 1000)
13.14.15 #       <user890104> i was close :)
13.14.35 #       <speachy> changed that to 5000, ought to do the trick.
13.15.05 #       <speachy> one of those little gotchas..
13.15.15 #       <user890104> also, there's no easy way to save the progress midway, or load a partially translated patch. but that would bring additional complications
13.15.49 #       <user890104> i trusted my laptop not to crash between sleep/wakeup a couple of times, and it worked in the end
13.17.01 #       <speachy> there's no good way to download/supply a work-in-process.  I guess the original intnt was that translators wouldn't let things get so far out of date.  :(
13.17.27 #       <speachy> at least with the infrastructure currently partial translations are still worth committing as the "same as english" stuff properly flags these as needing to be looked at.
13.18.18 #       <speachy> MAybe I could add a checkbox at the top that would tell it to supply the full .lang file instead of just the diff?
13.18.23 #       <user890104> i see, there also isn't an obvious way to run a non-commited translation file through the website checks
13.18.54 #       <speachy> yeah, and the mechanism that does exist isn't as strict as what the main tooling uses.
13.19.28 #       <speachy> ...all that is on the todo (eventually) list.
13.19.32 #       <user890104> I would do it as two separate buttons at the bottom, Download patch for submission, and Download complete file
13.19.44 #       <speachy> good call.
13.20.19 #       <user890104> well, as a PHP developer, I can easily make a bunch of improvements
13.22.41 #       <user890104> I would make the input file selectable between the master branch, a specific commit and a user-provided file, so that manual or partial translations can be checked and improved
13.23.25 #       <speachy> feel free to make any improvements you see fit
13.23.58 #       <speachy> it's not fair to say that I loathe PHP, but I'm usually very non-motivated to work on it.  :D
13.26.40 #       <speachy> Hmm.  Master or user-provided file only, I think.
13.29.40 #       <user890104> fair enough
13.33.20 #       <user890104> g#5831 has the changes beyond input variable 1000
13.33.24 #       <rb-bluebot> 3Gerrit review #5831 at https://gerrit.rockbox.org/r/c/rockbox/+/5831 : 3Update Bulgarian translation, part 2 by Vencislav Atanasov
13.34.00 #       <user890104> also the Italian is missing just a single voice file in order to be at 100%
13.35.03 #       <speachy> yep, it's something that needs someone that knows the right way to correct it.
13.41.18 #       <speachy> I hacked together a full/patch submit thing.  
13.43.53 #       <rb-bluebot> Build Server message: 3New build round started. Revision dd58b05f61, 337 builds, 7 clients.
13.43.53 #       <rb-bluebot> 3Update Bulgarian translation, part 2 by Vencislav Atanasov
13.45.52 #       <user890104> how do I generate the voice files locally? is is a configure/make option?
13.49.53 #       <speachy> yeah, advanced/voice
13.50.01 #       <speachy> or you can use rbutil
13.50.26 #       <speachy> could do it completely manually if you wanted, but this is far simpler.  :D
13.51.46 #       <user890104> I found tools/voice.pl but I'm unsure how to use it
13.52.20 #       <user890104> also there seems to be a typo in gtts_lang_map
13.52.37 #       <user890104>     'bulgarian' => 'l bg',
13.52.37 #       <user890104> it looks like this needs a dash before l
13.53.06 #       <speachy> oops.  clearly that's untested.  :D
13.53.37 #       <speachy> it's a little convoluted, set up a voice build using configure, then run make V=1 voice and it will show you the various command line invocations
13.53.50 #       <user890104> i see, thanks
13.54.06 #       <user890104> pushing to prod like CrowdStrike last week
13.54.43 #       <speachy> I'm not sure if anyone other than me (and the build infra) has run this stuff with anything other than the default settings.  
13.56.00 ***     Saving seen data "./dancer.seen"
13.56.33 #       <user890104> well, the fact that the script is in Perl speaks for itself
13.58.07 #       <speachy> I think it's more a reflection of the user/developer base than choice of implementation language.  non-technical users either use rbutil or download a pre-built voice.
13.58.52 #       <rb-bluebot> Build Server message: 3Build round completed after 899 seconds.
13.58.53 #       <rb-bluebot> Build Server message: 3Revision dd58b05f61 result: All green
13.58.57 #       <rb-bluebot> Build Server message: 3New build round started. Revision 6dd9b9786c, 337 builds, 7 clients.
13.58.57 #       <rb-bluebot> 3voice: Fix typo in Bulgarian gtts default cmdline by Solomon Peachy
13.59.08 #       <user890104> yes, but when making a translation, you need to be able to hear if it sounds well before submitting
13.59.12 #       <speachy> ...and even then hardly any feedback about how well this stuff works.  Before I started hacking on it so heavily it was clear that it was barely working even for just english.  :D
13.59.28 #       <user890104> ouch
14.01.24 #       <speachy> yeah.  now we're up to... 16 voice sets being generated nightly, and the overall translation (and localization) state has never been better.
14.02.24 #       <speachy> granted nobody knows how widely any of this stuff is used...
14.03.05 #       <user890104> at some point during the weekend I was wondering how  good of an idea would be to have a container image with the toolchains prebuilt, and running CI in GitHub
14.05.03 #       <speachy> various folks have created containers/VMs with the toolchains present, but they're never kept maintained.
14.06.24 #       <user890104> also there's 2k minutes per month per account of free CPU time for CI
14.07.03 #       <speachy> that would be exhausted in days.
14.08.29 #       <speachy> 337 separate build targets (that will be upped by another 30-ish once I add the rest of the manuals)
14.10.01 #       <user890104> ah, rbutil needs a player connected in order to generate the voice files (or do anything whatsoever)
14.10.31 #       <speachy> GitHub CI won't give us any _new_ capabilities and make it any easier for random other folks to meaningfully contribute
14.10.53 #       <user890104> well, it can save the old artifacts
14.12.47 #       <speachy> just the ones we keep amount to ... ~4.6GB/day
14.13.26 #       <speachy> and ~80GB for our current ~3wk retention.
14.15.23 #       <speachy> (with the additional voices added recently that's going to be nearly 100GB...)
14.15.33 #       <rb-bluebot> Build Server message: 3Build round completed after 997 seconds.
14.15.34 #       <rb-bluebot> Build Server message: 3Revision 6dd9b9786c result: All green
14.35.28 #       <user890104> speachy: can you please fix the last two errors on the Bulgarian voice strings manually, by prefixing them with a ~? I wonder how I missed these...
14.56.53 #       <speachy> will do
15.48.16 #       <rb-bluebot> Build Server message: 3New build round started. Revision 8d436561e2, 337 builds, 7 clients.
15.48.16 #       <rb-bluebot> 3plugins: playing time: use cpu_boost + simplify a bit by Christian Soffke
15.56.01 ***     No seen item changed, no save performed.
16.03.28 #       <rb-bluebot> Build Server message: 3Build round completed after 912 seconds.
16.03.29 #       <rb-bluebot> Build Server message: 3Revision 8d436561e2 result: All green
16.03.33 #       <rb-bluebot> Build Server message: 3New build round started. Revision 254eaf509e, 337 builds, 7 clients.
16.03.33 #       <rb-bluebot> 3Fix the final two Bulgarian translation same-as-English issues by Solomon Peachy
16.05.03 #       <speachy> ouch, the bulgarian translation bumped our language table size up by over 2KB.  gotta love multi-byte encodings.
16.16.56 #       <rb-bluebot> Build Server message: 3Build round completed after 804 seconds.
16.16.57 #       <rb-bluebot> Build Server message: 3Revision 254eaf509e result: All green
16.45.55 Join    lebellium [0] (~lebellium@2a01cb0405d07f00d33a0da62bfc2206.ipv6.abo.wanadoo.fr)
17.56.02 ***     No seen item changed, no save performed.
18.11.39 Quit    lebellium (Quit: Leaving)
18.19.59 Quit    IPG (Ping timeout: 260 seconds)
18.35.53 #       <rb-bluebot> Build Server message: 3New build round started. Revision c0ac043c6d, 337 builds, 6 clients.
18.35.54 #       <rb-bluebot> 3get_volume_name generate volume names once then reuse by William Wilgus
18.49.20 #       <_bilgus> So far so good on data aborts on USB unplug I hope thats the last of them, I think I made them more likely by exiting the file browser since there was no longer something to fallback to to take its place when the redraw event loop ran again
18.50.16 #       <_bilgus> not that it was guaranteed anyway..
18.55.38 #       <speachy> _bilgus: I have a feeling this is going to generate a pile of red.
18.56.13 #       <_bilgus> shouldn't
18.56.15 Join    Moriar [0] (~moriar@107-200-193-159.lightspeed.stlsmo.sbcglobal.net)
18.56.24 #       <speachy> no wait, MULTIDRIVE ==> MULTIVOLUME These days
18.56.39 #       <speachy> (after that last massive sea of red debacle of mine)
18.56.52 #       <_bilgus> but I do always worry when this stuff runs way long
18.57.18 #       <speachy> our build cycles are roughly 2x longer due to losing b0h00n's beefy builders
18.57.53 #       <_bilgus> oh wasn't aware
18.58.02 #       <speachy> but changing headers that half the world depends on means ccache isn't as effective
18.58.49 #       <rb-bluebot> Build Server message: 3Build round completed after 1376 seconds.
18.58.50 #       <rb-bluebot> Build Server message: 3Revision c0ac043c6d result: All green
18.58.54 #       <rb-bluebot> Build Server message: 3New build round started. Revision 96f564fe64, 337 builds, 6 clients.
18.58.54 #       <_bilgus> woot
18.58.54 #       <rb-bluebot> 3manual: update single mode by Christian Soffke
18.59.23 #       <speachy> glad to be wrong!
19.00.14 #       <speachy> would be nice to have other builders capable of manual builds
19.00.39 #       <speachy> those don't parallelize to additional cores, unlike everything else.  
19.01.42 #       <_bilgus> so dir create is ~3% faster diropen 33% dirscan 12% dirscan w/ info 6% and delete 11%
19.03.49 #       <_bilgus> I didn't see anything on the file create but its probably in the wash
19.03.49 #       <speachy> nice.
19.03.49 #       <speachy> gonna look into making the volume string voiced finally too.
19.03.49 #       <_bilgus> I should try a few tag cache builds the two changes I've made (not copying file path, and this) should equate to a few minutes off the nearly hour build time
19.04.42 #       <_bilgus> I won't argue, lmk if you need anything
19.18.28 #       <rb-bluebot> Build Server message: 3Build round completed after 1175 seconds.
19.18.29 #       <rb-bluebot> Build Server message: 3Revision 96f564fe64 result: All green
19.19.47 #       <speachy> animal-petur also dropped (along with the b0hoon's three)
19.20.02 #       <speachy> so there are some cold ccaches in there
19.36.08 Join    massiveH [0] (~massiveH@2600:4040:a982:dc00:5d20:47ea:b433:b34e)
19.56.03 ***     Saving seen data "./dancer.seen"
20.43.55 #       <__builtin> speachy: no, I don't believe the font pack needs updating
20.56.07 #       <rb-bluebot> Build Server message: 3New build round started. Revision cd1922fe74, 337 builds, 6 clients.
20.56.07 #       <rb-bluebot> 3puzzles: update documentation and copyright year. by Franklin Wei
21.13.16 #       <rb-bluebot> Build Server message: 3Build round completed after 1030 seconds.
21.13.17 #       <rb-bluebot> Build Server message: 3Revision cd1922fe74 result: All green
21.37.53 Join    dconrad [0] (~dconrad@152.117.104.217)
21.38.54 #       <dconrad> _bilgus: thanks for the lead, I'll look into that
21.39.44 #       <dconrad> not quite sure where to run it, perhaps from the button checking code
21.40.14 Quit    Bobathan_ (Quit: ZNC 1.8.2+deb2+b1 - https://znc.in)
21.47.54 Join    Bobathan_ [0] (~admin@syn-065-029-248-157.res.spectrum.com)
21.55.42 #       <_bilgus> idk thats probably fine there are a few always running queue/threads you cn pick
21.56.05 ***     Saving seen data "./dancer.seen"
21.57.05 Quit    Bobathan_ (Quit: ZNC 1.8.2+deb2+b1 - https://znc.in)
21.57.20 #       <_bilgus> you could also hook the track skip stuff and grab the time remaining and set up a timer to hit in x seconds to check again i'm not sure how perfect thatd'd be though 
21.58.10 #       <_bilgus> in theory it'd probably be the least hairy with edge cases
22.03.44 Quit    acidsys (Ping timeout: 264 seconds)
22.10.34 Join    Bobathan_ [0] (~admin@syn-065-029-248-157.res.spectrum.com)
22.16.37 #       <rb-bluebot> Build Server message: 3New build round started. Revision 89b4cd59ef, 337 builds, 6 clients.
22.16.37 #       <rb-bluebot> 3[BugFix] extra slashes in recording directory by William Wilgus
22.21.54 Join    acidsys [0] (~crameleon@openSUSE/member/crameleon)
22.25.29 Quit    Moriar (Ping timeout: 255 seconds)
22.33.00 #       <rb-bluebot> Build Server message: 3Build round completed after 984 seconds.
22.33.01 #       <rb-bluebot> Build Server message: 3Revision 89b4cd59ef result: All green
22.47.17 Quit    bluebrother (Ping timeout: 248 seconds)
22.47.39 Quit    rb-bluebot (Ping timeout: 276 seconds)
22.48.58 Join    bluebrother [0] (~dom@user/bluebrother)
23.00.06 Join    rb-bluebot [0] (~rb-bluebo@rockbox/bot/utility)
23.02.03 Quit    massiveH (Quit: Leaving)
23.47.13 #       <dconrad> _bilgus: for the time being I chose to hook into the playback.c start/stop/pause functions, that gets rid of my ~4 second delay
23.48.02 #       <dconrad> so that's cool, though now it actually works too fast haha, in some circumstances it doesn't wait for the fade out to finish.... I'm sure it's an easy oversight on my part but I'm happy with some progress for the night
23.49.20 #       <dconrad> and it still feels like a pretty hacky solution but I'm sure I can clean it up
23.56.06 ***     Saving seen data "./dancer.seen"
23.58.24 Quit    dconrad ()