--- Log for 21.09.121  Server: silver.libera.chat  Channel: #rockbox
--- Nick: rb-logbot  Version: Dancer V4.16  Started: 18 days and 14 hours ago

01.45.08 ***     No seen item changed, no save performed.
03.45.10 ***     No seen item changed, no save performed.
05.25.17 Join    advcomp2019_ [0] (~advcomp20@user/advcomp2019)
05.27.25 Quit    advcomp2019 (Ping timeout: 252 seconds)
05.45.14 ***     Saving seen data "./dancer.seen"
06.53.52 #       <speachy> ok, updated build scores committed.  We should see better overall efficiency and perhaps more fairness in handing things out.
07.45.18 ***     No seen item changed, no save performed.
07.46.51 Join    MarcAndersen [0] (~no_znepna@193.169.154.231)
07.48.06 #       <MarcAndersen> Hi. On my xduoo x3 I can only get one of the sd slots mounted at once, is this a bug or am I doing something wrong?
07.48.44 Join    johnb2 [0] (~johnb2@p5b3af8b1.dip0.t-ipconnect.de)
07.49.45 #       <johnb2> MarcAndersen: I have the same. This seems to happen on Windows only, I have been told.
07.54.24 #       <MarcAndersen> Oh. I hope it can be fixed at some point though, but thanks for clearing it up.
07.54.30 Quit    johnb2 (Ping timeout: 265 seconds)
08.00.52 #       <braewoods> MarcAndersen: welcome back.
08.01.06 #       <braewoods> i take it you also bought an x3?
08.04.55 #       <MarcAndersen> Yes.
08.28.44 #       <speachy> MarcAndersen: It _should_ work, but it's not something I've personally checked in quite some time
08.30.52 #       <MarcAndersen> It only mounts Rockbox Internal Storage USB Device, not the sd
08.32.39 #       <speachy> mine is in the car right now, I'lll grab it later today and check things while waiting for another interminable $dayjob build cycle
08.33.41 #       <MarcAndersen> What is a interminable $dayjob build cycle
09.11.27 Join    massiveH [0] (~massiveH@ool-18e4e82f.dyn.optonline.net)
09.24.56 Quit    speachy (Quit: WeeChat 3.2)
09.25.38 Join    speachy [0] (~speachy@rockbox/developer/speachy)
09.25.39 Mode    "#rockbox +v speachy" by ChanServ (ChanServ@services.libera.chat)
09.26.31 #       <speachy> MarcAndersen: it takes about 6 hours for a complete build.  
09.27.32 #       <speachy> Also, I just plugged two SD cards into my X3, and hooked it up to my PC, and both SD cards appeared.
09.28.05 #       <speachy> 210831 daily build, plugged into a USB3 hub connected to a Fedora 34 system.
09.29.02 #       <MarcAndersen> It's on windows I think, like johnb2 said
09.29.06 #       <speachy> (also using an experimental bootloader but that has no effect on things)
09.29.21 #       <speachy> I did notice that the Linux USB stack says it's applying some sort of quirk to our USB ID
09.29.41 #       <MarcAndersen> Is it easy to fix?
09.36.03 #       <speachy> The quirk is US_FL_FIX_INQUIRY, which overrides our INQUIRY response.
09.36.45 #       <MarcAndersen> And what does that mean exactly?
09.36.59 #       <speachy> I don't know if that has any affect on Windows.  (And why it would manifest itself as one slot appearing but not the other)
09.37.45 #       <speachy> INQUIRY is essentially just the vendor/product identification
09.38.19 #       <MarcAndersen> I only get Rockbox Internal Storage USB Device
09.38.31 #       <MarcAndersen> and not sd
09.40.49 #       <speachy> that would indicate that windows is only querying the first LUN, for whatever reason
09.41.18 #       <MarcAndersen> but on all my other sandisks both devices shows up
09.43.33 #       <speachy> the usb mass storage code is the same either way
09.44.56 #       <braewoods> speachy: doesn't rockbox allow you to hide one of the devices
09.45.04 #       <braewoods> that might be a viable workaround
09.45.07 #       <speachy> not that I'm aware of
09.45.12 #       <MarcAndersen> Would another windows user be able to test this at some point?
09.45.18 #       <braewoods> my x3 had an option for "hide internal storage"
09.45.20 ***     Saving seen data "./dancer.seen"
09.45.38 #       <speachy> windows 10?
09.45.43 #       <braewoods> probably. i have a rock pi x set aside just for testing windows USB stuff
09.45.59 #       <speachy> I mean, you can always just not have both SD cards plugged in simultaneously...
09.46.21 #       <braewoods> it should still be checked out speachy
09.46.23 #       <braewoods> but yea
09.46.36 #       <braewoods> see if there's any solutions to be had
09.47.26 #       <MarcAndersen> What doesn't make sense to me is that it has always worked on my sansas.
09.47.42 #       <speachy> I have a Windows 7 VM at home I use for printer driver tests.  but that's not ideal for low-level USB trickery due to the passthrough nature of the VMs.
09.47.52 #       <braewoods> it could be an X3 related issue.
09.48.08 #       <braewoods> it sounds like a hardware quirk
09.48.20 #       <MarcAndersen> sure
09.48.57 #       <speachy> the difference with the sansas is that the two storage devices are on different ... well, devices. one is internal flash, the other is SD.  X3 has both on SD.  but our usb stack (and mass storage code) doesn't care.
09.49.51 #       <braewoods> yea and one that can be papered over if we get working MTP at some point
09.51.26 #       <braewoods> speachy: did you see this https://forums.rockbox.org/index.php/topic,54005.msg249272.html#msg249272
09.51.32 #       <braewoods> i wasn't sure how to advise them
09.53.50 #       <speachy> braewoods:  they need to be in DFU mode for itunes restore to work
09.54.11 #       <braewoods> it sounded like their drive might be having issues
09.54.15 #       <speachy> but that aside, manually partition the player and see what happens.
09.54.39 #       <speachy> if they're in "rockbox bootloader mode" then I don't know if the disk writes can be trusted to be robust
09.54.47 #       <speachy> (with our last released bootloaders)
09.55.44 Quit    michaelni (Remote host closed the connection)
09.56.22 #       <speachy> MarcAndersen:  we'd need to see usb sniffs to make sense of what's going on. 
09.56.58 #       <speachy> you can do that yourself, but you'll need to install wireshark (or at least usbpcap) 
09.58.11 Join    michaelni [0] (~michael@213-47-68-29.cable.dynamic.surfer.at)
10.01.32 #       <MarcAndersen> I installed wireshark now. What do I need to do in there?
10.01.52 #       <speachy> you have to follow the usbpap insructions
10.02.14 #       <speachy> https://desowin.org/usbpcap/tour.html
10.03.00 #       <speachy> and when you've identified what you need to sniff, you can remove and re-insert the device
10.03.23 #       <speachy> stop the capture, send us the file
10.09.50 Quit    bluebrother (Ping timeout: 265 seconds)
10.09.59 Join    bluebrother [0] (~dom@user/bluebrother)
10.11.18 Join    berber8 [0] (~berber@v2202101107577140883.nicesrv.de)
10.13.22 Quit    vup (Ping timeout: 252 seconds)
10.13.22 Quit    rasher (Ping timeout: 252 seconds)
10.13.22 Quit    berber (Ping timeout: 252 seconds)
10.13.22 Nick    berber8 is now known as berber (~berber@v2202101107577140883.nicesrv.de)
10.13.31 Join    rasher [0] (~rasher@user/rasher)
10.13.39 Join    vup [0] (~~~~@46.101.193.235)
10.14.22 #       <MarcAndersen> I'm not sure if I did it right, but could you check this: https://www.dropbox.com/s/mche8ba7pw5ekep/test.pcap?dl=1
10.16.34 #       <speachy> it doesn't look like there's anything "interesting" in that capture -- just looks like idle traffic on a hub without the X3 plugged in.
10.20.35 #       <MarcAndersen> Try this one: https://www.dropbox.com/s/24u1ktt536aqsev/test2.pcap?dl=1
10.22.45 #       <speachy> looks about the same
10.23.29 #       <MarcAndersen> What am I doing wrong? And why do we need this anyway if we already know it's the lun?
10.24.34 #       <speachy> if we're lucky we'll see something going wrong (or there being something out-of-spec) in the handshake
10.24.54 #       <speachy> but we don't know what (or why) Windows is deciding there's only one LUN worth talking to
10.25.32 #       <MarcAndersen> No, but can we do something about it at all, or should we forget about it entirely?
10.26.12 #       <speachy> I'm asusming we can, but we don't know what it is that needs doing yet.  Windows is mostly a black box, but like you said the Sansas work so it's presumably due to something on our end.
10.27.09 #       <speachy> as for what oyu're doing wrong, my guess is that you're capuring on the wrong usb "bus" (vs what the X3 is plugged into)
10.28.04 #       <MarcAndersen> The first time I tried hub 1 and that was empty, but hub 2 showed something here in test2.pcap, didn't it?
10.30.19 #       <speachy> it wsa the same chatter between the hub and the host, nothing else
10.31.10 #       <MarcAndersen> Then I don't know. Maybe another windows user can step in, but just take your time, If it gets fixed that's great, and if not, it's not that big a deal.
10.32.46 Quit    massiveH (Quit: Leaving)
10.34.37 #       <speachy> oh, one other question -- you said sansas work, did you them with roughly the same daily/nightly build as you're using on the X3?  (Just to make sure it's not a general regression)
10.35.14 #       <MarcAndersen> No, 3.15. I will try my c250 with daily.
10.40.16 #       <MarcAndersen> It works on c250. Both drives.
10.40.39 #       <speachy> ok.
10.40.43 #       <speachy> good data point, thanks
10.41.40 #       <MarcAndersen> But now I have an invalid voice. I still hate that commit with the voice files, could we revert it?
10.43.02 #       <speachy> there were a lot of hidden issues that were caused by voice (and language) files being out of sync with the core.
10.43.37 #       <MarcAndersen> Oh
10.44.08 #       <MarcAndersen> I like the new espeak voices much better that the old festival though
10.45.39 #       <speachy> it's also a _lot_ faster to generate.  I'm open to having things tweaked but we needed a good usable baseline
10.51.06 #       <MarcAndersen> Can you put a danish espeak in? That would be good because my rockbox utility sapi thing is broke on my main machine
10.54.48 #       <speachy> The threshold for adding more auto-generated voices was the completness of he translation.
10.55.13 #       <speachy> danish is only at ~75%
10.55.33 #       <speachy> I think I had the threshold for voice generation at 95%
10.55.53 #       <speachy> https://translate.rockbox.org/
10.56.31 #       <MarcAndersen> Oh, we also need the number speaking fixed first so it says five and twenty instead of twenty five, for example.
10.57.05 #       <MarcAndersen> Until that's done there is really no need for it anyway
11.13.39 Quit    tchan (Ping timeout: 265 seconds)
11.14.08 Join    tchan [0] (~tchan@c-98-206-141-238.hsd1.il.comcast.net)
11.29.00 #       <speachy> as the saying goes, perfection is the enemy of good.    Numeric ordering is a problem, but having a quarter of the UI strings untranslated is probably a bigger deal..  Alas, my language skills are limited to English and Bad English.
11.39.58 #       <speachy> oh, speaking of numeric voicing.  would "525000" be spoken as "five hundred twenty-five thousand" or "five hundred five and twenty thousand"
11.45.23 ***     Saving seen data "./dancer.seen"
11.48.53 #       <MarcAndersen> the last one
11.49.28 #       <MarcAndersen> I think what we need to do is reverse the ones and the tens every time
11.49.37 #       <speachy> so "625531" would be "six hundred five and twenty thousand five hundred one and thirty"
11.50.00 #       <MarcAndersen> Yes
11.50.01 #       <speachy> that does simplify things; we already group numbers in threes, 
11.50.20 #       <MarcAndersen> I do as well in my talking clock
11.50.22 #       <speachy> there's just no way to signifgy in a language file that this ordering matters
11.50.30 #       <MarcAndersen> Oh
11.50.39 #       <speachy> well, that's fixable too.  :)
11.51.21 #       <MarcAndersen> I have it set up so that it looks for a specific file, and if it exists, it plays it one way, and if it doesn't it does it another way
11.51.48 #       <MarcAndersen> But that is not ideal in rockbox at all
11.54.05 #       <MarcAndersen> Something I am a bit concerned with is when we say 2021, should it be twenty one and twenty or two thousand one and twenty?
11.54.33 #       <speachy> dates don't follow the same numerical rules (in general, not speaking about rockbox specifically)
11.54.52 #       <MarcAndersen> Ok
11.56.36 #       <speachy> but RB currently reads it out numerically if it's < 1100 or >= 2000
11.57.14 #       <MarcAndersen> I see
11.57.21 #       <speachy> if it's in the middle it splits out the century (1931 => 19  31  and reads out each separately as numbers)
11.58.37 #       <speachy> there's a ton of variation here, and... should we say "1931 as "nineteen one and thirty" for Danish?
11.58.49 #       <speachy> or one thousand nine hundred one and thirty
12.00.24 #       <speachy> or nineteen thirty
12.00.43 #       <speachy> (and repeat these questions for every language / locale)
12.01.47 #       <MarcAndersen> The first one for the year, and the second one for normal numbers
12.09.46 #       <braewoods> speachy: i prefer to read numbers myself but obvious not possible here :|
12.49.06 Join    ZincAlloy [0] (~Adium@2a02:8108:943f:d824:2484:efb9:530a:91a)
13.45.26 ***     No seen item changed, no save performed.
13.53.04 Join    johnb2 [0] (~johnb2@p5b3af8b1.dip0.t-ipconnect.de)
13.53.06 Quit    johnb2 (Client Quit)
14.02.57 Join    lebellium [0] (~lebellium@lfbn-idf2-1-1386-25.w92-169.abo.wanadoo.fr)
15.45.27 ***     Saving seen data "./dancer.seen"
17.13.58 Join    Moriar [0] (~moriar@107-200-193-159.lightspeed.stlsmo.sbcglobal.net)
17.14.50 Quit    lebellium (Quit: Leaving)
17.16.06 Quit    speachy (Quit: Connection closed)
17.43.39 Quit    Piece_Maker (Read error: Connection reset by peer)
17.44.45 Join    Piece_Maker [0] (~Piece_Mak@cpc95746-bolt17-2-0-cust360.10-3.cable.virginm.net)
17.45.28 ***     Saving seen data "./dancer.seen"
17.52.52 Quit    ZincAlloy (Quit: Leaving.)
17.54.22 Join    ZincAlloy [0] (~Adium@2a02:8108:943f:d824:3972:34eb:ec77:40bc)
17.55.25 Quit    ZincAlloy (Client Quit)
17.57.53 Quit    Piece_Maker (Quit: ZNC 1.8.2 - https://znc.in)
17.58.17 Join    Piece_Maker [0] (~Piece_Mak@cpc95746-bolt17-2-0-cust360.10-3.cable.virginm.net)
18.03.55 Quit    Piece_Maker (Quit: ZNC 1.8.2 - https://znc.in)
18.06.18 Join    Piece_Maker [0] (~Piece_Mak@cpc95746-bolt17-2-0-cust360.10-3.cable.virginm.net)
18.09.05 Quit    rb-bluebot (Ping timeout: 264 seconds)
18.09.41 Quit    bluebrother (Ping timeout: 264 seconds)
18.21.22 Join    bluebrother [0] (~dom@user/bluebrother)
18.22.41 Join    rb-bluebot [0] (~rb-bluebo@rockbox/bot/utility)
18.37.21 Quit    ufdm_ (Quit: Leaving)
18.37.33 Join    ufdm [0] (~ufdm@c-73-164-63-214.hsd1.mn.comcast.net)
19.45.31 ***     Saving seen data "./dancer.seen"
19.49.19 Quit    JanC (Remote host closed the connection)
19.49.32 Join    JanC [0] (~janc@user/janc)
20.19.00 Join    speachy [0] (~speachy@taster.shaftnet.org)
20.19.00 Quit    speachy (Changing host)
20.19.00 Join    speachy [0] (~speachy@rockbox/developer/speachy)
20.19.00 Mode    "#rockbox +v speachy" by ChanServ (ChanServ@services.libera.chat)
20.48.25 #       <speachy> so I'd appreciate some comments on g#3832
20.48.28 #       <rb-bluebot> 3Gerrit review #3832 at https://gerrit.rockbox.org/r/c/rockbox/+/3832 : 3talk:  Add support for languages that swap the tens position in numbers by Solomon Peachy
20.49.11 #       <speachy> one issue I need to fix still (marked in the code) but the basic approach is correct, I think,
21.06.06 #       <speachy> assuming that there aren't more ways to handle numeric readback
21.19.24 Join    dconrad [0] (~dconrad@208.38.228.17)
21.30.14 Quit    dconrad (Remote host closed the connection)
21.35.07 Join    dconrad [0] (~dconrad@208.38.228.17)
21.45.34 ***     Saving seen data "./dancer.seen"
21.56.37 Join    amiconn_ [0] (jens@p200300ea870b0f00305e95fffec66ff3.dip0.t-ipconnect.de)
21.56.37 Quit    amiconn (Killed (platinum.libera.chat (Nickname regained by services)))
21.56.37 Nick    amiconn_ is now known as amiconn (jens@p200300ea870b0f00305e95fffec66ff3.dip0.t-ipconnect.de)
21.57.05 Quit    pixelma (Ping timeout: 264 seconds)
21.57.14 Join    pixelma [0] (marianne@p200300ea870b0f00305e95fffec66ff3.dip0.t-ipconnect.de)
22.17.29 Quit    Moriar (Ping timeout: 264 seconds)
23.10.53 Quit    dconrad ()
23.45.36 ***     Saving seen data "./dancer.seen"