| 00:02:31 | _bilgus | hmm |
| 00:03:59 | _bilgus | shutdowns are now taking forever :/ |
| 00:04:30 | _bilgus | I'm going to have to figure out a better way |
| 01:00 |
| 01:17:32 | *** | No seen item changed, no save performed. |
| 01:28:09 | rb-bluebot | Build Server message: New build round started. Revision fe5c1c8126, 303 builds, 10 clients. |
| 01:28:54 | _bilgus | I'm reverting this for now, it appears to be a red herring as its clearly calling for boost and something else is killing that boost |
| 01:29:24 | _bilgus | we have something thats maybe over clearing the boost counter |
| 01:39:50 | rb-bluebot | Build Server message: Build round completed after 702 seconds. |
| 01:39:51 | rb-bluebot | Build Server message: Revision fe5c1c8126 result: All green |
| 01:53:20 | | Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:edd2:a638:e532:7c27) |
| 01:57:43 | | Quit ZincAlloy (Ping timeout: 252 seconds) |
| 02:00 |
| 02:03:01 | | Quit MarcAndersen (Ping timeout: 245 seconds) |
| 03:00 |
| 03:17:36 | *** | Saving seen data "./dancer.seen" |
| 03:53:29 | | Quit emacsomancer (Ping timeout: 264 seconds) |
| 03:57:09 | | Join emacsomancer [0] (~emacsoman@136.60.128.68) |
| 05:00 |
| 05:17:40 | *** | Saving seen data "./dancer.seen" |
| 05:33:13 | | Quit j-r (Remote host closed the connection) |
| 05:34:35 | | Join j-r [0] (~j-r@p2003000623793827404207fffefd0a65.dip0.t-ipconnect.de) |
| 06:00 |
| 06:01:00 | | Join MarcAndersen [0] (~no_znepna@85.218.172.116) |
| 06:17:27 | speachy | yay, nightly stuff all ran properly. |
| 06:29:57 | | Join tomato [0] (~tomato@user/tomato) |
| 07:00 |
| 07:17:41 | *** | Saving seen data "./dancer.seen" |
| 07:48:49 | | Join cockroach [0] (~blattodea@user/cockroach) |
| 07:57:35 | braewoods | _bilgus: maybe logging calls to the boost would be helpful, especially if you can identify the caller |
| 08:00 |
| 08:18:03 | speachy | mailing list rsync script isn't working properly... le sigh. |
| 08:29:42 | edhelas | scripts are meant to break, this is necessary to keep admins busy |
| 08:58:33 | | Join massiveH [0] (~massiveH@ool-18e4e82f.dyn.optonline.net) |
| 09:00 |
| 09:17:42 | *** | No seen item changed, no save performed. |
| 09:21:27 | | Join unmanbearpig [0] (~unmanbear@user/unmanbearpig) |
| 10:00 |
| 10:20:42 | | Quit cockroach (Quit: leaving) |
| 10:25:28 | | Quit massiveH (Quit: Leaving) |
| 10:31:19 | _bilgus | braewoods, yes we have boost logging i'll have to make a build doing so |
| 11:00 |
| 11:12:41 | spork | is this 'jonus' forum user one of those bots/trolls or more human ? |
| 11:13:32 | braewoods | definitely a bot |
| 11:13:55 | braewoods | they include random links to places that make no sense |
| 11:14:04 | spork | the questions seem almost legit |
| 11:14:21 | speachy | they've been known to scrape previous questions |
| 11:14:27 | braewoods | well, if they were legit, why would they have random links that aren't even topic related? |
| 11:15:09 | spork | that is why i asked |
| 11:15:23 | spork | scraping old questions is a good explanation |
| 11:15:39 | spork | much better than generating new ones |
| 11:15:52 | braewoods | either that or they secretly created an AI and they've enslaved it to create more plausible spam |
| 11:15:59 | braewoods | :D |
| 11:16:28 | speachy | another possibility is a browser with a malicious plugin, but the latter question is nonsensical on its face. |
| 11:17:44 | *** | Saving seen data "./dancer.seen" |
| 11:18:29 | speachy | banned the user, all posts removed. |
| 11:28:04 | | Quit _bilgus (Ping timeout: 265 seconds) |
| 11:41:10 | | Join cockroach [0] (~blattodea@user/cockroach) |
| 11:45:18 | | Join bilgus_ph [0] (~bilgus_ph@172.58.187.95) |
| 11:45:23 | | Part edhelas |
| 11:45:42 | | Join edhelas [0] (9d94237298@v2202101139504140605.quicksrv.de) |
| 11:46:03 | bilgus_ph | There is probably still a lot of posts that s d |
| 11:46:37 | bilgus_ph | That sound plausible then later edited to include those links |
| 11:47:28 | bilgus_ph | I catch them in the changes log occasionally I'm sure there are plenty more lying in wait |
| 11:49:42 | | Quit bilgus_ph (Client Quit) |
| 12:00 |
| 12:01:50 | speachy | this is why we can't have nice things... |
| 12:42:16 | | Quit akaWolf (Ping timeout: 252 seconds) |
| 13:00 |
| 13:00:40 | | Join _bilgus [0] (~bilgus@162.154.213.134) |
| 13:05:10 | | Join akaWolf [0] (~akaWolf@akawolf.org) |
| 13:17:47 | *** | Saving seen data "./dancer.seen" |
| 13:23:23 | | Join lebellium [0] (~lebellium@2a01cb04012c0900017145434e411ad5.ipv6.abo.wanadoo.fr) |
| 13:48:02 | MarcAndersen | Why isn't there a battery time indicator on the gigabeat s? It just says battery 94% and no time? |
| 14:00 |
| 14:02:26 | braewoods | speachy: https://xkcd.com/810/ :D |
| 14:19:54 | | Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:74fc:ee4:aefd:3fca) |
| 15:00 |
| 15:17:49 | *** | No seen item changed, no save performed. |
| 16:00 |
| 16:28:59 | speachy | MarcAndersen: I'd like to add a working example of the voiced number reworking, I think you said you wanted Dutch? What should I put in as the tens separator? (eg "three hundred five AND twenty") |
| 16:35:12 | spork | danish |
| 16:48:37 | MarcAndersen | speachy: I don't exactly know what you mean. |
| 16:48:59 | speachy | I need to know what word to put into the language file |
| 16:49:05 | speachy | (ie instead of the english AND) |
| 16:49:13 | MarcAndersen | og |
| 16:49:26 | speachy | and that's danish, as spork indicated? |
| 16:49:36 | MarcAndersen | It sure is |
| 16:49:44 | speachy | awesome, thank you |
| 16:52:06 | speachy | If you're willing to give this a test, what target should I build the firmware and voice files for? |
| 16:53:09 | MarcAndersen | C200, clip zip, nano 1g, gigabeat s or xduoo x3 |
| 16:54:03 | _bilgus | MarcAndersen, not too familar with the gigabeast but I'm guessing the battery time is either a setting or perhaps a lack of data for the calculation? |
| 16:58:32 | speachy | generating now. |
| 17:00 |
| 17:04:17 | speachy | MarcAndersen: https://www.shaftnet.org/~pizza/rb-x3-experiment.zip |
| 17:04:31 | speachy | I'd recommend backing up your existing .rockbox directory just in case |
| 17:04:47 | speachy | but this has the updated code (for the X3), and english + danish voice files. |
| 17:05:18 | speachy | (Danish is only 75% translated, so I make no promises...) |
| 17:08:47 | speachy | It's what I used, though I mangled the english-us voice so I could make sure the logic worked. |
| 17:14:07 | MarcAndersen | We have some errors: when it should say twenty, it says nothing, and at thirty it says "and thirty" as well as "and forty" and so on. |
| 17:17:08 | speachy | does (eg) 1-20, 21, 31, etc work? |
| 17:17:38 | MarcAndersen | 1 to 19 does, 20 is silent |
| 17:17:52 | *** | No seen item changed, no save performed. |
| 17:17:53 | MarcAndersen | 21 to 39 and 31 to 39 and so on works |
| 17:17:54 | speachy | ok, found the error with 20. |
| 17:18:27 | MarcAndersen | and at 30, 40, 50 you just need to remove the "and" at the beginning. |
| 17:18:38 | speachy | and .. I think I have the "and.." error too |
| 17:19:18 | MarcAndersen | Hey is there a way for me to set which sd it boots from? that way I could use one for the experiment. |
| 17:19:56 | speachy | no.. it'll boot from SD1 if both are populated. |
| 17:20:03 | speachy | otherwise it'll boot from whichever |
| 17:20:08 | MarcAndersen | Ok thanks |
| 17:21:03 | MarcAndersen | Does your one also make some clicks from the hardware itself when you turn it on and off? |
| 17:21:13 | speachy | yes |
| 17:21:22 | MarcAndersen | What's that, lights? |
| 17:22:13 | speachy | relays for the power supply, I think. |
| 17:22:30 | speachy | https://www.shaftnet.org/~pizza/rb-x3-experiment2.zip |
| 17:22:33 | MarcAndersen | I've never had a player that did that before |
| 17:22:35 | speachy | safe to unzip over the old one |
| 17:22:41 | speachy | over the first one |
| 17:26:15 | MarcAndersen | speachy: I think that's it. |
| 17:28:11 | speachy | ok, gonna commit it now. |
| 17:28:15 | rb-bluebot | Build Server message: New build round started. Revision 70e72e01d2, 303 builds, 11 clients. |
| 17:28:30 | MarcAndersen | Commit? wow |
| 17:29:03 | MarcAndersen | I just got the email |
| 17:29:43 | braewoods | speachy: my worker won't be contributing as much this time. i'm building chromium. |
| 17:31:19 | MarcAndersen | My friend who got my old classic will be excited about this |
| 17:35:03 | speachy | Danish voice files aren't part of the nightly set still, due to <75% coverage |
| 17:35:32 | speachy | but all other translations that need to use this new rule need to be updated appropriately. |
| 17:35:48 | MarcAndersen | Did you use the make voice command then in the experiment? |
| 17:36:14 | speachy | yes |
| 17:36:37 | MarcAndersen | Ok. I'm going to git pull now and try it out. |
| 17:37:19 | yang | speachy: where can I translate the audio files ? |
| 17:37:28 | speachy | yang: https://translate.rockbox.org/ |
| 17:38:03 | yang | cool |
| 17:38:06 | speachy | once this build round is finished I'll check to see if I need to tweak the translate site backend to handle the special case.. |
| 17:38:15 | speachy | but that won't interfere with anything else. |
| 17:40:39 | yang | speachy: so I don't neeed to record my language audio files? I only neeed to translate words/letters into my language by typing ? |
| 17:40:54 | | Quit ZincAlloy (Quit: Leaving.) |
| 17:41:04 | speachy | just written, yes. we use text-to-speech systems to synthesize things. |
| 17:41:09 | yang | ok nice |
| 17:41:24 | speachy | of course one could just record a large pile of sounds and assemble that up but it's hard to justify that level of effort. |
| 17:42:14 | MarcAndersen | I actually wanted to try it at some point, but gave up again. |
| 17:42:29 | yang | I did it for Wikipedia |
| 17:42:38 | rb-bluebot | Build Server message: Build round completed after 862 seconds. |
| 17:42:39 | rb-bluebot | Build Server message: Revision 70e72e01d2 result: All green |
| 17:42:48 | MarcAndersen | yang: Spoken articles? |
| 17:42:51 | speachy | I don't know offhand how well the exiting tooling handles the manual recording scenario |
| 17:42:52 | yang | yes |
| 17:43:10 | MarcAndersen | Yang: which ones? |
| 17:43:16 | yang | Slovenian towns |
| 17:43:46 | MarcAndersen | Nice |
| 17:43:46 | yang | so people can pronounce our towns |
| 17:43:49 | yang | yes |
| 17:44:30 | MarcAndersen | I actually make my own talk clips with rbspeexenc and my own voice |
| 17:44:46 | yang | it was a bit time consuming, but actually once you have the setup done, it's easy to record and do multiple words and save audio files, I mean it could even be faster with some scripting tools |
| 17:44:51 | MarcAndersen | So the normal voice file would be awesome too. |
| 17:45:26 | MarcAndersen | I use sox for almost all the editing |
| 17:45:43 | yang | I figured that I can record multiple names of the towns into a large .wav and then just cut the sound out with audacity and transcode them to .oga |
| 17:45:54 | yang | MarcAndersen: did you do it for PBX? |
| 17:45:58 | MarcAndersen | Oh, audacity is cool too |
| 17:46:16 | MarcAndersen | what's pbx? |
| 17:46:26 | yang | PBX VoIP like Asterisk or similar |
| 17:46:34 | yang | for guided menus |
| 17:47:01 | MarcAndersen | No, only marcs talking clock, my own little thing for windows |
| 17:47:31 | yang | ok |
| 17:47:49 | yang | it's midnight here, I better go to sleep |
| 17:47:56 | yang | good night |
| 17:47:59 | MarcAndersen | I think I'm getting a bit off topic here as well |
| 17:49:24 | yang | MarcAndersen: I even bought some more advanced studio microphone aand even a "stand" for microphone, but later I didn't proceed with recordings anymore....too much other stuff to do |
| 17:49:28 | | Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:74fc:ee4:aefd:3fca) |
| 17:49:48 | MarcAndersen | Yang: I see |
| 17:50:07 | yang | I don't do video blogs like on Youtube or such |
| 17:50:17 | yang | was o nly doing files for wikipedia |
| 17:51:01 | yang | I think if wikipedia had some kind of funding |
| 17:51:14 | yang | so the people could earn while submitting |
| 17:51:24 | yang | a lot more work could be done |
| 17:51:38 | yang | on the other side...anyone can contribute in free time |
| 17:51:40 | speachy | ok, fixed the translate site, now VOICE_NUMERIC_TENS_SWAP_SEPARATOR can be specified. |
| 17:51:43 | yang | but it's quite adictive |
| 17:52:03 | | Quit ZincAlloy (Client Quit) |
| 17:53:06 | yang | ok GN, afk |
| 17:54:22 | MarcAndersen | I forgot something, but it's not urgent: Is there a way to make the date said as date month year instead of month date year? |
| 17:55:09 | speachy | MarcAndersen: I don't know, but I'll look into it. |
| 17:55:28 | MarcAndersen | speachy: Thanks |
| 17:57:14 | speachy | can you give me an example where it speaks dates? |
| 17:57:30 | MarcAndersen | The rockbox info screen |
| 17:57:39 | speachy | ok, thanks |
| 17:58:01 | MarcAndersen | And if you hold down select in the time and date settings menu |
| 17:59:44 | | Quit lebellium (Quit: Leaving) |
| 18:00 |
| 18:04:42 | speachy | there are mutiple places where different logic is used for speaking dates |
| 18:04:44 | speachy | (lovely..) |
| 18:05:11 | speachy | but the good news is that this should be relatively easy to augment. |
| 18:06:35 | braewoods | speachy: maybe something akin to strftime would help |
| 18:07:03 | speachy | yep |
| 18:09:46 | | Join Moriar [0] (~moriar@107-200-193-159.lightspeed.stlsmo.sbcglobal.net) |
| 18:11:14 | speachy | I'll only implement enough to read out the full year, numeric and month name, and day. |
| 18:59:44 | | Join smithjd [0] (~enderw@node-1w7jra28yypte4b1rejm4lqiw.ipv6.telus.net) |
| 19:00 |
| 19:02:46 | smithjd | I'm having trouble building the sim for the fioo m3k. |
| 19:03:31 | smithjd | buffering.c:174:13: error: variably modified ‘buffering_stack’ at file scope |
| 19:04:50 | smithjd | i just upgraded my compiler and got rid of some old versions of gcc and clang. i tried restoring the old versions of gcc, no change. |
| 19:05:42 | smithjd | gcc ver: GCCVER=11 |
| 19:06:18 | smithjd | gcc num: GCCNUM=1102 |
| 19:06:41 | smithjd | i also tried trimming down the compile options. still no compile. |
| 19:06:45 | smithjd | any suggestions? |
| 19:07:37 | smithjd | p.s. also voice_thread.c:93:13: error: variably modified ‘voice_stack’ at file scope |
| 19:08:27 | MarcAndersen | smithjd: I'm compiling an m3k sim now to test |
| 19:08:53 | MarcAndersen | Hmm could the voice thing be remains from the reverted voice thread commit from yesterday? |
| 19:11:34 | MarcAndersen | I just compiled an m3k sim on ubuntu |
| 19:17:54 | *** | Saving seen data "./dancer.seen" |
| 19:19:43 | smithjd | what compiler? |
| 19:20:29 | MarcAndersen | gcc (Ubuntu 10.3.0-1ubuntu1) 10.3.0 |
| 19:22:11 | smithjd | thanks |
| 19:27:52 | _bilgus | nah but speaking of I need a boost logging build lets see how this goes |
| 19:40:20 | | Quit JanC (Remote host closed the connection) |
| 19:40:33 | | Join JanC [0] (~janc@user/janc) |
| 19:48:34 | smithjd | still no luck :( |
| 19:57:06 | _bilgus | smithjd you might wanna nuke everything and start over |
| 19:57:42 | _bilgus | delete the whole directory if make clean doesn't help |
| 19:58:43 | smithjd | just tried. same compile error. |
| 19:59:30 | smithjd | commit 67716c6b causes a panic on my player, though it worked flawlessly in the simulator. i don't have a sim build with the correct patch applied and my toolchain (at least for the sim build) doesn't compile master anymore, so I can't verify this. |
| 20:00 |
| 20:01:33 | smithjd | so i'm fixing this up without the simulator. i've got it partially working, but it loses the fd for the playlist currently viewed, which also never happened in the simulator. |
| 20:02:06 | smithjd | <* this aso only happens when the player is playing. |
| 20:10:49 | _bilgus | smithjd, the simulator works fine for some things but even if you are running something like the address sanitizer the sim is just no substitute for certain situations |
| 20:11:13 | speachy | smithjd: did you pick the native or hosted m3k when you ran configure? |
| 20:11:19 | smithjd | native |
| 20:11:41 | speachy | just had a clean m3k sim build with git master, gcc 11.2.1 |
| 20:12:11 | smithjd | tried the linux build, same error. |
| 20:12:33 | speachy | so something is janky with your tree or build evironment |
| 20:13:32 | speachy | .. I mean, the build farm has built everything fine, and while I can't speak for the m3k specifically quite a few of us have used builds since 67716c6b on various players... |
| 20:14:22 | | Quit _bilgus (Quit: Leaving) |
| 20:14:43 | smithjd | i had also installed the device-specific compiler. i tried moving the compiler outside my path, still nothing. |
| 20:17:33 | speachy | what's in the generated rockbox-info.txt? |
| 20:18:41 | | Join _bilgus [0] (~bilgus@162.154.213.134) |
| 20:19:19 | | Quit MarcAndersen (Quit: I was using NightOwl 0.2.) |
| 20:19:38 | _bilgus | well the boost log sucks i need to add a way to dump it to disk or trim the mile long director path list |
| 20:19:46 | _bilgus | directory* |
| 20:20:14 | smithjd | speachy: doesn't get that far. |
| 20:20:19 | speachy | (well, if you're moving away the cross-compiler and it's still "building" then there's sometihng _very_ wonky indeed... |
| 20:23:02 | smithjd | that looks like a fairly mundane gcc code error, anyway. is anyone else using gcc 11.2.1 with any success? |
| 20:23:09 | smithjd | how would i compile with llvm? |
| 20:27:37 | smithjd | (or can it be done?) |
| 20:27:42 | speachy | at least two of the builders are using gcc 11, and one is using gcc 10. However the cross compilers are all gcc 4.9.4 |
| 20:27:50 | braewoods | we don't really support CLANG. |
| 20:28:10 | speachy | we haven't tried to use clang, but there's no inherent reason why it shouldn't work for simulator builds. |
| 20:29:10 | speachy | we will eventually move to newer cross compilers but we still haven't knocked out all the bugs from the last bump. |
| 20:34:03 | smithjd | any idea why the playlist viewer loses the fd for the playlist while in the loop? |
| 20:53:12 | _bilgus | stack overflow? |
| 20:57:33 | _bilgus | mmm smithjd I recently added extra error checking in that area so its probably something that has been a problem.. |
| 21:00 |
| 21:04:32 | smithjd | when the properties view of an unrelated playlist is exited while a track is playing, i get the error 'accessing playlist control file', and then after repeated presses of the back button 'error accessing playlist file' which goes away back to the first error after a timeout. the view is stuck at the error message until a track change, at which the the screen is redrawn with multiple entries of the same track overlaid by the |
| 21:04:33 | smithjd | first message. everything works fine if there is no playback ocurring, OR the properties plugin was opened onto the currently playling playlist, provided a redraw is issued after the properties window is dismissed. |
| 21:16:31 | _bilgus | does error accessing playlist error go away on its own if you don't press any buttons? |
| 21:17:56 | *** | Saving seen data "./dancer.seen" |
| 21:19:44 | smithjd | only 'error accessing playlist file' |
| 21:24:25 | _bilgus | odd that the control file one shows, is this after a usb plug event or on a COLD BOOT? |
| 21:24:42 | smithjd | doesn't seem to matter |
| 21:30:02 | _bilgus | ah you recently changed g#3824 so i assume you never had this option prior? |
| 21:30:05 | rb-bluebot | Gerrit review #3824 at https://gerrit.rockbox.org/r/c/rockbox/+/3824 : Playlist Viewer: Add file properties to playlist entry menu. by James D. Smith |
| 21:31:47 | _bilgus | likely its got some assumption that isn't immediately obvious |
| 21:33:56 | smithjd | correct. iirc everything (inc. during playback) worked in the sim build. i can't verify this right now. there was a build fix https://gerrit.rockbox.org/r/c/rockbox/+/3827 because the sim build allowed defining variables in-switch which the cross-compilers apparently didn't like . |
| 21:34:53 | smithjd | g#3827 |
| 21:34:56 | rb-bluebot | Gerrit review #3827 at https://gerrit.rockbox.org/r/c/rockbox/+/3827 : Fix the build for 67716c6b. by James D. Smith |
| 21:37:14 | smithjd | perhaps someone could try to reproduce this in the simulator? that would be a big help. |
| 21:43:26 | _bilgus | might also want to check playlist_get_track_info() for the return value |
| 21:44:25 | _bilgus | ill try it on the fuze+ |
| 21:44:27 | _bilgus | sim |
| 21:50:16 | _bilgus | smithjd do i just file browse to the playlist and open it? |
| 21:50:25 | _bilgus | view excuse me |
| 21:51:04 | smithjd | yep. view and then there's the new properties item per entry. try with and without playing back a different playlist. |
| 21:56:09 | _bilgus | it works fine as far as I can tell while playing |
| 21:56:28 | _bilgus | no issues with current or other playlist |
| 21:56:35 | _bilgus | ill try with no music |
| 21:57:58 | _bilgus | still fine with music stopped |
| 21:58:09 | _bilgus | you said you issue was in the m3k sim? |
| 21:58:20 | smithjd | no, everything's fine in the sim |
| 21:58:41 | smithjd | i just can't build a sim build. |
| 21:59:27 | smithjd | you went into the main menu to playlist catalogue, not the currently playing playlist menu, correct? |
| 21:59:41 | smithjd | to a different playlist and opened the viewer? |
| 21:59:49 | _bilgus | no i went through file browser to playlist dir |
| 21:59:56 | smithjd | perfect. |
| 22:00 |
| 22:00:14 | smithjd | thanks very much. |
| 22:00:27 | _bilgus | sure.. |
| 22:01:41 | smithjd | now the next question i guess is whether to revert this for now, or patch it to the point where it works mostly except for during playback. |
| 22:02:25 | rb-bluebot | Build Server message: New build round started. Revision eeacffbd15, 303 builds, 10 clients. |
| 22:05:09 | | Quit Moriar (Quit: Leaving.) |
| 22:06:36 | speachy | what's building allows per-language overriding of the voiced date strings |
| 22:06:48 | speachy | making displayable dates localized will be a bit more work though |
| 22:07:24 | speachy | I'm sure after a few more of these thigns I'll have to go back and rework some stuff. |
| 22:12:47 | _bilgus | up to you whats going to be cleaner? |
| 22:12:51 | _bilgus | smithjd |
| 22:14:07 | _bilgus | MarcAndersen on the boost unboost during voice the reason I didn't hear it the first times I tried is that there are other threads boosting the device (dircache) so that masks the issue |
| 22:14:22 | _bilgus | until the cache is built |
| 22:14:35 | _bilgus | then it becomes v. apparent |
| 22:15:12 | smithjd | i'd hate to leave it knowing it's in a broken state. i also wonder if it's only broken on the m3k. |
| 22:16:19 | _bilgus | and I see lots of churn boost unboost between the button code and the voice thread but it doesn't appear to by in an improper state although I wonder if some might be due to the refresh rate of the loop being faster than the time it takes to do all this speaking |
| 22:16:52 | _bilgus | smithjd we always try to leave things in a working state in case you get hit by a bus tomorrow |
| 22:18:08 | rb-bluebot | Build Server message: Build round completed after 942 seconds. |
| 22:18:09 | rb-bluebot | Build Server message: Revision eeacffbd15 result: All green |
| 22:20:14 | smithjd | @bilgus: do you run a recent dev build on your fuze+? |
| 22:23:24 | _bilgus | no the device is currently in a box its usually in my truck but I haven't driven it for a few months |
| 22:23:49 | _bilgus | I use the clip zip daily though and its usually a few days to weeks behind |
| 22:24:26 | smithjd | can you try the steps you followed in the simulator on your clip zip? |
| 22:26:05 | smithjd | if it's not too much trouble... |
| 22:27:52 | _bilgus | yeah its got a logf build currently but its at head |
| 22:29:31 | _bilgus | smithjd, works well |
| 22:29:44 | smithjd | even during playback? |
| 22:29:55 | _bilgus | it was a bit slow to load the properties but thats all I noticed |
| 22:30:08 | _bilgus | yeah only during pb and with voice |
| 22:30:22 | smithjd | ok, well no sense reverting this then. |
| 22:30:48 | _bilgus | I do see a bug |
| 22:31:11 | _bilgus | yeah when it returns the lit is wiped out |
| 22:31:46 | _bilgus | and got a data abort |
| 22:31:53 | _bilgus | let me try again |
| 22:32:00 | | Join dconrad [0] (~dconrad@208.38.228.17) |
| 22:32:19 | | Quit cockroach (Quit: leaving) |
| 22:32:37 | _bilgus | works fine when playing |
| 22:32:47 | _bilgus | its when its not playing thatit appears |
| 22:33:37 | _bilgus | yeah it falls apart if I stop the music and then do it (different playlist too) |
| 22:34:20 | _bilgus | same if I start from a non playing state cold boot |
| 22:34:55 | _bilgus | it returns a mostly blank list with a few bits of data and a bunch of the no char unicode boxes |
| 22:35:12 | smithjd | i need to call playlist_buffer_load_entries_screen() for the fiio hardware to avoid a panic. |
| 22:35:14 | _bilgus | I think yu have a stack ovfl most likely and i'm seeing data |
| 22:35:49 | _bilgus | one of them says 0B I bet the playlist viewer is blowing the stack |
| 22:35:56 | _bilgus | sorry properties |
| 22:36:02 | _bilgus | properties viewer |
| 22:37:40 | _bilgus | is the properties viewer inbuilt to core or is it a plugin? |
| 22:37:48 | smithjd | plugin |
| 22:40:42 | _bilgus | haha it has a crap ton of stuff maxpath size |
| 22:41:02 | _bilgus | but i think that should be staticly allocd to the plugin buffer |
| 22:43:39 | _bilgus | I don't see anything else on the heap though |
| 22:46:05 | _bilgus | yeah I'm not seeing anything that jumps out |
| 22:47:47 | _bilgus | smithjd, the playlist viewer is using the plugin buffer |
| 22:48:06 | _bilgus | that bit me a few weeks ago |
| 22:48:22 | _bilgus | I hate that shit core shouldn't get the plugin buffer :/ |
| 22:49:29 | _bilgus | https://github.com/Rockbox/rockbox/blob/master/apps/playlist_viewer.c#L333 |
| 22:50:09 | _bilgus | one option would be to svae state dump the buffer and start over when you return |
| 22:53:03 | smithjd | playlist_buffer_load_entries_screen() seems to work for this. i wonder what's causing my error during playback on the m3k? |
| 22:54:27 | _bilgus | it just happens to overwrite something else |
| 22:56:30 | _bilgus | If you manually invoked it I could see that working basically saving state |
| 22:57:41 | _bilgus | IDk it looks like it has logic for when something is playing |
| 22:58:35 | _bilgus | also I think i'd try and figure out a way to leave the buffer in a clean state rather than trying to smear it back over the top |
| 23:00 |
| 23:00:27 | _bilgus | beyond that the plugin buffer is used when playing (if at all possible) so you might be running into some other bug too |
| 23:01:18 | _bilgus | like the plugin buffer improperly overlapping something in the (crt?) file |
| 23:01:52 | _bilgus | or it could be a improper aligned address even |
| 23:03:30 | smithjd | master works correctly on my agptek rocker. i don't use it so it slipped my mind to check with. |
| 23:03:40 | smithjd | no redraw issues. |
| 23:05:18 | _bilgus | thats just because the plugin buffer is bigger |
| 23:05:40 | smithjd | than the m3k? |
| 23:05:56 | _bilgus | no than my clipzip |
| 23:06:58 | _bilgus | im not sure the m3k one is the same, maybe but ATM that code is sharing the buffer and its kinda last one wins?\ |
| 23:07:25 | smithjd | could be. |
| 23:07:36 | _bilgus | oh no question |
| 23:07:44 | _bilgus | line 333 shows it |
| 23:08:47 | _bilgus | and I recieved the formatted string in the list pretty much told me where to look |
| 23:09:39 | _bilgus | now your work around of reloading the buffer sounds plausible |
| 23:10:28 | smithjd | yes, very frustrating to find working and non-working players with the same code. |
| 23:11:06 | smithjd | the m3k is a very young port tho. |
| 23:11:42 | smithjd | would you like an updated patch with the playlist reoad? |
| 23:12:35 | _bilgus | sure I'll test it if ya got it |
| 23:14:00 | smithjd | do yuo want a build or just the patch? |
| 23:14:20 | _bilgus | put it on gerrit i can build it |
| 23:14:51 | _bilgus | I just found where the clipzip is under boosting |
| 23:14:53 | smithjd | ok, i'm just goingto verify the build. |
| 23:16:02 | _bilgus | 0.) B 0 rockbox/apps/main.c:405 |
| 23:16:02 | _bilgus | 1.) U 1 rockbox/firmware/target/arm/as3525/usb-as3525.c:61 <the usb code unboosts without first boosting |
| 23:16:18 | _bilgus | wonder if amachronics patch as this |
| 23:17:57 | *** | Saving seen data "./dancer.seen" |
| 23:23:08 | | Quit rb-bluebot (Ping timeout: 252 seconds) |
| 23:24:05 | | Quit bluebrother (Ping timeout: 264 seconds) |
| 23:24:42 | smithjd | @bilgus: https://gerrit.rockbox.org/r/c/rockbox/+/3849 |
| 23:25:26 | | Join bluebrother [0] (~dom@user/bluebrother) |
| 23:25:28 | _bilgus | give me a few |
| 23:37:04 | | Join rb-bluebot [0] (~rb-bluebo@rockbox/bot/utility) |
| 23:49:44 | _bilgus | ok compiling now |
| 23:50:26 | _bilgus | mu only concern is that the entry prior could still get corrupted I imagine |
| 23:57:23 | _bilgus | smithjd, yes it appear to work |
| 23:57:37 | smithjd | great. |
| 23:58:13 | _bilgus | error accessimg playlist control file on exit |
| 23:58:27 | _bilgus | I wonder if its using the plugin buffer too lol |
| 23:58:32 | smithjd | yes, that's what i got during playback. |
| 23:58:57 | _bilgus | I wasn't getting that before |
icon identifies that the person is a core developer (has commit access).