--- Log for 24.01.119 Server: orwell.freenode.net Channel: #rockbox --- Nick: logbot Version: Dancer V4.16 Started: 1 month and 27 days ago 00.20.58 Quit jhMikeS (Ping timeout: 268 seconds) 00.22.57 # is there a reason it just removes all files when it errors 00.29.00 Quit this_is_a_nick (Remote host closed the connection) 00.40.08 Quit krabador (Remote host closed the connection) 00.43.02 Quit ZincAlloy (Quit: Leaving.) 01.03.14 Join this_is_a_nick [0] (~amofiuhr_@152.231.181.72) 01.04.48 Join krabador [0] (~krabador@unaffiliated/krabador) 01.19.11 # okay I found a possible solution in https://www.rockbox.org/tracker/task/13156 01.21.54 # so I changed that line in the talkfile.cpp, but I can't seem to build the rockboxutility. I get a lot of 'undefined reference to `typeinfo for CryptoPP:*' 01.22.15 # I did install libcrypto++-dev 01.25.08 # collect2: error: ld returned 1 exit status 01.25.08 # Makefile:373: recipe for target 'RockboxUtility' failed 01.25.08 # make: *** [RockboxUtility] Error 1 01.29.55 Join jhMikeS [0] (~jethead71@d192-24-173-177.try.wideopenwest.com) 01.35.06 *** Saving seen data "./dancer.seen" 01.41.58 Quit CH23 (Ping timeout: 250 seconds) 01.49.53 Join amofiuhr_ [0] (~amofiuhr_@152.231.181.72) 01.50.02 Quit this_is_a_nick (Ping timeout: 272 seconds) 01.59.32 Quit amofiuhr_ (Ping timeout: 272 seconds) 02.09.43 Quit krabador (Remote host closed the connection) 02.12.44 Join amofiuhr_ [0] (~amofiuhr_@ip167-132-50-179.ct.co.cr) 02.26.08 Quit amofiuhr_ (Ping timeout: 272 seconds) 02.36.55 Join amofiuhr_ [0] (~amofiuhr_@152.231.181.72) 03.23.26 Join this_is_a_nick [0] (~amofiuhr_@152.231.181.72) 03.24.24 Quit amofiuhr_ (Ping timeout: 272 seconds) 03.35.09 *** Saving seen data "./dancer.seen" 03.52.53 Quit pystar89 (Ping timeout: 250 seconds) 03.54.20 Quit jhMikeS (Ping timeout: 268 seconds) 04.09.41 Join englishm [0] (~englishm@unaffiliated/englishm) 04.31.10 # <_Bilgus> CH23 Its fuzzy for me but I remember there being something funky with libcrypto the last time I set it up 05.23.25 Quit TheSeven (Ping timeout: 240 seconds) 05.23.58 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven) 05.35.13 *** Saving seen data "./dancer.seen" 06.28.01 Quit englishm (Quit: My MacBook has gone to sleep. ZZZzzz…) 06.41.51 Join pystar89 [0] (~pystar89@ip5f590e6b.dynamic.kabel-deutschland.de) 07.04.51 Join CH23 [0] (~CH23@d83131.upc-d.chello.nl) 07.05.47 Quit CH23 (Client Quit) 07.07.59 Quit Huntereb (Ping timeout: 272 seconds) 07.18.37 Join Huntereb [0] (~Huntereb@d-69-161-102-241.cpe.metrocast.net) 07.19.54 Join CH23 [0] (~CH23@d83131.upc-d.chello.nl) 07.21.03 # _Bilgus, do you have any idea how I could solve this? am I missing some other dependencies maybe? 07.22.35 # <_Bilgus> are you using ubuntu? 07.35.06 Quit this_is_a_nick (Read error: Connection reset by peer) 07.35.18 *** Saving seen data "./dancer.seen" 07.40.24 # <_Bilgus> ch23 the dependency is from imxboot IIRC https://github.com/Rockbox/rockbox/blob/master/rbutil/mkimxboot/Makefile 07.41.30 # <_Bilgus> I don't remember if I altered the makefile to add the include directory or if I edited the source this was a year or two ago 07.44.49 # <_Bilgus> it might even have been adding the proper name libcrypto++6# or something 07.47.49 Join this_is_a_nick [0] (~amofiuhr_@ip140-237-64-186.ct.co.cr) 07.48.30 # so this does not happen to everyone? 07.58.17 # <_Bilgus> Probably? I don't have it installed on this current machine to tell you eactly what it was if you haven't gotten an answer or figured it out by tomorrow ill install qt ans attempt to build it so I can tell you for sure 08.24.54 Quit JohnyPsychac (Quit: Page closed) 08.35.29 Quit pixelma (Quit: .) 08.35.29 Quit amiconn (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) 08.38.27 Join amiconn [0] (jens@rockbox/developer/amiconn) 08.38.27 Join pixelma [0] (marianne@rockbox/staff/pixelma) 08.46.42 # If the problem is with me, not with the utility, then i'll look into it. I doubt I could fix it otherwise. 08.51.55 Join petur [0] (~petur@80.169.83.226) 08.51.55 Quit petur (Changing host) 08.51.55 Join petur [0] (~petur@rockbox/developer/petur) 08.52.02 # either way i'll let you know. 08.53.25 Quit dys (Ping timeout: 240 seconds) 09.05.59 Join TheSphinX_ [0] (~briehl@srv001.nosupamu.de) 09.07.53 Quit TheSphinX^ (Ping timeout: 250 seconds) 09.07.53 Nick TheSphinX_ is now known as TheSphinX^ (~briehl@srv001.nosupamu.de) 09.35.19 *** Saving seen data "./dancer.seen" 10.27.20 # CH23: with questions about building the RockboxUtility you could try to get bluebrother's attention, I guess. He's still logged in here but I haven't seen him talking in a very long while but he was involved with coding this thing 10.32.32 # ouch. his last message was 2017/04/29 15:51 I'll try to find some time for the zlib / cryptopp issues 11.35.23 *** No seen item changed, no save performed. 11.46.47 # Did anyone happen to see my message the other day about long OPUS files having a delay while resuming? 11.49.40 # aphirst: check the IRC logs on the site, I believe Bilgus fixed it for you 11.53.56 # CH23_, ah for some reason i never got that message 11.54.04 # maybe my ZNC server was playing up 11.54.17 # _Bilgus, i'll test out a new build 12.05.08 # that seemed to fix it! 12.05.47 # was it http://gerrit.rockbox.org/r/#/c/2058/ ? 12.06.04 # resuming opus audiobook content now works great 12.06.12 # 0.2s instead of almost 10 12.10.38 # ah wait, it's not quite right though 12.10.41 # now all opus tracks start at a few seconds in 12.10.55 # and refues to play the first few seconds even if i rewind them 12.11.33 # ok not all files, but definitely my lower bitrate ones 12.13.04 # it doesn't seem to do it for my 160kbps music but for my 48kbps audiobooks they all start a couple seconds in 12.13.26 # i can share some sample files if it would help? 12.35.52 Quit michaelni (Ping timeout: 245 seconds) 12.48.20 Join michaelni [0] (~michael@213-47-41-20.cable.dynamic.surfer.at) 12.51.53 Join vmx [0] (~vmx@p5B2A2956.dip0.t-ipconnect.de) 13.06.30 # aphirst, a sample file or two would go a long way. 13.13.04 # CH23_, the problem I'm seeing building mkimxboot is that libcryptopp requires libpthread but it's not included in its pkgconfig file. manually adding it to the dependency list allows the build to proceed. 13.14.25 # mknwzboot has cryptopp issues too... 13.28.38 # speachy: i wish i had remote access to my laptop, i'll try this tonight. 13.32.33 # CH23_: See this patch: http://gerrit.rockbox.org/r/#/c/2061/ 13.33.03 # fixes the compile on my system, at least. 13.33.24 # one of the joys is that crypto++ is called different things on different systems. 13.33.35 # so that patch may have broken it for others 13.34.24 # i believe it's (lib)crypto++ on my distro 13.35.25 *** Saving seen data "./dancer.seen" 13.37.16 # header files are /usr/inc/crypto++/ ? 13.37.35 # or rather, /usr/include/crypto++/ 13.40.53 Join Ruhan [0] (uid76353@gateway/web/irccloud.com/x-flfylpiwvrykczkb) 14.10.32 Join massiveH [0] (~massiveH@ool-18e4e27c.dyn.optonline.net) 14.17.51 Join jhMikeS [0] (~jethead71@d192-24-173-177.try.wideopenwest.com) 14.34.46 Join pamaury [0] (~pamaury@rockbox/developer/pamaury) 14.55.51 Quit GeekShadow (Ping timeout: 250 seconds) 15.23.26 Join GeekShadow [0] (~antoine@nzf.turmel.info) 15.23.26 Quit GeekShadow (Changing host) 15.23.26 Join GeekShadow [0] (~antoine@reactos/tester/GeekShadow) 15.35.27 *** Saving seen data "./dancer.seen" 15.43.04 Quit massiveH (Quit: Leaving) 15.47.32 # <_Bilgus> aphirst some samples of the failing files will be helpful I'm guessing the issue is not having enough "reference" material to get the stream normalized 15.49.13 Quit this_is_a_nick (Remote host closed the connection) 15.53.00 # only in obd 15.53.02 # oops 16.14.10 Quit petur (Quit: Connection reset by beer) 16.14.22 Quit Rower- (Ping timeout: 246 seconds) 16.16.27 Join Rower [0] (~husvagn@m83-190-1-24.cust.tele2.se) 17.10.37 Quit Ruhan (Quit: Connection closed for inactivity) 17.35.31 *** Saving seen data "./dancer.seen" 17.35.44 Quit vmx (Remote host closed the connection) 18.08.49 Join Ruhan [0] (uid76353@gateway/web/irccloud.com/x-zbiuimnfnvwfiulj) 18.30.22 Quit pamaury (Quit: Konversation terminated!) 18.48.40 Join thomasjfox [0] (~thomasjfo@rockbox/developer/thomasjfox) 19.07.49 Join smoke_fumus [0] (~smoke_fum@188.35.176.90) 19.18.27 Quit smoke_fumus (Ping timeout: 250 seconds) 19.25.57 Join this_is_a_nick [0] (~amofiuhr_@201.206.191.130) 19.35.35 *** Saving seen data "./dancer.seen" 19.36.40 Join dys [0] (~dys@tmo-109-1.customers.d1-online.com) 19.44.02 # speachy, i have both crypto++ and cryptopp in /usr/include/ 19.46.28 # huh. interesting. 19.46.34 # cryptopp is a link to crypto++ 19.48.41 # do you have any idea how i should add cryptopp to the rbutilqt/Makefile 19.50.21 # I updated my build patch for the components. Haven't got far enough to build rbutil 19.50.59 # would adding LDOPTS += -lpthread be enough? 19.52.31 # rbutilqt uses qmake to generate the makefile, so it needs to be added elsewhere.. 19.53.14 # building... 19.53.15 # ... 19.55.11 # I needed to add '-lcryptopp' to the end of LIBS in the generated makefile. 19.56.08 # been a long time since I tried to grok the qmake stuff. 19.58.10 # it works! 19.58.18 # so easy 20.00.43 # what's the distro you're using? 20.00.52 # almost. i'm getting some errors, i'll see if that's just because of me being dumb 20.01.03 # Trisquel GNU/Linux version 8 20.01.13 # #parabola 20.08.53 # i'm getting a few warnings during build, but nothing that breaks it 20.08.56 # g#2061 updated. 20.09.24 # that fixes all of the build issues (at least on my system) relating to libcrypto++ 20.20.45 Join smoke_fumus [0] (~smoke_fum@188.35.176.90) 20.22.10 Join krabador [0] (~krabador@unaffiliated/krabador) 20.30.45 # sadly it seems the issue why i wanted to build it in the first place was not fixed by this 20.32.20 # while generating voicefiles, every time during the encoding stage from wave to mp3, i get an error and it halts: 20.32.31 # [encoderrbspeex.cpp:82 INFO] Encoding "/tmp/rbvoice//LANG_5.wav" to "/tmp/rbvoice//LANG_5.mp3" 20.32.31 # [encoderrbspeex.cpp:103 ERROR] Error: invalid WAV file 20.33.15 # i searched online and found https://www.rockbox.org/tracker/task/13156 20.33.43 # the comment mentions changing talkfile.cpp, and so i did. still i get the error 20.35.01 # if the tool could do this incrementally, instead of just deleting all generated voicefiles, it might stand a chance of working 20.51.23 # okay this seems to be a problem with swift, because espeak works 21.34.29 Quit Acou_Bass (Quit: byeeeeeeeeeeeeeee) 21.35.37 *** Saving seen data "./dancer.seen" 21.45.55 Join lebellium [0] (~lebellium@89-92-69-17.hfc.dyn.abo.bbox.fr) 21.50.53 Join Acou_Bass [0] (~Acou_Bass@cpc97736-bolt17-2-0-cust152.10-3.cable.virginm.net) 22.03.34 # found the issue: rockbox tries to generate the voices as fast as possible, but for the swift voice engine, you can concurrently generate only as many voices as you have a license for. this makes it sometimes create empty wavefiles 22.22.46 # CH23: I'm still hanging around though :) 22.25.46 # IIRC the main problem was that different distros call cryptopp differently. Fedora uses cryptopp, Debian crypto++, upstream cryptopp. 22.26.20 # and I haven't found a good way to handle that back then. Then life came around and I ended up being busy with other stuff :) 22.27.06 # if we used autotools it would be straightforward. :P 22.27.24 # though it seems like this issue has been addressed in the meanwhile? At least looking at the Makefile it appears so 22.27.31 # I'd rather move to cmake these days 22.28.06 # yeah, but someone still has to write the rules for the cryptopp header (and libname) detection.. 22.29.16 # g#2061 tries to unify the cryptopp use, but in rbutil that patch hardcodes -lcryptopp instead of replicating the detection logic used in the other makefiles 22.29.49 # not familiar enough with qmake to port over that logic 22.31.03 # bluebrother, good to see that you are still around :) 22.33.33 # ah, right. One of the problems with the current approach is that it only works on Linux. And Windows is an important target ... 22.34.27 # windows schmindows 22.34.47 # heh. Still widely used. For some strange reason :) 22.35.01 # I think if you're building on a relatively modern linux distro the 'cryptopp' name (and #include path) is used. 22.36.02 # speachy: Debian Stretch still calls it crypto++. At least stable. 22.38.28 # don't have a debian system handy. Fedora 28+ & Ubuntu 16.04+ are okay, as is RHEL7. 22.39.37 # sorry guys, real life got in the way 22.39.44 # I have links to some audio files which reproduce the issue 22.39.45 # https://0x0.st/sCar.opus 22.40.07 # https://0x0.st/sCas.opus 22.40.43 # debian stretch does have symlinks to support ++ or pp for both the soname and include dir 22.41.10 # CH23: try this: https://pastebin.com/5Rt9pEkb 22.41.11 # specifically @ _Bilgus and speachy 22.41.15 # given that upstream uses pp.. 22.41.24 # works for me, but couldn't test against other setups yet. 22.41.51 # I consider that a bit ugly, but at least it works. Well, for me ;-) 22.42.17 # speachy: Ubuntu should be the same as Debian. 22.46.43 # so does that mean I can go ahead and commit g#2061? It's needed on Redhat-flavored systems, and shouldn't break others. 22.47.25 # woo, my 3+ hour compile finally finished. 22.47.55 # (on an 8-core box!) 22.49.02 Quit utrack (Ping timeout: 250 seconds) 22.51.26 # speachy: if you assume that always using cryptopp works (seems like it would on Debian) then you could get rid of this pkg-config detection you added to the other Makefiles. 22.52.01 # anyway, have to go now. Back tomorrow. Highlight me if you need my attention (so I won't forget to have a look here ;-) 22.55.27 # i'm trying to trick the rockboxutility in thinking it is generating the .wav for the voicefiles, because it fails to actually do that due to it being too fast. 22.55.43 Quit this_is_a_nick (Remote host closed the connection) 22.55.48 # using pkg-config is always the right way, but we can probably scrap the iterative attempt to find alternative sonames 22.56.22 # i have all needed .wav in the /tmp/rbvoice directory, and i set chattr +i on those files 22.56.50 # now when i run the voice generation, it throws the error: 22.56.59 # [talkgenerator.cpp:181 ERROR] wavtrim returned error on "/tmp/rbvoice//LANG_0.wav" 22.57.13 # Error 10: (null) 23.00.36 Join utrack [0] (~utrack@unaffiliated/utrack) 23.08.22 # this is the hackiest of hacks i've ever hacked together, but it works. i removed the "return eERROR;" lines from rockbox/rbutil/rbutilqt/base/talkgenerator.cpp 23.13.12 Join opensuse234 [0] (~opensuse@dslb-188-096-191-071.188.096.pools.vodafone-ip.de) 23.22.27 # <_Bilgus> aphirst those files get all garbled when i play them in the sim bot the original code and my patch do you experience issues with playback when you play them on your device (exculding the skipping issue mentioned) 23.24.03 Quit lebellium (Quit: Leaving) 23.24.25 # _Bilgus, as far as I can tell they play just fine for me 23.24.47 # if you want I can generate some "fresh" opus encodes from the FLAC originals and test that again (later, tomorrow) 23.24.57 Part opensuse234 23.24.58 # or I could upload the FLAC originals for you to play with 23.25.14 # <_Bilgus> weird, I'll have to test them on a target then I'll let you know when i've figured it out 23.25.39 Join martin_155 [0] (~opensuse@dslb-188-096-191-071.188.096.pools.vodafone-ip.de) 23.25.55 # <_Bilgus> shouldn't be necessary but i'll let you know 23.26.04 # well i'm going to bed now is all 23.26.15 # i'll just do it now 23.28.12 # https://0x0.st/sCaG.flac https://0x0.st/sCak.flac 23.28.23 # <_Bilgus> thanks. 23.28.47 # and i just used the "soundcoverter" gui to convert to opus, at "very low" (48kbps claimed) opus 23.29.07 # gnome-soundconverter on arch linux 23.29.34 Quit martin_155 (Client Quit) 23.29.57 Join martin_155 [0] (~opensuse@dslb-188-096-191-071.188.096.pools.vodafone-ip.de) 23.33.10 Quit martin_155 (Client Quit) 23.33.41 Join martin_155 [0] (~martin_15@dslb-188-096-191-071.188.096.pools.vodafone-ip.de) 23.35.38 *** Saving seen data "./dancer.seen" 23.38.46 Quit martin_155 (Quit: Leaving) 23.39.11 Join martin_155 [0] (~martin_15@dslb-188-096-191-071.188.096.pools.vodafone-ip.de) 23.42.15 # Hi guys. I followed your wiki guide to restore partition on Fuze+. Just wanted to say that it is incomplete, under "Troubleshooting under Linux", there should be another step "Reformat the partition using mkdosfs /dev/sdX1" and "reinstall rockbox files via utility" 23.43.01 Quit thomasjfox (Quit: Konversation terminated!) 23.43.03 # could somebody add that to this page (I don't have a wiki account): https://www.rockbox.org/wiki/STMP37xxRecovery 23.53.51 # <_Bilgus> thanks martin_155 will do 23.57.48 # thanks _Bilgus :)