Beirdobot, irc.freenode.net

Friday, November 9th, 2018, 07:06 UTC

[07:06:40]

Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has joined #mythtv

[07:06:40]

Mode for #mythtv by ChanServ!ChanServ@services. : +v Steve-Goodey

[08:18:59]

natanojl (natanojl!~jonatan@mythtv/developer/natanojl) has joined #mythtv

[08:19:00]

Mode for #mythtv by ChanServ!ChanServ@services. : +v natanojl

[08:58:29]

SteveGoodey (SteveGoodey!~steve@2a00:23c5:7da3:4501:7a24:afff:fe9d:c233) has joined #mythtv

[08:58:30]

Mode for #mythtv by ChanServ!ChanServ@services. : +v SteveGoodey

[09:01:15]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has joined #mythtv

[09:05:04]

Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has quit (Quit: Konversation terminated!)

[09:17:09]

stuarta:

jpabq: i want setup to be driven through the web server in the backend, we should be able to "bootstrap" the backend (to get an initial empty db) then everything else is done through the web interface

[09:46:56]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has quit (Quit: warpme_)

[09:48:51]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has joined #mythtv

[11:52:56]

zbot` (zbot`!~supybot@2601:647:4802:4c3:222:4dff:fe51:6728) has joined #mythtv

[11:52:56]

zbot (zbot!~supybot@2601:647:4802:4c3:222:4dff:fe51:6728) has quit (Read error: Connection reset by peer)

[12:21:33]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has quit (Quit: warpme_)

[14:15:03]

Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has joined #mythtv

[14:15:04]

Mode for #mythtv by ChanServ!ChanServ@services. : +v Steve-Goodey

[14:43:28]

Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has quit (Quit: Konversation terminated!)

[15:02:14]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has joined #mythtv

[16:46:10]

MythBuild:

build #258 of master-debian-jessie-rpi2-armv71 is complete: Failure [failed configure core compile core] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . 1/builds/258

[16:48:42]

stuarta:

sigh, raspbian libcec4-dev package is back to being libcec-dev

[16:57:58]

peterbennett:

I tried raspbian with the experimental OpenGL. At first it worked but was slow. Now after a raspbian update it segfaults every time.

[16:58:41]

stuarta:

i may try it for playback on Monday. for now i just want it to get caught up with the builds. i suspect that'll take all weekend

[16:58:41]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has quit (Quit: warpme_)

[16:59:33]

peterbennett:

I am a bit flummoxed, it seg faults in the QT event handling, so something is clobbering that.

[17:00:07]

peterbennett:

It happens after OpenGL is initialized.

[17:02:33]

stuarta:

ah, check which opengl libs you've symlinked into the mythtv installation

[17:03:01]

stuarta:

iirc, the workaround you put in to link to the original (custom) opengl libs causes things to break

[17:03:22]

stuarta:

if you swap those links out for the standard experimental opengl libs, it works

[17:03:36]

peterbennett:

stuarta: I removed those symlinks. I did some temporary changes to prevent creating them.

[17:04:47]

stuarta:

ah

[17:05:31]

peterbennett:

You refer to the symlinks in /usr/share/mythtv/lib ? You just pointed them to the other libs?

[17:05:48]

stuarta:

yes

[17:06:06]

peterbennett:

I think something else is happening, after building without them I ran ldd and found no reference to the /opt libs

[17:06:14]

stuarta:

ok

[17:06:23]

stuarta:

it's been a long time since i tried

[17:06:43]

peterbennett:

It was working and then stopped, I had upgraded to the latest raspbian

[17:07:04]

** stuarta wanders off for a bit **

[17:07:50]

peterbennett:

I think translations apart from Polish and English are way out of date.

[17:08:22]

peterbennett:

I also notice that Google translate has got a lot better at translating with correct syntax and meaning.

[17:17:45]

Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has joined #mythtv

[17:17:46]

Mode for #mythtv by ChanServ!ChanServ@services. : +v Steve-Goodey

[17:20:09]

stuarta:

new buildbot master testing -> https://bb.mythtv.org/buildbot

[17:37:46]

tgm4883:

well that's weird. Apparently I need more than --author when doing a git commit from my work laptop https://github.com/MythTV/packaging/commit/70 . . . d345689e5f7d

[17:41:13]

peterbennett:

It takes the committer from .gitconfig

[17:47:11]

tgm4883:

ah

[17:47:53]

peterbennett:

or maybe from your ssh id

[18:04:55]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has joined #mythtv

[18:06:06]

tgm4883:

peterbennett: does 30.1 really need to be released by april? I would think a dec 2019 release for 30.1 would be better

[18:06:17]

tgm4883:

IDK if we're trying to rush things here for some reason

[19:01:43]

peterbennett:

tgm4883: We have had problems before with the 0.28 release having bugs, so last year we did 29.0 release in July and 29.1 in time for Ubuntu.

[19:02:39]

peterbennett:

tgm4883: So I thought that we can create v30 then if there are any bug fixes, create 30.1 for Ubuntu

[19:03:30]

peterbennett:

tgm4883: What is the latest date for a version of MythTV for Ubuntu 19.4 ?

[19:07:17]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has quit (Quit: warpme_)

[19:18:12]

gregbert (gregbert!~gregbert@unaffiliated/gregbert) has joined #mythtv

[19:25:00]

tgm4883:

peterbennett: I think there's a misunderstanding. Generally those dates we're aiming for are to get into the LTS release of Ubuntu. LTS releases are only in April of Even numbered years

[19:26:08]

tgm4883:

and I think we normally aim for the .0 release to be in the previous release, so 30.0 would need to be released in 19.10

[19:26:22]

tgm4883:

which I think would probably have a cutoff of July 2019

[19:29:25]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has joined #mythtv

[19:43:35]

peterbennett:

tgm4883: I had thought so also, but other developers requested that we should release v30 now.

[19:47:49]

warpme_:

peterbennet: regarding raspbian with experimental opengl crash: how frequent it is?

[19:48:38]

peterbennett:

warpme_: I crash every time during startup, before the window is displayed.

[19:49:59]

peterbennett:

warpme_: It was working a few days, then no more

[19:50:56]

warpme_:

pls check what cma raspbian has. Default (in kernel defconfig) is 48M I believe. For 1080 screen I need 256M. Then it works perfectly....

[19:51:26]

peterbennett:

You mean GPU memory?

[19:52:00]

peterbennett:

or something else?

[19:53:32]

peterbennett:

My plan was to see if opengl output on raspbian can work, and change it so openmax decode can go to opengl display

[19:54:25]

Steve-Goodey (Steve-Goodey!~steve@2a00:23c5:7da3:4501:7a84:3cff:fedf:a99) has quit (Quit: Konversation terminated!)

[19:56:30]

warpme_:

sort of. I suspect ?experimental opengl? means gl provider based on mesa vc4. If yes – then memory alocation for opengl provider is done not by rpi libs (and gpu memory setting) but by mesa requesting mem from kernel MMU. As vc4 don?t have any MMU – kernel must provide big enough contineus block mem allocation. cma (contineous memory alocation) is telling kernel mmu what biggest contineous mem alocation it should support.

[19:58:07]

warpme_:

mine kernel command for u-boot booting rpi is folowing: setenv bootargs cma=256MB console=ttyS1,115200 console=tty0 root=PARTUUID=${uuid} rw rootwait smsc95xx.macaddr="${usbethaddr}"

[19:59:03]

peterbennett:

warpme_: The experimental opengl is a GPU feature with two options full KMS and fake KMS. There is some different firmware "overlay" used.

[20:00:30]

warpme_:

peterbennet: pls check who is providing opnegl. I think there are 2 options: mesa and brcm libs

[20:01:44]

peterbennett:

boot line is

[20:01:46]

peterbennett:

dwc_otg.lpm_enable=0 console=serial0,115200 console=tty1 root=/dev/sda3 rootfstype=ext4 elevator=deadline fsck.repair=yes rootwait

[20:02:35]

peterbennett:

How do I check who is providing opengl?

[20:04:36]

peterbennett:

There are three ways for opengl – the default which is software emulation and very slow, vc4-kms-v3d, vc4-fkms-v3d

[20:05:31]

warpme_:

peterbennet: changing a bit topic: recent commits for playback optimization are a bit regression for me. I started to see 2 issues: 1\sometimes starting livetv gives 5–10sec of slowed-down video with good audio. After 5–15sec it finally catches-up. 2\jump ffwd/back has frequently vibrant picture for 100–300msec right after jump. Before those commits AVsync2 was prefect for me. I?ll bisest and will try provide you exact commit causing this.

[20:05:45]

warpme_:

now returning to our opengl talk

[20:06:36]

MythBuild:

build #1121 of master-debian-jessie-64bit is complete: Failure [failed compile core] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . /builds/1121 blamelist: David Engel <dengel@mythtv.org >

[20:07:50]

warpme_:

to check who is opnegl provider ideally will be to check who provides opengl for qt. Reading debian building sources is misery for me (compared to archlinux).

[20:09:06]

warpme_:

can we somehow see what was qt configure used to build raspbian with ?experimental opengl??

[20:10:02]

peterbennett:

I amusing the prebuilt QT installed from debian

[20:11:08]

peterbennett:

for a test I built a sample opengl program from the qt doc. That works fine

[20:11:16]

peterbennett:

Also the glxgears works fine

[20:11:58]

warpme_:

peterbennet: right. It will be good to know how those prebuild-qt were built....

[20:12:22]

warpme_:

anyway: pls try add cma=256MB to kernel cms line and reboot

[20:12:31]

peterbennett:

ok

[20:12:34]

warpme_:

s/cms/cmd

[20:13:09]

peterbennett:

What should i use for gpu memory?

[20:13:26]

warpme_:

are you using: software emulation or vc4-kms-v3d or vc4-fkms-v3d?

[20:13:57]

peterbennett:

vc4-kms-v3d

[20:14:18]

peterbennett:

maybe I should first leave gpu as default

[20:16:53]

warpme_:

ok – let?s guess this means mesa (this is v.old naming used by Eric Anhold two or so years ago when he was experimenting with brcm 3D engine). Now he works for brcm and wrote full 3D suuport for vc4 in mesa. Sad that rasbian uses so outdate terminology....

[20:17:32]

MythBuild:

build #819 of master-archlinux-64bit is complete: Success [build successful] Build details are at http://code.mythtv.org/buildbot/builders/mast . . . t/builds/819

[20:17:33]

peterbennett:

So I rebooted with that I will try starting mythfrontend with OpenGL painter.

[20:18:54]

warpme_:

yes. I forgot You are using rpi bootloader – not u-boot. Forgive me. I?m avoiding propietary blobs if possible so on rpi i?m mailine u-boot + mainline kernel guy....

[20:21:27]

peterbennett:

It still crashes.

[20:21:36]

warpme_:

argh

[20:22:10]

warpme_:

what mesa version raspbian has?

[20:22:29]

peterbennett:

How can I tell that?

[20:23:36]

warpme_:

pls try list of installed packages. maybe version will be there?

[20:23:51]

peterbennett:

It dies right after initializing OpenGL.

[20:24:52]

peterbennett:

libegl1-mesa:armhf 13.0.6–1+rpi2

[20:25:28]

peterbennett:

Bunch of mease libs with version 13.0.6

[20:25:35]

peterbennett:

mesa

[20:26:31]

peterbennett:

libgles2-mesa:armhf 13.0.6–1+rpi2

[20:26:53]

warpme_:

i remeber I had similiar thing sometime ago when I was testing what full FOSS can do on rpi. But since some mesa ver. it started to work. Initially with crash after some ui browsing, later ui was ok but playback wasn?t possible. lastly ALL warks really perfect (albeit cpu can deal only with sd content; for hd is too weak :-( )

[20:27:43]

warpme_:

arh 13.0.6 is indeed 2y old. On his level forget any vc4 support as it started at 18.+

[20:28:10]

warpme_:

geez – current raspbian has mesa at v13 level?

[20:30:02]

warpme_:

are there any ?experimental? packages for raspbian? we need 18.2 at least as this ver. fully works (acordingly to my development of minimyth2 for arm world)

[20:30:08]

ijc:

Hello. Who is best to ping to look at https://code.mythtv.org/trac/ticket/13315 AKA https://github.com/MythTV/mythtv/pull/167?

[20:30:08]

ijc:

It's fixes to the mythprotoserver "FileTransfer" handler to handle URL quoting correctly and similarly in the mainprogram equivalent.

[20:30:08]

ijc:

hampton: you touched the files most recently but they look to be things unrelated to the FileTransfer side of things which just happened to touch those files so maybe not?

[20:30:08]

ijc:

jya: you seem to have been involved back in 2014 but not so much nowadays it seems?

[20:31:14]

jya:

jya: yeah, got a job from august 2014, so had to start to do real work

[20:31:31]

warpme_:

jya: :-)))))

[20:31:45]

ijc:

jya: Very belated congratulations ;-)

[20:32:06]

jya:

i hadn't been working for about 5 years prior that (early retirement), mythtv was just a hobby then.

[20:32:23]

jya:

funnily, I got that job because the manager was a mythtv user.

[20:33:30]

warpme_:

jya: I?m really impressed by current mozilla as plex media player runtime. Ejoying it much much more that bloated chrome!

[20:33:43]

jya:

that's good to hear

[20:34:02]

jya:

didn't know you could use firefox as plex media player though.

[20:34:19]

jya:

codec support is rather limited if you want to watch TV content

[20:34:29]

jya:

just h264, vp8 and v9

[20:35:11]

warpme_:

jya: https://github.com/warpme/plex-plugin-mythtv-livetv works perfectly for me :-)

[20:35:34]

ijc:

jya: It's funny how things work out!

[20:35:34]

ijc:

Anyway, no worries, I'll wait and see if one of the active devs will pick it up.

[20:36:26]

warpme_:

jya: what is missing is hw assisted h264 decode on mac. simply my rMBP becomnes a bit too hot....

[20:36:43]

jya:

?? on mac h264 is fully hw accelerated

[20:36:47]

jya:

all the way to the compositor

[20:37:01]

warpme_:

and mythfrontend consumes 2x less cpu than firefox with plax player

[20:37:15]

jya:

it may become hot, but that can't be because of the hw decoder

[20:37:32]

warpme_:

s/plax/plex

[20:37:33]

jya:

warpme_: install this: https://addons.mozilla.org/en-US/firefox/addo . . . media-panel/

[20:37:46]

jya:

now play your video, press Command-Option-I

[20:37:55]

warpme_:

jay: qll – will try it :-)

[20:38:06]

jya:

and go to the media tab in there.. the URL of the video playing will show. Click on it and expand it.

[20:38:12]

jya:

paste the result here

[20:39:40]

hampton:

ijc: I've touched most of the files in the last year doing code cleanups.

[20:40:39]

ijc:

hampton: yeah, that's what it looked like.

[20:40:40]

ijc:

I don't suppose you know which dev it's best to ping on that stuff?

[20:43:49]

natanojl:

ijc: I think I made a comment about cleaning up the history but I think I may have done so in the #mythtv-users channel. It looks to me that you are adding code in one commit and remove it in a later one which could be squashed together

[20:45:01]

ijc:

natanojl: that's right, I was waiting for a dev to comment on which approach they preferred before squashing or removing.

[20:46:25]

natanojl:

ijc: Okay, I hadn't realized/seen that

[20:46:28]

ijc:

You weren't on mythtv-users when I came back and saw that comment so I never managed to reply.

[20:46:49]

warpme_:

jya: here is is https://pastebin.com/t9UUMfw5

[20:47:10]

ijc:

natanojl: Since this is #mythtv I now realise you are a dev and will take that as more of an instruction to go with the end state (which I think is more correct) and do the squash, sound good?

[20:48:01]

jya:

"apple hardware VT decoder"

[20:48:08]

jya:

warpme_: so it's the HW decoder

[20:48:58]

warpme_:

firefox consumes 50–80% cpu (on 0–800% scle). mythfrontend on this content has 25–35%

[20:49:10]

warpme_:

s/scle/scale

[20:49:31]

hampton:

I'm ok with switching from a qurl to a qstring once you've determined that you're dealing with a local file path.

[20:50:52]

warpme_:

jya: yes i?m seeing this. probably these 50–80% isn?t by decode byt maybe by js player plex uses?

[20:51:06]

peterbennett:

ijc: Is that ticket only for music tracks with quotes in the file name?

[20:51:12]

natanojl:

ijc: I probably need to take another look. But I believe it should be fine to drop QUrl in this case, like hampton says

[20:51:19]

warpme_:

any way: fantastic work with mozilla!

[20:51:24]

jya:

mythtfrontend does like 5% of the work done by FF to display a frame. we do have known issue especially with intel based mac.

[20:51:40]

jya:

if you were to force your mac to use the dGPU if it has one, battery usage would be much better

[20:52:18]

warpme_:

jya: interesting. I?ll check this as I have nv750m

[20:52:51]

ijc:

peterbennett: music is where I noticed it. I think other plugins (gallery perhaps) might hit the same problem but I didn't test that other than to verify it still worked for me with the change (but I don't have any photos with odd characters in the name)

[20:53:04]

ijc:

hampton: natanojl: I'm doing the squash now

[20:53:59]

ijc:

peterbennett: oh, and mythvideo too I think, I tested that with and without odd characters (according to the commit message at least)

[20:54:36]

warpme_:

hampton: nay chance to fix https://lists.gt.net/mythtv/dev/620919

[20:54:45]

peterbennett:

OK I see it, If I put quotes in a file name I cannot play it remotely but local works fine.

[20:54:45]

warpme_:

s/nay/any

[20:55:05]

warpme_:

it?s as clear regression

[20:57:43]

peterbennett:

natanojl: Are you taking care of that ticket or would you like me to take a look at it?

[20:58:00]

ijc:

natanojl: peterbennett: It seems like I dropped the QUrl from mythtv/programs/mythbackend/mainserver.cpp but left mythtv/libs/libmythprotoserver/requesthandler/fileserverhandler.cpp with the previous more complex fix (to undo the URL encoding instead of just not doing it).

[20:58:01]

ijc:

I thought I had done both and think I should do both.

[20:58:55]

natanojl:

peterbennett: Feel free to handle it if you want

[20:59:12]

ijc:

I'll look into fixing fully ASAP and provide an update, but don't merge for now, it'll take me a day or two at least. I'll ping you again when I'm done.

[20:59:12]

ijc:

Sorry, I really thought I'd done both halves.

[21:02:08]

peterbennett:

ijc: natanojl: I don't use MythMusic but I can look at this and test it with video and commit it if all looks good.

[21:03:56]

ijc:

peterbennett: thanks, will let you know when I've applied the fix to the mythprotoserver side.

[21:05:38]

natanojl:

peterbennett: Thank you

[21:07:28]

hampton:

warpme_: Top of my list.

[21:07:50]

peterbennett:

ok

[21:08:21]

warpme_:

hampton: thx alot!

[21:08:49]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has quit (Quit: warpme_)

[21:57:43]

gigem:

stuarta: It looks like the commit hooks are stuck again.

[21:57:50]

SteveGoodey (SteveGoodey!~steve@2a00:23c5:7da3:4501:7a24:afff:fe9d:c233) has quit (Quit: Konversation terminated!)

[22:01:30]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has joined #mythtv

[22:02:49]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has quit (Client Quit)

[22:21:15]

hampton:

I just pushed your commit message.

[22:28:50]

amessina (amessina!~amessina@unaffiliated/amessina) has joined #mythtv

[22:30:15]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has joined #mythtv

[22:41:45]

warpme_ (warpme_!~warpme_@89-70-137-1.dynamic.chello.pl) has quit (Quit: warpme_)

[23:14:50]

IonTau (IonTau!~IonTau@ppp121-45-221-40.bras1.cbr2.internode.on.net) has joined #mythtv

[23:15:45]

nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has quit (Ping timeout: 252 seconds)

[23:39:05]

nephyrin (nephyrin!~neph@2601:600:817f:f46a:e41d:6f5c:e3be:b8ad) has joined #mythtv

[23:48:57]

gigem:

hampton: Was that intended for me? If so, it didn't make it to the ticket in trac.

[23:51:41]

natanojl (natanojl!~jonatan@mythtv/developer/natanojl) has quit (Ping timeout: 260 seconds)

[23:56:41]

hampton:

Yes, that was for you. I think I got it now.

MythLogBot@irc.freenode.net :: #mythtv

Current users (47):