Tuesday, 2023-09-26

dcalisteGood morning pvuorela, do you have time to discuss some open PRs this morning ?07:07
piggz[m]poetaster: for seaprint ... have you tried turning off mobile data while wlan is turned on? for noson, i couldnt get it to find my sonos system until i did that, and that uses mdns for discovery09:04
poetasterpiggz[m], yeah, I tried that. It doesn't find the printer even if you supply the url and the printer is connected to the same network. Connect to the DHCP server of the printer, general netowrk regardless, no connect.09:09
poetasterpiggz[m], it's presumably the same issue with qmllive not being able to talk back to the sdk.09:09
piggzpoetaster: did you try flushing and accepting all iptables rules?09:10
piggzKeto: hi, could you give my new device store acesss. pls09:10
piggzDevice model: Phone X23 (halium-vidofnir / halium-vidofnir)09:11
poetasterpiggz, nope. I just set: NetworkInterfaceBlacklist = p2p,usb,rmnet,rev_rmnet,ccmni-lan,wlan0,ifb0,ifb1 and rebooted.09:12
piggzwhen i get store access on the x23 ill test09:13
poetasterpiggz, I'll compare iptables -L on the sony to the GS5 and see if anything pops up.09:14
Ketopiggz: added09:14
piggzpoetaster: if it makes you want the x23 .. it found my printer just fine09:17
* piggz[m] uploaded an image: (200KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/wTqtueEZahxkrAYqTzlbmKat/sailtrix-file.png >09:18
poetasterpiggz, I do not want the x2309:18
poetasterpiggz[m], the gs5/rephone has a better camera / screen resolution than the x23.09:19
piggzit does, but the x23 is great for outdoors :)09:19
poetasterpiggz[m], also, the x23 is bulky in comparison.09:19
poetasterpiggz, outdoors, I don't use a cellphone (at least very rarely). I stopped using cellphones for hiking about 2012 :)09:19
poetasterok, there are a number of ports like, ACCEPT     udplite--  anywhere             anywhere             multiport dports 1024:6553509:21
poetasteron the sony that don't show up on the GS509:22
poetasterpiggz, https://paste.opensuse.org/pastes/7ab5961694ff09:23
piggzpoetaster: try pasting this lot in..09:23
piggziptables -P INPUT ACCEPT09:23
piggziptables -P FORWARD ACCEPT09:23
piggziptables -P OUTPUT ACCEPT09:23
piggziptables -t nat -F09:23
piggziptables -t mangle -F09:23
*** Ischwitch is now known as Ingvix09:23
piggziptables -F09:23
piggziptables -X09:23
piggzthat should flush all and accept all incoming i think09:23
poetasterpiggz, I use shorewall.09:24
piggzwhats that09:24
piggzeven so ... the able will clear all iptables rules and rule them out09:25
poetasterthat's a sysadmin joke.09:25
piggznetworking issues are no joke :)09:26
poetasterthe pastebin I posted shows a number of accepts that look probable: udplite, sctp, dccp09:26
poetasterpiggz, I also don't have networking issues on my servers now, do I :)09:27
piggzmal: is something in .24 compiled/packed much more efficiently? zypper dup says 129.0 MiB will be freed.09:32
piggzbut only 2 license packages are being removed09:32
poetasterpiggz, flushing the tables doesn't make a difference.10:02
poetasterpiggz, at least not where seaprint mdns is concerned (testing qmlive foo, now)10:02
piggzill test seaprint with my v22 later10:03
poetasterpiggz, okidoke. I'll poke about some more.10:03
poetasterpiggz, well, flushing the tables DOES get qmllive working. yeah!10:08
poetastera bit permissive iptables, though :)10:08
poetasterpiggz, sorry, spoke too soon.10:10
poetasterpiggz, no, I did not speak too soon, the SDK is just unspeakably slow. I wonder what I did. with no iptables rules, qmllive works.10:37
piggzpoetaster: just have a script run on startup to clear all your rules, and its be fine :)10:54
poetasterpiggz, nah, I just develop with the GS290 / 10iii In a pinch, iptables -F will suffice.12:38
poetasterpiggz, so, I just restarted the phone, ran your flush procedure as above and now Seaprint sees the printer, too. So both qmllive and seaprint works if the firewall is lowered.13:17
malrubdos[m]: some issue with newer rust, we either need to revert several commits or add support for statx to scratchbox213:20
rubdos[m]aaaaaaaargh yikes :)13:21
rubdos[m]do you have some pointers for me to work on that, or can you decide internally do carry out the statx impl?13:21
malwell that same issue happens with newer glibc so we might need to support statx13:22
rubdos[m]I would expect that it'll become a requirement indeed...13:22
* rubdos[m] opens man statx13:22
malthe reason why statx is used it to fix y2038 issue on 32-bit archs caused by limitation of 32-bit timestamp13:23
rubdos[m]Doesn't sound like it will be horribly difficult... just quite a bit of honest work13:23
malwith statx 64-bit timestamp can be used13:23
rubdos[m]makes sense13:24
rubdos[m]so it will probably be a copy of stat, and then s/stats/statx in the impl13:24
rubdos[m]as a first try, I mean13:25
malnot sure if that simple, discussing with person who knows more about sb213:25
rubdos[m]ack, thanks.  If I can help, let me know!13:25
rubdos[m]Maybe it's also worth putting these kind of findings on a public Github issue on the sfos rust repo13:26
piggz[m]poetaster: weird network issues are always the firewall :D14:10
malpiggz: not sure about the size difference, there should not be any obvious change affecting that14:12
piggzmal: strange ... that was a change from .19 to .2414:21
maloh, need to check the diff of those, I thougth that was .21 to .2414:41
poetasterpiggz, I don't need no esteenking firewall. I feelter my packets vith my teeth.14:51
poetasterpiggz, so, there is no differnce between the connman configs (10iii vs. volla22) but for the main-mtk file.15:03
piggz[m]poetaster: so, comparing the 22 and 23 ... seaprint works on 23, but not my 22 also15:17
piggz[m]on 23 i connected to wlan1, on 22 im connected to wlan015:17
piggz[m]the iptables rules are also different15:17
*** ggabriel is now known as Guest137015:19
Solrac[m]Pardon me, but please correct me, it's still \not possible to use Video out from the USB-C  on the Xperia 10 III, right?18:23
malnot possible at the moment, not sure how popular such a feature would be18:30
* Solrac[m] rotates UT's convergence use cases degrees18:32
Solrac[m]It could be18:32
attahI don't see convergence as realistic... it will be like trying to do work on a tablet - awful18:47
attahThough i see a case for playing videos, or doing presentations - perhaps video calls18:47
attahBut like one use case per session sort of stuff18:48
Solrac[m]attah: This is what I had in mind. Worse-Convergence Case Scenario would be LXC  Containers with a Desktop18:48
attahArguably DLNA does the video part better, and is probably equivalent on presentations with some tuning18:55
poetasterpiggz[m], yeah, I noticed that the iptables rules were different on the 10iii. the only config diff19:54
poetasteris PreferredTechnologies = bluetooth,ethernet,wifi,cellular19:54
poetasterDefaultAutoConnectTechnologies = ethernet19:54
poetasterThose rules cause a bit of havoc on the volla22.19:55
poetasterattah, by DLNA you mean like casting with Juppi?19:56
poetasterattah, ok, it does work well with my rinky dink dlna media server.19:57
attahgiven a good format match, video playback is great19:57
attahNot sure where the screen mirroring performance is now, but should be good for presentations19:58
poetasterI've never tried that screen mirroring like stuff. I mainly use it for music. and old bbc recordings of dr. who. best viewed on a phone because of the lousy quality :)20:01
* attah should use more flac over dlna and less Spotify connect20:02
poetasteryeah, that does depend on you originals. I have a ton of so, so mp3s still (made circa 1999). But I don't use spotify. I do click about randomly on bandcamp, though.20:03
poetasterSomeone on the forum said please do bandcamp and I found a nice functional scraper in python. the audio is all pretty so, so, but there's a lot of odd stuff there.20:04
attahAnd i thought i was the young one... with my shelf-meter of CDs20:04
direc85[m]I'm trying to compile Ruben's LLVM 16 PR - with patches fixed and applied correctly this time - but the compilation stops in the beginning.20:04
direc85[m]I built cmake from Sailfish master and installed that to aarch64 build shell, but it still claims I have too old cmake - any clue what could cause that? Should I compile it on the platform sdk level and install there perhaps?20:06
poetasterOff to bed, good luck with cmake!20:06
attahdirec85[m]: The advice i got was to use a shared "task" output directory, so taht any declared dependencies may get pulled from there automagically20:07
direc85[m]LLVM 15 compiled just fine with mb2 --target SailfishOS-latest-aarch64 -s rpm/llvm.spec build but LLVM 16 does not...20:14
direc85[m]It looks like there's llvm/llvm/build/CMakeFiles and llvm/llvm/build/SOURCES/CMakeFiles, and the error is CMake Error: The source directory "SOURCES/CMakeFiles/CMakeTmp" does not exist. with CMake Error at /usr/share/cmake/Modules/CMakeDetermineCompilerABI.cmake:57 (try_compile): Failed to configure test project build system. so that could be the cmake error indeed...20:15
direc85[m]Okay, I don't know what's going on any more:... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/fRnTQNArddJKclEJOJfgnGya>)20:36
direc85[m]I had to install the cmake in the SailfishOS-latest tooling, now I'm compiling!20:55
direc85[m]Got llvm+clang 16 + cmake compiled for i486, had to do a few fixes and a one duct tape patch: https://github.com/rubdos/llvm/pull/122:56

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!