OK, why SeaMonkey has four sound streams open according to Pipewire!?

maybe this could explain my video failures?

just opened pavucontrol / Easy Effects and noticed four instances of SeaMonkey having audio streams open despite the browser playing not a single note

it's like... something is not closing a stream after being done?

hadn't noticed this until now, when I was checking the volume controls for completely unrelated reasons

I think there was a var to change the backend, which might help

er, help getting more information

but first, have you found a good way to reproduce consistently?

at least the multiple stream part.

provided a way to reproduce, then comes the question, does it map to the failures in some way

(and I don't do much audio, so I'm not even sure that alsa is working well in the build I use)

I am pissed

.. what else is new

Basilisk-Dev has dmca's all the basilisk repos on github because they have the basilisk logo in em.. they are all five years out of date and predate the move to gitea

including repos owned by former Pale Moon/Basilisk/UXP contributors namely adeshkp and lootyhoof aka ryan .. the theme master

Moonchild has lost control.

njsg: just noticed it's happening on two different systems, but both are using Pipewire

next thing I'm going to try is to open several videos in a row and see where it breaks

njsg: OK, managed to trigger it again, and gathered more info

the sites I visit the most with videos (Danbooru, onemileatatime.com, and the JWZ blog), some have autoplaying video (OMAAT), others don't.

Sometimes, Seamonkey will fail to close audio streams even after navigating away or closing a tab with a video

pavucontrol / Easy Effects will list those

keyword here is "sometimes" - not always, sometimes SeaMonkey DOES close the stream just OK, but sometimes it fails for whatever reason

here is the kicker: leave 4 audio streams open, and THAT will break videos browserwide, even those without audio!

Fun fact: the autoplaying videos on OMAAT are silent, yet sometimes those trigger a new audio stream that may or may not close

Oh, and if you exit the browser with those audio streams still open, it will hang at exit and you'll have to kill the process

(oh, just got a crash... but that's unusual)

EasyEffects / pavucontrol identify the audio streams as "SeaMonkey: AudioStream"

but SeaMonkey is going through Pulseaudio... which on Pipewire is emulated via pipewire-pulse

hmmm, is "media.cubeb.backend" supposed to work on SM? Tried setting it to "alsa" to force ALSA (which is gonna be bridged over Pipewire anyway) to test something, but SM still outputs via Pulse...

in any case, could be my video woes related to Pipewire?

fun fact: there is no cubeb backend for Pipewire (yet), and even Pipewire themselves are telling devs "don't use our API directly yet, continue using Pulse"

tomman what hardware gfx?

blocked by default: Acceleration blocked by platform

unavailable by default: Hardware compositing is disabled

(GPU is a Sandy Bridge HD3000)

frg_Away: that's what I'm looking here, audio backend says "pulse-rust"

Well separate audio chip then. HD3000 under Windows is junk. Don't have anythingleft with this chip so can't test Linux. When I am done with js I will try to update media again but no eta. JS updating takes longetr than expected because of real life issues.

well, SeaMonkey recognizes that I've changed media.cubeb.backend to "alsa", but it defaults to "pulse-rust" anyway

too bad I had already updated all of my boxes to Pipewire (except for those that have real soundcards, on those I don't even have PulseAudio installed, but I don't run web browsers on those either)

OMG

The Wall Street Journal replaced its paywall by a captcha

...that instablocks me, blaming "suspicious behavior" or "you have a robot in your IP"

or simply "you browse at a superhuman speed"