Skip to content

Fix screencopy shm release and PipeWire version check#4217

Draft
coleleavitt wants to merge 2 commits into
niri-wm:mainfrom
coleleavitt:feat/screencopy-pipewire-fixes
Draft

Fix screencopy shm release and PipeWire version check#4217
coleleavitt wants to merge 2 commits into
niri-wm:mainfrom
coleleavitt:feat/screencopy-pipewire-fixes

Conversation

@coleleavitt

Copy link
Copy Markdown

Summary

  • Avoid calling pw_check_library_version, which is missing from older libpipewire runtimes, by reading and parsing pw_get_library_version instead.
  • Release accepted shm buffers when wlr-screencopy frames finish, after ready on success or after failed on failed/abandoned copies.

Issues

Testing

  • cargo fmt --all -- --check
  • cargo test -q pipewire_version_is_at_least
  • cargo check --message-format=short
  • pre-push hook: cargo check

@veneficium42

Copy link
Copy Markdown

This is maybe not the most appropriate place to have this discussion but i couldn't help but notice you opened roughly 12 PRs in the past 5 hours.
May I ask why you opened all of them (some with big changes) in such a short amount of time? Surely you didn't write them all in the past 5 hours.
Sorry for disturbing the PR comments but I am really curious about it.

@j0ni

j0ni commented Jun 21, 2026

Copy link
Copy Markdown

This is maybe not the most appropriate place to have this discussion but i couldn't help but notice you opened roughly 12 PRs in the past 5 hours. May I ask why you opened all of them (some with big changes) in such a short amount of time? Surely you didn't write them all in the past 5 hours. Sorry for disturbing the PR comments but I am really curious about it.

rust-lang/rust#156077 might be instructive.

NB I am not passing any judgement, other than agreeing that 12 PRs in 5 hours, with no prior contribution, is anomalous behaviour. I know people who go on coding binges like this though, and I haven't looked at any of this code. Anyway, that rust-lang pr was closed with a moderation warning, and that's Rust core, who aren't known to be LLM-hostile per se.

@coleleavitt

coleleavitt commented Jun 21, 2026

Copy link
Copy Markdown
Author

This is maybe not the most appropriate place to have this discussion but i couldn't help but notice you opened roughly 12 PRs in the past 5 hours. May I ask why you opened all of them (some with big changes) in such a short amount of time? Surely you didn't write them all in the past 5 hours. Sorry for disturbing the PR comments but I am really curious about it.

rust-lang/rust#156077 might be instructive.

NB I am not passing any judgement, other than agreeing that 12 PRs in 5 hours, with no prior contribution, is anomalous behaviour. I know people who go on coding binges like this though, and I haven't looked at any of this code. Anyway, that rust-lang pr was closed with a moderation warning, and that's Rust core, who aren't known to be LLM-hostile per se.

with no prior contribution

if you'd look at the contribution history this statement is false, rather than speculating and insisting threats, it may be valuable to actually go read before saying

  • cole

@coleleavitt

coleleavitt commented Jun 21, 2026

Copy link
Copy Markdown
Author

This is maybe not the most appropriate place to have this discussion but i couldn't help but notice you opened roughly 12 PRs in the past 5 hours. May I ask why you opened all of them (some with big changes) in such a short amount of time? Surely you didn't write them all in the past 5 hours. Sorry for disturbing the PR comments but I am really curious about it.

I maintain a private fork and I have took it beyond myself not to contribute upstream a couple months ago due to a disagreement between members in the unofficial discord server over topics unrelated to development which resulted in a timeout when the initial conversation was induced by a development request. Hence, I took it upon myself to fix issues and maintain my own version rather than contributing because of the hostile actions of the team. It also seems the owner has become rather preoccupied with other medial tasks including school and his research not being able to handle the backlog of issues and features which caused me furthermore to fix such issues on my own time. In inception, I admired this project but the discord server despite being unofficial has become rather propaganda making me deterred from contribution

@veneficium42

Copy link
Copy Markdown

I maintain a private fork and I have took it beyond myself not to contribute upstream a couple months ago due to a disagreement between members in the unofficial discord server over topics unrelated to development which resulted in a timeout when the initial conversation was induced by a development request. Hence, I took it upon myself to fix issues and maintain my own version rather than contributing because of the hostile actions of the team. It also seems the owner has become rather preoccupied with other medial tasks including school and his research not being able to handle the backlog of issues and features which caused me furthermore to fix such issues on my own time. In inception, I admired this project but the discord server despite being unofficial has become rather propaganda making me deterred from contribution

ah, glad to know these weren't just vibecoded in the timeframe between each PR request.
One last curiosity ':D
Why did you put these two, from my POV, unrelated changes into a single PR?

@coleleavitt coleleavitt marked this pull request as draft June 21, 2026 14:35
@Sempyos Sempyos added area:pipewire Issues with the PipeWire code area:screencopy Issues with wlr/ext-screencopy code pr kind:fix Issue fixes, code cleanups labels Jun 24, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

area:pipewire Issues with the PipeWire code area:screencopy Issues with wlr/ext-screencopy code pr kind:fix Issue fixes, code cleanups

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Older versions of libpipewire-0.3 don't have pw_check_library_version shm buffer not released on ready after screencopy

4 participants