On Fri, 10 Jul 2026 06:20:23 +0300 Daniil Ryvkin <mail@leinadr.com> wrote: > On Thu, Jul 09, 2026 at 08:15:45PM -0500, izzy Meyer wrote: > > On Thu, 9 Jul 2026 20:50:00 +0300 > > Daniil Ryvkin <mail@leinadr.com> wrote: > > > > > On Thu, Jul 09, 2026 at 10:51:41AM -0500, izzy Meyer wrote: > > > > On Tue, 7 Jul 2026 18:10:15 +0300 > > > > Daniil Ryvkin <mail@leinadr.com> wrote: > > > > > > > > > This adds a port for OBS Studio 32.1.2, a free and open source > > > > > application for live streaming and screen recording. > > > > > > > > > > Requires devel/simde (submitted separately). > > > > > > > > > > Disabled: browser source (CEF), WebRTC, MPEG-TS output > > > > > (SRT/RIST), obs-websocket, virtual camera output. > > > > > > > > > > Tested on X11 (spectrwm). Screen capture (XSHM and > > > > > Xcomposite), audio (sndio) recording, webcam capture, x264 > > > > > and VAAPI encoding, file recording and streaming (RTMP) all > > > > > work. > > > > > > > > > > PipeWire screen capture (Wayland sessions) is implemented but > > > > > could use testing from someone running a Wayland compositor. > > > > > > > > Tested on amd64, using StumpWM on Xenocara. > > > > > > > > I recorded a quick screencap of my session with my webcam in the > > > > corner in X264 mp4 format. Seemed to work fine. I had trouble > > > > getting the sndio mic input to work on my end, though. I tried > > > > setting it to snd/mon when my server.mode was play,mon. No mic > > > > was picked up. Its probably something on my end cos my audio > > > > situation is screwy right now. > > > > > > > > Thanks for the port. > > > > > > > > -- > > > > iz (she/her) > > > > > > > > > I say mundane things > > > > > so the uninteresting > > > > > just might get noticed. > > > > > > > > izder456 (dot) neocities (dot) org > > > > > > > > > > Hi, > > > > > > Thanks for testing! > > > > > > Regarding the sndio mic issue: snd/mon is a monitor device for > > > capturing desktop audio (playback loopback), not microphone input. > > > For microphone recording, use snd/0 (built in) or snd/1 (USB mic). > > > > > > Also make sure audio recording is enabled at the kernel level: > > > sysctl kern.audio.record=1 > > > > > > > I had that sysctl set. I had a brain fart when reading your > > pkg-message. Perhaps this information in the pkg-message would be > > better in a pkg-readme for easy access after install, and perhaps > > other quirks that may be useful to the end-user? > > > > -- > > iz (she/her) > > > > > I say mundane things > > > so the uninteresting > > > just might get noticed. > > > > izder456 (dot) neocities (dot) org > > > > Thanks for the suggestion, moved the setup details to pkg-readme. > MESSAGE now only shows the essential sysctl commands. No problem. I have a few other nits: The pkg-readme should follow the template. E.G.: you missed adding the header shown in the file here: /usr/ports/infrastructure/templates/README.template +------------------------------------------------------------------------------- | Running ${PKGSTEM} on OpenBSD +------------------------------------------------------------------------------- -- pkg-readme and pkg-message could be streamlined a bit. Additionally, it may be best just to point users to the multimedia FAQ as to not duplicate documentation in your pkg-readme. How this would be worded, should be up to you. But the idea is that end-users would have one central place for finding multimedia information not-specific to this port. Putting generic multimedia quirks about OpenBSD as a system, that are not specific to this port is redundant IMO. Same concept applies to the pkg-message. It is duplicate information from the FAQ multimedia page. But IMO *only* quirks specific to the obs-studio port should be in pkg-readme, everything else should point to the FAQ or the relevant manpage(s). No need to put in pkg-message notice to read the pkg-readme, pkg_add already informs the user that a pkg-readme was installed. Also- I don't see many ports in the tree that still use pkg-message. I'm fairly certain the general approach is to *just* have a pkg-readme in your port. But I could be mistaken. -- Looks like the plist wasn't re-rolled when you added the pkg-readme as well. The plist should be updated with `make update-plist` when you add a file to the package, like a pkg-readme for example. Attached is a port with the re-rolled plist and pkg-readme with the header added. Could you streamline pkg-readme and pkg-message? -- iz (she/her) > I say mundane things > so the uninteresting > just might get noticed. izder456 (dot) neocities (dot) org
No comments:
Post a Comment