Open xander-st-aubyn opened 3 months ago
Good question. Once I have more time (currently busy writing thesis) I am planning to use xcb to send more compositing requests in a bulk, thus reducing the time frame where tearing can occur. I suspect this will not be enough though and then I'll likely port wait_vblank from xfwm4 or another compositor to fastcompmgr. If you want to help, you could test, whether tearing is reduced on your machine when running dcompmgr from Dana Jansens, clonable via git clone git://git.openbox.org/dana/dcompmgr
and report your results back. Compilation may be cumbersome, so you could also run my attached binary ./dcompmgr --no-fade
.
dcompmgr.tar.gz
On this website https://www.vsynctester.com I get almost no tearing with dcompmgr, even when using no GLX backend.
This was on EL9 (rocky if it matters) and I am currently in a fully setup openbox environment on FC39. I have no issues being told to go compile something if its more or less clone (./configure) make && make install
. I was happy your make file included an uninstall ;)
I will do that and report back!
In picom to get a tear free accelerated experience I have to specify a glx backend and vsync.. How can I get the same on fastcompmgr?