Blank window with glutin about emacs-ng HOT 16 CLOSED

With --enable-winit-tao the WIP branch works for me.

Be aware, changed to --with-winit=tao.

--enable-wr-glutin changed to --with-wr-gl=glutin

from emacs-ng.

Do you mean the second frame or all frames?
Second frame being blank is a known issue, which I fixed in my branch.

from emacs-ng.

all frames. A breakpoint on flush shows it does get there, but stepping through never causes anything to render.

from emacs-ng.

from emacs-ng.

Probably Mutext dead lock from event_loop.rs. I changed some lock() to try_lock()... Are there any Failed to grab a lock from logs?

No, but hang on, that's a good shout, I'll do some digging on the event loop.

EMACSNG_LOG=wrterm=trace ./emacs

Call me stupid but I actually couldn't figure out how to filter by crate. This is enormously useful. Methinks it should be in the docs (or I just missed it).

from emacs-ng.

from emacs-ng.

I've PRd some docs. What caught me out was not spotting that tracing's EnvFilter works like env_logging. I was just using grep....

from emacs-ng.

Nope, not a mutex problem. None of my quick guesses were right either. I'll dig into this later and figure it out if it's not yet fixed.

from emacs-ng.

from emacs-ng.

I replaced all try_locks with lock. Diffing with the working code should show what it is.

from emacs-ng.

Would you like to give it a try for the code here #490.
I managed to git cherry-picked all the changes from my tao branch in one patch. @2e0byo

from emacs-ng.

That now loads. the scaling code is broken, but it always was with winit. (Causes some weird bug with the cursor as clearly two different values for scale are being used internally.)

I'm a bit out of touch atm as I'm madly finishing my thesis :D But I needed to get Emacs functional first... When the tao/winit split lands upstream I'll try to see what's going on.

I run hyprland mostly and it's definitely more of a problem than e.g. weston. E.g. it sends lots of resize events, so the queue needs to be filtered before processing them, or they take forever to process (and occasionally hang). But all that is on the list for when the next two chapters are in ;)

from emacs-ng.

Bother, actually read the source now.

With --enable-winit-tao the WIP branch works for me. (pace some rendering problems which existed before hand). In any case this issue is definitely closed by it.

from emacs-ng.

pace some rendering problems which existed before hand

Do you mean the flicking when frame changing focused stated? It can be improved with using tao underline gtk GLContext I think. I may add that later.

For other rendering problems which exist in the issues we can tackle step by step if anyone is interested.

from emacs-ng.

I'm a bit out of touch atm as I'm madly finishing my thesis :D But I needed to get Emacs functional first... When the tao/winit split lands upstream I'll try to see what's going on.

Good luck with you thesis. I am sticking to PGTK Emacs for now until WR Emacs are ready for daily use.

tao/winit split should happen in the #490. We are waiting for @harryfei 's review. I squashed multi commits into one. There's a lot of changes. Might take some time.
But I am pretty confident about the change though.

from emacs-ng.

Do you mean the flicking when frame changing focused stated? It can be improved with using tao underline gtk GLContext I think. I may add that later.

Yes. I don't know what you mean with tao, but I'd be glad if it did fix it. I spent some time in the debugger trying to figure out what was blanking the window, but couldn't. It seems to happen outside flush().

I am sticking to PGTK Emacs for now

PGTK was unuseable on this ancient machine---it just can't handle the screen resolution. It's not finished but I'm really happy with emacsng for now.

I'll close this as it's fixed in #490 and I'm still running my own fork.

from emacs-ng.