Johan Malm

> but that would take down the environment, wouldn't it? No. It just re-loads `rc.xml` and `environment`. When reading the latter is just does some `setenv()` based on the key/value...

Is it enough to just call wlr_text_input_manager_v3_create() from server_init()?

> This is my work-in-progress implementation: [link](https://github.com/tokyo4j/labwc/tree/input-method) Great stuff. Thanks for working on this.

Implemented with https://github.com/labwc/labwc/pull/1548

Thanks for reporting. These are not too picky at all. We care about getting these sort of details right. 1. Not sure about the first one, but sounds like we...

I'm not sure what's gone wrong there, but that's abnormal behavior. Maybe some socket/pipe/other resource has not been freed properly. Is the problem repeatable? Do you get it with default...

I'm at a loss with this. Any detective work much appreciated.

Marked as draft in response to https://github.com/labwc/labwc/issues/1454#issuecomment-1911046196

@droc12345 Any chance of an ASAN log? ``` meson setup build/ meson configure -Db_sanitize=address,undefined build/ meson compile -C build/ ./build/labwc 2>log.txt ``` Then force crash and paste `log.txt` here.

Thanks for thinking about this. Just having a quick look, `view_next` and `view_prev` actually stop at head which doesn't seem right for `desktop_cycle_view()`. We'll probably have to come up with...