On Sat, 30 Sep 2023 23:30:53 GMT, Thiago Milczarek Sayao <tsa...@openjdk.org> wrote:
>> It looks like a condition between calls from `set_visible` to >> `request_focus`. When `request_focus` is called, the window is not yet >> mapped (map is requested on `set_visible`). >> >> 200x200 is the default gtk window size if none is specified. Since the >> window is presented at the requested size and then changes to 200x200 I >> would guess the size request on `gtk_window_present` (which is called on >> `request_focus`) is being queued and released later. But I'm not really sure. >> >> With this change immediate `request_focus` after the window is created will >> be ignored. But the default is to request focus on map, so it will work. >> `request_focus` will only be useful when called after the window is mapped >> (presented on the screen). >> >> This also fixes [JDK-8316425](https://bugs.openjdk.org/browse/JDK-8316425) > > Thiago Milczarek Sayao has updated the pull request incrementally with two > additional commits since the last revision: > > - Re-enable tests > - Fix fullscreen position I strongly suspect of what is happening: `WindowContextTop` constructor calls `glass_gtk_configure_transparency_and_realize` which calls `gtk_widget_realize`. This is probably creating the underlying gdk window before gtk expects and is messing things up. ------------- PR Comment: https://git.openjdk.org/jfx/pull/1249#issuecomment-1763116193
