Evgenii Pashkin

@sdroege Thanks for information

Problem that "(not nullable)" annotation in code still saves as the absence of "nullable" attribute in .gir: (no nullable="0" in gtk-rs or gstreamer also no not_nullable or something like this),...

@sdroege Thanks for confirming this and adding issue. I seen [`if return_.nullable and not return_.not_nullable:`](https://gitlab.gnome.org/GNOME/gobject-introspection/commit/10cb665fee2cc378dd2f13bad16e6384836a8b16#80c64e250d9a3c544e5e178e6814a87ff9f9d299_218_217) but was to lazy to try really run introspection.

There also third case https://github.com/gtk-rs/gir/blob/master/src/analysis/return_value.rs#L100 seems it can changed ```diff +++ b/src/analysis/return_value.rs @@ -97,7 +97,7 @@ pub fn analyze( } parameter = Some(library::Parameter { typ: type_tid, - nullable: nullable_override.unwrap_or(Nullable(false)), +...

After checking on current gir files I see only 2 changes: https://developer.gnome.org/gdk3/stable/gdk3-Cursors.html#gdk-cursor-new-from-name https://developer.gnome.org/glib/stable/glib-GDateTime.html#g-date-time-new-from-iso8601 both is good

For signals we have `fill_empty_signals_c_types` in `library_postprocessing.rs`, maybe we need add this fields too, but better way fix it upstream, as seems it so frequently as for signals

*NOT frequently as for signals

While names are same this crate is not GObjectIntrospection, but code generator that used its results.

This crate is code generator

FromGlibBorrow used mainly in trampolines https://github.com/gtk-rs/gtk/blob/master/src/auto/tree_view.rs#L1070