Compile with -Wfloat-conversion

This means we'll get warnings whenever a floating point value looses
precision, e.g. gets implicitly casted to an integer. It also warns when
implicitly casting double's to float's, which arguably is less of a
problem, but there are no warning for just float/double to int.

This would have caught
https://gitlab.gnome.org/GNOME/mutter/-/issues/3530.

Part-of: <https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/3822>
This commit is contained in:
Jonas Ådahl
2024-06-18 10:23:00 +02:00
committed by Sebastian Wick
parent d5bc883712
commit 422ee4515d
157 changed files with 1313 additions and 1161 deletions

View File

@ -676,10 +676,10 @@ meta_input_settings_native_set_tablet_area (MetaInputSettings *settings,
gfloat offset_x;
gfloat offset_y;
scale_x = 1. / (1. - (padding_left + padding_right));
scale_y = 1. / (1. - (padding_top + padding_bottom));
offset_x = -padding_left * scale_x;
offset_y = -padding_top * scale_y;
scale_x = (float) (1.0 / (1.0 - (padding_left + padding_right)));
scale_y = (float) (1.0 / (1.0 - (padding_top + padding_bottom)));
offset_x = (float) (-padding_left * scale_x);
offset_y = (float) (-padding_top * scale_y);
gfloat matrix[6] = { scale_x, 0., offset_x,
0., scale_y, offset_y };