Opening a bug here for me to remember.

In the Dutch locl feature, there are substitutions to access ijacute forms:

sub iacute j by iacute_j.loclNLD;
sub Iacute J by Iacute_J.loclNLD;

Generally it is not ideal to substitute two user-entered glyphs for a single glyph in a monospace coding environment. A better implementation here is: sub iacute j' by jacute. This will mean removal of the loclNLD glyph mentioned above, and addition of jacute.

In Catalan there are substitutions to access the Ŀ and ŀ:

sub  l' periodcentered' l by ldot;
sub  L' periodcentered' L by Ldot;

While this is a standard way to access the Ŀ and ŀ, I think that in this case it would make more sense to remove these substitutions and just leave the glyphs to look like l·l. These are atomic characters which can be accessed directly.