I have adapted my application to use Unicode controls (Tnt) and the
Unicode Private Use Area (F100 to F8FF, in the E000 to F8FF 2 byte
Block) glyphs I have created for my WIN1252 encoded font.

I had presumed that in order to complete this evolution, I would also
have to change my Firebird Character Set to UNICODE_FSS.

Having been prevented from making this v1.5+ UNICODE_FSS change due to
index size restrictions, I have experimented with using my existing
WIN1252 Firebird Character Set and it seems that my PUA (Private Use
Area) characters are successfully making the round trip through FB
v.1.5+.

In evaluating why this might work, I am beginning to suspect that the
primary reason for my 2 byte characters returning correctly is due to
my Unicode controls.

Is it reasonable to expect that the employment of WIN1252 in my FB 1.5+
(or greater) will be sufficient to maintain reliable storage of my 2
byte PUA characters. I.E., I presume that Firebird is not going to
change any bytes, and that my use (or not) of reliable Unicode controls
is the only significant issue?