I suggest that U+0218, U+0219, U+021a, and U+021b be added
to 0x53, 0x73, 0x54, and 0x74, respectively, in def7_uni.tbl.
Originally, the Unicode standard included only the "s/t with
cedilla" characters (U+015e, U+015f, U+0162, U+0163, already
listed in def7_uni.tbl). Later, the Romanian "s/t with
comma" were split off and assigned separate positions.
Incidentally, the cedilla/comma variants share the same code
points in ISO-8859-2 (cedilla) and ISO-8859-16 (comma).
On the web, both encodings are common in Romanian text,
the older ones (along with ISO-8859-2) being better
supported, the newer ones (along with ISO-8859-16)
being considered more correct.