Legacy TSS in the old and in the new situation

When UTF-T was introduced, legacy TSS remained supported. However, the implementation of the support was drastically changed, together with the implementation of the new UTF-T support. In the table below the differences between the old and the new support of legacy TSS are indicated

Legacy TSS in the old situation	Legacy TSS in the new situation
Conversion from TSS to any other character set recognizes only TSS corresponding to the character set of the current locale. For example: In a Simplified Chinese locale only multibyte TSS in the range 9B 25 pp qq is recognized, and multibyte TSS in the range 9B 27 pp qq is not recognized.	A conversion from TSS to any other character set recognizes all TSS characters. Ambiguous TSS characters are interpreted according to the character set of the current locale (if that is a single byte character set) or according to the ISO 8859-1 character set (otherwise).
Conversion from any Unicode encoding to any other character set always contains a conversion from single word UTF-16 to the character set of the current locale, using an external table of 128 KB in the file $BSE/lib/Unicode/<charset>.U2N, loaded in shared memory.	No external Unicode tables are required. The amount and size of the internal tables is less than the amount and size of the previously used external tables. The internal tables are integrated as read-only data into the executables of the porting set.
Conversion from any other character set to a Unicode encoding always contains a conversion from the character set of the current locale to single word UTF-16, using an external table of 128 KB in the file $BSE/lib/Unicode/<charset>.N2U, loaded in shared memory	See above.
The supported character sets are defined in some (password protected) sessions, dumped in file $BSE/lib/tss6.2 and compiled to runtime file $BSE/lib/tss_c6.2. This file is used for conversion between TSS and the native character set of the current locale.	No external file is required for a conversion between TSS and any native character set. The conversion is almost completely algorithmic. Only for single byte Windows Code Pages small tables are required, integrated as read-only data into the executables of the porting set.
The file $BSE/lib/tss_c6.2 contains also references to $BSE/lib/locale/<locale name>, used for character width determination.	No external file is required for character width determination.