I've been meditating on [ Bug #127473 ] isql always ignores charset NONE in
metadata extraction and I've got a question.

I would appreciate any thoughts on the following scenario:

Database A has no default character set.

Database A has the table T1 with field F1 with no character set specified.

Database A has the table T2 with field F2 with character set NONE explicitly
specified.


Database B uses the default character set WIN1251.


Using ISQL to copy the table structure for T1 into Database B, what
character set should be used for the field F1?

Using ISQL to copy the table structure for T2 into Database B, what
character set should be used for the field F2?




And, for extra credit, how about:

Database C uses the default character set WIN1251

Database C has the table T3 with field F3 with no character set specified
(will be defined with WIN1251)

Database C has the table T4 with field F4 with character set NONE explicitly
specified.


Database D uses the default character set CYRL.


Using ISQL to copy the table structure for T3 into Database D, what
character set should be used for the field F3?

Using ISQL to copy the table structure for T4 into Database D, what
character set should be used for the field F4?



Should fields with no explicit character set in one database assume the
default character set of the target database?

Should fields explicitly or implicitly using the default character set of
one database assume the default character set of the target database?

Thanks,
...pat