Review all legacy issues which should be fixed by new chardata support

[pp-mo]

Additional to #6898

Check the status of issues we expect to be solved by #6898 , and close where possible

Key issues are listed in #6309
But there may be others.
That list repeated here ...

Contact original authors of Issues for potential datasets to test.

• Iris cannot read NetCDF4 strings stored in NetCDF variables #4101
• Cubes with string data won't save, or index #4412
• Error comparing cubes with string content. #5362
• Errors in netcdf save code relating to string data. #4502
• Cannot save non-ASCII characters to NetCDF #5125
• https://github.com/pp-mo/ncdata/issues/111

[ukmo-ccbunney]

After discussion, we think that just supporting ASCII, UTF8, 16 & 32 encoding is more than adequate, pending of course any other encodings we might find from users files raised in the issues listed above.

[ukmo-ccbunney]

UTF16 saving needs to be properly implemented, as per utf32. See notes on #6898 regarding Endianness!