A spreadsheet file with tab separated variables/fields, e.g.
$ od -a tt.csv
0000000 1 ht 2 nl 3 ht 4 nl
when opened in Calc 220.127.116.11 shows up as just cell 1A in Lohit Hindi
Calc Tools -> Options -> Language Settings -> Languages
shows everything as Default - English (USA)
This language shift problem doesn't occur if I use the .ods file type.
It was set for Unicode (UTF-16) apparently as the installation default.
I changed it to UTF-8 and now all is well.
Can we therefore close this issue?
Yes. I think it's settled - but people should be aware that installation can be set for UTF-16 (unusual in my experience) and they need to change to UTF-8.
UTF-16 is not an "installation default". In fact, in 4.3.5 the encoding defaults to UTF-8 even if system encoding is UTF-16 (Windows). The only way to make an encoding stick is to select it explicitly. The easiest way is not to touch the encoding selection at all and rely on detection--that works fine for UTF-16/UTF-8 anyway.
I still see this occasionally with v18.104.22.168
I need to manually change back to UTF-8 when I see the Chinese text in the preview.
There's clearly a bug in the way it determines the file format.