Open rebcabin opened 1 year ago
looks like the meaning of kind
is not always the number of bytes
.
The meaning of "kind" is not bytes, just the kind of integer. We used 1, 2, 4, 8, 16 for int sizes in bytes, because that is what gfortran uses. But we can change it.
We don't need to name it kind either.
Yes, utf8 is mostly 1 byte, but can be up to 4 bytes, variable encoding.
Let's keep this issue open, and document in our ASR docs. In fact I want to move all comments from ASR.asdl into our docs.
I did a nice job in MASR today to handle this in a flexible way.
The character kind is listed as 1 byte and utf8. Isn't utf8 a variable-length type, up to 4 bytes?
Also, as an aside, there is an awkwardness in the following comment: