|
|
|
|
|
by estebank
540 days ago
|
|
The UTF-32 thing was confusing me. It is mentioned in multiple places but never explicit if it was a temporary design decision to ease the transition, perfectly reasonable, or the design going forward in perpetuity, in which case I would love to understand why something like bstr (WTF-8 under the covers) doesn't fit the bill. I've gathered so far that it might be the former, but if it is the later I would still like to know more. |
|
Unfortunately it's pervasive throughout the entire codebase (which does a lot of string shuffling).
So it's a historical mistake, but not something to be fixed in the same step as switching the implementation language.
bstr was mentioned as a direction and is a possibility.