Jeff Clites <[EMAIL PROTECTED]> wrote:
1) My version handles the case of code points > 0xFFFF as well.
It's simple now to change that code to include this. I've not done it
yet to keep this patch smaller, which includes #28473 too. My version is
just smaller, cleaner, and faster ;)
Ha, ha, good!
One other thing occurred to me, to save a few bytes: When upscaling, rather than passing clength, we can pass (result->strlen + number of bytes left in cstring). You made the very good point that we know the max size possible, and since at that point we've just parsed over an escape sequence, we know the max final size is a little less than clength.
So config stuff applied + some bits from #28494.
Good!
JEff
