|
|
|
|
|
|
by adam-p
411 days ago
|
|
|
Valid question, and I think you're right in the abstract and most of the time. But I also think you end up with a mismatch. What's the concrete spec for the limit if you've only got 10x storage per grapheme cluster? Probably you end providing the limit in bytes. That's fine, but it's no longer the "hybrid counting" thing anymore. |
|
|