Dave Anderson on Nostr: In a less fancy implementation this would be easy to figure out, because the ...
In a less fancy implementation this would be easy to figure out, because the collation algorithm specified by Unicode TR10 has "normalize the input" as step one.
But x/text/collate is very fancy indeed, and implements an optimization where the normalization is interleaved with the rest of the collation logic. This is very fast and efficient, but means collate.Force just mysteriously sets 'ignore[colltab.Identity] = false' on the collator internals, which will do... something during compares.
Published at
2024-06-24 17:05:46Event JSON
{
"id": "3a951556e53e0d22e8c74b5227ed24c455024006e8fef6052ec88c23aa9f3abf",
"pubkey": "dad113f50e47d6bffb79f9a69006ffd969c26e636e5aed2ae597b0407093175f",
"created_at": 1719248746,
"kind": 1,
"tags": [
[
"proxy",
"https://hachyderm.io/@danderson/112672685862059273",
"web"
],
[
"p",
"dad113f50e47d6bffb79f9a69006ffd969c26e636e5aed2ae597b0407093175f"
],
[
"e",
"f6fec5364f95ac439e50a33f001af7f03c721fc30fbebd05056311a0558a5744",
"",
"root"
],
[
"proxy",
"https://hachyderm.io/users/danderson/statuses/112672685862059273",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://hachyderm.io/users/danderson/statuses/112672685862059273",
"pink.momostr"
],
[
"expiration",
"1721856494"
]
],
"content": "In a less fancy implementation this would be easy to figure out, because the collation algorithm specified by Unicode TR10 has \"normalize the input\" as step one.\n\nBut x/text/collate is very fancy indeed, and implements an optimization where the normalization is interleaved with the rest of the collation logic. This is very fast and efficient, but means collate.Force just mysteriously sets 'ignore[colltab.Identity] = false' on the collator internals, which will do... something during compares.",
"sig": "91f343f2565205cf282e2a2c91cde66d86c82d20721dd3bbb4889714f70c5ad00b3d8240188f25e8ea6b9a0392d78a4558a58e90ec7587be9e878d5110e8e980"
}
