{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreice4kt2fbodeasng7u2c4ufqelsjhto23y5uwdock52jx4mvspr5u",
    "uri": "at://did:plc:ivbknywyskln22er3nkssdhl/app.bsky.feed.post/3mllthvwftld2"
  },
  "path": "/t/interiorassign-trait-a-b/24291?page=2#post_26",
  "publishedAt": "2026-05-11T16:46:54.000Z",
  "site": "https://internals.rust-lang.org",
  "textContent": "I'll add to the general negative feeling here by pointing out that with RefCell this operation can _panic_ and so I might not want to use it anyway (because it makes the panic harder to spot). And then we're talking about syntactic sugar that only really works for Cell and maybe UnsafeCell—it doesn't, as proposed, cover the HashMap case; as described it scares me a little with RefCell; it would be _deceptively_ simple syntax for Mutex and RwLock even ignoring poisoning; and what other types, even in crates outside std, do we expect to use it? I know it sounds like \"sour grapes\", but sometimes a bit of friction for something that shouldn't be your first choice anyway is okay.",
  "title": "InteriorAssign trait { a := b }"
}