You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
rustdoc: Enabling mergeable CCI broke the search functionality for the stdlib and for our internal docs #162334
I don't know if it's a rustdoc regression, a bootstrap regression or if it's only exhibited now because we've changed the compiler crates in some way like adding some public re-exports of rustc_ast::ast or rustc_ast::ast::Attribute to some other compiler crate (CC #145870 which is only tangentially related).
Most likely due to #161716? That PR got merged 2 days ago, so part of nightly-2026-09-03 (rustc version) aka nightly-2026-09-04 (toolchain name) and it's very plausible that it messed with the search indices.
Given that the fix-forward is in rustdoc rather than bootstrap, I’m inclined to follow the usual default practice of rolling back to a known-OK state and then working forwards from there, to avoid the risk of having an even bigger mess to revert.
Activity
fmease commentedon Sep 5, 2026
This feels like a regression.
I don't know if it's a rustdoc regression, a bootstrap regression or if it's only exhibited now because we've changed the compiler crates in some way like adding some public re-exports of
rustc_ast::astorrustc_ast::ast::Attributeto some other compiler crate (CC #145870 which is only tangentially related).theemathas commentedon Sep 5, 2026
This also affects searching of the standard library (as of
0ed41eb41 2026-09-04). Searching for "PathBuf" doesn't find thePathBuftype. https://doc.rust-lang.org/nightly/std/index.html?search=PathBuffmease commentedon Sep 5, 2026
Most likely due to #161716? That PR got merged 2 days ago, so part of nightly-2026-09-03 (rustc version) aka nightly-2026-09-04 (toolchain name) and it's very plausible that it messed with the search indices.
cc @notriddle
3 remaining items
jieyouxu commentedon Sep 5, 2026
I'll try this locally with a revert 🤔 If it's due to #161716 I'd like to revert / reland that, it's not trivial certainly.
jieyouxu commentedon Sep 5, 2026
I looked at the
rustc-docslocally too when reviewing the PR but also did not notice the difference 😓jieyouxu commentedon Sep 5, 2026
Posted a revert: #162339
notriddle commentedon Sep 5, 2026
And a fix #162346
Do we still want to land the revert?
Zalathar commentedon Sep 6, 2026
Given that the fix-forward is in rustdoc rather than bootstrap, I’m inclined to follow the usual default practice of rolling back to a known-OK state and then working forwards from there, to avoid the risk of having an even bigger mess to revert.
notriddle commentedon Sep 6, 2026
Okay, let's do that. The fix-forward shouldn't conflict with the revert, anyway.
Rollup merge of rust-lang#162339 - jieyouxu:revert-mergeable-cci, r=K…
Unrolled build for #162339 in rollup 162362
Add regression test for PathBuf not showing up
Rollup merge of rust-lang#162346 - notriddle:rustdoc-merge-type, r=lo…
Rollup merge of rust-lang#162346 - notriddle:rustdoc-merge-type, r=lo…
Unrolled build for #162346 in rollup 162612