Back more metadata using per-query tables

[cjgillot]

r? @ghost

[cjgillot]

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

[bors]

⌛ Trying commit 4442ef44c027a775c29820c4906cd7ae295c1323 with merge 11295abb90a0b335673636cbcf583aee7b99d98b...

[bors]

☀️ Try build successful - checks-actions
Build commit: 11295abb90a0b335673636cbcf583aee7b99d98b (11295abb90a0b335673636cbcf583aee7b99d98b)

[rust-timer]

Queued 11295abb90a0b335673636cbcf583aee7b99d98b with parent b8c56fa, future comparison URL.

[rust-timer]

Finished benchmarking commit (11295abb90a0b335673636cbcf583aee7b99d98b): comparison url.

Summary: This benchmark run shows 38 relevant improvements 🎉 to instruction counts.

• Average relevant improvement: -0.6%
• Largest improvement in instruction counts: -0.9% on full builds of await-call-tree doc

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR led to changes in compiler perf.

@bors rollup=never
@rustbot label: +S-waiting-on-review -S-waiting-on-perf -perf-regression

[cjgillot]

r? rust-lang/compiler

[lcnr]

after a quick skim this lgtm

r? @petrochenkov

[petrochenkov]

@bors r+
This slightly increases max-rss though.

[bors]

📌 Commit 7afcf9f has been approved by petrochenkov

[eddyb]

Very nice! I wonder if some time soon this whole enum could be subsumed by just DefKind.

[bors]

⌛ Testing commit 7afcf9f with merge 7ccfe2f...

[bors]

☀️ Test successful - checks-actions
Approved by: petrochenkov
Pushing 7ccfe2f to master...

[rust-timer]

Finished benchmarking commit (7ccfe2f): comparison url.

Summary: This benchmark run shows 42 relevant improvements 🎉 but 4 relevant regressions 😿 to instruction counts.

• Average relevant regression: 0.8%
• Average relevant improvement: -0.6%
• Largest improvement in instruction counts: -1.0% on full builds of wf-projection-stress-65510 doc
• Largest regression in instruction counts: 0.9% on incr-unchanged builds of externs opt

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

Next Steps: If you can justify the regressions found in this perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please open an issue or create a new PR that fixes the regressions, add a comment linking to the newly created issue or PR, and then add the perf-regression-triaged label to this PR.

@rustbot label: +perf-regression

[cjgillot]

The perf results are overall green for rustdoc. Regression is localized to a single benchmark externs for debug/opt configurations, but not check. This regression is real, but on a stress test: loading native_library_kind is slower.
The size of on-disk metadata has grown: the encoding is less space-efficient than it used to be (5% larger for externs).