Perf summaries completely ignore rustdoc

[jyn514]

This perf run had over 100% regressions for rustdoc, but rust-timer said there were no significant changes.

Originally posted by @rust-timer in rust-lang/rust#88219 (comment)

[jyn514]

Relevant code seems to be

rustc-perf/site/src/github.rs

Line 704 in 5c8e64a

fn comparison_is_relevant(confidence: ComparisonConfidence, is_master_commit: bool) -> bool {

/

rustc-perf/site/src/comparison.rs

Line 270 in 5c8e64a

pub fn confidence(&self) -> ComparisonConfidence {

.

@rylev do you know when you'll get a chance to look into this? Posting "no relevant changes" when there's a regression is worse than not posting anything at all IMO.

[jyn514]

Or alternatively, how can I run that code locally so I can debug it? cargo run --bin site seems to run the site, not the bot.

[jyn514]

Ugh, they're being deliberately filtered out for some reason.

rustc-perf/site/src/comparison.rs

Lines 571 to 573 in 5c8e64a

	fn get_individual_comparisons(&self) -> impl Iterator<Item = &TestResultComparison> {
	self.statistics.iter().filter(|b| b.profile != Profile::Doc)
	}

We discussed removing this in https://rust-lang.zulipchat.com/#narrow/stream/247081-t-compiler.2Fperformance/topic/triage.202021.2009.2007/near/252361701 so I'm going to go ahead and make a PR that doesn't filter them out.