Run WPT with Layout 2020 on CI

[SimonSapin]

… and gate PRs on the result.

[jdm]

r? @nox

[nox]

@bors-servo r+

[bors-servo]

📌 Commit 858bc5a has been approved by nox

[SimonSapin]

@nox We’ve got intermittent failures :(

  ▶ TIMEOUT [expected CRASH] /css/CSS2/box-display/block-in-inline-relpos-001.xht
  │ 
  │ 
  │ 
  │ 
  │ No protocol specified
  │ assertion failed: self.font_size.map_or(true, |f| f == font_size) (thread LayoutThread PipelineId { namespace_id: PipelineNamespaceId(1), index: PipelineIndex(1) }, at components/style/rule_cache.rs:29)
  │ stack backtrace:
  │    0: servo::main::{{closure}}
  │    1: std::panicking::rust_panic_with_hook
  │              at src/libstd/panicking.rs:468
  │    2: std::panicking::begin_panic
  │    3: style::values::specified::length::FontRelativeLength::to_computed_value
  │    4: style::values::computed::length::<impl style::values::computed::ToComputedValue for style::values::specified::length::LengthPercentage>::to_computed_value
  │    5: style::properties::longhands::margin_left::cascade_property
  │    6: style::properties::cascade::Cascade::apply_properties
  │    7: _ZN5style10properties7cascade13cascade_rules17h96faf43c84170c77E.llvm.7578349383558476548
  │    8: style::stylist::Stylist::cascade_style_and_visited
  │    9: style::style_resolver::StyleResolverForElement<E>::cascade_primary_style
  │   10: style::style_resolver::StyleResolverForElement<E>::resolve_style_with_default_parents
  │   11: style::traversal::compute_style
  │   12: <layout::traversal::RecalcStyle as style::traversal::DomTraversal<E>>::process_preorder
  │   13: style::driver::traverse_dom
  │   14: layout_thread::LayoutThread::handle_reflow
  │   15: profile_traits::time::profile
  │   16: layout_thread::LayoutThread::handle_request_helper
  │   17: layout_thread::LayoutThread::start
  │   18: profile_traits::mem::ProfilerChan::run_with_memory_reporting
  │   19: std::sys_common::backtrace::__rust_begin_short_backtrace
  │   20: _ZN3std9panicking3try7do_call17ha66e0d91de444c54E.llvm.1363856088101530016
  │   21: __rust_maybe_catch_panic
  │              at src/libpanic_unwind/lib.rs:81
  │   22: core::ops::function::FnOnce::call_once{{vtable.shim}}
  │   23: <alloc::boxed::Box<F> as core::ops::function::FnOnce<A>>::call_once
  │              at /rustc/1bd30ce2aac40c7698aa4a1b9520aa649ff2d1c5/src/liballoc/boxed.rs:942
  │   24: <alloc::boxed::Box<F> as core::ops::function::FnOnce<A>>::call_once
  │              at /rustc/1bd30ce2aac40c7698aa4a1b9520aa649ff2d1c5/src/liballoc/boxed.rs:942
  │       std::sys_common::thread::start_thread
  │              at src/libstd/sys_common/thread.rs:13
  │       std::sys::unix::thread::Thread::new::thread_start
  │              at src/libstd/sys/unix/thread.rs:79
  │   25: start_thread
  │   26: __clone
  │ 
  │ [2019-11-26T21:59:23Z ERROR servo] assertion failed: self.font_size.map_or(true, |f| f == font_size)
  │ Redirecting call to abort() to mozalloc_abort
  └ 

[jdm]

@bors-servo r-
Removing from the queue so the authors can decide what to do about it at a reasonable hour of the day.

[nox]

@SimonSapin I've seen this exactly once and could never reproduce it again.

[SimonSapin]

I was able to reproduce with ./mach test-wpt --release --layout-2020 --repeat-until-unexpected, though after several repetitions. On CI however I didn’t get a single successful run:

• https://community-tc.services.mozilla.com/tasks/YffcdLk6QEaah8nyMKRVEQ
• https://community-tc.services.mozilla.com/tasks/Vb6JhK3KQC2E1KwGx4maKQ
• https://community-tc.services.mozilla.com/tasks/e4zH6iqVSN6C3gap3hsqWg
• https://community-tc.services.mozilla.com/tasks/c-boqrirREO7nQpppdPvcw

These may not all be with the same expectations, I don’t remember exactly. But the problem is that the result for block-in-inline-relpos-*.xht is sometimes "crash" and sometimes "timeout". In both cases there’s a stack trace assertion failed: self.font_size.map_or(true, |f| f == font_size), but maybe after that the process sometimes hangs instead of exiting?

So we have two issues (and would ideally fix both):

• A failing debug assertion
• Intermittent hangs after a LayoutThread panics

[nox]

I do not think this is an issue in the layout code, rather I suspect that we are doing something wrong when recalculating styles before we do the layout. I'm not familiar with this part of Stylo and have little idea where to start looking to debug the issue.

[nox]

I added a commit that makes the assertion more helpful, and I got the following panic:

 0:03.59 pid:11674 assertion failed: `(left == right)`
 0:03.59 pid:11674   left: `NonNegative(16.0 px)`,
 0:03.59 pid:11674  right: `NonNegative(20.0 px)` (thread LayoutThread PipelineId { namespace_id: PipelineNamespaceId(1), index: PipelineIndex(1) }, at components/style/rule_cache.rs:30)

[nox]

@bors-servo try=wpt

[bors-servo]

⌛ Trying commit 9b4baf2 with merge 0787371...

[SimonSapin]

Homu forgot about this try because I pushed another commit, but results will still be at https://community-tc.services.mozilla.com/tasks/groups/bMEFWdVlTkW2hTgOIlFDFA

(This reachable through the orange/green/red dot near “bors-servo added a commit that referenced this pull request”.)

[SimonSapin]

It’s green \o/

@bors-servo r=nox,SimonSapin

[bors-servo]

📌 Commit 9a4bc23 has been approved by nox,SimonSapin

[bors-servo]

⌛ Testing commit 9a4bc23 with merge 83bb967...

[bors-servo]

☀️ Test successful - status-taskcluster
Approved by: nox,SimonSapin
Pushing 83bb967 to master...