Bug 1409255 - Replace all uses of the heapsize crate with malloc_size_of

[SimonSapin]

Patch from Nicholas Nethercote on https://bugzilla.mozilla.org/show_bug.cgi?id=1409255

---

Servo currently uses heapsize, but Stylo/Gecko use malloc_size_of.
malloc_size_of is better -- it handles various cases that heapsize does not
-- so this patch changes Servo to use malloc_size_of.

This patch makes the following changes to the malloc_size_of crate.

• Adds MallocSizeOf trait implementations for numerous types, some built-in
  (e.g. VecDeque), some external and Servo-only (e.g. string_cache).

• Makes enclosing_size_of_op optional, because vanilla jemalloc doesn't
  support that operation.

• For HashSet/HashMap, falls back to a computed estimate when
  enclosing_size_of_op isn't available.

• Adds an extern "C" malloc_size_of function that does the actual heap
  measurement; this is based on the same functions from the heapsize crate.

This patch makes the following changes elsewhere.

• Converts all the uses of heapsize to instead use malloc_size_of.

• Disables the "heapsize"/"heap_size" feature for the external crates that
  provide it.

• Removes the HeapSizeOf implementation from hashglobe.

• Adds ignore annotations to a few Rc/Arc, because malloc_size_of
  doesn't derive those types, unlike heapsize.

---

This change is 

[highfive]

Heads up! This PR modifies the following files:

• @bholley: ports/geckolib/glue.rs, components/style/values/generics/flex.rs, components/style/stylesheet_set.rs, components/style/rule_tree/mod.rs, components/style/values/specified/length.rs and 86 more
• @jgraham: components/webdriver_server/Cargo.toml
• @canaltinova: components/style/values/generics/flex.rs, components/style/stylesheet_set.rs, components/style/rule_tree/mod.rs, components/style/values/specified/length.rs, components/style/stylesheets/origin.rs and 83 more
• @emilio: ports/geckolib/glue.rs, components/script/dom/webglframebuffer.rs, components/style/values/generics/flex.rs, components/style/stylesheet_set.rs, components/style/rule_tree/mod.rs and 97 more
• @fitzgen: components/script/dom/request.rs, components/script/dom/xmlhttprequest.rs, components/script/dom/webglframebuffer.rs, components/script/dom/crypto.rs, components/devtools_traits/Cargo.toml and 127 more
• @KiChjang: components/script/dom/request.rs, components/script/dom/xmlhttprequest.rs, components/script/dom/webglframebuffer.rs, components/script/dom/crypto.rs, components/script/dom/vrdisplaycapabilities.rs and 136 more
• @asajeffrey: components/webdriver_server/Cargo.toml
• @paulrouget: components/servo_arc/Cargo.toml, components/servo_arc/lib.rs

[highfive]

Warning

• These commits modify unsafe code. Please review it carefully!

[jdm]

./components/malloc_size_of/lib.rs:153: Line is longer than 120 characters

[SimonSapin]

CC @nnethercote

I split the long line to fix tidy.

Nit: size_of_is_0! out of context sounds like std::mem::size_of(). Should that macro be renamed malloc_size_of_is_0!?

@bors-servo try

---

Reviewed 269 of 269 files at r1, 1 of 1 files at r2.
Review status: all files reviewed at latest revision, 2 unresolved discussions, some commit checks failed.

---

components/malloc_size_of/lib.rs, line 345 at r1 (raw file):

    fn shallow_size_of(&self, ops: &mut MallocSizeOfOps) -> usize {
        if let Some(elem) = self.get(0) {
            // This assumes that element 0 is the start of the heap block.

Element 0 is not necessarily at the start. VecDeque is a ring buffer internally.

---

components/script/dom/eventtarget.rs, line 121 at r1 (raw file):

impl MallocSizeOf for EventListenerType {
    fn size_of(&self, _ops: &mut MallocSizeOfOps) -> usize {
        // FIXME: Rc<T> isn't MallocSizeOf and we can't ignore it due to #6870 and #6871

This doesn’t apply anymore, does it? Shouldn’t this be a #[derive] with ignore attributes where appropriate?

---

Comments from Reviewable

[bors-servo]

⌛ Trying commit 4c7e747 with merge 39d0c34...

[bors-servo]

☔ The latest upstream changes (presumably #18924) made this pull request unmergeable. Please resolve the merge conflicts.

[bors-servo]

☔ The latest upstream changes (presumably #18704) made this pull request unmergeable. Please resolve the merge conflicts.

[SimonSapin]

Looks like (in AppVeyor) this is missing some imports on Windows:

error[E0425]: cannot find function `GetProcessHeap` in this scope
   --> C:\projects\servo\components\malloc_size_of\lib.rs:163:16
    |
163 |     let heap = GetProcessHeap();
    |                ^^^^^^^^^^^^^^ not found in this scope
error[E0425]: cannot find function `HeapValidate` in this scope
   --> C:\projects\servo\components\malloc_size_of\lib.rs:165:8
    |
165 |     if HeapValidate(heap, 0, ptr) == 0 {
    |        ^^^^^^^^^^^^ not found in this scope
error[E0425]: cannot find function `HeapSize` in this scope
   --> C:\projects\servo\components\malloc_size_of\lib.rs:169:5
    |
169 |     HeapSize(heap, 0, ptr) as usize
    |     ^^^^^^^^ not found in this scope
error: aborting due to 3 previous errors
error: Could not compile `malloc_size_of`.

heapsize has this dependency:

[target.'cfg(windows)'.dependencies]
kernel32-sys = "0.2.1"

[nnethercote]

Nit: size_of_is_0! out of context sounds like std::mem::size_of(). Should that macro be renamed malloc_size_of_is_0!?

I filed #18932 to fix that.

[bors-servo]

☔ The latest upstream changes (presumably #18932) made this pull request unmergeable. Please resolve the merge conflicts.

[nnethercote]

#18938 has the updated version of this change.

[SimonSapin]

Re-opening temporarily to get Reviewable to show a diff.