Auto merge of #887 - photoszzt:derive_hash, r=fitzgen

Can derive hash analysis

r? @fitzgen  Fix: #876 I haven't enable too much test yet. Enable the option for all tests changes more than 200 of them. I want some input on which test is good to enable or write a new test.

Author: bors-servo

src/codegen/mod.rs

@@ -13,7 +13,7 @@ use ir::comp::{Base, BitfieldUnit, Bitfield, CompInfo, CompKind, Field,
                FieldData, FieldMethods, Method, MethodKind};
 use ir::comment;
 use ir::context::{BindgenContext, ItemId};
-use ir::derive::{CanDeriveCopy, CanDeriveDebug, CanDeriveDefault};
+use ir::derive::{CanDeriveCopy, CanDeriveDebug, CanDeriveDefault, CanDeriveHash};
 use ir::dot;
 use ir::enum_ty::{Enum, EnumVariant, EnumVariantValue};
 use ir::function::{Abi, Function, FunctionSig};
@@ -1440,6 +1440,10 @@ impl CodeGenerator for CompInfo {
             }
         }
 
+        if item.can_derive_hash(ctx) {
+            derives.push("Hash");
+        }
+
         if !derives.is_empty() {
             attributes.push(attributes::derives(&derives))
         }
@@ -3394,12 +3398,23 @@ mod utils {
         )
             .unwrap();
 
+        // The actual memory of the filed will be hashed, so that's why these
+        // field doesn't do anything with the hash.
+        let union_field_hash_impl = quote_item!(&ctx.ext_cx(),
+            impl<T> ::$prefix::hash::Hash for __BindgenUnionField<T> {
+                fn hash<H: ::$prefix::hash::Hasher>(&self, _state: &mut H) {
+                }
+            }
+        )
+            .unwrap();
+
         let items = vec![union_field_decl,
                          union_field_impl,
                          union_field_default_impl,
                          union_field_clone_impl,
                          union_field_copy_impl,
-                         union_field_debug_impl];
+                         union_field_debug_impl,
+                         union_field_hash_impl];
 
         let old_items = mem::replace(result, items);
         result.extend(old_items.into_iter());

src/ir/analysis/derive_hash.rs

@@ -0,0 +1,337 @@
+//! Determining which types for which we can emit `#[derive(Hash)]`.
+
+use super::{ConstrainResult, MonotoneFramework, generate_dependencies};
+use std::collections::HashSet;
+use std::collections::HashMap;
+use ir::context::{BindgenContext, ItemId};
+use ir::item::IsOpaque;
+use ir::traversal::EdgeKind;
+use ir::ty::RUST_DERIVE_IN_ARRAY_LIMIT;
+use ir::ty::TypeKind;
+use ir::comp::Field;
+use ir::comp::FieldMethods;
+use ir::derive::CanTriviallyDeriveHash;
+use ir::comp::CompKind;
+
+/// An analysis that finds for each IR item whether hash cannot be derived.
+///
+/// We use the monotone constraint function `cannot_derive_hash`, defined as
+/// follows:
+///
+/// * If T is Opaque and layout of the type is known, get this layout as opaque
+///   type and check whether it can be derived using trivial checks.
+/// * If T is Array type, hash cannot be derived if the length of the array is
+///   larger than the limit or the type of data the array contains cannot derive
+///   hash.
+/// * If T is a type alias, a templated alias or an indirection to another type,
+///   hash cannot be derived if the type T refers to cannot be derived hash.
+/// * If T is a compound type, hash cannot be derived if any of its base member
+///   or field cannot be derived hash.
+/// * If T is a pointer, T cannot be derived hash if T is a function pointer
+///   and the function signature cannot be derived hash.
+/// * If T is an instantiation of an abstract template definition, T cannot be
+///   derived hash if any of the template arguments or template definition
+///   cannot derive hash.
+#[derive(Debug, Clone)]
+pub struct CannotDeriveHash<'ctx, 'gen>
+    where 'gen: 'ctx
+{
+    ctx: &'ctx BindgenContext<'gen>,
+
+    // The incremental result of this analysis's computation. Everything in this
+    // set cannot derive hash.
+    cannot_derive_hash: HashSet<ItemId>,
+
+    // Dependencies saying that if a key ItemId has been inserted into the
+    // `cannot_derive_hash` set, then each of the ids in Vec<ItemId> need to be
+    // considered again.
+    //
+    // This is a subset of the natural IR graph with reversed edges, where we
+    // only include the edges from the IR graph that can affect whether a type
+    // can derive hash or not.
+    dependencies: HashMap<ItemId, Vec<ItemId>>,
+}
+
+impl<'ctx, 'gen> CannotDeriveHash<'ctx, 'gen> {
+    fn consider_edge(kind: EdgeKind) -> bool {
+        match kind {
+            // These are the only edges that can affect whether a type can derive
+            // hash or not.
+            EdgeKind::BaseMember |
+            EdgeKind::Field |
+            EdgeKind::TypeReference |
+            EdgeKind::VarType |
+            EdgeKind::TemplateArgument |
+            EdgeKind::TemplateDeclaration |
+            EdgeKind::TemplateParameterDefinition => true,
+
+            EdgeKind::Constructor |
+            EdgeKind::Destructor |
+            EdgeKind::FunctionReturn |
+            EdgeKind::FunctionParameter |
+            EdgeKind::InnerType |
+            EdgeKind::InnerVar |
+            EdgeKind::Method => false,
+            EdgeKind::Generic => false,
+        }
+    }
+
+    fn insert(&mut self, id: ItemId) -> ConstrainResult {
+        trace!("inserting {:?} into the cannot_derive_hash set", id);
+
+        let was_not_already_in_set = self.cannot_derive_hash.insert(id);
+        assert!(
+            was_not_already_in_set,
+            "We shouldn't try and insert {:?} twice because if it was \
+             already in the set, `constrain` should have exited early.",
+            id
+        );
+
+        ConstrainResult::Changed
+    }
+}
+
+impl<'ctx, 'gen> MonotoneFramework for CannotDeriveHash<'ctx, 'gen> {
+    type Node = ItemId;
+    type Extra = &'ctx BindgenContext<'gen>;
+    type Output = HashSet<ItemId>;
+
+    fn new(ctx: &'ctx BindgenContext<'gen>) -> CannotDeriveHash<'ctx, 'gen> {
+        let cannot_derive_hash = HashSet::new();
+        let dependencies = generate_dependencies(ctx, Self::consider_edge);
+
+        CannotDeriveHash {
+            ctx,
+            cannot_derive_hash,
+            dependencies,
+        }
+    }
+
+    fn initial_worklist(&self) -> Vec<ItemId> {
+        self.ctx.whitelisted_items().iter().cloned().collect()
+    }
+
+    fn constrain(&mut self, id: ItemId) -> ConstrainResult {
+        trace!("constrain: {:?}", id);
+
+        if self.cannot_derive_hash.contains(&id) {
+            trace!("    already know it cannot derive Hash");
+            return ConstrainResult::Same;
+        }
+
+        let item = self.ctx.resolve_item(id);
+        let ty = match item.as_type() {
+            Some(ty) => ty,
+            None => {
+                trace!("    not a type; ignoring");
+                return ConstrainResult::Same;
+            }
+        };
+
+        if item.is_opaque(self.ctx, &()) {
+            let layout_can_derive = ty.layout(self.ctx).map_or(true, |l| {
+                l.opaque().can_trivially_derive_hash()
+            });
+            return if layout_can_derive {
+                trace!("    we can trivially derive Hash for the layout");
+                ConstrainResult::Same
+            } else {
+                trace!("    we cannot derive Hash for the layout");
+                self.insert(id)
+            };
+        }
+
+        if ty.layout(self.ctx).map_or(false, |l| l.align > RUST_DERIVE_IN_ARRAY_LIMIT) {
+            // We have to be conservative: the struct *could* have enough
+            // padding that we emit an array that is longer than
+            // `RUST_DERIVE_IN_ARRAY_LIMIT`. If we moved padding calculations
+            // into the IR and computed them before this analysis, then we could
+            // be precise rather than conservative here.
+            return self.insert(id);
+        }
+
+        match *ty.kind() {
+            // Handle the simple cases. These can derive hash without further
+            // information.
+            TypeKind::Void |
+            TypeKind::NullPtr |
+            TypeKind::Int(..) |
+            TypeKind::Enum(..) |
+            TypeKind::Named |
+            TypeKind::UnresolvedTypeRef(..) |
+            TypeKind::BlockPointer |
+            TypeKind::Reference(..) |
+            TypeKind::ObjCInterface(..) |
+            TypeKind::ObjCId |
+            TypeKind::ObjCSel => {
+                trace!("    simple type that can always derive Hash");
+                ConstrainResult::Same
+            }
+
+            TypeKind::Complex(..) |
+            TypeKind::Float(..) => {
+                trace!("    float cannot derive Hash");
+                self.insert(id)
+            }
+
+            TypeKind::Array(t, len) => {
+                if self.cannot_derive_hash.contains(&t) {
+                    trace!("    arrays of T for which we cannot derive Hash \
+                            also cannot derive Hash");
+                    return self.insert(id);
+                }
+
+                if len <= RUST_DERIVE_IN_ARRAY_LIMIT {
+                    trace!("    array is small enough to derive Hash");
+                    ConstrainResult::Same
+                } else {
+                    trace!("    array is too large to derive Hash");
+                    self.insert(id)
+                }
+            }
+
+            TypeKind::Pointer(inner) => {
+                let inner_type = self.ctx.resolve_type(inner).canonical_type(self.ctx);
+                if let TypeKind::Function(ref sig) = *inner_type.kind() {
+                    if !sig.can_trivially_derive_hash() {
+                        trace!("    function pointer that can't trivially derive Hash");
+                        return self.insert(id);
+                    }
+                }
+                trace!("    pointers can derive Hash");
+                ConstrainResult::Same
+            }
+
+            TypeKind::Function(ref sig) => {
+                if !sig.can_trivially_derive_hash() {
+                    trace!("    function that can't trivially derive Hash");
+                    return self.insert(id);
+                }
+                trace!("    function can derive Hash");
+                ConstrainResult::Same
+            }
+
+            TypeKind::ResolvedTypeRef(t) |
+            TypeKind::TemplateAlias(t, _) |
+            TypeKind::Alias(t) => {
+                if self.cannot_derive_hash.contains(&t) {
+                    trace!("    aliases and type refs to T which cannot derive \
+                            Hash also cannot derive Hash");
+                    self.insert(id)
+                } else {
+                    trace!("    aliases and type refs to T which can derive \
+                            Hash can also derive Hash");
+                    ConstrainResult::Same
+                }
+            }
+
+            TypeKind::Comp(ref info) => {
+                assert!(
+                    !info.has_non_type_template_params(),
+                    "The early ty.is_opaque check should have handled this case"
+                );
+
+                if info.kind() == CompKind::Union {
+                    if self.ctx.options().unstable_rust {
+                        trace!("    cannot derive Hash for Rust unions");
+                        return self.insert(id);
+                    }
+
+                    if ty.layout(self.ctx)
+                        .map_or(true,
+                                |l| l.opaque().can_trivially_derive_hash()) {
+                        trace!("    union layout can trivially derive Hash");
+                        return ConstrainResult::Same;
+                    } else {
+                        trace!("    union layout cannot derive Hash");
+                        return self.insert(id);
+                    }
+                }
+
+                let bases_cannot_derive = info.base_members()
+                    .iter()
+                    .any(|base|  !self.ctx.whitelisted_items().contains(&base.ty) ||
+                         self.cannot_derive_hash.contains(&base.ty));
+                if bases_cannot_derive {
+                    trace!("    base members cannot derive Hash, so we can't \
+                            either");
+                    return self.insert(id);
+                }
+
+                let fields_cannot_derive = info.fields()
+                    .iter()
+                    .any(|f| {
+                        match *f {
+                            Field::DataMember(ref data) => {
+                                !self.ctx.whitelisted_items().contains(&data.ty()) ||
+                                    self.cannot_derive_hash.contains(&data.ty())
+                            }
+                            Field::Bitfields(ref bfu) => {
+                                bfu.bitfields()
+                                    .iter().any(|b| {
+                                        !self.ctx.whitelisted_items().contains(&b.ty()) ||
+                                            self.cannot_derive_hash.contains(&b.ty())
+                                    })
+                            }
+                        }
+                    });
+                if fields_cannot_derive {
+                    trace!("    fields cannot derive Hash, so we can't either");
+                    return self.insert(id);
+                }
+
+                trace!("    comp can derive Hash");
+                ConstrainResult::Same
+            }
+
+            TypeKind::TemplateInstantiation(ref template) => {
+                let args_cannot_derive = template.template_arguments()
+                    .iter()
+                    .any(|arg| self.cannot_derive_hash.contains(&arg));
+                if args_cannot_derive {
+                    trace!("    template args cannot derive Hash, so \
+                            insantiation can't either");
+                    return self.insert(id);
+                }
+
+                assert!(
+                    !template.template_definition().is_opaque(self.ctx, &()),
+                    "The early ty.is_opaque check should have handled this case"
+                );
+                let def_cannot_derive = self.cannot_derive_hash
+                    .contains(&template.template_definition());
+                if def_cannot_derive {
+                    trace!("    template definition cannot derive Hash, so \
+                            insantiation can't either");
+                    return self.insert(id);
+                }
+
+                trace!("    template instantiation can derive Hash");
+                ConstrainResult::Same
+            }
+
+            TypeKind::Opaque => {
+                unreachable!(
+                    "The early ty.is_opaque check should have handled this case"
+                )
+            }
+        }
+    }
+
+    fn each_depending_on<F>(&self, id: ItemId, mut f: F)
+        where F: FnMut(ItemId),
+    {
+        if let Some(edges) = self.dependencies.get(&id) {
+            for item in edges {
+                trace!("enqueue {:?} into worklist", item);
+                f(*item);
+            }
+        }
+    }
+}
+
+impl<'ctx, 'gen> From<CannotDeriveHash<'ctx, 'gen>> for HashSet<ItemId> {
+    fn from(analysis: CannotDeriveHash<'ctx, 'gen>) -> Self {
+        analysis.cannot_derive_hash
+    }
+}

src/ir/analysis/mod.rs

@@ -51,6 +51,8 @@ mod derive_copy;
 pub use self::derive_copy::CannotDeriveCopy;
 mod has_type_param_in_array;
 pub use self::has_type_param_in_array::HasTypeParameterInArray;
+mod derive_hash;
+pub use self::derive_hash::CannotDeriveHash;
 
 use ir::context::{BindgenContext, ItemId};
 use ir::traversal::{EdgeKind, Trace};