Auto merge of #178 - servo:static-sets, r=Manishearth

Make Atom generic over the set of static strings

The new `string_cache_codegen` crate is intended to be used by downstream users in their build scripts. Provided a list of static strings, it generates code that defines:
- A type alias for `Atom<_>` with the type parameter set,
- A macro like the former `atom!` macro.

Based on #136, original work by @aidanhs.

Fixes #22, fixes #136, closes #83.

r? @nox

<!-- Reviewable:start -->
---

This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/string-cache/178)

<!-- Reviewable:end -->

Author: bors-servo

.travis.yml

@@ -13,6 +13,7 @@ script:
   - cargo test --features log-events
   - "if [ $TRAVIS_RUST_VERSION = nightly ]; then cargo test --features unstable; fi"
   - cargo test --features heapsize
+  - "cd string-cache-codegen/ && cargo build && cd .."
   - "cd examples/event-log/ && cargo build && cd ../.."
   - "cd examples/summarize-events/ && cargo build && cd ../.."
 notifications:

Cargo.toml

@@ -1,14 +1,20 @@
 [package]
 
 name = "string_cache"
-version = "0.2.30"
+version = "0.3.0"  # Also update README.md when making a semver-breaking change
 authors = [ "The Servo Project Developers" ]
 description = "A string interning library for Rust, developed as part of the Servo project."
 license = "MIT / Apache-2.0"
 repository = "https://github.com/servo/string-cache"
-documentation = "http://doc.servo.org/string_cache/"
+documentation = "https://docs.rs/string_cache/"
 build = "build.rs"
 
+# Do not `exclude` ./string-cache-codegen because we want to include
+# ./string-cache-codegen/shared.rs, and `include` is a pain to use
+# (It has to be exhaustive.)
+# This means that packages for this crate include some unused files,
+# but they’re not too big so that shouldn’t be a problem.
+
 [lib]
 name = "string_cache"
 
@@ -26,21 +32,14 @@ heap_size = ["heapsize"]
 
 [dependencies]
 lazy_static = "0.2"
-serde = ">=0.6, <0.9"
+serde = "0.8"
 phf_shared = "0.7.4"
 debug_unreachable = "0.1.1"
+rustc-serialize = { version = "0.3", optional = true }
+heapsize = { version = "0.3", optional = true }
 
 [dev-dependencies]
 rand = "0.3"
 
-[dependencies.rustc-serialize]
-version = "0.3"
-optional = true
-
-[dependencies.heapsize]
-version = ">=0.1.1, <0.4"
-optional = true
-
 [build-dependencies]
-phf_generator = "0.7.4"
-phf_shared = "0.7.4"
+string_cache_codegen = { version = "0.3", path = "./string-cache-codegen" }

README.md

@@ -2,6 +2,77 @@
 
 [![Build Status](https://travis-ci.org/servo/string-cache.svg?branch=master)](https://travis-ci.org/servo/string-cache)
 
-[Documentation](http://doc.servo.org/string_cache/)
+[Documentation](https://docs.rs/string_cache/)
 
 A string interning library for Rust, developed as part of the [Servo](https://github.com/servo/servo) project.
+
+## Simple usage
+
+In `Cargo.toml`:
+
+```toml
+[dependencies]
+string_cache = "0.3"
+```
+
+In `lib.rs`:
+
+```rust
+extern crate string_cache;
+use string_cache::DefaultAtom as Atom;
+```
+
+## With static atoms
+
+In `Cargo.toml`:
+
+```toml
+[package]
+build = "build.rs"
+
+[dependencies]
+string_cache = "0.3"
+
+[build-dependencies]
+string_cache_codegen = "0.3"
+```
+
+In `build.rs`:
+
+```rust
+extern crate string_cache_codegen;
+
+use std::env;
+use std::path::Path;
+
+fn main() {
+    string_cache_codegen::AtomType::new("foo::FooAtom", "foo_atom!")
+        .atoms(&["foo", "bar"])
+        .write_to_file(&Path::new(&env::var("OUT_DIR").unwrap()).join("foo_atom.rs"))
+        .unwrap()
+}
+```
+
+In `lib.rs`:
+
+```rust
+extern crate string_cache;
+
+mod foo {
+    include!(concat!(env!("OUT_DIR"), "/foo_atom.rs"));
+}
+```
+
+The generated code will define a `FooAtom` type and a `foo_atom!` macro.
+The macro can be used in expression or patterns, with strings listed in `build.rs`.
+For example:
+
+```rust
+fn compute_something(input: &foo::FooAtom) -> u32 {
+    match *input {
+        foo_atom!("foo") => 1,
+        foo_atom!("bar") => 2,
+        _ => 3,
+    }
+}
+```

build.rs

@@ -1,73 +1,13 @@
-extern crate phf_shared;
-extern crate phf_generator;
-
-#[path = "src/shared.rs"] #[allow(dead_code)] mod shared;
-#[path = "src/static_atom_list.rs"] mod static_atom_list;
+extern crate string_cache_codegen;
 
 use std::env;
-use std::fs::File;
-use std::io::{BufWriter, Write};
-use std::mem;
 use std::path::Path;
-use std::slice;
 
 fn main() {
-    let hash_state = generate();
-    write_static_atom_set(&hash_state);
-    write_atom_macro(&hash_state);
-}
-
-fn generate() -> phf_generator::HashState {
-    let mut set = std::collections::HashSet::new();
-    for atom in static_atom_list::ATOMS {
-        if !set.insert(atom) {
-            panic!("duplicate static atom `{:?}`", atom);
-        }
-    }
-    phf_generator::generate_hash(static_atom_list::ATOMS)
-}
-
-fn write_static_atom_set(hash_state: &phf_generator::HashState) {
-    let path = Path::new(&std::env::var("OUT_DIR").unwrap()).join("static_atom_set.rs");
-    let mut file = BufWriter::new(File::create(&path).unwrap());
-    macro_rules! w {
-        ($($arg: expr),+) => { (writeln!(&mut file, $($arg),+).unwrap()) }
-    }
-    w!("pub static STATIC_ATOM_SET: StaticAtomSet = StaticAtomSet {{");
-    w!("    key: {},", hash_state.key);
-    w!("    disps: &[");
-    for &(d1, d2) in &hash_state.disps {
-        w!("        ({}, {}),", d1, d2);
-    }
-    w!("    ],");
-    w!("    atoms: &[");
-    for &idx in &hash_state.map {
-        w!("        {:?},", static_atom_list::ATOMS[idx]);
-    }
-    w!("    ],");
-    w!("}};");
-}
-
-fn write_atom_macro(hash_state: &phf_generator::HashState) {
-    let set = shared::StaticAtomSet {
-        key: hash_state.key,
-        disps: leak(hash_state.disps.clone()),
-        atoms: leak(hash_state.map.iter().map(|&idx| static_atom_list::ATOMS[idx]).collect()),
-    };
-
-    let path = Path::new(&env::var("OUT_DIR").unwrap()).join("atom_macro.rs");
-    let mut file = BufWriter::new(File::create(&path).unwrap());
-    writeln!(file, r"#[macro_export]").unwrap();
-    writeln!(file, r"macro_rules! atom {{").unwrap();
-    for &s in set.iter() {
-        let data = shared::pack_static(set.get_index_or_hash(s).unwrap() as u32);
-        writeln!(file, r"({:?}) => {{ $crate::Atom {{ unsafe_data: 0x{:x} }} }};", s, data).unwrap();
-    }
-    writeln!(file, r"}}").unwrap();
-}
-
-fn leak<T>(v: Vec<T>) -> &'static [T] {
-    let slice = unsafe { slice::from_raw_parts(v.as_ptr(), v.len()) };
-    mem::forget(v);
-    slice
+    string_cache_codegen::AtomType::new("atom::tests::TestAtom", "test_atom!")
+        .atoms(&[
+            "a", "b", "address", "area", "body", "font-weight", "br", "html", "head", "id",
+        ])
+        .write_to_file(&Path::new(&env::var("OUT_DIR").unwrap()).join("test_atom.rs"))
+        .unwrap()
 }

examples/event-log/src/main.rs

@@ -9,7 +9,7 @@
 
 extern crate string_cache;
 
-use string_cache::Atom;
+use string_cache::DefaultAtom as Atom;
 use string_cache::event;
 
 use std::io;

examples/summarize-events/src/main.rs

@@ -12,14 +12,15 @@ extern crate string_cache;
 extern crate rustc_serialize;
 extern crate phf_shared;
 
-#[path = "../../../src/shared.rs"]
+#[path = "../../../string-cache-codegen/shared.rs"]
 #[allow(dead_code)]
 mod shared;
 
-use string_cache::Atom;
+use string_cache::DefaultAtom as Atom;
 
 use std::{env, cmp};
 use std::collections::hash_map::{HashMap, Entry};
+use std::marker::PhantomData;
 use std::path::Path;
 
 #[derive(RustcDecodable, Debug)]
@@ -88,7 +89,7 @@ fn main() {
 
                     // FIXME: We really shouldn't be allowed to do this. It's a memory-safety
                     // hazard; the field is only public for the atom!() macro.
-                    _ => Atom { unsafe_data: ev.id }.to_string(),
+                    _ => Atom { unsafe_data: ev.id, phantom: PhantomData }.to_string(),
                 };
 
                 match summary.entry(string) {