Merge pull request #50 from data-pup/add-garbage-subcommand

Add a `twiggy garbage` command. Fixes issue #48.

README.md

@@ -36,6 +36,7 @@ Use `twiggy` to make your binaries slim!
         - [`twiggy monos`](#twiggy-monos)
         - [`twiggy dominators`](#twiggy-dominators)
         - [`twiggy diff`](#twiggy-diff)
+        - ['twiggy garbage'](#twiggy-garbage)
     - [🦀 As a Crate](#-as-a-crate)
     - [🕸 On the Web with WebAssembly](#-on-the-web-with-webassembly)
 - [🔎 Supported Binary Formats](#-supported-binary-formats)
@@ -364,6 +365,25 @@ and new versions of a binary.
         +145 ┊ <wee_alloc::neighbors::Neighbors<'a, T>>::remove::hc9e5d4284e8233b8
 ```
 
+#### `twiggy garbage`
+
+The `twiggy garbage` sub-command finds and display code and data that is not
+transitively referenced by any exports or public functions.
+
+```
+ Bytes │ Size % │ Garbage Item
+───────┼────────┼──────────────────────
+    11 ┊  5.58% ┊ unusedAddThreeNumbers
+     8 ┊  4.06% ┊ unusedAddOne
+     7 ┊  3.55% ┊ type[2]
+     5 ┊  2.54% ┊ type[1]
+     5 ┊  2.54% ┊ unusedChild
+     4 ┊  2.03% ┊ type[0]
+     1 ┊  0.51% ┊ func[0]
+     1 ┊  0.51% ┊ func[1]
+     1 ┊  0.51% ┊ func[2]
+```
+
 ### 🦀 As a Crate
 
 `twiggy` is divided into a collection of crates that you can use

analyze/Cargo.toml

@@ -15,3 +15,4 @@ path = "./analyze.rs"
 twiggy-ir = { version = "0.1.0", path = "../ir" }
 twiggy-opt = { version = "0.1.0", path = "../opt", default-features = false }
 twiggy-traits = { version = "0.1.0", path = "../traits" }
+petgraph = "0.4.11"

analyze/analyze.rs

@@ -3,6 +3,7 @@
 #![deny(missing_docs)]
 #![deny(missing_debug_implementations)]
 
+extern crate petgraph;
 extern crate twiggy_ir as ir;
 extern crate twiggy_opt as opt;
 extern crate twiggy_traits as traits;
@@ -723,10 +724,7 @@ impl traits::Emit for Diff {
         ]);
 
         for entry in &self.deltas {
-            table.add_row(vec![
-                format!("{:+}", entry.delta),
-                entry.name.clone(),
-            ]);
+            table.add_row(vec![format!("{:+}", entry.delta), entry.name.clone()]);
         }
 
         write!(dest, "{}", &table)?;
@@ -801,3 +799,79 @@ pub fn diff(
     let diff = Diff { deltas };
     Ok(Box::new(diff) as Box<traits::Emit>)
 }
+
+#[derive(Debug)]
+struct Garbage {
+    items: Vec<ir::Id>,
+}
+
+impl traits::Emit for Garbage {
+    fn emit_text(&self, items: &ir::Items, dest: &mut io::Write) -> Result<(), traits::Error> {
+        let mut table = Table::with_header(vec![
+            (Align::Right, "Bytes".to_string()),
+            (Align::Right, "Size %".to_string()),
+            (Align::Left, "Garbage Item".to_string()),
+        ]);
+
+        for &id in &self.items {
+            let item = &items[id];
+            let size = item.size();
+            let size_percent = (f64::from(size)) / (f64::from(items.size())) * 100.0;
+            table.add_row(vec![
+                size.to_string(),
+                format!("{:.2}%", size_percent),
+                item.name().to_string(),
+            ]);
+        }
+
+        write!(dest, "{}", &table)?;
+        Ok(())
+    }
+
+    fn emit_json(&self, items: &ir::Items, dest: &mut io::Write) -> Result<(), traits::Error> {
+        let mut arr = json::array(dest)?;
+
+        for &id in &self.items {
+            let item = &items[id];
+
+            let mut obj = arr.object()?;
+            obj.field("name", item.name())?;
+
+            let size = item.size();
+            let size_percent = (f64::from(size)) / (f64::from(items.size())) * 100.0;
+            obj.field("bytes", size)?;
+            obj.field("size_percent", size_percent)?;
+        }
+
+        Ok(())
+    }
+}
+
+/// Find items that are not transitively referenced by any exports or public functions.
+pub fn garbage(items: &ir::Items, opts: &opt::Garbage) -> Result<Box<traits::Emit>, traits::Error> {
+    fn get_reachable_items(items: &ir::Items) -> BTreeSet<ir::Id> {
+        let mut reachable_items: BTreeSet<ir::Id> = BTreeSet::new();
+        let mut dfs = petgraph::visit::Dfs::new(items, items.meta_root());
+        while let Some(id) = dfs.next(&items) {
+            reachable_items.insert(id);
+        }
+        reachable_items
+    }
+
+    let reachable_items = get_reachable_items(&items);
+    let mut unreachable_items: Vec<_> = items
+        .iter()
+        .filter(|item| !reachable_items.contains(&item.id()))
+        .collect();
+
+    unreachable_items.sort_by(|a, b| b.size().cmp(&a.size()));
+    unreachable_items.truncate(opts.max_items() as usize);
+
+    let unreachable_items: Vec<_> = unreachable_items.iter().map(|item| item.id()).collect();
+
+    let garbage_items = Garbage {
+        items: unreachable_items,
+    };
+
+    Ok(Box::new(garbage_items) as Box<traits::Emit>)
+}

opt/definitions.rs

@@ -33,7 +33,12 @@ pub enum Options {
 
     /// Diff the old and new versions of a binary to see what sizes changed.
     #[structopt(name = "diff")]
-    Diff(Diff)
+    Diff(Diff),
+
+    /// Find and display code and data that is not transitively referenced by
+    /// any exports or public functions.
+    #[structopt(name = "garbage")]
+    Garbage(Garbage)
 }
 
 /// List the top code size offenders in a binary.
@@ -409,3 +414,62 @@ impl Diff {
         self.max_items = n;
     }
 }
+
+/// Find and display code and data that is not transitively referenced by any
+/// exports or public functions.
+#[derive(Clone, Debug)]
+#[derive(StructOpt)]
+#[wasm_bindgen]
+pub struct Garbage {
+    /// The path to the input binary to size profile.
+    #[cfg(feature = "cli")]
+    #[structopt(parse(from_os_str))]
+    input: path::PathBuf,
+
+    /// The destination to write the output to. Defaults to `stdout`.
+    #[cfg(feature = "cli")]
+    #[structopt(short = "o", default_value = "-")]
+    output_destination: OutputDestination,
+
+    /// The format the output should be written in.
+    #[cfg(feature = "cli")]
+    #[structopt(short = "f", long = "format", default_value = "text")]
+    output_format: traits::OutputFormat,
+
+    /// The maximum number of items to display.
+    #[structopt(short = "n", default_value = "10")]
+    max_items: u32,
+}
+
+impl Default for Garbage {
+    fn default() -> Garbage {
+        Garbage {
+            #[cfg(feature = "cli")]
+            input: Default::default(),
+            #[cfg(feature = "cli")]
+            output_destination: Default::default(),
+            #[cfg(feature = "cli")]
+            output_format: Default::default(),
+
+            max_items: 10,
+        }
+    }
+}
+
+#[wasm_bindgen]
+impl Garbage {
+    /// Construct a new, default `Garbage`
+    pub fn new() -> Garbage {
+        Garbage::default()
+    }
+
+    /// The maximum number of items to display.
+    pub fn max_items(&self) -> u32 {
+        self.max_items
+    }
+
+    /// Set the maximum number of items to display.
+    pub fn set_max_items(&mut self, max: u32) {
+        self.max_items = max;
+    }
+}

opt/opt.rs

@@ -51,6 +51,7 @@ cfg_if! {
                     Options::Paths(ref paths) => paths.input(),
                     Options::Monos(ref monos) => monos.input(),
                     Options::Diff(ref diff) => diff.input(),
+                    Options::Garbage(ref garbo) => garbo.input(),
                 }
             }
 
@@ -61,6 +62,7 @@ cfg_if! {
                     Options::Paths(ref paths) => paths.output_destination(),
                     Options::Monos(ref monos) => monos.output_destination(),
                     Options::Diff(ref diff) => diff.output_destination(),
+                    Options::Garbage(ref garbo) => garbo.output_destination(),
                 }
             }
 
@@ -71,6 +73,7 @@ cfg_if! {
                     Options::Paths(ref paths) => paths.output_format(),
                     Options::Monos(ref monos) => monos.output_format(),
                     Options::Diff(ref diff) => diff.output_format(),
+                    Options::Garbage(ref garbo) => garbo.output_format(),
                 }
             }
         }
@@ -152,6 +155,20 @@ cfg_if! {
             }
         }
 
+        impl CommonCliOptions for Garbage {
+            fn input(&self) -> &path::Path {
+                &self.input
+            }
+
+            fn output_destination(&self) -> &OutputDestination {
+                &self.output_destination
+            }
+
+            fn output_format(&self) -> traits::OutputFormat {
+                self.output_format
+            }
+        }
+
         /// Where to output results.
         #[derive(Clone, Debug)]
         pub enum OutputDestination {

twiggy/tests/expectations/garbage

@@ -0,0 +1,11 @@
+ Bytes │ Size % │ Garbage Item
+───────┼────────┼──────────────────────
+    11 ┊  5.58% ┊ unusedAddThreeNumbers
+     8 ┊  4.06% ┊ unusedAddOne
+     7 ┊  3.55% ┊ type[2]
+     5 ┊  2.54% ┊ type[1]
+     5 ┊  2.54% ┊ unusedChild
+     4 ┊  2.03% ┊ type[0]
+     1 ┊  0.51% ┊ func[0]
+     1 ┊  0.51% ┊ func[1]
+     1 ┊  0.51% ┊ func[2]

twiggy/tests/expectations/garbage_json

@@ -0,0 +1 @@
+[{"name":"unusedAddThreeNumbers","bytes":11,"size_percent":5.583756345177665},{"name":"unusedAddOne","bytes":8,"size_percent":4.060913705583756},{"name":"type[2]","bytes":7,"size_percent":3.5532994923857872},{"name":"type[1]","bytes":5,"size_percent":2.5380710659898478},{"name":"unusedChild","bytes":5,"size_percent":2.5380710659898478},{"name":"type[0]","bytes":4,"size_percent":2.030456852791878},{"name":"func[0]","bytes":1,"size_percent":0.5076142131979695},{"name":"func[1]","bytes":1,"size_percent":0.5076142131979695},{"name":"func[2]","bytes":1,"size_percent":0.5076142131979695}]

twiggy/tests/expectations/garbage_top_2

@@ -0,0 +1,4 @@
+ Bytes │ Size % │ Garbage Item
+───────┼────────┼──────────────────────
+    11 ┊  5.58% ┊ unusedAddThreeNumbers
+     8 ┊  4.06% ┊ unusedAddOne

twiggy/tests/fixtures/garbage.wat

@@ -0,0 +1,43 @@
+(module
+    ;; -------------------------------------------------------------------------
+    ;; This is a WebAssembly text file that can be compiled in a wasm module to
+    ;; test the `twiggy garbage` command. This file contains exported functions,
+    ;; as well as unreachable functions of different sizes.
+    ;; -------------------------------------------------------------------------
+    ;; NOTE: The test cases expect that this module is compiled with debug
+    ;; names written to the binary file, which affects the size percentages.
+    ;; Compile this file using the following command:
+    ;;
+    ;; wat2wasm --debug-names garbage.wat -o garbage.wasm
+    ;; -------------------------------------------------------------------------
+
+    ;; This unused function is called by 'unusedAddOne'. Push 1 onto the stack.
+    (func $unusedChild (result i32)
+        i32.const 1)
+
+    ;; This unused function will call `unusedChild`, and return `val + 1`.
+    (func $unusedAddOne (param $val i32) (result i32)
+        get_local $val
+        call $unusedChild
+        i32.add)
+
+    ;; This unused function adds three numbers, and returns the result.
+    (func $unusedAddThreeNumbers
+        (param $first i32) (param $second i32) (param $third i32) (result i32)
+            get_local $first
+            get_local $second
+            i32.add
+            get_local $third
+            i32.add
+    )
+
+    ;; This function exists to test that reachable items are not shown.
+    (func $add (param $lhs i32) (param $rhs i32) (result i32)
+        get_local $lhs
+        get_local $rhs
+        i32.add
+    )
+
+    ;; Export only the `add` function.
+    (export "add" (func $add))
+)

twiggy/tests/tests.rs

@@ -233,3 +233,21 @@ test!(
     "-n",
     "5"
 );
+
+test!(garbage, "garbage", "./fixtures/garbage.wasm");
+
+test!(
+    garbage_top_2,
+    "garbage",
+    "./fixtures/garbage.wasm",
+    "-n",
+    "2"
+);
+
+test!(
+    garbage_json,
+    "garbage",
+    "./fixtures/garbage.wasm",
+    "-f",
+    "json"
+);

twiggy/twiggy.rs

@@ -34,6 +34,7 @@ fn run(opts: opt::Options) -> Result<(), traits::Error> {
         opt::Options::Dominators(ref doms) => analyze::dominators(&mut items, doms)?,
         opt::Options::Paths(ref paths) => analyze::paths(&mut items, paths)?,
         opt::Options::Monos(ref monos) => analyze::monos(&mut items, monos)?,
+        opt::Options::Garbage(ref garbo) => analyze::garbage(&mut items, garbo)?,
         opt::Options::Diff(ref diff) => {
             let mut new_items = parser::read_and_parse(diff.new_input())?;
             analyze::diff(&mut items, &mut new_items, diff)?