Auto merge of #23289 - mihneadb:rustdoc-search-by-type, r=alexcrichton

This adds search by type (for functions/methods) support to Rustdoc. Target issue is at https://github.com/rust-lang/rfcs/issues/658.

I've described my approach here: https://github.com/rust-lang/rfcs/issues/658#issuecomment-76484200. I'll copy the text in here as well:

---

Hi, it took me longer than I wished, but I have implemented this in a not-too-complex way that I think can be extended to support more complex features (like the ones mentioned [here](https://github.com/rust-lang/rust/issues/12866#issuecomment-66945317)).

The idea is to generate a JSON representation of the types of methods/functions in the existing index, and then make the JS understand when it should look by type (and not by name).

I tried to come up with a JSON representation that can be extended to support generics, bounds, ref/mut annotations and so on. Here are a few samples:

Function:

```rust
fn to_uppercase(c: char) -> char
```

```json
{
    "inputs": [
        {"name": "char"}
    ],
    "output": {
        "name": "char",
    }
}
```

Method (implemented or defined in trait):

```rust
// in struct Vec
// self is considered an argument as well
fn capacity(&self) -> usize
```

```json
{
    "inputs": [
        {"name": "vec"}
    ],
    "output": {
        "name": "usize"
    }
}
```

This simple format can be extended by adding more fields, like `generic: bool`, a `bounds` mapping and so on.

I have a working implementation in https://github.com/rust-lang/rust/compare/master...mihneadb:rustdoc-search-by-type. You can check out a live demo [here](http://data.mihneadb.net/doc/std/index.html?search=charext%20-%3E%20char).

![screenshot from 2015-02-28 00 54 00](7e5374ee-bee4-11e4-99a6-9aac3c9d5068.png)


The feature list is not that long:
- search by types (you *can* use generics as well, as long as you use the exact name - e.g. [`vec,t -> `](http://data.mihneadb.net/doc/std/index.html?search=vec%2C%20t%20-%3E))
- order of arguments does not matter
- `self` is took into account as well (e.g. search for `vec -> usize`)
- does not use "complex" annotations (e.g. you don't search for `&char -> char` but for `char -> char`)

My goal is to get a working, minimal "base" merged so that others can build upon it. How should I proceed? Do I open a PR (badly in need of code review since this is my first non "hello world"-ish rust code)?

---

src/librustdoc/html/render.rs

@@ -34,6 +34,7 @@
 //! both occur before the crate is rendered.
 pub use self::ExternalLocation::*;
 
+use std::ascii::OwnedAsciiExt;
 use std::cell::RefCell;
 use std::cmp::Ordering;
 use std::collections::{HashMap, HashSet};
@@ -239,6 +240,51 @@ struct IndexItem {
     path: String,
     desc: String,
     parent: Option<ast::DefId>,
+    search_type: Option<IndexItemFunctionType>,
+}
+
+/// A type used for the search index.
+struct Type {
+    name: Option<String>,
+}
+
+impl fmt::Display for Type {
+    /// Formats type as {name: $name}.
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        // Wrapping struct fmt should never call us when self.name is None,
+        // but just to be safe we write `null` in that case.
+        match self.name {
+            Some(ref n) => write!(f, "{{\"name\":\"{}\"}}", n),
+            None => write!(f, "null")
+        }
+    }
+}
+
+/// Full type of functions/methods in the search index.
+struct IndexItemFunctionType {
+    inputs: Vec<Type>,
+    output: Option<Type>
+}
+
+impl fmt::Display for IndexItemFunctionType {
+    /// Formats a full fn type as a JSON {inputs: [Type], outputs: Type/null}.
+    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
+        // If we couldn't figure out a type, just write `null`.
+        if self.inputs.iter().any(|ref i| i.name.is_none()) ||
+           (self.output.is_some() && self.output.as_ref().unwrap().name.is_none()) {
+            return write!(f, "null")
+        }
+
+        let inputs: Vec<String> = self.inputs.iter().map(|ref t| format!("{}", t)).collect();
+        try!(write!(f, "{{\"inputs\":[{}],\"output\":", inputs.connect(",")));
+
+        match self.output {
+            Some(ref t) => try!(write!(f, "{}", t)),
+            None => try!(write!(f, "null"))
+        };
+
+        Ok(try!(write!(f, "}}")))
+    }
 }
 
 // TLS keys used to carry information around during rendering.
@@ -409,6 +455,7 @@ fn build_index(krate: &clean::Crate, cache: &mut Cache) -> io::Result<String> {
                         path: fqp[..fqp.len() - 1].connect("::"),
                         desc: shorter(item.doc_value()).to_string(),
                         parent: Some(did),
+                        search_type: None,
                     });
                 },
                 None => {}
@@ -458,7 +505,11 @@ fn build_index(krate: &clean::Crate, cache: &mut Cache) -> io::Result<String> {
                 let pathid = *nodeid_to_pathid.get(&nodeid).unwrap();
                 try!(write!(&mut w, ",{}", pathid));
             }
-            None => {}
+            None => try!(write!(&mut w, ",null"))
+        }
+        match item.search_type {
+            Some(ref t) => try!(write!(&mut w, ",{}", t)),
+            None => try!(write!(&mut w, ",null"))
         }
         try!(write!(&mut w, "]"));
     }
@@ -872,12 +923,21 @@ impl DocFolder for Cache {
 
             match parent {
                 (parent, Some(path)) if is_method || (!self.privmod && !hidden_field) => {
+                    // Needed to determine `self` type.
+                    let parent_basename = self.parent_stack.first().and_then(|parent| {
+                        match self.paths.get(parent) {
+                            Some(&(ref fqp, _)) => Some(fqp[fqp.len() - 1].clone()),
+                            _ => None
+                        }
+                    });
+
                     self.search_index.push(IndexItem {
                         ty: shortty(&item),
                         name: s.to_string(),
                         path: path.connect("::").to_string(),
                         desc: shorter(item.doc_value()).to_string(),
                         parent: parent,
+                        search_type: get_index_search_type(&item, parent_basename),
                     });
                 }
                 (Some(parent), None) if is_method || (!self.privmod && !hidden_field)=> {
@@ -2307,6 +2367,52 @@ fn make_item_keywords(it: &clean::Item) -> String {
     format!("{}, {}", get_basic_keywords(), it.name.as_ref().unwrap())
 }
 
+fn get_index_search_type(item: &clean::Item,
+                         parent: Option<String>) -> Option<IndexItemFunctionType> {
+    let decl = match item.inner {
+        clean::FunctionItem(ref f) => &f.decl,
+        clean::MethodItem(ref m) => &m.decl,
+        clean::TyMethodItem(ref m) => &m.decl,
+        _ => return None
+    };
+
+    let mut inputs = Vec::new();
+
+    // Consider `self` an argument as well.
+    if let Some(name) = parent {
+        inputs.push(Type { name: Some(name.into_ascii_lowercase()) });
+    }
+
+    inputs.extend(&mut decl.inputs.values.iter().map(|arg| {
+        get_index_type(&arg.type_)
+    }));
+
+    let output = match decl.output {
+        clean::FunctionRetTy::Return(ref return_type) => Some(get_index_type(return_type)),
+        _ => None
+    };
+
+    Some(IndexItemFunctionType { inputs: inputs, output: output })
+}
+
+fn get_index_type(clean_type: &clean::Type) -> Type {
+    Type { name: get_index_type_name(clean_type).map(|s| s.into_ascii_lowercase()) }
+}
+
+fn get_index_type_name(clean_type: &clean::Type) -> Option<String> {
+    match *clean_type {
+        clean::ResolvedPath { ref path, .. } => {
+            let segments = &path.segments;
+            Some(segments[segments.len() - 1].name.clone())
+        },
+        clean::Generic(ref s) => Some(s.clone()),
+        clean::Primitive(ref p) => Some(format!("{:?}", p)),
+        clean::BorrowedRef { ref type_, .. } => get_index_type_name(type_),
+        // FIXME: add all from clean::Type.
+        _ => None
+    }
+}
+
 pub fn cache() -> Arc<Cache> {
     CACHE_KEY.with(|c| c.borrow().clone())
 }

src/librustdoc/html/static/main.js

@@ -209,6 +209,33 @@
                         break;
                     }
                 }
+            // searching by type
+            } else if (val.search("->") > -1) {
+                var trimmer = function (s) { return s.trim(); };
+                var parts = val.split("->").map(trimmer);
+                var input = parts[0];
+                // sort inputs so that order does not matter
+                var inputs = input.split(",").map(trimmer).sort();
+                var output = parts[1];
+
+                for (var i = 0; i < nSearchWords; ++i) {
+                    var type = searchIndex[i].type;
+                    if (!type) {
+                        continue;
+                    }
+
+                    // sort index inputs so that order does not matter
+                    var typeInputs = type.inputs.map(function (input) {
+                        return input.name;
+                    }).sort();
+
+                    // allow searching for void (no output) functions as well
+                    var typeOutput = type.output ? type.output.name : "";
+                    if (inputs.toString() === typeInputs.toString() &&
+                        output == typeOutput) {
+                        results.push({id: i, index: -1, dontValidate: true});
+                    }
+                }
             } else {
                 // gather matching search results up to a certain maximum
                 val = val.replace(/\_/g, "");
@@ -329,6 +356,11 @@
                     path = result.item.path.toLowerCase(),
                     parent = result.item.parent;
 
+                // this validation does not make sense when searching by types
+                if (result.dontValidate) {
+                    continue;
+                }
+
                 var valid = validateResult(name, path, split, parent);
                 if (!valid) {
                     result.id = -1;
@@ -573,7 +605,8 @@
                 //              (String) name,
                 //              (String) full path or empty string for previous path,
                 //              (String) description,
-                //              (optional Number) the parent path index to `paths`]
+                //              (Number | null) the parent path index to `paths`]
+                //              (Object | null) the type of the function (if any)
                 var items = rawSearchIndex[crate].items;
                 // an array of [(Number) item type,
                 //              (String) name]
@@ -598,7 +631,7 @@
                     var rawRow = items[i];
                     var row = {crate: crate, ty: rawRow[0], name: rawRow[1],
                                path: rawRow[2] || lastPath, desc: rawRow[3],
-                               parent: paths[rawRow[4]]};
+                               parent: paths[rawRow[4]], type: rawRow[5]};
                     searchIndex.push(row);
                     if (typeof row.name === "string") {
                         var word = row.name.toLowerCase();