Refactor using a buffer for reading and a buffer for writing

Cargo.lock

@@ -1,5 +1,7 @@
 # This file is automatically @generated by Cargo.
 # It is not intended for manual editing.
+version = 3
+
 [[package]]
 name = "ansi_term"
 version = "0.11.0"
@@ -9,6 +11,12 @@ dependencies = [
  "winapi",
 ]
 
+[[package]]
+name = "arrayvec"
+version = "0.5.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "23b62fc65de8e4e7f52534fb52b0f3ed04746ae267519eef2a83941e8085068b"
+
 [[package]]
 name = "atty"
 version = "0.2.14"
@@ -242,6 +250,7 @@ version = "1.0.1"
 dependencies = [
  "clap",
  "criterion",
+ "utf8-chars",
 ]
 
 [[package]]
@@ -519,6 +528,15 @@ version = "0.2.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "8ccb82d61f80a663efe1f787a51b16b5a51e3314d6ac365b08639f52387b33f3"
 
+[[package]]
+name = "utf8-chars"
+version = "1.0.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c1348d8face79d019be7cbc0198e36bf93e160ddbfaa7bb54c9592627b9ec841"
+dependencies = [
+ "arrayvec",
+]
+
 [[package]]
 name = "vec_map"
 version = "0.8.2"

Cargo.toml

@@ -15,6 +15,7 @@ categories = ["command-line-utilities"]
 
 [dependencies]
 clap = { version= "2.33.3", optional = true }
+utf8-chars = "1.0.0"
 
 [dev-dependencies]
 criterion = "0.3"

src/lib.rs

@@ -3,6 +3,10 @@
 //!
 //! It does not do anything more than that, which makes it so fast.
 
+use std::io::{BufReader, BufWriter, Write};
+use utf8_chars::BufReadCharsExt;
+use std::error::Error;
+
 ///
 /// Set the indentation used for the formatting.
 ///
@@ -24,16 +28,32 @@ pub enum Indentation<'a> {
 /// The default indentation is faster than a custom one
 ///
 pub fn format_json(json: &str, indentation: Indentation) -> String {
-    // at least as big as the input to avoid resizing
+    let mut reader = BufReader::new(json.as_bytes());
-    // this might be too big if the input string is formatted in a weird way, but that's not expected, and it will still be efficient
+    let mut writer = BufWriter::new(Vec::new());
-    let mut out = String::with_capacity(json.len());
+
+    format_json_buffered(&mut reader, &mut writer, indentation).unwrap();
+    String::from_utf8(writer.into_inner().unwrap()).unwrap()
+}
+
+///
+/// # Formats a json string
+///
+/// The indentation can be set to any value using [Indentation](jsonformat::Indentation)
+/// The default value is two spaces
+/// The default indentation is faster than a custom one
+///
+pub fn format_json_buffered<R, W>(reader: &mut BufReader<R>, writer: &mut BufWriter<W>, indentation: Indentation) -> Result<(), Box<dyn Error>>
+where
+  R: std::io::Read,
+  W: std::io::Write {
 
     let mut escaped = false;
     let mut in_string = false;
     let mut indent_level = 0usize;
     let mut newline_requested = false; // invalidated if next character is ] or }
 
-    for char in json.chars() {
+    for char in reader.chars() {
+        let char = char?;
         if in_string {
             let mut escape_here = false;
             match char {
@@ -49,7 +69,7 @@ pub fn format_json(json: &str, indentation: Indentation) -> String {
                 }
                 _ => {}
             }
-            out.push(char);
+            writer.write_all(char.encode_utf8(&mut [0; 4]).as_bytes())?;
             escaped = escape_here;
         } else {
             let mut auto_push = true;
@@ -71,14 +91,14 @@ pub fn format_json(json: &str, indentation: Indentation) -> String {
                     indent_level = indent_level.saturating_sub(1);
                     if !newline_requested {
                         // see comment below about newline_requested
-                        out.push('\n');
+                        writeln!(writer)?;
-                        indent(&mut out, indent_level, indentation);
+                        indent_buffered(writer, indent_level, indentation)?;
                     }
                 }
                 ':' => {
                     auto_push = false;
-                    out.push(char);
+                    writer.write_all(char.encode_utf8(&mut [0; 4]).as_bytes())?;
-                    out.push(' ');
+                    writer.write_all(" ".as_bytes())?;
                 }
                 ',' => {
                     request_newline = true;
@@ -89,33 +109,37 @@ pub fn format_json(json: &str, indentation: Indentation) -> String {
                 // newline only happens after { [ and ,
                 // this means we can safely assume that it being followed up by } or ]
                 // means an empty object/array
-                out.push('\n');
+                writeln!(writer)?;
-                indent(&mut out, old_level, indentation);
+                indent_buffered(writer, old_level, indentation)?;
             }
 
             if auto_push {
-                out.push(char);
+                writer.write_all(char.encode_utf8(&mut [0; 4]).as_bytes())?;
             }
 
             newline_requested = request_newline;
         }
     }
 
-    out
+    Ok(())
 }
 
-fn indent(buf: &mut String, level: usize, indent_str: Indentation) {
+fn indent_buffered<W>(writer: &mut BufWriter<W>, level: usize, indent_str: Indentation) -> Result<(), Box<dyn Error>>
+where
+  W: std::io::Write {
     for _ in 0..level {
         match indent_str {
             Indentation::Default => {
-                buf.push(' ');
+                writer.write_all(" ".as_bytes())?;
-                buf.push(' ');
+                writer.write_all(" ".as_bytes())?;
             }
             Indentation::Custom(indent) => {
-                buf.push_str(indent);
+                writer.write_all(indent.as_bytes())?;
             }
         }
     }
+
+    Ok(())
 }
 
 #[cfg(test)]

src/main.rs

@@ -1,10 +1,10 @@
 use clap::clap_app;
-use jsonformat::{format_json, Indentation};
+use jsonformat::{Indentation, format_json_buffered};
-use std::fs;
+use std::error::Error;
-use std::io;
+use std::fs::File;
-use std::io::Read;
+use std::io::{BufReader, BufWriter, Read, Write};
 
-fn main() -> Result<(), io::Error> {
+fn main() -> Result<(), Box<dyn Error>> {
     let matches = clap_app!(jsonformat =>
         (version: "1.1")
         (author: "nilstrieb <nilstrieb@gmail.com>")
@@ -16,13 +16,13 @@ fn main() -> Result<(), io::Error> {
     )
     .get_matches();
 
-    let str = match matches.value_of("input") {
+    let reader: Box<dyn Read> = match matches.value_of("input") {
-        Some(path) => fs::read_to_string(path)?,
+        Some(path) => {
+            let f = File::open(path)?;
+            Box::new(BufReader::new(f))
+        },
         None => {
-            let mut buf = String::new();
+            Box::new(std::io::stdin())
-            let stdin = std::io::stdin();
-            stdin.lock().read_to_string(&mut buf)?;
-            buf
         }
     };
 
@@ -41,8 +41,6 @@ fn main() -> Result<(), io::Error> {
         None => Indentation::Default,
     };
 
-    let formatted = format_json(&str, indent);
-
     let mut output = matches.value_of("output");
     let mut windows_output_default_file: Option<String> = None;
 
@@ -57,12 +55,16 @@ fn main() -> Result<(), io::Error> {
 
     output = windows_output_default_file.as_deref().or(output);
 
-    match output {
+    let writer : Box<dyn Write> = match output {
         Some(file) => {
-            fs::write(file, formatted)?;
+            Box::new(File::create(file)?)
         }
-        None => println!("{}", formatted),
+        None => Box::new(std::io::stdout()),
-    }
+    };
+
+    let mut reader = BufReader::new(reader);
+    let mut writer = BufWriter::new(writer);
+    format_json_buffered(&mut reader, &mut writer, indent)?;
 
     Ok(())
 }