From 57a7f96ba599693593b5be0ed6b43b826fb68837 Mon Sep 17 00:00:00 2001
From: Tom Parker-Shemilt <palfrey@tevp.net>
Date: Fri, 10 Jan 2020 23:07:36 +0000
Subject: [PATCH] Start writing a rust-based url checker

---
 .gitignore  |  3 +-
 .travis.yml |  6 ++--
 Cargo.toml  |  9 +++++-
 README.md   |  2 +-
 src/main.rs | 81 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 5 files changed, 94 insertions(+), 7 deletions(-)
 create mode 100644 src/main.rs

diff --git a/.gitignore b/.gitignore
index 008f177..409fcf2 100644
--- a/.gitignore
+++ b/.gitignore
@@ -3,4 +3,5 @@ Cargo.lock
 target/
 cleanup
 cleanup.exe
-ab-results*.json
\ No newline at end of file
+ab-results*.json
+results.yaml
\ No newline at end of file
diff --git a/.travis.yml b/.travis.yml
index a37dd5c..c6e77e5 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -1,4 +1,2 @@
-language: ruby
-before_script: gem install awesome_bot
-# `crates.io` needs to be whitelisted, because of https://github.com/rust-lang/crates.io/issues/788
-script: awesome_bot -f README.md --allow-dupe --allow-ssl --white-list crates.io,sciter.com
+language: rust
+script: cargo run
\ No newline at end of file
diff --git a/Cargo.toml b/Cargo.toml
index 38962d6..abbf4d2 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -2,6 +2,13 @@
 name = "awesome-rust"
 version = "0.1.0"
 authors = []
+edition = "2018"
 
 [dependencies]
-pulldown-cmark= "0.0.8"
+pulldown-cmark = "0.6"
+futures = "0.3"
+reqwest = "0.10"
+tokio = {version = "0.2", features = ["macros", "rt-core", "rt-threaded"] }
+serde = { version = "1.0", features = ["derive"] }
+serde_yaml = "0.8"
+anyhow = "1.0"
\ No newline at end of file
diff --git a/README.md b/README.md
index 5b1a04a..f62dbf7 100644
--- a/README.md
+++ b/README.md
@@ -1017,7 +1017,7 @@ See also [Are we game yet?](http://arewegameyet.com)
 * Corange
   * [lucidscape/corange-rs](https://github.com/lucidscape/corange-rs) — [Corange](https://github.com/orangeduck/Corange) bindings
 * Entity-Component Systems (ECS)
-  * [slide-rs/specs](https://github.com/amethyst/specs) — Specs Parallel ECS [<img src="https://api.travis-ci.org/slide-rs/specs.svg">](httpsL//github.com/travis-ci.org/slide-rs/specs)
+  * [slide-rs/specs](https://github.com/amethyst/specs) — Specs Parallel ECS [<img src="https://api.travis-ci.org/slide-rs/specs.svg">](https://github.com/travis-ci.org/slide-rs/specs)
 * Game Engines
   * [Amethyst](https://amethyst.rs) — Data-oriented game engine [<img src="https://jenkins.amethyst-engine.org/job/amethyst/job/master/badge/icon">](https://jenkins.amethyst-engine.org/blue/organizations/jenkins/amethyst/activity/)
   * [Piston](https://www.piston.rs/) — [<img src="https://api.travis-ci.org/PistonDevelopers/piston.svg?branch=master">](https://travis-ci.org/PistonDevelopers/piston)
diff --git a/src/main.rs b/src/main.rs
new file mode 100644
index 0000000..5691fca
--- /dev/null
+++ b/src/main.rs
@@ -0,0 +1,81 @@
+use pulldown_cmark::{Parser, Event, Tag};
+use std::fs;
+use futures::future::select_all;
+use futures::future::FutureExt;
+use std::collections::{BTreeSet, BTreeMap};
+use serde::{Serialize, Deserialize};
+use anyhow::Result;
+
+fn to_anyhow<T, E>(res: std::result::Result<T, E>) -> Result<T>
+    where E: std::error::Error + std::marker::Send + std::marker::Sync + 'static
+{
+    res.map_err(|x| Into::<anyhow::Error>::into(x))
+}
+
+async fn get_url(url: String) -> (String, Result<String>) {
+    let res = reqwest::get(&url).await;
+    (url, to_anyhow(res.map(|x| format!("{:?}", x))))
+}
+
+#[derive(Debug, Serialize, Deserialize)]
+struct Results {
+    working: BTreeSet<String>,
+    failed: BTreeMap<String, String>
+}
+
+impl Results {
+    fn new() -> Results {
+        Results {
+            working: BTreeSet::new(),
+            failed: BTreeMap::new()
+        }
+    }
+}
+
+#[tokio::main]
+async fn main() -> Result<()> {
+    let markdown_input = fs::read_to_string("README.md").expect("Can't read README.md");
+    let parser = Parser::new(&markdown_input);
+
+    let mut results: Results = to_anyhow(fs::read_to_string("results.yaml")).and_then(|x| to_anyhow(serde_yaml::from_str(&x))).unwrap_or(Results::new());
+    results.failed.clear();
+
+    let mut url_checks = vec![];
+
+    for (event, _range) in parser.into_offset_iter() {
+        if let Event::Start(tag) = event {
+            if let Tag::Link(_link_type, url, _title) = tag {
+                if !url.starts_with("http") {
+                    continue;
+                }
+                let url_string = url.to_string();
+                if results.working.contains(&url_string) {
+                    continue;
+                }
+                let check = get_url(url_string).boxed();
+                url_checks.push(check);
+            }
+        }
+    }
+
+    while url_checks.len() > 0 {
+        let ((url, res), _index, remaining) = select_all(url_checks).await;
+        url_checks = remaining;
+        match res {
+            Ok(_) => {
+                print!("\u{2714} ");
+                results.working.insert(url);
+            },
+            Err(err) => {
+                print!("\u{2718} ");
+                results.failed.insert(url, err.to_string());
+            }
+        }
+        fs::write("results.yaml", serde_yaml::to_string(&results)?)?;
+    }
+    println!("");
+    for (url, error) in &results.failed {
+        println!("Error: {} {}", url, error);
+    }
+    Ok(())
+}
\ No newline at end of file