From 95107b6a39edb34930b79184e8c36a3754e06385 Mon Sep 17 00:00:00 2001
From: Felix Ableitner <me@nutomic.com>
Date: Wed, 17 Mar 2021 03:04:57 +0100
Subject: [PATCH] Include number of failed instances

---
 src/crawl.rs | 10 +++++++---
 src/main.rs  | 16 +++++++++-------
 2 files changed, 16 insertions(+), 10 deletions(-)

diff --git a/src/crawl.rs b/src/crawl.rs
index c3d6a88..063d1e0 100644
--- a/src/crawl.rs
+++ b/src/crawl.rs
@@ -10,13 +10,14 @@ use std::collections::VecDeque;
 pub async fn crawl(
     start_instances: Vec<String>,
     max_depth: i32,
-) -> Result<Vec<InstanceDetails>, Error> {
+) -> Result<(Vec<InstanceDetails>, i32), Error> {
     let mut pending_instances: VecDeque<CrawlInstance> = start_instances
         .iter()
         .map(|s| CrawlInstance::new(s.to_string(), 0))
         .collect();
     let mut crawled_instances = vec![];
     let mut instance_details = vec![];
+    let mut failed_instances = 0;
     while let Some(current_instance) = pending_instances.pop_back() {
         crawled_instances.push(current_instance.domain.clone());
         if current_instance.depth > max_depth {
@@ -34,11 +35,14 @@ pub async fn crawl(
                     }
                 }
             }
-            Err(e) => eprintln!("Failed to crawl {}: {}", current_instance.domain, e),
+            Err(e) => {
+                failed_instances += 1;
+                eprintln!("Failed to crawl {}: {}", current_instance.domain, e)
+            },
         }
     }
 
-    Ok(instance_details)
+    Ok((instance_details, failed_instances))
 }
 
 #[derive(Serialize, Clone)]
diff --git a/src/main.rs b/src/main.rs
index 5742be8..4c16c2b 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -32,8 +32,8 @@ pub async fn main() -> Result<(), Error> {
     let start_instances = trusted_instances.iter().map(|s| s.to_string()).collect();
 
     eprintln!("Crawling...");
-    let instance_details = crawl(start_instances, max_crawl_depth).await?;
-    let total_stats = aggregate(instance_details);
+    let (instance_details, failed_instances) = crawl(start_instances, max_crawl_depth).await?;
+    let total_stats = aggregate(instance_details, failed_instances);
 
     println!("{}", serde_json::to_string_pretty(&total_stats)?);
     Ok(())
@@ -41,23 +41,25 @@ pub async fn main() -> Result<(), Error> {
 
 #[derive(Serialize)]
 struct TotalStats {
-    total_instances: i32,
+    crawled_instances: i32,
+    failed_instances: i32,
     total_users: i64,
     total_online_users: i32,
     instance_details: Vec<InstanceDetails>,
 }
 
-fn aggregate(instance_details: Vec<InstanceDetails>) -> TotalStats {
-    let mut total_instances = 0;
+fn aggregate(instance_details: Vec<InstanceDetails>, failed_instances: i32) -> TotalStats {
+    let mut crawled_instances = 0;
     let mut total_users = 0;
     let mut total_online_users = 0;
     for i in &instance_details {
-        total_instances += 1;
+        crawled_instances += 1;
         total_users += i.total_users;
         total_online_users += i.online_users;
     }
     TotalStats {
-        total_instances,
+        crawled_instances,
+        failed_instances,
         total_users,
         total_online_users,
         instance_details,