feat: [#675] tracker checker (HTTP tracker) supports more service address formats

josecelano · josecelano · commit faee02f25781 · 2024-09-12T17:37:52.000+01:00
Now it supports a path prefix. It will be remove by the client to build the "scrape" URLs. This type of URL is very common in tracker lists like in https://newtrackon.com/. ```console TORRUST_CHECKER_CONFIG='{ "udp_trackers": [], "http_trackers": [ "http://127.0.0.1:7070", "http://127.0.0.1:7070/", "http://127.0.0.1:7070/announce" ], "health_checks": [] }' cargo run --bin tracker_checker ```
diff --git a/src/console/clients/checker/checks/http.rs b/src/console/clients/checker/checks/http.rs
@@ -28,21 +28,24 @@ pub async fn run(http_trackers: Vec<Url>, timeout: Duration) -> Vec<Result<Check
     tracing::debug!("HTTP trackers ...");
 
     for ref url in http_trackers {
+        let mut base_url = url.clone();
+        base_url.set_path("");
+
         let mut checks = Checks {
             url: url.clone(),
             results: Vec::default(),
         };
 
         // Announce
         {
-            let check = check_http_announce(url, timeout).await.map(|_| ());
+            let check = check_http_announce(&base_url, timeout).await.map(|_| ());
 
             checks.results.push((Check::Announce, check));
         }
 
         // Scrape
         {
-            let check = check_http_scrape(url, timeout).await.map(|_| ());
+            let check = check_http_scrape(&base_url, timeout).await.map(|_| ());
 
             checks.results.push((Check::Scrape, check));
         }
diff --git a/src/console/clients/checker/config.rs b/src/console/clients/checker/config.rs
@@ -215,7 +215,7 @@ mod tests {
         }
 
         mod http_trackers {
-            use crate::console::clients::checker::config::{Configuration, PlainConfiguration};
+            use crate::console::clients::checker::config::{Configuration, PlainConfiguration, ServiceUrl};
 
             #[test]
             fn it_should_fail_when_a_tracker_http_url_is_invalid() {
@@ -227,6 +227,41 @@ mod tests {
 
                 assert!(Configuration::try_from(plain_config).is_err());
             }
+
+            #[test]
+            fn it_should_allow_the_url_to_contain_a_path() {
+                // This is the common format for HTTP tracker URLs:
+                // http://domain.com:7070/announce
+
+                let plain_config = PlainConfiguration {
+                    udp_trackers: vec![],
+                    http_trackers: vec!["http://127.0.0.1:7070/announce".to_string()],
+                    health_checks: vec![],
+                };
+
+                let config = Configuration::try_from(plain_config).expect("Invalid plain configuration");
+
+                assert_eq!(
+                    config.http_trackers[0],
+                    "http://127.0.0.1:7070/announce".parse::<ServiceUrl>().unwrap()
+                );
+            }
+
+            #[test]
+            fn it_should_allow_the_url_to_contain_an_empty_path() {
+                let plain_config = PlainConfiguration {
+                    udp_trackers: vec![],
+                    http_trackers: vec!["http://127.0.0.1:7070/".to_string()],
+                    health_checks: vec![],
+                };
+
+                let config = Configuration::try_from(plain_config).expect("Invalid plain configuration");
+
+                assert_eq!(
+                    config.http_trackers[0],
+                    "http://127.0.0.1:7070/".parse::<ServiceUrl>().unwrap()
+                );
+            }
         }
 
         mod health_checks {

Original file line number	Diff line number	Diff line change
`@@ -28,21 +28,24 @@ pub async fn run(http_trackers: Vec<Url>, timeout: Duration) -> Vec<Result<Check`
`28`	`28`	`tracing::debug!("HTTP trackers ...");`
`29`	`29`
`30`	`30`	`for ref url in http_trackers {`
	`31`	`+ let mut base_url = url.clone();`
	`32`	`+ base_url.set_path("");`
	`33`	`+`
`31`	`34`	`let mut checks = Checks {`
`32`	`35`	`url: url.clone(),`
`33`	`36`	`results: Vec::default(),`
`34`	`37`	`};`
`35`	`38`
`36`	`39`	`// Announce`
`37`	`40`	`{`
`38`		`- let check = check_http_announce(url, timeout).await.map(\|_\| ());`
	`41`	`+ let check = check_http_announce(&base_url, timeout).await.map(\|_\| ());`
`39`	`42`
`40`	`43`	`checks.results.push((Check::Announce, check));`
`41`	`44`	`}`
`42`	`45`
`43`	`46`	`// Scrape`
`44`	`47`	`{`
`45`		`- let check = check_http_scrape(url, timeout).await.map(\|_\| ());`
	`48`	`+ let check = check_http_scrape(&base_url, timeout).await.map(\|_\| ());`
`46`	`49`
`47`	`50`	`checks.results.push((Check::Scrape, check));`
`48`	`51`	`}`