Remove race condition from #388 - Use HTTP server for tests

Cargo.toml

@@ -56,6 +56,7 @@ schemars = "0.8.21"
 dirs = "5.0.1"
 once_cell = "1.20.2"
 opentelemetry = { version = "0.23.0", features = ["trace", "metrics", "logs", "otel_unstable"] }
+tiny_http = "0.12.0"
 
 # Features definition =========================================================
 [features]

crates/weaver_cache/src/lib.rs

@@ -531,6 +531,7 @@ impl RegistryRepo {
 #[cfg(test)]
 mod tests {
     use super::*;
+    use weaver_common::test::ServeStaticFiles;
 
     fn count_yaml_files(repo_path: &Path) -> usize {
         let count = walkdir::WalkDir::new(repo_path)
@@ -608,15 +609,25 @@ mod tests {
 
     #[test]
     fn test_semconv_registry_remote_tar_gz_archive() {
-        let registry_path = "https://github.com/open-telemetry/semantic-conventions/archive/refs/tags/v1.26.0.tar.gz[model]"
-            .parse::<RegistryPath>().unwrap();
+        let server = ServeStaticFiles::from("tests/test_data").unwrap();
+        let registry_path = format!(
+            "{}[model]",
+            server.relative_path_to_url("semconv_registry_v1.26.0.tar.gz")
+        )
+        .parse::<RegistryPath>()
+        .unwrap();
         check_archive(registry_path, Some("general.yaml"));
     }
 
     #[test]
     fn test_semconv_registry_remote_zip_archive() {
-        let registry_path = "https://github.com/open-telemetry/semantic-conventions/archive/refs/tags/v1.26.0.zip[model]"
-            .parse::<RegistryPath>().unwrap();
+        let server = ServeStaticFiles::from("tests/test_data").unwrap();
+        let registry_path = format!(
+            "{}[model]",
+            server.relative_path_to_url("semconv_registry_v1.26.0.zip")
+        )
+        .parse::<RegistryPath>()
+        .unwrap();
         check_archive(registry_path, Some("general.yaml"));
     }
 }

crates/weaver_common/Cargo.toml

@@ -16,7 +16,9 @@ paris = { version = "1.5.15", features = ["macros"] }
 serde.workspace = true
 serde_json.workspace = true
 miette.workspace = true
+thiserror.workspace = true
+tiny_http.workspace = true
 
 [dev-dependencies]
-thiserror.workspace = true
+ureq.workspace = true
 

crates/weaver_common/src/lib.rs

@@ -7,6 +7,7 @@ pub mod error;
 pub mod in_memory;
 pub mod quiet;
 pub mod result;
+pub mod test;
 
 use std::sync::atomic::{AtomicBool, AtomicUsize, Ordering};
 use std::sync::{Arc, Mutex};

crates/weaver_common/src/test.rs

@@ -0,0 +1,154 @@
+// SPDX-License-Identifier: Apache-2.0
+
+//! HTTP server for testing purposes.
+
+use paris::error;
+use std::ffi::OsStr;
+use std::fs::File;
+use std::path::PathBuf;
+use std::str::FromStr;
+use std::sync::Arc;
+use std::{collections::HashMap, thread::JoinHandle};
+use tiny_http::{Header, Response, Server, StatusCode};
+
+/// An error that can occur while starting the HTTP server.
+#[derive(thiserror::Error, Debug, Clone)]
+#[error("Internal HTTP server error: {error}")]
+pub struct HttpServerError {
+    error: String,
+}
+
+/// A struct that serves static files from a directory.
+pub struct ServeStaticFiles {
+    server: Arc<Server>,
+    port: u16,
+    request_handler: JoinHandle<()>,
+}
+
+impl Drop for ServeStaticFiles {
+    /// Stops the HTTP server.
+    fn drop(&mut self) {
+        // Test to see if we can force tiny_http to kill our thread, dropping the Arc
+        // before we continue to try to ensure `server` is dropped, cleaning
+        // open threads.
+        let mut attempts = 0;
+        while !self.request_handler.is_finished() && attempts < 10 {
+            self.server.unblock();
+            std::thread::yield_now();
+            attempts += 1;
+        }
+    }
+}
+
+impl ServeStaticFiles {
+    /// Creates a new HTTP server that serves static files from a directory.
+    /// Note: This server is only available for testing purposes.
+    pub fn from(static_path: impl Into<PathBuf>) -> Result<Self, HttpServerError> {
+        let server = Server::http("127.0.0.1:0").map_err(|e| HttpServerError {
+            error: e.to_string(),
+        })?;
+
+        let content_types: HashMap<&'static str, &'static str> = [
+            ("yaml", "application/yaml"),
+            ("json", "application/json"),
+            ("zip", "application/zip"),
+            ("gz", "application/gzip"),
+        ]
+        .iter()
+        .cloned()
+        .collect();
+
+        let static_path = static_path.into();
+        let server = Arc::new(server);
+        let server_clone = server.clone();
+        let port = server
+            .server_addr()
+            .to_ip()
+            .map(|ip| ip.port())
+            .unwrap_or(0);
+
+        let request_handler = std::thread::spawn(move || {
+            for request in server_clone.incoming_requests() {
+                let mut file_path = static_path.clone();
+                if request.url().len() > 1 {
+                    for chunk in request.url().trim_start_matches('/').split('/') {
+                        file_path.push(chunk);
+                    }
+                }
+
+                if !file_path.exists() {
+                    let status = StatusCode(404);
+                    request
+                        .respond(Response::empty(status))
+                        .expect("Failed to respond");
+                } else if let Ok(file) = File::open(&file_path) {
+                    let mut response = Response::from_file(file);
+                    let content_type = file_path
+                        .extension()
+                        .and_then(OsStr::to_str)
+                        .and_then(|ext| content_types.get(ext).copied())
+                        .unwrap_or("text/plain");
+                    response.add_header(
+                        Header::from_str(&format!("Content-Type: {}", content_type))
+                            .expect("Failed to parse header"),
+                    );
+                    request.respond(response).expect("Failed to respond");
+                } else {
+                    let status = StatusCode(500);
+                    request
+                        .respond(Response::empty(status))
+                        .expect("Failed to respond");
+                }
+            }
+        });
+
+        Ok(Self {
+            server,
+            port,
+            request_handler,
+        })
+    }
+
+    /// Returns the port of the server.
+    #[must_use]
+    pub fn port(&self) -> u16 {
+        self.port
+    }
+
+    /// Returns the URL of a file.
+    /// The file path should be relative to the static path.
+    #[must_use]
+    pub fn relative_path_to_url(&self, file: &str) -> String {
+        format!("http://127.0.0.1:{}/{}", self.port, file)
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use crate::test::ServeStaticFiles;
+
+    #[test]
+    fn test_http_server() {
+        let server = ServeStaticFiles::from("tests/test_data").unwrap();
+
+        assert!(server.port() > 0);
+
+        let content = ureq::get(&server.relative_path_to_url("file_a.yaml"))
+            .call()
+            .unwrap();
+        assert_eq!(content.status(), 200);
+        assert_eq!(content.header("Content-Type").unwrap(), "application/yaml");
+        assert_eq!(content.into_string().unwrap(), "file: A");
+
+        let content = ureq::get(&server.relative_path_to_url("file_b.yaml"))
+            .call()
+            .unwrap();
+        assert_eq!(content.status(), 200);
+        assert_eq!(content.header("Content-Type").unwrap(), "application/yaml");
+        assert_eq!(content.into_string().unwrap(), "file: B");
+
+        let result = ureq::get(&server.relative_path_to_url("unknown_file.yaml")).call();
+        assert!(result.is_err());
+        assert!(matches!(result.unwrap_err(), ureq::Error::Status(404, _)));
+    }
+}

crates/weaver_common/tests/test_data/file_a.yaml

@@ -0,0 +1 @@
+file: A

crates/weaver_common/tests/test_data/file_b.yaml

@@ -0,0 +1 @@
+file: B

crates/weaver_semconv/src/registry.rs

@@ -215,6 +215,7 @@ mod tests {
     use crate::group::{GroupSpec, GroupType};
     use crate::registry::SemConvRegistry;
     use crate::Error;
+    use weaver_common::test::ServeStaticFiles;
 
     #[test]
     fn test_try_from_path_pattern() {
@@ -237,9 +238,10 @@ mod tests {
 
     #[test]
     fn test_semconv_spec_from_url() {
-        let semconv_url = "https://raw.githubusercontent.com/open-telemetry/semantic-conventions/main/model/url/common.yaml";
+        let server = ServeStaticFiles::from("tests/test_data").unwrap();
+        let semconv_url = server.relative_path_to_url("url/common.yaml");
         let result =
-            SemConvRegistry::semconv_spec_from_url(semconv_url).into_result_failing_non_fatal();
+            SemConvRegistry::semconv_spec_from_url(&semconv_url).into_result_failing_non_fatal();
         assert!(result.is_ok());
     }
 

crates/weaver_semconv/src/semconv.rs

@@ -188,6 +188,7 @@ mod tests {
         InvalidAttribute, InvalidExampleWarning, InvalidSemConvSpec, RegistryNotFound,
     };
     use std::path::PathBuf;
+    use weaver_common::test::ServeStaticFiles;
 
     #[test]
     fn test_semconv_spec_from_file() {
@@ -303,26 +304,27 @@ mod tests {
 
     #[test]
     fn test_semconv_spec_from_url() {
+        let server = ServeStaticFiles::from("tests/test_data").unwrap();
         // Existing URL. The URL is a raw file from the semantic conventions repository.
         // This file is expected to be available.
-        let semconv_url = "https://raw.githubusercontent.com/open-telemetry/semantic-conventions/main/model/url/common.yaml";
-        let semconv_spec = SemConvSpec::from_url(semconv_url)
+        let semconv_url = server.relative_path_to_url("url/common.yaml");
+        let semconv_spec = SemConvSpec::from_url(&semconv_url)
             .into_result_failing_non_fatal()
             .unwrap();
         assert!(!semconv_spec.groups.is_empty());
 
         // Invalid semconv file
-        let semconv_url = "https://raw.githubusercontent.com/open-telemetry/semantic-conventions/main/model/README.md";
-        let semconv_spec = SemConvSpec::from_url(semconv_url).into_result_failing_non_fatal();
+        let semconv_url = server.relative_path_to_url("README.md");
+        let semconv_spec = SemConvSpec::from_url(&semconv_url).into_result_failing_non_fatal();
         assert!(semconv_spec.is_err());
         assert!(matches!(
             semconv_spec.unwrap_err(),
             InvalidSemConvSpec { .. }
         ));
 
-        // Non-existing URL (including both a leading underscore (which is not a valid domain) and a non-existing domain)
-        let semconv_url = "http://_unknown.com.invalid/unknown-semconv.yaml";
-        let semconv_spec = SemConvSpec::from_url(semconv_url).into_result_failing_non_fatal();
+        // Non-existing URL
+        let semconv_url = server.relative_path_to_url("unknown-semconv.yaml");
+        let semconv_spec = SemConvSpec::from_url(&semconv_url).into_result_failing_non_fatal();
         assert!(semconv_spec.is_err());
         assert!(matches!(semconv_spec.unwrap_err(), RegistryNotFound { .. }));
     }

crates/weaver_semconv/tests/test_data/README.md

@@ -0,0 +1,7 @@
+# YAML Model for Semantic Conventions
+
+The YAML descriptions of semantic convention contained in this directory are intended to
+be used by the various OpenTelemetry language implementations to aid in automatic
+generation of semantics-related code.
+
+...

crates/weaver_semconv/tests/test_data/url/common.yaml

@@ -0,0 +1,12 @@
+groups:
+  - id: url
+    brief: Attributes describing URL.
+    type: attribute_group
+    attributes:
+      - ref: url.scheme
+      - ref: url.full
+        tag: sensitive-information
+      - ref: url.path
+      - ref: url.query
+        tag: sensitive-information
+      - ref: url.fragment