Skip to content

armgit / moli

Public Imported from GitHub
0

Best headless browser for AI agents. Lite, Fast, High-Compatibility. Built in Rust

Longer DOM Debugger waiting period

ldm0d41bf8ff9b1d
diff --git a/moli-protocol/src/domains/dom_debugger.rs b/moli-protocol/src/domains/dom_debugger.rs
index 312409a4f..6371b7a82 100644
--- a/moli-protocol/src/domains/dom_debugger.rs
+++ b/moli-protocol/src/domains/dom_debugger.rs
@@ -393,8 +393,13 @@ mod tests {
use crate::{conn::BrowserContext, testing::TestContext};
+ // Full-workspace CI runs these renderer-owner tests alongside CPU-heavy
+ // suites. Keep the guard diagnostic, but allow the same scheduling
+ // headroom as the test scheduler's external-input waits.
+ const DOM_DEBUGGER_COMMAND_TIMEOUT: Duration = Duration::from_secs(30);
+
async fn command(ctx: &mut TestContext, message: Value, command_id: u64) -> Value {
- tokio::time::timeout(Duration::from_secs(5), ctx.process_async(message))
+ tokio::time::timeout(DOM_DEBUGGER_COMMAND_TIMEOUT, ctx.process_async(message))
.await
.unwrap_or_else(|_| {
panic!("timed out processing DOMDebugger test command {command_id}")
@@ -417,7 +422,7 @@ mod tests {
.collect::<Vec<_>>();
let (mut messages, scheduler_events) = tokio::time::timeout(
- Duration::from_secs(5),
+ DOM_DEBUGGER_COMMAND_TIMEOUT,
ctx.complete_command_task_step_for_test(evaluate_step),
)
.await
@@ -425,7 +430,7 @@ mod tests {
assert!(scheduler_events.is_empty(), "{scheduler_events:?}");
for resume_step in resume_steps {
let (mut resume_messages, scheduler_events) = tokio::time::timeout(
- Duration::from_secs(5),
+ DOM_DEBUGGER_COMMAND_TIMEOUT,
ctx.complete_command_task_step_for_test(resume_step),
)
.await