On Tue, Oct 6, 2026 at 3:10 PM <[email protected]> wrote:
>
> This is an automated email from the ASF dual-hosted git repository.
>
> markt-asf pushed a commit to branch 11.0.x
> in repository https://gitbox.apache.org/repos/asf/tomcat.git
>
>
> The following commit(s) were added to refs/heads/11.0.x by this push:
> new 713c0c61ac Skip test on MacOS
> 713c0c61ac is described below
>
> commit 713c0c61acde39c12c6ceb7afaed4196e47103ef
> Author: remm <[email protected]>
> AuthorDate: Wed Sep 30 15:09:35 2026 +0200
>
> Skip test on MacOS
>
> The explanation does sound convincing, so it seems best to skip it.
Oops. It was not failing on the other branches on CI ... I should have
double checked why.
Rémy
> ---
> .../apache/tomcat/security/TestSecurity2026WebSocket.java | 12
> ++++++++++++
> 1 file changed, 12 insertions(+)
>
> diff --git a/test/org/apache/tomcat/security/TestSecurity2026WebSocket.java
> b/test/org/apache/tomcat/security/TestSecurity2026WebSocket.java
> index 43a34f5ccd..9e852f08a1 100644
> --- a/test/org/apache/tomcat/security/TestSecurity2026WebSocket.java
> +++ b/test/org/apache/tomcat/security/TestSecurity2026WebSocket.java
> @@ -34,6 +34,7 @@ import jakarta.websocket.WebSocketContainer;
> import jakarta.websocket.server.ServerEndpointConfig;
>
> import org.junit.Assert;
> +import org.junit.Assume;
> import org.junit.Test;
>
> import org.apache.catalina.Context;
> @@ -91,6 +92,17 @@ public class TestSecurity2026WebSocket extends
> WebSocketBaseTest {
> */
> @Test
> public void testCVE_2026_77791() throws Exception {
> + /*
> + * Skipped on macOS. The test relies on a stalled (zero TCP receive
> window) connection keeping the server's
> + * pending write stuck for the whole blocking send timeout. The
> macOS TCP stack does not honor that premise:
> + * it tends to tear such connections down early (e.g. RST on
> segments overlapping the zero window) so the
> + * pending write fails after a few polling iterations. The server
> then closes long before the blocking send
> + * timeout expires (observed: ~126ms vs the expected ~2s), which
> trips the sanity check below without the
> + * wait under test ever having been exercised. This is an OS TCP
> behavior difference, not a Tomcat issue.
> + */
> + Assume.assumeFalse("macOS TCP stack does not keep the stalled write
> stuck. See comment in code.",
> + System.getProperty("os.name").contains("OS X"));
> +
> serverSendLatch = new CountDownLatch(1);
> // Never released. The server's pending data message write must
> never complete, forcing
> // acquireMessagePartInProgressSemaphore() to wait for the full
> blocking send timeout.
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
