Previously we did “sleep 0.1” to wait for the mock node daemon (synthetic-events) to delete the control socket, when it ends. This is too short for some environments where the scenario gets run.
Fix this by waiting for the socket to be deleted for some reasonably large time limit, but checking for it being removed every second.
Signed-off-by: Lars Wirzenius liw@liw.fi
Previously we did “sleep 0.1” to wait for the mock node daemon (synthetic-events) to delete the control socket, when it ends. This is too short for some environments where the scenario gets run.
Fix this by waiting for the socket to be deleted for some reasonably large time limit, but checking for it being removed every second.
Signed-off-by: Lars Wirzenius liw@liw.fi
No GiHub Actions Workflows found.
No GiHub Actions Workflows found.
Rebase on main.
No GiHub Actions Workflows found.
fix: teach synthetic-events to connect to its daemon
This means we don’t need the non-portable nc.
Signed-off-by: Lars Wirzenius liw@liw.fi
No GiHub Actions Workflows found.
fix(src/bin/synthetic-events.rs):
Fix client code to close its write end and read until end. This avoids crashing the daemon.
Also: Refactor into smaller functions, for comprehensibility. Log errors in the daemon to the log file. Write errors from other parts to stderr.
Signed-off-by: Lars Wirzenius liw@liw.fi
No GiHub Actions Workflows found.
feat(src/bin/synthetic-events.rs): show daemon log when client fails
Signed-off-by: Lars Wirzenius liw@liw.fi
No GiHub Actions Workflows found.
debug: increase logging in daemon
Signed-off-by: Lars Wirzenius liw@liw.fi
No GiHub Actions Workflows found.
debug: try without timeout setting
Signed-off-by: Lars Wirzenius liw@liw.fi
No GiHub Actions Workflows found.
Rebase the patch set with a more helpful commit message
And a less confusing history.