From 8e73fbb1d64665bee687c140e66471b8ac24ee38 Mon Sep 17 00:00:00 2001 From: Sam Tobin-Hochstadt Date: Mon, 14 Sep 2026 10:01:52 -0400 Subject: [PATCH] Raise the per-test timeout to 180 seconds "racket-test/tests/future/fsema-lock.rkt" has drifted into the timeout. Its median duration was 64.9s over pushes 73464-73505, 84.0s over 73506-73726, and 89.7s over 73735-73747 -- against a 90s limit. It now times out on 54% of pushes, up from 2%. 180s leaves room for that drift and for the variation in machine load that pushed it over in the first place. Note this does not help "gui-test/framework/tests/pasteboard.rkt", which does not time out because it is slow: across 348 runs it finishes in under 10s eight times and around 35s another 314 times, with nothing at all between 45s and 89s and 26 runs killed at the limit. That is a hang, and a longer limit only makes it take longer to report. --- config.rkt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/config.rkt b/config.rkt index 7eafa2f..b116e71 100644 --- a/config.rkt +++ b/config.rkt @@ -15,7 +15,7 @@ (vncviewer-path "/usr/bin/vncviewer") (current-make-install-timeout-seconds (* 5 60 60)) (current-make-timeout-seconds (* 5 60 60)) -(current-subprocess-timeout-seconds 90) +(current-subprocess-timeout-seconds 180) (current-monitoring-interval-seconds 60) (number-of-cpus 18)