summaryrefslogtreecommitdiff
path: root/build.txt
diff options
context:
space:
mode:
authorEric S. Raymond <esr@thyrsus.com>2014-08-24 15:29:50 -0400
committerEric S. Raymond <esr@thyrsus.com>2014-08-24 15:29:50 -0400
commitc0d020644bf925fb1010c86756e5051e50f18e51 (patch)
treeaeb20db9c4adb7a4f52b7e6206b73333a2d61d78 /build.txt
parent0ca4dac4db42ddf2f95bb60007c8f3f5b4f17619 (diff)
downloadgpsd-c0d020644bf925fb1010c86756e5051e50f18e51.tar.gz
Add slow=yes option for regression tests in an attempt to avoid flakeouts.
Diffstat (limited to 'build.txt')
-rw-r--r--build.txt1
1 files changed, 1 insertions, 0 deletions
diff --git a/build.txt b/build.txt
index 934c0230..f2fb8e05 100644
--- a/build.txt
+++ b/build.txt
@@ -328,6 +328,7 @@ framework shove data through the pty and socket layers *way* faster
than would ever occur in production. If you get regression-test
failures that aren't repeatable and look like the test framework is
sporadically failing to feed the last line or two of test loads, try
+using the slow=yes option with scons check. If that fails, try
increasing the delay constants in fake.py. If you have to do this,
please report your experience to the GPSD maintainers.