summaryrefslogtreecommitdiffstats
path: root/tests/socket.test
diff options
context:
space:
mode:
authormax <max@tclers.tk>2011-06-22 14:21:28 (GMT)
committermax <max@tclers.tk>2011-06-22 14:21:28 (GMT)
commitd63052263676891c816f2dbc51362eaa0e3dc048 (patch)
tree9dbc94f9cd63a91cad539c0cf5e1bb4e12588cfc /tests/socket.test
parenta2c0c5611d68ee996777ad68e480daae28488ad9 (diff)
downloadtcl-d63052263676891c816f2dbc51362eaa0e3dc048.zip
tcl-d63052263676891c816f2dbc51362eaa0e3dc048.tar.gz
tcl-d63052263676891c816f2dbc51362eaa0e3dc048.tar.bz2
complete a comment in socket.testrmax_ipv6_branch
Diffstat (limited to 'tests/socket.test')
-rw-r--r--tests/socket.test3
1 files changed, 2 insertions, 1 deletions
diff --git a/tests/socket.test b/tests/socket.test
index 85d4d6f..e36914f 100644
--- a/tests/socket.test
+++ b/tests/socket.test
@@ -74,7 +74,8 @@ proc randport {} { expr {int(rand()*16383+49152)} }
# Test the latency of tcp connections over the loopback interface. Some OSes
# (e.g. NetBSD) seem to use the Nagle algorithm and delayed ACKs, so it takes
# up to 200ms for a packet sent to localhost to arrive. We're measuring this
-# here, so that OSes that don't have this problem can
+# here, so that OSes that don't have this problem can run the tests at full
+# speed.
set server [socket -server {apply {{s a p} {set ::s1 $s}}} 0]
set s2 [socket localhost [lindex [fconfigure $server -sockname] 2]]
vwait s1; close $server