diff options
author | Ken Raeburn <raeburn@mit.edu> | 2005-10-19 22:42:20 +0000 |
---|---|---|
committer | Ken Raeburn <raeburn@mit.edu> | 2005-10-19 22:42:20 +0000 |
commit | ffbb446429fc1917ef3a815bfab52907683bec7d (patch) | |
tree | 9abda136fb0247c424fad2c5d25996077f916eb1 /src/tests/dejagnu | |
parent | 8e431547d07ebd8e9bd12993445285813c7ac206 (diff) | |
download | krb5-ffbb446429fc1917ef3a815bfab52907683bec7d.tar.gz krb5-ffbb446429fc1917ef3a815bfab52907683bec7d.tar.xz krb5-ffbb446429fc1917ef3a815bfab52907683bec7d.zip |
This fixes an all-too-frequent nightly-testing failure on Tru64
* default.exp (setup_krb5_conf): Don't include a non-listening KDC port in the
client's config file.
git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@17434 dc483132-0cff-0310-8789-dd5450dbe970
Diffstat (limited to 'src/tests/dejagnu')
-rw-r--r-- | src/tests/dejagnu/config/ChangeLog | 5 | ||||
-rw-r--r-- | src/tests/dejagnu/config/default.exp | 10 |
2 files changed, 13 insertions, 2 deletions
diff --git a/src/tests/dejagnu/config/ChangeLog b/src/tests/dejagnu/config/ChangeLog index 82028c7c8e..cc348ee325 100644 --- a/src/tests/dejagnu/config/ChangeLog +++ b/src/tests/dejagnu/config/ChangeLog @@ -1,3 +1,8 @@ +2005-10-19 Ken Raeburn <raeburn@mit.edu> + + * default.exp (setup_krb5_conf): Don't include a non-listening KDC + port in the client's config file. + 2005-09-07 Ken Raeburn <raeburn@mit.edu> * default.exp: Initialize can_get_root to yes. diff --git a/src/tests/dejagnu/config/default.exp b/src/tests/dejagnu/config/default.exp index c3fd29425e..abb7b794ab 100644 --- a/src/tests/dejagnu/config/default.exp +++ b/src/tests/dejagnu/config/default.exp @@ -913,8 +913,14 @@ proc setup_krb5_conf { {type client} } { puts $conffile "" puts $conffile "\[realms\]" puts $conffile " $REALMNAME = \{" - # I hope nothing's listening on 15294... - puts $conffile " kdc = $hostname:[expr 6 + $portbase]" + # There's probably nothing listening here. It would be a good + # test for the handling of a non-responsive KDC address. However, + # on some systems, like Tru64, we often wind up with the client's + # socket bound to this address, causing our request to appear in + # our incoming queue as if it were a response, which causes test + # failures. If we were running the client and KDC on different + # hosts, this would be okay.... + #puts $conffile " kdc = $hostname:[expr 6 + $portbase]" puts $conffile " kdc = $hostname:[expr 1 + $portbase]" puts $conffile " admin_server = $hostname:[expr 4 + $portbase]" puts $conffile " kpasswd_server = $hostname:[expr 5 + $portbase]" |