README.regress
1Overview.
2
3$ ./configure && make tests
4
5You'll see some progress info. A failure will cause either the make to
6abort or the driver script to report a "FATAL" failure.
7
8The test consists of 2 parts. The first is the file-based tests which is
9driven by the Makefile, and the second is a set of network or proxycommand
10based tests, which are driven by a driver script (test-exec.sh) which is
11called multiple times by the Makefile.
12
13Failures in the first part will cause the Makefile to return an error.
14Failures in the second part will print a "FATAL" message for the failed
15test and continue.
16
17OpenBSD has a system-wide regression test suite. OpenSSH Portable's test
18suite is based on OpenBSD's with modifications.
19
20
21Environment variables.
22
23SUDO: path to sudo command, if desired. Note that some systems (notably
24 systems using PAM) require sudo to execute some tests.
25TEST_SSH_TRACE: set to "yes" for verbose output from tests
26TEST_SSH_QUIET: set to "yes" to suppress non-fatal output.
27TEST_SSH_x: path to "ssh" command under test, where x=SSH,SSHD,SSHAGENT,SSHADD
28 SSHKEYGEN,SSHKEYSCAN,SFTP,SFTPSERVER
29OBJ: used by test scripts to access build dir.
30TEST_SHELL: shell used for running the test scripts.
31TEST_SSH_PORT: TCP port to be used for the listening tests.
32TEST_SSH_SSH_CONFOPTS: Configuration directives to be added to ssh_config
33 before running each test.
34TEST_SSH_SSHD_CONFOTPS: Configuration directives to be added to sshd_config
35 before running each test.
36
37
38Individual tests.
39
40You can run an individual test from the top-level Makefile, eg:
41$ make tests LTESTS=agent-timeout
42
43If you need to manipulate the environment more you can invoke test-exec.sh
44directly if you set up the path to find the binaries under test and the
45test scripts themselves, for example:
46
47$ cd regress
48$ PATH=`pwd`/..:$PATH:. TEST_SHELL=/bin/sh sh test-exec.sh `pwd` \
49 agent-timeout.sh
50ok agent timeout test
51
52
53Files.
54
55test-exec.sh: the main test driver. Sets environment, creates config files
56and keys and runs the specified test.
57
58At the time of writing, the individual tests are:
59agent-timeout.sh: agent timeout test
60agent.sh: simple agent test
61broken-pipe.sh: broken pipe test
62connect-privsep.sh: proxy connect with privsep
63connect.sh: simple connect
64exit-status.sh: remote exit status
65forwarding.sh: local and remote forwarding
66keygen-change.sh: change passphrase for key
67keyscan.sh: keyscan
68proto-mismatch.sh: protocol version mismatch
69proto-version.sh: sshd version with different protocol combinations
70proxy-connect.sh: proxy connect
71sftp.sh: basic sftp put/get
72ssh-com-client.sh: connect with ssh.com client
73ssh-com-keygen.sh: ssh.com key import
74ssh-com-sftp.sh: basic sftp put/get with ssh.com server
75ssh-com.sh: connect to ssh.com server
76stderr-after-eof.sh: stderr data after eof
77stderr-data.sh: stderr data transfer
78transfer.sh: transfer data
79try-ciphers.sh: try ciphers
80yes-head.sh: yes pipe head
81
82
83Problems?
84
85Run the failing test with shell tracing (-x) turned on:
86$ PATH=`pwd`/..:$PATH:. sh -x test-exec.sh `pwd` agent-timeout.sh
87
88Failed tests can be difficult to diagnose. Suggestions:
89- run the individual test via ./test-exec.sh `pwd` [testname]
90- set LogLevel to VERBOSE in test-exec.sh and enable syslogging of
91 auth.debug (eg to /var/log/authlog).
92
93
94Known Issues.
95
96- Similarly, if you do not have "scp" in your system's $PATH then the
97 multiplex scp tests will fail (since the system's shell startup scripts
98 will determine where the shell started by sshd will look for scp).
99
100- Recent GNU coreutils deprecate "head -[n]": this will cause the yes-head
101 test to fail. The old behaviour can be restored by setting (and
102 exporting) _POSIX2_VERSION=199209 before running the tests.
103
104$Id: README.regress,v 1.12 2011/05/05 03:48:42 djm Exp $
105