5 \___|\___/|_| \_\_____|
11 1.2 Port numbers used by test servers
14 1.5 Shell startup scripts
23 2.1 Test case numbering
35 ==============================================================================
41 perl (and a unix-style shell)
42 python (and a unix-style shell)
43 diff (when a test fails, a diff is shown)
44 stunnel (for HTTPS and FTPS tests)
45 OpenSSH or SunSSH (for SCP, SFTP and SOCKS4/5 tests)
46 nghttpx (for HTTP/2 tests)
47 nroff (for --manual tests)
49 1.2 Port numbers used by test servers
55 - TCP/8994 for HTTP IPv6
56 - TCP/8995 for FTP (2)
57 - TCP/8996 for FTP IPv6
59 - UDP/8998 for TFTP IPv6
60 - TCP/8999 for SCP/SFTP
63 - TCP/9002 for POP3 IPv6
65 - TCP/9004 for IMAP IPv6
67 - TCP/9006 for SMTP IPv6
69 - TCP/9008 for RTSP IPv6
71 - TCP/9010 for GOPHER IPv6
72 - TCP/9011 for HTTPS server with TLS-SRP support
73 - TCP/9012 for HTTPS IPv6 server with TLS-SRP support
74 - TCP/9013 for HTTP proxy server for CONNECT
75 - TCP/9014 for HTTP pipelining server
76 - TCP/9015 for HTTP/2 server
77 - TCP/9016 for DICT server
78 - TCP/9017 for SMB server
79 - TCP/9018 for SMBS server (reserved)
80 - TCP/9019 for TELNET server with negotiation support
84 The test suite runs simple FTP, POP3, IMAP, SMTP, HTTP and TFTP stand-alone
85 servers on the ports listed above to which it makes requests. For SSL tests,
86 it runs stunnel to handle encryption to the regular servers. For SSH, it
87 runs a standard OpenSSH server. For SOCKS4/5 tests SSH is used to perform
88 the SOCKS functionality and requires a SSH client and server.
90 The base port number (8990), which all the individual port numbers are
91 indexed from, can be set explicitly using runtests.pl' -b option to allow
92 running more than one instance of the test suite simultaneously on one
93 machine, or just move the servers in case you have local services on any of
96 The HTTP server supports listening on a Unix domain socket, the default
97 location is 'http.sock'.
101 './configure && make && make test'. This builds the test suite support code
102 and invokes the 'runtests.pl' perl script to run all the tests. Edit the top
103 variables of that script in case you have some specific needs, or run the
104 script manually (after the support code has been built).
106 The script breaks on the first test that doesn't do OK. Use -a to prevent
107 the script from aborting on the first error. Run the script with -v for more
108 verbose output. Use -d to run the test servers with debug output enabled as
109 well. Specifying -k keeps all the log files generated by the test intact.
111 Use -s for shorter output, or pass test numbers to run specific tests only
112 (like "./runtests.pl 3 4" to test 3 and 4 only). It also supports test case
113 ranges with 'to', as in "./runtests 3 to 9" which runs the seven tests from
114 3 to 9. Any test numbers starting with ! are disabled, as are any test
115 numbers found in the files data/DISABLED or data/DISABLED.local (one per
116 line). The latter is meant for local temporary disables and will be ignored
119 When -s is not present, each successful test will display on one line the
120 test number and description and on the next line a set of flags, the test
121 result, current test sequence, total number of tests to be run and an
122 estimated amount of time to complete the test run. The flags consist of
123 these letters describing what is checked in this test:
134 1.5 Shell startup scripts
136 Tests which use the ssh test server, SCP/SFTP/SOCKS tests, might be badly
137 influenced by the output of system wide or user specific shell startup
138 scripts, .bashrc, .profile, /etc/csh.cshrc, .login, /etc/bashrc, etc. which
139 output text messages or escape sequences on user login. When these shell
140 startup messages or escape sequences are output they might corrupt the
141 expected stream of data which flows to the sftp-server or from the ssh
142 client which can result in bad test behaviour or even prevent the test
145 If the test suite ssh or sftp server fails to start up and logs the message
146 'Received message too long' then you are certainly suffering the unwanted
147 output of a shell startup script. Locate, cleanup or adjust the shell
152 The test script will check that all allocated memory is freed properly IF
153 curl has been built with the CURLDEBUG define set. The script will
154 automatically detect if that is the case, and it will use the
155 'memanalyze.pl' script to analyze the memory debugging output.
157 Also, if you run tests on a machine where valgrind is found, the script will
158 use valgrind to run the test with (unless you use -n) to further verify
161 runtests.pl's -t option will enable torture testing mode, which runs each
162 test many times and makes each different memory allocation fail on each
163 successive run. This tests the out of memory error handling code to ensure
164 that memory leaks do not occur even in those situations. It can help to
165 compile curl with CPPFLAGS=-DMEMDEBUG_LOG_SYNC when using this option, to
166 ensure that the memory log file is properly written even if curl crashes.
170 If a test case fails, you can conveniently get the script to invoke the
171 debugger (gdb) for you with the server running and the exact same command
172 line parameters that failed. Just invoke 'runtests.pl <test number> -g' and
173 then just type 'run' in the debugger to perform the command through the
178 All logs are generated in the log/ subdirectory (it is emptied first in the
179 runtests.pl script). Use runtests.pl -k to force it to keep the temporary
180 files after the test run since successful runs will clean it up otherwise.
184 All test cases are put in the data/ subdirectory. Each test is stored in the
185 file named according to the test number.
187 See FILEFORMAT for the description of the test case files.
191 gcc provides a tool that can determine the code coverage figures for
192 the test suite. To use it, configure curl with
193 CFLAGS='-fprofile-arcs -ftest-coverage -g -O0'. Make sure you run the normal
194 and torture tests to get more full coverage, i.e. do:
199 The graphical tool ggcov can be used to browse the source and create
200 coverage reports on *NIX hosts:
204 The text mode tool gcov may also be used, but it doesn't handle object files
205 in more than one directory very well.
209 The runtests.pl script provides some hooks to allow curl to be tested on a
210 machine where perl can not be run. The test framework in this case runs on
211 a workstation where perl is available, while curl itself is run on a remote
212 system using ssh or some other remote execution method. See the comments at
213 the beginning of runtests.pl for details.
217 2.1 Test case numbering
224 500 - 599 libcurl source code tests, not using the curl command tool
226 700 - 799 SOCKS4 (even numbers) and SOCK5 (odd numbers)
230 1000 - 1299 miscellaneous
231 1300 - 1399 unit tests
232 1400 - 1499 miscellaneous
233 1500 - 1599 libcurl source code tests, not using the curl command tool
235 1600 - 1699 unit tests
236 2000 - x multiple sequential protocols per test case
238 There's nothing in the system that *requires* us to keep within these number
243 Here's a quick description on writing test cases. We basically have three
244 kinds of tests: the ones that test the curl tool, the ones that build small
245 applications and test libcurl directly and the unit tests that test
246 individual (possibly internal) functions.
250 Each test has a master file that controls all the test data. What to read,
251 what the protocol exchange should look like, what exit code to expect and
252 what command line arguments to use etc.
254 These files are tests/data/test[num] where [num] is described in section 2
255 of this document, and the XML-like file format of them is described in the
256 separate tests/FILEFORMAT document.
260 A test case that runs the curl tool and verifies that it gets the correct
261 data, it sends the correct data, it uses the correct protocol primitives
266 The libcurl tests are identical to the curl ones, except that they use a
267 specific and dedicated custom-built program to run instead of "curl". This
268 tool is built from source code placed in tests/libtest and if you want to
269 make a new libcurl test that is where you add your code.
273 Unit tests are tests in the 13xx sequence and they are placed in tests/unit.
274 There's a tests/unit/README describing the specific set of checks and macros
275 that may be used when writing tests that verify behaviors of specific
276 individual functions.
278 The unit tests depend on curl being built with debug enabled.
284 Add tests for TELNET, LDAP, DICT...
288 SOCKS4/5 test deficiencies - no proxy authentication tests as SSH (the
289 test mechanism) doesn't support them