1# The curl Test Suite
2
3# Running
4
5## Requires to run
6
7  - perl (and a unix-style shell)
8  - python (and a unix-style shell, for SMB and TELNET tests)
9  - python-impacket (for SMB tests)
10  - diff (when a test fails, a diff is shown)
11  - stunnel (for HTTPS and FTPS tests)
12  - OpenSSH or SunSSH (for SCP, SFTP and SOCKS4/5 tests)
13  - nghttpx (for HTTP/2 tests)
14  - nroff (for --manual tests)
15
16### Installation of python-impacket
17
18  The Python-based test servers support both recent Python 2 and 3.
19  You can figure out your default Python interpreter with python -V
20
21  Please install python-impacket in the correct Python environment.
22  You can use pip or your OS' package manager to install 'impacket'.
23
24  On Debian/Ubuntu the package names are:
25
26  -  Python 2: 'python-impacket'
27  -  Python 3: 'python3-impacket'
28
29  On FreeBSD the package names are:
30
31  -  Python 2: 'py27-impacket'
32  -  Python 3: 'py37-impacket'
33
34  On any system where pip is available:
35
36  -  Python 2: 'pip2 install impacket'
37  -  Python 3: 'pip3 install impacket'
38
39  You may also need to manually install the Python package 'six'
40  as that may be a missing requirement for impacket on Python 3.
41
42### Port numbers used by test servers
43
44  All test servers run on "random" port numbers. All tests should be written
45  to use suitable variables instead of fixed port numbers so that test cases
46  continue to work independent on what port numbers the test servers actually
47  use.
48
49  See [FILEFORMAT](FILEFORMAT.md) for the port number variables.
50
51### Test servers
52
53  The test suite runs simple FTP, POP3, IMAP, SMTP, HTTP and TFTP stand-alone
54  servers on the ports listed above to which it makes requests. For SSL tests,
55  it runs stunnel to handle encryption to the regular servers. For SSH, it
56  runs a standard OpenSSH server. For SOCKS4/5 tests SSH is used to perform
57  the SOCKS functionality and requires a SSH client and server.
58
59  The base port number (8990), which all the individual port numbers are
60  indexed from, can be set explicitly using runtests.pl' -b option to allow
61  running more than one instance of the test suite simultaneously on one
62  machine, or just move the servers in case you have local services on any of
63  those ports.
64
65  The HTTP server supports listening on a Unix domain socket, the default
66  location is 'http.sock'.
67
68### Run
69
70  `./configure && make && make test`. This builds the test suite support code
71  and invokes the 'runtests.pl' perl script to run all the tests. Edit the top
72  variables of that script in case you have some specific needs, or run the
73  script manually (after the support code has been built).
74
75  The script breaks on the first test that doesn't do OK. Use `-a` to prevent
76  the script from aborting on the first error. Run the script with `-v` for
77  more verbose output. Use `-d` to run the test servers with debug output
78  enabled as well. Specifying `-k` keeps all the log files generated by the
79  test intact.
80
81  Use `-s` for shorter output, or pass test numbers to run specific tests only
82  (like `./runtests.pl 3 4` to test 3 and 4 only). It also supports test case
83  ranges with 'to', as in `./runtests.pl 3 to 9` which runs the seven tests
84  from 3 to 9. Any test numbers starting with ! are disabled, as are any test
85  numbers found in the files `data/DISABLED` or `data/DISABLED.local` (one per
86  line). The latter is meant for local temporary disables and will be ignored
87  by git.
88
89  When `-s` is not present, each successful test will display on one line the
90  test number and description and on the next line a set of flags, the test
91  result, current test sequence, total number of tests to be run and an
92  estimated amount of time to complete the test run. The flags consist of
93  these letters describing what is checked in this test:
94
95    s stdout
96    d data
97    u upload
98    p protocol
99    o output
100    e exit code
101    m memory
102    v valgrind
103
104### Shell startup scripts
105
106  Tests which use the ssh test server, SCP/SFTP/SOCKS tests, might be badly
107  influenced by the output of system wide or user specific shell startup
108  scripts, .bashrc, .profile, /etc/csh.cshrc, .login, /etc/bashrc, etc. which
109  output text messages or escape sequences on user login.  When these shell
110  startup messages or escape sequences are output they might corrupt the
111  expected stream of data which flows to the sftp-server or from the ssh
112  client which can result in bad test behaviour or even prevent the test
113  server from running.
114
115  If the test suite ssh or sftp server fails to start up and logs the message
116  'Received message too long' then you are certainly suffering the unwanted
117  output of a shell startup script.  Locate, cleanup or adjust the shell
118  script.
119
120### Memory test
121
122  The test script will check that all allocated memory is freed properly IF
123  curl has been built with the `CURLDEBUG` define set. The script will
124  automatically detect if that is the case, and it will use the
125  'memanalyze.pl' script to analyze the memory debugging output.
126
127  Also, if you run tests on a machine where valgrind is found, the script will
128  use valgrind to run the test with (unless you use `-n`) to further verify
129  correctness.
130
131  runtests.pl's `-t` option will enable torture testing mode, which runs each
132  test many times and makes each different memory allocation fail on each
133  successive run.  This tests the out of memory error handling code to ensure
134  that memory leaks do not occur even in those situations. It can help to
135  compile curl with `CPPFLAGS=-DMEMDEBUG_LOG_SYNC` when using this option, to
136  ensure that the memory log file is properly written even if curl crashes.
137
138### Debug
139
140  If a test case fails, you can conveniently get the script to invoke the
141  debugger (gdb) for you with the server running and the exact same command
142  line parameters that failed. Just invoke `runtests.pl <test number> -g` and
143  then just type 'run' in the debugger to perform the command through the
144  debugger.
145
146### Logs
147
148  All logs are generated in the log/ subdirectory (it is emptied first in the
149  runtests.pl script). They remain in there after a test run.
150
151### Test input files
152
153  All test cases are put in the `data/` subdirectory. Each test is stored in
154  the file named according to the test number.
155
156  See [FILEFORMAT.md](FILEFORMAT.md) for a description of the test case file
157  format.
158
159### Code coverage
160
161  gcc provides a tool that can determine the code coverage figures for the
162  test suite.  To use it, configure curl with `CFLAGS='-fprofile-arcs
163  -ftest-coverage -g -O0`.  Make sure you run the normal and torture tests to
164  get more full coverage, i.e. do:
165
166    make test
167    make test-torture
168
169  The graphical tool ggcov can be used to browse the source and create
170  coverage reports on *NIX hosts:
171
172    ggcov -r lib src
173
174  The text mode tool gcov may also be used, but it doesn't handle object files
175  in more than one directory very well.
176
177### Remote testing
178
179  The runtests.pl script provides some hooks to allow curl to be tested on a
180  machine where perl can not be run.  The test framework in this case runs on
181  a workstation where perl is available, while curl itself is run on a remote
182  system using ssh or some other remote execution method.  See the comments at
183  the beginning of runtests.pl for details.
184
185## Test case numbering
186
187  Test cases used to be numbered by category ranges, but the ranges filled
188  up. Subsets of tests can now be selected by passing keywords to the
189  runtests.pl script via the make `TFLAGS` variable.
190
191  New tests are added by finding a free number in `tests/data/Makefile.inc`.
192
193## Write tests
194
195  Here's a quick description on writing test cases. We basically have three
196  kinds of tests: the ones that test the curl tool, the ones that build small
197  applications and test libcurl directly and the unit tests that test
198  individual (possibly internal) functions.
199
200### test data
201
202  Each test has a master file that controls all the test data. What to read,
203  what the protocol exchange should look like, what exit code to expect and
204  what command line arguments to use etc.
205
206  These files are `tests/data/test[num]` where `[num]` is just a unique
207  identifier described above, and the XML-like file format of them is
208  described in the separate [FILEFORMAT.md](FILEFORMAT.md) document.
209
210### curl tests
211
212  A test case that runs the curl tool and verifies that it gets the correct
213  data, it sends the correct data, it uses the correct protocol primitives
214  etc.
215
216### libcurl tests
217
218  The libcurl tests are identical to the curl ones, except that they use a
219  specific and dedicated custom-built program to run instead of "curl". This
220  tool is built from source code placed in `tests/libtest` and if you want to
221  make a new libcurl test that is where you add your code.
222
223### unit tests
224
225  Unit tests are placed in `tests/unit`. There's a tests/unit/README
226  describing the specific set of checks and macros that may be used when
227  writing tests that verify behaviors of specific individual functions.
228
229  The unit tests depend on curl being built with debug enabled.
230