xref: /openbsd-src/share/man/man5/bsd.regress.mk.5 (revision 1ad61ae0a79a724d2d3ec69e69c8e1d1ff6b53a0)
1.\" $OpenBSD: bsd.regress.mk.5,v 1.25 2023/09/25 05:43:22 claudio Exp $
2.\"
3.\" Copyright (c) 2002 Anil Madhavapeddy
4.\" Copyright (c) 2000 Marc Espie
5.\"
6.\" All rights reserved.
7.\"
8.\" Redistribution and use in source and binary forms, with or without
9.\" modification, are permitted provided that the following conditions
10.\" are met:
11.\" 1. Redistributions of source code must retain the above copyright
12.\"    notice, this list of conditions and the following disclaimer.
13.\" 2. Redistributions in binary form must reproduce the above copyright
14.\"    notice, this list of conditions and the following disclaimer in the
15.\"    documentation and/or other materials provided with the distribution.
16.\"
17.\" THIS SOFTWARE IS PROVIDED BY THE DEVELOPERS ``AS IS'' AND ANY EXPRESS OR
18.\" IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
19.\" OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
20.\" IN NO EVENT SHALL THE DEVELOPERS BE LIABLE FOR ANY DIRECT, INDIRECT,
21.\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
22.\" NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
23.\" DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
24.\" THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
25.\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
26.\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
27.\"
28.Dd $Mdocdate: September 25 2023 $
29.Dt BSD.REGRESS.MK 5
30.Os
31.Sh NAME
32.Nm bsd.regress.mk
33.Nd regression test master Makefile fragment
34.Sh SYNOPSIS
35.Fd .include <bsd.regress.mk>
36.Sh DESCRIPTION
37.Nm
38holds the standard routines used by the source regression tests.
39Some variables and targets are for its internal use only.
40The rest are documented here.
41.Pp
42Since this file also includes
43.Nm bsd.prog.mk ,
44all of the usual
45.Ox
46Makefile variables may be used to build the regression
47test programs.
48.Sh TARGETS
49.Bl -tag -width regress
50.It Cm depend
51Build any dependencies required to carry out the current set
52of regression tests.
53.It Cm regress
54Executes all of the regression targets defined in the Makefile.
55If one of the tests fails, the line
56.Qq FAILED
57is printed to the standard output.
58By default, execution continues with the next test and, after running
59all the tests,
60.Sy make Cm regress
61exits with an exit status of 0 even if some of the tests failed.
62.It Cm run-regress-*
63Runs an individual regression test.
64If the exit status of the program indicates an error or timeout,
65then a failure is logged, otherwise the test is marked as a success.
66.El
67.Sh VARIABLES
68.Bl -tag -width Ds
69.It Ev REGRESS_CLEANUP
70If this variable is set, it contains a target that is executed
71after all regress targets.
72.It Ev REGRESS_EXPECTED_FAILURES
73Targets which are expected to fail.
74Specifically: sometimes tests are added for features that should be
75supported eventually, but are somewhat complicated to fix right away.
76.It Ev REGRESS_FAIL_EARLY
77If this variable is set to anything but
78.Dq no ,
79the
80.Cm regress
81target will abort as soon as a test fails.
82Defaults to
83.Dq yes
84unless
85.Ev REGRESS_LOG
86is set.
87.It Ev REGRESS_LOG
88Points to the fully-qualified path of a file to which regression
89results are appended.
90Defaults to
91.Pa /dev/null .
92If set to any other path,
93.Ev REGRESS_FAIL_EARLY
94defaults to
95.Dq no .
96.It Ev REGRESS_ROOT_TARGETS
97Targets for which root access is required to run the test.
98The
99.Ev SUDO
100variable should be used in the test parts that require root.
101See also
102.Ev SUDO .
103.It Ev REGRESS_SETUP
104If this variable is set, it contains a target that is executed
105before each regress target.
106Regress targets depend on this one to enforce setup.
107.It Ev REGRESS_SETUP_ONCE
108If this variable is set, it contains a target that is executed
109once before all regress targets.
110Regress targets depend on a stamp file generated during setup.
111.It Ev REGRESS_SKIP_SLOW
112If this variable is not empty, skip over all the regression
113tests which have been marked as being 'slow' using the
114.Ev REGRESS_SLOW_TARGETS
115variable.
116.It Ev REGRESS_SKIP_TARGETS
117Targets which are unconditionally skipped.
118.It Ev REGRESS_SLOW_TARGETS
119Targets which are defined as 'slow'.
120All of these tests can be skipped by setting the
121.Ev REGRESS_SKIP_SLOW
122variable.
123.It Ev REGRESS_TARGETS
124Targets which are invoked to run the set of regression tests
125for this Makefile.
126Defaults to
127.Cm run-regress-${PROG} .
128.It Ev SUDO
129Location of a command used to switch to root for certain
130test targets which require it.
131See
132.Xr doas 1 .
133.El
134.Pp
135Some variables are intended to be set at runtime in the environment
136or in
137.Xr mk.conf 5 ,
138but not in the regress Makefile itself.
139.Sh GUIDELINES
140If an individual test passes,
141.Sy make Ar testname
142should return with an exit status of 0.
143If it fails, it should return with a non-zero exit status.
144.Pp
145If a test cannot be executed because a package is not installed or
146some environment variable is not set,
147.Sy make Ar testname
148should print
149.Qq SKIPPED
150to stdout and exit with status 0.
151To skip everything, implement the
152.Cm regress
153target with a command that prints
154.Qq SKIPPED .
155.Pp
156Some tests may require a special setup on the test machine that has
157to be done manually before testing.
158This requirement has to be documented in the Makefile or in a
159.Pa README
160file.
161The test should find out whether the setup exists before running
162and print
163.Qq SKIPPED
164and exit if it is missing.
165.Pp
166Tests should not fail because an intended feature has not been
167implemented yet.
168To avoid such false failures, a test should show the reason, print
169.Qq DISABLED
170to stdout and exit with status 0.
171.Pp
172Tests must be able to run with an
173.Pa obj
174directory.
175In case the test is not linked to the build or the tester forgot
176to run
177.Sy make Cm obj
178before, this directory or symlink may not exist.
179Then the test should run anyway.
180.Pp
181Tests are executed with
182.Sy make Cm regress ,
183but running
184.Sy make Cm all
185or
186.Sy make
187should have the same effect.
188Tests must be runnable by root, and may also succeed when run as a
189regular user.
190Tests must not assume they have a controlling tty,
191to allow them to be run by
192.Xr cron 8 .
193An individual regress test may create a pseudo tty if it needs one.
194.Pp
195Tests should use the binaries installed and the kernel running on
196the local system.
197They may use environment variables to test alternative binaries or
198remote kernels running on other machines.
199In some cases a test may need binaries or libraries or object files
200to be present in
201.Pa /usr/obj/
202that exist only after
203.Sy make Cm build
204was run in
205.Pa /usr/src/ .
206The test must not assume that they have already been built, but
207should run
208.Sy make
209in the appropriate source directory as a dependency.
210For missing generated source or header files a target called
211.Sy make Cm generated
212is common.
213The
214.Pa /usr/src/
215tree can be found with a relative path or with the
216.Ev BSDSRCDIR
217variable.
218.Pp
219Tests should generally not set or modify
220.Ev MALLOC_OPTIONS
221unless it is testing specific behaviour that depends on a particular
222flag or combination of flags.
223When running tests, the appropriate sysctl
224.Pq vm.malloc_conf
225could be used instead.
226.Sh SEE ALSO
227.Xr bsd.port.mk 5
228.Sh HISTORY
229The regression system originally came from
230.Nx ,
231with many tests added by
232.Ox
233since.
234The current Makefile framework was written by Artur Grabowski
235and Marc Espie for
236.Ox 3.1 .
237