File: test_programs_style.txt

package info (click to toggle)
xorp 1.8.5-4
  • links: PTS, VCS
  • area: main
  • in suites: jessie, jessie-kfreebsd
  • size: 28,560 kB
  • ctags: 54,995
  • sloc: cpp: 397,204; sh: 17,490; ansic: 17,029; python: 7,643; lex: 1,632; yacc: 1,474; awk: 956; makefile: 251; perl: 217; sed: 33
file content (134 lines) | stat: -rw-r--r-- 4,775 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
#
# $XORP: xorp/devnotes/test_programs_style.txt,v 1.1.1.1 2002/12/11 23:55:54 hodson Exp $
#

			XORP test programs style

The style described below describes a set of rules that should be
applied to test programs we write.  Those rules should be considered as
"work-in-progress".  Hopefully, as we write more test programs, we will
have some clues about what those programs should be about.

- DOCTRINE

  Write test programs to test your code!  If you don't write them,
  then your code does not work, or will not work when it is modified
  in the future. Period.

- TEST PROGRAM GRANULARITY

  "Shall I write one big program to test everything, or shall I write
  many small programs, one for each particular test?"

  In general, prefer smaller test programs.  However, if several tests
  are tightly related, they can be combined into one test program.
  Also, if there is a relatively large number of tests, and each test is
  just few lines, then it is fine to combine them in one test program.

- TEST PROGRAM INTERNAL IDENTIFIERS

  * PROGRAM_NAME: the program filename (excluding the path to it).
    TODO: shall we include the path as well, so we don't have to worry
    whether each program has an unique name?

  * PROGRAM_VERSION_ID: should be in the `number-dot' format, with no
    blank spaces.  Typically, two or three digits would be sufficient.
    Examples: 0.2 or 1.20.1

    Optionally, one of the letters 'a' or 'b' may be added to the end to
    indicate 'alpha' or 'beta' status of the program.
    Examples: 0.2a or 1.20.1b

- RETURN STATUS

  Return status is 0 if all tests succeed, 1 if any test fails, and 2 if
  a test fails due to an internal error.

- TEST PROGRAM OUTPUT

  By default, the test-related output is printed to the standard output,
  and is generated only if the test program is run with the -v or
  --verbose command-line option (see below).  This option does not apply
  to the internal error messages, which by default are printed to the
  standard error stream.

- COMMAND-LINE OPTIONS

  Each test program should support the following command-line options:

	-h
	--help
		Print help information on the standard output, and then
		exit with return status of 0.

	-v
	--verbose
		Print information about successful tests as well as
		unsuccessful tests.

	-V
	--version
		Print the program version ID on the standard output, and
		then exit with return status of 0.

  At minimum, a test program should support at least -h and -v
  options (i.e., their LONG_OPTIONS format may be omitted).

  A test program may support some of the following command-line options:

	--info
		Print detailed information about the program on the
		standard output, and then exit with return status of 0.
		The information to print should be readable by human,
		and does not have specific format.  Preferably, it
		should contain the following information about the
		program: name, description of its purpose, list of
		supported version (lastest first), last modification
		date, list of authors, copyright status, etc.
		Example:
	Name:		test_ipvx
	Description:	Test IPv4 and IPv6 address manipulation
	Versions:	0.5.1 0.3 0.2.2
	Date:		Oct 10, 2002
	Copyright:	See file LICENSE
	Return:		0 on success, 1 if test error, 2 if internal error

	--verbose-level=<level>
		Print information about successful tests as well
		as unsuccessful tests.  The level of verbosity is
		defined by <level>.  The particular assignment of the
		verbosity levels is program-specific, but higher level
		should correspond to more verbosity.  However, if
		the level is 0, then the program should not print any
		test-related information (i.e., if there are no internal
		errors, it should return the appropriate status only).
		The verbose level does not apply to the messages related
		to internal errors.

	--with-version=<version>
		Run the program with version <version>, if supported.
		If the specified version is not supported, the program
		should print the list of supported versions to the
		standard error stream, and then exit with return status
		of 2.


  A test program may support additional test-specific command-line
  options.  Each command-line option should be specified in the
  LONG_OPTIONS format (e.g., --help), but it may have a short version
  as well (e.g., -h).

  If a program is called with the wrong arguments, this should be
  treated as an internal error: the program should print the help
  information (the same one produced with -h) to the standard error
  stream, and then exit with return status of 2.

- TEST SCRIPTS

  TODO: adapt Eddie's email about file format of test scripts, etc.

       Command-line arguments for test scripts:
	
	--preserve-temporaries
		Preserve the temporary directory created for the test.