Commit | Line | Data |
---|---|---|
75e0e026 PH |
1 | $Cambridge: exim/doc/doc-txt/test-harness.txt,v 1.1 2006/02/16 10:05:33 ph10 Exp $ |
2 | ||
3 | EXIM'S BEHAVIOUR CHANGES WHEN RUNNING IN THE TEST HARNESS | |
4 | --------------------------------------------------------- | |
5 | ||
6 | When Exim is running in its test harness, via the scripts in the exim-testsuite | |
7 | distribution, its behaviour is altered in a few ways, in order to make the | |
8 | regression testing work properly. The documentation for the test suite | |
9 | describes how a copy of the Exim binary is taken and patched in order to get it | |
10 | to run in the test harness. This document briefly lists the behavioural changes | |
11 | that result. They come into play when the Boolean variable running_in_test_ | |
12 | harness is true. | |
13 | ||
14 | ||
15 | Privilege | |
16 | --------- | |
17 | ||
18 | Exim does not give up its root privilege when called with -C or -D, nor does it | |
19 | insist on the caller being an admin user when starting a daemon, a queue | |
20 | runner, or requesting debug output. | |
21 | ||
22 | ||
23 | Small Pauses | |
24 | ------------ | |
25 | ||
26 | In a number of places, typically when a subprocess has been forked, there are | |
27 | short pauses of half or one second in one of the processes. This allows the | |
28 | other process to "go first"; it ensures that debugging or logging output always | |
29 | appears in the same order. | |
30 | ||
31 | ||
32 | Daemon | |
33 | ------ | |
34 | ||
35 | The daemon always writes a pid file when running in the test harness. | |
36 | ||
37 | ||
38 | CRAM-MD5 | |
39 | -------- | |
40 | ||
41 | The cram_md5 authenticator always uses the same challenge string. | |
42 | ||
43 | ||
44 | Appendfile | |
45 | ---------- | |
46 | ||
47 | After a quota error, the "time since last read" for the file is forced to 10s, | |
48 | for repeatability. | |
49 | ||
50 | ||
51 | Memory management | |
52 | ----------------- | |
53 | ||
54 | Memory management debugging output contains only the store pool and the size | |
55 | (other information is too variable). New memory is initialized to contain F0 in | |
56 | all bytes. | |
57 | ||
58 | ||
59 | Queue running | |
60 | ------------- | |
61 | ||
62 | There's a facility (-Tqt) for fudging queue times for testing retry logic. | |
63 | ||
64 | ||
65 | Syslog | |
66 | ------ | |
67 | ||
68 | Exim never writes to syslog in the test harness. Attempts to do so are silently | |
69 | ignored. None of the tests actually specify syslog logging for any actual log | |
70 | lines, but there is one that tests the inability to open the main and panic | |
71 | logs, which by default then tries to write to syslog. | |
72 | ||
73 | ||
74 | SMTP connection timeout | |
75 | ----------------------- | |
76 | ||
77 | In order to be able to test timeout handling, a "connection refused" error is | |
78 | converted into a timeout if the timeout value is set to 999999s. | |
79 | ||
80 | ||
81 | Random numbers | |
82 | -------------- | |
83 | ||
84 | The seed for the pseudo-random number generator is set to a fixed value in the | |
85 | test harness, to ensure repeatability. | |
86 | ||
87 | ||
88 | Bounce messages | |
89 | --------------- | |
90 | ||
91 | When Exim is submitting a bounce message to itself, unless the configuration | |
92 | has set queue_only, it uses -odi so that the bounce is delivered before the | |
93 | subprocess returns. This avoids a race that might put log lines in an arbitrary | |
94 | order. | |
95 | ||
96 | ||
97 | DNS lookups | |
98 | ----------- | |
99 | ||
100 | The real DNS resolver is never called. Instead, a fake resolver, which runs as | |
101 | a separate program, is used. It is part of the test suite and is documented | |
102 | there. This ensures complete control over the exact results of any DNS lookups. | |
103 | ||
104 | An attempt to look up a PTR record for 99.99.99.99 or an IP address for a host | |
105 | whose name ends with .test.again.dns always yields a "try again" error. | |
106 | ||
107 | A fake function is called instead of gethostbyname(). It recognizes the name | |
108 | "manyhome.test.ex" and generates a humungous number of IP addresses. It also | |
109 | recognizes an unqualified "localhost" and forces it to the appropriate loopback | |
110 | address (IPv4 or IPv6, as required). IP addresses are treated as literals. For | |
111 | other names, it does a DNS lookup (which of course actually calls the fake | |
112 | resolver) to find the host name. | |
113 | ||
114 | ||
115 | User names | |
116 | ---------- | |
117 | ||
118 | If unknown_login is set, it forces the login name, thus overriding the actual | |
119 | login for the test suite caller. When this happens, unknown_username provides a | |
120 | user name if it is set; otherwise an empty string is used. | |
121 | ||
122 | ||
123 | Ident | |
124 | ----- | |
125 | ||
126 | If -bh is used and both the sending host port and the incoming interface port | |
127 | are supplied, an ident (RFC 1413) call is made for testing purposes. | |
128 | ||
129 | ||
130 | Debug output | |
131 | ------------ | |
132 | ||
133 | Debugging output from the function that waits for the clock to tick at an | |
134 | appropriate resolution (before completing the arrival of a message, for | |
135 | example) is suppressed because the fractions of seconds that it contains will | |
136 | never be repeatable. | |
137 | ||
138 | ||
139 | Philip Hazel | |
140 | 15 February 2006 |