]> git.ipfire.org Git - thirdparty/systemd.git/blob - CODING_STYLE
treewide: auto-convert the simple cases to log_*_errno()
[thirdparty/systemd.git] / CODING_STYLE
1 - 8ch indent, no tabs
2
3 - Don't break code lines too eagerly. We do *not* force line breaks at
4 80ch, all of today's screens should be much larger than that. But
5 then again, don't overdo it, ~140ch should be enough really.
6
7 - Variables and functions *must* be static, unless they have a
8 prototype, and are supposed to be exported.
9
10 - structs in MixedCase (with exceptions, such as public API structs),
11 variables + functions in lower_case.
12
13 - The destructors always unregister the object from the next bigger
14 object, not the other way around
15
16 - To minimize strict aliasing violations, we prefer unions over casting
17
18 - For robustness reasons, destructors should be able to destruct
19 half-initialized objects, too
20
21 - Error codes are returned as negative Exxx. e.g. return -EINVAL. There
22 are some exceptions: for constructors, it is OK to return NULL on
23 OOM. For lookup functions, NULL is fine too for "not found".
24
25 Be strict with this. When you write a function that can fail due to
26 more than one cause, it *really* should have "int" as return value
27 for the error code.
28
29 - Do not bother with error checking whether writing to stdout/stderr
30 worked.
31
32 - Do not log errors from "library" code, only do so from "main
33 program" code. (With one exception: it is OK to log with DEBUG level
34 from any code, with the exception of maybe inner loops).
35
36 - Always check OOM. There is no excuse. In program code, you can use
37 "log_oom()" for then printing a short message, but not in "library" code.
38
39 - Do not issue NSS requests (that includes user name and host name
40 lookups) from PID 1 as this might trigger deadlocks when those
41 lookups involve synchronously talking to services that we would need
42 to start up
43
44 - Do not synchronously talk to any other service from PID 1, due to
45 risk of deadlocks
46
47 - Avoid fixed-size string buffers, unless you really know the maximum
48 size and that maximum size is small. They are a source of errors,
49 since they possibly result in truncated strings. It is often nicer
50 to use dynamic memory, alloca() or VLAs. If you do allocate fixed-size
51 strings on the stack, then it is probably only OK if you either
52 use a maximum size such as LINE_MAX, or count in detail the maximum
53 size a string can have. (DECIMAL_STR_MAX and DECIMAL_STR_WIDTH
54 macros are your friends for this!)
55
56 Or in other words, if you use "char buf[256]" then you are likely
57 doing something wrong!
58
59 - Stay uniform. For example, always use "usec_t" for time
60 values. Do not mix usec and msec, and usec and whatnot.
61
62 - Make use of _cleanup_free_ and friends. It makes your code much
63 nicer to read!
64
65 - Be exceptionally careful when formatting and parsing floating point
66 numbers. Their syntax is locale dependent (i.e. "5.000" in en_US is
67 generally understood as 5, while on de_DE as 5000.).
68
69 - Try to use this:
70
71 void foo() {
72 }
73
74 instead of this:
75
76 void foo()
77 {
78 }
79
80 But it is OK if you do not.
81
82 - Single-line "if" blocks should not be enclosed in {}. Use this:
83
84 if (foobar)
85 waldo();
86
87 instead of this:
88
89 if (foobar) {
90 waldo();
91 }
92
93 - Do not write "foo ()", write "foo()".
94
95 - Please use streq() and strneq() instead of strcmp(), strncmp() where applicable.
96
97 - Please do not allocate variables on the stack in the middle of code,
98 even if C99 allows it. Wrong:
99
100 {
101 a = 5;
102 int b;
103 b = a;
104 }
105
106 Right:
107
108 {
109 int b;
110 a = 5;
111 b = a;
112 }
113
114 - Unless you allocate an array, "double" is always the better choice
115 than "float". Processors speak "double" natively anyway, so this is
116 no speed benefit, and on calls like printf() "float"s get promoted
117 to "double"s anyway, so there is no point.
118
119 - Do not invoke functions when you allocate variables on the stack. Wrong:
120
121 {
122 int a = foobar();
123 uint64_t x = 7;
124 }
125
126 Right:
127
128 {
129 int a;
130 uint64_t x = 7;
131
132 a = foobar();
133 }
134
135 - Use "goto" for cleaning up, and only use it for that. i.e. you may
136 only jump to the end of a function, and little else. Never jump
137 backwards!
138
139 - Think about the types you use. If a value cannot sensibly be
140 negative, do not use "int", but use "unsigned".
141
142 - Do not use types like "short". They *never* make sense. Use ints,
143 longs, long longs, all in unsigned+signed fashion, and the fixed
144 size types uint32_t and so on, as well as size_t, but nothing else.
145
146 - Public API calls (i.e. functions exported by our shared libraries)
147 must be marked "_public_" and need to be prefixed with "sd_". No
148 other functions should be prefixed like that.
149
150 - In public API calls, you *must* validate all your input arguments for
151 programming error with assert_return() and return a sensible return
152 code. In all other calls, it is recommended to check for programming
153 errors with a more brutal assert(). We are more forgiving to public
154 users then for ourselves! Note that assert() and assert_return()
155 really only should be used for detecting programming errors, not for
156 runtime errors. assert() and assert_return() by usage of _likely_()
157 inform the compiler that he should not expect these checks to fail,
158 and they inform fellow programmers about the expected validity and
159 range of parameters.
160
161 - Never use strtol(), atoi() and similar calls. Use safe_atoli(),
162 safe_atou32() and suchlike instead. They are much nicer to use in
163 most cases and correctly check for parsing errors.
164
165 - For every function you add, think about whether it is a "logging"
166 function or a "non-logging" function. "Logging" functions do logging
167 on their own, "non-logging" function never log on their own and
168 expect their callers to log. All functions in "library" code,
169 i.e. in src/shared/ and suchlike must be "non-logging". Every time a
170 "logging" function calls a "non-logging" function, it should log
171 about the resulting errors. If a "logging" function calls another
172 "logging" function, then it should not generate log messages, so
173 that log messages are not generated twice for the same errors.
174
175 - Avoid static variables, except for caches and very few other
176 cases. Think about thread-safety! While most of our code is never
177 used in threaded environments, at least the library code should make
178 sure it works correctly in them. Instead of doing a lot of locking
179 for that, we tend to prefer using TLS to do per-thread caching (which
180 only works for small, fixed-size cache objects), or we disable
181 caching for any thread that is not the main thread. Use
182 is_main_thread() to detect whether the calling thread is the main
183 thread.
184
185 - Command line option parsing:
186 - Do not print full help() on error, be specific about the error.
187 - Do not print messages to stdout on error.
188 - Do not POSIX_ME_HARDER unless necessary, i.e. avoid "+" in option string.
189
190 - Do not write functions that clobber call-by-reference variables on
191 failure. Use temporary variables for these cases and change the
192 passed in variables only on success.
193
194 - When you allocate a file descriptor, it should be made O_CLOEXEC
195 right from the beginning, as none of our files should leak to forked
196 binaries by default. Hence, whenever you open a file, O_CLOEXEC must
197 be specified, right from the beginning.