2
0
mirror of https://gitlab.isc.org/isc-projects/bind9 synced 2025-08-22 01:59:26 +00:00
bind/doc/dev/style.md

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

910 lines
29 KiB
Markdown
Raw Normal View History

<!--
Copyright (C) Internet Systems Consortium, Inc. ("ISC")
SPDX-License-Identifier: MPL-2.0
This Source Code Form is subject to the terms of the Mozilla Public
License, v. 2.0. If a copy of the MPL was not distributed with this
file, you can obtain one at https://mozilla.org/MPL/2.0/.
See the COPYRIGHT file distributed with this work for additional
information regarding copyright ownership.
-->
## BIND 9 Coding Style
BIND 9 is principally written in [C](#cstyle), with some additional code
written in [Python](#pystyle), [Perl](#plstyle) and
[Bourne shell](#shstyle). Style guidelines for each of these are
below.
### <a name="cstyle"></a>C
#### Compiler
A C11 compiler, library with C11 extensions and POSIX:2001 are assumed. Feel
free to use any C11 feature, but make sure to provide compatibility shims for
all supported platforms that don't support all of the C11 features.
#### Warnings
Given a reasonable set of things to warn about (e.g. -W -Wall for gcc), the goal
is to compile with no warnings.
#### Copyright Notices
The license described in the ``COPYING`` file applies to the BIND 9 source as a
whole, though individual source files can have a different license which is
required to be compatible with the MPL-2.0.
Aside from that, individual files can be provided under a dual license,
e.g. MPL-2.0 license and alternatively under a permissive license like BSD, MIT
etc.
The common way of expressing the license of a source file is to add the matching
boilerplate text into the top comment of the file. Due to formatting, typos
etc. these “boilerplates” are hard to validate for tools which are used in the
context of license compliance.
An alternative to boilerplate text is the use of Software Package Data Exchange
(SPDX) license identifiers in each source file. SPDX license identifiers are
machine parsable and precise shorthands for the license under which the content
of the file is contributed. SPDX license identifiers are managed by the SPDX
Workgroup at the Linux Foundation and have been agreed on by partners throughout
the industry, tool vendors, and legal teams. For further information see
https://spdx.org/
The BIND 9 requires the precise SPDX identifier in all source files. The valid
identifiers used in the BIND 9 are explained in the section License identifiers
and have been retrieved from the official SPDX license list at
https://spdx.org/licenses/ along with the license texts.
#### Indentation
Use tabs for indentation. Spaces before statements are only allowed when
needed to line up a continued expression. In the following example, spaces
used for indentation are indicated with `"_"`:
if (i == 0) {
printf("this is going to be %s very long %s statement\\n",
_______"a", "printf");
}
Text editors should be configured with tab-stop set to 8 characters, and
tabs should not be expanded to into spaces. The following `vim` settings
conform well to BIND 9 C style:
set showmatch
set showmode
set autoindent
set expandtab
filetype plugin on
let c_syntax_for_h = 1
autocmd FileType c,cc,cpp set cindent
autocmd FileType c,cc,cpp set cino=(0:0l1
autocmd FileType c,cc,cpp set fo=rotcq
autocmd FileType c,cc,cpp set noexpandtab ts=8
autocmd FileType python set ts=4 sw=4
filetype indent on
#### Vertical Whitespace
Vertical whitespace is encouraged for improved code legibility: closely
related statements should be grouped, and then the groups separated with a
single empty line. There should never be two or more empty lines adjacent
to one another.
#### Line Length
Lines should be no longer than 80 characters, even if it requires violating
indentation rules to make them fit. Since C11 is assumed, the best way to
deal with strings that extend past column 80 is to break them into two or
more sections separated from each other by a newline and indentation:
puts("This string got very far to the "
"right and wrapped. ANSI "
"catenation rules will turn this "
"into one long string.");
The rule for string formatting can be violated in cases where breaking
the string prevents ability to lookup the string using grep. Also please
bear in mind that if you are too deeply nested, the code needs refactoring
and not more line breaks.
#### Comments
Comments should be used whenever they improve the readability or
comprehensibility of the code. Comments describing public functions are
usually in the header file below the function prototype; comments
describing static functions are above the function declaration.
Comments may be single-line or multi-line. A single-line comment should be
at the end of the line if there is other text on the line, and should start
in the same column as other nearby end-of-line comments. The comment
should be at the same indentation level as the code it is referring to.
Multi-line comments should start with `"/*"` on a line by itself.
Subsequent lines should have `" *"` lined-up with the `"*"` above. The end of
the comment should be `" */"` on a line by itself, again with the `"*"`
lined-up with the one above. Comments should start with a capital letter
and end with a period.
Good:
/*
* Private variables.
*/
static int a /* Description of 'a'. */
static int b /* Description of 'b'. */
static char * c /* Description of 'c'. */
The following macros should be used where appropriate:
FALLTHROUGH;
UNREACHABLE();
#### Header files
.h files should not rely on other files having been included. .h files
should prevent multiple inclusion. The OS is assumed to prevent multiple
inclusion of its .h files.
The `#pragma once` directive should be used instead of `#ifdef/#define`
combo, and the `#include <config.h>` should not be used anywhere, the
build system ensures that it's the first included file.
A header file defining a public interface is generally placed in the source
tree two levels below the C file that implements the interface. For
example, the include file defining the interface for `lib/dns/zone.c` is in
`lib/dns/include/dns/zone.h`. (The second "dns" in the path enables the file
to be included via `"#include <dns/zone.h>"`.)
Public header files should include interface documentation in Doxygen
format.
Private header files, describing interfaces that are for internal use
within a library but not for public use, are kept in the source tree at the
same level as their related C files, and often have `"_p"` in their names,
Refactor taskmgr to run on top of netmgr This commit changes the taskmgr to run the individual tasks on the netmgr internal workers. While an effort has been put into keeping the taskmgr interface intact, couple of changes have been made: * The taskmgr has no concept of universal privileged mode - rather the tasks are either privileged or unprivileged (normal). The privileged tasks are run as a first thing when the netmgr is unpaused. There are now four different queues in in the netmgr: 1. priority queue - netievent on the priority queue are run even when the taskmgr enter exclusive mode and netmgr is paused. This is needed to properly start listening on the interfaces, free resources and resume. 2. privileged task queue - only privileged tasks are queued here and this is the first queue that gets processed when network manager is unpaused using isc_nm_resume(). All netmgr workers need to clean the privileged task queue before they all proceed normal operation. Both task queues are processed when the workers are finished. 3. task queue - only (traditional) task are scheduled here and this queue along with privileged task queues are process when the netmgr workers are finishing. This is needed to process the task shutdown events. 4. normal queue - this is the queue with netmgr events, e.g. reading, sending, callbacks and pretty much everything is processed here. * The isc_taskmgr_create() now requires initialized netmgr (isc_nm_t) object. * The isc_nm_destroy() function now waits for indefinite time, but it will print out the active objects when in tracing mode (-DNETMGR_TRACE=1 and -DNETMGR_TRACE_VERBOSE=1), the netmgr has been made a little bit more asynchronous and it might take longer time to shutdown all the active networking connections. * Previously, the isc_nm_stoplistening() was a synchronous operation. This has been changed and the isc_nm_stoplistening() just schedules the child sockets to stop listening and exits. This was needed to prevent a deadlock as the the (traditional) tasks are now executed on the netmgr threads. * The socket selection logic in isc__nm_udp_send() was flawed, but fortunatelly, it was broken, so we never hit the problem where we created uvreq_t on a socket from nmhandle_t, but then a different socket could be picked up and then we were trying to run the send callback on a socket that had different threadid than currently running.
2021-04-09 11:31:19 +02:00
e.g. `lib/isc/mem_p.h`.
Header files that define modules should have a structure like the
following. Note that `<isc/lang.h>` MUST be included by any public header
file using the ISC_LANG_BEGINDECLS and ISC_LANG_ENDDECLS macros, so the
correct name-mangling happens for function declarations when C++ programs
include the file. `<isc/lang.h>` SHOULD be included for private header files
or for public files that do not declare any functions.
/*
* Copyright (C) 2016 Internet Systems Consortium, Inc. ("ISC")
*
* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, you can obtain one at https://mozilla.org/MPL/2.0/.
*/
#pragma once
/*****
***** Module Info
*****/
/*
* (Module name here.)
*
* (One line description here.)
*
* (Extended description and notes here.)
*
* MP:
* (Information about multiprocessing considerations
* here, e.g. locking requirements.)
*
* Reliability:
* (Any reliability concerns should be mentioned here.)
*
* Resources:
* (A rough guide to how resources are used by this module.)
*
* Security:
* (Any security issues are discussed here.)
*
* Standards:
* (Any standards relevant to the module are listed here.)
*/
/***
*** Imports
***/
/* #includes here. */
#include <isc/lang.h>
/***
*** Types
***/
/* (Type definitions here.) */
/***
*** Functions
***/
ISC_LANG_BEGINDECLS
/* (Function declarations here, with full prototypes.) */
ISC_LANG_ENDDECLS
#### Including Interfaces (.h files)
The first file to be included in a C source file must be config.h. The
config.h file must never be included by any public header file (that is,
any header file that will be installed by `"make install"`).
Try to include only necessary files, not everything under the sun.
Operating-system-specific files should not be included by most modules; if
they are needed, they should be used with `#ifdef` and controlled by
`configure`.
#### Statements
There should be at most one statement per line. The comma operator should
not be used to form compound statements.
Bad:
if (i > 0) {
printf("yes\\n"); i = 0; j = 0;
x = 4, y *= 2;
}
#### Functions
The use of ANSI C function prototypes is required.
The return type of the function should be listed on a line by itself when
specifying the implementation of the function. The opening curly brace
should occur on the same line as the argument list, unless the argument
list is more than one line long:
static void
func1(int i) {
/* whatever */
}
int
func2(int first_argument, int next_argument,
int last_argument)
{
/* whatever */
}
To suppress compiler warnings, unused function arguments must be
declared within the function via the `UNUSED()` macro.
In the function body, local variable declarations must be at the beginning
of the function, followed by any `REQUIRE()` statements, then `UNUSED()`
declarations, then all other code, in that order. These sections should be
separated by blank lines.
#### Curly Braces
Curly Braces do not get their own indentation.
An opening brace does not start a new line. The statements enclosed by the
braces should not be on the same line as the opening or closing brace. A
closing brace should be the only thing on the line, unless it's part of an
else clause.
If a controlling statement (e.g., `if`, `while`, or `for`) or a function
header at the start of a block of code is all on one line, then the opening
brace should at the end of that line. If the controlling statement occupies
multiple lines, then the opening brace should be on the next line by itself.
Historically, when a controlling statement such as `if` or `else` had
only a single action associated with it, then BIND style specified that
no bracing was to used around that action. This has been revised: in
newly added code, braces are now preferred around all control statement
code blocks. Note that legacy code has not yet been updated to adhere to
this.
Good:
static void
f(int i) {
if (i > 0) {
printf("yes\\n");
i = 0;
} else {
printf("no\\n");
}
}
Bad:
void f(int i)
{
if(i<0){i=0;printf("was negative\\n");}
if (i == 0)
printf("no\\n");
if (i > 0)
{
printf("yes\\n");
i = 0;
}}
#### Spaces
* DO put a space between operators like `=`, `+`, `==`, etc.
* DO put a space after `,`.
* DO put a space after `;` in a `for` statement.
* DO put spaces after C reserved words such as `if`, `for`, `while`, and `do`.
* DO put a space after `return`, and parenthesize the return value.
* Do NOT put a space between a variable or function name and `(` or `[`.
* Do NOT put a space after the `sizeof` operator name, and DO parenthesize its argument: `malloc(4 * sizeof(long))`.
* Do NOT put a space immediately after a `(` or immediately before a `)`, unless it improves readability. The same goes for `[` and `]`.
* Do NOT put a space before `++` or `--` when used in post-increment/ decrement mode, or after them when used in pre-increment/decrement mode.
* Do NOT put a space before `;` when terminating a statement or in a `for` statement.
* Do NOT put a space after `*` when used to dereference a pointer, or on either side of `->`.
* Do NOT put a space after `~`.
* The `|` operator may either have a space on both sides or it may have no spaces, depending on readability. Either way, if the `|` operator is used more than once in a statement, then the spacing must be consistent.
#### Return Values
If a function returns a value, it should be cast to `(void)` if you don't
care what the value is, except for `printf` and its variants, `fputc`,
`fwrite` (when writing text), `fflush`, `memmove`, `memset`, `strcpy`,
`strncpy`, and `strcat`.
Certain functions will return values or not depending on the operating
system or even compiler flags; these include `openlog` and `srandom`. The
return value of these should not be used nor cast to `(void)`.
All error conditions must be handled.
Mixing of error status and valid results within a single type should be
avoided.
Good:
os_result_t result;
os_descriptor_t s;
result = os_socket_create(AF_INET, SOCK_STREAM, 0, &s);
if (result != OS_R_SUCCESS) {
/* Do something about the error. */
return;
}
Not so good:
int s;
/*
* Obviously using interfaces like socket() (below) is allowed
* since otherwise you couldn't call operating system routines; the
* point is not to write more interfaces like them.
*/
s = socket(AF_INET, SOCK_STREAM, 0);
if (s < 0) {
/* Do something about the error using errno. */
return;
}
#### Integral Types
Careful thought should be given to whether an integral type should be signed or
unsigned, and to whether a specific size is required. The basic rule of thumb
is to use `size_t` for sizes, cardinalities or ordinal numbers (e.g. iteration
counters, array subscripts). Use unsigned type for small quantities that cant
be negative, use signed types for small quantities that bear a sign, and finally
use ptrdiff_t for large differences that bear a sign. Assignments and
comparisons between signed and unsigned integers should be avoided; suppressing
the warnings with casts is not desirable.
C99 standard integer types are generally preferred, and must be used when
`unsigned long` or `short` could be ambiguous, and `size_t` is preferred to
`unsigned int` variables.
#### Clear Success or Failure
A function should report success or failure, and do so accurately. It
should never fail silently. Use of [design by contract](dev.md#dbc)
can help here.
When a function is designed to return results to the caller by assigning
to caller variables through pointer arguments, it should perform the
assignment only if it succeeds, and leave the variables unmodified if it
fails. A `REQUIRE()` statement should be used to ensure that the pointer
is in a sane state when the function is called.
The `isc_result_t` is provided for use by result codes. See the
[results](dev.md#results) section of the [developer
information](dev.md) page for more details.
#### Testing Bits
Bit testing should be as follows:
Good:
/* Test if flag set. */
if ((flags & FOO) != 0) {
}
/* Test if flag clear. */
if ((flags & BAR) == 0) {
}
/* Test if both flags set. */
if ((flags & (FOO|BAR)) == (FOO|BAR)) {
}
Bad:
/* Test if flag set. */
if (flags & FOO) {
}
/* Test if flag clear. */
if (! (flags & BAR)) {
}
#### Testing for Zero or Non-zero
Explicit testing against zero is required for numeric, non-boolean
variables.
Good:
int i = 10;
/* ... */
if (i != 0) {
/* Do something. */
}
Bad:
int i = 10;
/* ... */
if (i) {
/* Do something. */
}
#### Null Pointer
The null pointer value should be referred to as `NULL`, not `0`.
Testing to see whether a pointer is `NULL` should be an explicit
comparison; do not treat a pointer variable as if it were a boolean.
Good:
char *c = NULL;
/* ... */
if (c != NULL) {
/* Do something. */
}
Bad:
char *c = NULL;
/* ... */
if (c) {
/* Do something. */
}
#### The Ternary Operator
The `?:` operator should mostly be avoided. It is tolerated when deciding
what value to pass as a parameter to a function, such as frequently happens
with printf, and also when a simple (non-compound) value is being used in
assignment or as part of a calculation.
If a statement containing a ternary operator spills over more than one
line, put the `?` and `:` at the beginning of the following lines with two
additional spaces of indent.
Using the ternary operator to specify a return value is very rarely
permissible, and never when returning result codes.
Good:
printf("%c is%s a number.\\n", c, isdigit(c) ? "" : " NOT");
l = (l1 < l2) ? l1 : l2;
s = (a_very_long_variable < an_even_longer_variable)
? "true"
: "false";
if (gp.length + (go < 16384 ? 2 : 3) >= name->length) {
/* whatever */
}
Okay:
return ((length1 < length2) ? -1 : 1);
Bad:
return (success ? ISC_R_SUCCESS : ISC_R_FAILURE);
#### Assignment in Parameters
Variables should not have their values assigned or changed when being
passed as parameters, except perhaps for the increment and decrement
operators.
Bad:
isc_mem_get(mctx, size = 20);
Okay:
fputc(c++, stdout);
#### Invalidating Pointers
When the data a pointer points to has been freed, or is otherwise no longer
valid, the pointer should be set to `NULL` unless the pointer is part of a
structure which is itself going to be freed immediately.
Good:
char *text;
/* text is initialized here. */
isc_mem_free(mctx, text);
text = NULL;
#### Variable Scopes
Always use minimal scopes for the variables, e.g. use block scope instead of
local scope whenever possible.
Bad:
void
foo() {
size_t i;
[...];
for (i = 0; i < 10; i++);
[...]
}
Good:
void
foo() {
[...];
for (size_t i = 0; i < 10; i++);
[...]
}
Bad:
void
foo() {
size_t j = 0;
[...] /* j not used here */
if (true) {
while (j < 10) ++j;
}
[...] /* j not used here */
return (0);
}
Good:
void
foo() {
[...]
if (true) {
size_t j = 0;
while (j < 10) ++j;
}
[...]
}
Integrating cppcheck with editor of your choice (f.e. flycheck with emacs) could
be a great help in identifying places where variable scopes can be reduced.
#### Initializing variables
Static initializers should be used instead of memset.
Good:
char array[10] = { 0 };
Bad:
char array[10];
memset(array, 0, sizeof(array));
Designated initializers should be used to initialize structures.
Good:
struct example {
int foo;
int bar;
int baz;
};
struct example x = { .foo = -1 };
Bad:
struct example {
int foo;
int bar;
int baz;
};
struct example x;
x.foo = -1;
x.bar = 0;
x.baz = 0;
Good:
struct example {
int foo;
int bar;
int baz;
};
struct example *x = isc_mem_get(mctx, sizeof(*x));
*x = (struct example){ .foo = -1 };
Bad:
struct example {
int foo;
int bar;
int baz;
};
struct example *x = isc_mem_get(mctx, sizeof(*x));
x->foo = -1;
x->bar = 0;
x->baz = 0;
#### Const
Declare variables as constant if they are not to be modified.
#### Variable-Length Arrays
Avoid using C99 variable length arrays From an attacker's point of view, a VLA declaration is essentially a primitive for performing arbitrary arithmetic on the stack pointer. If the attacker can control the size of a VLA they have a very powerful tool for causing memory corruption. To mitigate this kind of attack, and the more general class of stack clash vulnerabilities, C compilers insert extra code when allocating a VLA to probe the growing stack one page at a time. If these probes hit the stack guard page, the program will crash. From the point of view of a C programmer, there are a few things to consider about VLAs: * If it is important to handle allocation failures in a controlled manner, don't use VLAs. You can use VLAs if it is OK for unreasonable inputs to cause an uncontrolled crash. * If the VLA is known to be smaller than some known fixed size, use a fixed size array and a run-time check to ensure it is large enough. This will be more efficient than the compiler's stack probes that need to cope with arbitrary-size VLAs. * If the VLA might be large, allocate it on the heap. The heap allocator can allocate multiple pages in one shot, whereas the stack clash probes work one page at a time. Most of the existing uses of VLAs in BIND are in test code where they are benign, but there was one instance in `named`, in the GSS-TSIG verification code, which has now been removed. This commit adjusts the style guide and the C compiler flags to allow VLAs in test code but not elsewhere.
2022-03-18 14:50:36 +00:00
VLAs are unsafe when it is important to handle allocation failure in a
controlled manner rather than an uncontrolled crash. They are safer if the
array size is checked first, but then you lose a lot of their simplicity
and readability.
VLAs should not be used in most code in BIND. VLAs are OK in test code
where the lack of safety doesn't matter. The default compiler flags enforce
this rule.
#### <a name="public_namespace"></a>Public Interface Namespace
All public interfaces to functions, macros, typedefs, and variables
provided by the library, should use names of the form
`{library}_{module}_{what}`, such as:
isc_buffer_t /* typedef */
dns_name_setbuffer(name, buffer) /* function */
ISC_LIST_HEAD(list) /* macro */
isc_commandline_argument /* variable */
Structures which are `typedef`'d generally have the name of the typedef
sans the final `_t`:
typedef struct dns_rbtnode dns_rbtnode_t;
struct dns_rbtnode {
/* ... members ... */
}
In some cases, structures are specific to a single C file and are
opaque outside that file. In these cases, the `typedef` occurs in the
associated header file, but the structure definition in the C file
itself. Examples of this include the zone object `dns_zone_t`;
the structure is only accessible via get/set functions in
`lib/dns/zone.c`. Other times, structure members can be accessed
from outside the C file where they are implemented; examples include
`dns_view_t`. Which way to implement a particular object is up to
the developer's discretion.
Generally speaking, macros are defined with all capital letters, but this
is not universally consistent (eg, numerous `isc_buffer_{foo}` macros).
The `{module}` and `{what}` segments of the name do not have underscores
separating natural word elements, as demonstrated in
`isc_commandline_argument` and `dns_name_setbuffer` above. The `{module}`
part is usually the same as the basename of the source file, but sometimes
other `{module}` interfaces appear within one file, such as `dns_label_*`
interfaces in `lib/dns/name.c`. However, in the public libraries the file
name must be the same as some module interface provided by the file; e.g.,
`dns_rbt_*` interfaces would not be declared in a file named redblack.c (in
lieu of any other `dns_redblack_*` interfaces in the file).
The one notable exception to this naming rule is the interfaces provided by
`<isc/util.h>`. There's a large caveat associated with the public
description of this file that it is hazardous to use because it pollutes
the general namespace.
When the signature of a public function needs to change, the old function
name should be retained for backward compatibility, if at all possible.
For example, when `dns_zone_setfile()` needed to include a file format
parameter, it was changed to `dns_zone_setfile2()`; the original function
name became a wrapper for the new function, calling it with the default
value of the format parameter:
isc_result_t
dns_zone_setfile(dns_zone_t *zone, const char *file) {
return (dns_zone_setfile2(zone, file, dns_masterformat_text);
}
isc_result_t
dns_zone_setfile2(dns_zone_t *zone, const char *file,
dns_masterformat_t format)
{
...
}
#### <a name="private_namespace"></a>Shared Private Interfaces
When a module provides an interface for internal use by other modules in
the library or by unit tests, it should use the same naming convention
described for the public interfaces, except `{library}` and `{module}` are
separated by a double-underscore. This indicates that the name is
internal, its API is not as formal as the public API, and thus it might
change without any sort of notice. Examples of this usage include
`dns__zone_loadpending()` and `isc__mem_printallactive()`.
In some cases, a public interface is instantiated by a private back-end
implementation. The private interface implementations are typically
static functions that are pointed to by "method" tables. For example,
the `dns_db` interface is implemented in several places, including
`lib/dns/rbtdb.c` (the red-black tree database used for internal storage of
zones and cache data) and `lib/dns/sdlz.c` (an interface to DLZ modules).
An object of type `dns_dbmethods_t` is created for each of these,
containing function pointers to the local implementations of each of the
`dns_db` API functions. The `dns_db_findnode()` function is provided by
static functions called `findnode()` in each file, and so on.
#### Initialization
When an object is allocated from the heap, all fields in the object must be
initialized.
#### Dead Code Pruning
Source which becomes obsolete should be removed, not just disabled with
`#if 0 ... #endif`.
#### Portability
When using a C library function, consider whether all operating systems
support it. Is it in the POSIX standard? If so, how long has it been
there? Is its behavior the same on all platforms? Is its signature
the same? Are integer parameters the same size and signedness? Does it
always return the same values on success, and set the same `errno` codes
on failure?
If there is a chance the library call may not be completely portable,
edit `configure.in` to check for it on the local system and only call
it from within a suitable `#ifdef`. If the function is nonoptional,
it may be necessary to add your own implementation of it (or copy one
from a source with a BSD-compatible license).
BIND provides portable internal versions of many common library calls.
Some are designed to ensure that library calls have standardized
[ISC result codes](dev.md#results) instead of using potentially
nonwportable `errno` values; these include the file operations
in `isc_file` and `isc_stdio`. Others, such as `isc_tm_strptime()`,
are needed to ensure consistent cross-platform behavior.
Others simply provide needed functions on platforms that don't
have them: for example, `isc_string_strlcpy()` is an implementation
of the BSD-specific `strlcpy()` function. On Linux and systems
without a `strlcpy()` function, it is `#define`d to `isc_string_strlcpy()`
#### Some notes on standard functions
* Always use `memmove()` rather than `memcpy()`.
* If using `snprintf()` in a source file, be sure it includes `<isc/print.h>`
#### Log messages
Error and warning messages should be logged through the [logging
system](dev.md#logging). Debugging `printf`s may be used during
development, but must be removed when the debugging is finished.
Log messages do not start with a capital letter, nor do they end in a
period, and they are not followed by newlines.
When variable text such as a file name or domain name occurs as part of a
log message, it should be enclosed in single quotes, as in "zone '%s' is
lame".
When the variable text forms a separate phrase, such as when it separated
from the rest of the message by a colon, it can be left unquoted:
isc_log_write(... "open: %s: %s", filename, isc_result_totext(result));
File names (`__FILE__`), line numbers (`__LINE__`), function names,
memory addresses, and other references to program internals may be used
in debugging messages and in messages to report programming errors detected
at runtime. They may not be used in messages that indicate errors in the
program's inputs or operation.
### <a name="pystyle"></a>Python
Python is NOT required for building, installing, or using the BIND 9
name server. However, BIND 9 may use it for its system test
environment, and in some cases for generating source or documentation
files which are then committed to to the git repository.
For Python coding, we enforce a common codestyle using the tool
[black](https://black.readthedocs.io/en/stable/the_black_code_style/current_style.html)
There are also a few other requirements:
* The `__init__()` method should always be the first one declared in a
class definition, like so:
class Foo:
# constructor definition here
def __init__(self):
...
# other functions may follow
def bar(self):
...
Close all file and socket objects
* All Python standard library objects that have an underlying file
descriptor (fd) should be closed explicitly using the `.close()` method.
* In cases where a file is opened and closed in a single block, it
is often preferable to use the `with` statement:
with open('filename') as f:
do_something_with(f)
### <a name="plstyle"></a>Perl
Like Python, Perl is NOT required for building, installing, or using
the BIND 9 name server.
Perl 5 is assumed; Perl scripts do not need to work in Perl 4.
Perl source code should follow the conventions for C source code where
applicable.
### <a name="shstyle"></a>Bourne Shell
Shell scripts must be as portable as possible and should therefore conform
strictly to POSIX standards. Shell extensions such as those introduced in
Bash should be avoided. Some pitfalls to avoid:
* To capture the output of a command, use `` `backquotes` `` rather than
`$(parentheses)`
* For arithmetical computation, use `` `expr {expression}` ``, not
`$((expression))`
2018-09-11 10:59:09 -07:00
* To test string length use `` `expr $string : ".*"` `` rather than ``
`expr length $string` ``
* To test for the presence of a string in a file without printing anything
to stdout, use `"grep string filename > /dev/null 2>&1"`, rather than
`"grep -q string filename"`.
* To test for file existence use `"test -f"` rather than `"test -e"`
* Don't use newline (`\\n`) when calling `echo`. Either use another `echo`
statement, or use `"cat << EOF"`.
* To set a variable from outside awk, use `"awk '{...}' var=value"` rather
than `"awk -vvar=value '{...}'"`
* Don't close stdout/stderr descriptors (`>&-`), but redirect them to /dev/null
instead (`>/dev/null`) as the closed descriptor might get reused leading to
unpredictable behaviour when using `fprintf(stderr, ...)`