kumquat-buildroot/docs/manual/writing-rules.adoc

232 lines
6.2 KiB
Plaintext
Raw Normal View History

// -*- mode:doc; -*-
// vim: set syntax=asciidoc:
manual: use one-line titles instead of two-line titles (trivial) Asciidoc supports two syntaxes for section titles: two-line titles (title plus underline consisting of a particular symbol), and one-line titles (title prefixed with a specific number of = signs). The two-line title underlines are: Level 0 (top level): ====================== Level 1: ---------------------- Level 2: ~~~~~~~~~~~~~~~~~~~~~~ Level 3: ^^^^^^^^^^^^^^^^^^^^^^ Level 4 (bottom level): ++++++++++++++++++++++ and the one-line title prefixes: = Document Title (level 0) = == Section title (level 1) == === Section title (level 2) === ==== Section title (level 3) ==== ===== Section title (level 4) ===== The buildroot manual is currenly using the two-line titles, but this has multiple disadvantages: - asciidoc also uses some of the underline symbols for other purposes (like preformatted code, example blocks, ...), which makes it difficult to do mass replacements, such as a planned follow-up patch that needs to move all sections one level down. - it is difficult to remember which level a given underline symbol (=-~^+) corresponds to, while counting = signs is easy. This patch changes all two-level titles to one-level titles in the manual. The bulk of the change was done with the following Python script, except for the level 1 titles (-----) as these underlines are also used for literal code blocks. This patch only changes the titles, no other changes. In adding-packages-directory.txt, I did add missing newlines between some titles and their content. ---------------------------------------------------------------------------- #!/usr/bin/env python import sys import mmap import re for input in sys.argv[1:]: f = open(input, 'r+') f.flush() s = mmap.mmap(f.fileno(), 0) # Level 0 (top level): ====================== = # Level 1: ---------------------- == # Level 2: ~~~~~~~~~~~~~~~~~~~~~~ === # Level 3: ^^^^^^^^^^^^^^^^^^^^^^ ==== # Level 4 (bottom level): ++++++++++++++++++++++ ===== def replace_title(s, symbol, replacement): pattern = re.compile(r'(.+\n)\%s{2,}\n' % symbol, re.MULTILINE) return pattern.sub(r'%s \1' % replacement, s) new = s new = replace_title(new, '=', '=') new = replace_title(new, '+', '=====') new = replace_title(new, '^', '====') new = replace_title(new, '~', '===') #new = replace_title(new, '-', '==') s.seek(0) s.write(new) s.resize(s.tell()) s.close() f.close() ---------------------------------------------------------------------------- Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
2014-05-02 07:47:30 +02:00
== Coding style
Overall, these coding style rules are here to help you to add new files in
Buildroot or refactor existing ones.
If you slightly modify some existing file, the important thing is
to keep the consistency of the whole file, so you can:
* either follow the potentially deprecated coding style used in this
file,
* or entirely rework it in order to make it comply with these rules.
[[writing-rules-config-in]]
manual: use one-line titles instead of two-line titles (trivial) Asciidoc supports two syntaxes for section titles: two-line titles (title plus underline consisting of a particular symbol), and one-line titles (title prefixed with a specific number of = signs). The two-line title underlines are: Level 0 (top level): ====================== Level 1: ---------------------- Level 2: ~~~~~~~~~~~~~~~~~~~~~~ Level 3: ^^^^^^^^^^^^^^^^^^^^^^ Level 4 (bottom level): ++++++++++++++++++++++ and the one-line title prefixes: = Document Title (level 0) = == Section title (level 1) == === Section title (level 2) === ==== Section title (level 3) ==== ===== Section title (level 4) ===== The buildroot manual is currenly using the two-line titles, but this has multiple disadvantages: - asciidoc also uses some of the underline symbols for other purposes (like preformatted code, example blocks, ...), which makes it difficult to do mass replacements, such as a planned follow-up patch that needs to move all sections one level down. - it is difficult to remember which level a given underline symbol (=-~^+) corresponds to, while counting = signs is easy. This patch changes all two-level titles to one-level titles in the manual. The bulk of the change was done with the following Python script, except for the level 1 titles (-----) as these underlines are also used for literal code blocks. This patch only changes the titles, no other changes. In adding-packages-directory.txt, I did add missing newlines between some titles and their content. ---------------------------------------------------------------------------- #!/usr/bin/env python import sys import mmap import re for input in sys.argv[1:]: f = open(input, 'r+') f.flush() s = mmap.mmap(f.fileno(), 0) # Level 0 (top level): ====================== = # Level 1: ---------------------- == # Level 2: ~~~~~~~~~~~~~~~~~~~~~~ === # Level 3: ^^^^^^^^^^^^^^^^^^^^^^ ==== # Level 4 (bottom level): ++++++++++++++++++++++ ===== def replace_title(s, symbol, replacement): pattern = re.compile(r'(.+\n)\%s{2,}\n' % symbol, re.MULTILINE) return pattern.sub(r'%s \1' % replacement, s) new = s new = replace_title(new, '=', '=') new = replace_title(new, '+', '=====') new = replace_title(new, '^', '====') new = replace_title(new, '~', '===') #new = replace_title(new, '-', '==') s.seek(0) s.write(new) s.resize(s.tell()) s.close() f.close() ---------------------------------------------------------------------------- Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
2014-05-02 07:47:30 +02:00
=== +Config.in+ file
+Config.in+ files contain entries for almost anything configurable in
Buildroot.
An entry has the following pattern:
---------------------
config BR2_PACKAGE_LIBFOO
bool "libfoo"
depends on BR2_PACKAGE_LIBBAZ
select BR2_PACKAGE_LIBBAR
help
This is a comment that explains what libfoo is. The help text
should be wrapped.
http://foosoftware.org/libfoo/
---------------------
* The +bool+, +depends on+, +select+ and +help+ lines are indented
with one tab.
* The help text itself should be indented with one tab and two
spaces.
* The help text should be wrapped to fit 72 columns, where tab counts
for 8, so 62 characters in the text itself.
The +Config.in+ files are the input for the configuration tool
used in Buildroot, which is the regular _Kconfig_. For further
details about the _Kconfig_ language, refer to
http://kernel.org/doc/Documentation/kbuild/kconfig-language.txt[].
[[writing-rules-mk]]
manual: use one-line titles instead of two-line titles (trivial) Asciidoc supports two syntaxes for section titles: two-line titles (title plus underline consisting of a particular symbol), and one-line titles (title prefixed with a specific number of = signs). The two-line title underlines are: Level 0 (top level): ====================== Level 1: ---------------------- Level 2: ~~~~~~~~~~~~~~~~~~~~~~ Level 3: ^^^^^^^^^^^^^^^^^^^^^^ Level 4 (bottom level): ++++++++++++++++++++++ and the one-line title prefixes: = Document Title (level 0) = == Section title (level 1) == === Section title (level 2) === ==== Section title (level 3) ==== ===== Section title (level 4) ===== The buildroot manual is currenly using the two-line titles, but this has multiple disadvantages: - asciidoc also uses some of the underline symbols for other purposes (like preformatted code, example blocks, ...), which makes it difficult to do mass replacements, such as a planned follow-up patch that needs to move all sections one level down. - it is difficult to remember which level a given underline symbol (=-~^+) corresponds to, while counting = signs is easy. This patch changes all two-level titles to one-level titles in the manual. The bulk of the change was done with the following Python script, except for the level 1 titles (-----) as these underlines are also used for literal code blocks. This patch only changes the titles, no other changes. In adding-packages-directory.txt, I did add missing newlines between some titles and their content. ---------------------------------------------------------------------------- #!/usr/bin/env python import sys import mmap import re for input in sys.argv[1:]: f = open(input, 'r+') f.flush() s = mmap.mmap(f.fileno(), 0) # Level 0 (top level): ====================== = # Level 1: ---------------------- == # Level 2: ~~~~~~~~~~~~~~~~~~~~~~ === # Level 3: ^^^^^^^^^^^^^^^^^^^^^^ ==== # Level 4 (bottom level): ++++++++++++++++++++++ ===== def replace_title(s, symbol, replacement): pattern = re.compile(r'(.+\n)\%s{2,}\n' % symbol, re.MULTILINE) return pattern.sub(r'%s \1' % replacement, s) new = s new = replace_title(new, '=', '=') new = replace_title(new, '+', '=====') new = replace_title(new, '^', '====') new = replace_title(new, '~', '===') #new = replace_title(new, '-', '==') s.seek(0) s.write(new) s.resize(s.tell()) s.close() f.close() ---------------------------------------------------------------------------- Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
2014-05-02 07:47:30 +02:00
=== The +.mk+ file
* Header: The file starts with a header. It contains the module name,
preferably in lowercase, enclosed between separators made of 80 hashes. A
blank line is mandatory after the header:
+
---------------------
################################################################################
#
# libfoo
#
################################################################################
---------------------
+
* Assignment: use +=+ preceded and followed by one space:
+
---------------------
LIBFOO_VERSION = 1.0
LIBFOO_CONF_OPTS += --without-python-support
---------------------
+
Do not align the +=+ signs.
* Indentation: use tab only:
+
---------------------
define LIBFOO_REMOVE_DOC
$(RM) -r $(TARGET_DIR)/usr/share/libfoo/doc \
$(TARGET_DIR)/usr/share/man/man3/libfoo*
endef
---------------------
+
Note that commands inside a +define+ block should always start with a tab,
so _make_ recognizes them as commands.
* Optional dependency:
** Prefer multi-line syntax.
+
YES:
+
---------------------
ifeq ($(BR2_PACKAGE_PYTHON3),y)
LIBFOO_CONF_OPTS += --with-python-support
LIBFOO_DEPENDENCIES += python3
else
LIBFOO_CONF_OPTS += --without-python-support
endif
---------------------
+
NO:
+
---------------------
LIBFOO_CONF_OPTS += --with$(if $(BR2_PACKAGE_PYTHON3),,out)-python-support
LIBFOO_DEPENDENCIES += $(if $(BR2_PACKAGE_PYTHON3),python3,)
---------------------
** Keep configure options and dependencies close together.
* Optional hooks: keep hook definition and assignment together in one
if block.
+
YES:
+
---------------------
ifneq ($(BR2_LIBFOO_INSTALL_DATA),y)
define LIBFOO_REMOVE_DATA
$(RM) -r $(TARGET_DIR)/usr/share/libfoo/data
endef
LIBFOO_POST_INSTALL_TARGET_HOOKS += LIBFOO_REMOVE_DATA
endif
---------------------
+
NO:
+
---------------------
define LIBFOO_REMOVE_DATA
$(RM) -r $(TARGET_DIR)/usr/share/libfoo/data
endef
ifneq ($(BR2_LIBFOO_INSTALL_DATA),y)
LIBFOO_POST_INSTALL_TARGET_HOOKS += LIBFOO_REMOVE_DATA
endif
---------------------
[[writing-genimage-cfg]]
=== The +genimage.cfg+ file
+genimage.cfg+ files contain the output image layout that genimage utility
uses to create final .img file.
An example follows:
---------------------
image efi-part.vfat {
vfat {
file EFI {
image = "efi-part/EFI"
}
file Image {
image = "Image"
}
}
size = 32M
}
image sdimage.img {
hdimage {
}
partition u-boot {
image = "efi-part.vfat"
offset = 8K
}
partition root {
image = "rootfs.ext2"
size = 512M
}
}
---------------------
* Every +section+(i.e. hdimage, vfat etc.), +partition+ must be indented
with one tab.
* Every +file+ or other +subnode+ must be indented with two tabs.
* Every node(+section+, +partition+, +file+, +subnode+) must have an open
curly bracket on the same line of the node's name, while the closing one
must be on a newline and after it a newline must be added except for the
last one node. Same goes for its option, for example option +size+ +=+.
* Every +option+(i.e. +image+, +offset+, +size+) must have the +=+
assignment one space from it and one space from the value specified.
* Filename must at least begin with genimage prefix and have the .cfg
extension to be easy to recognize.
* Allowed notations for +offset+ and +size+ options are: +G+, +M+, +K+
(not +k+). If it's not possible to express a precise byte count
with notations above then use hexadecimal +0x+ prefix or, as last
chance, the byte count. In comments instead use +GB+, +MB+, +KB+
(not +kb+) in place of +G+, +M+, +K+.
* For GPT partitions, the +partition-type-uuid+ value must be +U+ for
the EFI System Partition (expanded to
+c12a7328-f81f-11d2-ba4b-00a0c93ec93b+ by _genimage_), +F+ for a FAT
partition (expanded to +ebd0a0a2-b9e5-4433-87c0-68b6b72699c7+ by
_genimage_) or +L+ for the root filesystem or other filesystems
(expanded to +0fc63daf-8483-4772-8e79-3d69d8477de4+ by
_genimage_). Even though +L+ is the default value of _genimage_, we
prefer to have it explicitly specified in our +genimage.cfg+
files. Finally, these shortcuts should be used without double
quotes, e.g +partition-type-uuid = U+. If an explicit GUID is
specified, lower-case letters should be used.
The +genimage.cfg+ files are the input for the genimage tool used in
Buildroot to generate the final image file(i.e. sdcard.img). For further
details about the _genimage_ language, refer to
https://github.com/pengutronix/genimage/blob/master/README.rst[].
manual: use one-line titles instead of two-line titles (trivial) Asciidoc supports two syntaxes for section titles: two-line titles (title plus underline consisting of a particular symbol), and one-line titles (title prefixed with a specific number of = signs). The two-line title underlines are: Level 0 (top level): ====================== Level 1: ---------------------- Level 2: ~~~~~~~~~~~~~~~~~~~~~~ Level 3: ^^^^^^^^^^^^^^^^^^^^^^ Level 4 (bottom level): ++++++++++++++++++++++ and the one-line title prefixes: = Document Title (level 0) = == Section title (level 1) == === Section title (level 2) === ==== Section title (level 3) ==== ===== Section title (level 4) ===== The buildroot manual is currenly using the two-line titles, but this has multiple disadvantages: - asciidoc also uses some of the underline symbols for other purposes (like preformatted code, example blocks, ...), which makes it difficult to do mass replacements, such as a planned follow-up patch that needs to move all sections one level down. - it is difficult to remember which level a given underline symbol (=-~^+) corresponds to, while counting = signs is easy. This patch changes all two-level titles to one-level titles in the manual. The bulk of the change was done with the following Python script, except for the level 1 titles (-----) as these underlines are also used for literal code blocks. This patch only changes the titles, no other changes. In adding-packages-directory.txt, I did add missing newlines between some titles and their content. ---------------------------------------------------------------------------- #!/usr/bin/env python import sys import mmap import re for input in sys.argv[1:]: f = open(input, 'r+') f.flush() s = mmap.mmap(f.fileno(), 0) # Level 0 (top level): ====================== = # Level 1: ---------------------- == # Level 2: ~~~~~~~~~~~~~~~~~~~~~~ === # Level 3: ^^^^^^^^^^^^^^^^^^^^^^ ==== # Level 4 (bottom level): ++++++++++++++++++++++ ===== def replace_title(s, symbol, replacement): pattern = re.compile(r'(.+\n)\%s{2,}\n' % symbol, re.MULTILINE) return pattern.sub(r'%s \1' % replacement, s) new = s new = replace_title(new, '=', '=') new = replace_title(new, '+', '=====') new = replace_title(new, '^', '====') new = replace_title(new, '~', '===') #new = replace_title(new, '-', '==') s.seek(0) s.write(new) s.resize(s.tell()) s.close() f.close() ---------------------------------------------------------------------------- Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com> Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
2014-05-02 07:47:30 +02:00
=== The documentation
The documentation uses the
https://asciidoc-py.github.io/[asciidoc] format.
For further details about the asciidoc syntax, refer to
https://asciidoc-py.github.io/userguide.html[].
=== Support scripts
Some scripts in the +support/+ and +utils/+ directories are written in
Python and should follow the
https://www.python.org/dev/peps/pep-0008/[PEP8 Style Guide for Python Code].