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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263
|
Developer Information
###############################################################################
This file provides some information for developers looking to submit patches to
OLA. The first section contains some general notes, and the latter
sections address particular languages.
Also take a look at README which includes some relevant information for
developers too.
General
===============================================================================
Code Reviews
------------
To have patches reviewed please push your changes to GitHub and create a Pull
Request from within your fork.
If you're using Google Code please enable comments on the repo. Under
Administer -> Source check:
Enable code reviews
Allow non-members to review code
Licensing
---------
Think carefully about the license for each file. Code to be used by clients
(./ola , ./python) should be under the LGPL, so that it may be used in
commercial projects.
Plugins and code used solely in olad should be under the GPL.
scripts/enforce_licence.py will look for a LICENCE file in each directory, and
ensure that all source files in the directory ( & subdirectories) start with
the LICENCE. You can pass --fix to automatically add the licence.
Unittests
---------
Unittests are *highly* encouraged. At the very least please make sure any
changes don't break the tests.
The tests are performed with `make check` in the root ola directory. The same
command can be run within a sub directory to only run a particular set of
tests (although you may experience issues with this method, running from the
root ola directory is guaranteed to work).
Branches, Versioning & Releases
-------------------------------
Version numbers take the form MAJOR.MINOR.PATCH based on http://semver.org/.
The release lifecycle is:
- New feature work occurs on the master branch.
- Once the new features are considered stable or enough time has passed, a new
minor release branch will be created, e.g. 0.10.
- The minor release branch will be stablized with bugfixes, these bug fixes
will also be merged back into master.
- Once declared stable, a new patch branch 0 will be created e.g. 0.10.0
- Release specific changes like the version number, debian files etc. will be
pushed to the patch branch.
- The release will occur.
- Changes on the release branch will be merged back into the minor version
branch.
- Bug fixes will continue on the minor version branch, leading to another patch
release.
What this means for you as a developer is that depending on the scope of your
change, we may request that you merge it into the minor release branch rather
than master. This allows us to get your change out to end uses faster than if
it was merged into master directly.
C++
===============================================================================
The bulk of OLA code is written in C++. This includes the daemon olad, and
plugins.
https://wiki.openlighting.org/index.php/OLA_developer_info describes many of
the underlying classes used.
C++ 11
------
As much as we'd like to start using C++11, Ubuntu 12.04 still comes with g++
4.6 which has partial support. Once 12.04 is end-of-lifed, we'll consider
using C++ 11 constructs.
Endian Issues
-------------
Be aware that OLA runs on big endian systems. When dealing with wire formats
always use the functions in include/ola/network/NetworkUtils.h to convert or
use the BigEndianOutputStream to manage this automatically for you.
Structure Packing
-----------------
Structure packing syntax differs between platforms and compilers. If you declare
a struct or class that needs to be packed, use the PACK() macro from
ola/base/Macro.h . See below for an example of PACK() usage.
Not that this macro doesn't work for enums. See plugins/espnet/EspNetPackets.h
for an example of how to pack an enum.
Non x86 platforms
-----------------
OLA also runs on more than just x86. Some platforms like ARM can't de-reference
pointers which aren't aligned.
For example:
PACK(struct f {
uint8_t value1;
uint16_t value2;
});
struct f foo = {1, 2};
uint16_t *ptr = &foo.value2;
// Bug! Will not be true on all platforms.
if (*ptr == 2)
http://www.aleph1.co.uk/chapter-10-arm-structured-alignment-faq has a good
explanation.
Style / Coding Standards
------------------------
We use the Google C++ Style Guide:
https://google.github.io/styleguide/cppguide.html
We provide formatting configuration files for some IDEs at
https://github.com/OpenLightingProject/editor-configs . If you use an IDE
which isn't listed there please consider submitting the formatting
configuration files.
Please run the cpplint.py script on all files before requesting a review,
alternatively it will be run automatically by Travis CI on pull requests.
cpplint.py can be downloaded here:
https://github.com/google/styleguide/tree/gh-pages/cpplint
Run it with:
--filter=-legal/copyright,-readability/streams,-runtime/arrays
Doxygen
-------
The code is documented using Doxygen. There is an automatically generated
version available from:
https://docs.openlighting.org/ola/doc/latest/
If you want to build it yourself, install Doxygen, run ./configure in the ola
root directory, then run make doxygen-doc. The files will be generated into
html/
Race Conditions
---------------
If possible, code should be tested with slower machines (embedded platforms,
virtual machines etc.). This has discovered race conditions in the past.
Valgrind
--------
All code must be tested with valgrind to ensure it doesn't leak memory.
Coverage (gcov)
---------------
To enable the coverage tools, you need lcov and gcov installed. To enable run
./configure --enable-gcov and then build as normal.
To generate the HTML pages run:
mkdir /tmp/lcov
tmpdir=`mktemp -d /tmp/lcov.XXXXXX`
coverage="${tmpdir}/coverage.info"
lcov --capture --directory ./ --output-file $coverage
genhtml $coverage --output-directory /tmp/lcov
Java
===============================================================================
An experimental Java API is provided.
Style / Coding Standards
------------------------
Please follow the Sun Java style guide.
Javascript
===============================================================================
Javascript is used for the olad web UI. Instructions for building the
javascript can be found in javascript/README.
Closure Compiler
----------------
The closure compiler catches many errors. The javascript code should build
cleanly.
Style / Coding Standards
------------------------
We use the Google Javascript style guide:
https://google.github.io/styleguide/javascriptguide.xml
A javascript linter can be downloaded from:
https://developers.google.com/closure/utilities/docs/linter_howto?csw=1
Please make sure all Javascript code is lint clean.
Python
===============================================================================
Python is used for tools like the RDM responder tests and the device model
collector. To enable these a OLA Python API is provided.
Style / Coding Standards
------------------------
We use the Google Python style guide:
https://google.github.io/styleguide/pyguide.html
However we deviate from the standard and use 2 space indents, so it's
consistent with the C++ code.
It's linted and static analysis is performed using flake8, you can either run
it yourself or await the Travis CI results of your pull request.
flake8 can be installed following the instructions here:
https://gitlab.com/pycqa/flake8/
Run it with:
--max-line-length 80 --exclude *_pb2.py,.git,__pycache \
--ignore E111,E121,E127,E129 data/rdm include/ola python scripts \
tools/ola_mon tools/rdm
Build System
===============================================================================
Autotools is a complex beast. Even after reading two books and a using it for a
decade I still feel like I hardly understand it.
Useful tips
-----------
* Run `make distcheck` to ensure you haven't broken anything.
* Use $(srcdir), $(top_srcdir) to reference files where required. This is to
support VPATH builds:
http://www.gnu.org/software/automake/manual/html_node/VPATH-Builds.html
* Prefer manual dependencies over BUILT_SOURCES where possible, see
http://www.gnu.org/software/automake/manual/html_node/Built-Sources-Example.html
However this doesn't work in some cases (libraries?) because it overrides the
automake generated rules.
|