root / README.hacking @ master
History | View | Annotate | Download (9.1 kB)
| 1 | # -*- Outline -*- |
|---|---|
| 2 | # |
| 3 | # Copyright 2004,2007,2008,2009 Free Software Foundation, Inc. |
| 4 | # |
| 5 | # This file is part of GNU Radio |
| 6 | # |
| 7 | # GNU Radio is free software; you can redistribute it and/or modify |
| 8 | # it under the terms of the GNU General Public License as published by |
| 9 | # the Free Software Foundation; either version 3, or (at your option) |
| 10 | # any later version. |
| 11 | # |
| 12 | # GNU Radio is distributed in the hope that it will be useful, |
| 13 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 14 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 15 | # GNU General Public License for more details. |
| 16 | # |
| 17 | # You should have received a copy of the GNU General Public License |
| 18 | # along with GNU Radio; see the file COPYING. If not, write to |
| 19 | # the Free Software Foundation, Inc., 51 Franklin Street, |
| 20 | # Boston, MA 02110-1301, USA. |
| 21 | # |
| 22 | |
| 23 | Random notes on coding conventions, some explanations about why things |
| 24 | aren't done differently, etc, etc, |
| 25 | |
| 26 | |
| 27 | * Boost 1.35 |
| 28 | |
| 29 | Until boost 1.35 or later is common in distributions, you'll need to |
| 30 | build boost from source yourself. See README.building-boost. |
| 31 | |
| 32 | |
| 33 | * C++ and Python |
| 34 | |
| 35 | GNU Radio is now a hybrid system. Some parts of the system are built |
| 36 | in C++ and some of it in Python. In general, prefer Python to C++. |
| 37 | Signal processing primitives are still built in C++ for performance. |
| 38 | |
| 39 | |
| 40 | * C++ namespaces |
| 41 | |
| 42 | In the cleanup process, I considered putting everything in the |
| 43 | gnuradio namespace and dropping the Gr|gr prefix. In fact, I think |
| 44 | it's probably the right idea, but when I tested it out, I ran into |
| 45 | problems with SWIG's handling of namespaces. Bottom line, SWIG |
| 46 | (1.3.21) got confused and generated bad code when I started playing |
| 47 | around with namespaces in a not particularly convoluted way. I saw |
| 48 | problems using the boost::shared_ptr template in combination with |
| 49 | classes defined in the gnuradio namespace. It wasn't pretty... |
| 50 | |
| 51 | |
| 52 | * Naming conventions |
| 53 | |
| 54 | Death to CamelCaseNames! We've returned to a kinder, gentler era. |
| 55 | We're now using the "STL style" naming convention with a couple of |
| 56 | modifications since we're not using namespaces. |
| 57 | |
| 58 | With the exception of macros and other constant values, all |
| 59 | identifiers shall be lower case with words_separated_like_this. |
| 60 | |
| 61 | Macros and constant values (e.g., enumerated values, |
| 62 | static const int FOO = 23) shall be in UPPER_CASE. |
| 63 | |
| 64 | |
| 65 | ** Global names |
| 66 | |
| 67 | All globally visible names (types, functions, variables, consts, etc) |
| 68 | shall begin with a "package prefix", followed by an '_'. The bulk of |
| 69 | the code in GNU Radio logically belongs to the "gr" package, hence |
| 70 | names look like gr_open_file (...). |
| 71 | |
| 72 | Large coherent bodies of code may use other package prefixes, but |
| 73 | let's try to keep them to a well thought out list. See the list |
| 74 | below. |
| 75 | |
| 76 | *** Package prefixes |
| 77 | |
| 78 | These are the current package prefixes: |
| 79 | |
| 80 | gr_ Almost everything |
| 81 | |
| 82 | gri_ Implementation primitives. Sometimes we |
| 83 | have both a gr_<foo> and a gri_<foo>. In that case, |
| 84 | gr_<foo> would be derived from gr_block and gri_<foo> |
| 85 | would be the low level guts of the function. |
| 86 | |
| 87 | atsc_ Code related to the Advanced Television |
| 88 | Standards Committee HDTV implementation |
| 89 | |
| 90 | qa_ Quality Assurance. Test code. |
| 91 | |
| 92 | |
| 93 | ** Class data members (instance variables) |
| 94 | |
| 95 | All class data members shall begin with d_<foo>. |
| 96 | |
| 97 | The big win is when you're staring at a block of code it's obvious |
| 98 | which of the things being assigned to persist outside of the block. |
| 99 | This also keeps you from having to be creative with parameter names |
| 100 | for methods and constructors. You just use the same name as the |
| 101 | instance variable, without the d_. |
| 102 | |
| 103 | class gr_wonderfulness {
|
| 104 | std::string d_name; |
| 105 | double d_wonderfulness_factor; |
| 106 | |
| 107 | public: |
| 108 | gr_wonderfulness (std::string name, double wonderfulness_factor) |
| 109 | : d_name (name), d_wonderfulness_factor (wonderfulness_factor) |
| 110 | {
|
| 111 | ... |
| 112 | } |
| 113 | ... |
| 114 | }; |
| 115 | |
| 116 | |
| 117 | ** Class static data members (class variables) |
| 118 | |
| 119 | All class static data members shall begin with s_<foo>. |
| 120 | |
| 121 | |
| 122 | ** File names |
| 123 | |
| 124 | Each significant class shall be contained in it's own file. The |
| 125 | declaration of class gr_foo shall be in gr_foo.h, the definition in |
| 126 | gr_foo.cc. |
| 127 | |
| 128 | |
| 129 | |
| 130 | * Storage management |
| 131 | |
| 132 | Strongly consider using the boost smart pointer templates, scoped_ptr |
| 133 | and shared_ptr. scoped_ptr should be used for locals that contain |
| 134 | pointers to objects that we need to delete when we exit the current |
| 135 | scope. shared_ptr implements transparent reference counting and is a |
| 136 | major win. You never have to worry about calling delete. The right |
| 137 | thing happens. |
| 138 | |
| 139 | See http://www.boost.org/libs/smart_ptr/smart_ptr.htm |
| 140 | |
| 141 | |
| 142 | * Unit tests |
| 143 | |
| 144 | Build unit tests for everything non-trivial and run them after every |
| 145 | change. Check out Extreme Programming: |
| 146 | http://c2.com/cgi/wiki?ExtremeProgrammingRoadmap |
| 147 | |
| 148 | Unit tests should also be written for all examples. This should kill |
| 149 | off the bit rot we've been plagued with. |
| 150 | |
| 151 | ** C++ unit tests |
| 152 | |
| 153 | For C++ we're using the cppunit framework. cppunit has its bad |
| 154 | smells, but it's mostly workable. http://cppunit.sf.net |
| 155 | |
| 156 | Currently each directory <dirname> contains files qa_<dirname>.{h,cc}
|
| 157 | that bring together all the qa_<foo> test suites in the directory. |
| 158 | We ought to be able to automate this without too much trouble. |
| 159 | |
| 160 | The directory gnuradio-core/src/tests contains programs that run |
| 161 | the tests. test_all runs all of the registered C++ unit tests. |
| 162 | |
| 163 | As far as I can tell, the cppunit TestFactoryRegistry maybe able to be |
| 164 | tricked into doing what we want. As is, I don't think it's enough by |
| 165 | itself, since there's nothing dragging the qa* files out of the |
| 166 | library and into the program. I haven't tested out this idea. |
| 167 | |
| 168 | ** Python unit tests |
| 169 | |
| 170 | We use the standard unittest package for unit testing of Python code. |
| 171 | |
| 172 | |
| 173 | |
| 174 | * Standard command line options |
| 175 | |
| 176 | When writing programs that are executable from the command line, |
| 177 | please follow these guidelines for command line argument names (short |
| 178 | and long) and types of the arguments. We list them below using the |
| 179 | Python optparse syntax. In general, the default value should be coded |
| 180 | into the help string using the "... [default=%default]" syntax. |
| 181 | |
| 182 | ** Mandatory options by gr_block |
| 183 | |
| 184 | *** Audio source |
| 185 | |
| 186 | Any program using an audio source shall include: |
| 187 | |
| 188 | add_option("-I", "--audio-input", type="string", default="",
|
| 189 | help="pcm input device name. E.g., hw:0,0 or /dev/dsp") |
| 190 | |
| 191 | The default must be "". This allows an audio module-dependent default |
| 192 | to be specified in the user preferences file. |
| 193 | |
| 194 | |
| 195 | *** Audio sink |
| 196 | |
| 197 | add_option("-O", "--audio-output", type="string", default="",
|
| 198 | help="pcm output device name. E.g., hw:0,0 or /dev/dsp") |
| 199 | |
| 200 | The default must be "". This allows an audio module-dependent default |
| 201 | to be specified in the user preferences file. |
| 202 | |
| 203 | |
| 204 | ** Standard options names by parameter |
| 205 | |
| 206 | Whenever you want an integer, use the "intx" type. This allows the |
| 207 | user to input decimal, hex or octal numbers. E.g., 10, 012, 0xa. |
| 208 | |
| 209 | Whenever you want a float, use the "eng_float" type. This allows the |
| 210 | user to input numbers with SI suffixes. E.g, 10000, 10k, 10M, 10m, 92.1M |
| 211 | |
| 212 | If your program allows the user to specify values for any of the |
| 213 | following parameters, please use these options to specify them: |
| 214 | |
| 215 | |
| 216 | To specify a frequency (typically an RF center frequency) use: |
| 217 | |
| 218 | add_option("-f", "--freq", type="eng_float", default=<your-default-here>,
|
| 219 | help="set frequency to FREQ [default=%default]") |
| 220 | |
| 221 | |
| 222 | To specify a decimation factor use: |
| 223 | |
| 224 | add_option("-d", "--decim", type="intx", default=<your-default-here>,
|
| 225 | help="set decimation rate to DECIM [default=%default]") |
| 226 | |
| 227 | |
| 228 | To specify an interpolation factor use: |
| 229 | |
| 230 | add_option("-i", "--interp", type="intx", default=<your-default-here>,
|
| 231 | help="set interpolation rate to INTERP [default=%default]") |
| 232 | |
| 233 | |
| 234 | To specify a gain setting use: |
| 235 | |
| 236 | add_option("-g", "--gain", type="eng_float", default=<your-default-here>,
|
| 237 | help="set gain in dB [default=%default]") |
| 238 | |
| 239 | |
| 240 | If your application specifies both a tx and an rx gain, use: |
| 241 | |
| 242 | add_option("", "--rx-gain", type="eng_float", default=<your-default-here>,
|
| 243 | help="set receive gain in dB [default=%default]") |
| 244 | |
| 245 | add_option("", "--tx-gain", type="eng_float", default=<your-default-here>,
|
| 246 | help="set transmit gain in dB [default=%default]") |
| 247 | |
| 248 | |
| 249 | To specify the number of channels of something use: |
| 250 | |
| 251 | add_option("-n", "--nchannels", type="intx", default=1,
|
| 252 | help="specify number of channels [default=%default]") |
| 253 | |
| 254 | |
| 255 | To specify an output filename use: |
| 256 | |
| 257 | add_option("-o", "--output-filename", type="string", default=<your-default-here>,
|
| 258 | help="specify output-filename [default=%default]") |
| 259 | |
| 260 | |
| 261 | To specify a rate use: |
| 262 | |
| 263 | add_option("-r", "--bit-rate", type="eng_float", default=<your-default-here>,
|
| 264 | help="specify bit-rate [default=%default]") |
| 265 | or |
| 266 | |
| 267 | add_option("-r", "--sample-rate", type="eng_float", default=<your-default-here>,
|
| 268 | help="specify sample-rate [default=%default]") |
| 269 | |
| 270 | |
| 271 | If your application has a verbose option, use: |
| 272 | |
| 273 | add_option('-v', '--verbose', action="store_true", default=False,
|
| 274 | help="verbose output") |
| 275 | |
| 276 | |
| 277 | If your application allows the user to specify the "fast USB" options, use: |
| 278 | |
| 279 | add_option("", "--fusb-block-size", type="intx", default=0,
|
| 280 | help="specify fast usb block size [default=%default]") |
| 281 | |
| 282 | add_option("", "--fusb-nblocks", type="intx", default=0,
|
| 283 | help="specify number of fast usb blocks [default=%default]") |