forked from GPSBabel/gpsbabel
-
Notifications
You must be signed in to change notification settings - Fork 0
/
README.contrib
125 lines (101 loc) · 5.55 KB
/
README.contrib
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
If you're interested in contributing to this program, here are some
guidelines. Mail patches to [email protected] for
consideration and integration.
Rules to Live By
----------------
Standards are good. ISO C++, Qt, and POSIX are greatly preferred to
adding more libraries or implementing your own hash maps or other common
functions.
You may find format_skeleton.c and filter_skeleton.c in the source tree
to be helpful examples. Just add meat!
Prefer Qt objects/classes to ISO C/POSIX services.
* QStrings are reference counted, implicitly shared, and have a
robust supporting library.
* QDateTime supports sub-second time and a range of dates far
beyond 1970->2038 and are much more pleasant to work with
than ctime/mktime/struct tm.
* QXmlStreamReader and Writer eliminate the need to explictly quote data.
"But I see strcpy, sprintf, and struct tm and such in the code!" It's
true; GPSBabel is a tenured project of well over ten years. We have
code that predates our move to C++/Qt that isn't well tested or has a
low payoff to modernize and uses old constructs. Our actively
maintained/strategic formats like GPX and KML tend to be better
examples of modern programming and are generally better models to
follow. New code shouldn't be using xstrdup and gbfprintf.
Compilers complain for a reason. Code shouldn't emit warnings.
The entire world doesn't run <your OS here>. We regularly test this code on
at least five different OSes. If you find yourself wanting to insert
compiler or OS specific magic, please resist.
We try to keep a consistent convention in code that's "ours". We aren't
interested in trying to convert zlib or jeeps to it, but in general we
prefer:
int
mumble(int whatever)
{
<2spaces>char* pointers_clump_to_type;
<2spaces>if (whatevever) {
<4spaces>return blah;
<2spaces>}
}
Formatting can be done with astyle using the included style file
astylerc with the command "astyle --options=astylerc {source files}".
astyle is available at http://astyle.sourceforge.net/
The corresponding astyle settings for this are provided in our source
tree in the file named 'astylerc'
Submitting Patches
------------------
If you are creating a new target you should submit patches (use
"svn diff" to create patches) to the following files:
* Yourcode.c and/or Yourcode.h - this is the code required to do your
conversions and any support files that your code requires.
* vecs.c - an updated vecs.c file implementing your conversion code into
GPSBabel.
* Makefile.in - an updated Makefile telling the compiler how to build and link
your conversion into GPSBabel
* testo.d/$TARGET - an updated script that tests your conversion (this should
produce no output if all is good, see the current testo script for examples)
* YourOutput - a sample file of code produced by your function (used in testo
and lives in a directory called "reference").
* Documentation - see below.
Please ensure that you are building and testing against the latest code
from the top of the SVN tree and that any code you modify is the latest
version from the SVN - Note: code changes sometimes occur frequently!
Documentation
-------------
HTML and text documentation are generated automatically from DocBook
source located in the "xmldoc" directory. That directory contains
two subdirectories of interest: "formats" and "filters". If your
contribution adds or affects a format, you'll want to be in the "formats"
directory. Otherwise, you'll want to be in the "filters" directory.
You should contribute a file called "yourname.xml", where "yourname" is the
name you would give on the command-line to invoke your new format or filter.
For example, the arc filter is documented in "filters/arc.xml".
This file contains a general description of your format or filter, any
limitations in your support for it, and anything else the end user should
know. For file formats, links to manufacturers' websites are encouraged.
The contents of this file are not valid or even well-formed XML on their own;
they are included into a larger framework. If you know DocBook, you should
ensure that the contents of this file will validate if included in a <section>.
If you do not know DocBook, see the other files in this directory for examples
or see http://docbook.org/tdg/en/html/docbook.html for the gory details. Tags
of interest will almost certainly include <para> for paragraphs,
<ulink url="..."> for web links, and <screen format="linespecific"> for
example command lines.
For each option supported by your format or filter, you should also contribute
a file in the "options" subdirectory called "yourname-youroption.xml", again
using the names you would use on the command line to invoke your format or
filter with that particular option. For example, the "distance" option to the
"arc" filter is documented in "filters/options/arc-distance.xml". These
files are similar to the general description above, and should meet the same
validation requirements.
As of this writing, there are two formats that violate this rule: Magellan
serial and Microsoft Streets & Trips. Because those formats have the same
names as other formats, their descriptions are located in "magellan1.xml" and
"msroute1.xml" respectively. These are special cases, and you should do your
best to ensure that they remain the only special cases.
Note that the automated framework already includes the name and description of
your format and its options as described in vecs.c and yourcode.c, so there is
no need to repeat that information in your documentation.
Enjoy!
Robert Lipe,