server-1.12/doc/Developers/programming_guide

217 lines
8.5 KiB
Plaintext

This document has 3 sections - section 1 is current programming style/hints for
developers to make things easier. Section 2 is a programming guide for new
addition. Section 3 contains notes for making patches.
-------------------------------------------------------------------------------
Section 1 - currently used conventions/hints for new code writers:
1) Variable abbreviations - op is short for object pointer, ob is for object,
and pl is for player.
2) Some functions are named using the conventions above - the naming reflects
what options they take (insert_ob_in_ob takes 2 object structures)
3) Identation is either 2 spaces or 4 spaces. This can be a pain to read, but
most functions should be consistent through the function.
4) Some structure elements should never be accessed directly - rather, there
are other functions to use the values.
object->owner: This contains the owner id for this object. Use
set_owner and get_owner instead. Directly using object->owner
is likely to get unpredictable results.
object->nrof: This contains the number of an object.
Since changing this will change the weight of an object, direct
access should also be avoided. Use decrease_ob_nr, split_ob,
and insert_ob_in_... - the later will merge the objects if
applicable.
5) If using insert_ob_in_map and plan to do further actions with the object,
check and make sure the object still exists after insertion - it is
possible that the object gets destroyed while being inserted.
-------------------------------------------------------------------------------
Section 2 - Style guide for new additions:
1) Use descriptive variable names. op and pl should only be used for
temporary variables (cycling through the list or the like). For variables
well defined, use an accurate name (ie, hitter, sack, etc).
2) Only add name options with #ifdef's to the config file if the behaviour
seriously changes the game. Adding a new spell does not warrant an #ifdef.
There are already too many options in the config.h file.
3) Log errors/diagnostics with the LOG function. When doing so, please
include the function name - this is especially true for errors.
4) If you want to add special debug code for certain compiles, generate a
unique #define for it - don't use the global DEBUG. For example,
NEWCS_DEBUG.
5) Try to use the [s/u]int[8/16/32] whenever possible. Use the one of
appropriate size/type. If not sure, go for the next size up. Do not ever
write code assuming that any of those will have an exact number of bits -
those types only mean that you will get at least that many bits - you may
get more.
6) The exception to #5 above is strings. Continue to use 'char', since the
signedness of functions that take string options can differ system to
system, and generate excessive warnings if the wrong sign is used.
7) When adding new function, include a comment of what the function is
supposed to do, what options it takes, and what if any value it returns.
This makes debugging of such functions easier, and also makes it better
known to other developers if that function might be useful to them.
8) Try to keep lines to less than 80 columns when possible. This is not a
strict requirement - don't break up some complex comparison because the
line would otherwise be 83 characters long. Xterms can be resized to most
any width. However, use your judgement on whether breaking up a long line
would make something more or less readable.
9) Assume all names use one namespace. For example, if there is a struct
called spell, don't make the name of an optional parameter spell. This
will break on ANSI C compilers that follow the spec strictly (gcc does not,
even with -strict -ansi)
10) As a followup on 9 above, don't use nonstandard gcc extensions (// for
comment lines, ability to nest functions, declare arrays with variable
bounds, etc.) Likewise, don't use special system functions - don't assume
the target system will be bsd or svr4 - if using a potentially non standard
function, add checks in the autoconf script and include a version of the
function in case it is not on that system. They key word here is
portability - don't assume everyone else has the same system as you do.
11) Write code that can easily be maintained in the future, not code that is
easiest to write at that second. This basically means don't do the quick
and ugly hack, but instead fix it properly.
12) Use 4 space indentation. While a lot of old code may have 2 space indents,
a move to 4 space will improve readability in the future.
13) /*
* do block
* comment like
* this
*/
/*
and not
like this
*/
/* If you are doing a single line comment, this method is fine */
It is much easier to spot the block comments if they all start with *,
and these comments tend to be worth noticing.
14) /**
* Functions should be commented like this.
*
* @param bla
* this is a parameter
* @return
* returns NULL
*/
This lets doxygen generate nice documentation.
15) As discussed on IRC, the preferred style for expressions is like this:
if (expression) {
statement;
statement;
}
If <space> (expression), the space between the if and expression is
required.
NOT like this:
if (expression)
{
statement;
statement;
}
16) The preferred style of formal parameters:
void myFooFunction(param1, param2, param3) {
statement;
statement;
}
There is no space after the left paren, no space before the right paren,
comma right after the formal param, and space right after the comma.
17) Local variable names. Just a rules of thumb.
These are ok:
int mylongvarname;
int my_long_var_name;
Please do NOT use caps except for typedefs, enums and defines.
-------------------------------------------------------------------------------
Section 3 - Submitting Patches:
1) Please send patches on a bug fix or feature enhancement basis individually,
and not make mega patches. A diff that changes 10 things is first more
difficult for me to look over and understand as unrelated changes might be
going on. It is also harder for me to reject part of a patch (feature X is
nice, but Y doesn't work).
2) Please state in the message included with the patch what it fixes/changes.
Too often, I get patches which is just a bunch of source code, and I have
no idea if I want to incorporate it, or even if the bug is still there.
Please also state what version of crossfire the diff is for.
3) I will assume any patches mailed directly to me are to be included. If
posting a patch on the mailing list (either source or ftp location), please
explicity state whether or not you want that patch incorporated into the
master source. Many times, a patch may be made available on an
expiremental basis which is not ready for widespread distribution.
4) When making patches, please make context diffs. Please also include the
directory that the file is in (run the diff in the top level directory).
Please make 5 line context diffs - large line context diffs are fine if you
think that may make it easier.
Example:
'diff -c5 (oldfile) (newfile)'
You can also do diffs of entire directories. Do do this, type:
'diff -c5 -r (old_directory) (new_directory)'
An example:
'diff -c5 -r crossfire-0.90.1 crossfire-0.90.2'
5) Gnu diff will include files that did not exist before. Other diff programs
may not do this.
6) If your diff looks excessively long and you made a lot of formatting
changes, you can add -w to the diff options to have it ignore whitespace.
Note that this will then mean that those formatting changes will then be
lost.
7) There is no need to make a seperate diff file for each file different (ie,
treasure.diff, player.diff, etc). Assuming you follow steps 1-6, all the
diffs can be contained in one file, and patch will deal with it just fine.
8) If you need to send a map, new archetypes, or other new files where a diff
doesn't make since, a uuencoded tar file will work just fine.
9) Submit all patches to the Sourceforge patch tracker on Sourceforge:
http://sourceforge.net/projects/crossfire
After posting the patch, mail an announcement to:
Crossfire Discussion Mailing List <crossfire@metalforge.org>