* socket.c (scm_fill_sockaddr): zero the address structure before
[bpt/guile.git] / HACKING
1 Here are some guidelines for members of the Guile developers team.
2
3 Contributing Your Changes ============================================
4
5 - If you have put together a change that meets the coding standards
6 described below, we encourage you to submit it to Guile. The best
7 place to post it is guile@sourceware.cygnus.com. Please don't send it
8 directly to me; I often don't have time to look things over. If you
9 have tested your change, then you don't need to be shy.
10
11 - Please submit patches using either context or unified diffs (diff -c
12 or diff -u). Don't include a patch for ChangeLog; such patches don't
13 apply cleanly, since we've probably changed the top of ChangeLog too.
14 Instead, provide the unaltered text at the top of your patch.
15
16 Please don't include patches for generated files like configure,
17 aclocal.m4, or any Makefile.in. Such patches are often large, and
18 we're just going to regenerate those files anyway.
19
20
21 CVS conventions ======================================================
22
23 - We use CVS to manage the Guile sources. The repository lives on
24 egcs.cygnus.com, in /cvs/guile; you will need an
25 account on that machine to access the repository. Also, for security
26 reasons, egcs presently only supports CVS connections via the SSH
27 protocol, so you must first install the SSH client. Then, you should
28 set your CVS_RSH environment variable to ssh, and use the following as
29 your CVS root:
30
31 :ext:USER@egcs.cygnus.com:/cvs/guile
32
33 Either set your CVSROOT environment variable to that, or give it as
34 the value of the global -d option to CVS when you check out a working
35 directory.
36
37 For more information on SSH, see http://www.cs.hut.fi/ssh.
38
39 The Guile sources live in several modules:
40
41 - guile-core --- the interpreter, QuickThreads, and ice-9
42 - guile-doc --- documentation in progress. When complete, this will
43 be incorporated into guile-core.
44 - guile-oops --- The Guile Object-Oriented Programming System (talk to mdj)
45 - guile-tcltk --- the Guile/Tk interface
46 - guile-tk --- the new Guile/Tk interface, based on STk's modified Tk
47 - guile-rgx-ctax --- the Guile/Rx interface, and the ctax implementation
48 - guile-scsh --- the port of SCSH to guile, talk to Gary Houston
49 - guile-www --- A Guile module for making HTTP requests.
50
51 There is a mailing list for CVS commit messages; see README for details.
52
53 - We check Makefile.in and configure files into CVS, as well as the
54 files they are built from (Makefile.am, configure.in); we do not check
55 in Makefiles or header files generated by configuration scripts. The
56 general rule is that you should be able to check out a working
57 directory of Guile from CVS, and then type "configure" and "make",
58 without running any other tools.
59
60 - (Automake 1.4 only) Be sure to run automake at the top of the tree
61 with no arguments. Do not use `automake Makefile' to regenerate
62 specific Makefile.in files, and do not trust the Makefile rules to
63 rebuild them when they are out of date. Automake 1.4 will add
64 extraneous rules to the top-level Makefile if you specify specific
65 Makefiles to rebuild on the command line. Running the command
66 `autoreconf --force' should take care of everything correctly.
67
68 - Make sure your changes compile and work, at least on your own
69 machine, before checking them into the main branch of the Guile
70 repository. If you really need to check in untested changes, make a
71 branch.
72
73 - Include each log entry in both the ChangeLog and in the CVS logs.
74 If you're using Emacs, the pcl-cvs interface to CVS has features to
75 make this easier; it checks the ChangeLog, and generates good default
76 CVS log entries from that.
77
78
79 Coding standards =====================================================
80
81 - As for any part of Project GNU, changes to Guile should follow the
82 GNU coding standards. The standards are available via anonymous FTP
83 from prep.ai.mit.edu, as /pub/gnu/standards/standards.texi and
84 make-stds.texi.
85
86 - The Guile tree should compile without warnings under the following
87 GCC switches, which are the default in the current configure script:
88 -O2 -Wall -Wpointer-arith -Wmissing-prototypes
89 The only warnings which can be tolerated are those about variables
90 being clobbered by longjmp/vfork in eval.c. The variables in question
91 are critical to the interpreter's performance; as far as I can tell,
92 it is difficult/annoying to avoid these warnings without slowing the
93 system down substantially. (If you can figure out a good fix, I'd be happy to see it.)
94
95 Note that the warnings generated vary from one version of GCC to the
96 next, and from one architecture to the next (apparently). To provide
97 a concrete common standard, Guile should compile without warnings from
98 GCC 2.7.2.3 in a Red Hat 5.2 i386 Linux machine. Furthermore, each
99 developer should pursue any additional warnings noted by on their
100 compiler. This means that people using more stringent compilers will
101 have more work to do, and assures that everyone won't switch to the
102 most lenient compiler they can find. :)
103
104 Note also that EGCS (as of November 3 1998) doesn't handle the
105 `noreturn' attribute properly, so it doesn't understand that functions
106 like scm_error won't return. This may lead to some silly warnings
107 about uninitialized variables. You should look into these warnings to
108 make sure they are indeed spurious, but you needn't correct warnings
109 caused by this EGCS bug.
110
111 - If you add code which uses functions or other features that are not
112 entirely portable, please make sure the rest of Guile will still
113 function properly on systems where they are missing. This usually
114 entails adding a test to configure.in, and then adding #ifdefs to your
115 code to disable it if the system's features are missing.
116
117 - When you make a user-visible change (i.e. one that should be
118 documented, and appear in NEWS, put an asterisk in column zero of the
119 start of the ChangeLog entry, like so:
120
121 Sat Aug 3 01:27:14 1996 Gary Houston <ghouston@actrix.gen.nz>
122
123 * * fports.c (scm_open_file): don't return #f, throw error.
124
125 When you've written a NEWS entry and updated the documentation, go
126 ahead and remove the asterisk. I will use the asterisks to find and
127 document changes that haven't been dealt with before a release.
128
129 - Please write log entries for functions written in C under the
130 functions' C names, and write log entries for functions written in
131 Scheme under the functions' Scheme names. Please don't do this:
132
133 * procs.c, procs.h (procedure-documentation): Moved from eval.c.
134
135 Entries like this make it harder to search the ChangeLogs, because you
136 can never tell which name the entry will refer to. Instead, write this:
137
138 * procs.c, procs.h (scm_procedure_documentation): Moved from eval.c.
139
140 Changes like adding this line are special:
141
142 SCM_PROC (s_serial_map, "serial-map", 2, 0, 1, scm_map);
143
144 Since the change here is about the name itself --- we're adding a new
145 alias for scm_map that guarantees the order in which we process list
146 elements, but we're not changing scm_map at all --- it's appropriate
147 to use the Scheme name in the log entry.
148
149 - There's no need to keep a change log for documentation files. This
150 is because documentation is not susceptible to bugs that are hard to
151 fix. Documentation does not consist of parts that must interact in a
152 precisely engineered fashion; to correct an error, you need not know
153 the history of the erroneous passage. (This is copied from the GNU
154 coding standards.)
155
156 - Make sure you have papers from people before integrating their
157 changes or contributions. This is very frustrating, but very
158 important to do right. From maintain.texi, "Information for
159 Maintainers of GNU Software":
160
161 When incorporating changes from other people, make sure to follow the
162 correct procedures. Doing this ensures that the FSF has the legal
163 right to distribute and defend GNU software.
164
165 For the sake of registering the copyright on later versions ofthe
166 software you need to keep track of each person who makes significant
167 changes. A change of ten lines or so, or a few such changes, in a
168 large program is not significant.
169
170 *Before* incorporating significant changes, make sure that the person
171 has signed copyright papers, and that the Free Software Foundation has
172 received them.
173
174 If you receive contributions you want to use from someone, let me know
175 and I'll take care of the administrivia. Put the contributions aside
176 until we have the necessary papers.
177
178 - When you make substantial changes to a file, add the current year to
179 the list of years in the copyright notice at the top of the file.
180
181
182 Helpful hints ========================================================
183
184 - [From Mikael Djurfeldt] When working on the Guile internals, it is
185 quite often practical to implement a scheme-level procedure which
186 helps you examine the feature you're working on.
187
188 Examples of such procedures are: pt-size, debug-hand and
189 current-pstate.
190
191 I've now put #ifdef GUILE_DEBUG around all such procedures, so that
192 they are not compiled into the "normal" Guile library. Please do the
193 same when you add new procedures/C functions for debugging purpose.
194
195 You can define the GUILE_DEBUG flag by passing --enable-guile-debug to
196 the configure script.
197
198 - You'll see uses of the macro SCM_P scattered throughout the code;
199 those are vestiges of a time when Guile was meant to compile on
200 pre-ANSI compilers. Guile now requires ANSI C, so when you write new
201 functions, feel free to use ANSI declarations, and please provide
202 prototypes for everything. You don't need to use SCM_P in new code.
203
204
205 Jim Blandy