3 (This is not a most actively maintained file, but recommended
9 Preferred patch format is output from "git format-patch", but unified
10 diffs and links to public git repos are acceptable as well.
12 Please include test-cases in your patch if possible.
14 Please format your submission for ease of reading: unless the change
15 is whitespace only, avoid re-indenting code you are not touching, etc.
16 Unless your patch is large and best understood as a series of
17 sequential changes, please send it in as single patch file.
19 If your patch includes algorithmic changes, explain them. If your
20 patch uses a published algorithm, please include a link to the paper.
21 We aren't always as well-educated as we'd like to...
23 Patches can be either mailed to sbcl-devel, or submitted via
24 Launchpad. If you submit a patch on Launchpad, please add the tag
25 "review" to the associated bug.
27 If you have any questions, feel free to ask them on sbcl-devel.
32 See also PRINCIPLES and TLA files.
34 Most of the style hints in the Lisp FAQ apply.
36 When porting code we prefer code which factors dependencies into a set
37 of interface functions and constants and includes implementations of
38 the interface for the different systems. Patches which require
39 conditional compilation (like all the old
40 #T+HPUX or #T-X86 tests in the sources inherited from CMUCL) might be
41 accepted if they're simple, in hopes of factoring out the differences
42 more cleanly later, but even if accepted, such code may not be
43 maintained for very long.
45 grammatical fussiness:
46 Phrases are not capitalized.
47 Sentences are capitalized.
48 Periods terminate sentences.
49 Periods separate phrases from succeeding sentences, e.g.
50 ;;; the maximum number of transformations we'll make before
51 ;;; concluding we're in an infinite loop and bailing. This can
52 ;;; be changed, but it is an error to change it while we're
54 (defvar *max-n-transformations* 10)
55 Lisp in comments is capitalized.
58 Function documentation can be a description of what the function
60 ;;; Parse the arguments for a BDEFSTRUCT call, and return
61 ;;; (VALUES NAME DEFSTRUCT-ARGS MAKE-LOAD-FORM-FUN BDEFSTRUCT-STYPE),
62 ;;; where NAME is the name of the new type, DEFSTRUCT-ARGS is the
63 ;;; munged result suitable for passing on to DEFSTRUCT,
64 ;;; MAKE-LOAD-FORM-FUN is the make load form function, or NIL if
65 ;;; there's none, and BDEFSTRUCT-SUPERTYPE is the direct supertype
66 ;;; of the type if it is another BDEFSTRUCT-defined type, or NIL
68 (defun parse-bdefstruct-args (nameoid &rest rest)
70 or a remark about the function, e.g.
71 ;;; a helper function for BDEFSTRUCT in the #+XC-HOST case
72 (defun uncross-defstruct-args (defstruct-args)
74 If you're talking about what the function does, ordinarily you
75 should just say what the function does, e.g.
76 ;;; Return the first prime number greater than or equal to X.
77 (defun primify (x) ..)
78 instead of telling the reader that you're going to tell him what
79 the function does, e.g.
80 ;;; PRIMIFY returns the first prime number greater than or
82 (defun primify (x) ..)
84 ;;; When you call this function on X, you get back the first
85 ;;; prime number greater than or equal to X.
86 (defun primify (x) ..)
88 In general, if you can express it in the code instead of the comments,
89 do so. E.g. the old CMUCL code has many comments above functions foo
92 If we were going to do something like that, we would prefer to do it by
95 (Instead, for various other reasons, we centralize all the exports
96 in package declarations.) The old "FOO -- interface" comments are bad
97 style because they duplicate information (and they illustrate one
98 of the evils of duplicating information by the way that they have
99 drifted out of sync with the code).
101 There are a number of style practices on display in the code
102 which are not good examples to follow:
103 * using conditional compilation to support different architectures,
104 instead of factoring the dependencies into interfaces and providing
105 implementations of the interface for different architectures;
106 * in conditional compilation, using a common subexpression over and
107 over again, e.g. #+(OR GENGC GENCGC), when the important thing is
108 that GENGC and GENCGC are (currently) the GCs which support scavenger
109 hooks. If you have to do that, define a SCAVHOOK feature,
110 write #+SCAVHOOK in many places, and arrange for the SCAVHOOK feature
111 to be set once and only once in terms of GENGC and GENCGC. (That way
112 future maintainers won't curse you.)
113 * putting the defined symbol, and information about whether it's
114 exported or not, into the comments around the definition of the symbol;
115 * naming anything DO-FOO if it isn't an iteration macro
116 * exposing a lot of high-level functionality not in the ANSI standard
117 to the user (as discussed above)
118 * not using a consistent abbreviation style in global names (e.g.
119 naming some things DEFINE-FOO and other things DEF-BAR, with
120 no rule to determine whether the abbreviation is used)
121 * using lots of single-colon package prefixes (distracting and hard
122 to read, and obstacles to reaching package nirvana where
123 package dependencies are a directed acyclic graph) or even
124 double-colon package prefixes (hard to understand and hard
125 to maintain). (One exception: I've sometimes been tempted to
126 add a CL: prefix to the definition of every CL symbol (e.g.
127 (DEFUN CL:CADDDR (..) ..) as reminders that they're required by
128 ANSI and can't be deleted no matter how obscure and useless some
129 of them might look.:-)
130 Most of these are common in the code inherited from CMUCL. We've
131 eliminated them in some places, but there's a *lot* of code inherited