style(9) manpage __P removal

ibotty bsd at ibotty.net
Tue Sep 23 07:00:23 PDT 2003


> :Speaking of style how do people feel about K&R declarations?
> :I'm wondering if I'm the only one that thinks that retaining
> :them is silly.
> :
> :-Kip
> 
>     I have been replacing them with their ANSI equivalents as I have come
>     across them.

i have update my patch (ehm, well...), to also reflect this cleanup.

matt, you will find some of the formulations familiar ;-).


>     int
>     fubar(int a, int b, int c, int d, int e, int f, int g, int h, int i,
> int j, int k, int l, int m)
>     { 
>     }
> 
>     But its up to you.  I have no strong opinion either way (other then
>     additional lines being indented somehow).

i have not included a guideline then.

~ibottyIndex: style.9
===================================================================
RCS file: /home/src/dcvs/src/share/man/man9/style.9,v
retrieving revision 1.3
diff -u -3 -p -r1.3 style.9
--- style.9	14 Aug 2003 18:28:48 -0000	1.3
+++ style.9	23 Sep 2003 13:54:57 -0000
@@ -252,17 +252,21 @@ Functions that are used locally in more 
 separate header file, e.g.\&
 .Qq Pa extern.h .
 .Pp
-Only use the
+Never use the
 .Dv __P
 macro from the include file
 .Aq Pa sys/cdefs.h
-if the source
-file in general is (to be) compilable with a K&R Old Testament compiler.
+in new source
+files, even if it is (to be) compilable with a K&R Old Testament compiler.
 Use of the
 .Dv __P
-macro in new code is discouraged, although modifications
-to existing files should be consistent with that file's conventions.
+macro in legacy code is also highly discouraged, even minor changes 
+to existing code should remove these macros.
 .Pp
+You should not use K&R declarations either. When replacing them with their 
+ANSI equivalent, beware of arguments, that are out of order.
+.Pp
+Nevertheless, modifications to existing files should be consistent with that file's conventions.
 In general code can be considered
 .Dq "new code"
 when it makes up about 50% or more of the file(s) involved.




More information about the Submit mailing list