From 9ab786915087da200bda071c9fe874fb6af7b43d Mon Sep 17 00:00:00 2001
From: lhecking <lhecking>
Date: Mon, 14 Jun 1999 19:17:37 +0000
Subject: [PATCH] Updated.

---
 CodeStyle  | 60 ++++++++++++++++++++++++++----------------------------
 PATCHLEVEL |  2 +-
 TODO       |  6 ++++--
 3 files changed, 34 insertions(+), 34 deletions(-)

diff --git a/CodeStyle b/CodeStyle
index abac33e26..46d2cd11e 100644
--- a/CodeStyle
+++ b/CodeStyle
@@ -1,10 +1,12 @@
-$Id: CodeStyle,v 1.7 1999/06/10 20:05:42 lhecking Exp $
+$Id: CodeStyle,v 1.8 1999/06/14 19:17:37 lhecking Exp $
 
 The following things have be observed when writing new code for gnuplot:
 (this file is currently under construction)  Some of the following is
 just personal bug-bears, so dont take any offense if something I
 moan about is something you think is good style...
 
+These rules were originally written up by David Denholm, and I keep
+updating them as I go along. -Lars
 
 
 MISSING FUNCTIONS, AND FUNCTIONS TO BE AVOIDED
@@ -90,39 +92,33 @@ prototypes have to be enclosed in a special macro, e.g.
 
 int strcmp __PROTO((char *s, char *t)); /* note the double ()'s */
 
-Function definitions, on the other hand, must be done in
-k&r style (which ANSI compilers accept).
+Probably from gnuplot release 3.7.1 on, I will require that
+all function declarations and definitions are in ANSI style.
+I see absolutely no point at all in ignoring the benefits of
+ANSI compilers, almost ten years after this language became
+an ISO standard.
 
-int strcmp(s,t)
-char *s, *t;
+int
+strcmp(char *s, char *t)
 {
   ...
 }
 
-While compilers do not require that explicit declarations be
-given for integer arguments, we do !
-
-
+On platforms which use the included configure script, the ansi2knr
+tool in the src subdirectory is invoked transparently if the compiler
+doesn't support prototypes (Ultrix, SunOS 4.x). Other platforms may
+require explicit rules or additional makefiles for non-ANSI/ANSI
+compilation. The man page for ansi2knr is included. Note that for
+ansi2knr to work, the function return type must be on a separate line,
+or, to quote from the ansi2knr manual, "ansi2knr recognizes functions
+by seeing a non-keyword  identifier at the left margin, followed by a
+left parenthesis, with a right parenthesis as the last character on
+the line."
 
-If a function has an argument which has a default promotion (char, short,
-float), this could cause problems with some strict ANSI conforming compilers.
-In this case an alternate function header must be used, e.g.
 
-#ifdef PROTOTYPES
-char *foo(char s)
-#else
-char *foo(s)
-char s;
-#endif
-
-but since there's no benefit from using a character as parameter, may
-as well use the int ie.
-  char *foo(s)
-  int s;
-
-gcc -Wtraditional  might help with this.
-[char parameters should be avoided. int can safely be used in this case.]
 
+While compilers do not require that explicit declarations be
+given for integer arguments, we do !
 
 
 While ANSI compilers can use prototypes for implicit typecasts, k&r
@@ -135,7 +131,7 @@ code on non-ansi compilers.
 
 [we will definitely give up k&r support in the near future, but
  since existing code seems to work with k&r, we're sticking
- with it.
+ with it for the time being.
 ]
 
 
@@ -202,9 +198,11 @@ putting   set ts=4  into your ~/.exrc file.
 
   Unfortunately, gnu indent does not seem to recognise this
 as a layout style. If it did, I'd have run all sources
-through it long ago. [GNU indent -kr -cp0 -l100 -lps comes
-pretty close, though. I need very little manual editing after
-running this, mostly for long (> 80 cols) lines. -Lars]
+through it long ago. [GNU indent -kr -cp0 -l132 -lps -br -psl
+is what I use. Very little manual editing is necessary after
+running this, mostly for long (> 80 cols) lines. Anyway,
+I don't care which indentation style _you_ use, I'm running
+all code through indent and be done with it. -Lars]
 
 
 
@@ -242,7 +240,7 @@ Unary operators should be writen as usual, ie i++.
 
 
 No C++ style comments (//).
-No comments on #ifdef/#ifndef lines.
+No trailing comments on #ifdef/#ifndef lines.
 No #error and #warning directives.
 
 [more to come]
diff --git a/PATCHLEVEL b/PATCHLEVEL
index aec258df7..b63ba696b 100644
--- a/PATCHLEVEL
+++ b/PATCHLEVEL
@@ -1 +1 @@
-0.8
+0.9
diff --git a/TODO b/TODO
index 360033799..d5f81bf2d 100644
--- a/TODO
+++ b/TODO
@@ -27,7 +27,8 @@ Lars' list
 - much-requested feature: read text/commands from data files
 - provide library of replacement functions instead of stdfn.c
 - possibly switch to automake
-  [PARTIALLY DONE] not done yet: docs lisp src tutorial
+  [PARTIALLY DONE] not done yet: docs lisp
+  [docs is next; won't do lisp]
 - completely rewrite the parser. Use yacc/bison if possible.
 - maybe rewrite the scanner in lex? The benefits are not so obvious,
   because the current scanner is extremely simple. This is probably
@@ -69,8 +70,9 @@ Lars' list
 - better lisp/ files integration (Makefile target to import Bruce's
   latest version). How to handle duplicates?
 - better documentation format; get rid of the doc2xxx utils
-- move src/* subdirs to toplevel?
+  [SGML. SGML. SGML]
 - update the FAQ!!!!
+  [Being worked on bei Juergen]
 - non-integer time value in 'pause' command
 - tilde and variable expansion to be done by 
   gp_fopen() - to be implemented