Minor cleanup before release.

README

@@ -1,39 +1,36 @@
 README                                          -*-text-*-
 
-This is a line-editing library.  It can be linked into almost any program to
+This is a line editing library.  It can be linked into almost any program to provide
-provide command-line editing and history.
+command-line editing and history.
 
-It is call-compatible with the FSF readline library, but it is a fraction of
+It is call-compatible with the FSF readline library, but it is a fraction of the size
-the size (and offers fewer features).  It does not use standard I/O.  It is
+(and offers fewer features).  It does not use standard I/O.  It is distributed under
-distributed under a "C News-like" copyright, see the file LICENSE for
+a "C News-like" copyright, see the file LICENSE for details.
-details.
 
-Configuration is made by supplying different options to the GNU configure
+Configuration is made by supplying different options to the GNU configure script.  In
-script.  In the examples/ directory you can find a couple of small & slow
+the examples/ directory you can find a couple of small & slow shell implementations
-shell implementations used for testing.
+used for testing.
 
-Before finding out about the Debian version I was on the lookout for a
+Before finding out about the Debian version I was on the lookout for a really small
-really small replacement for the GNU readline package. Not only was it large
+replacement for the GNU readline package. Not only was it large and GPL:ed (instead
-and GPL:ed (instead of LGPL:ed), it also depended on libncurses, so the
+of LGPL:ed), it also depends on libncurses, so the resulting size was a bit too much
-resulting size was a bit too much for my embedded system.  I eventually
+for my embedded system.  I eventually stubmled upon the BSD libedit library, which
-stubmled upon the BSD libedit library, which was sufficient for a while,
+was sufficient for a while, even though it too depends on libncurses.  I searched my
-even though it too depended upon libncurses.  I searched my soul and went
+soul and went back to where I, back in 1996, started out -- Minix.  And there it was,
-back to where I, back in 1996, started out -- Minix.  And there it was, a
+a really small readline replacement!
-really small readline replacement!
 
-In 2000 Jim Studt packaged libeditline for Debian[1], the exact origin of
+In 2000 Jim Studt packaged libeditline for Debian[1], the exact origin of the Debian
-the Debian code base is unclear, see the Sid package[2] for details. There
+code base is unclear, see the Sid package[2] for details. There were some notable
-were some notable differences between that version and the upstream Minix
+differences between that version and the upstream Minix sources, all of which have
-sources, all of which have now been merged here.
+now been merged here.
 
-An explanation of the version numbering may be in order.  I didn't know
+An explanation of the version numbering may be in order.  I didn't know about the
-about the Debian version for quite some time, so I kept a different name for
+Debian version for quite some time, so I kept a different name for the package and a
-the package and a different versioning scheme. Now, in June 2009, I decided
+different versioning scheme.  In June 2009, I decided to line up alongside Debian,
-to line up alongside Debian, with the intent of merging the efforts.  Sorry
+with the intent of merging the efforts.  Sorry for any confusion this might cause.
-for any confusion this might cause.
 
 Enjoy,
-	Joachim Nilsson <joachim.nilsson@vmlinux.org>
+	Joachim Nilsson <troglobit()vmlinux!org>
 
 [1] - http://lists.debian.org/debian-devel/2000/05/msg00548.html
 [2] - http://packages.debian.org/sid/libeditline0

TODO

@@ -1,5 +1,6 @@
 TODO
 
-* Add --enable-FEATURE for features defined in Makefile-minix.in
+* Add --enable-FEATURE for remaining features defined in Makefile-minix.in
+* Add support for rl_bind_key(), currently one needs to "hack" the Map[]
 * Add support for inhibiting completion: rl_inhibit_completion
 

doc/README

@@ -22,7 +22,7 @@ a small slow shell for testing.
 An earlier version was distributed with Byron's rc.  Principal
 changes over that version include:
 	Faster.
-	Is eight-bit clean (thanks to brendan@cs.widener.edu)
+	Is eight-bit clean (thanks to brendan()cs!widener!edu)
 	Written in K&R C, but ANSI compliant (gcc all warnings)
 	Propagates EOF properly; rc trip test now passes
 	Doesn't need or use or provide memmove.
@@ -44,7 +44,7 @@ There is one known bug:
 
 Enjoy,
 	Rich $alz
-	<rsalz@osf.org>
+	<rsalz()osf!org>
 
  Copyright 1992,1993 Simmule Turner and Rich Salz.  All rights reserved.
 

src/editline.c

@@ -20,8 +20,6 @@
 #define META(x)         ((x) | 0x80)
 #define ISMETA(x)       ((x) & 0x80)
 #define UNMETA(x)       ((x) & 0x7F)
-#define MAPSIZE         33
-#define METAMAPSIZE     17
 #if     !defined(HIST_SIZE)
 #define HIST_SIZE       20
 #endif  /* !defined(HIST_SIZE) */
@@ -86,8 +84,8 @@ static int        Point;
 static int        PushBack;
 static int        Pushed;
 static int        Signal;
-static el_keymap_t Map[MAPSIZE];
+static el_keymap_t Map[];
-static el_keymap_t MetaMap[METAMAPSIZE];
+static el_keymap_t MetaMap[];
 static SIZE_T     Length;
 static SIZE_T     ScreenCount;
 static SIZE_T     ScreenSize;
@@ -1401,7 +1399,7 @@ static el_status_t last_argument(void)
     return s;
 }
 
-static el_keymap_t Map[33] = {
+static el_keymap_t Map[] = {
     {   CTL('@'),       mk_set          },
     {   CTL('A'),       beg_line        },
     {   CTL('B'),       bk_char         },
@@ -1435,7 +1433,7 @@ static el_keymap_t Map[33] = {
     {   0,              NULL            }
 };
 
-static el_keymap_t   MetaMap[17]= {
+static el_keymap_t   MetaMap[]= {
     {   CTL('H'),       wipe            },
     {   DEL,            wipe            },
     {   ' ',            mk_set          },