[src/trunk]: src/bin/sh PR bin/48489 -- Shell "simple commands" are now not a...

[kre <kre%NetBSD.org@localhost>]

details:   https://anonhg.NetBSD.org/src/rev/588c28a1f5ef
branches:  trunk
changeset: 345136:588c28a1f5ef
user:      kre <kre%NetBSD.org@localhost>
date:      Mon May 09 20:36:07 2016 +0000

description:
PR bin/48489 -- Shell "simple commands" are now not allowed to be
completely empty, they must have something (var assignment, redirect,
or command, or multiple of those) to avoid a syntax error.  This
matches the requirements of the grammar in the standard.   Correct the
parser (using FreeBSD's sh as a guide) and update the man page to
remove text that noted a couple of places this was previously OK.

OK christos@

diffstat:

 bin/sh/parser.c |  32 ++++++++++++++++++++++----------
 bin/sh/sh.1     |  15 ++-------------
 2 files changed, 24 insertions(+), 23 deletions(-)

diffs (112 lines):

diff -r 494432d0ebca -r 588c28a1f5ef bin/sh/parser.c
--- a/bin/sh/parser.c   Mon May 09 20:28:08 2016 +0000
+++ b/bin/sh/parser.c   Mon May 09 20:36:07 2016 +0000
@@ -1,4 +1,4 @@
-/*     $NetBSD: parser.c,v 1.118 2016/05/03 03:16:55 kre Exp $ */
+/*     $NetBSD: parser.c,v 1.119 2016/05/09 20:36:07 kre Exp $ */
 
 /*-
  * Copyright (c) 1991, 1993
@@ -37,7 +37,7 @@
 #if 0
 static char sccsid[] = "@(#)parser.c   8.7 (Berkeley) 5/16/95";
 #else
-__RCSID("$NetBSD: parser.c,v 1.118 2016/05/03 03:16:55 kre Exp $");
+__RCSID("$NetBSD: parser.c,v 1.119 2016/05/09 20:36:07 kre Exp $");
 #endif
 #endif /* not lint */
 
@@ -498,20 +498,32 @@
                        synexpect(TEND, 0);
                checkkwd = 1;
                break;
-       /* Handle an empty command like other simple commands.  */
+
        case TSEMI:
+       case TAND:
+       case TOR:
+       case TPIPE:
+       case TNL:
+       case TEOF:
+       case TRP:
                /*
-                * An empty command before a ; doesn't make much sense, and
-                * should certainly be disallowed in the case of `if ;'.
+                * simple commands must have something in them,
+                * either a word (which at this point includes a=b)
+                * or a redirection.  If we reached the end of the
+                * command (which one of these tokens indicates)
+                * when we are just starting, and have not had a
+                * redirect, then ...
+                *
+                * nb: it is still possible to end up with empty
+                * simple commands, if the "command" is a var
+                * expansion that produces nothing
+                *      X= ; $X && $X
+                * -->          &&
+                * I am not sure if this is intended to be legal or not.
                 */
                if (!redir)
                        synexpect(-1, 0);
-       case TAND:
-       case TOR:
-       case TNL:
-       case TEOF:
        case TWORD:
-       case TRP:
                tokpushback++;
                n1 = simplecmd(rpp, redir);
                goto checkneg;
diff -r 494432d0ebca -r 588c28a1f5ef bin/sh/sh.1
--- a/bin/sh/sh.1       Mon May 09 20:28:08 2016 +0000
+++ b/bin/sh/sh.1       Mon May 09 20:36:07 2016 +0000
@@ -1,4 +1,4 @@
-.\"    $NetBSD: sh.1,v 1.121 2016/04/04 13:05:56 wiz Exp $
+.\"    $NetBSD: sh.1,v 1.122 2016/05/09 20:36:07 kre Exp $
 .\" Copyright (c) 1991, 1993
 .\"    The Regents of the University of California.  All rights reserved.
 .\"
@@ -31,7 +31,7 @@
 .\"
 .\"    @(#)sh.1        8.6 (Berkeley) 5/4/95
 .\"
-.Dd March 31, 2016
+.Dd May 9, 2016
 .Dt SH 1
 .ds flags abCEeFfhnuvxIimpqV
 .Os
@@ -330,11 +330,9 @@
 flag has no effect.
 .It Fl h Em trackall
 Bind commands in functions to file system paths when the function is defined.
-.\" They can seek me here (that dreaded filesystem)
 When off,
 the file system is searched for commands each time the function is invoked.
 (Not implemented.)
-.\" then can seek me there (the filesystem, once again)
 .It Fl p Em nopriv
 Do not attempt to reset effective uid if it does not match uid.
 This is not set by default to help avoid incorrect usage by setuid
@@ -589,7 +587,6 @@
 Otherwise, if the command name doesn't match a function or built-in, the
 command is searched for as a normal program in the file system (as
 described in the next section).
-.\" But the damned elusive filesystem shall never be defeated!
 When a normal program is executed, the shell runs the program,
 passing the arguments and the environment to the program.
 If the program is not a normal executable file (i.e., if it does
@@ -757,14 +754,6 @@
 .Dq baz
 and nothing else.
 This is not the way it works in C.
-Also, if you forget the left-hand side (for example when continuing lines but
-forgetting to use a backslash) it defaults to a true statement.
-This behavior is not useful and should not be relied upon.
-Similarly, if input to the
-.Nm
-reaches end of file immediately after one of these operators,
-it is assumed that a true statement follows.
-This behavior is also not useful and should not be relied upon.
 .Ss Flow-Control Constructs -- if, while, for, case
 The syntax of the if command is
 .Bd -literal -offset indent