From: cgf AT cygnus DOT com (Christopher G. Faylor)
Subject: Re: Is bash's "test" broken in a command substition ?
23 Nov 1998 00:04:59 GMT
Message-ID: <73a8rb$p9q$1@cronkite.cygnus.com>
References: <365730B5 DOT 83912514 DOT cygnus DOT gnu-win32 AT solfin DOT com>
X-Newsreader: trn 4.0-test63 (15 March 1998)

In article <365730B5 DOT 83912514 DOT cygnus DOT gnu-win32 AT solfin DOT com>,
Gabriel Galibourg <gabriel AT solfin DOT com> wrote:
>Hi,
>
>Is bash's 'test' broken (in a command substitution!) ? can someone else try this and tell me if they have a
>problem too ... this construct used to work fine in B19, works in ksh (under Solaris) ... quid?
>
>To reproduce:
>1) type this line in a file (call it foo.sh, but any other name will do)
>echo $( test 1 == 1 && /bin/echo yes || echo no $)
>
>2) under bash type:
>. foo.sh (YES! source it!)
>
>3) you should get the following (at least I do):
>bash.exe-2.02$ . foo.sh
>[proc] C:\usr\H-i586-cygwin32\bin\bash.exe 1607 (0) sig_send: wait for sig_complete event failed, sig 20, rc
>-1, error 6
>yes
>bash.exe-2.02$
>
>... ok, it does print 'yes' as expected but the error message is not really necessary!

The "error message" is a cygwin diagnostic.  It's a bug that will be fixed
in B20.2.
-- 
cgf AT cygnus DOT com
http://www.cygnus.com/