From: fujieda AT jaist DOT ac DOT jp (Kazuhiro Fujieda)
Subject: Re: script execution bug
22 Nov 1998 02:33:59 -0800
Message-ID: <uiug8jb02.fsf.cygnus.cygwin32.developers@parvati.will.or.jp>
References: <uww4qtsxk DOT fsf AT parvati DOT will DOT or DOT jp>
Mime-Version: 1.0 (generated by SEMI MIME-Edit 0.100 - "Shijima")
Content-Type: text/plain; charset=US-ASCII
To: cygwin32-developers AT cygnus DOT com

>>> On Fri, 20 Nov 1998 14:20:57 -0500
>>> Christopher Faylor <cgf AT cygnus DOT com> said:

> I agree that what cygwin is doing is not POSIX but we apparently
> now have a number of people on the mailing list using this
> "feature" so I don't believe that we can remove it without
> generating a lot of confusion and dissatisfaction.

I don't think so because this feature have been enabled just
only 10 days ago, but I can accept your decision.

> Does the patch below solve the problem?
(snip)

This patch solve the null pointer dereference, but can't solve
another problem pointed out in the latter half of my last mail
cited below.

>>> On 20 Nov 1998 22:06:15 +0900
>>> Kazuhiro Fujieda <fujieda AT jaist DOT ac DOT jp> said:

> Then, when a script of which interpreter is also a script is
> executed, the interpreter of the latter script is executed with
> odd arguments.
> 
> For example,
> $ echo '#!/usr/local/bin/bar' > /usr/local/bin/foo
> $ echo '#!/sh' > /usr/local/bin/bar
> $ /usr/local/bin/foo
> /usr/local/bin/C:\usr\local\bin\bar: Can't open /usr/local/bin/C:\usr\local\bin\bar
____
  | AIST      Kazuhiro Fujieda <fujieda AT jaist DOT ac DOT jp>
  | HOKURIKU  School of Information Science
o_/ 1990      Japan Advanced Institute of Science and Technology