Re: Assignment of $* to a var removes spaces on unset IFS.

[Robert Elz]

    Date:        Mon, 13 Aug 2018 21:36:23 -0400
    From:        Bize Ma <binaryzebra@gmail.com>
    Message-ID:  
<CAFra36gUtG8x8SoNP9fGNSp-X61EE0_yqebqHwqx9CXvhZ0S5w@mail.gmail.com>

  | You both are confused. The original table included

You misinterpreted my message.

I said two independant things ... first that the bug is already fixed.
I didn't say that the fixed version was released yet.

Second, that you should stop reporting bugs in 4.3 (which is
relatively ancient).   If you had reported it as a bug in 4.4
(which means that you had indicated that you had verified
it there) the second sentence would not have been in my
reply.

  | The report is valid for present releases.

That's true.   This is an insignificant issue really and not
even really worthy of a patch.   But it will be fixed in the next
release.


Eduardo A. Bustamante López <dualbus@gmail.com> said:
  | My point is that, in the context of bug reports, it's important that you
  | always test against the *unreleased* version of the code,

(first, sorry if your name gets messed up ... it looks correct in my
mail composition window, but the message is likely to be marked
as "charset=us-ascii" or something, and the accented character
is unlikely to survive correctly...)

I think that is going too far - if a bug is in the latest released
version, it is perfectly reasonable for people to report it.

Asking people to even know where the development version is,
let alone how to fetch it, build it, etc, is a bit too much to expect.
It is nice if that can be done, but certainly should not be a
requirement.

Sometimes (for any bug report) the submitter will just be told
"already fixed" which means that a later release will have the fix
- which is the best result that could be achieved if it was not
already fixed.

   | The fix for this bug breaks backwards compatibility, 

All bug fixes break backwards compat, but only for scripts
that incorrectly assumed that the broken behaviour was
correct - such scripts are also broken. 

I mean, what would you say to someone who said "xxx always
dumped core when I did ..., and I was relying on that to detect
the error by the presence of the core file, in the latest revision
it no longer dumps core, but instead gives an error message,
please put the core dump back."

There is no reason the fix for this could not be in a 4.4 patch. 
On the other hand this is such an obscure issue that getting
the fix distributed in a hurry is also not all that important.

kre