bug-libtool
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: echo vs. printf regression (darwin8)


From: Ralf Wildenhues
Subject: Re: echo vs. printf regression (darwin8)
Date: Tue, 17 Aug 2010 20:50:31 +0200
User-agent: Mutt/1.5.20 (2010-04-22)

* Paolo Bonzini wrote on Mon, Aug 16, 2010 at 10:28:33PM CEST:
> On Mon, Aug 16, 2010 at 21:53, Ralf Wildenhues wrote:
> > Whatever we do, NEWS needs a big warning note.
> 
> Why? It's just namespace pollution, which has always been there.

* Libtool configure macros do not use nor set the variables $RM, $MV,
  $ECHO any more.  These variables are also in use by some other
  packages for conflicting purposes, the values required by Libtool
  are in general only suitable for the generated `libtool' script
  but not for use inside configure or Makefile files.  Libtool macros
  now use the variables lt_RM, lt_MV, and lt_ECHO instead.
  If you use LT_OUTPUT inside configure and write tests that exercise
  one of the variables specifying libtool semantics such as
  $archive_cmds, you might need to temporarily set $ECHO and $RM from
  $lt_ECHO and $lt_RM.


I haven't looked in detail yet, but the above seems the most sane
solution to me, and yes, I think that anything that otherwise allows
users to cry "regression" warrants a big warning sign.

If somebody would like to take a stab at a patch (or only parts of it),
be my guest.  The patch would benefit from testing on all of GNU/Linux,
Solaris with a losing $CONFIG_SHELL forced, and one ksh or ash on a
different system (I can do the testing if needed).

Thanks,
Ralf



reply via email to

[Prev in Thread] Current Thread [Next in Thread]