bug-grep
[Top][All Lists]
Advanced

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

Re: grep-2.5.1a egrep/fgrep PATH problem


From: Charles Levert
Subject: Re: grep-2.5.1a egrep/fgrep PATH problem
Date: Thu, 23 Jun 2005 17:32:59 -0400
User-agent: Mutt/1.4.1i

* On Thursday 2005-06-23 at 08:30:48 -0300, Tony Abou-Assaleh wrote:
> Would it make sense to add switches to the configuration script that would
> determine how egrep/fgrep are installed? Current option seem to be:
> 
> 1) don't install (should use grep -E/F)
> 2) install wrapper (what we do now)
> 3) install symlinks (since against GNU standards, should not be default)
> 4) make them all binaries
> 
> For option 4, we would need to add defines to the source so that the
> operation of the binary doesn't depend on the name, but that different
> binaries are actually compiled.

I don't know.  I'm not a big fan of anything else
but the wrapper script approach.  Within that
option, the question is whether to hardcode the
bindir path.  E.g.,

   /usr/bin/grep
   /usr/bin/fgrep execs "/usr/bin/grep -F"

   /usr/beta/bin/grep
   /usr/beta/bin/fgrep execs "/usr/beta/bin/grep -F"

I checked what we now have in CVS:

   
<http://savannah.gnu.org/cgi-bin/viewcvs/grep/grep/src/Makefile.am.diff?r1=1.31&r2=1.32>

with this construct:

   dir=${0%/*}/

which isn't too bad, but has the following
disadvantages:

   -- The % thing is not too portable.

   -- Should for some reason /usr/bin/fgrep
      be copied elsewhere or symlinked to, it
      will fail to set dir=/usr/bin/ anymore.
      (That would remain the case if sed were
      used instead of the % construct.)




reply via email to

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