NetNews Usenet Archive 1992 #16/NN_1992_16.iso/spool/comp/unix/shell/3103

Internet Message Format | 1992-07-23 | 1.6 KB

Xref: sparky comp.unix.shell:3103 comp.unix.questions:9343
Newsgroups: comp.unix.shell,comp.unix.questions
Path: sparky!uunet!mcsun!sunic!aun.uninett.no!ugle.unit.no!news
From: [e-mail]
Subject: Re: What's wrong with this sed command?
In-Reply-To: [e-mail] (John Mackin)
Message-ID: <1992Jul23.192253*[e-mail]>
Keywords: sed' too much
Sender: [e-mail] (NetNews Administrator)
Organization: SINTEF DELAB, Trondheim, Norway.
References: <[e-mail]> <glaze.2520@glaze>
    <[e-mail]>
Date: 23 Jul 92 19:22:52
Lines: 28

John Mackin on `sed':
! What the original poster wants cannot be done in one step with the ed
! subset of regular expressions (as used by sed).  If sed accepted full regular
! expressions, a la egrep, then the command would be:
!
!     sed 's?: (\./)*?: ?'        WRONG WRONG WRONG

You can do that with GNU sed.  One more reason for everybody to rm /bin/sed
and install the GNU derived one..... :)  The *only* drawback is that it
probably runs a bit slower.

! but since sed doesn't do full RE's, that won't work.  You need
! another approach.  Since you can't do it in one step, do it
! in several:
!
!     sed -e ': loop' -e 's?: ./?: ?' -e 't loop'

I've always wondered why ":label; ....." was doomed to be unusable.
The way it is, I believe I may put it as

        echo 'rubbish.o: ././././././rubbish.c' | /bin/sed ':;
          s;: \./;: ;; t;'

for what it's worth :).  (Hey, that's a non-csh continuation there, you do
remember to make that   ....sed ':;\   before entering the 1.st newline in
csh/tcsh, dont you?).

   ~~harald E.