what about firing up grep as well ...
grep -i ^.TH | sed 's/4F/4/g'
steve
-----Original Message-----
From: Lee Howard [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, November 15, 2000 8:24 PM
To: Matthew Melvin
Cc: [EMAIL PROTECTED]
Subject: Re: sed -e "s/\(^.TH*\)4F/\14/g"
At 05:12 PM 11/15/00 +1000, Matthew Melvin wrote:
>Just to make sure i'm grokking this right...
>... so basically the F gets dropped.
Yes, that's correct - only on lines that begin with 'TH' as the second and
third characters. It's run during configure to change manpage numbers to
appropriate values on a per-OS basis (software development).
>Sounds like their sed is no fun all... do the same thing with perl?
It appears that the SCO, Solaris, and probably HP-UX versions of sed barf
when the \(match\) is empty or undefined. This is somewhat interesting,
because you can exploit this bug by doing something like:
echo 'hello world' | sed 's/\(\)/\1/g'
because the \(match\) is defined to be empty rather than undefined as empty.
>perl -p -e 's/(^.TH*)4F/${1}4/'
It will take a fair amount of work to change this to perl, unfortunately.
But so far it looks like my only hope. I can't expect all of the potential
users of this software to go out and change out their sed version.
Thanks.
Lee.
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list