Re: Replace 1 Char with 3Char's

Hi Dennis,

Okay... Well, I've been using FreeBSD for about 13 years. Personally, when I write Unix software, I try to go to the least common denominator so that it'll be portable.

I agree that for the purposes of inserting a CRLF into the data, the GNU version is more elegant. But is it _so_ elegant that it's worth restricting your software so that it'll only work in one environment? Or that you'll have to install a second version of sed in each place you want your script to work?

In this example, Jim Lowary would have to ensure that each time he migrates to another box, the new box gets the GNU version of sed installed and placed ahead of the standard version. If someone else besides Jim ever maintains the box, it'll have to be very carefully documented... otherwise people won't understand why they're installing sed on a system (or multiple systems) that already has sed.

I guess that's up to Jim to decide...

If it were me, I'd say the $(print '\\\\r\\\\n') isn't THAT bad.

Dennis Lovelady wrote:

The GNU version does give the option of being true POSIX-compliant (via its
--posix switch), which gives it an advantage over the older-style Berkley
derivatives, in my opinion. But it comes down to "which standard do you
want to choose? There are so many." I like the GNU stuff. But that's
probably because I've been using GNU for about twenty years. For example
SUN and AIX systems I've administered in the past (and now iOS) lacked the
GNU capabilities; so they were installed and placed in front of the
system-supplied stuff in the path.