Re: sed question

[R. G. Newbury]

> Date: Thu, 21 Jul 2016 08:00:44 -0400
> From: bruce 
> Subject: sed question
> Hey guys..
>
> Sed question.. should be simple, but after stack/net searches, lots or
> trials.. can't seem to get it..
>
> I've got a case
>
> any thoughts on how to handle the parens would be cool!
>
> just irks me that I couldn't see what I missed.

You cannot 'escape' the parens since '(' and ')' are defined 'codes' 
like ^ & and $.

( marks the start  of an internal variable to be recorded, and ) marks 
the end, substitution in the output side uses \1 (or \2,\3, or as many 
as needed)

So sed -i -e "s#('txt')#'(/dir/txt)'#g" foo

should work.

NOTE: 1) sed inplace replacement
2) using extended grep
3) for clarity, the usual use of '/' for the 's' replacement is changed 
to '#'
4) the global replacement marker, usually ', is changed to "

sed gets confused when dealing with mixed ' and " markers,  I suspect 
that this is your problem. So use the opposite marker outside the sed 
replacement, when you have one inside.