On Wed, 15 Jan 2014 19:11:56 -0600
"Darrell Anderson" <darrella@...> wrote:

> >> >I would like to inspect source tree text strings for instances 
> >of
> >> >KDE/kde rebranding issues and update to TDE/tde. Mostly in
> >> >tooltips
> >> >and What's This help strings. My challenge is being able to
> >> >efficiently grep only those strings, which often in the sources
> >> >are multiple lines.
> >> >
> >> >I would appreciate shell scripting advice to find such text
> >> >strings.
> >>
> >> Nobody knows?
> >
> >It would help to search whole words - grep option '-w'?
> 
> I don't understand how that helps. For example, I want to find all 
> TQWhatsThis strings, which often are multiple lines, and ensure I 
> am searching the entire string, which includes all the multiple 
> lines. About the only thing I know is the search will key on 
> something like TQWhatsThis and the entire glob of text to search 
> will end with an apostrophe (;). All of that constitutes the entire 
> string that must be searched for 'KDE/kde'.

I think people may have misinterpreted your use of the word "efficiently" 
in the original post to mean that you had the basics worked out and
were just worried about the run time (I know I did).

I don't think you can do this with grep, which operates primarily on
lines, unless you want to use shell script to read in each file and
concatenate the lines together without CR characters, then run grep 
on the result.  Much easier to forget about shell scripting and do it in 
Perl, in which case you need to recurse across the directory tree,
slurp in entire files, run each through a regular expression match 
like (from your example, and assuming you meant "semicolon" where 
you wrote "apostrophe"):
/(TQWhatsThis.*?kde.*?\;)/gim
and print out the matches.

E. Liddell