[Cialug] Perl regex

[Daniel A. Ramaley]

Since we're posting brain teasers, here's what i'm having trouble with:

Is there a way to design a Perl s/// regex such that after each 
replacement it starts searcing for the next replacement at the 
beginning of where it found the previous match instead of after the 
previous match?

What i'm trying to accomplish is to take files that look something like 
this:

# Comment
AB=some text
CDE=text
CDE= more text
CDE=even more
F=foo
GH=bar,
GH= baz

And concatenate the continued lines together, so after processing it 
looks like this:

# Comment
AB=some text
CDE=text more texteven more
F=foo
GH=bar, baz

Here's what i have so far, with the entire file contained in $data (i 
already had to slurp the whole thing for other processing that was done 
to it first):

1 while $data =~ s/^(([^#][^=]*)=.*)\n\2=/$1/gmo;

That works, but if there is a way to tell the regex parser to not 
advance its internal pointer after each match, then i can get rid of 
the useless while loop. (As it is now, without the while it only works 
on lines that have been split into 2 (such as "GH" in the example) but 
not lines split into 3 or more (such as "CDE").)

------------------------------------------------------------------------
Dan Ramaley                            Dial Center 118, Drake University
Network Programmer/Analyst             2407 Carpenter Ave
+1 515 271-4540                        Des Moines IA 50311 USA