Update: Sorry, I just realised, this is in essence what has been suggested by 1Nf3 right at the beginning of this thread ... apologies!

*********************************************************

Personally, I like the following two-liner because it is concise, simple and flexible (as illustrated, it will also work with an arbitrary number of indented lines):

( my $line = ( join "", <>  ) ) =~ s/\n\s+//g;    
print $line;



______DATA_______

firstline54324563213612531436436463246324634
         secondlidfndsfdsfdsfd
         thirdlinedsfsdf
firstline54324563213612531436436463246324634
         secondlidfndsfdsfdsfd
         thirdlinedsfsdf
         fourthlinedsfsdf
         fifthlinedsfsdf

_____OUTPUT______

firstline54324563213612531436436463246324634secondlidfndsfdsfdsfdthird
+linedsfsdf
firstline54324563213612531436436463246324634secondlidfndsfdsfdsfdthird
+linedsfsdffourthlinedsfsdffifthlinedsfsdf
[download]

Essentially, it compounds the entire data file into a single string from which any occurrence of \n\s+ is then removed. The latter string sequence is the one which occurs at every line end except for those line ends before a firstline.

Hope this helps.

Best regards -

Pat