wraps text lines at the specified width (90 here). -s option is to force to wrap on blank characters -b count bytes instead of columns

When you use a "for" construct, it cycles on every word. If you want to cycle on a line-by-line basis (and, well, you can't use xargs -n1 :D), you can set the IFS variable to . Show Sample Output

---- this line ends here but must be concatenated with this one "this line ends here" and should NOT be concatenated with this one

awk version of 7210. Slightly longer, but expanding it to catch blank lines is easier: awk 'BEGIN{RS="\0"}{gsub(/\n+/,"<SOMETEXT>");print}' file.txt Show Sample Output

Handle any bad named file which contains ",',\n,\b,\t,` etc Store the file name as null character separated list find . -print0 >name.lst and retrieve it using read -r -d "" Eg: find . -print0 >name.lst; cat name.lst| while IFS="" read -r -d "" file; do ls -l "$file"; done Show Sample Output

Let -p do it's voodoo and do absolutely nothing but set the output record separator :-)

Copying and pasting from Office documents open in Office:mac can dirty your files with Windows CRLF and (inexplicably) Classic Mac OS LF newlines, which can break some tools. This snippet replaces them with good ol' Unix LF newlines.

In case the line you want to join start with a char different than ", you may use \n.*"\n as regex.