The T-SQL PRINT statement is the way to output plain text from SQL scripts and stored procedures. Unfortunately, it has a built-in limitation of printing 8000 non-Unicode or 4000 Unicode characters, and the rest is truncated. I was writing some SQL that generated lengthy SQL scripts that exceeded this limitation, and I wanted to create my own version of PRINT that would split my string at the nearest CRLF, print the part preceding the CRLF, then repeat on the rest of the string to the right of the CRLF. The CRLF itself would be omitted, since each PRINT statement appends its own to the output. However, my own string would sometimes contain two consecutive CRLF sequences for formatting purposes. My printing procedure would correctly strip the second one and leave the first, passing it to PRINT. But instead of my CRLF plus PRINT's, I only got one, spoiling my formatting. I ran the following test:

declare @cr char(1), @lf char(1)
set @cr = char(13)
set @lf = char(10)

print 'cr + lf' + @cr + @lf
print '---'

print 'cr only' + @cr
print '---'

print 'lf only' + @lf
print '---'

print 'lf + cr' + @lf + @cr
print '---'

And got the following output:

cr + lf
---
cr only

---
lf only

---
lf + cr

---

Interesting results! Notice that when a string is followed by CRLF, there is only one, not one plus one from PRINT. A solitary CR or LF will result in a new line plus one from PRINT, and LFCR results in two new lines plus one from PRINT. Since SQL Server honors a single CR or LF and they work as I expected, I updated my SQL generation code to use LF only as new line characters.