End of Line Overflow in start of next line - bash

Question

So I have come across an AWK script that used to be working on HP-UX but has been ported over to RHEL6.4/6.5. It does some work to create headers and trailers in a file and the main script body handles the record formatting.
The problem I am seeing when it runs now is that the last letter from the first line flows onto the start of the next line. Then the last two letters of the second line flow into the start of the third and so on.
This is the section of the script that deals with the record formatting:
ls_buffer=ls_buffer $0;
while (length(ls_buffer)>99) {
if (substr(ls_buffer,65,6)=="STUFF") {
.....do some other stuff
} else {
if (substr(ls_buffer,1,1)!="\x01f" && substr(ls_buffer,1,1)!="^") {
printf "%-100s\n", substr(ls_buffer,1,100);
}
};
#----remove 1st 100 chars in string ls_buffer
ls_buffer=substr(ls_buffer,100);
}
To start with it looks like the file had picked up some LF,CR,FF so I removed them with gsub hex replacements further up the code but it is ending the line at 100 and then re-printing the last character at the start of the second line.
This is some sample test output just in case it helps:
1234567890123456789012345678901 00000012345TESTS SUNDRY PAYME130 DE TESTLLAND GROUP
P1234567890123456789012345678901 00000012345TESTS SUNDRY PAYME131 TESTS RE TESTSLIN
NS1234567890123456789012345678901 00000012345TESTS SUNDRY PAYME132 TESTINGS MORTGAG
GES1234567890123456789012345678901 00000012345TESTS SUNDRY PAYME937 TESTS SUNDRY PA
Can anyone offer any suggestions as to why this is happening? Any help would be appreciated.

Answer 1

The problem here seems to be that the offsets are incorrect in the manual buffer printing loop.
Specifically that the loop prints 100 characters from the buffer but then strips only 99 characters off the front of the buffer (despite the comments claim to the contrary).
The substr function in awk starts at the character position of its second argument. So to drop x characters from the front of the string you need to use x+1 as the argument to substr.
Example:
# Print the first ten characters from the string.
$ awk 'BEGIN {f="12345678901234567890"; print substr(f, 1, 10)}'
1234567890
# Attempt to chop off the first ten characters from the string.
$ awk 'BEGIN {f="12345678901234567890"; print substr(f, 10)}'
01234567890
# Correctly chop off the first ten characters from the string.
$ awk 'BEGIN {f="12345678901234567890"; print substr(f, 11)}'
1234567890
So the ls_buffer=substr(ls_buffer,100); line in the original script would seem to need to be ls_buffer=substr(ls_buffer,101); instead.
Given that you claim that the original script is working however I have to wonder if whatever version of awk is on that HP-UX machine had a slightly different interpretation of substr (not that I see how that could be possible).
The above aside this seems like a very odd way to go about this business (manually assembling a buffer and then chopping it up) but without seeing the input and the rest of the script I can't comment much more in that direction.