sed bug?

[Brent Wood]

I was using the following command to change all the "NaN"s in a file to
"0.0"


cat tmp.txt | sed 's/NaN/0.0/g' > $OUTFILE


This worked until the input file had some unusually long lines (about
8000+ characters) when sed seems to truncate the output lines, which lose
the line feed at the end so I get lines with 250,000+ chars. I was
unaware of sed having any such line length limitations. The command


cat tmp.txt | tr "a" "." | tr "N" "0" > $OUTFILE

seems to work with the longer lines OK.




If there is such a limitation for sed, I suggest the man page specify
this??


Thanks,

  Brent Wood