How to detect and handle different EOL in Ruby?

I am trying to process a CSV file that can either be generated with CF or LF as an EOL marker. When I try to read the file with

infile = File.open('my.csv','r')
whi开发者_如何学JAVAle line = infile.gets
...

The entire 20MB file is read in as one line.

How can I detect and handle properly?

TIA

I would slurp the file, normalize the input, and then feed it to CSV:

raw = File.open('my.csv','rb',&:read).gsub("\r\n","\n")
CSV.parse(raw) do |row|
  # use row here...
end

The above uses File.open instead of IO.read due to slow file reads on Windows Ruby.

When in doubt, use a regex.

> "how\r\nnow\nbrown\r\ncow\n".split /[\r\n]+/
=> ["how", "now", "brown", "cow"]

So, something like

infile.read.split(/[\r\n]+/).each do |line|
  . . .
end

Now, it turns out that the standard library CSV already understands mixed line endings, so you could just do:

CSV.parse(infile.read).each do |line|
  . . .