Takeaway (2)

[bpbond]

When you have a minute, please take a look at b664ee0.

The parse_8100A_file() basically looked like this:

parse_8100A_file <- function(file) {
  determine how many records there are in the file
  loop through them {
    parse each record
  }
  return(results)
}

This made the function overly long, and the "parse each record" part really hard to test. In the commit above I pulled it out, so the file now looks like this:

parse_8100A_record <- function(record_text) {
  parse the record_text
}

parse_8100A_file <- function(file) {
  determine how many records there are in the file
  loop through them {
    record <- isolate_record_text
    results <- results + parse_8100A_record(record)
  }
  return(results)
}

Because it's broken apart, the parsing code becomes MUCH easier to test--which we now do extensively!

That's it, close when ready, just wanted to make the point.