ruby gedcom parser EOF exception

Question

i need to parse gedcom 5.5 files for a analyziation project. The first ruby parser i found couses a stack level too deep error, so i tryed to find alternatives. I fount this project: https://github.com/jslade/gedcom-ruby

There are some samples included, but i don't get them to work either.

Here is the parser itself: https://github.com/jslade/gedcom-ruby/blob/master/lib/gedcom.rb

If i try the sample like this:

ruby ./samples/count.rb ./samples/royal.ged

i get the following error:

D:/rails_projects/gedom_test/lib/gedcom.rb:185:in `readchar': end of file reached (EOFError)

I wrote a "gets" in every method for better unterstanding, this is the output till the exception raises:

Parsing './samples/royal.ged'...
INIT
BEFORE
CHECK_PROC_OR_BLOCK
BEFORE
CHECK_PROC_OR_BLOCK
PARSE
PARSE_FILE
PARSE_IO
DETECT_RS

The exact line that causes the trouble is

while ch = io.readchar

in the detect_rs method:

# valid gedcom may use either of \r or \r\n as the record separator.
# just in case, also detects simple \n as the separator as well
# detects the rs for this string by scanning ahead to the first occurence
# of either \r or \n, and checking the character after it
def detect_rs io
puts "DETECT_RS"
  rs = "\x0d"
  mark = io.pos
  begin
    while ch = io.readchar
      case ch
      when 0x0d
        ch2 = io.readchar
        if ch2 == 0x0a
          rs = "\x0d\x0a"
        end
        break
      when 0x0a
        rs = "\x0a"
        break
      end
    end
  ensure
    io.pos = mark
  end
  rs
end

I hope someone can help me with this.

Answer 1

The readchar method of Ruby's IO class will raise an EOFError when it encounters the end of the file. http://www.ruby-doc.org/core-2.1.1/IO.html#method-i-readchar

The gedcom-ruby gem hasn't been touched in years, but there was a fork of it made a couple of years go to fix this very problem.

Basically it changes:

while ch = io.readchar

to

while !io.eof && ch = io.readchar

You can get the fork of the gem here: https://github.com/trentlarson/gedcom-ruby