csv.foreach stuck, not calling the block

I am trying to write a CSV "fixer". Unfortunately It seems that the csv.foreach instruction is not calling the lambda I have created. The CPU is used at 100%. Just wondering what ruby is doing in the meantime...开发者_如何转开发

Any ideas why my code is wrong?

  1 require "csv"
  2 
  3 ARGV.empty? do
  4     print "usage: fixcsv.rb <filename>"
  5     exit
  6 end
  7 
  8 filename_orig = Dir.pwd + "/" + ARGV[0]
  9 filename_dest = filename_orig.sub(/csv$/,"tmp.csv")
 10 topic = filename_orig.sub(/_entries.csv$/,"").sub(/.*\//,"")
 11 
 12 puts "topic:" + topic
 13 
 14 writer = CSV.open(filename_dest,"w",:col_sep=>";")
 15 #i=0
 16 cycler = lambda do |row|
 17     #i = i + 1
 18     #puts "row number:" + i.to_str
 19     #row[17] = topic
 20     puts "foo"
 21     writer << row
 22 end
 23 
 24 begin
 25     CSV.foreach(filename_orig,:col_sep=>",",&cycler)
 26 rescue
 27     puts "exception:" + $!.message
 28     exit
 29 else
 30     writer.close
 31 end

Here is the stack trace produced when I Ctrl-C it:

stab@ubuntu:~/wok$ ruby addtopic.rb civilpoliticalrights_entries.csv

topic:civilpoliticalrights
^C/usr/lib/ruby/1.8/csv.rb:914:in `buf_size': Interrupt
    from /usr/lib/ruby/1.8/csv.rb:825:in `[]'
    from /usr/lib/ruby/1.8/csv.rb:354:in `parse_body'
    from /usr/lib/ruby/1.8/csv.rb:227:in `parse_row'
    from /usr/lib/ruby/1.8/csv.rb:637:in `get_row'
    from /usr/lib/ruby/1.8/csv.rb:556:in `each'
    from /usr/lib/ruby/1.8/csv.rb:531:in `parse'
    from /usr/lib/ruby/1.8/csv.rb:311:in `open_reader'
    from /usr/lib/ruby/1.8/csv.rb:94:in `foreach'
    from addtopic.rb:25

EDIT: Ruby version is:

$ ruby --version
ruby 1.8.7 (2010-01-10 patchlevel 249) [i486-linux]

Your program worked fine for me in Ruby 1.9.

I have a few observations:

• If your input pathname does not end in csv, then the input and output file names will be the same. This could easily produce an infinite loop.

• You are definitely using the 1.9 flavor of csv. If this program needs to run on 1.8.7 it would need to have patches from the snippet below...

Mods for 1.8.7:

writer = CSV.open(filename_dest, "w", ?;)
#i=0
cycler = lambda do |row|
  #i = i + 1
  #puts "row number:" + i.to_str
  #row[17] = topic
  writer << row
end

begin
  CSV.open filename_orig, 'r', ?,, &cycler

The main problem with 1.8.7 csv is that the interfaces to CSV.open and CSV.foreach do not take Hash options. Worse, they are expecting numeric code points, a feature of Ruby that apparently didn't work out and was withdrawn in 1.9.