Class: Erector::UnicodeBuilder

Inherits:
Object
  • Object
show all
Defined in:
lib/erector/unicode_builder.rb

Overview

Note that this class is only used in building erector itself (and even then, only needs to be run when there is a new UnicodeData.txt file from unicode.org).

Instance Method Summary collapse

Constructor Details

#initialize(input, output) ⇒ UnicodeBuilder

Returns a new instance of UnicodeBuilder.



6
7
8
9
10
 
# File 'lib/erector/unicode_builder.rb', line 6

def initialize(input, output)
  @input = input
  @output = output
  @first = true
end

Instance Method Details

#generate ⇒ Object 



12
13
14
15
16
 
# File 'lib/erector/unicode_builder.rb', line 12

def generate()
  @output.puts "Erector::CHARACTERS = {"
  process_file
  @output.puts "}"
end

#namify(name) ⇒ Object 



62
63
64
 
# File 'lib/erector/unicode_builder.rb', line 62

def namify(name)
  name.downcase.gsub(/[- ]/, '_')
end

#output(name, code_point) ⇒ Object 



58
59
60
 
# File 'lib/erector/unicode_builder.rb', line 58

def output(name, code_point)
  output_line "  :#{namify(name)} => 0x#{code_point.downcase}"
end

#output_line(line) ⇒ Object 



32
33
34
35
36
37
38
39
40
 
# File 'lib/erector/unicode_builder.rb', line 32

def output_line(line)
  if (!@first)
    @output.puts(',')
  end
  
  @output.print(line)

  @first = false
end

#process_file ⇒ Object 



18
19
20
21
22
23
24
25
26
27
28
29
30
 
# File 'lib/erector/unicode_builder.rb', line 18

def process_file()
  while !@input.eof
    line = @input.gets.strip
    if (line == "")
      next;
    end
    
    process_line(line)
  end
  if (!@first)
    @output.puts
  end
end

#process_line(line) ⇒ Object 



42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
 
# File 'lib/erector/unicode_builder.rb', line 42

def process_line(line)
  fields = line.split(';')
  code_point = fields[0]
  name = fields[1]
  alternate_name = fields[10]

  if /^</.match(name)
    return ""
  end

  output name, code_point
  if (!alternate_name.nil? && alternate_name != "")
    output alternate_name, code_point
  end
end