RDoc and encoding

Ruby
1

Hi,

Running Ruby/JRuby 1.8.7 on Windows XP.

Until now I have been using the RDoc version built into the Ruby
Standard Library. That is version 1.0.1. Now I am trying out RDoc 3.4,
installed via a gem.

I have run into a problem with the double quote character. Example code:

RDoc 1.0.1

require ‘rdoc/markup/simple_markup’
require ‘rdoc/markup/simple_markup/to_html’

sm = SM::SimpleMarkup.new()
th = SM::ToHtml.new()
puts sm.convert(‘æææ"øøø"ååå’, th)

Output:

æææ"øøø"ååå

RDoc 3.4

require ‘rubygems’
require ‘rdoc/markup/to_html’

puts RDoc::Markup::ToHtml.new().convert(‘æææ"øøø"ååå’)

Output:

æææ“øøø”ååå

It seems as if RDoc 3.4 is adding a double quote in UTF-8 encoding
instead of “”". Running on Windows XP, the normal encoding is
Windows-1252. If I look at the HTML and tell the browser that it is
UTF-8 encoded, the double quotes are displayed correctly. Then, however,
the Danish national characters (æøå) are not displayed as they should.

Do you think I have hit a bug in Rdoc 3.4, or am I missing something?

Claus

2

On Jan 10, 2011, at 04:08, Claus Folke B. wrote:

RDoc 1.0.1

Output:

æææ“øøø”ååå

It seems as if RDoc 3.4 is adding a double quote in UTF-8 encoding
instead of “”". Running on Windows XP, the normal encoding is
Windows-1252. If I look at the HTML and tell the browser that it is
UTF-8 encoded, the double quotes are displayed correctly. Then, however,
the Danish national characters (æøå) are not displayed as they should.

Do you think I have hit a bug in Rdoc 3.4, or am I missing something?

Transcoding is not supported in RDoc on ruby 1.8.7. Upgrade to Ruby
1.9.

My primary platform for developing RDoc is Ruby 1.9. Ruby 1.8.6 is
unsupported and 1.8.7 gets second tier status and will not support
transcoding.

3

Eric H. wrote in post #973761:

Transcoding is not supported in RDoc on ruby 1.8.7. Upgrade to Ruby
1.9.

I don’t think it is a matter of transcoding. I would have thought the
output would remain in the Windows-1252 encoding of the input.

As I can figure out, RDoc always “thinks” the input is in UTF-8
encoding. This is probably rarely the case on Windows.

Can you explain the use of a double quote in UTF-8 encoding instead of
“”" in the generated HTML?

Claus

4

On Jan 10, 2011, at 13:20, Claus Folke B. wrote:

Eric H. wrote in post #973761:

Transcoding is not supported in RDoc on ruby 1.8.7. Upgrade to Ruby
1.9.

I don’t think it is a matter of transcoding. I would have thought the
output would remain in the Windows-1252 encoding of the input.

As I can figure out, RDoc always “thinks” the input is in UTF-8
encoding. This is probably rarely the case on Windows.

With Ruby 1.8 this is true. If you upgrade to Ruby 1.9 RDoc 3 can
automatically determine the output encoding and transcode for you. You
can also override it with --encoding.

Can you explain the use of a double quote in UTF-8 encoding instead of
“”" in the generated HTML?

RDoc now performs “prettier” replacements of characters such as matching
opening and closing quotes. Such characters are not available in all
output encodings so transcoding is performed.