Some kind of overlay texts generate urls which cause 400 Bad Request

[ttanimichi]

Some kind of overlay texts generate urls which cause 400 Bad Request

Describe the bug in a sentence or two.

require 'open-uri'

def foo(text)
  url = Cloudinary::Utils.cloudinary_url('sample', transformation: [
    {
      overlay: {
        font_size: 24,
        font_family: 'Sawarabi%20Gothic',
        text: text
      },
    }
  ])

  URI.open(url)
end

foo('hello') # No problem
foo('hello🍺') # => 400 Bad Request (OpenURI::HTTPError)

Issue Type (Can be multiple)

• Build - Cannot install or import the SDK
• Performance - Performance issues
• Behaviour - Functions are not working as expected (such as generate URL)
• Documentation - Inconsistency between the docs and behaviour
• Other (Specify)

Operating System

• Linux
• Windows
• macOS
• All

Environment and Libraries (fill in the version numbers)

$ bundle info cloudinary
  * cloudinary (1.21.0)
	Summary: Client library for easily using the Cloudinary service
	Homepage: http://cloudinary.com
	Path: /Users/ttanimichi/.rbenv/versions/3.0.3/lib/ruby/gems/3.0.0/gems/cloudinary-1.21.0

$ ruby -v
ruby 3.0.3p157 (2021-11-24 revision 3fb7d2cadc) [arm64-darwin21]

$ bin/rails -v
Rails 6.1.4.1

[michalkcloudinay]

Hi @ttanimichi, including emojis in text overlays can be done by double encoding the emoji as explained here:https://support.cloudinary.com/hc/en-us/articles/360003457091-How-to-overlay-emojis-in-images. Please note that not all emojis are supported. You can refer here for the list of supported emojis: https://www.piliapp.com/symbol/

[ttanimichi]

@michalkcloudinay Thank you for your reply.
Not only emojis but also some kind of chinese characters seem to cause the same error.

foo('漢字') # Most of chinese characters are No Problem
foo('あいうえお') # Japanese Hiragana letters are also No Problem

foo('𩸽') # => 400 Bad Request (OpenURI::HTTPError)
foo('𡈽') # => 400 Bad Request (OpenURI::HTTPError)
foo('𠀋') # => 400 Bad Request (OpenURI::HTTPError)

I found that the bytesize of 🍺, 𩸽 (Unicode: U+29E3D), 𡈽 and 𠀋 are four.

$ irb
irb(main):001:0> '𩸽'.bytesize
=> 4
irb(main):002:0> '🍺'.bytesize
=> 4
irb(main):003:0> '𡈽'.bytesize
=> 4
irb(main):004:0> '𠀋'.bytesize
=> 4

Probably, overlay texts of 4-bytesize Unicode characters cause this error. Isn't it a bug of Cloudinary API, is it?

[ttanimichi]

One of my best friend Mr. 𡈽田 (Tsuchida) can't display his name as overlay texts since his family name contains 𡈽, which is a 4-bytesize Unicode character. It's so much trouble for him...

[michalkcloudinay]

Hi @ttanimichi, for 4-bytesize Unicode characters you would need to use a double encoding. For example:
https://res.cloudinary.com/demo/image/upload/o_30/l_text:Arial_58:%2520%25F0%25A9%25B8%25BD/couple.jpg
Let us know if this works for you.

[ttanimichi]

@michalkcloudinay Thanks! I'll try it next week and let you know the result

[michalkcloudinay]

Great! Let us know how it goes.

[aleksandar-cloudinary]

Hi @ttanimichi - just wanted to follow up on this to see if you had a chance to test this and if it worked for you?