ruby 1.9 + sinatra incompatible character encodings: ASCII-8BIT and UTF-8

6
votes

I'm trying to migrate a sinatra application to ruby 1.9

I'm using sinatra 1.0, rack 1.2.0 and erb templates

when I start sinatra it works but when I request the web page from the browser I get this error:

Encoding::CompatibilityError at /
incompatible character encodings: ASCII-8BIT and UTF-8

all .rb files has this header:

#!/usr/bin/env ruby
# encoding: utf-8

I think the problem is in the erb files even if it shows that it's UTF-8 encoded

[user@localhost views]$ file home.erb
home.erb: UTF-8 Unicode text

any one had this problem before? is sinatra not fully compatible with ruby 1.9?

2
rubyencodingsinatraerbutf-8
Try temporarily changing the files to ascii only. – Adrian
the problem is I need to use utf-8 charterers in the templates. – John
If you are using HTML, you should replace them with entities. Otherwise, you might want to try temporarily taking them out just to see if they are the problem. – Adrian

2 Answers

15
votes

I'm not familiar with the specifics of your situation, but this kind of error has come up in Ruby 1.9 when there's an attempt to concatenate a string in the source code (typically encoded in UTF-8) with a string from outside of the system, e.g., input from an HTML form or data from a database.

ASCII-8BIT is basically a synonym for binary. It suggests that the input string was not tagged with the actual encoding that has been used (for example, UTF-8 or ISO-8859-1).

My understanding is that exception messages are not seen in Ruby 1.8 because it treats strings as binary and silently concatenates strings of different encodings. For subtle reasons, this often isn't a problem.

I ran into a similar error yesterday and found this excellent overview.

One option to get your error message to go away is to use force_encoding('UTF-8') (or some other encoding) on the string coming from the external source. This is not to be done lightly, and you'll want to have a sense of the implications.

0
votes

I had the same issue. The problem was a utf8 encoded file which should be us-ascii.

I checked using the file command (on OSX):

$ file --mime-encoding somefile
somefile: utf-8

After removing the weird characters from the file:

$ file --mime-encoding somefile
somefile: us-ascii

This fixed the issue for me.