Bug #13738: Active Scaffold image upload using CarrierWave gem

• Status: Third Party's Issue
• Assignee: -
• Target version: -
• ruby -v: 2.3.1
• Backport: 2.2: UNKNOWN, 2.3: UNKNOWN, 2.4: UNKNOWN

[karanamsandhya (Karanam Sandhya)]

getting Encoding::UndefinedConversionError ("\xFF" from ASCII-8BIT to UTF-8): this error while uploding the image. and image params are like this:

"photo"=>#<ActionDispatch::Http::UploadedFile:0x007fbcc2925c80 @tempfile=#Tempfile:/tmp/RackMultipart20170712-7555-c67kvs.jpg, @original_filename="1.jpg", @content_type="image/jpeg", @headers="Content-Disposition: form-data; name="record[photo]"; filename="1.jpg"\r\nContent-Type: image/jpeg\r\n">

[k0kubun (Takashi Kokubun)]

• Status changed from Open to Third Party's Issue

Did you see https://github.com/carrierwaveuploader/carrierwave/issues/1404?
As far as I can see from your description, it's unlikely to be a bug in Ruby-side. Please confirm the issue and then could you report it to carrierwave or Rails?

[duerst (Martin Dürst)]

karanamsandhya (Karanam Sandhya) wrote:

getting Encoding::UndefinedConversionError ("\xFF" from ASCII-8BIT to UTF-8): this error while uploding the image.

There's definitely no error in Ruby itself, in the encoding "ASCII-8BIT", only the lower half is convertible; the upper half is okay in the source, but undefined and so not convertible to any other encoding.

I don't know the CarrierWave gem. But if they are trying to convert image data (which very easily can contain a 0xFF byte, and for which a String with an encoding of ASCII-8BIT (alias: BINARY) is appropriate) to UTF-8, then they are definitely making a mistake. If the error happens for some other data (e.g. for some text data extracted from an image header), then I can understand that such an error happens. If the image format actually defines what encoding to use for the text, then String#force_encoding should be applied first. If not, I'd suggest to change the conversion from something like

string_with_unknown_encoding_labeled_as_ascii_8bit.encode('UTF-8')

to something like

string_with_unknown_encoding_labeled_as_ascii_8bit.encode('UTF-8', invalid: :replace)

which will convert 0xFF bytes (and similar) to '?'.