Bug with non-breaking spaces in 0.3.0? about html2text HOT 6 CLOSED

This definitely sounds like sadness, some weird interaction between UTF-8 and str_replace. Maybe there needs to be some utf8_encode()/utf8_decode(), maybe there needs to be some iconv/mbstring, maybe html2text simply can't support this... What was the hex output of html2text 0.2.3 on the same input?

Can you create a pull request with a new test (in tests/), both with the before and expected after text encoded as UTF-8, and we can go from there?

from html2text.

Sure thing :)

from html2text.

I have the same problem. All the content that was converted correctly before the upgrade from 0.2.3 to 0.3.0 now brings question marks characters, not recognized as UTF8, breaking the output.

Removing the line

$html = str_replace("\xa0", " ", $html);

from "Html2Text" class resolves the problem

from html2text.

Can you all check out the new non-breaking-space branch and see if this resolves the issue for your tests? (PHP + character encodings = much sadness!)

If so, I'll merge into master and release 0.3.1 right away. If not, can you please contribute more pull requests off this branch, with additional tests?

Thank you @ulrichsg for the test contribution!

from html2text.

The new commit in the non-breaking-space branch solves the problem for my use cases

from html2text.

Resolved with PR #32 and released as 0.3.1

from html2text.