SOLVED: Catch fails when illegal letters/bytes in text

Hi there,

in the last few weeks I got errors from I generated a feed via Creator and let FTR catch the full article.

In FreshRSS I see literal html code in some articles instead of the article text, only the title is correct. Today I did some investigation in this issue. I think Creator, is not the problem. But FTR cannot deal with the faulty text of the article.


In the first sentence there are two single bytes with value 0x02 within the text. Marked here with X:

  • Fünf FahrzeuglenXkern wurde im dortigen Bereich der Führerschein an Ort und Stelle abgenomXmen.

These illegal chars do not replace a regular char of the text.

When fetching this article by (v3.10) I got this error:


When trying this with my self-hosted FTR 3.9.13, I got this red errorbox on yellow ground but without the black and red text like in the screenshot above.

Can you intercept this somehow, @fivefilters ?

The places in the words look as if they are hidden hyphenations/syllabifications

I resolved this by just replacing the illegal character:


You don’t see the 0x02 char between brackets in the example, but it is there: