[bug] validation of chars

[drwetter]

Hi,

often I paste content into the wiki which has either a charset not in the scope of the document or it's not printable.

Upon opening with libreoffice I encounter a popup box and after acknowledging it closes.

It would be great if the dokuwiki plugin would either omit such characters or try to sanitize them. This bug maybe also has a security implication (http://cwe.mitre.org/data/definitions/19)

I worked my way around that by editing the content.xml and put the archive together but it's kind of tedious.

Thanks, Dirk

[Klap-in]

Could you provide examples to make this reproducible for others?

DokuWiki should store it, and show it if relevant. And should not break on the exotic stuff. But I expect that everything that's utf-8 should be handled properly (is this assumption right?) so is my interpretation right that DokuWiki does not need change?

The troubles you mentioned are when the data is converted to another output format (odt). So direct solution would be that the odt plugin sanatizes the data more? And are other export plugins also affected?

[Klap-in]

Ah, this issue is already submitted for the odt plugin. Sorry I'm reacting via my phone, resulting in less overview.

[drwetter]

I can't tell whether it was UTF-8 as it was a binary display from an image in a client side proxy. That I encounter such chars is a use case for me. Those hieroglyphs I wanted to show up in the final document. I assumed though that somebody in the chain from dokuwiki, odt plugin, libreoffice is cleaning the mess up.

Let me know If you want a hexdump of some sections.

[drwetter]

e.g.