There is an edge case when parsing export snippets:

@@html:fo aksjdaksjd ajs d

askdjas aksj daj sao@@

Our current parser recognises the opening "@@html:" as a standalone
snippet.

Unless I misunderstand something, it is not intentional.

The fix is attached.

Nicolas, unless you have any objections, I will install the attached
patch to bugfix.

Best,
Ihor