Comment 3 for bug 1366580

While working on this, I found a probably better solution:
- write the listitem into its own paragraph, i.e. return "<p line-spacing=3 font-size=12>"..symbol.." "..text.."<br></p><p font-size=8> <br></p>" (the size of the empty line can be adjusted)
- to avoid further problems, remove the p(body) resp. p(text) from speech() and from objective_text(). Instead, make the scripter open paragraphs himself.

The latter point has two consequences:
These two functions now have the same syntax as message_box, i.e. the same string can be passed to speech or directly be shown.
The downside: It is not possible anymore to place translatable quotation marks around the text in speech(). I do not think that that is grave; there is an image of the speaking person next to the text.

Are there things I have overlooked? Otherwise, I consider this the cleanest solution and will implement it.