sometimes you want to have a slide with a picture with no text at all, and have some speaker notes to remember what you have to say at that point.

but if the speaker notes come right below the slide delimiter, pinpoint fails to parse them and add the first line of speaker as if they were slide text.

For example, in the following slide, pinpoint would put "# cat" as the slide text, and just the second line as the actual speaker notes.

--------------------------8<------------------------
-- [linus.jpg]
# a cat
# everybody loves cats, right?
--------------------------8<------------------------

A workaround is adding a blank line between slide delimiter and comments, but that breaks the possibility to have them close to each other as a unique block.

--------------------------8<------------------------
-- [linus.jpg]

# a cat
# everybody loves cats, right?
--------------------------8<------------------------

the problem is that the code that parses the slide delimiter line sets startofline to FALSE, even though it actually consumes the current line through its end and the next character that will be read actualy *is* the first of its line.