Writing GNU Emacs Extensions

Changing the Definition of a Paragraph

First, we must arrange for a line consisting of %% not to be considered part of a paragraph. This means we must change the variable paragraph-separate, whose value is a regexp that describes lines that separate paragraphs. We'll also have to change paragraph-start, a regexp that describes lines that serve as either the first line of a paragraph or (despite the name) as a line that separates paragraphs.^[35]

Emacs uses the regexps in paragraph-start and paragraph-separate to match at the beginnings of lines, even though the regexps do not explicitly begin with the magic ^ ("match at the beginning of a line") character.

In Text mode, the value of paragraph-start is "[ \t\n\^L]", which means that if a line starts with a space, tab, newline,^[36] or Control-L (the ASCII "formfeed" character), it's either the first line of a paragraph or a line that separates paragraphs.

Text mode's value for paragraph-separate is "[ \t\^L]*$", which means that a line containing zero or more spaces, tabs, or formfeeds, or some combination thereof, and nothing else, is not part of any paragraph.

What we'd like to do is augment these patterns to say "a line containing %% is a paragraph separator, too."

The first step is to make these variables have separate values for the current buffer when in Quip mode. (That is, setting these variables, which are global, should not affect other buffers that may not be in Quip mode.) Therefore, in addition to the basic skeleton described ...