Text copied out of a PDF or an email arrives broken at every line. This rejoins it into flowing paragraphs, and keeps the paragraph breaks that actually mean something.
A PDF, a plain-text email and an old-style code comment all store a line ending at the end of every visual line, because they were laid out for a fixed column width. Your browser does not know those endings were decorative. So a paragraph that looked fine in the source arrives in your text box broken after roughly every seventy characters, and it will stay broken when you publish it. The distinction that matters is between a line break that was a layout artefact and one the author meant. The blank line between two paragraphs. Strip both and you get a wall of text; strip neither and you get a ragged column.
Leave the paragraph toggle on and the tool treats a blank line as a real boundary: the wrapped lines inside each paragraph are joined, and the paragraph divisions survive. Turn it off and everything collapses to a single line, which is what you want for a bio field or a spreadsheet cell that has to hold one value. Each line is trimmed before joining, so the double spaces that appear where an indent used to be do not survive either. If the source used a hyphen to split a word across two lines, that hyphen is still there afterwards, no tool can reliably tell a line-break hyphen from a real one in a compound word.
Not with the paragraph toggle on. A blank line between two blocks is treated as a real break and kept. Turn the toggle off only when you want the whole thing on one line.
There should not be. Every line is trimmed before it is joined, so trailing spaces at the end of a wrapped line are dropped rather than added to the join.
No. Rejoining "inter-" and "national" correctly would need a dictionary, and it would wrongly merge genuine compounds like "well-known". The hyphen is left exactly as written.