Not true at all. Many people have trouble tracking long lines, partially because eyes are jumpy (not scrolling). Some people have this very seriously, but are otherwise fine (far as humans go).

This is also something testable. There's empirical research. According to this 

"An empirical demonstration carried out by Morrison and Rayner (1981) confirmed that saccade size is consistent in terms of number of characters, and not visual angle"

meaning, you plop your eyes a certain amount of characters - not angle. You read a certain amount of chars per peek, and variable width might confuse you (maybe).

"If the lines are too short, readers cannot make use of much information in each fixation. If line lengths are too long, the return sweeps to the beginning of the next line are difficult."

Pretty straightforward/as expected

"(..) 2.2A + 21 ms (Carpenter 1977), which means that the greater number of return sweeps with shorter lines will add more to the time than longer lines."

Not sure about the unit of A (amplitude?) but it means you take 21 ms, plus some measure of distance, to move your eyes from x to y.

The authors seem to be fans of the idea that reading speed simply improves from longer lines, up to about 130 characters, after which the line becomes unwieldy. Some in-print research showed 70 or 52 (?) characters to be the ideal width. I find it amusing that these relate roughly to coders' standards of 60, 80 or 120 characters.

Ultimately, a bit offtopic and I apologize.

Postultimate, write a preprocessor if you hate malformatted text so much?