The MC documentation is wrong
Posted: July 7th, 2021, 10:07 pm
The documentation on the MC format states that '.'s at the end of a row are suppressed, but doesn't say anything about '$'s at the end of a line being suppressed. It also gives "$$..*$...*$.***$$$$" as an example node, which clearly implies that they are not, and that an MC decoder can safely assume that 2-state leaf lines always contain 8 $s.
Maybe that sounds nitpicky, but I just uncovered a bug in my program which arose from assuming that all 8 $s are included. (The relevant line in Golly's code, if you're curious, seems to be this. Notice the early exit condition.)
The documentation should state that $s can be and are suppressed. It should also clarify whether decoders can assume that there's at least one redundant $ at the end (which I'm still not sure about myself), and whether encoders are obligated to provide them.
Maybe that sounds nitpicky, but I just uncovered a bug in my program which arose from assuming that all 8 $s are included. (The relevant line in Golly's code, if you're curious, seems to be this. Notice the early exit condition.)
The documentation should state that $s can be and are suppressed. It should also clarify whether decoders can assume that there's at least one redundant $ at the end (which I'm still not sure about myself), and whether encoders are obligated to provide them.