Parsers likely contain list span mistakes

[raehik]

#276 fixes what appears to be a mistake in the F77 parser, where a span isn't set properly: a list is reversed before being placed in the AST, but the span uses the yet-reversed list. The problematic call was getTransSpan:

fortran-src/src/Language/Fortran/Util/Position.hs

Lines 160 to 163 in 6388025

	instance {-# OVERLAPPABLE #-} (Spanned a, Spanned b) => SpannedPair a b where
	getTransSpan x y = SrcSpan l1 l2'
	where SrcSpan l1 _ = getSpan x
	SrcSpan _ l2' = getSpan y

which calls getSpan on a list:

fortran-src/src/Language/Fortran/Util/Position.hs

Lines 105 to 109 in 6388025

	instance (Spanned a) => Spanned [a] where
	getSpan [] = error "Trying to find how long an empty list spans for."
	getSpan [x] = getSpan x
	getSpan (x:xs) = getTransSpan x (last xs)
	setSpan _ _ = error "Cannot set span to an array"

And following the recursive getTransSpan, we see that by calling this on a reversed list, we accidentally span to the end of the first element instead of the last. Spanning requires things to go from left to right, it doesn't re-order automagically.

I noted another likely mistake in the F2003 parser. This probably happens a handful of times in the parsers.