Per-area coverage, and what a skip is not

sectiontally is read now, paired with section by identifier, and orbistoun probe reports each area of the platform with how much of it came out green.

The point is that one total cannot say what is understood. Ninety passes spread thinly across every area and ninety concentrated in one are the same number and completely different situations, and only the second means a subsystem can be relied on. The section's purpose line - the probe's own words about what it establishes - is carried rather than summarised.

On the real report:

areas     4 of 8 wholly green
  + 010-kernel           Kernel core            3 pass
    035-libc             C runtime              5 pass, 1 partial, 12 fail, 2 skip
    040-file             Filesystem             0 pass

Surprises

The area view tells the story of that run in three lines. Boot, kernel, memory and threads are wholly green; the C runtime is badly broken; and the filesystem section reports no passes at all - because the check that never concluded is inside it. A single total would have shown neither where the strength is nor where the run stopped.

A skip is not green, and saying so took a deliberate decision. A skip is a check that did not run, so the section did not establish what it claims to. Rounding one up is how a subsystem gets relied on for something nobody tested - the same error as reading a death as a return value, one level up. It stays in the denominator.

A half-reported section still appears. A section with no tally, and a tally naming no section, are incomplete reports rather than absent ones. Dropping either would shrink the denominator, which flatters the result in the one direction nobody should be flattered.

Two clippy rules asked for real improvements rather than appeasement this time. parse_record had grown past a hundred lines, and the right split was to lift out the outcome parsing - which carries the rule this whole crate is shaped around, so a rule worth stating turned out to be worth being able to find. And pass is now printed even at zero, because a section reporting no passes is the interesting case and omitting the number would make it look like a section with no checks rather than one where nothing worked.