cb6047ea

By: Tom Sydney Kerckhove <syd@cs-syd.eu>

hopinion: check a repository against one rule

A tool that reads a repository's Haskell and reports what a rule has to say
about it, and a Nix layer that runs it as one check.

One rule ships: HsNoCustomShowRead, from haskell-style.md. It is here so that
the machinery is exercised by something real rather than by a demonstration.

The shape it is built on:

A rule says there is something to fix or there is not. There is no third
answer, because a checker whose silence can mean "I could not tell" is a
checker whose silence means nothing. A module that did not parse is a failure
of the run rather than a module with nothing wrong in it, and a module whose
source is a preprocessor's input is counted as unread.

A rule is a value in a registry rather than an instance of a class: no generic
code is polymorphic in a check type, so a record of functions is the right
encoding. Adding a rule is a module and a line in a list.

Producing a report always succeeds and only judging one fails, so a check that
found something still leaves the report where a person can read it.

Nothing here is built for a rule that does not exist yet. There is one level,
no fact store, no cross-process split and no reading of compiler output,
because one module-level rule needs none of them. The build enforces that:
-Wunused-packages, -Wunused-top-binds and weeder all run as errors, so
machinery that stops being used stops compiling.

What comes next, each with the rule that pays for it: suppressions and comment
attachment with the first comment rule, the fact store with the first package
rule, and the artifacts and the cross-process split with the first project
rule.

Timeline