awakesecurity / language-ninja Goto Github PK

Title is self-explanatory

This is especially useful once #12 is done.

The title says it all.

We should remove IR.ninjaPhonys :: HashMap Target (HashSet Target) and replace it by switching IR.buildRule :: Rule to Maybe Rule.

It would probably be good to throw an error (or display a warning) if a parsed Ninja file uses a feature inconsistent with its ninja_required_version variable.

The Cabal file says this library has only been tested with GHC 8.0, which was a long time ago! Additionally, many dependencies (e.g. megaparsec) have jumped a few major versions.

• Automatic release workflow
  • One command that will git tag -s release-<version> -m "Version <version>", git push --tags, and make a request against the GitHub API to add data from the changelog to the GitHub release.
• Linter for @since attributes
  • Currently we check that everything has an @since attribute, but it would also be nice to warn the maintainer if an @since attribute has not been bumped even though the underlying definition has changed. Of course, there will be false positives due to changes that are backwards-compatible, but this is better than nothing.
• Public Hydra jobset

The lexer/parser/types should be updated for compatibility with implicit outputs, added in Ninja 1.7 and documented here.

• Update Lexeme to support implicit outputs
• Update AST.Ninja to support implicit outputs
• Add implicit output support to parser
• Create tests for implicit outputs

The current lexer/parser is well-optimized, but it is unclear that it is correct. We should have another parser, written using megaparsec, that is more obviously correct and easier to test. This also requires adding polymorphic annotation fields to all the datatypes. For example, this datatype:

data Foo = MkFoo { _fooBar :: Bar, _fooBaz :: Baz }

would become this datatype:

data Foo ann = MkFoo { _fooAnn :: ann, _fooBar :: Bar, _fooBaz :: Baz }
instance Annotated Foo where { ... }

where Annotated is defined by:

class (Functor t) => Annotated (t :: * -> *) where
  annotation :: Lens' (t ann) ann

The old, fast parser would then have type Text -> Parser (PNinja ()), whereas the new megaparsec parser would have type Text -> Parser (PNinja (Maybe Pos)), for example. This would allow pretty-printing such that the output is identical to the parsed input, which is important for diagnostics, testing, and formatting tools.

I'm installing on Ubuntu 14.04 but it seems like Cabal doesn't like any of the available versions of Haskell for the requirement even though the versions appear to fall within the range. What am I doing wrong?

$ cabal install
Warning: language-ninja.cabal: Ignoring unknown section type: custom-setup
Resolving dependencies...
cabal: Could not resolve dependencies:
trying: language-ninja-0.1.0
rejecting: base-4.6.0.1/installed-8aa... (conflict: language-ninja =>
base>=4.8 && <5.0)
rejecting: base-4.10.0.0, 4.9.1.0, 4.9.0.0, 4.8.2.0, 4.8.1.0, 4.8.0.0,
4.7.0.2, 4.7.0.1, 4.7.0.0, 4.6.0.1, 4.6.0.0, 4.5.1.0, 4.5.0.0, 4.4.1.0,
4.4.0.0, 4.3.1.0, 4.3.0.0, 4.2.0.2, 4.2.0.1, 4.2.0.0, 4.1.0.0, 4.0.0.0,
3.0.3.2, 3.0.3.1 (global constraint requires installed instance)

The title says it all.

Release checklist:

• Library:
  • Address all FIXME comments.
  • Full documentation.
  • Add @since annotations to all public functions.
  • Get the code to be clean with respect to -Weverything.
  • Remove unneeded dependencies.
• Tests:
  • Set up doctest.
  • Test all optics with tasty-lens.
  • Remove unneeded dependencies.
• Executables:
  • Move ninja2nix into a separate project.
  • Write executables:
    • ninja-lex
    • ninja-parse
    • ninja-evaluate
  • Proper command-line interface for any executables.
  • Remove unneeded dependencies.
• Blockers:
  • smallcheck-lens and tasty-lens need base version bumps.
    • jdnavarro/smallcheck-lens#4
    • jdnavarro/tasty-lens#1
• Miscellaneous:
  • Add README.md et al. to extra-*-files
  • Add .github/CONTRIBUTING.md
  • Add changelog section to Cabal file.
  • Signed git tag (non-lightweight) at the appropriate version.

Control.Exception.Lens provides a nice approach for modeling exception subtyping using prisms. Instead of an ether type like this:

(MonadExcept IOException m, MonadException CompileError m) => ...

... you would have:

(AsIOException e, AsCompileError e, MonadException e m) => ...

See this thread for more details:

https://www.reddit.com/r/haskell/comments/2utxua/prisms_and_monaderrors/?st=j4aqwbyx&sh=1c57491b

... and:

https://hackage.haskell.org/package/lens-4.15.3/docs/Control-Exception-Lens.html