| Age | Commit message (Collapse) | Author |
|---|
|
I don't care about ghcjs anymore, the most javascript I want to do is jQuery.
|
|
I had thought that I needed some unmerged commits to nixpkgs that enable Guile
3.0, but turns out I just needed to use a different target name :P
|
|
asdf seems to always be necessary for any other packages to work, so I just
include that in the call to nix-shell, and swank is included because it's just
useful to start a repl server.
|
|
I ended up deleting miso, and consequently all files under Hero/ and Miso/,
because I couldn't get miso to build with GHC 9.2.
Other things:
- Niv has been wrapped by Biz/Bild/Deps.hs, so I can extend it to my liking.
- Apply-refact is gone because I couldn't get it to build.
- Disabled python stuff.
|
|
Support for packages and third-party imports will need to come later once I
figure out how to lookup rust packages by their import statements. Until then,
this works to compile "hello world".
|
|
|
|
|
|
|
|
|
|
|
|
This is basically building on top of bild's analysis, which is great becuase it
means that bild is becoming a useful tool rather than a monolithic do-everything
job runner. The eventual goal is for the bild analysis to be much more useful,
maybe even provide data to remote repls or language-server or whatever.
|
|
|
|
This means my call to ghc-pkg will look at the full package set from Hoogle.
|
|
|
|
|
|
Still need to add this to the CLI, and there should be other features like
delete and so on, but this works for now.
|
|
|
|
|
|
|
|
This parses the files contents for imports, then uses ghc-pkg to lookup the
package that provides the module. Now I can do that analysis in Haskell instead
of nix, which is much easier to code with.
|
|
|
|
Getting me closer to the latest GHC. This release also includes my own packages
that I submitted some time ago.
GHCJS is not present in 21.05 for some reason, but I think it's back in master,
so I might do another upgrade soon, but for now I just disabled my GHCJS
support. I'm not really using it anyway.
I also had to bring it string-quote, update nixos-mailserver, and a few other
things.
|
|
|
|
This allows us to use nix-build as a check that bild is working.
I think bild sometimes hangs because it doesn't get input from nix-build? I'm
not sure, but one workaround is to run nix-build on the target, and then bild
will just get the cached build.
|
|
|
|
|
|
This is supposed to be how to cleanup the database and any other local files.
Should only be used before/after test, so maybe I can find a way to enforce this
constraint somehow in the code.
|
|
|
|
|
|
Also improved the test situation, did some refactors, and now listing the user's
past analyses on their account page.
|
|
- structured log messages
- compact logs rewrite the line on bild completion
- using conduit for streaming output (although I think this isn't working quite
right)
- new Proc type for running subprocesses
- general code cleanup and refactoring
|
|
|
|
I really could use those structured logs.
|
|
|
|
This makes scanning logs *much* easier. I figure keep it as simple as possible,
just red, yellow, and green.
I also added two spaces between labels in the log messages. It would be nice to
have a more structured logging system, but for now this works.
|
|
This argument will run the tests for an output after building. It's active in
'ci' so running that will ensure tests are passing. This way testing a
namespace and building a namespace are as close together as possible, so
presumably it will be that much easier to write good tests.
|
|
Also changes the --test option to a 'test' command. This is because running the
tests for a namespace/exe should never be combined with anything else: you
either want to run the tests, or not.
|
|
Wraps docopt rather nicely. It's much nicer than optparse-applicative and runs
tests with the --test argument automatically. Next I just need to implement a
test framework.
|
|
Now bild knows how to determine between modules that require ghcjs and ghc. It
also knows what *not* to build, meaning it won't try to build non-buildable nix
targets, for example (unfortunately this is just hardcoded for now), but it also
won't build scm or py targets that I haven't implemented yet. It just silently
fails, which is fine, because it means I can do `bild **/*` and everything just
works.
Of course, if I want to build scm code then I will have to implement that, but
that's not a priority right now.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Move them under the Biz root so that we know they are specific to Biz stuff. Biz
is for proprietary stuff that we own.
I also had to refactor the bild namespace parsing code because it couldn't
handle a namespace with 3 parts. I really need to get that namespace library
written and tested.
|
|
|
|
|
|
|
|
Also had to capitalize some stuff, and move some nix files around and rename
the metadata directive from 'exe' to 'out' because that just makes more sense,
and fix some compiler errors. But now bild treats both nix and hs files as
buildable things. So that's cool.
One interesting example is Biz/Pie.{nix,hs} - I can either create a dev build of
the hs file with ghc, or I can create a fully-encapsulated nix build. Its nice
to have both options because a dev build with ghc takes half the amount of time,
and I can rely on my locally cached hi and ho files. I think this shows the
power of bild, but also can be a somewhat subtle thing.
The issue really is with the separate command calls in nix builds vs dev builds.
I figure there are a few ways to fix this:
1. Try to use bild inside the nix rules. That could be interesting, but could
also lead to some weird behavior or worm holes forming.
2. Extract the command line invocation into a separate file, some kind of
really simple template that gets pulled into both programs.
It is important to consider that in the future I might want to have bild do a
module-by-module nix build of programs, but I'm not sure how that would effect
my choice here.
|
