John Wiegley
a627d885e8
Implement builtins.fetchurl
...
Squashed commit of the following:
commit 15b10d898e0457237f07cda9e5e9525bac0e95f6
Author: John Wiegley <johnw@newartisans.com>
Date: Wed May 2 16:33:30 2018 -0700
Update Exec.hs
commit d4a886dccf2715f1c1790e01adc242c352e7f427
Merge: 02afb27 4caacc113f3ebd555ce95431006f
2018-05-02 21:38:39 -07:00
hsloan
c1561ea711
Implement builtins.currentTime
2018-05-02 21:32:00 -07:00
John Wiegley
ee8ba02dfd
Much more work on type inference; report multiple possible types
...
This is needed because of a function like this:
x: y: x + y
This is polymorphic, but over exactly four possibilities:
int -> int
int -> float
float -> int
float -> float
The "type" is really the combination of the four, since we don't yet have a
mechanism like type classes, which could have rendered this as a single type:
(Num a, Num b) => (x :: a): (y :: b): x + y
Going this route will require that we manage an implicit type classes
hierarchy, and perform unification on constraints as well as types. In the
interim, I just lifted the unification algorithm into the LogicT monad, and
use back-tracking search to find all the possible types an expression could
be.
The main problem with using LogicT, however, is that there are many types
it *couldn't* be, and in the case of a unification failure, it not yet clear
what the type should have been. For example:
"foo" + 2
Should the string have been a float or an integer or a path? Or should the
integer have been a string? So for now we report all the possibilities, since
it's not obvious which part of the expression is in error:
hnix: Type error: TypeInferenceErrors
[ UnificationFail (TCon "integer") (TCon "string")
, UnificationFail (TCon "string") (TCon "path")
, UnificationFail (TCon "string") (TCon "float")
, UnificationFail (TCon "string") (TCon "integer")
]
This is a case where enumerating types rather than trying to make them compact
using type classes might actually be an improvement, since the errors here
would have been only slightly less numerous:
string != Num a => a
string != path
integer != string
Clearly a better reporting mechanism is needed to clarify these problems. I
can imagine that in an IDE, there would be a squiggly under both sides of the
expression, each suggesting the type that was expected for that argument under
the assumption that the other argument (the one not be inspected) was the
correct one.
2018-05-01 20:49:10 -04:00
John Wiegley
f662d44c56
Hindley-Milner is close to working, just the MonadFix problem again
2018-04-30 21:52:25 -04:00
John Wiegley
bda103bc11
Don't include ghc-datasize if --flags=profiling is set
2018-04-30 21:48:45 -04:00
John Wiegley
7f01cc8bb8
valueSize will not work with GHC 8.4 right now
...
ghc-heap-view hasn't been updated yet
2018-04-28 18:29:37 -07:00
John Wiegley
8d0a67b52c
Only turn on -Werror for Travis
2018-04-28 18:21:24 -07:00
John Wiegley
b24598b3a3
Add top-level __valueSize
...
Fixes #253
2018-04-28 17:12:32 -07:00
John Wiegley
ad18b950ac
Turn warnings into errors
2018-04-28 16:26:32 -07:00
John Wiegley
100c41db6d
Run the repl in the same MonadNix environment as Main.hs
...
Fixes #204
2018-04-28 16:26:32 -07:00
John Wiegley
05ca87a732
Several improvements to error reporting; add new --thunks option
2018-04-25 13:00:41 -07:00
John Wiegley
2de8c7cf84
More fixes to error reporting
2018-04-24 12:25:40 -07:00
John Wiegley
01bcb86d3f
Further improvements to error rendering, but not all done yet
2018-04-24 11:12:20 -07:00
John Wiegley
9864a8c7a5
Rework the way that info frames and error are handled
...
There is still work to be done to render the new "typed frames", and to
convert all the current string based frame into typed frames. This will pave
the way forward to smarter error messages that can be browsed in intelligent
environments like Emacs and the browser.
2018-04-24 02:15:28 -07:00
John Wiegley
a60181df3a
Add --flags=optimize for cabal configure
2018-04-22 14:33:38 -07:00
John Wiegley
96a0c387dd
Merge Nix.Trace and Nix.Reduce into Nix.Reduce
...
Since almost all of that code had to do with program reduction. addTracing and
tracingEvalExpr have moved into Exec.hs, since they depend on MonadNix.
2018-04-21 10:28:09 -07:00
John Wiegley
8a93068102
Rename Nix.Core back to Nix.Eval, since Nix.Eval was now almost empty
2018-04-21 10:11:34 -07:00
John Wiegley
86b09103d6
Move production of a reduced test case to --reduce, improve --trace
2018-04-20 22:36:40 -07:00
John Wiegley
47e8829a28
Rename Nix.StringOperations to Nix.Strings
2018-04-20 17:05:21 -07:00
John Wiegley
986a3043da
Add a constant-folding tree transformer: Reduce.reduceExpr
2018-04-20 02:57:23 -07:00
John Wiegley
0ff6c7a1e4
Move tracing code into its own file
2018-04-18 21:43:18 -07:00
John Wiegley
92217059b5
Add --flags=tracing, and document it in README.md
2018-04-18 16:28:21 -07:00
John Wiegley
8d25aa461d
Add --json option to hnix, which can render both expressions and values
2018-04-17 14:24:52 -07:00
John Wiegley
f4fbbe7903
Make NKeyName and NString instances of Hashable
2018-04-17 11:33:20 -07:00
John Wiegley
6c9cbf241a
Rewrite ToNix/FromNix, although Builtins.hs is not using it yet
2018-04-13 18:09:12 -07:00
John Wiegley
99189e9536
The basic REPL now works; doesn't allow establishing definitions yet
2018-04-12 16:46:34 -07:00
John Wiegley
709cc5247e
Add code for doing Hindley-Milner type inference, and a repl
...
This code was written by Stephen Diehl and Kwang Yul Seo, which they released
under an MIT license as the project poly_constraints:
https://github.com/kseo/poly_constraints
The code is added to hnix right now as mainly a placeholder, and will need to
be modified before it can be used.
2018-04-12 10:53:50 -07:00
John Wiegley
96df58fb29
Split off command-line options processing into a new Nix.Options module
2018-04-11 20:53:38 -07:00
John Wiegley
b621ffee0a
Move the caching support code into its own Nix.Cache module
2018-04-11 20:21:02 -07:00
John Wiegley
22175aa927
Use the serialise library on non-Linux systems, compact on Linux
2018-04-11 17:59:42 -07:00
John Wiegley
941d4f9fd3
Add testing code for writing out Nix parses as compact regions
2018-04-11 14:35:17 -07:00
John Wiegley
cd4aa33abd
Run the tests with threading enabled, for parallel execution
2018-04-11 13:53:55 -07:00
John Wiegley
74de14dc7d
Restore the linter, to some degree, at least for the language tests
2018-04-10 21:02:01 -07:00
John Wiegley
5efa46e516
Add Binary instances for NExpr
2018-04-10 18:57:17 -07:00
John Wiegley
dcd081d899
Add directory dependency for the tests
2018-04-10 12:38:14 -07:00
John Wiegley
a7e7e84795
Drop dependency on insert-ordered-containers
...
Fixes #161
2018-04-10 11:58:08 -07:00
John Wiegley
81c63a98d5
Comment out -fexpose-all-unfoldings while profiling
2018-04-10 09:02:59 -07:00
John Wiegley
8f37f37986
Down to 16 basic test failures with megaparsec
2018-04-09 21:35:46 -07:00
John Wiegley
9d7f25d368
All is building with Megaparsec now, but many parsing tests fail
2018-04-09 17:12:16 -07:00
John Wiegley
b1fb73d363
Remove Trifecta and parsec, begin switch to megaparsec
2018-04-09 10:16:25 -07:00
John Wiegley
4c2eebbe40
Refactor evaluation using a type class
2018-04-09 00:52:10 -07:00
John Wiegley
3681d51dd6
Create modules Nix.Monad.Lazy, Nix.Monad.Lint, Nix.Monad.Context
2018-04-07 23:14:08 -07:00
John Wiegley
9dcd7b37b6
Factor some code into new module Nix.Normal
2018-04-07 15:25:39 -07:00
John Wiegley
c8e9012778
Some minor fixups for compilation
2018-04-07 15:13:58 -07:00
John Wiegley
deb592b11d
Undo changes to hnix.cabal
2018-04-07 15:08:48 -07:00
Doug Beardsley
1b7dac071d
Test to catch a bug in builtins.match
...
Also added infrastructure for comparing against current nix behavior,
not just current nix tests.
2018-04-07 15:07:09 -07:00
John Wiegley
ea6a98f602
Implement builtins.tryEval
2018-04-07 14:33:15 -07:00
John Wiegley
935c7c5ee1
Remove some unnecessary code from package.yaml
2018-04-07 14:17:17 -07:00
John Wiegley
954ffccebb
Ensure in the tests that hnix.cabal is not edited directly
2018-04-07 12:09:49 -07:00
John Wiegley
443129315d
Use hpack and move all the Nix sources into a src/ directory
2018-04-07 12:02:33 -07:00
