Jeroen van Rijn
Jeroen van Rijn
(It's fine if those are two separate PRs, if you prefer. One for the bindings, one for better complex support in `core:math`.)
I'll let @gingerBill decide whether he wants the parser to emit a warning on ignoring a tag.
> Not sure about the usefulness of `expect_not` as you can always write `expect(, !value)`, can't you? > > You get my full support for `likely` and `unlikely`. No, `!value`...
Maybe add `likely` and `unlikely` as builtins (so you don't have to import `intrinsics`) and add a deprecation tag to `intrinsics.expect`.
 Fixes were made yesterday. Do you still have these issues on HEAD?
I consider it closed provisionally. Will test it tomorrow with `tests\core\math\big`'s library for the Linux CI. But it's promising.
> Should I close this? I'll close it tomorrow after I verify it successfully builds and runs the big maths tests on Linux.
I'd prefer `odin test -build-only` to splitting it up into two verbs, `odin test-build` & `odin test-run`. Too much noise for a rare use case.
And while we're at it, we could add a `odin test -keep-test-executable` to keep the executable after running the test, with it being deleted by default.
Another victim: Commit 0db86a0638d417583a042ea7b3c2e5304d653a17 exposed a bug in DLL global. When compiling `core:math/big` as part of an EXE, `initialize_constants();` gets called from the global scope, correctly setting globals including `MUL_KARATSUBA_CUTOFF`....