Run aya integration tests #69
Conversation
tamird
force-pushed
the
test-aya-integration
branch
16 times, most recently
from
f893927 to
2a0dc4a
Compare
|
@alessandrod @ajwerner this is ready for review (see aya-rs/aya#641 for changes required to make this pass tests) |
tamird
force-pushed
the
test-aya-integration
branch
from
f16a830 to
2a0dc4a
Compare
src/linker.rs
Outdated
| @@ -313,7 +313,7 @@ impl Linker { | |||
| in_type: Option<InputType>, | |||
| ) -> Result<(), LinkerError> { | |||
| let mut data = Vec::new(); | |||
| reader | |||
| let _: usize = reader | |||
There was a problem hiding this comment.
do we need the type annotation here?
src/llvm/mod.rs
Outdated
| @@ -220,7 +220,7 @@ pub unsafe fn optimize( | |||
| } | |||
|
|
|||
| debug!("running function passes"); | |||
| LLVMInitializeFunctionPassManager(fpm); | |||
| let _modified = LLVMInitializeFunctionPassManager(fpm) != 0; | |||
There was a problem hiding this comment.
what is the advantage of doing this? the compiler isn't complaining about it, we never check it, the code seems worse to read to me after the change
src/lib.rs
Outdated
| @@ -1,4 +1,5 @@ | |||
| #![deny(clippy::all)] | |||
| #![deny(unused_results)] | |||
There was a problem hiding this comment.
Oh, this. But why? Did you run into some LLVM stuff that's supposed to be #[must_use]? If so let's annotate that instead?
There was a problem hiding this comment.
Nah, this doesn't really belong in this PR specifically. It's a useful lint that avoids things like partial reads and writes which report the number of bytes and folks often forget to check; see src/linker.rs where I changed read to read_exact because the length read wasn't being checked -- this lint surfaced that.
I'll move this to another PR so we can discuss separately from this change.
Fixes #65.
tamird
force-pushed
the
test-aya-integration
branch
from
2a0dc4a to
3418abc
Compare
Fixes #65.