This patch does several things.
1. For the crate `wasmer-c-api`, the library name is modified from
`wasmer_c_api` to `wasmer` in `Cargo.toml`. That way, the new
library files are named `libwasmer.*` rather than
`libwasmer_c_api.*`. That's the primaly goal of this patch. The
rest is a consequence of this point. Why do we want that? Because
the `build.rs` script of the `wasmer-c-api` crate will configure
the `soname` (on Linux), the `install_name` + `current_version` +
`compatibility_version` (on macOS), and the `out-implib` +
`output-def` (on Windows) for a library named `libwasmer`, which is
the name we provide in the Wasmer package for the Wasmer
libraries. If we want everything to be testable, we cannot use
`libwasmer` in `soname` for a file named `libwasmer_c_api` for
example. If we rename the file when packaging (as it's done prior
this patch), we would need to re-update all those information in
the `Makefile`. It implies to duplicate the code in 2 places. So
let's do things right directly and only once: We want the library
to be named `libwasmer`, let's do that.
2. For the crate `wasmer-c-api`, since the library name has been
renamed to `wasmer`, it creates a conflict with the `wasmer` crate
which is a dependency. Consequently, this patch also updates the
`Cargo.toml` file to modifiy the dependency name with the special
`package` attribute (see
https://doc.rust-lang.org/cargo/reference/specifying-dependencies.html#renaming-dependencies-in-cargotoml
to learn more). So now, the `wasmer` refers to the `wasmer_c_api`
crate, and `wasmer-api` refers to the `wasmer` crate.
3. The code of the `wasmer-c-api` crate is updated accordingly. The
`WasmerEnv` derive procedural macro fails because it expects a
crate named `wasmer` (which is now `wasmer_api`), so we implement
the `WasmerEnv` trait by hand.
4. The patch updates the `build.rs` script of the `wasmer-c-api`
crate:
1. In the `build_cdylib_link_arg` function: The dependency to the
`cdylib-link-lines` crate has been removed because the output is
not exactly the one we expect. So we compute all the
`cargo:rustc-cdylib-link-arg=…` lines by hand. The version number
no longer appears in the library file name for example.
2. In the `build_inline_c_env_vars` function: Values passed to
`LDFLAGS` have been updated to be `libwasmer` rather than
`libwasmer_c_api`.
3. A new `shared_object_dir` helper function has been created
because it's used in `build_inline_c_env_vars` and in
`build_cdylib_link_arg`.
5. The `Makefile` has been updated:
1. In `package-capi`, we no longer rename `libwasmer_c_api` to
`libwasmer` since the name is correctly defined since the
beginning now.
Calling `install_name_tool` on macOS is no longer required since
`install_name` is correctly set by the linker in the `build.rs`
script of `wasmer-c-api`.
2. In `package-docs`, some stuffs have been fixed, like the
`SOURCE_VERSION` variable that didn't exist, so removed, or the
`mkdir` command that was incorrect etc.
3. In `build-docs`, the `wasmer-c-api` crate is excluded from the
list of crates to generate the documentation for. Mostly because
the `build-docs-capi` recipe exists, and we must use it to
generate the documentation of `wasmer-c-api` now.
4. In `build-docs-capi`, we generate the documentation for the
`wasmer-c-api` crate. But `rustdoc` uses the library name for the
directory name in the `target/doc/` directory. Since the library
name is now `wasmer`, it creates a conflict with the `wasmer`
crate. Consequently, we update the library name by using `sed` on
the `Cargo.toml` file before running `cargo doc`, to finally
restore `Cargo.toml` as it was previously.
`wasm_trap_new` expects a `wasm_message_t`. It's a type alias to
`wasm_name_t` with the exception that it represents a null-terminated
string.
When calling `wasm_trap_new`, no check was present to ensure the
string was well-formed. That's a first issue. But in the best
scenario, the string was correctly formed and was
null-terminated. This string was transformed to a Rust `String` —with
the null byte!— and passed to `RuntimeError`.
Then in `wasm_trap_message`, another null byte was pushed at the end
of the message. It's been introduced in
https://github.com/wasmerio/wasmer/pull/1947. It results in a
doubly-null-terminated string, which is incorrect.
This patch does the following:
1. It checks that the string given to `wasm_trap_new` contains a
null-terminated string or not, and will act accordingly. Note that
it's possible to pass a non-null-terminated string, and it will
still work because this detail is vicious. The idea is to get a
well-formed `RuntimeError` in anycase.
* If no null byte is found, the string is passed to `RuntimeError`
as a valid Rust string,
* If a null byte is found at the end of the string, a new string is
passed to `RuntimeError` but without the final null byte,
* If a null byte is found but not at the end, it's considered as an
error,
* If the string contains invalid UTF-8 bytes, it's considered as an
error.
2. It updates `wasm_trap_message` to always add a null byte at the end
of the returned owned string.
3. It adds test cases when passing a null-terminated or a
non-null-terminated string to `wasm_trap_new` and to compare the
results to `wasm_trap_message`.
1. Use single tabs rather than three, it fits in a small terminal
window,
2. `CAPI_WASMER_TESTS` has been removed in
https://github.com/wasmerio/wasmer/pull/2375, but partially. This
patch finishes the work.
3. The `test-%: %.o` rule is a remainder of some older code, it's no
longer needed.
Until this patch, our C API comes in 2 flavors: `deprecated` and
`wasm_c_api`. With the coming 2.x version of Wasmer, we would like to
remove the `deprecated` API, and keep the `wasm_c_api` only.
This patch removes the `deprecated` API from the `wasmer-c-api`
crate. It also cleans up the `Makefile` and the documentation
system. Previously, the documentation for the `deprecated` API was
relying on Doxygen, which was one new dependency the user had to
install. For the `wasm_c_api`, it relies on `rustdoc`, which is way
better because all examples are run and tested as part of our test
suite.
This clean up also removes the need to deal with `system-libffi` both
in the crate itself and in the `Makefile`, which was an edge case for
macOS on aarch64, and a needle in the foot for some of our users.
Finally, the `build.rs` is now simplified because we no longer need to
exclude symbols from one header to another. It also means that we only
provide the `wasmer_wasm.h` header file now; the `wasmer.h` and
`wasmer.hh` headers are removed.
2364: feat: Rename `wasmer-engine-object-file` to `wasmer-engine-staticlib`. r=Hywan a=Hywan
# Description
This PR renames the `wasmer-engine-object-file` crate into `wasmer-engine-staticlib`. We believe it conveys a better meaning of what it does. This PR is similar to its siblings, #2340 and #2356.
Co-authored-by: Ivan Enderlin <ivan@mnt.io>
2161: feat(llvm): Make NaN canonicalization configurable r=jubianchi a=jubianchi
compiler-llvm now uses the experimental.constrained intrinsics to ensure
correct behavior on FP operations when full-canonicalization is
disabled.
This patch requires TheDan64/inkwell#247
# Review
- [ ] Add a short description of the the change to the CHANGELOG.md file
Co-authored-by: jubianchi <julien@wasmer.io>
compiler-llvm now uses the experimental.constrained intrinsics to ensure
correct behavior on FP operations when full-canonicalization is
disabled.
This patch requires TheDan64/inkwell#247
2092: Add reference types r=MarkMcCaskey a=MarkMcCaskey
This PR adds support for reference types in the API and in our compilers.
Wasm C API support should probably be a separate PR after this one lands.
# Status
- [x] spectests passing
- [x] compiler-cranelift
- [x] compiler-singlepass
- [x] compiler-llvm
- [x] Tables work with references
- [x] Globals work with references
- [x] FuncRefs for local functions
- [x] FuncRefs for imported functions
- [ ] ExternRefs
- [x] Work as values in Wasm
- [x] Tests showing it being passed to/from host functions
- [x] Top level API for interacting with ExternRefs
- [ ] Reference counting works as expected, fully tested
- [ ] ~~Wasm C API support (should be done on a PR branched off of this branch)~~ (not going to do this in this PR)
- [x] ~Host info (see Wasm C API and wasmer Rust API)~ (not going to do this in this PR. The value of this feature seems low and the cost seems high; unclear if we even want it at the Rust API level)
# Review
<!--
## High priority things to review
- [ ] ExternRef reference counting behavior (blocked on implementation, but feedback of what's there is welcome)
- [ ] That we're counting in all the places that matter (we're not yet)
- [ ] The manual `ref_clone` and `ref_drop` in order to make putting `VMExternRef` in a `union` work.
- [ ] FuncRef storage
- [ ] Distinction between imported and local FuncRef
- [x] Duplicating data rather than storing it inline in the dynamic part of the `Instance`
- [ ] Changes to `wasmer_types::Value`
- [ ] ExternRef
- [ ] FuncRef / ValueEnumType
- [ ] Passing a store via `&dyn Any` to some methods and the garbage placeholder values we pass when we don't have a store.
- [ ] The location of the `Option` for `FuncRef`
- [x] Implementation of new instructions (bug check)
- [ ] API changes
- [ ] Changes in the wasmer crate
- [ ] Changes in the core traits for example `vm::Table`
- [ ] Use of `repr(C)` union `vm::table::TableElement` in `extern "C"` / FFI contexts. I just assumed `repr(C)` on a union of all usize sized things would just be passed like a usize.
- [ ] WAST test runner updates
## Lower priority things to review
These are things that I think we can fix with follow-up PRs. Let me know if you think anything on this list is too important to fix later.
- [ ] Implementation of new instructions (calling functions rather than doing it inline)
- [ ] Use of libcalls distinguishing between imported vs local in cases where it probably doesn't matter
-->
# misc
- [ ] Add a short description of the the change to the CHANGELOG.md file
Co-authored-by: Mark McCaskey <mark@wasmer.io>
Co-authored-by: Nick Lewycky <nick@wasmer.io>
2210: Fix memory leak in C API in some uses of `wasm_name_t` r=MarkMcCaskey a=MarkMcCaskey
This PR fixes a bug relating to `wasm_importtype_t` leaking memory, particularly on the path where `String` is converted into `wasm_name_t`. It fixes this by introducing a new type, `owned_wasm_name_t`, which is identical to `wasm_name_t` but has RAII and calls the destructor if it goes out of scope.
I'm not confident that the use of `owned_wasm_name_t` on the FFI boundary is correct though, we'll have to thoroughly review that before shipping this PR.
# Review
- [x] Add a short description of the the change to the CHANGELOG.md file
Co-authored-by: Mark McCaskey <mark@wasmer.io>
