The recent version of the rust toolchain will emit ref types opcodes,
which needs to enable this feature in the `iwasm` build.
The vector format parsing logic has some errors in the current version.
I disabled the check for now and am waiting for further investigation.
As suggested in #3829, in release CI, we add zip/tar.gz artifacts named
iwasm-gc-eh-{version}-{platform} for `iwasm` which supports features
garbage collection and exception handling(classic interpreter only).
Also, add a command line option to control GC heap size for `iwasm` on
the Windows platform.
- For Windows, llvm libs need to cache more directories, so use a multi-line
environment variable for paths
- Remove conditionally build directories `win32build`, just use `build` for all platform
- Add Windows wamrc and iwasm(disable lib pthread semaphore and fast jit for now)
build in release CI
Add table64 extension(in Memory64 proposal) support in classic-interp
and AOT running modes, currently still use uint32 to represent table's
initial and maximum size to keep AOT ABI unchanged.
In the AOT compiler, allow the user to control stack boundary check when the boundary
check is enabled (e.g. `wamrc --bounds-checks=1`). Now the code logic is:
1. When `--stack-bounds-checks` is not set, it will be the same value as `--bounds-checks`.
2. When `--stack-bounds-checks` is set, it will be the option value no matter what the
status of `--bounds-checks` is.
In classic interpreter, fast interpreter and fast-jit running modes, set the local
variables' default value to NULL_REF (0xFFFFFFFF) rather than 0 if they are type
of externref or funcref.
The issue was reported in #3390 and #3391.
- Move CodeQL scripts to the scripts directory
- Only report error in CI if it's a CodeQL reported issue and was not dismissed
before and is likely to be an actual error
In the release CI and related scripts, when comparing and printing to the CI,
use the most recent **ancestor** tag(s) for the release branch rather than
the most recent one(s).
And fix the build_wamr_sdk.yml and build_wamr_lldb.yml CIs.
Add GC, exception handling and memory64 to README.md post MVP feature list,
and update build_wamr.md for how to build them.
And remove `Non-trapping float-to-int conversions`, `Sign-extension operators`,
`Multi-value` links in README since they are in wasm MVP and very common now.
This PR fixes the random failing test case `nofollow_errors` mentioned in
https://github.com/bytecodealliance/wasm-micro-runtime/issues/3222
```C
// dirfd: This is the file descriptor of the directory relative to which the pathname is interpreted.
int openat(int dirfd, const char *pathname, int flags, ...);
```
The value should be a directory handle instead of a file handle (which is always -1 in this context)
returned from `openat`.
- Merge unused field `used_to_be_wasi_ctx` in `AOTModuleInstance` into `reserved` area
- Add field `memory_lock` in `WASMMemoryInstance` for future refactor
- Go binding: fix type error
https://github.com/bytecodealliance/wasm-micro-runtime/issues/3220
- Python binding:
type annotation uses the union operator "|", which requires Python version >=3.10
- Temporarily disable the deque test for the VS Code extension, as the
Rust formatter seems to malfunction after a recent VS Code update.
- Add configuration for iwasm host managed heap size, allowing users
to flexibly `malloc` memory. This also fixes the current bug that when
default size is 0, it can't run and debug.
- Apply coding style formatting for WAMR IDE source code and add a
format check for it in CI.
- Update document and some screenshots.
Using `CHECK_BULK_MEMORY_OVERFLOW(addr + offset, n, maddr)` to do the
boundary check may encounter integer overflow in `addr + offset`, change to
use `CHECK_MEMORY_OVERFLOW(n)` instead, which converts `addr` and `offset`
to uint64 first and then add them to avoid integer overflow.
For shared memory, runtime should get the memories pointer from
module_inst first, then get memory instance from memories array,
and then get the fields of the memory instance.
The LLVM zext IR may be inserted after the terminator of a basic block
when popping the arguments of a wasm block. Change to insert the
zext IR before the terminator of the basic block to resolve the issue.
Reported in #2620.
Unaligned store v128 value to the AOT function argument of the pointer for
the extra return value may cause segmentation fault.
Fix the issue reported in #2556.
- Provide a Dockerfile.old to fix issue of ESP32 custom linker scripts not working
properly with the newer version of Zephyr, as reported in #2263
- Provide a Dockerfile with newer Zephyr for other boards
- Update the corresponding document
Change main thread hangs when encounter debugger encounters error to
main thread exits when debugger encounters error
Change main thread blocks when debugger detaches to
main thread continues executing when debugger detaches, and main thread
exits normally when finishing executing
Allow to wait for a new debugger connection once the previous one
is disconnected:
- when receiving a detach command
- when the client socket is closed (for example, lldb process is killed)
