docs: update README and Cargo.toml for build alignment and new features

Author: ryancinsight

snmalloc-rs/Cargo.toml

@@ -34,3 +34,9 @@ notls = ["snmalloc-sys/notls"]
 stats = ["snmalloc-sys/stats"]
 usewait-on-address = ["snmalloc-sys/usewait-on-address"]
 libc-api = ["snmalloc-sys/libc-api"]
+tracing = ["snmalloc-sys/tracing"]
+fuzzing = ["snmalloc-sys/fuzzing"]
+vendored-stl = ["snmalloc-sys/vendored-stl"]
+check-loads = ["snmalloc-sys/check-loads"]
+pageid = ["snmalloc-sys/pageid"]
+gwp-asan = ["snmalloc-sys/gwp-asan"]

snmalloc-rs/README.md

@@ -18,20 +18,14 @@ a global allocator for rust. snmalloc is a research allocator. Its key design fe
 Some old benchmark results are available in
 the [`snmalloc` paper](https://github.com/microsoft/snmalloc/blob/master/snmalloc.pdf). Some recent benchmark results
 are listed at
-[bench_suite](https://github.com/SchrodingerZhu/bench_suite). There are three features defined in this crate:
+[bench_suite](https://github.com/SchrodingerZhu/bench_suite). There are following features defined in this crate:
 
-- `debug`: Enable the `Debug` mode in `snmalloc`.
-- ~~`1mib`: Use the `1mib` chunk configuration. From `0.2.17`, this is set as a default feature~~ (removed since 0.3.0)
-- ~~`16mib`: Use the `16mib` chunk configuration.~~ (removed since 0.3.0)
-- ~~`cache-friendly`: Make the allocator more cache friendly (setting `CACHE_FRIENDLY_OFFSET` to `64` in building the
-  library).~~ (removed since 0.3.0)
+- `debug`: Enable the `Debug` mode in `snmalloc`. This is also automatically enabled if Cargo's `DEBUG` environment variable is set to `true`.
 - `native-cpu`: Optimize `snmalloc` for the native CPU of the host machine. (this is not a default behavior
   since `0.2.14`)
 - `qemu`: Workaround `madvise` problem of QEMU environment
-- ~~`stats`: Enable statistics~~ (removed since 0.3.0)
 - `local_dynamic_tls`: Workaround cannot allocate memory in static tls block
 - `build_cc`: Use of cc crate instead of cmake (cmake still default) as builder (more platform agnostic)
-- ~~`usecxx20`: Enable C++20 standard if available~~ (removed since 0.3.0)
 - `usecxx17`: Use C++17 standard
 - `check`: Enable extra checks to improve security, see upstream [security docs](https://github.com/microsoft/snmalloc/tree/main/docs/security).
   Note that the `memcpy` protection is not enabled in Rust.
@@ -40,6 +34,25 @@ are listed at
 - `notls`: Enables to be loaded dynamically, thus disable tls.
 - `stats`: Enables allocation statistics.
 - `libc-api`: Enables libc API backed by snmalloc.
+- `usewait-on-address`: Enable `WaitOnAddress` support on Windows (enabled by default).
+- `tracing`: Enable structured tracing/logging.
+- `fuzzing`: Enable fuzzing support.
+- `vendored-stl`: Use self-vendored STL.
+- `check-loads`: Enable check loads feature.
+- `pageid`: Enable page ID feature.
+- `gwp-asan`: Enable GWP-ASan integration. Requires `SNMALLOC_GWP_ASAN_INCLUDE_PATH` and `SNMALLOC_GWP_ASAN_LIBRARY_PATH`.
+
+## Build Configuration
+
+The build script ensures architectural alignment between the Rust profile and the underlying `snmalloc` allocator:
+
+### Environment Variables
+The following environment variables are automatically detected and propagated:
+- `DEBUG`: Synchronizes the `snmalloc` build type with the Cargo profile. If `true`, `snmalloc` is built in `Debug` mode.
+- `OPT_LEVEL`: Propagated to the C++ compiler to ensure optimization parity between Rust and C++ components.
+
+### Windows CRT Consistency
+On Windows, the build script enforces static CRT linking (`/MT` or `/MTd`) across both `cc` and `cmake` builders. This prevents linker errors and ensures consistency when `snmalloc` is used as a global allocator.
 
 **To get the crates compiled, you need to choose either `1mib` or `16mib` to determine the chunk configuration**