Rollup merge of #125011 - diondokter:opt-for-size, r=Amanieu,kobzol

Add opt-for-size core lib feature flag

Adds a feature flag to the core library that enables the possibility to have smaller implementations for certain algorithms.

So far, the core lib has traded performance for binary size. This is likely what most people want since they have big simd-capable machines. However, people on small machines, like embedded devices, don't enjoy the potential speedup of the bigger algorithms, but do have to pay for them. These microcontrollers often only have 16-1024kB of flash memory.

This PR is the result of some talks with project members like `@Amanieu` at RustNL.
There are some open questions of how this is eventually stabilized, but it's a similar question as with the existing `panic_immediate_abort` feature.

Speaking as someone from the embedded side, we'd rather have this unstable for a while as opposed to not having it at all. In the meantime we can try to use it and also add additional PRs to the core lib that uses the feature flag in areas where we find benefit.

Open questions from my side:
- Is this a good feature name?
  - `panic_immediate_abort` is fairly verbose, so I went with something equally verbose
  - It's easy to refactor later
- I've added the feature to `std` and `alloc` as well as they might benefit too. Do we agree?
  - I expect these to get less usage out of the flag since most size-constraint projects don't use these libraries often.

5 files changed, 12 insertions, 1 deletions

diff --git a/library/alloc/Cargo.toml b/library/alloc/Cargo.toml
index e8afed6b35a..2e7fcb9dbd3 100644
--- a/library/alloc/Cargo.toml
+++ b/library/alloc/Cargo.toml
@@ -37,4 +37,6 @@ compiler-builtins-no-asm = ["compiler_builtins/no-asm"]
 compiler-builtins-mangled-names = ["compiler_builtins/mangled-names"]
 compiler-builtins-weak-intrinsics = ["compiler_builtins/weak-intrinsics"]
 # Make panics and failed asserts immediately abort without formatting any message
-panic_immediate_abort = []
+panic_immediate_abort = ["core/panic_immediate_abort"]
+# Choose algorithms that are optimized for binary size instead of runtime performance
+optimize_for_size = ["core/optimize_for_size"]
diff --git a/library/core/Cargo.toml b/library/core/Cargo.toml
index a02fcf50416..11d33971f25 100644
--- a/library/core/Cargo.toml
+++ b/library/core/Cargo.toml
@@ -31,6 +31,8 @@ rand_xorshift = { version = "0.3.0", default-features = false }
 [features]
 # Make panics and failed asserts immediately abort without formatting any message
 panic_immediate_abort = []
+# Choose algorithms that are optimized for binary size instead of runtime performance
+optimize_for_size = []
 # Make `RefCell` store additional debugging information, which is printed out when
 # a borrow error occurs
 debug_refcell = []
diff --git a/library/std/Cargo.toml b/library/std/Cargo.toml
index a77d438e1f1..5b36867fe24 100644
--- a/library/std/Cargo.toml
+++ b/library/std/Cargo.toml
@@ -78,6 +78,8 @@ system-llvm-libunwind = ["unwind/system-llvm-libunwind"]
 
 # Make panics and failed asserts immediately abort without formatting any message
 panic_immediate_abort = ["core/panic_immediate_abort", "alloc/panic_immediate_abort"]
+# Choose algorithms that are optimized for binary size instead of runtime performance
+optimize_for_size = ["core/optimize_for_size", "alloc/optimize_for_size"]
 
 # Enable std_detect default features for stdarch/crates/std_detect:
 # https://github.com/rust-lang/stdarch/blob/master/crates/std_detect/Cargo.toml
diff --git a/library/sysroot/Cargo.toml b/library/sysroot/Cargo.toml
index 6ff24a8db59..1ddacd92e6b 100644
--- a/library/sysroot/Cargo.toml
+++ b/library/sysroot/Cargo.toml
@@ -22,6 +22,7 @@ llvm-libunwind = ["std/llvm-libunwind"]
 system-llvm-libunwind = ["std/system-llvm-libunwind"]
 panic-unwind = ["std/panic_unwind"]
 panic_immediate_abort = ["std/panic_immediate_abort"]
+optimize_for_size = ["std/optimize_for_size"]
 profiler = ["std/profiler"]
 std_detect_file_io = ["std/std_detect_file_io"]
 std_detect_dlsym_getauxval = ["std/std_detect_dlsym_getauxval"]
diff --git a/src/ci/docker/scripts/x86_64-gnu-llvm.sh b/src/ci/docker/scripts/x86_64-gnu-llvm.sh
index 2eb751ca376..876b300d35c 100755
--- a/src/ci/docker/scripts/x86_64-gnu-llvm.sh
+++ b/src/ci/docker/scripts/x86_64-gnu-llvm.sh
@@ -23,6 +23,10 @@ if [[ -z "${PR_CI_JOB}" ]]; then
     # Run `ui-fulldeps` in `--stage=1`, which actually uses the stage0
     # compiler, and is sensitive to the addition of new flags.
     ../x.py --stage 1 test tests/ui-fulldeps
+
+    # The tests are run a second time with the size optimizations enabled.
+    ../x.py --stage 1 test library/std library/alloc library/core \
+        --rustc-args "--cfg feature=\"optimize_for_size\""
 fi
 
 # When running gcc backend tests, we need to install `libgccjit` and to not run llvm codegen