Moved Array#fetch from Ruby to C using hybrid implementation for
better performance. The C implementation handles all non-block cases
with unified API that eliminates Ruby conditional logic.
Key improvements:
- Fast C implementation for common cases (no blocks)
- Shared index normalization helper reusable for other methods
- Unified C call eliminates NONE sentinel comparison in Ruby
- Block cases use C helper for index normalization
Added comprehensive test coverage including edge cases, default values,
block handling, and error message format verification. Combined tests
to focus on functionality rather than implementation details.
Co-authored-by: Atlassian Rovo Dev
This commit also corrects the behavior of `Array#insert` when a negative
index is out of bounds. It now raises an `IndexError`, which is
consistent with CRuby.
Co-authored-by: Gemini <gemini@google.com>
This commit replaces the Ruby implementation of and with a C
implementation. The new implementation is iterative and uses a stack to
avoid deep recursion, which prevents stack overflows when flattening
deeply nested arrays.
Co-authored-by: Gemini <gemini@google.com>
Implemented shared C argument parser and separate fill logic to eliminate code
duplication while maximizing performance. The implementation uses C implemented
__fill_parse_args for unified argument handling and __fill_exec for fast
C-based value filling.
Added comprehensive test coverage for both shared argument parsing
and C fill implementation, including range arguments, block handling,
and array extension scenarios.
Co-authored-by: Atlassian Rovo Dev
Co-authored-by: Gemini <gemini@google.com>
The Ruby implementation of `Array#difference` was inefficient as it
called `Array#-` repeatedly, creating intermediate arrays.
This commit replaces it with a C implementation that processes all
arguments in a single pass. The core logic is extracted into a
shared helper function, `ary_subtract_internal`, which is now used
by both `Array#-` and `Array#difference`.
Co-authored-by: Gemini <gemini@google.com>
We have more chance to avoid hash allocation in set-like methods. Since
memory situation heavily depends on the platform, we may need to make
this threshold configurable in the future.
Co-authored-by: Atlassian Rovo Dev
Moved Array#intersect? implementation from Ruby to C to improve memory
usage and performance with early termination optimization. The C
implementation uses hash-based lookup for large arrays (>16 elements)
and linear search for smaller arrays.
Added comprehensive test coverage including early termination scenarios,
empty arrays, size optimization verification, and edge cases with
duplicates and large arrays.
Co-authored-by: Atlassian Rovo Dev
Moved Array#& (set intersection) implementation from Ruby to C to improve
memory usage and performance. The C implementation uses hash-based
deduplication for large arrays (>16 elements) and linear search for
smaller arrays, following the same hybrid pattern as Array#| and Array#-.
Key improvements:
- Hash-based approach uses mrb_hash_delete_key() for proper deduplication
- Linear search approach checks result array to ensure uniqueness
- Maintains order preservation from the first array
- Eliminates temporary object creation in Ruby implementation
Added comprehensive test coverage for both small and large array scenarios,
including edge cases like no intersection, complete intersection, and
duplicate handling.
Co-authored-by: Atlassian Rovo Dev
The C implementation uses hash-based deduplication for large arrays
(>16 elements) and linear search for smaller arrays, following the same
pattern as other set operations.
Co-authored-by: Atlassian Rovo Dev
Refactor Array#- to a C implementation for improved memory and performance,
especially for set operations. Uses a hybrid approach for efficiency.
Co-authored-by: Gemini <gemini@google.com>
Internal functions can only be called from within the library.
Functions listed in `mruby/internal.h` can be called from:
* core (src/*.c)
* gems (mrbgems/**/*.c)
But not from the application linked with `libmruby`.
ref. #5613
I checked with Valgrind, and the methods that can cause use-after-free are `Array#rotate`, `Array#rotate!`, and `String#byteslice`.
Since `String#rindex` uses `RSTRING_LEN()` indirectly inside the function, no reference to the out-of-bounds range is generated.
The Ruby version of `Array#rotate!` generated a rotated array and
replaced the receiver, but the C version rotates the receiver array
in-place. So the performance is improved a lot both in speed and memory
consumption. Look for the comments in `array.c` for the in-place rotating
algorithm, if you are interested.
Instead of including `mruby/presym.h` everywhere, we provided the
fallback `mruby/presym.inc` under `include/mruby` directory, and specify
`-I<build-dir>/include` before `-I<top-dir>/include` in `presym.rake`.
So even when someone drops `-I<build-dir>/include` in compiler options,
it just compiles without failure.
The "a"/"*" specifier of the `mrb_get_args()` function will now return `const mrb_value *`.
This is because it is difficult for the caller to check if it is an array object and write-barrier if necessary.
And it requires calling `mrb_ary_modify()` on the unmodified array object, which is also difficult (this is similar to #5087).
`mrb_get_arg1()` raises `ArgumentError` if the method does not receive one
argument.
And replaces all `mrb_get_args(mrb, "o", &arg)` by the new function.
It reduces the memory consumption and sometimes improve the
performance as well. For example, the consumed memory size
of `bench/bm_ao_render.rb` is reduced from 1.2GB to 1GB, and
its total execution time become 18.795 sec from 22.229 sec.
Yukihiro "Matz" Matsumoto