| Age | Commit message (Collapse) | Author |
|
Stirs internal state when `seed` is set in `Random`
|
|
This is to avoid the approximation of `Random#rand` when `seed`s are close together.
For example, the first `#rand` values after doing `Random#srand(0)` and `Random#srand(1)` are very similar.
Below is the sequence of mruby before this patch was given a `seed` of `0...20`:
```console
% bin/mruby -e 'r = Random.new; 20.times { |i| r.srand(i); puts "seed=%2d %s" % [i, 10.times.map { "%0.3f" % r.rand }.join(" ")] }'
seed= 0 0.643 0.585 0.198 0.732 0.087 0.605 0.548 0.468 0.573 0.966
seed= 1 0.643 0.585 0.198 0.607 0.370 0.605 0.633 0.593 0.395 0.439
seed= 2 0.643 0.585 0.197 0.981 0.652 0.730 0.875 0.713 0.529 0.269
seed= 3 0.643 0.585 0.198 0.857 0.934 0.730 0.960 0.216 0.286 0.523
seed= 4 0.643 0.585 0.197 0.231 0.217 0.605 0.959 0.958 0.478 0.482
seed= 5 0.643 0.585 0.197 0.106 0.249 0.605 0.044 0.330 0.925 0.047
seed= 6 0.643 0.585 0.197 0.481 0.781 0.731 0.285 0.960 0.804 0.725
seed= 7 0.643 0.585 0.197 0.356 0.813 0.731 0.370 0.711 0.937 0.448
seed= 8 0.643 0.585 0.198 0.732 0.329 0.108 0.243 0.974 0.766 0.936
seed= 9 0.643 0.585 0.198 0.607 0.611 0.108 0.827 0.102 0.962 0.597
seed=10 0.643 0.585 0.198 0.981 0.393 0.233 0.569 0.723 0.472 0.805
seed=11 0.643 0.585 0.198 0.857 0.676 0.233 0.154 0.222 0.603 0.371
seed=12 0.643 0.585 0.198 0.231 0.458 0.108 0.654 0.979 0.928 0.577
seed=13 0.643 0.585 0.198 0.106 0.490 0.108 0.239 0.355 0.749 0.831
seed=14 0.643 0.585 0.198 0.481 0.523 0.233 0.981 0.486 0.505 0.131
seed=15 0.643 0.585 0.198 0.356 0.555 0.234 0.565 0.233 0.011 0.666
seed=16 0.643 0.585 0.197 0.730 0.573 0.611 0.904 0.512 0.971 0.153
seed=17 0.643 0.585 0.197 0.605 0.855 0.611 0.240 0.636 0.041 0.509
seed=18 0.643 0.585 0.196 0.979 0.137 0.736 0.229 0.765 0.674 0.832
seed=19 0.643 0.585 0.197 0.855 0.420 0.736 0.566 0.268 0.183 0.219
```
|
|
|
|
|
|
|
|
Embedding reduce memory consumption, sacrificing precision. It clips least
significant 2 bits from `mrb_float`, so if you need to keep float precision,
define `MRB_USE_FLOAT_FULL_PRECISION`.
`MRB_WORD_BOXING` and `MRB_INT64`:
`mrb_float` (`double`) is embedded in `mrb_value` clipped last 2 bits.
`MRB_WORD_BOXING` and `MRB_INT64` and `MRB_USE_FLOAT_FULL_PRECISION`:
`mrb_float` is allocated in the heaps wrapped by `struct RFloat`.
`MRB_WORD_BOXING` and `MRB_INT32` and `MRB_USE_FLOAT32`:
`mrb_float` (`float`) is embedded in `mrb_value` clipped last 2 bits.
In addition, to reserve bit space in the `mrb_value`, maximum inline
symbol length become 4 (instead of 5) in the configuration.
`MRB_WORD_BOXING` and `MRB_INT32`:
Assume `MRB_USE_FLOAT_FULL_PRECISION` and allocate Float values in heap.
|
|
Add SerenityOS crossbuild
|
|
|
|
|
|
Cross compile mruby for SerenityOS (https://github.com/SerenityOS/serenity)
|
|
In additions:
* use bitmap for flags
* BER integer compression for length
|
|
|
|
|
|
Probably we should add conditions for FreeBSD, etc.
|
|
|
|
- mrb_packed_int_len()
- mrb_packed_int_encode()
- mrb_packed_int_decode()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Add instructions for installing doxygen on RHEL/Fedora/CentOS
|
|
|
|
Check the class with `I` specifier of `mrb_get_args()`
|
|
Previously, the `I` specifier only checked if the object was `MRB_TT_ISTRUCT`.
So it was at risk of getting pointers to different C structs if multiple classes were to use the `MRB_TT_ISTRUCT` instance.
Change this behavior and change the C argument corresponding to the `I` specifier to `(void *, struct RClass)`.
This change is not compatible with the previous mruby.
Please note that if the user uses the previous specifications, `SIGSEGV` may occur or the machine stack may be destroyed.
resolve #5527
|
|
|
|
`rewind_pc` should not be called when `s->pc` is `0` (top).
|
|
Last commit made `mrb_decode_insn()` to take `NULL` as `pc`.
|
|
|
|
|
|
Fix #5528
This reverts commit 59201b59046b9e73c309508350cd3c0fafd20e4d; #5497
|
|
|
|
|
|
|
|
Recent peephole optimization made `ADDI/SUBI` destinations possibly
local variables.
|
|
`a=10; a+=1` to generate:
```
1 000 OP_LOADI R1 11 ; R1:a
```
instead of:
```
1 000 OP_LOADI R1 10 ; R1:a
1 003 OP_MOVE R2 R1 ; R1:a
1 006 OP_ADDI R2 1
1 009 OP_MOVE R1 R2 ; R1:a
```
|
|
mruby/dependabot/github_actions/github/super-linter-4.6.1
build(deps): bump github/super-linter from 4.6.0 to 4.6.1
|
|
Bumps [github/super-linter](https://github.com/github/super-linter) from 4.6.0 to 4.6.1.
- [Release notes](https://github.com/github/super-linter/releases)
- [Commits](https://github.com/github/super-linter/compare/v4.6.0...v4.6.1)
---
updated-dependencies:
- dependency-name: github/super-linter
dependency-type: direct:production
update-type: version-update:semver-patch
...
Signed-off-by: dependabot[bot] <[email protected]>
|
|
Expose all header files of the gems for small compilation
|
|
After building mruby, if the user compiles and links to `libmruby.a`, expose the included directory of the captured gems.
This is a change to the `<build-dir> /lib/libmruby.flags.mak` file and will result in being added to `bin/mruby-config --cflags`.
This eliminates the need for the user to look up the path and add the compiler flag if the user wants to take advantage of her gems publishing features.
In the main build with `rake CONFIG=...`, there is no problem because it can only be seen from the gems that depends directly and indirectly as before.
However, when compiling independently by the user using `bin/mruby-config`, a header file name collision may occur if a unique header directory is added.
|
|
pre-commit autoupdate
|
|
https://pre-commit.com/#pre-commit-autoupdate
|
|
Fixed compilation error of "mruby-pack" with `MRB_NO_STDIO`
|
|
docs: fix case of mruby
|
|
|
|
|