| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2021-07-29 | parse.y: revert some changes to be compiled by old `bison`. | Yukihiro "Matz" Matsumoto | |
| 2021-07-29 | y.tab.c: remove auto generated `y.tab.c` from repository. | Yukihiro "Matz" Matsumoto | |
| It used to be generated in the build, but `bison` required to generation was too hard to install correctly on MacOS and Windows (version mismatch and package maneger issues). This is retry. | |||
| 2021-07-28 | codegen.c: labels should be `uint32_t`. | Yukihiro "Matz" Matsumoto | |
| 2021-07-28 | codegen.c: fix `loopinfo` structure members. | Yukihiro "Matz" Matsumoto | |
| - `pc0`: `next` destination - `pc1`: `redo` destination (renamed from `pc2`) - `pc3`: `break` destination (renamed from `pc3`) old `pc1` was unused so removed. | |||
| 2021-07-28 | codegen.c: removed bytecode when the value of `while` is not used. | Yukihiro "Matz" Matsumoto | |
| 2021-07-28 | codegen.c: fixed a bug when value is taken from `while` and `until`. | Yukihiro "Matz" Matsumoto | |
| ```ruby p ((while true; p 1; break; end)) ``` should print `1 nil` but was `1`. | |||
| 2021-07-26 | parse.y: fix an integer cast warning. | Yukihiro "Matz" Matsumoto | |
| 2021-07-26 | parse.y: `mrb_int_read()` returns `mrb_int` not `unsigned long`. | Yukihiro "Matz" Matsumoto | |
| Also, now too big capture group index just gives `nil`, not error. | |||
| 2021-07-26 | parse.y: unify redundant functions `yywarn()` and `yywarning()`. | Yukihiro "Matz" Matsumoto | |
| 2021-07-25 | parse.y: replace `strtoul()` by `mrb_int_read()`. | Yukihiro "Matz" Matsumoto | |
| 2021-07-25 | Remove redundant include headers. | Yukihiro "Matz" Matsumoto | |
| - stdlib.h - stddef.h - stdint.h - stdarg.h - limits.h - float.h | |||
| 2021-07-23 | codegen.c: fixed a bug in `OP_LOADI32` peephole optimization. | Yukihiro "Matz" Matsumoto | |
| 2021-07-22 | codegen.c: `get_int_operand()` to support `OP_LOADL` (int in pool). | Yukihiro "Matz" Matsumoto | |
| 2021-07-22 | codegen.c: add constant folding for unary numeric operators (+, -, ~). | Yukihiro "Matz" Matsumoto | |
| 2021-07-22 | codegen.c: compare symbol names directly avoiding string conversion. | Yukihiro "Matz" Matsumoto | |
| 2021-07-21 | codegen.c: skip `-@` call if the argument is a literal integer. | Yukihiro "Matz" Matsumoto | |
| 2021-07-21 | codegen.c: move `gen_setxv()` after `new_sym()`. | Yukihiro "Matz" Matsumoto | |
| 2021-07-21 | codegen.c: introduce `gen_int()` to generate integer instructions. | Yukihiro "Matz" Matsumoto | |
| 2021-07-21 | codegen.c: add peephole optimization for `OP_LOADI32` before `OP_MOVE`. | Yukihiro "Matz" Matsumoto | |
| 2021-07-21 | codegen.c: add peephole optimization for `OP_LOADI16` before `OP_MOVE`. | Yukihiro "Matz" Matsumoto | |
| 2021-07-21 | codegen.c: a new function `get_int_operand`. | Yukihiro "Matz" Matsumoto | |
| 2021-07-20 | codegen.c: negative zero equals to positive zero. | Yukihiro "Matz" Matsumoto | |
| `OP_LOADI Rn -0` should be `OP_LOADI_0`. | |||
| 2021-07-20 | codegen.c: allow `OP_EXT` before `OP_ADDI` and `OP_SUBI`. | Yukihiro "Matz" Matsumoto | |
| This is preparation for integer constant folding. | |||
| 2021-07-17 | codegen.c: optimize variable assignments after `OP_MOVE`. | Yukihiro "Matz" Matsumoto | |
| - `OP_SETGV` - `OP_SETIV` - `OP_SETCV` - `OP_SETCONST` | |||
| 2021-07-17 | codegen.c: optimize `OP_SETUPVAR` after `OP_MOVE`. | Yukihiro "Matz" Matsumoto | |
| 2021-07-16 | codegen.c: add new peephole optimization for `OP_MOVE`. | Yukihiro "Matz" Matsumoto | |
| If `OP_MOVE` comes after `OP_GETUPVAR`, you can skip move and redirect the destination register of `OP_GETUPVAR`. | |||
| 2021-07-16 | codegen.c: add new peephole optimization for `OP_GETUPVAR`. | Yukihiro "Matz" Matsumoto | |
| When `OP_GETUPVAR` is generated right after `OP_SETUPVAR`, there is no need to read the upvar back to the register, e.g. 3 008 OP_ADDI R2 1 3 011 OP_SETUPVAR R2 1 0 4 015 OP_GETUPVAR R2 1 0 4 019 OP_LOADI_2 R3 `OP_GETUPVAR` at the address `015` is useless. We can skip it like: 3 008 OP_ADDI R2 1 3 011 OP_SETUPVAR R2 1 0 4 015 OP_LOADI_2 R3 | |||
| 2021-07-15 | codegen.c: `gen_jmpdst` always needs to generate relative address. | Yukihiro "Matz" Matsumoto | |
| 2021-07-09 | codegen.c: avoid uninitialized local variable. | Yukihiro "Matz" Matsumoto | |
| 2021-07-04 | codegen.c: jump address should be generated by `gen_jmpdst()`. | Yukihiro "Matz" Matsumoto | |
| 2021-07-03 | vm.c: `OP_DEF` to push a symbol to `a` register. | Yukihiro "Matz" Matsumoto | |
| The code generator no longer need to emit `OP_LOADSYM` after `OP_DEF`. `doc/opcode.md` is also updated. | |||
| 2021-06-30 | Revert "Remove `OP_EXT[123]` from operands." | Yukihiro "Matz" Matsumoto | |
| This reverts commit fd10c7231906ca48cb35892d2a86460004b62249. I thought it was OK to restrict index value within 1 byte, but in some cases index value could be 16 bits (2 bytes). I had several ideas to address the issue, but reverting `fd10c72` is the easiest way. The biggest reason is `mruby/c` still supports `OP_EXT[123]`, so that they don't need any additional work. | |||
| 2021-06-28 | Merge pull request #5493 from dearblue/binding.2 | Yukihiro "Matz" Matsumoto | |
| Fixed finding variables from `proc` in `binding.eval` failed | |||
| 2021-06-26 | Fixed finding variables from `proc` in `binding.eval` failed | dearblue | |
| Previously the following code did not produce the expected results: ```ruby bx = binding block = bx.eval("a = 1; proc { a }") bx.eval("a = 2") p block.call # Expect 2 but return 1 due to a bug ``` The previous implementation of `Binding#eval` evaluated the code and then merged the top layer variables. This patch will parse and expand the variable space before making a call to `eval`. This means that the call to `Binding#eval` will do the parsing twice. In addition, the following changes will be made: - Make `mrb_parser_foreach_top_variable()`, `mrb_binding_extract_proc()` and `mrb_binding_extract_env()` functions private global functions. - Remove the `posthook` argument from `mrb_exec_irep()`. The `posthook` argument was introduced to implement the `binding` method. This patch is unnecessary because it uses a different implementation method. ref #5362 fixed #5491 | |||
| 2021-06-26 | Fixed finding variables defined in the upper proc failed | dearblue | |
| If no new variable was defined in the `eval` method, the variable was hidden from the nested `eval` method. ```ruby a = 1 p eval %(b = 2; eval %(a)) # => 1 (good) p eval %(eval %(a)) # => undefined method 'a' (NoMethodError) ``` This issue has occurred since mruby 3.0.0. | |||
| 2021-06-19 | codegen.c: stop `uninitialized` warning. | Yukihiro "Matz" Matsumoto | |
| 2021-06-11 | codegen.c: refactor `readint()` | Yukihiro "Matz" Matsumoto | |
| * renamed from redundant `readint_mrb_int()` * supports only base upto 16 * no base validation (already done in parser) * no negative read (negate after read) * overflow detection using `mrb_int_{mul,add}_overflow()` | |||
| 2021-05-29 | codegen.c: just raise the exception on broken node for assignments. | Yukihiro "Matz" Matsumoto | |
| 2021-05-18 | parse.y: allow "command" syntax in endless method definition. | Yukihiro "Matz" Matsumoto | |
| This change allows `def hello = puts "Hello"` without parentheses. This syntax has been introduced since Ruby3.1. | |||
| 2021-05-18 | parse.y: endless singleton method definition can omit parens. | Yukihiro "Matz" Matsumoto | |
| 2021-04-16 | feat(CI): add the GitHub Super Linter | John Bampton | |
| The GitHub Super Linter is a more robust and better supported tool than the current GitHub Actions we are using. Running these checks: ERROR_ON_MISSING_EXEC_BIT: true VALIDATE_BASH: true VALIDATE_BASH_EXEC: true VALIDATE_EDITORCONFIG: true VALIDATE_MARKDOWN: true VALIDATE_SHELL_SHFMT: true VALIDATE_YAML: true https://github.com/marketplace/actions/super-linter https://github.com/github/super-linter Added the GitHub Super Linter badge to the README. Also updated the pre-commit framework and added more documentation on pre-commit. Added one more pre-commit check: check-executables-have-shebangs Added one extra check for merge conflicts to our GitHub Actions. EditorConfig and Markdown linting. Minor grammar and spelling fixes. Update linter.yml | |||
| 2021-03-31 | codegen.c: `s->ainfo` should be zero along with `OP_ENTER(0)`. | Yukihiro "Matz" Matsumoto | |
| 2021-03-31 | codegen.c: `yield` outside of method is now `SyntaxError`. | Yukihiro "Matz" Matsumoto | |
| 2021-03-31 | codegen.c: remove `codegen error:` prefix from error messages. | Yukihiro "Matz" Matsumoto | |
| 2021-03-31 | codegen.c: `ainfo` may be negative. | Yukihiro "Matz" Matsumoto | |
| When argument information is not available. So it should not happen for `yield` (error). In contrast, the error from `super` should be handled in run time (ignored). | |||
| 2021-03-22 | codegen.c: fix integer size warnings. | Yukihiro "Matz" Matsumoto | |
| along with repeated calls of `strlen()`. | |||
| 2021-03-13 | codegen.c: fix memory leak from `new_litbn()`. | Yukihiro "Matz" Matsumoto | |
| 2021-03-12 | codegen.c: need to preserve `neg` info. | Yukihiro "Matz" Matsumoto | |
| 2021-03-12 | codegen.c: avoid integer comparison of different sign. | Yukihiro "Matz" Matsumoto | |
| 2021-03-12 | codegen.c: no integer overflow error in `codegen`; close #5376 | Yukihiro "Matz" Matsumoto | |
| Add new pool value type `IREP_TT_BIGINT` and generate integer overflow error in the VM. In the future, `mruby` will support `Bignum` for integers bigger than `mrb_int` (probably using `mpz`). | |||
