6b457d2c00
Merge branch 'work_for_merge' of https://github.com/zubycz/mruby into zubycz-work_for_merge
2020-10-15 18:35:13 +09:00
bec4d05340
Introduce endless range (a part of #5085 )
...
Co-Authored-By: n4o847 <22975590+n4o847@users.noreply.github.com >
Co-Authored-By: smallkirby <ssmallkirby@gmail.com >
2020-10-13 14:09:36 +09:00
2b188ed8a1
Reorganize Integer system.
...
- Integrate `Fixnum` and `Integer`
- Remove `Integral`
- `int / int -> int`
- Replace `mrb_fixnum()` to `mrb_int()`
- Replace `mrb_fixnum_value()` to `mrb_int_value()`.
- Use `mrb_integer_p()` instead of `mrb_fixnum_p()`
2020-10-12 18:19:54 +09:00
2a92fb2516
Make division by zero cause ZeroDivisionError.
...
As described in ISO 15.2.30.
2020-10-12 16:21:48 +09:00
80fe9838d2
Integrate Fixnum class into Integer class
...
* The `Fixnum` constant is now an alias for the `Integer` class.
* Remove `struct mrb_state::fixnum_class` member.
If necessary, use `struct mrb_state::integer_class` instead.
2020-10-12 16:21:44 +09:00
52507b1083
Generate C struct from irep instead of binary dump.
2020-10-12 16:21:10 +09:00
0bb4d10ca7
Define Array#to_a to avoid unnecessary loops.
2020-07-22 14:56:56 +09:00
914da3d712
Small comment fix in mrblib/array.c.
2020-07-22 14:56:36 +09:00
33c8d6af1e
Move definition of BasicObject#!= to mrblib.
...
C implementation used `mrb_funcall()` that bypassed many optimization.
2020-06-20 14:09:23 +09:00
181f7b97d6
Fix some Hash methods are inconsistent with values
...
Inconsistent when hash has duplicate key.
### Example
```ruby
# example.rb
keys = (1..3).map{[_1]}
h = keys.to_h{[_1, _1[0]]}
keys[0][0] = 2
p h.values
p h.each_value.to_a
p h
```
#### Before this patch:
```console
$ bin/mruby example.rb
[1, 2, 3]
[1, 1, 3]
{[2]=>1, [2]=>1, [3]=>3}
```
#### After this patch (same as Ruby)
```console
$ bin/mruby example.rb
[1, 2, 3]
[1, 2, 3]
{[2]=>1, [2]=>2, [3]=>3}
```
2020-05-09 15:27:42 +09:00
26e6e75ba6
Use Rake DSL instead of commands of FileUtils
...
- Respect `--verbose(-v)` and `--dry-run(-n)` options.
- Silence warnings to keyword arguments on Ruby 2.7.
2019-12-27 16:49:32 +09:00
6bbdb97e75
Merge pull request #4527 from lopopolo/string-each-line-paragraph-mode
...
Add paragraph mode to String#each_line in mrblib
2019-09-12 21:19:42 +09:00
5b1f25a4e0
Implement Array#each using inline mruby bytecode.
2019-08-17 14:46:32 +09:00
61763129db
Implement Class#new using inline mruby bytecode.
2019-08-16 22:30:03 +09:00
603005ba65
Integrate kazuho/mruby-class-new-fiber-safe in the master.
...
Avoid calling `initialize` via `mrb_funcall`, which cause `cross C
boundary` error from Fibers started in the method.
2019-08-14 14:09:56 +09:00
9f18aced9f
Enumerable#reject, etc. should return Enumerable without block
2019-08-10 22:18:54 +09:00
d5939879cc
Fixed a bug in #4034
2019-07-17 10:39:17 +09:00
d605b72c1d
Merge branch 'master' into i110/inspect-recursion
2019-07-17 10:35:41 +09:00
b4cdced22f
Enumerable#detect {and #find} should call ifnone; fix #4484
...
It's an error in ISO specification; 15.3.2.2.4 and 15.3.2.2.7
2019-07-13 18:05:01 +09:00
32c2aa10c7
Fix Numeric#step to infinity; ref. #4555
2019-07-07 22:10:32 +09:00
0d452073f4
Replace String#[]= method by C implements
...
The purpose is to eliminate string objects that are temporarily created during processing.
2019-06-29 14:41:43 +09:00
46c972f746
Unify loops to minimize bytecode size
2019-06-25 00:01:51 +02:00
6b92acf361
Use explicit block parameter
2019-06-23 03:40:39 +01:00
8ea45b862a
Optimize String#each_line
2019-06-23 03:29:47 +01:00
a932b8c96a
Speed up base case by 2x
...
Make non-paragraph mode twice as fast. Performance is within a factor of 2
of the original implementation.
2019-06-22 12:55:02 +01:00
0f516bbe1d
Add paragraph mode to String#each_line in mrblib
...
mruby/mruby#4511 demonstrated an infinite loop in `String#each_line` when
given an empty string separator. In MRI, an empty separator places
String#each_line in paragraph mode, where the String is separated on
successive runs of newlines. In paragraph mode, the String
`"abc\n\n\ndef\nxyz"` is split into `["abc\n\n\n", "def\nxyz"]`.
This commit makes the String#each_line implementation as close to
ruby/spec compliant as possible given the limitations of mruby core.
With this patch, the following specs fail for `String#each_line`:
- uses `$/` as the separator when none is given (can be fixed by
aliasing and redefining the method to use $/ as the default value
of separator in mruby-io)
- when no block is given returned Enumerator size should return nil
(`Enumerator#size` is not supported on mruby)
- tries to convert the separator to a string using to_str (`String#to_str`
is not implemented on mruby)
This patch has similar memory consumption compared to the prior
implementation and is takes 4x the time the prior implementation takes to
execute:
```console
/usr/bin/time -l ./bin/mruby -e '("aaa\n\nbbbbb\n\n\n\n\ncccc" * 100000).each_line("\n") { }';
```
2019-06-22 12:31:01 +01:00
c75402a57b
Fix typo in mrblib/range.rb [ci skip]
2019-06-04 10:45:13 +09:00
780342eddb
Add Enumerator support to String#each_byte.
...
`String#each_byte` is not defined in ISO Ruby but it is implemented in
the core mruby because it's useful.
2019-05-15 14:59:48 +09:00
fd37bc53de
Remove String#=~ and String#match that requires Regexp.
2019-05-15 09:55:21 +09:00
270131253f
Remove duplicated String#each_char
2019-04-27 12:50:02 +09:00
cdb458ed4e
Commented out String#scan because it is not implemented yet
2019-04-21 20:34:39 +09:00
4a8b88f775
Add type check (conversion) in String#[]=
...
Before this patch:
'a'[0] = 1 #=> 1
'a'[:a] = '1' #=> ArgumentError
'a'[:a, 0] = '1' #=> ArgumentError
'a'[0, :a] = '1' #=> ArgumentError
'a'[0, 1] = 1 #=> 1
After this patch / Ruby:
'a'[0] = 1 #=> TypeError
'a'[:a] = '1' #=> TypeError
'a'[:a, 0] = '1' #=> TypeError
'a'[0, :a] = '1' #=> TypeError
'a'[0, 1] = 1 #=> TypeError
2019-04-19 20:14:23 +09:00
916045921e
Remove duplicated include Comparable in mrblib/string.rb
2019-04-18 19:58:05 +09:00
292bffb32e
Avoid infinite loop when no Regexp class is available; fix #4363
2019-04-08 20:27:28 +09:00
af51119d3a
Raise error on failed comparison in sort; ref #4307
2019-03-02 15:46:50 +09:00
6da3cd5d22
Move NONE to mrblib/enum.rb
2019-02-01 16:02:01 +09:00
8969edf03b
Avoid runtime evaluation for MRB_WITHOUT_FLOAT
2019-01-16 19:32:29 +09:00
de5da4b7f6
Fix coercing for first step counter in Numeric#step
...
Before:
a=[]; 7.step(4, -3.0) { |c| a << c }; p a #=> [7, 4.0]
After / Ruby:
a=[]; 7.step(4, -3.0) { |c| a << c }; p a #=> [7.0, 4.0]
2019-01-15 21:41:54 +09:00
b30ca87bd0
Integrate mrblib/float.rb into src/numeric.c
...
- Avoid hack for `MRB_WITHOUT_FLOAT` in build scripts
- Avoid runtime dispatch for `MRB_WITHOUT_FLOAT`
2019-01-04 20:31:05 +09:00
cca19532c5
Remove Kernel#class_defined? which is not available in CRuby; #3829
2019-01-03 11:34:35 +09:00
c022e4643f
Avoid assignments from type checking String#__to_str.
2018-11-19 12:08:28 +09:00
610bcc88c2
Removed to_hash conversion method.
2018-11-19 12:08:28 +09:00
ff08856fe3
Remove implicit conversion using to_str method; fix #3854
...
We have added internal convenience method `__to_str` which
does string type check.
The issue #3854 was fixed but fundamental flaw of lack of stack
depth check along with fibers still remains. Use `MRB_GC_FIXED_ARENA`
for workaround.
2018-11-19 12:05:46 +09:00
afca99a40b
Remove implicit conversion using to_int method.
...
The ISO standard does not include implicit type conversion using
`to_int`. This implicit conversion often causes vulnerability.
There will be no more attacks like #4120 .
In addition, we have added internal convenience method `__to_int` which
does type check and conversion (from floats).
2018-11-19 11:28:51 +09:00
b37b312081
Rename libmruby stuff to avoid confusion
2018-10-29 14:07:16 +09:00
34872e90d4
Re-implement Array#_inspect and Hash#_inspect without blocks.
...
To reduce the env object allocation; ref #4143
2018-10-29 11:55:49 +09:00
91444c4747
replace quicksort with mergesort.
2018-10-18 13:07:47 +09:00
54ec08c7f1
Implement Hash#rehash in C using sg_compact().
2018-09-26 23:09:48 +09:00
d78acc7afe
Add index to larger segment lists for performance
2018-09-26 23:09:48 +09:00
8ffd4e47fb
Use mrb_undef_value for delete mark instead of shifting Hash entry table.
...
That means entry table should be compacted periodically by `sg_compact()`.
2018-09-26 23:09:48 +09:00
Yukihiro "Matz" Matsumoto