Yukihiro "Matz" Matsumoto
e2c628583e
Fix mrb_ary_splat() to copy the array always.
2020-08-30 19:14:34 +09:00
Yukihiro "Matz" Matsumoto
6b8664b951
Fix the bug caused by to_a returning a frozen array.
...
Reported by @shuujii.
2020-08-29 23:00:36 +09:00
Yukihiro "Matz" Matsumoto
f08f3c34be
Fix a bug introduced by the last commit.
...
Should have handled the case `to_a` returns `nil`.
2020-08-29 22:58:07 +09:00
Yukihiro "Matz" Matsumoto
564372d7b9
Fix mrb_obj_ceqq to avoid array copying.
2020-08-29 18:43:47 +09:00
Yukihiro "Matz" Matsumoto
a66d86cfc1
Fix mrb_ary_splat to copy the array; fix #5067
2020-08-29 18:43:26 +09:00
Yukihiro "Matz" Matsumoto
5e55b61c44
Fix mrb_int and size_t combination warnings.
2020-08-11 15:06:51 +09:00
Yukihiro "Matz" Matsumoto
df969b163a
Return NaN for 0/0; d8e060d
2020-08-08 15:36:55 +09:00
Yukihiro "Matz" Matsumoto
ad8b36daa1
Should not use mrb_float_value() with MRB_WITHOUT_FLOAT.
2020-08-08 14:14:17 +09:00
Yukihiro "Matz" Matsumoto
c849b894ed
Reintroduce mrb_static_assert; #5051
...
Note that the home brew version of `mrb_static_assert` only works within
the function body. This reverts commit 8f99689 .
2020-08-08 10:39:26 +09:00
Yukihiro "Matz" Matsumoto
d8e060d2d3
Avoid division by zero undefined behavior.
2020-08-07 10:51:33 +09:00
Yukihiro "Matz" Matsumoto
0b95c8a1e8
Use memcpy to copy stack.
2020-08-06 17:02:15 +09:00
Yukihiro "Matz" Matsumoto
57a798e687
Use memset() to clear stack unless MRB_NAN_BOXING.
2020-08-06 17:01:59 +09:00
Yukihiro "Matz" Matsumoto
755233ef7d
Remove block clear of callinfo.
2020-08-06 17:01:28 +09:00
Yukihiro "Matz" Matsumoto
8f99689ba3
Remove mrb_static_assert from the core; #5051
2020-08-06 16:01:45 +09:00
Yukihiro "Matz" Matsumoto
11cc7bed96
Merge pull request #5052 from dearblue/cistacks
...
Extend the `cipush()` and `cipop()` functions
2020-08-05 17:38:51 +09:00
Yukihiro "Matz" Matsumoto
faed5054f8
Initialized local variables in mrb_hash_shift().
2020-08-03 15:53:50 +09:00
dearblue
d96be5c1b2
Extend the cipush() and cipop() functions
...
- Returns the updated call info.
- Unify the processing around `cipush()`.
- `cipop()` restores the stack.
2020-07-26 13:12:01 +09:00
Yukihiro "Matz" Matsumoto
425b5f32d6
Merge pull request #5049 from shuujii/use-type-tag-for-hash-code-in-ht_hash_func
...
Use type tag for hash code in `ht_hash_func()`
2020-07-25 19:40:05 +09:00
KOBAYASHI Shuji
0c88c71786
Use type tag for hash code in ht_hash_func()
...
The function corresponding to `ht_hash_func()` was as follows in the days of
khash implementation (before d78acc7a ).
```c
mrb_hash_ht_hash_func(mrb_state *mrb, mrb_value key)
{
enum mrb_vtype t = mrb_type(key);
...
switch (t) {
...
default:
hv = mrb_funcall(mrb, key, "hash", 0);
h = (khint_t)t ^ (khint_t)mrb_fixnum(hv);
break;
}
...
}
```
When switched to the segmented list implementation (d78acc7a ), this function
was changed as follows.
```c
sg_hash_func(mrb_state *mrb, seglist *t, mrb_value key)
{
enum mrb_vtype tt = mrb_type(key);
...
switch (tt) {
...
default:
hv = mrb_funcall(mrb, key, "hash", 0);
h = (size_t)t ^ (size_t)mrb_fixnum(hv);
break;
}
...
}
```
Since the argument `t` was added, the variable for type tag was changed from
`t` to `tt`, but the variable used in the expression of `h` remained `t`.
Probably this is an omission of change, so fixed it.
2020-07-25 16:59:02 +09:00
Yukihiro "Matz" Matsumoto
f868d7d357
Change the logic to calculate object (iv_tbl) size; #5045
2020-07-24 13:13:06 +09:00
Yukihiro "Matz" Matsumoto
c69ca2c7f8
Merge pull request #5045 from dearblue/memsize_of
...
Improve `mruby-os-memsize`
2020-07-24 12:02:05 +09:00
dearblue
ccd5f203bf
Improve prototype for mrb_objspace_page_slot_size(); ref #5032
...
If it qualify a return type that is not a pointer with `const`, the
compiler ignores it.
2020-07-24 11:48:09 +09:00
dearblue
5c2b11d215
Avoid using FPU with mruby-os-memsize; ref #5032
...
And, in the calculation of the instance variable size, the fraction was
always rounded down because of division of integers, so fix it.
At the same time, test items that are no longer passed due to this
change are deleted.
2020-07-24 11:42:37 +09:00
Yukihiro "Matz" Matsumoto
e7599050dc
Fix a bug with ht_index called with size==0; fix #5046
...
It happens when a hash made empty calls `rehash`.
2020-07-23 07:06:35 +09:00
Yukihiro "Matz" Matsumoto
b920270508
Use more mrb_field_write_barrier for instance variables.
2020-07-22 15:01:55 +09:00
Yukihiro "Matz" Matsumoto
491aa1bfe5
Use more local variables.
...
To make debugging easy, and to improve the performance little bit.
2020-07-22 15:01:54 +09:00
Yukihiro "Matz" Matsumoto
c99bb756c4
Move gray_list update from gc_mark_children.
...
The responsibility moved to caller to avoid confusion. Currently the
function is called from only 2 places, so it is relatively easy to
ensure not to update `gray_list` in the caller. But the assumption
may change in the future.
2020-07-22 15:01:54 +09:00
Yukihiro "Matz" Matsumoto
097f525817
Avoid using mrb_ary_modify from the internal function.
...
`mrb_ary_modify` calls `mrb_write_barrier`, so can cause the same
problem of the past `push`. It is provided for use-level API.
2020-07-22 15:01:54 +09:00
Yukihiro "Matz" Matsumoto
5533c29833
Use mrb_field_write_barrier instead of mrb_write_barrier for push.
...
When the array is very big, the simpler `mrb_write_barrier` causes
calling `gc_mark_children` for big arrays repeatedly. That would hinder
performance very badly.
2020-07-22 15:01:53 +09:00
Yukihiro "Matz" Matsumoto
670622f45d
Skip unnecessary mark_context if mrb->c == mrb->root_c.
2020-07-22 14:56:03 +09:00
Yukihiro "Matz" Matsumoto
6334949ba6
Fix the VM stack handling bug in 'mrb_yield_with_class()`; fix #5042
2020-07-21 12:47:35 +09:00
Yukihiro "Matz" Matsumoto
f76defd310
Use c specifier for mrb_get_args.
2020-07-19 07:47:07 +09:00
Yukihiro "Matz" Matsumoto
d023adcb12
Add new specifier c to mrb_get_args.
...
`C` retrieves a `mrb_value` that refers a class/module.
`c` retrieves a `struct RClass*` pointer to a class/module.
2020-07-19 07:45:16 +09:00
Yukihiro "Matz" Matsumoto
1952004d5a
Use proc->env to check block_given? if possible; fix #5039
...
This bug has been there since mruby 1.4.0 (2018-04).
2020-07-17 16:15:47 +09:00
Yukihiro "Matz" Matsumoto
b5bf9510ee
Merge pull request #5032 from RoryO/add-objspace-memsize-of
...
Add ObjectSpace.memsize_of
2020-07-17 07:10:58 +09:00
Rory O'Connell
f74d370c15
mrb_ prefix convention
2020-07-15 19:57:22 -07:00
Rory OConnell
6f945a09b4
Use object iv table size in calculation
2020-07-13 15:59:24 -07:00
Rory OConnell
e7bd7d0eaf
Use size of hash's table in calculation
2020-07-13 15:58:50 -07:00
Rory OConnell
41e3220539
All values use page slot size in calculation
2020-07-13 15:56:27 -07:00
dearblue
28e39419a5
Remove the prototype declaration mrb_free_backtrace()
...
This function is removed by 9644ad5 .
2020-07-11 16:15:17 +09:00
Yukihiro "Matz" Matsumoto
01a8d8498f
Avoid infinite loop when converting objects to strings.
2020-07-06 23:11:51 +09:00
Rory OConnell
5746958470
fix object_id of true, false, and undef all 0
2020-07-04 20:33:38 -07:00
Yukihiro "Matz" Matsumoto
97319697c8
Cancel 9cdf439
...
Should not free the pointer in `realloc` since it can cause
use-after-free problem.
2020-07-02 10:41:03 +09:00
Yukihiro "Matz" Matsumoto
233cb49903
Remove unnecessary stack adjustment in OP_CALL.
2020-06-25 06:57:43 +09:00
Yukihiro "Matz" Matsumoto
9cdf439db5
Free the original pointer if realloc failed.
...
The POSIX `realloc` keep the original pointer untouched, so it can
easily leads to memory leakage. `mrb_realloc()` should handle those
bookkeeping, while `mrb_realloc_simple()` keeps the original `realloc`
behavior.
2020-06-25 06:57:43 +09:00
Yukihiro "Matz" Matsumoto
95360a17f2
Remove unnecessary break from numeric.c.
2020-06-25 06:57:42 +09:00
Yukihiro "Matz" Matsumoto
c07f24cd18
Change flag names in preparation of REnv refactoring.
2020-06-25 06:57:41 +09:00
Yukihiro "Matz" Matsumoto
03c2b8656b
Use mrb_get_argc() to improve performance.
2020-06-25 06:57:41 +09:00
Yukihiro "Matz" Matsumoto
c2b37e76b3
Skip mrb_get_args() in mrb_ary_{aget,aset} unless necessary.
...
Use simpler `mrb_get_argc()` and `mrb_get_arg1()` instead.
2020-06-22 21:47:23 +09:00
Yukihiro "Matz" Matsumoto
28e793496a
Support integer and float combination in mrb_equal().
2020-06-20 18:56:05 +09:00