| Age | Commit message (Collapse) | Author |
|---|
|
The lookup in the table is using the wrong key when converting generic
instance variables to too complex, which means that it never looks up
the entry which leaks memory when the entry is overwritten.
|
|
|
|
|
|
|
|
Add a special treatment for when the argument of self is an
integral multiple of 45 degrees.
1i ** (10 ** 100) #=> 1+0i
1i ** (10 ** 100 + 1) #=> 0+1i
(1+1i) ** (10 ** 100) # warning: in a**b, b may be too big
#=> (Infinity+0.0i)
(1+1i) ** (10 ** 100 + 1) # warning: in a**b, b may be too big
#=> (Infinity+Infinity*i)
|
|
When transitioning generic instance variable objects to too complex, we
set the shape first before performing inserting the new gen_ivtbl. The
st_insert for the new gen_ivtbl could allocate and cause a GC. If that
happens, then it will crash because the object will have a too complex
shape but not yet be backed by a st_table.
This commit changes the order so that the insert happens first before
the new shape is set.
The following script reproduces the issue:
```
o = []
o.instance_variable_set(:@a, 1)
i = 0
o = Object.new
while RubyVM::Shape.shapes_available > 0
o.instance_variable_set(:"@i#{i}", 1)
i += 1
end
ary = 1_000.times.map { [] }
GC.stress = true
ary.each do |o|
o.instance_variable_set(:@a, 1)
o.instance_variable_set(:@b, 1)
end
```
|
|
|
|
Prior to this commit, we weren't recursing up scopes to look for
the local definition. With this commit, we do so, fixing local writes
within blocks
|
|
This PR implements the once node on interpolated regexes.
There is a bug in Prism where the interpolated regex with the once flag
only works when there is not a local variable so the test uses a "1".
We'll need to fix that.
|
|
|
|
|
|
Commit e87d0882910001ef3b0c2ccd43bf00cee8c34a0c introduced a
regression where the keyword splat object passed by the caller
would be directly used by callee as keyword splat parameters,
if it implemented #to_hash. The return value of #to_hash would be
ignored in this case.
|
|
Reproduction script:
```
o = Object.new
10.times { |i| o.instance_variable_set(:"@a#{i}", i) }
i = 0
a = Object.new
while RubyVM::Shape.shapes_available > 2
a.instance_variable_set(:"@i#{i}", 1)
i += 1
end
o.remove_instance_variable(:@a0)
puts o.instance_variable_get(:@a1)
```
Before this patch, it would incorrectly output `2` and now it correctly
outputs `1`.
|
|
due to a failure on a CI
http://ci.rvm.jp/results/trunk-iseq_binary@ruby-sp2-docker/4779277
```
expected:
== disasm: #<ISeq:prism_test_forwarding_arguments_node1@<compiled>:2 (2,8)-(4,11)>
local table (size: 1, argc: 0 [opts: 0, rest: -1, post: 0, block: -1, kw: -1@-1, kwrest: 1])
[ 1] "..."@0
0000 putself ( 3)
0001 getlocal_WC_0 ?@-2
0003 splatarray false
0005 getblockparamproxy ?@-1, 0
0008 send <calldata!mid:prism_test_forwarding_arguments_node, argc:1, ARGS_SPLAT|ARGS_BLOCKARG|FCALL>, nil
0011 leave ( 2)
actual:
== disasm: #<ISeq:prism_test_forwarding_arguments_node1@<compiled>:2 (2,8)-(4,11)>
local table (size: 1, argc: 0 [opts: 0, rest: -1, post: 0, block: -1, kw: -1@-1, kwrest: 1])
[ 1] "..."@0
0000 putself ( 3)
0001 getlocal_WC_0 ?@-2
0003 splatarray false
0005 getblockparamproxy "!"@-1, 0
0008 send <calldata!mid:prism_test_forwarding_arguments_node, argc:1, ARGS_SPLAT|ARGS_BLOCKARG|FCALL>, nil
0011 leave ( 2)
/tmp/ruby/src/trunk-iseq_binary/tool/lib/iseq_loader_checker.rb:36:in `exit': exit (SystemExit)
```
|
|
|
|
|
|
`test_thread_trace` is also flaky due to the same reason as #8916.
https://rubyci.s3.amazonaws.com/centos7/ruby-master/log/20231114T213002Z.fail.html.gz
https://rubyci.s3.amazonaws.com/wsl2/ruby-master/log/20231114T090003Z.fail.html.gz
|
|
This test suite is flaky on CI, because the test asserts events from
finalizers also. Tracing events from finalizers is not deterministic
and is not a part of test interests, so this commit changes the test
to assert only events originated from the test file itself.
|
|
Other objects may be using the shape, so we can't change the capacity
otherwise the other objects may have a buffer overflow.
|
|
For better assert failure diagnostics.
|
|
|
|
|
|
Some code out there blind calls `force_encoding` without checking
what the original encoding was, which clears the coderange uselessly.
If the String is big, it can be a rather costly mistake.
For instance the `rack-utf8_sanitizer` gem does this on request
bodies.
|
|
This commit adds tests for the compilation of the
OptionalKeywordParameterNode, and fixes cases on the
RequiredKeywordParameterNode
|
|
|
|
This commit adds tests for BlockParameterNode, RequiredParameterNode,
RequiredKeywordParameterNode and RestParameterNode
|
|
Co-authored-by: Alan Wu <alansi.xingwu@shopify.com>
Co-authored-by: Maxime Chevalier-Boisvert <maxime.chevalierboisvert@shopify.com>
|
|
Apply Nobu's suggestions which improve style, memory handling and error correction.
Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
|
|
This drops the total size of a Time object from 86 bytes to 80 bytes.
Running the benchmark benchmark/time_now.yml, this commit improves
performance of Time.now by about 30%:
```
Time.now
Branch: 13159405.4 i/s
Master: 10036908.7 i/s - 1.31x slower
Time.now(in: "+09:00")
Branch: 2712172.6 i/s
Master: 2138637.9 i/s - 1.27x slower
```
It also decreases memory usage by about 20%:
```
ary = 10_000_000.times.map { Time.now }
puts `ps -o rss= -p #{$$}`
```
Branch: 961792
Master: 1196544
Co-Authored-By: Jean Boussier <byroot@ruby-lang.org>
|
|
This commit compiles most parameter types, setting appropriate values on
the ISEQ_BODY. It also adds tests for callers and callees of methods,
using many versions of tests from bootstraptest
|
|
|
|
Compilation now works for MultiWriteNodes and MultiTargetNodes, with
nesting on MultiWrites. See the tests added in this commit for example
behavior.
|
|
|
|
`Kernel#require` converts feature name objects that have the `to_path`
method such as `Pathname`, but had used the original object on error
and had resulted in an unexpected `TypeError`.
|
|
first_column and last_column return byte positions, but existing implementations
did not consider multibyte.
|
|
When a begin node is popped it only needs to putnil if that nil is going
to be the return value, otherwise it can successfully be optimised out.
|
|
It was assuming only objects can be complex.
|
|
Previously emitting a call node with an argument followed by another
node would cause the argument to be mistakenly omitted from the argument
list causing a stack underflow.
```
PRISM: **************************************************
-- raw disasm--------
0000 putself ( 0)
0001 send <calldata:puts, 1>, nil ( 0)
* 0004 pop ( 0)
0005 putobject 1 ( 0)
0007 leave ( 0)
---------------------
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Unfortunately, 84dfa0fa5e451adb87beaf497165cb5a1bc93770 introduced
a bug where we were no longer testing the "popped" case because the
"; 1" meant to be appended to the source was no longer functioning
as intended.
This commit re-introduces the popped case, and comments out all
now failing tests.
|
