```
../gc.c:2342:45: warning: comparison of integers of different signs: 'short' and 'size_t' (aka 'unsigned long') [-Wsign-compare]
GC_ASSERT(size_pools[pool_id].slot_size == slot_size);
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~
```
Add cast to short, because `GC_ASSERT`s in `size_pool_for_size`
already use cast to short.
```
../gc.c:2342:25: warning: array subscript is of type 'char' [-Wchar-subscripts]
GC_ASSERT(size_pools[pool_id].slot_size == slot_size);
^~~~~~~~
```
This change fixes `-v --yjit-stats`. Previously in this situation,
YJIT._print_stats wasn't defined as yjit.rb is not evaluated when there
is only "-v" and no Ruby code to run.
On -DRUBY_DEVEL builds, `ruby -v` can print extra info about the last
commit on a separate line, breaking some tests that expect a single
line. Assert only the first line instead.
Lets consider the following scenario:
~~~
irb(#<Test::Unit::AutoRunner::Runner:0x0000560f68afc3c8>):001:0> p suite
OpenSSL::TestEC
=> OpenSSL::TestEC
irb(#<Test::Unit::AutoRunner::Runner:0x0000560f68afc3c8>):002:0> p all_test_methods
["test_ECPrivateKey", "test_ECPrivateKey_encrypted", "test_PUBKEY", "test_check_key", "test_derive_key", "test_dh_compute_key", "test_dsa_sign_asn1_FIPS186_3", "test_ec_group", "test_ec_key", "test_ec_point", "test_ec_point_add", "test_ec_point_mul", "test_generate", "test_marshal", "test_sign_verify", "test_sign_verify_raw"]
=>
["test_ECPrivateKey",
"test_ECPrivateKey_encrypted",
"test_PUBKEY",
"test_check_key",
"test_derive_key",
"test_dh_compute_key",
"test_dsa_sign_asn1_FIPS186_3",
"test_ec_group",
"test_ec_key",
"test_ec_point",
"test_ec_point_add",
"test_ec_point_mul",
"test_generate",
"test_marshal",
"test_sign_verify",
"test_sign_verify_raw"]
irb(#<Test::Unit::AutoRunner::Runner:0x0000560f68afc3c8>):003:0> p filter
/\A(?=.*)(?!.*(?-mix:(?-mix:memory_leak)|(?-mix:OpenSSL::TestEC.test_check_key)))/
=> /\A(?=.*)(?!.*(?-mix:(?-mix:memory_leak)|(?-mix:OpenSSL::TestEC.test_check_key)))/
irb(#<Test::Unit::AutoRunner::Runner:0x0000560f68afc3c8>):004:0> method = "test_check_key"
=> "test_check_key"
~~~
The intention here is to exclude the `test_check_key` test case.
Unfortunately this does not work as expected, because the negative filter
is never checked:
~~~
irb(#<Test::Unit::AutoRunner::Runner:0x0000560f68afc3c8>):005:0> filter === method
=> true
irb(#<Test::Unit::AutoRunner::Runner:0x0000560f68afc3c8>):006:0> filter === "#{suite}##{method}"
=> false
irb(#<Test::Unit::AutoRunner::Runner:0x0000560f68afc3c8>):007:0> filter === method || filter === "#{suite}##{method}"
=> true
~~~
Therefore always filter against the fully qualified method name
`#{suite}##{method}`, which should provide the expected result.
However, if plain string filter is used, keep checking also only the
method name.
This resolves [Bug #16936].
This is a follow-up for commit 265c002239 ("Do not allocate
ractor-local storage in dfree function during GC", 2021-02-09).
The comparison with the default rb_random_mt_t is useless in the first
place, since it is never equal: no actual Random object is associated
with it.
[Bug #17653] https://bugs.ruby-lang.org/issues/17653
Even if it's newer than the running versions. Dev versions are not
released to rubygems.org, so the warning message suggests a command that
doesn't work. And dev versions are currently non deterministic
(2.3.0.dev can be many different versions), so the warning doesn't
really make sense at the moment.
https://github.com/rubygems/rubygems/commit/6f31af27ef
Now the formatter configuration is per Ractor. DefaultFormatter is used
if not set.
DefaultFormatter#message_for is now a class method to allow sub-Ractors
to call the method.
https://github.com/ruby/error_highlight/commit/9fbaa8ab7c
```ruby
r, w = IO.pipe
r.close
IO.for_fd(w.fileno).close
```
This code closes a file descriptor `w.fileno`, but `w` doesn't know
the closing. Another code can open same file descriptor with opening
file (`f`). After that, the `w` will GCed and `w.fileno` is closed
again, and `f.fileno` is closed too, so IO operations for `f` (`f.close`)
will cause EBADF.
To fix this issue, do this test in another process.
It is relatively well known that mac does not update its command line
tools, and make is no exception. They ship GNU make 3.x, which didn't
yet implemented GNUMAKEFLAGS.
Resort to MAKEFLAGS there.
In an effort to minimize build issues on non x64 platforms, we can
decide at build time to not build the bulk of YJIT. This should fix
obscure build errors like this one on riscv64:
yjit_asm.c:137:(.text+0x3fa): relocation truncated to fit: R_RISCV_PCREL_HI20 against `alloc_exec_mem'
We also don't need to bulid YJIT on `--disable-jit-support` builds.
One wrinkle to this is that the YJIT Ruby module will not be defined
when YJIT is stripped from the build. I think that's a fair change as
it's only meant to be used for YJIT development.
`duphash` showed up in the top-20 most frequent exit ops for @jhawthorn's benchmark that renders github.com/about
The implementation was almost exactly the same as `duparray`
Co-authored-by: John Hawthorn <john@hawthorn.email>
Co-authored-by: John Hawthorn <john@hawthorn.email>
Kazuhiro NISHIYAMA
Nobuyoshi Nakada
git
Alan Wu
John Hawthorn
Vít Ondruch
Satoshi Moris Tagomori
Kazuki Yamaguchi
David Rodríguez
Yusuke Endoh
Jean Boussier
Jeremy Evans
Peter Zhu
Steven Nunez
Shugo Maeda
Koichi Sasada
卜部昌平
Jenny Shen
Maxime Chevalier-Boisvert
Ian C. Anderson