This reverts commit a698181021.
It failed on macos for a unknown problem.
```
1) Error:
TestM17N#test_object_inspect_external:
Encoding::ConverterNotFoundError: code converter not found (US-ASCII to UTF-16BE)
/Users/runner/work/ruby/ruby/src/test/ruby/test_m17n.rb:312:in `encode'
/Users/runner/work/ruby/ruby/src/test/ruby/test_m17n.rb:312:in `inspect'
/Users/runner/work/ruby/ruby/src/test/ruby/test_m17n.rb:315:in `inspect'
/Users/runner/work/ruby/ruby/src/test/ruby/test_m17n.rb:315:in `<main>'
/Users/runner/work/ruby/ruby/src/test/ruby/test_m17n.rb:299:in `test_object_inspect_external'
```
4207871418
The test uses `Encoding.default_external = Encoding::UTF_16BE`, which
may add a wrongly UTF_16BE-encoded path to $LOADED_FEATURES (depending
on the order of tests). Unfortunately this breaks another test:
http://ci.rvm.jp/results/trunk-test@ruby-sky1/3711615
```
/tmp/ruby/v3/src/trunk-test/test/io/console/test_io_console.rb:11:in `===': incompatible encoding regexp match (US-ASCII regexp with UTF-16BE string) (Encoding::CompatibilityError)
```
According to @naruse-san, we don't pay effort to such a case, so this
change just avoids the issue by running the test in question under
another process.
Co-Authored-By: Koichi Sasada <ko1@atdot.net>
It reports "opt_regexpmatch2 insn is not included", but actually it is
included. This is due to a known bug of ISeq#to_a on which this check
depends.
https://bugs.ruby-lang.org/issues/18269
This commit fixes a memory leak introduced in an early part of the
variable width allocation project that would prevent the rb_classext_t
struct from being free'd when the class is swept.
52ab9bbee9 makes a failure.
```
1) Failure:
HTTPRequestTest#test_initialize_HEAD [/tmp/ruby/v3/src/trunk/test/net/http/test_http_request.rb:53]:
<{"accept"=>["*/*"], "user-agent"=>["Ruby"]}> expected but was
<{"accept-encoding"=>["gzip;q=1.0,deflate;q=0.6,identity;q=0.3"],
"accept"=>["*/*"],
"user-agent"=>["Ruby"]}>.
```
I'm not sure it is correct, so please review later.
test_connect_certificate_verify_failed_exception_message occasionally
fails. Is it possible that OpenSSL sees a different clock from Ruby by
more than 10 seconds?
http://ci.rvm.jp/logfiles/brlog.trunk-random0.20211111-072828
Let's give more time after the certificate expiration date to see if
this fixes the flakiness. We had similar occasional failures in
test_x509store.rb before, which disappeared after ruby/ruby commit
7930a352a5 and ruby/openssl commit fb2fcbb13734.
Follow up of 428227472f. The previous fix
uses `rb_ary_new_from_values` to create the result array, but it may
trigger the GC.
This second try is to create the result array by `rb_ary_new_capa`
before the second iteration, and assume that `rb_ary_push` does not
trigger GC. This assumption is very fragile, so should be improved in
future.
[Bug #18282] [Feature #14394]
The `parse_make_command_line` in `assert_contains_make_command` fails to get
the make targets correctly, when the make command is set with make options by
environment variable such as `export make='make -j2'` at
lib/rubygems/ext/builder.rb::make.
So, we include the make options (eg, -XX) as a part of the command to fix the
case. Note that this commit still doesn't fix the case of
`export make='make -j 2'`.
7730ef3fa0
20211109T063003Z.log.html.gz
```
[ 9898/21145] TestISeq#test_super_with_block_and_kwrest/home/chkbuild/chkbuild/tmp/build/20211109T063003Z/ruby/test/ruby/test_iseq.rb:1: warning: method redefined; discarding old touch
/home/chkbuild/chkbuild/tmp/build/20211109T063003Z/ruby/test/ruby/test_iseq.rb:1: warning: previous definition of touch was here
= 0.00 s
```
20211109T063003Z.log.html.gz
```
[ 1960/21145] TestEnumerable#test_to_a_keywords/home/chkbuild/chkbuild/tmp/build/20211109T063003Z/ruby/test/ruby/test_enum.rb:138: warning: method redefined; discarding old each
/home/chkbuild/chkbuild/tmp/build/20211109T063003Z/ruby/test/ruby/test_enum.rb:11: warning: previous definition of each was here
= 0.00 s
```
GC must not be triggered during callback of rb_class_foreach_subclass.
To prevent GC, we can not use rb_ary_push. Instead, this changeset calls
rb_class_foreach_subclass twice: first counts the subclasses, then
allocates a buffer (which may cause GC and reduce subclasses, but not
increase), and finally stores the subclasses to the buffer.
[Bug #18282] [Feature #14394]
... because readline library (or libtinfo) seems to cause SEGV
internally even with Readline 8.1.
This SEGV is caught as SystemStackError, which leads to a test failure.
20211109T030008Z.fail.html.gz
```
1) Failure:
TestReadline#test_interrupt_in_other_thread [/home/chkbuild/chkbuild/tmp/build/20211109T030008Z/ruby/test/readline/test_readline.rb:568]:
Unknown failure with exit status pid 13097 exit 1
Log: ** START **Readline::VERSION is 8.1.
/home/chkbuild/chkbuild/tmp/build/20211109T030008Z/tmp/interrupt_in_other_thread20211109-6272-kcfqxo:13:in `readline': stack level too deep (SystemStackError)
from /home/chkbuild/chkbuild/tmp/build/20211109T030008Z/tmp/interrupt_in_other_thread20211109-6272-kcfqxo:13:in `block in <main>'
/home/chkbuild/chkbuild/tmp/build/20211109T030008Z/tmp/interrupt_in_other_thread20211109-6272-kcfqxo:13:in `readline': stack level too deep (SystemStackError)
from /home/chkbuild/chkbuild/tmp/build/20211109T030008Z/tmp/interrupt_in_other_thread20211109-6272-kcfqxo:13:in `block in <main>'
```
```ruby
def foo(*); ->{ super }; end
```
This code makes anonymous parameters which is not registered as an
ID. The problem is that when Ractors try to scan `getlocal`
instructions, it puts the Symbol corresponding to the parameter
in to a hash. Since it is not registered, we end up with a
strange exception. This commit wraps the unregistered ID in an
internal ID so that we get the same exception for `...` as `*`.
Co-Authored-By: Aaron Patterson <tenderlove@ruby-lang.org>
Co-Authored-By: John Hawthorn <john@hawthorn.email>
Standard usage should be tested by default. I will add a test for
`--prefix` once I fix that option since it's not working correctly at
the moment.
50a7e34586
The test randomly gets stuck on Solaris:
20211105T060014Z.fail.html.gz
```
1) Error:
TestIRB::TestInit#test_recovery_sigint:
Timeout::Error: execution of assert_in_out_err expired timeout (100.0 sec)
pid 3195 killed by SIGTERM (signal 15)
| Switch to inspect mode.
| exit
|
```
I investigated the issue but I couldn't figure it out.
This random failure is noisy, which makes it difficult to track the CI.
So I skip the test on Soalris. Contribution is welcome.
Prior to this patch, if I ran:
ruby setup.rb --destdir /foo
Then Bundler files would be written into /foo/foo, because destdir was
being prepended, even though `bundler_spec.bin_dir` already included
destdir.
9e857ffb52
The compaction tests get stuck randomly on s390x for unknown reason.
20211104T030003Z.fail.html.gz
```
[13715/21145] TestGCCompact#test_gc_compact_statstimeout: output interval exceeds 1800.0 seconds.
```
We spent some time to investigate this issue, but we can't figure out
why, and it is unlikely that we'll be able to fix it anytime soon.
This random failure makes the CI unuseful, so tentatively we suppress
this test for a while. A contribution from those who are familiar with
s390x is welcome.
The last_communicated timestamp is for HTTP persistent connection, to
decide whether the current TCP connection may be reused for the
subsequent requests or not. Naturally, the timer must be reset if the
connection is recreated since it is no longer relevant.
0a013de42d
