* This partially reverts commit dead7478748a828c45e16134fca812bc7771344e.
* Windows will not allow a file to be unlinked if any file handles exist,
see https://github.com/ruby/ruby/pull/3597
The added test case crashes the interpreter because it makes
ObjectSpace.internal_class_of return the second VALUE slot of an AST
imemo object. The second VALUE slot of `struct rb_ast_struct` is
not a VALUE and not a pointer to a Ruby object.
by explicitly setting `Warning[:deprecated] = true`.
I removed "capture_io" at 79063d8cbfb7ce4740774289252a2a20dc9a5dc1, but
it printed the warning when `RUBYOPT=-w`.
This change makes the warnings enabled explicitly, capture and check the
warning.
The deprecation warning was disabled, and the code to check the warning
was removed at 996af2ce086249e904b2ce95ab2fcd1de7d757be, thus capture_io
is no longer needed.
* Add buffer protocol
* Modify for some review comments
* Per-object buffer availability
* Rename to MemoryView from Buffer and make compilable
* Support integral repeat count in memory view format
* Support 'x' for padding bytes
* Add rb_memory_view_parse_item_format
* Check type in rb_memory_view_register
* Update dependencies in common.mk
* Add test of MemoryView
* Add test of rb_memory_view_init_as_byte_array
* Add native size format test
* Add MemoryView test utilities
* Add test of rb_memory_view_fill_contiguous_strides
* Skip spaces in format string
* Support endianness specifiers
* Update documentation
* Support alignment
* Use RUBY_ALIGNOF
* Fix format parser to follow the pack format
* Support the _ modifier
* Parse count specifiers in get_format_size function.
* Use STRUCT_ALIGNOF
* Fix test
* Fix test
* Fix total size for the case with tail padding
* Fix rb_memory_view_get_item_pointer
* Fix rb_memory_view_parse_item_format again
The only remaining use of this function was to get the internal
message object from an exception's hidden `mesg` instance
variable to allow it to be dumped wiithout converting to a string.
As discussed in #103, this exposes internal implementation details
of CRuby, and ultimately does not provide any real utility to the
user since they can't directly inspect this hidden variable. The
test change here is to reflect CRuby behavior that denies equality
if the internal message objects do not match, as is the case after
the exception has been loaded and now has a simple String value.
The impact to users is that exceptions with special hidden message
objects will convert those objects to String during marshaling
through YAML. I believe this only affects NameError and its
descendants, since users can't set this field directly on their
own exception types.
Fixes#103.
JRuby's environment variables are provided by the Java Development
Kit's (JDK's) classes, which present them as a map from string to
string. In order to do this, those environment variable names and
values must be decoded into characters, which breaks any variables
that are intended to be "raw" bytes not necessarily decodable with
the default system encoding.
This issue is detailed in jruby/jruby#6248. The only solution on
the JRuby side will be to bypass the JDK environment variable API
and go directly to the native getenv/setenv system calls. This is
not likely to happen in the near future, due to the complexity of
such a change and the rarity of undecodable environment values.
The exclude here was added due to the Windows platform also having
a similar sensitivity to character encodings when working with
environment variables. It seems appropriate to expand this skip
to the "java" platform, as the root issue is largely the same.
https://github.com/ruby/webrick/commit/dc453e5c3c
On some environments that uses OpenSSL 1.1.1h, the two tests now fail.
http://rubyci.s3.amazonaws.com/android29-x86_64/ruby-master/log/20200924T062352Z.fail.html.gzhttps://github.com/ruby/ruby/runs/1159288773?check_suite_focus=true
```
1) Failure:
TestNetHTTPS#test_get [/data/data/com.termux/files/home/cb/tmp/build/20200924T062352Z/ruby/test/net/http/test_https.rb:47]:
<"0\x82\x03\xED0\x82\x02\xD5\xA0\x03..."> expected but was
<"0\x82\x03\xE30\x82\x02\xCB\xA0\x03...">.
```
Not sure why, but verify_callback now seems to receive only SERVER_CERT
but not CA_CERT.
It would be good to investigate the issue furthermore, but tentatively,
I want to stop the failures.
Receiving UnixSocket works fine if you don't provide a mode, and
I think it is reasonable to expect that you should not provide
a mode if klass.for_fd would not accept a mode.
Fixes [Bug #11778]
This reverts commit eeef16e190cdabc2ba474622720f8e3df7bac43b.
This also reverts the spec change.
Preventing the SystemStackError would be nice, but there is valid
code that the fix breaks, and it is probably more common than cases
that cause the SystemStackError.
Fixes [Bug #17182]
