* Fix [Bug #19632]: Disable external iterator for frozen enumerator
Currently, methods to manipulate an external iterator like `#next`
and `#feed` can be called even if a receiver of an enumerator is
frozen. However, these methods change the state of an external
iterator in an enumerator. Therefore, it seems a BUG to me, and
these methods should raise FrozenError if the receiver is frozen.
This fixed the following methods to raise FrozenError if the receiver is
frozen.
- `Enumerator#next`
- `Enumerator#next_values`
- `Enumerator#peek`
- `Enumerator#peek_values`
- `Enumerator#feed`
- `Enumerator#rewind`
* Fix a typo in the document
Thanks @Maumagnaguagno.
There is no longer a limit on the number of IVs you can store.
SHAPE_MAX_NUM_IVS was used to work around the IV10K problem (the well
known problem where setting 10k instance variables in a row would be too
slow). The redblack tree works well at any shape depth, even depths
greater than 80, and solves the IV10K problem.
String locking with locktmp is not really part of the public API,
and the test relies in a side effect of using it to protect the
buffer. On other implementations without locktmp this does not
fail. Separate into its own test so it can be excluded from public
API expectations.
Because the `&` call checks for interrupts, the test was accidentally
timing dependent. Stop checking for exits.
[Bug #19921]
Reported-by: Vít Ondruch <vondruch@redhat.com>
Reported-by: Mamoru Tasaka <mtasaka@fedoraproject.org>
The for-human stats summaries are not relevant for the children
`test_yjit.rb` spawns. Avoid compiling and running the printing code.
On a -O0 dev build this halves the time for `test_yjit.rb` on my machine.
For <https://bugs.ruby-lang.org/issues/19921>, I suspect the test is
failing due to a timing related interrupt, which on paper could
happen with slow-enough GC runs.
In any case, it's helpful for debugging to have more information when
tests fail.
Co-authored-by: Takashi Kokubun <takashikkbn@gmail.com>
They are too unstable on the machine.
```
1) Failure:
TestRegexp#test_timeout_shorter_than_global [/home/chkbuild/chkbuild/tmp/build/20231018T230003Z/ruby/test/ruby/test_regexp.rb:1788]:
Expected |0.2 - 0.962938869| (0.7629388690000001) to be <= 0.15000000000000002.
```
20231018T230003Z.fail.html.gz
```
1) Failure:
TestRegexp#test_timeout_longer_than_global [/home/chkbuild/chkbuild/tmp/build/20231017T140006Z/ruby/test/ruby/test_regexp.rb:1788]:
Expected |0.5 - 1.040696078| (0.5406960780000001) to be <= 0.375.
```
20231017T140006Z.fail.html.gz
When we eval the iseqs generated by prism, they can have side effects
like defining methods. In this case, we were defining the method "foo",
but other tests were expecting that the foo method would _not_ be
defined.
We changed ScopeNodes to point to their parent (previous) ScopeNodes.
Accordingly, we can remove pm_compile_context_t, and store all
necessary context in ScopeNodes, allowing us to access locals from
outer scopes.
These are necessary to get the tests passing with LibreSSL 3.8.1+,
which dropped support for TLSv1.0 and TLSv1.1 for security reasons.
This updates the tests to use TLSv1.2 on OpenBSD. This is only
strictly necessary on OpenBSD 7.4+, but it will work fine in previous
versions as well.
32707b2db5
Right now, our Call{Operator,And,Or}WriteNode nodes represent two
different concepts:
```ruby
foo.bar += 1
foo[bar] += 1
```
These two statements are different in what they can support. The
former can never have arguments (or an opening_loc or closing_loc).
The former can also never have a block. Also, the former is a
variable method name.
The latter is always going to be []/[]=, it can have any number of
arguments including blocks (`foo[&bar] ||= 1`), and will always
have an opening_loc and closing_loc.
Furthermore, these statements end of having to take different paths
through the various compilers because with the latter you have to
consider the arguments and the block, whereas the former can
perform some additional peephole optimizations since there are
fewer values on the stack.
For these reasons, I'm introducing Index{Operator,And,Or}WriteNode.
These nodes never have a read_name or write_name on them because
they are always []/[]=. They also support blocks, which the previous
write nodes didn't. As a benefit of introducing these nodes, I've
removed the opening_loc, closing_loc, and arguments from the older
write nodes because they will always be null.
For the serialized format, both of these nodes end up being
smaller, and for in-memory we're storing fewer things in general,
so we have savings all around.
I don't love that we are introducing another node that is a call
node since we generally want consumers to only have to handle a
single call, but these nodes are so specific that they would have
to be handled separately anyway since in fact call 2 methods.
70155db9cd
Currently, you can install multiple versions of the same gem just fine:
```
$ gem install simplecov:0.19.0 simplecov:0.22.0
Fetching simplecov-0.19.0.gem
Successfully installed simplecov-0.19.0
Parsing documentation for simplecov-0.19.0
Installing ri documentation for simplecov-0.19.0
Done installing documentation for simplecov after 0 seconds
Fetching simplecov-0.22.0.gem
Successfully installed simplecov-0.22.0
Parsing documentation for simplecov-0.22.0
Installing ri documentation for simplecov-0.22.0
Done installing documentation for simplecov after 0 seconds
2 gems installed
```
But to uninstall both of them, you need to run the equivalent uninstall
command twice:
```
~$ gem uninstall simplecov:0.19.0 simplecov:0.22.0
Successfully uninstalled simplecov-0.22.0
~$ gem uninstall simplecov:0.19.0 simplecov:0.22.0
Gem 'simplecov' is not installed
Successfully uninstalled simplecov-0.19.0
```
This resolves that problem by using the gem's full name (which includes
the version) when tracking which ones have already been uninstalled so
when it gets to the second version listed it doesn't think it was
already uninstalled.
d96101b753
If `assert_equal(backtrace_locations.size, profile_frames.size)` in
`TestProfileFrames#test_matches_backtrace_locations_main_thread`
failed, we do not have enough information about it like that:
```
1) Failure:
TestProfileFrames#test_matches_backtrace_locations_main_thread [/home/runner/work/ruby/ruby/src/test/-ext-/debug/test_profile_frames.rb:148]:
<31> expected but was
<30>.
```
This patch shows both `backtrace_locations` and `profile_frames`
if failed.
Some people encounter an issue that test_yjit uses the installed Ruby
instead of the currently-running Ruby. It's fixed when they remove the
installed Ruby.
However, test_yjit should run the currently-running Ruby for testing
YJIT in subprocesses. EnvUtil is unfortunately used outside tests as
well, so for compatibility reasons, this commit only changes the
argument given to EnvUtil.invoke_ruby to always use RbConfig.ruby.
Co-authored-by: Alan Wu <XrXr@users.noreply.github.com>
If `RUBY_MN_THREADS=1` is given, this patch shows `+MN` in
`RUBY_DESCRIPTION` like:
```
$ RUBY_MN_THREADS=1 ./miniruby --yjit -v
ruby 3.3.0dev (2023-10-17T04:10:14Z master 908f8fffa2) +YJIT +MN [x86_64-linux]
```
Before this patch, a warning is displayed if `$VERBOSE` is given.
However it can make troubles with tests (with `$VERBOSE`), do not
show any warning with a MN threads configuration.
This comment previously specified TLS 1.2, but actually set the
version to TLS 1.0. LibreSSL 3.8.1 (included in OpenBSD 7.4)
dropped support for TLS 1.0/1.1 for security reasons, which
broke this test. Switch the test to use TLS 1.2 as documented
so it will continue to work on OpenBSD 7.4+.
97be4de53a
