[Bug #17516]
`fork(2)` only leave the calling thread alive in the child.
Because of this forking from the non-main ractor can easily
leave the VM in a corrupted state.
It may be possible in the future to carefully allow forking from non-main
Ractor, but shot term it's preferable to add this restriction.
The `FL_FREEZE` flag is redundant with `SHAPE_ID_FL_FROZEN`, so
ideally it should be eliminated in favor of the later.
Doing so would eliminate the risk of desync between the two, but
also solve the problem of the frozen status being global in namespace
context (See Bug #21330).
There are few legitimate use cases for duplicate keys, and can
in some case be exploited.
Rather to always silently accept them, we should emit a warning,
and in the future require to explictly allow them.
06f00a42e8
Since we no longer pass ruby CLI flags in our spec commands, we no
longer need the previous workaround and can get the realworld code
tested.
fd92c855fb
Before the patch:
```
$ ./miniruby -e '[1, 2].inject(:tap)'
-e:1:in '<main>': wrong number of arguments (given 1, expected 0) (ArgumentError)
from -e:1:in 'Enumerable#inject'
from -e:1:in '<main>'
```
After the patch:
```
$ ./miniruby -e '[1, 2].inject(:tap)'
-e:1:in 'Kernel#tap': wrong number of arguments (given 1, expected 0) (ArgumentError)
from -e:1:in 'Enumerable#inject'
from -e:1:in '<main>'
```
Fixes https://bugs.ruby-lang.org/issues/20968#change-113811
Follow up to fix 3b7373fd00.
In that commit, the line number in the first frame was overwritten after
the whole backtrace was created. There was a problem that the line
number was overwritten even if the location was backpatched.
Instead, this commit uses first_lineno if the frame is
VM_FRAME_MAGIC_DUMMY when generating the backtrace.
Before the patch:
```
$ ./miniruby -e '[1, 2].inject(:tap)'
-e:in '<main>': wrong number of arguments (given 1, expected 0) (ArgumentError)
from -e:1:in 'Enumerable#inject'
from -e:1:in '<main>'
```
After the patch:
```
$ ./miniruby -e '[1, 2].inject(:tap)'
-e:1:in '<main>': wrong number of arguments (given 1, expected 0) (ArgumentError)
from -e:1:in 'Enumerable#inject'
from -e:1:in '<main>'
```
* [Bug #21449] Fix Set#divide{|a,b|} using Union-find structure
Implements Union-find structure with path compression.
Since divide{|a,b|} calls the given block n**2 times in the worst case, there is no need to implement union-by-rank or union-by-size optimization.
* Avoid internal arrays from being modified from block passed to Set#divide
Internal arrays can be modified from yielded block through ObjectSpace.
Freeze readonly array, use ALLOCV_N instead of mutable array.
Pop two values from the stack, return the first if it is a string,
otherwise return string coercion of the second
Also piggybacks a fix for string subclasses skipping `to_s` for
`objtostring`.
Co-authored-by: composerinteralia <composerinteralia@github.com>
Even when namespaces are enabled, only a few core classes created
during init will eventually be namespaced.
For these it's OK to allocate a 320B slot to hold the extra namespace
stuff.
But for any class created post init, we know we'll never need the
namespace and we can fit in a 160B slot.
Rather than to lazily check the env using a trinary
value, we can more straightforwardly check for the
env during the VM boot.
This allow `rb_namespace_available` to just be a pointer
dereference.
As showed by the unskiped spec, on Windows trying to use the 0.0.0.0
interface raises this error, and it's raised as a generic system error
when trying to create a `bundler.lock` file. Here's is a better place to
handle that.
e32c5a9e5c
The mirror probing spec file was moved to our regular suite, which runs
in parallel, recently. These specs rely on starting and stopping actual
servers in localhost, but this does not play nice with parallelization,
because a spec may kill the webserver another spec has created.
This commit moves mirror probing specs to not need to start servers in
localhost and do something more similar to what the other specs do.
ca9a19706f
If all nodes in the array are safe, then it is safe to avoid
allocation for the positional splat:
```ruby
m(*a, kw: [:a]) # Safe
m(*a, kw: [meth]) # Unsafe
```
This avoids an unnecessary allocation in a Rails method call.
Details: https://github.com/rails/rails/pull/54949/files#r2052645431
This is a follow-up to commit b120f5e38d (avoid fork-unsafe arc4random
implementations, 2018-09-04).
Avoid defining a no-op fill_random_bytes_syscall() if arc4random_buf(3)
exists, but we are unsure if it is fork-safe. Check for other options
instead. IOW, see if getrandom(2) is available.
glibc 2.36, released in 2022, started to provide arc4random_buf(3) on
Linux. This causes fill_random_bytes_syscall() to use neither of them
and makes Random.urandom solely rely on getentropy(3) via
fill_random_bytes_urandom().
While the glibc implementation is safe, I did not add it to the list
because using getrandom(2) directly is preferable on Linux.
If this is not a system call, then it is using getrandom (which would
have been tried already), and cannot be used as a replacement for the
random devices.
