Activity
From 06/17/2025 to 06/23/2025
06/23/2025
-
11:55 PM Revision 74e6bddf (git): ZJIT: Parse putspecialobject(VMCore) into Const (#13683)
- This way we get more information in HIR for the optimizer.
Fix https://github.com/Shopify/ruby/issues/587 -
- If we have a frozen array `[..., a, ...]` and a compile-time fixnum index `i`,
we can do the array load at compile-time. -
08:17 PM Misc #21399: DevMeeting-2025-07-10
- * [Feature #17473] Make Pathname to embedded class of Ruby (eregon)
* Related to that issue and discussed there, I made a PR to define most of Pathname in Ruby: https://github.com/ruby/pathname/pull/53
* It's faster, easier to read a... -
- I (finally) made a PR to define most of Pathname in Ruby: https://github.com/ruby/pathname/pull/53
-
07:07 PM Bug #21398: Ractor.select hangs when multiple threads submit heavy jobs concurrently
- I sent a [PR](https://github.com/ruby/ruby/pull/13682) for this but I don't know if this is the proper fix or if we should use a different mutex to wait for condvar for DNTs. I did, however, figure out why it's happening. @ko1 what are y...
-
05:43 AM Bug #21398: Ractor.select hangs when multiple threads submit heavy jobs concurrently
I'm experiencing a hang when combining `Ractor::Port` and multiple `Thread`s in Ruby 3.5.0dev (ruby-head). The script below is a modified version of the one shared in comment #1 of this issue, with adjusted parameters.
When I increase...-
06:20 PM Bug #20714: Handle optional dependencies in `bundled_gems.rb`
- For what it's worth, I did end up considering my PR a strict improvement over what we have now, but I just closed it because it was ignored for a long time. Hopefully we'll have a proper solution in the future.
-
05:56 PM Bug #21449 (Closed): Set#divide is order dependent and not working correctly
- Applied in changeset commit:git|67346a7d94b101acc00c177b01ad0aabfef605a8.
----------
[Bug #21449] Fix Set#divide{|a,b|} using Union-find structure (#13680)
* [Bug #21449] Fix Set#divide{|a,b|} using Union-find structure
Implements Uni... -
- pull request https://github.com/ruby/ruby/pull/13680
-
- ~~~ruby
Set[0,1,2,3].divide{(_1 - _2).abs == 1}
#=> #<Set: {#<Set: {0, 1, 2, 3}>}>
Set[0,2,3,1].divide{(_1 - _2).abs == 1}
#=> #<Set: {#<Set: {0, 1, 2, 3}>}> (ruby 3.4)
#=> #<Set: {#<Set: {0, 1}>, #<Set: {2, 3, 1}>}> (ruby 3.5.0de... -
- * [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 o... -
05:28 PM Revision db6f3979 (git): Tweaks for String#bytesize
-
03:20 PM Revision b1ce569f (git): ZJIT: `anytostring` to HIR (GH-13658)
- 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 <compose... -
03:15 PM Bug #21450 (Closed): Inconsistent `upcase` between `String` and `Symbol`
- Behavior for `Symbol#upcase` and `String#upcase` differs for `i` character if the `:turkic` option is present
I'd expect `val.upcase(:turkic)` behaves consistently for both cases:
```ruby
'i'.upcase(:turkic)
# "İ" with dot
:i.... -
11:03 AM Revision af6b98f7 (git): Make the critical level an enum
-
09:04 AM Revision c6dd07d6 (git): Allocate singleton classes as namespaceable if their parent are
-
- 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'l... -
-
-
-
- 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. -
-
02:07 AM Revision a1d62a3b (git): [rubygems/rubygems] Handle RubyGems installing to custom dir with non-existent parent dirs
- https://github.com/rubygems/rubygems/commit/4701123601
-
- 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.
https://gi... -
- https://github.com/rubygems/rubygems/commit/bfa6770e39
-
- https://github.com/rubygems/rubygems/commit/627a7615f2
-
- https://github.com/rubygems/rubygems/commit/57e8ae7aa6
-
- https://github.com/rubygems/rubygems/commit/380c95ce05
-
- https://github.com/rubygems/rubygems/commit/b556167793
-
- 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 ki... -
- rhenium (Kazuki Yamaguchi) wrote in #note-2:
> FWIW `getentropy()` has recently made it into POSIX (with a slight difference in errno behavior compared to existing implementations). I think supporting it as an alternative makes sense, n...
06/22/2025
-
- @mame Here is an example bug due to this change, based on the `inject` case I mentioned in https://bugs.ruby-lang.org/issues/20968#note-6.
I think this is strong enough to revert that change.
ruby-master:
```
$ ruby -e '[1,2].injec... -
06/21/2025
-
09:43 PM Revision 353fa6f0 (git): Avoid allocation for positional splat for literal array keyword argument
- 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: ht... -
- First try dedicated system calls, such as `getrandom` or `getentropy`,
next possible libraries, then fallback to `/dev/urandom`. -
- - `ractor_sync_terminate_atfork` is unused unless fork is working
- `cr` in `vm_lock_leave` is only for debugging -
04:07 PM Bug #21445: [BUG] push_mark_stack() called for broken object raised since cd9f447be247478d2eb3da985295735cce20cb23
- Cool, that works (for me at least). You can also close https://bugs.ruby-lang.org/issues/21444, since it was the same problem.
-
- Applied in changeset commit:git|edbd9ed46842b8cd811fba45a373bb39ee155bfb.
----------
variable.c: avoid out of bound write in `generic_field_set`
[Bug #21445] -
- Thanks for the even simpler repro. I got a fix: https://github.com/ruby/ruby/pull/13675
-
- I was just trying to reduce my issue in https://bugs.ruby-lang.org/issues/21444 and while I did manage to do so, I now ended up at the assertion from this bug report. So I guess they are the same issue:
```rb
class Foo
def initialize
... -
- > So there must be something more than just a T_DATA.
The common denominator between the two impacted gems seem to be that they don't set: `RUBY_TYPED_FREE_IMMEDIATELY`. So something must be happening during the zombie phase. -
- Much smaller repro:
```ruby
require "sqlite3"
10.times do
db = SQLite3::Database.new(":memory:")
db.__id__
db = nil
GC.stress = true
Object.new
GC.stress = false
end
``` -
- Another observation:
- It reproduce with `sqlite3` and `mysql2`, but not with `postgresql` nor `trilogy`. So there must be something more than just a `T_DATA`. -
- I had a bit of time to look at that this morning.
A few observations:
- It's triggered by a very specific test, that calls `__id__` on `T_DATA` objects.
- I can make it consistently fail with `GC.stress`: https://github.com/by... -
- [Bug #21445]
-
- Applied in changeset commit:git|1181a682a6c314c92686e3701defa1eb44068c4e.
----------
[Bug #21448] Use `getentropy(2)` only on macOS
If this is not a system call, then it is using getrandom (which would
have been tried already), and can... -
08:24 AM Bug #21448: Random.urandom may fail to fall back to reading /dev/urandom on Linux < 3.17
- FWIW `getentropy()` has recently made it into POSIX (with a slight difference in errno behavior compared to existing implementations). I think supporting it as an alternative makes sense, not just for macOS.
https://pubs.opengroup.org... -
- Unfortunately, despite the same signature, `getentropy` is different thing in different layers depending on the platform.
* `getentropy(3)` - a library function on Linux:
> The `getentropy()` function is implemented using `getr... -
- This is a follow-up to commit b120f5e38d9c (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... -
- Probably the same as https://bugs.ruby-lang.org/issues/21445, during reduction I started hitting `push_mark_stack() called for broken object` instead.
-
- 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. -
-
- Don't confuse them with other nearby instructions.
06/20/2025
-
11:47 PM Bug #21447 (Closed): Fix handling of PM_CONSTANT_PATH_NODE node in keyword arguments with ARGS_SPLAT
- Fixed by commit:1d94a9e1a4351e01f851dad250ba97dad859ee70
-
- I added an optimization to reduce implicit allocations when using parse.y in commit:3de20efc308cccc38bf9dacfffca6c626d039a06. An implementation was added for prism in commit:14bb376b79ca9140bfd36dbbad06be045f31f92e. Unfortunately, when...
-
- This was handled correctly in parse.y (NODE_COLON2), but not in
prism. This wasn't caught earlier, because I only added tests for
the optimized case and not the unoptimized case. Add tests for
the unoptimized case.
In code terms:
```ru... -
- Add methods for setting supported signature algorithms, corresponding
to SSL_CTX_set1_sigalgs_list() and SSL_CTX_set1_client_sigalgs_list(),
respectively.
https://github.com/ruby/openssl/commit/6bbe58c492
Co-authored-by: Markus Jung <m... -
- https://github.com/ruby/openssl/commit/54f22395e7
-
- https://github.com/ruby/openssl/commit/9c9333c07d
-
05:37 PM Bug #21375: Set[] does not call #initialize
- Ethan (Ethan -) wrote in #note-3:
> My feeling is that the base expectation is that #initialize is called on new objects.
I agree.
`Array[*args]` is equivalent to `Array.new.concat(*args)` and in that case there was no need to call #ini... - 05:13 PM Revision edfcbaca (git): thread_cleanup: set CFP to NULL before clearing ec's stack
- We clear the CFP first so that if a sampling profiler interrupts the current thread during `rb_ec_set_vm_stack`,
`thread_profile_frames` returns early instead of trying to walk the stack that's no longer set on the ec.
The early return ... -
04:52 PM Revision b6babd9a (git): ZJIT: Typofix (#13665)
-
- Assigning to myself, but I might not have time to look into it before Tuesday so feel free to investigate if you so desire.
-
- Origianlly reported for tmpdir: https://github.com/ruby/tmpdir/issues/50
On Linux, `Random.urandom` is expected to first attempt the `getrandom(2)` syscall (Linux >= 3.17), and fall back to reading from `/dev/urandom` if it is not sup... -
-
- It's something we don't expect and might be coming from somewhere
else.
https://github.com/ruby/io-console/commit/f0646b2b6a -
- v1.245.0 is not working with the following issue:
https://github.com/ruby/ruby/actions/runs/15769771616/job/44464179119?pr=13661
```
D:/a/ruby/ruby/src/test/rubygems/mock_gem_ui.rb:83:in 'Gem::MockGemUi#terminate_interaction': Gem::Moc... -
-
- As well as compilers/entrypoint.sh.
-
-
-
- 04:52 AM Revision 8ce65463 (git): Update bundled gems list as of 2025-06-20
-
-
-
-
-
-
- We can fuse linked lists of blocks. This can be run any time, improves
future analyses, improves codegen, and also makes the HIR output look
nicer.
Inspired by my implementation of CleanCFG for Cinder, which was itself
inspired by Brett... -
- It does not seem to need to reopen the refinement.
Reproduces just by changing visibility twice.
```ruby
module R
refine B do
private :a
public :a
end
end
``` -
06/19/2025
-
06:21 PM Bug #21445: [BUG] push_mark_stack() called for broken object raised since cd9f447be247478d2eb3da985295735cce20cb23
- GitHub is also seeing a test failure that we bisected to cd9f447be247478d2eb3da985295735cce20cb23. We're still working on narrowing it down (it's only failing in CI), so not sure yet how related it is to this crash.
-
01:23 AM Bug #21445 (Closed): [BUG] push_mark_stack() called for broken object raised since cd9f447be247478d2eb3da985295735cce20cb23
- Rails CI against Ruby master branch fails
https://buildkite.com/rails/rails-nightly/builds/2398#019784d7-fbc2-4ff4-aee7-6e74eb560369
According to git bisect, this issue is triggered by cd9f447be247478d2eb3da985295735cce20cb23 .
H... -
- ```ruby
class A
def a
:a
end
end
class B < A
end
module R
refine B do
private :a
end
end
module R
refine B do
public :a
end
end
using R
B.new.a # StackOverflow
```
I would expect it to cha... -
- Add a fast path for known strings at compile time, otherwise calls method id to_s using Insn::SendWithoutBlock
Co-authored-by: composerinteralia <[email protected]>
More specific test name in zjit/src/hir.rs
Co-authored-by:... -
02:29 PM Revision 34eaa641 (git): ZJIT: Add `dupn` support
-
-
-
-
-
-
- - Send outputs from debugger to stderr
- Use `%W` to interpolate the pid -
- https://github.com/ruby/ruby/actions/runs/15751511003/job/44397451542?pr=13649
```
1) Failure:
TestLastThread#test_last_thread [/Users/runner/work/ruby/ruby/src/test/-ext-/gvl/test_last_thread.rb:18]:
Expected |1.0 - 1.167141| (... -
-
-
03:54 AM Revision 912edb47 (git): Fix missing write barrier on class fields
- Found by wbcheck
klass = Class.new
200.times do |iv|
klass.instance_variable_set("@_iv_#{iv}", Object.new)
end -
-
- RDoc markdown parser requires exact 4 spaces or tab as indentation.
Also the first nested bullet list must be separated from the enclosing
bullet list item by a blank line. - 01:54 AM Revision a4c4a347 (git): Update default gems list at c2d5d2969a501b9db93ed66db9017a [ci skip]
-
- https://github.com/ruby/weakref/commit/f6bd03ed54
-
-
01:23 AM Revision 471da0f0 (git): [rubygems/rubygems] Use `persist-credentials: false` in template
- `actions/checkout` defaults this value to `true`, causing credentials to
be written to `.git/config`. By setting it to `false`, we lessen the
likelihood of secrets being written to disk.
https://github.com/rubygems/rubygems/commit/a751d... -
- They should only load plugins from gems in the bundle.
https://github.com/rubygems/rubygems/commit/a229507820 -
- https://github.com/rubygems/rubygems/commit/5ad0737e77
-
- https://github.com/rubygems/rubygems/commit/d23b3d61ac
-
- https://github.com/rubygems/rubygems/commit/28b6a7cf5e
-
- Currently to test Bundler 3 mode we have to actually edit the version
file to simulate we're running a future version. This is inconvenient.
Instead, allow passing an environment variable, `BUNDLER_3_MODE`, to set
the "working mode" Bun... -
- The "ignore" attribute is a RubyGems thing to mark when a installed gem
should be ignored for activation because its extensions are not properly
compiled.
In the case of gems from path sources, the warning is not accurate
because extens...
06/18/2025
-
- * `invokebuiltin`
* `invokebuiltin_delegate`
* `invokebuiltin_delegate_leave`
These instructions all call out to a C function, passing EC, self, and
some number of arguments. `invokebuiltin` gets the arguments from the
stack, whereas th... - 06:09 PM Revision c57efbfb (git): Fix `heap-use-after-free` in `free_fast_fallback_getaddrinfo_entry` (#13231)
- This change addresses the following ASAN error:
```
==36597==ERROR: AddressSanitizer: heap-use-after-free on address 0x512000396ba8 at pc 0x7fcad5cbad9f bp 0x7fff19739af0 sp 0x7fff19739ae8
WRITE of size 8 at 0x512000396ba8 thread T0
... - 06:09 PM Revision 584365af (git): Ensure that memory is not freed before calling `free_fast_fallback_getaddrinfo_*` (#12661)
- Ensure that `getaddrinfo_entry` and `getaddrinfo_shared` exist before free them in the main thread.
-
- Fix generic_ivar_set_shape_ivptr for table rebuild
[Bug #21438]
Previously GC could trigger a table rebuild of the generic ivar
st_table in the middle of calling the st_update callback. This could
cause entries to be reallocated or rea... -
- Found by wbcheck
-
- Found by wbcheck
-
- Found by wbcheck
-
- Found by wbcheck
-
- Found by wbcheck
-
- Found by wbcheck
-
- Found by wbcheck
-
- We want to always use time_set_timew, as timew is 64-bit even on 32-bit
platforms so we need to be careful to both write that size, but still
trigger write barriers if we end up with a heap object. -
-
-
- We are inserting directly into the st_table, so we need to issue a write
barrier from the hash. -
- This object was newly allocated on move_enter, so some GC may happen and
it may have been marked by move_leave, so we need to issue an
rb_gc_writebarrier_remember so that any new references are seen afer the
memcpy from the old object. -
- Ractor::Port will mark the ractor, so we must issue a write barrier.
This was detected by wbcheck, but we've also seen it in CI:
verify_internal_consistency_reachable_i: WB miss (O->Y) 0x000071507d8bff80 ractor/port/Ractor::Port ra... -
- https://github.com/ruby/ruby/pull/13652
-
12:40 PM Bug #21443 (Closed): Backport: Fix heap-use-after-free in free_fast_fallback_getaddrinfo_entry
- Please backport the following AddressSanitizer error into Ruby 3.4
https://github.com/ruby/ruby/pull/13231
I did not find a ticket related to the original fix.
Thank you! -
- ```rb
require 'bundler/inline'
gemfile do
source 'https://rubygems.org'
gem 'rubocop'
end
require 'objspace'
require 'tmpdir'
3.times { GC.start }
GC.disable
Dir.mktmpdir do |dir|
Dir.chdir(dir) do
File.open... -
-
- https://github.com/ruby/ruby/pull/13649
-
- In last year, I gave up to make `tsort` to bundled gems because `set` depends `tsort`.
https://github.com/ruby/set/blob/29a253d2259e8d6d9b313dec8418ec5ed0caab08/lib/set.rb#L786
Fortunately, `set` is now core class. So, there is no ... -
08:44 AM Misc #20968: `Array#fetch_values` unexpected method name in stack trace
- I have updated the behavior to (mostly) suppress the display of `<internal:...>` frames in a backtrace. The algorithm works as follows:
* When internal frames appear in a sequence, it retains only one frame -- the one directly called ... -
- Applied in changeset commit:git|10767283dd0277a1d780790ce6bde67cf2c832a2.
----------
Exclude internal frames from backtrace
This changeset suppresses backtrace locations like
`<internal:array>:211` as much as possible.
Before the patc... -
-
- 07:06 AM Revision 13a2b9fa (git): Update bundled gems list as of 2025-06-18
-
06:39 AM Bug #21210: IO::Buffer gets invalidated on GC compaction
- I think @alanwu means the mutable buffer path (IO::Buffer.for w/ block) uses rb_str_locktmp to pin the String's malloc'ed content memory (not RString), while the immutable buffer path (IO::Buffer.for w/o block) utilizes CoW to obtain a f...
-
-
-
- This changeset suppresses backtrace locations like
`<internal:array>:211` as much as possible.
Before the patch:
```
$ ruby -e '[1].fetch_values(42)'
<internal:array>:211:in 'Array#fetch': index 42 outside of array bounds: -1...1 (Index... -
- In preparation for using it to update not only cfunc frames but also
internal frames, the function (and related variable names) are chagned.
I felt that the word "backpatch" is more appropriate than the more
general verb "update" here. -
- Applied in changeset commit:git|ecf1746fa4b899692514e68719227f04f9d6c1cc.
----------
Use the original gemspec from release package of RDoc
[Bug #21312]
https://bugs.ruby-lang.org/issues/21312
https://github.com/ruby/rdoc/pull/1379 -
- Because they are required markdown.rb provided by release package.
-
- [Bug #21312]
https://bugs.ruby-lang.org/issues/21312
https://github.com/ruby/rdoc/pull/1379
06/17/2025
-
10:51 PM Revision 3cfd71e7 (git): Fix minor typos in comments, specs, and docs
- Just a bit of minor cleanup
Signed-off-by: Tim Smith <[email protected]> -
- This write barrier occurred before the entry was added to the table, so
if GC occurred when inserting into the table, the write could be missed. -
- After returning from the callback in st_update is the point that the
hash table may be resized, which could trigger a GC and mark the table
being used for the tally.
RUBY_GC_LIBRARY=wbcheck WBCHECK_VERIFY_AFTER_WB=1 ./miniruby -e '(... -
10:03 PM
Bug #21441 (Closed): SEGV during thread cleanup if profiler calls thread_profiles_frames at wrong time
- Applied in changeset commit:git|e3ec101cc21613550ef87b7bd8432a69c7e639de.
----------
thread_cleanup: set CFP to NULL before clearing ec's stack
We clear the CFP first so that if a sampling profiler interrupts the current thread during ... -
- I opened a PR [here](https://github.com/ruby/ruby/pull/13643).
This may need to be backported. -
- During thread cleanup, if a ruby sampling profiler calls `thread_profile_frames` at just the wrong time, the ec's stack is cleared but the CFP still exists on the EC, so `thread_profile_frames` assumes it can walk the stack and crashes.
... - 10:03 PM Revision e3ec101c (git): thread_cleanup: set CFP to NULL before clearing ec's stack
- We clear the CFP first so that if a sampling profiler interrupts the current thread during `rb_ec_set_vm_stack`,
`thread_profile_frames` returns early instead of trying to walk the stack that's no longer set on the ec.
The early return ... -
- * Update method names.
* Sort shuffled tests by names. -
- This reverts commit 980f61935f6e8331e0908dc963e60fb727ab4d8c, which
seems no longer needed. -
02:41 PM
Bug #20998: rb_str_locktmp() changes flags of frozen strings and string literals
- ______________________________________________
-
- @hanazuki Thanks for the link, I wish I found that earlier. It confirms that it was already broken which is the same conclusion I came to.
hanazuki (Kasumi Hanazuki) wrote in #note-20:
> FYI: IO::Buffer is now broken (by a recent change... -
- Do not clear the commit-wide flags per file.
-
-
- Followup: https://github.com/ruby/ruby/pull/13589
This simplify a lot of things, as we no longer need to manually
manage the memory, we can use the Read-Copy-Update pattern and
avoid numerous race conditions.
Co-Authored-By: Étienne Ba... -
-
-
-
- Issue a call to rb_vm_opt_getconstant_path() like the interpreter, but
since that allocates the IC, we need to save the PC before calling. Add
FrameState to GetConstPath to get access to the PC. -
- * `opt_hash_freeze`
* `opt_ary_freeze`
* `opt_str_freeze`
* `opt_str_uminus`
Similar to `opt_neq`, but there are no args for `freeze`
Co-authored-by: ywenc <[email protected]>
Co-authored-by: Max Bernstein <[email protected]> -
- ruby/ruby#13636
https://github.com/ruby/prism/commit/e13d4f19db
Co-Authored-By: Tim Smith <[email protected]> -
- ```
compare-ruby: ruby 3.5.0dev (2025-06-17T08:45:40Z master e9d35671d2) +PRISM [arm64-darwin24]
last_commit=[ruby/json] Fix a typo
built-ruby: ruby 3.5.0dev (2025-06-17T09:27:05Z opt-getivar-for-cl.. ed1d7cd778) +PRISM [arm64-darwin24]
... -
- Halfway on the way to 3.5 now. I want to help out if I am able to, but right now I'm not sure what the holdup is.
-
- ruby/ruby#13636
https://github.com/ruby/tempfile/commit/366d9ccb8f
Co-Authored-By: Tim Smith <[email protected]> -
- ruby/ruby#13636
https://github.com/ruby/json/commit/6fc2c4b6ab
Co-Authored-By: Tim Smith <[email protected]> -
- The remaining issue was fixed in 055fef00a1c27fdc8293114dc134ca7910b1dc79.
Backport PRs:
- 3.4: https://github.com/ruby/ruby/pull/13637
- 3.3: https://github.com/ruby/ruby/pull/13638
Note that the bug while possible is quit... -
-
- alanwu (Alan Wu) wrote in #note-5:
> Looks like the mutable string buffer code paths pin using rb\_str\_locktmp().
Does `rb_str_locktmp()` pin? I couldn't see anything like that from the usage of the flag it sets. -
- Applied in changeset commit:git|8aac19d5987150cf5c45fee73c7a949ca472f488.
----------
io_buffer: Reimplement dcompact for IO::Buffer
The `source` field in IO::Buffer can have a String or an IO::Buffer
object, if not nil.
- When the `so... -
-
- https://github.com/rubygems/rubygems/commit/228f59e3ab
-
- https://github.com/rubygems/rubygems/commit/e28b5e306f
-
- Could potentially fix some flakies we're using and make the specs more
"modern" and simplifies them because less fallbacks are involved.
https://github.com/rubygems/rubygems/commit/30da9a1a93 -
- I don't think the indirection improve things.
https://github.com/rubygems/rubygems/commit/b408b28844 -
- If a previous copy of a gem is already installed, RubyGems will not
reinstall the gem but only recompile its extensions. This seems like a
good idea, but only if the gem is being installed from the registry.
If we are installing a local... -
- https://github.com/rubygems/rubygems/commit/9a859078ab
-
- I don't see any warnings.
https://github.com/rubygems/rubygems/commit/395df777a2 -
- https://github.com/rubygems/rubygems/commit/945a29a477
-
- Previously we were performing a realloc and then inserting the new value
into the table. If the table was flagged as requiring a rebuild, this
could trigger GC work and marking within that GC could access the fields
freed by realloc. -
- The `source` field in IO::Buffer can have a String or an IO::Buffer
object, if not nil.
- When the `source` is a String object. The `base` field points to the
memory location of the String content, which can be embedded in
RSTRING, ... -
- This reverts commit 6012145299cfa4ab561360c78710c7f2941a7e9d.
-
- - Put shell commands in code blocks so they can easily be copied from
the GitHub UI directly
- Fix a few typos
- Fix a dead link to MSDN
Signed-off-by: Tim Smith <[email protected]>
https://github.com/ruby/win32-registry/commit/61a4... -
-
-
-
- 01:07 AM Revision 50c6bd47 (git): Update vm->self location and mark it in vm.c for consistency