Activity
From 11/23/2024 to 11/29/2024
11/29/2024
-
11:53 PM Feature #20770: A *new* pipe operator proposal
- austin (Austin Ziegler) wrote in #note-53:
> I don't entirely agree. The ambiguity still exists because there is (more or less) an implicit block behaviour. If `_` already exists in the current scope, *both* the use of a pipe operator an... -
07:25 PM Feature #20770: A *new* pipe operator proposal
- AlexandreMagro (Alexandre Magro) wrote in #note-52:
> austin (Austin Ziegler) wrote in #note-50:
> ...
Not necessarily. I don't know much about how the parser works to produce the AST, but if it is able to do a *small* bit of backtrack... -
- austin (Austin Ziegler) wrote in #note-50:
> As I said, I don't think Ruby needs a pipe operator, but I wonder if a *different* approach might be taken. In irb, `_` is automatically assigned the result of the previous expression (well, ... -
- lpogic (Łukasz Pomietło) wrote in #note-49:
> I wanted _str_, _a_simple_json_ and _a_ to be treated as variables, maybe method arguments but impossible to reduce to one line.
If these variables need to be explicit and cannot be reduc... -
- lpogic (Łukasz Pomietło) wrote in #note-49:
> The notation `a |>= b` could be considered syntactic sugar for `a = a |> b`. Please remember that most binary operators in Ruby have similar syntactic sugar (including `||` and `&&`). I do... -
03:52 PM Feature #20770: A *new* pipe operator proposal
- AlexandreMagro (Alexandre Magro) wrote in #note-47:
> The pipe operator is well-known, but this type of operation with assignment is something I haven’t seen before, and I’m not sure of its precedence. The examples also don’t illustrate... -
11:42 PM Bug #20915 (Closed): Segfault with `TracePoint#parameters` and aliased C method
- Applied in changeset commit:git|660b995365f719fa59ed6f2809bb1527e6470d14.
----------
[Bug #20915] Fix SEGV with `TracePoint#parameters` and aliased C method
The following snippet results with a SEGV:
```ruby
C = Class.new do
alias_m... -
- The following snippet results with a SEGV:
```ruby
C = Class.new do
alias_method :new_to_s, :to_s
end
TracePoint.new(:c_call, &:parameters).enable { C.new.new_to_s }
```
at MRI 3.3.6 and ruby 3.4.0dev
The root cause of the issue li... -
11:33 PM Revision 88764dde (git): [DOC] Rewrite docs for rb_sym2str()
- Explaining this by reference to rb_id2str() obscures a few important
details because IDs and symbols don't map to each other perfectly (you
can have a dynamic symbol without an ID!) Also, it used to take 2
redirections to get to concrete... -
-
- Fewer allocations on boot, too.
Suggested-by: https://github.com/ruby/ruby/pull/12217 - 07:29 PM Revision 36a98307 (git): [rubygems/rubygems] make pretty_print tests whitespace agnostic
- https://github.com/rubygems/rubygems/commit/23e9396d22
-
07:07 PM Revision 7e02410b (git): [ruby/reline] Don't skip start_with check on encoding-incompatible
- candidates
(https://github.com/ruby/reline/pull/787)
https://github.com/ruby/reline/commit/8588be652f -
06:15 PM Revision c06dcba9 (git): [ruby/reline] Fix RELINE_TEST_ENCODING
- (https://github.com/ruby/reline/pull/743)
* Fix RELINE_TEST_ENCODING
It was not working because it was not environment variable.
* Fix Encoding::CompatibilityError: Shift_JIS is not compatible with UTF-8
Error: test_completion_append... -
06:15 PM Revision 7f34c75b (git): Use `RSTRING_PTR` instead of `StringValuePtr`
- ... since it is certain to be a String in this context.
Also, I want to avoid the anxious use of `StringValuePtr(str)` and
`RSTRING_LEN(str)` as arguments in the same function call. -
04:35 PM Bug #20924: IO#readline ignores the limit argument when the encoding is UTF-32LE and the limit would split a character
- Note that replacing `#readline` with `#gets` in the examples causes the issue to be seen in all Ruby versions, including 3.3.6.
-
- ```ruby
require 'tempfile'
Tempfile.open(binmode: true, encoding: 'utf-32le') do |f|
f.write('0123456789')
f.rewind
# A limit that would truncate a character becomes completely ignored
f.readline(3).bytesize # => 40; s... -
03:46 PM Bug #20921 (Closed): Use-after-free in constant cache
- Applied in changeset commit:git|f65a6c090c229de1665af49f2e51fc1d6397ab72.
----------
Fix use-after-free in constant cache
[Bug #20921]
When we create a cache entry for a constant, the following sequence of
events could happen:
- vm_t... -
- [Bug #20921]
When we create a cache entry for a constant, the following sequence of
events could happen:
- vm_track_constant_cache is called to insert a constant cache.
- In vm_track_constant_cache, we first look up the ST table for th... -
03:29 PM Revision 73c4023c (git): [rubygems/rubygems] Enable pending exec spec
- This is how Ruby works, and it's not going to change. So let's run the
test since it makes no sense to keep it pending.
https://github.com/rubygems/rubygems/commit/5f78d3c737 -
- https://github.com/rubygems/rubygems/commit/abb658757f
-
- They had slightly different behavior in Bundler 2 and Bundler 3 because
Bundler 3 will auto-clean after `bundle install`, so one of them system
installed gems won't be present when `Bundler.setup` fails.
We can unify the specs by moving... -
- https://github.com/rubygems/rubygems/commit/9dbfce76cf
-
- And make sure `bundle install --local` can install from it without git.
https://github.com/rubygems/rubygems/commit/7d6b631620 -
- https://github.com/rubygems/rubygems/commit/33536aa61c
-
- These are never equal.
https://github.com/rubygems/rubygems/commit/69e369da74 -
- The "revision" attribute is actually getting the revision from the
Gemfile.lock file. So I think "locked" is a better term here to avoid
confusion with the revision checked out in `vendor/cache`.
https://github.com/rubygems/rubygems/com... -
03:15 PM Misc #20923 (Closed): Add builds for existing shared GC's to CI
-
- @peterzhu2118 has brought to my attention that CI builds for the shared GC's, specifically MMTk were already proposed and approved as part of the discussion of #20860 providing they are non-blocking.
Closing this ticket as a duplicate. -
- ## Summary
I'd like to add CI builds that exercise the GC API implementation with the currently supported known GC modules.
Implemented in [[GH-#12212]](https://github.com/ruby/ruby/pull/12212)
## Background
Ruby now ships w... -
- `len` is always at least 1 here.
-
02:19 PM Revision 705714be (git): prefer ruby_memerror instead
- This could be out of GVL
-
- Let there be rooms for each GC implementations how to handle multi
threaded situations. They can be totally reentrant, or can have
their own mutex, or can rely on rb_thread_call_with_gvl.
In any ways the allocator (has been, but now of... -
11:47 AM Feature #20922: Should not we omit parentheses in assert calls?
- I generally agree, and this change would be welcome, but there's still a common case that is an error:
```ruby
assert_equal {}, something
```
-
- FYI:
https://github.com/ruby/ruby/commit/6a39e6fc2d0db590ad605f7af9c99d32c64c6a22
https://github.com/ruby/ruby/commit/fc67091fc8dc4c7122499cd0d287cd34ff01bf30
https://github.com/ruby/ruby/commit/5b146eb5a15cce4c7a6ce279bd53e75a61d4a... -
- I often see the style of omitting parentheses in assert calls, but it leads to annoying warnings in the following case:
```
assert_equal -1, val #=> warning: ambiguous first argument; put parentheses or a space even after `-` oper... - 10:28 AM Revision 5067a465 (git): Update default gems list at ae85fc72717c5b6d13be101781d41a [ci skip]
- 10:27 AM Revision ae85fc72 (git): [ruby/set] Bump VERSION to 1.1.1
- https://github.com/ruby/set/commit/1c3cded76a
- 09:49 AM Revision 49d2e79f (git): Ensure to close pipes when `TCPSocket.new` finishes processing (#12181)
- `TCPSocket.new` with HEv2 uses three threads.
The last of these threads to exit closed pipes.
However, if pipes were open at the end of the main thread, they would leak.
This change avoids this by closing pipes at the end of the main thr... -
08:38 AM Bug #20506: YJIT build error on aarch64 with Rust 1.78.0
- I tested with the recent 1.84 beta and it seems resolved with that version.
-
08:25 AM Feature #20912: Add warning when redefining __id__ as well as object_id
- I am strongly discourage redefining `object_id`, so we should keep the warning. I am not against warning `__id__` redefinition.
Matz.
- 05:18 AM Revision 22e1a8c4 (git): Allow disable to `fast_fallback` of `TCPSocket.new` (#12210)
- with `Socket.tcp_fast_fallback=`
The functions that `Socket.tcp` had are now also available in `TCPSocket.new`. -
- If `slave` is negative, neither `dup2(slave,0)` or `close(slave)` should
be executed. I believe this check is completely useless. -
- Here `nb` should never be NULL. If it were, the following
`nb->buffer_list` would be strange.
A follow-up to ddd8da4b6ba3dfcca21ca710e7cef2fa3b9632d7 -
02:48 AM Revision 43dd9c72 (git): [ruby/date] Fix mixed declarations and code
- This still support ruby 2.6 which does not require C99.
https://github.com/ruby/date/commit/61d849758f -
- It was used intentionally.
https://github.com/ruby/date/commit/291b40f939 -
- https://github.com/ruby/time/commit/c668704413
-
12:49 AM Revision 84b14c03 (git): Dont't need to run nmake test
-
-
-
- This reverts commit 743a31d639d4c96827b1b17f59c829c1ec31abc3,
"[ruby/date] [DOC] Make document coverage 100%". This file is only
for rdoc coverage in the date gem, and useless when the built-in Time
is documented.
11/28/2024
-
- > That does not have too long history, seeing this NEWS entry:
You can implement a polyfill [such as this one in io-console](https://github.com/ruby/io-console/blob/b152f5b681895e53fce10e02ced43e6d35c19e82/ext/io/console/console.c#L15... -
03:19 PM Feature #20265: Deprecate and remove rb_newobj and rb_newobj_of
- That does not have too long history, seeing this NEWS entry:
~~~
$ grep -R -A 1 rb_io_open_descriptor
doc/NEWS/NEWS-3.3.0.md: * `rb_io_open_descriptor()` is introduced to make an IO object from a file
doc/NEWS/NEWS-3.3.0.md- d... -
08:04 PM Revision a5119a3f (git): YJIT: Add missing prepare before calling str_dup
-
- GitHub PR: https://github.com/ruby/ruby/pull/12203
When we create a cache entry for a constant, the following sequence of events could happen:
- vm_track_constant_cache is called to insert a constant cache.
- In vm_track_constant_... -
- ```
$ cat foo.rb
puts __FILE__
$ ruby foo.rb
foo.rb
$ ruby ./foo.rb
./foo.rb
$ ruby -e 'load "foo.rb"'
foo.rb
$ ruby -e 'load "./foo.rb"'
/full/path/to/foo.rb
```
More than an issue, this... -
- When loading a crafted marshal data of Random, a pointer to an illegal
address was created. I don't think there is any harm since the data is
normalized before access, but just to be safe, I add a check to make it
an error. -
-
04:56 PM Feature #20914: Constructor Parameter Shortcuts
- Thank you! This appears to be literally the same proposal, so it can be closed.
I will review #5825 and follow for updates. -
- > What do you think about this option:
Then this proposal is almost the same as #5825. -
- You’re absolutely right; this part does seem ambiguous. I initially added it as an empty block, modeled after the short syntax, like this:
```ruby
def initialize(name, age, email = nil) = (@name, @age, @email = name, age, email)
```
Wh... -
- ```ruby
def initialize(@name, @age, @email = nil) =
```
What is the equal sign at the end?
`nil` is missed? -
-
- .git-blame-ignore-revs will be updated after this commit is merged.
-
- https://github.com/rubygems/rubygems/commit/23d06195fa
- 03:23 PM Revision 3e308971 (git): Update default gems list at afea58a0ef07b05846dbb24ca8afbf [ci skip]
-
- (https://github.com/ruby/reline/pull/786)
https://github.com/ruby/reline/commit/d4f6741e7e -
- @nobu, I just checked with a fresh build of the master branch, and the second failure case I mentioned in my previous message still exists. I opened #20919 to cover it.
-
- When transcoding characters, `IO#seek` and `IO#pos=` only clear the internal character buffer if `IO#getc` is called first:
```ruby
require 'tempfile'
Tempfile.open(encoding: 'utf-8:utf-16le') do |f|
f.write('0123456789')
f.... - 01:39 PM Revision 47f8a552 (git): Ensure to free fast_fallback_getaddrinfo_shared with single family (#12199)
- With https://github.com/ruby/ruby/pull/12156,
the memory of the `struct fast_fallback_getaddrinfo_shared`
is now allocated even if there is only one address family.
This change will always free it when `TCPSocket.new` finishes. -
- This is not important because of the path of `assert(false)`, but just
in case.
Coverity Scan found this issue.
https://github.com/ruby/prism/commit/7335c62f9d -
12:18 PM Revision 36ed9c2f (git): [ruby/rdoc] Improve how gemspec's files are defined
- (https://github.com/ruby/rdoc/pull/1212)
Currently, the gemspec's files are defined by hand, which is error-prone.
For example: https://github.com/ruby/rdoc/pull/1211
This commit uses `Dir.glob` where possible to reduce the risk of th... -
11:55 AM Bug #20917: redo/next in nested begin block causes wrong order of execution
- Indeed, it seems like a bug, I think as well the "redo jump/unwind/exception" should run ensure's and get to the "break jump/unwind/exception" which should override the `redo` like when a Ruby exception overrides another.
FWIW, this i... -
03:26 AM Bug #20917 (Open): redo/next in nested begin block causes wrong order of execution
- It seems that `redo`/`next` in a nested `begin` block can cause the wrong order of execution.
For example:
```ruby
for _ in [0]
puts 0
begin
puts 1
begin
puts 2
redo
ensure
puts 3
end
... -
- > It seems used a lot, yes:
I meant in context where it is called on an unknown object where `object_id` may have been redefined and cause the "serious problem" the warning hints at. I've dealt with tons of meta-programing messes over... -
- mame (Yusuke Endoh) wrote in #note-5:
> I think it's a good idea to warn against redefining `__id__`, but I think we should keep warning against redefining `object_id` as well.
> ...
I agree and I think this a well thought out argument.
... -
-
-
09:19 AM Revision 31a3e877 (git): [DOC] typo fix in NEWS.md
- follow up: https://github.com/ruby/ruby/pull/12196/files#r1861730753
-
08:36 AM Revision da2c083f (git): add a NEWS entry about extra rescue/ensure frames
- from backtraces.
-
-
- @matz how about `Ractor.local_storage_once(key){ ... }`? It is from [pthread_once](https://man7.org/linux/man-pages/man3/pthread_once.3p.html).
`Ractor.once(key){ ... }` seems too short?
```ruby
def inc
Ractor.local_storage_once(:mtx... -
- After any `LINK_ELEMENT` sequence is added, `LINK_ANCHOR` must not
loop. -
-
-
-
-
- A follow-up to ef59175a68c448fe334125824b477a9e1d5629bc. That commit
uses `&body->local_table[...]` but `body->local_table` is already freed.
I think it is an undefined behavior to calculate a pointer that exceeds
the bound by more than ... -
- ```
$ ruby -e 'foo('
-e: -e:2: syntax error found (SyntaxError)
1 | foo(
> 2 |
| ^ unexpected end-of-input; expected a `)` to close the arguments
```
It says `-e:2`, but there is no line 2 in `-e 'foo('`.
`eval("foo(")`... -
- The same issue as https://github.com/ruby/prism/pull/3246
Coverity Scan found this issue. - 03:25 AM Revision d94ae8b8 (git): Update bundled gems list at e0b68b2dc93da31f93232fe0b545ba [ci skip]
-
- Co-authored-by: Sutou Kouhei <[email protected]>
-
-
-
- The Prism compiler raises an exception against the following code.
```
$ ./miniruby -e '# shareable_constant_value: experimental_everything
C = { **{ } }'
-e:2: Ractor constant writes do not support **
-e: node type not implement... -
- When mmap fails for any reason, the fd must be closed.
Coverity Scan found this issue.
https://github.com/ruby/prism/commit/c06e9c400f -
- I'm getting a segfault in the following example
```ruby
C = Class.new do
alias_method :new_to_s, :to_s
end
TracePoint.new(:c_call, &:parameters).enable { C.new.new_to_s }
```
Proposed fix: https://github.com/ruby/ruby/pull... -
01:08 AM Misc #20913: Proposal: Adding Jeremy Evans and Burdette Lamar to www.ruby-lang.org's English Editorial Team
- I would also like to be added to both.
11/27/2024
-
- # Constructor Parameter Shortcuts
I propose adding a new syntax to simplify the constructor definition in the classes. Defining a constructor often involves repetitive assignments of parameters to instance variables. This pattern is v... -
* [Misc #20913] Proposal: Adding Jeremy Evans and Burdette Lamar to www.ruby-lang.org's English Editorial Team
* They have expressed enthusiasm for improving the website during recent discussions at RubyConf in Chicago.
* Adding ...-
- Hi everyone, I'd like to propose that we invite @jeremyevans0 and @burdettelamar to join the editorial team for [ruby-lang.org](https://ruby-lang.org)'s English content. Specifically, this means giving them permission to merge English co...
-
- Is `object_id` really called that much? I can't really think of something that really depend on it from the top of my head.
-
- I think it's a good idea to warn against redefining `__id__`, but I think we should keep warning against redefining `object_id` as well.
The reason is that it is relatively well known that "we should use `__send__` instead of `send` f... -
- > do you want to redefine it?
Shopify has case like this where one of the GraphQL attribute is `object_id` and that cause a warning to be emitted. I also know the Facebook API has (or used to have) an `object_id` field.
Overall, it... -
- nobu (Nobuyoshi Nakada) wrote in #note-2:
> As for `object_id`, do you want to redefine it?
I think it's best not to redefine, but a warning is probably more severe than is needed. I don't think it causes "serious problems", especial... -
- Agree that `__id__` should be warned.
As for `object_id`, do you want to redefine it?
-
- https://github.com/ruby/ruby/pull/12177
-
- Currently if you create a class and redefine or remove either `object_id` or `__send__` it will issue a warning. There's no such warning on `__id__`.
```
❯ ruby -we 'class << Object.new; def object_id = 1; end'
-e:1: warning: redefi... -
- fill_lines is passed -1 for offset, which causes it to read the -1 index
of traces. This is not valid memory as -1 is reading before the trace
global variable in rb_print_backtrace. This code comes from commit
99d1f5f88b9b6de3124e3101990... -
- (https://github.com/ruby/reline/pull/785)
https://github.com/ruby/reline/commit/85e20f0031 -
- (https://github.com/ruby/reline/pull/784)
https://github.com/ruby/reline/commit/a6d1ac54e6 -
- macOS clang 16 generates DWARF5, which have Mach-O section names that
are limited to 16 characters, which causes sections with long names to
be truncated and not match above.
See: https://wiki.dwarfstd.org/Best_Practices.md#Mach-2d-O -
-
- It's not used outside of error.c.
-
- > Just FTR, there are also project which are using NEWOBJ, such as:
> ...
That code is using internal data structures which are subject to change. It should use an API such as `rb_io_open_descriptor` instead. -
- Just FTR, there are also project which are using `NEWOBJ`, such as:
https://github.com/search?q=repo%3Ahparra%2Fruby-serialport%20NEWOBJ&type=code
So what is recommended instead? -
- GCC 13 prints the following warning.
https://rubyci.s3.amazonaws.com/ubuntu/ruby-master/log/20241127T001003Z.log.html.gz
```
compiling generator.c
generator.c: In function ‘raise_generator_error’:
generator.c:91:5: warning: function ‘ra... -
- While profiling `strscan`, I noticed `rb_must_asciicompat` was quite
slow, as more than 5% of the benchmark was spent in it: https://share.firefox.dev/49bOcTn
By checking for the common 3 ASCII compatible encoding index first,
we can sk... -
12:03 PM Revision 43b059b6 (git): [DOC] Fix a link and sort links in NEWS.md
-
09:11 AM Bug #20910: leaked-globals not happy about dtrace related symbols
- Hi. Yes, this works as well.
-
- Or does this help?
```diff
diff --git a/tool/leaked-globals b/tool/leaked-globals
index 05ce02658b7..6118cd56e81 100755
--- a/tool/leaked-globals
+++ b/tool/leaked-globals
@@ -90,6 +90,7 @@
next if !so and n.start_with?("___a... -
-
-
-
-
-
-
- Possible fix for recent crashes seen on CI.
[BUG] rb_sys_fail_str(<STDIN>) - errno == 0
rb_io_path() calls rb_obj_dup(), which could call initialize_dup in Ruby
and clobber errno before rb_sys_fail_str() gets to read errno. So
sav... -
- This reverts commit 5bd144c1bb20e22e4d9f5e5e0264820fd3ef8137.
-
-
- (https://github.com/ruby/strscan/pull/115)
Fix: https://github.com/ruby/strscan/issues/113
This allows to directly parse an Integer from a String without needing
to first allocate a sub string.
Notes:
The implementation is limited by... -
12:24 AM Revision a041a6c1 (git): [ruby/strscan] Add scan and search benchmark
- (https://github.com/ruby/strscan/pull/111)
# Why?
To improve the parsing process, I would like to add benchmarks for all
parsing processes.
## scan
- scan_full(regexp, false, true) == StringScanner#check
- scan_full(regexp, false, fals...
11/26/2024
- 10:04 PM Revision 092a48de (git): [rubygems/rubygems] [SpecFetcher] If candidates include {name}-ruby or ruby-{name}, recommend those.
- https://github.com/rubygems/rubygems/commit/d7d33172c1
- 09:49 PM Revision 8f9b9aec (git): YJIT: Implement opt_reverse insn (#12175)
-
- The sync script defaults to `master` when analysing which commits to
pick, but the default for new repos now is `main`. - 07:31 PM Revision 1dd40ec1 (git): Optimize instructions when creating an array just to call `include?` (#12123)
- * Add opt_duparray_send insn to skip the allocation on `#include?`
If the method isn't going to modify the array we don't need to copy it.
This avoids the allocation / array copy for things like `[:a, :b].include?(x)`.
This adds a BOP ... -
- `Array#max` accepts a block that compares two values.
```
p ["1", "2", "3"].max {|i, j| i.to_i <=> j.to_i } #=> "3"
```
You may want to use `Array#max_by`.
```
p ["1", "2", "3"].max_by(&:to_i) #=> "3"
``` -
06:09 PM Bug #20911 (Rejected): Array#max doesn't take block if using &:
- ```
["1","2","3"].max { |i| i.to_i }
=> "3"
["1","2","3"].max(&:to_i)
(irb):6:in `to_i': no implicit conversion of String into Integer (TypeError)
["1","2","3"].max(&:to_i)
^^^^^^
from (irb):6:in `max'... -
- (https://github.com/ruby/reline/pull/713)
https://github.com/ruby/reline/commit/5a8da85f2b -
- (https://github.com/ruby/irb/pull/1038)
Fixes indent calculation of this input
```
if false
p `ls`
end
```
https://github.com/ruby/irb/commit/4217a46f5d -
05:07 PM Feature #17566: Tune thread QoS / efficiency on macOS
- I just ran across this article that touches on the sordid history of Thread priority and on how Apple Silicon is using QoS dispatch, and thought it might be nice to add here for context. https://blog.xoria.org/macos-tips-threading/
Th... -
- I am not sure if this is Solaris specific, but when I build Ruby 3.3.6 with dtrace enabled and run its test suite, the leaked-globals checker fails with the following:
```
Checking leaked global symbols...48 un-prefixed symbols leaked
... -
-
- [Feature #19236]
Ref: https://github.com/ruby/ruby/pull/10357 - 10:36 AM Revision ffbfec39 (git): [MinGW] - Fix NET_LUID check
-
-
-
- Since mkmf.rb is for extension libraries after installation, it cannot
work alone in build directory and needs to run from extmk.rb. -
- - Add flags to appropriate variables.
- Use `append_cflags` to append a flag safely, instead of appending
blindly. -
- As common.mk is used by nmake.exe, the commands there need to be
accepted also by cmd.exe. -
- (https://github.com/ruby/uri/pull/140)
A follow-up to https://github.com/ruby/uri/commit/bd2e4be9d0fa
Also, leave a comment that the use of `URI::REGEXP` is intentional
https://github.com/ruby/uri/commit/bdf765e44a -
09:41 AM Feature #13820: Add a nil coalescing operator
- Here are my 2 cents regarding this subject, if it can help moving forward with this feature.
Wkipedia has a great article explaining and listing null coalescing operators in other languages: https://en.wikipedia.org/wiki/Null_coalesci... -
- ```
NoMethodError: undefined method 'flush' for an instance of JsonWrite
``` -
- https://github.com/rubygems/rubygems/commit/9ea1539b08
-
- When converging locked specifications to select the ones that should be
preserved while resolving, we can avoid having to do a second pass to
ignore the ones that have been explicitly unlocked.
https://github.com/rubygems/rubygems/commi... -
- As long as some spec in the materialization is complete.
https://github.com/rubygems/rubygems/commit/9a673b0bbb -
- https://github.com/rubygems/rubygems/commit/3041b3d784
-
- This gives more flexibility to allow further improvements.
https://github.com/rubygems/rubygems/commit/f11a890f5e -
- So there's no need to pass it around in so many places.
https://github.com/rubygems/rubygems/commit/784ab7481b -
- https://github.com/rubygems/rubygems/commit/0c3a65871a
-
- https://github.com/rubygems/rubygems/commit/d8d68cc00e
-
- When converging specification to pass the set of versions that should be
preserved from the lockfile during resolution, we should make sure all
top level gems are considered, and only exclude those gems themselves
(and not their dependen... -
- https://github.com/rubygems/rubygems/commit/d0f789970f
-
- For consistency with other specs.
https://github.com/rubygems/rubygems/commit/a5b2449896 -
- https://github.com/rubygems/rubygems/commit/c519830d4d
-
- When used with `LazySpecification` objects, `SpecSet#for` was
incorrectly considering development dependencies. This did not cause any
issues because all usages of this method with `LazySpecification`'s are
not strict, meaning the pass `... -
- https://github.com/rubygems/rubygems/commit/a2bb68a29b
-
- https://github.com/rubygems/rubygems/commit/6dc64f9851
-
- https://github.com/rubygems/rubygems/commit/e1caeecdf8
-
- https://github.com/rubygems/rubygems/commit/6a6041d073
-
- https://github.com/rubygems/rubygems/commit/df2c9eb52f
-
- https://github.com/rubygems/rubygems/commit/c382b606bd
-
- Fix: https://github.com/ruby/json/issues/710
Makes it easier to debug why a given tree of objects can't
be dumped as JSON.
Co-Authored-By: Étienne Barrié <[email protected]> -
- It's using `rb_gc_mark_maybe` under the hood, which isn't what
we need.
https://github.com/ruby/json/commit/e10d0bffcd -
- Ref: https://github.com/ruby/json/issues/524
Rather than to buffer everything in memory.
Unfortunately Ruby doesn't provide an API to write into
and IO without first allocating a string, which is a bit
wasteful.
https://github.com/rub... -
- Ah, that could be a nice idea. Not sure if that breaks other parts though.
-
- Can't we prioritize the condition `defined(rb_define_method)` for `RUBY_METHOD_FUNC`?
-
- The extension library is confusing our C level API. `rb_define_singleton_method()` and friends do not intend to take `RUBY_METHOD_FUNC`. That had accidentally worked because of the past C specifications you described, but not any longer...
-
05:00 AM Revision 51970165 (git): Update NEWS.md for Bug #20620
-
-
- commit:02b70256b5171d4b85ea7eeab836d3d7cfb3dbfc
commit:6b4f8945d600168bf530d21395da8293fbd5e8ba -
- https://github.com/ruby/uri/commit/bd2e4be9d0
-
- (https://github.com/ruby/uri/pull/138)"
This reverts commit https://github.com/ruby/uri/commit/c00726a20a00.
https://github.com/ruby/uri/commit/22f5a7a790 -
-
- https://bugs.ruby-lang.org/issues/20108
https://bugs.ruby-lang.org/issues/20564
https://bugs.ruby-lang.org/issues/20705
https://bugs.ruby-lang.org/issues/20782
https://bugs.ruby-lang.org/issues/20811 -
-
- * [Feature #20884] reserve "Ruby" toplevel module for Ruby language
* It's better to reserve this at Ruby 3.4, not 3.5. -
-
-
- https://github.com/ruby/resolv/actions/runs/12021339854/job/33511655994?pr=64
```
=> 705: with_udp_and_tcp('127.0.0.1', 0) do |u1, t1|
706: with_udp_and_tcp('127.0.0.1', 0) do |u2,t2|
707: u2.close # XXX: u... -
02:13 AM Revision 4e59ae6e (git): Add news about MatchData#bytebegin and MatchData#byteend
- 01:29 AM Revision 56fbf64a (git): Fix vm_objtostring optimization for Symbol
- Co-authored-by: John Hawthorn <[email protected]>
-
- https://github.com/ruby/resolv/commit/83f18e2a49
11/25/2024
- 10:31 PM Revision 7be0d505 (git): [ruby/logger] include license texts in gem
- https://github.com/ruby/logger/commit/1a64cb552f
- 08:38 PM Revision 544d6844 (git): Use footnote links instead
- 08:38 PM Revision 96766517 (git): Update maintainer GH handles and add links when possible
- 08:38 PM Revision c3144e4f (git): Escape Module in the header
- 08:38 PM Revision 7c200a46 (git): Add new lines after headers
- 08:38 PM Revision 9362753d (git): Update RDoc's maintainers list
-
07:32 PM Revision 081bdc51 (git): YJIT: fix small typo in command line options help (#12167)
- 06:35 PM Revision cdd47436 (git): [rubygems/rubygems] Bump the rb-sys group across 2 directories with 1 update
- Bumps the rb-sys group with 1 update in the /test/rubygems/test_gem_ext_cargo_builder/custom_name/ext/custom_name_lib directory: [rb-sys](https://github.com/oxidize-rb/rb-sys).
Bumps the rb-sys group with 1 update in the /test/rubygems/t... -
- Due to a typo in the spec, the issue was not caught initially. If
Bundler does not need to re-resolve, `bundle lock` is a noop so Bundler
does not add checksums.
To fix the issue, we do something similar to what `bundle install` does,
j... -
- I always found the `resolve_if_necessary` method pretty confusing
because by reading it, it suggests that resolution always happens, and
the point is whether that needs to be local or remote. This commit tries
to make that more clear.
h... -
- All we need is to setup remote or local sources appropriately.
https://github.com/rubygems/rubygems/commit/3ceff46a2a -
05:56 PM Bug #20906: Segmentation Fault in compile_keyword_arg
- Not sure if this is the same, but here's a shorter one that also crashes in 3.3.5.
*a[a:]=0
-
- (https://github.com/ruby/reline/pull/782)
https://github.com/ruby/reline/commit/cbf213291c -
- The man page is clear that every `va_start` call MUST be succeeded by
the corresponding `va_end` call.
So `rb_raise` can't call `rb_exc_raise` before `va_end`, otherwise
`va_end` is never called.
Co-Authored-By: Étienne Barrié <etienne... -
04:40 PM Bug #20908: Ruby extension builds fail with GCC 15 which defaults to -std=gnu23
- cc @shyouhei who made other C23 porting changes
-
- Hi!
Upcoming GCC 15 defaults to C23 (`-std=gnu23`). One thing C23 changes is removing unprototyped functions, so `void foo()` now means `void foo(void)`, rather than "any arguments".
Ruby extensions fail to build as a result with G... - 04:20 PM Revision 1d510a95 (git): [rubygems/rubygems] Fix test task name on generated readme when using test-unit
- * bundler/lib/bundler/templates/newgem/README.md.tt (Development): Use the
test_task value to get the correct test task name ("test", not "test-unit").
* bundler/spec/commands/newgem_spec.rb (README.md): Add tests for test task
names for... -
- https://github.com/rubygems/rubygems/commit/585a6a89d4
-
- https://github.com/rubygems/rubygems/commit/09e0971ab8
-
- That's indeed the ideal behavior but it's a mess to maintain because the
version of RubyGems shipped with each patchlevel of Ruby changes. We
could try looking at the `VERSION` constant in `
RbConfig::CONFIG["rubylibdir"` but for now I c... -
- We already do this check in `setup.rb` itself, which is run earlier.
https://github.com/rubygems/rubygems/commit/160cc3f1c5 -
- https://github.com/rubygems/rubygems/commit/31fadaf2d2
-
-
- So that it doesn't get included in the generated binaries for builds
that don't support loading shared GC modules
Co-Authored-By: Peter Zhu <[email protected]> -
- Co-Authored-By: Peter Zhu <[email protected]>
-
- ```
for (int i = 0; i < arg->family_size; i++) {
arg->getaddrinfo_entries[i] = allocate_fast_fallback_getaddrinfo_entry();
if (!(arg->getaddrinfo_entries[i])) rb_syserr_fail(errno, "calloc(3)");
```
If the allocation fails in th... -
- I checked if it is feasible with parse.y: https://github.com/ydah/ruby/commit/50100c136ac4bbe379b2c360994b739227610e17
I have been able to confirm that the following are feasible with the modifications shown above.
```ruby
... -
- `make test-all RUBYOPT=-w` in the ruby/ruby repo printed the warning.
This change prevents the warning by specifying the content-type
explicitly.
A follow-up to https://github.com/rubygems/rubygems/commit/b70c1bb1503d
https://github.co... -
-
- https://github.com/ruby/open-uri/commit/04067e54b9
-
- (https://github.com/ruby/uri/pull/138)
https://github.com/ruby/uri/commit/c00726a20a -
-
- 05:10 AM Revision ff5fc4b5 (git): Do not save the last error without sockets in the connection attempt (#12153)
- * Do not save the last_error if there are no sockets waiting to be connected
In this implementation, the results of both name resolution and connection attempts are awaited using select(2).
When it returned, the implementation attempted... -
02:02 AM Feature #20861: Add an environment variable for tuning the default thread quantum
- byroot (Jean Boussier) wrote in #note-6:
> Yes, quite literally: `Thread.current.quantum = 20` or something like that.
> ...
What happened to this idea? According to the latest comments it seems we're back to the idea of an environment... -
11/24/2024
-
-
- bug
(https://github.com/ruby/reline/pull/763)
https://github.com/ruby/reline/commit/d3ba7216eb -
-
- ```
compiling gc.c
In file included from gc.c:80:
/usr/include/sys/prctl.h:88:8: error: redefinition of 'struct prctl_mm_map'
88 | struct prctl_mm_map {
| ^~~~~~~~~~~~
In file included from gc.c:79:
/usr/include/linux/prc... -
08:26 AM Bug #20904: 3.4.0-preview2: Building miniruby.exe fails for mswin32
- nobu (Nobuyoshi Nakada) wrote in #note-1:
> jun66j5 (Jun Omae) wrote:
> ...
Thanks for the fix!
> > Also, adding temporarily `-w24013` to `WARNFLAGS` in `win32/Makefile.sub`, the following warning is received.
> ...
Weird. I get th...
11/23/2024
-
11:54 PM
Bug #20907 (Closed): Fiber scheduler does not correctly re-lock mutex if `Mutex#sleep` is interrupted.
- Applied in changeset commit:git|a8c2d5e7bee5fad0965baeb58d312ddc5932ec26.
----------
Ensure fiber scheduler re-acquires mutex when interrupted from sleep. (#12158)
[Bug #20907] -
11:26 PM Bug #20907 (Closed): Fiber scheduler does not correctly re-lock mutex if `Mutex#sleep` is interrupted.
- The following test fails with a `ThreadError` instead of a `RuntimeError`:
```ruby
def test_condition_variable
condition_variable = ::Thread::ConditionVariable.new
mutex = ::Thread::Mutex.new
error = nil
thr... - 11:54 PM Revision a8c2d5e7 (git): Ensure fiber scheduler re-acquires mutex when interrupted from sleep. (#12158)
- [Bug #20907]
- 03:44 PM Revision 31997661 (git): UBF is also required for synchronous name resolution (#12156)
- `rb_thread_call_without_gvl2` is used to wait for the results of name resolution and connection attempts.
When there is only one address family to resolve, the necessary resources were not being passed to the UBF.
With this change, the h... - 02:04 PM Revision 8d575e49 (git): Save the error that occurred during name resolution (#12155)
- even if a system call error happens after the name resolution failure in the child thread.
pipe and write(2) are used to notify the main thread of the name resolution results from the child thread.
After name resolution is completed in ... -
- jun66j5 (Jun Omae) wrote:
> Investigating it, Windows 8 is required after #20563 but `NTVER` is still `0x0600` in `win32/Makefile.sub`. I think it should be `0x0602`. Workaround is to invoke `win32\configure.bat` with `--with-ntver=0x06... -
- ASAN_OPTIONS="detect_leaks=1" RUBY_FREE_AT_EXIT=1 ./miniruby -e 'trace_var(:$x){}'
-