Activity
From 12/29/2021 to 01/04/2022
01/04/2022
-
11:18 PM Revision 13241b71 (git): Fix failures on ruby 3.1
- Because https://github.com/ruby/ruby/pull/5148 merged after Ruby 3.1.0 released.
https://github.com/ruby/actions/runs/4705986643?check_suite_focus=true#step:18:144
```
1)
Method#parameters adds * rest arg for "star" argument FAILED
... -
10:53 PM Feature #18181: Introduce Enumerable#min_with_value, max_with_value, and minmax_with_value
- For a single value, a "then" suffix might be nice.
``` ruby
%w(abcde fg hijk).min_by(&:size)
#=> "fg"
%w(abcde fg hijk).min_by_then(&:size)
#=> 2
```
It mirrors #min_by and #then with the same block.
``` ruby
%w(abcde fg... -
10:25 PM Revision d1a91076 (git): [rubygems/rubygems] Fix race conditon on JRuby
- On JRuby, sometimes we get the following error in CI when running a
realworld test that checks that `gem install rails` succeeds:
```
ERROR: While executing gem ... (NoMethodError)
undefined method `ignored=' for nil:NilClass
... -
08:42 PM Bug #18436 (Closed): Fix Pathname dot directory globbing
- Applied in changeset commit:git|5917f49ad03cd3b2c13d3eea85c969a113b83bf8.
----------
Fix documentation for Dir.glob
Dir.glob("*", File::FNM_DOTMATCH) no longer includes ".." in the
returned array as of Ruby 3.1.
Fixes [Bug #18436] -
08:38 PM Bug #18436: Fix Pathname dot directory globbing
- The glob difference between Ruby 3.0 and 3.1 is expected, as the change was made to fix bug #17280. Looks like the documentation needs to be updated to reflect the change. I'll take care of that shortly.
-
- Dir.glob("*", File::FNM_DOTMATCH) no longer includes ".." in the
returned array as of Ruby 3.1.
Fixes [Bug #18436] -
-
-
07:46 PM Revision 44379805 (git): Fix crash on bootup when RGENGC_CHECK_MODE=2 with GC stress
- We need to unshare the array loaded_features because the shared root
could be old gen (while loaded_features is young) which causes a WB
miss. -
07:28 PM Feature #18459 (Closed): IRB autocomplete dropdown colour options
- It would be great to be able to specify bg/fg colours for the new autocomplete dropdown in irb in ruby 3.1. This could help for accessibility purposes, or for anyone who just wants to make it look more personalised for their terminal 😎
... -
06:57 PM Revision 9c15da0f (git): Enhanced RDoc for IO (#5381)
- Treats:
#lineno
#lineno=
#readline
#readlines
#each_line
#each_byte
#each_char
#each_codepoint -
- Applied in changeset commit:git|9e0a91d0640600f2dfd7fc1d5fae6667019c9ca5.
----------
Don't segfault if Warning.warn is undefined
Check that there is a method entry for the method before passing
it to rb_method_entry_arity.
Fixes [Bug ... -
- Pull requested submitted to fix this: https://github.com/ruby/ruby/pull/5391
-
02:21 PM Bug #18458 (Closed): Segmentation fault when missing Warning#warn method
- If you remove the `Warning#warn` method and call `Kernel#warn`, you get a segmentation fault:
ruby 3.1.0:
```
$ ruby -e 'Warning.undef_method(:warn); warn ""' 2>&1 | head -2 ... -
- Check that there is a method entry for the method before passing
it to rb_method_entry_arity.
Fixes [Bug #18458] - 03:28 PM Revision e7806049 (git): * 2022-01-05 [ci skip]
-
03:27 PM Revision ad007bc6 (git): Switch `is_pointer_to_heap` to use library bsearch
- This commit switches from a custom implemented bsearch algorithm to
use the one provided by the C standard library.
Because `is_pointer_to_heap` will only return true if the pointer
being searched for is a valid slot starting address wi... -
- Applied in changeset commit:git|615e9b28658c5b44a4474e04a53b84ae83b8e3fd.
----------
[Feature #18364] Add GC.stat_heap to get stats for memory heaps
GC.stat_heap will return stats for memory heaps. This is
used for the Variable Width A... -
- GC.stat_heap will return stats for memory heaps. This is
used for the Variable Width Allocation feature. -
02:18 PM Revision e9a4cc02 (git): [ruby/error_highlight] Add a test to check if it is robust against a spoofed filename
- Previously, RubyVM::AST.of opened a wrong file if the iseq's file path is spoofed.
```
module Dummy
binding.irb
end
```
```
$ ruby test.rb
From: test.rb @ line 2 :
1: module Dummy
=> 2: binding.irb
3: end
irb(Dummy):001... - 02:09 PM Revision 11b1ebe6 (git): [ruby/error_highlight] Fix the spurious TypeError.
- When we have no backtrace locations, we can't have the highlight,
so just return the message.
https://github.com/ruby/error_highlight/commit/9f5c639494 - 02:09 PM Revision e22218b5 (git): [ruby/error_highlight] Reproduce the error seen when calling .to_s in embedded Ruby
- The test fails with the following error:
Error: test_simulate_funcallv_from_embedded_ruby(ErrorHighlightTest): TypeError: wrong argument type nil (expected method)
https://github.com/ruby/error_highlight/commit/52943c9cd2 -
11:09 AM Feature #18136: take_while_after
- `take_until`, from reading the method name, clearly implies to me including the element for which the condition became true, i.e., it includes that "last" element, like `1..3` includes `3`.
But if that's too confusing for others then it... -
- I also think the approach of yielding an enumerable while smart looks really complicated to implement properly, and I could imagine it might cause issues in some VMs.
Also if I understand correctly, that would mean `enum.each` would star... -
-
08:45 AM Bug #18454: YJIT slowing down key Discourse benchmarks
- What about YJIT_STATS=1 ? That's the part that will tell us why it's slower.
> Oddly bootsnap may be causing an RSS increase 338 -> 370 (on 2.7.5 290 -> 315)
RSS isn't necessarily the best mesure, but yes bootsnap incur increased m... -
02:29 AM Bug #18454: YJIT slowing down key Discourse benchmarks
- @k0kubun already doing 20 iterations as a warmup, so it is not likely :cry: https://github.com/discourse/discourse/blob/bbca25e875c36d3541b22c520b6d56ae6322dfbe/script/bench.rb#L192-L192
@byroot @noahgibbs this is a tough one Ruby ... -
- This fixes the following failure.
```
1) Error:
TestMethod#test_method_list:
NoMethodError: undefined method `<=>' for #<BasicObject:0x00007f7757e7eb60>
mods = mods.sort_by {|m| m.name }
^^^^^^^^
```
https://gith... -
08:25 AM Revision 47bf64a2 (git): Use omit instead of skip: test/ruby/**/*.rb
-
- https://github.com/ruby/optparse/commit/e4f2682128
-
- * [Feature #18438] Add `Exception#additional_message` to show additional error information (mame)
* I'd like to apply error_highlight more exceptions than NameError, but it will break some existing tests that check the return value of... -
05:59 AM Revision 5074aa19 (git): [rubygems/rubygems] Test the actual checksums of the mock gems
- https://github.com/rubygems/rubygems/commit/2b42630959
-
- "NUL.*" means the NUL device on Windows, as well as mere "NUL",
and no data is read.
https://github.com/rubygems/rubygems/commit/e2c7d22745 -
- https://github.com/rubygems/rubygems/commit/48ea2778e9
-
- Calculate the checksum of the content, not the given pathname at
the build time itself.
https://github.com/rubygems/rubygems/commit/b60ee97ee9
01/03/2022
-
- Server Name Indication does not allow IP addresses (RFC 6066, section 3:
`Literal IPv4 and IPv6 addresses are not permitted in "HostName".`).
Recent versions of LibreSSL enforce this restriction, which raises
when setting the hostname t... - 08:11 PM Revision 8d6e4f15 (git): * 2022-01-04 [ci skip]
-
07:36 PM Feature #18136: take_while_after
- `until` is the negative of `while`, so I would expect `take_until{cond}` (or the synonym `take_till`) to behave like `take_while{!cond}`. So `(1..9).take_until{_1==5}` would/should produce `[1,2,3,4]`, unlike the current proposal to get ...
-
- I recently wanted to have something like this for advent of code, and worked around by using `Enumerator.produce + find + side effect` but it feels clearly less elegant.
https://github.com/eregon/adventofcode/blob/a582daa1f2c549d459d6b0... -
- * [Bug #15928] Constant declaration does not conform to JIS 3017:2013 (jeremyevans0)
* This was discussed at the May 2021 developer meeting, and it was decided to wait until after 3.1.
* This makes constant assignment consistent wi... -
- headius (Charles Nutter) wrote in #note-15:
> Yes I botched that part of my example. The Enumerable would have to be yielded to a block and used within that block. My example done right would be more like:
> ...
I think this approach o... -
- I see, I agree more with GitHub assessment.
If I understand correctly, ruby 2.6 will be declared as EOL soon, so it seems fine to me to way for that to happen and don't take any action. Users who want to stick to Ruby 2.6 but use a fi... -
10:19 AM Bug #18431: Ruby 2.6.9, bundler 1.17.2 and CVE-2021-43809
- deivid (David Rodríguez) wrote in #note-3:
> How is the score for vulnerabilities calculated? I tried to set the score myself to "Low" in the Github Advisory, because the chances that this issue is ever explored seemed very low to me. I... -
05:43 PM Revision 3ae8b115 (git): YJIT: Fix confusing self reference when initializing variable
-
- Correct method name, fix copypasta, and add an ASCII illustration.
[ci skip] -
02:13 PM Bug #18454: YJIT slowing down key Discourse benchmarks
- Possibly related: YJIT released 3.1 was breaking Rails caching: https://bugs.ruby-lang.org/issues/18453
-
01:33 PM Revision 980bf94f (git): Kernel#=~: delete
- Has been deprecated since ebff9dc10e6e72239c23e50acc7d3cbfdc659e7a.
-
01:00 PM Bug #18457 (Closed): ruby 2.7.5 fiddle/types.rb use uint32_t but fiddle/cparser.rb lacks uint32_t
- ```
C:\src\git\reline>bundle exec rake test_yamatanooroti
Traceback (most recent call last):
17: from C:/Ruby27-x64/lib/ruby/gems/2.7.0/gems/rake-13.0.6/lib/rake/rake_test_loader.rb:6:in `<main>'
16: from C:/Ruby27-x6...
01/02/2022
- 10:06 PM Revision 3ff762cc (git): * 2022-01-03 [ci skip]
-
10:06 PM Revision 66e92561 (git): [ruby/reline] Remove test TODO comments that have been confirmed by E2E tests
- https://github.com/ruby/reline/commit/2ed77b693f
-
11:24 AM Revision c8694303 (git): Update RBS to skip testing `Kernel#=~` (#5386)
- 08:47 AM Revision 03d97cc9 (git): Remove UTF-8 from documentation.
-
08:43 AM Revision e7c2d5cc (git): [DOC] Adjust IO::Buffer docs (#5374)
-
08:22 AM Bug #18132: TODO: fix ccan/list thread safety
- Can you explain what VM list has a problem?
-
- https://github.com/ruby/etc/commit/f87fe7ad93
- 06:06 AM Revision 30c03f9e (git): Update default gems list at 6f53425825eb71d71c7fdd424801a2 [ci skip]
-
-
- https://github.com/ruby/etc/commit/143edbd215
-
- https://github.com/ruby/etc/commit/ca41f182f6
-
- https://github.com/ruby/etc/commit/82dd69a4cf
-
- https://github.com/ruby/etc/commit/17b99fe42e
-
- https://github.com/ruby/etc/commit/106429f10b
-
- https://github.com/ruby/etc/commit/8c8c6198e2
-
- https://github.com/ruby/etc/commit/959416b18a
-
- Incorrect arguments can cause SEGV.
https://github.com/ruby/reline/commit/a58748bcf9 -
- https://github.com/ruby/reline/commit/1ca779740a
- 02:29 AM Revision b2a88063 (git): Update default gems list at 6d1b406dc8dea6f618ae14899dc6b7 [ci skip]
-
- https://github.com/ruby/reline/commit/9ab5850444
-
01:29 AM Misc #18399: DevMeeting-2022-01-13
- - [Bug #18294] error when parsing regexp comment (Martin Dürst)
- We should decide whether this is spec, or to make parsing easier, or a bug. -
- https://github.com/ruby/reline/commit/695212d5d2
01/01/2022
-
11:15 PM Bug #18456 (Closed): rdoc non-determinism: module includes can be added once or twice to generated .ri
- We're observing a sporadically occurring non-determinism in generated .ri files:
```
[ak@localhost ruby-repro]$ ri --dump=./oe-reproducible-20211214-gj4vndgo/packages-excluded/reproducibleA/tmp/deploy/rpm/core2_64/usr/share/ri/3.0.0/... -
- 03:40 PM Revision fa5b32f1 (git): * 2022-01-02 [ci skip]
-
- https://github.com/ruby/reline/commit/af4d77ba09
-
- https://github.com/ruby/reline/commit/62eec42e4a
-
- https://github.com/ruby/reline/commit/abc1e4ee88
-
- https://github.com/ruby/reline/commit/0451ed7a28
-
12:30 PM Feature #12901: Anonymous functions without scope lookup overhead
- One more reason for having an isolated kind of Proc is using it with an implicit block argument - for `define_method` or for some other metaprogramming. For example, `yield` in this code is incorrect due to the scope capturing problem:
... -
- Ah, the difference is thread `t` won't be interrupted anymore, but `close(2)` probably stops the blocking read.
@ioquatix So what's the result of the two snippets after this patch? -
- Nice, I thought this logic was already per IO instance.
I don't see how it will change anything about `this leads to strange behaviour` though.
Which is fine because it seems unavoidable as long as `IO.for_fd(fd)` exists.
`IO.for_fd(fd... -
- Here is the PR: https://github.com/ruby/ruby/pull/5384
-
- `IO#close` should be responsible for closing the file descriptor referred to by the IO instance. When dealing with buffered IO, one can also expect this to flush the internal buffers if possible.
Currently, all blocking IO operations ... -
- > Any ideas on what we should test here and how to narrow this down?
You could compile Ruby with `YJIT_STATS=1` to have a break down of why and how much YJIT exit back to the interpreter. It's possible that you have some hotspots usin... -
- @mame Thanks for making this proposal, I believe it goes in the right direction.
Dan0042 (Daniel DeLorme) wrote in #note-6:
> I had the same thought initially but there's a hiccup: the description needed by full_message should includ... -
-
- The respond_to expectation just suffice as duck-typing.
-
-
-
- 07:23 AM Revision 243320d2 (git): Fix incorrect minimum read length. (#5353)
-
- Pretty much anything will be better than the current behaviour. I think your proposal makes sense.
-
- Thank you for such a comprehensive bug report!
-
01:32 AM Bug #18453 (Closed): YJIT breaks Rails collection caching
- Applied in changeset commit:git|5414de4b6e4372af832e338f8eb7a9fe8de17c84.
----------
YJIT: Fix SP index with optarg and unordered kwarg
Previously when we were calling a method with an optional argument and
multiple keywords arguments ... -
- Previously when we were calling a method with an optional argument and
multiple keywords arguments which weren't in the order the receiver
expected we would use the wrong SP index to rearrange them.
Fixes Bug #18453
12/31/2021
-
09:23 PM Bug #18454: YJIT slowing down key Discourse benchmarks
- To see what's going on, could you also try running some iterations (at least 10, the default --yjit-call-threashold) before starting performance measurement and/or increasing the number of iterations?
-
4 out of 6 key Discourse benchmarks used in: https://github.com/discourse/discourse/blob/main/script/bench.rb are slower under YJIT.
Vanilla (median) -> YJIT (median)
Categories: 47ms -> 51
Home page: 85 -> 86
Topic: 42 -> 38
...-
- Yes, supporting Window's x64 calling convention is non-trivial.
In addition to what John already mentioned we also need to uphold [unwindability](https://docs.microsoft.com/en-us/cpp/build/x64-calling-convention?view=msvc-170#unwindabil... -
05:39 PM Feature #14982: Improve namespace system in ruby to avoiding top-level names chaos
- I really like this idea and can really reduce the name clashes and global namespace pollution.
- 03:00 PM Revision 2fac066b (git): * 2022-01-01 [ci skip]
-
-
- 11:13 AM Revision a493732c (git): re-enabled YJIT on x64-mswin64
- note that YJIT does not work correctly on the platform now.
- 10:48 AM Revision 1034b6e7 (git): merge revision(s) 0c5f8c62766afe4605172800063e63fe36996658: [Backport #10961]
- [ruby/zlib] Resume zstream if available [Bug #10961]
---
ext/zlib/zlib.c | 6 ++++++
1 file changed, 6 insertions(+) - 10:48 AM Revision 1cc5567b (git): merge revision(s) 0846c2da457e7523819236ac7da492029b3ef73d,6c7cb00c094332a208cf36e5cd723a9ba60c41b8: [Backport #16376]
- Check backref number buffer overrun [Bug #16376]
---
regcomp.c | 21 ++++++++++++---------
test/ruby/test_regexp.rb | 6 ++++++
2 files changed, 18 insertions(+), 9 deletions(-)
... - 10:48 AM Revision 426266af (git): merge revision(s) fdf39963490cf2cf95b30d91bb9b35964c2c2350: [Backport #18421]
- Empty and return the buffer if zero size is given [Bug #18421]
In `IO#readpartial` and `IO#read_nonblock`, as well as `IO#read`.
---
io.c | 8 ++++++--
test/ruby/test_io.rb | 21 ++++++++... - 10:48 AM Revision 7d3cff6e (git): merge revision(s) a2d4e1cda68a49980a4f9f353f400efbde7e7884,d6c5a30cfdf658280338dbb8c8b17fab3190b928: [Backport #18392]
- Fixed the check order in wmap_live_p [Bug #18392]
Check if the object is a pointer to heap before check the flag in
that object.
---
gc.c | 35 ++++++++++++++++++++++-------------
1 file changed,... - 10:48 AM Revision b8298111 (git): merge revision(s) 43a3c88187dcd54ba008834a865e50eb28a78de9: [Backport #18372]
- [win32] get rid of redefinition of reserved macro
```
warning C4117: macro name '_INTEGRAL_MAX_BITS' is reserved, '#define' ignored
```
---
win32/Makefile.sub | 2 ++
1 file changed, 2 in... - 10:47 AM Revision c24f436c (git): merge revision(s) 9f8a50723f8a84b3e4755b418570148f422d1b28: [Backport #17836]
- Specify -c to emit pch with clang (#4423)
[Bug #17836]
---
mjit_worker.c | 1 +
1 file changed, 1 insertion(+) -
10:44 AM Bug #10961: Zlib corrupts data when receive signal
- ruby_2_7 50a986ca582bf16ed643cab33bad9d3d7ebb815d merged revision(s) 0c5f8c62766afe4605172800063e63fe36996658.
-
- ruby_2_7 c4a5eaf37b9bf87f8324a7070c8dbc597b89c4fc merged revision(s) fdf39963490cf2cf95b30d91bb9b35964c2c2350.
-
- ruby_2_7 836aca4fee4d27a55d56044620a00866a2c279ef merged revision(s) a2d4e1cda68a49980a4f9f353f400efbde7e7884,d6c5a30cfdf658280338dbb8c8b17fab3190b928.
-
- ruby_2_7 e42f8fc5a6ae869652cd1feb3d33353f0ad6a4ec merged revision(s) 43a3c88187dcd54ba008834a865e50eb28a78de9.
-
- ruby_2_7 57f8bbcf63f8e2ef5c2cea488ba21c3e7c34f862 merged revision(s) 9f8a50723f8a84b3e4755b418570148f422d1b28.
-
12/30/2021
-
- Adds sections to class RDoc:
Lines
Line Separator
Line Limit
Line Number
Revises example text file t.txt to:
Include paragraphs (separated by double line separator).
Avoid being too long.
Revises e... -
- Applied in changeset commit:git|f53dfab95c30e222f67e610234f63d3e9189234d.
----------
Add support for anonymous rest and keyword rest argument forwarding
This allows for the following syntax:
```ruby
def foo(*)
bar(*)
end
def baz(**)... - 10:37 PM Revision a2f47cf8 (git): * 2021-12-31 [ci skip]
-
- This allows for the following syntax:
```ruby
def foo(*)
bar(*)
end
def baz(**)
quux(**)
end
```
This is a natural addition after the introduction of anonymous
block forwarding. Anonymous rest and keyword rest arguments were
alrea... -
- I think only enabling on explicit --jit/--yjit/--mjit is better because it's more intuitive.
If e.g. ` --yjit-exec-mem-size=32` is in RUBYOPT it seems unexpected that it would enable YJIT without any `--yjit` anywhere. -
- https://github.com/ruby/ruby/pull/5379
-
- Tracked this down to `cache_fragment_name` (https://github.com/rails/rails/blob/main/actionview/lib/action_view/helpers/cache_helper.rb#L240) getting the wrong args, so probably a bug in kwarg handling.
-
05:54 PM Bug #18453 (Closed): YJIT breaks Rails collection caching
- A minimal app that demonstrates this issue is available [on GitHub](https://github.com/georgeclaghorn/yjit-rails-collection-caching-bug). It has:
* One model, `Post`, with a string `title` attribute.
* One controller action, `posts#ind... -
-
04:59 PM Bug #18452: Dramatic performance regression in Zeitwerk with 3.1
- > I think your original assumption is more reasonable, and Bootsnap's assumption is not.
OK, thanks for the hint.
> ...
In projects managed by Zeitwerk, subdirectories represent namespaces. If the namespace has a file defining it, all ... -
- fxn (Xavier Noria) wrote in #note-2:
> I thought about different options, and believed that Bootsnap made a reasonable assumption, so just a few days ago I published a new version that pushed the directory to `$LOADED_FEATURES`.
> ...
... -
- BTW, Zeitwerk does not call the original `require` either. But it is a fact that if you comment out the `push` call, the regression disappears.
-
- > Is there a reason Zeitwerk is adding directories to $LOADED_FEATURES, instead of using its own data structure?
It didn't do that since it first shipped. The lib intercepted the `require` and did its thing without invoking the original... -
- The bug fixed in commit:79a4484a072e9769b603e7b4fbdb15b1d7eccb15 is always going to result in some slow down, since Ruby now has to determine the realpath for all entries in `$LOADED_FEATURES`. The commit doesn't change when the loaded ...
-
- Hi!
In Ruby 3.1, Zeitwerk loads implicit namespaces way slower than in previous versions. In this benchmark:
```ruby
require 'fileutils'
require 'tmpdir'
require 'zeitwerk'
Dir.mktmpdir do |dir|
Dir.chdir(dir)
for i i... -
- Cast to `void *` first to use the definition of `intptr_t`.
-
12:16 PM Revision d5fc3fa4 (git): win32/win32.c Fix PROT_EXEC bit flag check for FlushInstrucitonCache()
-
- https://github.com/ruby/win32ole/commit/27d0fdc622
-
- https://github.com/ruby/reline/commit/c559d0f7a9
-
- Some tokens in Ruby 2.6 have difference information than in 2.7 and later,
but 2.6 will soon be out of support.
https://github.com/ruby/reline/commit/f3bc698385 -
- https://github.com/ruby/reline/commit/b7536dc224
-
11:23 AM Revision f589242e (git): [ruby/reline] Use unix_line_discard when Ctrl-u is entered
- The kill-line was called when C-u was entered, so it is now called unix-line-discard.
In readline(3):
> unix-line-discard (C-u)
> ...
https://github.com/ruby/reline/commit/27570d195e -
10:33 AM Feature #18376: Version comparison API
- pocke (Masataka Kuwabara) wrote in #note-7:
> `::Version` will be conflict with `optparse`'s convention. Probably we need to consider the naming.
Maybe `Semver`?
This is a bit clunky but would arguably indicate the core functional... -
-
- Thanks, @peterzhu2118!
-
-
- mame (Yusuke Endoh) wrote in #note-3:
> Looks like `\c?` in a regexp literal was changed for #14367.
> ...
As @janosch-x mentioned, the matched codepoints are the same. The fact that #source returns a different result does not seem lik...
12/29/2021
-
- i really like this in Swift so i'll try to give another example for how this addition might improve both readability and conciseness.
currently there are two approaches: using verbose names for keyword arguments or short ones. both have... -
- regexps with these escapes can still be constructed with the `Regexp::new` constructor, they are only pre-processed to hex escapes in Regexp literals.
```ruby
/\cC/.source == Regexp.new('\cC').source # false iff Ruby >= 3.1
```
a... -
05:24 AM Bug #18449: Bug in 3.1 regexp literals with \c
- I was just coming back to point at:
```
Jeremy Evans: Fix handling of control/meta escapes in literal regexps [Wed May 12 12:37:55 2021 -0700 (8 months ago)]
```
found in https://github.com/ruby/ruby/commit/11ae581a4a7f5d5f5ec637... -
- Looks like `\c?` in a regexp literal was changed for #14367.
```
p(/\cC/.source) #=> "\\cC" in Ruby 3.0
p(/\cC/.source) #=> "\\x03" in Ruby 3.1
```
@jeremyevans0 What do you think? -
- It looks like `tokadd_escape` has drastically changed and dropped the `\c`, `\M-`, and `\C-` forms...
This isn't mentioned in the release notes, and seems a backwards incompatibility that should be reserved for 4.0: https://www.ruby-... -
- This file passes on 2.7, 3.0, and fails (if you remove the `skip` line) on 3.1:
``` ruby
#!/usr/bin/env ruby -w
require "minitest/autorun"
class TestRegexpCreation < Minitest::Test
R31 = RUBY_VERSION > "3.1"
def test_li... -
- this affects:
- all String escapes that can be invalid (`\x`, `\u`, `\u{...}`, `\M`, `\C`, `\c`)
- only invalid escapes (e.g. `\x7F` is fine)
- no Regexp-specific escapes such as `\p{...}`, `\g<...>`, `\k<...>`
- Regexp literals (`Synt... -
- The current way works, but error messages show duplicate "/" in paths,
which is weird.
https://github.com/rubygems/rubygems/commit/9123deb4fa -
09:54 PM Bug #18447: Potential performance regression with String#lines in large strings
- Thanks everyone! I wondered if it wasn't due to memory allocations, but the "build a massive string" operation was successful and quite fast. I looked through `rb_str_enumerate_lines` in the C source, and got lost rather quickly, aside f...
-
- Applied in changeset commit:git|002fa2859962f22de8afdbeece04966ea57b7da9.
----------
On 64bit macOS, enlarge heap pages to reduce mmap calls [Bug #18447] -
- I think we should consider making the page sizes 64KB by default. Many systems (e.g. PowerPC) have 64 KB page sizes, so we can't use mmap on those systems meaning that we can't use compaction. See issue #18133.
I opened a PR for this ... -
- This seems to make it faster nearly to 3.0.
https://github.com/nobu/ruby/tree/mac-heap_page
```diff
diff --git a/gc.c b/gc.c
index a5369c28f13d..aaae009b2ee7 100644
--- a/gc.c
+++ b/gc.c
@@ -847,8 +847,15 @@ typedef struct rb_objs... -
- Thank you for your great report. I could reproduce the issue on M1 macOS. It does not repro on Ubuntu.
According to `git bisect`, this regression is caused by commit:785f5eb8f0f994c755ac5f1e1479f6e2ec6d0ef3.
```
% git checkout 09f284c0... -
-
- I added documentation for these keys in commit [52525b527b4605435110bd61fb9e04d82ca26ff7](https://github.com/ruby/ruby/commit/52525b527b4605435110bd61fb9e04d82ca26ff7).
-
- Applied in changeset commit:git|52525b527b4605435110bd61fb9e04d82ca26ff7.
----------
[Bug #18451] [ci skip] Update documentation for GC.stat
Adds documentation for keys `time`, `compact_count`,
`read_barrier_faults`, `total_moved_objec... -
- As of 3.1 these keys aren't [documented](https://docs.ruby-lang.org/en/master/GC.html#method-c-stat):
* `:compact_count` (2.7)
* `:read_barrier_faults`, `:total_moved_objects` (3.0)
* `:time` (new in 3.1)
(I am just judging by vers... - 09:21 PM Revision 7304e357 (git): * 2021-12-30 [ci skip]
-
- Adds documentation for keys `time`, `compact_count`,
`read_barrier_faults`, `total_moved_objects`. -
- So this serves roughly the same purpose as #18296
I also have a slight preference for a `description` method that includes the `message`. It seem simpler for everyone (including Sentry, Datadog, etc.) to use `e.description` instead of... -
- ivoanjo (Ivo Anjo) wrote in #note-3:
> 1. Would `additional_message` only be something that gems like did_you_mean and error_highlight use (e.g. better debugging/developer experience tools), or do you see normal user code using it as we... -
- What if instead of having two distinct `#message` and `#additional_message` that need to be concatenated, there would simply be an alternative, more descriptive method?
e.g.
`Exception#description` which would be:
```ruby
def m... -
04:06 PM Bug #18429: Configure ruby-3.0.3 on Solaris 10 Unknown keyword 'URL' in './ruby.tmp.pc'
- # gmake
BASERUBY = echo executable host ruby is required. use --with-baseruby option.; false
CC = /usr/local/gcc4/bin/gcc -std=gnu99
LD = /usr/ccs/bin/ld
LDSHARED = /usr/local/gcc4/bin/gcc -std=gnu99 ... -
- We are using newer version of gcc now :
/usr/local/gcc4/bin/gcc --version
gcc (GCC) 4.5.1
-
- dklein (Dmitri Klein) wrote in #note-20:
> ```shell-session
> ...
It should be `CC=/opt/csw/bin/gcc-5.5 ./configure`.
> ```shell-session
> ...
Seems the necessary library is not installed. -
02:02 PM Feature #18450: Force break in prettyprint
- mame (Yusuke Endoh) wrote in #note-1:
> Could you explain a use case?
I am writing a code generator where in some cases I need to force-break the output code because it makes more sense for a human reader/reviewer:
Here's an example tak... -
- Could you explain a use case?
-
- # Abstract
Support force-breaking a group in the std's [prettyprint](https://github.com/ruby/prettyprint)
# Background
There is a need to forcibly break a group and transform breakables into
newlines. The library doesn't provid... -
-
- https://github.com/ruby/reline/commit/a538de421f
-
-
-
- Looks like there is something wrong in your gcc installation. Maybe you are using GCC with Solaris's as commend. I have no idea that GCC supports such a setup.
-
-
- The log contains:
> ld.so.1: conftest: fatal: libssp.so.0: open failed: No such file or directory
I think this means your GCC installation is not working well. -
- # /usr/local/gcc4/bin/gcc --version
gcc (GCC) 4.5.1
# which gcc
/usr/local/gcc4/bin/gcc
#
# ./configure
...
checking size of long long... configure: error: in `/export0/home/telegraf/fluentd/ruby_versions/ruby3/ruby-3.0.3':... -
- Can you use a single ticket for this failure? Please use https://bugs.ruby-lang.org/issues/18429
- 07:02 AM Revision 6d572902 (git): Update bundled gems list at 2021-12-29
-
-
- `??` is a character literal.
As omitting the `:` and after makes the syntax value conflicting of course, I don't think this is acceptable. -
-
-
- 01:03 AM Revision 7fbd93e6 (git): Update default gems list at d75f7078c831d45ab5ba2fae4fbb30 [ci skip]
-
- https://github.com/ruby/io-console/commit/ad3805200c
- 12:37 AM Revision e4cd319f (git): [ruby/irb] Prefer `IO#wait_readable` over `IO#select`. (https://github.com/ruby/irb/pull/323)
- https://github.com/ruby/irb/commit/1c03bd3373
- 12:27 AM Revision f27eb814 (git): Sync io-console gem.