Activity
From 06/14/2023 to 06/20/2023
06/20/2023
-
05:09 PM Misc #19599: DevMeeting-2023-05-10 @ Matsumoto, Japan
- > Is a log going to be posted?
I've just posted it at https://github.com/ruby/dev-meeting-log/blob/master/2023/DevMeeting-2023-05-10.md. -
03:13 PM Revision e25403d0 (git): [ruby/irb] Improve indentation: bugfix, heredoc, embdoc, strings
- (https://github.com/ruby/irb/pull/515)
* Implement heredoc embdoc and string indentation with bugfix
* Fix test_ruby_lex's indentation value
* Add embdoc indent test
* Add workaround for lines==[nil] passed to auto_indent when exit I... -
03:06 PM Bug #19739: String coderange not cleared by String#slice!
- I suspect https://github.com/ruby/ruby/pull/5867 is what fixed it, but don't have time to confirm just yet.
-
- That is very interesting.
The two strings are identical (aside from one being frozen):
```ruby
require "objspace"
hash = {}
key = "ABC OÜ"
p key.hash
key.slice!(/ oü$/i)
p key.hash
puts ObjectSpace.dump(key)
hash[key] ... -
02:07 PM Bug #19739 (Closed): String coderange not cleared by String#slice!
- Hello,
Key cannot be found in a Hash when slice! method is applied to the key. I cannot reproduce this behaviour with sub! or gsub! methods:
``` ruby
hash = {}
key = "ABC OÜ"
key.slice!(/ oü$/i)
# key.sub!(/ oü$/i, '')
# key.gs... -
- Opening this as Misc, as at this stage I don't have a fully formed feature request.
Ref: https://github.com/ruby/ruby/commit/1a3bcf103c582b20e9ea70dfed0ee68b24243f55
Ref: https://github.com/ruby/timeout/pull/30
Ref: https://github.... - 02:02 PM Revision 9ce6e096 (git): [ruby/reline] Omit constant under Struct
- (https://github.com/ruby/reline/pull/554)
https://github.com/ruby/reline/commit/8761a11fa5 -
01:28 PM Revision 932dd9f1 (git): [DOC] Regexp doc (#7923)
-
12:24 PM Bug #19736 (Closed): Ripper.tokenize ignores token after heredoc_beg when embexpr inside heredoc is unclosed
- Applied in changeset commit:git|6be402e172a537000de58a28af389cb55dd62ec8.
----------
[Bug #19736] Recover after unterminated interpolation -
-
-
- `rb_ivar_foreach` disallows `ST_DELETE`.
-
-
04:35 AM Revision 1edbaa85 (git): Merge rubygems/rubygems HEAD
- Pick from https://github.com/rubygems/rubygems/commit/880dd95996c93adc1e032399816931b243c5fe17
-
-
02:29 AM Bug #19738: `ObjectSpace.each_object.to_a` crashes in `make runirb`.
- In order to get the assertion failure, you must compile with RUBY_DEBUG=1:
```
./configure -C --disable-install-doc --prefix=/Users/samuel/.rubies/ruby-head cppflags=-DRUBY_DEBUG=1
``` -
- ```
> make runirb
RUBY_ON_BUG='gdb -x ./.gdbinit -p' ./miniruby -I./lib -I. -I.ext/common ./tool/runruby.rb --extout=.ext -- --disable-gems -r irb -e 'IRB.start("make runirb")'
irb(main):001:0> ObjectSpace.each_object.to_a
Assert... - 01:48 AM Revision 504f37af (git): more fix
- 01:32 AM Revision 48b00c4d (git): Fix previous commit
-
12:49 AM Feature #19737: Add `IO::Buffer#cat` for concat `IO::Buffer` instances
- My use case is this: concat given two buffers, then modify (copy, slice, xor) it.
(More concretely, I want to apply/remove header protection in the QUIC protocol.)
I understood that using copy method is a better way than concat buffe...
06/19/2023
-
- Thanks for working on this.
So, in short, I'm against the name (even if it is cute), but I'm okay with considering the functionality.
I'd like to know more the use case.
We shouldn't encourage users to create lots of temporary b... -
07:48 PM Feature #19737: Add `IO::Buffer#cat` for concat `IO::Buffer` instances
- Could you include your actual concrete use case?
See https://github.com/ruby/ruby/wiki/How-To-Request-Features
Regarding naming, `cat` or `concat` feels wrong, `+` seems a much better match for the proposed semantics (no mutation).
... -
- Opend here <https://github.com/ruby/ruby/pull/7960>
-
- ## motivation
In my use case, I want to concat two IO::Buffer instances. But current implementation doesn't have that way.
Then I created a patch. Opend here: TBD
## concern
I have two concerns about it.
### 1. Should we provide... -
- Dan0042 (Daniel DeLorme) wrote in #note-32:
> Ah interesting, I didn't realize it was already used that way. Since the ruby main repo already demonstrates how to "lie" about the class name in a sensible and useful way, I believe this is... -
05:46 PM Feature #19521: Support for `Module#name=` and `Class#name=`.
- > However, this will prevent `labeled_class`/`labeled_module` from using `set_temporary_name`, at least until those labels are updated to be "not a constant name" or we remove this restriction. When I did this change, ~35 tests failed. I...
-
- I am okay with restricting names to be not-constant-names, e.g.
```c
if (rb_is_const_name(name)) {
rb_raise(rb_eArgError, "name must not be valid constant name");
}
```
However, this will prevent `la... -
03:34 PM Feature #19521: Support for `Module#name=` and `Class#name=`.
- ioquatix (Samuel Williams) wrote in #note-25:
> And in any case, your example is already trivially possible without this PR/change:
> ...
There is a big difference with what's possible today and what would be possible after this change... -
- @ioquatix, how would you feel about raising an error if the temporary name is a valid constant name?
-
- ioquatix (Samuel Williams) wrote in #note-25:
> @Eregon I understand your point, but why would anyone do that? Isn't that kind of shooting yourself in the foot?
Well, even the docs and specs you added in that PR do that, it's inviting p... -
07:19 PM Revision 9ff4399d (git): CI: work around vcpkg openssl issue with OpenSSL providers
- Currently, the openssl package from vcpkg uses an incorrect, hard-coded
default location for OpenSSL providers under C:\vcpkg\packages, which
is supposed to be a temporary directory.
Override it with the OPENSSL_MODULES environment vari... - 06:16 PM Revision 179f0e53 (git): [rubygems/rubygems] Bump rb-sys
- Bumps [rb-sys](https://github.com/oxidize-rb/rb-sys) from 0.9.78 to 0.9.79.
- [Release notes](https://github.com/oxidize-rb/rb-sys/releases)
- [Commits](https://github.com/oxidize-rb/rb-sys/compare/v0.9.78...v0.9.79)
---
updated-depende... -
- https://github.com/rubygems/rubygems/commit/97dbe4cabd
-
- * Make TAINT and UNTRUSTED flags zero
These flags do nothing already, and should break nothing.
* Remove TAINT and UNTRUSTED macros same as functions
These macros had been defined to use with `#ifdef`, but should not be
used anymore. - 11:49 AM Revision 6302bf2e (git): merge revision(s) d2520b7b76759118071a16e6bca22726a5de9fb4: [Backport #19439]
- Marshal.load: restore instance variables on Regexp
[Bug #19439]
The instance variables were restore on the Regexp source,
not the regexp itself.
Unfortunately we have a bit of a chicken and egg problem.... -
11:28 AM Bug #19323: Integer overflow in `Integer#<<`
- backported into `ruby_3_1` at 344249674f7a7835445a44695664897fa4b83f00
- 11:21 AM Revision 34424967 (git): merge revision(s) 1cdf8ab07b24ebd16e93621957196e8b1d67f2ba: [Backport #19323]
- [Bug #19323] Raise `RangeError` instead of integer overflow
---
bignum.c | 5 ++++-
test/ruby/test_integer.rb | 18 ++++++++----------
2 files changed, 12 insertions(+), 11 deletions(-) -
- (https://github.com/ruby/irb/pull/608)
https://github.com/ruby/irb/commit/1cd3b45402 -
- I expect Ripper.tokenize(code) to tokenize `+1+2+3\n` part.
```ruby
code = <<'RUBY'
<<A+1+2+3
#{4
RUBY
Ripper.tokenize code
# => ["<<A", "\#{", "4", "\n"]
```
It will be easy for IRB's internal code to handle indent and othe... -
-
08:45 AM Revision e9f62fa5 (git): [ruby/openssl] add OpenSSL Provider support
- https://github.com/ruby/openssl/commit/189c167e40
[rhe: tool/update-deps --fix to update ext/openssl/depend] -
- https://github.com/rubygems/rubygems/commit/657d57621e
-
- https://github.com/rubygems/rubygems/commit/6d445a85d7
06/18/2023
-
- My understanding is, because it's not strictly an assignment, it shouldn't look like that of the user. Also, it's more convenient to use:
```ruby
my_class = Class.new do
set_temporary_name "my_class"
end
```
Using `name=` or ... -
- Aside from the feature itself, I must say I'm very surprised by the `set_` prefix of that method, it's quite rare in Ruby.
-
- @Eregon I understand your point, but why would anyone do that? Isn't that kind of shooting yourself in the foot?
And in any case, your example is already trivially possible without this PR/change:
```
c = Class.new
String = c
c.... -
- @matz Are you really OK with things like:
```ruby
c = Class.new
c.set_temporary_name "String"
c.new.foo # => undefined method 'foo' for #<String:0x00007efc38711fc0> (NoMethodError)
c = Class.new
c.set_temporary_name "Object"
c.n... -
12:10 PM Feature #19521: Support for `Module#name=` and `Class#name=`.
- Accepted. Thank you.
Matz.
-
- clang generates a warning:
../../../../ext/openssl/ossl_pkey.c:326:22: error: implicit truncation from 'int' to a one-bit wide bit-field changes value from 1 to -1 [-Werror,-Wsingle-bit-bitfield-constant-conversion]
... -
- Ruby 2.6 has reached EOL on 2022-03. Requiring Ruby 2.7 allows us to
use C99 syntax.
https://github.com/ruby/openssl/commit/f6ba75e51e - 04:57 PM Revision bc0539a9 (git): [ruby/openssl] Fix modular square root test with LibreSSL >= 3.8
- If x is a modular square root of a (mod p) then so is (p - x). Both
answers are valid. In particular, both 2 and 3 are valid square roots
of 4 (mod 5). Do not assume that a particular square root is chosen by
the algorithm. Indeed, the a... -
- RUBY_OPENSSL_EXTCFLAGS and RUBY_OPENSSL_EXTLDFLAGS have been added for
the primary purpose of appending custom warning flags during
development and CI.
Since checking programs generated by mkmf may not be completely
warning-free, we don... - 04:57 PM Revision 366d8005 (git): [ruby/openssl] CI: Add the test/openssl/test_pkey.rb on the FIPS mode case.
- It's to test the `OpenSSL::PKey.read` in the `test/openssl/test_pkey.rb`.
I added the pending status to the following tests failing on the FIPS mode
case in the `test/openssl/test_pkey.rb`.
* `test_ed25519`
* `test_x25519`
* `test_comp... - 04:57 PM Revision 920bc712 (git): [ruby/openssl] Drop a common logic disabling the FIPS mode in the tests.
- We want to run the unit tests in the FIPS mode too.
https://github.com/ruby/openssl/commit/ab92baff34 - 04:57 PM Revision 2a483405 (git): [ruby/openssl] Workaround: Fix OpenSSL::PKey.read that cannot parse PKey in the FIPS mode.
- This commit is a workaround to avoid the error below that the
`OpenSSL::PKey.read` fails with the OpenSSL 3.0 FIPS mode.
```
$ openssl genrsa -out key.pem 4096
$ ruby -e "require 'openssl'; OpenSSL::PKey.read(File.read('key.pem'))"
-e:... -
- The vast majority have no reference so it's just a matter of setting the flags.
For the couple exception, they have very little references so it's
easy.
https://github.com/ruby/openssl/commit/2c7c6de69e - 04:57 PM Revision 0a84bd6b (git): [ruby/openssl] Append flags from environment variables.
- According to the `mkmf.rb#init_mkmf`, there are command line options below.
* `--with-cflags` to set the `cflags`
* `--with-ldflags` to set the `ldflags`
For example the following command compiles with the specified flags. Note that
`M... -
02:51 PM Bug #18686 (Closed): Regexp supporting unexpected age properties in Ruby < 3.2
-
12:38 PM Bug #18686: Regexp supporting unexpected age properties in Ruby < 3.2
- this no longer occurs in ruby 3.2 or 3.3 so IMO this bug can be closed @duerst @jeremyevans0
```ruby
RUBY_DESCRIPTION
# => "ruby 3.3.0preview1 (2023-05-12 master a1b01e7701) [arm64-darwin22]"
[*0..1000].product([*0..100]).each_wi... -
- I am negative for several reasons.
* IO (File) and Dir have different role and implementation in Unix, which Ruby design based on
* In Ruby, we cannot have multiple class parents, case for your proposal, File's super class must be ei... -
06:55 AM Bug #19701: The rb_classext_t::classpath field is not marked for T_ICLASS
- jeremyevans0 (Jeremy Evans) wrote in #note-2:
> There is a third option, set "FrozenCore" as an fstring that doesn't get garbage collected (via `rb_gc_register_address` or something). That reduces the cost to 1 object marking per major... -
05:39 AM Revision 1740482d (git): Fix rb_compile_option_t comments [ci skip]
-
-
- `yyparse` never changes the value of `coverage_enabled`.
`coverage_enabled` depends on only return value of `e_option_supplied`.
Therefore `parser_params` doesn't need to have `coverage_enabled.
06/17/2023
-
05:18 PM Feature #19735: Add support for UUID version 7
- I edited the description: the draft RFC URL previously pointed to the 6th revision of draft-ietf-uuidrev-rfc4122bis (https://www.ietf.org/archive/id/draft-ietf-uuidrev-rfc4122bis-06.html). But the 7th revision has already been posted. ...
-
- Although the specification for UUIDv7 is still in draft, the UUIDv7 algorithm has been stable as the RFC progresses to completion.
Version 7 UUIDs can be very useful, because they are lexographically sortable, which can improve e.g: d... -
02:53 PM Feature #19734 (Rejected): Let `Dir` methods be available to `File`
- I propose to let `Dir` singleton methods be available to `File` in some way. Motivations are as follows.
When we want to do something with a file (for example `File.read` or `File.write`), we quite often want to achieve the home direc... -
- sunc treats int bitfield as unsigned int.
This commit will fix build failure on sunc.
* http://rubyci.s3.amazonaws.com/solaris10-sunc/ruby-master/log/20230617T100003Z.fail.html.gz
* http://rubyci.s3.amazonaws.com/solaris11-sunc/ruby-mas... -
- This commit reduces dependency to CRuby object.
- 06:59 AM Revision e5ae7a16 (git): Update bundled gems list at 2023-06-17
-
06:48 AM Bug #19732 (Closed): Possible missing header (stdint.h) in event.h
- Applied in changeset commit:git|b943e9c7b9d9cc8ba4bbf043414ab1ed4e1a8b5f.
----------
Fixes [Bug #19732]. Add missing stdint.h header to event.h. -
-
-
- Include the functions which are only used for
`rb_hash_bulk_insert_into_st_table`. -
-
-
- I've updated the PR with documentation and tests: https://github.com/ruby/ruby/pull/7483
-
- Based on the discussion with Matz, here is the updated proposal:
```ruby
c = Class.new
# => #<Class:0x00000001031b99d8>
c.set_temporary_name("Hello")
# => Hello
C = c
# => C
C.set_temporary_name("Hello")
# `set_temporary_name': can't c...
06/16/2023
-
-
-
- Thanks for pointing that out. Updated the PR (https://github.com/ruby/ruby/pull/7945) with those adjustments.
-
- Sorry, I got wrongly if it were about to add it to debug.h.
The file is correct, but just the `include` should be after the other `include`s with `HAVE_STDINT_H` check. -
- Although I don't think it is a big issue as we don't assume or guarantee all our headers can be usable individually, welcome the improvement.
It seems ruby/internal/event.h is which really needs `uint32_t`.
```shell-session
$ clan... -
08:54 AM Revision 80dfe2bc (git): [ruby/rdoc] Fix missing closing colon for `:notnew:`
- https://github.com/ruby/rdoc/commit/1e550b7d02
-
- https://github.com/rubygems/rubygems/commit/b696edcd73
-
- Maybe it was already obvious to everyone, but I feel that something important has not been mentioned in this discussion: `(a >> b).call(x)` is equivalent to `b.call(a.call(x))` but **neither a nor b needs to be a Proc**. Either could be ...
06/15/2023
-
- While preparing that diff I realized what happened. I'm using `clang-format` which sorts includes and ordered https://github.com/banister/debug_inspector/blob/5424c4094df30adfecd961a4e77d95f1fea9bc48/ext/debug_inspector/debug_inspector.c...
-
- Could you share the code to reproduce?
Is it https://github.com/banister/debug_inspector/tree/master/ext/debug_inspector/debug_inspector.c? - 05:38 PM Revision 598ad56a (git): * 2023-06-15 [ci skip]
-
- * pry is not needed for test-bundled-gems
* Run test-unit test without rake task to avoid yard dependency
* Use power_assert-1.2.1 to avoid pry dependency - 05:36 PM Revision e031f2b4 (git): * 2023-06-15 [ci skip]
-
- * pry is not needed for test-bundled-gems
* Run test-unit test without rake task to avoid yard dependency
* Try to skip Prime_test.rb -
- add new features easily
(https://github.com/ruby/irb/pull/500)
* Add nesting level parser for multiple use (indent, prompt, termination check)
* Rewrite RubyLex using NestingParser
* Add nesting parser tests, fix some existing tests
... -
- @mame It seems that you were not aware of the fact that the documentation is wrong. And I hope you share with me the idea that we want the documentation to be correct. I think the reason why there was a mistake in the documentation, and ...
-
- mame (Yusuke Endoh) wrote in #note-2:
> "Confusing" is not necessarily a bug. What Kernel#Rational accepts is clearly stated in the documentation in BNF style.
> ...
I had read that documentation, and thought that perhaps someone is goin... -
12:57 AM Bug #19733 (Feedback): Kernel#Rational does not accept prefix 0
- "Confusing" is not necessarily a bug. What Kernel#Rational accepts is clearly stated in the documentation in BNF style.
https://docs.ruby-lang.org/en/master/Kernel.html#method-i-Rational
There may be room for discussion to make `Kernel... -
- `Integer` and `Rational` literals accept prefix `0`. There is no difference in this respect.
```ruby
0b10 # => 2
0b10r # => (2/1)
```
However, when it comes to `Kernel#Integer` and `Kernel#Rational`, the former accepts prefix `0... -
08:06 AM Feature #19315: Lazy substrings in CRuby
- Hanmac (Hans Mackowiak) wrote in #note-11:
> it confused me too, i thought Copy On Write was default for shared strings
> ...
Pat Shaughnessy in his blog describes exactly the same thing as Benoit Daloze above: Ruby shares string data ... -
07:38 AM Feature #19315: Lazy substrings in CRuby
- it confused me too, i thought Copy On Write was default for shared strings
https://patshaughnessy.net/2012/1/18/seeing-double-how-ruby-shares-string-values -
- Recently Cirrus-CI does not start over 10 hours. Maybe we have too
many runs. -
- 03:39 AM Revision 1e55ac49 (git): Bump ruby/setup-ruby from 1.151.0 to 1.152.0
- Bumps [ruby/setup-ruby](https://github.com/ruby/setup-ruby) from 1.151.0 to 1.152.0.
- [Release notes](https://github.com/ruby/setup-ruby/releases)
- [Commits](https://github.com/ruby/setup-ruby/compare/bc1dd263b68cb5626dbb55d5c89777d793...
06/14/2023
-
11:47 PM Revision 0c55ef11 (git): [ruby/rdoc] Use flat_map for better performance
- https://github.com/ruby/rdoc/commit/76192a280d
-
- Pick from https://github.com/rubygems/rubygems/commit/7a7b234721c375874b7e22b1c5b14925b943f04e
-
- https://github.com/rubygems/rubygems/commit/6d9e8025dc
-
- https://github.com/rubygems/rubygems/commit/225fdf0b2e
-
- https://github.com/rubygems/rubygems/commit/522b5f1ecd
-
10:01 PM Revision 3e7f5b02 (git): [rubygems/rubygems] Improve error message in frozen mode edge case
- When a top level dependency is missing from the lockfile, and we're in
frozen mode, we should also print a "frozen error".
https://github.com/rubygems/rubygems/commit/3e82b835e3 -
- This error message is also printed when using `bundler/setup` in frozen
model, so we're not necessarily installing any gems when it happens.
This new message play nicer with all situations.
https://github.com/rubygems/rubygems/commit/6... -
- I think it communicates better what's going on.
https://github.com/rubygems/rubygems/commit/07a25767a4 -
- https://github.com/rubygems/rubygems/commit/b86caaa3d5
-
- https://github.com/rubygems/rubygems/commit/1ff8626571
-
- https://github.com/rubygems/rubygems/commit/4180008c2d
-
- For conciseness and consistency with printing the relative path to the
lockfile just before.
https://github.com/rubygems/rubygems/commit/74cf6aea7c -
- https://github.com/rubygems/rubygems/commit/22b8caf42f
-
- https://github.com/rubygems/rubygems/commit/251e80aec6
-
- https://github.com/rubygems/rubygems/commit/ad52f840f2
-
- https://github.com/rubygems/rubygems/commit/d91c245921
-
-
08:53 PM Feature #19714: Add Refinement#refined_module
- matz (Yukihiro Matsumoto) wrote in #note-2:
> I am not 100% satisfied. Probably (unlike `class_eval` and `module_eval` whose class/module distinction is for the receiver), the method returns a class or a module, unrelated to the name.
... -
- I am not 100% satisfied. Probably (unlike `class_eval` and `module_eval` whose class/module distinction is for the receiver), the method returns a class or a module, unrelated to the name.
Is there any idea?
Matz.
-
- Ruby's event.h (https://github.com/ruby/ruby/blob/813a5f4fc46a24ca1695d23c159250b9e1080ac7/include/ruby/internal/event.h#L103) is using type aliases from stdint.h, however it's not directly included.
An example where this causes issue... -
- I found that `enc-unicode.rb` deals with some inconsistent unicode data (i.e. some data which uses short property names and some data which uses long names), so it doesn't provide much useful context. I've made a PR to create documentati...
-
02:42 PM Revision 813a5f4f (git): Directly allocate FrozenCore as an ICLASS
- It's a bad idea to overwrite the flags as the garbage collector may have
set other flags. -
-
- I don't feel this proposed extension readable (than explicit one). Is there any example of readable usage of this notation?
Matz.
-
- Currently it is an expected behavior because it is parsed as `(def show = puts 'ok') if params`. Related to #19392.
-
07:35 AM Bug #19731 (Closed): Can’t call an instance method inside an Endless method definition
- Can’t call an instance method inside an Endless method definition.
```
$ ruby -v
ruby 3.2.2 (2023-03-30 revision e51014f9c0) [arm64-darwin22]
$ cat withend.rb
class A
def show
puts 'ok' if params
end
def params
... - 05:25 AM Revision 58b8e217 (git): Update bundled gems list at bd019ac87aa47c481348ec6b7e3773 [ci skip]
-
- It's needless with racc-1.7.1
-
-
-
- Separate the directly from the customized test-unit, since it may not
work with bundled gems. -
- Bumping this because it's kinda shocking to me that strings don't already work this way. My mental model of ruby strings has always been that
```
m = rx.match(very_large_string)
before, match, after = m.pre_match, m[0], m.post_match
... -