Activity
From 08/14/2022 to 08/20/2022
08/20/2022
- 11:33 PM Revision e85db849 (git): * 2022-08-21 [ci skip]
-
11:33 PM Misc #18968 (Closed): Run MJIT under a forked Ruby process
- Applied in changeset commit:git|dc8d70e4615cdf12378322fbcd4396486270ddbe.
----------
Execute MJIT in a forked Ruby p... -
- Yeah. Ever since I created llrb.gem, I've sometimes thought about it. Like what the description of JEP 243 says, JIT ...
-
10:08 AM Misc #18968: Run MJIT under a forked Ruby process
- An interesting project if you are changing the interface could be to write a generic JIT interface, which any JIT cou...
-
- # Changes
* When MJIT compiles a method, it spawns a child Ruby process and lets it generate a C code and compile it... -
- [Misc #18968]
-
10:01 PM Feature #18959: Handle gracefully nil kwargs eg. **nil
- I have another use case.
I'm receiving options from a web protocol.
Sometimes there are no options.
So I end up wi... -
07:52 PM Feature #18969: Compare only method definitions for Method#== and UnboundMethod#==
- #18798 made me realize the proposed behavior makes sense for UnboundMethod#== but probably less for Method#== which p...
-
- @ko1 Indeed, I'll mark this one as duplicate and close it, and I'll add your issue to the dev meeting.
-
04:21 PM Feature #18969: Compare only method definitions for Method#== and UnboundMethod#==
- Same as https://bugs.ruby-lang.org/issues/18798 ?
-
- From https://bugs.ruby-lang.org/issues/18751#note-16
During the last dev meeting, @ko1 said Method#== should be "i... -
- * [Feature #18798] `UnboundMethod#==` with inherited classes (eregon)
* OK to change UnboundMethod#== to check if ... -
- I think we should do this, because I think the main purpose of UnboundMethod#== is to find out if two methods have th...
-
- Ah, maybe this cryptic header actually works for Bash too?
Then the problem is that CRuby adds it instead of RubyGem... -
- Same as https://github.com/oneclick/rubyinstaller2/issues/299, but I figured it's extremely likely to be a bug in CRu...
-
01:06 PM Bug #18580: Range#include? inconsistency for beginless String ranges
- TBH, my position is that:
1. I don't see why non-numeric `Range#include?` should be prohibited if it is possible t... -
03:32 AM Bug #18580: Range#include? inconsistency for beginless String ranges
- This was my attempt.
https://github.com/nobu/ruby/tree/nonnumeric-range-include_p -
03:21 AM Bug #18580: Range#include? inconsistency for beginless String ranges
- nobu (Nobuyoshi Nakada) wrote in #note-6:
> About the implementation, how about splitting the function by `string_us... -
- jeremyevans0 (Jeremy Evans) wrote in #note-5:
> Assuming @matz only wants this behavior change for beginless/endless... -
12:48 PM Feature #18965: Further Thread::Queue improvements
- `pop(nonblock: true)` that wouldn't raise definitely make sense to me, so I'll for sure propose it at the next meetin...
-
- Thank you for the benchmark.
Given the results, I think it's currently not worth it to add batch push/pop, because i... -
- FWIW, TruffleRuby and JRuby don't have ZSUPER methods, so it's like they always had the fix of https://github.com/rub...
-
- @nagachika wrote in #note-14:
> I agree to backport the revert to ruby_3_1, but I think removing the existing method... -
11:44 AM Bug #18435 (Closed): Calling `protected` on ancestor method changes result of `instance_methods(false)`
- Applied in changeset commit:git|8212aab81a77a2a91fb7c1681b4968171193b48f.
----------
Make Object#method and Module#i... -
- matz (Yukihiro Matsumoto) wrote in #note-24:
> Instead, each class should have lists of methods for each visibility.... -
- (I reopen because the new methods or equivalent functionality is not present on current master.)
Since https://git... -
- Eregon (Benoit Daloze) wrote in #note-16:
> @ko1 said Method#== should be "is it the same definition?".
> I agree, ... -
- Fixed in https://github.com/ruby/ruby/pull/6242
-
- Applied in changeset commit:git|8212aab81a77a2a91fb7c1681b4968171193b48f.
----------
Make Object#method and Module#i... -
- * Fixes https://bugs.ruby-lang.org/issues/18751
-
- Based on https://github.com/jeremyevans/ruby/commit/c95e7e5329140f640b6497905485761f3336d967
Among other things, thi... -
- > The main purpose of the new class is immutability, not "frozen", not "plain", not simplicity: FrozenStruct, SimpleS...
-
-
- * Rename mjit_exec to jit_exec
* Rename mjit_exec_slowpath to mjit_check_iseq
* Remove mjit_exec references from co... -
- Applied in changeset commit:git|6f3857f6a7b3cd6bd7e62e4efdbb1b841544e053.
----------
Support Encoding::Converter new... -
03:25 AM Revision fc5382d4 (git): Reuse rb_class_new_instance_kw function
-
03:24 AM Revision 01e8d393 (git): Fix gdb incompatibilies in rp_class, rb_ps_vm
- Other changes are needed to bring .gdbinit up-to-date with current ruby.
It looks like lldb is the preferred approach... -
- Applied in changeset commit:git|1a2f99275be28fb0e8ff4cfc0165966e99898d70.
----------
[Bug #18958] format string must... -
- Previously, newline: :lf was accepted but ignored. Where it
should have been used was commented out code that didn't... -
-
-
- Applied in changeset commit:git|d2483393cbcb4dcfa0000fa8166bb7fa7ed9f7b4.
----------
[Bug #18956] Negative codepoint... -
08/19/2022
-
11:22 PM Feature #18159: Integrate functionality of syntax_suggest gem into Ruby
- Thank you all for your help, this is exciting!
-
- nobu (Nobuyoshi Nakada) wrote in #note-4:
> matz (Yukihiro Matsumoto) wrote in #note-3:
> > So I decided to make `i... -
-
- Related to https://github.com/Shopify/yjit-bench/pull/109
-
- Applied in changeset commit:git|ce384ef5a95b809f248e089c1608e60753dabe45.
----------
[Bug #18955] Check length of ar... -
- It consists of multiple members, so calling it data itself doesn't seem like a problem to me. For documentation, you ...
-
- Umm wait.
`Data` is actually a [plural form](https://www.learnenglish.de/mistakes/data.html). While using it as si... -
10:19 AM Feature #16122: Data: simple immutable value object
- I am not 100% satisfied with any of the candidates, but Struct::Value and Data are better than others.
Struct::Value... -
- > At the dev meeting, @matz rejected all name candidates except `Struct::Value` and `Data`.
So, as far as I can un... -
08:30 AM Feature #16122: Data: simple immutable value object
- This is my personal opinion. I think `Data` is a good choice since there are few compatibility issues at this time de...
-
- At the dev meeting, @matz rejected all name candidates except `Struct::Value` and `Data`.
* He wants to avoid the ... - 06:57 PM Revision 745de85a (git): * 2022-08-20 [ci skip]
-
-
-
- Thank you for your report. At least we need the full trace, especially. "C level backtrace information" section. Can ...
-
03:42 PM Bug #18967 (Third Party's Issue): Segmentation fault in stackprof with Ruby 2.7.6
- Ruby 2.7.6 appears to have broken the stackprof gem, crashing on every run with a segmentation fault.
Please see t... -
- So I implemented `pop(:count)` https://github.com/ruby/ruby/pull/6257
And benchmarked it against my somewhat reali... -
- Also note that I did a single ticket for 3 somewhat related features, but if `concat` (aka batch push) is deemed unde...
-
01:18 PM Revision 4177f60e (git): Write interface instead of interfact
-
12:49 PM Feature #5825: Sweet instance var assignment in the object initializer
- I would really love to see native support for an initializer instance variable assignment shorthand without adding th...
-
11:23 AM Bug #18941: Endless loop with a memory leak in FileUtils::mkpath
- I proposed a fix at https://github.com/ruby/fileutils/pull/100.
-
- https://github.com/ruby/error_highlight/commit/defcaf1beb
-
08:51 AM Bug #18966 (Open): Strange behavior when numbered parameters and method definition are both used in a block
- ~~~ruby
'a'.tap { p _1; def f()=42 } #=> "a"
'a'.tap { p _1; def f(a)=42 } #=> nil
'a'.tap { def f()=42; p _1 } #=... -
- We still need to discuss about the implementation detail, but the baseline is what @eregon explained above.
Matz. -
- I still don't think `non` (or `itself_if` or `itself_unless`) gain readability. It may work well in some cases, but m...
-
04:20 AM Revision 17d0e5be (git): syntax_suggest moved to under the ruby organization.
-
- Not only `J` is called military timezone.
- 01:03 AM Revision 61ab06fe (git): Update default gems list at 490af8dbdb66263f29d0b4e43752fb [ci skip]
-
- ```
$ tool/sync_default_gems.rb syntax_suggest
``` -
- Adds the `syntax_suggest` syntax error display tool to Ruby through the same mechanism as `error_highlight` and `did_...
-
-
08/18/2022
-
- > That would mean basically ignoring the timeout, not great.
Maybe my writing wasn't clear. I don't suggest to ign... -
- chrisseaton (Chris Seaton) wrote in #note-3:
> I was going to comment that adding or removing multiple items from a ... -
- > Then there's the question of how SizedQueue would behave if it's not full, but still doesn't have space for all the...
-
- I was going to comment that adding or removing multiple items from a queue is likely not great for implementation, as...
-
- @p8, indeed. Thanks for the heads up.
-
05:06 PM Feature #18965: Further Thread::Queue improvements
- Hi Jean,
Should that second code example use `items` instead of `item`?
``` ruby
queue.concat(items)
```
-
- Following the recent addition of a `timeout` parameter to `Queue#pop`, there are a handful of other improvements I'd ...
-
- > With shapes you could start with {a,b} and then add the hash keys to get shape {a,b,c,d} and pass all values [1,2,3...
-
06:49 PM Feature #18776: Object Shapes
- tenderlovemaking (Aaron Patterson) wrote in #note-21:
> It's currently implemented this way. :)
Ok, thank you for th... -
- Reference keyword arguments - what we're doing is using the same idea of shapes, but applying them to the keyword arg...
-
06:06 PM Feature #18776: Object Shapes
- Dan0042 (Daniel DeLorme) wrote in #note-19:
> Thank you for this important work. In particular I think shapes will b... -
- Dan0042 (Daniel DeLorme) wrote in #note-19:
> In particular I think shapes will be very useful in the future to impr... -
- Thank you for this important work. In particular I think shapes will be very useful in the future to improve the perf...
-
04:50 AM Feature #18776: Object Shapes
- About the general principle, since Ruby is used on many production environment by many companies, non-optional featur...
-
-
-
-
- I can live with `DataClass` too, but I still can't forget the beautiful shortness of `Data`. `DataClass.new` feels li...
-
05:05 PM Feature #16122: Data: simple immutable value object
- If “data” is the naming direction folks like, I think the class name should be DataClass. This aligns with kotlin (wh...
-
- > I'm not sure if @zverok likes Data over his own proposal, but note that data appears in his local variable name as...
-
06:59 AM Feature #16122: Data: simple immutable value object
- k0kubun (Takashi Kokubun) wrote in #note-37:
> My enthusiastic +1 for `Data`.
>
> I've used Kotlin and its [Data... -
- My enthusiastic +1 for `Data`.
I've used Kotlin and its [Data classes](https://kotlinlang.org/docs/data-classes.htm... -
05:25 PM Feature #18910 (Closed): Improve maintainability of LLDB helpers
- Applied in changeset commit:git|f1ccfa0c2c200c9443fbfc3f1ac3acbdd3e35559.
----------
[ci-skip][Feature #18910][lldb]... - 05:25 PM Revision 59da2678 (git): * remove trailing spaces. [ci skip]
-
- by moving it fully into RbBaseCommand
-
- Push the newly refactored lldb files into a sub-directory so that we're
not cluttering up the misc directory -
-
-
- `lldb_cruby.py` manages lldb custom commands using functions. The file
is a large list of Python functions, and an in... -
- 03:17 PM Revision 8c44b07f (git): * 2022-08-19 [ci skip]
-
-
-
-
- PR: https://github.com/ruby/ruby/pull/6251
-
- To explain my mental model:
I think Method/UnboundMethod captures a given method, and that should never change (even... -
- @matz agreed to remove ZSUPER methods, because the current illusion is incomplete.
In practice the semantics are the ... -
- Thanks @schneems and @matz
I'll merge this and invite @schneems to our team as a Ruby committer tomorrow.
-
- The new name `syntax_suggest` looks good.
Matz.
-
- @naruse, @nobu, and @matz agreed that it should raise an exception if the format string is not ASCII-compatible.
-
- I agree with raising exception when the format string is not ASCII compatible.
Matz.
-
- @naruse, @nobu, and @matz agreed that a negative integer to `%c` should raise an explicit exception.
-
- At the dev-meeting, @akr proposed that the format `%c` behaves like `%s` (with the one-codepoint restriction) and @ma...
-
- Copying from the dev meeting log:
* We will try to change the encoding lookup for stings so that the fast path get... -
- We could not finish all the agenda in time. We will continue on 25th.
-
- * [Feature #18159] Integrate functionality of syntax_suggest gem into Ruby
* Is it ok for `syntax_suggest`? -
- We discussed this at the dev meeting.
@ko1 said Method#== should be "is it the same definition?".
I agree, I'll tak... -
- @matz agreed to remove ZSUPER methods, because the current illusion is incomplete.
In practice the semantics are the ... -
- matz (Yukihiro Matsumoto) wrote in #note-3:
> So I decided to make `include?` to raise exception for beginless/endle... -
- I once thought about removing `each`+`succ` semantics from `include?` altogether for simplicity, but it is too big in...
-
- If we are expanding the string or only stripping extra capacity
then coderange won't change, so clearing it is wasteful. -
- Applied in changeset commit:git|fe61cad7490da8a879597f851f4a89856d44838e.
----------
Implement SizedQueue#push(timeo... -
- It looks good to me.
Matz.
-
- [Feature #18944]
If both `non_block=true` and `timeout:` are supplied, ArgumentError
is raised. -
- What about `ObjectSpace`?
-
- An API to notify "end of boot" seems useful beyond just fork COW optimizations, as you say.
For instance a JIT might... -
- Thank you Matz.
> RubyVM is not globally available (e.g., not for JRuby or TruffleRuby).
Yes, what was on purp... -
- I am OK with adding this feature, but I have some concerns with the place and the name.
`RubyVM` is not globally ava... - 07:04 AM Revision b3718ede (git): Update bundled gems list at 2022-08-18
-
- I reject this proposal to make the feature built-in for several reasons:
* I still think it should be done in libr... -
- Applied in changeset commit:git|725626d8905fe1ac4a2cf1c3e2db6412bf8f381f.
----------
[Bug #18964] Update the code ra... -
- These functions do not update the code range of the destination string, and the string still caches outdated flags.
... -
-
- icc now seems to provide libutil.so that is not related to pty.
This extconf.rb wrongly finds it and adds `-lutil`, b... -
- Check out the revisions for testing as "detached" from the beginning.
08/17/2022
-
08:32 PM Feature #18776: Object Shapes
- It's unfortunate that there are spills there and there might be ways to reduce that by reorganizing the code a bit, b...
-
07:51 PM Feature #18776: Object Shapes
- After running the YJIT benchmarks in release mode, we found that setting instance variables is, indeed, slower on our...
-
- > I think it would be a good idea to open a draft pull request so that it's easier to look at the diff and comment on...
-
- > These are our proposed code changes to implement Object Shapes in CRuby.
I think it would be a good idea to open... -
- # Object Shapes Update
We are writing with an update on the Object Shapes implementation, and to ask what needs to b... - 05:26 PM Revision bfefaf47 (git): * 2022-08-18 [ci skip]
-
-
- jeremyevans0 (Jeremy Evans) wrote in #note-2:
> This type of optimization is not safe unless the entire class hierar... -
- Dan0042 (Daniel DeLorme) wrote in #note-1:
> I'm also excited by the possibility this could allow to memoize methods... -
- The limitations of `@var ||=` memoization are brought up from time to time (e.g. #17316, #6023) so I think that's an ...
-
02:54 PM Revision b4daf6e2 (git): Fix flaky test for GC thrashing
- GC could be in an intermediate state after creating the objects, so we
should finish GC by running a minor GC. -
- ujihisa (Tatsuhiro Ujihisa) wrote in #note-19:
> `itself_if`
> `itself_unless`
zverok (Victor Shepelev) wrote in... -
05:36 AM Feature #17330: Object#non
- I had another idea regarding a #non method that I haven't voiced before, but may result in cleaner code overall:
`... -
- TBH, I believe that core method names should be short, recognizable and elegant in context in the first place, not ex...
-
04:46 AM Feature #17330: Object#non
- ujihisa (Tatsuhiro Ujihisa) wrote in #note-19:
[H]ow's the following 2-words names instead?
>
> `itself_if`
> `i... -
03:23 AM Feature #17330: Object#non
- > non() and oui()
Since what these 2 methods do aren't trivial from the names, how's the following 2-words names ins... -
-
- @sam.saffron it's my fault for forgetting to update the commit message. `CGI.url_encode` was never implemented, what ...
-
08:06 AM Feature #18822: Ruby lack a proper method to percent-encode strings for URIs (RFC 3986)
- @byroot,
I am not sure the name is right here:
`CGI.path_encode`
with an alias of
`CGI.params_encode`
is far ... -
-
-
08/16/2022
-
- mame (Yusuke Endoh) wrote in #note-15:
> ```
> params[:name]&.non(&:empty?)
> ```
>
> Frankly speaking, this above i... -
- I understand the logic, but the result of your generalization, `&.non(&:empty?)`, is hard to read anyway. @matz says ...
-
- @mame One thing that is really frequent in Rails codebases is code like this:
```ruby
params.presence || DEFAUT_PAR... -
- ```
params[:name]&.non(&:empty?)
```
Frankly speaking, this above is dead hard to read and write for me. Do you ... -
- I also think 2) is the best
while reiterating that we can and should keep the behavior of `("a".."z").step(num)` for... -
- Thanks, @mame, for a summary.
I'd like to emphasize, though, that the "undesirability" of undesirable behavior loo... -
- IMHO 2) is the best. I think trying to prevent the "undesirable behavior" is too artificial.
Also that behavior seem... -
- Let me summarize the problem and possible solutions.
Desirable behavior:
```
(timestamp1...timestamp2).step(3.... -
- Set `date` member to `RUBY_RELEASE_DATE` instead of the date at the
build time, to make installed files reproducible. -
- - Split into `Ext` and `Lib` classes.
- `Ext#files` should not include built extension libraries.
- `Ext#files` shoul... -
- mame (Yusuke Endoh) wrote in #note-35:
> BTW, I personally wanted Struct to store the field values simply in instance... -
- BTW, I personally wanted Struct to store the field values simply in instance variables rather than the hidden storage...
-
- @nobu proposed `Data`, which used to be a class for extension library authors, but deprecated since ruby 2.5 and remo...
-
- Applied in changeset commit:git|3850113e20b8c031529fc79de7202f61604425dd.
----------
[ruby/cgi] Implement `CGI.url_e... -
- Maybe, but that would need approval. If you feel strongly about it please open a dedicated issue.
-
- Should we also add aliases for `escape_html` and so on?
-
- Here we go: https://github.com/ruby/cgi/pull/27
-
- Ah right, I forgot to add the alias that were agreed upon. I'll open a followup: `GCI.escape_uri_component` and `GCI....
-
- I'm positive on this feature, but I'm negative on the naming convention being introduced. Are we going to add aliases?
-
- It's mostly just waiting for review. I'm not certain who's the maintainer though. @hsbt @nobu I added you both for re...
-
- Since we just finished working around a nightmare scenario here @byroot, I think it is rather instructive to see a re...
- 03:45 PM Revision 5528648a (git): * 2022-08-17 [ci skip]
-
-
- [Feature #18822]
Ruby is somewhat missing an RFC 3986 compliant escape method.
https://github.com/ruby/cgi/commit/c... -
- A codepoint is expected for `%c`, then the former examples are currently expected behaviors, I think.
The latter... -
08/15/2022
-
11:14 PM Revision 0608a9a0 (git): Optimize Marshal dump/load for large (> 31-bit) FIXNUM (#6229)
- * Optimize Marshal dump of large fixnum
Marshal's FIXNUM type only supports 31-bit fixnums, so on 64-bit
platforms t... -
10:41 PM Revision e49db0f7 (git): Do not clone method entries when bind_call is used
- I noticed that this site unconditionally clones the method entry, which
means that `bind_call` always allocates a `T_... -
- https://github.com/ruby/rdoc/commit/488f89aee4
- 05:05 PM Revision e34720dc (git): * 2022-08-16 [ci skip]
-
- * Simplify around `USE_YJIT` macro
- Use `USE_YJIT` macro only instead of `YJIT_BUILD`.
- An intermediate macro `YJI... -
- https://github.com/ruby/ruby/pull/6242 now makes resolved-through-zsuper methods equal for compatibility.
That means ... -
- Here is a simplified script which does not depend on Class.new == Class#new:
```ruby
class C
class << self
alia... -
- This test will prevent performance regressions like [Bug #18929].
-
01:16 PM Revision 8d40ede2 (git): [ruby/date] [DOC] Enhanced intro for Date (https://github.com/ruby/date/pull/72)
- https://github.com/ruby/date/commit/59a6673221
-
- PR to fix this by no longer hiding ZSUPER methods: https://github.com/ruby/ruby/pull/6242
-
- Also as @ufuk says in https://bugs.ruby-lang.org/issues/18435#note-2, owner and instance_methods should be consistent...
-
- This is still an issue.
We need to try to no longer hide ZSUPER methods and estimate if there is any incompatibility... -
- Reopening this as because of the revert it is still inconsistent and unsolved.
PR to fix this by no longer hiding ... -
- My understanding of this issue is it's the same as #18729.
The bug is not in `instance_methods` which is correct.
I... -
- protected methods like in this issue are part of the method table, so I think it makes no sense to hide them.
What w... -
- * [Bug #18729] Method#owner and UnboundMethod#owner are incorrect after using Module#public/protected/private (eregon...
-
- https://github.com/rubygems/rubygems/commit/0d321c9e3a
- 05:12 AM Revision cbcf2dad (git): * 2022-08-15 [ci skip]
-
- Applied in changeset commit:git|89aa09afaf77920fd748aefaba99fe4b0f19e684.
----------
[ruby/rinda] Handle situations ... -
- Fixes [Bug #13864]
https://github.com/ruby/rinda/commit/3cd620f38c
08/14/2022
Also available in: Atom