Activity
From 05/10/2020 to 05/16/2020
05/16/2020
-
06:21 PM Feature #16891: Restore Positional Argument to Keyword Conversion
- Dan0042 (Daniel DeLorme) wrote in #note-1:
> IMHO there's one consideration that's more important than all others. The following code _must_ work in ruby 2.8/3.0
> ...
You are correct that behavior I proposed doesn't support this. How... -
08:37 AM Revision 2c3c6c96 (git): Defer initialization
- Defer initialization of extension libraries, loading prelude files
and requiring files, and skip if dump options are given. -
- Source code encoding is defaulted to UTF-8 now too.
-
- When multiple autoclose IO objects are leaked too.
-
08:28 AM Feature #16781: alias :fold :reduce
- shan (Shannon Skipper) wrote in #note-2:
> I was surprised to hear this, since I think of "reduce" as most popular.
https://en.wikipedia.org/wiki/Fold_(higher-order_function)
Article 'Reduce_(higher-order_function)' - just redirect ... -
08:20 AM Misc #16775: DevelopersMeeting20200514Japan
- We've hold a dev-meeting at 14th, but could go through only half topics. We'll hold an extra meeting at 26th.
Log: https://github.com/ruby/dev-meeting-log/blob/master/DevelopersMeeting20200514Japan.md -
06:30 AM Feature #16894: Integer division for Ruby 3
- The following is meant to be an objective assessment of the proposal, taking no stand on approval, or not.
The proposal is for inclusion of the "feature" in **Ruby 3** . By definition, Ruby 3 will be a major version, which will incorp... -
- Just FYI, I've made these patches for division recently.
https://github.com/ruby/ruby/compare/master...nobu:feature/fract-slash?expand=1
https://github.com/ruby/ruby/compare/master...nobu:feature/division?expand=1
-
01:43 AM Feature #16894: Integer division for Ruby 3
- fwiw, array indexing code likely wouldn't need changed.
``` ruby
a = ["zero", "one", "two"]
a[1/2] # "zero"
a[1/2.0] # "zero"
```
Some code would break, but we have major versions to account for this. It seems like a smaller ... -
01:35 AM Feature #16897: General purpose memoizer in Ruby 3 with Ruby 2 performance
- Yes I can get this to work with hacks like this:
```ruby
def memoize_26(method_name)
cache = {}
uncached = "#{method_name}_without_cache"
alias_method uncached, method_name
m = (define_method(method_name) do |*arguments|
f... -
- You should be able to have similar performance by using `ruby2_keywords` and keeping the method definition the same. Have you tried that?
Note that with such an approach, you would not be able to differentiate between providing keywo... -
- ```ruby
require 'benchmark/ips'
module Memoizer
def memoize_26(method_name)
cache = {}
uncached = "#{method_name}_without_cache"
alias_method uncached, method_name
define_method(method_name) do |*arguments|
foun...
05/15/2020
-
11:51 PM Misc #16895: Request for cooperation: Try your applications/libraries with master branch and debug options
- Make first paragraph read a bit better.
-
- Updated from latest edits.
-
02:36 PM Misc #16895: Request for cooperation: Try your applications/libraries with master branch and debug options
- Samuel and Eregon rewrote a description. Thank you!
-
- ## Summary
If you maintain a Ruby application or library, please consider testing with the Ruby `master` branch, including the *debug* build. This should be in addition to testing with all supported stable releases of Ruby.
To make... -
10:08 PM Feature #16894: Integer division for Ruby 3
- IMHO this is completely unrealistic for compatibility, and very clearly not worth breaking all the code (even more so for floats which lose precision).
-
05:39 PM Feature #16894: Integer division for Ruby 3
- > It'd be great if division in Ruby matched what we all learned in school.
In my time it was C++ or Java at school and in those language (like Ruby) if the two number are Integer, then the result is an Integer. 1 / 2 in those language... -
- I was just pointing out that other languages recognized this issue and only fairly recently chose to make syntax changes to explicitly deal with it.
I think using ``//`` is probably the ``least worst`` choice because it's already being ... -
- Here's a good read on the thoughts and motivation behind Python changing it: https://www.python.org/dev/peps/pep-0238/
-
09:22 AM Feature #16894: Integer division for Ruby 3
- jzakiya (Jabari Zakiya) wrote in #note-7:
> Ruby already has ``a.div b`` for explicit integer division.
Thanks, I had forgotten about that. But while the remainder can be achieved by the infix operator `%`, it would be strange if there ... -
- jzakiya (Jabari Zakiya) wrote in #note-7:
> Ruby already has ``a.div b`` for explicit integer division.
Also `a.quo b` for rational division, and `a.fdiv b` for float division. -
- Ruby already has ``a.div b`` for explicit integer division.
Nim also uses ``a div b`` for integer division.
Crystal v0.31 switched to ``a // b``, so now ``9 // 2 = 4`` and ``9 / 2 = 4.5``.
Python, et al, also use ``//``.
-
- Thanks for the responses!
Re 0.5 vs 1/2r: In my experience, developers use floats way more than rationals.
Re "is it worth the incompatibility?": I like to imagine it's not confusing new developers 20 years for now. If it's a chang... -
- What would you do when you want integer division? Have you thought of it? ~~`(a / b).round`~~ `(a / b).floor`is not a solution. That is like saying, "write `"a"[0, 0]` when you want an empty string."
-
03:43 AM Feature #16894: Integer division for Ruby 3
- > 1) Why 0.5 and not 1/2r
I'm for 1/2r. -
03:42 AM Feature #16894: Integer division for Ruby 3
- I would bet that a majority of Rubyists would agree that `1/2 == 0` is an unfortunate choice. Either `0.5` or `1/2r` would probably be a better solution.
Your proposal has at least two issues.
1) Why `0.5` and not `1/2r`
0) Much mor... -
- Sorry, for the edits, first issue! Changed normal to floating point.
-
- Hi Ruby team,
It'd be great if division in Ruby matched what we all learned in school.
``` ruby
1 / 2 == 0.5
```
New developers wouldn't immediately be confused when they try to do division and experienced developers could sto... -
09:14 PM Revision cc525d76 (git): [ci skip] Enhanced rdoc for String.new (#3067)
- * Per @nobu review
* Enhanced rdoc for String.new
* Respond to review -
- * Per @nobu review
* Rdoc enhancements for Array
* Responses to review - 09:12 PM Revision a3cd0152 (git): * 2020-05-16 [ci skip]
-
- * Per @nobu review
* [CI skip] Enhance rdoc intro for Hash
* Tweak call-seq for Hash.new
* Tweak call-seq for Hash.new
* Minor corrections
* Respond to review
* Respond to review
* Respond to review
* Respond to review
* Fix cha... -
- Dan0042 (Daniel DeLorme) wrote in #note-6:
> `exec` only works as a tail-call at the end of the script. Most of the time it's not applicable:
> ...
``` ruby
exec 'cmd1 && cmd2 && cmd3'
```
-
- Right now they are public, and since mkmf.rb does `include MakeMakefile` it defines a lot of public methods on all objects.
```
$ ruby -e 'a=1.public_methods; require "mkmf"; b=1.public_methods; puts (b-a).sort'
append_cflags
appen... -
03:06 PM Feature #16786: Light-weight scheduler for improved concurrency.
- Hi, I am a Ruby user that would probably vote for
``` ruby
Fiber do
end
```
only if I see the poll back then.
The result should be taken as a grain of salt.
I second @ko1 on the point that the people who voted for the choice j... -
03:10 AM Feature #16786: Light-weight scheduler for improved concurrency.
- @ioquatix I was well represented by Martin-sensei (@duerst).
The fiber created by `Fiber() do ...end` does context-switch on I/O operations. The traditional (or original) fibers don't.
So naming the function `Fiber` may indicate that *... -
- @sylvain.joyeux
Hi! I read [your comment in the discuss RubyOnRails](https://discuss.rubyonrails.org/t/new-2-7-3-0-keyword-argument-pain-point/74980/13). I'm sorry for bothering you about this change. Let me explain a bit.
Thi... -
- First of all, I'm really sorry for bothering many people about this change. And I'd like to really thank you @jeremyevans0 for the great work to this issue. Honestly, I'm sorry to see this change reverted because I still believe this c...
-
- Dan0042 (Daniel DeLorme) wrote in #note-1:
> Also I believe there's a fifth aspect that deserves consideration for reverting:
> ...
Totally agree. I probably missed that conversation, but I am very curious as to what the use-cases were... -
- Dan0042 (Daniel DeLorme) wrote in #note-1:
> ## Non-Symbol Keyword Argument
> ...
The core of that problem actually has nothing to do with allowing non-symbol keyword argument. That developer ad hocly used string vs. symbol distinction i... -
01:23 AM Feature #16891: Restore Positional Argument to Keyword Conversion
- IMHO there's one consideration that's more important than all others. The following code _must_ work in ruby 2.8/3.0
```ruby
def foo(h={}, **kw)
[h, kw]
end
foo({x:1}) #=> [{x:1}, {}]
foo(x:1) #=> [{}, {x:1}]
```
This is ... -
08:45 AM Revision a9b2014d (git): Fix a typo [ci skip]
-
- Applied in changeset commit:git|71c166e11e761dcaaa943f9b94da38f86ada1bdb.
----------
Fixed argument forwarding in reserved word method [Bug #16854] -
- It results in a syntax error, right?
https://github.com/ruby/ruby/pull/3112 -
-
- TestAutoload#test_source_location can't run multiple test-run so
that use assert_separately().
repro command:
make yes-test-all TESTS='--repeat-count=50 ruby/test_autoload -n test_source_location' -
- I've been doing some [Rosetta Code](http://rosettacode.org/wiki/Rosetta_Code) tasks recently, and thought it would be a good idea to have a page/link to a list of all the Ruby done tasks on the website. These are good teaching aides, esp...
-
01:26 AM
Bug #16892 (Closed): Reconsider the test directory name for scheduler
- Applied in changeset commit:git|6fa8455ebbf457e5d8752295a8d6380146636c0c.
----------
Move `test/scheduler` -> `test/fiber` [Bug #16892][ruby-core:98366]. -
- I am happy to move it to `test/fiber`.
-
12:14 AM Bug #16892 (Closed): Reconsider the test directory name for scheduler
- `test/scheduler` is broke the convention of test directory. We use only `class` or `module` name excepts `ruby` and `-ext-` for test/foo directory.
Should we use `test/fiber` or other?
- 01:26 AM Revision 6fa8455e (git): Move `test/scheduler` -> `test/fiber` [Bug #16892][ruby-core:98366].
-
- You should comment it to https://discuss.rubyonrails.org/t/new-2-7-3-0-keyword-argument-pain-point/
We are going to propose the solution like [this](https://bugs.ruby-lang.org/issues/16891) after hearing pain points.
05/14/2020
-
09:00 PM Feature #16851: Ruby hashing algorithm could be improved using Tabulation Hashing
- @byroot
> I'd suggest running the hash related benchmarks included in ruby's repo: https://github.com/ruby/ruby/tree/master/benchmark
For most of the benchmarks, this changes doesn't make much of a different because they are either too... -
- Based on feedback from Rails-core, Matz has decided to postpone full separation of keyword and positional arguments (see https://discuss.rubyonrails.org/t/new-2-7-3-0-keyword-argument-pain-point/74980). I think most Ruby-core developers...
-
07:28 PM Misc #16890: [Ruby Keywords and Ruby 3.0 release] Feedback to matz and the ruby core team
- The discussion at https://discuss.rubyonrails.org/t/new-2-7-3-0-keyword-argument-pain-point/74980/2
is already quite long, and I do not intend to add to the discussion there, largely because I think
it may be difficult for matz and th... -
- As some folks may already have read/heard, matz is asking for feedback.
He specifically asked this in regards to the rails devs, and the rails ecosystem
but this may be relevant to all ruby users.
You can read his ideas here:
h... -
04:23 PM Bug #16787: [patch] allow Dir.home to work for non-login procs when $HOME not set
- Attaching version "v7-rebased-2020-05-14" of the patch. This version corresponds to the **rebase-only** changes pushed to my [ads/b.r-l.o-issue-16787](https://github.com/salewski/ruby/tree/ads/b.r-l.o-issue-16787) branch for [PR 3034](ht...
-
- Add `--suppress-backtrace=num` option to limit the backtrace length
- 04:22 PM Revision 531e4a35 (git): * 2020-05-15 [ci skip]
-
- https://rubyci.s3.amazonaws.com/debian/ruby-master/log/20200514T123004Z.log.html.gz
```
/home/chkbuild/chkbuild/tmp/build/20200514T123004Z/ruby/test/scheduler/scheduler.rb:29: warning: assigned but unused variable - fiber
/home/chkbuild... -
- An even better example, `dir.rb` could be used in TruffleRuby & other implementations easily:
https://github.com/ruby/ruby/blob/d7d0d01401a8082e514eb2cb3cec5410e7acba7d/dir.rb#L14-L25
There is already quite a bit of logic there.
All... -
- As an example, https://github.com/ruby/ruby/blob/d7d0d01401a8082e514eb2cb3cec5410e7acba7d/trace_point.rb could be used as-is in TruffleRuby if using the `Primitive.name` syntax.
It might not be a lot, but it's a start and it's already n... -
- Some examples of `Primitive.name` used in TruffleRuby:
https://github.com/oracle/truffleruby/blob/af97b03a55b757688a62455cd56672c53cd56d0d/lib/truffle/weakref.rb#L36
https://github.com/oracle/truffleruby/blob/af97b03a55b757688a62455cd56... -
- Maybe `enable(&block)` should behave like `enable(target: block); disable`?
-
- ```ruby
threads = []
inspects = []
trace = TracePoint.new(:line) do |tp|
threads << Thread.current
inspects << tp.inspect
end
done = false
thread = Thread.new do
Thread.pass until done
end
trace.enable do
line_eve... -
- IMHO it's not nice to replace a nice Ruby-level API by a C API one.
Is there a benchmark of how much is it slower with `rb_funcallv_public`?
Also, will it still work when passing a custom object understanding `rand`?
If so, there ar... -
- Sounds OK. Go ahead.
Matz.
-
- ko1 (Koichi Sasada) wrote in #note-12:
> Please respect current convention `rb_`/`RB_` are called in ruby interpreter (such as `rb_str_new()`) and `ruby_`/`RUBY_` can be called from outside of ruby interpreter (such as `ruby_xmalloc()`)... -
-
-
- Are you happy adding the following to `LEGAL`:
```
coroutine::
Where specified, these files are licensed under the {MIT License}[rdoc-label:label-MIT+License].
``` -
- @mame I have merged the light weight concurrency patch, and it included some changes to these tests to make them less flaky, by putting it in separate test file. In my experience it seems much more reliable now. Just FYI.
-
- Using latest master:
```ruby
class Scheduler
def fiber(&block)
fiber = Fiber.new(blocking: false, &block)
fiber.resume
return fiber
end
end
Thread.current.scheduler = Scheduler.new
f1 = Fiber d... -
09:25 AM Feature #16786: Light-weight scheduler for improved concurrency.
- ioquatix (Samuel Williams) wrote in #note-34:
> Thanks Matz.
> ...
I can't speak for Matz, but my guess is that he meant "not the original type of fiber", i.e. not the same as you'd get e.g. with `Fiber.new`. -
- Thanks Matz.
> since the fiber created by the method is not the original fiber at all.
Can you clarify "not the original fiber at all"? It's the same way `Integer(...)` creates instance of `class Integer`. -
- Accepted for experimentation.
We still have some concerns, for example, mixture with blocking and non-blocking fibers. @mame will describe the concern.
In addition, I don't like the method name `Fiber`, since the fiber created by the... -
- > I don't think we should refer this kind of result because the voted people does not know concerns.
I think you are underestimating the collective knowledge of the community. That poll had almost 300 responses. I've also been working o... - 10:10 AM Revision 0e3b0fcd (git): Thread scheduler for light weight concurrency.
-
09:32 AM Feature #16822 (Rejected): Array slicing: nils and edge cases
-
- I don't think the benefit of changing outweighs the pain of incompatibility. Rejected.
Matz.
-
- Julia uses `~/.julia/compiled/vX.Y` directory to store precompiled cache files.
-
- I think python changed that default mostly because it can be annoying to see all .py files give rise to .pyc
files in the very same working directory.
(I know way too little to comment on the content of the issue at hand here, so thi... -
- Sounds OK. Let's see how it works.
Matz.
-
- LGTM (including `--backtrace-limit`).
Matz.
-
- `cache_dir` is an optinal argument but not a keyword argument, and
the return value is a `Pathname`. -
-
- Accepted.
Matz.
-
- We still wait for the PoC of this issue.
-
04:29 AM Feature #15921: R-assign (rightward-assignment) operator
- hi.
I have summarized the expected behavior and the actual behavior with right assignment.
see: https://gist.github.com/osyo-manga/ef1db68fcb62a6fce7dace0f655c0b17
I think there is another issue of priority, apart from the fact th... -
03:27 AM Revision 4a620aff (git): Restore class variable setting for tests
-
-
-
-
-
- please write a specification of your proposal. there is only a code, test and motivation.
-
-
05/13/2020
-
- ko1 (Koichi Sasada) wrote in #note-10:
> * (1) how to manage the code?
> ...
Yes, that's fine, at least for now.
Maybe later we can share such code in a shared repository, but one step at a time, and probably not needed (other Rubies ... -
08:18 PM Bug #16853: calling bla(hash, **kw) with a string-based hash passes the strings into **kw (worked < 2.7)
- > This change was deliberate and mentioned in the 2.7.0 release announcement:
I missed this specific implication of that change ... Sorry for the noise. Thanks. -
- This change was deliberate and mentioned in the 2.7.0 release announcement:
```
Non-symbols are allowed as keyword argument keys if the method accepts arbitrary keywords. [Feature #14183]
def foo(**kw); p kw; end; foo("str" => 1) ... -
- Hmm. I am confused about the example though.
Isn't
bla "some" => "string"
actually
bla( { "some" => "string" } )
? -
- The following code
~~~
def bla(hash = {}, **kw)
puts "H: #{hash}"
puts "K: #{kw}"
end
bla "some" => "string"
~~~
**silently** outputs the following (no warnings about deprecation of keyword parameters-from-hash)
... -
- Eregon (Benoit Daloze) wrote in #note-3:
> Maybe another angle for this would be to make the `instance variable @foobar not initialized` shown even with the default `$VERBOSE` being `false` (i.e. `rb_warn` instead of `rb_warning`).
> .... -
- Maybe another angle for this would be to make the `instance variable @foobar not initialized` shown even with the default `$VERBOSE` being `false` (i.e. `rb_warn` instead of `rb_warning`).
Then people would likely be a lot more eager to... -
06:58 PM Feature #16855 (Rejected): Add a tracepoint for warnings
- Eregon (Benoit Daloze) wrote in #note-1:
> I'm afraid I don't see the point of a TracePoint for this if warnings can already be hooked via `Warning.prepend SomeModuleWithWarn`.
> ...
Yep, I agree. I'm closing this :)
-
- I'm afraid I don't see the point of a TracePoint for this if warnings can already be hooked via `Warning.prepend SomeModuleWithWarn`.
Invoking the TracePoint regardless of $VERBOSE would be a large performance cost, and would prevent ... -
- I would like to add a tracepoint for warnings. I want to do this so that DidYouMean can suggest fixes for instance variables. I noticed did you mean [has experimental support](https://github.com/ruby/did_you_mean/blob/master/lib/did_yo...
-
- Really looking forward to this API, it's _very_ promising.
What exactly are the implications of enter_blocking_region/exit_blocking_region? Does it mean the scheduler should not resume fibers even if IO is ready? Since the scheduler is ... -
- note that I missed:
> If any mutex is acquired by a fiber, then a scheduler is not called; the same behaviour as blocking Fiber.
in description, so I agree there is no issue if Mutex is used correctly.
checking such situation (io ops ... -
- ioquatix (Samuel Williams) wrote in #note-26:
> > non-blocking fiber creation API
> ...
I don't think we should refer this kind of result because the voted people does not know concerns.
> > Scheduler class
> ...
It can be, Maybe we sho... -
- Add clarification about introducing new hooks.
-
- Specify the root fiber is also blocking.
-
- > non-blocking fiber creation API
It was voted by community, strongly in favour of `Fiber do ... end`. If you think your suggestion is better, we should confirm with community.
> ...
We can introduce `Fiber::Scheduler` however I don't ... -
- I believe this was fixed in #16801, which is already marked as required for backport to 2.7. With the master branch, there is no warning for the code.
-
06:11 PM Bug #16856 (Closed): Structs accepting keyword arguments issue a warning with Ruby 2.7
- This code is issuing a warning with Ruby 2.7, but should not:
``` ruby
class NotFine < Struct.new(:x)
def initialize(x, arg:)
end
end
class Fine
def initialize(x, arg:)
end
end
NotFine.new(1, arg: 1) # This is cau... -
- Interesting find and explanation.
-
04:42 PM Bug #16854 (Closed): Using (...) when the method name is a keyword generates error
- Although I wouldn't generally recommend it, naming a method `true` is allowed. However, in combination with the new argument forwarding keyword `...`, a syntax error is generated which does not occur with other argument choices.
This ... -
- Thanks for the insight into bypass vs deoptimization.
Of course if you override #hash and #eql? to return true for previously false cases then it wouldn't work, but anyway I don't see much use or reason for doing this on Integer or St... -
- Dan0042 (Daniel DeLorme) wrote in #note-4:
> Would it be feasible to raise some sort of error when trying to redefine Integer#hash ?
> ...
It's probably feasible, but I don't think it is a good idea. You can do it in pure Ruby if you ... -
01:14 AM Bug #16850 (Closed): Object#hash doesn't behave as documented
- Applied in changeset commit:git|de29a022acb93691dfc50db852cb04f763565072.
----------
Document that #hash is not called for certain core classes [ci skip]
Fixes [Bug #16850] -
- Would it be feasible to raise some sort of error when trying to redefine Integer#hash ?
Not just for this case, but I feel there were other cases where ruby bypasses the method call to use the built-in implementation directly. Can't rec... -
- +1 The updates LGTM.
Matz.
- 03:51 PM Revision 65c5a395 (git): * 2020-05-14 [ci skip]
-
- Ragel generates a code `0 <= (*p)` where `*p` is char.
As char is unsigned by default on arm and RISC-V, it is warned by gcc:
```
compiling parser.c
parser.c: In function ‘JSON_parse_string’:
parser.c:1566:2: warning: comparison is alwa... -
- > * A method that seems to return a new array that is directly related to the receiver, should return an instance of the receiver's class.
> ...
So this is the old thinking?
> I used to think methods should honor subclasses, but I ch... -
- That optimization is nice to have, but I think the point of this ticket is that it's currently not possible to have an arbitrary string separator.
```ruby
str = "aaabababbbabbabaabaaabbbabab"
sep = "x" #or anything except " "
str.g... -
08:49 AM Feature #16847: Cache instruction sequences by default
- Also I forgot to mention. Python 3 no longer store this cache alongside source files, but in a subdirectory.
Python 2 use to store `path/foo.py` cache as `path/foo.pyc`, Python 3 now stores it as `path/__pycache__/foo.pyc`. -
- > How cache is effective and how large cache storage is needed depend an application.
Of course there is some variance, but from my testing the improvement is relatively constant. The cache is approximately 1.5 times as large as the s... -
- How cache is effective and how large cache storage is needed depend an application.
But Ruby don't have enough example of such applications other than Rails. -
06:47 AM Revision 87662134 (git): [ruby/openssl] Ruby/OpenSSL 2.2.0
- https://github.com/ruby/openssl/commit/41587f69e1
-
- Let's revert the changes for now, as it cannot be included in the 2.2.0
release.
My comment on #257:
> A blocker is OpenSSL::SSL::SSLContext#add_certificate_chain_file. It
> ...
This effectively reverts the following commits:
- dacd0... -
- Implement OpenSSL::PKey::PKey#oid as a wrapper around EVP_PKEY_id().
This allows user code to check the type of a PKey object.
EVP_PKEY can have a pkey type for which we do not provide a dedicated
subclass. In other words, an EVP_PKEY t... -
06:47 AM Revision a7145c3d (git): [ruby/openssl] Fix signing example to not use Digest instance
- https://github.com/ruby/openssl/commit/033fb4fbe4
-
- https://github.com/ruby/openssl/commit/b08ae7e73d
-
- This is not present in the referenced files anymore, and not useful to most users
https://github.com/ruby/openssl/commit/eae30d2b96 -
- https://github.com/ruby/openssl/commit/b28fb2f05c
-
- https://github.com/ruby/openssl/commit/c4374ff041
-
05:50 AM Revision fcd25762 (git): Stop always inlining not-optimized get/setivar
- As we have the optimization in _mjit_compile_ivar.erb, we don't use
these functions if we successfully optimize ivars. Therefore it's
consuming code size where we can't optimize it well.
To decrease code size, we'd better avoid inlining... -
- `#define NDEBUG` produces "macro redefined" warnings when it is already
defined via cppflags -
- It produces "unused variable" warnings in NDEBUG mode
-
- Sure.
-
- ver is [3, 1, 0] which is not less then or equal to [3, 1]
-
- Interesting!! @ana06 Do you need any help?
Matz.
-
- Fixes [Bug #16850]
05/12/2020
-
- Maybe due to e1855100e46040e73630b378974c17764e0cccee, CentOS, RHEL, and
Fedora CIs have started failing with SEGV. Try to avoid
ffi_closure_alloc on those environments.
https://rubyci.org/logs/rubyci.s3.amazonaws.com/centos8/ruby-mast... -
-
-
-
-
-
-
-
-
-
-
-
-
- I think it seems difficult to implement it.
If an interpreter manages everything, it is easy (at least I can image how to implement it).
(1) API
I'm not sure we can implement Mutex scheduling with the hooks introduced at #10.
If there ... -
- Sorry for late response.
First of all, I agree to merge it and try before next release (please wait Matz's comment for merging).
There are several considerations.
# non-blocking fiber creation API
For me, the name `Fiber()` is ... -
08:52 PM Bug #16852 (Closed): Refining Enumerable fails with ruby 2.7
- When using [rspec](https://github.com/rspec/rspec-core) and ruby 2.7 I am unable to refine `Enumerable`. I have created an [issue in rspec](https://github.com/rspec/rspec-core/issues/2727) but I'm wondering if there's an underlying ruby ...
-
- Please respect current convention `rb_`/`RB_` are called in ruby interpreter (such as `rb_str_new()`) and `ruby_`/`RUBY_` can be called from outside of ruby interpreter (such as `ruby_xmalloc()`). `RUBY3_...` are already renamed, but I w...
-
- > This kind of technique can cause something strange behavior
Can you tell me what kind of behavior you are thinking of? I don't think that was ever a problem with bootsnap (it has a few corner cases, but none that I can think of for ... -
- isn't it enough to use libraries such as bootsnap for it?
This kind of technique can cause something strange behavior and users may know what they are doing. Using a library is good opt-in method, IMO. -
- > I mean, that can also be achieved when you compile your ruby script into an shared object ahead-of-time.
Sure, but then it's no longer "by-default", as in all users get it for free without having to configure anything nor changing ... -
08:39 AM Feature #16847: Cache instruction sequences by default
- byroot (Jean Boussier) wrote in #note-4:
> > isn't that also true for instruction sequence caches?
> ...
I mean, that can also be achieved when you compile your ruby script into an shared object ahead-of-time. Basically there must be... -
- > isn't that also true for instruction sequence caches?
No, caching instruction sequence can be done without any functional change for users.
> ...
That depend of the storage mechanism. But assuming you have 1 cache file for each ... -
- byroot (Jean Boussier) wrote in #note-2:
> @shyouhei that would be a radical change in how Ruby is deployed.
Agreed, however isn't that also true for instruction sequence caches?
> ...
This point is valid, and makes me wonder how ... -
- @shyouhei that would be a radical change in how Ruby is deployed.
Also boot performance is particularly important during development, where AOT doesn't make sense.
-
- Why not compile into native binary then? I have recently learnt that Emacs is moving from its .elc bitecodes to .eln native shared object https://arxiv.org/abs/2004.02504
Because we already have JIT it seems now possible for us to AOT. -
- > How to implement the gem_path_resolver?
By simply walking down the directories in the loadpath.
In the context of Bundler/Rubygems it could directly be done post install, in the case of Bootsnap it's done on the fly on cache mis... -
- Just curious:
How to implement the `gem_path_resolver`? `gem_path_resolver` should know the set of existing files in the gem install directories.
What happens if a new file is created in a gem directory after installation? I understand... -
- @mame yes, your example fits my proposal.
But note that I'm also open to other solutions, as long as it allows to resolve paths of the `$LOAD_PATH`. -
- As far as my understanding, this is a self-contained simple example to demonstrate the proposal (to make dev-meeting discussion easy)
```ruby
gem_path_resolver = -> feature do
case feature
when "bar" then "/gems/bar/bar.rb"
... -
- @nobu indeed. I initially only focused on solving the performance problem of having a large `$LOAD_PATH`, but maybe it's the occasion to also allow for alternative storage mechanisms.
In such case it can be interesting to look at how ... -
- byroot (Jean Boussier) wrote:
> ### Proposal
> ...
This seems similar to my proposal in several years ago, to allow loading from archived libraries.
It was not callable but calling a particular method, I don't remember accurately now ... -
- > But it doesn't seem right for other NameError
@eregon Absolutely, but my ticket is only about `uninitialized constant`. If my description or my PR made you think otherwise, it's a misunderstanding and / or a bug in my PR. -
- Right, I agree it makes sense for missing constants NameError's, since the user expect a "constant path" including the missing constant.
But it doesn't seem right for other NameError such as:
```
$ ruby -e foobar
Traceback (most re... -
- +1 for using #name and falling back to #inspect
Rails is correct in extending inspect to return more useful human-readable information, and a NameError should tell you about the problem name (including namespace), not about extra info... -
- > (also #name can be nil, and calling #name on arbitrary objects could easily be worse than #inspect).
So the issue title is no longer relevant. Since then I dug more into this issue, and I think the behavior should be similar to `rb_... -
- Nice work on that PR :)
I think whoever overrides `#inspect` should make sure it's easy to identify which object it is.
In other words, I don't think we should special-case here for Module & Class (also #name can be `nil`, and callin... -
- https://github.com/ruby/ruby/pull/3090 is simply a weird limitation I discovered while studying that behavior.
I still think we should try to display the actual class path. -
- https://github.com/ruby/ruby/pull/3090 was merged. Can we close this ticket? (using `#name` instead of `#inspect`).
-
- This seems interesting. I'd suggest running the hash related benchmarks included in ruby's repo: https://github.com/ruby/ruby/tree/master/benchmark
-
- I have implemented Linear Probing and Simple tabulation in Ruby: https://github.com/Ana06/ruby/compare/master...Ana06:tabulation
I tested it using the following code:
https://github.com/Ana06/ruby-tabulation/blob/master/benchmark_tab... -
- That looks good to me :+1: I may add String as well (such as Integer and String).
-
- I consider this an implementation detail. It never makes sense to override the hash calculation for the cases where Ruby uses the built-in calculation instead of calling #hash, so I don't think Ruby should change behavior to increase con...
-
- From [Ruby 2.7 Object class documentation](https://ruby-doc.org/core-2.7.0/Object.html#method-i-hash):
>The hash value is used along with eql? by the Hash class to determine if two objects reference the same hash key.
From this I exp... - 04:24 PM Revision 68224651 (git): * 2020-05-13 [ci skip]
-
- to define HAVE_FFI_CLOSURE_ALLOC.
The macro is used in closure.c, so have_func check is needed.
If pkg-config is not installed, extconf.rb fails to detect the version
of libffi, and does not add "-DUSE_FFI_CLOSURE_ALLOC=1" even when sys... -
-
- As this rule does not use a separate timestamp file, the target
will be rebuilt every time once it is outdated. -
- | |compare-ruby|built-ruby|
|:--------------|-----------:|---------:|
|re_chars-1 | 169.230k| 973.855k|
| | -| 5.75x|
|re_chars-10 | 25.536k| 107.598k|
| | ... -
- Optimized `String#split` with `/ /` (single space regexp) as
simple string splitting. [ruby-core:98272]
| |compare-ruby|built-ruby|
|:--------------|-----------:|---------:|
|re_space-1 | 432.786k| 1.539M|
| ... -
- It has the same issue as RHEL 8. k0kubun said he will fix later
-
-
- znz (Kazuhiro NISHIYAMA) wrote in #note-8:
> How about optimization `split(/ /)` (when regexp is single space only) instead of changing `split(" ")`?
Sounds nice. https://github.com/ruby/ruby/pull/3103
-
- How about optimization `split(/ /)` (when regexp is single space only) instead of changing `split(" ")`?
-
- My guess is that, perhaps, even now, it is very rare to use a single space string argument with the expectation to match single or multiple spaces. In such use cases, normally, `split` is used without an argument, or with a regex argumen...
-
- naruse (Yui NARUSE) wrote in #note-12:
> NDEBUG is not acceptable.
NDEBUG is not my invention. Please file a bug report to upstream (ISO/IEC JTC1/SC22/WG14).
I'm not against defining it by default, though.
-
- This reverts commit 946dadd3f479198e87873a863d15c7660a8e2b56,
which broke `TestGemRequire` and others. -
-
- Allow to create symbolic links (if available) to share same or
updated files. -
- Reduce duplicate parts such as package name and version numbers.
-
- `mainsrc` is WIP name for migrating this issue. nobu said we can rename `src` from `mainsrc` after that.
-
-
03:56 AM
Feature #9758 (Closed): Allow setting SSLContext#extra_chain_cert in Net::HTTP
- Applied in changeset commit:git|31af0dafba6d3769d2a39617c0dddedb97883712.
----------
Expose SSLContext#extra_chain_cert in Net::HTTP
Currently, Net::HTTP can only send a single SSL certificate when it
establishes a connection. Some use... - 03:55 AM Revision 31af0daf (git): Expose SSLContext#extra_chain_cert in Net::HTTP
- Currently, Net::HTTP can only send a single SSL certificate when it
establishes a connection. Some use-cases involve sending an entire
certificate chain to the destination; for this, SSLContext supports
assigning to #extra_chain_cert=.
... -
03:25 AM Revision 946dadd3 (git): Sync did_you_mean
-
-
- +1
-
- Argument of RUBY_ASSERT_FAIL is already stringized message, so no
more extra stringization should be applied. -
- Functions using `rb_thread_t` and `rb_execution_context_t` are
internal use only.
05/11/2020
-
- Also alternatively the callable could be responsible for actually loading the code rather than return an absolute path.
This would allow for more flexibility in implementing alternative code loaders, e.g. read the source from an archi... -
- Right.
I always thought RubyGems should be smarter and when requiring `foo/bar.rb` it could look for a gem named `foo` first, and not look in all gems if the file exist.
This feature could help achieve some of that, instead of going ... -
- @eregon only if you did actually require it. My explanation is probably a bit confusing.
```ruby
$LOAD_PATH = [
"stdlib/"
]
Bootsnap.setup
# Bootsnap see `delegate.rb` in the stdlib, so cache as `{"delegate.rb" => "/stdlib/delegate.rb... -
- There is a now a spec for that as we replicated that caching in TruffleRuby:
https://github.com/oracle/truffleruby/commit/30024ed8c8f0900cf03bdc7fdf3fa7b4776837ac -
- > The other, more important difficulty, is that you also have to invalidate the cache whenever you add or delete a file in one of the $LOAD_PATH members, otherwise if you shadow or unshadow another file that is farther in the $LOAD_PATH,...
-
- I just realized I made a mistake in my benchmark, I also both instruction sequence and load path caching enabled.
With instruction sequence disabled, the speedup is only 17% for Redmine:
```
$ RAILS_ENV=production time bin/rails r... -
- Make it easier to implement `$LOAD_PATH` caching, and speed up application boot time.
I benchmarked it on Redmine's master using bootsnap with only the optimization enabled:
```ruby
if ENV['CACHE_LOAD_PATH']
require 'bootsnap'
... -
- Just returns column 1 for ambiguous width because this I/O is not tty and can't
seek. -
-
- It would be a relatively frequent case. It is still slower than
`float * float` because `*` has a dedicated VM instruction (opt_mult),
though. -
- I want Ruby 2.8/3.0 is faster than 2.7 by default.
NDEBUG is not acceptable.
I think Microsoft's _DEBUG approach is more reasonable. - 04:15 PM Revision 95ac2355 (git): * 2020-05-12 [ci skip]
-
- This limit was introduced on Nov 20 1996
in 554b989ba1623b9f6a0b76f00824c83a23fbcbc1
Apparently to protect from a buffer overflow:
* eval.c (f_missing): オブジェクトの文字列表現が長すぎる時バッファ
を書き潰していた
However I tested that path with very ... -
- The error is easy to reproduce:
e.g. on Ruby 2.3:
```
$ ruby -robjspace -e 'ObjectSpace.trace_object_allocations_stop'
-e:1:in `trace_object_allocations_stop': wrong argument type false (expected tracepoint) (TypeError)
from -e... -
- * [Feature #16847] Cache instruction sequences by default (byroot)
* This make code loading about 30% faster.
* All it's needed is to agree on where and how to store the cache.
* [Feature #16848] Allow callables in `$LOAD_PATH` (b... -
- * [Misc #16803] Discussion: those internal macros reside in public API headers (ko1)
* confirm the discussion.
-
- Instruction sequence caching is available since Ruby 2.3, and on recent rubies it speeds up code loading by about 30%.
I just benchmarked it on Redmine's master, using bootsnap with only that optimization enabled:
```ruby
if ENV['... -
- Function pointers are not void*. See also
115fec062ccf7c6d72c8d5f64b7a5d84c9fb2dd8
ce4ea956d24eab5089a143bba38126f2b11b55b6
8427fca49bd85205f5a8766292dd893f003c0e48 -
- SunPro's #pragma does_not_return(...) needs an argument. That does not
fit the attribute syntax we employ. -
- This header file is simply out of date (for decades since at least
1989). It's the 21st century. Just stop using it. -
- https://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html says:
> GCC never returns 1 when you call the inline function with a string
> ...
Because rb_str_new and familiy take string constants, it was a bad idea
for them to contain RB_CONS... -
- To fix build failures.
-
- Devs do not love "impl".
-
-
-
- Devs do not love "3". The only exception is RUBY3_KEYWORDS in parse.y,
which seems unrelated to our interests. -
- This shall fix compile errors.
-
- Devs do not love "3".
05/10/2020
-
- The Ruby interpreter of metasploit framework is embedded package. So, you can ask your issue to the tracker of metasploit at first.
https://github.com/rapid7/metasploit-framework -
- How did you install ruby there? One-click installer? If so then I am not sure
why you attempt to install rvm. I don't think that is necessary; back when I
used ruby on windows sometimes, the one-click installer worked.
Either way, this ... -
- Applied in changeset commit:git|c89c3801b985916b6fb6726aab966d28371dfaaa.
----------
BASERUBY have to be 1.9 or later at least [Bug #16845]
Many tools under tool directory haven't worked with ruby 1.8. -
- OK, since e1f62d7f0e3, if `HAVE_BASERUBY=yes` then `file2lastrev.rb` must succeed.
-
03:58 PM Bug #16845: Building Ruby with system Ruby 1.8.7 results in make failing due to syntax error in ./tool/lib/vcs.rb
- Maybe I am missing something, but I get the same result with `-p`. Log is attached.
The `tar` manpage suggests `-p` is about preserving permissions, not timestamps.
And it looks like the timestamps are preserved even if I don't u... -
- The timestamps in tarball are aligned to avoid to remake files which need BASERUBY.
But extracting **without** `-p` option disables it.
```
04:48 PM [eswan@rndssh1(RnD):/tmp/eswan]$ tar -xJvf ruby-2.7.1.tar.xz
```
-
- It sounds to me that the root cause of this problem is released tarball having BASERUBY dependency.
-
- I have updated the title and description to correct where the syntax error is located. It is located in `./tool/lib/vcs.rb`, which is ***called*** by `./tool/file2lastrev.rb`.
From looking at the related issue that Shibata-san linked, i... -
- Many tools under tool directory haven't worked with ruby 1.8.
-
-
-
-
03:22 PM Bug #16846 (Closed): Commit - win32ole: separate global variable declarations and definitions - backport?
- The below commit fixes building win32ole using MinGW gcc 10.1.0. New Ruby releases (patch/teeny) are normally done with a current MSYS2 installation, so backporting this would be helpful for 2.5, 2.6, & 2.7. Not sure about the status o...
- 03:19 PM Revision 27efe3f7 (git): * 2020-05-11 [ci skip]
-
- GNU make does not export it by default.
-
02:17 PM Feature #16786: Light-weight scheduler for improved concurrency.
- I recently did a deep dive into this approach and how it would fit into the Ruby ecosystem as part of my work on TruffleRuby.
I think what is being proposed here looks like a very practical idea for improving concurrency on Ruby, for ... -
- https://gcc.gnu.org/gcc-10/changes.html#c
> * GCC now defaults to `-fno-common`. As a result, global
> ...
> silently merged during linking. -
-
- `RubyVM.each_builtin` is not defined when cross compiling.
-
- not only when !w but also when w == WAITPID_LOCK_ONLY.
See also: f7c0cc36920a4ed14a3ab1ca6cfdf18ceff1e5d5 and a2264342063260d660b99872eaf5080f6ab08e81.
We thought this change was an oversight in the latter commit.
Without this change, t... -
- Fix MSVC build error.
-
- To generate what is necessary via generic_erb.rb instead.
-
- To generate what is necessary via generic_erb.rb instead.
-
- with MJIT worker enabled
The problem:
```
$ ruby -ropen3 --jit -e 'Open3.capture2e("cmake") rescue nil;binding.irb'
irb(main)[01:0]> Process.waitall
=> [[10656, #<Process::Status: pid 10656 exit 127>]]
$ ruby -ropen3 -e 'Open3.capture... - 05:19 AM Revision 18f22490 (git): * 2020-05-10 [ci skip]
-
- just for a case. In addition, this change suppresses unused variable
warning. -
01:53 AM Bug #16831: Running `Pathname#glob` with `File::FNM_DOTMATCH` option loses `.` and `..`
- Thank you for your explanation. With your hint and reading [the implementation code](https://github.com/ruby/ruby/blob/761528e8aa7c54ec92c90335fe26a584b992918b/ext/pathname/pathname.c#L1131-L1161) solved my question.
Dan0042 (Daniel D... -
- Just as an add-on:
Go has similar effects to those reported here for Ruby, see e.g. https://stackoverflow.com/questions/33640943/golang-float-arithmetic.
@FedorKK, if you can find out *why* e.g. `fmt.Println(0.29 * 100 == 29.0)` prints... -
-
- Additional information: Unicode is still recovering from a VM crash last month. Mirroring of the Unicode Character Database on github might eventually happen, but not direct hosting.