Activity
From 09/16/2020 to 09/22/2020
09/22/2020
-
11:08 PM Bug #16455: coroutine ucontext uses deprecated POSIX getcontext/swapcontext/makecontext, absent in musl and uclibc
- ioquatix (Samuel Williams) wrote in #note-7:
> It is fixed: https://github.com/ruby/ruby/pull/2995
Hello, you have provided good fix but it is related to armv7 only.
For example: aarch64-gentoo-linux-musl and aarch64_be-gentoo-linux... -
10:44 PM Bug #11778 (Closed): recv_io with mode breaks when klass arg is kind of BasicSocket
- Applied in changeset commit:git|ed27c2514c2f4140546dce2e8f5f64cb91712c47.
----------
Update UnixSocket#recv_io tests to handle receiving a UnixSocket
Receiving UnixSocket works fine if you don't provide a mode, and
I think it is reason... -
09:11 PM Bug #11778: recv_io with mode breaks when klass arg is kind of BasicSocket
- I don't think this is a bug. If you pass a non-nil mode argument to `recv_io`, you should ensure that `klass.for_fd` accepts the mode argument. The example works correctly if you don't pass a mode argument to `recv_io`. I have submitt...
-
- Receiving UnixSocket works fine if you don't provide a mode, and
I think it is reasonable to expect that you should not provide
a mode if klass.for_fd would not accept a mode.
Fixes [Bug #11778] -
09:32 PM Revision b904b729 (git): Enhanced RDoc for String (#3565)
- Makes some methods doc compliant with https://github.com/ruby/ruby/blob/master/doc/method_documentation.rdoc. Also, other minor revisions to make more consistent.
Methods:
try_convert
+string
-string
concat
<<
pr... -
08:40 PM Revision 3d474e19 (git): Rudimentary support for disassembling rb_iseq_t
- I need to disassemble instruction sequences while debugging, so I wrote
this.
Usage is like this:
```
(lldb) p iseq
(rb_iseq_t *) $147 = 0x0000000101068400
(lldb) rbdisasm iseq
0000 putspecialobject( 3 )
0002 putnil
0003 defineclass( I... -
-
- The OpenStruct changes were reverted starting around commit:e026e186f4a01aa3f6cd02ae6ef33f44f129361c, so you once again get `/` output for the program.
-
-
- To restore compatibility with Ruby 2.7 and fix #17182, I've reverted commit:eeef16e190cdabc2ba474622720f8e3df7bac43b at commit:179384a66862d5ef7413b6f4850b97d0becf4ec9.
@shugo, I would appreciate your input here on how refinements sho... -
- Applied in changeset commit:git|179384a66862d5ef7413b6f4850b97d0becf4ec9.
----------
Revert "Prevent SystemStackError when calling super in module with activated refinement"
This reverts commit eeef16e190cdabc2ba474622720f8e3df7bac43b.... -
- After some more thought, I think it's better to revert my fix and reopen #17007. Here's a PR for that, I plan on merging it after it passes CI: https://github.com/ruby/ruby/pull/3564.
-
- This is discussed in https://bugs.ruby-lang.org/issues/17007#note-4. I'm not sure whether the previous behavior is spec or just an implementation detail. We could revert commit:eeef16e190cdabc2ba474622720f8e3df7bac43b, but then we end ...
-
03:47 PM Bug #17182 (Closed): Refinements behavior is broken
- Calling the methods defined by Refinements via `super` from a singleton method breaks the behavior.
```ruby
class X
def hoge
["X#hoge"]
end
end
using Module.new {
refine X do
def hoge
["refine #hoge"] ... -
- This reverts commit eeef16e190cdabc2ba474622720f8e3df7bac43b.
This also reverts the spec change.
Preventing the SystemStackError would be nice, but there is valid
code that the fix breaks, and it is probably more common than cases
that... -
- I can't test cygwin, but I've added a pull request to logger that should fix this by treating cygwin like mswin and mingw in regards to log shifting: https://github.com/ruby/logger/pull/50
-
-
- 04:01 PM Revision 3c808898 (git): * 2020-09-23 [ci skip]
-
- If an object has a finalizer flag set on it, prevent it from moving.
This partially reverts commit 1a9dd31910699c7cd69f2a84c94af20eacd5875c. -
12:45 PM Bug #17175: Ruby 2.5: OpenSSL related test failures
- > As a information, openssl 2.1.2 is used on Ruby 2.5.8.
openssl 2.1.2 is used for both Ruby 2.5.8 and Ruby 2.7.1.
I see the SSL test failures in Ruby 2.5.8, but I do not see the failures in Ruby 2.6.6, Ruby 2.7.1.
Maybe we can back... -
07:40 AM Feature #17135: Improve performance of Integer#size method
- I try to improve performance of `Integer#size` in C code with refer to `TrueClass#to_s`
```c
static VALUE rb_cInteger_fix_size;
static VALUE
int_size (VALUE num)
{
if (FIXNUM_P (num)) {
return rb_cInteger_fix_size;
... -
- Applied in changeset commit:git|df14c758fc705c49c2aaf4c9276a8f7229438fbf.
----------
Make hash returned by Hash#transform_values not have a default
This sets an explicit default of nil. There is probably a better
approach of removing ... -
- PR: https://github.com/ruby/ruby/pull/3563
-
- This sets an explicit default of nil. There is probably a better
approach of removing the default.
Fixes [Bug #17181]
09/21/2020
-
- jeremyevans0 (Jeremy Evans) wrote in #note-1:
> Thanks for the report. I can confirm the issue. It looks like `rb_hash_transform_values` is missing `COPY_DEFAULT(result, hash);`. I'll update the tests and submit a PR to fix this toni... -
- Thanks for the report. I can confirm the issue. It looks like `rb_hash_transform_values` is missing `COPY_DEFAULT(result, hash);`. I'll update the tests and submit a PR to fix this tonight.
-
04:57 PM Bug #17181 (Closed): Hash with default block breaks after transforming values
- I found a problem with Hashes that use a default proc.
```ruby
x = Hash.new { |h, k| h[k] = [] }
x[:a] << 1
x[:b] << 2
```
I expected that transforming the values as follows:
```ruby
y = x.transform_values { |arr| arr.map(&... - 04:28 PM Revision 7ee166ed (git): * 2020-09-22 [ci skip]
-
- Methods:
::new
#length
#bytesize
#empty?
#+
#*
#% -
- As my experiment, I tried to run the tests on the Ruby 2.5.8 by regular user after changing the crypto policy by root on Fedora 34, and all the above tests passed. There are only new 2 failures possibly coming from the weak (= legacy) cr...
-
- Hi Greg L,
Thanks for taking a look at this!
> Sorry for the delay. I used to get pinged here, but it stopped a while ago.
It's okay. I will keep in mind to add your email as To or CC in this kind of situation for next time.
> ... -
01:55 AM Bug #17175: Ruby 2.5: OpenSSL related test failures
- I suspect that the ruby/ruby commit mentioned above will help, also see the commits @rhenium did listed in https://github.com/ruby/openssl/issues/341
-
02:22 PM Revision 82998918 (git): Make Thread#join always convert its argument, as before 70f08f1eed
-
- * This shows block() with a timeout is similar to #kernel_sleep and also
does not need to change `@blocking`. -
- * And not some list of sort.
-
-
01:50 PM Feature #17145: Ractor-aware `Object#deep_freeze`
- ko1 (Koichi Sasada) wrote in #note-8:
> For Ractor, it is very clear. Another candidate is `to_shareable`
Methods `to_...` should be reserved for conversion methods, i.e. methods that may return a different object than the receiver. ... -
- What's `Fiber#annotate`?
-
12:22 AM Feature #16786: Light-weight scheduler for improved concurrency.
- With the help of @eregon, I have tidied up the implementation so that:
- We found and fixed a few bugs.
- Changing the current scheduler invokes `#close` on the current scheduler (before did nothing).
- At thread & process exit, the cur... - 05:04 AM Revision 10d795cf (git): Add NEWS about `Fiber#backtrace` and `Fiber#backtrace_locations`.
- 03:28 AM Revision f7aa51b2 (git): Update NEWS & documentation relating to scheduler.
09/20/2020
- 11:48 PM Revision 70f08f1e (git): Make `Thread#join` non-blocking.
- 09:51 PM Revision 59617315 (git): * 2020-09-21 [ci skip]
- 09:51 PM Revision 501fff14 (git): When setting current thread scheduler to nil, invoke `#close`.
-
- Maybe `obj.shareable`? And that would then return as-is if already shareable, and make it shareable if not (deep freezing until it reaches already-shareable objects)
I don't like anything with "ractor" in the name, that becomes not de... -
- For the Struct example above, probably it needs the "Matz" String to be frozen too, or `#deep_freeze` (#17145) to mark it as a shareable value.
-
02:13 PM Revision b6d599d7 (git): Update heap_pages_himem after freeing pages
- 02:10 PM Revision be2efb11 (git): Fulfill missing tests and stabilize tests
- 02:10 PM Revision 398da711 (git): Validate name during initialization
- 02:10 PM Revision d5fa6615 (git): Add status to Ractor#inspect
- 11:39 AM Revision 84c4c7be (git): * 2020-09-20 [ci skip]
-
- * Previously this could lead to an invalid waiter entry and then trying
to wake up that waiter would result in various issues in rb_mutex_unlock_th(). -
11:23 AM Revision c10326d9 (git): bump patchlevel to 128
-
- This reverts commit a6b50f379d4b71acac73ac6b22cbe2d2c5b50b3f.
-
07:19 AM Bug #17179: Unexpected warning during value assignment to setter
- jeremyevans0 (Jeremy Evans) wrote in #note-3:
> This is expected behavior in Ruby 2.7, due to keyword argument separation. The warning is valid because the code will cause an ArgumentError in Ruby 3.
> ...
Thank you, Jeremy! -
- @jaruga
Sorry for the delay. I used to get pinged here, but it stopped a while ago. Anyway, I'm not sure what the 'policy' is, as this may be a backport of a default gem, which hasn't been common.
I'll have a look at this tomorro...
09/19/2020
-
11:19 PM Feature #17177: Include the current file name and the line number in the output of `p`
- What about the method name `inform`?
-
- This is expected behavior in Ruby 2.7, due to keyword argument separation. The warning is valid because the code will cause an ArgumentError in Ruby 3.
You should not use keywords parameters if you want to assign a hash. Instead, yo... -
- ```ruby
some_instance = Class.new { def some_setter=(a:, b:); end }.new
setter_params = { a: 1, b: 2 }
some_instance.some_setter = setter_params # => {:a=>1, :b=>2}
# warning: Using the last argument as keyword parameters is deprec... -
09:33 PM Misc #17180: Ractor and constant referencing
- Another example:
```ruby
Person = Struct.new(:name)
CREATOR = Person.new("Matz").freeze
parser = Ractor.new { CREATOR }
Ractor.select(parser)
```
```
./test.rb:12:in `block in <main>': can not access non-sharable objects in constant O... -
- Hey there.
From a high level, this bug report is describing my experience building something simple with Ractor without any shared state. I hope this is helpful as a feedback for the stable Ractor implementation in Ruby 3.0.
I've b... -
02:34 PM Feature #17047: Support parameters for MAIL FROM and RCPT TO
- I am new to Ruby development. Please let me know if I need to do anything else to get this patch accepted.
-
08:47 AM Revision 73a626c0 (git): [ruby/io-wait] update required_ruby_version
- As d387029f39d976565c955377117103499d47ff09 made io/wait
unbuildable with older versions, drop older versions tentatively.
Also the change seems to instroduce behavioral incompatibilities,
may address them or bump up the version later. -
-
- This is a temporary commit to try to find a GC issue. It seems like
mjit is pointing at a moved address in the call cache. I want to assert
that they aren't TMOVED or garbage objects at the time they get copied
09/18/2020
-
- If a new method is to be created for this purpose, then one possibility is to combine this with the idea that was once proposed by a Ruby committer to define a public method on `Object` so that we can insert the method in or at the end o...
-
07:16 PM Feature #17177: Include the current file name and the line number in the output of `p`
- https://github.com/AndyObtiva/puts_debuggerer
-
02:24 PM Feature #17177 (Rejected): Include the current file name and the line number in the output of `p`
- Rejected. I agree this behavior is convienient in some cases, but changing `p` is too drastic and incompatible.
Please retry with another name.
Matz.
-
- In many debugging situations, we want to inspect some objects in more than one source location throughout the code. To be identified where it is called from, it is very common that the method `p` is used together with a `puts` method tha...
-
- * [Bug #17178] Procs with kw:def/**kw lose elements when called with a single Array (eregon)
* Does it sound buggy to you too? Do you think we can experiment and try changing it for Ruby 3? Current behavior seems useless, inconsistent... -
08:49 AM Misc #17138: DevelopersMeeting20200925Japan
- * [Feature #13381] Expose rb_fstring and its family to C extensions
* Still an extremely useful feature long awaited, it would be very disappointing if it didn't make it to 3.0.
* Both `json` and `msgpack-ruby` agreed to a feature th... -
- jeremyevans0 (Jeremy Evans) wrote in #note-3:
> I'm not sure the loss of backwards compatibility is worth it, and this will break backwards compatibility.
It would only break code that explicitly loses arguments that way, which I wou... -
- Eregon (Benoit Daloze) wrote in #note-2:
> Is there a reason that keywords in procs cause autosplatting?
> ...
Previously, the autosplatted argument could be used as keywords if they were a hash or implicitly convertible to a hash, tho... -
- Is there a reason that keywords in procs cause autosplatting?
It seems counter-intuitive.
I would expect autosplatting depends on declared positional parameters only.
I'd argue it's a bug, because we lose part of the arguments for s... -
- Eregon (Benoit Daloze) wrote:
> ```
> ...
Are we sure this is a bug, and not just the result of a different behavior of `proc{|a|}` `proc{|a,|}`, with keywords being treated like the latter (autosplatting) as a comma is used?
Keywor... -
- ```
$ ruby -ve '{a: 1}.each_pair { |a, kw: :def| p [a,kw] }'
ruby 3.0.0dev (2020-09-18T09:12:58Z master e1cca08a6d) [x86_64-linux]
[:a, :def] # bug, should be [[1, 2], :def]
$ ruby -ve '{a: 1}.each_pair { |a, **kw| p [a,kw] }'
ru... -
08:13 PM Revision e193dd1e (git): [ruby/irb] Version 1.2.7
- https://github.com/ruby/irb/commit/0eaa06838b
-
- https://github.com/ruby/irb/commit/4c89b0775b
- 07:32 PM Revision 7d9b4d3c (git): * 2020-09-19 [ci skip]
-
- When finalizers run (in `rb_objspace_call_finalizer`) the table is
copied to a linked list that is not managed by the GC. If compaction
runs, the references in the linked list can go bad.
Finalizer table shouldn't be used frequently, s... -
09:47 AM Revision 6dd257ed (git): Added sync task for bigdecimal
-
-
-
- * When there is a scheduler, the Fiber that would be blocked has already
been rescheduled and there is no point to interrupt something else.
That blocked Fiber will be rescheduled as the next call to the scheduler
(e.g., IO, sleep,... -
08:35 AM Feature #16746: Endless method definition
- Why do we even need a ``` def ```. ? Why not just do it like Haskell?
``` ruby
triple(x) = 3*x
```
Since this is an assignment if triple hasn't already been assigned it should create a function otherwise it should pose a syntax error.... -
07:55 AM Feature #17176: GC.auto_compact / GC.auto_compact=(flag)
- Also do you have any performance results, for example memory consumption or GC time and so on?
-
- Another idea is `GC.enable(compact: true/false)`.
BTW as I mentioned in slack, GC.compact may have an issue (can access to T_MOVED objects from Ruby level) so please fix it before merge this large commit.
(and I need to learn about... -
- https://github.com/ruby/rdoc/commit/7cb5c3611f
- 05:57 AM Revision 0d56aec1 (git): [ruby/rdoc] Fix spelling error in parser comment
- https://github.com/ruby/rdoc/commit/f237c9e223
-
- Cross references need parse all files which define the subject
names. This commit makes `--force-update` option enforce to parse
all files if any file is newer than the previous parse, not only
updated files.
https://github.com/ruby/rd... -
- https://github.com/ruby/rdoc/commit/41db49c485
-
- ruby_global_symbols can be accessed with multiple ractors so that
the accesses should be synchronized. -
-
-
-
- Ractors can access this table simultaneously so we need to sync
accesses.
09/17/2020
- 03:30 PM Revision 609e6ac0 (git): * 2020-09-18 [ci skip]
-
- * Mutex operations no longer disable the Fiber scheduler.
-
- * To still remove the lock from the Thread's list of acquired locks.
* Also to not wake up other waiters and preserve blocking behavior. -
- * It would hit "[BUG] unexpected THREAD_STOPPED" before.
-
-
- * scheduler.unblock was already already called before but with no corresponding scheduler.block
* add test that Queue#pop makes the scheduler wait until it gets an element. -
- * Move #kernel_sleep next to #block as it is similar
-
-
-
11:34 AM Revision e4dbb91f (git): Fix typos [ci skip]
- s/overriden/overridden/
-
-
-
- Hello @MSP-Greg
Do you have any idea to fix it on ruby_2_5 branch?
Maybe that is same with what you did for puma last month on this commit.
https://github.com/puma/puma/commit/aac47086f1a6e38baf69159a2bc7e64b0495f080
-
-
- I agree too that it would be much cleaner, I keep seeing newer Ruby developers expecting it to work that way. That and `def self.something`.
> Maybe we can make an experiment and gather feedback from gems testing against ruby-head?
...
09/16/2020
-
- Eregon (Benoit Daloze) wrote in #note-1:
> API-wise, `GC.autocompact = true/false` as suggested by @ioquatix on the PR, or `GC.auto_compact = true/false`, sounds nicer to me than 2 separate methods.
I mainly went with `GC.(enable|disabl... -
- API-wise, `GC.autocompact = true/false` as suggested by @ioquatix on the PR, or `GC.auto_compact = true/false`, sounds nicer to me than 2 separate methods.
-
- Hi,
I'd like to make compaction automatic eventually. As a first step, I would like to introduce two functions:
* GC.enable_autocompact
* GC.disable_autocompact
One function enables auto compaction, the other one disables it. ... -
- mame (Yusuke Endoh) wrote in #note-3:
> I did not introduce `private_constant` with no receiver because I wanted to avoid a new module state (as I recall, matz now dislikes the module state), but I agree that it is redundant.
Interes... - 08:52 PM Revision 2e6c4840 (git): * 2020-09-17 [ci skip]
- 08:52 PM Revision 8e173d8b (git): Warn on a finalizer that captures the object to be finalized
- Also improve specs and documentation for finalizers and more clearly
recommend a safe code pattern to use them. -
- I am facing OpenSSL test failures when building the Ruby 2.5 latest version 2.5.8 + some patches with OpenSSL 1.1.1g on Fedora 34 and 33.
AS this issue does not happen for Ruby 2.6, Ruby 2.7, and does not happen on Ruby 2.5.8 on Fedora ... - 12:31 PM Revision f75009c1 (git): Prefer to use `prep_io` for temporary IO instances.
-
-
-
-
-
10:27 AM Bug #17172 (Closed): Passing a complex number to Complex.polar occurs assertion failed when NDEBUG is undefined - Applied in changeset commit:git|68b5f14d536c1a81c63412a9f3701380c9bc116c.
----------
Fix assertion failed in Complex.polar without NDEBUG (#3551)
Fixes [Bug #17172]. -
04:31 AM Bug #17172: Passing a complex number to Complex.polar occurs assertion failed when NDEBUG is undefined
- A patch is https://github.com/ruby/ruby/pull/3551.
-
- ```
$ build-debug/ruby -e "p Complex.polar(1, Complex(1, 0))"
Assertion Failed: ../complex.c:637:f_complex_polar:!RB_TYPE_P(y, T_COMPLEX)
ruby 3.0.0dev (2020-09-15T22:45:01Z fix_complex_polar 344304c89f) [x86_64-linux]
``` - 10:27 AM Revision 68b5f14d (git): Fix assertion failed in Complex.polar without NDEBUG (#3551)
- Fixes [Bug #17172].
-
-
- * That is, for plain string literals, not interpolated.
* The test below is very similar and uses the same check. -
- This reverts commit 889bbbaf527372c5cc229b56dca9a2322e325f26.
-
07:45 AM Bug #17152: Missing TracePoint return event for String#unpack
- Yes Shun, I expect the return calls.
Thanks for the help! -
- 9b535f3ff7 does not support ISeq#to_binary.
To make CI green, skip this test temporarily.
Please fix this issue and revert this patch. -
05:29 AM Misc #17174 (Feedback): "Error relocating, symbol not found" error when compiling a native extension on Alpine with Ruby >=2.4
- My native extension gem compiles fine with all versions of Ruby on macOS but only with Ruby 2.3 on Alpine. It throws various `Error relocating`, `symbol not found` when I require the `.o` file. Here I made a branch to reproduce it in Doc...
-
- Debian GNU/Linux 10 (buster) の OpenSSL 1.1.1d の環境だと https://www.famitsu.com で `dh key too small` になってつながらないのですが、 `ciphers` に `DEFAULT:!DH` を設定するとつながるので、 `open-uri` 経由でも `ciphers` を設定したいです。
curl での確認:
```
% curl --head https://www.... -
12:11 AM Feature #17170 (Feedback): Numeric.zero, Numeric.one
- ```ruby
% rbenv exec irb
irb(main):001:0> require 'bigdecimal'
=> true
irb(main):002:0> require 'bigdecimal/util'
=> true
irb(main):003:0> require 'complex'
=> false
irb(main):004:0> 0.to_i
=> 0
irb(main):005:0> 0.to_f
=> 0.0
...