Activity
From 12/15/2021 to 12/21/2021
12/21/2021
-
10:52 PM Revision fff9b45f (git): Reverts the accidental commits for rubygems/bundler.
- We stay to the stable version for releasing Ruby 3.1.0.
Revert commits:
b86a7ba49252e00e19c22bbcdc821b5acc31ff92
ef973aa7aa65143a4f05d623c016d459426770d0 - 10:04 PM Revision d9570437 (git): Remove incorrectly added `update.rb` file.
- 09:57 PM Revision e3092035 (git): Extended interface for IO::Buffer & documentation.
-
07:22 PM Feature #14975: String#append without changing receiver's encoding
- `IO::Buffer` makes this less important.
-
06:12 PM Revision b86a7ba4 (git): [rubygems/rubygems] Fix error when gem specified twice in gemfile under different platforms
- https://github.com/rubygems/rubygems/commit/83bc87ca98
-
- https://github.com/rubygems/rubygems/commit/04b1ac72b9
- 05:49 PM Revision ad353773 (git): * 2021-12-22 [ci skip]
-
05:49 PM Revision 779039a4 (git): add ref to the ticket
-
04:22 PM Bug #18380: TestAddressResolve#test_socket_getnameinfo_domain_blocking test failures
- Still issue with commit:git|cdb7d699d0
-
- This might be outfall of the did_you_mean update, which was reverted. It seems that commit:git|cdb7d699d0 passes fine. I'll keep observing.
-
- Testing Ruby 3.1 commit:git|a6f76122a2, I have started to observe deadlock in `TestObjectSpace#test_finalizer_thread_raise`. I hope I am not just impatient, but I have not observer any issues testing previously with commit:git|74b58dd690
-
- nobu (Nobuyoshi Nakada) wrote in #note-19:
> Could you show `rbconfig.rb` file too?
Done -
12:44 PM Bug #18373: Bundled gem (RBS, debug) extensions are not properly built
- vo.x (Vit Ondruch) wrote in #note-18:
> No, unfortunately. Please see the attached Makefile. I think that the culprit is still the `arch_hdrdir = $(extout)/include/$(arch)`
Yes, it's the problem.
Could you show `rbconfig.rb` file too? -
- nobu (Nobuyoshi Nakada) wrote in #note-17:
> Does this fix it?
No, unfortunately. Please see the attached Makefile. I think that the culprit is still the `arch_hdrdir = $(extout)/include/$(arch)`
-
- Does this fix it?
https://github.com/ruby/ruby/pull/5318 -
- Note that we can reproduce the issue with `--includedir` configure option like that: `./configure --includedir=/usr/include && make -j && make install DESTDIR=/tmp/foo` also on Ubuntu with git repo.
@nobu is checking. -
01:10 PM Revision cdb7d699 (git): Revert commits for did_you_mean
- This reverts commit 4560091b1c99ab33db0d653b9dd2d977fe4676d5.
This reverts commit a6f76122a2395bd914daa0aa04fb5a6ce4e0c045.
This reverts commit e59b18a6379c55f15ccda85c27d6997d44ef5293.
This reverts commit 505dfae05d56d844ea150676edb8785... -
-
12:12 PM Feature #14394: Class.descendants
- Sorry for the last minute change. The biggest reason is that we still have several options, so I didn't want to restrict the future possibility.
For example, `Class#descendants` can either:
* behave as it was first introduced. `Modul... -
11:38 AM Feature #14394: Class.descendants
- Just curious about the motivation of the revert though. I sin mainly because of the potential `Module#descendants`?
Because unless I'm missing something, the potential future inclusion of `Module#descendants` would change the behavior... -
- Rails fixed in https://github.com/rails/rails/pull/43951, I'll see if we can get a 7.0.1 release before 3.1.0 is out.
-
- 10:08 AM Revision e59b18a6 (git): Update default gems list at 505dfae05d56d844ea150676edb878 [ci skip]
-
-
- [Bug #17866][ruby-core:103866]
-
09:21 AM Revision 48a0905d (git): Add missing '%' in format string
-
-
-
- I discussed this with @naruse. We decided to bundle Psych 4 with Ruby 3.1. And we should note this incompatibility and migration way with Psych 3.3.2 to NEWS.
-
06:50 AM Revision 1855f901 (git): [ruby/irb] Check colorize option correctly to clear char attr and don't use it for tests
- https://github.com/ruby/irb/commit/de561cafeb
-
- https://github.com/ruby/irb/commit/b5f953dc33
- 06:27 AM Revision 10a2273e (git): Update default gems list at f51811ca75894f5227fb2dd9f2f61f [ci skip]
-
-
- `overloaded_cme_table` keeps cme -> monly_cme pairs to manage
corresponding `monly_cme` for `cme`. The lifetime of the `monly_cme`
should be longer than `monly_cme`, but the previous patch losts the
reference to the living `monly_cme`.
... -
06:15 AM Bug #18414 (Closed): win32ole.gemspec installed on Linux?
- Applied in changeset commit:git|3c7e95966d6b9efd201170eb59b0e4a3c4eb6134.
----------
rbinstall.rb: install ext only when it's configured [Bug #18414] -
-
-
-
- 04:22 AM Revision 02a9a72f (git): Tidy up fiber scheduler interface documentation for `address_resolve` and `timeout_after`.
-
- `def` (`rb_method_definition_t`) is shared by multiple callable
method entries (cme, `rb_callable_method_entry_t`).
There are two issues:
* old -> young reference: `cme1->def->mandatory_only_cme = monly_cme`
if `cme1` is young and `m...
12/20/2021
- 11:32 PM Revision 711342d9 (git): Update cont.c
-
11:32 PM Revision eae5a34b (git): Add SchedulerInterface#timeout_after and #address_resolve docs
-
-
- All the issues here have been addressed, except for:
(1) `get_value` / `set_value` type names have not changed, but the method name has changed.
(2) Using flags to indicate allocation strategy. I will review it and document the expec... -
- Yes that all makes sense.
After spending some time today working on it, I decided I don't like the mutable/immutable terminology. The problem is, even if the buffer is "immutable", the underlying data can still change. This is the unf... -
- > We need to be a little bit careful because string operations are not the primary point
Interesting! Because I started to document most of the usage from the point of view of some code mostly using `get` and `set`, but then I looked ... - 11:25 PM Revision 617687df (git): Rename IMMUTABLE to READONLY.
- 11:25 PM Revision 9fbf94ff (git): Improve interface for get/set/copy.
- 11:25 PM Revision 2de5dc53 (git): Fix handling of clear with offset and without length.
- 11:25 PM Revision c86bcd43 (git): Mark non-private mapped files as external.
- 11:25 PM Revision da46b8d8 (git): Default `IO::Buffer#get_string` to use BINARY encoding.
- 11:25 PM Revision c3d8d26a (git): Add tests for `IO::Buffer` `get`/`set`.
- 11:25 PM Revision 71bbc40f (git): Rename `to_str` -> `get_string` and add support for encodings.
- 11:25 PM Revision 49166fc7 (git): Improved exception usage/classes.
- 11:25 PM Revision 71bf5cef (git): Fix handling of frozens strings.
- 11:25 PM Revision 9de5c4ec (git): Fix mapping invalid non-file object.
-
10:28 PM Revision 6ad8cf70 (git): [DOC] Enhanced RDoc for IO (#5307)
- Treated:
#sync
#sync=
#fsync
#fdatasync
#fileno
#pid
#inspect
#to_io -
- https://github.com/rubygems/rubygems/commit/d6df0b7de0
Co-authored-by: André Arko <andre@arko.net> -
- The `BUNDLE_` prefix should be reserved to first class settings that
should be listed when running `bundle config`. This one is just a hacky
environment variable that has not corresponding documented setting.
https://github.com/rubygems... -
- https://github.com/rubygems/rubygems/commit/d647ab5607
-
09:13 PM Revision 39cf0b53 (git): Show whether object is garbage in rb_raw_obj_info()
- When using `rp(obj)` for debugging during development, it may be
useful to know that an object is soon to be swept. Add a new letter to
the object dump for whether the object is garbage. It's easy to forget
about lazy sweep. -
09:08 PM Feature #18402: Argument Labels
duerst (Martin Dürst) wrote in #note-6:
> Does this mean that both names can be used inside the method? And in that case, would they alias the same variable, or would there be two variables? Or can only one of the names be used inside...- 07:57 PM Revision 8b1b4e38 (git): [rubygems/rubygems] Skip find_in_unresolved_tree test for TruffleRuby due to longer runtime
- https://github.com/rubygems/rubygems/commit/36b8fbc508
-
- That's unfortunate. The very recently released Rails 7.0.0 will need to be updated, some code paths won't expect `Class#subclasses` being present, but not `Class#descendants`.
I'll update Rails tomorrow. -
07:04 PM Feature #14394 (Open): Class.descendants
- matz (Yukihiro Matsumoto) wrote in #note-33:
> Considering those facts, let up postpone introducing `Class#descendats` for 3.1.
Class#descendants removed at commit:3bd5f27f737c7d365b7d01c43d77a958c224ab16. We can potentially reintro... -
- `Module#descendants` has been proposed too in #9779.
In addition, upon the existence of newly introduced `Class#subclasses`, the need for `Class#descendants` is decreased.
Considering those facts, let up postpone introducing `Class#d... -
07:02 PM Revision 3bd5f27f (git): Remove Class#descendants
-
06:34 PM Feature #18408: Allow pattern match to set instance variables
- This worked originally (when this feature was called "rightward assignment"). But since it has been transformed into a pattern match (in the final 3.0 release), we lost the ability to assign to anything but local vars—that's what pattern...
-
- Install bundled gem extension files to the gem extension directory
under DESTDIR, when static-linked-ext as well as non-static case. -
-
- https://github.com/ruby/reline/commit/1f8a3aee43
Co-authored-by: Alex Gittemeier <me@a.lexg.dev> -
- In some tests, the LineEditor#reset method is always called, but doesn't
need to set the signal handlers there, so cuts it out to a separate
method.
https://github.com/ruby/reline/commit/b143c4f5f9 -
- https://github.com/ruby/reline/commit/7a758e73dc
- 03:11 PM Revision e54e99bd (git): * 2021-12-21 [ci skip]
-
03:05 PM Revision 6b67f063 (git): test/fiber/test_io_buffer.rb: fix file descriptor leaks
- I got the warning while running "make test-all":
Leaked file descriptor: TestFiberIOBuffer#test_write_nonblock: 9 : #<UNIXSocket:fd 9>
Closed file descriptor: TestFiberIOBuffer#test_read_write_blocking: 9
Leaked ... -
- So the basic scenarios seem to work as expected (unexpectedly for me, sorry). But this is the failing scenario, where I am using expanded tarball:
~~~
$ history
1 cd /builddir/build/BUILD/ruby-3.1.0-74b58dd690/
2 autoconf
... -
- mame (Yusuke Endoh) wrote in #note-12:
> Please answer to https://bugs.ruby-lang.org/issues/18373#note-3. I couldn't reproduce the issue by the following commands
> ...
This ^^ is one issue. The starting point should be something like:... -
09:57 AM Bug #18373: Bundled gem (RBS, debug) extensions are not properly built
- Please answer to https://bugs.ruby-lang.org/issues/18373#note-3. I couldn't reproduce the issue by the following commands
```
$ cd /tmp
$ git clone https://github.com/ruby/ruby.git
$ cd ruby
$ ./autogen.sh
$ ./configure
$ make
... -
- Replace EVP_PKEY_CTX_new_id() with the new EVP_PKEY_CTX_new_from_name()
which takes the algorithm name in a string instead of in an NID.
https://github.com/ruby/openssl/commit/d6535d13d1 -
- OpenSSL 3.0's EVP_PKEY_get0() returns NULL for provider-backed pkeys.
This causes segfault because it was supposed to never return NULL
before.
We can't check the existence of public key components in this way on
OpenSSL 3.0. Let's just... -
- Use the combination of TLS 1.2 and TLS 1.3 instead of TLS 1.1 and TLS
1.2 so that will the test case will be run on latest platforms.
https://github.com/ruby/openssl/commit/e168df0f35 -
- Use a different invalid data example to prevent SSLSocket#accept from
reaching EOF.
https://github.com/ruby/openssl/commit/2e089c1916 -
- Add all SSL_OP_* constants defined in OpenSSL 3.0.0 which are not
specific to DTLS.
https://github.com/ruby/openssl/commit/b1ee2f23b2 -
- The entire ENGINE API is deprecated in OpenSSL 3.0 in favor of the new
"Provider" concept.
OpenSSL::Engine will not be defined when compiled with OpenSSL 3.0.
We would need a way to interact with providers from Ruby programs, but
since ... -
- EVP_MD_CTX_copy() doesn't seem to work as intended on HMAC EVP_MD_CTX
on OpenSSL 3.0.0 and causes a double free. I haven't found the root
problem yet, but let's skip the test case for now.
https://github.com/ruby/openssl/commit/4699581639 -
- According to the manpage, the "siglen" parameter must be initialized
beforehand.
https://github.com/ruby/openssl/commit/6a60c7b2e7 -
- Do not attempt to actually use all algorithms. Not all algorithms listed
in OpenSSL::Cipher.ciphers are always available.
https://github.com/ruby/openssl/commit/91d04f991f -
- OpenSSL 3.0 made EVP_PKEY immutable. This means we can only have a const
pointer of the low level struct and the following methods can no longer
be provided when linked against OpenSSL 3.0:
- OpenSSL::PKey::RSA#set_key
- OpenSSL::PKey... -
- OpenSSL::PKey::EC#generate_key! will not work on OpenSSL 3.0 because
keys are made immutable. Users should use OpenSSL::PKey.generate_key
instead.
https://github.com/ruby/openssl/commit/5e2e66cce8 -
- OpenSSL::PKey::DH#generate_key! will not work on OpenSSL 3.0 because
keys are made immutable. Users should use OpenSSL::PKey.generate_key
instead.
https://github.com/ruby/openssl/commit/8ee6a582c7 -
- Similarly to DH#compute_key, work around it by constructing a
SubjectPublicKeyInfo. This should be considered as a temporary
implementation.
https://github.com/ruby/openssl/commit/fc9aabc18d -
- DH#set_key will not work on OpenSSL 3.0 because keys are immutable.
For now, let's reimplement DH#compute_key by manually constructing a
DER-encoded SubjectPublicKeyInfo structure and feeding it to
OpenSSL::PKey.read.
Eventually, we sho... -
- We can use it to implement OpenSSL::PKey::PKey#initialize_copy. This
should work on all key types, not just DH/DSA/EC/RSA types.
https://github.com/ruby/openssl/commit/66cd8cbaaf -
- Allocate an EVP_PKEY when the content is ready: when #initialize
or #initialize_copy is called, rather than when a T_DATA is allocated.
This is more natural because the lower level API has been deprecated
and an EVP_PKEY is becoming the ... -
- Passing NULL to ossl_pkey_new() makes no sense in the first place, and
in fact it is ensured not to be NULL in all cases.
https://github.com/ruby/openssl/commit/316cb2a41f -
- OpenSSL 3.0 has rewritten routines to load pkeys (PEM_read_bio_* and
d2i_* functions) around the newly introduced OSSL_DECODER API.
This comes with a slight behavior change. They now decrypt and parse
each encountered PEM block, then ch... -
- PEM-encoded private keys are sometimes stored together with irrelevant
PEM blocks, such as the corresponding X.509 certificate.
PEM_read_bio_*() family automatically skips unknown PEM blocks, but on
OpenSSL 3.0 we will be using the new ... -
- baweaver (Brandon Weaver) wrote:
> ### Current Viable Workarounds
> ...
In Ruby 3.1, we can omit parens:
```ruby
list_of_people.select { _1 in first_name: 'something' | /another/ }
```
Much readable, WDYT?
Can we go further ... -
12:39 PM Feature #17339: Semantic grouping with BigDecimal#to_s
- I will follow the consequence of #18410.
-
- This was fixed at https://github.com/ruby/bigdecimal/issues/94.
-
-
- 09:50 AM Revision 7700e9a5 (git): Update bundled gems list at 2021-12-20
-
-
- https://github.com/ruby/psych/commit/e7bbf26cb2
- 09:26 AM Revision 4db88173 (git): Update default gems list at 706c7a27fa00f72ec4f5d2f3af8151 [ci skip]
-
- https://github.com/ruby/psych/commit/75ab76e788
-
-
- https://github.com/ruby/logger/commit/cdeddb2f67
-
- https://github.com/ruby/irb/commit/c24a8e2483
-
- https://github.com/ruby/irb/commit/4cade4b7e5
- 07:05 AM Revision 145f7c09 (git): Update bundled gems list at 2021-12-20
-
- https://github.com/ruby/reline/commit/ba97f3bd87
-
- https://github.com/ruby/reline/commit/2ce2696d6e
-
- https://github.com/ruby/reline/commit/96067bc13c
-
- https://github.com/ruby/reline/commit/9bf2ca692e
-
- https://github.com/ruby/reline/commit/16aa20c380
-
- https://github.com/ruby/reline/commit/7e5dbe4750
-
- https://github.com/ruby/reline/commit/c4bb1bfd79
-
- https://github.com/ruby/reline/commit/4b7fa6b213
-
05:51 AM Revision 2a8ff602 (git): [ruby/reline] windows jruby issue
- jruby needs terminal control with Windows API on classic console
https://github.com/ruby/reline/commit/b61bc43374 -
- Windows Terminal does smart screen clearing when \e 2 J (not clear entire screen but scrolls down just needed)
On consoles not support sequences, ruby still converts it to API call.
https://github.com/ruby/reline/commit/c00930dab9 -
- ScrollConsoleScreenBuffer can't scroll window of Windows Terminal.
Use LF to sctoll.
Microsoft says
```In the virtual terminal sequences world, the size of the window and the size of the screen buffer are fixed to the same value.
```
ht... -
- This reverts commit https://github.com/ruby/reline/commit/6009b3ef7ab7.
To merge a Pull Request...
https://github.com/ruby/reline/commit/83021f4267 -
- This reverts commit https://github.com/ruby/reline/commit/1bb071bcf5a9.
To merge a Pull Request...
https://github.com/ruby/reline/commit/e9fe1b1305 -
- Treated:
::try_convert
#write
#<<
#flush
#tell
#seek
#pos=
#rewind
#eof -
-
- https://github.com/ruby/pp/commit/c66d1b9cdc
12/19/2021
-
- If `bundler _<version>_` is given, I guess the most reasonable approach
is to completely skip version switching, because the user is technically
opting out of it. But since binstubs completely remove this argument
from `ARGV` after proce... -
- https://github.com/rubygems/rubygems/commit/23178f7d7b
-
- > (2) I am not sure that type names (:U8, :S8, :U16, :u16 and so on) introduced "from scratch" (?) is a good thing. From one point, they are kinda nice, and consistent. From another, we already have abbreviated names for the types in Arr...
-
- Just for reference, link to the latest draft of the absolutely amazing documentation: https://github.com/ruby/ruby/pull/5302
-
- Thanks so much for this awesome list of things we should try to address.
https://github.com/ruby/ruby/pull/5303
We are considering how `to_str` should work. Implicit conversion is useful but it could also introduce issues.
> Ano... -
- (Got the latest `master`, updated the problems list :))
-
- Hello. While working for documentation for `IO::Buffer` (still WIP), I uncovered these problems. I believe at least some of them should be fixed before the release, though I understand not much time left.
The list is not final, probab... -
- 03:50 PM Revision 1dd10e18 (git): [DOC] Add documentation for Random.rand and Random.seed
- 03:37 PM Revision 12ad53f4 (git): * 2021-12-20 [ci skip]
-
- Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
-
-
- Applied in changeset commit:git|e2ec97c4b823a0b2e0c31e7a6d77b1dcdc0dfada.
----------
[DOC] How to get the longest last match [Bug #18415] -
- I agree that this is intentional and a documentation matter.
-
10:44 AM Bug #18415: String#rpartition is not sufficiently greedy compared to String#partition
- I think the actual result is expected. It is `partition`'s responsibility to match the left-most position possible, and it is `+`'s responsibility to match as long as possible from that position. Likewise, it is `rpartition`'s responsibi...
-
02:02 AM Bug #18415 (Closed): String#rpartition is not sufficiently greedy compared to String#partition
- ``` ruby
3.1.0-preview1 :043 > "...999...".partition /\d+/
=> ["...", "999", "..."]
3.1.0-preview1 :044 > "...999...".rpartition /\d+/
=> ["...99", "9", "..."]
```
Using the regex /\d+/, `partition` gives you all consecutiv... -
12:37 PM Revision 3f2b5818 (git): Add news about allowing omission of parentheses in one-line pattern matching
-
11:58 AM Feature #18418 (Open): Add Net::HTTP#security_level=
- ### Use Case:
I want to connect to an HTTP(S) server (using `Net::HTTP`) that cannot connect with security level[^1] 2 and needs to specify security level 1.
My server is a newer debian (docker image `ruby:3.0.3`, based on debian bul... -
-
11:26 AM Revision 0eb1c4ea (git): [ruby/irb] Add information about --extra-doc-dir option in the comments
- https://github.com/ruby/irb/commit/ac3d4b9e79
- 11:17 AM Revision 56811617 (git): Improve IO::Buffer resize and introduce ownership transfer.
-
05:56 AM Bug #18416 (Closed): Backport 6bef1ac62850be6d83a069ac3d3be0b8e2f3afcf
- This is a ticket for the backport management.
I don't know 6bef1ac62850be6d83a069ac3d3be0b8e2f3afcf is necessary in ruby_3_0. Just for memorandum. I will take a look into it later. -
- When merged to ruby/ruby, reline.gemspec file is located under
lib/reline, as the same as reline/version.rb. That is the latter
path relative from the former differs from the ruby/reline case,
and the reline/version.rb in the default lo... -
- https://github.com/ruby/reline/commit/dabf5313e0
12/18/2021
- 11:25 PM Revision f3e30b26 (git): Default size for IO::Buffer.
-
- and it fixes compaction issue:
http://rubyci.s3.amazonaws.com/freebsd12/ruby-master/log/20211218T203001Z.fail.html.gz -
08:59 PM Revision 45f21824 (git): YJIT: Implement intern
- `intern` showed up in the top 20 most frequent exit ops (granted with a
fairly small percentage) in a benchmark run by @jhawthorn on
github/github.
This implementation is similar to gen_anytostring, but with 1
stack pop instead of 2.
C... -
- `iseq_type_sym()` returns `ID` (surprisingly!) so rename it
to `iseq_type_id()`. -
- It is shorthand `ISeq#to_a[9]`.
-
-
- Now the following code works without an exception.
```
RubyVM.keep_script_lines = true
eval(<<END)
def foo
end
END
p RubyVM::AbstractSyntaxTree.of(method(:foo))
``` -
- This check is needed to fix a bug of error_highlight when NameError
occurred in eval'ed code.
https://github.com/ruby/error_highlight/pull/16
The same check for proc/method has been already introduced since
64ac984129a7a4645efe5ac57c168... -
- [DOC] What's Here for ENV
- 04:07 PM Revision 1ed520a9 (git): * 2021-12-19 [ci skip]
-
- fix this failure:
```
configure ... cppflags=-DMJIT_FORCE_ENABLE
...
make test-spec
1)
The -v command line option when used alone prints version and ends FAILED
Expected
"ruby 3.1.0dev (2021-12-18T10:10:42Z master 78c175280b) +MJIT [x8... -
03:34 PM Bug #17568 (Rejected): Thread.handle_interrupt is per-Thread but should probably be per-Fiber
- Closing since it seems intentional for some cases (Enumerator fibers).
- 11:19 AM Revision 235f523a (git): Fix spelling of DECLAIR_TYPE.
- 10:10 AM Revision 78c17528 (git): Fix code formatting.
- 10:10 AM Revision 1560cc16 (git): Introduce setup instructions for better parallelism.
- 06:15 AM Revision 75b5a480 (git): Initial hacking guide.
- 05:19 AM Revision 42d32311 (git): Introduce io_result wrapper for passing `[-errno, size]` in VALUE.
-
-
-
-
- https://github.com/ruby/reline/commit/7b50638e24
-
- https://github.com/ruby/reline/commit/6009b3ef7a
-
- https://github.com/ruby/reline/commit/d014cc85b7
-
- https://github.com/ruby/reline/commit/1bb071bcf5
12/17/2021
-
11:38 PM Revision 66ec3217 (git): [ruby/pp] Add context to bind_call and skip 2.7+ test
- The bind_call definition here is added primarily to support
running the tests on JRuby 9.3, which only supports Ruby 2.6
features. The excluded test appears to depend on inspect behavior
that changed after Ruby 2.6.
With these two chang... -
- * UnboundMethod#bind_call
* ruby2_keywords gem for testing
https://github.com/ruby/pp/commit/721d8cbd10 -
- The use of `etc.so` here requires that etc is always implemented
as a C extension on-disk. However at least one impl – JRuby –
currently implements it as an internal extension, loaded via a
Ruby script. This require should simply use the... -
- This class does not exist in any implementation except CRuby.
I would recommend moving this code somewhere else, like a separate
file loaded only on CRuby or into CRuby itself. For now this
change is sufficient to load the library on ot... -
11:38 PM Revision de6ef2ca (git): [ruby/benchmark] Freeze VERSION
- https://github.com/ruby/benchmark/commit/89b889b109
-
- https://github.com/ruby/logger/commit/5ee0fd6e51
- 11:38 PM Revision 43755468 (git): [ruby/logger] Test cases
- https://github.com/ruby/logger/commit/40adb645e2
- 11:38 PM Revision a8b11b5c (git): [ruby/logger] Changes to datetime formatting
- Formatting a datetime should only pertain to itself and valid datetimes do not contain a space. Should there be a desire to show show a space between the datetime and the process pid in the formatted log, this formatting logic should tak...
-
- Inline and remove iseq_supported_args_p(iseq) to remove a potentially
dangerous double check on `iseq->body->param.flags.has_block` and
`iseq->body->local_iseq == iseq`. Double checking should be fine at the
moment as there should be no ... -
11:26 PM Revision c2197bf8 (git): YJIT: Fix check for required kwargs
- Previously, YJIT would not check that all the required keywords were
specified in the case that there were optional arguments specified. In
this case YJIT would incorrectly call the method with invalid arguments. -
- Previously we mirrored the fast paths the interpreter had for having
only one of kwargs or optional args. This commit aims to combine the
cases and reduce complexity.
Though this allows calling iseqs which have have both optional and
ke... -
- * What's Here for Symbol
- 07:21 PM Revision 6a1365d7 (git): Update bundled gems list at 2021-12-17
-
-
- That explanation makes sense and the "Ignoring debug-1.3.4 because its extensions are not built. Try: gem pristine debug --version 1.3.4" is just red herring there and therefore the PR is legit.
Nevertheless, the point that win32ole.g... -
- #18373 is completely different from the issue that the pull request tried to fix. The pull request is about default gems, and #18373 is about bundled gems.
I wonder if you misunderstand `--with-static-linked-ext`. The configure option... -
- While win32ole is not built on Linux and it used to not be installed. But the .gemspes is installed now. I believe that this is due to https://github.com/ruby/ruby/pull/5206.
Besides the win32ole.gemspec being installed now, I think t... -
- Dan0042 (Daniel DeLorme) wrote in #note-21:
> Wait a sec... actually there's precedent for this. `super` does this. It forwards all arguments up the inheritance chain. And it's possible to do `eval("super")`. `__params__` is like the f... -
01:54 PM Feature #17785: Allow named parameters to be keywords
- Dan0042 (Daniel DeLorme) wrote in #note-19:
> Eregon (Benoit Daloze) wrote in #note-17:
> ...
Wait a sec... actually there's precedent for this. `super` does this. It forwards all arguments up the inheritance chain. And it's possible t... -
- BTW, this causes also different issues as far as I can tell:
https://github.com/ruby/ruby/pull/5206 -
- I was able to use this little hack to compile the extensions:
~~~
find .bundle -name extconf.rb -exec \
sed -i \
-e '/create_makefile/i \$arch_hdrdir = "$(hdrdir)/../.ext/include/$(arch)"' \
-e '/create_makefile/i \$DLDF... -
04:49 PM Bug #18413: Segfault in `ripper/lexer.rb`
- Encountered a segfault in the `ripper/lexer.rb` file when attempting to compile Ruby `master` this morning.
### Steps to reproduce:
* `make -j4`
### Console dump:
```shell
BASERUBY = /usr/bin/ruby --disable=gems
... -
- Was attempting to build using 3.0.3 today and encountered a segfault with `ripper`.
- 04:18 PM Revision 74b58dd6 (git): Update bundled gems list at 2021-12-17
- 04:18 PM Revision 55a93533 (git): * 2021-12-18 [ci skip]
-
-
-
- Adds file doc/case_mapping.rdoc, which describes case mapping and provides a link target that methods doc can link to.
Revises:
String#capitalize
String#capitalize!
String#casecmp
String#casecmp?
String#downcase
... -
-
10:37 AM Bug #17866: Incompatible changes with Psych 4.0.0
- Can I get a final decision if Psych 4 will be shipped with Ruby 3.1 as a standard library?
I think this is one of the biggest incompatibilities which will be introduced with Ruby 3.1.
As far as I understand, Psych is a standard libra... -
- https://github.com/rubygems/rubygems/commit/f481e8f41a
-
- https://github.com/rubygems/rubygems/commit/e4a1a9663d
-
- https://github.com/rubygems/rubygems/commit/84e2d6a955
-
- If we are resolving a dependency against a particular platform, and
there are no platform specific variants of the candidates that match
that platform, we should not consider those candidates.
https://github.com/rubygems/rubygems/commit... -
- https://github.com/rubygems/rubygems/commit/077e3c2e4d
-
- Use a more standard naming for gems.
https://github.com/rubygems/rubygems/commit/75121e83f1 -
-
- pointed by @nobu.
-
- The `ENV` object can have instance variables like other objects,
but they should be accessed only on the main ractor.
fix https://github.com/ruby/ruby/pull/5263#issuecomment-995585766 -
04:54 AM Revision d524b9de (git): NEWS.md: Add link to `rbs collection` documentation
-
-
04:23 AM Bug #18412: Segfault in test_ractor.rb
- Presumably the same failure observed in CI for x86_64-linux-gnu: http://rubyci.s3.amazonaws.com/ubuntu/ruby-3.0/log/20211217T005904Z.log.html.gz
-
- Parallel worker's stdout is captured as the control protocol.
12/16/2021
-
- I agree with the proposal, but `choice` is not a good name for a public method.
Matz.
- 05:58 PM Revision c4c78343 (git): Update stdlib version at 693a561cfa19dc35ff0d199a1934c9 [ci skip]
-
- 04:56 PM Revision b746b713 (git): * 2021-12-17 [ci skip]
-
- I am not sure what to really blame. The place where it goes quite wrong is actually mkmf, somewhere around [here](https://github.com/ruby/ruby/blob/fff058a8d6c2e5d7c3529484a1c9a636dbf66dc2/lib/mkmf.rb#L242-L249):
~~~ruby
if not $extmk... -
-
02:44 PM Bug #18133: LTO: TestGCCompact#test_ast_compacts segfaults on i686
- Thank you for checking the patch! Historically, Ruby has been using 16KB pages, so there's assumptions in the GC about this. This wasn't a problem on 64KB page size systems when we were using `posix_memalign`, but we can no longer use th...
-
- Thx, the latest version passes the test suite everywhere and your explanation why it does not work makes sense.
Nevertheless, this makes me wonder in case there are 64 KB pages used for mmap, why Ruby wont use 64 KB pages as well? I a... -
11:10 AM Revision fff058a8 (git): Add description of ruby header files to extension.rdoc
-
11:03 AM Revision 9b187fec (git): Add `rb_parser_set_pos` function
- Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
Co-authored-by: Marivaldo Cavalheiro <marivaldo@gmail.com> -
11:02 AM Revision 41d4902c (git): Improve wording on NEWS.md
-
- During build of Ruby 3.0.3 for Fedora, I sometimes bump into this issue:
~~~
... snip ...
test_attr.rb ..
test_autoload.rb ........
test_block.rb ...................................................... -
09:07 AM Feature #18410: Proposal to make inspect include underscores on numerics
- sawa (Tsuyoshi Sawada) wrote in #note-4:
> In agreement with mame's comment, even if we can come to an agreement on the spec, I think it should be a property of pretty print rather than normal inspection. If there is need to make the nu... -
- mame (Yusuke Endoh) wrote in #note-2:
> Thousands separators may vary depending on locales. See the discussion of #16011.
> ...
As mentioned by Martin, I was proposing always using _, as you can in Ruby code, not any kind of locale spe... -
- In agreement with mame's comment, even if we can come to an agreement on the spec, I think it should be a property of pretty print rather than normal inspection. If there is need to make the numbers particularly readable to human, in suc...
-
03:54 AM Feature #18410: Proposal to make inspect include underscores on numerics
- mame (Yusuke Endoh) wrote in #note-2:
> Thousands separators may vary depending on locales. See the discussion of #16011.
Yes indeed. The advantage of using 964_218_442 is that a) I don't think there's actually a locale that uses thi... -
- Thousands separators may vary depending on locales. See the discussion of #16011.
Even if we can come to an agreement on the spec, changing `Integer#inspect` may have a big impact. We need to estimate the impact of the incompatibility... -
-
- Run only checks for source code. Currently, our CIs do almost
nothing about the documents. -
-
- Actually used methods are all instance method, not the singleton
method. -
-
- @matz this isn't strictly necessary but makes pure Ruby implementation of IO read/write hooks more efficient. I'm not sure if there is valid case for general user code. Can we consider it for Ruby 3.1?
-
- When implementing pure-ruby IO scheduler, we may need to invoke some Ruby IO operations without entering the scheduler.
```ruby
def io_write(fiber, io, buffer, length)
offset = 0
while length > 0
# From offset until th... -
- Thanks!
-
- When building with --with-static-linked-ext, some exts without rb file
doesn't produce neither .so or .rb under .ext/common. Therefore, change
rbinstall.rb to install gemspec even if there is no .so or .rb for that
case. -
- `IO#reopen` internally uses dup syscall but some platforms don't support
the syscall. re-assigning `$stderr` is enough to capture the interpreter's
errors and warnings. -
- `to_s` has the explict specification while `inspect` is often
vague. -
- Lomig (Lomig Enfroy) wrote in #note-5:
> The main idea / need is to use different names for variables/arguments when calling the method and inside the method itself.
Does this mean that both names can be used inside the method? And... -
-
04:02 AM Revision 02ba0bda (git): Remove RubyVM::JIT (#5275)
- [Feature #18349] reverts [Feature #17490]
12/15/2021
-
- The documentation for inspect reads "Returns a string containing a human-readable representation of obj." but for large numerical values, while 964218442 is human-readable, 964_218_442 is a lot more readable. The language allows the use ...
- 09:13 PM Revision 40cc8e92 (git): * 2021-12-16 [ci skip]
-
- No one reads it at the moment and it's heap allocated.
-
- @Dan0042 I basically agree with your assessment and support (a) and in addition I'd prefer if we don't introduce inconsistent formatting for quoted messages, i.e. `#<NameError:"...">` should definitely have a space between the `:` and `"...
-
- I'm interested in this feature because I'm uncomfortable with the current situation of sticking everything into `Exception#message` and losing access to the original message.
In #18194 I think everyone has agreed that
> an error message... -
- I noticed that commit:737e4432b978eb4b9f5b10fb6cc6d9c883a5d17a is not backported to Ruby 3.0.x series. So before `./configure` you may need to do something like
```bash
export LDFLAGS="-Wl,--no-as-needed"
```
Make sure that the `... -
05:04 PM Bug #18409: Crash (free(): invalid pointer) if LD_PRELOAD doesn't explicitly include libjemalloc.so.2
- How can I try to trace it? I tried attaching GDB to the process with little success.
-
- Seems like someone (who?) is calling libc `free()` directly, maybe?
-
- I recently installed Ruby 3.0.2 using rbenv with:
```
RUBY_CONFIGURE_OPTS='--with-jemalloc' rbenv install 3.0.2
```
After running our app it always crashed with:
```
free(): invalid pointer
[1] 51364 IOT instruction (core... -
- headius (Charles Nutter) wrote in #note-15:
> ```ruby
> ...
That looks somewhat complex to implement, but it's such an elegant API, and doesn't even require a new method.
+1 -
- @jeremeyevans0:
> It sounds like you want iteration, and you are recommending with Thread.each_backtrace yielding Thread::Backtrace::Location objects. Is that correct?
Yes to iteration. Capturing a single frame is only one use case... -
- On another semi-unrelated note, isn't weird that this method returns `nil` sometimes (e.g. `Struct.new(:a).keyword_init? # => nil`)?
Is there any precedence for a question mark method returning `nil`? I know `nil` is falsey, but this mig... -
- OK, I'll try to fight it, there are quite a few weird tricks that might help
-
- Gotcha! Thanks. I couldn't make it show up on the docs, though.
The docs are definitely there (struct.c:288)
```c
/*
* call-seq:
* StructClass.keyword_init? -> true or false
*
* Returns true if the class was initialized with +keyw... -
- I am able to repro this. Since ppc64 uses 64KB pages for mmap, we can't use mmap to allocate memory for Ruby pages (since they are 16KB). Because we can't use mmap, we can't use mprotect used by the read barriers of compaction, so we can...
-
- vo.x (Vit Ondruch) wrote in #note-18:
> Checking on CentOS Steam 9, it is passing on ppc:
> ...
Hups, scratch that, it fails also on c9s:
~~~
1) Error:
TestGCCompact#test_ast_compacts:
NotImplementedError: Compaction isn't avail... -
- Checking on CentOS Steam 9, it is passing on ppc:
https://kojihub.stream.rdu2.redhat.com/koji/taskinfo?taskID=848077
The configure checks:
~~~
checking for mmap... yes
checking for sys/user.h... yes
checking whether PAGE_SIZE... -
- peterzhu2118 (Peter Zhu) wrote in #note-16:
> I debugged this today. Can you try with commits [0130e17a410d60a10e7041ce98748b8de6946971](https://github.com/ruby/ruby/commit/0130e17a410d60a10e7041ce98748b8de6946971) and [32b7dcfb56a417c1... -
01:30 PM Bug #17826: Ractor#take hangs if used in multiple Threads
- Thank you for the review, and I'm glad to provide a confirmed bug report.
Volodya Sveredyuk and I invite you to give me a talk in our great and cozy local Ruby community in Lviv, Ukraine - Pivorak. Aaron Patterson and José Valim alrea... - 12:52 PM Revision fbd21a81 (git): Update stdlib version at 333865e56f594f365f2b3b4ef0f86f [ci skip]
-
- This file searches "lib/reline/version.rb" in the same directory,
"lib/reline". As no such path, the file in the default load path
is loaded unexpectedly. -
-
11:09 AM Bug #18381: Default vs Bundled gems
- :+1:
Documentation would solve probably my question about:
https://bugs.ruby-lang.org/issues/18400 -
-
- hsbt (Hiroshi SHIBATA) wrote in #note-3:
> You should add `prime` into your `Gemfile` for using prime via bundler environment.
Thanks for the hint, it resolves issue fully.
Can be closed. - 10:47 AM Revision 1fc80754 (git): Update stdlib version at 9f87c0cc6dff70e8a1c4d204d1d459 [ci skip]
-
- The `File::BINARY` flag is apparently ignored due to a ruby bug, and
thus writing can cause encoding issues.
https://github.com/rubygems/rubygems/commit/db4efbebf2 -
- https://github.com/rubygems/rubygems/commit/93051fd2aa
-
- https://github.com/rubygems/rubygems/commit/c6ef75424d
-
-
-
- Merge from https://github.com/rubygems/rubygems/commit/793ad95ecb40e84a1dcb4cb60f2686843ed90de5
- 09:00 AM Revision 583e06e2 (git): Update stdlib version at ded33ed5b8d923b0bb707b4e617ba0 [ci skip]
-
-
- 07:01 AM Revision b2395539 (git): Update bundled_gems at 2021-12-15
-
-
- `ENV[key] = long_str` can raise `Errno::EINVAL` also on ucrt env.
-
-
- We found that we need to make Ruby objects while locking the environ
to ENV operation atomically, so we decided to use `RB_VM_LOCK_ENTER()`
instead of `env_lock`. -
-
-
-
-
- * --braces-after-func-def-line
* --space-after-for -
-
-
-
-
-
-
-
- Previously, YJIT incorrectly discarded the upper 32 bits of the object
pointer when writing out VALUEs to setup default keyword arguments.
In addition to incorrectly truncating, the output pointers were not
properly tracked for handling...