Activity
From 04/02/2020 to 04/08/2020
04/08/2020
-
11:55 PM Revision cdd613b2 (git): configure.ac: Skip C++ compiler of Sun OpenStudio
- It fails to compile ext/-test-/cxxanyargs/cxxanyargs.cpp.
Need work to support it. Contribution is welcome. -
- I'm unsure why, but it fails on icc.
https://rubyci.org/logs/rubyci.s3.amazonaws.com/icc-x64/ruby-master/log/20200408T220004Z.log.html.gz
```
In file included from ../../.././include/ruby/3/core.h(32),
from ../../.././i... - 10:44 PM Revision a4021ecd (git): * 2020-04-09 [ci skip]
-
- I don't find how to suppress a false positive "unreachable" warning of
the Sun compiler:
```
"./include/ruby/3/core/rtypeddata.h", line 163: warning: statement not
reached
"./include/ruby/3/memory.h", line 256: warning: statement not re... -
10:32 PM Bug #16771 (Closed): Segmentation fault when inspecting a bound method
- This piece of code leads to a crash in ruby 2.7.1.
```ruby
bound_method = Kernel.instance_method(:respond_to?).bind(Object.new)
100000.times { bound_method.inspect }
```
I cannot reproduce the error in 2.7.0 so this must be new.... -
10:24 PM Bug #16770: Commit: RUBY3_HAS_BUILTIN: fix for nonexistent builtin - large increase in MinGW & mswin compile times
- Is it due to that specific commit, or rather due to https://github.com/ruby/ruby/pull/2991 ?
-
01:20 PM Bug #16770 (Closed): Commit: RUBY3_HAS_BUILTIN: fix for nonexistent builtin - large increase in MinGW & mswin compile times
- Inspecting logs both here and in ruby-loco, the following commit greatly increased compile time in MinGW builds:
https://github.com/ruby/ruby/commit/6163a4e1b33ed7df27d0c37e835ce04407857a16
In ruby-loco, the make/compile time jumpe... -
- @ko1 wrote in #note-5:
> Eregon (Benoit Daloze) wrote:
> ...
It would be great to unify the style primitives are written between Ruby implementations.
`__builtin_foo(1, 2)` is MRI-specific and not really pretty in Ruby code.
@ko1 Could ... -
- Agreed, and @ioquatix and @headius seemed positive too in some recent discussion.
-
10:45 AM Feature #16769: Struct.new(..., immutable: true)
- Makes sense.
-
09:00 AM Feature #16769 (Rejected): Struct.new(..., immutable: true)
- ## Background
We've discussed interface to pass Struct attributes (like `immutable: true`, which is actually not added yet) at once. But I believe just adding `immutable: true` alone is really helpful in various cases. Thus I've spun ou... -
04:49 PM Bug #16658: `method__cache__clear` DTrace hook was dropped without replacement
- @Koichi could you please take a look, since your commit removed the hook. Thx
-
04:41 PM Feature #16688: Allow #to_path object as argument to system()
- I think what would be useful is when interpolating Path like objects (`to_path`) that the path objects are shell escaped or cmd escaped when used with `system()` or other cmd running objects. There are always errors on windows with dire...
-
01:41 PM Revision e7128aa7 (git): Disable deprecation error on mingw
-
01:32 PM Revision dd04b803 (git): RARRAY_EMBED_LEN/RSTRING_EMBED_LEN: add RUBY3_CAST
- RUBY3_CAST is a macro to suppress g++/clang++ warnings about C-style
casts. Though Ruby core don't have to bother C++ situations, extension
libraries can benefit from this. -
- It turned out that compilation errors in icc were due to their having
broken __has_builtin. Let's just skip such situations. -
09:33 AM Revision 8d622a58 (git): [ruby/weakref] Update TODO on README
- https://github.com/ruby/weakref/commit/a67073856e
-
09:32 AM Bug #16337: kernel_gem.rb:68 - ThreadError
- Great, let me know if you run into issues.
-
02:20 AM Bug #16337: kernel_gem.rb:68 - ThreadError
- Sorry, I never got notifications about your posts. I was acting as the maintainer of google-fluentd, and trying to upgrade it to Ruby 2.7. My colleague will try upgrading rubygem. Thanks.
-
- * [Feature #16769] Struct.new(..., immutable: true) (k0kubun)
* I wanted it today, and found the discussion of [Feature #16122] stuck with designing helpers to set a combination of attributes. Is there any objection to introduce `immu... -
-
-
- Trying to fix icc breakage:
https://rubyci.org/logs/rubyci.s3.amazonaws.com/icc-x64/ruby-master/log/20200408T050004Z.fail.html.gz
It seems the macro had problems when a builtin does not exist for the
target compiler. Force evaluete to ... -
-
- https://github.com/ruby/net-ftp/commit/a56ba121ee
-
- https://github.com/ruby/net-ftp/commit/bed4bc31db
-
-
-
- * prime gem didn't use `version.rb` file.
* Removed development_dependency because they are handled by Gemfile. -
- * It couldn't use `English` because `english` is already reserved.
* Bump version to 0.7.0 because versions < 0.7 are reserved. -
- Split ruby.h
-
-
- 1. Are they completely binary compatible?
2. Only the last digit changes?
If it is expected from now on, I prefer to avoid hardcoding the list in Makefile.sub.
---
```diff
diff --git c/win32/Makefile.sub w/win32/Makefile.sub
... -
04/07/2020
-
06:40 PM Bug #13671: Regexp with lookbehind and case-insensitivity raises RegexpError only on strings with certain characters
- FYI The issue has been addressed in Onigmo https://github.com/k-takata/Onigmo/pull/116 and has already been released in version 6.2.0. I tried it by applying the changes using Ruby 2.6.6 and it works as expected.
-
- Applied in changeset commit:git|2f1895fa15386fb3cdc91c5604171290828b9da8.
----------
Fixed formatted substring expansion [Bug #16767] -
12:45 PM Bug #16767 (Closed): Time#strftime not working correctly with a big precision on recurred formats like "%c"
- Confirmed on: Ruby 2.6.3 and 2.7.0 on Mac OS X 10.13.
```ruby
Time .now .strftime("%28c")
```
Expected:
```
" Tue Apr 7 12:44:16 2020"
```
Actual:
```
" \u0000ue Apr 7 12:44:16 2020"
```
Appendix:
* For `"%c"... -
-
- Prior to 2.6 don't have `Module#const_source_location`, but redefinition warning message has this location.
-
- Applied in changeset commit:git|927308108cced69cae478798004524b9a5d2f252.
----------
Fix source location of autoloaded constant [Bug #16764] - 03:06 PM Revision 17112581 (git): * 2020-04-08 [ci skip]
-
- Confirmed on: Ruby 2.7.0 on Mac OS X 10.13.
```
require 'date'
DateTime._strptime("aus\neastern standard time", "%z")
```
Expected (Ruby 2.6.3 works like this):
```
=> {:zone=>"aus\neastern standard time", :offset=>36000}
`... -
-
-
06:44 AM Bug #16760 (Closed): backport #67305 / e39f7e64 to 2.6?
- Thank you for pointing this out. I have confirmed the snippet reproduces SEGV with 2.6.6 on my laptop.
I will fill the Backport field and close this ticket to notice stable branch maintainers. -
-
- 03:50 AM Revision bdf769bc (git): * 2020-04-07 [ci skip]
-
- Fixed up a58bbd6a512d95ca010d8bebae4fe590400c1413.
-
01:16 AM Bug #16669: Rails Action Text unit tests fail since b9007b6c548f91e88fd3f2ffa23de740431fa969
- https://bugs.ruby-lang.org/projects/ruby-master/repository/git/revisions/2943ff9d4441485a18773aa745bab7f47767dde2 fixes some of failures in Rails CI, but "undefined method `root' for Rails:Module" still exist I explained above.
https:...
04/06/2020
-
09:56 PM Bug #16765 (Feedback): Crash when use sass image-url in email layout
- Unfortunately, the crash report does not provide the relevant debugging information:
```
Thread 22 Crashed:: thread_pool.rb*
0 libsystem_kernel.dylib 0x00007fff6cd2c33a __pthread_kill + 10
1 libsystem_pthread.dylib ... -
09:37 PM Bug #16765: Crash when use sass image-url in email layout
- doctype html
html
head
meta(charset='utf-8')
meta(http-equiv="Content-Type" content="text/html")
scss:
.footer-box:after {
content: '';
width: 100%;
height: 100%;
opacity: 0... -
- I have the following layout I use for my emails:
mailer.html.slim
doctype html
html
head
meta(charset='utf-8')
meta(http-equiv="Content-Type" content="text/html")
css: -
08:02 PM Bug #16764 (Closed): Module.const_source_location does not work on autoloaded constants
- As we found when trying out `const_source_location` on a Rails project, it breaks when used on autoloaded constants.
Example:
```
$ cat foo.rb
autoload :Bar, './bar'
Bar
puts Module.const_source_location("Bar")
```
Retu... -
07:13 PM Feature #16763 (Closed): MSVC: allow ranges for MSVC 2017 and 2019 support in win32/Makefile.sub
- `_MSC_VER` differs from previous version in 2017 and 2019, in the sense that updates have their own last digit changing.
For a list of `_MSC_VER` see for example https://dev.to/yumetodo/list-of-mscver-and-mscfullver-8nd
config.h en... -
- The issue is similar to #15335 and the fix is also similar:
~~~
diff --git a/coroutine/arm32/Context.S b/coroutine/arm32/Context.S
index b66db29622..596284b5cf 100644
--- a/coroutine/arm32/Context.S
+++ b/coroutine/arm32/Context.S... -
-
-
-
-
-
- _mjit_compile_send.erb doesn't use _mjit_compile_insn_body.erb
-
- to unify the duplicated declarations and to make sure it's not used
until set properly.
Also changed it from legacy TRUE/FALSE to stdbool. -
-
- using sp_inc_of_sendish for consistency and to make it easier to
understand -
- and simplify `v` variable references a little.
There's no CALL_METHOD anymore, and the original code lives in
vm_sendish instead of insns.def now. -
08:30 AM Feature #16761 (Open): Add an API to move the entire heap, as to make testing GC.compact compatibility easier
- We recently started testing GC.compact effectiveness in production, and one challenge we faced was to ensure that C extensions were compatible with it.
Here's two examples of C-extensions which caused various issues, and their respect... -
08:10 AM Bug #16760 (Closed): backport #67305 / e39f7e64 to 2.6?
- https://bugs.ruby-lang.org/projects/ruby-master/repository/git/revisions/e39f7e64b73f0506def7adc88226d6821608da54
aka
https://github.com/ruby/ruby/commit/e39f7e64b73f0506def7adc88226d6821608da54
still crashes in the latest 2.6. ... -
- in mjit_compile_send to clarify it's not that deeply branched.
-
- Just FTR, this is from the original description:
~~~
0x00003fffb79413a8 - 0x00003fffb7941f78 is __libc_IO_vtables in /lib64/power8/libc.so.6
0x00003fffb74213c0 - 0x00003fffb7421f90 is __libc_IO_vtables in /lib64/libc.so.6
~... -
- shyouhei (Shyouhei Urabe) wrote in #note-10:
> > >Loding by SONAME is the only safe option.
> ...
This issue is specifically about glibc branches and they should be modified. There should not be provided any path when glibc is detected... -
- Thank you for confirmation.
vo.x (Vit Ondruch) wrote in #note-9:
> Well, we don't have PPC64 just PPC64LE on Fedora. I have run 5 builds and all passed just fine.
OK then, let me close this. Don't hesitate to reopen when somethin... -
- Thanks to the work by @Eregon in ruby/setup-ruby, the issue first appeared in 2.5.6.
Passed
ruby 2.4.10p364
ruby 2.5.0p0
ruby 2.5.1p57
ruby 2.5.3p105
ruby 2.5.5p157
ruby 2.6.0p0
Failed
ruby 2.5.6p201
See:
https://git...
04/05/2020
- 11:29 PM Revision a7e70003 (git): * 2020-04-06 [ci skip]
- 11:29 PM Revision ba3eccd9 (git): fix typo in documentation
-
- A failure with MinGW 2.5.8 appeared in ruby/spec CI from the below in core/binding/local_variable_set_spec.rb. I believe the specs passed on 2.5.7:
```ruby
it "raises a NameError on global access" do
bind = binding
-> { bind.loc... -
04:18 PM Feature #16746: Endless method definition
- To be honest, I'm a little confused if this is serious proposal now, since it probably started as a joke and got some attention already at GitHub pull request (https://github.com/ruby/ruby/pull/2996) with mixed attitude.
Anyway for one-... -
- The document in these external libraries may not be written in
RDoc, ignore the whole directory. -
- These libraries may not be written in RDoc.
-
09:24 AM Revision b25ef4bf (git): Suppress warnings: reserved for numbered parameter
-
04:35 AM Bug #16743: problem with multi threading [BUG] Segmentation fault
- > If it is a bug in the library, shouldn't Ruby show it differently?
Do you mean that the stack trace should be different? Not necessarily. When data corruption happens the crash site is usually far away from the code doing the wrong ... -
-
- It can fail due to different prefixes, e.g., drive letters or UNC
paths on DOSish platform. -
- This reverts commit 4a6571dbc14ee4e88c12cd9931f7695077a3ee6e,
because chkbuild does not follow. -
-
-
04/04/2020
- 11:28 PM Revision 5e2463b2 (git): * 2020-04-05 [ci skip]
-
-
09:55 PM Bug #16743: problem with multi threading [BUG] Segmentation fault
- alanwu (Alan Wu) wrote in #note-10:
> > what kind of bug in third-party library did you imagine caused the crash in method_entry_get?
> ...
If it is a bug in the library, shouldn't Ruby show it differently?
k0kubun (Takashi Kokubun)... -
- > what kind of bug in third-party library did you imagine caused the crash in method_entry_get?
Forgetting to mark objects can make this happen. It looks like the library is not marking the [self data member](https://github.com/zombieca... -
- According to the C backtrace, the place which caused SEGV was https://github.com/ruby/ruby/blob/v2_6_5/vm_method.c#L812.
> It is most likely a bug in the chingu-pathfinding library
I haven't taken a close look at it yet, but out of... -
- paulorja (paulo jansen) wrote in #note-6:
> alanwu (Alan Wu) wrote in #note-3:
> ...
It is most likely a bug in the chingu-pathfinding library, and should be filed upstream: https://github.com/zombiecalypse/chingu-pathfinding/issues -
- alanwu (Alan Wu) wrote in #note-4:
> Also, do you happen to be running on Ubuntu Bionic Beaver?
Im using Linux Mint
And tested on amazon ec2 instance with Ubuntu Server too -
- alanwu (Alan Wu) wrote in #note-3:
> I see that you use `chingu-pathfinding` which includes a native extension.
> ...
It's only happening when I use that lib.
I am a beginner with multi threading. I imagine the bug is happening because ... -
-
02:16 PM Revision 33260d40 (git): Enhance pattern matching introduction
-
10:00 AM Misc #16693: DevelopersMeeting20200410Japan
- * [Feature #16684] Use the word "to" instead of "from" in backtrace (sawa)
* Be free of wondering about the printed backtrace direction by using the word "to" in most-recent--call-last situation. -
- * [Bug #14413] `-n` and `-p` flags break when stdout is closed
* only in the loop / always signal on `EPIPE`
* `STDOUT` only / + `STDERR` / generic for `IO`
-
07:07 AM Misc #16693: DevelopersMeeting20200410Japan
- * [Bug #6087] How should inherited methods deal with return values of their own subclass? (greggzst)
* It's been a long time and matz said it was to be fixed inn 3.0 -
- * [Feature #16740] Deprecating and removing the broken `Process.clock_getres` (jeremyevans0)
* Is it OK to deprecate `Process.clock_getres` in 3.0 and remove it in 3.1? -
03:15 AM Misc #16693: DevelopersMeeting20200410Japan
- * [Feature #16688] Allow `#to_path` object as argument to `system()` (Dan0042)
* `system`/`exec` should be compatible with `Pathname` objects -
-
- I think this proposal has been realized under the name `Enumerable#filter_map` following the duplicate proposal #15323. So it should be closed.
-
- Includes a duplicate of #14788.
-
-
- The crash happens in `rb_mysql_connect` in `mysql2.bundle`. You should probably file an issue on their GitHub: https://github.com/brianmario/mysql2/issues
-
02:14 AM Bug #16758 (Third Party's Issue): Unable to run task that require bundle exec
- I'm trying to run migrations or start my server on OSX. I was previously using an Ubuntu vagrant box. Anytime I try it crashes.
Attached is the crash report. I've also created a gist
https://gist.github.com/antarr/e4789657a9ca9dad3... -
- Yes, this uses `RUBY_PAGER` if it is defined, or `PAGER` if defined.
The pager is used if stdin and stdout are tty both, redirecting stdin stops it.
```shell
ruby --help <&-
```
```shell
ruby --help </dev/null
```
And now empty `PAGER`... -
- Too much, no longer needed output.
-
- On platform where searchs shared libraries by `PATH` environment
variable (i.e., Windows), the shared library is installed in
`bindir`. On other platforms, the library directory is directed
by `libdirname` indirectly. -
- This reverts commit 34b0a7be0ed2fd4ca4d1d509a22964b5e61dfe34.
04/03/2020
-
- > You can filter the warnings by overriding Warning.warn, or using the warning gem:
require 'warning'
Warning.ignore(/`(els)?if' at the end of line without an expression/)
That's pretty cool! I did not know that was possible in... -
- As you can see from the commit message in commit:ba35c14325ebbf1da8f200df83c45ee9937ff8a1, this is a new feature and expected behavior in Ruby 2.7, though it may still be considered experimental.
You can filter the warnings by overrid... -
02:47 PM Feature #16755 (Open): warning: `if' at the end of line without an expression
- I'm using this notation in a lot of scripts:
``` ruby
if
true
then
puts :a
else
puts :b
end
```
Using ruby 2.6.5 I'm getting expected response:
``` shell
(0) 2.6.5 mpapis@mpapis-linux:~/tmp>ruby -w test.rb
a
```
... -
- Should be fixed by https://github.com/ruby/spec/commit/7289ea3e3c22fb41b58d072e966420845cca6e9e
-
- I'm unsure how to resolve this.
This means on Linux we have `File.respond_to?(:lchmod) == true` but it's basically unusable for its main purpose.
I guess we'll have to do similar as https://github.com/ruby/ruby/commit/a19228f878d955eaf2... -
05:01 PM Bug #16756: File.chmod does not work on links.
- This is deliberate because Linux does not support changing the mode of a symbolic link (except for file systems with bugs). The error code ENOTSUP is mandated by POSIX for this case. You can still use lchmod to avoid following symbolic l...
-
- I have reported this against glibc in Fedora:
https://bugzilla.redhat.com/show_bug.cgi?id=1820722 -
- Now I observe different error after fix of #16749:
~~~
1)
File.lchmod changes the file mode of the link and not of the file ERROR
Errno::ENOTSUP: Operation not supported @ apply2files - /builddir/build/BUILD/ruby-2.8.0-810d66f3e7... -
-
-
09:29 PM Feature #16757 (Open): Add intersection to Range
- It would be great if there was a way to do an intersection with ranges. I wrote a method recently to solve this but it gets complicated and doesn't solve for all edge cases.
The example I was using it for was getting the intersection of... -
08:44 PM Bug #14413: `-n` and `-p` flags break when stdout is closed
- For anyone curious, these are the examples I went with instead: https://twitter.com/josh_cheek/status/1246175226213404672
-
- I saw flipflop was readded and was wanting to explain to people how to use it. I feel like if people knew how to use it, they wouldn't have wanted to remove it. So, I made this example (shell is fish, not sure if this example would work ...
-
06:35 PM Bug #16148: bugs.ruby-lang.org is not sending email notifications for watched issues
- Evidently the redmine_bugs_ruby_lang plugin redirects notifications only to the mailing list.
Is this what you actually want? To deny bug reporters a chance to participate even in their own issues? -
- Well, we don't have PPC64 just PPC64LE on Fedora. I have run 5 builds and all passed just fine. Nevertheless, this is what glibc maintainers said about the issue [1]:
>(In reply to Florian Weimer from comment #17)
> ...
IOW, I don't ... - 03:36 PM Revision 6be57cfc (git): * 2020-04-04 [ci skip]
-
03:35 PM Revision f70ca6a1 (git): Remove unnecessary escape
-
- Oh... I just noticed in the code, I think that is the environment variable?
My C knowledge is so bad, but I guess `getenv()` indicates that. :D
```c
const char *pager_env = getenv("RUBY_PAGER");
``` -
- I think the basic idea for the suggestion is fine.
There may be cases where the ruby user may prefer to not use (or depend) on a page
and may be fine with a long message output from --help. This may also depend on
the terminal and s... -
- The help message by `--help` is getting longer and longer, and it often exceeds usual terminal windows.
How about to show the message via pager when running on a tty?
And although the message is sectioned, they are less remarkable in... -
- https://github.com/ruby/ruby/pull/2998
-
07:50 AM Bug #16753 (Closed): ruby -run -e httpd . -p 8080 , not support CJK directorys .
- httpd not support CJK directorys .
-
-
- @nobu implemented `def foo(arg) = expression` in one night. Also, it allows `private def: foo = 42`. Perfect.
https://github.com/ruby/ruby/pull/2996
@matz We are ready to perform the experiment. -
- [ci skip]
-
- [ci skip]
-
- [ci skip]
-
-
- By using spec/mspec/tool/remove_old_guards.rb.
-
- * Allow spaces around `...`
* Matches quotes more precisely
* Matches 0 teeny -
- I somewhat agree with sawa here.
This may still leave the issue over whether one may prefer a single line/method
call, but I believe this may largely depend on the use case more (if there is
one). I can understand it to some extent ...
04/02/2020
-
10:21 PM Feature #16746: Endless method definition
- Have you considered adopting Python's whitespace sensitive indentation?
```ruby
def hello(name):
puts("Hello, #{ name }")
hello("endless Ruby") #=> Hello, endless Ruby
```
-
- https://github.com/ruby/ruby/pull/2996
-
03:28 PM Feature #16746: Endless method definition
- shevegen (Robert A. Heiler) wrote in #note-1:
> > I'm paranoid that the ends end Ruby. I hope Ruby is endless.
Like - basically - a snake biting its own tail amirite ;-) -
03:14 PM Feature #16746: Endless method definition
- matz (Yukihiro Matsumoto) wrote in #note-7:
> I totally agree with the idea **seriously**, (far better than #5054, #5065 and #12241) but don't like the syntax. First I thought
> ...
How about this:
```python
def foo(a):
expression
``` -
- Fixed with da05c1552ee519d8f180b48d97148d28501acb35, thanks for the report.
-
-
-
07:28 PM Feature #16752: :private param for const_set
- It might be a little helpful *only* if we are defining constants dynamically.
-
- To be precise, actually,
```ruby
class A
FOO = 1
private_constant :FOO
end
```
is not two method calls; it is one assignment and one method call.
Besides that, I don't see that writing:
```ruby
class A
const_se... -
- Defining a private constant should not require two method calls.
You want to encourage private declarations because they communicate intent better and are easier to refactor, two statements discourage it.
Ideally there should be co... -
-
-
- 04:08 PM Revision 061eba02 (git): * 2020-04-03 [ci skip]
-
03:49 PM Revision 0f57d66f (git): webrick/ssl: More keyUsage for self-signed certs
- Chrome 75+ started to strictly enforce X.509 keyUsage against TLS server
certificates. Webrick supports generating instant self-signed
certificates for debugging purpose and these certificates lacks required
keyUsage for modern TLS. So a... -
-
- No tags in Redmine now.
-
- Looks like I'm not able to add tags so... April fools :-)
-
04:01 AM Misc #16750: Change typedef of VALUE for better type checking
- The current opaque approach has several reasons:
* portability
The bit layout of bit fields in `struct` is somewhat implementation dependant, especially mixed with endianness. The `union` definition in OP may not work on big-endian... -
- FWIW ruby internally has `struct RVALUE` inside of gc.c, which is something similar to what is proposed in this ticket. It is at lest intentional to avoid the approach. The intention itself is not necessarily clear to me, though.
-
-
- https://github.com/ruby/etc/commit/25c538120c
-
- This reverts commit 69f698d85ada116db5747873e53c597991bedff8.
Again. -
- This reverts commit 5885550c8cb80fb8784598d3af1d90823f4f6604.
It didn't work without `--enable-shared`. - 12:33 PM Revision 736695ce (git): * 2020-04-02 [ci skip]
-
-
- https://github.com/ruby/etc/commit/25c538120c
-
- If this is possible by `-rbundle/setup`, bundler could try the benefit without changing ruby implementations.
And then, if it is significant but the option is too long, let's add the new option.