Activity
From 06/08/2021 to 06/14/2021
06/14/2021
-
11:21 PM Feature #17950: Unable to pattern-match against a String key
- The fact that only symbol keys are supported is documented (https://docs.ruby-lang.org/en/master/doc/syntax/pattern_matching_rdoc.html#label-Patterns), so this is expected and not a bug. Switching to feature request.
-
09:17 PM Bug #17984: [BUG] try to mark T_NONE object
- Can you send me a core file and I can take a look?
-
09:50 AM Bug #17984 (Closed): [BUG] try to mark T_NONE object
- About 1% of our CI jobs fail with this error.
If necessary I can try to find a repro, but these GC issues are hard to pin down because they require GC.stress which is very slow.
```
<OBJ_INFO:gc_mark_ptr@gc.c:6687> 0x00007f6717407... -
06:11 PM Bug #17985 (Closed): Ractor + GC: [BUG] not single ractor mode
- The attached script (named `bug17882-gcmod.rb`) is a modification of `bootstraptest/test_ractor.rb:224` that reads:
```ruby
def test n
# (...snip...)
end
GC.disable # newly added
p 30.times.map{|i|
test i
}
GC.start # ... - 04:28 PM Revision a09ddfc4 (git): [lib/ostruct] Fix YAML test
- 04:28 PM Revision 52369fc5 (git): [lib/ostruct] Allow overriding of `block_given?`
- 03:04 PM Revision fc98602e (git): * 2021-06-15 [ci skip]
-
03:03 PM Revision edbe0e22 (git): Suppress exception report in inner thread
-
-
- https://github.com/ruby/ostruct/commit/bb253be3e9
-
- 01:53 PM Revision 931ea7cf (git): Add fallback block to `OpenStruct#delete_field` (#1409)
-
01:37 PM Revision 90cad6e1 (git): prefer cc/gcc over clang on solaris
- requested by tankf33der at https://bugs.ruby-lang.org/issues/17949#change-92430
-
- These methods are not !-suffixed, and the messages were very
confusing. - 10:09 AM Revision 1ab71a5b (git): Fix fiber scheduler address resolve solaris tests
- 10:07 AM Revision 25921fe1 (git): Revert "Suppress gcc11 clobbered warning"
- This reverts commit f0f9e77b65990001bd2acb42e1c6b673f6324425.
-
08:35 AM Bug #17951: Collisions in Proc#hash values for blocks defined at the same line
- It was there since 1.9 as far as I can see https://github.com/ruby/ruby/commit/a3e1b1ce7ed7e7ffac23015fc2fde56511b30681#diff-2672918174f926386106967d117f11da8aa1905772dcf48fce53694386e4a666R658-R668
-
- A possible fix:
```diff
--- a/proc.c
+++ b/proc.c
@@ -1451,7 +1451,7 @@ rb_hash_proc(st_index_t hash, VALUE prc)
GetProcPtr(prc, proc);
hash = rb_hash_uint(hash, (st_index_t)proc->block.as.captured.code.val);
hash... - 05:56 AM Revision 050a8954 (git): Wake up join list within thread EC context. (#4471)
- * Wake up join list within thread EC context.
* Consume items from join list so that they are not re-executed.
If `rb_fiber_scheduler_unblock` raises an exception, it can result in a
segfault if `rb_threadptr_join_list_wakeup` is not w... -
-
-
- 04:21 AM Revision 2792acc8 (git): Add scheduler hook `Addrinfo.getaddrinfo`. (#4375)
- Co-authored-by: Bruno Sutic <code@brunosutic.com>
-
01:42 AM Bug #17947 (Closed): [SOLARIS11, SPARC] time.c compilation failed
- Applied in changeset commit:git|688b217706546c2bc9a0926de246dc29d0935261.
----------
time.c: Check if defined(RUBY_MSVCRT_VERSION) to build on Solaris
Fixes [Bug #17947] -
- Fixes [Bug #17947]
- 01:09 AM Revision 09ea2bb0 (git): * 2021-06-14 [ci skip]
-
- ```
p RubyVM::AbstractSyntaxTree.parse("::Foo += 1").children
#=> before: [[], nil, (OP_CDECL@1:0-1:10 (COLON3@1:0-1:10 :Foo) :+ (LIT@1:9-1:10 1))]
#=> after: [[], nil, (OP_CDECL@1:0-1:10 (COLON3@1:0-1:5 :Foo) :+ (LIT@1:9-1:10 1))]
```
06/13/2021
-
- A separate ticket for `-Wparentheses` warning issue: #17946
-
- Or should `extconf.rb` be improved to notice deprecation?
```
../../../ext/date/date_core.c:7803:17: warning: 'timezone' is deprecated: Only provided for source compatibility; this variable might not always be accurate when linking t... -
- `_WIN64` is not defined on i386-mingw32 + ucrt.
-
07:49 AM Feature #17370: Introduce non-blocking `Addrinfo.getaddrinfo` and related methods.
- Here is PR.
https://github.com/ruby/ruby/pull/4375 -
- I ran that patch on our CI, and confirm that the crash is entirely gone. Thank you!
-
06:38 AM Bug #17949: [SOLARIS11, SPARC, SEGFAULT] generating encdb.h
- Please disable default clang-llvm selection by default on Solaris.
All tests above passed with `CC=gcc ./configure --enable-shared --disable-rpath`. -
- You were right, this is llvm's issue.
Ruby successfully passed all tests under GCC on Solaris SPARC.
```
...
...
PASS all 1488 tests
exec ./miniruby -I./lib -I. -I.ext/common ./tool/runruby.rb --extout=.ext -- --disable-gems "./b... -
-
- As well as 2366c681166a1dab95de6b9ca8ffcaae18aadd39.
-
-
- jemalloc (5.2.1 at least) cannot compile in C++ on macOS SDK, due
to conflicts on exception specification. -
05:09 AM Bug #17948: Backport d74e5d5b4fba41a9120b3ed2762cf765478605ad (Crash more nicely when the VM isn't fully set up)
- ruby_3_0 f0a1c063490d69f0868876deed74ad52db63cf4b merged revision(s) d74e5d5b4fba41a9120b3ed2762cf765478605ad.
-
- Crash more nicely when the VM isn't fully set up
If we crash but the VM isn't fully alive, we can get an infinite loop.
---
vm_dump.c | 62 ++++++++++++++++++++++++++++++++------------------------------
... - 04:45 AM Revision cba92b32 (git): * 2021-06-13 [ci skip]
-
04:44 AM Revision d54f74a4 (git): Refactor rb_block_call function
- rb_block_call and rb_block_call_kw have similar code.
So, using rb_block_kw function in rb_block_call function for refactoring.
06/12/2021
-
- ```ruby
require 'set'
def capture(&block)
block
end
# it creates 1k of same blocks
blocks = Array.new(1000) { capture { :foo } }
hashes = blocks.map(&:hash).uniq
ids = blocks.map(&:object_id).uniq
equality = blocks.map {... -
- From #17947 I guess that Clang/LLVM is used as the compiler. There are some possible workarounds that I can think of, from my experience with Clang/LLVM and `setjmp`/`longjmp`.
* `__builtin_setjmp` is enabled for `sparc-sun-solaris` ... -
- The simplest way to crash miniruby is exec and press Ctrl-D inside it as input.
-
- The same latest Solaris 11 and ruby, next crash while compiling:
```
$ gmake
...
...
compiling time.c
compiling version.c
compiling vm.c
compiling vm_trace.c
compiling coroutine/ucontext/Context.c
compiling enc/trans/newline.c
... -
03:36 PM Feature #17950 (Open): Unable to pattern-match against a String key
- I'm unable to parse against an internal hash, when the internal hash contains strings as keys:
```ruby
case {status: 200, headers: {"content-type" => "application/json"}, body: "bla"}
in { status: , headers: {"content-type" => type}... -
- Confirmed. Fixed.
-
- Can you confirm that the following patch fixes the issue?
```diff
diff --git a/time.c b/time.c
index a5c00e2aa2..a7ef9d2f47 100644
--- a/time.c
+++ b/time.c
@@ -1623,7 +1623,7 @@ localtime_with_gmtoff_zone(const time_t *t, struct... -
- I can not imagine how llvm-clang ecosystem can mess define values as subject of this report.
-
- I suppose that the Ruby build system is not maintained for the combination of Solaris + Clang/LLVM. You may want to use Sun C or GCC instead unless intended to use Clang specifically. Of course patches are always welcome.
-
- Compilation failed on file time.c, added line to skip error helped to see later compilation progress.
`miniruby` compiled and crashing, Solaris journey will be long.
```
#define RUBY_MSVCRT_VERSION 0
```
Latest ruby cant be comp... - 11:17 AM Revision 146721ef (git): * 2021-06-12 [ci skip]
-
-
- d74e5d5b4fba41a9120b3ed2762cf765478605ad should be backported into ruby_2_7/ruby_3_0 branches.
06/11/2021
-
06:49 PM Bug #17880: [BUG] We are killing the stack canary set by `opt_setinlinecache`
- Aaron and I worked on a fix today: https://github.com/ruby/ruby/pull/4565
We simplified the script and were able to reproduce locally before the change and unable to reproduce after. -
- I re-ran the benchmarks. I accidentally switched them when I pasted them into here.
Before: Rails 6.1 / Ruby master
```
ruby 3.1.0dev (2021-06-04T00:24:57Z master 91c542ad05) [x86_64-darwin19]
-- create_table(:posts, {:force=>tru... -
03:41 PM Bug #17933: `Net::HTTP#write_timeout` doesn't work with `body_stream`
- I've created a pull request in the [Ruby Net::HTTP Github project](https://github.com/ruby/net-http) with the proposed solution: https://github.com/ruby/net-http/pull/27
-
10:01 AM Feature #17930: Add column information into error backtrace
- mame (Yusuke Endoh) wrote in #note-14:
> If needed, we can implement `Thread::Backtrace::Location#code_location` by using `AST.of`.
> ...
I think it is good to add, as basically `Thread::Backtrace::Location`-backtraces (caller_location... -
- Pattern matching for `target_os` in configure script should be permissive if we consider suffixing something onto "mingw32".
-
- A warning in [win32/win32.c:2600](https://github.com/ruby/ruby/blob/9210f8df7ff23f06c03dedf9f89fd8ab902043e3/win32/win32.c#L2600):
```
../win32/win32.c:2606:88: warning: & has lower precedence than ==; == will be evaluated first [-Wp...
06/10/2021
- 11:31 PM Revision 9210f8df (git): * 2021-06-11 [ci skip]
-
- If we crash but the VM isn't fully alive, we can get an infinite loop.
-
05:35 PM Bug #17945 (Closed): Date::Infinity comparison <=> with Float::INFINITY not symmetric
- **Example**
``` ruby
require 'date'
p Float::INFINITY <=> Date::Infinity.new
p Date::Infinity.new <=> Float::INFINITY
```
**Expected Result**
I expected the comparison to have the same result when the order of values is switched... -
02:59 PM Revision 929cc615 (git): Finish GC before calling gc_set_initial_pages
- If we are during incremental sweeping when calling gc_set_initial_pages
there is an assertion error. The following patch will artificially
produce the bug:
```
diff --git a/gc.c b/gc.c
index c3157dbe2c..d7282cf8f0 100644
--- a/gc.c
+++ ... -
12:09 PM Misc #17936: Run Travis at push on forked repository
- > I sent the PR here.
> ...
The PR was merged. I expect that now you can run Travis on any branches on your forked repo without changing `.travis.yml` as well as GitHub Actions.
-
08:26 AM Feature #17944 (Open): Remove Socket.gethostbyaddr and Socket.gethostbyname
- It is marked as unsafe by our static analysis. Can it be removed now?
It was deprecated in:
https://bugs.ruby-lang.org/projects/ruby-master/repository/trunk/revisions/60266
https://bugs.ruby-lang.org/issues/13097 -
08:15 AM Feature #10320: require into module
- texpert (Aurel Branzeanu) wrote in #note-11:
> And another similar gem is modulation ( https://rubygems.org/gems/modulation/versions/0.25 )
It looks much more developed and actively maintained! Thank you! :) -
06:18 AM Feature #10320: require into module
- jaesharp (J Lynn) wrote in #note-10:
> I'd like to note that there exists a gem called modules ( https://rubygems.org/gems/modules ) which uses `Kernel#load` with the `wrap=true` option in order to implement a module import/export resol... -
- I'd like to note that there exists a gem called modules ( https://rubygems.org/gems/modules ) which uses `Kernel#load` with the `wrap=true` option in order to implement a module import/export resolution system similar in nature to the on...
-
07:53 AM Revision d281fc52 (git): Add missing dependencies
- https://github.com/ruby/ruby/runs/2791163586?check_suite_focus=true
-
- https://ci.appveyor.com/project/ruby/ruby/builds/39542385/job/8b7aq951f9t01x4x
-
- by a race condition by multiple Ractors.
Atmically incrementing body->total_calls may have its own cost, so for
now we intentionally leave the unreliable total_calls. So we allow an
ISeq to be never pushed when you use multiple Ractors.... -
- [ruby/net-ftp] Bump version to 0.1.2
https://github.com/ruby/net-ftp/commit/895ba44b3c
---
lib/net/ftp.rb | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
- test/net/ftp/test_ftp.rb - fix intermittent MinGW failure
Fixes intermittent error as below:
[242/838] 5316=test_ftp
#<Thread:0x0000020aa8733f20 D:/a/ruby/ruby/src/test/net/ftp/test_ftp.rb:2532 run> terminated w... -
- test/net/ftp/test_ftp.rb: remove unused variable
---
test/net/ftp/test_ftp.rb | 1 -
1 file changed, 1 deletion(-) -
- [ruby/net-ftp] Replace "iff" with "if and only if"
iff means if and only if, but readers without that knowledge might
assume this to be a spelling mistake. To me, this seems like
exclusionary language that is unn... -
- [ruby/net-ftp] Close the passive connection data socket if there is
an error setting up the transfer
Previously, the connection leaked in this case. This uses
begin/ensure and checking for an error in the ensur... -
- [ruby/net-ftp] Replace Timeout.timeout with socket timeout
Timeout.timeout is inefficient since it spins up a new thread for
each invocation, use Socket.tcp's connect_timeout option instead
when we aren't using S... -
-
- If needed, we can implement `Thread::Backtrace::Location#code_location` by using `AST.of`.
```
class Thread::Backtrace::Location
def code_location
node = RubyVM::AbstractSyntaxTree.of(self)
[
node.first_lineno,
... -
- ruby_3_0 d4f4ae599f650feb0a9da6fa486ce425dbe8a81a merged revision(s) 73f9831a57e133ab5645221df862a176a42f6685.
-
- POSIX timer cannot be shared in forked process [Bug #17941]
---
thread_pthread.c | 1 +
1 file changed, 1 insertion(+)
06/09/2021
-
11:27 PM Feature #17930: Add column information into error backtrace
- > Another thought: for the case of NoMethodError, maybe showing the arguments would be enough to differentiate in most cases which of the two calls with the same name was problematic?
> ...
I don't think it is ideal. If you chain `#map` ... -
- Another thought: for the case of NoMethodError, maybe showing the arguments would be enough to differentiate in most cases which of the two calls with the same name was problematic?
`NoMethodError#args` already exists, so it would be th... -
- > I'll try to implement it again in prelude.rb, or think of creating a private, MRI-specific default gem.
Why not exposing column information on Thread::Backtrace::Location?
That would be a proper public API.
Given that Thread::Back... -
- > did_you_mean is a public gem, so I think it should avoid using such a internal, unstable (the AST will evolve and differs per Ruby implementation), and MRI-specific API.
That makes sense.
First I tried to implement the feature in... -
- Move the non-heap related configurations to gc_marks_start.
-
- Applied in changeset commit:git|f7741cda323e759432f34c954d7da0ac3d865ab1.
----------
Enable Travis on the specific branches or forked repositories.
Align the Travis enabling timing with GitHub Actions.
For the syntax, see <https://doc... - 03:17 PM Revision fa340dd9 (git): * 2021-06-10 [ci skip]
- 03:16 PM Revision f7741cda (git): Enable Travis on the specific branches or forked repositories.
- Align the Travis enabling timing with GitHub Actions.
For the syntax, see <https://docs.travis-ci.com/user/conditions-v1>.
We use `repo` syntax rather than `fork = true/false` syntax to show a general usage in any repositories on GitHub... -
-
- FTR commit:e03bf76b312e9cba0e44488fcd8f67099af816cb by user:nobu should solve this in a more general way. I think we can close this now.
-
- Reordered iseq_inline_constant_cache_entry members not to exceed
the size of RValue. -
- Use MACHINE for x86 CPU family, ARCH is CPU model name (i386) and
cannot be x86. -
08:50 AM Bug #17943: Segmentation fault in dispatch()
- mame (Yusuke Endoh) wrote in #note-1:
> It is a static function in protobuf. Please contact on them first.
Thanks for you analysis and quick response. I submitted https://github.com/protocolbuffers/protobuf/issues/8716.
-
- It is a static function in protobuf. Please contact on them first.
https://github.com/protocolbuffers/protobuf/blob/v3.11.0/ruby/ext/google/protobuf_c/upb.c#L7015 -
- Ruby is crashing due to a segfault in dispatch() while decoding a protobuf message on s390 platform. This occurred with Ruby 2.4.6 and reproduces with 2.7.3.
I believe this is in Ruby, since `nm -D /opt/rbenv/versions/2.7.3/lib/ruby/g... -
- You should probably read @matz's response to a previous request for instance variable parameters: https://bugs.ruby-lang.org/issues/8563#note-3 (which he confirmed had not changed as of 2017: https://bugs.ruby-lang.org/issues/8563#note-1...
-
- FYI: This is (partially) a duplicate of #5825, #8563, #12023, #12578, #12820, and #15192.
-
04:55 AM Feature #17942 (Open): Add a `initialize(public @a, private @b)` shortcut syntax for defining public/private accessors for instance vars as part of constructor
- This proposal builds on the proposed `initialize(@a, @b)` instance var assignment shortcut syntax described in #15192.
1. It allows you to add an *optional* `public`/`protected`/`private` modifier before any instance var parameter. Do... -
- ## Same feature in TypeScript
It's worth mentioning that other languages have a shortcut for assignment var assignment directly from constructor parameters. So it seems especially painful that Ruby, despite being *so* beautifully elegan... -
- Applied in changeset commit:git|73f9831a57e133ab5645221df862a176a42f6685.
----------
POSIX timer cannot be shared in forked process [Bug #17941] -
-
- A patch is attached for the configure script to set `rb_cv_msvcrt=ucrt` and `RT_VER=140` when UCRT is used.
-
12:31 AM Feature #17763: Implement cache for cvars
- Thank you for the benchmark.
How about that?
>> We also measured database queries in Rails and with the cvar cache they are about ~9% faster.
> ...
-
12:08 AM Feature #17785: Allow named parameters to be keywords
- This feels related to [this proposal](https://bugs.ruby-lang.org/issues/16460) I submitted ~1.5 years ago concerning external/internal names for keyword parameters as it would solve this problem somewhat more elegantly:
```ruby
# Onl...
06/08/2021
-
- I have not seen this issue for a long time so it must be fixed.
-
- I agree we should introduce an abstract rich model for this.
https://bugs.ruby-lang.org/issues/13383 is a similar problem.
It makes sense to me to introduce some kind of `Backtrace::Location#source_location`.
We should stop thin... -
- mame (Yusuke Endoh) wrote in #note-5:
> In fact there is no enough information because `Thread::Backtrace::Location` only knows path and node_id.
This seems an implementation detail, not something that should decide what the API is.
... -
- Ahh, I've tried but I found that the current prototype draws an underline from "the end column of the receiver of the call" to "the end column of the call", so four integers are clearly not enough. Hardcoding such a location looks very a...
-
- I've created a PR for did_you_mean: https://github.com/ruby/did_you_mean/pull/151
Eregon (Benoit Daloze) wrote in #note-4:
> There is no point to use `RubyVM::AbstractSyntaxTree` when all the information is already in the `Thread::Ba... -
- > I talked with ko1 (Koichi Sasada), and we agreed that allowing RubyVM::AbstractSyntaxTree.of(Thread::Backtrace::Location) is a good approach to allow this feature.
I strongly disagree, RubyVM is a CRuby-specific API.
This effective... -
- Thank you @matz .
I talked with @ko1, and we agreed that allowing `RubyVM::AbstractSyntaxTree.of(Thread::Backtrace::Location)` is a good approach to allow this feature.
```ruby
def target
RubyVM::AbstractSyntaxTree.of(caller_lo... -
-
05:20 PM Bug #17767 (Closed): `Cloned ENV` inconsistently returns `ENV` or `self`
- Applied in changeset commit:git|117310bdc00236c0a7676616ce25b5106775dabc.
----------
Make ENV.clone warn and ENV.dup raise
ENV.dup returned a plain Object, since all of ENV's behavior is
defined in ENV's singleton class. So using dup ... - 05:19 PM Revision f9e13cda (git): * 2021-06-09 [ci skip]
-
- ENV.dup returned a plain Object, since all of ENV's behavior is
defined in ENV's singleton class. So using dup makes no sense.
ENV.clone works and is used in some gems, but it doesn't do what
the user expects, since modifying ENV.clone...