Activity
From 05/18/2021 to 05/24/2021
05/24/2021
- 11:03 PM Revision 274d0aa7 (git): * 2021-05-25 [ci skip]
-
11:02 PM Revision 45ddefb1 (git): add rb_id2str to lldb debugging scripts
-
08:00 PM Bug #17885: require_relative and require should be compatible with each other when symlinked files are used
- I totally support this. And there is more to this #16978
-
04:40 PM Bug #17885: require_relative and require should be compatible with each other when symlinked files are used
- This causes [an issue with rules_ruby for Bazel](https://github.com/bazelruby/rules_ruby/issues/100). Bazel uses symlinks extensively to implement hermetic build sandboxing.
-
- This is similar to #10222 (which was fixed), but it's with symlinked files. #10222 was with symlinked directories.
If files are symlinked, then mixing `require` and `require_relative` can result in the same file being loaded twice.
... -
07:32 PM Feature #17883: Load bundler/setup earlier to make `bundle exec ruby -r` respect Gemfile
- Hei!
I see how this is inconsistent, and if it was to be implemented from scratch, I think it would make sense to always process options in RUBYOPT first so that CLI args can always override them "Override" can have multiple meanings he... -
05:09 PM Feature #17883: Load bundler/setup earlier to make `bundle exec ruby -r` respect Gemfile
- I'm clearly against changing the order because of compatibility. Also, the current order is somewhat reasonable. In principle, explicit command-line arguments are "stronger" than RUBYOPT. For `RUBYOPT=-W2 ruby -W0`, `-W0` wins; `-W2` is ...
-
02:27 PM Feature #17883: Load bundler/setup earlier to make `bundle exec ruby -r` respect Gemfile
- Instead of adding `BUNDLE_BIN_PATH`, which would be specific to bundler, I would prefer to instead add something more generic such as `RUBYEARLYOPT` or `RUBYPREOPT`, which is just like `RUBYOPT` but processed before command line options.
-
10:25 AM Feature #17883: Load bundler/setup earlier to make `bundle exec ruby -r` respect Gemfile
- Personally I find the order confusing, in many/most command-line executables, direct arguments have precedence over environment variables.
For `-W`, direct arguments actually have precedence, as expected (command line args should alwa... -
- To reproduce the issue, prepare a Gemfile and run `bundle install --path=vendor/bundle`.
```
$ cat Gemfile
source "https://rubygems.org"
gem "activesupport"
$ bundle install --path=vendor/bundle
```
`Kernel#require` respects... -
- This ticket was discussed in the previous dev meeting. The use case is understandable, but @matz was not completely convinced that WeakMap is a suitable solution for that. He said he would reply.
In addition, @ko1 said that it would b... -
05:08 PM Bug #17887 (Closed): Missed constant lookup after prepend
- **Description**
The following shows that the constant lookup from B does not find the constant in the prepended M module. I would expect this lookup to behave like "B.include M" which does print the constant from module M.
**Examp... -
-
- # The next dev meeting
**Date: 2021/06/17 13:00-17:00**
Place/Sign-up/Agenda/Log: https://github.com/ruby/dev-meeting-log/blob/master/DevelopersMeeting20210615Japan.md
- Dev meeting *IS NOT* a decision-making place. All decisions ... -
03:41 PM Feature #17884 (Assigned): locindex for profiling tools
- (MRI internals)
Profiling tools need to record the code location, mainly a pair of file path and line number ("file:line").
To record this pair in 64bit CPU, 8B (VALUE) + 4B (int) = 12B is needed. In general, the number of pairs (fil... -
10:55 AM Feature #17853: Add Thread#thread_id
- Created a proposed implementation:
https://github.com/ruby/ruby/compare/master...nurse:native_thread_id -
09:37 AM Revision 4aaa98e6 (git): xibbar is retired from cgi.rb maintainer
-
- @knu I have no strong opinion to remove `syslog` in Ruby 3.1. I withdraws it from this proposal.
-
03:13 AM Revision f57968e3 (git): Set USE_COLORIZE to the default value in a test
-
02:56 AM Revision d06d4560 (git): [ruby/irb] update test/irb/test_init.rb to avoid useless eval
- Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
https://github.com/ruby/irb/commit/2dfdc031ec -
- When `NO_COLOR` is set to any non-nil value, output is not colorized.
See https://no-color.org/
https://github.com/ruby/irb/commit/401d0916fe - 12:41 AM Revision e16a6429 (git): * 2021-05-24 [ci skip]
-
12:41 AM Revision ce360977 (git): Refactor num_zero_p function (#4522)
05/23/2021
-
10:42 PM Bug #17866: Incompatible changes with Psych 4.0.0
- naruse (Yui NARUSE) wrote in #note-10:
> Since I want Ruby 3.1 to keep as much as compatibility for Ruby 3.0 to ensure application/library developer adopt Ruby 3, I’m negative to go Ruby 3.1 with Psych 4.0.0.
Unfortunately, users wil... -
08:23 PM Bug #17792: make notes and make test fail with Ruby3.0.1p64 RaspberryPI 4B Ubuntu 20.10 ARM64
- Interruption in `copy_stream` is what this test is intended for, and so the result of the test in #note-4 is a success. But taking about 2 minutes to finish is unexpected to me. Possibly due to a bug in the pthread library?
FWIW, the ... -
12:10 PM Feature #17881: Add a Module#const_added callback
- > Would trigger on Foo::Bar = Class.new. (This one being a current limitation with TracePoint if I'm not mistaken)
Correct. Nowadays, you cannot define an explicit namespace with a constant assignment like that ([docs](https://github.co... -
10:38 AM Feature #17881: Add a Module#const_added callback
- > What I had in mind in the conversation in the Zeitwerk ticket was not const_added. It was a callback that gets called when a new class/module is created and after it gets its name set if defined with the class/module keywords.
I see... -
- > If I'm not mistaken, such usage could cause the TracePoint to stay active at runtime, which isn't a huge deal, but will still trigger every time a singleton_class or anonymous class is created, which isn't a good practice at runtime, b...
-
- Oh, and while Zeitwerk can disable the TracePoint, it wouldn't be able to disable `on_new_module(&block)`, at most have a flag.
My personal take is: As far as Zeitwerk is concerned, I'd study some API if core believes we should not use ... -
- Another gotcha I see is that we cannot assume `const_added` is exclusive to Zeitwerk. Users could have their own. So we would need to prepend our own, to make sure it is not rewritten. And on reload, we would need to keep track of which ...
-
08:16 AM Feature #17853: Add Thread#thread_id
- `native_id` sounds better to me. I agree with it.
As for `m:n thread model`, in my opinion, either volatile value or `nil` (this can happen when it's waiting an event in `select/epoll/kqueue` without assigning any thread?) sounds good... -
07:10 AM Bug #17736: Destructive methods inconsistently handle `receiver frozen state` in given block
- ruby_3_0 44b87adc07621b6a8eddfcf4aaff34ce634179d4 merged revision(s) e019dd24df4ed7063ad80d4c2e4070141793f598,7954bb056be30e86c419fe3792064d28990a4999,7d3fdfb27dac456827b004d9e66a44b15f8cd762.
-
- ruby_3_0 5af5ea7f860ed64062796e54e73274e7a56c7280 merged revision(s) cfd162d535c7a4f8b1f95255cc6be696a8b75557.
-
- Make String#{strip,lstrip}{,!} strip leading NUL bytes
The documentation already specifies that they strip whitespace
and defines whitespace to include null.
This wraps the new behavior in the appropriate guards... -
- Ensure the receiver is modifiable before shrinking [Bug #17736]
* Ensure the receiver is modifiable before shinking [Bug #17736]
* Assert the receivers are not modified
---
array.c | 1 ... -
- [ruby/rdoc] Vertical-bar is disallowed in path names on Windows
No risk of remote code execution, when the file cannot be created.
https://github.com/ruby/rdoc/runs/2565343916?check_suite_focus=true#step:5:58
``... -
06:04 AM Bug #17827: Monitor is not fiber safe
- @nagachika thank you so much for your effort.
-
- ruby_3_0 f9196de1dee6f5ab8b6fe115070b92775a3500fe merged revision(s) 3a3b19b2bba49e5d6f1cf13764eb6dd701397be9.
-
- ruby_3_0 fc73fc79cbaee6674256589b7cfe53d20b638a05 merged revision(s) 1ad222477344597038d7ec08885a41f547c2a3b4.
-
- ruby_3_0 410216c9db033a72f4dd119c366bc58dd54383ec merged revision(s) d8a13e504992a45d52063f7c925408d7aad3595a.
-
- [Bug #17780] Fix Method#super_method for module alias
Method#super_method crashes for aliased module methods because they are
not defined on a class. This bug was introduced in
c60aaed1856b2b6f90de0992c3477183001... -
- [ruby/openssl] Fixed the results of
OpenSSL::Timestamp::Response#failure_info
Made stored values `Symbol`s instead of `ID`s.
Fixes https://bugs.ruby-lang.org/issues/17625
Co-Authored-By: xtkoba (Tee KO... -
- Fix Monitor to lock per Fiber, like Mutex [Bug #17827]
---
ext/monitor/monitor.c | 10 +++++-----
test/monitor/test_monitor.rb | 7 +++++++
2 files changed, 12 insertions(+), 5 deletions(-) -
12:52 AM Bug #17584 (Assigned): Segmentation fault with Ruby 3.0
-
- mjit_compile.c:525 of Ruby 3.0.0 is the only hint about this error, but it seems like a race condition with GC of child ISeqs. I guess this could happen because mjit_mark only marks `compiling_iseq`. We probably need to mark children as ...
-
-
- It's unfortunate that `rb_ractor_sharable_p` may call arbitrary methods, but maintaining leafness of opt_setinlinecache is probably not as important as opt_getinlinecache's. We could just set `attr bool leaf = false;` for fixing this. We...
05/22/2021
-
- Option 1 would be useful, might avoid delaying finalizers if the main thread is for a while in native code, and it is already what JRuby (IIRC) and TruffleRuby currently do. There is potentially some incompatibility, but it seems pretty ...
-
- mame (Yusuke Endoh) wrote in #note-14:
> Just FYI. According to @ko1, masking all exceptions in an ensure clause is possible in pure Ruby, as long as you use MRI.
It does not seem to be the case: https://gist.github.com/eregon/9022f5... -
- > If I'm not mistaken, such usage could cause the TracePoint to stay active at runtime, which isn't a huge deal, but will still trigger every time a singleton_class or anonymous class is created, which isn't a good practice at runtime, b...
-
- > If the TP is not needed ever, it is not even enabled. If at some point you needed it, but no longer, Zeitwerk disables it.
True, but what worries me would be gems using Zeitwerk and not fully eagerloading, there's even an explicit A... -
- Let me also add something for context. @byroot knows it and probably @Eregon too, just for anybody else following.
Something the TP on the :class event allows you to do is to enable only if you need it. Allows you to be precise.
If... -
- Thanks a lot for working on this Jean.
If we add this, Zeitwerk would be using a somewhat more "normal" API. TracePoint is correct, but seems a bit off to me conceptually (I do not know if I should revise that perception, maybe should I... -
- Honestly I shouldn't have mentioned current performance at all in the first place, it's really not the main reason we'd like this new API.
However you do have a point that this new API should be at least as performant, otherwise it wi... -
- One potential overhead of the :class TracePoint is that it allows accessing the binding through TracePoint#binding:
```
$ ruby -e 'TracePoint.new(:class) { |tp| p tp; p tp.self; p tp.binding.local_variables }.enable; a=1; class C; b=2;... -
- Ugh, Redmine doesn't handle posting a comment concurrently and now incorrectly reverted the description changes even though I didn't touch the description.
@byroot Could you re-apply your changes? -
- > A JIT that strip tracing code, and simply de-optimize when tracing is enabled doesn't seem that far fetched to me.
Yes, and I think it makes a lot of sense to deoptimize/recompile for e.g., :line and :call events (those have a huge ... -
- > This would trigger for all constants
Indeed, as it seemed the most consistent API to me. But I'm open to an alternative that would only trigger for `class` / `module`. Not sure how it could be named though.
> ...
I haven't measur... -
- This would trigger for all constants, which might affect startup quite a bit by having all those extra calls (I already see it as a problem for `method_added`, but maybe we can optimize for the common case that it's not redefined).
Mayb... -
- ### Use case
Autoloaders like `zeitwerk` need a callback when a new class or module is registered in the constant table.
Currently this is implemented with TracePoint's `:class` event. It works, but it is a bit unfortunate to have ... -
- The attached test code is excerpted from `bootstraptest/test_ractor.rb:224`. This code causes a segmentation fault every time when run on `x86_64-cygwin`. There are at least 3 types of dying messages, as shown below.
I have no idea wh... -
11:37 AM Bug #17878: bootstraptest/test_ractor.rb:224 a random failing test with "The outgoing-port is already closed (Ractor::ClosedError)"
- I am trying to skip the assertion on Travis CI arm64 cases by https://github.com/ruby/ruby/pull/4518 6th commit.
-
- ruby_3_0 31816356eee6313fe968eecd4cd3ad9ac4848819 merged revision(s) 2bc293e899c9d32dcd794a73de8925c49ecf8f15,d0e6c6e682b9ba2b0309a5177933a0628e8ef316,cc0dc67bbbe1951ff90004bc987f78545625d772,e1eff837cf12a8e813de9d4ff2db50c9b68b86b5,0ab0...
-
- cdhash_cmp: can take rational literals
Rational literals are those integers suffixed with `r`. They tend to
be a part of more complex expressions like `123/456r`, but in theory
they can live alone. When such "b... -
- I commented out the faulty assertions in https://bugs.ruby-lang.org/issues/17879 to see if some other bug is lurking. See: https://github.com/Shopify/ruby/commit/f698294cc419e696f1cabd8d05e7576c3e490a7a (directly after current master)
... -
- ruby_3_0 af9de56c6fde7f9adb81d6ed0ef3067af81f90e5 merged revision(s) 110f242ef9b495037f59e4972ee102a8b8b372d5.
-
- Also `\U` after control/meta is invalid [Bug #17861]
As well as `\u`, `\U` should be invalid there too.
And highlight including `u`/`U` not only the backslash before it.
---
parse.y | 12 ... -
- ruby_3_0 5e21726cda22e3cb34127751aec7e9babb4308b3 merged revision(s) 799ea1d1540aca03542ee31647052343507a4b41.
-
- mkmf.rb: convert also arch_hdrdir [Bug #16651]
---
lib/mkmf.rb | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-) -
- 43a9a974e276dc49b03ec81ccace0adb534a3d20 is included in ruby_3_0 branch.
-
- Ah right, I missed it in the output, here you go:
```
<OBJ_INFO:rb_ractor_confirm_belonging@ractor_core.h:327> 0x00007fe2577028f8 [0 U] T_STRING (String)63
``` -
02:01 AM Bug #17879: [BUG] id == 0 but not shareable in `rb_ractor_confirm_belonging`
- That message is here:
```C
if (UNLIKELY(!rb_ractor_shareable_p(obj))) {
rp(obj);
rb_bug("id == 0 but not shareable");
}
```
What was dumped before the message? -
- ruby_3_0 ddd720f8dcd98de6a250a152fa40c3c044d62383 merged revision(s) 5026f9a5d5012248729a0052cd6cec811748291b.
-
- compile.c: stop the jump-jump optimization if the second has any
event
Fixes [Bug #17868]
---
compile.c | 3 ++-
test/ruby/test_settracefunc.rb | 15 +++++++++++++++
... -
- I'd like to retry this kind of error, but showing no backtrace is hard
to deal with.
https://github.com/ruby/ruby/runs/2644908002 -
-
- ruby_3_0 9c0df2e81c22e6e35f3c5d69a070c2a3cf67e320 merged revision(s) 9edc162583a4f685332239f6249745ad9b518cbe.
-
- [ruby/resolv] Fix confusion of received response message
This is a follow up for commit 33fb966197f1 ("Remove sender/message_id
pair after response received in resolv", 2020-09-11).
As the @senders instance vari... -
- ruby_3_0 e175917f66f0b51638d9d9e143918d4109d489f4 merged revision(s) dfe21ef7a15a30fc33f3a5cc089e15d09e905eb4.
-
- ruby_3_0 e175917f66f0b51638d9d9e143918d4109d489f4 merged revision(s) dfe21ef7a15a30fc33f3a5cc089e15d09e905eb4.
-
- Do not block JIT with pending_stale_p
Because we want to flush pending stale units before unloading units, the
pending_stale_p check is implemented in this waiting loop.
However, once all methods are called more... -
- I backported this with some other revisions about lib/rdoc/ and test/rdoc/ to synchronize rdoc bundled in ruby_3_0 branch with upstream repo's v6.3.1.
https://github.com/ruby/ruby/commit/2a70d2148453aeb39443f4c84147ace722d53a9a
-
- Synchronize rdoc with upstream version v6.3.1.
[ruby/rdoc] Use File.open to fix the OS Command Injection
vulnerability in CVE-2021-31799
https://github.com/ruby/rdoc/commit/a7f5d6ab88
---
lib/r... -
- The order of arguments to callback of qsort is not defined.
That means `a` may not be 3 at all. -
01:34 AM Bug #17739 (Closed): Array#sort! changes the order even if the receiver raises FrozenError in given block
- Applied in changeset commit:git|8b00bfb7c2c33827490c78a16c44b102cb0d724b.
----------
Do not allow array modifications after freeze inside sort!
If freezing an array inside sort!, previously the array could be
modified after the freeze.... -
- If freezing an array inside sort!, previously the array could be
modified after the freeze. This checks whether the receiver is
frozen after every yield and potential call to #> or #<,
preventing modifications if the receiver is frozen ...
05/21/2021
-
- Ractor setup is invoked even if it is never used, so it is not weird.
I guess this is a false alarm due to incorrect debugging assertions. It should be fixed of course. -
- I recently re-enabled our ruby-head CI with `-DRUBY_DEBUG`, and we're hitting this bug consistently, and early in the test suite.
What's very weird is that we're not using Ractors at all.
```
/app/lib/active_support/cache/versione... -
- I still believe neither of these proposals (fork callbacks or Kernel.fork delegation) should be accepted. In general, the proposals are for the same feature (wrapping behavior on fork), so in my opinion there is not much difference betwe...
-
- @matz and some committers discussed this ticket for about an hour in today's dev-meeting, but we didn't reach any conclusion again.
Mainly, @akr -san still agrees with @jeremyevans0 and feels a bad smell about the proposed hook API. I... -
- Because I know how busy @ko1 is, I thought I'd save him some work and tried to implement this myself: https://github.com/ruby/ruby/pull/4519
-
06:53 PM Feature #17849: Fix Timeout.timeout so that it can be used in threaded Web servers
- > (2) Some applications can run their program body in ensure clauses.
@ko1 first, thank you for looking at this issue. I agree this is a problem, it is the pathological case to what I was describing by the inability to detect a halti... -
- Just FYI. According to @ko1, masking all exceptions in an ensure clause is possible in pure Ruby, as long as you use MRI.
```
begin
...
ensure
Thread.handle_interrupt(Object => :never) do
..
end
end
```
No interru... -
06:11 AM Feature #17849: Fix Timeout.timeout so that it can be used in threaded Web servers
- I doubt if there's a clear solution.
* `Timeout` uses asynchronous exceptions and it **is** difficult to handle correctly.
* for example, if some calls `Timeout.timeout` from within `ensure` clause directly/indirectly
* or what if t... -
- I sent the PR to revive Travis. The i686 case on GitHub Actions would be another PR later.
https://github.com/ruby/ruby/pull/4518
-
- We discussed this issue in today's dev-meeting. We agreed that, if we can find a good enough threshold, `Regexp#backtrack_limit=` is better than `Regexp#timeout=`. For example,
* the threshold should stop almost all practical Regexps ... -
04:13 PM Revision b2fc592c (git): Build CDHASH properly when loading iseq from binary
- Before this change, CDHASH operands were built as plain hashes when
loaded from binary. Without setting up the hash with the correct
st_table type, the hash can sometimes be an ar_table. When the hash is
an ar_table, lookups can call the... -
- Applied in changeset commit:git|636d4f7eb9f3fcb088e1a44af4181c4aa36789b4.
----------
Avoid setting the visibility of refinement method entries
Since refinement search is always performed, these entries should always
be public. The meth... -
- Applied in changeset commit:git|50a534a1526e2b9f4ea41e44b802bd73f9cebbeb.
----------
ripper: wrap endless method in bodystmt [Bug #17869] - 04:12 PM Revision 0772f1b5 (git): * 2021-05-22 [ci skip]
-
- Since refinement search is always performed, these entries should always
be public. The method entry that the refinement search returns decides
the visibility.
Fixes [Bug #17822] -
02:36 PM Bug #17875: Segfault when reusing a closed Fiber scheduler inside a Thread
- Oh, thanks! I installed 3.1.0-dev and it worked fine for me, too. (That is, it failed with a nice error message and backtrace.) I thought I was on the latest version, but I guess I overlooked the `-dev` version.
I can't figure out how ... -
- Seems already fixed?
```
$ ruby scheduler_bug.rb
#<Thread:0x00007f933e4f7760 scheduler_bug.rb:240 run> terminated with exception (report_on_exception is true):
scheduler_bug.rb:241:in `block in <main>': undefined method `run' for n... -
- I was running Travis several times I am trying to revive based on the master commit: `50a534a1526e2b9f4ea41e44b802bd73f9cebbeb`.
Then I got the following failure on Travis arm64 Ubuntu focal environment. The failure happened for the fir... -
- > FWIW, ActiveRecord::Base.logger no longer uses class variables since https://github.com/rails/rails/commit/dcc2530af74cf6355a9206bb1d0b084a734fae3e
Yes, we've talked about changing it since, it's not really related to this ticket. I... -
- FWIW, `ActiveRecord::Base.logger` no longer uses class variables since
https://github.com/rails/rails/commit/dcc2530af74cf6355a9206bb1d0b084a734fae3e (nice!)
Maybe we should really just deprecate class variables since their semantics... -
-
10:12 AM Feature #17873: Update of default gems in Ruby 3.1
- I don't see a good reason to remove the syslog library from ruby.
It is so small, simple, and almost maintenance-free because the syslog facility is unlikely to change anymore.
Syslog has long been a standard logging facility in Unix... -
- I added dbm and gdbm to removal list with other commiter's opinion.
-
- >I'm not sure about ostruct.
Thanks, I'm also aware this. So, I removed `ostruct` from this proposal. -
05:39 AM Feature #17873: Update of default gems in Ruby 3.1
- I agree about net/ftp and net/imap.
-
04:34 AM Feature #17873: Update of default gems in Ruby 3.1
- Thanks, I agree about `prime`.
I imagine `matrix` users are limited and will be ok. I'm not sure about `ostruct`.
Could we consider to bundle them in next major version but with some warning? If you "update" the gem then the conver... -
- >I would like prime library to remain in the default gems or the bundled gems.
Thanks. I moved prime to the bundled gems list.
And I removed fcntl from this proposal. It's also widely used in other gems. -
- Two concerns.
(1) name should include `native`, `os` or something to represent it is platform value. Maybe `native_id` same as Python's API is enough. Ruby can set internal ID for threads.
(2) In future the thread model can be change... -
08:38 AM Feature #17853: Add Thread#thread_id
- +1 This would be useful for Datadog's [ddtrace](https://rubygems.org/gems/ddtrace) gem as well: currently we need to do quite a bit of [hackery and monkey-patching](https://github.com/DataDog/dd-trace-rb/blob/master/lib/ddtrace/profiling...
-
-
-
-
- 609de71f043e8ba34f22b9993e444e2e5bb05709 fixes the issue by using
`throw` insn if `ensure` is used. However, that patch introduce
additional `throw` even if it is not needed. This patch solves
the issue.
This issue is pointed by @mame. -
- user:NuriYuri #note-5 seems to be a distinct bug, as Ractor is not used during a build process (I believe). Can you create a separate ticket about that?
-
06:58 AM Bug #17864: [BUG] try to mark T_NONE object (in Ractors)
- Hi, I tried to build Ruby with last version of master. I got this result during build (with nmake):
```
<OBJ_INFO:gc_mark_ptr@gc.c:6576> 0x0099df40 [2 M ] T_NONE
./tool/generic_erb.rb:7: [BUG] try to mark T_NONE object
ruby 3.1.0de... -
- Since I want Ruby 3.1 to keep as much as compatibility for Ruby 3.0 to ensure application/library developer adopt Ruby 3, I’m negative to go Ruby 3.1 with Psych 4.0.0.
-
- Accepted. I have small concerns about performance (in microbenchmarks). But I don't think I have another choice.
Matz.
-
- Traditionally, `real` just means being floating-point numbers in Computer Science. Other languages treat infinity and `NaN` in similar manner.
Matz.
-
- https://github.com/ruby/rdoc/commit/f480b970cc
-
- Use test-unit assertions instead of minitest.
https://github.com/ruby/rdoc/commit/d6a6209d7f -
- No risk of remote code execution, when the file cannot be created.
https://github.com/ruby/rdoc/runs/2565343916?check_suite_focus=true#step:5:58
```
Error: test_remove_unparseable_CVE_2021_31799(TestRDocRDoc): Errno::EINVAL: Invalid arg... -
- RDoc seems still supporting outdated 2.4.
https://github.com/ruby/rdoc/runs/2565344070?check_suite_focus=true#step:5:64
```
Error: test_remove_unparseable_CVE_2021_31799(TestRDocRDoc): NoMethodError: undefined method `children' for Dir:... -
- Fixes ruby/rdoc#808
https://github.com/ruby/rdoc/commit/80766fd389 -
- You beat me :) Thanks!
-
- I will close this ticket to trigger backport request notice.
To tell the truth, I already file a backport ticket for the changests ;) https://bugs.ruby-lang.org/issues/17874
Anyway Thank you for your clarify. -
- Please consider backporting commit:dfe21ef7a15a30fc33f3a5cc089e15d09e905eb4 to ruby_3_0. This is a fix for a bug introduced in Ruby 3.0.
-
- I'm neutral for this proposal, but just like to add two comments.
1. The proposed change will break a program that uses `Timeout.timeout` in a finalizer. Such a program may get stuck. I don't think that it is a good program, though.
... -
- In the Ruby repository: https://github.com/ruby/ruby/commit/b1c73f239fe9af97de837331849f55d67c27561e
-
04:45 AM Bug #17877 (Closed): Backport a fix of RDoc for CVE-2021-31799
-
- https://github.com/ruby/rdoc/commit/a7f5d6ab88632b3b482fe10611382ff73d14eed7.diff
-
- https://github.com/ruby/rdoc/commit/9307f932b7
-
- https://github.com/ruby/rdoc/commit/a7f5d6ab88
-
- Applied in changeset commit:git|9edad0df74c6ad39281852cca9793fc7dba5c81f.
----------
Remove short options with argument [Bug #17870]
Remove GNU make `-O` and `-W` options which are short but followed
by an argument, so that `$mflags.se... -
- Remove GNU make `-O` and `-W` options which are short but followed
by an argument, so that `$mflags.set?(?n)` does not return `true`
wrongly. -
- So that no longer disturbed by C comment delimiters.
05/20/2021
-
- This issue comes from Ripper, and is due to the fact that in irb, ripper is being called line-by-line and not all at once. When ripper parses the `s2 <<s1` line, it doesn't know that `s2` is a local variable, so it treats `<<s1` as a he...
-
- Ah, yes, thanks for linking back to that. It sure looks like the same thing. I searched for "fiber scheduler segfault" and didn't find it, but I see it calls it "segv" instead.
Please feel free to close this one if it's not useful (I... -
- The symptom seems similar to that of #17618.
-
- Hi! I've been trying to learn the new `Fiber.scheduler` API. I ran into this segfault when I was misusing scheduler instances in Minitest. I was able to reproduce the bug in isolation and I thought I'd share it here in case it's helpful ...
-
- Applied in changeset commit:git|2e919e4a643814735c68cf6724972cf7aa0cd91e.
----------
Remove extra word in heredoc documentation
From Thibault Jouan
Fixes [Misc #17872] -
05:28 AM Misc #17872 (Closed): Remove extra word from Here Documents syntax doc
- When reading the documentation for here documents literal, for example
when executing `ri ruby:syntax/literals`, or directly the file at
`doc/syntax/literals.rdoc', then looking at "Here Documents" section
there is a sentence that see... - 07:25 PM Revision a7fae2af (git): * 2021-05-21 [ci skip]
-
- From Thibault Jouan
Fixes [Misc #17872] -
- user:mame wrote in #note-4:
> I'm unsure why the fix addressed the issue of #17306.
This is because after #17306 was fixed the commit commit:32b7dcfb56a417c1d1c354102351fc1825d653bf changed the behavior of `{,un}lock_page_body` so th... -
- Running the each test in test/ruby/test_gc_compact.rb on the above environment, here is the result. The failed test is not only the test_ast_compacts test.
* test_enable_autocompact : ok
* test_disable_autocompact : ok
* test_major_... -
- xtkoba (Tee KOBAYASHI) wrote in #note-1:
> Possibly `sysconf(3)` does not report the correct page size?
I think it reports 64k correctly.
```
$ ruby -retc -e'p Etc.sysconf(Etc::SC_PAGE_SIZE)'
65536
```
The fix for #17306 dis... -
03:17 PM Feature #17873: Update of default gems in Ruby 3.1
- I would like `prime` library to remain in the default gems or the bundled gems.
I do competitive programming as a hobby, and I sometimes use it and find it useful.
Since external libraries are usually not available for competitive prog... -
- >In my opinion mutex_m is widely used. Even Rails uses it: https://github.com/rails/rails/search?q=mutex_m
Thanks for your information. I agreed your opinion and removed mutex_m from this proposal. -
- > They are no longer actively maintained and not widely used today.
> ...
In my opinion `mutex_m` is widely used. Even Rails uses it: https://github.com/rails/rails/search?q=mutex_m
-
- >So I guess you mean no longer including them with Ruby, `and gem install` NAME would be needed to use them?
Yes. I meant to remove them ruby repository and release. -
- > And I also propose to extract the following libraries from the default gems.
Those are already default gems.
So I guess you mean no longer including them with Ruby, and `gem install NAME` would be needed to use them?
Seems fine ... -
- I promoted many standard libraries to the default gems. You can see them at https://stdgems.org/#default-gems-ruby-301
I propose the following libraries to promote the bundled gems from the default gems.
```
matrix
prime
net/ftp... -
-
- This is a ticket for backport management.
dfe21ef7a15a30fc33f3a5cc089e15d09e905eb4 should be backported into ruby_3_0. -
- I noticed `-Onone` is also troublesome. Other `-O[type]`'s (type being `target` and `recurse`) seem safe to use.
Proposed patch:
```diff
--- a/tool/rbinstall.rb
+++ b/tool/rbinstall.rb
@@ -113,6 +113,7 @@
$make, *rest = Shellw... -
- * [Feature #17873] Update of default gems in Ruby 3.1
* Does anyone have any objections? -
- Because we want to flush pending stale units before unloading units, the
pending_stale_p check is implemented in this waiting loop.
However, once all methods are called more than --jit-min-calls,
mjit_worker_wakeup will not be signaled ... -
- Applied in changeset commit:git|5026f9a5d5012248729a0052cd6cec811748291b.
----------
compile.c: stop the jump-jump optimization if the second has any event
Fixes [Bug #17868] -
- Thank you for the quick response!
I have confirmed that the original code can be measured with this patch. -
- @ko1 agreed with my patch. I'll commit it soon.
-
- Fixes [Bug #17868]
-
- I think (2) is rare, and IMHO such programs should be fixed.
What would happen is then those programs can't be interrupted (only by killing the process), but then that also becomes a clear motivation to fix those programs so they can be... -
- There are two issues on masking interrupts in `ensure` clauses:
(1) performance: manipulating masks in each `ensure` clauses introduces overhead
(2) Some applications can run their program body in `ensure` clauses.
Before we discussed ... -
- FYI:
```ruby
p Float::INFINITY.real #=> Infinity
p Float::INFINITY.imag #=> 0
p Float::NAN.real #=> NaN
p Float::NAN.imag #=> 0
```
Python and Octave have same behavior. Probably MATLAB does too.
If we change the behavior of `r... -
- FYI: numpy also says inf is a real.
```
>>> import numpy
> ...
True
``` -
- Cloned `ENV` accesses/modifies the process global environment variables, as well as `ENV` itself, however the codes we found expect these are different.
So I propose `ENV.clone` and `ENV.dup` should raise an exception. -
- Fixed by commit:9edc162583a4f685332239f6249745ad9b518cbe.
-
- Fixed by commit:9edc162583a4f685332239f6249745ad9b518cbe.
-
- Thank you the patch. It seems good. Could you commit with a test?
-
- > Your example always allocates a new instance first, then deduplicates it.
Yes, for more complex cases that's kind of the only way. What I'm after here is mostly memory retention, not so much allocations.
> ...
Yes, I used this in... -
- Your example always allocates a new instance first, then deduplicates it.
Why not:
```ruby
def new(x, y, z)
REGISTRY[-"#{self}-#{x},#{y},#{z}"] ||= super
end
```
This would work even in 2.7 too. - 05:35 AM Revision 821e3c12 (git): * 2021-05-20 [ci skip]
-
05/19/2021
-
- +1
I second the proposal.
```ruby
p [1, 2, 3] #=> [1, 2, 3]
p "hello" #=> "hello"
p 1.0 #=> 1.0
p({"x"=>1}) #=> {"x"=>1}
p 1..2 #=> 1..2
p nil #=> nil
p true #=> true
p :name #=> :name
p Array #=> Array
... -
- I had a resistance to upload the full log check.log, because the file size is big. However I would upload the log file now.
```
$ du -sh check.log
1.2M check.log
$ wc -l check.log
22963 check.log
```
-
- Some googling told me that the page size defaults to 64k on powerpc64le for some Linux distros. Possibly `sysconf(3)` does not report the correct page size?
-
- This issue was found by @mame yesterday on our new Power 9 server. I would like to open the ticket.
The test failure was reported and fixed on the #17306 6 months ago.
However on the latest master `adcbae8d49ec04d365ce13274783b1495c3... -
- @byroot For the blocking C extension code without an unblocking function case, I think there is no other way to unstuck it than to kill the entire process.
That's probably a good last measure to have, but obviously it's very costly so s... -
- > Assignee set to matz (Yukihiro Matsumoto)
Maybe we want to tag in Koichi. Last I talked to Matz (EuRuKo in 2019) I believe he suggested having Koichi take a look at the issue. I'm not familiar with the interface here I don't know if... -
- FWIW the work around is to capture the status and re-exec with an exit. It works, but the downside is that shell invocations are not cheap:
``` ruby
irb(main):042:0> a, b, c = Open3.capture3("ls /does/not/exist")
irb(main):043:0> `e... -
- It is reported in https://bugs.gentoo.org/show_bug.cgi?id=790272 that `make --output-sync=line install` installs nearly nothing, and I can confirm this for Ruby 3.0.1p64.
The argument `-Oline` seems to be somehow troublesome in `--mfl... -
04:32 PM Bug #17869 (Closed): `Ripper.sexp`'s S-expression when using endless method definition
- ## Summary
The result of `Ripper.sexp` differs between normal method definition and endless method definition.
### Normal method definition
```ruby
% ruby -rripper -e "pp Ripper.sexp('def m(a) foo end')"
[:program,
[[:def,
... -
- Thank you for your report. This is caused by the peephole optimization.
```
% ruby --dump=i while_in_while.rb
...
0010 jump 23 ( 3)[Li]
...
0033 getlocal_WC_0 ... -
- ```ruby
n = 3
while n > 0
n -= 1 while n > 0
end
```
This `while_in_while.rb` is the file to be measured.
```ruby
require "coverage"
Coverage.start
load "while_in_while.rb"
pp Coverage.result
# {"while_in_while.rb"=>[1, ... -
- This seems to be fixed by backporting commit:ea8f7d4 + `lib/rubygems/openssl.rb` (created in commit:69fa0d6 and changed in commit:4aca77e).
-
- This is reproducible for me with Ruby 2.7.3p183 when passing `--with-baseruby=no` to `./configure`, with OpenSSL header files removed (`mv /usr/include/openssl{,.bak}`).
And this seems to be not reproducible with Ruby 3.0.1p64. Probab... -
-
- https://github.com/flori/json/commit/e2ad91fc2094d3cc2f76adc6c55d420cd06f34d8
-
12:21 AM Bug #17781: Resolv::DNS RequestID table allocations are never freed, causing DNS lookups to eventually hang
- update... https://github.com/ruby/resolv/pull/9/files got merged.
05/18/2021
-
11:52 PM Bug #16475: [Installation process - compiling ruby 2.7.0 from source} "make install" step oddity for ruby 2.7.0 on linux mint (debian base)
- @shevegen: Can this be closed? Looks like nobody else has the same problem.
- 09:53 PM Revision 637683c9 (git): * 2021-05-19 [ci skip]
-
- Since compaction can be concurrent, the machine stack is allowed to
change while compaction is happening. When compaction finishes, there
may be references on the machine stack that need to be reverted so that
we can remove the read bar... -
- byroot (Jean Boussier) wrote in #note-8:
> > I'm ok with enabling aliases by default.
> ...
Ya, that's what I was considering. `safe_load` shouldn't change behavior whatsoever. -
- > I'm ok with enabling aliases by default.
Just to clarify: `load` would have them on by default, but `safe_load` wouldn't, right? -
- Eregon (Benoit Daloze) wrote in #note-4:
> byroot (Jean Boussier) wrote in #note-3:
> ...
No, you can definitely make recursive data structures:
```ruby
irb(main):006:0> x = []
=> []
irb(main):007:0> x << x
=> [[...]]
irb(main)... -
- So indeed I was wrong about the circular reference.
As for Rails I fixed it this morning: https://github.com/rails/rails/commit/179d0a1f474ada02e0030ac3bd062fc653765dbe, but yes, if `load` would allow aliases by default it would make ... -
- > It kinda is, as aliases allow for circular references which can cause some programs to end up in infinite loop.
I thought of the same thing and I tested what Eregon wrote before writing the comment. So, unless you know of a YAML that ... -
- byroot (Jean Boussier) wrote in #note-3:
> It kinda is, as aliases allow for circular references which can cause some programs to end up in infinite loop.
Doesn't the YAML implementation handle/avoid such recursive/circular reference... -
- > and disabling aliases wasn't really important.
It kinda is, as aliases allow for circular references which can cause some programs to end up in infinite loop.
> ...
Maybe we should have `YAML.load_file` use a more relaxed default... -
- I noticed `Psych.unsafe_load` uses `aliases: false` by default. This means that default `config/database.yml` of Rails will stop working if you use Psych 4.0.0.
https://github.com/rails/rails/blob/v6.1.3.2/railties/lib/rails/generators/... -
- NuriYuri (Youri Nouri) wrote in #note-3:
> I've run this `ruby -e"200.times { system('ruby src/IODigester.ytest_copy.rb') }"` to check the probability, 88% success rate with `ruby 3.0.1p64`.
> ...
Can you test with the master branch an... -
- Sorry for my typo.
> So, I plan to not use only free builds without credits on Travis.
So, I plan to use only free builds without credits on Travis.
-
- Let me share the current status and the plan for ruby/ruby's Travis.
> As a side note from my experience, the concept of the credits might be not considered on travis-ci.org.
The travis-ci.org shut down will occur on May 31, 2021.... -
- * type-limits when plain-char is unsigned
* unused-const-variable for NFA constants -
01:16 PM Revision 6b32f1f3 (git): Object whats here (#4503)
- What's Here section for class Object.
-
-
- http://rubyci.s3.amazonaws.com/debian9/ruby-master/log/20210518T093002Z.log.html.gz
```
[ 9396/21186] TestGemCommandsOpenCommand#test_execute-e:1: warning: possibly useless use of a literal in void context
= 0.06 s
``` -
- http://rubyci.s3.amazonaws.com/debian9/ruby-master/log/20210518T093002Z.log.html.gz
```
/home/chkbuild/chkbuild/tmp/build/20210518T093002Z/ruby/test/psych/test_coder.rb:277: warning: ambiguity between regexp and two divisions: wrap regex... -
09:51 AM Revision 6674dd61 (git): Remove ubuntu-16.04 (already commented out)
- https://github.blog/changelog/2021-04-29-github-actions-ubuntu-16-04-lts-virtual-environment-will-be-removed-on-september-20-2021
-
-
-
-
- Unfortunately, Ruby 2.0.0 is EOL today.
-
- I think the ruby 2.0.0p648 is the ruby installed as a RPM package on the CentOS 7 by default.
If you do not have any reason to continue to use it, I would recommend a newer ruby installed by Software Collections or one built from the so... -
- A quick workaround would be not to use such an old version as 2.0.0p648.
-
07:21 AM Bug #17867 (Rejected): split.rb:30: [BUG] Segmentation fault
- Hello,
We hit a "Segmentation fault" of a small ruby script (split.rb).
OS is CentOS7.7 and ruby 2.0.0p648 is used.
# ruby --version
ruby 2.0.0p648 (2015-12-16) [x86_64-linux]
I'd like to know is this is a known issue of ruby ... -
-
- Create a file for test under the temporary directory.
- 03:48 AM Revision 79717f81 (git): [ruby/fiddle] windows: link to ws2_32 for WSAGetLastError()
- https://github.com/ruby/fiddle/commit/e9955d74ae
- 03:48 AM Revision 71d4a493 (git): [ruby/fiddle] windows: add Fiddle.win32_last_socket_error{,=}
- GitHub: fix GH-72
Users can't use WSAGetLastError() with Ruby 3.0 or later because
rb_funcall() resets the last socket error internally.
Users can get the last socket error by Fiddle.win32_last_socket_error.
Reported by Kentaro Hayash... -
- Fixes [Bug #12666]
https://github.com/ruby/fiddle/commit/a267a40be7 - 03:48 AM Revision ab5212b3 (git): [ruby/fiddle] Add support for "const" in type
- GitHub: fix #68
Reported by kojix2. Thanks!!!
https://github.com/ruby/fiddle/commit/d7322c234a - 03:48 AM Revision 4d1bb460 (git): [ruby/fiddle] Add --enable-debug-build option to extconf.rb
- https://github.com/ruby/fiddle/commit/e0498e60ea
- 03:48 AM Revision 791e8eec (git): [ruby/fiddle] win32types: sort
- https://github.com/ruby/fiddle/commit/35dec6c5a5
- 03:48 AM Revision 8758b07b (git): [ruby/fiddle] Fix more Win32Types definitions
- https://docs.microsoft.com/en-us/windows/win32/winprog/windows-data-types
https://github.com/ruby/fiddle/commit/805c1a595a -
03:48 AM Revision 25e56fe3 (git): [ruby/fiddle] Fix Win32Types for Windows 64-bit (#63)
- https://docs.microsoft.com/en-us/windows/win32/winprog/windows-data-types
https://github.com/ruby/fiddle/commit/28ee5b1608 - 03:48 AM Revision 8d63b1dc (git): [ruby/fiddle] Bump version
- https://github.com/ruby/fiddle/commit/0cbd370fd6
- 03:48 AM Revision 881b2dc8 (git): [ruby/fiddle] closure: add support for const char *
- GitHub: fix GH-62
Reported by Cody Krieger. Thanks!!!
https://github.com/ruby/fiddle/commit/284b820f2d - 03:48 AM Revision b2de5999 (git): [ruby/fiddle] closure: accept symbol as type
- https://github.com/ruby/fiddle/commit/dc2da6633e
- 03:48 AM Revision 212d836c (git): [ruby/fiddle] Remove wrong comment
- https://github.com/ruby/fiddle/commit/831522e768
- 03:48 AM Revision 7bd3d6d2 (git): [ruby/fiddle] Bump version
- https://github.com/ruby/fiddle/commit/63e5f98412