[#105104] [Ruby master Bug#18141] Marshal load with proc yield strings before they are fully initialized — "byroot (Jean Boussier)" <noreply@...>

Issue #18141 has been reported by byroot (Jean Boussier).

10 messages 2021/09/01

[#105114] [Ruby master Feature#18143] Add a new method to change GC.stress only in the given block such as GC.with_stress(flag) {...} — "kou (Kouhei Sutou)" <noreply@...>

Issue #18143 has been reported by kou (Kouhei Sutou).

8 messages 2021/09/02

[#105180] [Ruby master Bug#18156] 3.0.2 configuration checks by default for C++ compiler instead of C? — "vo.x (Vit Ondruch)" <noreply@...>

Issue #18156 has been reported by vo.x (Vit Ondruch).

11 messages 2021/09/08

[#105191] [Ruby master Bug#18159] Integrate functionality of dead_end gem into Ruby — duerst <noreply@...>

Issue #18159 has been reported by duerst (Martin Dst).

37 messages 2021/09/11

[#105269] [Ruby master Bug#18169] Local copies of gemified libraries are being released out of sync with their gems — "headius (Charles Nutter)" <noreply@...>

Issue #18169 has been reported by headius (Charles Nutter).

15 messages 2021/09/15

[#105276] [Ruby master Bug#18170] Exception#inspect should not include newlines — "mame (Yusuke Endoh)" <noreply@...>

Issue #18170 has been reported by mame (Yusuke Endoh).

29 messages 2021/09/16

[#105310] [Ruby master Misc#18174] DevelopersMeeting20211021Japan — "mame (Yusuke Endoh)" <noreply@...>

Issue #18174 has been reported by mame (Yusuke Endoh).

14 messages 2021/09/16

[#105313] [Ruby master Misc#18175] Propose Jean Boussier (@byroot) as a core committer — "tenderlovemaking (Aaron Patterson)" <noreply@...>

Issue #18175 has been reported by tenderlovemaking (Aaron Patterson).

11 messages 2021/09/16

[#105354] [Ruby master Feature#18181] Introduce Enumerable#min_with_value, max_with_value, and minmax_with_value — "kyanagi (Kouhei Yanagita)" <noreply@...>

Issue #18181 has been reported by kyanagi (Kouhei Yanagita).

16 messages 2021/09/20

[#105361] [Ruby master Feature#18183] make SecureRandom.choose public — "olleicua (Antha Auciello)" <noreply@...>

Issue #18183 has been reported by olleicua (Antha Auciello).

17 messages 2021/09/21

[#105377] [Ruby master Bug#18187] Float#clamp() returns ArgumentError (comparison of Float with 1 failed) — "SouravGoswami (Sourav Goswami)" <noreply@...>

Issue #18187 has been reported by SouravGoswami (Sourav Goswami).

7 messages 2021/09/22

[#105391] [Ruby master Bug#18189] `rb_cString` can be NULL during `Init_Object` — "ioquatix (Samuel Williams)" <noreply@...>

Issue #18189 has been reported by ioquatix (Samuel Williams).

9 messages 2021/09/23

[#105428] [Ruby master Bug#18194] No easy way to format exception messages per thread/fiber scheduler context. — "ioquatix (Samuel Williams)" <noreply@...>

Issue #18194 has been reported by ioquatix (Samuel Williams).

16 messages 2021/09/26

[#105450] [Ruby master Feature#18228] Add a `timeout` option to `IO.copy_stream` — "byroot (Jean Boussier)" <noreply@...>

Issue #18228 has been reported by byroot (Jean Boussier).

11 messages 2021/09/27

[#105452] [Ruby master Feature#18229] Proposal to merge YJIT — "maximecb (Maxime Chevalier-Boisvert)" <noreply@...>

Issue #18229 has been reported by maximecb (Maxime Chevalier-Boisvert).

21 messages 2021/09/27

[#105500] [Ruby master Feature#18231] `RubyVM.keep_script_lines` — "ko1 (Koichi Sasada)" <noreply@...>

Issue #18231 has been reported by ko1 (Koichi Sasada).

19 messages 2021/09/30

[#105504] [Ruby master Bug#18232] Ractor.make_shareable is broken in code loaded with RubyVM::InstructionSequence.load_from_binary — "byroot (Jean Boussier)" <noreply@...>

Issue #18232 has been reported by byroot (Jean Boussier).

7 messages 2021/09/30

[ruby-core:105402] [Ruby master Bug#17885] require_relative and require should be compatible with each other when symlinked files are used

From: "jeremyevans0 (Jeremy Evans)" <noreply@...>
Date: 2021-09-23 23:28:14 UTC
List: ruby-core #105402
Issue #17885 has been updated by jeremyevans0 (Jeremy Evans).

Status changed from Closed to Open

Reopening this since the patch was reverted.

After multiple previous attempts to fix this were shown to still have the same issue by the AWS Graviton2 gcc 11 CI (thank you @fkorotkov), I was finally able to reproduce the problem locally by calling `Thread.pass` inside `load_lock`.  The issue is a race-condition in `ruby_verbose` checking in `load_lock`.  The simplest way to fix it is to allow silencing the warning when loading encoders and transcoders.  The encoder loading code already implicitly set `ruby_verbose` to false before loading, and back to its original setting after loading.  The transcoder was also being required, and could also trigger the warning.  If the transcoder require happened after the encoder finished, the warning was generated. I think the code was always vulnerable to a race condition, it just took the use of realpath when loading to expose the race condition between the encoder loading and transcoder loading.

To fix this, I adding a `warn` argument to `require_internal` and `load_lock`, and I added `rb_require_internal_silent`, which never issues the warning.  Then I had encoder and transcoder loading use `rb_require_internal_silent` to avoid the warning.

Revised pull request at https://github.com/ruby/ruby/pull/4887

----------------------------------------
Bug #17885: require_relative and require should be compatible with each other when symlinked files are used
https://bugs.ruby-lang.org/issues/17885#change-93817

* Author: john_firebaugh (John Firebaugh)
* Status: Open
* Priority: Normal
* ruby -v: ruby 2.6.6p146 (2020-03-31 revision 67876) [x86_64-darwin19]
* Backport: 2.6: UNKNOWN, 2.7: UNKNOWN, 3.0: UNKNOWN
----------------------------------------
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.

```
mkdir a
echo "require_relative 'b'" > a/a.rb
echo "p 'b loaded'" > a/b.rb
mkdir b
ln -s ../a/a.rb b
ln -s ../a/b.rb b
echo "$: << File.expand_path('../b', __FILE__); require 'b'; require 'a'" > c.rb
ruby c.rb
```

Expected behavior is that this prints "b loaded" once. Actual behavior is that it prints "b loaded" twice.



-- 
https://bugs.ruby-lang.org/

Unsubscribe: <mailto:ruby-core-request@ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>

In This Thread

Prev Next