Activity
From 10/27/2022 to 11/02/2022
11/02/2022
-
10:27 PM Revision 59a6caf8 (git): Run only daily schedule because CodeQL provides a lot of false-positive results for Ruby code
-
08:18 PM Feature #19090: Do not duplicate an unescaped string in CGI.escapeHTML
- We might be able to make it efficient for a non-String buffer too, so I'll give it a try first. Thanks for the idea.
-
- > You know, if speed is that much a concern, I think optimized_escape_html2 should seek the first escapable character and only if found then do the ALLOCV_N (and memcpy from cstr to buf).
`ALLOCV_N` is essentially `alloca`. I'd say this... -
03:02 PM Feature #19090: Do not duplicate an unescaped string in CGI.escapeHTML
- Actually an even better approach may be to append the escaped bytes directly to the final output buffer, instead of generating any intermediary string at all. As such I would like to propose:
`CGI.escapeHTML(str, "")` append to new strin... -
- You know, if speed is *that* much a concern, I think `optimized_escape_html2` should seek the first escapable character and only if found then do the ALLOCV_N (and memcpy from cstr to buf).
-
- > Isn't rb_str_dup copy-on-write and so should be fairly cheap?
As I wrote in the description, calling `rb_str_dup` is making the whole method 1.34x slower no matter how efficient CoW is. Whether the 34% slowdown is cheap or not depends... -
05:34 PM Revision b777408c (git): Fix crash in test runner on timeout
- When a test worker hangs and timeouts, the test runner crashes with the
following stack trace:
ruby/tool/lib/test/unit.rb:1747:in `puke': undefined method `backtrace' for Timeout::Error:Class (NoMethodError)
from ruby/tool/lib/t... -
- * YJIT: Support invokeblock
* Update yjit/src/backend/arm64/mod.rs
* Update yjit/src/codegen.rs
Co-authored-by: Maxime Chevalier-Boisvert <[email protected]> -
04:10 PM Bug #19100 (Closed): Ruby 3 PRNG values diverge from Ruby 2 for some initial values
- The outputs of the Mersenne Twister implementation diverged from their expected results in Ruby 2 for initial pseudo-random number generator seeds between 2^32 and 2^33-1, inclusive. I used ruby versions 2.7.4 vs. 3.0.3, 3.0.4, and 3.1.2...
- 03:16 PM Revision ee7c031d (git): YJIT: don't show a full crash report if mmap is only out of memory (#6659)
-
- Co-Authored-By: Alan Wu <[email protected]>
Co-Authored-By: Maxime Chevalier-Boisvert <[email protected]>
Co-authored-by: Alan Wu <[email protected]>
Co-authored-by: Maxime Chevalier-Boisvert <maxime.c... -
- The ELTS_SHARED flag is generic, so we should prefer to use the flags
specific of the type (STR_SHARED for strings and RARRAY_SHARED_FLAG
for arrays). -
01:11 PM Feature #19089: Load bundler/setup in gem_prelude.rb when "bundle exec" is used
- I agree with @mame that using `--disable-gems` or `--disable-did_you_mean` to workaround this is far from ideal, because these flags are intended to disable the functionality, not to use a different version of it. So I think this should ...
- 10:31 AM Revision 875b7b33 (git): [rubygems/rubygems] Bump rb-sys in /test/rubygems/test_gem_ext_cargo_builder/custom_name
- Bumps [rb-sys](https://github.com/oxidize-rb/rb-sys) from 0.9.34 to 0.9.35.
- [Release notes](https://github.com/oxidize-rb/rb-sys/releases)
- [Commits](https://github.com/oxidize-rb/rb-sys/compare/v0.9.34...v0.9.35)
---
updated-depende... -
07:01 AM Feature #19099: Support `private_constant` for an undefined constant
- `const_set` is already exist. How about `private_const_set`?
-
05:12 AM Feature #19099 (Open): Support `private_constant` for an undefined constant
- All the following discussion applies to `public_constant` too. Maybe `deprecate_constant` as well.
## Problem
```ruby
class C
X = ...
private_constant :X
end
```
The above idiom usually works fine, but when `...` part i... -
06:03 AM Misc #19098: Time#strftime: %z and width
- ```c
#include <stdio.h>
#include <time.h>
void
test_strftime(const char *fmt, const struct tm *t)
{
char buf[1024];
int n = strftime(buf, sizeof(buf), fmt, t);
printf("%-10s=>\t[%s]\n", fmt, buf);
}
int
main(vo... - 05:05 AM Revision d9cf0388 (git): Update Regexp.timeout doc (#6658)
- * Correct Regexp.timeout doc, Timeout.timeout= => Regexp.timeout=
* add link Regexp Timeout section -
- Nice summary. This seems like a solid proposal.
-
04:39 AM Bug #19082: Recent gRPC gem fails to build from the source in already released versions
- monfresh (Moncef Belyamani) wrote in #note-6:
> nobu (Nobuyoshi Nakada) wrote in #note-5:
> ...
Note that this issue is also related to https://bugs.ruby-lang.org/issues/19005. The fixes for Ruby to include `-undefined dynamic_lookup` to... -
- When `--suppress_not_found` option is given, no revision information
is available. And remove extraneous newline, when result is empty or
ends with a newline. -
-
- So the `--srcdir` option in this file can override the same option in
`VPath`. -
-
11/01/2022
-
10:38 PM Revision 70173a72 (git): Ivar copy needs to happen _before_ setting the shape
- When we copy instance variables, it is possible for the GC to be kicked
off. The GC looks at the shape to determine what slots to mark inside
the object. If the shape is set too soon, the GC could think that there
are more instance var... -
-
06:16 PM Revision cbf15e5c (git): YJIT: Add an assert to help with Context changes
- While experimenting I found that it's easy to change Context and forget
to also change the copying operation in limit_block_versions(). Add an
assert to make sure we substitute a compatible generic context when
limiting the number of ver... -
- Context::new() is the same as Context::default() and
Context::new_with_stack_size() was only used in tests. -
05:49 PM Misc #19098 (Open): Time#strftime: %z and width
- It seems `%z` behaves in some surprising way when it is combined with a width - sign `+` is placed at the beginning of the result string:
```ruby
Time.now.strftime("%10z")
=> "+000000200"
Time.now.strftime("%_10z")
=> " +20... -
- Indeed. But I assumed this feature only affects `#strftime`'s `%-z` directive, not treating a time zone offset.
I've checked on Ruby 3.0.3 and `-00:00` wasn't treated as UTC:
```ruby
Time.new(2022, 1, 1, 0, 0, 0, "-00:00").zone
# => ni... -
04:41 PM Misc #19096: [Question] Time with `-00:00` offset is in UTC
- Oh, it is actually deliberate, see #17544:
> In RFC 3339, -00:00 is used for the time in UTC is known, but the offset to local time is unknown. -
- On my computer (GMT+2, Europe/Kyiv timezone), the same behavior is reproducing since 3.1 (and on the current head).
```ruby
[Time.new(2022, 1, 1, 0, 0, 0, "+00:00").utc?, Time.new(2022, 1, 1, 0, 0, 0, "-00:00").utc?]
# => [false, true... -
03:28 PM Misc #19096: [Question] Time with `-00:00` offset is in UTC
- On my computer, which is set to JST (+0900), the three examples make no difference:
```rb
Time.new(2022, 1, 1, 0, 0, 0, "-00:00").utc? # => false
```
-
- It's a bit unexpected but
```ruby
Time.new(2022, 1, 1, 0, 0, 0, "-00:00").utc?
# => true
```
But time with `+00:00` or `0` offset is treated as not UTC time:
```ruby
Time.new(2022, 1, 1, 0, 0, 0, "+00:00").utc? # => false
T... -
03:21 PM Feature #19097 (Open): Improve performance some Float and IO class constants
- Improve performance some Float and IO class constants written by Ruby.
benchmark:
```yaml
benchmark:
Float_RADIX: |
Float::RADIX
Float_MANT_DIG: |
Float::MANT_DIG
Float_DIG: |
Float::DIG
Float_MIN_EXP: |... -
01:17 PM Bug #18380: TestAddressResolve#test_socket_getnameinfo_domain_blocking test failures
- ioquatix (Samuel Williams) wrote in #note-2:
> This should be fixed by <https://github.com/ruby/ruby/pull/6652>. Can you please confirm? Thanks :)
This also works on Fedora infrastructure:
https://src.fedoraproject.org/rpms/ruby/c... -
11:20 AM Bug #18380 (Closed): TestAddressResolve#test_socket_getnameinfo_domain_blocking test failures
- Closing as our CI on Fedora 36 passed http://rubyci.s3.amazonaws.com/fedora36/ruby-master/recent.html . Thank you!
-
04:11 AM Bug #18380: TestAddressResolve#test_socket_getnameinfo_domain_blocking test failures
- This should be fixed by <https://github.com/ruby/ruby/pull/6652>. Can you please confirm? Thanks :)
-
- I agree the dup is unnecessary since 99.99% of the time you're just doing `buf << CGI.escapeHTML(str)`. Not sure the performance gain would be measurable. A new method like `CGI.escapeHTML!` would make sense to me, as it indicates the da...
-
-
- Also:
- format -> time_format
- output -> formatter -
- - writing to a file or stdout
- touching timestamp files
- overwriting only if changed
- colorizing -
-
- 04:10 AM Revision 16953867 (git): We don't care about actual hostname resolution. (#6652)
- https://bugs.ruby-lang.org/issues/18380
-
04:06 AM Feature #19064: UDPSocket#bind does not take AddrInfo, despite documentation saying it should
- The linked documentation is for Socket, not UDPSocket. The documentation for UDPSocket does not suggest that it accepts an Addrinfo. https://docs.ruby-lang.org/en/master/UDPSocket.html
The difference between Socket/UDPSocket is confusi... -
-
- I want to add annotations to default gems like `Net::HTTP` so it would need to be in core, or at least some kind of hook. If it's too much to have a specific implementation, we could just make a general annotation method and then hook it...
-
-
- Applied in changeset commit:git|2d86e79fe6a8eaea85edb4b9ab59d12228dbd8b9.
----------
Always lookup IV buffers when iterating
Always look up instance variable buffers when iterating. It is possible
for the instance variable buffer to c... - 12:10 AM Revision 4c59808a (git): [rubygems/rubygems] Bump rb-sys
- Bumps [rb-sys](https://github.com/oxidize-rb/rb-sys) from 0.9.34 to 0.9.35.
- [Release notes](https://github.com/oxidize-rb/rb-sys/releases)
- [Commits](https://github.com/oxidize-rb/rb-sys/compare/v0.9.34...v0.9.35)
---
updated-depende... -
- Always look up instance variable buffers when iterating. It is possible
for the instance variable buffer to change out from under the object
during iteration, so we cannot cache the buffer on the stack.
In the case of Bug #19095, the t...
10/31/2022
-
09:05 PM Revision 02f15542 (git): Implement object shapes for T_CLASS and T_MODULE (#6637)
- * Avoid RCLASS_IV_TBL in marshal.c
* Avoid RCLASS_IV_TBL for class names
* Avoid RCLASS_IV_TBL for autoload
* Avoid RCLASS_IV_TBL for class variables
* Avoid copying RCLASS_IV_TBL onto ICLASSes
* Use object shapes for Class and Module IVs -
- * YJIT: Add RubyVM::YJIT.code_gc
* Rename compiled_page_count to live_page_count -
06:29 PM Revision 5e6633fc (git): YJIT: reduce default `--yjit-exec-mem-size` to 128MiB instead of 256 (#6649)
- Reduce default --yjit-exec-mem-size to 128MiB instead of 256
-
01:49 PM Misc #19074: DevMeeting-2022-12-01 (postponed from 11-17)
- * [Bug #18899] Inconsistent argument handling in IO#set_encoding (@eregon)
* Could someone review Jeremy's fix and merge it?
* The logic for IO set_encoding is really complicated (also following the C code for it is hard), I wish w... -
12:24 PM Feature #18996: Proposal: Introduce new APIs to reline for changing dialog UI colours
- > > Do you think it'd make sense to have one for the colours too?
> ...
Yes `completier` was a typo. I've corrected it.
I did a quick search and it looks like `completer_*` methods, `completer_word_break_characters` and `completer_qu... -
- `Mutex#sleep` accepts an explicit nil argument and it's documented.
+1, this inconsistency is rather weird (and rather messy to implement in TruffleRuby).
I don't know about `false` though, that's `can't convert FalseClass into time ... -
- By the way, my motivation was to make this easier for the user: https://github.com/socketry/async/issues/161
-
- Haha, great, but I think it's odd you can't just write `sleep(nil)`. I guess Ruby uses `Qundef` internally, but user can't write `sleep(undef)`, but `sleep(*nil)` is close. Actually, I didn't even know you can write `sleep(*nil)` to get ...
-
- I don't think it is hard to implement. All you need is a twinkle little star.
```rb
class Sleeper
def initialize(time = nil)
@time = time
end
def sleep
Kernel::sleep(*@time)
end
end
```
-
- ```
> sleep(nil)
(irb):1:in `sleep': can't convert NilClass into time interval (TypeError)
```
However, I feel that this makes implementing a compatible sleep method a little difficult.
```ruby
def sleep(time = nil)
if time
... -
-
- Compiling with the clang address sanitizer and the following patch, `make BASERUBY='$(MINIRUBY)' golf_prelude.c` aborts by use-after-poison.
I can't make minimum code yet.
```patch
diff --git i/template/prelude.c.tmpl w/template/pre... -
04:58 AM Revision 350d0aa0 (git): [ruby/error_highlight] Support nodes in `spot`
- Fixes a bug where `spot` was using the wrong local variable.
We want to use error highlight with code that has been eval'd,
specifically ERB templates. We can recover the compiled source code of
the ERB template but we need an API to pa... -
- Thank you all! I have merged the PR at commit:3391c51effcd61f9a718adf59740661d99f57b5b
-
02:04 AM Feature #19083: Add `node_id_for_backtrace_location` to ast.c
- Sounds OK. Please merge.
Matz.
-
- We want to use error highlight with eval'd code, specifically ERB
templates. We're able to recover the generated code for eval'd templates
and can get a parse tree for the ERB generated code, but we don't have a
way to get the node id fr... -
- I created the backport pull-request
https://github.com/ruby/ruby/pull/6648
10/30/2022
-
-
10:39 PM Bug #19092: gem uninstall exit code when multiple gem are installed and no option is given on prompt
- Hello! This seems a valid problem report, I tried locally and I can reproduce. Can you open your issue at rubygems issue tracker please? (https://github.com/rubygems/rubygems/issues/new?assignees=&labels=RubyGems&template=rubygems-relate...
-
09:17 PM Bug #19092 (Third Party's Issue): gem uninstall exit code when multiple gem are installed and no option is given on prompt
- When there are several versions of a gem to uninstall and there is a prompt with all the installed versions and it is waiting for the number of the option to uninstall, if you press ENTER then it displays an error message like this
> ... -
-
- please follow https://bugs.ruby-lang.org/issues/19092#note-1
-
- Scenario: calling `gem uninstall some_gem` and there is more than 1 version of that gem installed
A list of the n versions of the gem is displayed and a prompt is waiting for a number of the item to uninstall.
When using ctrl+d on the ... -
- `HAVE_` macros by autoconf are defined as 1.
https://github.com/ruby/bigdecimal/commit/cd35868aa6 -
01:25 PM Bug #19091 (Closed): backport reline-0.3.1 to ruby-3.1
- I am ashamed to say, last year I committed a work-in-progress file instead of the correct one to ruby/reline tree.
There is a bug that breaks scrolling on windows (with terminals other than command prompt?).
That mistake was included i... -
- The source directory may be read-only.
-
- The CI for MinGW has used it.
-
-
- Now revision.tmp will be regenerated always and every times, even if
the recent file exists in the source directory, as far as using git.
On the other hand, VirtualBox mounts shared folders as root, and git
rejects the repository there a... -
- 07:04 AM Revision 1acbcf0e (git): Update bundled gems list at 2022-10-30
10/29/2022
-
07:47 PM Revision 91c28ab2 (git): [DOC] Enhanced RDOc for IO (#6642)
- In io.c treats:
#close
#close_read
#close_write
#closed -
-
- I've update the function name from node_id_for_location to node_id_for_backtrace_location. Also I attached the patch here.
-
- https://github.com/ruby/optparse/commit/51f7e060ee
- 01:46 PM Revision 37291df9 (git): [ruby/optparse] #load() into hash
- (https://github.com/ruby/optparse/pull/42)
OptionParser#load learns .load(into: Hash)
https://github.com/ruby/optparse/commit/2ea626fcff
Co-authored-by: Nobuyoshi Nakada <[email protected]> -
- https://github.com/ruby/optparse/commit/fb34a1d3a3
-
- Isn't `rb_str_dup` copy-on-write and so should be fairly cheap?
-
- ## Proposal
Stop guaranteeing that `GGI.escapeHTML` returns a new string even if there's nothing to be escaped.
More specifically, stop calling this `rb_str_dup` https://github.com/ruby/cgi/blob/v0.3.3/ext/cgi/escape/escape.c#L72 for... -
- https://github.com/ruby/digest/commit/cce9ada85e
-
- As clang on macOS defines this macro as 0 internally when a sanitizer
option is given, clear it before definition to suppress redefinition
warnings. -
- According to @k0kubun, some people actually use the hack to work around this problem by using `--disable-gems` or `--disable-error_highlight`.
However, this is arguably far from ideal. The option `--disable-gems` is not recommended fo...
10/28/2022
-
-
- jeremyevans0 (Jeremy Evans) wrote in #note-6:
> vo.x (Vit Ondruch) wrote in #note-2:
> ...
Ah, right. I have not realized this. Thx for pointing this out.
Eregon (Benoit Daloze) wrote in #note-8:
I completely agree with your as... -
- I believe this feature hurts startup time optimizations, so from that point of view it's not great.
For instance TruffleRuby loads `did_you_mean` ahead of time, during context pre-initialization.
Similarly, CRuby could for instance sto... -
- Thank you all.
All I ask is that did_you_mean, etc. be loaded by default and that their versions be controllable by Gemfile. I don't have a strong preference on how to fix it.
Now I agree that modifying Ruby is overkill for this is... -
02:50 PM Feature #19089: Load bundler/setup in gem_prelude.rb when "bundle exec" is used
- vo.x (Vit Ondruch) wrote in #note-2:
> From that POV, it would be much better if it was possible to disable did_you_mean autoloading.
That's already possible via `--disable-did_you_mean`
> ...
I agree with this 100%.
Is this fi... -
- @nobu pointed out that we can load `bundler/setup` in lib/rubygems.rb instead of gem_prelude.rb.
```ruby
diff --git a/lib/rubygems.rb b/lib/rubygems.rb
index 915a899f38..b61d21050d 100644
--- a/lib/rubygems.rb
+++ b/lib/rubygems.rb
@@ -... -
- @deivid Thank you for your reply!
deivid (David Rodríguez) wrote in #note-3:
> I'm not sure the proposed approach will work when Bundler is installed as a default gem though?
Oops, you are right. My current patch does not work wit... -
- @mame That's indeed a [long standing issue](https://github.com/rubygems/rubygems/blob/master/bundler/spec/runtime/setup_spec.rb#L1305-L1308) for us and I'm very happy if it can get fixed! I'm not sure the proposed approach will work when...
-
- Let me provide you different perspective. did_you_mean is development dependency, which have no benefit for runtime, where there should be error free code. From that POV, it would be much better if it was possible to disable did_you_mean...
-
- ### Problem
Currently, we cannot specify the version of did_you_mean by using Gemfile.
For example, Ruby master bundles with did_you_mean 1.6.1 by default:
```
$ ruby -e 'p DidYouMean::VERSION'
"1.6.1"
```
Consider that we... -
- Got it. Thank you.
-
- `exception:false` is effective only for the content.
ASCII incompatible encoding strings are not subjects of parsing.
-
- https://github.com/ruby/ruby/pull/6645
-
-
-
- (https://github.com/ruby/irb/pull/357)
https://github.com/ruby/irb/commit/298b134792 -
- https://no-color.org has been updated (jcs/no_color#83):
> Command-line software which adds ANSI color to its output by default
should check for a `NO_COLOR` environment variable that, when present
and **not an empty string** (regardles... -
- https://github.com/ruby/irb/commit/a7097c5b80
-
- https://github.com/ruby/irb/commit/565eeb3c19
-
- This was merged at https://github.com/ruby/ruby/pull/6187.
-
-
02:44 AM Bug #19037: [Ruby 3.1.2] irb display color character not correct(as literal form) when set IRB.conf[:USE_COLORIZE] = true (it is default)
- zw963 (Wei Zheng) wrote in #note-4:
> mame (Yusuke Endoh) wrote in #note-2:
> ...
Any update? -
- > Often times, this is done to avoid holding a reference to the object in the thread, effectively making it immortal.
I think the correct way to make opaque per-instance keys is something like this:
```ruby
class ConnectionPool
...
10/27/2022
-
- This code used to roll its own heap object check before we made a better
version in guard_known_class(). The improved version uses one fewer
comparison, so let's use that. -
- * Update docs wrt YJIT limitations and building YJIT
* Update building_ruby.md
Fix relative link -
- I've noticed a minor inconsistency of the `Complex()` method. When it is called with `exception: false` option I suppose it shouldn't raise any exception.
For instance it doesn't raise error in the following cases:
```ruby
Complex... -
- Fixes this warning:
warning: assigned but unused variable - err
https://github.com/ruby/irb/commit/298fcb57a3 - 07:29 PM Revision b260c1e8 (git): [ruby/irb] Remove unnecessary test setup
-
-
- * Specifically the second one causes `$HOME` to be unset, which breaks `File.expand_path('~')`.
https://github.com/ruby/irb/commit/61963305f5 -
-
05:47 PM Bug #19086: Dir#entries and Dir#glob fail in stand-alone CRuby app
- mame (Yusuke Endoh) wrote in #note-1:
> I guess you need to call `ruby_init` and `ruby_options` instead of `ruby_setup`. I think `Dir.glob` is defined in `ruby_options`.
=_=¡ I was hoping the solution wouldn't be that simple because it ... -
- I guess you need to call `ruby_init` and `ruby_options` instead of `ruby_setup`. I think `Dir.glob` is defined in `ruby_options`.
-
- I've noticed a minor inconsistent behaviour of `String#to_c` method. It does not treat a sequence of `_` character as an end of meaningful characters:
```ruby
"123".to_c # => (123+0i)
"12_3".to_c # => (123+0i)
"12__3".to_c # => (12... -
- Applied in changeset commit:git|c6f439a6a8df582416e756d7511aa4d9c72071a9.
----------
Improve performance some `Integer` and `Float` methods [Feature #19085] (#6638)
* Improve some Integer and Float methods
* Using alias and Remove unn... -
- * Improve some Integer and Float methods
* Using alias and Remove unnecessary code
* Remove commentout code - 03:36 PM Revision 7cf7e6c3 (git): [ruby/irb] Add missing require
-
- This proc displays rdoc document when the input matches certain symbols
perfectly, like "String". It's commonly triggered with autocompletion
but only has 1 test case. So this commit increases its test coverage.
https://github.com/ruby/... -
10:15 AM Feature #19078: Introduce `Fiber#storage` for inheritable fiber-scoped variables.
- > I feel like this encourages bad practices because people are forced to convert objects to the hash.
Often times, this is done to avoid holding a reference to the object in the thread, effectively making it immortal.
If we take th... -
08:16 AM Revision 5129ca3e (git): [ruby/rdoc] Delay `require "readline"` in case the terminal is in raw mode
-
-