ruby-core@ruby-lang.org archive (unofficial mirror)
 help / color / mirror / Atom feed
From: "pabloh (Pablo Herrero) via ruby-core" <ruby-core@ml.ruby-lang.org>
To: ruby-core@ml.ruby-lang.org
Cc: "pabloh (Pablo Herrero)" <noreply@ruby-lang.org>
Subject: [ruby-core:118965] [Ruby master Bug#20701] Hash argument passed as keyword splat can be mutated inside method
Date: Tue, 27 Aug 2024 06:04:38 +0000 (UTC)	[thread overview]
Message-ID: <redmine.issue-20701.20240827060437.5077@ruby-lang.org> (raw)
In-Reply-To: <redmine.issue-20701.20240827060437.5077@ruby-lang.org>

Issue #20701 has been reported by pabloh (Pablo Herrero).

----------------------------------------
Bug #20701: Hash argument passed as keyword splat can be mutated inside method
https://bugs.ruby-lang.org/issues/20701

* Author: pabloh (Pablo Herrero)
* Status: Open
* ruby -v: ruby 3.4.0dev (2024-08-27T00:39:24Z master e5295644c9) [x86_64-linux]
* Backport: 3.1: UNKNOWN, 3.2: UNKNOWN, 3.3: UNKNOWN
----------------------------------------
If you call a method with a splatted kwargs that also uses a splatted array followed by at least one positional argument, the called method can mutate the splatted kwargs.

``` ruby
def foo(a, h)
  h[:splat_modified] = true
end

a = []
b = {splat_modified: false}

foo(*a, :x, **b) 
b[:splat_modified] # Should be 'false' but is 'true'

# It makes no difference if the called method expects a kwargs
def foo2(a, **h)
  h[:splat_modified] = true
end

b[:splat_modified] = false
foo2(*a, :x, **b)
b[:splat_modified] # Should be 'false' but is 'true'


# The case when a block is also pass is broken too

c = proc { }
b[:splat_modified] = false
foo(*a, :x, **b, &c) 
b[:splat_modified] # Should be 'false' but is 'true'


# When the extra positional argument after the array splat is removed the correct behavior is restored
b[:splat_modified] = false
foo(*a, **b)
foo(*a, :x, **b, &c)
b[:splat_modified] # Is 'false'
```

Apparently, this problem is due to the `KW_SPLAT_MUT` flag being set at the method's call site:


`ruby --dump=insns,-optimize -e 'a, b, c = [4,6,4],{a: 3},:ds; foo(*a,9,**c)` outputs
```
0000 duparray                               [4, 6, 4]                 (   1)[Li]
0002 duphash                                {:a=>3}
0004 putobject                              :ds
0006 setlocal                               c@2, 0
0009 setlocal                               b@1, 0
0012 setlocal                               a@0, 0
0015 putself
0016 getlocal                               a@0, 0
0019 splatarray                             true
0021 putobject                              9
0023 pushtoarray                            1
0025 getlocal                               c@2, 0
0028 send                                   <calldata!mid:foo, argc:2, ARGS_SPLAT|ARGS_SPLAT_MUT|FCALL|KW_SPLAT|KW_SPLAT_MUT>, nil
0031 leave
```

The error is present in both 3.3 and 3.4:
- ruby 3.4.0dev (2024-08-27T00:39:24Z master e5295644c9) [x86_64-linux]
- ruby 3.3.4 (2024-07-09 revision be1089c8ec) [x86_64-linux]

PR for test: https://github.com/ruby/ruby/pull/11467



-- 
https://bugs.ruby-lang.org/
 ______________________________________________
 ruby-core mailing list -- ruby-core@ml.ruby-lang.org
 To unsubscribe send an email to ruby-core-leave@ml.ruby-lang.org
 ruby-core info -- https://ml.ruby-lang.org/mailman3/lists/ruby-core.ml.ruby-lang.org/

       reply	other threads:[~2024-08-27  6:04 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-27  6:04 pabloh (Pablo Herrero) via ruby-core [this message]
2024-08-27 14:23 ` [ruby-core:118966] " jeremyevans0 (Jeremy Evans) via ruby-core
2024-09-02 10:19 ` [ruby-core:119014] " k0kubun (Takashi Kokubun) via ruby-core

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=redmine.issue-20701.20240827060437.5077@ruby-lang.org \
    --to=ruby-core@ml.ruby-lang.org \
    --cc=noreply@ruby-lang.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).