Magic v2: Viral value magic by leonerd · Pull Request #24508 · Perl/perl5

leonerd · 2026-06-25T11:15:25Z

This PR builds on top of the general "Magic v2" to add viral value magic; as outlined in PPC0036.

tonycoz · 2026-06-29T05:26:12Z

look at only the final 3 commits

There's a few commits before those that aren't in #24163 (I wondered where the op.c change came from.)

tonycoz · 2026-06-29T05:36:46Z

+    }

    SV *ex = vmess(pat, args);
+    PERL_ARGS_ASSERT_VWARN;


The assert should be above the vmess() call, it's pointless below it.

Oh oops. I don't recall moving that but maybe that was a recent(ish) blead fix that has been broken by some rebase operation over the past few months. I shall fix that.

leonerd · 2026-06-29T20:56:31Z

look at only the final 3 commits

There's a few commits before those that aren't in #24163 (I wondered where the op.c change came from.)

Ahyes, those are some post-phase1 fixups to other files to use the new magic system in static locations, just as a demonstration that it works. I intend to submit those as little fixes after the first branch is merged, before this one. github doesn't easily allow me to make even a draft PR between two moving points in a branch, so that was the best I could do. Those will be out of this PR by the time it becomes ready.

leonerd · 2026-07-08T17:45:10Z

Since it seems nobody's going to review this while it's drafted I shall undraft it. I don't imagine there's a danger anyone will accidentally merge it ahead of time anyway. ;)

leonerd · 2026-07-13T16:51:20Z

Are there any reviewers for this?

tonycoz · 2026-07-13T23:42:12Z

    /* ideally we could just remove the magic from the SV but we don't get the SV here */
-    SvREFCNT_dec(mg->mg_obj);
-    mg->mg_obj = NULL;
-
-    return 0;
+    MgAUXSV_set(mg, NULL);


comment out of date

Ahyes; though that's part of #24547. I'll see what I can do there.

tonycoz · 2026-07-13T23:54:38Z

+GCC_DIAG_IGNORE_DECL(-Wmissing-field-initializers);
 static const struct MagicFunctions magicfuncs_perlencoding = {
    .ver   = 2,
    .shape = MGv2s_BASE,
    .debug_name = "PerlIO::encoding/tag",
 };
+GCC_DIAG_RESTORE_DECL;


Given the discussion I'm not sure this still needs to be here,

Ahyes; also part of #24547

tonycoz · 2026-07-14T00:35:14Z

 /* Macros for valuemagic support */
-/* TODO: allow these to be conditional like the taint ones are */
-#define VALUEMAGIC_CLEAR \
+#if defined(NO_VALUEMAGIC_SUPPORT)


Should -DNO_TAINT_SUPPORT implicitly -DNO_VALUEMAGIC_SUPPORT? This would simplify checks like in magic2_value.t

It would also later allow value magic without taint

tonycoz · 2026-07-14T04:13:20Z

Except for the minor issues I mentioned it looks good to me.

tonycoz · 2026-07-14T23:42:44Z

+#
+#        unless ($defines =~ m/\b(NO_VALUEMAGIC_SUPPORT)\b/) {
+#            my $valuemagic_support = 'define';
+#            s/^(valuemagic_support=['"])(["'])/$1$valuemagic_support$2/m;
+#        }


I thought there was some discussion of this recently: on the fragility of using compiler -D switches to control this type of feature instead of Configure defines.

But I couldn't find it.

I don't know if you saw it, or if you agree or not... but it seems relevant to this (wishing I could find it)

leont enlightened me, it was #23306

Ahyes, I vaguely recalled it. That's why for example, the newly-added variable in intrpvar.h isn't #ifdef-guarded. That way the structure always has the same shape. The only thing that the macro guards are the actual bits of code that implement the behaviour.

If the value magic feature is compiled out, stub implementations of the propagate and unpropagate functions are still built anyway, which just do nothing. So I think it should be fairly safe if, for example, a CPAN module gets built with a different set of flags than the core binary. Structure layouts all match, but the relevant behaviour just won't happen unless all the parts line up.

tonycoz · 2026-07-22T01:32:32Z

+/* we can't make this one `static` because -Wc++-compat gets upset if we do */
+extern const struct ScalarValueMagicFunctions magicfuncs_value;


You should be able to predeclare the function instead. Not that it's a big deal.

Leont · 2026-07-28T11:22:13Z

Much like taint support, this probably should be controlled by a configuration flag.

Leont · 2026-07-28T11:24:03Z

Oh I now see there is a NO_VALUEMAGIC_SUPPORT define (which is half of the work). Much like taint that design is a mistake and really really needs a configuration flag. And probably it should be an opt-in feature instead of opt-out.

tonycoz · 2026-07-30T01:20:14Z

I saw some IRC discussion related to adding a ./Configure flag - is that happening?

Leont · 2026-07-30T09:33:31Z

I saw some IRC discussion related to adding a ./Configure flag - is that happening?

The real discussion there was about whether is should be enabled by default or not. In my opinion it should not (or at least for now).

leonerd · 2026-07-30T10:13:53Z

Yes, I'm looking into both of these today. It seems to be a short excursion via learning how metaconfig works ;)

My plan is to end up with two PRs, to add two new configuration options; one for this and one for taint while I'm there. Likely the defaults will be taint on, valuemagic off; but either can be adjusted:

./Configure -[D|U]usevaluemagic -[D|U]usetaint ...

(It probably doesn't make sense to enable valuemagic while disabling taint as I suspect that in itself will break, so there might need to be some interlocking logic for that situation.)

Currently unused by the rest of the code. This commit is extracted separately simply so it can be placed at the very start of the branch. A full rebuild is needed either side of this commit, so by being at the beginning this makes rebase, bisect, and other operations easier as now you don't have to `make clean`.

Adds a new SVs_VMG flag + associated macros; use the SvVMAGICAL flag to fast-path sv_has_valuemagic() if appropriate Use the sv_has_valuemagic() function to know whether to call mg_propagate() in sv_setsv_flags()

Much like -DNO_TAINT_SUPPORT, this is optional and not default, but will disable value magic support in exchange for getting that little extra performance boost by not having to worry about it.

leonerd added the defer-next-dev This PR should not be merged yet, but await the next development cycle label Jun 25, 2026

leonerd changed the title ~~Magic v2 phase2~~ Magic v2: Viral value magic Jun 25, 2026

leonerd force-pushed the magic-v2-phase2 branch 3 times, most recently from f6f60ec to e7256b0 Compare June 26, 2026 14:13

tonycoz reviewed Jun 29, 2026

View reviewed changes

Comment thread perl.h Outdated

tonycoz reviewed Jun 29, 2026

View reviewed changes

Comment thread op.c

tonycoz reviewed Jun 29, 2026

View reviewed changes

leonerd force-pushed the magic-v2-phase2 branch from e7256b0 to 676058c Compare June 29, 2026 20:49

leonerd force-pushed the magic-v2-phase2 branch 2 times, most recently from 9d35f64 to bc28a9c Compare June 30, 2026 15:01

leonerd mentioned this pull request Jul 7, 2026

Use C99 designated initialisers for PerlIO_funcs #24390

Merged

leonerd force-pushed the magic-v2-phase2 branch from d277e05 to 3a50b1d Compare July 8, 2026 15:38

leonerd mentioned this pull request Jul 8, 2026

Magic v2 Internal Code Conversions #24547

Closed

leonerd added the do not merge Don't merge this PR, at least for now label Jul 8, 2026

leonerd marked this pull request as ready for review July 8, 2026 17:45

leonerd force-pushed the magic-v2-phase2 branch 2 times, most recently from d2cdd7c to 39731be Compare July 13, 2026 13:00

tonycoz reviewed Jul 13, 2026

View reviewed changes

tonycoz reviewed Jul 14, 2026

View reviewed changes

Comment thread perl.h

tonycoz reviewed Jul 14, 2026

View reviewed changes

leonerd force-pushed the magic-v2-phase2 branch from 39731be to dfee0d1 Compare July 14, 2026 12:56

tonycoz reviewed Jul 14, 2026

View reviewed changes

github-actions Bot added the hasConflicts label Jul 15, 2026

github-actions Bot added the hasConflicts label Jul 21, 2026

tonycoz reviewed Jul 22, 2026

View reviewed changes

github-actions Bot removed the hasConflicts label Jul 22, 2026

leonerd force-pushed the magic-v2-phase2 branch 6 times, most recently from b2bea2c to 9ba1201 Compare July 27, 2026 10:30

github-actions Bot added the hasConflicts label Jul 27, 2026

leonerd force-pushed the magic-v2-phase2 branch from 9ba1201 to 821948d Compare July 27, 2026 17:07

github-actions Bot removed the hasConflicts label Jul 27, 2026

leonerd force-pushed the magic-v2-phase2 branch 2 times, most recently from 8585c94 to 1475f03 Compare July 28, 2026 11:02

leonerd force-pushed the magic-v2-phase2 branch from 1475f03 to 42230a9 Compare July 28, 2026 12:56

leonerd removed the do not merge Don't merge this PR, at least for now label Jul 28, 2026

github-actions Bot added the hasConflicts label Jul 30, 2026

leonerd force-pushed the magic-v2-phase2 branch from 42230a9 to 2547676 Compare July 30, 2026 18:05

leonerd mentioned this pull request Jul 30, 2026

[PPC0036] Add document about Viral Value Magic Perl/PPCs#90

Open

github-actions Bot removed the hasConflicts label Jul 30, 2026

leonerd added 4 commits August 3, 2026 16:33

Add TAINT_IF_SV() helper macro

19f383d

Create value magic

496bdd3

Adds a new SVs_VMG flag + associated macros; use the SvVMAGICAL flag to fast-path sv_has_valuemagic() if appropriate Use the sv_has_valuemagic() function to know whether to call mg_propagate() in sv_setsv_flags()

Recognise a -DNO_VALUEMAGIC_SUPPORT to disable value magic

a37b467

Much like -DNO_TAINT_SUPPORT, this is optional and not default, but will disable value magic support in exchange for getting that little extra performance boost by not having to worry about it.

leonerd force-pushed the magic-v2-phase2 branch from 2547676 to a37b467 Compare August 3, 2026 15:34

		/* we can't make this one `static` because -Wc++-compat gets upset if we do */
		extern const struct ScalarValueMagicFunctions magicfuncs_value;

Uh oh!

Conversation

leonerd commented Jun 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tonycoz commented Jun 29, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

leonerd commented Jun 29, 2026

Uh oh!

leonerd commented Jul 8, 2026

Uh oh!

leonerd commented Jul 13, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

tonycoz commented Jul 14, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Leont commented Jul 28, 2026

Uh oh!

Leont commented Jul 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tonycoz commented Jul 30, 2026

Uh oh!

Leont commented Jul 30, 2026

Uh oh!

leonerd commented Jul 30, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

leonerd commented Jun 25, 2026 •

edited

Loading

Leont commented Jul 28, 2026 •

edited

Loading