prepare perldelta for the release of v5.42.0 #23366

book · 2025-06-11T17:07:35Z

This is a manual merge of the perl541*delta.pod files in preparation for the 5.42.0 release.

Review and comments (especially about omissions and inconsistencies) are very welcome.

We wonder if the (currently)sections about apostrophe_as_package_separator and smartmatch really belong to "Incompatible Changes" and not to "Core Enhancements". The "unfeatures" are backward-compatible by design.

Leont · 2025-06-11T17:14:37Z

The cloning of dirhandle bug isn't mentioned (probably because I neglected to write that for the dev release). That should absolutely be in the stable release though.

richardleach · 2025-06-11T22:13:57Z

pod/perldelta.pod

+
+=head2 New C<any> and C<all> operators
+
+Two new experimental features have been added, which add two new list-processing


Rather than have 2x "two new", how about "Two new experimental features add the list-processing operators, C and C."?

richardleach · 2025-06-11T22:16:16Z

pod/perldelta.pod

+This can be controlled with the C<apostrophe_as_package_separator>
+feature which is enabled by default, but is disabled from the 5.41
+feature bundle onwards.


For this perldelta, should it read "disabled from the 5.42 feature bundle onwards"?

Technically, :5.41 is an alias to :5.42. The documentation for feature also says: "This is enabled by default, but disabled from the 5.41 feature bundle
onwards."

However, the use of development bundles is not very consistent in feature.pm, so maybe we should just mention stable versions in there.

richardleach · 2025-06-11T22:19:12Z

pod/perldelta.pod

+They were deprecated in Perl v5.38.0 and scheduled for removal in Perl
+5.42.0, and entirely removed in Perl 5.41.3.
+
+After some discussion these have been re-instated in 5.41.9.
+
+Using them no longer produces a deprecation warning.


Probably too much detail for this perldelta. How about something along the lines of:

"They were deprecated in Perl v5.38.0 and scheduled for removal in Perl 5.42.0, but after some discussion their removal has been indefinitely postponed. Using them no longer produces a deprecation warning."

richardleach · 2025-06-11T22:20:38Z

pod/perldelta.pod

+Note that calling C<eval EXPR> in a function unconditionally causes a function
+to reference its enclosing functions as it always has.
+
+=head2 Switch and Smart Match operator reinstated


Needs merging with the previous version of this notice.

richardleach · 2025-06-11T22:21:50Z

pod/perldelta.pod


 =item *

-XXX
+C<Perl_leave_adjust_stacks>: additional efficiency for mortal copies.


I don't think this needs to be included. It's unlikely to be very user-noticeable.

richardleach · 2025-06-11T22:23:42Z

pod/perldelta.pod

+=item *
+
+Subroutines in packages no longer need to be stored in typeglobs:
+declaring a subroutine will now put a simple sub reference directly in the
+stash if possible, saving memory.  The typeglob still notionally exists,
+so accessing it will cause the stash entry to be upgraded to a typeglob
+(i.e. this is just an internal implementation detail).
+This optimization does not currently apply to XSUBs or exported
+subroutines, and method calls will undo it, since they cache things in
+typeglobs.
+[L<GH #23001|https://github.com/Perl/perl5/issues/23001>]
+
+(This optimization was originally announced in L<perl5220delta>, but due to a
+bug it only worked for subroutines in package C<main>, not in modules.)


IIRC this got reverted

richardleach · 2025-06-11T22:27:21Z

pod/perldelta.pod

+=item *
+
+pp_reverse: don't COW buffer just to then un-COW it [L<GH #22729|https://github.com/Perl/perl5/issues/22729>]


This doesn't need to be included. There's no API change, it doesn't change the visible operation of the OP and the performance change won't be noticeable to end users.

book · 2025-06-12T08:27:00Z

The cloning of dirhandle bug isn't mentioned (probably because I neglected to write that for the dev release). That should absolutely be in the stable release though.

Please take the time to write it in: you're the best positioned to do it.

book · 2025-06-12T08:29:04Z

@richardleach Thanks a lot for your review! I'll go over it soon.

book · 2025-06-18T15:48:59Z

Note to self: mention the doc change from fdbbfbf.

Co-authored-by: Thibault Duponchelle <[email protected]> Co-authored-by: Eric Herman <[email protected]>

book · 2025-06-18T16:44:01Z

Note to self: move apostrophe_as_package_separator and smartmatch unfeatures in the "Core Enhancements" section. Look at the deltas for 5.36 and 5.38 for examples.

I pressed the wrong button

robrwo · 2025-06-18T18:49:32Z

Note to self: mention the doc change from fdbbfbf.

"The recommended alternatives to the rand() function were updated to modern modules recommended by the CPAN Security Group."

khwilliamson · 2025-06-18T18:32:43Z

pod/perldelta.pod

+invoking the code block.
+
+The feature flags enabling those keywords have been named L<keyword_any>
+and L<keyword_all> to avoid confusion with to the ability of the


with to the doesn't make sense to me

khwilliamson · 2025-06-18T18:33:50Z

pod/perldelta.pod

+This can be controlled with the C<apostrophe_as_package_separator>
+feature which is enabled by default, but is disabled from the 5.41
+feature bundle onwards.


khwilliamson · 2025-06-18T18:35:19Z

pod/perldelta.pod

+
+Along with the ability to declare methods lexically, this release also permits
+invoking a lexical subroutine as if it were a method, bypassing the usual
+name-based method resolution by name.


name...name sounds redundant

khwilliamson · 2025-06-18T18:37:35Z

pod/perldelta.pod

+They were deprecated in Perl v5.38.0 and scheduled for removal in Perl
+5.42.0, and entirely removed in Perl 5.41.3.
+
+After some discussion these have been re-instated in 5.41.9.


s/some/extensive/
seems better to me

khwilliamson · 2025-06-18T18:40:14Z

pod/perldelta.pod

+It is believed that this vulnerability can enable Denial of Service or
+Arbitrary Code Execution attacks on platforms that lack sufficient defenses.
+
+Discovered by: Nathan Mills.


Maybe add that the patch to fix this is applicable to all perls that are vulnerable, including those out-of-support

khwilliamson · 2025-06-18T18:48:55Z

pod/perldelta.pod

-=head3 L<XXX>
+C<gv_autoload4> is now documented with C<gv_autoload_pv> and additional notes added.
+
+=back


Maybe add that the long Perl_ forms are now listed when available

khwilliamson · 2025-06-18T18:51:12Z

pod/perldelta.pod

+=item *
+
+Normalized alignment of verbatim sections, fixing how they are displayed by
+some Pod viewers that strip indentation.


This doesn't make sense to me. A viewer that strips indentation doesn't care about the original spacing.

I'd like to know of any viewers that don't strip the indentation and substitute their own.

The content was using tabs for alignment of column data. But it was using 4 spaces for indents. If the indentation you strip is not aligned with a tab stop, then it changes the alignment of the later tabs.

You can see the result of this in the precedence table on https://perldoc.perl.org/perlop#Operator-Precedence-and-Associativity

khwilliamson · 2025-06-18T18:52:38Z

pod/perldelta.pod

+Normalized alignment of verbatim sections, fixing how they are displayed by
+some Pod viewers that strip indentation.
+
+=back


There were several very recent changes to perlop that probably should be mentioned.

khwilliamson · 2025-06-18T18:55:29Z

pod/perldelta.pod

 =over 4

 =item *

-XXX
+A new F<t/run/todo.t> test script was added as a place for TODO tests
+for known unfixed bugs.


Maybe add that patches are welcome to add to this file

khwilliamson · 2025-06-18T18:59:28Z

pod/perldelta.pod

+
+=item *
+
+Enable removing most of mathoms.c and stub functions [L<GH #22691|https://github.com/Perl/perl5/issues/22691>] and [L<GH #22714|https://github.com/Perl/perl5/issues/22714>]


This has been reverted

khwilliamson · 2025-06-18T19:00:03Z

pod/perldelta.pod

+=item *
+
+put apostrophe as package separator behind a default enabled feature,
+which is disabled from feature bundle 5.41.


Isn't this already covered enough earlier?

khwilliamson · 2025-06-18T19:05:23Z

pod/perldelta.pod

+
+However, the separate numeric/string bitwise operators provided by L<the
+C<bitwise> feature|feature/The 'bitwise' feature>, C<< &= ^= |= &.= ^.= |.= >>,
+did not do so:


did is confusing with the above, and not fully explained below

khwilliamson · 2025-06-18T19:06:13Z

pod/perldelta.pod


 =item *

-XXX
+Fix C<POSIX::strftime()> [L<GH #22369|https://github.com/Perl/perl5/issues/22369>].


below we also mention strftime. The two should be merged

khwilliamson · 2025-06-18T19:08:36Z

pod/perldelta.pod

+=item *
+
+On threaded builds on POSIX-like systems, if the perl signal handler
+receives we now resend the signal to the main perl thread.  Previously


receives what?

khwilliamson · 2025-06-18T19:09:25Z

pod/perldelta.pod

+=item *
+
+L<perlfunc/shmread> and L<perlfunc/shmwrite> are no longer limited to 31-bit
+values for their POS and SIZE arguments.


they still are on a 32 bit box, no?

khwilliamson · 2025-06-18T19:11:10Z

pod/perldelta.pod

+
+=item *
+
+Save debugger lines as C<PVIV> SVs rather than as C<PVMG> SVs as they


This needs expansion to make sense. And probably less detail

And is this change actually visible

I'm not capable of expanding that, though.

khwilliamson · 2025-06-18T19:12:45Z

pod/perldelta.pod

+=item *
+
+Prevent a signature parameter of the form C<$ => from leaking an OP at
+compile-time. [L<GH #23187|https://github.com/Perl/perl5/issues/23187>]


Is this important enough to include?

khwilliamson · 2025-06-18T19:13:40Z

pod/perldelta.pod

+
+L<perlfunc/readline> now clears the error flag if an error occurs when
+reading and that error is C<EAGAIN> or C<EWOULDBLOCK>.  This allows
+old code that depended on C<readline> to clear all errors to ignore


why does the code have to be "old"

khwilliamson · 2025-06-18T19:14:10Z

pod/perldelta.pod

+
+=item *
+
+L<C<open>|perlfunc/open> automatically creates an anonymous temporary file


This sounds similar to a previous entry.

thibaultduponchelle · 2025-06-19T09:26:25Z

@haarg It would be great if you could double check aac338e (I followed your 17182bb) 🙏🏼

haarg · 2025-06-19T09:38:36Z

@haarg It would be great if you could double check aac338e (I followed your 17182bb) 🙏🏼

The commit message isn't quite right, because this just cleans up the old delta files, it doesn't do the merge.

thibaultduponchelle · 2025-06-19T10:51:10Z

The commit message isn't quite right, because this just cleans up the old delta files, it doesn't do the merge.

I fixed that, thank you :)

Co-authored-by: Thibault Duponchelle <[email protected]>

richardleach reviewed Jun 11, 2025

View reviewed changes

prepare perldelta for the release of v5.42.0

71e2823

Co-authored-by: Thibault Duponchelle <[email protected]> Co-authored-by: Eric Herman <[email protected]>

khwilliamson previously approved these changes Jun 18, 2025

View reviewed changes

khwilliamson reviewed Jun 18, 2025

View reviewed changes

thibaultduponchelle force-pushed the book/perl5420delta branch 2 times, most recently from e7c160d to 66fe524 Compare June 19, 2025 09:23

thibaultduponchelle force-pushed the book/perl5420delta branch from 66fe524 to aac338e Compare June 19, 2025 09:31

thibaultduponchelle force-pushed the book/perl5420delta branch from ae3b850 to 10396cf Compare June 19, 2025 10:26

book and others added 4 commits June 19, 2025 16:02

review the perldelta against individual perl541*delta

65284d4

Co-authored-by: Thibault Duponchelle <[email protected]>

perldelta: Simplify the recommended alternatives to rand()

33c0fe4

perldelta: Fix syntax

a91e2ce

Clean up BLEAD-POINT perldelta files

d4bd789

Fix podcheck errors related to perldelta

82c3b20

book force-pushed the book/perl5420delta branch from 10396cf to 82c3b20 Compare June 19, 2025 22:28


		=head2 New C<any> and C<all> operators

		Two new experimental features have been added, which add two new list-processing

		=item *

		pp_reverse: don't COW buffer just to then un-COW it [L<GH #22729\|https://github.com/Perl/perl5/issues/22729>]


		=item *

		Enable removing most of mathoms.c and stub functions [L<GH #22691\|https://github.com/Perl/perl5/issues/22691>] and [L<GH #22714\|https://github.com/Perl/perl5/issues/22714>]


		=item *

		Save debugger lines as C<PVIV> SVs rather than as C<PVMG> SVs as they


		=item *

		L<C<open>\|perlfunc/open> automatically creates an anonymous temporary file

prepare perldelta for the release of v5.42.0 #23366

Are you sure you want to change the base?

prepare perldelta for the release of v5.42.0 #23366

Conversation

book commented Jun 11, 2025

Uh oh!

Leont commented Jun 11, 2025

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!

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!

Choose a reason for hiding this comment

Uh oh!

book commented Jun 12, 2025

Uh oh!

book commented Jun 12, 2025

Uh oh!

book commented Jun 18, 2025

Uh oh!

book commented Jun 18, 2025

Uh oh!

robrwo commented Jun 18, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

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!

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!

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!

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!

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!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

robrwo commented Jun 18, 2025 •

edited

Loading

thibaultduponchelle commented Jun 19, 2025 •

edited

Loading