perl.perl5.porters

Re: Magic v2 - Review for Merge and next steps by Leon Timmermans

Wed, 30 Jul 2025 23:55:35 +0000

On Mon, Jul 21, 2025 at 7:51â€¯AM Ovid <[email protected]> wrote:
>
> On Sun, Jul 20, 2025 at 3:11â€¯PM Leon Timmermans <[email protected]> wrote:
>>
>> >> I don't think it's quite clear why we would want value hooks.
>
> <snip>
>>
>> Value magic as an annotation sounds useful and has precedent (e.g. vstrings).
>
> <snip>
>>>
>>> I think that [COW would] require an unholy amount of backreferences to
>>> everything that owns a reference to the current value.
>
> <snip>
>>>
>>> How would [not printing sensivive information] work? ...
>
> <snip>
>
> I think this is getting into implementation details, which is great, but your original concern was "I don't think it's quite clear why we would want value hooks." That's the question I was trying to answer.

Fair point, my apologies.

Leon

Re: no 'declaration after statement' error by bulk 88

Wed, 30 Jul 2025 22:39:23 +0000

Dave Mitchell wrote:
> In the context of updating perlxs.pod, I wanted to find which version of
> perl stopped requiring that all C variable declarations come before
> statements. So I wrote a short 'bad' XSUB and tested it against older
> perls, expecting it to give a C compiler error; but it worked back to at
> least perl 5.8.9. I'm probably doing something obviously wrong, but I
> can't see why.
>
> My two questions are:
>
> 1) Which perl release relaxed this restriction?
>
> 2) Why doesn't this XSUB give a C compiler error on older perls:
>
> void
> foo(int i = 0)
> CODE:
> int j = 0;

Get a copy of Visual C or aCC. FSF/GNU don't own the intellectual property rights to C.

In theory GCC can be commanded with a cmd line option to stop uses its GNU flavored C dialect. But I GCC impossible to use when its forced into ISO C mode. I find no legacy SW works with that run mode of GCC. None of my private SW works with GCC in ISO C mode, and nothing from CPAN works unmodified with GCC in ISO C mode. Its similar or more of a disaster than sending "C23" or "C99" or "C89" files into a C++ compiler.

If you PORT that C89-C23 CPAN XS lib to GCC in ISO mode, and leave the cmd line args inside your Makefile.PL, then yes GCC is an excellent way to validate that Visual C will compile your module from a Linux env. But no unmodified src code ever works with GCC ISO C mode.

Its not worth my time to spend 30 minutes coding+30 minutes ticket bureaucracy per CPAN tarball over and over again to get all of CPAN compatible with GCC ISO C mode.

Its a coin toss for me time wise if I want to port a CPAN .c/.xs to Strawberry GCC in ISO C mode, or start a Linux VM and transfer files. YMMV.

-DUSE_CPLUSPLUS fails 50% of CPAN tarballs IME. GCC in ISO C is worse since its the only reference unit test framework for the ISO C TC, and Clang/LLVM need GCC in ISO C to be strict enough to be the CI framework for Clang since there would be no other way to cross check the spec against the compiler (COUGH COUGH PERL 5).

GCC ISO C mode is a useful tool, but its exact settings doesn't represent any other known Windows or Commercial Unix C compiler for unmodified source code compatibility verification. But you get a guarantee your .c will work on VC and all legacy Commercial Unix CCs you never heard of.

Re: Magic v2 - Review for Merge and next steps by bulk 88

Wed, 30 Jul 2025 21:51:45 +0000

Leon Timmermans wrote:> It does point out an important challenge with the idea of value magic:
> our current set magic is *after* the assignment (which makes sense for
> variable magic), but value magic generally wants it to happen *before*
> the assignment. This alone would mean a vast amount of change in the
> codebase (various places that set SVs).

This is just a speed question.

If a MG* consumer was negligent to save a copy of SvPVX into mg->mg_ptr during the last GMG event sent to it, why would it be the interp's problem if contents inside SvPVX got nuked between the GMG event and SMG event?

If you single step TIEHASH and TIEARRAY, you will see they reversed the GMG, nuke SvPVX(), SMG, design pattern.

And created more Newx() blocks and mortal SV* heads vs the number of fingers I have for a single "my $zipcode = $hash->{zipcode};" statement.

>> Or maybe we could declare some values as immutable, without having to use constants? Here are a couple of examples that are annoying:
>>
>> use Readonly;
>> use Hash::Util qw(lock_hash);
>>
>> Readonly my $x => "hello";
>> my $y = $x; # $y is NOT read-only, infection doesn't happen
>>
>> my %h = (a => 1, b => 2);
>> lock_hash(%h);
>> my %copy = %h; # %copy is a regular hash, not locked
>
> I suspect that would fundamentally break the language. What should this do?

This makes no sense in Perl. A Perl $scalar can't be a copy paste of JavaScript's Object() class or String() class. A Perl $scalar is JavaScript's https://developer.mozilla.org/en-US/docs/Glossary/Primitive . JS's primitives are POD, you never need to write a try{}/catch{} around them. If they quack like a duck, they are a duck.

In JS the = operator usually does, and the .toString()/.toNumber()/.toValue() calls must strip dynamic-ism from the value. Value infection is supposed to be a C level secret. Its not the next BlockChain CryptoCurreny IDL SV VC startup.

There are enough of them already https://en.wikipedia.org/wiki/Interface_description_language

Perl has enough of them already https://metacpan.org/pod/Moose

Value infection is some kind of invisible "tag" set on a C level, and at some point later on some C code (XSUB/PPSUB/Perl's & sigil) can retrieve that invisible "tag" to separate 2 different SV*s addrs with identical contents.

Value infection is how nodeJS/JavaScript can automatically close your Unix FD or Unix socket when the last Pure JS reference falls out of scope, even though the ECMAScript spec forbids adding Perl's sub DESTROY {}. sub DESTROY {} definitely exists inside FireFox/SpiderMonkey on the JS-XS level. You are just simple prohibited from setting that C struct member from JS code.

If JavaScript didn't have a sub DESTROY {}, nodeJS and Deno and Bun would be unusable because you'd have to call libc's "free(ptr);" symbol manually everywhere. Javascript's mark and sweep algorithm runs randomly and unpredictably. I've read its either 300 millisecond wall clock timer running 24/7, but thats obsolete >= 2010 according to some blogs. The post-2010 algorithm is 30% or 70% of all physical memory according to the OS kernel's metrics. That can be 36 milliseconds away, 36 minutes away, or 36 days away. STDIO FDs/Unix FDs are value magic/value tag infection/sub DESTROY {}, along with Perl's SAVETMPS/FREETMPS/pp_nextstate fn ptr.

SpiderMonkey simply doesn't call SAVETMPS until you pass a an tag infected "POD" value through "=". The SM SAVETMPS also has to now de-optimize and swap the function pointer for pp_nextstate from the NOOP version to the heavyweight Perl version until the next scope leave so that the RC LL is flushed instantly, and not 36 hours from now, and the sub DESTROY{} gets fired on the C level that you don't know about.

Also Perl should stop using the words Mortal and Mortalize. The correct API terms to use are Rooted and Rooter.

C code using Value infection also has to be fault tolerant to get Value infection stripped. Value infection can't be a replace for Digital Rights Management libraries or the ultimate SW fix for https://en.wikipedia.org/wiki/Analog_hole

Good or best practices Perl code won't strip the infectious Value tag. Mediocre and student Perl code will instantly strip the infectious Value tag, and whatever was the producer/consumer library that made that tag has to tolerate those accidents happening.

> my %a
> my %b : Readonly = (a => 1, b => 2);
> %a = %b;
> %a = ();
>
>> Or someone tries to print a password or PII and it's printed as [Redacted]? Also, lots of debugging opportunities here.

Stripping Value infection tag logic can never lead to a CVE/heap corruption/SEGVs if there is an accidental stripping.

>> Readonly my $x => "hello";
>> my $y = $x; # $y is NOT read-only, infection doesn't happen

If this worked as written, "$y = 1;" represents a croak(PL_no_modify); or a SEGV. And this:

eval("\$z = '$y';");

Must flawlessly represent a croak(PL_no_modify); or a SEGV. Nothing else.

my $y = $x;
my $z;
$SIG{__DIE__) = sub {
warn 'inside the trap frame: probe='.$z.' err='.$@."\n";
$logger->WriteException(Carp::longmess(), $y, $z, $@, $!);
};
eval {
$z = $y;
$z = 1;
};

This also has to be removed from the Perl syntax because this is now a security exploit according to some InfoSec SME on X.
> How would that work? As an annotation that print specifically looks
> for? How would it work if someone appends that password to another
> string first?

use ---;
--- my $x => "hello";

is a great idea for TPC talk if someone uploads that module to CPAN.

Re: Magic v2 - Review for Merge and next steps by bulk 88

Wed, 30 Jul 2025 20:07:32 +0000

Paul "LeoNerd" Evans wrote:> TL;DR: Please help review my "magic v2" so far so I can add more useful
> abilities that will provide a good foundation for implementing a
> whole bunch of features on top of.
>
> I have a slowly-growing collection of ideas and partial experiments,
> some of which are written up in PPC documents, that all involve adding
> extra features, extra abilities, that would be useful in various
> situations. A lot of these are partly stalled for want of a decent way
> to encapsulate "extension behaviour". For example, a module or core
> feature that wanted to provide an attribute for subroutine signature
> parameters will need some way to provide extra code that may be
> required at various points during compilation or runtime.
If strategic goals are the purpose, or 1 API that is forever flexible for everyone. I have to ask, what are the design requirements and what use cases are specifically excluded from the proposed design?

Who is the intended customer of Magic v2?

P5P to P5P, P5P to CPAN XS, or P5P to PP users or CPAN XS vs CPAN XS (3P to 3P)?

Who or what group will coordinate and write interface specs for the producer and consumer? Producer and consumer are 2 different independent authors or maintainers.

Do producers need get a GUID/namespace from P5P to bless/announce their magic v2 class? Everyone knows the current system ran out of lower ASCII printable chars to use as class names or base class names.

#define PERL_MAGIC_uvar 'U' /* Available for use by extensions */
#define PERL_MAGIC_uvar_elem 'u' /* Reserved for use by extensions */
#define PERL_MAGIC_tied 'P' /* Tied array or hash */
#define PERL_MAGIC_tiedelem 'p' /* Tied array or hash element */
#define PERL_MAGIC_tiedscalar 'q' /* Tied scalar or handle */
#define PERL_MAGIC_extvalue '^' /* Value magic available for use by extensions */
#define PERL_MAGIC_ext '~' /* Variable magic available for use by extensions */

Only 7 classes or "interfaces" were reserved for end users.

The ptr to vtable trick allows finding out the class name/producer name for an unknown MG*, but it is currently impossible to find out, what interface is the producer. Yes, many TIMTOWTDI 1-off tricks exist for a dispatcher or consumer to find out what features/base class, this producer class is offering to the interp, or to PP, or to other CPAN XS code. U16 mg_private, or magic or sentinel SV* addrs, or magic void *s. The MAGIC struct has 2.5-3.25 void *s of dynamic data storage in it.

struct magic {
MAGIC* mg_moremagic;
MGVTBL* mg_virtual;
U16 mg_private;
char mg_type;
U8 mg_flags;
SSize_t mg_len;
SV* mg_obj;
char* mg_ptr;
};

I've never done it, and I've never seen anyone do this. MAGIC struct becomes 4 void *s of data storage, if there are extra const void *s hidden before or after the first 5 or 8 official fn ptrs, of the end user controlled MGVTBL* mg_virtual ptr.
> I have a bunch of existing modules that try to do variations on this
> kind of idea, by providing some sort of C-level data structure
> definition that extension modules can provide into the base system, to
> contain those extensions. There's a converging design among many of
> these, that it'd be most handy to have a data structure that contains
> function pointers, each of which is run at a particular phase of
> operation. These structures then define a kind of behaviour, provided
> by a module, that can be attached to various target instances.
Who dictates/specs out the behavior?

How many behaviors can there be in a perl process?

does the interp need to be recompiled?

or wait 1 year until the next stable release to "get a new behavior"?

can new behaviors be "backported" to stale/out of date interp versions?

can someone backport a new behavior from Perl 5.48 to Perl 5.46 and publish it on CPAN?

can 1 producer obj announce 2 behaviors?

or there must be 2 producer objs created by the same class factory (CPAN module)?

Can 1 data variable (SV*, AV*, HV*, or CV*) have 2 or more producers for same interface or same behavior attached to it simultaneously? AKA hooking.

Can producer objs legally (design-wise/API docs-wise, not real life) see how many other producer objs are nested/stacked above and below themselves on the linked list? A.K.A. caller() or PadWalker or @ISA.

Or is it that prohibited by API docs from day 1?
> In many ways this sounds exactly like Perl's "MAGIC". MAGIC is useful
> for adding trigger options to scalar variables, but has some
> shortcomings when trying to apply it to other situations. Perl core
> does attempt with some limited success to reuse the same magic virtual
> table definition on non-scalars (i.e. arrays, hashes and subroutines),
> but this is of limited utility. Furthermore, there's no ability to
> create kinds of magic that attach to the semantic idea of a "value",
> and follow it around as it gets copied and assigned to various other
> scalars.
Perl has this already

https://perldoc.perl.org/overload#Mathemagic,-Mutators,-and-Copy-Constructors

but overload.pm is as cool as smoking cigarettes with Joe Camel post year 2000. IDK why Perl Ecosystem/end users hate overload.pm with a passion.

> The reason that existing magic is rather limited in ability is because
> the entire set of behaviours provided by a kind of magic are
> encapsulated in a bunch of functions whose addresses are given in
> pointers in the "Magic virtual table" structure (MGVTBL), but that
> virtual table doesn't have any other fields to store metadata of any
> kind. It defines a fixed set of 8 different functions for various
> scalar-variable related scenarios, and makes it difficult to extend to
> other situations, as would apply to arrays, hashes, or other things.
Yep, no versioning, no length member at the top, no GUID at the top, what little versioning there was inside U8 mg_flags ran out very quick.

The current function pointers inside MGVTBL can be reworked, there is just no buy-in from anyone.

struct mgvtbl {
int (*svt_get) (pTHX_ SV *sv, MAGIC* mg);
int (*svt_set) (pTHX_ SV *sv, MAGIC* mg);
U32 (*svt_len) (pTHX_ SV *sv, MAGIC* mg);
int (*svt_clear) (pTHX_ SV *sv, MAGIC* mg);
int (*svt_free) (pTHX_ SV *sv, MAGIC* mg);
int (*svt_copy) (pTHX_ SV *sv, MAGIC* mg,
SV *nsv, const char *name, I32 namlen);
int (*svt_dup) (pTHX_ MAGIC *mg, CLONE_PARAMS *param);
int (*svt_local)(pTHX_ SV *nsv, MAGIC *mg);
};

Fn ptrs svt_len, svt_clear, and svt_copy, are dead code basically. They serve no purpose for $SCALARS, and svt_len, svt_clear, and svt_copy are exclusive to aggregates like TIEARRAY/TIEHASH, and PP's tie()'s API for TIEARRAY/TIEHASH is UNREACHABLE from XS.

https://github.com/Perl/perl5/blob/blead/hv.c#L3051
https://github.com/Perl/perl5/blob/9ef5300f9a0c33a7b865a307d84e06ada598c612/mg.c#L2280
https://github.com/Perl/perl5/blob/9ef5300f9a0c33a7b865a307d84e06ada598c612/mg.c#L2286
https://github.com/Perl/perl5/blob/9ef5300f9a0c33a7b865a307d84e06ada598c612/mg.c#L2106
https://github.com/Perl/perl5/blob/9ef5300f9a0c33a7b865a307d84e06ada598c612/mg.c#L2147

Don't have a CV*? Don't have a GV*? no HV* stash?

Tough luck. TIEHASH/TIEARRAY is a PP only API, C code is not allowed in.

I could modify the current TIEARRAY/TIEHASH CV* based API to be as fast as a "int(*svt_get)(pTHX_ SV *sv, MAGIC* mg);" style API with no visible C level changes for CPAN XS modules if someone was interested.

> The main idea of my "Magic v2" redesign therefore is to extend the
> definition of this vtable, with a whole new structure shape that has
> metadata fields for what "shape" it is, a version number, some flags,
> etc... Various different "shapes" are then defined for various
> situations. There's a "scalar variable" shape that mostly matches the
> uses of existing magic, with plenty of space for other shapes to be
> added with different sets of function pointers.
Are you using the word "shapes" the same way Chrome/Node use that word or a different meaning?
> Right now, I have this redesign in a branch that contains scalar, array
> and hash variable shapes, and a shape intended to apply to scalar
> *values*, with code around that tries to copy the magic attachment
> around when the value itself is copied.
Even though SvROK()/overload.pm already enable what you are describing, would any Perl users have real or imaginary "security" concerns over viral values being added to Perl 5?

I don't see any security concerns myself, but I can imagine some dissent over not being able to copy or serialize or "strip" the getter/setter methods away from the viral value, without round tripping it through JSON or a disk file. The complaints would be about whether viral values can see the my $vars of upper scopes with string eval, or that viral values can't be detected by a human developer using perl5db.pl's TUI interface.

Nobody wants all the time stamp $scalars in their %{@{$logging}} array to keep going up after calling " $start = time(); " and the IV inside $start continues to tick 1x a second.
> Of these, the very last one - creating the concept of scalar value
> magic - still has a number of performance-related questions that have a
> measurable impact on the operation of the interpreter even in programs
> that do not make use of it. So that part at least isn't ready for
> consideration of merge into blead at this time.
Why would it be a performance problem? there are only 2 places inside libperl to add that. newSVsv() and sv_setsv().

Is your performance problem actually the dozens or low 100s of places where the interp or XS does " newSVpvn(SvPVX(old), SvCUR()); " instead of " newSVsv(old) "?

>
> However, now that we can start considering things to look into bringing
> into development for (5.)43.1, I would like to propose the basic part
> of Magic v2 (the base mechanism with the ability to have different
> shapes), as well as the stable parts of scalar/array/hash variable
> magic. All of these are just neatened-up rewrites of existing
> behaviours, but made in a way as to allow a lot of extension in future.
>
> Once these parts are merged, we'll be in a much better place to consider
> ability to add magic in other situations around subroutines, sub
> signature parameters, object fields, and a whole bunch of other places.
> Several of these overlap some existing PPC documents I have open, and
> would be a good step towards implementing them, so there's a lot that
> can be built on top of it.
>
> Now that the next development cycle can start, I would like to begin
> with a review and discussion on the current set of code as it stands. It
> lives here on this branch, which I periodically rebase onto `blead`:
>
> https://github.com/leonerd/perl5/tree/hooks
mg.h:

enum HookShape {
HKs_BASE,
HKs_SCALARVAR,
HKs_SCALARVALUE,
HKs_ARRAYVAR,
HKs_HASHVAR,
};

How do you or future P5P devs EOL/retire 1 or 2 but not all 5 base class/interface from this V2 enum, when MAGIC v3 is created from MAGIC V2?

This also means your implementation can't grow just like Perl_sv_magic() can't grow beyond 128-32. And 2 3rd party CPAN XS libs with 2 different author can't add their own types to the HookShape registry.

/* common to all hook function structs */
#define _PERL_HOOKFUNCTIONS_COMMON_FIELDS \
MGVTBL _v1_vtbl; /* reserve space */ \
U32 ver; \
enum HookShape shape; \
U32 flags; \
const char *debug_name; \
size_t user_size; \
void (*free) (pTHX_ SV *sv, MAGIC *mg); \
void (*clone)(pTHX_ SV *osv, MAGIC *omg, SV *nsv, MAGIC *nmg, CLONE_PARAMS *params)

are you going to drop out func ptr "clone" struct member on no thd perls?

Why not declare these fn ptrs as varargs from day 1?

That was a huge mistake Larry made with MAGIC V1 and the CV * prototypes.

+static int
+XS_SDBM_File_sdbm_DESTROY(ix, sp, items)
+register int ix;
+register int sp;
+register int items;
+{
+ if (items < 1 || items > 1) {
+ croak("Usage: SDBM_File::DESTROY(db)");
+ }
+ {
+ SDBM_File db;
+
+ if (sv_isa(ST(1), "SDBM_File"))
+ db = (SDBM_File)(unsigned long)SvNV((SV*)SvANY(ST(1)));
+ else
+ croak("db is not of type SDBM_File");
+ sdbm_close(db);
+ }
+ return sp;
+}

Actually it someone after Larry, got rid of Larry's original design, and made it worse by transporting dSP and dAXMARK in C global variables, instead of transporting them on the C stack/C's ABI CPU registers.

Arg 3 items from Larrys design can be removed tho. Perl's I32 items, is inefficient on all 3 modern C compilers.

83 CC 23 8B ADD X3, X4, W3,SXTW#3 << 3 shift left ax to SV**
42 00 03 CB SUB X2, X2, X3 SP - MARK
42 FC 43 93 ASR X2, X2, #3 shift right >> 3
5F 04 00 71 CMP W2, #1
C1 0A 00 54 B.NE loc_20031C
loc_20031C
E2 0E 00 F0 ADRP X2, #Perl_croak_xs_usage_ptr@PAGE
42 B0 45 F9 LDR X2, [X2,#Perl_croak_xs_usage_ptr@PAGEOFF]
E0 03 01 AA MOV X0, X1
21 01 00 90 ADRP X1, #aLibref@PAGE ; "libref"
21 00 25 91 ADD X1, X1, #aLibref@PAGEOFF ; "libref"
40 00 3F D6 BLR X2

MGVTBL _v1_vtbl; /* reserve space */ \ 8
U32 ver; \4
enum HookShape shape; \4
U32 flags; \4
ALIGNMENT PROBLEM
const char *debug_name; 8

I think I see 4 missing bytes in this struct.

const char *debug_name; \

Bad idea for Perl. Okay idea for -DDEBUGGING Perl, but these 8 bytes are useless for production PP/XS apps. Only ChatGPT and Monkeys with Mice and Monitors can decrypt the data stored inside
"char *debug_name;". Field "char *debug_name;" should be a SV* or HV* or a ptr to a static RO or static RW C struct.

"char *debug_name;" needs to be an analog of SV*'s SvSTASH(), or be a C struct with much more broken down data. It needs to be machine readable. Basically "char *debug_name;" is how to programmatically find out which class/XS module created this MG V2 object.

"char *debug_name;" is going to be as useless as GvFILE() or CvFILE() is right now. Unreachable from PP, and Devel::Peek's Dump() is the only user of the CvFILE() member in the Perl Ecosystem.

size_t user_size; \

It can be turned into a U32 or U16 to fix the alignment problem. A 6 GBs long vtable is insane.

typedef struct {
MAGIC _magic;
IV keyiv;
} MAGICWithKeyIV;

typedef struct {
MAGIC _magic;
SV *keysv;
} MAGICWithKeySV;

Why are keyiv and keysv on the bottom and not on the top? Type IV is larger than type SV* on all i386 Perls except Visual C Perl.

I see a back-forwards compatibility problem here.

You also aren't solving the alignment problem with

struct magic {
MAGIC* mg_moremagic;
MGVTBL* mg_virtual; /* pointer to magic functions */
U16 mg_private;
char mg_type;
U8 mg_flags;
SSize_t mg_len;
SV* mg_obj;
char* mg_ptr;
};

This post is long enough. MAGIC V2 needs to follow the design of C++ getter/setter operator overload methods. Not repeat every single C level mistake that was done in Perl 5.000 Alpha 001.

MAGIC V1 was not even whiteboard-ed as a public API for random CPAN C ppl to use. It was a private API between libperl.so and DBM_File.xs, NDBM_File.xs, ODBM_File.xs, GDBM_File.xs, SDBM_File.xs and Curses.xs after these 6 TUs/.c/.xs/.us files were kicked out of libperl.so and stopped being "keywords with 2 :: characters" and started being plugins.

I only spent 2 hours reading the MAGIC V2 branch and stopped, but I see MAGIC V2 does nothing about call_sv()/"TIESCALAR" being 14x slower than writing:

PUSHMARK(SP);
CvXSUB(cv)(aTHX_ cv);

All these problems need to be fixed for any MAGIC V2 attempt.

The best practices are whatever Python/PHP/Ruby/Rust/Chrome/Firefox/Dalvik/Hotspot/CLR/C++ STL/ObjC/Swift are doing right now on the .i file level.

MAGIC V1 currently looks like Windows 95 attempting to link C function written in i386 language to C function written in i286 language inside the same address space. Win95 never had access to DBT or https://en.wikipedia.org/wiki/Rosetta_(software) technology.

Perl 5 Commit Summary by Perl 5 commit summary

Wed, 30 Jul 2025 03:11:48 +0000

Perl 5 commit summary, activity since Saturday

Current branch blead
26 commits. 5 unique authors. 4 unique committers.
28 files changed, 803 insertions(+), 1144 deletions(-)
Thanks, applied: Karl Williamson (1)
Snapshot: http://github.com/Perl/perl5/tarball/e31d3ed5601e07f2

Change core-team email address for jkeenan
James E Keenan 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/e31d3ed5601e07f2

Convert RT ticket reference to GH
James E Keenan 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/148136c313ee439b

sv.c: make PUSH_EXTEND_MORTAL__SV_C an inline function
Lukas Mai 1 file changed, 16 insertions(+), 15 deletions(-
https://github.com/Perl/perl5/commit/97d0d863372c5eff

perldelta for 'catch (my $e)' diagnostic
Lukas Mai 1 file changed, 16 insertions(+)
https://github.com/Perl/perl5/commit/d1bb39457a8845dd

toke.c: explicitly diagnose 'catch (my $e)'
Lukas Mai 3 files changed, 65 insertions(+), 8 deletions(-
https://github.com/Perl/perl5/commit/88ff8dafb1a058b1

perly.y, op.c: PL_in_my is an enum (kind of), not a bool
Lukas Mai 5 files changed, 9 insertions(+), 9 deletions(-)
https://github.com/Perl/perl5/commit/9e492c255d5e10e2

Reapply "Attempt to remove deprecated sv_locking/unlocking"
Karl Williamson 6 files changed, 3 insertions(+), 68 deletions(-
https://github.com/Perl/perl5/commit/2b8cbf5b9e4d6ee9

Make sure mathoms gets loaded
Karl Williamson 5 files changed, 25 insertions(+)
https://github.com/Perl/perl5/commit/0f6279cac483c346

Rm Perl_sv_pv(byte)? from mathoms
Karl Williamson 4 files changed, 11 insertions(+), 49 deletions(
https://github.com/Perl/perl5/commit/5df7a2e2277b1803

Rm Perl_newHV() from mathoms
Karl Williamson 4 files changed, 7 insertions(+), 15 deletions(-
https://github.com/Perl/perl5/commit/c13b3b863b3836be

Rm Perl_sv_pvutf8() from mathoms
Karl Williamson 4 files changed, 7 insertions(+), 16 deletions(-
https://github.com/Perl/perl5/commit/10e38d94b58fffcb

Rm Perl_sv_utf8_upgrade() from mathoms
Karl Williamson 4 files changed, 6 insertions(+), 18 deletions(-
https://github.com/Perl/perl5/commit/a810de6508814172

Convert Perl_hv_foo to hv_foo and rm from mathoms
Karl Williamson 4 files changed, 25 insertions(+), 98 deletions(
https://github.com/Perl/perl5/commit/5144e0d9e6751711

Make Perl_sv_setsv a macro and rm from mathoms
Karl Williamson 4 files changed, 6 insertions(+), 18 deletions(-
https://github.com/Perl/perl5/commit/c1ae415246f1e0bd

Make Perl_sv_mortalcopy a macro and rm from mathoms
Karl Williamson 4 files changed, 7 insertions(+), 12 deletions(-
https://github.com/Perl/perl5/commit/ca4ce29c796a6752

mathoms.c Remove functions that just call their macro
Karl Williamson 4 files changed, 249 insertions(+), 572 deletion
https://github.com/Perl/perl5/commit/a5fc9ab124f4f767

mathoms.c: Convert Perl_foo definitions to call foo
Karl Williamson 1 file changed, 41 insertions(+), 49 deletions(-
https://github.com/Perl/perl5/commit/5b36fc695957614a

mathoms.c Remove functions that just call their macro
Karl Williamson 4 files changed, 88 insertions(+), 164 deletions
https://github.com/Perl/perl5/commit/b7a7434345a06e5d

Move pod from mathoms.c to respective files
Karl Williamson 2 files changed, 50 insertions(+), 62 deletions(
https://github.com/Perl/perl5/commit/a3bd1cdbc1851618

regen/embed.pl: Fix auto longname gen when no arglist
Karl Williamson 1 file changed, 3 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/876fe29b1de46f7d

embed.fnc: Update E flag description
Karl Williamson 1 file changed, 2 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/96a7727df0813ba5

embed.fnc: _invlist_union() not publicly accessible
Karl Williamson 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/a81c01f5c1cd0880

op.h/op.c/peep.c - Optimise away empty if{} blocks
Richard Leach 5 files changed, 148 insertions(+), 13 deletions
https://github.com/Perl/perl5/commit/e7778069bc401e57

op.c/peep.c - Optimise away empty else{} blocks
Richard Leach 6 files changed, 57 insertions(+), 8 deletions(-
https://github.com/Perl/perl5/commit/75a310dd6d6a8cfc

Silence autodoc warning
Karl Williamson 1 file changed, 2 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/613101d7fb08cecc

todo test for 16008 (truncate scalar fh)
brian d foy 1 file changed, 15 insertions(+)
https://github.com/Perl/perl5/commit/ca40375cd6658b54

Current branch maint-5.38
1 commit. 1 unique author. 1 unique committer.
1 file changed, 1 insertion(+), 1 deletion(-)

Snapshot: http://github.com/Perl/perl5/tarball/3b9631c726a54ceb

Minor correction to perldelta
Steve Hay 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/3b9631c726a54ceb

Current branch maint-5.40
1 commit. 1 unique author. 1 unique committer.
1 file changed, 1 insertion(+), 1 deletion(-)

Snapshot: http://github.com/Perl/perl5/tarball/f7d5d15fdf279c83

Minor correction to perldelta
Steve Hay 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/f7d5d15fdf279c83

New branch haarg/db-sub-overload-fixes
3 commits. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/a2b1ab5d3784e8f9

bump perl5db VERSION
Graham Knop 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/a2b1ab5d3784e8f9

perl5db: only check sub names for non-references
Graham Knop 1 file changed, 4 insertions(+), 4 deletions(-)
https://github.com/Perl/perl5/commit/02494a5da6d05d6a

perl5db: use a lexical copy of $DB::sub inside DB::sub
Graham Knop 1 file changed, 6 insertions(+)
https://github.com/Perl/perl5/commit/a71385d4b292a5f6

New branch parsexs-length-t_pv
3 commits. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/e2140899e7e27d98

Explicitly reject length(NAME) with typemaps other than T_PV
Leon Timmermans 3 files changed, 30 insertions(+), 2 deletions(-
https://github.com/Perl/perl5/commit/e2140899e7e27d98

Bump ExtUtils::ParseXS to 3.59
Leon Timmermans 12 files changed, 12 insertions(+), 12 deletions
https://github.com/Perl/perl5/commit/975d8393e11814c6

Pull in ExtUtils::ParseXS Changes file from CPAN
Leon Timmermans 1 file changed, 9 insertions(+)
https://github.com/Perl/perl5/commit/7c58ce3da0a7371e

Deleted branch sv_dup_doc_fixes

Martian commit 1db8709adb69c0e7dfaa52d3819ab0523cdef945
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/1db8709adb69c0e7

sv.c: make PUSH_EXTEND_MORTAL__SV_C an inline function
Lukas Mai 1 file changed, 16 insertions(+), 15 deletions(-
https://github.com/Perl/perl5/commit/1db8709adb69c0e7

Martian commit 6cb2c8f258ffca17db140e2b47b17f4d986a4c28
3 commits. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/6cb2c8f258ffca17

perldelta for 'catch (my $e)' diagnostic
Lukas Mai 1 file changed, 16 insertions(+)
https://github.com/Perl/perl5/commit/6cb2c8f258ffca17

toke.c: explicitly diagnose 'catch (my $e)'
Lukas Mai 3 files changed, 65 insertions(+), 8 deletions(-
https://github.com/Perl/perl5/commit/c43d77915505bb24

perly.y, op.c: PL_in_my is an enum (kind of), not a bool
Lukas Mai 5 files changed, 9 insertions(+), 9 deletions(-)
https://github.com/Perl/perl5/commit/5cbe29418e1b34e2

Martian commit 2a93cd9a80f473c04a9873fc803232afe83b6632
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/2a93cd9a80f473c0

Add ASSERT_() a legal synonym for __ASSERT_()
Karl Williamson 1 file changed, 18 insertions(+), 7 deletions(-)
https://github.com/Perl/perl5/commit/2a93cd9a80f473c0

Martian commit c6a413a86e0ed5161e133164762f4e98cb7a0931
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/c6a413a86e0ed516

Convert RT ticket reference to GH
James E Keenan 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/c6a413a86e0ed516

Martian commit e7a6bb12a8feeb315e209446ad63cfacc8c9a39f
2 commits. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/e7a6bb12a8feeb31

Reapply "Attempt to remove deprecated sv_locking/unlocking"
Karl Williamson 6 files changed, 3 insertions(+), 68 deletions(-
https://github.com/Perl/perl5/commit/e7a6bb12a8feeb31

Make sure mathoms gets loaded
Karl Williamson 5 files changed, 25 insertions(+)
https://github.com/Perl/perl5/commit/abaa2d40e327e00e

Martian commit 9a66bc24e378a29d00d9ec158a22e550f4f8826c
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/9a66bc24e378a29d

Mark do_open(), do_openn() as internal, not public
Karl Williamson 1 file changed, 2 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/9a66bc24e378a29d

Martian commit cea3b6314d395edc0319814cd22f1d641faeef55
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/cea3b6314d395edc

Silence autodoc warning
Karl Williamson 1 file changed, 2 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/cea3b6314d395edc

Martian commit e5b6b48ff2e159436f9e73a94f833e9f12d63321
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/e5b6b48ff2e15943

todo test for 16008 (truncate scalar fh)
brian d foy 1 file changed, 15 insertions(+)
https://github.com/Perl/perl5/commit/e5b6b48ff2e15943

Martian commit 63442f8646cf2635aec9cb8bdbe92c34519450f5
12 commits. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/63442f8646cf2635

Rm Perl_sv_pv(byte)? from mathoms
Karl Williamson 4 files changed, 11 insertions(+), 49 deletions(
https://github.com/Perl/perl5/commit/63442f8646cf2635

Rm Perl_newHV() from mathoms
Karl Williamson 4 files changed, 7 insertions(+), 15 deletions(-
https://github.com/Perl/perl5/commit/d8da93e4c791c238

Rm Perl_sv_pvutf8() from mathoms
Karl Williamson 4 files changed, 7 insertions(+), 16 deletions(-
https://github.com/Perl/perl5/commit/7a875276716123b6

Rm Perl_sv_utf8_upgrade() from mathoms
Karl Williamson 4 files changed, 6 insertions(+), 18 deletions(-
https://github.com/Perl/perl5/commit/56e05b3dafa20957

Convert Perl_hv_foo to hv_foo and rm from mathoms
Karl Williamson 4 files changed, 25 insertions(+), 98 deletions(
https://github.com/Perl/perl5/commit/303b7a57dafc8caf

Make Perl_sv_setsv a macro and rm from mathoms
Karl Williamson 4 files changed, 6 insertions(+), 18 deletions(-
https://github.com/Perl/perl5/commit/1f8422479bd038d4

Make Perl_sv_mortalcopy a macro and rm from mathoms
Karl Williamson 4 files changed, 7 insertions(+), 12 deletions(-
https://github.com/Perl/perl5/commit/8785e0c90c04ce2b

mathoms.c Remove functions that just call their macro
Karl Williamson 4 files changed, 249 insertions(+), 572 deletion
https://github.com/Perl/perl5/commit/2b75c8bbd6114346

mathoms.c: Convert Perl_foo definitions to call foo
Karl Williamson 1 file changed, 41 insertions(+), 49 deletions(-
https://github.com/Perl/perl5/commit/234ef7879274c293

mathoms.c Remove functions that just call their macro
Karl Williamson 4 files changed, 88 insertions(+), 164 deletions
https://github.com/Perl/perl5/commit/c73cbe9258896a08

Move pod from mathoms.c to respective files
Karl Williamson 2 files changed, 50 insertions(+), 62 deletions(
https://github.com/Perl/perl5/commit/c85bd5195f7e30a3

regen/embed.pl: Fix auto longname gen when no arglist
Karl Williamson 1 file changed, 3 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/5f86eaf38bd5857c

Ignored 106 GitHub auto-generated merge commits

Re: no 'declaration after statement' error by Dave Mitchell

Tue, 29 Jul 2025 13:17:44 +0000

On Tue, Jul 29, 2025 at 12:54:29PM +0100, Paul "LeoNerd" Evans wrote:
> On Tue, 29 Jul 2025 12:42:44 +0100
> Dave Mitchell <[email protected]> wrote:
>
> > In the context of updating perlxs.pod, I wanted to find which version
> > of perl stopped requiring that all C variable declarations come before
> > statements. So I wrote a short 'bad' XSUB and tested it against older
> > perls, expecting it to give a C compiler error; but it worked back to
> > at least perl 5.8.9. I'm probably doing something obviously wrong,
> > but I can't see why.
>
> It'll depend on the C compiler. gcc is very lax with allowing that even
> if you didn't explicitly pass `-std=c99`.

Having looked further, its appears that when building *perl*, we pass these
flags to gcc:

-Werror=declaration-after-statement pre 5.36.0
-Wno-declaration-after-statement post 5.36.0

but these flags *aren't* passed on to the C compiler by default when
building XS modules. So XS modules won't complain on gcc, but might on
other compilers.

Does that sound right?

--
"But Sidley Park is already a picture, and a most amiable picture too.
The slopes are green and gentle. The trees are companionably grouped at
intervals that show them to advantage. The rill is a serpentine ribbon
unwound from the lake peaceably contained by meadows on which the right
amount of sheep are tastefully arranged." -- Lady Croom, "Arcadia"

Re: no 'declaration after statement' error by Paul "LeoNerd" Evans

Tue, 29 Jul 2025 11:54:42 +0000

On Tue, 29 Jul 2025 12:42:44 +0100
Dave Mitchell <[email protected]> wrote:

> In the context of updating perlxs.pod, I wanted to find which version
> of perl stopped requiring that all C variable declarations come before
> statements. So I wrote a short 'bad' XSUB and tested it against older
> perls, expecting it to give a C compiler error; but it worked back to
> at least perl 5.8.9. I'm probably doing something obviously wrong,
> but I can't see why.

It'll depend on the C compiler. gcc is very lax with allowing that even
if you didn't explicitly pass `-std=c99`.

--
Paul "LeoNerd" Evans

[email protected]
http://www.leonerd.org.uk/ | https://metacpan.org/author/PEVANS

no 'declaration after statement' error by Dave Mitchell

Tue, 29 Jul 2025 11:42:51 +0000

In the context of updating perlxs.pod, I wanted to find which version of
perl stopped requiring that all C variable declarations come before
statements. So I wrote a short 'bad' XSUB and tested it against older
perls, expecting it to give a C compiler error; but it worked back to at
least perl 5.8.9. I'm probably doing something obviously wrong, but I
can't see why.

My two questions are:

1) Which perl release relaxed this restriction?

2) Why doesn't this XSUB give a C compiler error on older perls:

void
foo(int i = 0)
CODE:
int j = 0;

which gets translated to this C code:

XS_EUPXS(XS_Foo__Bar_foo)
{
dVAR; dXSARGS;
if (items < 0 || items > 1)
croak_xs_usage(cv, "i= 0");
{
int i;

if (items < 1)
i = 0;
else {
i = (int)SvIV(ST(0))
;
}
#line 22 "Bar.xs"
int j = 0;
#line 188 "Bar.c"
}
XSRETURN_EMPTY;
}

where as AFAICT, the 'int j' declaration *does* follow a statement.

--
It's not that I'm afraid to die, I just don't want to be there when it
happens.
-- Woody Allen

Re: Wierd behaviour of map when trying to fill a hash by Ovid

Mon, 28 Jul 2025 16:06:37 +0000

On Sat, Jul 26, 2025 at 9:19â€¯PM Russ Allbery <[email protected]> wrote:

> Jan Jansen via perl5-porters <[email protected]> writes:
>
> > Running the map construct to create a hash behaves oddly under specific
> > circumstances.
>
> > Typical expected behaviour to create a hash reference for a list
> > $hashref = { map { <instructions>; $key => $value } @list };
>
> > What work :
> > $hashref = { map { 'key_'.$_ => $raw->{$_} } keys %$raw };
> > $hashref = { map { ("key_.$_" => $raw->{$_}) } keys %$raw };
>
> > What generates a syntax error :
> > $hashref = { map { "key_$_" => $raw->{$_} } keys %$raw };
>
> I believe this is the known edge case in the Perl parser that is
> documented in perldoc -f map:
>
> <snip>
>
> my %hash = map {; "\L$_" => 1 } @array # this also works
>

Putting the semicolon after the opening curly, to me, is the most
consistent and understandable approach to this problem. It's also how I
learned to spot Abigail's code when I used to work with them. I thought it
was silly at the time. Once again, Abigail was right and I was not.

Best,
Ovid
--
https://curtispoe.org/

Re: Wierd behaviour of map when trying to fill a hash by Uri Guttman

Mon, 28 Jul 2025 15:19:58 +0000

On 7/23/25 06:58, Jan Jansen via perl5-porters wrote:
> Typical expected behaviour to create a hash reference for a list
> $hashref = { map { <instructions>; $key => $value } @list };
>
> What work :
> $hashref = { map { 'key_'.$_ => $raw->{$_} } keys %$raw };
> $hashref = { map { ("key_.$_" => $raw->{$_}) } keys %$raw };
>
> What generates a syntax error :
> $hashref = { map { "key_$_" => $raw->{$_} } keys %$raw };
>
> **Steps to Reproduce**
> 
>
> $ perl -e '{ map { "key_$_" => $raw->{$_} } (keys %$raw) };'
> syntax error at -e line 1, near "} ("
> Execution of -e aborted due to compilation errors.
>
> **Expected behavior**
> 
> Similar to
> $ perl -e '{ map { 'key_'.$_ => $raw->{$_} } (keys %$raw) };'
> No syntax error !
>

both of these work have no syntax errors. perl has an issue determining
whether you want map to call a code block or build an anon hash where
you have {}. putting a comma after the block says you want a hash ref.
putting the code inside () and a comma tells map it is code to run.

perl -e '{ map { "key_$_" => $raw->{$_} }, (keys %$raw) };' # makes a
hash ref inside the anon hash
perl -e '{ map ( "key_$_" => $raw->{$_} ), (keys %$raw) };' # makes what
you want

uri

--
https://uriguttman.blogspot.com/
A Long Strange Trip
A blog about computers, food, my life and silliness.

Re: purpose of t/run/todo.t by Karl Williamson

Sun, 27 Jul 2025 18:08:43 +0000

On 7/27/25 05:35, Ovid wrote:
> On Sun, Jul 27, 2025 at 11:59 AM Dave Mitchell <[email protected]
> <mailto:[email protected]>> wrote:
>
> I still don't "get" the purpose of t/run/todo.t. What's the difference
> between adding a TODO to that generic file for a reported but not yet
> fixed issue, as opposed to the traditional approach of adding a TODO
> test
> to the relevant .t file which tests that aspect of perl?
>
> When should we add to which file?

The purpose of this file is to encourage newcomers to contribute without
the additional headache of their having to figure out our internal
structure and conform to it.

It seems to me that this is a way to get people started. Just look
through the open tickets, and convert the failure that's in the text of
the ticket to something that is in our code base of tests. You don't
have to know very much to do this. And maybe they will get hooked on
contributing.

I figured that once fixes were added, the fixer could move the test to
the appropriate .t.

I hadn't thought of the possibility of these duplicating an existing
todo test; I rarely see any of those. What I think could be done is to
grep for them, find their ticket numbers, and add that list to
t/run/todo.t as comments, so that people would see that and not
duplicate what's already been done.

It would be nice if some people would take it on themselves to convert
the text of failures in a ticket into an actual executable test, as
those tickets come in. A high skill level isn't generally required.

But we have 2K open tickets already, and so I hoped to get people to
start on those, without worrying about new ones coming along.

I would expect an experienced contributor to add tests in the files they
deem best; I can't remember the last time I saw a TODO added for an old
ticket by any of our team. Then there's no conflict.

I'm open to other possibilities that don't impose an extra cognitive
load on the people I'm hoping to get to contribute.

>
>
> On Sun, Jul 27, 2025 at 11:59 AM Dave Mitchell <[email protected]
> <mailto:[email protected]>> wrote:
>
> I still don't "get" the purpose of t/run/todo.t. What's the difference
> between adding a TODO to that generic file for a reported but not yet
> fixed issue, as opposed to the traditional approach of adding a TODO
> test
> to the relevant .t file which tests that aspect of perl?
>
> When should we add to which file?
>
>
> I get tired of walking into shops and watching developers grep to see if
> they can find tests for a feature. I can't say it applies here, but in a
> number of my consulting contracts, I've found that reorganizing test
> suites to mirror main file structure (instead of functionality) made it
> dead easy to find where the tests are for a given thing.
>
> .
> ├── lib
> │ ├── Foo.pm
> │ ├── Bar.pm
> │ └── Foo
> │ └── Baz.pm
> │
> └── t
> └── tests
> ├── bar.t
>    ├── foo.t
>    └── foo
>    └── baz.t
>
> That works great for many tests and would make it clear where the TODO
> tests would be (though I use xUnit instead of .t tests, but it's the
> same principle).
>
> Best,
> Ovid
> --
> https://curtispoe.org/ <https://curtispoe.org/>

Re: Perl 5.40.3-RC1 is now available! by Steve Hay via perl5-porters

Sun, 27 Jul 2025 15:28:01 +0000

On Sun, 27 Jul 2025 at 11:32, Dan <[email protected]> wrote:
>
> Small note, in the perldelta for 5.40.3 RC1 and 5.38.5 RC1, they reference "Perl 5.42 will no longer chdir to each handle." This should be updated to reference these maintenance releases which presumably will also exhibit this behavior.
>

Fixed. Thanks for spotting this!

Re: purpose of t/run/todo.t by Ovid

Sun, 27 Jul 2025 11:35:52 +0000

On Sun, Jul 27, 2025 at 11:59â€¯AM Dave Mitchell <[email protected]> wrote:

> I still don't "get" the purpose of t/run/todo.t. What's the difference
> between adding a TODO to that generic file for a reported but not yet
> fixed issue, as opposed to the traditional approach of adding a TODO test
> to the relevant .t file which tests that aspect of perl?
>
> When should we add to which file?
>

On Sun, Jul 27, 2025 at 11:59â€¯AM Dave Mitchell <[email protected]> wrote:

> I still don't "get" the purpose of t/run/todo.t. What's the difference
> between adding a TODO to that generic file for a reported but not yet
> fixed issue, as opposed to the traditional approach of adding a TODO test
> to the relevant .t file which tests that aspect of perl?
>
> When should we add to which file?

I get tired of walking into shops and watching developers grep to see if
they can find tests for a feature. I can't say it applies here, but in a
number of my consulting contracts, I've found that reorganizing test suites
to mirror main file structure (instead of functionality) made it dead easy
to find where the tests are for a given thing.

.
â”œâ”€â”€ lib
â”‚ â”œâ”€â”€ Foo.pm
â”‚ â”œâ”€â”€ Bar.pm
â”‚ â””â”€â”€ Foo
â”‚ â””â”€â”€ Baz.pm
â”‚
â””â”€â”€ t
â””â”€â”€ tests
â”œâ”€â”€ bar.t
â”œâ”€â”€ foo.t
â””â”€â”€ foo
â””â”€â”€ baz.t

That works great for many tests and would make it clear where the TODO
tests would be (though I use xUnit instead of .t tests, but it's the same
principle).

Best,
Ovid
--
https://curtispoe.org/

Re: Perl 5.40.3-RC1 is now available! by Dan

Sun, 27 Jul 2025 10:32:49 +0000

On Mon, Jul 21, 2025 at 4:26â€¯PM Steve Hay via perl5-porters <
[email protected]> wrote:

> In 1522, thirty years after Columbus' landfall in the Bahamas, another
> ship sailing west in the service of Spain completed the first voyage
> round the world. It had set out under the command of a Portuguese,
> Magellan, who had got as far as the Philippines, where he was killed.
> By then he had discovered and sailed through the straits named after him
> and with this voyage and its demonstration that the great oceans were
> all interconnected, the prologue to the European age can be considered
> over.
>
> -- J. M. Roberts, "The Penguin History of Europe"
>
> We are pleased to announce the first release candidate of version 40.3,
> the third maintenance release of version 40 of Perl 5.
>
> You will soon be able to download Perl 5.40.3-RC1 from your favourite
> CPAN mirror or find it at:
>
> https://metacpan.org/release/SHAY/perl-5.40.3-RC1/
>
> SHA256 digests for this release are:
>
> *perl-5.40.3-RC1.tar.gz:
> f0771e49db212191347589da4315ca5e002105493e08372b525201f015d15ec1
>
> *perl-5.40.3-RC1.tar.xz:
> e0d3e4555458cbfbbd994d024d676150286d8f8dcd48a1deeafcab21a64783d6
>
> You can find a full list of changes in the file "perldelta.pod" located
> in the "pod" directory inside the release and on the web.
>
> ***Please, please, please*** test your code against perl-5.40.3-RC1.
> This is your last chance to point out any critical regressions before
> v5.40.3 ships "for real," and we'd rather fix things for v5.40.3 than
> for v5.40.4!
>
> The final v5.40.3 is expected to ship on 3 August.
>
> Steve Hay
>

Small note, in the perldelta for 5.40.3 RC1 and 5.38.5 RC1, they reference
"Perl 5.42 will no longer chdir to each handle." This should be updated to
reference these maintenance releases which presumably will also exhibit
this behavior.

-Dan

purpose of t/run/todo.t by Dave Mitchell

Sun, 27 Jul 2025 09:59:57 +0000

Hi Karl,
I still don't "get" the purpose of t/run/todo.t. What's the difference
between adding a TODO to that generic file for a reported but not yet
fixed issue, as opposed to the traditional approach of adding a TODO test
to the relevant .t file which tests that aspect of perl?

When should we add to which file?

--
My Dad used to say 'always fight fire with fire', which is probably why
he got thrown out of the fire brigade.

Re: New XS features? by Leon Timmermans

Sun, 27 Jul 2025 01:28:57 +0000

On Mon, Jul 7, 2025 at 7:45â€¯PM Leon Timmermans <[email protected]> wrote:
>
> Given Dave's recent refactoring of ExtUtils::ParseXS I suspect we are
> now in a much better position to actually add new features, instead of
> being stuck in essentially the feature set we had in the late 90s.
>
> In particular I would like to propose two small features.
>
> 1. I want a way to declare that all XSUBs with a CODE section should
> be assumed to be OUTPUT: RETVAL, unless their return type is void. I
> understand why it isn't the default conceptually, but in my experience
> it's what people want 99% of the time. It should probably be
> overridable with an empty OUTPUT section for the rare case when it
> isn't.
>
> 2. I want a simple way to declare constants. We currently have
> ExtUtils::Constant for this sort of thing, but it's overly complicated
> for most use-cases, and putting the constants in Makefile.PL instead
> of the XS files feels entirely wrong. This is an easy problem that
> should have an easy solution. Something like:
>
> CONST: UV MY_CONSTANT
> or
> CONST: IV MY_CONSTANT = 3

Currently I'm experiencing two problems around strings that could use
better solutions than what is currently available.

1. The char pointer typemaps currently do not do SV_CHECK_THINK_FIRST,
and hence can very easily mess up COW strings. But obviously should
only do so selectively. Fortunately this scenario is probably uncommon
(given one often has to resize or some such, one usually needs the
full SV anyway).

2. The char pointer typemaps predate unicode support, and therefore
code using them has the unicode bug. One really needs to handle these
separately (SvPVbyte versus SvPVutf8). Using two macro-redefines one
can sort-of fix this for the whole module, but that becomes a problem
when you need both kinds in the same module.

One approach out of this predicament would be to use separate typemaps
for all these permutations. A major problem with that is that the
length() feature is hardcoded to only work with the `T_PV` typemap.
Making that work would probably require adding a new type of map
handling input with length.

Another approach might be modifiers on the types (much like IN/OUT/â€¦
already exist nowadays). This feels more flexible (given that the
above approach gives increasingly more permutations) but puts more
weight on ParseXS.

Leon

Re: Wierd behaviour of map when trying to fill a hash by Russ Allbery

Sat, 26 Jul 2025 19:19:53 +0000

Jan Jansen via perl5-porters <[email protected]> writes:

> Running the map construct to create a hash behaves oddly under specific
> circumstances.

> Typical expected behaviour to create a hash reference for a list
> $hashref = { map { <instructions>; $key => $value } @list };

> What work :
> $hashref = { map { 'key_'.$_ => $raw->{$_} } keys %$raw };
> $hashref = { map { ("key_.$_" => $raw->{$_}) } keys %$raw };

> What generates a syntax error :
> $hashref = { map { "key_$_" => $raw->{$_} } keys %$raw };

I believe this is the known edge case in the Perl parser that is
documented in perldoc -f map:

"{" starts both hash references and blocks, so "map { ..." could be
either the start of map BLOCK LIST or map EXPR, LIST. Because Perl
doesn't look ahead for the closing "}" it has to take a guess at which
it's dealing with based on what it finds just after the "{". Usually
it gets it right, but if it doesn't it won't realize something is
wrong until it gets to the "}" and encounters the missing (or
unexpected) comma. The syntax error will be reported close to the "}",
but you'll need to change something near the "{" such as using a unary
"+" or semicolon to give Perl some help:

my %hash = map { "\L$_" => 1 } @array # perl guesses EXPR. wrong
my %hash = map { +"\L$_" => 1 } @array # perl guesses BLOCK. right
my %hash = map {; "\L$_" => 1 } @array # this also works
my %hash = map { ("\L$_" => 1) } @array # as does this
my %hash = map { lc($_) => 1 } @array # and this.
my %hash = map +( lc($_) => 1 ), @array # this is EXPR and works!

my %hash = map ( lc($_), 1 ), @array # evaluates to (1, @array)

or to force an anon hash constructor use "+{":

my @hashes = map +{ lc($_) => 1 }, @array # EXPR, so needs
# comma at end

to get a list of anonymous hashes each with only one entry apiece.

If you instead write:

$hashref = { map { +"key_$_" => $raw->{$_} } keys %$raw };

that will unconfuse the parser.

--
#!/usr/bin/perl -- Russ Allbery, Just Another Perl Hacker
$^=q;@!>~|{>krw>yn{u<$$<[~||<Juukn{=,<S~|}<Jwx}qn{<Yn{u<Qjltn{ > 0gFzD gD,
00Fz, 0,,( 0hF 0g)F/=, 0> "L$/GEIFewe{,$/ 0C$~> "@=,m,|,(e 0.), 01,pnn,y{
rw} >;,$0=q,$,,($_=$^)=~y,$/ C-~><@=\n\r,-~$:-u/ #y,d,s,(\$.),$1,gee,print

Wierd behaviour of map when trying to fill a hash by Jan Jansen via perl5-porters

Sat, 26 Jul 2025 19:07:15 +0000

This is a bug report for perl from [email protected],
generated with the help of perlbug 1.43 running under perl 5.40.2.

-----------------------------------------------------------------


**Description**


Running the map construct to create a hash behaves oddly under specific circumstances.

Typical expected behaviour to create a hash reference for a list
$hashref = { map { <instructions>; $key => $value } @list };

What work :
$hashref = { map { 'key_'.$_ => $raw->{$_} } keys %$raw };
$hashref = { map { ("key_.$_" => $raw->{$_}) } keys %$raw };

What generates a syntax error :
$hashref = { map { "key_$_" => $raw->{$_} } keys %$raw };

**Steps to Reproduce**


$ perl -e '{ map { "key_$_" => $raw->{$_} } (keys %$raw) };'
syntax error at -e line 1, near "} ("
Execution of -e aborted due to compilation errors.

**Expected behavior**

Similar to
$ perl -e '{ map { 'key_'.$_ => $raw->{$_} } (keys %$raw) };'
No syntax error !




---
**Flags**
- category=core
- severity=low
---
**Perl configuration**
```
Site configuration information for perl 5.40.2:

Configured by Red Hat, Inc. at Mon Apr 14 00:00:00 UTC 2025.

Summary of my perl5 (revision 5 version 40 subversion 2) configuration:

Platform:
osname=linux
osvers=6.11.0
archname=x86_64-linux-thread-multi
uname='linux localhost 6.11.0 #1 smp preempt_dynamic 6.11.0 x86_64 gnulinux '
config_args='-des -Doptimize=none -Dccflags=-O2 -flto=auto -ffat-lto-objects -fexceptions -g -grecord-gcc-switches -pipe -Wall -Wno-complain-wrong-lang -Werror=format-security -Wp,-U_FORTIFY_SOURCE,-D_FORTIFY_SOURCE=3 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -march=x86-64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -mtls-dialect=gnu2 -fno-omit-frame-pointer -mno-omit-leaf-frame-pointer -Dldflags=-Wl,-z,relro -Wl,--as-needed -Wl,-z,pack-relative-relocs -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -Wl,--build-id=sha1 -Dccdlflags=-Wl,--enable-new-dtags -Wl,-z,relro -Wl,--as-needed -Wl,-z,pack-relative-relocs -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld
-specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -Wl,--build-id=sha1 -Dlddlflags=-shared -Wl,-z,relro -Wl,--as-needed -Wl,-z,pack-relative-relocs -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -Wl,--build-id=sha1 -Dshrpdir=/usr/lib64 -DDEBUGGING=-g -Dversion=5.40.2 -Dmyhostname=localhost -Dperladmin=root@localhost -Dcc=gcc -Dcf_by=Red Hat, Inc. -Dprefix=/usr -Dman1dir=/usr/share/man/man1 -Dman3dir=/usr/share/man/man3 -Dvendorprefix=/usr -Dsiteprefix=/usr/local -Dsitelib=/usr/local/share/perl5/5.40 -Dsitearch=/usr/local/lib64/perl5/5.40 -Dprivlib=/usr/share/perl5 -Dvendorlib=/usr/share/perl5/vendor_perl -Darchlib=/usr/lib64/perl5 -Dvendorarch=/usr/lib64/perl5/vendor_perl -Darchname=x86_64-linux-thread-multi -Dlibpth=/usr/local/lib64 /lib64 /usr/lib64 -Duseshrplib -Dusethreads -Duseithreads -Dusedtrace=/usr/bin/dtrace
-Duselargefiles -Dd_semctl_semun -Di_db -Ui_ndbm -Di_gdbm -Di_shadow -Di_syslog -Dman3ext=3pm -Duseperlio -Dinstallusrbinperl=n -Ubincompat5005 -Uversiononly -Dpager=/usr/bin/less -isr -Dd_gethostent_r_proto -Ud_endhostent_r_proto -Ud_sethostent_r_proto -Ud_endprotoent_r_proto -Ud_setprotoent_r_proto -Ud_endservent_r_proto -Ud_setservent_r_proto -Dscriptdir=/usr/bin -Dusesitecustomize -Duse64bitint'
hint=recommended
useposix=true
d_sigaction=define
useithreads=define
usemultiplicity=define
use64bitint=define
use64bitall=define
uselongdouble=undef
usemymalloc=n
default_inc_excludes_dot=define
Compiler:
cc='gcc'
ccflags ='-D_REENTRANT -D_GNU_SOURCE -O2 -flto=auto -ffat-lto-objects -fexceptions -g -grecord-gcc-switches -pipe -Wall -Wno-complain-wrong-lang -Werror=format-security -Wp,-U_FORTIFY_SOURCE,-D_FORTIFY_SOURCE=3 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -march=x86-64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -mtls-dialect=gnu2 -fno-omit-frame-pointer -mno-omit-leaf-frame-pointer -fwrapv -fno-strict-aliasing -I/usr/local/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64'
optimize=' -g'
cppflags='-D_REENTRANT -D_GNU_SOURCE -O2 -flto=auto -ffat-lto-objects -fexceptions -g -grecord-gcc-switches -pipe -Wall -Wno-complain-wrong-lang -Werror=format-security -Wp,-U_FORTIFY_SOURCE,-D_FORTIFY_SOURCE=3 -Wp,-D_GLIBCXX_ASSERTIONS -specs=/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -march=x86-64 -mtune=generic -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection -mtls-dialect=gnu2 -fno-omit-frame-pointer -mno-omit-leaf-frame-pointer -fwrapv -fno-strict-aliasing -I/usr/local/include'
ccversion=''
gccversion='14.2.1 20250110 (Red Hat 14.2.1-7)'
gccosandvers=''
intsize=4
longsize=8
ptrsize=8
doublesize=8
byteorder=12345678
doublekind=3
d_longlong=define
longlongsize=8
d_longdbl=define
longdblsize=16
longdblkind=3
ivtype='long'
ivsize=8
nvtype='double'
nvsize=8
Off_t='off_t'
lseeksize=8
alignbytes=8
prototype=define
Linker and Libraries:
ld='gcc'
ldflags ='-Wl,-z,relro -Wl,--as-needed -Wl,-z,pack-relative-relocs -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -Wl,--build-id=sha1 -fstack-protector-strong -L/usr/local/lib'
libpth=/usr/local/lib64 /lib64 /usr/lib64 /usr/local/lib /usr/lib
libs=-lpthread -lresolv -lgdbm -ldb -ldl -lm -lcrypt -lutil -lc -lgdbm_compat
perllibs=-lpthread -lresolv -ldl -lm -lcrypt -lutil -lc
libc=/lib/../lib64/libc.so.6
so=so
useshrplib=true
libperl=libperl.so
gnulibc_version='2.40'
Dynamic Linking:
dlsrc=dl_dlopen.xs
dlext=so
d_dlsymun=undef
ccdlflags='-Wl,--enable-new-dtags -Wl,-z,relro -Wl,--as-needed -Wl,-z,pack-relative-relocs -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -Wl,--build-id=sha1 '
cccdlflags='-fPIC'
lddlflags='-lpthread -shared -Wl,-z,relro -Wl,--as-needed -Wl,-z,pack-relative-relocs -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld -specs=/usr/lib/rpm/redhat/redhat-annobin-cc1 -Wl,--build-id=sha1 -L/usr/local/lib -fstack-protector-strong'

Locally applied patches:
Fedora Patch1: Removes date check, Fedora/RHEL specific
Fedora Patch2: support for libdir64
Fedora Patch3: use libresolv instead of libbind
Fedora Patch4: USE_MM_LD_RUN_PATH
Fedora Patch5: Provide MM::maybe_command independently (bug #1129443)
Fedora Patch6: Dont run one io test due to random builder failures
Fedora Patch8: Define SONAME for libperl.so
Fedora Patch9: Install libperl.so to -Dshrpdir value
Fedora Patch10: Make *DBM_File desctructors thread-safe (RT#61912)
Fedora Patch11: Replace EU::MakeMaker dependency with EU::MM::Utils in IPC::Cmd (bug #1129443)
Fedora Patch12: Link XS modules to pthread library to fix linking with -z defs
Fedora Patch13: Pass the correct CFLAGS to dtrace
Fedora Patch200: Link XS modules to libperl.so with EU::CBuilder on Linux
Fedora Patch201: Link XS modules to libperl.so with EU::MM on Linux
Fedora Patch202: Add definition of OPTIMIZE to .ph files

---
@INC for perl 5.40.2:
/usr/local/lib64/perl5/5.40
/usr/local/share/perl5/5.40
/usr/lib64/perl5/vendor_perl
/usr/share/perl5/vendor_perl
/usr/lib64/perl5
/usr/share/perl5

---
Environment for perl 5.40.2:
HOME=/home/jan
LANG=en_US.UTF-8
LANGUAGE (unset)
LD_LIBRARY_PATH (unset)
LOGDIR (unset)
PATH=/home/jan/.local/bin:/home/jan/bin:/usr/local/bin:/usr/local/sbin:/usr/bin:/usr/sbin:/var/lib/snapd/snap/bin
PERL_BADLANG (unset)
SHELL=/bin/bash
```

Perl 5 Commit Summary by Perl 5 commit summary

Sat, 26 Jul 2025 03:05:41 +0000

Perl 5 commit summary, activity since Wednesday

Current branch blead
15 commits. 6 unique authors. 5 unique committers.
19 files changed, 281 insertions(+), 89 deletions(-)
Thanks, applied: Karl Williamson (2)
Snapshot: http://github.com/Perl/perl5/tarball/f3a5b2b22f875a08

Revert "Attempt to remove deprecated sv_locking/unlocking"
Karl Williamson 6 files changed, 68 insertions(+), 3 deletions(-
https://github.com/Perl/perl5/commit/f3a5b2b22f875a08

Perl_more_bodies - figure out sizing from sv_type
Richard Leach 5 files changed, 30 insertions(+), 22 deletions(
https://github.com/Perl/perl5/commit/b40ae72cfcfd0eec

regmatch: Use new utf8_to_uv
Karl Williamson 1 file changed, 7 insertions(+), 6 deletions(-)
https://github.com/Perl/perl5/commit/9c39a33a0e817782

perlapi: Combine is_safe_syscall, IS_SAFE_SYSCALL
Karl Williamson 2 files changed, 6 insertions(+), 15 deletions(-
https://github.com/Perl/perl5/commit/194d4eac2f635f5c

Skip one test under miniperl; can't handle threads
James E Keenan 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/5d7576a6ed24ecd8

Add Ryan Carsten Schmidt to AUTHORS file
Ryan Carsten Schmidt 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/f5b140f3b87fe090

Resolve any symlinks to cwd in dylib install name
Ryan Carsten Schmidt 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/1f30e27f2871696d

Note when functions were first deprecated
Karl Williamson 4 files changed, 20 insertions(+), 6 deletions(-
https://github.com/Perl/perl5/commit/604d1d0444885ed2

Attempt to remove deprecated sv_locking/unlocking
Karl Williamson 6 files changed, 3 insertions(+), 68 deletions(-
https://github.com/Perl/perl5/commit/3afafdb6cda45b88

perl.h: include locale related headers after we decide to use locales
Tony Cook 1 file changed, 9 insertions(+), 9 deletions(-)
https://github.com/Perl/perl5/commit/bc5a3e960d095c85

op.c: factor out repeated entersub code
Lukas Mai 1 file changed, 39 insertions(+), 27 deletions(-
https://github.com/Perl/perl5/commit/404e37b805ecb843

perldelta for the 2-var 'for' crash (#23405)
Lukas Mai 1 file changed, 23 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/7aa27d7640d9bdff

perf/opcount.t: ensure imported builtins are optimized
Lukas Mai 1 file changed, 124 insertions(+), 5 deletions(-
https://github.com/Perl/perl5/commit/55ce815fbe0b4bd9

newFOROP: fix crash when optimizing 2-var for over builtin::indexed
Lukas Mai 2 files changed, 24 insertions(+), 3 deletions(-
https://github.com/Perl/perl5/commit/96673a4bb36a973a

regcomp_trie: stringify *tmp before accessing string fields
Lukas Mai 1 file changed, 3 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/d83bd2549fce92d1

Current branch smoke-me/tonyc/22125-openbsd-embed
3 commits. 1 unique author. 1 unique committer.
561 files changed, 14646 insertions(+), 10580 deletions(-)

Snapshot: http://github.com/Perl/perl5/tarball/3e2c4b36ca1032ea

perldelta for fix embedding test on openbsd
Tony Cook 1 file changed, 10 insertions(+)
https://github.com/Perl/perl5/commit/3e2c4b36ca1032ea

openbsd: ensure we link to the built libperl.a, not the system libperl.a
Tony Cook 1 file changed, 17 insertions(+), 4 deletions(-)
https://github.com/Perl/perl5/commit/1f7d43dcf66c46e3

add API to test that libperl and the current executable are compatible
Tony Cook 7 files changed, 72 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/f4362aa19b05fb4d

Martian commit 6f324ad9fffabd8f516a475f168e954d9be9c9df
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/6f324ad9fffabd8f

Revert "Attempt to remove deprecated sv_locking/unlocking"
Karl Williamson 6 files changed, 68 insertions(+), 3 deletions(-
https://github.com/Perl/perl5/commit/6f324ad9fffabd8f

Martian commit f84aec142de0fb5bcc37fee323755b9e5b1559ee
2 commits. 1 unique author. 1 unique committer.
Net code removal: HungryStackOf Pancakes (118599 lines)

Snapshot: http://github.com/Perl/perl5/tarball/f84aec142de0fb5b

Prune to all non-win32 README files and platform code for Windows ARM only
HungryStackOf Pancakes 205 files changed, 118615 deletions(-)
https://github.com/Perl/perl5/commit/f84aec142de0fb5b

Add Windows ARM64 support
HungryStackOf Pancakes 30 files changed, 16 insertions(+), 10038 deleti
https://github.com/Perl/perl5/commit/8fa8910e17bf6fe9

Martian commit a2d30e0c57349460e1189042377b406c4c5cbc3e
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/a2d30e0c57349460

Skip one test under miniperl; can't handle threads
James E Keenan 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/a2d30e0c57349460

Martian commit 40ce3d5f6ecefb541ac468ab5da37798907335ef
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/40ce3d5f6ecefb54

Note when functions were first deprecated
Karl Williamson 4 files changed, 20 insertions(+), 6 deletions(-
https://github.com/Perl/perl5/commit/40ce3d5f6ecefb54

Martian commit cc1e873b9f0cd05da90a4c30a9aefb723a576ee8
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/cc1e873b9f0cd05d

Attempt to remove deprecated sv_locking/unlocking
Karl Williamson 6 files changed, 3 insertions(+), 68 deletions(-
https://github.com/Perl/perl5/commit/cc1e873b9f0cd05d

Martian commit 63844f08efcede56076ea6512de0eba7ccd53476
2 commits. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/63844f08efcede56

Add Ryan Carsten Schmidt to AUTHORS file
Ryan Carsten Schmidt 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/63844f08efcede56

Resolve any symlinks to cwd in dylib install name
Ryan Carsten Schmidt 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/f825d24a4c585a2f

Ignored 94 GitHub auto-generated merge commits

Re: State of XSpp / C++ integration by Steffen Mueller

Fri, 25 Jul 2025 12:14:38 +0000

Hi Johan,

to be honest I just came across this email randomly because I peek at my
p5p subscription every six months or so. Maybe that's already the answer to
your question about support. I've been largely away from Perl and OSS due
to work & family demands in the past decade. I took a quick look at a few
of the outstanding issues and they seem to be largely either trivially
fixable (eg. classic ". not in @INC") or self-notes/feature requests.

I'd love to promise to help, but I've become quite rusty and my time comes
in random small increments. So while I'd be happy to take a look at
whatever you're running into, I'm also leery of asking you to do a lot of
work to explain the issues because I have low confidence in my usefulness.

Mattia is the original author of ExtUtils::XSpp and wxPerl, so I'm CCing
him. Hi Mattia! Hope you're well!

--Steffen

On Tue, Jul 22, 2025 at 12:27â€¯PM Johan Vromans <[email protected]> wrote:

> While trying to port wxPerl to a more recent version of wxWidgets I
> stumbled upon a number of XSpp issues that I find hard to solve.
>
> Before spending a lot of energy I would first like to ask if XSpp is still
> a viable toolchain for C++ embedding. It hasn't been updated for many years
> and has a load of old issues that have never looked at.
>
> If someone with practical XSpp experience would be able to provide some
> help I would welcome it.
>
> Thanks.
>

Perl 5 Commit Summary by Perl 5 commit summary

Wed, 23 Jul 2025 03:11:56 +0000

Perl 5 commit summary, activity since Saturday

Current branch blead
37 commits. 8 unique authors. 8 unique committers.
61 files changed, 1393 insertions(+), 1043 deletions(-)
Thanks, applied: James E Keenan (1) Tony Cook (1)
Snapshot: http://github.com/Perl/perl5/tarball/96784b083ca7e05f

win32_checkTLS(): nothing to do without at least multiplicity
Tony Cook 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/96784b083ca7e05f

win32_checkTLS: add a panic message and break into the debugger
Tony Cook 1 file changed, 11 insertions(+)
https://github.com/Perl/perl5/commit/0144414218630c60

win32_checkTLS: eliminate no longer used variable
Tony Cook 1 file changed, 2 deletions(-)
https://github.com/Perl/perl5/commit/43dfd9c1c1f6d33b

perldelta for Win32 -Dusethreads builds fix
Tony Cook 1 file changed, 3 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/4a9179a6e6e60831

win32: build without threads but still with implicit sys
Tony Cook 1 file changed, 8 insertions(+)
https://github.com/Perl/perl5/commit/7b87a86b61a2855b

GH #23390 re/Makefile.PL silence uninit $args{"LINKTYPE"} warning
Daniel Dragan 1 file changed, 3 insertions(+), 3 deletions(-)
https://github.com/Perl/perl5/commit/872e7cbc2803403b

UCD.pm: Add check for parameter type
Karl Williamson 7 files changed, 22 insertions(+), 7 deletions(-
https://github.com/Perl/perl5/commit/ea25adf22cbf3d9a

perlre: Clarify wording
Karl Williamson 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/434c9cd5f17004da

perlop: Clarify empty pattern section not valid for qr//
Karl Williamson 1 file changed, 3 insertions(+)
https://github.com/Perl/perl5/commit/652110bce8cbe422

perl.h: Make sure USE_LOCALE is off if no locales available
Karl Williamson 2 files changed, 5 insertions(+)
https://github.com/Perl/perl5/commit/9054d08e4819d1f2

regen/locale.pl: Use 'USE_LOCALE" not 'NO_LOCALE'
Karl Williamson 2 files changed, 15 insertions(+), 15 deletions(
https://github.com/Perl/perl5/commit/03e269a3672fae76

Fix typo in epigraph for 5.43.1
Steve Hay 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/bbfbc490422f71c8

Add epigraphs for 5.38.5-RC1/5.40.3-RC1
Steve Hay 1 file changed, 25 insertions(+)
https://github.com/Perl/perl5/commit/ab34963a6d7d97b9

regen/embed.pl Use /xx to clarify regex
Karl Williamson 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/4d32d2c598ed4085

5.38.5-RC1/5.40.3-RC1 today
Steve Hay 1 file changed, 2 insertions(+)
https://github.com/Perl/perl5/commit/93253b7b9371211f

Add 5.40.3 to release schedule
Steve Hay 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/77413f095185658e

croak_xs_usage pod: remove unnecessary diag_listed_as
Tony Cook 1 file changed, 1 insertion(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/63b30a217f0590c2

Prepare Module::Corelist for 5.43.2
Richard Leach 3 files changed, 31 insertions(+), 2 deletions(-
https://github.com/Perl/perl5/commit/2904b5b0342335f0

Bump the perl version in various places for 5.43.2
Richard Leach 20 files changed, 124 insertions(+), 124 deletio
https://github.com/Perl/perl5/commit/4a431cdfdca44f0d

Archive perldelta 5.43.1
Richard Leach 10 files changed, 536 insertions(+), 84 deletion
https://github.com/Perl/perl5/commit/a1ca12c1f6991b57

Update release_schedule.pod following 5.43.1 release
Richard Leach 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/cad4d402f55a5a60

Epigraph for 5.43.1
Richard Leach 1 file changed, 11 insertions(+)
https://github.com/Perl/perl5/commit/df1bffbbf477c356

Add new release to perlhist
Richard Leach 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/89012989f3104d6b

Finalize perldelta for 5.43.1
Richard Leach 1 file changed, 46 insertions(+), 326 deletions(
https://github.com/Perl/perl5/commit/6cf1e366599aaf3f

Update Module::CoreList for 5.43.1
Richard Leach 2 files changed, 6 insertions(+), 3 deletions(-)
https://github.com/Perl/perl5/commit/25bed50932dd736e

Improve documentation of sv_dup and sv_dup_inc
Leon Timmermans 1 file changed, 10 insertions(+), 10 deletions(-
https://github.com/Perl/perl5/commit/83b32ab641017de8

Revert marking sv_dup and sv_dup_inc as core-only
Leon Timmermans 1 file changed, 2 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/166fcd923517d162

cpan/experimental - Update to version 0.036
Leon Timmermans 3 files changed, 21 insertions(+), 6 deletions(-
https://github.com/Perl/perl5/commit/cf7cc0a627da2ab1

Convert uv_to_utf8_msgs() to a macro
Karl Williamson 5 files changed, 10 insertions(+), 13 deletions(
https://github.com/Perl/perl5/commit/9f64cab5b47dac7c

Allow automatic long name macro generation
Karl Williamson 2 files changed, 34 insertions(+), 2 deletions(-
https://github.com/Perl/perl5/commit/93f23f07eb96405f

Change #define Perl_foo Perl_bar" to plain "foo bar"
Karl Williamson 5 files changed, 45 insertions(+), 45 deletions(
https://github.com/Perl/perl5/commit/6277a14c3a20e1c9

regen/embed.pl: Separate conditional into its own clause
Karl Williamson 1 file changed, 7 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/763cee0e4c3df09d

regen/embed.pl: Expand tabs to blanks
Karl Williamson 1 file changed, 15 insertions(+), 15 deletions(-
https://github.com/Perl/perl5/commit/808719b4a26e583c

add missing releaser names
Philippe Bruhat (BooK) 1 file changed, 2 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/48b684594b711eb2

remove duplicate consecutive releaser names
Philippe Bruhat (BooK) 1 file changed, 119 insertions(+), 119 deletions
https://github.com/Perl/perl5/commit/6d2e3b2b1a2fe4b2

reindent the whole perl history table
Philippe Bruhat (BooK) 1 file changed, 754 insertions(+), 740 deletions
https://github.com/Perl/perl5/commit/9fea74fc5b04f962

perldelta entry for ExtUtils::ParseXS work
David Mitchell 1 file changed, 3 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/dae10d9a000cb3eb

Current branch maint-5.38
81 commits. 6 unique authors. 1 unique committer.
28 files changed, 1916 insertions(+), 1244 deletions(-)
Thanks, applied: Steve Hay (68)
Snapshot: http://github.com/Perl/perl5/tarball/b2a5a9dfa0810ab6

Fix typo in epigraph for 5.43.1
Steve Hay 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/b2a5a9dfa0810ab6

Add epigraphs for 5.38.5-RC1/5.40.3-RC1
Steve Hay 1 file changed, 25 insertions(+)
https://github.com/Perl/perl5/commit/c153240585b970da

Disarm patchlevel.h and regen META.json
Steve Hay 2 files changed, 1 insertion(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/db49170007ffc1a6

Bump to RC1 and update META.json
Steve Hay 2 files changed, 2 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/195905f9cb5b7b52

5.38.5-RC1/5.40.3-RC1 today
Steve Hay 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/a9e82d4af2faf156

Finalize perldelta
Steve Hay 1 file changed, 24 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/38a99c32d3b2bd92

Add 5.40.3 to release schedule
Steve Hay 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/6a778c2355d85fba

Update release_schedule.pod following 5.43.1 release
Richard Leach 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/b274abe21b704fb4

Epigraph for 5.43.1
Richard Leach 1 file changed, 11 insertions(+)
https://github.com/Perl/perl5/commit/3f04fa72877c1eb7

Add new release to perlhist
Richard Leach 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/a6f3674851036cf3

Update Module::CoreList for 5.43.1
Richard Leach 2 files changed, 6 insertions(+), 3 deletions(-)
https://github.com/Perl/perl5/commit/bea7c06263655bfb

Add delta related to CVE-2025-40909
Steve Hay 1 file changed, 12 insertions(+), 4 deletions(-)
https://github.com/Perl/perl5/commit/84894d57c2006dad

perldelta: Update modules section, remove boilerplate, copy-editing
Steve Hay 1 file changed, 16 insertions(+), 367 deletions(
https://github.com/Perl/perl5/commit/5209daa1696fbe63

Update Module::CoreList with data so far for the imminent 5.38.5
Steve Hay 3 files changed, 31 insertions(+), 2 deletions(-
https://github.com/Perl/perl5/commit/3eaaaad74aba6ee3

Bump version to 5.38.5, ahead of its impending release
Steve Hay 21 files changed, 112 insertions(+), 112 deletio
https://github.com/Perl/perl5/commit/e227c9edc05a9100

add missing releaser names
Philippe Bruhat (BooK) 1 file changed, 2 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/bafbfea9cf536ac1

remove duplicate consecutive releaser names
Philippe Bruhat (BooK) 1 file changed, 119 insertions(+), 119 deletions
https://github.com/Perl/perl5/commit/39c00cb0bc84e767

reindent the whole perl history table
Philippe Bruhat (BooK) 1 file changed, 754 insertions(+), 740 deletions
https://github.com/Perl/perl5/commit/4e092ac7fac2335a

perlhist: fix some little historical inaccuracies and ordering
Thibault Duponchelle 1 file changed, 6 insertions(+), 6 deletions(-)
https://github.com/Perl/perl5/commit/a8f0122d77d26bb4

update perlhist for 5.43.0
Philippe Bruhat (BooK) 1 file changed, 2 insertions(+)
https://github.com/Perl/perl5/commit/ea4cdaf4f3eb6e62

perlhist: Small update to selected release size for 5.42.0
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/1a9733441e828f6d

perldelta/perlhist: small update of metrics
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/8df120ec5f003af7

Update perlhist
Thibault Duponchelle 1 file changed, 2 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/c5026eca99b4e464

Add 5.42.0 to perlhist
Thibault Duponchelle 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/1add257b75da0708

Update perlhist with release size
Thibault Duponchelle 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/8eae3094bd6b08f6

Add new release to perlhist
Thibault Duponchelle 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/71a1b02889cd47ce

Update perlhist.pod
Thibault Duponchelle 1 file changed, 2 insertions(+)
https://github.com/Perl/perl5/commit/33aa757401e4c61c

Add new release to perlhist
Philippe Bruhat (BooK) 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/22aca67be3b3f5b3

Add new release to perlhist
Karen Etheridge 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/f30817c439ceaac7

Add new release to perlhist
Karen Etheridge 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/0e6a6a07c9d342b0

epigraph for v5.42.0
Philippe Bruhat (BooK) 1 file changed, 6 insertions(+)
https://github.com/Perl/perl5/commit/9ec7e3915ce53884

Add epigraph for 5.42.0-RC3
Thibault Duponchelle 1 file changed, 6 insertions(+)
https://github.com/Perl/perl5/commit/a5229da5dc97e003

Add epigraph for 5.42.0-RC2
Thibault Duponchelle 1 file changed, 6 insertions(+)
https://github.com/Perl/perl5/commit/9ea70cd6ccf5a069

Add epigraph for 5.42.0-RC1
Thibault Duponchelle 1 file changed, 7 insertions(+)
https://github.com/Perl/perl5/commit/2658f0c0bb6a7f13

add epigraph for 5.41.13
Philippe Bruhat (BooK) 1 file changed, 6 insertions(+)
https://github.com/Perl/perl5/commit/71b72ea2df75d7a8

add 5.41.12 epigraph
Karen Etheridge 1 file changed, 7 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/0a219c63570ba60d

add epigraph for 5.41.11
Karen Etheridge 1 file changed, 8 insertions(+)
https://github.com/Perl/perl5/commit/e0a7a1e1ecd6125e

release_schedule: Add a new victim
Thibault Duponchelle 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/bb2beca25b04d408

volunteering for 5.43.3
Philippe Bruhat (BooK) 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/5c62bbf426cda1d6

release_schedule: Add release manager to 5.43.6
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/9744341037ce58de

release_schedule: normalize release manager to the common practice
Thibault Duponchelle 1 file changed, 5 insertions(+), 5 deletions(-)
https://github.com/Perl/perl5/commit/aeac8405006c3eb9

release_schedule: Add release manager to 5.43.4
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/c82d0a23fa2995b2

I volunteer as tribute
Karen Etheridge 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/78f41d2803fd7151

release_schedule: Tick 5.43.0 - Add release manager to 5.43.1, 5.43.5, 5.43.
Thibault Duponchelle 1 file changed, 4 insertions(+), 4 deletions(-)
https://github.com/Perl/perl5/commit/3ded3aa9597bcc91

release_schedule: 5.44 in 2026
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/7392d54f57fc3936

Add future 5.44.0 to schedule
Thibault Duponchelle 1 file changed, 4 insertions(+)
https://github.com/Perl/perl5/commit/3ee04d2dda4b0e6c

new release schedule for the v5.43 series
Philippe Bruhat (BooK) 1 file changed, 15 insertions(+), 24 deletions(-
https://github.com/Perl/perl5/commit/2edcfdf4d5471705

add 5.41.13 to release history
Karen Etheridge 1 file changed, 2 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/47bff22bd3cc0da9

tick off the release
Karen Etheridge 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/2d41c44a50732124

tick off 5.41.11 release
Karen Etheridge 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/234508ecd6edb53e

improve the GitHub links regexp in the RMG
Philippe Bruhat (BooK) 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/213b86fdce56c46c

RMG: be more explicit about Perl Releasers team requirement
Lukas Mai 1 file changed, 11 insertions(+), 8 deletions(-)
https://github.com/Perl/perl5/commit/5bcd768c493925f1

RMG - For commit message related to preparation of Module::Corelist for next
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/43bb2cf50121b6cf

RMG - Run perlivp in install path
Thibault Duponchelle 1 file changed, 3 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/d97210670f7e3be3

RMG - Edit commit message for step of archiving perldelta
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/ea8212101f03b269

Porting/release_managers_guide.pod: document version placeholders
Lukas Mai 1 file changed, 29 insertions(+)
https://github.com/Perl/perl5/commit/c47ea663fb7b5a28

Fix merge error
Steve Hay 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/014c1fb047d91df4

Replace placeholder of Module::CoreList release date for 5.43.1
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/14c73a3eec36cdc6

Module::CoreList version will be 5.20250720
Thibault Duponchelle 3 files changed, 8 insertions(+), 3 deletions(-)
https://github.com/Perl/perl5/commit/558b232b15f75f99

Prepare Module::CoreList for 5.43.1
Thibault Duponchelle 2 files changed, 309 insertions(+)
https://github.com/Perl/perl5/commit/47c7c279da070c6a

Update Module::CoreList for 5.43.0
Thibault Duponchelle 1 file changed, 15 insertions(+), 3 deletions(-)
https://github.com/Perl/perl5/commit/5aacab020a0eab5b

prepare Module::CoreList for perl 5.43.0
Philippe Bruhat (BooK) 3 files changed, 8 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/b2ff80e6ac171952

prepare Module::CoreList for 5.43.0
Philippe Bruhat (BooK) 2 files changed, 20 insertions(+), 2 deletions(-
https://github.com/Perl/perl5/commit/6cc753a7d46f8273

Update Module::CoreList for 5.42.0-RC3
Thibault Duponchelle 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/193cf18bd42e3c3d

Update Module::CoreList for 5.42.0
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/3e3fde573f5f4cda

Update Module::CoreList for 5.42.0
Thibault Duponchelle 3 files changed, 5 insertions(+), 5 deletions(-)
https://github.com/Perl/perl5/commit/b1b8549688632fd2

Update Module::CoreList for 5.42.0 after fix in ExtUtils::CBuilder and doc u
Thibault Duponchelle 1 file changed, 15 insertions(+)
https://github.com/Perl/perl5/commit/247f99982a7932d8

Update Module::CoreList version
Thibault Duponchelle 3 files changed, 6 insertions(+), 6 deletions(-)
https://github.com/Perl/perl5/commit/91287c3282b70e5d

Remove the removal of unicore::Name from Module::CoreList 5.41.13
Thibault Duponchelle 1 file changed, 1 deletion(-)
https://github.com/Perl/perl5/commit/1f381dcd6ca6a5a2

Update Module::CoreList version to actual RC1 date
Thibault Duponchelle 3 files changed, 6 insertions(+), 6 deletions(-)
https://github.com/Perl/perl5/commit/037bbfb91eb18af2

Update Module::CoreList Changes
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/6f7baffcc44a1bc8

Update Module::CoreList for 5.42.0-RC1. Regen Op_private. Update META.json
Thibault Duponchelle 2 files changed, 7 insertions(+), 6 deletions(-)
https://github.com/Perl/perl5/commit/e8164e571f78efac

Update Module::CoreList for 5.41.14
Philippe Bruhat (BooK) 3 files changed, 31 insertions(+), 2 deletions(-
https://github.com/Perl/perl5/commit/f94b668600256627

Update Module::CoreList for 5.41.13
Philippe Bruhat (BooK) 3 files changed, 65 insertions(+), 8 deletions(-
https://github.com/Perl/perl5/commit/eed56b2a642b80e5

Update Module::CoreList for 5.41.13
Karen Etheridge 3 files changed, 32 insertions(+), 3 deletions(-
https://github.com/Perl/perl5/commit/c910f1e314e4fdaf

Update Module::CoreList for 5.41.12
Karen Etheridge 3 files changed, 3 insertions(+), 3 deletions(-)
https://github.com/Perl/perl5/commit/972e7b1698d696f8

Unicode 16.0.0 was actually added in 5.41.11
Karen Etheridge 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/37d262ab63320793

Update Module::CoreList for 5.41.12
Karen Etheridge 3 files changed, 30 insertions(+), 2 deletions(-
https://github.com/Perl/perl5/commit/318c37bc22f2f29d

Update Module::CoreList for 5.41.11
Karen Etheridge 3 files changed, 263 insertions(+), 4 deletions(
https://github.com/Perl/perl5/commit/6ea33a50e0dd1bef

Update Module-CoreList with data for 5.38.4/5.40.2
Steve Hay 3 files changed, 29 insertions(+)
https://github.com/Perl/perl5/commit/6e5ed5f5d15a7e94

Prepare Module::Corelist for 5.41.11
Lukas Mai 2 files changed, 24 insertions(+)
https://github.com/Perl/perl5/commit/8d2a763e7ed49476

Current branch maint-5.40
81 commits. 6 unique authors. 1 unique committer.
31 files changed, 2034 insertions(+), 1243 deletions(-)
Thanks, applied: Steve Hay (68)
Snapshot: http://github.com/Perl/perl5/tarball/ca8b3320a147ff9e

Fix typo in epigraph for 5.43.1
Steve Hay 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/ca8b3320a147ff9e

Add epigraphs for 5.38.5-RC1/5.40.3-RC1
Steve Hay 1 file changed, 25 insertions(+)
https://github.com/Perl/perl5/commit/29d24b1fc61b9d9d

Disarm patchlevel.h and regen META.json
Steve Hay 2 files changed, 1 insertion(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/60d4710b863b9022

Bump to RC1 and update META.json
Steve Hay 2 files changed, 2 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/57d9c1bc123e937a

5.38.5-RC1/5.40.3-RC1 today
Steve Hay 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/16115e3dd33a1748

Finalize perldelta
Steve Hay 1 file changed, 24 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/768fb243dd16fbe3

Add 5.40.3 to release schedule
Steve Hay 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/a575be635479a854

Update release_schedule.pod following 5.43.1 release
Richard Leach 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/f0661ba866efb9de

Epigraph for 5.43.1
Richard Leach 1 file changed, 11 insertions(+)
https://github.com/Perl/perl5/commit/025bc7016bc7cc43

Add new release to perlhist
Richard Leach 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/a8c8f3e5b853faa2

Update Module::CoreList for 5.43.1
Richard Leach 2 files changed, 6 insertions(+), 3 deletions(-)
https://github.com/Perl/perl5/commit/cc70981d343f93e3

Add delta related to CVE-2025-40909
Steve Hay 1 file changed, 12 insertions(+), 4 deletions(-)
https://github.com/Perl/perl5/commit/79bcc385c433af3c

Import perl5384delta.pod
Steve Hay 4 files changed, 119 insertions(+)
https://github.com/Perl/perl5/commit/6a90bc1f78b64d05

perldelta: Update modules section, remove boilerplate, copy-editing
Steve Hay 1 file changed, 16 insertions(+), 367 deletions(
https://github.com/Perl/perl5/commit/d57f2a6b562d4370

Update Module::CoreList with data so far for the imminent 5.40.3
Steve Hay 3 files changed, 31 insertions(+), 2 deletions(-
https://github.com/Perl/perl5/commit/6786f7b2129445db

Bump version to 5.40.3, ahead of its impending release
Steve Hay 20 files changed, 111 insertions(+), 111 deletio
https://github.com/Perl/perl5/commit/9bd367d4eb6f24b5

add missing releaser names
Philippe Bruhat (BooK) 1 file changed, 2 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/167e8f3024ae7425

remove duplicate consecutive releaser names
Philippe Bruhat (BooK) 1 file changed, 119 insertions(+), 119 deletions
https://github.com/Perl/perl5/commit/89837aa96ce277f3

reindent the whole perl history table
Philippe Bruhat (BooK) 1 file changed, 754 insertions(+), 740 deletions
https://github.com/Perl/perl5/commit/b90936c81c75a946

perlhist: fix some little historical inaccuracies and ordering
Thibault Duponchelle 1 file changed, 6 insertions(+), 6 deletions(-)
https://github.com/Perl/perl5/commit/65a8edfd789a46a2

update perlhist for 5.43.0
Philippe Bruhat (BooK) 1 file changed, 2 insertions(+)
https://github.com/Perl/perl5/commit/76a163725b569dde

perlhist: Small update to selected release size for 5.42.0
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/0b2e5acd63d4e376

perldelta/perlhist: small update of metrics
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/2a29f78a2cb65570

Update perlhist
Thibault Duponchelle 1 file changed, 2 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/7bfe9bd2ab996f7f

Add 5.42.0 to perlhist
Thibault Duponchelle 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/73706e61aec22f71

Update perlhist with release size
Thibault Duponchelle 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/7f4802ab7568270e

Add new release to perlhist
Thibault Duponchelle 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/a887b19f6694430d

Update perlhist.pod
Thibault Duponchelle 1 file changed, 2 insertions(+)
https://github.com/Perl/perl5/commit/14c1fbf95dee17ca

Add new release to perlhist
Philippe Bruhat (BooK) 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/b669c637ea61a8b3

Add new release to perlhist
Karen Etheridge 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/f977f54c9e11283e

Add new release to perlhist
Karen Etheridge 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/6198ae6bfa5aa635

epigraph for v5.42.0
Philippe Bruhat (BooK) 1 file changed, 6 insertions(+)
https://github.com/Perl/perl5/commit/e3b2c96d89f3c5eb

Add epigraph for 5.42.0-RC3
Thibault Duponchelle 1 file changed, 6 insertions(+)
https://github.com/Perl/perl5/commit/3c6308ad1c8d828b

Add epigraph for 5.42.0-RC2
Thibault Duponchelle 1 file changed, 6 insertions(+)
https://github.com/Perl/perl5/commit/9bb7787554d85b43

Add epigraph for 5.42.0-RC1
Thibault Duponchelle 1 file changed, 7 insertions(+)
https://github.com/Perl/perl5/commit/656981c4df0881a4

add epigraph for 5.41.13
Philippe Bruhat (BooK) 1 file changed, 6 insertions(+)
https://github.com/Perl/perl5/commit/9f1340e21a00b52b

add 5.41.12 epigraph
Karen Etheridge 1 file changed, 7 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/01d00b8f596e0ee4

add epigraph for 5.41.11
Karen Etheridge 1 file changed, 8 insertions(+)
https://github.com/Perl/perl5/commit/9aac9b08140bdffb

release_schedule: Add a new victim
Thibault Duponchelle 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/6845ff6c6b34f164

volunteering for 5.43.3
Philippe Bruhat (BooK) 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/367669c356abd88b

release_schedule: Add release manager to 5.43.6
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/7b8702e84790bb9e

release_schedule: normalize release manager to the common practice
Thibault Duponchelle 1 file changed, 5 insertions(+), 5 deletions(-)
https://github.com/Perl/perl5/commit/e933573795b6e0e6

release_schedule: Add release manager to 5.43.4
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/552da45f87324d3c

I volunteer as tribute
Karen Etheridge 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/9fdf6c22303e6db4

release_schedule: Tick 5.43.0 - Add release manager to 5.43.1, 5.43.5, 5.43.
Thibault Duponchelle 1 file changed, 4 insertions(+), 4 deletions(-)
https://github.com/Perl/perl5/commit/0df157460539752c

release_schedule: 5.44 in 2026
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/04dbda0b550953d6

Add future 5.44.0 to schedule
Thibault Duponchelle 1 file changed, 4 insertions(+)
https://github.com/Perl/perl5/commit/2a7adb131e82f0e4

new release schedule for the v5.43 series
Philippe Bruhat (BooK) 1 file changed, 15 insertions(+), 24 deletions(-
https://github.com/Perl/perl5/commit/a2d088fb0453c043

add 5.41.13 to release history
Karen Etheridge 1 file changed, 2 insertions(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/1e5f4dd5ec0fa61b

tick off the release
Karen Etheridge 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/9499e2073dd049ab

tick off 5.41.11 release
Karen Etheridge 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/975f2309b49d9036

improve the GitHub links regexp in the RMG
Philippe Bruhat (BooK) 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/07120f5d407cec0d

RMG: be more explicit about Perl Releasers team requirement
Lukas Mai 1 file changed, 11 insertions(+), 8 deletions(-)
https://github.com/Perl/perl5/commit/767937c8806998d1

RMG - For commit message related to preparation of Module::Corelist for next
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/419b06fe3b8ad549

RMG - Run perlivp in install path
Thibault Duponchelle 1 file changed, 3 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/2ce6be322e403e50

RMG - Edit commit message for step of archiving perldelta
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/51fa8c41eb91ad58

Porting/release_managers_guide.pod: document version placeholders
Lukas Mai 1 file changed, 29 insertions(+)
https://github.com/Perl/perl5/commit/02165875d93b678f

Replace placeholder of Module::CoreList release date for 5.43.1
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/a5194daa715891b1

Module::CoreList version will be 5.20250720
Thibault Duponchelle 3 files changed, 8 insertions(+), 3 deletions(-)
https://github.com/Perl/perl5/commit/ed4d83668e975666

Prepare Module::CoreList for 5.43.1
Thibault Duponchelle 2 files changed, 309 insertions(+)
https://github.com/Perl/perl5/commit/a8b84cbf3e0d9c7f

Update Module::CoreList for 5.43.0
Thibault Duponchelle 1 file changed, 15 insertions(+), 3 deletions(-)
https://github.com/Perl/perl5/commit/5010f514fd4b6885

prepare Module::CoreList for perl 5.43.0
Philippe Bruhat (BooK) 3 files changed, 8 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/b4c92023fb466700

prepare Module::CoreList for 5.43.0
Philippe Bruhat (BooK) 2 files changed, 20 insertions(+), 2 deletions(-
https://github.com/Perl/perl5/commit/1bd39f7ca6b0bc8b

Update Module::CoreList for 5.42.0-RC3
Thibault Duponchelle 1 file changed, 1 insertion(+)
https://github.com/Perl/perl5/commit/886a8a842455194f

Update Module::CoreList for 5.42.0
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/e86142504f962452

Update Module::CoreList for 5.42.0
Thibault Duponchelle 3 files changed, 5 insertions(+), 5 deletions(-)
https://github.com/Perl/perl5/commit/535cb3c9d8d5701a

Update Module::CoreList for 5.42.0 after fix in ExtUtils::CBuilder and doc u
Thibault Duponchelle 1 file changed, 15 insertions(+)
https://github.com/Perl/perl5/commit/4eb6e6ccbb7c01cb

Update Module::CoreList version
Thibault Duponchelle 3 files changed, 6 insertions(+), 6 deletions(-)
https://github.com/Perl/perl5/commit/b5621d6255daecca

Remove the removal of unicore::Name from Module::CoreList 5.41.13
Thibault Duponchelle 1 file changed, 1 deletion(-)
https://github.com/Perl/perl5/commit/82747f198a96f8a2

Update Module::CoreList version to actual RC1 date
Thibault Duponchelle 3 files changed, 6 insertions(+), 6 deletions(-)
https://github.com/Perl/perl5/commit/e59833de4e03359f

Update Module::CoreList Changes
Thibault Duponchelle 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/d2eb7c84e9417d67

Update Module::CoreList for 5.42.0-RC1. Regen Op_private. Update META.json
Thibault Duponchelle 2 files changed, 7 insertions(+), 6 deletions(-)
https://github.com/Perl/perl5/commit/042bec5f3eb12d56

Update Module::CoreList for 5.41.14
Philippe Bruhat (BooK) 3 files changed, 31 insertions(+), 2 deletions(-
https://github.com/Perl/perl5/commit/616bae77834f23c0

Update Module::CoreList for 5.41.13
Philippe Bruhat (BooK) 3 files changed, 65 insertions(+), 8 deletions(-
https://github.com/Perl/perl5/commit/4174744e8dfe484b

Update Module::CoreList for 5.41.13
Karen Etheridge 3 files changed, 32 insertions(+), 3 deletions(-
https://github.com/Perl/perl5/commit/00e92f261842b076

Update Module::CoreList for 5.41.12
Karen Etheridge 3 files changed, 3 insertions(+), 3 deletions(-)
https://github.com/Perl/perl5/commit/a4533fb56ce9df0a

Unicode 16.0.0 was actually added in 5.41.11
Karen Etheridge 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/d11a98c5859da817

Update Module::CoreList for 5.41.12
Karen Etheridge 3 files changed, 30 insertions(+), 2 deletions(-
https://github.com/Perl/perl5/commit/ac2bb7f9e9769781

Update Module::CoreList for 5.41.11
Karen Etheridge 3 files changed, 263 insertions(+), 4 deletions(
https://github.com/Perl/perl5/commit/d6c83a7497aace53

Update Module-CoreList with data for 5.38.4/5.40.2
Steve Hay 3 files changed, 29 insertions(+)
https://github.com/Perl/perl5/commit/9a58914b744b7874

Prepare Module::Corelist for 5.41.11
Lukas Mai 2 files changed, 24 insertions(+)
https://github.com/Perl/perl5/commit/757c2bf2ec268d71

New branch sv_dup_doc_fixes
2 commits. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/d7a4c9f2592f6bad

Improve documentation of sv_dup and sv_dup_inc
Leon Timmermans 1 file changed, 10 insertions(+), 10 deletions(-
https://github.com/Perl/perl5/commit/d7a4c9f2592f6bad

Revert marking sv_dup and sv_dup_inc as core-only
Leon Timmermans 1 file changed, 2 insertions(+), 2 deletions(-)
https://github.com/Perl/perl5/commit/a8ce676243573e3e

Deleted branch davem/xsdelta

Deleted branch smoke-me/jkeenan/cpan-socket-synch-20250704

Deleted branch smoke-me/jkeenan/experimental-sync-with-cpan-20250501

Deleted branch smoke-me/jkeenan/sync-math-bigint-20250413-4

Deleted branch smoke-me/jkeenan/sync-pod-simple-20250523

Martian commit ab3f47b8fbe075082aed1bbd5e82608409df0dab
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/ab3f47b8fbe07508

UCD.pm: Add check for parameter type
Karl Williamson 7 files changed, 22 insertions(+), 7 deletions(-
https://github.com/Perl/perl5/commit/ab3f47b8fbe07508

Martian commit 3a88cfc24c0cd92ec3849b337287670e0bd55cf0
1 commit. 1 unique author. 1 unique committer.

Snapshot: http://github.com/Perl/perl5/tarball/3a88cfc24c0cd92e

regen/embed.pl Use /xx to clarify regex
Karl Williamson 1 file changed, 1 insertion(+), 1 deletion(-)
https://github.com/Perl/perl5/commit/3a88cfc24c0cd92e

Martian commit 40a06973c4b5c541a716e34aedfe88a1995cd361
1 commit. 1 unique author. 1 unique committer.
Thanks, applied: GitHub (1)
Snapshot: http://github.com/Perl/perl5/tarball/40a06973c4b5c541

Merge branch 'blead' into auto_full_name
Karl Williamson 2 parents
https://github.com/Perl/perl5/commit/40a06973c4b5c541

Ignored 98 GitHub auto-generated merge commits

Re: How do I configure so as to exclude '-lpthread' from 'libs' and'perllibs' in 'perl -V'? by James E Keenan

Tue, 22 Jul 2025 23:01:36 +0000

On 7/22/25 08:45, Craig A. Berry wrote:
> On Sat, Jul 19, 2025 at 6:00 PM James E Keenan <[email protected]> wrote:
>
>> To repeat, the focus of this message is to learn what switches I would
>> supply to |./Configure| so as to *not* have |-lpthread| appear in |libs|
>> or |perllibs|.
>
> Have you tried clobbering libswanted with only the values you want to
> have end up in libs? E.g.,
>
> -A define:libswanted='dl m crypt util'
>
> Not sure it will give you what you want, and it doesn't make sense to
> me to tell the linker to use a library (via ldflags) that you have
> otherwise asserted you don't have, but that does seem to be what you
> are asking for.

Craig, thanks for looking into this.

My first attempt at implementing your suggestion had ./Configure barfing
very quickly:
#####
sh ./Configure -des -Dusedevel \
-Duseithreads \
-Doptimize="-O2 -pipe -fstack-protector-strong -fno-strict-aliasing" \
-A define:libs='dl m crypt util' \
-A define:perllibs='dl m crypt util' \
&& make test_prep

Checking your choice of C compiler and flags for coherency...
I've tried to compile and run the following simple program:

#include <stdio.h>
int main() { printf("Ok\n"); return(0); }

I used the command:

cc -o try -O2 -pipe -fstack-protector-strong -fno-strict-aliasing
-DHAS_FPSETMASK -DHAS_FLOATINGPOINT_H -DNO_POSIX_2008_LOCALE
-fno-strict-aliasing -pipe -fstack-protector-strong -I/usr/local/include
-pthread -Wl,-E -fstack-protector-strong -L/usr/local/lib try.c dl m
crypt util
./try

and I got the following output:

cc: error: no such file or directory: 'dl'
cc: error: no such file or directory: 'm'
cc: error: no such file or directory: 'crypt'
cc: error: no such file or directory: 'util'
I can't compile the test program.
You have a BIG problem. Shall I abort Configure [y]
Ok. Stopping Configure.
#####

I then added '-l' to each of the libraries in the value strings for
'libs' and 'perllibs'.

##########

sh ./Configure -des -Dusedevel \
-Duseithreads \
-Doptimize="-O2 -pipe -fstack-protector-strong -fno-strict-aliasing" \
-A define:libs='-ldl -lm -lcrypt -lutil' \
-A define:perllibs='-ldl -lm -lcrypt -lutil' \
&& make test_prep

$ ./perl -Ilib -V
Summary of my perl5 (revision 5 version 43 subversion 1) configuration:
Commit id: 89012989f3104d6b9f30be98d2060b806df0f95e
Platform:
osname=freebsd
osvers=14.2-release-p1
archname=amd64-freebsd-thread-multi
uname='freebsd perlmonger.nycbug.org 14.2-release-p1 freebsd
14.2-release-p1 generic amd64 '
config_args='-des -Dusedevel -Duseithreads -Doptimize=-O2 -pipe
-fstack-protector-strong -fno-strict-aliasing -A define:libs=-ldl -lm
-lcrypt -lutil -A define:perllibs=-ldl -lm -lcrypt -lutil'
...
useithreads=define
...
Linker and Libraries:
...
ldflags ='-pthread -Wl,-E -fstack-protector-strong -L/usr/local/lib'
...
libs=-ldl -lm -lcrypt -lutil
perllibs=-ldl -lm -lcrypt -lutil
...
#####

So, one *can* configure so that `-lpthread` does not appear in `libs` or
`perllibs`. I personally don't want to do that, but it *may* explain
the curious `perl -V` entries we're seeing in the vendor perl on
FreeBSD. Thanks again for your investigation.

Re: New XS features? by Michael Conrad

Tue, 22 Jul 2025 13:19:01 +0000

On 7/20/25 5:36 AM, Dave Mitchell wrote:
>>> Secondly, as a way of quickly creating a wrapper for a C library by
>>> parsing its .h file: should we still be recommending this?
>> I'm not sure it's working well enough for that. Possibly we should try
>> it on a bunch of them and see what it produces.
> For our readers: has anyone following this thread ever successfully used
> h2xs to generate a basic .xs file from a C header file for some library?
>
> If it turns out that h2xs is no longer up to the job, should be just
> drop all mention of it in the XS docs apart from creating an empty
> skeleton distro?

I got my feet wet with XS starting with Inline::C (around 2015). I used
that happily in my local projects before deciding to make some of it
into a CPAN module (around 2017) when I learned about Inline::Module,
and then later decided to fully upgrade to a standard XS dist to get the
last little bits of integration I wanted. With that direction, I
already had most of a distribution assembled before I needed XS
tooling. I think I looked into h2xs for the *next* module I wrote, but
I was interested in a Dist::Zilla pattern and the output didn't seem
useful, so I disregarded that and just copy/pasted a skeleton from my
previous project. Ever since then I've copy/pasted from whichever of my
previous XS modules is closest in structure to the new one I want to
write. I prefer high-integration with the library I'm wrapping, so
boilerplate calls to the C API doesn't usually provide any value to me.

While this doesn't say much about whether h2xs *can* be useful (my goal
was to write new C code, not just wrap a pre-existing C library) I think
my opinion on the topic is that the best recommendation to new XS users
is to start with Inline::C, and then if they wish to release it on cpan,
package it with Inline::Module. This has significantly less learning
curve than immediately jumping into XS module tooling.

While XS tooling isn't as bad as Autoconf, it has the same
barrier-to-entry effect of "ok so here's 2 weeks of boring complicated
arcane learning curve you need to wrestle with before you can get to the
fun stuff, and BTW none of the cool kids are doing it this way anymore".

-Mike C

Re: How do I configure so as to exclude '-lpthread' from 'libs' and'perllibs' in 'perl -V'? by Craig A. Berry

Tue, 22 Jul 2025 12:45:36 +0000

On Sat, Jul 19, 2025 at 6:00â€¯PM James E Keenan <[email protected]> wrote:

> To repeat, the focus of this message is to learn what switches I would
> supply to |./Configure| so as to *not* have |-lpthread| appear in |libs|
> or |perllibs|.

Have you tried clobbering libswanted with only the values you want to
have end up in libs? E.g.,

-A define:libswanted='dl m crypt util'

Not sure it will give you what you want, and it doesn't make sense to
me to tell the linker to use a library (via ldflags) that you have
otherwise asserted you don't have, but that does seem to be what you
are asking for.

Re: State of XSpp / C++ integration by James E Keenan

Tue, 22 Jul 2025 12:21:07 +0000

On 7/22/25 07:27, Johan Vromans wrote:
> While trying to port wxPerl to a more recent version of wxWidgets I
> stumbled upon a number of XSpp issues that I find hard to solve.
>
> Before spending a lot of energy I would first like to ask if XSpp is still
> a viable toolchain for C++ embedding. It hasn't been updated for many years
> and has a load of old issues that have never looked at.
>
> If someone with practical XSpp experience would be able to provide some
> help I would welcome it.
>
> Thanks.

#####
$ corelist ExtUtils::XSpp

Data for 2025-03-21
ExtUtils::XSpp was not in CORE (or so I think)
#####

So XSpp is not in the core distribution or maintained by Perl 5 Porters.
I cannot recall hearing of it on this list before. So my hunch is
that your impression of its viability is correct.

https://metacpan.org/dist/ExtUtils-XSpp

State of XSpp / C++ integration by Johan Vromans

Tue, 22 Jul 2025 11:27:55 +0000

While trying to port wxPerl to a more recent version of wxWidgets I
stumbled upon a number of XSpp issues that I find hard to solve.

Before spending a lot of energy I would first like to ask if XSpp is still
a viable toolchain for C++ embedding. It hasn't been updated for many years
and has a load of old issues that have never looked at.

If someone with practical XSpp experience would be able to provide some
help I would welcome it.

Thanks.

Re: New XS features? by Jon Gentle

Tue, 22 Jul 2025 03:46:32 +0000

On Sun, Jul 20, 2025 at 5:36â€¯AM Dave Mitchell <[email protected]> wrote:
>
> On Fri, Jul 18, 2025 at 02:45:01AM +0200, Leon Timmermans wrote:
> > > Secondly, as a way of quickly creating a wrapper for a C library by
> > > parsing its .h file: should we still be recommending this?
> >
> > I'm not sure it's working well enough for that. Possibly we should try
> > it on a bunch of them and see what it produces.
>
> For our readers: has anyone following this thread ever successfully used
> h2xs to generate a basic .xs file from a C header file for some library?
>
> If it turns out that h2xs is no longer up to the job, should be just
> drop all mention of it in the XS docs apart from creating an empty
> skeleton distro?

I attempted to use h2xs a little over 2 years ago to create a module
around a C header; specifically I attempted to use it for WebGPU
native, a single .h file that allows languages to interface with
WebGPU through a C API. This was my third XS module, and in each case
I did not feel that h2xs lived up to my expectations. At best it
helped me to know how to structure an XS file. For WebGPU, I pretty
quickly ended up abandoning h2xs in favor of a perl script that would
parse webgpu.h and generate the required files for a couple of
reasons:

* All of the functions and constants are dumped into a single package,
no options to separate them somehow by prefix for instance.
* It created all of the structs in unprefixed packages: i.e.
WGPUBindGroupDescriptor instead of
WebGPU::Direct::WGPUBindGroupDescriptor.
* Structs ended up with 2 packages: WGPUBindGroupDescriptor and
WGPUBindGroupDescriptorPtr, which map to the C types
WGPUBindGroupDescriptor and WGPUBindGroupDescriptor*.
* The package that ends with Ptr actually had the struct manipulation
functions in it, so if I wanted to set or get `entryCount` on
WGPUBindGroupDescriptor, I have to call new on WGPUBindGroupDescriptor
first, then transform it into a WGPUBindGroupDescriptorPtr object by
using the `_to_ptr` method. This conversion appears to be one-way, you
cannot get the non-Ptr version from the Ptr version.
* I could not find enums mapped to constants. Also it mapped the enum
types to T_PTROBJ in the typemap
* It mapped several common types to T_PTROBJ in the typemap, including
char const *, double, int and size_t to name a few.
* I saw no option for rerunning h2xs on an existing module that wasn't
a complete overwrite. This is an issue since the webgpu API is still
actively being developed, but even in general this would prohibit
using h2xs again when an API adds a function.
* Not that I expected it to, but h2xs had no option to organize groups
of functions and types into seperate objects. This ultimately became
the biggest pain point for WebGPU since all but one function operates
off an opaque object type as its first parameter, and those types need
to be separated into different packages in order for the API to make
sense.

I did attempt to rerun h2xs on webgpu.h to see if the file I was
looking at was manipulated before being committed, but I could not get
h2xs to run correctly on webgpu.h. I'm not sure how I originally
coaxed h2xs to produce the file that got committed. The simple fact
that it takes quite a bit of effort to get h2xs to run correctly on
what is, in my opinion, a straightforward C header is concerning.

Overall it feels like h2xs was created for a different era, that its
purpose is to allow access to C libraries directly from perl. I would
think the use case is more like if there was some interesting vendor
library that you wanted to prototype quickly with, this would get you
access to it from perl instead of having to write a C project. The
fact that you can start a module always seemed like an unrelated
feature that was bolted on and ultimately became the defining feature.
It has always been strange to me that the suggestion in perlmodlib is
to use h2xs to start a project when I have no C header nor xs in a
project. But by all accounts h2xs does generate a usable module
skeleton. I personally would prefer that the tool that perl core ships
for generating a simple, modern module skeleton have a name that
reflects its purpose and was well documented about the upgrade path to
more advanced tools.

-Jon

Perl 5.40.3-RC1 is now available! by Steve Hay via perl5-porters

Mon, 21 Jul 2025 20:25:24 +0000

In 1522, thirty years after Columbus' landfall in the Bahamas, another
ship sailing west in the service of Spain completed the first voyage
round the world. It had set out under the command of a Portuguese,
Magellan, who had got as far as the Philippines, where he was killed.
By then he had discovered and sailed through the straits named after him
and with this voyage and its demonstration that the great oceans were
all interconnected, the prologue to the European age can be considered
over.

-- J. M. Roberts, "The Penguin History of Europe"

We are pleased to announce the first release candidate of version 40.3,
the third maintenance release of version 40 of Perl 5.

You will soon be able to download Perl 5.40.3-RC1 from your favourite
CPAN mirror or find it at:

https://metacpan.org/release/SHAY/perl-5.40.3-RC1/

SHA256 digests for this release are:

*perl-5.40.3-RC1.tar.gz:
f0771e49db212191347589da4315ca5e002105493e08372b525201f015d15ec1

*perl-5.40.3-RC1.tar.xz:
e0d3e4555458cbfbbd994d024d676150286d8f8dcd48a1deeafcab21a64783d6

You can find a full list of changes in the file "perldelta.pod" located
in the "pod" directory inside the release and on the web.

***Please, please, please*** test your code against perl-5.40.3-RC1.
This is your last chance to point out any critical regressions before
v5.40.3 ships "for real," and we'd rather fix things for v5.40.3 than
for v5.40.4!

The final v5.40.3 is expected to ship on 3 August.

Steve Hay

Perl 5.38.5-RC1 is now available! by Steve Hay via perl5-porters

Mon, 21 Jul 2025 20:25:22 +0000

Columbus set sail in 1492 and after 69 days his three little ships made
a landfall in the Bahamas. A fortnight later he came to Cuba, which he
named Hispaniola, before returning to Spain. The following year he came
back to explore the islands known ever since as the West Indies (whose
name commemorates Columbus' tenacious belief that he had reached Asia).
His confident leap in the dark had changed world history. Without
knowing it he had discovered a new world.

-- J. M. Roberts, "The Penguin History of Europe"

We are pleased to announce the first release candidate of version 38.5,
the fifth maintenance release of version 38 of Perl 5.

You will soon be able to download Perl 5.38.5-RC1 from your favourite
CPAN mirror or find it at:

https://metacpan.org/release/SHAY/perl-5.38.5-RC1/

SHA256 digests for this release are:

*perl-5.38.5-RC1.tar.gz:
a7cd66e3b82a1740bdfb39097d143da1effa4c772ae5686185373a8f94ecbe72

*perl-5.38.5-RC1.tar.xz:
ef59ce5ca010beac656709c3b73ffc8d43ca7819b3c119e8bd5ddf0d983c33db

You can find a full list of changes in the file "perldelta.pod" located
in the "pod" directory inside the release and on the web.

***Please, please, please*** test your code against perl-5.38.5-RC1.
This is your last chance to point out any critical regressions before
v5.38.5 ships "for real," and we'd rather fix things for v5.38.5 than
for v5.38.6!

The final v5.38.5 is expected to ship on 3 August.

Steve Hay

Re: New XS features? by "George S. Baugh" via perl5-porters

Mon, 21 Jul 2025 16:00:57 +0000

Tried h2xs out the other day to see if I could wrap libpg_query

The primary problem I saw was that it generated totally useless
typemaps (e.g. everything is OPAQUE_STRUCT, couldn't even figure out
basic char* -> SV either)

All the XS examples in perlxstut basically showed writing raw XS and
ignoring typemaps for everything but the most trivial types

So I tried protobuf-perlxs, which also does not work (looks like it
needs another layer of code generator atop it to read the protobuf and
build the xs/pm files).

Current strategy for "struct -> hash" is "crib from one of the
libraries deserializing arbitrary data into hashes, e.g. JSON::XS", but
that will take a good deal longer, which I don't have time for right
now.

Maybe there's some good off-the-shelf typemaps somewhere but I haven't
found them.

On Sun, 2025-07-20 at 13:01 +0100, Dave Mitchell wrote:
> On Fri, Jul 18, 2025 at 02:45:01AM +0200, Leon Timmermans wrote:
> > > Secondly, as a way of quickly creating a wrapper for a C library
> > > by
> > > parsing its .h file: should we still be recommending this?
> >
> > I'm not sure it's working well enough for that. Possibly we should
> > try
> > it on a bunch of them and see what it produces.
>
> I had a quick play with h2xs -x (x = extract info from header file
> and
> generate XSUBs) and didn't get far. Not encouraging. This is after
> installingÂ C::Scan.
>
> $ h2xs5.43.1 -n MySSH2 -x /usr/include/libssh2.h
> Defaulting to backwards compatibility with perl 5.43.1
> If you intend this module to be compatible with earlier perl
> versions, please
> specify a minimum perl version with the -b option.
>
> Writing MySSH2/ppport.h
> Scanning typemaps...
> Â Scanning /home/davem/perl5/git/bleed.out/lib/5.43.1/ExtUtils/typemap
> Scanning /usr/include/libssh2.h for functions...
> panic: multiple types without intervening comma in
> void (*libssh2_trace_handler_func)(LIBSSH2_SESSION*,
> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â void *,
> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â const char *,
> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â size_t)
> whited-out as
> Â Â Â Â Â (*libssh2_trace_handler_func)(LIBSSH2_SESSION*,
> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â *,
> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â *,
> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â size_t)
> Cannot process declaration `__attribute__((deprecated("since libssh2
> " "1.11.1" ". " "Use libssh2_session_callback_set2()")))
> Â void *libssh2_session_callback_set(LIBSSH2_SESSION *session,
> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â int cbtype, void
> *callback)' without an identifier at
> /home/davem/perl5/git/bleed.out/lib/site_perl/5.43.1/C/Scan.pm line
> 720.
>
>
> I've also now noticed that perlxstut.pod doesn't actually include an
> example of using 'h2xs -x'
>

Re: New XS features? by Maxim

Mon, 21 Jul 2025 07:49:13 +0000

Dave Mitchell, 2025-07-20 10:36 +0100:
> For our readers: has anyone following this thread ever successfully used
> h2xs to generate a basic .xs file from a C header file for some library?

My experience is a drop in the ocean, but just in case...

I have seriously tried XS only once when I was wrapping a C library in
Perl a long time ago. I don't remember using h2xs, but looking at its
output now, I suspect I might have gotten too intimidated back then.

- It doesn't just "convert .h C header files to Perl extensions",
but creates an entire distribution.
- To generate XSUBs (flag -x) a non-core module C::Scan is required.
- The generated code doesn't compile.
- The generated code is also a lot to understand and handle on top
of already non-trivial subject of XS.

I did use ExtUtils::Constant back then, which is documented to be based
on the code of h2xs, whatever that means.

Re: Magic v2 - Review for Merge and next steps by Ovid

Mon, 21 Jul 2025 05:51:14 +0000

On Sun, Jul 20, 2025 at 3:11â€¯PM Leon Timmermans <[email protected]> wrote:

> >> I don't think it's quite clear why we would want value hooks.
>
<snip>

> Value magic as an annotation sounds useful and has precedent (e.g.
> vstrings).

<snip>

> I think that [COW would] require an unholy amount of backreferences to
>> everything that owns a reference to the current value.
>>
> <snip>

> How would [not printing sensivive information] work? ...
>>
> <snip>

I think this is getting into implementation details, which is great, but
your original concern was "I don't think it's quite clear why we would want
value hooks." That's the question I was trying to answer.

The beauty of this idea is, as Paul pointed out, is that instead of ten
different cool features having ten different implementations, we have one
underlying idea that all makes all of those features possible, plus many
more that we've never thought of. To me, instead of a red flag, it's a
green flag. It doesn't *prove* the idea is good, but it's a great indicator.

Best,
Ovid
--
https://curtispoe.org/

Release announcement for perl v5.43.1 by Richard Leach

Sun, 20 Jul 2025 21:56:56 +0000

The water from the brine is evaporated in enormous hot vessels,
which leave the salt about the same consistency as wet sand.
After a run through another oven, what emerges are glistening
pure crystals of salt. I run my hands through the grains as
they emerge, surprised for a moment by how hot they are to touch.
'This is ultra fine,' says Joe. 'It goes into Wotsits.'

-- Ed Conway, "Material World"

We are as keen as mustard to announce version 5.43.1,
the first development release of version 5.43 of Perl.

You will soon be able to download Perl 5.43.1 from your
favorite CPAN mirror or find it at:

https://metacpan.org/release/HYDAHY/perl-5.43.1/

SHA256 digests for this release are:

5221ebf5badfbb943d168ff589ce93456a11f219105c930cc01e8a82a62adb65
perl-5.43.1.tar.gz
260fa2f8cae4a700083f48db70c2eb56abc3e45a166a6eb22df3319aef7eb141
perl-5.43.1.tar.xz

You can find a full list of changes in the file "perldelta.pod" located in
the "pod" directory inside the release and on the web at

https://metacpan.org/release/HYDAHY/perl-5.43.1/view/pod/perldelta.pod

Perl 5.43.1 represents approximately 2 weeks of development since Perl
5.43.0 and contains approximately 15,000 lines of changes across 410 files
from 21 authors.

Excluding auto-generated files, documentation and release tools, there were
approximately 12,000 lines of changes to 360 .pm, .t, .c and .h files.

Perl continues to flourish into its fourth decade thanks to a vibrant
community of users and developers. The following people are known to have
contributed the improvements that became Perl 5.43.1:

Alexander Karelas, Chad Granum, Chris 'BinGOs' Williams, Chris Prather,
Dagfinn Ilmari MannsÃ¥ker, Dan Book, Daniel Dragan, Daniel LaÃ¼gt, Dave Cross,
David Mitchell, James E Keenan, Karen Etheridge, Karl Williamson, Leon
Timmermans, Paul Evans, Philippe Bruhat (BooK), Richard Leach, Å tÄ›pÃ¡n NÄ›mec,
TAKAI Kousuke, Thibault Duponchelle, Tony Cook.

The list above is almost certainly incomplete as it is automatically
generated from version control history. In particular, it does not include
the names of the (very much appreciated) contributors who reported issues to
the Perl bug tracker.

Many of the changes included in this version originated in the CPAN modules
included in Perl's core. We're grateful to the entire CPAN community for
helping Perl to flourish.

For a more complete list of all of Perl's historical contributors, please
see the F<AUTHORS> file in the Perl source distribution.

We expect to release version 5.43.2 on 2025-08-20.
The next major stable release of Perl should appear in the first half of 2026.

Om nom nom nom

Re: New XS features? by Leon Timmermans

Sun, 20 Jul 2025 14:45:16 +0000

On Sun, Jul 20, 2025 at 11:36â€¯AM Dave Mitchell <[email protected]> wrote:
>
> On Fri, Jul 18, 2025 at 02:45:01AM +0200, Leon Timmermans wrote:
> > > First, for creating a skeleton disto when starting a new project (or when
> > > following tutorial examples): what should be be recommended instead.
> >
> > Modern authoring tools have built-in support for creating a new
> > distribution. Personally I usually use Dist::Zilla with my TemplateXS
> > plugin, it has a minimal but useful XS template. There's more than one
> > way to do it though; that makes this rather tricky.
>
> Well, for the specific case of creating a skeleton distro for playing with
> when following along in perlxstut.pod, do you think its still ok to to
> tell the reader to use "h2xs -A -n Foo::Bar"?

It's a valid option, but I'm not sure it's a good option (e.g.
defaulting to requiring the current perl version at the minimum is
extremely unhelpful).

Honestly I think any of us could write a more helpful tool in an
afternoon, because there's a general pattern of h2xs trying to do way
too much. Even a tool doing the absolute minimum would be more usable.

> For the more general case case of creating a skeleton distro when
> starting a new project, perhaps we could say that 'h2xs -n' is bundled and
> will produce a basic skeleton, but see Dist::Zilla (and other examples??)
> for more modern ways of doing it?

Maybe. I think Module::Starter can also be configured to do something
like this, I just have never done so. I'm sure there are more
possibilities that aren't even on my radar (Minilla?). And to be
honest, I'm not sure we even need a tool.

> PS - are you happy in principle with my "'REQUIRE 4.0' implies strict"
> proposal? I don't think anyone's commented on it yet.

I'm good with "'REQUIRE 4.0' implies strict", I'm just still not sure
if "REQUIRE" is the best approach here (as opposed to having only
"STRICT: ENABLED" or some such). I'm worried we're over-engineering it
now to compensate for the historic under-engineering.

Explicit versioning does have a certain mental cost, and I suspect we
actively want to keep that cost low here. That means that if we do
this, we should pack a bunch of features so that we don't have to
release a 5.0 any time soon.

Leon

Re: Magic v2 - Review for Merge and next steps by Leon Timmermans

Sun, 20 Jul 2025 13:11:16 +0000

On Sun, Jul 20, 2025 at 8:59â€¯AM Ovid <[email protected]> wrote:
>
> On Sun, Jul 20, 2025 at 2:32â€¯AM Leon Timmermans <[email protected]> wrote:
>>
>> On Sat, Jul 12, 2025 at 4:06â€¯PM Paul "LeoNerd" Evans
>> <[email protected]> wrote:
>> > TL;DR: Please help review my "magic v2" so far so I can add more useful
>> > abilities that will provide a good foundation for implementing a
>> > whole bunch of features on top of.
>> <snip>
>> I don't think it's quite clear why we would want value hooks. I don't
>> think there is a good description of what problems we have that this
>> would solve. I don't think it's obvious upfront what kind of changes
>> are required in the codebase.
>
>
> This is one of those areas where I think we open up a whole host of capabilities and if people start playing with the idea, we'll find use cases we never dreamed of.
>
> I've said in the past that I'd like value hooks to validate if a variable is undefined or uninitialized. Having "infectious" values (like taint) would also help here.

Value magic as an annotation sounds useful and has precedent (e.g. vstrings).

> COW magic might be possible? my $foo :Cow = $bar; could assign a reference to a huge data structure, but if anyone tried to change $foo, it would try to clone it first (not always possible, I know).

I think that would require an unholy amount of backreferences to
everything that owns a reference to the current value. COW works for
strings because they don't nest.

It does point out an important challenge with the idea of value magic:
our current set magic is *after* the assignment (which makes sense for
variable magic), but value magic generally wants it to happen *before*
the assignment. This alone would mean a vast amount of change in the
codebase (various places that set SVs).

> Or maybe we could declare some values as immutable, without having to use constants? Here are a couple of examples that are annoying:
>
> use Readonly;
> use Hash::Util qw(lock_hash);
>
> Readonly my $x => "hello";
> my $y = $x; # $y is NOT read-only, infection doesn't happen
>
> my %h = (a => 1, b => 2);
> lock_hash(%h);
> my %copy = %h; # %copy is a regular hash, not locked

I suspect that would fundamentally break the language. What should this do?

my %a
my %b : Readonly = (a => 1, b => 2);
%a = %b;
%a = ();

> Or someone tries to print a password or PII and it's printed as [Redacted]? Also, lots of debugging opportunities here.

How would that work? As an annotation that print specifically looks
for? How would it work if someone appends that password to another
string first?

Leon

Re: New XS features? by Dave Mitchell

Sun, 20 Jul 2025 12:01:36 +0000

On Fri, Jul 18, 2025 at 02:45:01AM +0200, Leon Timmermans wrote:
> > Secondly, as a way of quickly creating a wrapper for a C library by
> > parsing its .h file: should we still be recommending this?
>
> I'm not sure it's working well enough for that. Possibly we should try
> it on a bunch of them and see what it produces.

I had a quick play with h2xs -x (x = extract info from header file and
generate XSUBs) and didn't get far. Not encouraging. This is after
installing C::Scan.

$ h2xs5.43.1 -n MySSH2 -x /usr/include/libssh2.h
Defaulting to backwards compatibility with perl 5.43.1
If you intend this module to be compatible with earlier perl versions, please
specify a minimum perl version with the -b option.

Writing MySSH2/ppport.h
Scanning typemaps...
Scanning /home/davem/perl5/git/bleed.out/lib/5.43.1/ExtUtils/typemap
Scanning /usr/include/libssh2.h for functions...
panic: multiple types without intervening comma in
void (*libssh2_trace_handler_func)(LIBSSH2_SESSION*,
void *,
const char *,
size_t)
whited-out as
(*libssh2_trace_handler_func)(LIBSSH2_SESSION*,
*,
*,
size_t)
Cannot process declaration `__attribute__((deprecated("since libssh2 " "1.11.1" ". " "Use libssh2_session_callback_set2()")))
void *libssh2_session_callback_set(LIBSSH2_SESSION *session,
int cbtype, void *callback)' without an identifier at /home/davem/perl5/git/bleed.out/lib/site_perl/5.43.1/C/Scan.pm line 720.

I've also now noticed that perlxstut.pod doesn't actually include an
example of using 'h2xs -x'

--
Diplomacy is telling someone to go to hell in such a way that they'll
look forward to the trip

Re: New XS features? by Dan

Sun, 20 Jul 2025 10:35:26 +0000

On Thu, Jul 17, 2025 at 8:46â€¯PM Leon Timmermans <[email protected]> wrote:

> Modern authoring tools have built-in support for creating a new
> distribution. Personally I usually use Dist::Zilla with my TemplateXS
> plugin, it has a minimal but useful XS template. There's more than one
> way to do it though; that makes this rather tricky.
>

I have no experience in creating XS dists, but
https://metacpan.org/pod/Dist::Zilla::MintingProfile::Starter describes how
to easily customize it and I would love to be able to include an
XS-specific minting profile as an example if you have any suggestions.

-Dan

Re: New XS features? by Dave Mitchell

Sun, 20 Jul 2025 09:36:25 +0000

On Fri, Jul 18, 2025 at 02:45:01AM +0200, Leon Timmermans wrote:
> > First, for creating a skeleton disto when starting a new project (or when
> > following tutorial examples): what should be be recommended instead.
>
> Modern authoring tools have built-in support for creating a new
> distribution. Personally I usually use Dist::Zilla with my TemplateXS
> plugin, it has a minimal but useful XS template. There's more than one
> way to do it though; that makes this rather tricky.

Well, for the specific case of creating a skeleton distro for playing with
when following along in perlxstut.pod, do you think its still ok to to
tell the reader to use "h2xs -A -n Foo::Bar"?

For the more general case case of creating a skeleton distro when
starting a new project, perhaps we could say that 'h2xs -n' is bundled and
will produce a basic skeleton, but see Dist::Zilla (and other examples??)
for more modern ways of doing it?

> > Secondly, as a way of quickly creating a wrapper for a C library by
> > parsing its .h file: should we still be recommending this?
>
> I'm not sure it's working well enough for that. Possibly we should try
> it on a bunch of them and see what it produces.

For our readers: has anyone following this thread ever successfully used
h2xs to generate a basic .xs file from a C header file for some library?

If it turns out that h2xs is no longer up to the job, should be just
drop all mention of it in the XS docs apart from creating an empty
skeleton distro?

> > > We could just disallow all-caps labels in strict mode?
> >
> > Not sure. Seems a bit too extreme?
>
> > Not sure. Seems a bit too extreme?
> Maybe. I never write all-caps labels so to me it feels like a
> reasonable constraint. We have worse constraints to be honest
> (especially around #ifdefs).

Fair enough.

PS - are you happy in principle with my "'REQUIRE 4.0' implies strict"
proposal? I don't think anyone's commented on it yet.

--
Diplomacy is telling someone to go to hell in such a way that they'll
look forward to the trip

Re: Magic v2 - Review for Merge and next steps by Ovid

Sun, 20 Jul 2025 06:59:58 +0000

On Sun, Jul 20, 2025 at 2:32â€¯AM Leon Timmermans <[email protected]> wrote:

> On Sat, Jul 12, 2025 at 4:06â€¯PM Paul "LeoNerd" Evans
> <[email protected]> wrote:
> > TL;DR: Please help review my "magic v2" so far so I can add more useful
> > abilities that will provide a good foundation for implementing a
> > whole bunch of features on top of.
> <snip>
> I don't think it's quite clear why we would want value hooks. I don't
> think there is a good description of what problems we have that this
> would solve. I don't think it's obvious upfront what kind of changes
> are required in the codebase.
>

This is one of those areas where I think we open up a whole host of
capabilities and if people start playing with the idea, we'll find use
cases we never dreamed of.

I've said in the past that I'd like value hooks to validate if a variable
is undefined or uninitialized. Having "infectious" values (like taint)
would also help here.

COW magic might be possible? my $foo :Cow = $bar; could assign a reference
to a huge data structure, but if anyone tried to change $foo, it would try
to clone it first (not always possible, I know). Or maybe we could declare
some values as immutable, without having to use constants? Here are a
couple of examples that are annoying:

use Readonly;
use Hash::Util qw(lock_hash);

Readonly my $x => "hello";
my $y = $x; # $y is NOT read-only, infection doesn't happen

my %h = (a => 1, b => 2);
lock_hash(%h);
my %copy = %h; # %copy is a regular hash, not locked

Or someone tries to print a password or PII and it's printed as [Redacted]?
Also, lots of debugging opportunities here.

So, tons of use cases I could think of. Some of this we can do with overload,
but it's painful, limited, and not really transparent.

Best,
Ovid

--
https://curtispoe.org/

Re: Magic v2 - Review for Merge and next steps by Leon Timmermans

Sun, 20 Jul 2025 00:32:38 +0000

On Sat, Jul 12, 2025 at 4:06â€¯PM Paul "LeoNerd" Evans
<[email protected]> wrote:
> TL;DR: Please help review my "magic v2" so far so I can add more useful
> abilities that will provide a good foundation for implementing a
> whole bunch of features on top of.
>
> I have a slowly-growing collection of ideas and partial experiments,
> some of which are written up in PPC documents, that all involve adding
> extra features, extra abilities, that would be useful in various
> situations. A lot of these are partly stalled for want of a decent way
> to encapsulate "extension behaviour". For example, a module or core
> feature that wanted to provide an attribute for subroutine signature
> parameters will need some way to provide extra code that may be
> required at various points during compilation or runtime.
>
> I have a bunch of existing modules that try to do variations on this
> kind of idea, by providing some sort of C-level data structure
> definition that extension modules can provide into the base system, to
> contain those extensions. There's a converging design among many of
> these, that it'd be most handy to have a data structure that contains
> function pointers, each of which is run at a particular phase of
> operation. These structures then define a kind of behaviour, provided
> by a module, that can be attached to various target instances.
>
> In many ways this sounds exactly like Perl's "MAGIC". MAGIC is useful
> for adding trigger options to scalar variables, but has some
> shortcomings when trying to apply it to other situations. Perl core
> does attempt with some limited success to reuse the same magic virtual
> table definition on non-scalars (i.e. arrays, hashes and subroutines),
> but this is of limited utility. Furthermore, there's no ability to
> create kinds of magic that attach to the semantic idea of a "value",
> and follow it around as it gets copied and assigned to various other
> scalars.
>
> The reason that existing magic is rather limited in ability is because
> the entire set of behaviours provided by a kind of magic are
> encapsulated in a bunch of functions whose addresses are given in
> pointers in the "Magic virtual table" structure (MGVTBL), but that
> virtual table doesn't have any other fields to store metadata of any
> kind. It defines a fixed set of 8 different functions for various
> scalar-variable related scenarios, and makes it difficult to extend to
> other situations, as would apply to arrays, hashes, or other things.
>
> The main idea of my "Magic v2" redesign therefore is to extend the
> definition of this vtable, with a whole new structure shape that has
> metadata fields for what "shape" it is, a version number, some flags,
> etc... Various different "shapes" are then defined for various
> situations. There's a "scalar variable" shape that mostly matches the
> uses of existing magic, with plenty of space for other shapes to be
> added with different sets of function pointers.
>
> Right now, I have this redesign in a branch that contains scalar, array
> and hash variable shapes, and a shape intended to apply to scalar
> *values*, with code around that tries to copy the magic attachment
> around when the value itself is copied.
>
> Of these, the very last one - creating the concept of scalar value
> magic - still has a number of performance-related questions that have a
> measurable impact on the operation of the interpreter even in programs
> that do not make use of it. So that part at least isn't ready for
> consideration of merge into blead at this time.
>
> However, now that we can start considering things to look into bringing
> into development for (5.)43.1, I would like to propose the basic part
> of Magic v2 (the base mechanism with the ability to have different
> shapes), as well as the stable parts of scalar/array/hash variable
> magic. All of these are just neatened-up rewrites of existing
> behaviours, but made in a way as to allow a lot of extension in future.

I'm very much in favor of us gaining better variable magic, but I have
some serious reservations about value magic.

For variable magic, I believe it's fairly clear what we would want to
achieve. Essentially it's ties without some of the limitations of ties
(performance and otherwise). The rough shapes of what that would look
like are obvious (e.g. a vtable with a method for every array
operation). Given ties are already there, I expect that the required
changes in the rest of the code are also straightforward (refactor
array/hash tie code to use the hooks as a layer of indirection
instead).

I don't think it's quite clear why we would want value hooks. I don't
think there is a good description of what problems we have that this
would solve. I don't think it's obvious upfront what kind of changes
are required in the codebase.

We can't meaningfully discuss a solution before we have agreed on the
problem it's meant to solve. Even a basic question like "what kind of
hooks would we need" and "where would those hooks be called" do not
have an answer right now.

Leon