Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Missing newSVsv_nomg macro variant #16461

Closed
p5pRT opened this issue Mar 12, 2018 · 27 comments

Comments

@p5pRT
Copy link
Collaborator

commented Mar 12, 2018

Migrated from rt.perl.org#132964 (status was 'pending release')

Searchable as RT132964$

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Mar 12, 2018

From @pali

Hi! In perl XS api is missing function which creates a copy of the
scalar without processing get magic and without destroying source scalar
when is mortal.

It is needed in situation when you want to call other perl function (via
call_pv()) with scalar argument which comes from the argument passed to
the current XS function (e.g. ST(0)) on which was already called
SvGETMAGIC() and it is needed to prevent modification of it argument
(e.g. because it is used also after call_pv() call).

Table of available functions​:

  magic steal mortal
  target source source target
sv_setsv() X X
sv_setsv_mg() X X X
sv_setsv_nomg() X
SvSetMagicSV() X X X
SvSetMagicSV_nosteal() X X
SvSetSV() X X
SvSetSV_nosteal() X
sv_mortalcopy() X X X
newSVsv() X
sv_setsv_flags() CAN CAN CAN

Function which would create a copy of the scalar without processing get
magic and without destroying (stealing) source scalar can be caller
newSVsv_nomg() (according to to other _nomg names).

Primitive implementation is there​:

static SV *newSVsv_nomg(SV *sv) {
  SV *ret = newSV(0);
  sv_setsv_flags(ret, sv, SV_NOSTEAL);
  return ret;
}

For above case with call_pv() can be useful function like
newSVsv_nomg(), but which would return mortal copy. Based on above used
naming scheme, I would propose sv_mortalcopy_nosteal_nomg().

Primitive implementation​:

static SV *sv_mortalcopy_nosteal_nomg(SV *sv) {
  SV *ret = sv_newmortal();
  sv_setsv_flags(ret, sv, SV_NOSTEAL);
  return ret;
}

Please consider implementing these two functions into perl.h/perlapi.

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Aug 9, 2018

From @pali

On Monday 12 March 2018 10​:59​:55 pali@​cpan.org wrote​:

Hi! In perl XS api is missing function which creates a copy of the
scalar without processing get magic and without destroying source scalar
when is mortal.

It is needed in situation when you want to call other perl function (via
call_pv()) with scalar argument which comes from the argument passed to
the current XS function (e.g. ST(0)) on which was already called
SvGETMAGIC() and it is needed to prevent modification of it argument
(e.g. because it is used also after call_pv() call).

Table of available functions​:

                         magic       steal    mortal
                    target  source   source   target

sv_setsv() X X
sv_setsv_mg() X X X
sv_setsv_nomg() X
SvSetMagicSV() X X X
SvSetMagicSV_nosteal() X X
SvSetSV() X X
SvSetSV_nosteal() X
sv_mortalcopy() X X X
newSVsv() X
sv_setsv_flags() CAN CAN CAN

Function which would create a copy of the scalar without processing get
magic and without destroying (stealing) source scalar can be caller
newSVsv_nomg() (according to to other _nomg names).

Primitive implementation is there​:

static SV *newSVsv_nomg(SV *sv) {
SV *ret = newSV(0);
sv_setsv_flags(ret, sv, SV_NOSTEAL);
return ret;
}

For above case with call_pv() can be useful function like
newSVsv_nomg(), but which would return mortal copy. Based on above used
naming scheme, I would propose sv_mortalcopy_nosteal_nomg().

Primitive implementation​:

static SV *sv_mortalcopy_nosteal_nomg(SV *sv) {
SV *ret = sv_newmortal();
sv_setsv_flags(ret, sv, SV_NOSTEAL);
return ret;
}

Please consider implementing these two functions into perl.h/perlapi.

Now I found that in blead is undocumented macro sv_mortalcopy_flags and
function Perl_sv_mortalcopy_flags() which could potentially replace that
idea for sv_mortalcopy_nosteal_nomg(). So what about documenting macro
sv_mortalcopy_flags and making it public?

Also there is still a need for newSVsv_nomg()-like function or macro.

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Jan 21, 2019

From @pali

On Thursday 09 August 2018 13​:22​:18 pali@​cpan.org wrote​:

On Monday 12 March 2018 10​:59​:55 pali@​cpan.org wrote​:

Hi! In perl XS api is missing function which creates a copy of the
scalar without processing get magic and without destroying source scalar
when is mortal.

It is needed in situation when you want to call other perl function (via
call_pv()) with scalar argument which comes from the argument passed to
the current XS function (e.g. ST(0)) on which was already called
SvGETMAGIC() and it is needed to prevent modification of it argument
(e.g. because it is used also after call_pv() call).

Table of available functions​:

                         magic       steal    mortal
                    target  source   source   target

sv_setsv() X X
sv_setsv_mg() X X X
sv_setsv_nomg() X
SvSetMagicSV() X X X
SvSetMagicSV_nosteal() X X
SvSetSV() X X
SvSetSV_nosteal() X
sv_mortalcopy() X X X
newSVsv() X
sv_setsv_flags() CAN CAN CAN

Function which would create a copy of the scalar without processing get
magic and without destroying (stealing) source scalar can be caller
newSVsv_nomg() (according to to other _nomg names).

Primitive implementation is there​:

static SV *newSVsv_nomg(SV *sv) {
SV *ret = newSV(0);
sv_setsv_flags(ret, sv, SV_NOSTEAL);
return ret;
}

For above case with call_pv() can be useful function like
newSVsv_nomg(), but which would return mortal copy. Based on above used
naming scheme, I would propose sv_mortalcopy_nosteal_nomg().

Primitive implementation​:

static SV *sv_mortalcopy_nosteal_nomg(SV *sv) {
SV *ret = sv_newmortal();
sv_setsv_flags(ret, sv, SV_NOSTEAL);
return ret;
}

Please consider implementing these two functions into perl.h/perlapi.

Now I found that in blead is undocumented macro sv_mortalcopy_flags and
function Perl_sv_mortalcopy_flags() which could potentially replace that
idea for sv_mortalcopy_nosteal_nomg(). So what about documenting macro
sv_mortalcopy_flags and making it public?

Also there is still a need for newSVsv_nomg()-like function or macro.

Hi! Any comments for making Perl_sv_mortalcopy_flags() function public?
And for providing newSVsv_nomg()-like function?

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 7, 2019

From @pali

In attachment is implementation of newSVsv_nomg() variant macro which
does not process get magic on input scalar.

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 7, 2019

From @pali

0001-Add-newSVsv_nomg-macro-which-is-like-newSVsv-but-doe.patch
From 1a69aa8329d53d6edfc31b23f829cf6b7135547b Mon Sep 17 00:00:00 2001
From: Pali <pali@cpan.org>
Date: Thu, 7 Feb 2019 14:10:35 +0100
Subject: [PATCH] Add newSVsv_nomg() macro which is like newSVsv() but does not
 process get magic

Both newSVsv() and newSVsv_nomg() are now implemented via new Perl_newSVsv_flags() function.
---
 embed.fnc | 4 +++-
 embed.h   | 2 +-
 mathoms.c | 6 ++++++
 proto.h   | 8 ++++++++
 sv.c      | 9 +++++++--
 sv.h      | 3 +++
 6 files changed, 28 insertions(+), 4 deletions(-)

diff --git a/embed.fnc b/embed.fnc
index ffdc04e5e8..aed1f0f23f 100644
--- a/embed.fnc
+++ b/embed.fnc
@@ -1210,7 +1210,9 @@ ApdR	|SV*	|newSVpv_share	|NULLOK const char* s|U32 hash
 AfpdR	|SV*	|newSVpvf	|NN const char *const pat|...
 ApR	|SV*	|vnewSVpvf	|NN const char *const pat|NULLOK va_list *const args
 Apd	|SV*	|newSVrv	|NN SV *const rv|NULLOK const char *const classname
-ApdR	|SV*	|newSVsv	|NULLOK SV *const old
+ApmbdR	|SV*	|newSVsv	|NULLOK SV *const old
+ApmdR	|SV*	|newSVsv_nomg	|NULLOK SV *const old
+ApR	|SV*	|newSVsv_flags	|NULLOK SV *const old|I32 flags
 ApdR	|SV*	|newSV_type	|const svtype type
 ApdR	|OP*	|newUNOP	|I32 type|I32 flags|NULLOK OP* first
 ApdR	|OP*	|newUNOP_AUX	|I32 type|I32 flags|NULLOK OP* first \
diff --git a/embed.h b/embed.h
index d35508ac44..c863c2ec15 100644
--- a/embed.h
+++ b/embed.h
@@ -546,7 +546,7 @@
 #define newSVpvn_flags(a,b,c)	Perl_newSVpvn_flags(aTHX_ a,b,c)
 #define newSVpvn_share(a,b,c)	Perl_newSVpvn_share(aTHX_ a,b,c)
 #define newSVrv(a,b)		Perl_newSVrv(aTHX_ a,b)
-#define newSVsv(a)		Perl_newSVsv(aTHX_ a)
+#define newSVsv_flags(a,b)	Perl_newSVsv_flags(aTHX_ a,b)
 #define newSVuv(a)		Perl_newSVuv(aTHX_ a)
 #define newUNOP(a,b,c)		Perl_newUNOP(aTHX_ a,b,c)
 #define newUNOP_AUX(a,b,c,d)	Perl_newUNOP_AUX(aTHX_ a,b,c,d)
diff --git a/mathoms.c b/mathoms.c
index 8b003d3538..b8dcb8913d 100644
--- a/mathoms.c
+++ b/mathoms.c
@@ -1755,6 +1755,12 @@ Perl_instr(const char *big, const char *little)
     return instr((char *) big, (char *) little);
 }
 
+SV *
+Perl_newSVsv(pTHX_ SV *const old)
+{
+    return newSVsv(old);
+}
+
 #endif /* NO_MATHOMS */
 
 /*
diff --git a/proto.h b/proto.h
index 2023f5c75d..8720564f14 100644
--- a/proto.h
+++ b/proto.h
@@ -2516,8 +2516,16 @@ PERL_CALLCONV SV*	Perl_newSVpvn_share(pTHX_ const char* s, I32 len, U32 hash)
 PERL_CALLCONV SV*	Perl_newSVrv(pTHX_ SV *const rv, const char *const classname);
 #define PERL_ARGS_ASSERT_NEWSVRV	\
 	assert(rv)
+#ifndef NO_MATHOMS
 PERL_CALLCONV SV*	Perl_newSVsv(pTHX_ SV *const old)
 			__attribute__warn_unused_result__;
+#endif
+
+PERL_CALLCONV SV*	Perl_newSVsv_flags(pTHX_ SV *const old, I32 flags)
+			__attribute__warn_unused_result__;
+
+/* PERL_CALLCONV SV*	Perl_newSVsv_nomg(pTHX_ SV *const old)
+			__attribute__warn_unused_result__; */
 
 PERL_CALLCONV SV*	Perl_newSVuv(pTHX_ const UV u)
 			__attribute__warn_unused_result__;
diff --git a/sv.c b/sv.c
index ad238cdb3b..fe7e0a64fe 100644
--- a/sv.c
+++ b/sv.c
@@ -9748,11 +9748,15 @@ Perl_newRV(pTHX_ SV *const sv)
 Creates a new SV which is an exact duplicate of the original SV.
 (Uses C<sv_setsv>.)
 
+=for apidoc newSVsv_nomg
+
+Like C<newSVsv> but does not process get magic.
+
 =cut
 */
 
 SV *
-Perl_newSVsv(pTHX_ SV *const old)
+Perl_newSVsv_flags(pTHX_ SV *const old, I32 flags)
 {
     SV *sv;
 
@@ -9763,7 +9767,8 @@ Perl_newSVsv(pTHX_ SV *const old)
 	return NULL;
     }
     /* Do this here, otherwise we leak the new SV if this croaks. */
-    SvGETMAGIC(old);
+    if (flags & SV_GMAGIC)
+        SvGETMAGIC(old);
     new_SV(sv);
     /* SV_NOSTEAL prevents TEMP buffers being, well, stolen, and saves games
        with SvTEMP_off and SvTEMP_on round a call to sv_setsv.  */
diff --git a/sv.h b/sv.h
index f3392b08ec..3136fef243 100644
--- a/sv.h
+++ b/sv.h
@@ -2175,6 +2175,9 @@ struct clone_params {
   AV *unreferenced;
 };
 
+#define newSVsv(sv) newSVsv_flags((sv), SV_GMAGIC)
+#define newSVsv_nomg(sv) newSVsv_flags((sv), 0)
+
 /*
 =for apidoc Am|SV*|newSVpvn_utf8|const char* s|STRLEN len|U32 utf8
 
-- 
2.11.0

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 7, 2019

From @pali

And here is patch which makes sv_mortalcopy_flags() function public.

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 7, 2019

From @pali

0001-Make-sv_mortalcopy_flags-public.patch
From 71918e3013329a0120b37905e8046f06ab003eb0 Mon Sep 17 00:00:00 2001
From: Pali <pali@cpan.org>
Date: Thu, 7 Feb 2019 14:22:55 +0100
Subject: [PATCH] Make sv_mortalcopy_flags() public

---
 embed.fnc | 2 +-
 embed.h   | 2 +-
 sv.c      | 5 +++++
 3 files changed, 7 insertions(+), 2 deletions(-)

diff --git a/embed.fnc b/embed.fnc
index aed1f0f23f..9fadb1c7d7 100644
--- a/embed.fnc
+++ b/embed.fnc
@@ -1634,7 +1634,7 @@ Ein	|bool	|sv_only_taint_gmagic|NN SV *sv
 : exported for re.pm
 EXp	|MAGIC *|sv_magicext_mglob|NN SV *sv
 ApdbmR	|SV*	|sv_mortalcopy	|NULLOK SV *const oldsv
-XpR	|SV*	|sv_mortalcopy_flags|NULLOK SV *const oldsv|U32 flags
+ApdR	|SV*	|sv_mortalcopy_flags|NULLOK SV *const oldsv|U32 flags
 ApdR	|SV*	|sv_newmortal
 Apd	|SV*	|sv_newref	|NULLOK SV *const sv
 Ap	|char*	|sv_peek	|NULLOK SV* sv
diff --git a/embed.h b/embed.h
index c863c2ec15..e3eaca3d1a 100644
--- a/embed.h
+++ b/embed.h
@@ -773,6 +773,7 @@
 #define sv_len_utf8(a)		Perl_sv_len_utf8(aTHX_ a)
 #define sv_magic(a,b,c,d,e)	Perl_sv_magic(aTHX_ a,b,c,d,e)
 #define sv_magicext(a,b,c,d,e,f)	Perl_sv_magicext(aTHX_ a,b,c,d,e,f)
+#define sv_mortalcopy_flags(a,b)	Perl_sv_mortalcopy_flags(aTHX_ a,b)
 #define sv_newmortal()		Perl_sv_newmortal(aTHX)
 #define sv_newref(a)		Perl_sv_newref(aTHX_ a)
 #define sv_nosharing(a)		Perl_sv_nosharing(aTHX_ a)
@@ -1581,7 +1582,6 @@
 #define sv_del_backref(a,b)	Perl_sv_del_backref(aTHX_ a,b)
 #define sv_free_arenas()	Perl_sv_free_arenas(aTHX)
 #define sv_len_utf8_nomg(a)	Perl_sv_len_utf8_nomg(aTHX_ a)
-#define sv_mortalcopy_flags(a,b)	Perl_sv_mortalcopy_flags(aTHX_ a,b)
 #define sv_resetpvn(a,b,c)	Perl_sv_resetpvn(aTHX_ a,b,c)
 #define sv_sethek(a,b)		Perl_sv_sethek(aTHX_ a,b)
 #ifndef PERL_IMPLICIT_CONTEXT
diff --git a/sv.c b/sv.c
index fe7e0a64fe..43b906585c 100644
--- a/sv.c
+++ b/sv.c
@@ -9221,6 +9221,11 @@ The new SV is marked as mortal.  It will be destroyed "soon", either by an
 explicit call to C<FREETMPS>, or by an implicit call at places such as
 statement boundaries.  See also C<L</sv_newmortal>> and C<L</sv_2mortal>>.
 
+=for apidoc sv_mortalcopy_flags
+
+Like C<sv_mortalcopy>, but the extra C<flags> are passed to the
+C<sv_setsv_flags>.
+
 =cut
 */
 
-- 
2.11.0

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 13, 2019

From @tonycoz

On Thu, 07 Feb 2019 05​:22​:52 -0800, pali@​cpan.org wrote​:

In attachment is implementation of newSVsv_nomg() variant macro which
does not process get magic on input scalar.

Why not make SV_NOSTEAL significant for newSVsv_flags() ?

Otherwise it will be messy adding it in the future.

Tony

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 13, 2019

The RT System itself - Status changed from 'new' to 'open'

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 13, 2019

From @tonycoz

On Thu, 07 Feb 2019 05​:29​:14 -0800, pali@​cpan.org wrote​:

And here is patch which makes sv_mortalcopy_flags() function public.

+=for apidoc sv_mortalcopy_flags
+
+Like C<sv_mortalcopy>, but the extra C<flags> are passed to the
+C<sv_setsv_flags>.
+

SV_GMAGIC is also meaningful for sv_mortalcopy_flags().

Tony

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 13, 2019

From @pali

On Tuesday 12 February 2019 20​:13​:42 Tony Cook via RT wrote​:

On Thu, 07 Feb 2019 05​:29​:14 -0800, pali@​cpan.org wrote​:

And here is patch which makes sv_mortalcopy_flags() function public.

+=for apidoc sv_mortalcopy_flags
+
+Like C<sv_mortalcopy>, but the extra C<flags> are passed to the
+C<sv_setsv_flags>.
+

SV_GMAGIC is also meaningful for sv_mortalcopy_flags().

So... any suggestion how to improve documentation?

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 14, 2019

From @pali

On Tuesday 12 February 2019 20​:08​:47 Tony Cook via RT wrote​:

On Thu, 07 Feb 2019 05​:22​:52 -0800, pali@​cpan.org wrote​:

In attachment is implementation of newSVsv_nomg() variant macro which
does not process get magic on input scalar.

Why not make SV_NOSTEAL significant for newSVsv_flags() ?

Otherwise it will be messy adding it in the future.

Ok. In attachment is updated V2 patch which supports all flags like
sv_setsv_flags() function. So not only SV_NOSTEAL, but also
SV_COW_SHARED_HASH_KEYS.

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 14, 2019

From @pali

v2-0001-Add-newSVsv_nomg-macro-which-is-like-newSVsv-but-doe.patch
From f8bca093b3ceb8a7e22fbbc7b2b50a30e894c592 Mon Sep 17 00:00:00 2001
From: Pali <pali@cpan.org>
Date: Thu, 7 Feb 2019 14:10:35 +0100
Subject: [PATCH v2] Add newSVsv_nomg() macro which is like newSVsv() but does
 not process get magic

Both newSVsv() and newSVsv_nomg() are now implemented via new Perl_newSVsv_flags() function.
---
 embed.fnc |  4 +++-
 embed.h   |  2 +-
 mathoms.c |  6 ++++++
 proto.h   |  8 ++++++++
 sv.c      | 13 ++++++++-----
 sv.h      |  5 +++++
 6 files changed, 31 insertions(+), 7 deletions(-)

diff --git a/embed.fnc b/embed.fnc
index 9d4a8461f5..80d90b985f 100644
--- a/embed.fnc
+++ b/embed.fnc
@@ -1209,7 +1209,9 @@ ApdR	|SV*	|newSVpv_share	|NULLOK const char* s|U32 hash
 AfpdR	|SV*	|newSVpvf	|NN const char *const pat|...
 ApR	|SV*	|vnewSVpvf	|NN const char *const pat|NULLOK va_list *const args
 Apd	|SV*	|newSVrv	|NN SV *const rv|NULLOK const char *const classname
-ApdR	|SV*	|newSVsv	|NULLOK SV *const old
+ApmbdR	|SV*	|newSVsv	|NULLOK SV *const old
+ApmdR	|SV*	|newSVsv_nomg	|NULLOK SV *const old
+ApR	|SV*	|newSVsv_flags	|NULLOK SV *const old|I32 flags
 ApdR	|SV*	|newSV_type	|const svtype type
 ApdR	|OP*	|newUNOP	|I32 type|I32 flags|NULLOK OP* first
 ApdR	|OP*	|newUNOP_AUX	|I32 type|I32 flags|NULLOK OP* first \
diff --git a/embed.h b/embed.h
index 4df6fa0b0f..cd58193fb5 100644
--- a/embed.h
+++ b/embed.h
@@ -546,7 +546,7 @@
 #define newSVpvn_flags(a,b,c)	Perl_newSVpvn_flags(aTHX_ a,b,c)
 #define newSVpvn_share(a,b,c)	Perl_newSVpvn_share(aTHX_ a,b,c)
 #define newSVrv(a,b)		Perl_newSVrv(aTHX_ a,b)
-#define newSVsv(a)		Perl_newSVsv(aTHX_ a)
+#define newSVsv_flags(a,b)	Perl_newSVsv_flags(aTHX_ a,b)
 #define newSVuv(a)		Perl_newSVuv(aTHX_ a)
 #define newUNOP(a,b,c)		Perl_newUNOP(aTHX_ a,b,c)
 #define newUNOP_AUX(a,b,c,d)	Perl_newUNOP_AUX(aTHX_ a,b,c,d)
diff --git a/mathoms.c b/mathoms.c
index 8b003d3538..b8dcb8913d 100644
--- a/mathoms.c
+++ b/mathoms.c
@@ -1755,6 +1755,12 @@ Perl_instr(const char *big, const char *little)
     return instr((char *) big, (char *) little);
 }
 
+SV *
+Perl_newSVsv(pTHX_ SV *const old)
+{
+    return newSVsv(old);
+}
+
 #endif /* NO_MATHOMS */
 
 /*
diff --git a/proto.h b/proto.h
index adf1ef5d40..5c5a0c7d66 100644
--- a/proto.h
+++ b/proto.h
@@ -2516,8 +2516,16 @@ PERL_CALLCONV SV*	Perl_newSVpvn_share(pTHX_ const char* s, I32 len, U32 hash)
 PERL_CALLCONV SV*	Perl_newSVrv(pTHX_ SV *const rv, const char *const classname);
 #define PERL_ARGS_ASSERT_NEWSVRV	\
 	assert(rv)
+#ifndef NO_MATHOMS
 PERL_CALLCONV SV*	Perl_newSVsv(pTHX_ SV *const old)
 			__attribute__warn_unused_result__;
+#endif
+
+PERL_CALLCONV SV*	Perl_newSVsv_flags(pTHX_ SV *const old, I32 flags)
+			__attribute__warn_unused_result__;
+
+/* PERL_CALLCONV SV*	Perl_newSVsv_nomg(pTHX_ SV *const old)
+			__attribute__warn_unused_result__; */
 
 PERL_CALLCONV SV*	Perl_newSVuv(pTHX_ const UV u)
 			__attribute__warn_unused_result__;
diff --git a/sv.c b/sv.c
index 0a4a2e531a..5107202d31 100644
--- a/sv.c
+++ b/sv.c
@@ -9750,11 +9750,15 @@ Perl_newRV(pTHX_ SV *const sv)
 Creates a new SV which is an exact duplicate of the original SV.
 (Uses C<sv_setsv>.)
 
+=for apidoc newSVsv_nomg
+
+Like C<newSVsv> but does not process get magic.
+
 =cut
 */
 
 SV *
-Perl_newSVsv(pTHX_ SV *const old)
+Perl_newSVsv_flags(pTHX_ SV *const old, I32 flags)
 {
     SV *sv;
 
@@ -9765,11 +9769,10 @@ Perl_newSVsv(pTHX_ SV *const old)
 	return NULL;
     }
     /* Do this here, otherwise we leak the new SV if this croaks. */
-    SvGETMAGIC(old);
+    if (flags & SV_GMAGIC)
+        SvGETMAGIC(old);
     new_SV(sv);
-    /* SV_NOSTEAL prevents TEMP buffers being, well, stolen, and saves games
-       with SvTEMP_off and SvTEMP_on round a call to sv_setsv.  */
-    sv_setsv_flags(sv, old, SV_NOSTEAL);
+    sv_setsv_flags(sv, old, flags & ~SV_GMAGIC);
     return sv;
 }
 
diff --git a/sv.h b/sv.h
index f3392b08ec..3a648e4971 100644
--- a/sv.h
+++ b/sv.h
@@ -2175,6 +2175,11 @@ struct clone_params {
   AV *unreferenced;
 };
 
+/* SV_NOSTEAL prevents TEMP buffers being, well, stolen, and saves games
+   with SvTEMP_off and SvTEMP_on round a call to sv_setsv.  */
+#define newSVsv(sv) newSVsv_flags((sv), SV_GMAGIC|SV_NOSTEAL)
+#define newSVsv_nomg(sv) newSVsv_flags((sv), SV_NOSTEAL)
+
 /*
 =for apidoc Am|SV*|newSVpvn_utf8|const char* s|STRLEN len|U32 utf8
 
-- 
2.11.0

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 24, 2019

From @tonycoz

On Thu, 14 Feb 2019 03​:46​:17 -0800, pali@​cpan.org wrote​:

On Tuesday 12 February 2019 20​:08​:47 Tony Cook via RT wrote​:

On Thu, 07 Feb 2019 05​:22​:52 -0800, pali@​cpan.org wrote​:

In attachment is implementation of newSVsv_nomg() variant macro which
does not process get magic on input scalar.

Why not make SV_NOSTEAL significant for newSVsv_flags() ?

Otherwise it will be messy adding it in the future.

Ok. In attachment is updated V2 patch which supports all flags like
sv_setsv_flags() function. So not only SV_NOSTEAL, but also
SV_COW_SHARED_HASH_KEYS.

Thanks, applied as 238f2c1.

Tony

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 24, 2019

@tonycoz - Status changed from 'open' to 'pending release'

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Feb 26, 2019

From @pali

On Wednesday 13 February 2019 20​:16​:03 pali@​cpan.org wrote​:

On Tuesday 12 February 2019 20​:13​:42 Tony Cook via RT wrote​:

On Thu, 07 Feb 2019 05​:29​:14 -0800, pali@​cpan.org wrote​:

And here is patch which makes sv_mortalcopy_flags() function public.

+=for apidoc sv_mortalcopy_flags
+
+Like C<sv_mortalcopy>, but the extra C<flags> are passed to the
+C<sv_setsv_flags>.
+

SV_GMAGIC is also meaningful for sv_mortalcopy_flags().

So... any suggestion how to improve documentation?

Tony, ping. What do you prefer or how do you want to improve this change?

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Mar 28, 2019

From @pali

On Tuesday 26 February 2019 10​:00​:58 pali@​cpan.org wrote​:

On Wednesday 13 February 2019 20​:16​:03 pali@​cpan.org wrote​:

On Tuesday 12 February 2019 20​:13​:42 Tony Cook via RT wrote​:

On Thu, 07 Feb 2019 05​:29​:14 -0800, pali@​cpan.org wrote​:

And here is patch which makes sv_mortalcopy_flags() function public.

+=for apidoc sv_mortalcopy_flags
+
+Like C<sv_mortalcopy>, but the extra C<flags> are passed to the
+C<sv_setsv_flags>.
+

SV_GMAGIC is also meaningful for sv_mortalcopy_flags().

So... any suggestion how to improve documentation?

Tony, ping. What do you prefer or how do you want to improve this change?

ping

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Mar 28, 2019

@tonycoz - Status changed from 'pending release' to 'open'

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Mar 29, 2019

From @tonycoz

On Thu, 28 Mar 2019 05​:58​:25 -0700, pali@​cpan.org wrote​:

On Tuesday 26 February 2019 10​:00​:58 pali@​cpan.org wrote​:

On Wednesday 13 February 2019 20​:16​:03 pali@​cpan.org wrote​:

On Tuesday 12 February 2019 20​:13​:42 Tony Cook via RT wrote​:

On Thu, 07 Feb 2019 05​:29​:14 -0800, pali@​cpan.org wrote​:

And here is patch which makes sv_mortalcopy_flags() function public.

+=for apidoc sv_mortalcopy_flags
+
+Like C<sv_mortalcopy>, but the extra C<flags> are passed to the
+C<sv_setsv_flags>.
+

SV_GMAGIC is also meaningful for sv_mortalcopy_flags().

So... any suggestion how to improve documentation?

Tony, ping. What do you prefer or how do you want to improve this change?

ping

Sorry, I thought this was all resolved (and I was sort of wrong.)

Since sv_mortalcopy_flags()'s treatment of S_GMAGIC is equivalent to sv_setsv_flags() treatment, I no longer think there's any extra documentation needed.

I mis-read the code, sorry for my misunderstanding and the delay in the response.

Tony

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Mar 29, 2019

@tonycoz - Status changed from 'open' to 'pending release'

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Apr 2, 2019

From @pali

On Thursday 28 March 2019 17​:01​:47 Tony Cook via RT wrote​:

On Thu, 28 Mar 2019 05​:58​:25 -0700, pali@​cpan.org wrote​:

On Tuesday 26 February 2019 10​:00​:58 pali@​cpan.org wrote​:

On Wednesday 13 February 2019 20​:16​:03 pali@​cpan.org wrote​:

On Tuesday 12 February 2019 20​:13​:42 Tony Cook via RT wrote​:

On Thu, 07 Feb 2019 05​:29​:14 -0800, pali@​cpan.org wrote​:

And here is patch which makes sv_mortalcopy_flags() function public.

+=for apidoc sv_mortalcopy_flags
+
+Like C<sv_mortalcopy>, but the extra C<flags> are passed to the
+C<sv_setsv_flags>.
+

SV_GMAGIC is also meaningful for sv_mortalcopy_flags().

So... any suggestion how to improve documentation?

Tony, ping. What do you prefer or how do you want to improve this change?

ping

Sorry, I thought this was all resolved (and I was sort of wrong.)

Since sv_mortalcopy_flags()'s treatment of S_GMAGIC is equivalent to sv_setsv_flags() treatment, I no longer think there's any extra documentation needed.

I mis-read the code, sorry for my misunderstanding and the delay in the response.

Ok, so are there any other changes needed for this patch?

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented Apr 23, 2019

From @pali

On Tuesday 02 April 2019 11​:51​:28 pali@​cpan.org wrote​:

On Thursday 28 March 2019 17​:01​:47 Tony Cook via RT wrote​:

On Thu, 28 Mar 2019 05​:58​:25 -0700, pali@​cpan.org wrote​:

On Tuesday 26 February 2019 10​:00​:58 pali@​cpan.org wrote​:

On Wednesday 13 February 2019 20​:16​:03 pali@​cpan.org wrote​:

On Tuesday 12 February 2019 20​:13​:42 Tony Cook via RT wrote​:

On Thu, 07 Feb 2019 05​:29​:14 -0800, pali@​cpan.org wrote​:

And here is patch which makes sv_mortalcopy_flags() function public.

+=for apidoc sv_mortalcopy_flags
+
+Like C<sv_mortalcopy>, but the extra C<flags> are passed to the
+C<sv_setsv_flags>.
+

SV_GMAGIC is also meaningful for sv_mortalcopy_flags().

So... any suggestion how to improve documentation?

Tony, ping. What do you prefer or how do you want to improve this change?

ping

Sorry, I thought this was all resolved (and I was sort of wrong.)

Since sv_mortalcopy_flags()'s treatment of S_GMAGIC is equivalent to sv_setsv_flags() treatment, I no longer think there's any extra documentation needed.

I mis-read the code, sorry for my misunderstanding and the delay in the response.

Ok, so are there any other changes needed for this patch?

So if not, can you apply that patch?

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented May 14, 2019

From @tonycoz

On Tue, 23 Apr 2019 06​:27​:36 -0700, pali@​cpan.org wrote​:

On Tuesday 02 April 2019 11​:51​:28 pali@​cpan.org wrote​:

On Thursday 28 March 2019 17​:01​:47 Tony Cook via RT wrote​:

On Thu, 28 Mar 2019 05​:58​:25 -0700, pali@​cpan.org wrote​:

On Tuesday 26 February 2019 10​:00​:58 pali@​cpan.org wrote​:

On Wednesday 13 February 2019 20​:16​:03 pali@​cpan.org wrote​:

On Tuesday 12 February 2019 20​:13​:42 Tony Cook via RT wrote​:

On Thu, 07 Feb 2019 05​:29​:14 -0800, pali@​cpan.org wrote​:

And here is patch which makes sv_mortalcopy_flags()
function public.

+=for apidoc sv_mortalcopy_flags
+
+Like C<sv_mortalcopy>, but the extra C<flags> are passed
to the
+C<sv_setsv_flags>.
+

SV_GMAGIC is also meaningful for sv_mortalcopy_flags().

So... any suggestion how to improve documentation?

Tony, ping. What do you prefer or how do you want to improve
this change?

ping

Sorry, I thought this was all resolved (and I was sort of wrong.)

Since sv_mortalcopy_flags()'s treatment of S_GMAGIC is equivalent
to sv_setsv_flags() treatment, I no longer think there's any extra
documentation needed.

I mis-read the code, sorry for my misunderstanding and the delay in
the response.

Ok, so are there any other changes needed for this patch?

So if not, can you apply that patch?

https://rt-archive.perl.org/perl5/Ticket/Display.html?id=132964#txn-1614862

It was applied.

Tony

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented May 15, 2019

From @tonycoz

On Mon, 13 May 2019 17​:09​:09 -0700, tonyc wrote​:

On Tue, 23 Apr 2019 06​:27​:36 -0700, pali@​cpan.org wrote​:

On Tuesday 02 April 2019 11​:51​:28 pali@​cpan.org wrote​:

On Thursday 28 March 2019 17​:01​:47 Tony Cook via RT wrote​:

On Thu, 28 Mar 2019 05​:58​:25 -0700, pali@​cpan.org wrote​:

On Tuesday 26 February 2019 10​:00​:58 pali@​cpan.org wrote​:

On Wednesday 13 February 2019 20​:16​:03 pali@​cpan.org wrote​:

On Tuesday 12 February 2019 20​:13​:42 Tony Cook via RT wrote​:

On Thu, 07 Feb 2019 05​:29​:14 -0800, pali@​cpan.org wrote​:

And here is patch which makes sv_mortalcopy_flags()
function public.

+=for apidoc sv_mortalcopy_flags
+
+Like C<sv_mortalcopy>, but the extra C<flags> are passed
to the
+C<sv_setsv_flags>.
+

SV_GMAGIC is also meaningful for sv_mortalcopy_flags().

So... any suggestion how to improve documentation?

Tony, ping. What do you prefer or how do you want to improve
this change?

ping

Sorry, I thought this was all resolved (and I was sort of wrong.)

Since sv_mortalcopy_flags()'s treatment of S_GMAGIC is equivalent
to sv_setsv_flags() treatment, I no longer think there's any extra
documentation needed.

I mis-read the code, sorry for my misunderstanding and the delay in
the response.

Ok, so are there any other changes needed for this patch?

So if not, can you apply that patch?

https://rt-archive.perl.org/perl5/Ticket/Display.html?id=132964#txn-1614862

Your reply didn't make it to the ticket.

You're right, I missed a patch, I'll apply it post 5.31.

Tony

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented May 15, 2019

@tonycoz - Status changed from 'pending release' to 'open'

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented May 29, 2019

From @tonycoz

On Tue, 14 May 2019 17​:02​:17 -0700, tonyc wrote​:

On Mon, 13 May 2019 17​:09​:09 -0700, tonyc wrote​:

On Tue, 23 Apr 2019 06​:27​:36 -0700, pali@​cpan.org wrote​:

On Tuesday 02 April 2019 11​:51​:28 pali@​cpan.org wrote​:

On Thursday 28 March 2019 17​:01​:47 Tony Cook via RT wrote​:

On Thu, 28 Mar 2019 05​:58​:25 -0700, pali@​cpan.org wrote​:

On Tuesday 26 February 2019 10​:00​:58 pali@​cpan.org wrote​:

On Wednesday 13 February 2019 20​:16​:03 pali@​cpan.org wrote​:

On Tuesday 12 February 2019 20​:13​:42 Tony Cook via RT wrote​:

On Thu, 07 Feb 2019 05​:29​:14 -0800, pali@​cpan.org wrote​:

And here is patch which makes sv_mortalcopy_flags()
function public.

+=for apidoc sv_mortalcopy_flags
+
+Like C<sv_mortalcopy>, but the extra C<flags> are passed
to the
+C<sv_setsv_flags>.
+

SV_GMAGIC is also meaningful for sv_mortalcopy_flags().

So... any suggestion how to improve documentation?

Tony, ping. What do you prefer or how do you want to improve
this change?

ping

Sorry, I thought this was all resolved (and I was sort of wrong.)

Since sv_mortalcopy_flags()'s treatment of S_GMAGIC is equivalent
to sv_setsv_flags() treatment, I no longer think there's any extra
documentation needed.

I mis-read the code, sorry for my misunderstanding and the delay in
the response.

Ok, so are there any other changes needed for this patch?

So if not, can you apply that patch?

https://rt-archive.perl.org/perl5/Ticket/Display.html?id=132964#txn-1614862

Your reply didn't make it to the ticket.

You're right, I missed a patch, I'll apply it post 5.31.

Thanks, applied as c6dcb9e.

Sorry for the mix-up.

Tony

@p5pRT

This comment has been minimized.

Copy link
Collaborator Author

commented May 29, 2019

@tonycoz - Status changed from 'open' to 'pending release'

@p5pRT p5pRT closed this May 29, 2019
@p5pRT p5pRT added the Severity Low label Oct 19, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
1 participant
You can’t perform that action at this time.