This is a live mirror of the Perl 5 development currently hosted at
23 months agoUpdate epigraph for v5.31.7
Nicolas R [Fri, 20 Dec 2019 22:09:31 +0000 (15:09 -0700)]
Update epigraph for v5.31.7

23 months agoFinalise perldelta and update perlhist v5.31.7
Nicolas R [Fri, 20 Dec 2019 18:59:42 +0000 (11:59 -0700)]
Finalise perldelta and update perlhist

23 months agoAdjust release_managers_guide for GH tracker
Nicolas R [Fri, 20 Dec 2019 22:03:36 +0000 (15:03 -0700)]
Adjust release_managers_guide for GH tracker

23 months agoUpdate Module::CoreList for 5.31.7
Nicolas R [Fri, 20 Dec 2019 21:39:37 +0000 (14:39 -0700)]
Update Module::CoreList for 5.31.7

23 months agoMention CI smokers in release_managers_guide
Nicolas R [Fri, 20 Dec 2019 18:47:20 +0000 (11:47 -0700)]
Mention CI smokers in release_managers_guide

List and link to the Continuous Integrations currently setup.

23 months agoAdd a basic linux smoke GitHub Action
Nicolas R [Fri, 20 Dec 2019 17:55:53 +0000 (10:55 -0700)]
Add a basic linux smoke GitHub Action

In addition to Travis CI, it will not hurt
to have some additional linux smokers running.

We can see overtime if we want to keep one or the

♪ Let it Smoke, Let it Smoke ♫♫♫♬♭

Note: GitHub allow a total of 20 concurrent jobs.

23 months agoBump Maintainers to point to CPAN 2.27
Nicolas R [Fri, 20 Dec 2019 17:46:59 +0000 (10:46 -0700)]
Bump Maintainers to point to CPAN 2.27

Avoid listing CPAN in modules discrepancies list
as the trial version we were using, after exclusions
is the same as the most recent 2.27 one.

23 months agoUpdate Net::Ping to upstream version 2.72
Nicolas R [Fri, 20 Dec 2019 17:14:19 +0000 (10:14 -0700)]
Update Net::Ping to upstream version 2.72

This retains blead customizations:
1a58b39af8 remove of 'use vars'
7bfdd8260c 500_ping_icmp.t: remove sudo code

This is mainly a unit test changes / noop changes.

Customized files are:

Files not ported to blead:

Do not list Net::Ping customized files as mentioned by 48a109d29

Note: we could have considered doing nothing for this update,
but this would avoid updating to the CPAN version, and preserved
the existing patches.

23 months agofixing usally and patern
Bryan Stenson [Thu, 19 Dec 2019 21:15:33 +0000 (13:15 -0800)]
fixing usally and patern

Typos in documentation and inline comments.

Pursuant to discussion on openbsd-tech list:


Bryan Stenson is now a Perl author.
Regenerate via: perl regen/

23 months agot/ Avoid some work on ASCII platforms
Karl Williamson [Thu, 19 Dec 2019 18:28:05 +0000 (11:28 -0700)]
t/ Avoid some work on ASCII platforms

This sets things up so that functions on ASCII simply return their
arguments, and don't have extra stuff in them for EBCDIC.  It also moves
an array initialization out of a function, so it is only done once, and
doesn't initialize it on ASCII, as it isn't needed there.

23 months agot/re/regexp.t: Speed up many regex tests on ASCII platform
Karl Williamson [Thu, 19 Dec 2019 16:59:54 +0000 (09:59 -0700)]
t/re/regexp.t: Speed up many regex tests on ASCII platform

This commit:

 commit 0cd59ee9ca0f0af3c0c172ecc27bb3f02da6db08
 Author:     Karl Williamson <>
 AuthorDate: Fri Sep 6 10:23:26 2019 -0600
 Commit:     Karl Williamson <>
 CommitDate: Mon Nov 11 21:05:13 2019 -0700

     t/re/regexp.t: Only convert to EBCDIC once

     Some tests get added as we go along, and those added tests have already
     been converted to EBCDIC if necessary.  Don't reconvert, which messes
     things up.

caused a huge slowdown in regex tests.  The most noticeable on my
platform was regexp_qr_embed_thr.t which doubled in wall clock time

It turns out that it was because a function was now always being called,
and that does nothing on ASCII platforms besides return its argument,
which then was copied over the argument.

This new commit causes the function to be a constant { 1; } on ASCII
platforms, so should be completely optimized out, returning the time
spent in that .t to 5.30 levels.

23 months agoregen/ Move inversion list adjacent to similar
Karl Williamson [Thu, 19 Dec 2019 03:25:34 +0000 (20:25 -0700)]
regen/ Move inversion list adjacent to similar

This will lessen any paging that might occur.  Further, on most builds,
it and another table are identical, so only one is actually needed.

23 months agoperlfunc: link to "Plain Old Comments" section of perlsyn from __FILE__, __LINE__...
Dan Book [Mon, 18 Mar 2019 22:29:26 +0000 (18:29 -0400)]
perlfunc: link to "Plain Old Comments" section of perlsyn from __FILE__, __LINE__, and caller

23 months agoMention that you cannot return from do BLOCK
Dan Book [Wed, 18 Dec 2019 21:39:38 +0000 (16:39 -0500)]
Mention that you cannot return from do BLOCK

23 months agoPATCH: GH #17371
Karl Williamson [Wed, 18 Dec 2019 18:46:40 +0000 (11:46 -0700)]
PATCH: GH #17371

This was caused by a character being counted as both the first delimiter
of a pattern, and the final one, which led to the pattern's length being
negative, which was turned into a very large unsigned number.

23 months agoAdd memCHRs() macro and use it
Karl Williamson [Sat, 7 Dec 2019 20:47:05 +0000 (13:47 -0700)]
Add memCHRs() macro and use it

This replaces strchr("list", c) calls throughout the core.  They don't
work properly when 'c' is a NUL, returning the position of the
terminating NUL in "list" instead of failure.  This could lead to
segfaults or even security issues.

23 months agoUpdate ExtUtils-MakeMaker to CPAN version 7.42
Chris 'BinGOs' Williams [Wed, 18 Dec 2019 09:16:30 +0000 (09:16 +0000)]
Update ExtUtils-MakeMaker to CPAN version 7.42


7.42    Tue 17 Dec 22:02:25 GMT 2019

    No changes since v7.41_01

7.41_01 Mon 16 Dec 21:36:24 GMT 2019

    Test fixes:
    - README.pod warnings are suppressed during testing
      so no need to skip when PERL_CORE is defined

    Doc fixes:
    - Update referenced modules for new distributions
    - Updated SEE ALSO section accordingly

23 months agoPATCH: GH #17363 Reserve two sequences
Karl Williamson [Wed, 18 Dec 2019 00:59:16 +0000 (17:59 -0700)]
PATCH: GH #17363 Reserve two sequences

We can't guarantee what people might do in the future with these two
currently unused illegal pattern sequences.  What I did was put a
comment and code where things would have to change to use them to refer
to this ticket.

23 months agoPATCH GH #17276 New compilation diagnostic
Karl Williamson [Tue, 17 Dec 2019 22:21:26 +0000 (15:21 -0700)]
PATCH GH #17276 New compilation diagnostic

The previous patch in commit 830b3eb245d5dbcf095fbd4b5d59764c697c20df
didn't work generally.  This uses #ifdef instead.

23 months agoActually fix GH #17370
Karl Williamson [Tue, 17 Dec 2019 21:59:29 +0000 (14:59 -0700)]
Actually fix GH #17370

I only added a test, but not the change in
9f16475a53933d1d9c547f871b97b3b12160cece.  The test passes except when
run under address sanitizer or valgrind.

23 months agodon't use -Werror=declaration-after-statement with C++
Tony Cook [Tue, 17 Dec 2019 00:32:40 +0000 (11:32 +1100)]
don't use -Werror=declaration-after-statement with C++

declaration after statement is normal for C++ and C++ compilers
rightly complain if we try to warn (or error) on them, so don't
try to.

fixes #17353

23 months agoPATCH: GH #17370, read beyond buffer in grok_inf_nan
Karl Williamson [Tue, 17 Dec 2019 21:23:07 +0000 (14:23 -0700)]
PATCH: GH #17370, read beyond buffer in grok_inf_nan

Like GH #17367, this was caused by a failure to check that we aren't at
the end of the buffer after advancing the ptr to it.

23 months agoNote that certain flags are documented
Karl Williamson [Tue, 17 Dec 2019 00:02:53 +0000 (17:02 -0700)]
Note that certain flags are documented

This is useful in Devel::PPPort for generating its api-info data.  That
useful feature of D:P allows someone to find out what was the first
release of Perl to have a function, macro, or flag.  And whether using
ppport.h backports it further.

I went through apidoc.pod and looked for flags that were documented but
that D:P didn't know about.  This commit adds entries for each so that
D:P can find them.

23 months agoPATCH: GH #17367 read 1 beyond end of buffer
Karl Williamson [Wed, 13 Nov 2019 22:39:15 +0000 (15:39 -0700)]
PATCH: GH #17367 read 1 beyond end of buffer

This is a bug in grok_infnan() in which in one place it failed to check
that it was reading within bounds.

23 months agoLocal variable 'o' hides a parameter of the same name.
James E Keenan [Sun, 15 Dec 2019 22:38:35 +0000 (17:38 -0500)]
Local variable 'o' hides a parameter of the same name.

This sub-optimal code has been reported by LGTM static code analysis.
There are actually two such instances very close to one another within
op.c.  This commit handles only the first of them, renaming a variable
and regularizing the indents to make the relevant scope more self-evident.

23 months agosv.h: Fix typo in pod
Karl Williamson [Tue, 17 Dec 2019 00:06:29 +0000 (17:06 -0700)]
sv.h: Fix typo in pod

23 months agoUpdate ExtUtils-MakeMaker to CPAN version 7.40
Chris 'BinGOs' Williams [Mon, 16 Dec 2019 20:36:21 +0000 (20:36 +0000)]
Update ExtUtils-MakeMaker to CPAN version 7.40


7.40    Mon 16 Dec 19:33:13 GMT 2019

    No changes since v7.39_05

7.39_05 Thu 21 Nov 11:45:13 GMT 2019

    Bug fixes:
    - Always assume that libraries are shared on AIX

7.39_04 Mon 18 Nov 14:54:46 GMT 2019

    Test Fixes:
    - Skip "merged /usr" tests on Cygwin

    QA Fixes:
    - Added cygwin testing with Github actions
    - Added macos testing with Github actions

7.39_03 Sun 17 Nov 19:53:04 GMT 2019

    Doc fixes:
    - Fix typo in link to MakeMaker
    - Link to referenced modules from MM_Cygwin maybe_command

7.39_02 Thu  7 Nov 09:33:29 GMT 2019

    Cygwin fixes:
    - Removed MM_Cygwin all_target() override

    Doc fixes:
    - Add crosslinks to various referenced documentation

7.39_01 Mon 16 Sep 07:19:37 BST 2019

    Test fixes:
    - README.pod warnings suppressed during testing
    - Don't parallise dynamic/static tests

23 months agoTick off 5.31.6, released 2019-11-20
Matthew Horsfall [Mon, 16 Dec 2019 13:56:41 +0000 (08:56 -0500)]
Tick off 5.31.6, released 2019-11-20

23 months agoharden IO's cachepropagate-tcp
Tony Cook [Thu, 12 Dec 2019 00:30:59 +0000 (11:30 +1100)]
harden IO's cachepropagate-tcp

This failed on Win32 like in the #17351 CI checks with:

../dist/if/t/if.t .................................................. ok
Can't call method "sockdomain" on an undefined value at t/cachepropagate-tcp.t line 46.
# Looks like your test exited with 9 just after 5.
../dist/IO/t/cachepropagate-tcp.t ..................................
Dubious, test returned 9 (wstat 2304, 0x900)
Failed 3/8 subtests

I suspect what happened is there was a race between the parent
accepting the connection and the child exiting and closing the

The Microsoft documentation for accept() indicates one possible
reason for failure is:

An incoming connection was indicated, but was subsequently
        terminated by the remote peer prior to accepting the call.

which I suspect happened here.

So I've:

- added a basic error check for the result of accept()
- made the child to wait for the parent to close the socket
- the parent explicitly closes the socket

23 months agotreat TAINT_get and TAINTING_get as unlikely
Tony Cook [Thu, 12 Dec 2019 05:49:11 +0000 (16:49 +1100)]
treat TAINT_get and TAINTING_get as unlikely

While testing #17359 it appeared that inlining of SvTRUE was being
suppressed (indicated by -Winline) by being used in the statement:

  if (TAINT_get || SvTRUE(error)) {

but making TAINT_get unlikely allowed it to be inlined.

I expect even in a program that does use taint the vast majority
of data will be untainted, so I think it's safe to make TAINT_get

TAINTING_get is a harder case, but it's only used in a relatively
much smaller number of cases, and I expect most runs of a system
perl will have neither -T nor -t.

23 months agoRevert "Move PL_check to the interp vars to fix threading issues"
Tony Cook [Sun, 15 Dec 2019 21:56:52 +0000 (08:56 +1100)]
Revert "Move PL_check to the interp vars to fix threading issues"

and the associated commits, at least until a way to make
wrap_op_checker() work is available.

23 months agoUpdate release schedule
Sawyer X [Sun, 15 Dec 2019 20:39:01 +0000 (22:39 +0200)]
Update release schedule

23 months agot/re/pat.t: White-space only
Karl Williamson [Thu, 5 Dec 2019 20:16:22 +0000 (13:16 -0700)]
t/re/pat.t: White-space only

23 months agot/re/pat.t: Fix skip count for limited mem platforms
Karl Williamson [Fri, 6 Dec 2019 00:03:43 +0000 (17:03 -0700)]
t/re/pat.t: Fix skip count for limited mem platforms

And rearrange so is easier to see the correct value.

23 months agot/re/pat.t: Skip tests that don't work on EBCDIC
Karl Williamson [Wed, 4 Dec 2019 20:45:24 +0000 (13:45 -0700)]
t/re/pat.t: Skip tests that don't work on EBCDIC

These are fuzzer generated, and don't translate well to EBCDIC

23 months agoperlreref: Fix typos
Karl Williamson [Sat, 14 Dec 2019 04:36:49 +0000 (21:36 -0700)]
perlreref: Fix typos

Pointed out in GH #17363

23 months agoavoid identical stack traces
David Mitchell [Fri, 13 Dec 2019 13:48:25 +0000 (13:48 +0000)]
avoid identical stack traces

GH #15109

The output of caller() (e.g. as produced by carp::Confess) produces
multiple identical outputs when within a nested use/require. This is
because at the time of calling the 'BEGIN { require ... }', PL_curcop is
set to &PL_compiling, which is a fixed buffer within the interpreter,
whose individual file and line fields are saved and restored when doing
a new require/eval. This means that within the innermost require,
PL_compiling has file:lineno of the innermost source file, and multiple
saved PL_curcop values in the context stack frames all point to the same
&PL_copmpiling.  So all levels of the stack trace appear to come from the
innermost file.

This commit fixes this (after a fashion) by, at the start of calling a
BEGIN, making PL_curcop point to a temporary copy of PL_compiling

This is all a bit of a hack.

23 months agoEnable Cygwin smoking on all kinds of branches
Max Maischein [Fri, 13 Dec 2019 18:05:47 +0000 (19:05 +0100)]
Enable Cygwin smoking on all kinds of branches

23 months agoRemove bad "git switch" command
Max Maischein [Fri, 13 Dec 2019 18:05:35 +0000 (19:05 +0100)]
Remove bad "git switch" command

"git switch" isn't supported by Github git anyway

23 months agoIO: update the ChangeLog after a (trial) release of IO to CPAN
Ricardo Signes [Thu, 12 Dec 2019 18:23:34 +0000 (13:23 -0500)]
IO: update the ChangeLog after a (trial) release of IO to CPAN

23 months agoperldelta for a559786348e0
Chris 'BinGOs' Williams [Thu, 12 Dec 2019 01:03:52 +0000 (01:03 +0000)]
perldelta for a559786348e0

23 months agoperldelta for e6c7056ba6f1
Tony Cook [Thu, 12 Dec 2019 00:40:14 +0000 (11:40 +1100)]
perldelta for e6c7056ba6f1

23 months agoun-TODO pl_check.t and remove io/handle.t
Tony Cook [Mon, 9 Dec 2019 00:22:25 +0000 (11:22 +1100)]
un-TODO pl_check.t and remove io/handle.t

io/handle.t depends on IO::Handle using the PL_check hack, but
Nicholas's back portable fix no longer uses that

Fix threaded perl detection, thanks to James Keenan.

23 months agoMove PL_check to the interp vars to fix threading issues
Stefan Seifert [Wed, 30 Oct 2019 15:23:01 +0000 (16:23 +0100)]
Move PL_check to the interp vars to fix threading issues

Fixes issue #14816

23 months agoprovide a test for the PL_check mis-behaviour
Tony Cook [Sun, 8 Dec 2019 23:51:19 +0000 (10:51 +1100)]
provide a test for the PL_check mis-behaviour

Nicholas Clark's fix for IO makes the test in niner's patch
meaningless, so test it separately.

23 months agoNote that G_RETHROW is documented
Karl Williamson [Wed, 11 Dec 2019 18:37:58 +0000 (11:37 -0700)]
Note that G_RETHROW is documented

This is for Devel::PPPort.

23 months agoFix where we look for the Sun/Solaris/Developer Studio compiler
Chris 'BinGOs' Williams [Wed, 11 Dec 2019 18:11:15 +0000 (18:11 +0000)]
Fix where we look for the Sun/Solaris/Developer Studio compiler

The movable feast that is the sun/solaris/developer compiler

23 months agoperlpacktut: fix broken url
Ask Bjørn Hansen [Wed, 11 Dec 2019 09:12:54 +0000 (01:12 -0800)]
perlpacktut: fix broken url

This is my best guess as for what it's meant to point to.

Reported in

23 months agoregexec.c: Clarify comment
Karl Williamson [Tue, 10 Dec 2019 03:28:24 +0000 (20:28 -0700)]
regexec.c: Clarify comment

23 months agoRmv leading underscore from macro name
Karl Williamson [Tue, 10 Dec 2019 03:27:02 +0000 (20:27 -0700)]
Rmv leading underscore from macro name

These are illegal in C, but we have plenty of them around; I happened
to be looking at this function, and decided to fix it.  Note that only
the macro name is illegal; the function was fine, but to change the
macro name means changing the function one.

23 months agoOnly allow punct delimiter for regex subpattern
Karl Williamson [Tue, 10 Dec 2019 02:37:28 +0000 (19:37 -0700)]
Only allow punct delimiter for regex subpattern

The experimental feature that allows wildcard subpatterns in finding
Unicode properties, is supposed to only allow ASCII punctuation for
delimitters.  But if you preceded the delimitter by a backslash, the
check was skipped.  This commit fixes that.

It may be that we will eventually want to loosen the restriction and
allow a wider range of delimiters.  But until we have valid use-cases
that would push us in that direction, I don't want to get into
supporting stuff that we might later regret, such as invisible
characters for delimitters.  This feature is not really required for
programs to work, so I don't view it as necessary to be as general as

23 months agoClarify in documentation that sv_isa_sv does not invoke magic
Paul "LeoNerd" Evans [Wed, 11 Dec 2019 12:05:06 +0000 (12:05 +0000)]
Clarify in documentation that sv_isa_sv does not invoke magic

23 months agoAPItest.xs: Increase version
Pali [Tue, 10 Dec 2019 16:48:10 +0000 (17:48 +0100)]
APItest.xs: Increase version

23 months agoAPItest.xs: Fix compile warning: format ‘%d’ expects argument of type ‘int’, but...
Pali [Tue, 10 Dec 2019 14:26:19 +0000 (15:26 +0100)]
APItest.xs: Fix compile warning: format ‘%d’ expects argument of type ‘int’, but argument 2 has type ‘line_t {aka long unsigned int}’

Use "%u" format modifier instead of "%d" as line_t is unsigned. Type line_t
is of U32 type and therefore casting it to unsigned int is safe on most

23 months agoEnable smoke-testing on all branches
Max Maischein [Tue, 10 Dec 2019 17:29:39 +0000 (18:29 +0100)]
Enable smoke-testing on all branches

Previously, branch names that contain a "/" (as in "smoke-me/Corion-foo")
were skipped

23 months agoFix superfluous article and comma splice
Felipe Gasper [Tue, 10 Dec 2019 15:08:11 +0000 (10:08 -0500)]
Fix superfluous article and comma splice

23 months agoAdd the `isa` operator
Paul "LeoNerd" Evans [Wed, 23 Oct 2019 18:00:38 +0000 (19:00 +0100)]
Add the `isa` operator

Adds a new infix operator named `isa`, with the semantics that

  $x isa SomeClass

is true if and only if `$x` is a blessed object reference that is either
`SomeClass` directly, or includes the class somewhere in its @ISA
hierarchy. It is false without warning or error for non-references or
non-blessed references.

This operator respects `->isa` method overloading, and is intended to
replace boilerplate code such as

  use Scalar::Util 'blessed';

  blessed($x) and $x->isa("SomeClass")

23 months agoregcomp.c: Make some refcnt changes NN
Karl Williamson [Sun, 8 Dec 2019 19:25:04 +0000 (12:25 -0700)]
regcomp.c: Make some refcnt changes NN

to save a few cycles

23 months agore/regexp_unicode_prop.t: Add comment
Karl Williamson [Sun, 8 Dec 2019 19:19:08 +0000 (12:19 -0700)]
re/regexp_unicode_prop.t: Add comment

23 months agoregcomp.c: White-space only
Karl Williamson [Sun, 8 Dec 2019 19:18:11 +0000 (12:18 -0700)]
regcomp.c: White-space only

Indent after the previous commit introduced a new block around this

23 months agoPATCH GH #17025 \p{user-defined} overrides official Unicode
Karl Williamson [Sun, 8 Dec 2019 19:16:29 +0000 (12:16 -0700)]
PATCH GH #17025 \p{user-defined} overrides official Unicode

Prior to this patch, they only sometimes overrode.

23 months agoregcomp.c: Replace a loop with strcspn()
Karl Williamson [Sat, 7 Dec 2019 23:08:15 +0000 (16:08 -0700)]
regcomp.c: Replace a loop with strcspn()

A later commit will add a terminating character for the span, so we use
this function instead of memchr.

23 months agoCorrect typo: the warning has 'statement' in singular
James E Keenan [Mon, 9 Dec 2019 00:27:39 +0000 (19:27 -0500)]
Correct typo:  the warning has 'statement' in singular

23 months agofix Github URL in message posted to #p5p-qa
Tony Cook [Tue, 3 Dec 2019 22:27:49 +0000 (09:27 +1100)]
fix Github URL in message posted to #p5p-qa

the repository key is deprecated, and we now produce the correct
full URL, so the comment is meaningless

23 months agore-work travis cases to test more variety
Tony Cook [Tue, 3 Dec 2019 21:50:26 +0000 (08:50 +1100)]
re-work travis cases to test more variety

I expect a -DPERL_GLOBAL_STRUCT_PRIVATE build to catch most
-DPERL_GLOBAL_STRUCT issues, and these are very rarely used options
anyway, so eliminate the second.

Expand the usesitecustomize case to also check long doubles and
an English Unicode locale.

Replace the removed case to check quadmath, cbacktrace and a non-English
Unicode locale.

23 months agoregcomp.c: Replace a loop with memchr
Karl Williamson [Sat, 7 Dec 2019 22:39:18 +0000 (15:39 -0700)]
regcomp.c: Replace a loop with memchr

23 months agoregcomp.c: After AvARRAY(), don't use av_store
Karl Williamson [Thu, 27 Jun 2019 15:49:33 +0000 (09:49 -0600)]
regcomp.c: After AvARRAY(), don't use av_store

We are manimpulating the array directly here, so don't revert to going
through the regular interface, as not all the hand-offs have been done

23 months agoregcomp.c: White space only
Karl Williamson [Thu, 27 Jun 2019 15:45:52 +0000 (09:45 -0600)]
regcomp.c: White space only

23 months agoregcomp.c: Use inRANGE macro
Karl Williamson [Thu, 3 Oct 2019 03:19:22 +0000 (21:19 -0600)]
regcomp.c: Use inRANGE macro

This is faster and clearer

23 months agot/op/die.t: 'use utf8'
Karl Williamson [Thu, 3 Oct 2019 03:09:24 +0000 (21:09 -0600)]
t/op/die.t: 'use utf8'

This file is encoded in UTF-8, even though it didn't say it was.

23 months agohandy.h: Fix typo in little-used macro
Karl Williamson [Wed, 4 Dec 2019 20:46:35 +0000 (13:46 -0700)]
handy.h: Fix typo in little-used macro

This only affected an EBCDIC build where no other tools are around.
This is no longer really used.  It was used to bootstrap EBCDIC when
first porting to it (after the 5.8 series).

23 months agoext/XS-APItest/t/ Fix for EBCDIC
Karl Williamson [Sat, 7 Dec 2019 18:31:48 +0000 (11:31 -0700)]
ext/XS-APItest/t/ Fix for EBCDIC

There were some flaws here that showed up in EBCDIC testing.

23 months agoext/XS-APItest/t/ Don't recalculate value
Karl Williamson [Sat, 7 Dec 2019 18:27:23 +0000 (11:27 -0700)]
ext/XS-APItest/t/ Don't recalculate value

This value is already available in a different variable

23 months agoext/XS-APItest/t/ Fix comment typos
Karl Williamson [Sat, 7 Dec 2019 18:21:52 +0000 (11:21 -0700)]
ext/XS-APItest/t/ Fix comment typos

23 months agoregen/ Fix bug in decoding some high code points
Karl Williamson [Wed, 4 Dec 2019 20:50:12 +0000 (13:50 -0700)]
regen/ Fix bug in decoding some high code points

This affected non-allocated or little-used high code points.

23 months agoregen/ Comments only: typo, add column headings
Karl Williamson [Sat, 7 Dec 2019 17:38:06 +0000 (10:38 -0700)]
regen/ Comments only: typo, add column headings

23 months agoFix UTF8_IS_START on EBCDIC
Karl Williamson [Thu, 5 Dec 2019 19:26:21 +0000 (12:26 -0700)]

23 months agoUpdate IO-Compress to CPAN version 2.093
Chris 'BinGOs' Williams [Sat, 7 Dec 2019 20:10:12 +0000 (20:10 +0000)]
Update IO-Compress to CPAN version 2.093

23 months agoUpdate Compress-Raw-Bzip2 to CPAN version 2.093
Chris 'BinGOs' Williams [Sat, 7 Dec 2019 20:09:35 +0000 (20:09 +0000)]
Update Compress-Raw-Bzip2 to CPAN version 2.093

23 months agoUpdate Compress-Raw-Zlib to CPAN version 2.093
Chris 'BinGOs' Williams [Sat, 7 Dec 2019 20:08:41 +0000 (20:08 +0000)]
Update Compress-Raw-Zlib to CPAN version 2.093

23 months agoUpdate IO-Compress to CPAN version 2.092
Chris 'BinGOs' Williams [Fri, 6 Dec 2019 14:44:28 +0000 (14:44 +0000)]
Update IO-Compress to CPAN version 2.092


  2.092 4 December 2019

      * No changes

23 months agoUpdate Compress-Raw-Bzip2 to CPAN version 2.092
Chris 'BinGOs' Williams [Fri, 6 Dec 2019 14:42:54 +0000 (14:42 +0000)]
Update Compress-Raw-Bzip2 to CPAN version 2.092


  2.092 4 December 2019

      * No changes

23 months agoUpdate Compress-Raw-Zlib to CPAN version 2.092
Chris 'BinGOs' Williams [Fri, 6 Dec 2019 14:41:39 +0000 (14:41 +0000)]
Update Compress-Raw-Zlib to CPAN version 2.092


  2.092 4 December 2019

      * No Changes

23 months agoLocal variable 'last' hides a parameter of the same name
James E Keenan [Tue, 3 Dec 2019 14:16:07 +0000 (09:16 -0500)]
Local variable 'last' hides a parameter of the same name

LGTM static analysis identified a situation where it recommended
renaming a variable.


Better variable name per khw review

23 months agoUpdate link to Birrell's threads paper
brian d foy [Thu, 5 Dec 2019 18:49:36 +0000 (13:49 -0500)]
Update link to Birrell's threads paper

23 months agoLocal variable 'imode' hides a parameter of the same name
James E Keenan [Tue, 3 Dec 2019 02:50:17 +0000 (21:50 -0500)]
Local variable 'imode' hides a parameter of the same name

LGTM static code analysis of Perl 5 source code issued this
recommendation.  Implement fix.

23 months agoperldelta for 5e91ca38cafa2b006f7671e941287e05d31c2d97
James E Keenan [Thu, 5 Dec 2019 02:58:47 +0000 (21:58 -0500)]
perldelta for 5e91ca38cafa2b006f7671e941287e05d31c2d97

23 months agoperldelta for 41fb88e3240, b34c1a7ee25
Tony Cook [Thu, 5 Dec 2019 00:06:08 +0000 (11:06 +1100)]
perldelta for 41fb88e3240b34c1a7ee25

23 months agoupdates to $^H{feature_validname} now update cop_features
Tony Cook [Mon, 2 Dec 2019 04:50:49 +0000 (15:50 +1100)]
updates to $^H{feature_validname} now update cop_features

This removes $^{FEATURE_BITS}, since it's no longer needed.

23 months agotests for the faster features problem sprout pointed out
Tony Cook [Mon, 2 Dec 2019 04:22:48 +0000 (15:22 +1100)]
tests for the faster features problem sprout pointed out

23 months agoCorrect inconsistent version numbers in
James E Keenan [Wed, 4 Dec 2019 15:05:35 +0000 (10:05 -0500)]
Correct inconsistent version numbers in

Per commit message by David Mitchell in 142a37fd (Jul 24 2013), the code
changes affecting the performance of $`, $&, $' weren't fully enabled
until 5.20.  Hence, the documentation should caution for 5.18 and

As reported by peterdd++ in

23 months agoperldelta for 62fc08893a06fdb5633e7656e3c783d705682fd1
James E Keenan [Wed, 4 Dec 2019 14:59:38 +0000 (09:59 -0500)]
perldelta for 62fc08893a06fdb5633e7656e3c783d705682fd1

23 months agoUpdate Test-Simple to CPAN version 1.302170
Chad Granum [Wed, 4 Dec 2019 14:36:07 +0000 (09:36 -0500)]
Update Test-Simple to CPAN version 1.302170

From Changes:

    1.302170  2019-12-02 13:25:48-08:00 America/Los_Angeles

    - Fix unwanted END phase event (#840)

23 months agoskip code that requires dynamic loading and minitest works
Tony Cook [Mon, 18 Nov 2019 03:34:13 +0000 (14:34 +1100)]
skip code that requires dynamic loading and minitest works

23 months agofix skip for loop over locales
Tony Cook [Mon, 18 Nov 2019 03:33:31 +0000 (14:33 +1100)]
fix skip for loop over locales

Any skips inside this loop was skipping over the entire loop,
not just the current iteration, from the conditions tested in
the loop this seemed incorrect (besides messing up the test count.)

23 months agoPOSIX isn't loadable in miniperl
Tony Cook [Sun, 17 Nov 2019 21:29:47 +0000 (08:29 +1100)]
POSIX isn't loadable in miniperl

23 months agomove the implementation of %-, %+ into core
Tony Cook [Wed, 13 Nov 2019 21:02:34 +0000 (08:02 +1100)]
move the implementation of %-, %+ into core

Previousl this could cause problems during minitest.

Fixes #17293

23 months agoRegenerate Configure and friends after metaconfig changes
Karl Williamson [Mon, 2 Dec 2019 16:52:14 +0000 (09:52 -0700)]
Regenerate Configure and friends after metaconfig changes

A probe for wcrtomb() and one for attribute always inline have been
added, and the ones for:
1) checking if there is a C backtrace facility; and
2) character data alignedness
have been revised

23 months agometaconfig.h: Add entry for wcrtomb()
Karl Williamson [Mon, 2 Dec 2019 15:42:20 +0000 (08:42 -0700)]
metaconfig.h: Add entry for wcrtomb()

23 months agometaconfig.h: Rmv temporary entries
Karl Williamson [Mon, 2 Dec 2019 15:41:43 +0000 (08:41 -0700)]
metaconfig.h: Rmv temporary entries

These are no longer needed