This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Karl Williamson [Wed, 22 Jul 2020 12:02:56 +0000 (06:02 -0600)]
embed.fnc: Mark doing_taint as Core only
This appears to be for internal use, and there are no cpan usages
Karl Williamson [Wed, 22 Jul 2020 11:48:53 +0000 (05:48 -0600)]
embed.fnc: Mark cxinc as Core only
The macro CXINC is what one is supposed to use for this functionality.
(though it is currently undocumented)
Karl Williamson [Wed, 22 Jul 2020 11:46:35 +0000 (05:46 -0600)]
embed.fnc: Mark moreswitches as Core only
This is an internal function used to parse command line options; there
are no cpan uses.
Karl Williamson [Wed, 22 Jul 2020 11:38:37 +0000 (05:38 -0600)]
embed.fnc: Mark runops_debug, runops_standard as Core only
These are internal functions to run the program; there are no cpan uses
Karl Williamson [Wed, 22 Jul 2020 11:29:01 +0000 (05:29 -0600)]
embed.fnc: Mark regnext as Core only
This is used internally by the regex engine; there is one use in cpan,
an internals module, Devel::RegExp by Ilya Zakharevich, with nothing
dependent on it, and last updated in 1995.
Karl Williamson [Wed, 22 Jul 2020 11:19:41 +0000 (05:19 -0600)]
embed.fnc: Mark cx_dup as Core only
Similar cx functions are supposed to be called with a macro for their
functionality, but no such macro exists for this, and there are no uses
on CPAN
Karl Williamson [Mon, 14 Dec 2020 02:41:24 +0000 (19:41 -0700)]
perlapi: Document and consolidate SvPV functions
Karl Williamson [Mon, 14 Dec 2020 15:51:45 +0000 (08:51 -0700)]
POSIX::asctime pod: Note it always returns English
And give locale's alternative
Karl Williamson [Mon, 14 Dec 2020 02:07:54 +0000 (19:07 -0700)]
perlfunc: Improve localtime entry
This rearranges some paragraphs that really belong together, and
clarifies the result is always in English
Karl Williamson [Sun, 20 Dec 2020 13:28:38 +0000 (06:28 -0700)]
regexec.c: Link to github issue in comment
Max Maischein [Sun, 20 Dec 2020 14:44:10 +0000 (15:44 +0100)]
Prepare Module::CoreList for 5.33.6
Max Maischein [Sun, 20 Dec 2020 14:32:28 +0000 (15:32 +0100)]
Bump version to 5.33.6
Max Maischein [Sun, 20 Dec 2020 14:28:55 +0000 (15:28 +0100)]
new perldelta for 5.33.6
Max Maischein [Sun, 20 Dec 2020 14:23:02 +0000 (15:23 +0100)]
Update release schedule
Max Maischein [Sun, 20 Dec 2020 14:22:54 +0000 (15:22 +0100)]
Update epigraphs.pod with quote from Max Weber
Max Maischein [Sun, 20 Dec 2020 14:12:21 +0000 (15:12 +0100)]
Merge branch 'release-5.33.5' into blead
# Conflicts:
# pod/perldelta.pod
Max Maischein [Sun, 20 Dec 2020 10:19:16 +0000 (11:19 +0100)]
add new release to perlhist
Max Maischein [Sun, 20 Dec 2020 09:53:09 +0000 (10:53 +0100)]
Update Module::CoreList for 5.33.5
Max Maischein [Sun, 20 Dec 2020 09:45:48 +0000 (10:45 +0100)]
Finalize perldelta for 5.33.5
Max Maischein [Sun, 20 Dec 2020 09:32:01 +0000 (10:32 +0100)]
Finalise perldelta for 5.33.5
Karl Williamson [Sun, 6 Dec 2020 22:01:14 +0000 (15:01 -0700)]
Fix broken PERL_MEM_LOG under threads
This fixes GH #18341
There are problems with getenv() on threaded perls wchich can lead to
incorrect results when compiled with PERL_MEM_LOG.
Commit
0b83dfe6dd9b0bda197566adec923f16b9a693cd fixed this for some
platforms, but as Tony Cook, pointed out there may be
standards-compliant platforms that that didn't fix.
The detailed comments outline the issues and (complicated) full solution.
Karl Williamson [Tue, 1 Dec 2020 14:44:18 +0000 (07:44 -0700)]
Add GETENV_LOCK
get_env() needs to lock other threads from writing to the environment
while it is executing. It may need to have an exclusive lock if those
threads can clobber its buffer before it gets a chance to save them.
The previous commit has added a Configure probe which tells us if that
is the case. This commit uses it to select which type of mutex to use.
Karl Williamson [Sat, 28 Nov 2020 16:20:46 +0000 (09:20 -0700)]
Add Configure probe for getenv() buffer race
Most implementations do not have a problem with two getenv()'s running
simultaneously in different threads. But Posix doesn't require such
good behavior. This adds a simple probe to test the current system.
Karl Williamson [Tue, 1 Dec 2020 14:17:21 +0000 (07:17 -0700)]
perl.h: Add capability for many-reader ENV mutex locking
There are several places where there could be a problem in the
environment were changed by another thread when a function is executing,
but otherwise if another thread were reading the environment at the same
time, there isn't a problem. This adds mutex for that situation.
Future commits will take advantage of it.
Karl Williamson [Wed, 9 Dec 2020 01:30:56 +0000 (18:30 -0700)]
Make many-reader mutexes more resilient
These mutexes rely on a counter being accurate to work. If for some
reason that I can't imagine happening, the count goes below 0, this
commit resets it to zero, which may be enough to cause the program to
continue.
Karl Williamson [Sat, 14 Nov 2020 20:55:38 +0000 (13:55 -0700)]
regexec.c: White-space, comments only
Mostly indent because the prior commit created a new block
Karl Williamson [Fri, 13 Nov 2020 16:38:21 +0000 (09:38 -0700)]
regexec.c: Revamp S_setup_EXACTISH_ST() loop end conditions
Consider the pattern /A*B/ where A and B are arbitrary. The pattern
matching code tries to make a tight loop to match the span of A's. The
logic of this was not really updated when UTF-8 was added. I did
revamp it some releases ago to fix some bugs and to at least consider
UTF-8.
This commit changes it so that Unicode is now a first class citizen.
Some details are listed in the ticket GH #18414
Karl Williamson [Sun, 18 Oct 2020 16:51:22 +0000 (10:51 -0600)]
regexec.c: Change name of static function
The new name reflects its new functionality coming in future commits
Karl Williamson [Sun, 18 Oct 2020 16:46:14 +0000 (10:46 -0600)]
regexec.c: Trim trailing blanks
Karl Williamson [Sun, 18 Oct 2020 16:20:38 +0000 (10:20 -0600)]
regcharclass.pl: Get code point folding to a seq
Previously regcharclass.pl could tell if an input string was a
multi-character fold of some Unicode code point. This commit adds the
ability to return what that code point is. This capability will be used
in a later commit.
Karl Williamson [Wed, 11 Nov 2020 23:37:23 +0000 (16:37 -0700)]
regen/regcharclass.pl: Rmv special case
This avoided checking for optimizations. Whatever its original use, it
doesn't do any good, and the optimizations are actually useful.
James E Keenan [Sat, 19 Dec 2020 20:41:56 +0000 (15:41 -0500)]
ext/POSIX/t/waitpid.t: Correct one typo
TAKAI Kousuke [Fri, 18 Dec 2020 09:55:50 +0000 (18:55 +0900)]
MANIFEST: Add ext/POSIX/t/iv_const.t
TAKAI Kousuke [Thu, 17 Dec 2020 14:27:31 +0000 (23:27 +0900)]
ext/POSIX: Added tests to check integer constants are defined as IV
TAKAI Kousuke [Wed, 16 Dec 2020 17:17:31 +0000 (02:17 +0900)]
ext/POSIX: Change integer constants to IV.
Some floating-point related constants, such as (FLT|L?DBL)_DIG,
FLT_RADIX, FP_*, etc. had been defined as NV, but these are
actually integer constants.
This change should not affect the behaviour of user programs.
James E Keenan [Sat, 19 Dec 2020 17:42:48 +0000 (17:42 +0000)]
Jason McIntosh [Sat, 19 Dec 2020 14:37:40 +0000 (14:37 +0000)]
Add pod/perldocstyle.pod
Satisfies https://news.perlfoundation.org/post/grant_proposal_documentation_standards_perl7
Committer: Adding a new file underneath pod/ entails adding (i) an entry for that file
in pod/perl.pod and (ii) entries in win32/pod.mak.
(i) In turn requires, figuring out an appropriate category within
perl.pod in which to place the new file. The entry should match that in
MANIFEST and be consistent with the style of entries in MANIFEST.
(ii) Requires running Porting/pod_rules.pl so that entries in
win32/pod.mak are also generated for the corresponding *.man, *.html and
*.tex files.
Supersedes https://github.com/Perl/perl5/pull/18275.
H.Merijn Brand [Fri, 18 Dec 2020 15:41:30 +0000 (16:41 +0100)]
Update Config::Perl::V to 0.33
James E Keenan [Fri, 18 Dec 2020 14:21:37 +0000 (09:21 -0500)]
Merge branch 'thibaultduponchelle-fix_rt_number_in_perl5260delta' into blead
James E Keenan [Fri, 18 Dec 2020 14:20:42 +0000 (09:20 -0500)]
Merge branch 'fix_rt_number_in_perl5260delta' of https://github.com/thibaultduponchelle/perl5 into thibaultduponchelle-fix_rt_number_in_perl5260delta
Thibault DUPONCHELLE [Fri, 18 Dec 2020 13:27:53 +0000 (14:27 +0100)]
Fix RT number in perl5260delta.pod
Chris 'BinGOs' Williams [Fri, 18 Dec 2020 12:54:10 +0000 (12:54 +0000)]
Update ExtUtils-Install to CPAN version 2.20
[DELTA]
2.20
- Exercise _is_prefix() method more in tests
- Optimisations for file comparisons
- Optimisations for directory creation
- Typo fix in POD
- Optimisations for tree traversal
Max Maischein [Fri, 18 Dec 2020 09:44:02 +0000 (10:44 +0100)]
Also update Maintainers.pl with the updated PerlIO::via::QuotedPrint version
Max Maischein [Fri, 18 Dec 2020 09:40:09 +0000 (10:40 +0100)]
Update PerlIO::via::QuotedPrint from 0.08 to 0.09
This should be picked up by the auto-generated perldelta
Max Maischein [Fri, 18 Dec 2020 09:12:49 +0000 (10:12 +0100)]
Update Text::Balanced from 2.03 to 2.04
No entry in Perldelta because that will be generated automatically
Leon Timmermans [Thu, 17 Dec 2020 20:02:46 +0000 (21:02 +0100)]
Correctly set LC_CTYPE back in POSIX/t/time.t
James E Keenan [Tue, 15 Dec 2020 16:38:30 +0000 (11:38 -0500)]
t/lib/croak/toke: Correct test label
The RT ticket number cited in the test's label is wrong. The body of
the commit message for commit
7259f4194f3131957240f6b3dba47b74f53ac660
shows the correct RT number. Provide GH number as well.
Karl Williamson [Mon, 14 Dec 2020 20:38:47 +0000 (13:38 -0700)]
autodoc.pl: Adjust max permissible line length
perldoc takes up 4 column indent, not 7.
Karl Williamson [Mon, 14 Dec 2020 20:33:37 +0000 (13:33 -0700)]
podcheck.t: Combine duplicated code into a sub
Karl Williamson [Mon, 14 Dec 2020 20:29:19 +0000 (13:29 -0700)]
podcheck.t: perldoc standard indent is 4.
Before, it was using an expected nroff indent of 7. But current perldoc
places the =head1 lines at the left margin, and the text within them
indented by 4.
Karl Williamson [Mon, 14 Dec 2020 20:41:07 +0000 (13:41 -0700)]
perlapi: 'ie' should be 'i.e.', and italicized
Tony Cook [Mon, 14 Dec 2020 05:00:33 +0000 (16:00 +1100)]
bump $XS::APItest::VERSION
Tony Cook [Mon, 14 Dec 2020 04:54:25 +0000 (15:54 +1100)]
consistently parenthesise the RW lock macro arguments
and fix the mixed up pointer vs non-pointer use of the mutex
argument.
Without the parentheses (or without the new tests in the previous
commit) for code like;
PERL_WRITE_LOCK(&PL_some_mutex);
the MUTEX_LOCK(mutex.lock) in that code would expand to:
MUTEX_LOCK(&PL_some_mutex.lock);
and happen to work, even though the next line has:
if ((mutex)->readers_count)
treating the mutex parameter as a pointer.
With the parentheses the MUTEX_LOCK(mutex.lock) becomes:
MUTEX_LOCK((&PL_some_mutex).lock);
which as expected fails to compile in a similar way to the pointer
test code in the previous commit.
Tony Cook [Mon, 14 Dec 2020 04:20:17 +0000 (15:20 +1100)]
build test for the read/write mutex macros
This fails to compile
Karl Williamson [Thu, 27 Aug 2020 13:47:02 +0000 (07:47 -0600)]
perl.h: White-space only
Properly indent some nested #defines
Karl Williamson [Sun, 6 Sep 2020 16:55:28 +0000 (10:55 -0600)]
perlapi: Remove duplicate text
This came about in a rebasing error
Karl Williamson [Mon, 14 Sep 2020 16:21:49 +0000 (10:21 -0600)]
Document and consolidate SvPV..force functions
Karl Williamson [Wed, 22 Jul 2020 22:47:53 +0000 (16:47 -0600)]
Consolidate and document all get_cvFOO() variants
Karl Williamson [Thu, 10 Dec 2020 17:11:34 +0000 (10:11 -0700)]
many-reader mutexes: Change structure element name
The old name did not encompass all the possible reasons for the
mutex signal condition to be invoked
Karl Williamson [Thu, 10 Dec 2020 18:07:15 +0000 (11:07 -0700)]
PERL_WRITE_UNLOCK: add missing condition signal
The mutex is locked, and any contenders are awaiting a signal, which
until this commit was missing.
Karl Williamson [Thu, 10 Dec 2020 15:58:53 +0000 (08:58 -0700)]
dist/threads/t/libc.t: Add timer to avoid deadlock
This test file can deadlock if there are bugs. Add a timeout to keep
the test from hanging indefinitely.
Karl Williamson [Fri, 11 Dec 2020 21:59:28 +0000 (14:59 -0700)]
locale.c: Work around a z/OS limitation/feature
Without per-thread locales, a multi-thread application is inherently
unsafe. IBM solves that by allowing you to set up the locale any way
you want, but after you've created a thread, all future locale changes
are ignored, and return failure.
But Perl itself changes the locale in a couple of cases. Recent changes
have surfaced this issue in one case, leading to a panic. And this
commit works around it, so that messages will be displayed in the locale
in effect before the threads were created.
The remaining case requires further investigation. Nothing in our suite
is failing.
Karl Williamson [Thu, 10 Dec 2020 23:09:31 +0000 (16:09 -0700)]
perlapi: Add markup
Tom Hukins [Wed, 2 Dec 2020 20:01:58 +0000 (20:01 +0000)]
Combine the two separate smoke test sections
Tom Hukins [Tue, 17 Nov 2020 16:43:29 +0000 (16:43 +0000)]
Simplify the release management instructions
Getting access to GitHub issues and commit privileges takes one action,
not two, so merge the separate sections.
Tom Hukins [Tue, 17 Nov 2020 13:58:40 +0000 (13:58 +0000)]
dromedary no longer has a public_html directory
The old users.perl5.git.perl.org machine has been offline for some time.
A replacement exists at dromedary.p5h.org, but this does not provide
per-user public_html directories for sharing files.
Steve Hay [Thu, 10 Dec 2020 14:15:51 +0000 (09:15 -0500)]
Synch with CPAN distribution libnet-3.12
James E Keenan [Thu, 10 Dec 2020 14:01:03 +0000 (09:01 -0500)]
Correct POD formatting error
Paul "LeoNerd" Evans [Sat, 27 Jun 2020 17:10:47 +0000 (18:10 +0100)]
Implement SAVEt_STRLEN_SMALL
Most uses of SAVEt_STRLEN actually store small values; often zero.
Rather than using an entire U64-sized element for these values, it saves
space to use the same "SMALL" mechanism as other numerical values, like
SAVEt_INT_SMALL.
Karl Williamson [Wed, 9 Dec 2020 00:09:04 +0000 (17:09 -0700)]
Fix freebsd/netbsd builds
These have an inconsistent configuration in which several functions exist
for thread-safe locales, but the crucial one doesn't show up in our
Configure probe. The code this commit fixes assumed that all or nothing
would be present.
Karl Williamson [Wed, 9 Dec 2020 00:07:17 +0000 (17:07 -0700)]
Fix cygwin build
A macro name changed, and I didn't see it in the grep
Tony Cook [Tue, 8 Dec 2020 23:43:12 +0000 (10:43 +1100)]
bump perl5db.pl's $VERSION
Tony Cook [Tue, 8 Dec 2020 23:42:59 +0000 (10:42 +1100)]
jkahrman is now a perl author
Tony Cook [Tue, 8 Dec 2020 23:19:26 +0000 (10:19 +1100)]
test - and . at beginning of debugger alias name
jkahrman [Thu, 16 Apr 2020 06:00:04 +0000 (02:00 -0400)]
Allow debugger aliases that start with '-' and '.'
Since the '.' and '-' commands don't take any arguments and don't run if any are provided, don't treat commands starting with these characters as the single commands '.' and '-'. Restores behavior that existed prior to https://github.com/Perl/perl5/commit/
7fdd4f080863703d44282c6988834455d129040 (v5.27) at least back to v5.8.8
https://github.com/Perl/perl5/commit/
7fdd4f080863703d44282c6988834455d129040
Karl Williamson [Tue, 8 Dec 2020 20:25:36 +0000 (13:25 -0700)]
locale.c: Remove some unnecessary mutex locks
These aren't necessary as the called function has its own lock until
done copying into the local structure.
And these were breaking blead on Windows, as they are no longer defined.
The smoke I ran included more commits beyond the breaking one, so I
didn't catch it.
Karl Williamson [Sun, 6 Dec 2020 23:29:11 +0000 (16:29 -0700)]
locale.c: Unlock mutex before croaking
These cases aren't supposed to happen, but unlock the mutex first;
we could get into deadlock in trying to output the death message.
Karl Williamson [Mon, 30 Nov 2020 17:11:01 +0000 (10:11 -0700)]
Name individual locale locks
These locks for different functions all use the same underlying mutex;
but that may not always be the case. By creating separate names
used only when we think they will be necessary, the compiler will
complain if the conditions in the code that actually use them are the
same. Doing this showed a misspelling in an #ifdef, fixed in
9289d4dc7a3d24b20c6e25045e687321ee3e8faf
Karl Williamson [Mon, 30 Nov 2020 16:46:34 +0000 (09:46 -0700)]
Refactor locale mutex setup
This was prompted by my realization that even on a locale thread-safe
platform, there are functions we call that may not be thread-safe in
that they return their results in an internal static buffer, which may
be process-wide instead of per-thread. Tomasz Konojacki++ briefly
looked at Windows source code for localeconv() and this indeed did
appear to be the case.
If we thought a platform was thread-safe, no locale mutexes were set up,
and instead the calls in the code to lock were no-oops. This would lead
to potential races, the most likely candidate being localeconv(). None
have been reported, at least as far as we know. Likely that function
isn't called frequently. This would be true on both Posix 2008 and
Windows platforms, except possibly for FreeBSD, which may be the only
platform that we support that has a localeconv_l() function, which is
supposed to be immune from this issue..
The solution adopted here is to test for all the possible functions that
the Perl core uses that may be susceptible to this, and to set up the
mutex if any are found. Thus there won't be no-ops where there should
be a lock.
Karl Williamson [Mon, 30 Nov 2020 00:05:13 +0000 (17:05 -0700)]
Change name of mutex macro.
This macro is for localeconv(); the new name is clearer as to the
meaning, and this preps for further changes.
Karl Williamson [Sun, 6 Dec 2020 22:45:20 +0000 (15:45 -0700)]
locale.c: Add debugging info to panic message
Karl Williamson [Sun, 6 Dec 2020 22:43:45 +0000 (15:43 -0700)]
duplocale() is part of Posix 2008 locales
Thus if we know we have the Posix versions, we have duplocale(), and
hence don't need to test separately for it.
Karl Williamson [Sun, 6 Dec 2020 22:16:41 +0000 (15:16 -0700)]
Fix up many-reader mutex typedef
This previously worked on some compilers, but not others.
Jae Bradley [Sat, 5 Dec 2020 20:57:16 +0000 (15:57 -0500)]
Add author name in alphabetically correct place
Jae Bradley [Wed, 2 Dec 2020 00:35:30 +0000 (19:35 -0500)]
Update the Xcode version in the testsuite.yml workflow
Xcode v12 is now the default Xcode version: https://github.com/actions/virtual-environments/issues/2056
Jae Bradley [Tue, 1 Dec 2020 23:48:55 +0000 (18:48 -0500)]
Use macOS-10.15 explicitly
Jae Bradley [Tue, 1 Dec 2020 23:40:12 +0000 (18:40 -0500)]
Add Jae Bradley as an Author
Jae Bradley [Tue, 1 Dec 2020 20:04:51 +0000 (15:04 -0500)]
Explicitly use Catalina when running the test with Xcode 11
Tom Hukins [Wed, 2 Dec 2020 17:00:09 +0000 (17:00 +0000)]
Remove empty "#ifdef"s
Chris 'BinGOs' Williams [Tue, 8 Dec 2020 11:58:44 +0000 (11:58 +0000)]
Here be camels
David Mitchell [Tue, 8 Dec 2020 09:50:47 +0000 (09:50 +0000)]
perly.y: avoid <0 test on unsigned value
Coverity CID 313707
Moving to a newer version of Bison has changed how the YYTRANSLATE()
macro is defined: in particular it now has a <0 test, which
Coverity is complaining about, since the arg is an unsigned value.
This commit just casts the arg back to a signed value. In more detail:
we formerly had:
yytoken = YYTRANSLATE(NATIVE_TO_UNI(parser->yychar));
yychar is of type int, but NATIVE_TO_UNI returns a UV. So just cast the
result back to an int:
yytoken = YYTRANSLATE((int)NATIVE_TO_UNI(parser->yychar));
TAKAI Kousuke [Tue, 1 Dec 2020 00:50:35 +0000 (09:50 +0900)]
perldelta.pod: Document new "0o" octal syntax.
TAKAI Kousuke [Sat, 31 Oct 2020 05:15:27 +0000 (14:15 +0900)]
perlfunc.pod: add a description for octal strings for oct()
These sentences are borrowed from the description for hex().
TAKAI Kousuke [Fri, 30 Oct 2020 12:55:03 +0000 (21:55 +0900)]
perldata.pod: mention new octal integer format (0o12_345).
TAKAI Kousuke [Thu, 22 Oct 2020 17:14:57 +0000 (02:14 +0900)]
pp.c: oct() now skips "0o" and "o" prefix.
t/op/oct.t: Add test for oct() with 0o and o prefix.
TAKAI Kousuke [Thu, 22 Oct 2020 16:07:16 +0000 (01:07 +0900)]
toke.c: Eliminate temporary variables base, Base and max.
These variables are only used on emitting diagnostic messages.
Calculating them on-demand will make the code slightly faster
on normal cases.
Note: previously bases[], Bases[] and maxima[] may be completely
optimized out by fusing array accesses into if-brances.
Now they become real arrays, and will slightly increase the number
of dynamic relocations on PIC build.
TAKAI Kousuke [Thu, 22 Oct 2020 14:56:58 +0000 (23:56 +0900)]
toke.c: Preserve "0o" prefix on warnings and strings passed to overload hook.
TAKAI Kousuke [Thu, 22 Oct 2020 14:23:57 +0000 (23:23 +0900)]
toke.c: Recognize "0odddd" octal literals.
t/base/num.t: Add some test for "0odddd" octals.
Branislav Zahradník [Thu, 6 Aug 2020 07:18:15 +0000 (09:18 +0200)]
Use GNU Bison's named references
Usage of Bison's named references makes actions little bit easier
to read and maintain.
Branislav Zahradník [Thu, 6 Aug 2020 17:42:08 +0000 (19:42 +0200)]
Raise minimal supported GNU Bison version to 2.5
which added support for named references