| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
| |
Export lock_hashref_recurse and unlock_hashref_recurse
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
t/pod.t no longer exists on CPAN, and the MAP is no longer required
|
| | |
|
| | |
|
| |
|
|
| |
newXSUB hasn't been around for a long time
|
| |
|
|
|
|
|
|
| |
nytprof using full (not mini) Win32 perl reported 44 calls on line
"$ENV{PERL_CORE} = 1;" at 623 us spent on line for make_ext.pl --dynamic
Move the env var setting out from build_extension() to before the foreach
loop starts.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
[DELTA]
2.11 2015-07-17
- Change argument check error from croak to carp since there are some
dependent modules using non-standard options. The error check will
move back to croak when dependent modules have the opportunity to
correct the argument.
2.10_005 2015-07-17
- Better argument checking and reporting of failure for unrecognized
options.
- RT 71562 Document automount race condition limitation and workaround
- RT 99230 Document multithreaded application limitation and that the
limitation may be removed in a future release.
2.10_004 2015-07-10
- Remove use of English.pm since it breaks many older Perls
- Fix a unit test skip count for users who have not installed
Test::Output
2.10_003 2015-07-08
- Administrative changes to MANIFEST and MANIFEST.SKIP
- Style changes to Path.pm for easier reading of code
- Removal of pod.t
- Use English.pm to make variables like $! more fluent for code readers
2.10_002 2015-06-26
- RT 42139. Add better SKIP test emit for VMS.
- RT 85360. Fix typos and better .gitignore and MANIFEST.SKIP settings.
- RT 51588. Added patch elements but unable to verify on VMS.
2.10_001 2015-06-24
- RT 39949. Report errors on lstat failure.
- RT 53178. Deprecate UNIVERSAL::isa usage
- RT 70657. Test added.
- RT 70938. Documentation fix for Windows native relative usage
- RT 72256. Option added for setting permission (chmod)
- RT 73840. Fix taint/untaint bug.
- RT 95150. Add CPAN metadata for source tree location.
- RT 103512. Documentation language fix.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
[DELTA]
3.07 2015-07-17
- Net::FTP::rmdir() has been made more robust by making use of the MLSD
command in addition to the NLST command since the latter is known not to
be processed correctly by some FTP servers.
[Chris Lindee, CPAN RT#100694]
- Net::FTP, Net::NNTP, Net::POP3 and Net::SMTP can now restrict domain to
IPv4 even if IPv6 is available by using the new Domain or Family argument.
Net::NNTP now supports the LocalPort argument in addition to LocalAddr.
Net::POP3 now supports the LocalAddr and LocalPort arguments in addition
to ResvPort (which is retained for backwards compatibility).
[Steffen Ullrich, PR#18]
- Fixed a bug in Net::Cmd::datasend() which caused octets in [\x80-\xFF]
stored in a "binary string" to be replaced with their UTF-8 encodings if
the string happened to be stored internally in an "upgraded" state (i.e.
with the UTF-8 flag on). (As noted below, strings passed to datasend()
should always be encoded first, and therefore not stored in such a state
anyway, but it is all too easy for perl to change this internal state
unless the encodeing is done at the very last minute before calling
datasend(), so it helps if datasend() plays more nicely in this case. In
particular, it was wrong of datasend() to treat upgraded and downgraded
strings differently when their contents were identical at the Perl level.)
This bugfix results in a breaking change to the case of a "text string"
with characters in U+0080..U+00FF stored internally in an upgraded state
since those characters are likewise no longer encoded to UTF-8 by
datasend(), but callers of datasend() should not have been relying on this
behaviour anyway: In general, datasend() has no idea what encoding is
required for output so callers should always encode the data to be output
to whatever encoding is required first. This has now been clarified in
the documentation.
Finally, a text string with characters >= U+0100 will now cause a "Wide
character in print" warning from datasend() since such characters cannot
be output as bytes and datasend() no longer encodes to UTF-8. In this
case, UTF-8 bytes will still be output as before since that happens to be
the internal representation of such characters, but the warning is new.
Callers should heed this warning and encode such strings to whatever
encoding is required before calling datasend(), as noted above.
[Ricardo Signes, CPAN RT#104433]
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
In 04ce8649990 I inadvertently created conflicting definitions for
PerlIO_ungetc on the non-VMS branch of an #ifdef. This broke a
NetWare build that was still based on the stdio layer. See
<http://nntp.perl.org/group/perl.perl5.porters/228427>. Even
though stdio is no longer supported, we really ought to ship
something that compiles up until the point where we remove it,
which will hopefully be soon for perlsdio.h.
N.B. This commit is not sufficient to get stdio as the base
layer working again. Notably XS::APITest is broken for unrelated
reasons.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
| |
sv_catpvf() and friends ultimately end up calling sv_vcatpvfn_flags() with a
C-style va_list argument (rather than with an array of SV pointers). When
the sprintf implementation in sv_vcatpvfn_flags() is called with a va_list
it always ignores any attempt by the format string to reorder the arguments.
This reasonable limitation is now documented, and the implementation throws
an exception when it encounters this situation.
Minimal tests for these exceptions have been added to XS::APItest.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The printf builtins gained support for argument reordering in October 2000,
in eb3fce905f8436bbc374998ec8c7c34ce2b73e4e, as part of the 5.7.x blead
release cycle. However, a simple typo meant that it never worked: the
initial implementation stored the argument index in the variable "ewix"
(explicit width-argument index) instead of "epix" (for "precision").
The bug was noticed in July 2002, two days before the 5.8.0 release, so
fixing it at that stage was deemed riskier than documenting it. The commit in
question is 7b8dd722af72d0ca45650fb784c09763c0732e34, which was originally
posted as http://marc.info/?l=perl5-porters&m=102683138220786&w=2 .
For whatever reason, the obvious fix was never subsequently applied. This
commit remedies that, and extends it to also skip all redundant-argument
warnings when the precision is supplied with via an explicit index.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- RT#125469 points out that no "redundant argument" warning should be emitted
for code like C<< printf '<%*2$s>', "a", 6 >>; that's now fixed.
- We no longer emit a "missing argument" warning for invalid format strings,
so C<< printf '%4$K %d', 17 >> now emits one "invalid" warning, and no
other warnings. (Perl 5.12 and subsequent versions have inappropriately
emitted a "missing argument" warning in this case.)
- We no longer treat the invalid format string in C<< printf '%1$$d', 17 >>
as containing an explicit index, so (a) we emit an "invalid" warning for
the double "$", and (b) we emit a "redundant argument" warning for the
trailing argument. The "redundant argument" warning is new in this
situation.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
This moves the definition to before the function it is used in, rather
than disrupting the flow of code within the function.
|
| |
|
|
|
|
|
|
|
|
|
| |
See http://nntp.perl.org/group/perl.perl5.porters/229168
Also, the documentation has been updated beyond this change to clarify
related matters, based on some experimentation.
Previously, spaces couldn't be in variable names; now ASCII control
characters can't be either. The remaining permissible ASCII characters
in a variable name now must be all graphic ones.
|
| |
|
|
|
| |
Caret variable names don't have to be limited to $^A through $^Z. $^],
etc. are also valid.
|
| |
|
|
|
| |
Add some clarifying comments, and properly indent some lines to
prevailing level.
|
| |
|
|
|
|
|
|
| |
These show a bug in perl parsing where utf8ness makes a difference.
in what happens. In this case, a syntax error is accompanied by warning
messages when in 'use utf8', and no warnings when not. I'm not filing a
bug report, as I don't think it is worth fixing, as it is a syntax error
after all. But I did make tests for it, as TODOs.
|
| |
|
|
| |
This helps debug when the test fails.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
The feature still exists, for compatibility with code that tries to enable
it, but it has no effect. The postderef_qq feature still exists, however.
|
| |
|
|
|
|
|
|
| |
There is no problem with setting or deleting HOME as there was
back in 58277c145607d, but there is still no way to delete
SYS$LOGIN from the job table. So we do have to skip testing for
what happens with chdir() when the environment is clear, but it's
not conditional on $Config{d_setenv}.
|
| |
|
|
|
|
|
| |
$OLD_PERL_VERSION was added commented-out in Perl 5.6.0, when $PERL_VERSION
became $^V. It remained unimplemented and undocumented until it was added to
perlvar.pod, in the deprecated section, in 5.13.10. It was then implemented in
English.pm in 5.19.10, but remained undocumented until 5.22.0.
|
| |
|
|
|
|
| |
The CRTL has supplied getlogin since v7.0, so let's not use the
home-made one anymore. Plus the CRTL als has a reentrant version,
so we'll use that under threads.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
In order to address weaknesses in the documentation of find_version() in
CoreList.pod, we first had to address its lack of unit tests. It was
previously only exercised inside the 'corelist' utility. Refactored
function a bit to expose all its branches for testing. Then added tests
for the function with and without arguments. Corrected documentation
for find_version() in CoreList.pod.
Documentation subsequently clarified per feedback from Steve Parker.
For: RT # 125563 (1st of 2 problems cited in that ticket).
|
| |
|
|
|
|
|
| |
I.e., a string like 5.010001. This makes documentation of deprecated_in()
consistent with that of other Module::CoreList functions.
For: RT # 125563 (2nd of two problems raised)
|
| |
|
|
|
|
|
|
|
|
|
| |
There used to be some version of GCC 2.x that ran only on VAX that
was capable of building Perl. But that was 18-20 years ago and
there have been no recent reports of building Perl with gcc on VMS.
If and when a modern version of GCC is ported to VMS (and the
underpinnings are reportedly there as part of GNAT Pro Ada, just
nothing complete or publicly available), these ancient workarounds
are at least as likely to do harm as good. So get rid of them.
|
| |
|
|
|
|
|
|
|
| |
It turns out it's quite legal but then causes other mayhem, such
as confusing things that are looking for the "." directory (because
there is no such thing as a file without an extension so passing
an empty string to fopen creates ".;1" on disk).
Also make this test clean up its test files.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
xpadl_alloc should really be pointer to a struct with a flexible array
member, but flexible array members aren't portable enough among CCs. While
debugging the padlist API for memory corruption (caused by an unrelated
XS module), I saw that the pointer in the first slice of xpadl_alloc
pointed to an AV head of gibberish but 2nd slice was fine. This was
confusing and led me to belive the memory corruption was a bad write to
the array in xpadl_alloc. PadlistARRAY's POD a couple pages down mentions
that index 0 is not an AV *, but the struct comments just said
"pointer to beginning of array of AVs " and didnt mention index 0.
Fix the comments to make it clear what xpadl_alloc is. Add a union so it
is easier to analyze a crash dump/breakpoint with a C debugger, without
writing new code "PADNAMELIST * pnl = PadlistNAMES(pl);" in many places
and recompiling the interp with -O0, just to be able to inspect the
padnamelist struct.
|
| | |
|
| |
|
|
|
|
|
|
| |
I noticed AV *s in my C debugger didn't the sv_debug_* members, but if I
casted the AV * to a SV * in my C debugger using a expession in my watch
tab, all the info was there. Put the DEBUG_LEAKING_SCALARS in
_SV_HEAD_UNION so the sv_debug_* members appear everywhere (CV/HV/AV/etc),
not just for SVs.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit a24d8dfd08 "Make IPC::Open3 work without fork()" from 5.003 created
an eval block, and if that eval block threw an exception, instead of
propagating $@, the code propagated $!, even though no system call was done
and $! is effectivly unintialized data. In one case for me, a taint
exception inside system was turned into open3() throwing an exception
about "Inappropriate I/O control operation" or "Bad file descriptor", which
had nothing to do with the real fault which was a Perl C level croak with
the message "Insecure $ENV{PATH} while running with -T switch at ..."
which was called as Perl_pp_system->Perl_taint_env->Perl_taint_proper->
Perl_croak->Perl_vcroak. This patch does not try to fix the ambiguity of
the error messages between the !DO_SPAWN and IO::Pipe
branches/implementations of _open3.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
commit 50e5165b96 "stop T_IN/OUT/INOUT/STDIO typemaps leaking" changed
newRV to newRV_noinc, but the GV * returned by newGVgen() is owned by the
package tree, like the SV * returned by get_sv(). Now when the RV to GV is
freed on mortal stack, the GV * in the package tree is freed, and now there
is a freed GV * in the package tree, if you turn on "PERL_DESTRUCT_LEVEL=2"
(and perhaps DEBUGGING is needed too), the package tree is destroyed SV *
by SV *, and perl will eventually warn with
"Attempt to free unreferenced scalar" which a very bad panic type warning.
commit 50e5165b96 was reverted in commit bae466e878
"Revert "stop T_IN/OUT/INOUT/STDIO typemaps leaking" for 5.22's release
to stop the panic, but reintroduced the SV/RV leak. So fix the RV leak (the val
passed as source arg of sv_setsv) by freeing it after the copying. In a very
unlikely scenario, the RV could still leak if sv_setsv dies.
Also fix the problem, that if this OUTPUT: type is being used for an
incoming arg, not the outgoing RETVAL arg, you can't assign a new SV*
ontop of the old one, that only works for perl stack return args, so
replace "$arg = &PL_sv_undef;" with "sv_setsv($arg, &PL_sv_undef);" if its
not RETVAL, this way OUTPUT on incoming args also works if it goes down the
error path. For efficiency, in a RETVAL siutation, let the undef original
SV* in $arg which is typically obtained from sv_newmortal() by xsubpp pass
through if we error out.
Also for efficiency, if it is RETVAL (which is more common) dont do the
sv_setsv/SvREFCNT_dec_NN stuff (2 function calls), just mortalize
(1 function call) the ex-temp RV and arrange for the RV to wind up on
perl stack.
Also, the GV * already knows what HV * stash it belongs to, so avoid the
stash lookup done by gv_stashpv() and just use GvSTASH which are simple
pointer derefs.
|
