Craig A. Berry [Mon, 28 Oct 2013 02:43:55 +0000 (21:43 -0500)]
getnameinfo, getaddrinfo, and inet_xxx fix-ups for configure.com.
This is mostly a very belated follow-up to
d7fba2bf7c68c, but in
the case of inet_aton, is a stupendously belated follow-up to
570bb66edbdc8.
Craig A. Berry [Mon, 28 Oct 2013 02:31:13 +0000 (21:31 -0500)]
Don't make bool an int on VMS.
The special case has been there since
61bb59065bf1b12edab3, most
likely because the VMS C++ compiler, like a lot of other C++
compilers in the 1990s implemented a bool as an int, and making
the type in C compatible seemed like a good idea. But no C++
compiler that's likely to build Perl on VMS has a bool type that
occupies more than one byte now, so remove the special case. We're
unlikely to even see this code since we've had stdbool.h since
DEC C 6.4, released in 2001.
Chris 'BinGOs' Williams [Sun, 27 Oct 2013 22:06:43 +0000 (22:06 +0000)]
Update Unicode-Collate to CPAN version 1.00
[DELTA]
1.00 Sun Oct 27 13:22:17 2013
- When a subroutine by 'overrideOut' taking a out-of-range value and
returning undef, now the value is treated as if it were U+FFFD.
* 0.99 wrongly calculates implicit weights based on out-of-range values.
- Assertion using unpack 'U' is added. If not only pack('U') but also
unpack('U') of CORE:: don't work as expected, this module will die.
Chris 'BinGOs' Williams [Sat, 26 Oct 2013 20:44:25 +0000 (21:44 +0100)]
Update Term-Cap to CPAN version 1.15
[DELTA]
1.15 Sat Oct 26 21:28:29 BST 2013
- meta data patch from David Steinbrunner
- Change email in metadata
1.14 Sat Oct 26 19:20:03 BST 2013
- Update versions everywhere
- Change email address and add the github location
1.13 Sat Oct 26 19:08:27 BST 2013
- Bring everything up to date with all patches from RT
- Sync version with core
Father Chrysostomos [Sat, 26 Oct 2013 19:28:11 +0000 (12:28 -0700)]
Don’t let gv.c:gv_try_downgrade touch PL_statgv
PL_statgv remembers the handle last used by stat, for the sake of -T _
and -B _. If a glob is freed when PL_statgv points to it, PL_statgv
is set to null.
gv_try_downgrade exists to remove globs and subs that were (possibly
temporarily) vivified by bareword lookup. It is called whenever a
gvop is freed and its gv looks like a candidate for downgrading. That
means it applies, not only to potential sub calls, but also to *foo
and *bar. gv_try_downgrade may delete a glob from the stash alto-
gether if it is empty. So eval "*foo if 0" may delete the *foo glob.
If PL_statgv is pointing to *foo, then eval "*foo if 0" may change the
behaviour, which is not supposed to happen:
$ ./perl -Ilib -le 'stat *{"foo"}; -T _; print $!; -T _; print $!'
Bad file descriptor
Bad file descriptor
$ ./perl -Ilib -le 'stat *{"foo"}; -T _; print $!; eval "*foo if 0"; -T _; print $!'
Bad file descriptor
No such file or directory
Father Chrysostomos [Sat, 26 Oct 2013 13:00:59 +0000 (06:00 -0700)]
Make PL_DBgv and other db interp vars refcounted
$ PERL5DB=1 perl -de '*DB::DB = sub {} if 0; sub DB::DB{}'
No DB::DB routine defined at -e line 1.
When the ‘if’ gets folded and *DB::DB is freed, it triggers
gv_try_downgrade, causing $DB::{DB} to be deleted and PL_DBgv to point
to a freed SV. A higher refcount would prevent gv_try_downgrade from
touching it.
Father Chrysostomos [Sat, 26 Oct 2013 12:42:08 +0000 (05:42 -0700)]
perl.c: Avoid redundant gv_AVadd and GvAVn
We don’t need to call gv_AVadd if follewed by GvAVn, which itself
calls gv_AVadd when the AV is null.
Also, we don’t need to keep checking whether GvAV(PL_argvgv) is still
defined (via GvAVn with the final n) when pushing on to it.
This code goes back at least to perl 5 alpha 2.
Father Chrysostomos [Sat, 26 Oct 2013 12:40:10 +0000 (05:40 -0700)]
gv:gv_add_by_type: Don’t check @ISA when not vivifying
I added the code (in
986d39eeb4) to magicalise @ISA when vivifying it
via @{*ISA}. But we don’t need to do that every time gv_add_by_type
is called, just when it is autovivifying.
Matthew Horsfall [Thu, 3 Oct 2013 14:09:22 +0000 (10:09 -0400)]
Add op_other to B::Concise -debug output for LOGOPs
Father Chrysostomos [Fri, 25 Oct 2013 21:40:11 +0000 (14:40 -0700)]
Regen META.{yml,json}
(If t/porting/regen.t fails without this regeneration, why doesn’t
‘make regen’ do this?)
Father Chrysostomos [Fri, 25 Oct 2013 21:36:31 +0000 (14:36 -0700)]
Add Kevin Falcone to AUTHORS
Kevin Falcone [Tue, 22 Oct 2013 18:42:49 +0000 (14:42 -0400)]
No longer point to rt.perl.org/perlbug/ as a landing page
We'll use the homepage of rt.perl.org instead, as it's much more likely
to get text updates.
Kevin Falcone [Tue, 22 Oct 2013 18:39:16 +0000 (14:39 -0400)]
update RTSERVER
This didn't work properly as is (because of the redirect from
http://rt.perl.org/rt3/ to https://rt.perl.org/rt3/) but at least it'll
work if someone tries it going forward.
Kevin Falcone [Tue, 22 Oct 2013 18:38:53 +0000 (14:38 -0400)]
Allow newer bug URLs on rt.perl.org
Kevin Falcone [Tue, 22 Oct 2013 18:38:22 +0000 (14:38 -0400)]
Fixup rt.perl.org/rt3 => rt.perl.org
These links all would redirect properly, but give out the top URL.
Chris 'BinGOs' Williams [Fri, 25 Oct 2013 18:49:52 +0000 (19:49 +0100)]
Update AutoLoader to CPAN version 5.74
[DELTA]
5.74 Fri Oct 25 19:08 2013
- Makefile.PL: Set installdirs to "site" on Perl >= 5.12
Father Chrysostomos [Fri, 25 Oct 2013 13:26:16 +0000 (06:26 -0700)]
Make sure truncated in-memory files have trailing null
Just spotted this....
If the target scalar contains something already, then setting SvCUR to
0 is not sufficient.
Father Chrysostomos [Fri, 25 Oct 2013 13:18:55 +0000 (06:18 -0700)]
t/porting/dual-life.t: Another exception
In parallel testing, it might fail on files like
cpan/Module-Metadata/MB-9dzadSLU/Simple/bin/simple.plx.
Father Chrysostomos [Fri, 25 Oct 2013 13:16:02 +0000 (06:16 -0700)]
Increase PerlIO::scalar::VERSION to 0.18
Father Chrysostomos [Fri, 25 Oct 2013 13:15:30 +0000 (06:15 -0700)]
Better fix for #119529
<$fh> used to loop if $fh was opened to an in-memory handle containing
a reference.
Commit
552908b174 fixed the loop by forcing a reference to a string
when the handle was created. It did not take into account that the
reference might be read-only. It was also insufficient, in that
the target scalar could be set to a reference after the handle
was created.
The real reason it was looping was that the code for returning and
setting the size of the buffer was not handling non-PVs properly
(unless they were globs, which were special-cased). It might return
0, and it might not, depending on what the internal SV field hap-
pened to hold.
This caused looping under 5.12 and onwards, but even in 5.10 <$fh>
returned nothing.
In this case, deleting code makes things just work.
Reverting the hunk from
552908b174 stops appending to refs from work-
ing, so I tweaked PerlIOScalar_pushed to fix that (which also reduced
the amount of code).
Steve Hay [Wed, 23 Oct 2013 09:32:25 +0000 (10:32 +0100)]
Remove some unnecessary Makefile.PLs from cpan/ and ext/
These modules all work with the default Makefile.PL written by make_ext.pl.
(There are more Makefile.PLs in dist/ that could be removed too, but the
general policy with dist/ is to leave such files in place to make it easier
to roll CPAN releases from the blead source, which is canonical.)
Karl Williamson [Fri, 25 Oct 2013 04:35:51 +0000 (22:35 -0600)]
Silence some Win32 compiler warnings
These get rid of some "possible loss of data" warnings
Karl Williamson [Fri, 18 Oct 2013 02:33:30 +0000 (20:33 -0600)]
PATCH [perl #120314]: fold_grind.t spews tons of warnings
It turns out that fixing the compiler warnings about possible loss of data
(by adding casts) silences these warnings.
Father Chrysostomos [Thu, 24 Oct 2013 21:45:34 +0000 (14:45 -0700)]
Propagate lvalue context to && and ||
I did this for the last statement in an lvalue sub (and the argument
to ‘return’ in an lvalue sub) in commit
2ec7f6f2428.
It turns out that it needs to apply in other lvaluish contexts, too:
$ ./perl -Ilib -le 'for($a||$b){$_++} print $b'
1
$ ./perl -Ilib -le 'for(pos $a || pos $b){$_++} print pos $b'
Modification of a read-only value attempted at -e line 1.
If I can assign to $b through this construct, then why not pos?
Father Chrysostomos [Thu, 24 Oct 2013 19:53:21 +0000 (12:53 -0700)]
pp_hot.c:pp_rv2av: Remove superfluous SPAGAIN
Commit perl-5.8.0-9908-g5e17dd7, in March of 2007, added a SPAGAIN
here to account for the stack shifting when hv_scalar calls tied
hashes’ SCALAR method.
It was never necessary, because commit perl-5.8.0-3008-ga3bcc51,
which added hv_scalar and magic_scalarpack in December of 2003, made
magic_scalarpack push a new stack, protecting the old one.
Father Chrysostomos [Thu, 24 Oct 2013 19:46:52 +0000 (12:46 -0700)]
rv2hv does not use its TARG
rv2hv has had a TARG since perl 5.000, but it has not used it since
hv_scalar was added in perl-5.8.0-3008-ga3bcc51.
This commit removes it, saving a tiny bit of space in the pad.
Chris 'BinGOs' Williams [Thu, 24 Oct 2013 19:42:28 +0000 (20:42 +0100)]
Update Package-Constants to CPAN version 0.04
[DELTA]
0.04 Thu Oct 24 20:34:00 BST 2013
=================================================
* Update Makefile.PL to add repository et al
* Install into site if >= v5.12.0
* Add use deprecate due to core scheduled removal
Chris 'BinGOs' Williams [Thu, 24 Oct 2013 19:20:03 +0000 (20:20 +0100)]
Schedule Package::Constants for removal from core
Chris 'BinGOs' Williams [Thu, 24 Oct 2013 19:13:11 +0000 (20:13 +0100)]
Update Archive-Tar to CPAN version 1.96
[DELTA]
1.96 24/10/2013
- integrate Package::Constants into Constant module
and remove requirement on it.
1.94 24/10/2013
- install into site if >= 5.012
1.93_02 22/10/2013 (XLAT)
- [rt.cpan.org #78030] symlinks resolution on MSWin32
Father Chrysostomos [Thu, 24 Oct 2013 17:53:46 +0000 (10:53 -0700)]
Restore prev. behaviour of @a||... in lv sub
$ perl5.18.1 -lwe 'my @a; sub i:lvalue {@a||@b} @a=1; (i())=3'
Name "main::b" used only once: possible typo at -e line 1.
Useless assignment to a temporary at -e line 1.
Bleadperl:
$ ./perl -Ilib -lwe 'my @a; sub i:lvalue {@a||@b} @a=1; (i())=3'
Name "main::b" used only once: possible typo at -e line 1.
Can't return array to lvalue scalar context at -e line 1.
I accidentally changed it in commit
2ec7f6f242 by propagating the
lvalue context. This commit changes it back by only flagging the
rv2av op as being in an lvalue sub if it is not already flagged as
being in scalar context.
The old behaviour was inconsistent, and this commit does restore it
(see the tests), but resolving that discrepancy is for a future commit
(if I ever get to it).
In any case, ‘Can't return array to lvalue scalar context’ is wrong.
Father Chrysostomos [Thu, 24 Oct 2013 12:41:17 +0000 (05:41 -0700)]
Fix bare blocks in lvalue subs
If a bare block is the last thing in an lvalue sub, OP_LEAVELOOP needs
to propagate lvalue context and handle returned arrays properly, just
as OP_LEAVE has done since yesterday.
This is a follow-up to
2ec7f6f24289. This came up in ticket #119797.
Father Chrysostomos [Thu, 24 Oct 2013 12:32:17 +0000 (05:32 -0700)]
Make the details array in universal.c static
Otherwise we can have linking problems when a global ‘details’ symbol
is defined elsewhere. (That happened to me with an XS module.)
Father Chrysostomos [Thu, 24 Oct 2013 03:23:21 +0000 (20:23 -0700)]
Increase $B::Concise::VERSION to 0.991
I don’t know, is it time for version 1 yet?
Father Chrysostomos [Thu, 24 Oct 2013 03:21:04 +0000 (20:21 -0700)]
[perl #119797] Fix if/else in lvalue sub
When if/else/unless is the last thing in an lvalue sub, the lvalue
context is not always propagated properly and scope exit tries to
copy things, including arrays, resulting in ‘Bizarre copy of ARRAY’.
This commit fixes the bizarre copy by flagging any leave op that is
part of an lvalue sub’s return sequence, using the OPpLEAVE flag added
for this purpose in the previous commit. Then pp_leave uses that flag
to avoid copying return values, but protects them via the mortals
stack just like pp_leavesublv (actually pp_ctl.c:S_return_lvalues).
For ‘if’ and ‘unless’ without ‘else’, the lvalue context was not being
propagated, resulting in arrays’ getting flattened despite the lvalue
context. op_lvalue_flags in op.c needed to handle AND and OR ops,
which ‘if’ and ‘unless’ compile to, to make this work.
Father Chrysostomos [Thu, 24 Oct 2013 02:33:50 +0000 (19:33 -0700)]
Add OPpLVALUE flag
This flag will be used by the next commit.
Father Chrysostomos [Thu, 24 Oct 2013 00:59:54 +0000 (17:59 -0700)]
Increase $bignum::VERSION to 0.37
Father Chrysostomos [Thu, 24 Oct 2013 00:56:05 +0000 (17:56 -0700)]
Add Marc Simpson to AUTHORS
Marc Simpson [Wed, 23 Oct 2013 14:19:23 +0000 (15:19 +0100)]
Minor bignum POD fix (capitalisation).
David Mitchell [Wed, 23 Oct 2013 18:21:35 +0000 (19:21 +0100)]
bump Tie::StdHandle version
after recent fixes
David Mitchell [Wed, 23 Oct 2013 18:14:36 +0000 (19:14 +0100)]
Tie::StdHandle::WRITE: handle offset
the WRITE method was ignoring its offset argument.
Anno Siegel [Wed, 23 Oct 2013 17:06:45 +0000 (18:06 +0100)]
Tie::StdHandle appends extra copies of $\ to output
[perl #120202]
The following code demonstrates the problem:
use Tie::Handle;
my $out = do { no warnings 'once'; \ local *HANDLE };
tie *$out, 'Tie::StdHandle', '>&', \ *STDOUT or die;
$\ = "haha\n";
print $out "hihi\n";
which prints
hihi
haha
haha
The string in $\ has been added twice, once explicitly by
Tie::Handle::PRINT and another time implicitly by the use of
(CORE::) print in Tie::StdHandle::WRITE.
The bug also affects the use of say() with tied handles where a spurious
newline is added by the same effect.
[ test added by davem ]
David Mitchell [Wed, 23 Oct 2013 15:40:55 +0000 (16:40 +0100)]
stdhandle.t: unlink the right file.
It writes to "afile", but initially deleted "afile.new" if it already
existed.
David Mitchell [Wed, 23 Oct 2013 15:37:39 +0000 (16:37 +0100)]
stdhandle.t: convert to use Test::More
print "ok 13\n" is *so* 1990's.
David Mitchell [Wed, 23 Oct 2013 14:12:18 +0000 (15:12 +0100)]
Make B/Deparse handle unicode regexes
[perl #120182]
the precomp B::OP method was returning a non-UTF8 string even if the
regex was utf8
Daniel Dragan [Wed, 23 Oct 2013 10:36:45 +0000 (11:36 +0100)]
correct POD for SvIsCOW, it returns a U32
Daniel Dragan [Wed, 23 Oct 2013 10:28:15 +0000 (11:28 +0100)]
refactor sv_add_backref
-don't assign NULL twice to var mg, seen in machine code
-use sv_magicext to get MAGIC * instead of a 2nd call to mg_find, less
passes through the linked list is faster
-move the mg_flags MGf_REFCOUNTED statement to before av, to reduce scope
of liveness for var mg, it is never used again in the function
-av can not be null since it was just created, use
SvREFCNT_inc_simple_void_NN
-on a tsv, that is being upgraded to be an AV backref holder, call
av_extend only once with the correct slice count based on whether tsv was
a SV backref holder or not previously
-since on a newly AV backref tsv, the number of backrefs will only be 1 or
2, and if we made an AV type backref holder, doing the SV type
optimization is now impossible, so don't check av for NULL when we know
it wont be if we just made the AV, and don't do the bounds check and 2nd
av_extend either since the AV is new and under our control
-if tsv is already an AV type backref, the AV can have 1 to infinity
elements so do the bounds check and av_extend as before
It is intentional that on the new AV type holder branch, that after
"av_extend(av, *svp ? 2 : 1);" there are no further function calls in the
execution path until the return.
Steve Hay [Tue, 22 Oct 2013 17:05:17 +0000 (18:05 +0100)]
Add support for building with Visual C++ 2013
Two tests (t/io/fs.t and cpan/HTTP-Tiny/t/110_mirror.t) fail on my system,
currently in Daylight Saving Time, both due to times written/read by
utime()/stat() being off by one hour...
Not sure what the issue is yet, but I've reproduced it in this C program:
#include <stdio.h>
#include <sys/stat.h>
#include <sys/utime.h>
void main(void) {
struct _utimbuf ut;
struct _stat st;
time_t t;
t =
760060800;
printf("Setting: %ld\n", t);
ut.actime = t;
ut.modtime = t;
if (_utime("test.c", &ut) == -1) {
perror("_utime failed\n");
return;
}
if (_stat("test.c", &st) != 0) {
perror("_stat failed\n");
return;
}
printf(" atime: %ld\n", st.st_atime);
printf(" mtime: %ld\n", st.st_mtime);
}
which outputs
Setting:
760060800
atime:
760057200
mtime:
760057200
with VC++ 2013, instead of
Setting:
760060800
atime:
760060800
mtime:
760060800
like Visual C++ 6.0 through 2012 all do.
David Mitchell [Tue, 22 Oct 2013 15:48:16 +0000 (16:48 +0100)]
spurious "Use of "-x" without parentheses" mesg
RT #120288.
Something like '-X' at the start of a line (where X is one char, but not a
filetest operator) was giving a spurious
Warning: Use of "-a" without parentheses is ambiguous
error. This was because the toker was provisionally marking the -a as the
last encountered unary operator (PL_last_uni), then not unmarking it when
it found that that it wasn't in fact a filetest operator.
Father Chrysostomos [Tue, 22 Oct 2013 12:36:38 +0000 (05:36 -0700)]
Unbreak -bareword under strict+integer
Commit
077da62ff9 was not supposed to change behaviour, but only
remove logic rendered unnecessary two commits earlier in
1c2b3fd6f10.
But the special stricture exception for negation was in the same func-
tion (S_op_integerize) which applied it to OP_NEGATE, but now needed
to apply it to OP_I_NEGATE, too.
Steve Hay [Tue, 22 Oct 2013 07:26:19 +0000 (08:26 +0100)]
Daniel Dragan [Tue, 22 Oct 2013 01:59:32 +0000 (21:59 -0400)]
WinCE Perl EVC4 support
- CCTYPE has always been empty string in Makefile.ce, makedef.pl does not
complain about this, but it is a bug, so fix it
- create a profile for a EVC4 arm build
- put -GS- and -GL flags into compiler version specific defs so VC6 era
EVC4 will work
- don't delete ..\lib\Config.pm, it was made by win32/Makefile, for a
Desktop Perl, deleting this breaks miniperl when it loads its pure perl
modules for cross MakeMaker module building, and requires rebuilding the
Desktop Perl so miniperl works again. CE Perl has its Config.pm in the
xlib\$(MACHINE) dir
Steve Hay [Mon, 21 Oct 2013 17:31:34 +0000 (18:31 +0100)]
Porting/Maintainers.pl - Remove redundant EXCLUDED files
Also add extra verbose output to Porting/core-cpan-diff to identify such
cases.
Daniel Dragan [Sun, 20 Oct 2013 01:26:29 +0000 (21:26 -0400)]
.gitignore Win32's miniperl object files
Comments from the committer:
win32/mini/*.obj (and hence win32/mini/ since it contains nothing other than
*.obj) should be .gitignored by the top-level .gitignore file anyway, but
git clean -dfX doesn't remove them without this patch. (git clean -dfx does,
but some developers want to keep untracked files other than build products
explicitly listed by .gitignore files.)
Daniel Dragan [Wed, 9 Oct 2013 07:33:47 +0000 (03:33 -0400)]
WinCE Makefile and make_ext.pl general and XS fixes
On a WinCE build. On the 2nd nmake run, using Makefile.ce, eventually calls
the Extensions target which calls make_ext.pl. What happens is nmake for CE
for each module is called on the Desktop per module makefile from the
earlier Desktop build. Since the Desktop Perl already was built
sucessfully, all rules/deps are met in the Desktop per module makefile, and
nothing happens during the module building phase for a CE build. Previously
I used external file management tools to delete the per module Makefiles
before running Makefile.ce.
*make_ext.pl
- implement deleting and rebuilding the per module makefile on a Cross
build
- use constants for constant folding, there are opportunities for other
variables to be converted to constants in the future
- fix a bug from commit
baff067e71 where unlink() on a file with an open
handle ($mfh) didn't delete the file from disk and a new per module
makefile would be not be built by make_ext.pl later since the per module
makefile was still on disk. This was observed on Win32. Also harden the
unlink code with a new _unlink sub that is fatal if the file is still on
disk after unlink supposedly deleted it.
- var $header and the quotemeta is because of an issue in Perl #119793
*Makefile.ce
- bring the debugging symbol generation flags and optimization flags
to be closer to a Dekstop VC Perl build
- ICWD is obsolete as of commit
f6b3c354c9 , remove it
- MINIMOD is obsolete as of commit
7b4d95f74b , remove it
- make a poisoned config.h so if there is a XS building mixup between
a desktop and CE perl, the poisoned config.h for CE will stop the build
gracefully
- $(MINIPERL) has never been defined in Makefile.ce from day 1 (10 years)
replace with $(HPERL) everywhere, this was causing things to not run
silently since $(MINIPERL) was empty string. Use HPERL instead of
MINIPERL to allow flexibility to use the full perl binary if necessery
one day
- better cleaning on root makefile clean target
*win32/win32.h
*win32/win32iop.h
- silence alot of redefinition warnings which gave pages of warnings on
each WinCE compliand
*mg.c
- win32_get_errno is only on WIN32 build not WINCE
a "nmake -f Makefile.ce all" will now build the CE interp and all modules
in 1 shot with no user intervention
Daniel Dragan [Mon, 21 Oct 2013 14:43:06 +0000 (15:43 +0100)]
make sv_2bool_flags() non-recursive on overload
When Perl_sv_2bool_flags() has an overloaded arg, it calls SvTRUE()
on the SV returned from the overload method. This indirectly calls
sv_2bool_flags() again.
Change it so that sv_2bool_flags() just iterates the new overload value
each time.
2 callsites were converted to gotos. A SvTRUE_common was expanded so goto
can be used. This function's machine code size on VC2003 32 bits dropped
by 0x24 bytes after this patch.
David Mitchell [Mon, 21 Oct 2013 12:45:36 +0000 (13:45 +0100)]
Benchmark.t: avoid ultra-lightweight code
Some tests benchmarked the code '$i++', which is so lightweight that it
could trigger the die "Timing is consistently zero in estimation loop"
in Benchmark.pm.
So make the code slightly more heavyweight.
David Mitchell [Mon, 21 Oct 2013 12:07:19 +0000 (13:07 +0100)]
Benchmark.pm: bail out earlier on zero delta
countit(), which runs code for $n seconds, performs an initial
set of calibration loops to find out roughly how many iterations
are required to burn 0.1 CPU secs more in the code loop than in the empty
loop (which is then used to guesstimate how many iterations to do for the
main loop).
This calibration is designed to bail out if we consistently get <= 0
difference between the empty loop and the code loop; however, this
bailout can take too long: it looks for a run of 17 *consecutive* zero
deltas during 2048, 4096,8192,... iterations, which means it won't fail
until approx 131 million empty + code iterations are run, even if
the zeroes are consecutive. Of course if just by random noise one delta
is infinitesimally > 0, then the count is reset and the exponential count
continues.
The net effect of this is that the calibration loop can loop forever on
'small' code.
This commit makes it additionally bail out of the calibration loop if
running the iterations for a particular value of N takes more than 8
CPU seconds, while still giving zero delta.
Chris 'BinGOs' Williams [Sun, 20 Oct 2013 17:41:58 +0000 (18:41 +0100)]
Skip 'BAD LOCALE' test on DragonflyBSD
Steve Hay [Sun, 20 Oct 2013 15:43:59 +0000 (16:43 +0100)]
Prepare Module::CoreList for Perl 5.19.6
Steve Hay [Sun, 20 Oct 2013 15:39:13 +0000 (16:39 +0100)]
Module::CoreList 3.00 is now on CPAN
Steve Hay [Sun, 20 Oct 2013 14:42:47 +0000 (15:42 +0100)]
Bump version for Perl 5.19.6
Steve Hay [Sun, 20 Oct 2013 14:40:05 +0000 (15:40 +0100)]
Create new perldelta for Perl 5.19.6
Steve Hay [Sun, 20 Oct 2013 14:39:39 +0000 (15:39 +0100)]
Add epigraph for Perl 5.19.5
Steve Hay [Sun, 20 Oct 2013 10:08:27 +0000 (11:08 +0100)]
Update perlhist for Perl 5.19.5
Steve Hay [Sun, 20 Oct 2013 10:06:10 +0000 (11:06 +0100)]
Finalize perldelta for Perl 5.19.5
Steve Hay [Sun, 20 Oct 2013 10:00:16 +0000 (11:00 +0100)]
Update Module::CoreList for Perl 5.19.5
James E Keenan [Sun, 13 Oct 2013 17:27:23 +0000 (19:27 +0200)]
Add descriptions for tests in dist/autouse/t/autouse.t lacking them.
Steve Hay [Sat, 19 Oct 2013 22:24:15 +0000 (23:24 +0100)]
perldelta - Correct sort order of upgraded modules
Steve Hay [Sat, 19 Oct 2013 20:11:32 +0000 (21:11 +0100)]
Porting/Maintainers.pl - Move B and B::Concise to _PERLLIB
Now that B::Concise has no special MAINTAINER of its own (following commit
b3dcf77564) there is no reason to have it separated from the rest of ext/B
and hence no reason for that not to move to _PERLLIB along with every other
ext/ module.
Steve Hay [Sat, 19 Oct 2013 19:09:06 +0000 (20:09 +0100)]
Note IPC-Cmd upgrade from commit
6d1e091186
Steve Hay [Sat, 19 Oct 2013 19:08:08 +0000 (20:08 +0100)]
Move B-Deparse from dist/ to lib/ since it is non-dual-lived and pure-Perl
Steve Hay [Sat, 19 Oct 2013 16:02:49 +0000 (17:02 +0100)]
Porting/Maintainers.pl - Remove the remaining MAINTAINER fields
This changes all six modules concerned (B::Concise, B::Deparse, VMS::DCLsym,
VMS::Stdio, warnings and Win32CORE) to P5P maintainership, but that is the
case in practice anyway (since they all live in ext/ or lib/, apart from
B::Deparse, which is surely in error in being in dist/ since it has no CPAN
dual-life distribution).
This was discussed on p5p and #onionsketch:
http://www.nntp.perl.org/group/perl.perl5.porters/2013/06/msg203768.html
http://www.nntp.perl.org/group/perl.perl5.porters/2013/10/msg208704.html
http://irclog.perlgeek.de/onionsketch/2013-10-17
Steve Hay [Sat, 19 Oct 2013 15:14:49 +0000 (16:14 +0100)]
perldelta - Remove all but one XXX notices
Steve Hay [Sat, 19 Oct 2013 15:08:16 +0000 (16:08 +0100)]
perldelta - Corrections and updates
Chris 'BinGOs' Williams [Sat, 19 Oct 2013 15:05:35 +0000 (16:05 +0100)]
Fix IPC-Cmd tests for Solaris (and potentially other SVR* variants)
Steve Hay [Sat, 19 Oct 2013 14:11:39 +0000 (15:11 +0100)]
Upgrade Scalar-List-Utils from version 1.34 to 1.35
(None of the files listed as EXCLUDED are actually in the CPAN distribution
any more anyway, so remove them from Porting/Maintainers.pl.)
Steve Hay [Fri, 18 Oct 2013 22:00:31 +0000 (23:00 +0100)]
perldelta updates for recent changes
Steve Hay [Fri, 18 Oct 2013 21:59:54 +0000 (22:59 +0100)]
Remove some IGNORABLE files from ext/ and lib/
There is surely no point in having any of these modules' MANIFEST files in
the core distribution, but other IGNORABLE files should generally stay in
ext/ and lib/ in case they ever get dual-lived (and likewise for modules in
dist/, which are already dual-lived and some of which include even MANIFEST
files, presumably used to roll CPAN releases from).
However, XS-APItest's README file and DBM_Filter's Changes file offer
nothing useful, so remove them.
Steve Hay [Fri, 18 Oct 2013 17:40:26 +0000 (18:40 +0100)]
Add the rt.cpan.org ticket number for the Digest::MD5 customization
Steve Hay [Fri, 18 Oct 2013 17:20:31 +0000 (18:20 +0100)]
Remove further IGNORABLE-like files under cpan/
These are all examples, demos and oddly named READMEs.
Steve Hay [Fri, 18 Oct 2013 14:10:31 +0000 (15:10 +0100)]
Remove all IGNORABLE files under cpan/
This was discussed on p5p:
http://www.nntp.perl.org/group/perl.perl5.porters/2013/10/msg208635.html
One LICENSE file (Locale-Codes) has been removed (contrary to one suggestion
in that thread) since it was the only one, and only says "This module is
free software; you can redistribute it and/or modify it under the same
terms as Perl itself." anyway.
Two files that I had previously suggested retaining for the sake of tests
have also been removed. In one case, a test script has been modified (in
Digest-MD5; I will send a ticket upstream); in the other case (in
Test-Harness) I have removed the test which depended on the MANIFEST files,
as suggested in the above thread.
David Mitchell [Fri, 18 Oct 2013 16:00:29 +0000 (17:00 +0100)]
add comment to foreach toking code
to make it clear what its doing.
Hio [Fri, 18 Oct 2013 15:48:47 +0000 (16:48 +0100)]
Correctly parse class name in 'for my class $foo
The code in toke.c on encountering 'for my' or for our',
skips past any optional package name, then checks that the next thing is a
'$', so that it can give a "Missing $ on loop variable" error if
necessary.
However, the code to skip the package name was using scan_ident() rather
than scan_word(); the latter assumes that the first char was the sigil,
and starts scanning from the second char onwards. So in something like
for my a1b $x (...)
scan_ident() starts scanning at '1b' in 'a1b', thinks it scanning $1 or
similar, and stops at the first non-digit char, in this case the b.
The fix is to use parse_word() instead.
Eric Brine [Fri, 18 Oct 2013 15:13:48 +0000 (16:13 +0100)]
Call select's 4th arg's magic
[perl #120102] CORE::select ignoring timeout var's magic
Patch by Eric, with tested added by davem.
Philip Guenther [Fri, 18 Oct 2013 14:38:55 +0000 (15:38 +0100)]
fix off-by one error in a2p
The str_gets() function, when encountering a newline character, checked to
see if the previous char was a \ escape. For a blank line, the check would
read the char at the position one before the start of the buffer. There
was a test to avoid this, but it was off-by-one.
Steve Hay [Fri, 18 Oct 2013 07:53:37 +0000 (08:53 +0100)]
Porting/Maintainers.pl - Remove most UPSTREAM flags and generate from FILES
With only one exception (version, and only until an outstanding rt.cpan.org
ticket is resolved), UPSTREAM is now 'cpan' for all FILES in cpan/, and
'blead' for everything else, so there is no point in explicitly stating
this: remove them all and auto-generate the same information instead.
The was discussed on p5p in the following thread:
http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2013-10/msg00321.html
Steve Hay [Fri, 18 Oct 2013 07:11:33 +0000 (08:11 +0100)]
Remove s2p from Porting/Maintainers.pl
As per RJBS's instructions in the onionsketch:
http://irclog.perlgeek.de/onionsketch/2013-10-17
Steve Hay [Fri, 18 Oct 2013 00:35:11 +0000 (01:35 +0100)]
Regen t/porting/customized.dat
Steve Hay [Fri, 18 Oct 2013 00:34:47 +0000 (01:34 +0100)]
Upgrade Test-Simple from version 0.98 to 0.99
This incorporates the CUSTOMIZED files t/fail-more.t, lib/Test/Builder.pm,
lib/Test/Builder/Module.pm, lib/Test/More.pm and lib/Test/Simple.pm.
Jesse Luehrs [Thu, 17 Oct 2013 20:24:58 +0000 (16:24 -0400)]
isa should fall back to checking @UNIVERSAL::ISA in all cases
Jesse Luehrs [Thu, 17 Oct 2013 19:43:54 +0000 (15:43 -0400)]
fix compiler warning
Steve Hay [Thu, 17 Oct 2013 17:13:44 +0000 (18:13 +0100)]
Steve Hay [Thu, 17 Oct 2013 17:05:40 +0000 (18:05 +0100)]
Skip two test modules from the cmpVERSION.pl check
Steve Hay [Thu, 17 Oct 2013 16:50:06 +0000 (17:50 +0100)]
Porting/Maintainers.pl - Move most of ext/ and lib/ into _PERLLIB
Most ext/ and lib/ entries in %Modules only list a single folder or a couple
of files sharing the same basename and specify no more information other
than UPSTREAM => 'blead'. This information can be more succinctly conveyed
by a line in _PERLLIB.
Only the following ext/ and lib/ entries in %Modules now remain:
B (has EXCLUDED files)
B::Concise (has a MAINTAINER)
VMS::DCLsym (has a MAINTAINER)
VMS::Stdio (has a MAINTAINER)
Win32CORE (has a MAINTAINER)
warnings (has a MAINTAINER and files from various places)
The middle four can probably move too if it is deemed that the MAINTAINER
information is not useful.
David Mitchell [Thu, 17 Oct 2013 14:35:14 +0000 (15:35 +0100)]
PerlIO::scalar: stringify refs
If $s in
open my $fh, "<", \$s
or similar is a ref, then
stringify that ref: i.e. convert it from a ref into the string
"SCALAR(0x....)" or whatever.
This fixes
RT #119529 Filehandle opened from ref to ref hangs on reading
which in this case was looping forever, since it kept thinking that
the var was a string ("SCALAR.."), but whose length was 0.
I haven't gone for a complete "always force var into a string" approach,
since PerlIO::scalar has quite a tolerance for "bad" vars; e.g.
it won't warn if $var is undef or a read-only constant number etc;
and it already normalises under some circumstances and not others.
So I've just increased the cases somewhat where it normalises.
Also, I didn't look to closely at the code that was looping (or to put it
another way, I looked at it but didn't understand it), so it could
conceivably still behave badly on some other strange type of variable that
manages to avoid getting normalised.
Steve Hay [Thu, 17 Oct 2013 13:27:08 +0000 (14:27 +0100)]
Porting/Maintainers.pl - Remove unnecessary Maintainers following
0cd7b4f8d7
Steve Hay [Thu, 17 Oct 2013 13:22:13 +0000 (14:22 +0100)]
Porting/Maintainers.pl - Remove remaining UPSTREAM => undef entries
There is some agreement that the things left with UPSTREAM => undef (and
the only two other pod/ files that were tracked -- perldtrace and
perlreftut) are not worth tracking in this file anyway, so just remove
them.
I've left s2p and made UPSTREAM => 'blead', as per comments in the first
thread cited below, for now. Maybe it should go too? Other than
regen/warnings.pl there are no other files covered thaty aren't under
cpan/, dist/, ext/ and lib/ ...
http://www.nntp.perl.org/group/perl.perl5.porters/2013/06/msg203768.html
http://www.nntp.perl.org/group/perl.perl5.porters/2013/10/msg208670.html
Steve Hay [Thu, 17 Oct 2013 12:51:37 +0000 (13:51 +0100)]
perldelta updates for recent changes
Steve Hay [Thu, 17 Oct 2013 12:51:21 +0000 (13:51 +0100)]
Upgrade JSON::PP from version 2.27202_01 to 2.27203
Commit
be08498a8d updated Maintainers.pl and perldelta.pod, but missed the
JSON/PP.pm file itself.
Steve Hay [Thu, 17 Oct 2013 12:29:28 +0000 (13:29 +0100)]
Fix Win32 build with MinGW/gcc-4.8+ which define some errno.h values >= 100
Commit
ea95436966 made changes to how errno.h constants are handled for
VC++ 2010 and above, which have added new values in the range 100-140.
Some versions of gcc-4.8.0 and above are now catching up and provide some
of these new values too (e.g. binaries from http://mingw-w64.sourceforge.net/
but not currently those from http://www.mingw.org/), but they don't provide
all of them. EADDRINUSE is provided, so convert_errno_to_wsa_error() gets
included, but the compilation fails because the following #defines which
VC++ 2010 and above provide are missing:
EBADMSG
EIDRM
ENODATA
ENOLINK
ENOMSG
ENOSR
ENOSTR
ENOTRECOVERABLE
EOTHER
ETIME
ETXTBSY
Simply ignore (#ifdef away) these constants for those compilers that don't
provide them.