In perl.git, the annotated tag v5.21.4 has been created
<http://perl5.git.perl.org/perl.git/commitdiff/29c9523a87e5e7fb44278288b7662e0318b71f8b?hp=0000000000000000000000000000000000000000>
at 29c9523a87e5e7fb44278288b7662e0318b71f8b (tag)
tagging 84c122ba1d2b7b856521ceffc2a095716a739335 (commit)
replaces v5.21.3
tagged by Steve Hay
on Sat Sep 20 11:09:51 2014 +0100
- Log -----------------------------------------------------------------
Perl 5.21.4
Alberto Simões (1):
Update ExtUtils::CBuilder to version 0.280219. Correct one test.
Alexandre (Midnite) Jousset (2):
magic.t: android: bypass two $0 tests and add one
perl.c: fix small bug on Android from commit 9054c81
Andy Dougherty (1):
Correct usage of memEQs in attributes.xs [perl #122701]
Anthony Heading (1):
stat Makefile.PL to get values for utime.
Brian Fraser (3):
Configure: Improve the d_prctl_set_name detection for Android
Android hints: define d_procselfexe and procselfexe if needed
Android hints: Always append /system/lib and /vendor/lib in libpth
Chris 'BinGOs' Williams (21):
Module-CoreList CPAN version synchronisation
Update experimental to CPAN version 0.010
Update Test-Harness to CPAN version 3.33
Update Socket version in Maintainers.pl
Update Pod-Perldoc to CPAN version 3.24
Update ExtUtils-CBuilder version for CPAN release
Update HTTP-Tiny to CPAN version 0.048
Update ExtUtils-Manifest to CPAN version 1.66
Update Time-Piece to CPAN version 1.29
Update HTTP-Tiny to CPAN version 0.049
Update Test-Simple to CPAN version 1.001006
Update Scalar-List-Utils to CPAN version 1.40
Update IO-Socket-IP to CPAN version 0.32
Bump Module-CoreList versions to new date stamped version
Bump Module-CoreList versions again
Remove the corevers.t from Module-CoreList and amend corelist porting test
Changed Module-CoreList version to how it should be
Synchronise Maintainers.pl for Module-CoreList
Update ExtUtils-Install to CPAN version 2.04
Update ExtUtils-Manifest to CPAN version 1.68
Document customisation introduced by 991b13da7
Craig A. Berry (8):
VMS has the isfinite macro on all architectures.
Include fp.h with math.h on VMS.
fix fp_classify in perl.h for VMS.
Some missing c99_xxx functions for POSIX.xs on VMS.
isnormal macro in fp.h broken on VMS.
The VMS CRTL has lrint, but not llrint.
Disable isfinite on VMS.
Disable libquadmath in configure.com.
Daniel Dragan (4):
more NAN FP fixs for VC6/NAN_COMPARE_BROKEN compilers
in Perl_sv_copypv_flags move mg_get call to callee
factor out TAINT_NOT into S_adjust_stack_on_leave
remove duplicate SvNV calls in pp_enteriter
David Mitchell (22):
dump.c: eliminate DUMP_OP_FLAGS/PRIVATE macros
Revert "cpan/Time-HiRes/t/itimer.t: better diagnostics"
embed.fnc: fix comment
fix 'make' header dependency generation
better document OA_ flags
Automate processing of op_private flags
mask VMS hints bits in COPs
Assert valid op_private bits in op_free()
add missing op_private labels to Concise etc
[MERGE] add regen/op_private and B::Op_private
add $VERSION to B::Op_private
perldelta: mention the op_private work
fix B::PADOP->sv and ->gv
B.xs: add class names to method branch comments
remove vestigal threadsv/threadsv_named
comment pp_foo aliases in pp*.c
regen/op_private: fix assorted typos
pp.h: remove spurious comment about OPpTARGET_MY
regen/op_private: update TARGLEX comments.
move OPpTARGET_MY comments from opcode.pl
regen/op_private: tidy up rv2cv entry
B.xs: fix syntax err in PERL_VERSION < 17
Doug Bell (2):
sig() should be kill()
add missing close paren in pod
Father Chrysostomos (331):
perldiag grammar tweaks
Partial minitest fix-up
Increase $_charnames::VERSION to 1.42
toke.c: Donât set PL_expect when emitting USE tokens
Set PL_expect less often when parsing semicolons
Expand a comment in toke.c (eliminate âwhy?â)
toke.c:S_lop: Donât set PL_expect if PL_nexttoke
toke.c: Rework LOOPX macro
toke.c: Skip PL_expect assignment under KEY_require
toke.c: Correct S_ao description
[perl #80368] Fix implicit assignop in qq"a\U="
toke.c: Touch PL_expect less for implicit [.,] in quotes
toke.c: For plugins, donât set PL_expect if PL_nexttoke
Stop using the value of PL_expect
toke.c: Consolidate some PL_expect assignments
Stop setting PL_lex_expect
toke.c: Remove PL_lex_expect define
parser.h: Comment that lex_expect is unused
Set PL_expect only once after curly subscripts
In perly.y, change PL_parser to parser
[Merge] Eliminate PL_lex_expect; fix one bug; other clean-up
Remove compile-time checking of rv2?v with const kid
Stop ck_rvconst from treating GV constants as strings
toke.c: Remove unnecessary condition
Stop trying to disambiguate {} after $
perldiag: reword and rewrap an entry
perlfunc: consistent spaces after dots
toke.c: Remove unnecessary PL_expect assignment
Allow package name in âuse constantâ constants
Increase $constant::VERSION to 1.32
constant.pm: outdated comment
constant.pm: Remove redundant truth checks
constant.pm: Remove doc reference to compile-time checking
constant.pm: Document fully-qualified constant names
Typo in pad.c apidocs
[perl #122556] Make undef $s free refs forthwith
Fix crash when lex subs are used for AUTOLOAD
Fix crash when lex subs are used for overload
Fix crash in leave_scope when my sub has CvGV
Fix crash with lex subs under -d
[Merge] Fix crashes with lexical subs
t/porting/readme.t: Get list from MANIFEST
perldiag grammar
Alphabetise perldiag
pad.c:pad_reset: check the name
pad.c:pad_reset: Also make sure the thing is not PADMY
pad.c: More clearly separate targ/constant allocation
op.c:ck_rvconst: Allocate GV pad slots like constants
pad.c:pad_free: Donât turn off the PADTMP flag
op.c:newPADOP: Allocate GV pad slots like constants
uniprops.t: Correct command line in comment
intrpvar.h: Correct comment
Split PL_padix into two variables
pad.c: Use PadnamelistMAXNAMED in another place
pad.c:pad_findmy_pvn: Skip âourâ hack for subs
pad.c:pad_free: Donât reset PL_padix if using pad_reset
pad.c: Remove BROKEN from USE_BROKEN_PAD_RESET
Fix up Peek.t
pad.c: pad_reset comments
Make Peek.t easier to run directly
[Merge] pad_reset
Avoid vivifying stuff when looking up barewords
Update B::Concise for OPpMAY_RETURN_CONSTANT
Increase $B::Concise::VERSION to 0.993
Stop turning off bit 1 in rv2cv private flags
Fix up Peek.t after priv flag twiddling
Porting/bisect.pl: Add missing full stop
Stop anon sub compilation from clobbering __ANON__
Stop prototype declaration from clobbering constants
perldelta for #80368/e4916dd1b3
perldelta for eff754733a9
perldelta for acb187b49 (part of #69456)
perldelta for 8380b6901
perldelta: A to-do item
perldelta for 4dda930be/#122556
perldelta for f20d0df0f
perldelta for 211a4342
perldelta for 8b57a5fa
perldelta for 172a6a76e8c
perldelta for 59e6df9f1cf
perlsec: Consistent spaces after dots
perlsec: Typo
perlsec: Another typo
Get â./miniperl -Ilib -Tâ working
taint.t: Fix skip count
Let taint.t run under miniperl
[perl #122669] Donât taint at compile time
perl.c: Improve buildcustomize error message
Fix assertion failure with our subs and constants
Revert "sv_grow: performance improvement for short strings"
Fix thinko in toke.c in 211a434
Add test for \p{} failing silently
Get chdir.t working under minitest
Get closure.t working under minitest
Get eval.t working under minitest
Get lc.t working under minitest
Get sub.t working under minitest
Get taint.t working under minitest
Get tr.t working under minitest
Get pat_re_eval.t working under minitest
Get reg_eval_scope.t working under minitest
Get rxcode.t working under minitest
Get t/uni/cache.t working under minitest
Get t/uni/class.t working under minitest
Get t/uni/fold.t working under minitest
Get t/uni/gv.t working under minitest
Get t/uni/labels.t working under minitest
Get t/uni/lower.t working under minitest
Get t/uni/title.t working under minitest
Get t/uni/upper.t working under minitest
Get t/comp/hints.t working under miniperl
Get t/io/fs.t working under miniperl
Get t/io/open.t working under miniperl
Get avhv.t working under miniperl
Get blocks.t working under miniperl
Really get chdir.t working under miniperl
Get coreamp.t working under miniperl
Get filetest.t working under miniperl
Get gv.t working under miniperl
Get hash.t working under miniperl
Get inccode.t working under miniperl
Get index.t working under miniperl
Get kvaslice.t working under miniperl
Get kvhslice.t working under miniperl
Skip lc.t under miniperl unless uni tables exist
Get length.t working under miniperl
Get lexsub.t working under miniperl
Get method.t working under miniperl
Get mkdir.t working under minitest
Get my_stash.t working under minitest
Get override.t working under minitest
Get pack.t working under minitest
Get packagev.t working under minitest
Get push.t working under minitest
Get readline.t working under minitest
Get reverse.t working under minitest
Get smartkve.t working under minitest
Get smartmatch.t working under minitest
Get sort.t working under minitest
Get split.t working under minitest
Get sprintf.t working under minitest
Get substr.t working under minitest
Get tie.t working under minitest
Get tie_fetch_count.t working under minitest
Get tiehandle.t working under minitest
Get universal.t working under minitest
Get utftaint.t working under minitest
Get vec.t working under minitest
Get warn.t working under minitest
Get write.t working under minitest
Get t/re/charset.t working under minitest
Skip t/re/pat.t under miniperl unless uni tables exist
Skip t/re/reg_mesg.t under miniperl unless uni tables exist
Get t/re/subst.t working under minitest
Skip t/re/regex_sets.t under miniperl unless uni tables exist
Skip t/re/regexp.t under miniperl unless uni tables exist
Skip uniprops.t under miniperl; add error checking
Get t/run/fresh_perl.t working under minitest
Get t/run/switches.t working under minitest
Really get t/uni/cache.t working under miniperl
Really get t/uni/class.t working under miniperl
Really get t/uni/fold.t working under miniperl
Really get t/uni/gv.t working under miniperl
Really get t/uni/labels.t working under miniperl
Put miniperl @INC logic in test.pl
skip_all_without_unicode_tables
Move @INC setup to t/uni/case.pl
Really get t/uni/ case tests working under miniperl
Get t/uni/method.t working under minitest
t/uni/opcroak.t: Die if eval fails
Get t/uni/opcroak.t working under minitest
Skip t/uni/parser.t in absence of Unicode tables
Get t/uni/readline.t working under minitest
Get t/uni/stash.t working under minitest
Get t/uni/universal.t working under minitest
Skip t/uni/variables.t in absence of Unicode tables
[Merge] minitest fixes
Fix refcounting in rv2gv when it calls newGVgen
Fix t/op/taint.t on Windows
toke.c: Stop using len to indicate trailing ::
MAD leftovers in toke.c
Donât call intuit_method twice for the same barewords
Avoid duplicate GV lookup for barewords
toke.c: Combine two identical chunks via goto
Fix our-sub method confusion
toke.c: Avoid extra sv_setpv for foo <newline> =>
Increase $bigrat::VERSION to 0.37
Increase $bignum::VERSION to 0.38
Increase $bigint::VERSION to 0.37
Remove -foo ambiguity warning
Fix my constant $var utf8 confusion
lex_utf8.t can run under miniperl
op.h: Correct PERL_LOADMOD_IMPORT_OPS comment
t/uni/parser.t: Make a test more robust
make_ext.pl: Typo in error message
mro.c: Donât inline sv_sethek
Remove GV_ADDINEVAL
perldelta for 64ff300b/#122669
perldelta for 7bdb4ff09
perldelta for a14c24d0
perldelta for 7950e9cf3
perldelta: Move a bug fix
perldelta: Move compile-time const deref checking
perldelta for 7950e9cf39
Precompute hash for require() arg
op.c: Calculate hash for CvNAME_HEK
Get mkdir.t working under miniperl *again*
Stop barewords from trumping subs with (*) proto
[perl #122728] Make sub(){undef} inlinable
Skip warnable const folding outside warnings scope
Remove redundant check from overload.pm
Increase $overload::VERSION to 1.23
op.c: ck_rvconst: Remove âdoubleâ gv lookup
IO.xs: OP_SIBLING is too new
Increase $IO::VERSION to 1.34
Wrap long lines in B::Op_private pod
Note where B::Op_privateâs pod is to be edited
Update rv2cv flags comment following 211a4342c9
perldiag: Rewrap for better splain output
Wrap long pod lines in B::Op_private for real
Gut Perl_save_re_context
Donât call save_re_context
Remove obsolete comment from utf8.c
utf8.c: Move an #ifndef for clarity
pp_ctl.c: Remove PL_curcop assignment
Mathomise save_re_context
pp_ctl.c: Remove junk from #endif
Deparse our(LIST) correctly
Deparse typed vars
Deparse.pm:pp_list: Donât call $lop->name repeatedly
Increase $B::Deparse::VERSION to 1.28
concise-xs.t is overly chummy with B::Deparse
toke.c: Remove redundant PL_expect check
Obsolete comment in sv.c:utf8_mg_pos_cache_update
perldelta for cfc7ef1
perldelta for 1542cb4/#122728
perldelta for c4c61c6
perldelta for 2c1f00b/#122747
perldelta for b4fa55d and 0ddd4a5
perldelta for 6678689 and 56cd2ef
Consistent spaces after dots in perlsub
Clone lex sub names properly
perldelta for d205f0c
Add CVf_LEXICAL flag
Turn on CVf_LEXICAL for lexical subs
cv_name
sv_cathek
pad.c: Document cv_name
Test cv_name
Increase $XS::APItest::VERSION to 0.64
For lexical subs, reify CvGV from CvSTASH and CvNAME_HEK
Remove bogus gv-handling code from toke.c
Avoid creating GVs when subs are declared
sv.h: Expand comment about potential SVf_UTF8 conflict
Fix-ups for Attribute::Handlers
Tweak Peek.t
Increase $Attribute::Handlers::VERSION to 0.97
op.c: ck_rvconst: Inline the noexpand var
Avoid reifying GVs in rv2cv
pad.c:cv_name: Reword docs for future extensibility
pad.c: Avoid struct name followed by colon
Add cv_set_call_checker_flags
Use cv_name in pp_hot.c:sub_crush_depth
op.c:ck_subr: reify GVs based on call checker
Teach dump.c about CVf_NAMED
Teach dump.c about CVf_LEXICAL
Inline op.c:too_few_arguments_sv into its only caller
Inline op.c:too_many_arguments_sv into its only caller
Teach ck_entersub_args_proto about non-GV names
Remove no-longer-used op.c:S_gv_ename
Add safename() func to B
Increase $B::VERSION to 1.51
Teach Deparse about coderefs in stashes
dump.c: Missing comma from CVf_HASEVAL output
Tweak Peek.t again
concise-xs.t needs to know about B::safename
[Merge] CVs without GVs
universal.c:croak_xs_usage: Donât fetch context twice
Term::ReadLine: Check $^O before -e
Increase $Term::ReadLine::VERSION to 1.15
perldelta: Insert missing =item
perldelta for f9d9e965
dump.c: Donât dump GvFLAGS as part of GP
Peek.t needs to know about the prev. change
Fix assertion failure with undef &my_sub/&anon
Avoid reifying GV when lex stub is used as method
Stop undef &foo from temporarily anonymising
Avoid reifying GVs when lex subs are used for overload
perldelta for 7fcb4126
Fix compiler warnings
Avoid reifing GVs in gv.c:gv_try_downgrade
Make âSIG handler not definedâ UTF8- and null-safe
mg.c: Avoid reifying GV for undefined sig handler
Exclude lex sub package name from (caller $n)[3]
toke.c: Remove obsolete comment
lexsub.t: To-do tests for sort lex_sub
Make âNo comma allowedâ respect lex subs
Make sort bareword respect lexical subs
Fix $DB::sub in DB::goto for lexical subs
perldelta for a5f477419
perldelta for 2872f9187
op.c: new is a keyword in C++
pad.c:pad_alloc: Remove redundant IS_PADGV
pad.c:pad_tidy: Remove redundant IS_PADGV
pad.c:pad_tidy: Remove IS_PADCONST and another IS_PADGV
pad.c:cv_clone_pad: Donât bother with IS_PADGV
pad.c:pad_push: Donât bother with IS_PADGV
pad.c:padlist_clone: Donât bother with IS_PADGV
Change IS_PADGV to isGV in op.c:newPADOP
Remove !IS_PADGV assertions
pad.c: Stop turning on GvIN_PAD
Only #define IS_(PADGV|CONST) if !PERL_CORE
Remove GVf_IN_PAD
[Merge] Eliminate IS_PADGV etc.
perlop tweak
Teach dump.c:debop about CV refs in stashes
gv.h: Comment typo
gv.h: Add a gp_flags field
Skip no-common-vars optimisation for aliases
Dump GPFLAGS
toke.c: Fix bad read in S_scan_heredoc
Rename S_adjust_stack_on_leave
Consistent indentation in AUTHORS
[perl #121638] Make âGlobal symbolâ message more newbie-friendly
Update perldiag to reflect âGlobal symbolâ change
perldelta for ff2a62e0/#15667
perldelta for 343573e55 (the 25-year-old bug)
perldelta for d8c6310a4f0/#121638
gv.c: For âGlobal symbolâ msg, donât check utf8ness twice
Teach diag.t about Perl_mess
perldiag: Reunite âperhaps you forgot to loadâ
George Greer (1):
ListUtil.xs: Temporary fix removing declaration after statement.
H.Merijn Brand (3):
Configure backports cause reorder
Sync up with meta after backports
regen after backport truncl probe
James E Keenan (8):
Check validity of keys in hash passed as first argument.
Add perldelta entries for updated modules and pragmata.
Add perldelta entry for ExtUtils::Manifest upgrade.
Correct typo reported by Hauke D. (RT #41359)
Add perldelta entry for Getopt::Std.
Add perldelta entry for 8a384d3a99 (ParseXS and locales).
Semicolon before ellipsis inside block disambiguates.
Increment $VERSION in Math::BigInt.
Jarkko Hietaniemi (267):
Comment: scan_num also handles hexadecimal floats now.
Modify the hexfp exponent explicitly in each long double case.
double-double needs the first mantissa bits of doubles.
Reduce double-double %a to single-double for now.
Quadruple float BE test triggering had a typo.
Add todo comment about special hexfp values.
Configure for isfinitel / isinfl.
Perl_isfinite() enhancements.
OSX does have isfinite(), and 10.9+ prefers it over finite().
Perl_isinf enhancements.
Use Perl_isfinite().
Win32 apparently has _isnan() and _finite(), and _fpclass().
Perl_signbit should return non-zero for -0.
Separate grok_infnan() from grok_number().
perldelta for hexadecimal floats.
Pass $ENV{HARNESS_TIMER} to Test::Harness from t/harness.
Try addressing perl #122578, low-order fp diffs.
Mention the possibility of fp rounding modes' effects.
IEEE quadruple exponents were off.
Do not anchor the matches, because of sprintf padding.
The quest for the hexfp exponent continues.
Partial revert of c6d2504a: do not use isfinite().
Use grok_infnan() if NV_INF and NV_NAN are defined.
Unify inf/nan printf output to Inf, -Inf, and NaN.
Unify the Inf/-Inf/NaN also in basic NV stringify.
Add inf/nan tests.
Clean up the fpclass forest.
Perl_isinf and Perl_isfinite enhancements.
Introduce Perl_strtod (macro) to call strtold if long doubles.
Implement POSIX::strtold().
Add some basic math tests.
For Darwin MacPorts add the gcc/g++ libdir.
Darwin can have libfoo.0.dylib instead of libfoo.dylib.0
If this format ever happens, it is probably left-aligned.
Remove obsolete comment.
Fix the PEEK_INFNAN (wrong macro arg name).
Fix infnan.t Win32 failure.
Old HP-UXen had a non-standard strtold().
Do not test memcmp et al, now also Linux inlined it.
Test 9**9**9 for Inf and sin(9**9**9) for NaN.
The less-than-zero branch needs to be the default always.
Revert "Test 9**9**9 for Inf and sin(9**9**9) for NaN."
Test 9**9**9 for Inf and sin(9**9**9) for NaN.
Skip strtold in concise-xs.t, "platform varying".
Fix bad quoting.
Explicitly cast to MANTISSATYPE since some compilers worry.
More robust inf/nan recognition and generation.
More is_inf() et alia dance.
Configure scan for fp_classify().
Configure scan for fp_classl().
Reorder the *fp*class* in preference order, add comments.
Test fpclassify() with full compile.
Allow eg ./perl t/harness t/op/lc.t
Sprinkle the fact that long doubles aren't standardized.
Only one successful exit from grok_infnan().
Comment why hexfp is set here, not earlier.
Simplify the hexfp space computation.
Post-VAX VMS has fp_classify().
*fp*class* comment tweaks.
Avoid useless warning, remove debug code.
Add more infnan tests.
printf inf/nan should be inf/nan.
Detect false infinities.
Comment tweak.
NVgf/NVff agree with -Duselongdouble.
Make sprintf %c and chr() on inf/nan return the U+FFFD.
pack c/C on inf/nan.
De-prefix Perl_isinfnan.
Under C99 isfinite/isinf/isnan are sizeof-aware macros.
Don't use isfinitel/isinfl/isnanl if the C99 macros exist.
Add a simple atan2() test.
apidoc for Perl_isinfnan.
Configure: add i_stdint.
HAS_C99 was oddly defined, and oddly used.
Another way to detect C99.
Avoid using function pointers for math functions.
Do not test sin/cos at [+-]pi, either.
S_infnan_copy needs at most 5 bytes of output size.
Wrong result for double-double sqrt(2).
The fp classification is a royal mess.
Detect HAS_AIX_POWL_NEG_BASE_BUG in hints/aix.sh
AIX: if long doubles do not mix with infinity, disable long doubles.
sqrt() disappeared in OSX, assuming it became pow().
Couple more infnan tests.
=for apidoc wrong api
POSIX math: Add M_ constants like M_PI.
POSIX math: Implement glue for the C99 math functions.
POSIX math: Add FP_ILOGB0 and FP_ILOGBNAN.
POSIX math: Configure scan for j0 and j0l.
POSIX math: Hopefully better use of Configure symbols.
POSIX math: Add more C99 math tests.
POSIX math: Configure scan for acosh.
POSIX math: Skip the POSIX C99 math tests if no acosh().
POSIX math: Use HAS_ACOSH to enable/disable the C99 math.
POSIX math: Win32 shuffling.
POSIX math: Have only one not_here for each API.
POSIX math: HP-UX exceptions.
POSIX math: Portability emulations and constants.
POSIX math: fpclassify emulation.
POSIX math: Configure scan d_fegetround and i_fenv.
POSIX math: More handling of legacy implementations.
POSIX math: Add fegetround() and fesetround().
POSIX math: lrint and rint emulations.
POSIX math: Configure scan fpgetround
POSIX math: todo for emulating fegetround() with fpgetround().
POSIX math: Bessel fixes.
POSIX math: fpclassify continues.
POSIX math: use POSIX now imports Inf and NaN.
POSIX math: Solaris needs -lsunmath.
AIX: allow one to keep long doubles, even if broken.
POSIX math: M_SQRT2 and M_2_SQRTPI were misspelled.
Do not use HUGE_VAL/VALL for NV_INF.
Perl_isinf and Perl_isfinite fallbacks.
Use Perl_isfinite() to guard the frexpl.
HP-UX: 10.20 math functions don't generate infinities.
HP-UX: 10.X doesn't really have isfinite().
HP-UX: missing C99 math.
POSIX math: M_LN10 was missing.
POSIX math: Initialize RETVAL to NV_NAN/-1.
POSIX math: define the fpclassify returns always if needed.
POSIX math: Perl_ldexp, not Perl_ldexpl.
POSIX math: First cut at C99 math docs.
Static initialization using 1/0 or 0/0 not C89.
HP-UX: more robust (10.X) and verbose test skippage.
Ignore warning about unused result since it is desperation time.
POSIX math: problems found in Win32.
Make certain pp_sin result is always initialized.
POSIX math: lgammal, not gammal.
POSIX math: Comment tweaks.
Basic string-to-nv should equal floating point literals.
POSIX math: Define my_fpclassify only if no c99_classify already.
POSIX math: erf implementation from http://www.johndcook.com/
POSIX math: better hypot.
POSIX math: isless etc tests.
POSIX math: more hypot tests.
POSIX math: looser error bound for erf test.
POSIX math: erfc implementation.
POSIX math: add isinf/isnan/isfinite tests.
POSIX math: add tests.
POSIX math: define the isless et al if necessary and feasible.
POSIX math: nan(NULL) leads to...
The 'nv' here is not always a NV, so have our own format.
POSIX math: Use 1.0 - erf(x) for erfc().
POSIX math: isunordered emulation was all broken.
POSIX math: make erf emulation C89 compliant.
POSIX math: s = "0" is not happy with -Wwrite-strings.
POSIX math: simplify the fpclassify emulation.
POSIX math: Win32 does not have erf and erfc.
POSIX math: c99_erfc, not plain erfc.
POSIX math: have the Perl_func wrappers for the C89 math, too.
POSIX math: if no fesetround, try fpsetround.
Some <termios.h> #define WRAP, undef it.
Grab the warnflags and stdflags from cflags [perl #122694].
Use sizeof() in UNUSED_ARG and UNUSED_VAR to avoid accessing the values.
Avoid gcc warning.
POSIX math: There's no #elifdef, unfortunately.
POSIX math: Cygwin lacks at least nexttoward.
POSIX math: let's not override the real lrint().
POSIX math: Use rounding macros, instead of the c99_ functions.
%ld expected long int, but got ssize_t.
Version multibump.
POSIX math: Use the Perl_ math APIs (C89) consistently.
fpclass snan detection errors (cut-n-paste).
POSIX math: Add the opengroup URL for math.h
NV_INF/NV_NAN fallback where int32 is coerced to IEEE754 float.
The .i target should depend at least on perl.h and config.h.
Long double might have INFINITYL or NANL.
Configure: scan for truncl.
modfl emulation via truncl (C99) and copysignl.
5df8ece2 didn't really allow one to keep long doubles.
POSIX math: HP-UX support varies depending on HP cc.
POSIX math: NV_NAN can be the right answer for nan().
POSIX math: implement lround().
POSIX math: stupidness in 9e010b89
POSIX math: AIX fixes.
The recent pad changes needed dVAR love.
Not nice to die in the middle of a test.
Document the IBM admission of weirdness of AIX long doubles.
hexfp: Implement double-double hexdigit extraction.
hexfp: double-double correct test goldens.
Add gtags and ctags files into .gitignore
Rename S_infnan_copy as S_infnan_2pv.
POSIX math: log1p emulation was wrong.
Sort the Perl_ math wrappers for easier eyeballing.
Comment tweaks.
POSIX math: log1p (emulation) needs arg > -1.0.
Rename the 'nv' as 'fv' since it's not NV.
POSIX math: log1p is more subtle.
Separate S_my_atof_infnan().
hexfp: 0 is 0x0p+0.
If USE_LONG_DOUBLE, NVff is PERL_PRIldbl.
Floating point should always output more than nothing.
Whitespace only.
Comment tweaks.
hexfp: Fix zero in IEEE 754 quadruples, cleanups, safety.
No need to print an int as a %0.f
Use the Perl_ math in time64.c, too.
Remove obsolete comment.
darwin data sections can have wild prefixes and suffixes.
undef doesn't accept args for macros.
Also protect against elen < 0.
Have a Perl_isfinitel() separate from USE_LONG_DOUBLE.
long double / NV schizophrenia continues.
Use -fstack-protector-strong, if available.
Unfinished business from e4c957f4.
More ways to Perl_isfinitel.
Perl_isfinitel should not use the fpclass macross blindly.
Use the multiplication test instead of the DBL_MAX/LDBL_MAX.
Comment tweak for HAS_C99.
The space computation for hexfp was overovershooting.
expl will not be used if long double is double.
Need to have frexpl for long doubles.
HAS_C99 comment tweak.
Use HAS_ISFINITE for Perl_isfinitel.
Mention inf/nan and POSIX C99 math improvements.
Make the link not a link.
Untangle the fp value retrieval.
Warning about the va_arg vs long doubles.
Use Perl_frexp, and cast arg to NV as appropriate.
Define Perl_isfinite only if not already defined.
Darwin: if cc is explicitly gcc/g++, use it also as ld.
share_hek_hek "return value" is unused here.
Darwin: set ld to 'cc' if $cc is 'cc' or empty, $cc otherwise.
[perl #122788] broken nan on win32
Do not test inf/nan if utterly broken.
Use the Perl_isfinite multiplication fallback last.
If numtype is inf/nan, don't do Atof and iv/uv cast.
If numtype is inf/nan, don't bother with Atof.
Use more NAN_COMPARE_BROKEN.
POSIX pod tweaks.
Document the double-double format a little.
Add comment about the double-double extraction.
hexfp: use the bytewise extraction also for plain doubles.
Make certain we have the top-level "lib" in @INC.
Add dVAR:s for -DPERL_GLOBAL_STRUCT.
Comment detypo.
Conditionally unused variable.
Unused return value.
POSIX math: Tru64 has rint()
POSIX math: the pa-risc gcc missing tgamma and nexttoward
POSIX math: AIX 7 has nexttoward
infnan: ++ or -- on infnan doesn't lose precision.
Do not fall into Gconvert or F0convert on inf/nan.
infnan: more tests.
Parenthesize & and | a bit.
Mention more C99 math interfaces.
Configure for <quadmath.h>
Configure -Dusequadmath.
quadmath __float128 for longdblkind.
quadmath __float128 as NVTYPE.
quadmath needs gcc, at least 4.6.
quadmath interfaces and constants
quadmath_snprintf formats.
quadmath INSTALL.
quadmath NV formatted I/O.
quadmath Q format specifier and gcc -Wformat do not mix.
quadmath helpers export list.
quadmath has expq instead of expl.
quadmath and gconvert do not mix.
quadmath doesn't do locale radixes.
quadmath strtoflt128 accepts false infinities.
quadmath NVs cannot be cast to double etc.
darwin can have libfoo.0.dylib, as opposed to libfoo.dylib.0
Editorializing quadmath entry.
quadmath: add test actually exercising this fancy precision.
perldelta: quadmath
Revert "infnan: more tests."
Jerry D. Hedden (1):
Upgrade to threads 1.96
Karen Etheridge (1):
replace uses of Test.pm in dist/ distributions (IO, Math-BigInt)
Karl Williamson (69):
perlfunc: Update -B, -T descriptions
unicode_constants.h: Add definitions for ESC and VT
pp_sys.c: Generalize -B, -T for EBCDIC, clean up
Use constant now defined for ESC
regex: Use #define for number of bits in ANYOF
Move _get_regclass_nonbitmap_data() to regcomp.c
regcomp.c: Reorder cases: in switch
Avoid redundant text -in -Dr output
regcomp.c: Swap 'else' clauses.
regcomp.c: Avoid function call overhead
perlop: Note that negative rep now warns
perlrecharclass: Remove now-irrelevant text
Add and use macros for case-insensitive comparison
regcomp.c: Use symbolic constant instead of number
Bump version of POSIX.pm
perlxs: Add text about dealing with locales, etc
embed.fnc: Clarify m flag behavior comment
sv.c: Silence VMS compiler warning
regcomp.c: Use available mnemonics for dumping regex
regcomp.c: white-space only
regcomp.c: Extract some code into a macro
handy.h, regcomp.c: Add, clarify comments
regen/unicode_constants.pl: Find max ascii print cp
perldiag: Add clarification.
regcomp.c: Add 'const' to parameter
regcomp.c: Use symbolic constant instead of number
Improve -Dr output of bracketed char classes
perlrecharclass: Use more common property synonym
ext/re/t/regop.t: Use eq instead of == for strings
regcomp.c: Silence Win32 compiler warning
PATCH: [perl #122655] 5.20 regression: '"X" !~ /[x]/i'
perldelta: Fix typo
lib/utf8_heavy.pl: Don't return empty under miniperl
regexec.c: Simplify a short code section
regcomp.c: Remove unnecessary test
regcomp.h: Fix comment that said the opposite of the truth
regcomp.h: Use unsigned 1 in left shift
regcomp.h: Remove some no-longer used #defines
Rename some internal regex #defines
regcomp.c: Swap if/else clauses
Fix -Dr output to work for larger ANYOF node size
Allow for changing size of bracketed regex char class
regcomp.h: Comment nits
numeric.c: Comment tweak
regcomp.c: Extract functionality into a static function
regcomp.c: Make macro a lookup
regcomp.c: Remove unnecessary cast
regcomp.c: Move comment closer to code it applies to
regcomp.c: Don't prematurely skip error checking
regcomp.c: Vertically stack ternary
recomp.c: output given warning in only one pass
PATCH: [perl #122671] Many warnings in regcomp.c can occur twice
regcomp.c: Refactor one area to use common subroutine
regcomp.c: Comment clarifications, nits
regcomp.c: Refactor func so caller handles anomalies
regcomp.c: White-space only
regcomp.c: Extract out functionality into a function
Allow \N{named seq} in qr/[...]/
regcomp.c: Remove extraneous tests
regcomp.c: Don't doubly do 'use encoding'
perldelta: Mention -T -B changes
perldelta: Note bug fix to suppress repeated warnings
perldelta: Entry for [perl #122655].
bisect-runner.pl: Correct pod misspelling
perlop: Nits
regcomp.c: White-space only
perlsyn: Remove reference to potential Unicode op
perldelta: Grammar nit
perllocale: Nit
Olivier Mengué (3):
big{num,rat,int}: use eval block to load Math::BigInt::Lite
bignum tests: use eval block to load Math::BigInt::Lite
bignum tests: use eval-block instead of eval-string
Peter Martini (9):
perldelta: fix my bad pod
Scowl. 19 should have been '[VERSION]'
Merge branch 'release-5.21.3' into blead
Epigraph for 5.21.3
New perldelta.
Bump version number from 5.21.3 to 5.21.4
Version bump addendum; changed a 5.21.1 to .3
Module::CoreList version number bump
Fix documented return type for PAD_SV
Reini Urban (1):
sv_grow: performance improvement for short strings
Ricardo Signes (6):
todo: expand the list of OS access builtins to update
release schedule: put Abigail on for October
release schedule: add managers for 5.21.6, 5.21.8
release schedule: Max M. will do December 2014
document how to graduate from experimental status
delete/exists on array elements is discouraged, not deprecated
Steffen Mueller (1):
Data::Dumper version bump for CPAN release
Steve Hay (46):
Add new pumpkin keeper victim following release of 5.21.3
Update Module-CoreList for 5.20.1
Add epigraph for 5.20.1-RC1
Add release date of 5.20.1-RC1
Update release schedule
Upgrade version from version 0.9908 to 0.9909
Porting/Maintainers.pl - Fix ExtUtils-CBuilder entry
5.20.1-RC2 today
Add epigraph for 5.20.1-RC2
Upgrade CPAN::Meta::Requirements from version 2.126 to 2.128
Upgrade ExtUtils::CBuilder from version 0.280217 to 0.280219
Scalar-List-Utils has been upgraded from version 1.40 to 1.41
Porting/Maintainers.pl - Tidy up @IGNORABLE
Perl 5.20.1 today
Perl 5.20.1 today
Add epigraph for 5.20.1
Copy perl5201delta into blead
Update 5.20 release schedule: Plan for 5.20.2 in January
Upgrade Archive-Tar from version 2.00 to 2.02
Remove IGNORABLE files from ExtUtils-Install as per 8db6555210
Note that one MakeMaker test is currently CUSTOMIZED
Add new Digest-SHA file missed by 207902b174, and note a CUSTOMIZED file
Make "NEWS" an IGNORABLE file
Teach Porting/cmpVERSION.pl about Scalar-List-Utils (broken by cb8c84586a)
Remove reference to obsolete threads version in Porting/cmpVERSION.pl
perldelta - Start updating modules
perldelta - Add links for rt.perl.org bugs
Module::CoreList - re-order so that Porting/corelist.pl runs ok
Remove pointless variable in win32/config_sh.PL
Eliminate pointless win32/makefile.mk macro
regen_config_h for VC++
Define PERL_PRI[efg]ldbl and PERL_SCNfldbl in Win32 canned config files
regen_config_h for GCC
perldelta - Fix a typo and remove a trailing space
Revert a231c375cc and fix 61cd60c6c5 differently instead
Bump all Math::BigInt $VERSIONs as per 945313f0ae following a0732aaa4b
perldelta - Data::Dumper has been upgraded
perldelta - Add links for rt.perl.org bugs
perldelta for 1645b83c5c
Update release announcement template
perldelta - Remove boilerplate, fill in modules changes and other minor
edits
$VERSION bump cpan/ExtUtils-MakeMaker/lib/ExtUtils/Liblist/Kid.pm
Blead is in sync with Data::Dumper 2.154
Update Module::CoreList for 5.21.4
Finalize perldelta
Perl 5.21.4 today
Sullivan Beck (1):
Update Locale-Codes to CPAN version 3.32
Tadeusz SoÅnierz (1):
Make eval_pv documentation more precise
Tony Cook (1):
don't recurse infinitely in Data::Dumper
Yves Orton (8):
Rename S_stashpvn to S_gv_stashpvn_internal and add to embed.fnc
perl #122747: localize PL_curpm to null in _core_swash_init
add test for rt122747
Eliminate the duplicative regops BOL and EOL
change NODE_ALIGN_FILL to set flags to 0
Perl RT #122761 - split /\A/ should not behave like split /^/m
only produce diagnostic warnings in regen/regcomp.pl when STDERR is a
terminal
silence diagnostics in regen/regcomp.pl for now
syber (4):
Make gv_stashpvn() use PL_stashcache
Make S_method_common use gv_stashpvn instead of copypasted cache usage
introduce gv_stashsvpvn_cached()
avoid local *f = \&foo resetting the method cache
Ãvar Arnfjörð Bjarmason (1):
perlfunc: Improve the pointer from "elseif" to "elsif"
-----------------------------------------------------------------------
--
Perl5 Master Repository
